paperID
stringlengths 36
36
| pwc_id
stringlengths 8
47
| arxiv_id
stringlengths 6
16
⌀ | nips_id
float64 | url_abs
stringlengths 18
329
| url_pdf
stringlengths 18
742
| title
stringlengths 8
325
| abstract
stringlengths 1
7.27k
⌀ | authors
stringlengths 2
7.06k
| published
stringlengths 10
10
⌀ | conference
stringlengths 12
47
⌀ | conference_url_abs
stringlengths 16
198
⌀ | conference_url_pdf
stringlengths 27
199
⌀ | proceeding
stringlengths 6
47
⌀ | taskID
stringlengths 7
1.44k
| areaID
stringclasses 688
values | embedding
stringlengths 9.26k
12.5k
| umap_embedding
stringlengths 29
44
|
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
9c5bd677-8a2a-48b6-82b1-9d3e1c16e45f
|
deepring-learning-roto-translation-invariant
|
2210.11029
| null |
https://arxiv.org/abs/2210.11029v1
|
https://arxiv.org/pdf/2210.11029v1.pdf
|
DeepRING: Learning Roto-translation Invariant Representation for LiDAR based Place Recognition
|
LiDAR based place recognition is popular for loop closure detection and re-localization. In recent years, deep learning brings improvements to place recognition by learnable feature extraction. However, these methods degenerate when the robot re-visits previous places with large perspective difference. To address the challenge, we propose DeepRING to learn the roto-translation invariant representation from LiDAR scan, so that robot visits the same place with different perspective can have similar representations. There are two keys in DeepRING: the feature is extracted from sinogram, and the feature is aggregated by magnitude spectrum. The two steps keeps the final representation with both discrimination and roto-translation invariance. Moreover, we state the place recognition as a one-shot learning problem with each place being a class, leveraging relation learning to build representation similarity. Substantial experiments are carried out on public datasets, validating the effectiveness of each proposed component, and showing that DeepRING outperforms the comparative methods, especially in dataset level generalization.
|
['Yue Wang', 'Rong Xiong', 'Li Tang', 'Xuecheng Xu', 'Sha Lu']
|
2022-10-20
| null | null | null | null |
['loop-closure-detection', 'one-shot-learning']
|
['computer-vision', 'methodology']
|
[-3.16408984e-02 -3.72694552e-01 -4.09305930e-01 -5.11694252e-01
-8.15176845e-01 -4.72333312e-01 6.60908043e-01 1.98933125e-01
-5.47124982e-01 5.84470689e-01 2.17480555e-01 6.31167144e-02
-2.44733766e-01 -8.34961414e-01 -9.36829269e-01 -7.03001499e-01
-1.93835303e-01 3.27122301e-01 6.55768216e-02 -1.82657897e-01
4.48600292e-01 8.25333297e-01 -1.47311151e+00 -2.52327055e-01
7.67515540e-01 1.02401233e+00 2.15409681e-01 6.76557273e-02
2.01276585e-01 5.01707733e-01 -1.07125901e-01 2.69115210e-01
6.02884650e-01 1.82665601e-01 -6.39595449e-01 -2.15288386e-01
5.88055313e-01 -2.75292397e-01 -6.05621040e-01 9.04329062e-01
5.29668748e-01 3.22610140e-01 7.25255668e-01 -1.30217373e+00
-9.31837320e-01 1.35557398e-01 -5.68043411e-01 2.57662088e-02
5.83166778e-01 -1.22437604e-01 9.78660345e-01 -1.03334475e+00
4.43361580e-01 1.30543900e+00 9.66233075e-01 -8.88444076e-04
-1.07215154e+00 -7.87050962e-01 2.10295975e-01 4.05531347e-01
-1.54927862e+00 -3.37520719e-01 7.23151386e-01 -3.67516339e-01
8.94349813e-01 -2.28487432e-01 4.87160742e-01 1.04027748e+00
3.83367807e-01 8.50842476e-01 1.06116796e+00 -1.04068164e-02
2.63196796e-01 -2.63268948e-01 -1.54059768e-01 6.02525711e-01
4.49103355e-01 1.82264104e-01 -4.95455861e-01 1.05954990e-01
7.07677484e-01 6.91383421e-01 -1.95896506e-01 -9.30627942e-01
-1.45977592e+00 7.03950047e-01 1.07981277e+00 1.83743000e-01
-2.29633778e-01 2.75077224e-01 2.35444218e-01 3.94167423e-01
7.43498281e-02 4.04314250e-01 -2.55711555e-01 2.00507522e-01
-5.05683661e-01 2.26614356e-01 4.76439804e-01 1.32014894e+00
1.21887767e+00 -1.48020610e-01 8.02977830e-02 6.10823929e-01
4.66838628e-01 7.84230471e-01 8.00229728e-01 -5.46503961e-01
4.68032449e-01 7.35390425e-01 3.07473000e-02 -1.27494371e+00
-5.54188669e-01 -3.41340542e-01 -9.40993547e-01 6.51915297e-02
-4.91344891e-02 1.15525588e-01 -9.91376638e-01 1.60811114e+00
2.08146602e-01 4.66348559e-01 1.75018743e-01 7.68016219e-01
6.62164092e-01 5.27769566e-01 -1.31831035e-01 2.06654653e-01
1.18290114e+00 -9.36489940e-01 -5.20595372e-01 -5.83439350e-01
7.09961593e-01 -4.93474424e-01 5.99925697e-01 7.62346089e-02
-2.30486095e-01 -6.92367852e-01 -1.47867644e+00 -4.65524226e-01
-5.98657668e-01 1.94113970e-01 7.75256991e-01 1.86306357e-01
-7.73424029e-01 6.72212422e-01 -8.26737702e-01 -5.93963206e-01
3.42805833e-01 4.24845636e-01 -8.08623493e-01 -2.61284381e-01
-9.22993481e-01 7.53994167e-01 4.30446893e-01 3.46913449e-02
-8.20281744e-01 -2.96544105e-01 -1.29664862e+00 -2.58785576e-01
1.43627459e-02 -4.54171002e-01 9.96371269e-01 -2.40861520e-01
-1.21048105e+00 7.94004142e-01 -2.70780802e-01 -7.02468574e-01
4.79212701e-01 -4.82825160e-01 -4.55469966e-01 -2.46964544e-01
6.31479263e-01 7.03196585e-01 7.74640441e-01 -1.20065176e+00
-8.78372192e-01 -7.97533512e-01 -1.38862491e-01 3.19982111e-01
1.09207116e-01 -6.44791722e-01 -7.99517408e-02 -3.36564362e-01
1.13016117e+00 -1.14327359e+00 -2.06190914e-01 1.49517134e-01
-3.23179811e-01 -3.46988320e-01 8.65701735e-01 -1.89752519e-01
4.99309897e-01 -2.49040151e+00 -2.02386782e-01 9.90053825e-03
-8.23511928e-03 -2.05492005e-01 -1.51662871e-01 5.01379132e-01
-6.23957776e-02 -1.45045489e-01 -1.76274583e-01 -3.61321986e-01
4.57136109e-02 2.81909853e-01 -7.30807722e-01 1.00619423e+00
1.74586326e-01 8.52791548e-01 -1.18161047e+00 -1.84587345e-01
3.31212223e-01 2.60286361e-01 -3.80024612e-01 -1.18529825e-02
4.17102545e-01 5.96242726e-01 -5.16116440e-01 8.16137314e-01
9.83874142e-01 1.34168446e-01 -1.11987635e-01 -1.25840008e-01
-3.08851153e-01 5.18133640e-01 -1.20805693e+00 2.16617584e+00
-4.90855396e-01 6.37109876e-01 -5.03115714e-01 -9.72378135e-01
1.36046195e+00 -1.66322231e-01 4.39052194e-01 -9.92307663e-01
1.14119914e-03 4.74874049e-01 -3.25552851e-01 -3.51903528e-01
6.19929314e-01 6.03885949e-02 -3.89603615e-01 6.69377809e-03
8.99908319e-02 -2.01154724e-01 -2.89153665e-01 -3.82516831e-01
9.20038044e-01 2.66522586e-01 5.12001574e-01 -8.52979571e-02
3.48971099e-01 -7.07391426e-02 6.54679120e-01 7.24895537e-01
-4.76211101e-01 6.78753316e-01 -1.16861746e-01 -6.22417152e-01
-6.34261787e-01 -1.20339000e+00 -2.66107291e-01 8.61348093e-01
8.66240382e-01 7.29859481e-03 3.51506993e-02 -4.50977564e-01
4.43019867e-01 4.06034261e-01 -5.79634964e-01 -4.85117227e-01
-6.97620511e-01 -2.26543233e-01 6.02859020e-01 7.64838099e-01
9.58133996e-01 -7.22956240e-01 -8.35764647e-01 4.75304686e-02
-1.31292641e-01 -1.05704427e+00 -2.27026835e-01 5.54684520e-01
-9.27577555e-01 -1.03251803e+00 -6.15103900e-01 -1.20953012e+00
5.02263546e-01 9.67536986e-01 3.60415936e-01 -4.10091192e-01
-1.29404485e-01 2.05874413e-01 -2.88462937e-01 -2.39648774e-01
4.55570847e-01 1.23347215e-01 4.73480970e-01 -1.04740568e-01
5.44884384e-01 -8.33678246e-01 -5.67678332e-01 2.71832496e-01
-4.11933511e-01 -4.13086295e-01 9.70892847e-01 7.81331182e-01
8.22482526e-01 -1.96427882e-01 2.55852878e-01 -2.61261642e-01
3.40954900e-01 -6.28083587e-01 -4.22016799e-01 -3.17411534e-02
-5.44581831e-01 3.58484536e-01 3.93911153e-01 -3.60311717e-01
-5.27490616e-01 4.24380124e-01 1.49531841e-01 -4.04347122e-01
-4.14577276e-01 4.66259003e-01 -2.21481562e-01 -1.75934926e-01
6.47754669e-01 4.93195832e-01 -1.99232489e-01 -5.05426526e-01
4.26916391e-01 4.66644526e-01 6.53079093e-01 -3.92340422e-01
1.10121119e+00 8.10478926e-01 1.32572711e-01 -6.97794020e-01
-6.38790488e-01 -7.95780480e-01 -1.02322781e+00 4.27184731e-01
6.65484428e-01 -1.35288107e+00 -6.73986375e-01 3.70808959e-01
-1.12173140e+00 1.88674659e-01 -2.31833994e-01 7.97583997e-01
-5.57191610e-01 3.70314926e-01 -9.34461802e-02 -6.60171509e-01
2.79081687e-02 -1.01029491e+00 1.38949597e+00 4.12774473e-01
1.41935423e-01 -5.19472957e-01 5.36443233e-01 -1.91711038e-01
-5.80837503e-02 4.38685238e-01 6.35258615e-01 -8.15956831e-01
-7.19810247e-01 -3.83667201e-01 -3.57884288e-01 -1.02445327e-01
2.64577568e-01 -4.36934531e-01 -1.06975889e+00 -5.00548005e-01
2.52156742e-02 -2.84797043e-01 1.05619669e+00 -3.18895206e-02
8.44859660e-01 -1.36701420e-01 -7.05516696e-01 9.50501680e-01
1.43765545e+00 2.41056621e-01 5.39686441e-01 7.81466126e-01
6.44382656e-01 2.54972845e-01 7.90899873e-01 2.23931596e-01
6.75927877e-01 5.50293922e-01 6.26149356e-01 1.40052095e-01
1.29845053e-01 -6.93894386e-01 4.24328744e-01 5.56245148e-01
-2.36500055e-03 3.07085872e-01 -9.99853611e-01 7.62427092e-01
-2.02791524e+00 -9.51618135e-01 1.17565624e-01 2.33237171e+00
2.01348528e-01 -6.00324795e-02 -3.14474612e-01 6.76045641e-02
6.32360578e-01 3.77599567e-01 -7.23989189e-01 1.05319940e-01
-1.57300875e-01 -8.52117017e-02 7.89311230e-01 2.64275968e-01
-1.43703222e+00 1.01678896e+00 5.73486567e+00 3.00471038e-01
-1.61100268e+00 -1.19670138e-01 -6.15555868e-02 3.35978240e-01
-2.48208586e-02 9.71470103e-02 -7.82978833e-01 9.47264284e-02
3.60906214e-01 -3.39157172e-02 2.11479113e-01 1.28024554e+00
-2.34237432e-01 -4.83637489e-02 -1.38622105e+00 1.32264447e+00
1.56223550e-01 -9.48628664e-01 -4.03675064e-02 8.48007202e-02
5.44468701e-01 5.74332774e-01 2.51982689e-01 7.55029738e-01
3.17273773e-02 -9.64575648e-01 8.26847196e-01 4.20362920e-01
6.06912732e-01 -5.77049196e-01 7.56461561e-01 4.29739028e-01
-1.65056229e+00 -3.32777113e-01 -7.54990697e-01 -2.04778656e-01
-1.55360252e-01 2.11516857e-01 -1.13165927e+00 8.92059982e-01
7.44120955e-01 1.37325013e+00 -6.27957582e-01 1.31792223e+00
-3.16720188e-01 -1.86681807e-01 -4.15146738e-01 1.23916470e-01
3.25405359e-01 -2.79761195e-01 5.17615855e-01 9.21294093e-01
6.50617003e-01 -2.34980673e-01 4.02653188e-01 8.40515792e-01
6.03931360e-02 -8.92633572e-02 -1.19961083e+00 1.60741672e-01
8.20937037e-01 9.96199548e-01 -5.41775584e-01 4.63727564e-02
-1.71875164e-01 1.06089389e+00 4.36773360e-01 3.01142126e-01
-6.70082152e-01 -6.34530902e-01 7.95346737e-01 -1.92984074e-01
3.60715091e-01 -6.74391448e-01 -2.84980714e-01 -1.10447061e+00
1.80524930e-01 -3.85605335e-01 7.15736151e-02 -6.08015180e-01
-1.04711342e+00 3.32655966e-01 -1.13245994e-01 -1.79943943e+00
-6.95192814e-02 -7.09787667e-01 -5.38668692e-01 5.94258010e-01
-1.99010050e+00 -1.30017352e+00 -5.38155258e-01 6.01948857e-01
5.31440914e-01 1.59415994e-02 7.74916410e-01 3.10434014e-01
-2.84499586e-01 5.56583703e-01 2.83119649e-01 2.69869804e-01
9.64917183e-01 -1.11450291e+00 5.07829547e-01 7.05288589e-01
3.68885815e-01 9.46665823e-01 4.17263746e-01 -5.43315291e-01
-1.61144888e+00 -1.31829786e+00 9.20882225e-01 -4.05183733e-01
4.73890632e-01 -4.14274156e-01 -6.80146933e-01 8.28929782e-01
-2.84936488e-01 2.89172739e-01 4.45065290e-01 1.24908783e-01
-6.98283851e-01 -2.26046100e-01 -1.14858186e+00 4.77153331e-01
1.35523307e+00 -9.60979998e-01 -9.88014758e-01 2.01433629e-01
7.13588834e-01 -5.49731493e-01 -5.23846328e-01 6.76472306e-01
6.14771843e-01 -8.54852080e-01 1.05638576e+00 -3.83905262e-01
1.48956180e-01 -6.78442121e-01 -6.37653947e-01 -1.26270139e+00
-5.42350948e-01 -1.95468113e-01 2.83572078e-01 9.60505486e-01
8.22975487e-02 -9.43894684e-01 3.98466647e-01 -1.92917421e-01
-2.69752264e-01 -5.39052248e-01 -1.15806043e+00 -1.07662868e+00
6.92177564e-03 -3.15791070e-01 8.47936869e-01 9.81298566e-01
-1.41881689e-01 4.17068958e-01 -1.87203199e-01 7.04320669e-01
3.17588776e-01 4.40170616e-01 1.00620949e+00 -1.61273193e+00
2.72801280e-01 -1.40352547e-01 -1.01902199e+00 -1.51290834e+00
1.61284864e-01 -1.06700051e+00 4.28498417e-01 -1.52348018e+00
-4.94119301e-02 -5.03206730e-01 -4.92164522e-01 6.06219828e-01
1.73949957e-01 -4.60088588e-02 1.43914327e-01 5.64795554e-01
-6.19191587e-01 9.17704940e-01 9.73590195e-01 -4.31166649e-01
-1.36868492e-01 -3.78352404e-02 -4.52195227e-01 7.61239290e-01
6.30156398e-01 -4.70502704e-01 -2.64414757e-01 -6.27904832e-01
1.09560758e-01 -4.16057080e-01 5.25523186e-01 -1.50801480e+00
4.49823976e-01 -1.61191344e-01 5.17978072e-01 -8.85535777e-01
3.84388715e-01 -1.10655463e+00 -1.29122943e-01 6.13294065e-01
-1.64918274e-01 2.32021019e-01 1.39010876e-01 1.05354464e+00
-2.62000829e-01 1.41390607e-01 5.88233709e-01 9.52075198e-02
-1.26988876e+00 5.03402531e-01 7.19984900e-03 -2.13294119e-01
9.21536386e-01 -5.10942817e-01 -1.68476164e-01 -1.96288034e-01
-2.81555295e-01 2.26004437e-01 6.21366978e-01 7.77313054e-01
6.86070025e-01 -1.69938862e+00 -3.52015257e-01 5.22306919e-01
6.44855082e-01 4.16622818e-01 -4.15865779e-02 9.26779151e-01
-4.28849757e-01 5.67530513e-01 -4.33898836e-01 -9.59271848e-01
-6.48917913e-01 5.42626441e-01 2.88003892e-01 1.49468943e-01
-7.64010012e-01 7.55731463e-01 1.01090379e-01 -9.47687209e-01
2.79575616e-01 -7.87802935e-01 -1.30049631e-01 2.03589916e-01
2.80311316e-01 2.49271378e-01 2.39553396e-02 -8.41182411e-01
-6.81192279e-01 9.88142192e-01 -4.17713001e-02 7.95220211e-02
1.31785965e+00 -2.64830858e-01 5.59567474e-02 7.47459471e-01
1.36656046e+00 -3.49397138e-02 -1.23800993e+00 -3.95230055e-01
2.75656551e-01 -5.72218478e-01 -2.59923249e-01 -2.35198915e-01
-6.23753369e-01 9.82556462e-01 9.87398386e-01 -1.19149782e-01
5.17267227e-01 -2.53250480e-01 7.70143509e-01 1.15780461e+00
9.36554432e-01 -8.47749472e-01 2.35996228e-02 1.10642183e+00
8.91042888e-01 -1.57479680e+00 4.10464555e-02 -9.81411114e-02
-3.68661314e-01 1.02089322e+00 5.07524788e-01 -6.89896584e-01
5.80840290e-01 -2.26859376e-01 -2.81987526e-02 -6.84229657e-02
-1.34600177e-01 -2.83035815e-01 3.61686081e-01 7.09802032e-01
1.94598228e-01 1.49360433e-01 1.25864789e-01 3.24074864e-01
-5.17631233e-01 -2.62167573e-01 1.57065123e-01 1.10629165e+00
-7.83457816e-01 -6.80176258e-01 -1.91237807e-01 -7.49241188e-02
3.58926982e-01 1.60275042e-01 -3.93620640e-01 8.40961099e-01
2.65168846e-01 6.78984344e-01 2.07236931e-01 -6.17257237e-01
4.06213194e-01 -5.49008474e-02 3.54977489e-01 -6.54622197e-01
5.21093234e-02 -2.70101070e-01 -4.39888090e-01 -7.68019438e-01
-2.26403922e-01 -5.49178779e-01 -1.54329324e+00 -5.90632334e-02
-9.93985459e-02 4.94522713e-02 7.56846011e-01 1.04366195e+00
4.94776428e-01 2.88088650e-01 9.02939498e-01 -1.12508011e+00
-7.85564125e-01 -7.61085808e-01 -5.84776044e-01 3.58182520e-01
8.84492040e-01 -1.06138074e+00 -3.79471451e-01 -3.42880398e-01]
|
[7.551284313201904, -2.133492946624756]
|
7a1cb48d-0553-4287-b9a1-c3f0d5701aa7
|
compressed-video-quality-assessment-for-super
|
2305.04844
| null |
https://arxiv.org/abs/2305.04844v1
|
https://arxiv.org/pdf/2305.04844v1.pdf
|
Compressed Video Quality Assessment for Super-Resolution: a Benchmark and a Quality Metric
|
We developed a super-resolution (SR) benchmark to analyze SR's capacity to upscale compressed videos. Our dataset employed video codecs based on five compression standards: H.264, H.265, H.266, AV1, and AVS3. We assessed 17 state-ofthe-art SR models using our benchmark and evaluated their ability to preserve scene context and their susceptibility to compression artifacts. To get an accurate perceptual ranking of SR models, we conducted a crowd-sourced side-by-side comparison of their outputs. The benchmark is publicly available at https://videoprocessing.ai/benchmarks/super-resolutionfor-video-compression.html. We also analyzed benchmark results and developed an objective-quality-assessment metric based on the current bestperforming objective metrics. Our metric outperforms others, according to Spearman correlation with subjective scores for compressed video upscaling. It is publicly available at https://github.com/EvgeneyBogatyrev/super-resolution-metric.
|
['Dmitriy Vatolin', 'Ivan Molodetskikh', 'Evgeney Bogatyrev']
|
2023-05-08
| null | null | null | null |
['video-quality-assessment', 'video-quality-assessment']
|
['computer-vision', 'time-series']
|
[ 2.10907146e-01 -4.86987591e-01 -2.51953751e-01 -2.47959048e-01
-1.25988507e+00 -5.21043897e-01 4.32311803e-01 -1.40092447e-01
-1.60153046e-01 6.10315800e-01 7.70506203e-01 -1.76970303e-01
-7.68088968e-03 -5.88994682e-01 -7.11020350e-01 -3.19915086e-01
-4.69970345e-01 -3.02056521e-01 6.64063752e-01 -3.25534999e-01
5.28710842e-01 2.56924927e-01 -1.73354530e+00 9.13171947e-01
7.87696600e-01 1.10187018e+00 5.35104096e-01 1.19114101e+00
4.21497017e-01 1.30542326e+00 -4.92779911e-01 -4.18476284e-01
3.44195873e-01 -6.29870594e-01 -7.28564739e-01 -1.57990932e-01
5.17094612e-01 -6.15773857e-01 -5.35760462e-01 1.13449216e+00
6.05643153e-01 -1.86848771e-02 1.97302967e-01 -1.05220020e+00
-9.81492817e-01 6.96587265e-01 -4.54563022e-01 9.87282336e-01
1.02475119e+00 2.76588678e-01 1.08739853e+00 -1.04819894e+00
1.02530146e+00 1.20450771e+00 4.71082270e-01 4.11762923e-01
-1.10844707e+00 -7.54085839e-01 -6.50466025e-01 6.95242226e-01
-1.53401864e+00 -8.60383034e-01 4.40722138e-01 -2.08617494e-01
8.97428095e-01 4.18185055e-01 3.89383167e-01 1.20192587e+00
4.39878762e-01 3.71051282e-01 1.10828447e+00 -1.16188928e-01
3.24994415e-01 -1.63895175e-01 -2.98628658e-01 4.01181310e-01
1.99181244e-01 3.15216362e-01 -9.22735035e-01 1.39699608e-01
1.05911601e+00 -5.30509651e-01 -5.12440681e-01 -1.39834806e-01
-1.30276811e+00 4.91991609e-01 1.99517950e-01 3.14614952e-01
-1.94226310e-01 1.07498743e-01 5.06878376e-01 5.87934375e-01
3.91228229e-01 2.42306188e-01 -2.03267708e-01 -4.83236790e-01
-1.22882605e+00 3.45079213e-01 3.91217172e-01 1.13457680e+00
2.92672634e-01 2.24829108e-01 -2.42398828e-01 1.00900090e+00
-1.69988330e-02 3.97362739e-01 4.58040118e-01 -1.91949344e+00
6.15166426e-01 -9.45932418e-02 2.16529042e-01 -1.01197004e+00
4.38548275e-04 -2.46203288e-01 -7.09456980e-01 4.42960769e-01
2.29734229e-03 1.70401618e-01 -3.70096475e-01 1.35051906e+00
-2.41565093e-01 2.69754142e-01 1.53897792e-01 1.17919672e+00
1.06146121e+00 5.89304268e-01 3.89685179e-03 -3.70866746e-01
1.06988180e+00 -9.43815172e-01 -7.84515977e-01 2.63278425e-01
3.77441883e-01 -9.32091296e-01 1.08190143e+00 4.00932342e-01
-1.65098786e+00 -9.42335308e-01 -1.21300161e+00 -2.48348311e-01
2.69878566e-01 -2.96897441e-01 4.63772006e-02 3.88178408e-01
-1.60713255e+00 8.69644046e-01 -6.80991590e-01 -1.30650878e-01
4.54281479e-01 -8.42112973e-02 -4.42085266e-01 -1.24617800e-01
-1.19534993e+00 7.13672757e-01 2.73152530e-01 -7.78123856e-01
-1.19233465e+00 -7.23870933e-01 -7.55771041e-01 -1.77378617e-02
1.97920159e-01 -5.59099019e-01 1.35181212e+00 -1.09292161e+00
-1.32007289e+00 7.66016006e-01 -1.49934873e-01 -5.89314044e-01
5.49130261e-01 -2.86220104e-01 -9.80731785e-01 7.56583393e-01
6.20222068e-04 6.04341507e-01 7.16966629e-01 -1.47561622e+00
-8.49578559e-01 -2.24724952e-02 2.57168144e-01 2.14099854e-01
7.20743760e-02 5.73571742e-01 -4.36627060e-01 -1.00785100e+00
-1.89313982e-02 -5.36555707e-01 1.29851282e-01 2.39164345e-02
-7.07104951e-02 3.88544291e-01 7.50142038e-01 -1.03028214e+00
1.54897404e+00 -2.42671609e+00 1.87106818e-01 -2.82330751e-01
3.64767015e-01 2.54218504e-02 -3.74301642e-01 3.12303960e-01
-1.61538422e-01 4.43376899e-01 -8.28147586e-03 -2.05031767e-01
-3.98286462e-01 -2.41603926e-01 -9.58727524e-02 2.63371080e-01
-7.42099211e-02 5.04649878e-01 -1.00452006e+00 -8.30764055e-01
2.66168028e-01 8.30166161e-01 -8.08096051e-01 2.53993541e-01
8.67096856e-02 3.69440824e-01 7.33312219e-02 8.44948053e-01
5.85469246e-01 -3.42747271e-01 -4.01164107e-02 -4.29438025e-01
-1.80632994e-01 9.93534848e-02 -1.17793453e+00 1.77445650e+00
-3.57547790e-01 9.50351417e-01 -8.12535211e-02 -2.30131015e-01
6.84176624e-01 4.12713468e-01 5.22753179e-01 -9.72233534e-01
-9.13380682e-02 3.28544557e-01 -2.71935880e-01 -5.00964701e-01
1.10555315e+00 3.75639766e-01 2.70303428e-01 -3.96195278e-02
4.16526645e-02 -5.34069054e-02 5.65051258e-01 3.57693702e-01
1.36060297e+00 1.88297659e-01 4.14932549e-01 -3.44743192e-01
6.32399440e-01 -2.51177520e-01 3.69080037e-01 5.00297427e-01
-4.43995655e-01 1.25837922e+00 3.52386922e-01 -2.58137017e-01
-1.46722257e+00 -1.45415711e+00 -8.50684792e-02 1.16351366e+00
4.22694623e-01 -6.96147382e-01 -8.37461352e-01 5.37447690e-04
-4.27673489e-01 8.45545173e-01 -3.16761553e-01 7.60811493e-02
-2.41921112e-01 -3.20275724e-02 5.05356193e-01 4.63798225e-01
7.24883080e-01 -9.54933703e-01 -8.29170167e-01 -5.72921298e-02
-5.05980492e-01 -1.39566576e+00 -4.21585441e-01 -4.23051149e-01
-8.75311434e-01 -1.03199589e+00 -8.15031707e-01 -5.33971786e-01
2.26556119e-02 5.28677464e-01 1.40823233e+00 1.81248143e-01
-5.12070656e-02 2.75148422e-01 -8.49970937e-01 2.84644544e-01
-8.07834923e-01 -2.93707937e-01 1.69546902e-01 -4.47223693e-01
1.24958819e-02 -6.35869026e-01 -9.87934649e-01 5.72965682e-01
-1.05557108e+00 1.56439930e-01 1.97432324e-01 2.82579839e-01
8.17280531e-01 2.28131860e-01 1.54942453e-01 -3.44899595e-01
4.96841401e-01 -5.39110541e-01 -5.53011119e-01 5.06295590e-03
-3.65054637e-01 -1.31441832e-01 5.70339143e-01 -1.51783586e-01
-1.00709665e+00 -5.38457155e-01 -1.96611702e-01 -7.10366726e-01
-5.19794039e-02 2.10491940e-01 -4.66697803e-03 5.42483339e-03
7.85196841e-01 2.23677069e-01 -3.52979958e-01 -5.08871317e-01
2.25529775e-01 6.69376314e-01 8.96706700e-01 -3.62467885e-01
7.42804289e-01 3.80424559e-01 -2.57054031e-01 -7.81687438e-01
-4.03711230e-01 -2.23055080e-01 -3.05819929e-01 -3.48817557e-01
9.09299314e-01 -1.30411041e+00 -1.91868737e-01 2.19903350e-01
-9.14036334e-01 -3.89404982e-01 -3.37921947e-01 4.54243511e-01
-9.21551347e-01 6.52295768e-01 -9.83730912e-01 -4.51280534e-01
-2.55428672e-01 -1.26326811e+00 8.75691891e-01 4.34815437e-02
-1.83173299e-01 -6.65784717e-01 1.87404782e-01 5.33225298e-01
8.17573845e-01 2.08108515e-01 3.24909240e-01 -1.36308372e-01
-4.60247666e-01 3.87321591e-01 -5.25217235e-01 5.28136313e-01
-8.35315734e-02 2.34732985e-01 -7.46541917e-01 -5.03408074e-01
-1.03189386e-01 -2.72511363e-01 7.32481241e-01 4.25659657e-01
1.44575644e+00 -4.37502921e-01 2.54035354e-01 9.94229853e-01
1.82012355e+00 8.15737918e-02 1.15432000e+00 6.46267116e-01
3.11408788e-01 1.10651143e-01 7.86004007e-01 7.13531494e-01
2.67714024e-01 7.87612796e-01 4.67879951e-01 2.06410602e-01
-6.10693932e-01 -1.86573386e-01 6.64638460e-01 8.72962713e-01
-7.59710848e-01 -4.02009159e-01 -7.08158374e-01 2.77445942e-01
-1.35603869e+00 -1.35723555e+00 -3.88135500e-02 2.01063037e+00
7.17314363e-01 3.01764816e-01 1.96280822e-01 2.81561792e-01
6.93768919e-01 4.56069738e-01 -2.41631716e-01 -4.32408810e-01
-5.43751359e-01 8.31316933e-02 5.65551639e-01 4.99054909e-01
-8.85414660e-01 7.75637507e-01 6.37547970e+00 9.56432283e-01
-8.10336888e-01 2.70433903e-01 6.97424114e-01 -3.61267030e-01
-3.05451632e-01 -1.97474405e-01 -4.50687438e-01 6.94058716e-01
1.52247012e+00 -4.54075187e-01 6.93659842e-01 7.86917865e-01
5.80482185e-01 -7.40618557e-02 -8.99592936e-01 1.12895167e+00
1.37000397e-01 -1.66085827e+00 2.85394073e-01 -1.18661135e-01
8.60929489e-01 2.80413955e-01 2.48776063e-01 -1.79321747e-02
7.96429217e-02 -9.84269381e-01 1.12164879e+00 3.12150270e-01
1.53129888e+00 -5.41155338e-01 7.14539886e-01 -5.05327523e-01
-1.55495548e+00 -7.67410621e-02 -3.93869877e-01 2.00362980e-01
2.65105367e-01 2.47978970e-01 2.77768038e-02 6.15769506e-01
1.20193839e+00 8.65155995e-01 -8.06535065e-01 9.57726717e-01
-1.01286164e-02 4.60275561e-01 2.61441082e-01 7.34057307e-01
-3.41305822e-01 1.69979662e-01 7.76455820e-01 1.53262973e+00
7.41502404e-01 2.59340525e-01 -2.14207143e-01 5.58218420e-01
-4.49195579e-02 1.85096145e-01 -3.59223276e-01 1.32994667e-01
7.49079108e-01 8.75342369e-01 -5.06076634e-01 -4.97396111e-01
-5.60802758e-01 1.01162374e+00 -2.25224018e-01 3.39913011e-01
-1.13381135e+00 -1.53630823e-01 7.83036470e-01 3.58292907e-01
3.48802984e-01 -2.55083054e-01 -1.39630497e-01 -1.26242411e+00
-1.27134144e-01 -1.35229981e+00 4.66518700e-01 -1.27598810e+00
-8.17364037e-01 8.65881085e-01 2.30042696e-01 -1.70377254e+00
-1.44558534e-01 -2.92203695e-01 -1.49167746e-01 4.81697410e-01
-1.56233704e+00 -4.77426082e-01 -6.34405375e-01 7.97872365e-01
1.05505514e+00 -3.08080614e-01 5.07924676e-01 4.92535830e-01
-2.20161468e-01 5.72838545e-01 7.55428523e-02 -9.99019966e-02
8.04515004e-01 -8.47662151e-01 3.18132013e-01 1.15042865e+00
-2.43691072e-01 8.42325389e-02 1.12613261e+00 -4.71328080e-01
-1.22525406e+00 -9.36268449e-01 4.48355138e-01 -2.17319742e-01
7.62223363e-01 1.77208200e-01 -9.91249859e-01 4.63994712e-01
3.75012755e-01 1.54001400e-01 5.58595181e-01 -5.04742026e-01
-6.08226895e-01 -1.26090944e-01 -1.29271972e+00 5.14327645e-01
1.43941391e+00 -5.60940683e-01 -3.97522271e-01 -7.93516915e-03
1.01435530e+00 -4.68298495e-01 -1.42351747e+00 3.94159466e-01
5.37492990e-01 -1.69741225e+00 1.21661139e+00 -2.33586542e-02
1.39816630e+00 -3.73175710e-01 -9.17838395e-01 -9.94560838e-01
-6.18846953e-01 -3.95566225e-01 -4.32637036e-01 8.99237394e-01
3.38734210e-01 -1.87304839e-01 3.33648622e-01 1.63615033e-01
-9.20603946e-02 -3.25265437e-01 -7.73087621e-01 -1.11315167e+00
-2.34269127e-01 -4.25177813e-01 5.43325365e-01 7.62807190e-01
-4.85979393e-03 -9.43387747e-02 -3.58700067e-01 2.53180206e-01
8.79881382e-01 -2.58004129e-01 5.31958759e-01 -5.00865340e-01
-4.68072802e-01 -5.08159220e-01 -5.81045449e-01 -6.32015705e-01
-2.80923158e-01 -6.87741697e-01 -3.17172050e-01 -1.44931781e+00
4.50048834e-01 1.24984793e-01 -4.92436439e-01 1.44653795e-02
4.82005626e-02 7.69422829e-01 4.63009655e-01 5.28203368e-01
-1.09701002e+00 1.38593808e-01 1.11275291e+00 2.12707132e-01
-1.36022931e-02 -5.13459325e-01 -4.62684989e-01 5.39331079e-01
1.30951369e+00 -1.94801256e-01 -2.32111081e-01 -6.57823086e-01
6.18839934e-02 5.26127756e-01 3.67044300e-01 -1.55266201e+00
-2.36608181e-02 -1.44622967e-01 3.00794005e-01 -2.88836479e-01
3.76219332e-01 -6.20334804e-01 5.85887015e-01 4.96666670e-01
-5.01438260e-01 4.93010432e-01 2.91772094e-02 2.13960961e-01
-3.40082020e-01 6.36425912e-02 1.04659748e+00 -7.86754265e-02
-1.30319333e+00 1.51122585e-01 -1.84989333e-01 2.68187761e-01
7.91570187e-01 -4.10331875e-01 -6.40035748e-01 -6.10003352e-01
-5.19816220e-01 -1.58083096e-01 1.20267892e+00 6.27389550e-01
1.10794854e+00 -1.23573112e+00 -1.22677970e+00 -8.68595466e-02
1.77389234e-01 -7.48138070e-01 4.29622710e-01 6.07786298e-01
-1.13844573e+00 1.20139316e-01 -7.43827164e-01 -5.48275173e-01
-1.60735762e+00 8.29059005e-01 1.09288827e-01 -7.79242367e-02
-6.29358351e-01 5.45131743e-01 -9.55962464e-02 5.13035357e-01
1.04106434e-01 -2.57578105e-01 -2.16281235e-01 -4.04450566e-01
9.32662427e-01 8.75319719e-01 -3.14976245e-01 -9.65598941e-01
-3.76061499e-01 5.72129607e-01 3.09229285e-01 -2.58734792e-01
1.21660852e+00 -4.88733083e-01 7.63977990e-02 2.45280817e-01
1.23658681e+00 1.71758845e-01 -1.19906211e+00 -2.85515357e-02
-1.09052151e-01 -1.06738436e+00 7.57939816e-02 -6.87790215e-01
-1.16258049e+00 2.63761789e-01 8.02427173e-01 1.99224919e-01
1.48805356e+00 -1.06031112e-01 7.55873442e-01 -2.87593484e-01
6.34406388e-01 -1.00410938e+00 9.29819942e-02 2.03991815e-01
1.19134080e+00 -1.33590877e+00 3.41798097e-01 -4.77926940e-01
-8.13287497e-01 1.03176296e+00 3.10845375e-01 -1.34038493e-01
4.21058953e-01 4.40655798e-01 2.31626313e-02 2.55477339e-01
-1.04255199e+00 7.78382272e-02 2.26008475e-01 6.07402146e-01
7.40960836e-01 2.44160015e-02 -3.74380261e-01 5.61263710e-02
-4.56511945e-01 3.79541099e-01 8.56965542e-01 6.87064588e-01
-4.88939911e-01 -7.36287713e-01 -3.65669191e-01 1.68929636e-01
-8.40968370e-01 -2.05294877e-01 1.67767242e-01 5.10738611e-01
-1.48517132e-01 1.27830374e+00 -3.82129438e-02 -7.61417508e-01
1.93971768e-01 -5.40575683e-01 4.49431062e-01 -2.08226442e-01
-2.62412399e-01 -9.33159962e-02 3.05910915e-01 -1.24113417e+00
-7.01758981e-01 -6.69162691e-01 -9.73439038e-01 -7.35140324e-01
3.08348984e-01 -1.48401195e-02 5.65019786e-01 1.75698563e-01
4.34823722e-01 5.87442696e-01 6.57243431e-01 -1.00049698e+00
-2.71776736e-01 -7.23945558e-01 -3.97045016e-01 6.63089991e-01
4.23975706e-01 -2.81276226e-01 -5.41258931e-01 4.80406702e-01]
|
[11.503279685974121, -1.8006539344787598]
|
3409fe84-e697-45d5-acb3-3495de3aa357
|
taxonomy-completion-via-triplet-matching
|
2101.01896
| null |
https://arxiv.org/abs/2101.01896v3
|
https://arxiv.org/pdf/2101.01896v3.pdf
|
Taxonomy Completion via Triplet Matching Network
|
Automatically constructing taxonomy finds many applications in e-commerce and web search. One critical challenge is as data and business scope grow in real applications, new concepts are emerging and needed to be added to the existing taxonomy. Previous approaches focus on the taxonomy expansion, i.e. finding an appropriate hypernym concept from the taxonomy for a new query concept. In this paper, we formulate a new task, "taxonomy completion", by discovering both the hypernym and hyponym concepts for a query. We propose Triplet Matching Network (TMN), to find the appropriate <hypernym, hyponym> pairs for a given query concept. TMN consists of one primal scorer and multiple auxiliary scorers. These auxiliary scorers capture various fine-grained signals (e.g., query to hypernym or query to hyponym semantics), and the primal scorer makes a holistic prediction on <query, hypernym, hyponym> triplet based on the internal feature representations of all auxiliary scorers. Also, an innovative channel-wise gating mechanism that retains task-specific information in concept representations is introduced to further boost model performance. Experiments on four real-world large-scale datasets show that TMN achieves the best performance on both taxonomy completion task and the previous taxonomy expansion task, outperforming existing methods.
|
['Lei LI', 'Yuning Mao', 'Jiaming Shen', 'Jiaze Chen', 'Ying Zeng', 'Xiangchen Song', 'Jieyu Zhang']
|
2021-01-06
| null | null | null | null |
['taxonomy-expansion']
|
['natural-language-processing']
|
[ 2.49250978e-01 -3.03104281e-01 -5.20940840e-01 -6.83493793e-01
-2.39216000e-01 -5.83338022e-01 3.06277096e-01 4.96598810e-01
-3.21166307e-01 2.46532172e-01 2.03524381e-01 -1.71002507e-01
-6.37952447e-01 -1.04653454e+00 -7.97221716e-03 -2.87624121e-01
7.66972601e-02 7.86450267e-01 1.39942184e-01 -5.13969898e-01
1.85352355e-01 1.45476535e-01 -1.98181760e+00 5.04318476e-01
8.61115396e-01 1.35242927e+00 4.61832672e-01 -1.66572511e-01
-5.82170010e-01 2.51808822e-01 -5.92056334e-01 -5.04234493e-01
2.92794496e-01 -1.55100957e-01 -7.85019994e-01 -2.54947215e-01
1.45941734e-01 2.30916306e-01 -1.14860579e-01 1.33058238e+00
2.34955817e-01 2.58372098e-01 2.99577564e-01 -1.48314178e+00
-4.28288996e-01 6.19190812e-01 -5.31566262e-01 2.48965204e-01
4.19755548e-01 -4.47678000e-01 1.79529524e+00 -8.98375809e-01
6.21914804e-01 1.30949092e+00 2.76066810e-01 2.85800248e-01
-1.12059450e+00 -1.26766086e+00 3.45575243e-01 5.32260001e-01
-1.56917000e+00 2.79299557e-01 6.66768610e-01 -2.88891464e-01
1.05498755e+00 3.20303708e-01 6.94651663e-01 7.71294773e-01
-1.67588741e-01 3.78281415e-01 7.62322724e-01 -2.79110253e-01
2.16584280e-01 1.95392057e-01 3.19720209e-01 3.22482824e-01
3.80561084e-01 1.29343554e-01 -7.02394724e-01 -3.95571083e-01
5.37451565e-01 4.20612067e-01 -4.75733913e-02 -3.69071424e-01
-1.02787507e+00 1.06310380e+00 7.90810406e-01 4.23052847e-01
-6.19697154e-01 -2.56172568e-01 4.20652568e-01 4.77143228e-01
5.46219014e-02 7.99319804e-01 -7.72490978e-01 2.31021896e-01
-6.72178924e-01 3.29745382e-01 7.53639519e-01 1.14577532e+00
1.06677771e+00 -4.41558391e-01 -9.86997932e-02 9.52212930e-01
2.88739502e-01 1.31347194e-01 9.51178253e-01 -6.67700410e-01
5.19778192e-01 1.46638727e+00 -2.18746305e-01 -9.53405142e-01
-6.28477097e-01 -6.50739431e-01 -6.84865057e-01 -4.71342713e-01
-3.18632692e-01 5.77806354e-01 -8.76445353e-01 1.69571018e+00
4.80028480e-01 2.05920517e-01 6.48690527e-03 8.91341865e-01
1.24545562e+00 4.79549468e-01 3.31454873e-01 -2.62259185e-01
1.79242229e+00 -5.40885031e-01 -5.21980286e-01 -1.76854998e-01
6.50438607e-01 -6.89808309e-01 1.09978735e+00 3.84191334e-01
-3.82740170e-01 -4.17784244e-01 -9.09212053e-01 2.56009083e-02
-7.30903268e-01 -1.36150420e-01 1.10661852e+00 5.09891748e-01
-4.78237659e-01 3.32639039e-01 5.03765084e-02 -6.06164336e-01
1.26599297e-01 6.31527364e-01 -3.17288578e-01 -4.36715305e-01
-1.74294341e+00 7.11444199e-01 9.50788558e-01 -2.37762570e-01
-5.72855175e-01 -6.93066955e-01 -8.86228085e-01 4.49426115e-01
5.72496772e-01 -6.32460654e-01 1.21557522e+00 -6.20425224e-01
-7.30330944e-01 9.54825461e-01 -2.20868617e-01 -4.06108320e-01
-4.84084815e-01 2.85016775e-01 -1.09909678e+00 -7.70721734e-02
4.76427078e-01 6.66029572e-01 3.83499980e-01 -1.12648857e+00
-1.26343942e+00 -6.26984894e-01 1.20182656e-01 5.23106039e-01
-6.93729758e-01 1.29669487e-01 -4.04349208e-01 -6.34089947e-01
7.09525108e-01 -4.52335685e-01 -1.86870649e-01 -5.46336532e-01
-1.12660110e-01 -6.72680914e-01 5.59556127e-01 -2.57722080e-01
1.55337322e+00 -1.82628548e+00 -9.46108922e-02 7.51560390e-01
5.99458277e-01 2.45434910e-01 -3.56302112e-01 4.81219947e-01
-3.24624300e-01 -4.28016596e-02 -2.28989609e-02 3.28908533e-01
5.40181547e-02 5.31736314e-01 -4.45945829e-01 -2.57510573e-01
-2.02280462e-01 1.08418238e+00 -9.78592098e-01 -3.72476548e-01
9.39281285e-03 -3.22386660e-02 -5.58805048e-01 6.01769611e-02
-2.40319192e-01 -8.20624270e-03 -4.81929779e-01 8.74012351e-01
5.13692856e-01 -3.94594312e-01 4.60495770e-01 -3.93329382e-01
2.85318822e-01 5.18885732e-01 -1.08815944e+00 1.80058527e+00
-6.47161365e-01 -1.15557924e-01 -2.33713210e-01 -1.21236718e+00
1.38227034e+00 2.93249905e-01 7.84181476e-01 -1.14416933e+00
1.49806127e-01 6.18372083e-01 3.07535291e-01 -4.56099749e-01
4.82569098e-01 -3.99746478e-01 -2.62959242e-01 3.14602852e-01
3.53144497e-01 1.37957916e-01 3.25255454e-01 2.18537107e-01
1.05091619e+00 -2.80765802e-01 6.83775485e-01 -2.18002483e-01
5.76216578e-01 -1.18531967e-02 9.16119397e-01 4.41681415e-01
5.68198524e-02 8.30018595e-02 2.45721281e-01 -5.71592093e-01
-7.45376766e-01 -1.04416132e+00 -8.49148706e-02 1.48856068e+00
5.31816542e-01 -9.62086678e-01 -2.14605317e-01 -5.59724629e-01
2.94633567e-01 6.34105802e-01 -3.77977937e-01 -4.53157395e-01
-3.17381859e-01 -7.82278538e-01 3.23985010e-01 4.11178857e-01
4.10865009e-01 -1.27224696e+00 -6.17964447e-01 4.47471231e-01
-4.69996810e-01 -1.12668753e+00 -3.74942333e-01 3.34630013e-01
-9.10558641e-01 -1.01872551e+00 -1.23830073e-01 -8.41670573e-01
2.59838015e-01 4.72473651e-01 1.34994066e+00 2.37699732e-01
-4.21581417e-01 6.21580705e-03 -6.66454494e-01 -3.58807147e-01
1.66432619e-01 2.71056652e-01 4.16655540e-02 8.40482190e-02
1.17846000e+00 -7.62987196e-01 -4.39882606e-01 5.72120905e-01
-9.65264261e-01 -2.42724746e-01 6.18804574e-01 9.17113721e-01
6.40105665e-01 4.20154423e-01 7.97648251e-01 -9.28376973e-01
9.14327562e-01 -6.76618516e-01 -4.16825563e-01 3.87274504e-01
-1.14132392e+00 -3.55812982e-02 4.70532656e-01 -4.78913128e-01
-8.59274030e-01 -9.37484205e-02 -2.33031018e-03 -3.52351159e-01
3.99960577e-02 1.02874851e+00 -5.34789443e-01 2.43891627e-01
6.36058569e-01 1.90463871e-01 -4.91987526e-01 -6.35648191e-01
3.97106975e-01 7.12254941e-01 6.38028622e-01 -5.11885226e-01
8.11638355e-01 3.37200910e-01 -1.31761460e-02 -3.81212622e-01
-1.01388407e+00 -1.26746523e+00 -5.19547582e-01 3.28512549e-01
4.59335119e-01 -9.28515196e-01 -1.01032400e+00 -2.36253291e-01
-1.04707849e+00 6.43450499e-01 -4.70834851e-01 5.02206743e-01
-2.52292216e-01 1.38736486e-01 -2.93107443e-02 -4.94268835e-01
-5.54955721e-01 -7.70371675e-01 1.01201546e+00 2.64101863e-01
-4.01846319e-01 -6.30652130e-01 -1.44497260e-01 4.07256901e-01
2.24012703e-01 -3.21437836e-01 1.38216078e+00 -1.20729101e+00
-3.83285135e-01 -4.70993996e-01 -5.23364186e-01 -3.02037243e-02
1.95480078e-01 -7.47733414e-01 -7.77170420e-01 -1.73010558e-01
1.53833210e-01 -2.14180529e-01 1.05164695e+00 2.32109372e-02
1.14649332e+00 -4.18259889e-01 -5.43868482e-01 6.47547543e-01
1.29713082e+00 4.85428870e-01 4.14364666e-01 3.26407403e-01
6.28796458e-01 8.00316751e-01 9.60712671e-01 4.88689542e-01
5.01826108e-01 7.80491531e-01 4.74609494e-01 1.40011698e-01
2.73940533e-01 -4.67397064e-01 -1.68844759e-01 6.21546030e-01
2.43135333e-01 -8.72261524e-02 -8.15084517e-01 4.35077667e-01
-1.92157674e+00 -9.01943266e-01 7.92962760e-02 2.35159326e+00
8.41835558e-01 2.68836971e-02 2.41221534e-03 1.94274709e-01
7.23733485e-01 -2.04419672e-01 -8.04438233e-01 -9.34688747e-02
-9.06862691e-02 4.66032028e-01 1.20843321e-01 1.41403139e-01
-8.42071652e-01 1.20892429e+00 4.86838150e+00 1.05631292e+00
-8.01482677e-01 1.24242298e-01 -1.72856711e-02 6.46928186e-03
-7.28046119e-01 3.50884467e-01 -1.01698971e+00 2.66679138e-01
4.52184886e-01 -4.87069428e-01 4.49245751e-01 8.90971005e-01
-1.46027312e-01 1.61511883e-01 -9.73975122e-01 1.29658365e+00
4.19968814e-02 -1.15071344e+00 5.58083355e-01 1.64742023e-01
4.14555371e-01 -4.60057966e-02 -1.82898864e-01 6.35452867e-01
3.39768767e-01 -1.07019138e+00 1.61966950e-01 -1.64922606e-03
9.66359138e-01 -6.22308314e-01 6.94439411e-01 3.21179569e-01
-1.81876159e+00 -6.09962881e-01 -5.69752216e-01 -4.39254008e-03
-1.61045734e-02 6.29126370e-01 -8.64304066e-01 8.11625838e-01
7.48510659e-01 7.08768666e-01 -5.42788446e-01 1.25466347e+00
-3.73149306e-01 2.55998254e-01 -3.75452965e-01 -1.59276333e-02
2.06061199e-01 -1.26345620e-01 3.73438448e-01 8.58772457e-01
3.79298925e-01 2.06876606e-01 4.56529677e-01 8.84458184e-01
-2.71960855e-01 4.26554739e-01 -2.76002556e-01 -5.59483133e-02
9.32860255e-01 1.60574305e+00 -6.45395398e-01 -3.91643047e-01
-3.03103983e-01 6.38614178e-01 1.52998060e-01 4.02523637e-01
-4.74013716e-01 -4.99787390e-01 8.83705795e-01 -5.23864962e-02
4.22959290e-02 2.99762547e-01 -3.00641865e-01 -1.21782279e+00
1.51601568e-01 -9.16152358e-01 1.13310301e+00 -5.85683048e-01
-1.33744621e+00 5.17475784e-01 -4.01469693e-02 -1.25783145e+00
-2.16299400e-01 -5.14656365e-01 -5.74999034e-01 1.02371097e+00
-1.50175726e+00 -1.27823901e+00 -4.82556015e-01 6.02296054e-01
4.01395738e-01 -4.90891129e-01 1.11080968e+00 4.54749435e-01
-1.24220513e-01 5.78117073e-01 -2.02993512e-01 -7.19095170e-02
4.12652761e-01 -9.26761627e-01 3.66980702e-01 3.64547014e-01
5.35070360e-01 1.00264347e+00 3.51132393e-01 -8.10890615e-01
-9.27811503e-01 -9.46630239e-01 1.40886486e+00 -1.17092274e-01
5.24758101e-01 -4.85537797e-01 -1.01637220e+00 4.16584939e-01
-3.01182002e-01 -4.56042200e-01 8.85569930e-01 6.08390689e-01
-1.07687259e+00 -4.00671989e-01 -1.05754960e+00 2.71975338e-01
1.26431394e+00 -5.09142399e-01 -7.59727001e-01 3.08074713e-01
1.21229017e+00 2.49563269e-02 -8.61635447e-01 7.01334655e-01
6.53937757e-01 -5.57062924e-01 1.13428450e+00 -9.54574287e-01
-6.63842559e-02 -3.87863427e-01 -5.53387403e-01 -1.19800627e+00
-4.43394005e-01 -3.24132353e-01 -2.04185471e-01 1.08891082e+00
3.75072122e-01 -6.66195869e-01 8.58867049e-01 2.51761526e-01
7.39516914e-02 -8.46405506e-01 -1.09267795e+00 -9.96467233e-01
-2.63396949e-01 -4.03853238e-01 1.23502660e+00 1.15025151e+00
4.06975538e-01 8.61910522e-01 -4.58201356e-02 -1.18879072e-01
3.71522456e-01 6.63996637e-01 2.40363762e-01 -2.05791640e+00
3.46999546e-03 -6.84600830e-01 -6.39730513e-01 -8.87261331e-01
4.26469967e-02 -1.35279393e+00 -4.04212683e-01 -1.60329700e+00
3.84066582e-01 -4.63967681e-01 -7.19571829e-01 7.19250321e-01
-1.00407265e-01 -1.14200354e-01 1.06006913e-01 4.52953547e-01
-5.85434437e-01 6.16293669e-01 1.00516784e+00 -2.72052228e-01
-2.67625809e-01 2.68758625e-01 -1.06203508e+00 4.62472230e-01
6.50059760e-01 -5.92460454e-01 -7.09854960e-01 -1.06359929e-01
6.61051989e-01 -1.77680716e-01 2.58239031e-01 -6.74624681e-01
3.53823811e-01 -1.71419248e-01 1.55800670e-01 -5.38250864e-01
4.02087718e-01 -9.38754678e-01 -9.18811113e-02 4.38698173e-01
-4.22260255e-01 1.34737551e-01 -4.19067144e-02 5.34080625e-01
-5.32839715e-01 -3.63642722e-01 4.69490141e-01 -1.78728312e-01
-1.12719238e+00 5.37227511e-01 2.32072949e-01 -3.57691534e-02
7.03142583e-01 -3.32868159e-01 -1.36496395e-01 -1.34145066e-01
-5.99959254e-01 6.35956466e-01 -3.52639928e-02 9.63974953e-01
8.14422846e-01 -1.51606572e+00 -4.12975758e-01 2.26125002e-01
8.03926229e-01 -1.50873065e-01 8.09033662e-02 3.55206132e-01
2.50604093e-01 7.85788417e-01 -1.38420343e-01 -3.27470005e-01
-1.18975925e+00 7.71024287e-01 1.32995069e-01 -4.85177726e-01
-3.33383858e-01 9.60715353e-01 6.56930208e-01 -8.58696938e-01
3.74275714e-01 -2.17412010e-01 -6.60062253e-01 2.50706315e-01
4.60923016e-01 1.20320641e-01 3.26421231e-01 -6.71501637e-01
-4.85574841e-01 5.40217340e-01 -2.09565625e-01 4.65272181e-02
1.01806259e+00 6.99054450e-03 -3.75444174e-01 2.86862075e-01
1.13571382e+00 -7.01378465e-01 -1.21025726e-01 -7.94764400e-01
5.19328713e-01 -5.23326933e-01 -2.50653654e-01 -1.00450313e+00
-9.53821182e-01 8.73682737e-01 5.80526412e-01 9.29774642e-02
1.47468352e+00 1.69858620e-01 9.00136530e-01 4.87657219e-01
6.31445467e-01 -1.05232954e+00 6.60672709e-02 5.94972730e-01
7.38931000e-01 -1.02193689e+00 -1.33049503e-01 -6.76584184e-01
-5.74391246e-01 8.18022311e-01 8.92765880e-01 3.01718682e-01
6.57885909e-01 -3.10141146e-01 -6.50737882e-02 -6.45609200e-01
-7.55734146e-01 -5.13745308e-01 5.33589542e-01 6.07721508e-01
3.93919766e-01 2.79440641e-01 -5.96045196e-01 1.04424644e+00
-4.21603262e-01 -1.26814082e-01 -1.16864495e-01 5.18725038e-01
-6.56309426e-01 -1.31366980e+00 -2.43964344e-02 9.24181938e-01
7.31579587e-02 -4.80744869e-01 -3.16469818e-01 3.58270228e-01
5.60795069e-01 9.64143276e-01 -2.68589854e-02 -8.20873678e-01
4.78140354e-01 1.17386259e-01 9.57476869e-02 -1.00496733e+00
-6.88408077e-01 -4.32932526e-02 -2.16627091e-01 -5.67614496e-01
-7.43741766e-02 -3.43947470e-01 -1.50150800e+00 7.99133629e-02
-5.29328048e-01 3.98874491e-01 5.79167843e-01 9.41714883e-01
4.95385319e-01 2.64905930e-01 4.15534377e-01 -3.24831307e-02
-6.97439969e-01 -9.17959213e-01 -8.69504154e-01 7.20348179e-01
-2.79354453e-01 -9.85080063e-01 5.53467609e-02 -4.15682703e-01]
|
[9.224791526794434, 8.004592895507812]
|
27214a65-b7dc-4074-a81f-877f39479206
|
sfcw-gpr-tree-roots-detection-enhancement-by
|
2204.02594
| null |
https://arxiv.org/abs/2204.02594v1
|
https://arxiv.org/pdf/2204.02594v1.pdf
|
SFCW GPR tree roots detection enhancement by time frequency analysis in tropical areas
|
Accurate monitoring of tree roots using ground penetrating radar (GPR) is very useful in assessing the trees health. In high moisture tropical areas such as Singapore, tree fall due to root rot can cause loss of lives and properties. The tropical complex soil characteristics due to the high moisture content tends to affect penetration depth of the signal. This limits the depth range of the GPR. Typically, a wide band signal is used to increase the penetration depth and to improve the resolution of the GPR. However, this broad band frequency tends to be noisy and selective frequency filtering is required for noise reduction. Therefore, in this paper, we adapt the stepped frequency continuous wave (SFCW) GPR and propose the use of a Joint time frequency analysis (JTFA) method called short time Fourier transform (STFT), to reduce noise and enhance tree root detection. The proposed methodology is illustrated and tested with controlled experiments and real tree roots testing. The results show promising prospects of the method for tree roots detection in tropical areas.
|
['Mohamed Lokman Mohd Yusof', 'Genevieve Ow', 'Abdulkadir C. Yucel', 'Yee Hui Lee', 'Wenhao Luo']
|
2022-04-06
| null | null | null | null |
['gpr', 'gpr']
|
['computer-vision', 'miscellaneous']
|
[ 6.43328547e-01 -3.23016077e-01 2.48103544e-01 1.74360290e-01
-4.46545452e-01 -2.78507054e-01 -1.14944004e-01 2.49734864e-01
1.76964805e-01 9.02012825e-01 -1.31313488e-01 -4.65317369e-01
-5.03355742e-01 -1.21460414e+00 1.12975113e-01 -9.50677454e-01
-4.41751450e-01 -2.63093174e-01 3.11648130e-01 -1.17017843e-01
3.85421097e-01 7.63567626e-01 -1.73241520e+00 -6.13786057e-02
1.01588166e+00 7.65249848e-01 1.08434057e+00 5.07429659e-01
1.33106131e-02 4.78197902e-01 -6.50379062e-01 4.47275549e-01
-1.81951046e-01 -6.86013997e-02 -5.42941153e-01 -1.52106076e-01
-3.86171579e-01 -3.88283998e-01 3.06740642e-01 8.66223633e-01
5.50874174e-01 1.28801195e-02 5.48671126e-01 -5.11833012e-01
-8.68553221e-02 3.49172771e-01 -1.11873043e+00 3.04792464e-01
5.52175939e-01 -4.52220351e-01 4.90311354e-01 -7.44652092e-01
2.43814260e-01 9.79591131e-01 9.25591171e-01 -2.19646711e-02
-1.00521755e+00 -7.81515539e-01 -4.29780930e-01 3.23428661e-01
-8.99419427e-01 -2.01667264e-01 7.32343316e-01 -4.49911803e-02
6.31715357e-01 2.81444669e-01 8.80349994e-01 5.81773341e-01
7.55747736e-01 1.58007115e-01 1.26141477e+00 -6.48099005e-01
4.08052921e-01 -4.35156822e-01 -3.14661339e-02 2.79880255e-01
5.81318498e-01 3.93793166e-01 5.29397465e-02 -3.32392544e-01
5.97463906e-01 -2.72383630e-01 -5.43877602e-01 1.31598055e-01
-4.19731885e-01 9.66812432e-01 5.01344562e-01 7.39371419e-01
-9.18963909e-01 -6.89835995e-02 3.65355611e-01 -7.47195408e-02
6.65277004e-01 2.26456210e-01 -3.11228544e-01 -3.70370001e-02
-9.32462275e-01 1.95499852e-01 9.06132340e-01 1.02186307e-01
1.39913484e-01 3.10910791e-01 3.23091805e-01 9.81385887e-01
4.72485989e-01 8.94321263e-01 2.72326291e-01 -7.98545301e-01
-2.42679656e-01 -2.32711315e-01 1.16823336e-04 -1.23180962e+00
-2.57779956e-01 -5.63593924e-01 -6.10775054e-01 2.94837523e-02
1.50366828e-01 -1.29356384e-01 -8.23353529e-01 1.24499798e+00
4.78970855e-01 3.11292037e-02 3.49855274e-02 5.23475885e-01
7.65618861e-01 1.05680466e+00 4.32942688e-01 -6.24215126e-01
1.51014602e+00 1.70658715e-02 -9.34923530e-01 -3.11237842e-01
6.09465480e-01 -1.14788616e+00 5.98811626e-01 5.41413426e-01
-6.01035178e-01 -2.29867861e-01 -1.07515192e+00 9.00456071e-01
-1.03005350e-01 8.94208699e-02 8.78429472e-01 9.11897123e-01
-4.90961730e-01 5.36413014e-01 -1.13362920e+00 -6.82888210e-01
2.68945575e-01 -8.61436725e-02 -8.22089463e-02 -6.23099864e-01
-1.04327285e+00 9.90247428e-01 -1.02102429e-01 4.74993527e-01
2.86535341e-02 -6.37113392e-01 -7.06133842e-01 9.66911484e-03
-1.98931396e-01 -8.63935682e-04 1.07265568e+00 8.19965377e-02
-1.53255379e+00 3.64663005e-01 1.52970374e-01 -2.42398694e-01
-6.04814552e-02 -4.55278367e-01 -6.77656591e-01 4.55387771e-01
5.87722838e-01 -1.90680668e-01 5.36355138e-01 -1.11042786e+00
-6.56591654e-01 -6.42735243e-01 -7.84995675e-01 -3.04748893e-01
2.45608225e-01 5.33204898e-02 6.86782658e-01 -5.08094549e-01
8.98143589e-01 -5.56448996e-01 -3.91023606e-01 -5.54484844e-01
-1.27318054e-01 2.65422970e-01 1.45739675e+00 -1.06239235e+00
9.14964795e-01 -2.20403290e+00 -7.43466675e-01 4.20067221e-01
-3.38144958e-01 1.63437098e-01 2.34612301e-01 6.97957575e-01
-1.42975762e-01 -1.03417970e-01 -4.25406873e-01 5.94824791e-01
-8.79429102e-01 3.06724876e-01 -2.45212436e-01 7.07608223e-01
-1.08183116e-01 -4.11718898e-03 -7.88303196e-01 -3.16272914e-01
2.81013519e-01 7.26323962e-01 -1.66673996e-02 -1.47247732e-01
5.99231362e-01 1.87039703e-01 -9.39273953e-01 1.12886763e+00
1.40715694e+00 6.13872588e-01 -1.32713199e-01 -4.42987502e-01
-6.88407958e-01 9.71944183e-02 -1.15445948e+00 7.58636117e-01
-5.82497299e-01 2.45200589e-01 5.43517113e-01 -1.27559471e+00
1.41505563e+00 4.20751661e-01 6.29252374e-01 -7.74523914e-01
-1.51370421e-01 5.18298984e-01 -9.30480585e-02 -6.71679676e-01
3.02214593e-01 -4.45158094e-01 3.98290187e-01 -8.51117596e-02
-4.86076206e-01 -5.93015790e-01 -1.43068478e-01 -1.76913723e-01
1.13373542e+00 2.20688015e-01 4.76414740e-01 -5.13409495e-01
2.37791345e-01 1.14664370e-02 5.63996911e-01 2.46871501e-01
-9.09506530e-02 -6.29539648e-03 1.22824023e-02 -6.21864991e-03
-5.61989367e-01 -9.63903964e-01 -8.06971431e-01 7.13886440e-01
-1.22841701e-01 7.76554644e-02 -2.78174937e-01 2.39154026e-01
1.26562729e-01 9.45083797e-01 -3.22132200e-01 -2.63967842e-01
-4.15810227e-01 -1.17292750e+00 3.91313434e-01 3.02920818e-01
7.51886666e-01 -1.30619395e+00 -1.23804772e+00 9.67341721e-01
-3.46096009e-01 -8.80313694e-01 5.40082932e-01 7.79468417e-01
-1.22726119e+00 -8.54147196e-01 -6.52127326e-01 -5.91185391e-01
-4.41246592e-02 5.12402534e-01 6.91797018e-01 -7.23214075e-02
-9.19100821e-01 1.96653157e-01 -9.29835081e-01 -5.18360674e-01
-1.34224981e-01 -1.77844718e-01 -4.27175999e-01 -6.72371864e-01
3.22224140e-01 -1.07328928e+00 -4.34767932e-01 2.09070116e-01
-5.61923683e-01 -6.79681003e-01 7.24389553e-01 8.63113582e-01
3.89360309e-01 9.38586414e-01 1.02665985e+00 -7.46302366e-01
7.25108206e-01 -5.78387141e-01 -3.57313544e-01 -1.89574361e-01
-3.31204563e-01 -5.84119081e-01 1.45341471e-01 -2.69011259e-01
-1.07886672e+00 -2.23465800e-01 -1.07715547e-01 6.13447487e-01
-2.45099649e-01 9.34337556e-01 -4.48724143e-02 -2.55097568e-01
6.96885586e-01 -1.19738448e-02 -1.16621368e-01 -5.60320795e-01
-2.40305096e-01 5.12239337e-01 6.33714139e-01 -4.71359968e-01
6.70974135e-01 2.70405918e-01 3.81037533e-01 -1.61741805e+00
-6.31829500e-01 -5.31742811e-01 9.92213190e-03 -3.02258015e-01
6.73036337e-01 -4.97850597e-01 -4.09076035e-01 6.10581338e-01
-7.84771740e-01 -5.15045226e-02 -1.89581942e-02 9.98284459e-01
-3.88829887e-01 4.51755345e-01 -5.51242054e-01 -1.42225528e+00
-7.16484666e-01 -4.51591611e-01 6.87062263e-01 4.17402655e-01
-1.54648498e-01 -6.06567919e-01 5.35049662e-02 -7.36543909e-02
6.05437577e-01 7.60651708e-01 7.77880371e-01 2.46153802e-01
1.42021596e-01 -4.31696951e-01 -1.85937092e-01 1.17188528e-01
4.39903110e-01 4.08407599e-01 -7.71950424e-01 -2.04679325e-01
4.51426804e-01 -5.56777744e-03 7.12652981e-01 1.12270701e+00
7.10369885e-01 3.20057243e-01 -5.90686202e-01 4.28169072e-01
1.83193636e+00 5.34525752e-01 9.34566259e-01 5.55375040e-01
5.06625548e-02 8.53087604e-01 1.24768710e+00 7.06195474e-01
-4.02248263e-01 1.65767428e-02 5.78982294e-01 -1.94103584e-01
2.92174637e-01 1.26213834e-01 2.03277662e-01 5.51648736e-01
-4.32320207e-01 -1.36760071e-01 -6.97705090e-01 6.42180800e-01
-1.00695956e+00 -1.08715034e+00 -6.40282810e-01 2.13589096e+00
3.41075629e-01 -1.17578208e-02 -1.80822134e-01 1.08753848e+00
8.84083748e-01 -1.54154405e-01 -2.85343695e-02 -5.65651834e-01
4.96442243e-02 1.09426582e+00 9.35030162e-01 2.93073207e-01
-5.10741591e-01 5.14102221e-01 5.93406534e+00 5.77661812e-01
-1.33698678e+00 -1.89386979e-01 -1.21581189e-01 7.41639256e-01
-1.65402163e-02 1.61124423e-01 -3.04871708e-01 8.39101300e-02
6.29603446e-01 3.42808664e-02 -1.94466606e-01 7.44525671e-01
4.84138578e-01 -1.03818166e+00 -1.48059940e-03 4.70439881e-01
-8.49957883e-01 -3.98315966e-01 -5.33463478e-01 1.91961452e-01
-2.17142269e-01 -1.89413428e-01 -2.19221130e-01 1.93426907e-01
1.04315512e-01 -8.14247191e-01 7.89072514e-02 3.33179124e-02
5.50370455e-01 -1.18939149e+00 8.17474365e-01 2.93260306e-01
-1.66836095e+00 -1.18672170e-01 -6.17385626e-01 -9.77922231e-02
4.88328993e-01 1.46352911e+00 -7.07481623e-01 7.48127580e-01
1.03947854e+00 4.26865608e-01 -1.60343945e-01 1.37147081e+00
-1.07805394e-01 1.11240768e+00 -9.84747648e-01 2.06391290e-01
1.96753666e-01 -2.15183422e-01 6.35381103e-01 7.65669882e-01
1.12356913e+00 4.57534194e-01 5.00291213e-02 6.11246288e-01
9.19817030e-01 3.53106648e-01 -5.99108160e-01 -1.37391776e-01
8.34119260e-01 1.07297277e+00 -1.05445898e+00 3.81262898e-01
1.27774607e-02 4.06566024e-01 -7.58947849e-01 1.32653430e-01
-4.66763258e-01 -6.01775289e-01 4.57546785e-02 4.42194521e-01
5.86540520e-01 -2.19972044e-01 -1.16934665e-01 -4.14148301e-01
-2.21791252e-01 -4.66821641e-01 3.04816902e-01 -7.96639144e-01
-1.04523528e+00 2.71146744e-01 9.42066386e-02 -8.70761096e-01
5.57477549e-02 -1.61824703e-01 -9.16356623e-01 1.10669184e+00
-1.37589419e+00 -9.80976522e-01 -4.08549547e-01 4.20526713e-01
2.64113605e-01 3.66176873e-01 1.30935490e+00 2.48945639e-01
-1.28144071e-01 -8.27427283e-02 1.25834629e-01 -6.40402794e-01
1.45534024e-01 -7.36815929e-01 -2.33334973e-01 6.22420907e-01
-6.93706274e-01 2.21617147e-02 1.37522602e+00 -1.12582660e+00
-1.14706266e+00 -7.01152980e-01 6.75696373e-01 8.50264072e-01
3.68372560e-01 1.49947733e-01 -9.92365181e-01 -3.68301272e-02
-3.24383318e-01 -8.39286372e-02 6.64692521e-01 5.12736253e-02
2.43388608e-01 6.90569878e-02 -1.83680665e+00 -9.51019954e-03
3.59646857e-01 -1.07065879e-01 -3.34233493e-01 3.23957838e-02
6.68687746e-02 1.49233028e-01 -1.02358210e+00 9.48007286e-01
9.69668388e-01 -8.65595579e-01 1.04435921e+00 3.38991791e-01
1.00466982e-01 3.13422196e-02 -5.89393318e-01 -1.26320267e+00
-6.80939555e-01 -6.68600127e-02 3.96605760e-01 1.34724545e+00
-5.42498007e-02 -9.57067490e-01 5.31231642e-01 -4.47085679e-01
1.24283589e-01 -4.94587362e-01 -9.05174077e-01 -5.70189774e-01
-3.01538914e-01 -2.46882826e-01 1.29931122e-01 6.81650460e-01
-8.62422436e-02 7.23324493e-02 4.10126597e-02 5.10478795e-01
1.04925060e+00 5.16404286e-02 3.96833336e-03 -1.82470739e+00
-1.49825573e-01 2.63922840e-01 -4.90573019e-01 -2.73038536e-01
-2.76613414e-01 -5.92369549e-02 1.42278761e-01 -1.68591821e+00
-4.17056143e-01 -5.53846836e-01 1.46884724e-01 4.54886109e-02
3.42022963e-02 6.79280683e-02 -2.20008239e-01 8.71555321e-03
9.19813395e-01 3.64913940e-01 1.25315058e+00 2.40534201e-01
-4.25340265e-01 5.90563655e-01 -3.71573001e-01 6.61655247e-01
9.69417632e-01 -6.30406499e-01 -4.92400885e-01 2.79834718e-01
4.22081910e-02 6.72425985e-01 2.14650825e-01 -1.16706717e+00
-4.07970816e-01 -1.68874532e-01 3.54804605e-01 -1.25869179e+00
2.85384119e-01 -9.40916896e-01 3.14179301e-01 1.03050148e+00
5.28603375e-01 -3.11372668e-01 2.99320936e-01 5.03340781e-01
-6.31074980e-02 -3.63287598e-01 1.14856577e+00 -7.51247481e-02
-4.56329495e-01 -2.39919901e-01 -1.03904951e+00 -6.72297537e-01
8.10903728e-01 -7.49458969e-01 -7.85471052e-02 -3.66158426e-01
-8.62925112e-01 -1.51803493e-01 -1.12942941e-01 -2.98991978e-01
5.25747180e-01 -1.07912195e+00 -6.94401205e-01 1.28917053e-01
-2.46313602e-01 -1.77705824e-01 5.19050002e-01 9.53046918e-01
-7.43487597e-01 6.66100308e-02 -6.83581054e-01 -5.55961549e-01
-1.27093232e+00 -3.48607115e-02 1.30402043e-01 -2.87500143e-01
-9.81714249e-01 6.98439658e-01 -1.63900420e-01 1.56131592e-02
-3.06462765e-01 -4.11722809e-01 -8.18575740e-01 8.97340104e-02
3.63761216e-01 6.39650881e-01 2.51358509e-01 -3.10846299e-01
-4.85504180e-01 7.60740101e-01 1.32285550e-01 -5.18072806e-02
1.77090979e+00 -1.28613994e-01 -2.55642176e-01 4.15600002e-01
6.58594728e-01 3.05928677e-01 -5.02699614e-01 1.60015941e-01
5.22820391e-02 -5.24024546e-01 5.84380507e-01 -5.39555609e-01
-9.60335135e-01 4.90372539e-01 7.80856609e-01 7.62804091e-01
1.53912127e+00 -3.96793813e-01 8.66604805e-01 2.53701180e-01
4.84042466e-01 -1.02689016e+00 -5.22794664e-01 1.57685801e-01
7.23371923e-01 -4.37288642e-01 3.21720600e-01 -1.07226598e+00
2.04303473e-01 1.28912818e+00 1.90198958e-01 -3.88549715e-01
1.05166698e+00 8.77376735e-01 -7.41410702e-02 -2.72552460e-01
-3.07866812e-01 -2.51089305e-01 -7.54711211e-01 1.18077040e+00
7.66744614e-01 2.34664023e-01 -8.96347106e-01 2.96109408e-01
-4.87104386e-01 2.14273289e-01 5.49282193e-01 1.48759401e+00
-1.06435668e+00 -9.44703519e-01 -8.90419543e-01 6.06494188e-01
-7.67015338e-01 1.84693307e-01 1.72099456e-01 8.46916139e-01
-9.36547667e-02 1.21617377e+00 -3.51103485e-01 5.73021248e-02
3.28147352e-01 -1.23921514e-01 7.59320676e-01 -5.00699818e-01
-1.08910762e-01 3.67033541e-01 2.19458431e-01 -1.87635362e-01
-6.89739704e-01 -6.79668248e-01 -1.20593500e+00 -4.03824955e-01
-1.13813400e+00 4.02745456e-01 1.14352584e+00 7.86721706e-01
-4.90771443e-01 7.09194064e-01 9.02949333e-01 -5.10244548e-01
-4.33138669e-01 -1.29045820e+00 -1.47742283e+00 -4.29815173e-01
-2.65107527e-02 -1.14087772e+00 -5.86714208e-01 -2.75261402e-01]
|
[6.831613540649414, 1.392909049987793]
|
bf8323e1-2f14-4561-9052-1a0e16672234
|
neuralpci-spatio-temporal-neural-field-for-3d
|
2303.15126
| null |
https://arxiv.org/abs/2303.15126v1
|
https://arxiv.org/pdf/2303.15126v1.pdf
|
NeuralPCI: Spatio-temporal Neural Field for 3D Point Cloud Multi-frame Non-linear Interpolation
|
In recent years, there has been a significant increase in focus on the interpolation task of computer vision. Despite the tremendous advancement of video interpolation, point cloud interpolation remains insufficiently explored. Meanwhile, the existence of numerous nonlinear large motions in real-world scenarios makes the point cloud interpolation task more challenging. In light of these issues, we present NeuralPCI: an end-to-end 4D spatio-temporal Neural field for 3D Point Cloud Interpolation, which implicitly integrates multi-frame information to handle nonlinear large motions for both indoor and outdoor scenarios. Furthermore, we construct a new multi-frame point cloud interpolation dataset called NL-Drive for large nonlinear motions in autonomous driving scenes to better demonstrate the superiority of our method. Ultimately, NeuralPCI achieves state-of-the-art performance on both DHB (Dynamic Human Bodies) and NL-Drive datasets. Beyond the interpolation task, our method can be naturally extended to point cloud extrapolation, morphing, and auto-labeling, which indicates its substantial potential in other domains. Codes are available at https://github.com/ispc-lab/NeuralPCI.
|
['Changjun Jiang', 'Guang Chen', 'Fan Lu', 'Ruisi Lu', 'Danni Wu', 'Zehan Zheng']
|
2023-03-27
| null |
http://openaccess.thecvf.com//content/CVPR2023/html/Zheng_NeuralPCI_Spatio-Temporal_Neural_Field_for_3D_Point_Cloud_Multi-Frame_Non-Linear_CVPR_2023_paper.html
|
http://openaccess.thecvf.com//content/CVPR2023/papers/Zheng_NeuralPCI_Spatio-Temporal_Neural_Field_for_3D_Point_Cloud_Multi-Frame_Non-Linear_CVPR_2023_paper.pdf
|
cvpr-2023-1
|
['3d-point-cloud-interpolation']
|
['computer-vision']
|
[-1.40467107e-01 -3.89964074e-01 -1.65602267e-01 -3.84916633e-01
-6.81668758e-01 -2.28960931e-01 4.42602098e-01 -2.89341182e-01
-1.57757252e-01 5.22009730e-01 -3.83361727e-02 -1.72125116e-01
2.62247294e-01 -6.24132693e-01 -1.03176546e+00 -3.76783162e-01
5.08507267e-02 5.23844838e-01 3.56438696e-01 -3.10400903e-01
-1.06755391e-01 6.64612234e-01 -1.53902590e+00 -7.78419198e-03
1.11804175e+00 8.65444839e-01 3.24767679e-01 4.82134670e-01
1.58184692e-01 4.53066140e-01 1.41599402e-02 -4.39049989e-01
4.68743324e-01 -1.01199210e-01 -5.83902180e-01 1.60922140e-01
9.17324722e-01 -6.76923335e-01 -6.45507812e-01 8.76168966e-01
4.73106503e-01 2.87665635e-01 3.45779091e-01 -1.36577976e+00
-8.06623638e-01 -5.94204217e-02 -7.87915826e-01 1.60517008e-03
2.81680793e-01 5.90332448e-01 5.64267397e-01 -1.14465296e+00
6.87811434e-01 1.40869021e+00 1.04382718e+00 5.50882101e-01
-1.21535659e+00 -8.12998533e-01 7.99881741e-02 3.37098330e-01
-1.29018831e+00 -4.11223412e-01 7.82023430e-01 -5.99470377e-01
7.83109605e-01 -4.65400666e-02 9.51872051e-01 9.00956869e-01
-4.12606038e-02 9.28341925e-01 8.00553501e-01 -9.93778855e-02
-2.18434446e-02 -4.66398865e-01 -2.12043136e-01 5.31432450e-01
-2.96355281e-02 3.90813380e-01 -4.68218178e-01 8.67891535e-02
1.21073508e+00 8.35720077e-02 -3.03236991e-01 -5.11395156e-01
-1.56764305e+00 5.58333874e-01 7.17780590e-01 -1.43985644e-01
-4.01605606e-01 5.56523979e-01 2.19014093e-01 -1.55463383e-01
7.22202599e-01 1.55602582e-02 -2.91625261e-01 -2.15738535e-01
-1.06310570e+00 8.35230708e-01 2.81713873e-01 1.45043814e+00
7.53963649e-01 1.65009305e-01 2.32658554e-02 8.13887060e-01
3.39127511e-01 7.82148361e-01 -6.73390776e-02 -1.60678542e+00
6.86824262e-01 2.70334661e-01 3.68269563e-01 -9.49128628e-01
-1.92222163e-01 -3.10887277e-01 -1.02696574e+00 3.37774038e-01
5.42006135e-01 9.94952992e-02 -7.53876150e-01 1.53124273e+00
6.66505635e-01 6.00742996e-01 -2.59343445e-01 1.32093847e+00
6.30706966e-01 8.59091580e-01 4.62599099e-02 2.46604353e-01
1.05292916e+00 -1.23401165e+00 -5.10157287e-01 -2.89382249e-01
3.23755175e-01 -5.39679825e-01 1.15251541e+00 2.72493362e-01
-1.29558480e+00 -8.42373312e-01 -8.17726195e-01 -7.60018528e-01
1.27019912e-01 1.06264483e-02 7.16843486e-01 -8.03597420e-02
-1.00558889e+00 5.47953129e-01 -1.20878434e+00 -2.11004928e-01
5.98464489e-01 2.43903756e-01 -3.19437146e-01 -4.74699527e-01
-9.37239826e-01 6.42563045e-01 1.03432514e-01 4.05425727e-01
-6.20784223e-01 -1.01711559e+00 -9.30175185e-01 -3.36795002e-01
1.45299733e-01 -1.17471147e+00 1.29993069e+00 -4.95126039e-01
-1.05967975e+00 7.06426978e-01 -4.93174225e-01 -5.21446645e-01
9.61694896e-01 -3.56008172e-01 -1.72267914e-01 -8.11495930e-02
3.61326814e-01 1.24590921e+00 5.34221232e-01 -1.33003664e+00
-7.30511487e-01 -5.17152667e-01 -1.66096985e-01 3.21454257e-01
1.35802194e-01 -2.91333258e-01 -9.26428020e-01 -7.71246135e-01
9.78535041e-02 -1.32139444e+00 -4.51876581e-01 6.32644475e-01
-1.46451607e-01 -2.67454743e-01 1.01688719e+00 -6.25082672e-01
7.52214611e-01 -2.18095732e+00 3.21023256e-01 -3.21524918e-01
1.00279599e-01 7.48849437e-02 -6.30393103e-02 7.30209351e-02
1.90110102e-01 -5.19853421e-02 -5.78981698e-01 -7.23082304e-01
-3.76785249e-02 3.86327893e-01 -3.00544500e-01 5.66083491e-01
2.50945777e-01 1.07523489e+00 -1.01383471e+00 -5.31917512e-01
5.53670406e-01 8.15080225e-01 -7.50613153e-01 -6.53925538e-02
-2.93737382e-01 1.13467598e+00 -3.97472858e-01 8.40484023e-01
1.00511432e+00 -3.39360237e-01 -5.65154552e-01 -2.12469324e-01
-2.66068637e-01 -1.22504979e-01 -8.81581545e-01 2.37235451e+00
-4.25194144e-01 8.35171402e-01 1.41639318e-02 -4.58301663e-01
5.60465753e-01 3.16732638e-02 7.12327719e-01 -4.88415718e-01
-3.32495064e-01 2.40892977e-01 -3.33584487e-01 -5.15133500e-01
7.70413101e-01 6.99686818e-03 1.98338762e-01 -3.55729073e-01
-1.67711690e-01 -4.17977482e-01 3.49791236e-02 9.70137343e-02
6.19169712e-01 7.08408654e-01 -1.32843122e-01 -3.30737978e-02
3.60879928e-01 3.83225292e-01 7.33659863e-01 2.73615986e-01
-2.78883845e-01 1.02788556e+00 -2.17312336e-01 -5.30311525e-01
-1.36864066e+00 -1.15367043e+00 -3.46299857e-01 7.05595613e-01
5.42257488e-01 -1.39519900e-01 -4.37427461e-01 -1.28317520e-01
5.12398899e-01 6.31607056e-01 -2.29318172e-01 2.99220681e-01
-9.35876369e-01 -3.18856627e-01 4.45283443e-01 7.38478541e-01
7.30433464e-01 -9.03699756e-01 -3.71097624e-01 3.11940312e-01
-6.32531703e-01 -1.49312270e+00 -7.52063751e-01 -4.48330343e-01
-1.28230929e+00 -6.29008651e-01 -9.88493443e-01 -7.30184317e-01
4.37046796e-01 5.25257766e-01 1.21067297e+00 -1.30100420e-03
-1.11173652e-01 3.21434915e-01 -5.66092022e-02 -4.41310763e-01
-3.44490141e-01 -1.73466448e-02 8.35664198e-03 -3.06564987e-01
2.20388904e-01 -7.25572109e-01 -9.33871269e-01 5.44393420e-01
-6.93487465e-01 3.83787662e-01 2.89019167e-01 5.81183374e-01
9.99819994e-01 -3.43374252e-01 2.43817419e-01 -3.03627819e-01
8.26491788e-02 -5.07528663e-01 -7.64046431e-01 -2.79641807e-01
-2.66700685e-01 -2.01379448e-01 3.42468649e-01 -3.44921231e-01
-9.48508084e-01 3.61963928e-01 -4.91296887e-01 -8.98695886e-01
-2.34220728e-01 2.97319680e-01 -2.26529557e-02 -1.28943101e-01
5.84776402e-01 1.59403294e-01 1.53924373e-03 -4.05262858e-01
4.68738168e-01 4.37895089e-01 1.06815028e+00 -6.19451940e-01
8.87622535e-01 8.40700686e-01 3.65769453e-02 -9.44347322e-01
-5.90331078e-01 -5.06293952e-01 -7.66083837e-01 -5.10297716e-01
1.09616983e+00 -1.49580181e+00 -5.96747458e-01 5.88903069e-01
-1.40423596e+00 -6.92116141e-01 -2.63716638e-01 4.95181084e-01
-7.75095642e-01 6.29880905e-01 -6.78326786e-01 -6.37432694e-01
-2.48551801e-01 -1.37704170e+00 1.49522090e+00 9.11950916e-02
-2.15615034e-02 -7.88796008e-01 -1.54382676e-01 6.33504272e-01
1.10778674e-01 4.15182471e-01 4.56174433e-01 4.15854424e-01
-1.10379422e+00 -8.42028037e-02 -1.91361263e-01 2.15944275e-01
-8.04445371e-02 2.52919085e-02 -8.21019888e-01 -2.37089843e-01
-2.23064542e-01 -2.33044222e-01 7.84975171e-01 5.65339327e-01
1.35743332e+00 3.93865034e-02 -2.62118727e-01 1.12321651e+00
1.17304909e+00 1.44902762e-04 7.71872401e-01 2.65913188e-01
1.19868505e+00 4.07244295e-01 7.79333413e-01 4.00457144e-01
9.25306857e-01 1.08488822e+00 6.86912835e-01 -2.14704469e-01
-3.67321461e-01 -4.43143249e-01 1.12485126e-01 7.64519811e-01
-5.19476235e-01 -5.88029847e-02 -1.10448861e+00 7.19458878e-01
-1.89510167e+00 -1.02925551e+00 -5.90932548e-01 2.04116678e+00
6.81741714e-01 -1.68116122e-01 2.19434932e-01 -1.37652993e-01
6.91927314e-01 1.20846242e-01 -9.55426812e-01 1.24030426e-01
-6.13466799e-02 -2.58111537e-01 6.32039547e-01 4.06424433e-01
-1.05833542e+00 1.01818645e+00 5.24808455e+00 9.01852667e-01
-1.02179670e+00 1.87276617e-01 6.77718759e-01 3.85010540e-02
-2.68815666e-01 -6.34756982e-02 -8.63721073e-01 7.23105669e-01
5.84486306e-01 5.23672700e-02 5.84036410e-01 8.98259521e-01
5.01813531e-01 -4.82000634e-02 -1.05939949e+00 1.23194301e+00
-3.95047992e-01 -1.50115561e+00 -9.74283814e-02 2.06762403e-01
9.07313764e-01 6.93582356e-01 1.64356068e-01 2.17471525e-01
1.92869380e-01 -7.40131855e-01 1.00903153e+00 4.19182986e-01
1.12496603e+00 -5.90628088e-01 2.62908667e-01 6.77649438e-01
-1.33392203e+00 1.99761972e-01 -5.62801480e-01 -1.19612522e-01
5.68313658e-01 5.69558144e-01 -3.44873875e-01 5.53562224e-01
9.30481613e-01 1.09453118e+00 -3.51982206e-01 1.25159776e+00
-9.39297229e-02 3.73519033e-01 -4.33521301e-01 4.52942133e-01
2.04235598e-01 -3.93489629e-01 5.27379215e-01 1.06448007e+00
5.17924368e-01 2.02050343e-01 1.71164438e-01 1.03805804e+00
-6.24493100e-02 -2.79894799e-01 -5.88365495e-01 3.94256711e-01
4.27531570e-01 9.70503092e-01 -3.50230038e-01 -3.05153638e-01
-4.88328069e-01 1.05308735e+00 1.67614877e-01 5.23134410e-01
-1.14480078e+00 3.48955691e-02 9.36753035e-01 3.73240888e-01
1.56361744e-01 -7.93084443e-01 -4.91793752e-01 -1.28272879e+00
2.62975812e-01 -5.10793447e-01 -1.75807402e-01 -1.02905035e+00
-1.30676961e+00 4.37921494e-01 1.83819801e-01 -1.67224956e+00
-2.56351292e-01 -2.80736119e-01 -3.29611093e-01 1.00967932e+00
-1.60515368e+00 -1.29103601e+00 -8.41163993e-01 6.40100121e-01
8.96312118e-01 3.19548279e-01 3.25236142e-01 6.30854726e-01
-1.95562884e-01 2.23746240e-01 1.34110600e-01 6.26253933e-02
7.12388992e-01 -9.42581773e-01 1.11859751e+00 9.07633603e-01
-2.25346804e-01 3.11095625e-01 6.46495998e-01 -7.31714308e-01
-1.58758211e+00 -1.57643378e+00 5.96455097e-01 -6.62596583e-01
4.50290233e-01 -2.88855821e-01 -1.05505013e+00 7.28151143e-01
-1.05182216e-01 5.49507976e-01 1.27006337e-01 -3.83847594e-01
-6.15048930e-02 -9.14031863e-02 -1.07555044e+00 7.85327077e-01
1.50929308e+00 -3.16496342e-01 -1.32299870e-01 4.05607343e-01
9.76812482e-01 -1.09934413e+00 -1.04065681e+00 6.05201781e-01
5.06251991e-01 -9.08063173e-01 1.36005330e+00 -1.17070109e-01
8.04686010e-01 -6.18866026e-01 -8.10933933e-02 -1.20039141e+00
-4.26261008e-01 -4.46191430e-01 -2.18131423e-01 8.79796565e-01
9.03240144e-02 -4.46435243e-01 9.51856494e-01 7.98477530e-01
-5.77993691e-01 -7.78245986e-01 -1.08479965e+00 -9.16538000e-01
3.04665506e-01 -6.60313547e-01 6.32971406e-01 8.81013632e-01
-4.70010161e-01 -4.80066016e-02 -5.67843199e-01 2.04236537e-01
9.04910624e-01 1.24616593e-01 1.23321784e+00 -1.07047975e+00
-2.40303800e-01 -1.94809556e-01 -4.72919285e-01 -1.84352410e+00
1.62785783e-01 -8.86798918e-01 2.19256714e-01 -1.66740620e+00
-2.88513243e-01 -7.75087714e-01 2.75815129e-01 2.22817034e-01
-2.90211856e-01 5.54392457e-01 4.14018244e-01 6.89733267e-01
-2.61501729e-01 7.94411540e-01 1.73226070e+00 -2.44736820e-01
-2.34517008e-01 1.13465182e-01 -2.05811679e-01 7.35673249e-01
6.62160754e-01 -2.19860658e-01 -3.26364815e-01 -1.03294814e+00
-6.98680431e-02 4.79460090e-01 8.25757027e-01 -1.36092544e+00
2.91792303e-01 -3.04536968e-01 4.55663800e-01 -1.05676985e+00
7.43432999e-01 -9.11843956e-01 4.43962365e-01 2.75912076e-01
-3.91657017e-02 3.20999354e-01 3.58593315e-01 7.58932412e-01
-1.53865606e-01 2.99344867e-01 7.25461841e-01 -6.13136888e-02
-8.88638794e-01 9.73598540e-01 1.32599086e-01 2.98704803e-02
8.62934351e-01 -5.43533385e-01 8.95562582e-03 -1.88322887e-01
-5.96430659e-01 5.37032068e-01 8.76016736e-01 4.19322073e-01
7.18069255e-01 -1.51349866e+00 -8.31024051e-01 1.36676073e-01
2.50370264e-01 1.01492429e+00 6.09280169e-01 1.01584625e+00
-9.14203525e-01 1.66352704e-01 -1.11289583e-01 -1.22333729e+00
-1.00274062e+00 5.09856403e-01 2.01952726e-01 2.37251475e-01
-1.03503942e+00 7.56313980e-01 3.06812674e-01 -3.86836648e-01
6.98587671e-02 -7.31260896e-01 3.55154693e-01 -5.66616774e-01
1.44364923e-01 5.40821195e-01 -9.47151855e-02 -9.94254649e-01
-1.56294838e-01 7.82734573e-01 2.76534766e-01 -9.49672759e-02
1.18222344e+00 -2.33992174e-01 1.45105287e-01 3.75115544e-01
1.11423111e+00 -1.58983469e-01 -1.92035067e+00 -1.31116033e-01
-3.32588404e-01 -7.38051236e-01 -7.07178786e-02 -3.85263294e-01
-1.17460179e+00 1.06324542e+00 4.69772369e-01 -3.62378120e-01
9.99301612e-01 -8.57812613e-02 1.21052158e+00 1.92824274e-01
7.49084532e-01 -6.87658548e-01 -3.42000157e-01 4.78015333e-01
1.03744888e+00 -1.37625802e+00 2.09560581e-02 -6.59940124e-01
-6.65991426e-01 9.56461012e-01 6.73671305e-01 -3.11500043e-01
4.34105217e-01 1.32149681e-01 -1.50860548e-01 1.18498027e-01
-4.49528277e-01 4.71941866e-02 2.11893216e-01 8.00282240e-01
2.50649482e-01 1.58053953e-02 -1.34410681e-02 -2.72315629e-02
-3.21589440e-01 5.89825332e-01 3.57141018e-01 6.84164584e-01
-1.27873123e-01 -8.39342773e-01 -4.44658846e-01 1.20373599e-01
-1.02011114e-01 -7.40409503e-03 1.98752999e-01 8.05440128e-01
2.95501649e-01 7.17654169e-01 1.39295191e-01 -2.63747454e-01
4.69500244e-01 -4.67566043e-01 4.33913380e-01 -2.68096983e-01
-1.62155032e-01 9.04569961e-03 -1.84672371e-01 -7.42163181e-01
-5.76630950e-01 -8.29497039e-01 -1.43402100e+00 -5.83504558e-01
-1.40397638e-01 -1.97474316e-01 7.64268100e-01 6.22774899e-01
5.70800304e-01 3.46065938e-01 2.15369701e-01 -1.60994625e+00
-2.67017603e-01 -7.73889601e-01 -2.68997639e-01 5.14212430e-01
4.78781968e-01 -7.79662430e-01 -9.65915918e-02 4.27223265e-01]
|
[8.456883430480957, -2.115225076675415]
|
22c06c68-c00b-438e-9baf-42ce196c7d77
|
neural-gas-network-image-features-and
|
2301.12176
| null |
https://arxiv.org/abs/2301.12176v1
|
https://arxiv.org/pdf/2301.12176v1.pdf
|
Neural Gas Network Image Features and Segmentation for Brain Tumor Detection Using Magnetic Resonance Imaging Data
|
Accurate detection of brain tumors could save lots of lives and increasing the accuracy of this binary classification even as much as a few percent has high importance. Neural Gas Networks (NGN) is a fast, unsupervised algorithm that could be used in data clustering, image pattern recognition, and image segmentation. In this research, we used the metaheuristic Firefly Algorithm (FA) for image contrast enhancement as pre-processing and NGN weights for feature extraction and segmentation of Magnetic Resonance Imaging (MRI) data on two brain tumor datasets from the Kaggle platform. Also, tumor classification is conducted by Support Vector Machine (SVM) classification algorithms and compared with a deep learning technique plus other features in train and test phases. Additionally, NGN tumor segmentation is evaluated by famous performance metrics such as Accuracy, F-measure, Jaccard, and more versus ground truth data and compared with traditional segmentation techniques. The proposed method is fast and precise in both tasks of tumor classification and segmentation compared with other methods. A classification accuracy of 95.14 % and segmentation accuracy of 0.977 is achieved by the proposed method.
|
['S. Muhammad Hossein Mousavi']
|
2023-01-28
| null | null | null | null |
['tumor-segmentation']
|
['computer-vision']
|
[ 2.44302735e-01 -1.84456795e-01 -5.89101240e-02 -1.37280554e-01
-4.35599744e-01 -1.10908821e-01 2.28109956e-01 4.90709603e-01
-9.70779598e-01 8.55723977e-01 -3.87777746e-01 -1.38886020e-01
-5.99736989e-01 -9.46484149e-01 7.21285567e-02 -1.28689384e+00
-1.61639795e-01 8.30207467e-01 1.79695576e-01 4.37608138e-02
6.12128198e-01 7.60758519e-01 -1.51608968e+00 -7.92532936e-02
1.25949252e+00 1.27711487e+00 4.17040229e-01 5.67335069e-01
-3.25308114e-01 8.17189097e-01 -6.59575224e-01 1.90602511e-01
9.18941945e-02 -1.77985221e-01 -9.17279243e-01 2.86257528e-02
-4.84957844e-01 4.13863748e-01 2.25034818e-01 1.34398067e+00
3.99100989e-01 2.21445799e-01 9.25314963e-01 -1.34055543e+00
-8.63972157e-02 4.77046698e-01 -8.39893639e-01 3.43866497e-01
-3.89399886e-01 -7.68396854e-02 2.68950611e-01 -4.11529809e-01
4.93129224e-01 4.97512847e-01 6.53347433e-01 4.77062017e-01
-6.02282822e-01 -7.55387545e-01 -7.02052653e-01 3.81158948e-01
-1.33747673e+00 1.40810028e-01 6.52387738e-01 -5.24135172e-01
7.61142910e-01 6.76023185e-01 1.13564968e+00 3.28592896e-01
7.09728301e-01 5.98908186e-01 1.26993704e+00 -5.59630632e-01
5.83996952e-01 -1.48251742e-01 5.07447660e-01 9.71132457e-01
5.21657050e-01 1.88822895e-01 -6.16284795e-02 4.40330012e-03
4.75711435e-01 1.79212213e-01 -4.28139530e-02 1.53842911e-01
-1.19971597e+00 1.08914149e+00 8.31737638e-01 8.00161958e-01
-7.05705523e-01 1.33709759e-01 5.42178094e-01 -2.37651005e-01
4.30442274e-01 6.95634902e-01 -1.50474370e-01 -1.39807602e-02
-1.22591460e+00 -1.19403914e-01 4.06386793e-01 3.44321072e-01
2.00627461e-01 2.39227228e-02 -2.40623832e-01 8.41925979e-01
1.57248870e-01 2.27363735e-01 1.27071095e+00 -6.53455138e-01
-3.24441284e-01 8.52140248e-01 -3.35788935e-01 -8.99673939e-01
-1.03484225e+00 -6.06363952e-01 -1.32036436e+00 3.33230436e-01
2.19982326e-01 -4.05056067e-02 -1.22803807e+00 1.05414724e+00
3.04540217e-01 8.25128481e-02 2.92784590e-02 8.54869306e-01
9.49362814e-01 3.52150619e-01 1.89625651e-01 -5.24271369e-01
1.40101111e+00 -1.35270429e+00 -7.71526873e-01 3.71156745e-02
7.78394699e-01 -7.21606016e-01 4.59250689e-01 6.46373570e-01
-8.03146362e-01 -1.10534169e-01 -7.38068700e-01 4.75057662e-01
-4.95930344e-01 5.97662143e-02 1.03259134e+00 9.79393363e-01
-1.02959836e+00 3.92268300e-01 -1.22386849e+00 -5.28453529e-01
8.83959949e-01 6.64743900e-01 -1.74692377e-01 4.18305583e-02
-7.63803959e-01 7.14119494e-01 7.17269719e-01 -2.03806609e-02
-5.32593787e-01 -5.84019840e-01 -4.75081533e-01 -2.37858504e-01
-1.63556516e-01 -4.38019961e-01 7.42224932e-01 -1.14010513e+00
-1.24113226e+00 9.73848581e-01 2.21259952e-01 -7.81039834e-01
4.49326932e-01 5.69461286e-01 -1.90192416e-01 3.91229987e-01
1.94746745e-03 8.59180391e-01 3.56796533e-01 -8.44319522e-01
-7.09351838e-01 -6.74503565e-01 -6.08126938e-01 2.34490380e-01
-3.43126267e-01 5.53503968e-02 2.06471041e-01 -7.11486220e-01
5.19449055e-01 -7.53493547e-01 -5.05067050e-01 -2.39647254e-01
-4.11826998e-01 -2.18349904e-01 7.94654310e-01 -9.04808402e-01
9.76833403e-01 -1.72101712e+00 -7.28382319e-02 8.79864335e-01
3.68652135e-01 1.76167339e-01 3.24878395e-01 -3.18753004e-01
-1.25756577e-01 9.05184522e-02 -6.17394686e-01 2.95335114e-01
-4.07373697e-01 2.53439248e-01 8.93561065e-01 5.96535325e-01
-1.38373420e-01 9.25013840e-01 -7.47717440e-01 -8.04287255e-01
2.24361107e-01 2.69506425e-01 -2.74507701e-01 1.14084873e-02
1.55419171e-01 5.21293342e-01 -2.48562217e-01 1.07815111e+00
6.15601182e-01 -6.42267466e-02 -3.23290020e-01 -1.64047718e-01
1.31047750e-02 -7.73368955e-01 -7.01703429e-01 1.47512674e+00
-8.52962360e-02 6.11497223e-01 -1.57153323e-01 -1.45948470e+00
1.03724897e+00 1.32884622e-01 1.26479304e+00 -7.68988788e-01
7.04531491e-01 2.00511292e-01 3.80611688e-01 -9.43908215e-01
1.70218959e-01 9.69460458e-02 4.58777040e-01 4.36559469e-01
-1.41344920e-01 -7.69848600e-02 2.97513336e-01 -2.40636423e-01
1.08837414e+00 -5.94791830e-01 2.69198626e-01 -5.46238542e-01
5.64586401e-01 7.56349444e-01 4.45132136e-01 5.66529989e-01
-4.52326328e-01 2.86029875e-01 -8.27842113e-03 -4.58863735e-01
-7.43981242e-01 -5.56114376e-01 -4.91704226e-01 6.29758596e-01
6.73027188e-02 1.22005053e-01 -1.26371145e+00 -4.73718941e-01
-2.55756080e-01 8.03212643e-01 -8.79786611e-01 -1.57005548e-01
-3.02606225e-01 -1.58374739e+00 4.08014923e-01 2.98518717e-01
9.73774374e-01 -1.21538591e+00 -9.56880450e-01 1.21201627e-01
-1.73365355e-01 -7.01716244e-01 1.44826576e-01 2.74291903e-01
-1.09373343e+00 -1.32788420e+00 -7.71347344e-01 -1.01189685e+00
1.08039629e+00 -4.02181856e-02 8.14543605e-01 5.05924165e-01
-1.03941143e+00 -7.16299936e-02 -5.37830234e-01 -4.78454769e-01
-4.40403342e-01 5.84165473e-03 -2.85131454e-01 -1.47068262e-01
4.08352494e-01 -2.50019222e-01 -6.95825517e-01 2.11600989e-01
-9.42017376e-01 8.60003084e-02 6.90929472e-01 1.01262498e+00
8.53221357e-01 7.43744731e-01 1.14237972e-01 -8.77113163e-01
6.89041674e-01 -3.81692559e-01 -5.46126902e-01 2.48262286e-01
-9.56892848e-01 -3.31539243e-01 2.45092645e-01 -1.70472980e-01
-5.38846493e-01 -2.65012514e-02 -1.59889892e-01 -1.72006473e-01
-1.82204440e-01 7.11982787e-01 4.63229060e-01 -5.81889689e-01
8.61505926e-01 2.28876293e-01 4.30997372e-01 8.39833692e-02
-4.21649098e-01 6.73414767e-01 2.87002951e-01 -1.23861082e-01
1.81535020e-01 4.37014997e-01 2.82911956e-01 -6.38498247e-01
-2.98512697e-01 -5.02408743e-01 -4.19549555e-01 -4.76576328e-01
1.11324823e+00 -8.85655060e-02 -7.39555299e-01 8.81356478e-01
-4.98088330e-01 -1.07044972e-01 -1.39758691e-01 7.43703485e-01
-3.21885705e-01 4.52151932e-02 -3.94244939e-01 -8.76114726e-01
-9.99446154e-01 -1.53284299e+00 4.12597388e-01 5.93214452e-01
1.05855085e-01 -9.47443843e-01 -3.32380086e-01 6.22155845e-01
7.75952339e-01 6.20138109e-01 9.11916077e-01 -9.53010082e-01
8.32323730e-02 -4.46354151e-01 -1.90577775e-01 1.60788015e-01
2.59230047e-01 3.53475481e-01 -6.75678015e-01 -2.74819285e-01
2.85161823e-01 -3.11487429e-02 7.29452193e-01 1.00865161e+00
1.64982176e+00 -2.19526723e-01 -7.81489193e-01 7.31555402e-01
1.68527329e+00 9.44894850e-01 6.55921996e-01 8.23708355e-01
5.30009091e-01 4.34742063e-01 5.77911437e-01 3.09621781e-01
1.49704237e-02 1.07559808e-01 5.91812074e-01 -1.82854980e-01
1.00948557e-01 7.28018939e-01 -4.61755365e-01 8.62770677e-01
-3.82064372e-01 -5.36571145e-02 -1.52771246e+00 3.46777141e-01
-1.57033980e+00 -9.41022575e-01 -2.39069954e-01 1.87824059e+00
5.52043080e-01 8.81778300e-02 -1.89734367e-03 7.57009745e-01
9.00943637e-01 -5.97397029e-01 -4.94881392e-01 -2.53060281e-01
-4.28714156e-02 5.62305629e-01 7.37133980e-01 1.08747035e-01
-1.20725751e+00 6.14858806e-01 5.24244452e+00 1.21933663e+00
-1.41122520e+00 2.77358651e-01 1.22675729e+00 1.10314637e-01
2.03918755e-01 -5.44950247e-01 -9.87614468e-02 6.23217940e-01
6.63572550e-01 3.14666480e-02 4.42410707e-01 5.80044448e-01
2.06182465e-01 -6.85443938e-01 -2.89658219e-01 1.27423561e+00
1.76958125e-02 -1.36801207e+00 -2.58166850e-01 -3.68798822e-02
7.81361639e-01 7.79081434e-02 -4.06011492e-02 -1.39557227e-01
8.20941031e-02 -1.31013644e+00 2.80679196e-01 6.59858346e-01
4.69846427e-01 -1.18135715e+00 1.06760418e+00 3.72107327e-01
-6.36256576e-01 -2.87552178e-01 -1.24376751e-01 4.55795109e-01
-4.14716691e-01 6.96985602e-01 -9.42869723e-01 4.70977694e-01
8.40458632e-01 2.78235316e-01 -5.56674838e-01 1.67358863e+00
3.61706257e-01 7.12155700e-01 -4.15843576e-01 -4.83950585e-01
4.67568964e-01 -3.07388663e-01 3.59680235e-01 8.35309029e-01
3.84170502e-01 3.00404459e-01 2.69925185e-02 5.11021972e-01
2.45917335e-01 4.83440995e-01 5.33864833e-02 -7.34362900e-02
3.85465980e-01 1.53395319e+00 -1.67751145e+00 -2.68265516e-01
2.49501929e-01 5.37635446e-01 -1.53041512e-01 -1.42819181e-01
-1.02646923e+00 -6.09555125e-01 1.45854242e-02 -6.17400259e-02
-3.40149671e-01 1.16606310e-01 -7.14954019e-01 -5.30770123e-01
-5.09451330e-01 -4.93385166e-01 5.08388400e-01 -7.04257011e-01
-9.81855452e-01 1.00396991e+00 -2.20451429e-01 -1.06046426e+00
1.58716291e-01 -6.32897437e-01 -6.60146236e-01 4.17073995e-01
-1.37398601e+00 -8.53488564e-01 -8.20975900e-01 7.40340352e-01
4.80536371e-01 -4.69294161e-01 7.06755579e-01 -6.55365661e-02
-8.46416295e-01 4.82200801e-01 3.95934165e-01 3.62353534e-01
1.38849333e-01 -1.09200227e+00 -5.93443811e-01 6.49384081e-01
-2.57502794e-01 -1.03957616e-01 5.04668355e-01 -4.69671994e-01
-8.90136838e-01 -1.07489836e+00 3.61785889e-01 5.17379820e-01
2.63007969e-01 3.05019915e-01 -4.00656909e-01 6.99117826e-03
1.97547778e-01 8.55109915e-02 8.86413515e-01 -6.20227873e-01
6.70389235e-01 -1.51123226e-01 -1.93446732e+00 1.51649386e-01
3.93007666e-01 4.50493991e-01 -1.45380601e-01 9.40876961e-01
1.99179187e-01 -6.13587379e-01 -1.04417348e+00 5.72954535e-01
3.04044843e-01 -9.54605043e-01 8.33454013e-01 -3.50734115e-01
2.60845095e-01 -1.64165765e-01 5.19326441e-02 -1.16365063e+00
-2.57900923e-01 5.39152548e-02 5.40290833e-01 6.60279036e-01
3.87288213e-01 -7.34639466e-01 1.22490048e+00 7.06821799e-01
-2.57158637e-01 -1.18303788e+00 -9.26850379e-01 -5.75795114e-01
-6.80825263e-02 -3.58574122e-01 6.83857262e-01 1.23080063e+00
-2.11528793e-01 -4.96131212e-01 4.86076146e-01 -1.40035972e-01
9.75919306e-01 -1.20873213e-01 -3.53582948e-02 -1.39994276e+00
2.33062744e-01 -9.59236503e-01 -8.85999382e-01 4.03062105e-01
-4.51276712e-02 -1.17701435e+00 -2.03624308e-01 -1.74125743e+00
1.73546270e-01 -9.14100826e-01 -7.16878057e-01 5.28562486e-01
1.22129783e-01 3.92091811e-01 -1.99894711e-01 4.05031711e-01
-7.84124658e-02 1.38710141e-01 1.37584567e+00 -5.69548786e-01
-7.34181777e-02 1.61641091e-01 -3.52402002e-01 6.85657561e-01
1.09481978e+00 -7.28188157e-01 -2.60365993e-01 -2.89330363e-01
-2.55533844e-01 1.04339913e-01 1.92177683e-01 -1.46265566e+00
6.82698011e-01 -2.91090757e-01 7.35463500e-01 -5.48438966e-01
-2.84181572e-02 -9.42169189e-01 3.52226764e-01 9.83861923e-01
-3.09318513e-01 1.87316522e-01 -4.64748479e-02 7.96280950e-02
-3.03855866e-01 -6.67674780e-01 1.05788267e+00 -3.01252753e-01
-7.02181578e-01 3.87381971e-01 -5.12011111e-01 -2.86039203e-01
1.62157285e+00 -8.31750751e-01 -1.90791562e-01 3.37325603e-01
-8.66896868e-01 2.66350452e-02 1.09657288e-01 -2.48106405e-01
7.53314435e-01 -1.27001357e+00 -6.23180628e-01 1.29746825e-01
-1.30702788e-02 1.28992274e-01 1.32345065e-01 1.46485567e+00
-1.32059455e+00 1.19756803e-01 -7.35534370e-01 -7.84046769e-01
-1.31343472e+00 2.30581805e-01 6.56318784e-01 -3.24231476e-01
-3.00977737e-01 1.31372416e+00 -5.40477335e-01 -2.09520876e-01
1.80006459e-01 -3.02889585e-01 -8.09734166e-01 -3.53568845e-04
2.91521937e-01 6.78924143e-01 3.84158224e-01 -6.86966419e-01
-5.37645042e-01 4.66031373e-01 -2.34392312e-04 1.36929199e-01
1.48819876e+00 4.87613201e-01 -7.73638725e-01 -1.59743398e-01
1.04644477e+00 -5.98402858e-01 -3.55636299e-01 2.17047378e-01
7.39201829e-02 -1.55234814e-01 6.27291024e-01 -1.22214127e+00
-1.80711710e+00 6.13408566e-01 1.25605083e+00 2.83399880e-01
1.62263525e+00 -3.94838959e-01 8.49201620e-01 3.13562334e-01
1.63954273e-01 -1.29529548e+00 -2.48599142e-01 1.17975891e-01
4.37727451e-01 -1.26903164e+00 -1.45164043e-01 -1.91606954e-01
-5.08595049e-01 1.43913805e+00 6.01249576e-01 -3.27230603e-01
9.02224243e-01 5.29408216e-01 6.05454221e-02 -3.80075544e-01
-2.76224524e-01 -5.19803017e-02 1.69172473e-02 6.28088713e-01
2.06338897e-01 2.96442866e-01 -6.65989637e-01 6.37368500e-01
-3.75095844e-01 2.59349018e-01 2.06650808e-01 1.09374475e+00
-7.90061414e-01 -5.70299327e-01 -6.57962143e-01 1.05962634e+00
-5.90931773e-01 -1.02277286e-01 -1.92064550e-02 7.00405777e-01
2.50741243e-01 8.92481089e-01 2.01408461e-01 -5.03623962e-01
-2.80945480e-01 -3.47352654e-01 4.61304337e-01 2.40005795e-02
-8.45066667e-01 -4.35946397e-02 -3.44323993e-01 -3.39325249e-01
-7.21957147e-01 -3.96100461e-01 -1.62834799e+00 -1.28522038e-01
-7.27974951e-01 5.84231138e-01 1.31573308e+00 1.03348923e+00
-2.44262487e-01 8.41529727e-01 7.18162835e-01 -6.07374728e-01
5.26175871e-02 -7.64796376e-01 -6.70569539e-01 2.08020315e-01
-3.10893118e-01 -7.80406296e-01 -1.44749209e-01 -1.77964076e-01]
|
[14.753894805908203, -2.553692579269409]
|
f2d6b59b-d4cd-4a62-9368-ef656c730d8d
|
applying-information-extraction-to-storybook
| null | null |
https://aclanthology.org/2022.rocling-1.36
|
https://aclanthology.org/2022.rocling-1.36.pdf
|
Applying Information Extraction to Storybook Question and Answer Generation
|
For educators, how to generate high quality question-answer pairs from story text is a time-consuming and labor-intensive task. The purpose is not to make students unable to answer, but to ensure that students understand the story text through the generated question-answer pairs. In this paper, we improve the FairyTaleQA question generation method by incorporating question type and its definition to the input for fine-tuning the BART (Lewis et al., 2020) model. Furthermore, we make use of the entity and relation extraction from (Zhong and Chen, 2021) as an element of template-based question generation.
|
['Chia-Hui Chang', 'Kai-Yen Kao']
| null | null | null | null |
rocling-2022-11
|
['question-generation', 'answer-generation']
|
['natural-language-processing', 'natural-language-processing']
|
[ 6.24635555e-02 3.87739658e-01 3.32629621e-01 -4.32909161e-01
-9.99087393e-01 -9.69426036e-01 4.03028846e-01 5.25384367e-01
-3.22577804e-01 1.06371260e+00 2.91001797e-01 -8.30632985e-01
-2.94805974e-01 -1.25068700e+00 -4.46345866e-01 2.95330554e-01
8.12699676e-01 3.81985039e-01 7.96195924e-01 -7.09607244e-01
4.99925315e-01 7.74960369e-02 -1.62671828e+00 4.98946488e-01
1.47570181e+00 4.60590750e-01 6.48373961e-02 1.03042006e+00
-8.96698236e-01 1.38574219e+00 -8.96434367e-01 -8.68781149e-01
-3.95967811e-01 -1.13619590e+00 -1.41771591e+00 -2.23056018e-01
5.18637598e-01 -2.81652361e-01 7.17495531e-02 6.31992221e-01
3.56642425e-01 3.37357461e-01 5.92443585e-01 -9.69380558e-01
-7.57076025e-01 8.77251923e-01 8.31753388e-02 1.94410160e-01
9.27155018e-01 -1.31007910e-01 9.51698363e-01 -6.38828158e-01
6.31889999e-01 6.91562176e-01 5.12135744e-01 7.84948289e-01
-8.67710173e-01 -4.97062802e-01 -3.06933731e-01 6.47593021e-01
-9.74026024e-01 -3.37258577e-01 6.13750815e-01 -5.15654802e-01
6.83720410e-01 6.48965299e-01 9.64744091e-01 6.93103313e-01
-8.38505924e-02 6.51670218e-01 1.00845397e+00 -9.82122540e-01
-3.14217806e-02 4.02285129e-01 3.65314156e-01 6.42965376e-01
2.40673065e-01 -3.75004560e-01 -4.52457339e-01 4.34025889e-03
5.40637016e-01 -6.34558380e-01 -2.56607771e-01 2.49055512e-02
-6.42186582e-01 1.01841450e+00 -1.36474699e-01 4.32000607e-01
3.93939279e-02 -1.84596568e-01 6.46304041e-02 5.89347363e-01
-1.08504675e-01 1.17132068e+00 -4.78490204e-01 -7.17881680e-01
-8.14380527e-01 7.51804113e-01 1.39166284e+00 1.03663504e+00
3.34180862e-01 -3.34192961e-01 -6.16489589e-01 7.03564703e-01
4.38654333e-01 8.14945847e-02 5.58615267e-01 -9.94098961e-01
4.75009859e-01 1.05511475e+00 1.48829803e-01 -9.14846599e-01
1.12069033e-01 -4.29273337e-01 -1.36011407e-01 -2.13032454e-01
8.98898125e-01 -4.02869761e-01 -5.42706847e-01 1.38396859e+00
5.29155552e-01 -2.18234673e-01 2.35184893e-01 5.25882900e-01
1.71682489e+00 5.43548107e-01 1.96685374e-01 1.21038228e-01
1.65244365e+00 -1.04951894e+00 -9.79299128e-01 -7.89161325e-02
9.66203809e-01 -1.19433129e+00 1.17278492e+00 1.17861070e-01
-1.44863927e+00 -5.20509183e-01 -7.99736202e-01 -6.09335184e-01
-2.53630221e-01 4.68348339e-02 2.97365874e-01 1.01271760e+00
-7.40499318e-01 2.38425344e-01 7.56579712e-02 -2.53798962e-01
6.56438172e-02 -8.41353238e-02 -4.12078835e-02 -1.99170500e-01
-1.54959238e+00 9.40613508e-01 8.41869190e-02 -3.88941139e-01
-5.77757992e-02 -1.04342484e+00 -9.79905725e-01 3.29727829e-01
5.35532773e-01 -9.31638241e-01 1.77941322e+00 -4.58009183e-01
-1.85447168e+00 6.44938231e-01 1.42744845e-02 -4.84361053e-02
3.49302530e-01 -9.07501280e-02 -9.42539126e-02 1.54543683e-01
1.07397079e-01 5.26429057e-01 2.63745874e-01 -9.87144530e-01
-5.54241955e-01 1.59310713e-01 5.14551580e-01 4.79736090e-01
-1.14314139e-01 1.00569114e-01 -2.36188814e-01 -5.94627261e-01
1.54925272e-01 -4.36867714e-01 3.05821281e-02 -2.95730799e-01
-7.63586760e-02 -6.97881579e-01 5.02220392e-01 -1.09006655e+00
1.53464365e+00 -1.53915226e+00 -4.44995046e-01 2.18150407e-01
1.65162608e-01 4.75098610e-01 -1.97784379e-01 6.83488727e-01
-9.02845711e-03 3.39683056e-01 1.89598650e-02 2.01666608e-01
1.74294665e-01 6.03488423e-02 -2.21833706e-01 -4.42872703e-01
1.75360084e-01 1.03822756e+00 -1.04869354e+00 -9.08372700e-01
7.69426674e-02 1.42771140e-01 -6.82617009e-01 6.17529392e-01
-4.76988286e-01 2.45936021e-01 -4.60441113e-01 3.15859526e-01
2.10461877e-02 -2.71761473e-02 -1.93655118e-01 2.85418391e-01
-7.59243369e-02 8.89497638e-01 -1.19800162e+00 1.45713103e+00
-7.71114469e-01 6.60531938e-01 -3.38644296e-01 -4.56413031e-01
1.07357502e+00 5.97343445e-01 -2.38611568e-02 -7.82603920e-01
1.41213939e-01 2.90864468e-01 4.72801663e-02 -1.00940418e+00
8.97967041e-01 -2.50920773e-01 -2.49652162e-01 8.09230268e-01
1.09467112e-01 -9.88998413e-01 8.04463089e-01 4.44648683e-01
1.13679016e+00 3.75838578e-01 3.60018164e-01 2.83143874e-02
8.94145429e-01 2.73348957e-01 1.10577993e-01 8.23853195e-01
1.64105266e-01 6.18724763e-01 5.95663965e-01 1.46831512e-01
-7.66306341e-01 -7.60842621e-01 3.65144312e-02 8.77461016e-01
-5.72829723e-01 -5.53016245e-01 -1.05381382e+00 -7.65142024e-01
-4.64141577e-01 1.42764401e+00 -4.29583758e-01 -1.52392939e-01
-6.63115978e-01 -3.71198729e-02 6.03380322e-01 2.47311950e-01
4.49583173e-01 -1.03027678e+00 -7.33590186e-01 5.02010286e-01
-8.47954512e-01 -9.09945786e-01 -5.75923860e-01 -1.75487429e-01
-4.54883486e-01 -1.05975449e+00 -6.36530042e-01 -8.50595653e-01
5.51591992e-01 -5.84765226e-02 1.55645561e+00 5.23047268e-01
-6.04656115e-02 6.23998940e-01 -7.85635591e-01 -5.14666438e-01
-5.88859499e-01 2.74048656e-01 -9.75317001e-01 -6.08761668e-01
4.98273313e-01 -4.26248610e-01 -3.45708966e-01 1.46750629e-01
-1.16204655e+00 2.81543523e-01 1.61794901e-01 5.31002760e-01
1.76559374e-01 -1.97525546e-01 8.51298571e-01 -1.03563201e+00
1.27596295e+00 -3.34342450e-01 -5.81284344e-01 6.10357463e-01
-4.45443481e-01 1.42711625e-01 6.85864568e-01 -2.69114643e-01
-1.08256149e+00 -2.52104580e-01 -7.90421605e-01 4.65002447e-01
-1.08702272e-01 5.66518247e-01 -2.79104203e-01 -3.00518900e-01
7.12598801e-01 2.35200584e-01 -2.25412950e-01 -3.71420950e-01
4.73899215e-01 4.56678122e-01 2.72830248e-01 -7.39254355e-01
7.93539405e-01 -4.85736996e-01 -1.85992807e-01 -5.23226798e-01
-1.04676282e+00 -4.17840481e-01 -5.36649883e-01 -5.54230630e-01
7.11301506e-01 -4.90425050e-01 -7.54079878e-01 1.15340389e-01
-1.29964685e+00 -3.50283116e-01 -7.54284322e-01 2.12380752e-01
-3.05777013e-01 2.05070242e-01 -4.88238901e-01 -5.80828190e-01
-1.24201074e-01 -8.27149332e-01 3.29282820e-01 7.74923742e-01
-8.05126488e-01 -1.01951671e+00 2.58315295e-01 1.17599201e+00
4.03079212e-01 5.23179211e-02 1.35215914e+00 -9.05983925e-01
-4.38919961e-01 -1.71628475e-01 -6.56630620e-02 2.84042954e-01
-2.38957349e-02 1.52923986e-01 -5.13793826e-01 4.96141315e-01
2.51937062e-01 -4.51401532e-01 1.55942664e-01 -1.53131574e-01
9.11338747e-01 -7.05741525e-01 3.93576354e-01 9.57328007e-02
1.28226769e+00 1.90122858e-01 6.68243170e-01 2.50968933e-01
5.87100744e-01 1.17796075e+00 3.78734440e-01 -3.87345254e-02
1.10046017e+00 4.37697977e-01 -2.13992506e-01 4.30628061e-01
-4.48705286e-01 -4.62556124e-01 5.58268987e-02 1.25021672e+00
3.14462304e-01 -2.23906085e-01 -9.77593124e-01 9.86602068e-01
-1.58464909e+00 -1.04299402e+00 -9.68191147e-01 1.79272580e+00
1.33623064e+00 -7.69987032e-02 -6.90074265e-02 2.90311188e-01
1.49129733e-01 -3.54830235e-01 1.55031577e-01 -6.90559268e-01
2.70876378e-01 7.96618700e-01 -1.72581568e-01 8.49693716e-01
-2.24695101e-01 8.60099077e-01 5.69474602e+00 7.88022518e-01
-5.08098960e-01 3.22338678e-02 3.24207395e-01 2.70077199e-01
-8.81957054e-01 1.48468077e-01 -8.12340379e-01 2.24357858e-01
1.10350287e+00 -6.60172462e-01 9.61845815e-02 2.75494754e-01
-6.73079537e-03 -4.84743536e-01 -9.08614337e-01 4.89581198e-01
3.51263344e-01 -1.27202034e+00 5.89757487e-02 -4.64065671e-01
7.71056294e-01 -1.07912099e+00 -4.12496805e-01 6.22135699e-01
5.77009380e-01 -1.11555827e+00 6.79699898e-01 5.95944047e-01
1.87463239e-01 -7.89118826e-01 8.30096126e-01 4.16542470e-01
-1.01641047e+00 1.47532746e-01 -7.32690841e-02 -1.92593887e-01
7.98256993e-02 3.88750404e-01 -1.11004186e+00 4.83622998e-01
4.60485995e-01 -1.45698905e-01 -8.79599214e-01 1.25942755e+00
-9.27275538e-01 8.18384588e-01 -6.81399554e-02 -7.06464946e-01
8.27042684e-02 -7.27395490e-02 2.85223424e-01 9.12631452e-01
4.92659807e-01 7.89176822e-01 -8.37122574e-02 9.00604606e-01
-1.99315444e-01 4.62386340e-01 -1.73054069e-01 5.36657199e-02
5.49951315e-01 1.21301293e+00 -4.36184645e-01 -1.95581242e-01
-2.73118436e-01 6.31390095e-01 1.06046349e-01 1.61607713e-01
-3.29039812e-01 -9.82314229e-01 9.31523517e-02 4.41042513e-01
1.53496236e-01 -1.05115891e-01 -6.03777170e-01 -8.45471084e-01
1.61904246e-01 -1.27766728e+00 2.88561940e-01 -1.06844127e+00
-1.04374516e+00 2.91796356e-01 2.16968060e-02 -9.44259405e-01
-5.96544743e-01 -2.12982357e-01 -8.77797186e-01 1.14949715e+00
-1.28341508e+00 -9.08278584e-01 -3.53256166e-01 3.98021042e-01
4.44385618e-01 4.33584005e-01 8.13575327e-01 3.53605330e-01
-1.23129807e-01 7.31742382e-01 -4.04330760e-01 3.07397097e-01
6.47438824e-01 -1.51062179e+00 3.02367568e-01 7.97747493e-01
1.13738939e-01 5.57543993e-01 8.88184428e-01 -6.94714129e-01
-1.13079393e+00 -8.54785442e-01 1.76682377e+00 -6.64323688e-01
7.26139784e-01 -7.93689638e-02 -1.28978562e+00 2.85461605e-01
6.53276801e-01 -6.22944415e-01 1.24125314e+00 -1.71636954e-01
1.36255659e-02 1.55508816e-01 -1.24501717e+00 7.31025934e-01
5.12953997e-01 -4.84645486e-01 -1.32682216e+00 1.42431989e-01
7.84742594e-01 -6.06252849e-01 -9.94140327e-01 8.03875476e-02
2.75373369e-01 -7.43392706e-01 6.21172905e-01 -5.96800327e-01
7.28287101e-01 -2.10607156e-01 2.26619154e-01 -1.27906525e+00
1.34222150e-01 -5.08238435e-01 8.14901516e-02 1.61215639e+00
7.83641696e-01 -3.02747965e-01 7.17526674e-01 1.05056763e+00
-1.67968720e-01 -7.75619507e-01 -7.75975049e-01 -3.25431883e-01
3.23583245e-01 -3.61041665e-01 8.35921466e-01 9.34777915e-01
2.42530674e-01 7.04218805e-01 3.17050785e-01 -2.96221584e-01
2.63723850e-01 7.79369846e-02 7.45680392e-01 -1.16251528e+00
-2.36343682e-01 -3.86340946e-01 3.18155795e-01 -1.08908391e+00
-1.26271948e-01 -6.46506071e-01 2.50234127e-01 -2.13986897e+00
-2.62187898e-01 -4.42108363e-01 4.09542352e-01 3.11720669e-01
-5.63930750e-01 5.05443402e-02 1.88161895e-01 -2.45070145e-01
-5.08138418e-01 4.66776043e-01 1.62022817e+00 3.64490569e-01
-2.54910160e-02 2.01636091e-01 -1.00148273e+00 5.46186149e-01
7.80052960e-01 -5.71114421e-01 -6.35706484e-01 -4.20648128e-01
8.98951828e-01 3.97420406e-01 2.43423223e-01 -9.23539519e-01
5.00653088e-01 -3.70182246e-01 2.87603289e-01 -6.03193164e-01
-1.13780476e-01 -7.28573740e-01 -2.89405704e-01 2.70126820e-01
-5.66226006e-01 4.43649530e-01 3.71045500e-01 -2.74251461e-01
-4.66289133e-01 -1.29108167e+00 3.99528801e-01 -2.59340495e-01
-1.93106353e-01 -2.84901470e-01 -7.21144497e-01 5.79913914e-01
9.64572787e-01 -2.39203319e-01 -5.37797749e-01 -7.50284731e-01
-2.63233155e-01 5.25574505e-01 8.49218443e-02 3.00284445e-01
6.51124239e-01 -1.27712440e+00 -9.69179690e-01 -1.57732561e-01
-1.00174189e-01 1.58312097e-01 3.20422769e-01 2.91927904e-01
-7.30913520e-01 6.15690231e-01 -1.45295933e-01 7.36556873e-02
-1.58289075e+00 -1.36046156e-01 2.23053783e-01 -7.14367390e-01
5.93174715e-03 1.02627254e+00 -6.70978785e-01 -8.38139653e-01
-1.98260248e-02 -3.80483329e-01 -9.61644828e-01 2.65273631e-01
6.61867738e-01 3.30168009e-01 2.10960492e-01 -4.58857697e-03
2.26036459e-01 2.47139826e-01 6.31755888e-02 -4.38615739e-01
1.14762127e+00 -1.04237743e-01 -9.27193090e-02 2.39100739e-01
8.39077830e-01 4.80824471e-01 -4.36632365e-01 -4.95754667e-02
3.35013330e-01 -2.32681021e-01 -2.66674429e-01 -1.20770788e+00
-4.49385583e-01 7.41591215e-01 7.32099637e-02 5.34563184e-01
9.41908896e-01 -1.46434754e-01 1.10559702e+00 3.08846235e-01
-1.03489205e-01 -1.00513971e+00 1.37518123e-01 9.32719052e-01
9.24227655e-01 -9.59565878e-01 -1.57202810e-01 -6.39594853e-01
-2.50590384e-01 1.22234046e+00 1.04265416e+00 4.49039608e-01
1.19643144e-01 -5.23933731e-02 2.28351489e-01 -1.27354354e-01
-8.08038950e-01 -1.42597005e-01 7.15884030e-01 4.94760364e-01
9.16266739e-01 -2.59721965e-01 -9.24415708e-01 9.29106593e-01
-8.50470483e-01 1.93851039e-01 1.02603269e+00 1.29903913e+00
-6.40287757e-01 -1.56594551e+00 -5.88650644e-01 6.98658466e-01
-4.65248913e-01 -4.46439415e-01 -8.72256279e-01 5.90195835e-01
9.08429176e-02 1.38790464e+00 -2.32584268e-01 -1.86117053e-01
7.28117764e-01 5.63665926e-01 7.08329439e-01 -1.08125842e+00
-1.41018319e+00 -8.08259487e-01 5.87467670e-01 1.44012153e-01
-1.18755754e-02 -5.74177086e-01 -1.15322328e+00 -2.80498147e-01
-4.01697546e-01 7.58082330e-01 5.87265849e-01 1.26401258e+00
1.05780117e-01 8.59850943e-01 2.22799137e-01 3.55150104e-01
-5.85489213e-01 -1.06444347e+00 1.49002180e-01 2.38498330e-01
-5.02086505e-02 -2.00114518e-01 -2.30124757e-01 7.34365284e-02]
|
[11.48165225982666, 8.040923118591309]
|
d0550c61-220a-4e60-87b8-f23d7f9e1443
|
skellam-rank-fair-learning-to-rank-algorithm
|
2306.06607
| null |
https://arxiv.org/abs/2306.06607v1
|
https://arxiv.org/pdf/2306.06607v1.pdf
|
Skellam Rank: Fair Learning to Rank Algorithm Based on Poisson Process and Skellam Distribution for Recommender Systems
|
Recommender system is a widely adopted technology in a diversified class of product lines. Modern day recommender system approaches include matrix factorization, learning to rank and deep learning paradigms, etc. Unlike many other approaches, learning to rank builds recommendation results based on maximization of the probability of ranking orders. There are intrinsic issues related to recommender systems such as selection bias, exposure bias and popularity bias. In this paper, we propose a fair recommender system algorithm that uses Poisson process and Skellam distribution. We demonstrate in our experiments that our algorithm is competitive in accuracy metrics and far more superior than other modern algorithms in fairness metrics.
|
['Hao Wang']
|
2023-06-11
| null | null | null | null |
['selection-bias']
|
['natural-language-processing']
|
[-5.05434692e-01 -5.02598166e-01 -4.40150917e-01 -4.47659433e-01
-1.43143296e-01 -4.83547688e-01 6.05149329e-01 1.27818882e-01
-9.25923511e-02 9.72438931e-01 3.72041255e-01 -4.65152979e-01
-6.37396514e-01 -9.17696655e-01 -2.19018638e-01 -3.20349932e-01
-2.58969218e-01 8.97702694e-01 -5.14762476e-02 -6.98831141e-01
8.58046591e-01 4.10431772e-01 -1.42131233e+00 3.83268982e-01
9.33998644e-01 1.00539076e+00 -3.81533116e-01 6.85620666e-01
-1.67356819e-01 9.72708583e-01 -4.38335955e-01 -8.32135439e-01
5.43145359e-01 -1.96625248e-01 -4.01229411e-01 -6.96619570e-01
3.51559222e-01 -5.34736097e-01 -4.25332725e-01 8.82616222e-01
6.08511150e-01 5.25690854e-01 1.06156087e+00 -1.42155969e+00
-1.30737007e+00 1.00833392e+00 -8.90161276e-01 4.38580513e-01
2.22842246e-01 -6.47777855e-01 1.51155686e+00 -1.07130241e+00
9.64183882e-02 1.34926200e+00 6.18066847e-01 2.45734259e-01
-8.65613341e-01 -9.82502878e-01 2.28871256e-01 2.51581848e-01
-1.09431827e+00 2.32394263e-02 1.28881529e-01 -4.46228296e-01
3.48939538e-01 2.63245612e-01 4.25423115e-01 7.25230038e-01
6.60198152e-01 7.46107161e-01 1.10833788e+00 -7.87914619e-02
3.41631770e-01 2.54484624e-01 5.00871837e-01 1.15382433e-01
4.93252814e-01 6.70583129e-01 -5.89570642e-01 -6.57980740e-01
8.45333517e-01 5.80476642e-01 3.21391106e-01 -2.27301031e-01
-6.12614810e-01 1.37568033e+00 1.39009684e-01 -1.19801000e-01
-4.44435537e-01 3.02165896e-01 2.63716429e-01 7.66264737e-01
4.55340743e-01 4.19345081e-01 -5.73891342e-01 1.40318662e-01
-1.02347422e+00 4.31620330e-01 1.11119068e+00 7.50125170e-01
2.26724759e-01 7.63597861e-02 -3.90246928e-01 9.44537699e-01
6.56242371e-01 6.77580774e-01 5.22203147e-01 -9.74861801e-01
-1.91784993e-01 -3.76088396e-02 3.27492118e-01 -9.65838611e-01
-7.55061880e-02 -8.21269929e-01 -9.07274544e-01 2.40046680e-01
2.70577103e-01 -2.81037748e-01 -4.17120010e-01 1.10085213e+00
3.93340248e-04 2.36069247e-01 -1.90742910e-01 9.31910932e-01
6.27552330e-01 8.03378403e-01 -9.36293416e-03 -3.56064290e-01
7.97749221e-01 -8.80538404e-01 -6.00507438e-01 6.68839753e-01
2.04748008e-02 -1.29253256e+00 7.06101000e-01 1.05485761e+00
-1.04889023e+00 -5.17562628e-01 -7.84253001e-01 3.85911077e-01
-8.96075815e-02 -1.82264492e-01 1.29377532e+00 9.95712101e-01
-9.99651015e-01 1.11293936e+00 2.09982157e-01 -2.48425081e-01
5.46334922e-01 7.93650925e-01 1.80225343e-01 9.29151997e-02
-1.31023824e+00 7.95293152e-01 -2.44727537e-01 -1.87468216e-01
-8.95493686e-01 -9.12284672e-01 1.54651329e-01 2.53930420e-01
2.26506829e-01 -7.05837011e-01 1.35457695e+00 -9.51359928e-01
-1.88247907e+00 1.08080186e-01 3.82640004e-01 -5.51446736e-01
4.76788640e-01 -7.66747832e-01 -7.32664764e-01 -6.11720443e-01
-3.88987690e-01 -8.56182724e-02 8.74225974e-01 -1.02059793e+00
-1.27489221e+00 -2.50222653e-01 2.10290685e-01 2.06764996e-01
-1.62569478e-01 1.46408945e-01 4.37102586e-01 -4.59606946e-01
-2.65749127e-01 -8.55472028e-01 -6.11532807e-01 -4.11385298e-01
-2.93636951e-03 -5.33485293e-01 2.53027886e-01 -1.28822714e-01
1.25009096e+00 -1.80040324e+00 -4.38307613e-01 6.88264608e-01
4.20959651e-01 9.77661759e-02 -1.62195414e-01 5.70492744e-01
2.05761120e-01 4.77621965e-02 8.86277497e-01 4.19468045e-01
2.63090253e-01 -7.16514885e-02 -6.22393906e-01 5.52417219e-01
-6.97972596e-01 5.22376359e-01 -9.23989892e-01 -1.73788279e-01
1.24036036e-01 2.91946441e-01 -7.47941196e-01 1.87258929e-01
9.30482000e-02 -1.74726418e-03 -4.10404235e-01 4.80275661e-01
8.17393184e-01 -2.20982030e-01 1.71761557e-01 -6.29735086e-03
4.68233647e-03 1.80807218e-01 -1.33284163e+00 1.00487792e+00
-5.57526767e-01 2.92478800e-01 -2.93437183e-01 -6.94022119e-01
9.93452609e-01 1.66667446e-01 5.51384926e-01 -4.29220915e-01
5.15744507e-01 2.05694929e-01 4.41807330e-01 7.18629137e-02
1.10784984e+00 -2.55241930e-01 2.51653552e-01 9.23155487e-01
1.87766422e-02 3.69819760e-01 4.76631299e-02 6.26655877e-01
7.98162043e-01 -3.44046444e-01 2.21389428e-01 -7.06426322e-01
3.99998605e-01 -5.07987797e-01 4.03470367e-01 1.18921781e+00
-2.42327172e-02 3.03984344e-01 3.74883533e-01 -6.37605011e-01
-1.01155305e+00 -1.18131495e+00 -4.86276597e-02 1.96733892e+00
2.86979437e-01 -3.04595917e-01 4.99646068e-02 -7.79502451e-01
5.92922568e-01 6.41950369e-01 -5.60096800e-01 1.06600821e-02
-8.32404941e-02 -7.77719140e-01 -9.13409144e-02 4.12612796e-01
-2.03276783e-01 -8.90663981e-01 3.07025760e-01 2.20200479e-01
6.29151642e-01 -1.40938982e-01 -6.02367103e-01 4.45305631e-02
-9.81463432e-01 -8.71809959e-01 -6.72744870e-01 -2.68691957e-01
4.62486982e-01 5.69471776e-01 1.40872455e+00 1.98177189e-01
1.12168461e-01 1.88175455e-01 -5.25476933e-01 -4.00473654e-01
-1.38770536e-01 -6.22897372e-02 4.40954626e-01 -1.20074391e-01
4.46905792e-01 -6.15850806e-01 -1.02907705e+00 5.27315438e-01
-3.23379964e-01 -7.38242149e-01 6.88978732e-01 8.25825155e-01
3.11222196e-01 2.88368583e-01 8.53287637e-01 -1.66601253e+00
1.35040450e+00 -8.47548544e-01 -5.28561831e-01 1.15830235e-01
-1.40171862e+00 1.51208770e-02 6.17307663e-01 -5.61564565e-01
-1.06797361e+00 -4.55877274e-01 6.22111745e-02 -1.69785544e-02
2.87119359e-01 3.52486849e-01 2.01547071e-01 3.27100679e-02
7.46469200e-01 -3.04995775e-01 -1.75610736e-01 -4.26228970e-01
8.81658792e-01 7.01835692e-01 3.02277822e-02 -6.44188941e-01
7.75778413e-01 2.74228305e-01 -1.00238854e-02 -1.02074899e-01
-9.82671201e-01 -5.20738542e-01 -2.59595573e-01 -2.70598769e-01
-7.94199035e-02 -7.25536585e-01 -1.05756891e+00 -1.51604295e-01
-6.38199389e-01 4.15284149e-02 -4.29958314e-01 8.00706387e-01
-2.20043808e-01 4.02902626e-02 -8.55689466e-01 -8.92399907e-01
-8.60057473e-01 -6.11182094e-01 2.68298060e-01 5.82942724e-01
-2.35865086e-01 -9.36303377e-01 4.16792035e-01 8.86142403e-02
7.66437292e-01 -4.32099730e-01 6.40808463e-01 -1.27594841e+00
-3.00888062e-01 -3.27862114e-01 -3.73430431e-01 2.48763144e-01
-9.14956555e-02 3.17746609e-01 -4.93757904e-01 -4.07855660e-01
-3.65148425e-01 6.24227896e-02 7.35766113e-01 6.81171715e-01
1.02402437e+00 -9.75110307e-02 -1.95018739e-01 6.54444098e-02
1.46621537e+00 4.14299756e-01 6.42280638e-01 -9.02548134e-02
6.09479189e-01 4.36893940e-01 9.03824329e-01 8.02685380e-01
1.56243607e-01 1.45674050e-01 1.25545129e-01 8.69534537e-02
3.32718223e-01 -1.14930779e-01 2.86139905e-01 9.02386069e-01
-2.74636656e-01 -4.35229480e-01 -2.35898525e-01 8.49437192e-02
-2.25161242e+00 -1.25021112e+00 -3.19819182e-01 2.44714880e+00
3.01208258e-01 1.22102223e-01 4.16931331e-01 -1.23442560e-01
1.00442719e+00 -3.65675569e-01 -4.53175068e-01 -8.88246477e-01
8.46097097e-02 4.26582366e-01 8.65423918e-01 4.36987460e-01
-7.45230198e-01 7.83791184e-01 7.89956999e+00 8.97368133e-01
-6.75231814e-01 3.04442257e-01 8.17753494e-01 -5.48629403e-01
-5.36145389e-01 5.26855886e-02 -8.11768055e-01 4.71215069e-01
9.63402808e-01 -6.92527831e-01 4.38616425e-01 1.10747278e+00
1.94927588e-01 -3.68568674e-02 -1.09725332e+00 1.01591253e+00
1.26560614e-01 -1.28393805e+00 1.42874956e-01 5.78994870e-01
1.18761933e+00 -1.23063661e-01 5.52032351e-01 5.69518805e-01
1.32724178e+00 -1.22130883e+00 3.38535339e-01 7.46829271e-01
2.87310600e-01 -1.27080452e+00 1.04259753e+00 -1.90591589e-01
-6.49644792e-01 -3.48751843e-01 -1.10606420e+00 -4.34278160e-01
1.47090107e-01 1.07884157e+00 -1.88439772e-01 2.64664114e-01
5.25834143e-01 7.13024318e-01 -5.19304536e-02 1.47255075e+00
1.09607384e-01 7.79678464e-01 -2.97887381e-02 -4.51922178e-01
-9.15667042e-02 -4.64403957e-01 7.88253024e-02 7.48872697e-01
5.37846804e-01 2.51519531e-01 2.33561009e-01 2.89737761e-01
-3.02286118e-01 6.68962955e-01 -3.70288491e-01 2.61325419e-01
5.43045461e-01 1.67905033e+00 -6.46631658e-01 -3.55756313e-01
-3.90372425e-01 4.45931226e-01 -1.16099000e-01 1.62972789e-02
-7.55479634e-01 -5.78938276e-02 6.71328247e-01 3.53284329e-01
3.26281041e-01 1.54151455e-01 -2.60025412e-01 -8.07453632e-01
-8.92423332e-01 -9.36821818e-01 6.03585601e-01 -3.30189228e-01
-2.16881275e+00 3.87530714e-01 -4.11961466e-01 -1.18819165e+00
3.25187504e-01 -6.49727762e-01 -8.76791835e-01 7.84644127e-01
-1.22943854e+00 -6.36670411e-01 9.47997496e-02 6.69285595e-01
4.62793410e-01 -1.08023345e+00 4.16305274e-01 6.48202419e-01
-2.30852067e-01 7.51930356e-01 9.11001027e-01 -5.66473961e-01
1.13165629e+00 -1.48775995e+00 4.01280038e-02 1.95703596e-01
3.21446806e-01 9.81163979e-01 8.01521957e-01 -5.22148192e-01
-1.30512464e+00 -7.02546418e-01 4.57113922e-01 -5.26811779e-01
9.12121892e-01 1.58717737e-01 -3.89042228e-01 1.61121815e-01
3.27812254e-01 -1.88965991e-01 1.40348768e+00 1.06803560e+00
-3.74808252e-01 -4.50279236e-01 -1.02293193e+00 3.56699675e-01
7.61508822e-01 -1.04413010e-01 -3.42337310e-01 5.85272789e-01
2.51200408e-01 1.06327564e-01 -9.76066053e-01 1.86732076e-02
1.08537054e+00 -1.01764536e+00 7.78285384e-01 -9.75916028e-01
5.75337768e-01 -1.82935610e-01 -2.33960256e-01 -1.31629646e+00
-1.11377668e+00 -9.00259793e-01 -2.89213300e-01 1.24534202e+00
5.53718567e-01 -5.00244439e-01 1.01434886e+00 4.95664001e-01
4.54568774e-01 -6.25507891e-01 -3.49483579e-01 -6.80377841e-01
4.07449514e-01 3.18752858e-03 7.99312949e-01 8.02205920e-01
2.52161957e-02 7.60735869e-01 -1.00047886e+00 -1.99118048e-01
6.30798995e-01 4.42141742e-01 7.27335155e-01 -1.92896450e+00
-4.38076645e-01 -4.64520246e-01 3.16000846e-03 -9.27223384e-01
-4.78197671e-02 -8.77503037e-01 -2.09248558e-01 -1.34107983e+00
4.52511340e-01 -9.34853554e-01 -1.32639837e+00 -3.14491838e-01
8.24229941e-02 3.18058342e-01 -8.10370687e-03 3.75283957e-01
-1.10914326e+00 2.83460915e-01 1.11719966e+00 1.50470048e-01
-6.31711185e-02 5.69968224e-01 -1.35380197e+00 5.34697056e-01
1.01676059e+00 -6.18169367e-01 -6.03066862e-01 6.99176192e-02
8.58139038e-01 4.75084260e-02 -4.47852492e-01 -5.00225842e-01
2.45422050e-01 -4.24367487e-01 5.45934916e-01 -7.99763560e-01
-3.08182180e-01 -6.74361050e-01 2.18831435e-01 3.35380554e-01
-6.73307300e-01 2.96288431e-01 -4.34124321e-01 7.69029081e-01
2.43823886e-01 -3.53783399e-01 5.95033884e-01 -2.04773471e-02
-4.16525304e-01 6.44202828e-01 -5.40854216e-01 -1.62055388e-01
6.99013054e-01 7.42916837e-02 -3.75938654e-01 -8.26046824e-01
-6.10774696e-01 1.64312854e-01 1.16886355e-01 6.40533328e-01
4.60224360e-01 -1.43789029e+00 -9.19488549e-01 -3.89346063e-01
-2.44257376e-01 -1.13949251e+00 1.10240437e-01 5.71379423e-01
-4.45979476e-01 2.50331968e-01 -4.66018677e-01 1.69674441e-01
-1.20534301e+00 5.66204906e-01 -2.13491116e-02 -3.78061742e-01
-1.34305969e-01 9.83970821e-01 1.65185452e-01 -3.26554149e-01
1.33375779e-01 2.41426200e-01 -5.64787865e-01 4.95350324e-02
6.92621112e-01 9.30300653e-01 -4.20372307e-01 -3.27498138e-01
-5.89637682e-02 2.27808490e-01 -6.75864577e-01 -9.31048468e-02
1.15138674e+00 -1.83810607e-01 -2.50197560e-01 3.01430374e-01
5.44351637e-01 3.83251071e-01 -7.17368424e-01 -2.67539769e-01
-9.95457247e-02 -9.79726017e-01 1.75538972e-01 -9.76109087e-01
-1.24567044e+00 5.53103209e-01 9.71148431e-01 5.65134287e-01
5.29163659e-01 -3.85118067e-01 8.31361890e-01 3.14388335e-01
4.51283485e-01 -1.35804093e+00 5.47220558e-02 6.36206031e-01
4.79733199e-01 -1.33414316e+00 7.23815918e-01 -1.42192543e-01
-6.99705839e-01 9.92590308e-01 5.52821457e-01 -7.88327575e-01
1.36883950e+00 6.08804263e-02 9.31774154e-02 -6.17834600e-03
-1.04645157e+00 3.66363861e-02 3.69592309e-01 6.03519380e-01
1.00011480e+00 2.70826876e-01 -9.62194204e-01 9.30784762e-01
-3.57521355e-01 -1.13708628e-02 6.59204721e-01 3.01785290e-01
-7.37343669e-01 -1.49883068e+00 -2.09722757e-01 1.27285480e+00
-8.04341316e-01 -4.30825084e-01 -1.01484179e-01 1.81554183e-01
9.87355188e-02 1.19701064e+00 -1.93730695e-03 -5.84888756e-01
1.04087718e-01 -5.63907921e-01 4.28608030e-01 -5.49183547e-01
-9.17275846e-01 9.03568789e-03 -7.22101051e-03 -3.19374442e-01
-8.78104344e-02 -5.64545929e-01 -7.52594888e-01 -1.08801663e+00
-8.94437730e-01 7.63799608e-01 4.96185809e-01 7.21685767e-01
2.27056265e-01 4.26621020e-01 8.97396564e-01 -3.54905784e-01
-8.53860378e-01 -7.04610944e-01 -1.34819591e+00 2.52540797e-01
-4.55120176e-01 -8.41930985e-01 -2.26993039e-01 -3.20493013e-01]
|
[9.955678939819336, 5.709502696990967]
|
a50ae757-4931-479e-9697-00711daabe1a
|
grassmann-pooling-as-compact-homogeneous
| null | null |
http://openaccess.thecvf.com/content_ECCV_2018/html/Xing_Wei_Grassmann_Pooling_for_ECCV_2018_paper.html
|
http://openaccess.thecvf.com/content_ECCV_2018/papers/Xing_Wei_Grassmann_Pooling_for_ECCV_2018_paper.pdf
|
Grassmann Pooling as Compact Homogeneous Bilinear Pooling for Fine-Grained Visual Classification
|
Designing discriminative and invariant features is the key to visual recognition. Recently, the bilinear pooled feature matrix of Convolutional Neural Network (CNN) has shown to achieve state-of-the-art performance on a range of fine-grained visual recognition tasks. The bilinear feature matrix collects second-order statistics and is closely related to the covariance matrix descriptor. However, the bilinear feature could suffer from the visual burstiness phenomenon similar to other visual representations such as VLAD and Fisher Vector. The reason is that the bilinear feature matrix is sensitive to the magnitudes and correlations of local CNN feature elements which can be measured by its singular values. On the other hand, the singular vectors are more invariant and reasonable to be adopted as the feature representation. Motivated by this point, we advocate an alternative pooling method which transforms the CNN feature matrix to an orthonormal matrix consists of its principal singular vectors. Geometrically, such orthonormal matrix lies on the Grassmann manifold, a Riemannian manifold whose points represent subspaces of the Euclidean space. Similarity measurement of images reduces to comparing the principal angles between these ``homogeneous" subspaces and thus is independent of the magnitudes and correlations of local CNN activations. In particular, we demonstrate that the projection distance on the Grassmann manifold deduces a bilinear feature mapping without explicitly computing the bilinear feature matrix, which enables a very compact feature and classifier representation. Experimental results show that our method achieves an excellent balance of model complexity and accuracy on a variety of fine-grained image classification datasets.
|
['Yue Zhang', 'Xing Wei', 'Yihong Gong', 'Nanning Zheng', 'Jiawei Zhang']
|
2018-09-01
| null | null | null |
eccv-2018-9
|
['fine-grained-visual-recognition']
|
['computer-vision']
|
[-2.70843208e-01 -5.14569700e-01 -6.49563447e-02 -3.35288703e-01
-2.68084973e-01 -6.68695807e-01 7.82863081e-01 -2.09611356e-01
-5.24920166e-01 2.57502764e-01 9.96060967e-02 1.27195999e-01
-5.17786980e-01 -6.03274345e-01 -5.67054570e-01 -1.08692682e+00
-6.13601543e-02 -3.38425428e-01 1.03508785e-01 -1.17846683e-01
2.88753778e-01 8.60650420e-01 -1.68681967e+00 2.37055764e-01
3.79259437e-01 1.50430954e+00 -2.60175802e-02 3.20923299e-01
1.22128829e-01 4.24558043e-01 -2.50292361e-01 -7.45883361e-02
3.42940360e-01 -1.45598039e-01 -6.95447564e-01 -1.96386538e-02
5.11497676e-01 6.34578317e-02 -6.08436942e-01 1.24118471e+00
1.64540172e-01 1.54521957e-01 9.16981816e-01 -1.30294764e+00
-9.10853088e-01 -8.78315195e-02 -2.58573115e-01 2.84558088e-01
-3.53532936e-03 -1.40555874e-01 1.26909268e+00 -1.29420674e+00
4.82611716e-01 1.07565916e+00 4.82346416e-01 1.10764354e-01
-1.11683536e+00 -2.39329189e-01 -1.84871275e-02 3.65906090e-01
-1.93610346e+00 -1.93512127e-01 6.74389243e-01 -6.26005232e-01
7.81350434e-01 5.56391358e-01 5.37671030e-01 6.45780325e-01
4.16666627e-01 3.32274646e-01 9.88635063e-01 -2.37613067e-01
9.00149867e-02 5.22174723e-02 1.03171475e-01 9.07296896e-01
2.29847327e-01 1.82657633e-02 -4.13036287e-01 -7.86214545e-02
9.61188376e-01 4.84695733e-01 -3.97085965e-01 -7.82341421e-01
-1.56561923e+00 8.85918975e-01 8.90976846e-01 6.59930706e-01
-3.41423631e-01 -1.05863484e-02 7.61845708e-02 3.29687536e-01
-1.01195097e-01 2.58600682e-01 -1.59136221e-01 1.76725298e-01
-4.15891260e-01 -5.79611324e-02 5.91029882e-01 5.96260488e-01
1.00284159e+00 -6.54224828e-02 -2.55873591e-01 6.36384428e-01
4.78666306e-01 4.99814630e-01 6.46615028e-01 -6.81689680e-01
2.72481740e-01 9.09117520e-01 -2.22887740e-01 -1.59108722e+00
-4.36427683e-01 -3.95653546e-01 -1.22476268e+00 1.23723112e-01
5.55905938e-01 4.36694503e-01 -3.57849121e-01 1.63755250e+00
-4.07754555e-02 -1.79100931e-01 -2.15743437e-01 1.12181437e+00
7.31807888e-01 3.60370696e-01 -2.76861101e-01 3.11701950e-02
1.38452578e+00 -4.32646543e-01 -3.81955117e-01 1.99825615e-01
4.27989572e-01 -6.61491394e-01 9.90898132e-01 -4.03533988e-02
-6.07274771e-01 -5.98628342e-01 -1.26975369e+00 -7.92360380e-02
-6.15098655e-01 2.95323402e-01 5.50141752e-01 4.53055471e-01
-1.00346816e+00 7.81060576e-01 -6.75849617e-01 -3.47950995e-01
1.39323652e-01 1.93816811e-01 -9.38051105e-01 4.00050916e-03
-9.55066025e-01 7.19202757e-01 1.67995274e-01 4.87620264e-01
-4.79083359e-01 -3.84412646e-01 -7.61100233e-01 2.49080181e-01
-3.09807956e-01 -3.97649944e-01 4.90455955e-01 -5.80804288e-01
-1.21707666e+00 7.43724465e-01 -1.05710380e-01 1.26499012e-01
1.59583420e-01 1.58625990e-01 -3.44899207e-01 8.43441561e-02
-5.37497103e-02 1.50576144e-01 1.02584600e+00 -7.71818161e-01
-2.73700297e-01 -5.53244591e-01 8.85686055e-02 4.03412655e-02
-7.02667356e-01 -8.40317011e-02 -1.41828254e-01 -5.34298718e-01
6.25433981e-01 -9.80069458e-01 -1.02243498e-01 1.43380657e-01
-3.02379668e-01 -3.00869346e-01 7.74303675e-01 -1.98140278e-01
1.19733429e+00 -2.53403473e+00 2.00098813e-01 5.74469924e-01
3.37630242e-01 -8.08228478e-02 -2.04809785e-01 4.20061499e-01
-2.82715112e-01 7.66393542e-02 1.44081086e-01 3.63532186e-01
1.21329561e-01 -3.19469981e-02 -1.87721014e-01 9.07387614e-01
4.68899280e-01 8.75035405e-01 -6.91828489e-01 -2.47464925e-01
1.38543710e-01 6.44097805e-01 -4.48157400e-01 6.92419633e-02
6.18136883e-01 2.38092422e-01 -7.18359947e-01 3.27217847e-01
6.97443604e-01 -3.92522216e-01 7.24875880e-03 -8.14201951e-01
-2.51752496e-01 -1.28782419e-02 -1.31506515e+00 1.37779069e+00
-1.42629027e-01 6.43715739e-01 -2.99701065e-01 -1.15405262e+00
1.15252745e+00 4.03322689e-02 2.38766119e-01 -4.94064629e-01
3.76984119e-01 1.03575006e-01 1.44060403e-01 -2.52345055e-01
2.17727557e-01 1.26345932e-01 -1.02319740e-01 1.16393477e-01
2.64882624e-01 7.66895860e-02 2.15796288e-04 -4.94413413e-02
9.18446481e-01 -2.36694857e-01 5.79176426e-01 -6.33305609e-01
7.94861436e-01 -5.96196353e-01 4.22293603e-01 5.49128294e-01
-2.41186187e-01 5.89011312e-01 5.46709001e-01 -6.59263372e-01
-6.89173996e-01 -1.33150589e+00 -4.65503335e-01 7.53664672e-01
1.10941857e-01 -4.66583461e-01 -6.44765019e-01 -4.60483938e-01
-4.52324115e-02 -1.51279017e-01 -7.23450303e-01 -4.83366072e-01
-2.10373446e-01 -5.49822390e-01 3.35233122e-01 3.83869082e-01
6.71596408e-01 -6.76705599e-01 -4.41428959e-01 -1.60467014e-01
-6.00655098e-03 -8.74566555e-01 -6.20690703e-01 1.78065121e-01
-8.26899529e-01 -1.18530560e+00 -4.65453058e-01 -7.84589887e-01
8.37247491e-01 6.96887612e-01 5.95465481e-01 -1.14355320e-02
-5.61440289e-01 3.50428700e-01 -2.81669289e-01 1.25384703e-01
1.73701733e-01 -1.79143637e-01 3.87730092e-01 7.50559390e-01
5.37627339e-01 -5.87477267e-01 -7.41521060e-01 5.02606392e-01
-9.90603566e-01 -4.39618140e-01 6.94940627e-01 1.08672273e+00
7.11444616e-01 -7.98866302e-02 1.54428303e-01 -5.81867471e-02
5.18551290e-01 -3.01207006e-01 -4.01418865e-01 3.41288567e-01
-1.50325865e-01 3.33581656e-01 7.91774154e-01 -2.12430343e-01
-4.55292463e-01 8.23588297e-02 3.19079012e-01 -4.42004174e-01
-5.57892546e-02 5.09379566e-01 -2.64913946e-01 -6.14552259e-01
6.75003707e-01 3.97320032e-01 -1.30345076e-01 -4.42946821e-01
3.97419810e-01 6.23864830e-01 3.06859791e-01 -3.25681567e-01
8.94784927e-01 6.58781171e-01 3.21368456e-01 -1.11366212e+00
-7.22777367e-01 -5.79450428e-01 -1.05892229e+00 -1.45933837e-01
8.04472983e-01 -7.41273046e-01 -1.06740725e+00 4.64735866e-01
-1.01665580e+00 3.99820298e-01 -1.37871832e-01 7.66228437e-01
-5.18160760e-01 4.23964530e-01 -5.60039341e-01 -3.06935787e-01
-2.34955344e-02 -1.00937462e+00 8.19404900e-01 1.21363096e-01
-2.17003450e-02 -8.82512212e-01 -6.36505112e-02 -2.00438231e-01
3.64320874e-01 1.84976056e-01 9.86872256e-01 -3.62059057e-01
-6.95507050e-01 -5.23429751e-01 -5.21994352e-01 5.59097409e-01
1.57081559e-01 3.15317996e-02 -9.78563309e-01 -4.25796956e-01
2.35254779e-01 -7.65551478e-02 9.20304656e-01 2.05200657e-01
1.13094985e+00 -2.32825309e-01 -6.22319952e-02 7.29757428e-01
1.46269226e+00 -7.45777935e-02 5.18593907e-01 3.07026058e-01
8.16108286e-01 4.93053794e-01 3.28496963e-01 4.29087400e-01
1.88475788e-01 7.05096066e-01 2.89364904e-01 1.49009421e-01
2.54152745e-01 -1.00793988e-01 3.19923520e-01 1.24164355e+00
-3.46645504e-01 4.79344517e-01 -6.30058587e-01 2.98544854e-01
-1.82183170e+00 -8.77748430e-01 6.38832375e-02 2.42557383e+00
4.23918337e-01 -2.12621510e-01 -2.88949072e-01 2.01864034e-01
7.06323445e-01 3.31828296e-01 -2.62541831e-01 -4.13725108e-01
-3.99916410e-01 3.30454148e-02 3.62672061e-01 1.12359785e-01
-1.22295952e+00 6.52399838e-01 5.71930885e+00 7.31834650e-01
-1.29774654e+00 2.36718506e-02 2.58953482e-01 1.76165417e-01
-6.69444725e-02 -2.07969412e-01 -5.53443789e-01 1.95887879e-01
5.44933021e-01 -1.69109449e-01 4.73027438e-01 7.52412438e-01
-1.26870826e-01 2.48687953e-01 -1.17530823e+00 1.34815919e+00
1.77738249e-01 -1.39856994e+00 4.25376087e-01 2.58824617e-01
6.92337453e-01 1.35062486e-01 3.26289445e-01 8.70840345e-03
-1.73408359e-01 -1.08109057e+00 6.63947701e-01 7.78506398e-01
7.98285782e-01 -6.75237000e-01 6.48699045e-01 9.14120525e-02
-1.44078207e+00 -4.64285128e-02 -7.75512695e-01 -1.61782965e-01
-4.77034956e-01 4.73004162e-01 -3.53613734e-01 3.82303745e-01
9.43341792e-01 7.93575406e-01 -6.64861739e-01 1.00495708e+00
1.34248242e-01 -8.19361508e-02 -7.68978819e-02 -1.53679594e-01
4.66921747e-01 -5.60292363e-01 3.19514751e-01 1.01158690e+00
4.92165565e-01 6.10826984e-02 5.51909581e-03 7.34098196e-01
-4.97399867e-02 3.33916932e-01 -8.91438127e-01 -1.42543077e-01
2.33661011e-01 1.59373319e+00 -6.21038079e-01 7.33652711e-02
-6.29947543e-01 9.13030505e-01 5.59163332e-01 4.76537049e-01
-3.53048772e-01 -6.83761775e-01 1.09335804e+00 -2.07904026e-01
5.42414904e-01 -4.65843111e-01 3.46991532e-02 -1.35817134e+00
3.31007749e-01 -5.12341201e-01 6.72974959e-02 -3.44491392e-01
-1.41844237e+00 8.81755233e-01 -4.08553451e-01 -1.59432149e+00
9.04911198e-03 -1.13557434e+00 -4.29017216e-01 8.49301696e-01
-1.27618062e+00 -8.53386283e-01 -4.82717842e-01 1.17812371e+00
-2.51226872e-01 -2.17439964e-01 1.05326068e+00 1.23036347e-01
-4.08955514e-01 6.79633498e-01 5.01343071e-01 4.46339130e-01
4.38868642e-01 -1.00762510e+00 1.82399467e-01 5.80936015e-01
2.76058763e-01 9.52677608e-01 2.50697106e-01 1.81129966e-02
-1.66886842e+00 -1.10120821e+00 9.68703985e-01 -3.90129626e-01
8.86805654e-01 -3.52515519e-01 -7.99723327e-01 4.54584986e-01
-4.01039720e-01 5.61674654e-01 7.42815256e-01 -8.74936059e-02
-9.02483702e-01 -4.03251797e-01 -8.36921513e-01 6.04830682e-01
9.90786195e-01 -1.09151137e+00 -3.85568023e-01 2.33413443e-01
3.06569248e-01 1.19273558e-01 -1.17726135e+00 3.08139503e-01
8.10826838e-01 -1.05714858e+00 9.06069577e-01 -8.07527721e-01
2.45984197e-02 -5.94101965e-01 -6.64078712e-01 -1.27410555e+00
-8.60741854e-01 -1.10382922e-01 2.10626245e-01 7.37677097e-01
6.72680512e-02 -8.54653358e-01 4.63329405e-01 2.32770711e-01
2.05246508e-01 -8.27761829e-01 -1.25960171e+00 -9.72745419e-01
-9.02221650e-02 -2.30216831e-01 3.76021057e-01 8.89116526e-01
1.91183880e-01 1.28428206e-01 -1.52563006e-01 1.70769572e-01
5.32996118e-01 3.01718861e-01 4.00966525e-01 -1.44722712e+00
1.17000066e-01 -5.41880846e-01 -1.20125711e+00 -9.42564547e-01
1.69789538e-01 -1.13692653e+00 -3.54893971e-03 -1.03644383e+00
3.53809953e-01 -2.78703183e-01 -7.70159662e-01 2.44612515e-01
2.81062573e-01 6.05934441e-01 3.11694145e-01 3.50000203e-01
-4.92787510e-01 6.57291234e-01 1.30683112e+00 -1.22918472e-01
5.56354336e-02 -2.40554392e-01 -4.87274796e-01 7.80776918e-01
6.34015024e-01 -1.54669791e-01 8.02970007e-02 -2.45070428e-01
1.48683593e-01 -3.64654928e-01 5.61010420e-01 -9.12546337e-01
3.20778280e-01 -1.11098818e-01 4.91993934e-01 -7.73192570e-02
3.37284505e-01 -7.61724114e-01 9.39280838e-02 4.22664881e-01
-6.34984970e-02 1.28034115e-01 -6.07301928e-02 4.80108082e-01
-6.47974014e-01 -1.87086333e-02 8.47864211e-01 -2.14532390e-02
-7.98917174e-01 6.04864001e-01 -2.17749819e-01 -1.44399643e-01
7.83989787e-01 -3.37552309e-01 -3.30193251e-01 -1.44796863e-01
-4.22832876e-01 -4.62272763e-01 3.99370193e-01 6.51288927e-01
7.65667737e-01 -1.93568194e+00 -5.07567763e-01 5.95122755e-01
3.92888457e-01 -5.72006106e-01 3.35223883e-01 1.03887939e+00
-2.47250780e-01 8.58556688e-01 -6.61905050e-01 -8.07371080e-01
-1.01861179e+00 5.07405221e-01 4.57418263e-01 -2.57595144e-02
-2.48533398e-01 6.83713138e-01 6.67210460e-01 -3.18742901e-01
-5.43098450e-02 -3.76648843e-01 -3.62619698e-01 1.42238200e-01
7.02266574e-01 2.30402693e-01 2.78456092e-01 -1.13078535e+00
-7.52476990e-01 1.04749393e+00 3.56124230e-02 3.09673637e-01
1.15778542e+00 -6.32448718e-02 -4.49409664e-01 4.35334802e-01
1.96685684e+00 -1.85993657e-01 -1.08370292e+00 -5.07630944e-01
1.78056806e-02 -6.07937872e-01 8.84521659e-03 -2.54883394e-02
-1.02819002e+00 9.58671331e-01 6.73827291e-01 3.54519367e-01
9.06964600e-01 1.24717079e-01 1.54188946e-01 7.06633985e-01
4.98107344e-01 -8.05848956e-01 4.04166244e-02 7.37959623e-01
1.29140735e+00 -1.25221610e+00 -2.42071256e-01 -2.52149731e-01
-2.90540427e-01 1.34369850e+00 2.61705101e-01 -4.43478912e-01
1.09147179e+00 -2.82409728e-01 8.11454505e-02 -2.66903281e-01
-3.05505931e-01 -1.81142822e-01 8.66646349e-01 3.82720262e-01
3.61582458e-01 3.43027204e-01 -1.75923750e-01 5.39400697e-01
-3.21945786e-01 -3.88856828e-01 9.65259522e-02 4.08959955e-01
-4.88455921e-01 -6.71735108e-01 -2.72572458e-01 3.25848222e-01
-2.51636416e-01 -1.22524455e-01 -1.39084011e-01 6.06031120e-01
9.99848470e-02 7.05866396e-01 1.98512211e-01 -5.88634729e-01
4.36504155e-01 -1.86978310e-01 5.53176999e-01 -3.40589762e-01
-2.07574502e-01 -2.76097119e-01 -5.32190323e-01 -1.02734458e+00
-2.65841275e-01 -5.82873106e-01 -9.37701225e-01 -2.80788928e-01
-1.41799122e-01 9.54110622e-02 6.86481237e-01 9.81309354e-01
3.38390648e-01 8.02414492e-02 9.86702800e-01 -1.04113626e+00
-7.29560614e-01 -6.86976910e-01 -1.03533089e+00 6.61202550e-01
3.57093066e-01 -8.88468087e-01 -5.69537640e-01 -1.05698958e-01]
|
[8.999458312988281, 2.137913465499878]
|
1f962528-6c6a-4792-b6b6-a4c3e7cd1df5
|
copula-density-neural-estimation
|
2211.15353
| null |
https://arxiv.org/abs/2211.15353v1
|
https://arxiv.org/pdf/2211.15353v1.pdf
|
Copula Density Neural Estimation
|
Probability density estimation from observed data constitutes a central task in statistics. Recent advancements in machine learning offer new tools but also pose new challenges. The big data era demands analysis of long-range spatial and long-term temporal dependencies in large collections of raw data, rendering neural networks an attractive solution for density estimation. In this paper, we exploit the concept of copula to explicitly build an estimate of the probability density function associated to any observed data. In particular, we separate univariate marginal distributions from the joint dependence structure in the data, the copula itself, and we model the latter with a neural network-based method referred to as copula density neural estimation (CODINE). Results show that the novel learning approach is capable of modeling complex distributions and it can be applied for mutual information estimation and data generation.
|
['Andrea M. Tonello', 'Nunzio A. Letizia']
|
2022-11-25
| null | null | null | null |
['mutual-information-estimation']
|
['methodology']
|
[-2.83508599e-01 -2.25938976e-01 -1.38952807e-01 -5.06263912e-01
-5.34550250e-01 -2.13514328e-01 6.45201683e-01 2.55362570e-01
-4.03281897e-01 1.19125509e+00 6.23698309e-02 -1.32285640e-01
-3.75861645e-01 -1.10835922e+00 -9.71246123e-01 -8.64861369e-01
-4.84213769e-01 9.04405355e-01 -2.02109218e-01 2.83486843e-01
-1.42094761e-01 6.19405210e-01 -1.37169671e+00 -2.84927070e-01
1.04638946e+00 9.55389917e-01 5.63677609e-01 5.27826011e-01
-4.70192641e-01 6.72490358e-01 -8.14026833e-01 -3.91279072e-01
-4.18579340e-01 -2.18046591e-01 -1.88190058e-01 -1.98129654e-01
-1.91426948e-01 -2.07402810e-01 9.28442776e-02 9.53072309e-01
2.73253798e-01 1.00399882e-01 1.40548611e+00 -1.27867818e+00
-5.23554742e-01 5.09714723e-01 -5.93105853e-01 2.44830009e-02
-8.83845910e-02 -5.33419788e-01 6.65192008e-01 -7.49112368e-01
2.31871843e-01 9.95952129e-01 7.18958974e-01 9.87128019e-02
-1.48796737e+00 -2.43210346e-01 -3.23408842e-01 2.57471830e-01
-1.70397973e+00 1.73756704e-02 5.76163590e-01 -4.84934360e-01
7.86786139e-01 -1.68669537e-01 5.28642714e-01 1.18307674e+00
4.50977504e-01 7.40516186e-01 9.77690518e-01 -3.38757813e-01
6.15055561e-01 2.79345721e-01 -1.12213783e-01 2.22215742e-01
3.39839995e-01 1.38707414e-01 -1.59495875e-01 -2.51024336e-01
8.80336285e-01 1.85525984e-01 1.19514216e-03 -4.23145473e-01
-7.00381696e-01 9.87338364e-01 2.90883958e-01 2.36551791e-01
-5.46049535e-01 1.92311451e-01 8.64090864e-03 -1.18439801e-01
7.87992358e-01 -7.93574378e-02 -4.58265066e-01 -3.91518325e-01
-9.31712449e-01 3.66799444e-01 1.27212524e+00 6.38383329e-01
1.08936167e+00 8.18417519e-02 6.00794889e-02 9.60689366e-01
2.40501150e-01 7.95562565e-01 3.28000695e-01 -5.30589700e-01
3.51460993e-01 3.82543325e-01 1.78193390e-01 -8.42880964e-01
-5.03066719e-01 -2.98626304e-01 -1.19823897e+00 -6.71147648e-03
6.59865677e-01 -6.28307223e-01 -8.09834898e-01 1.74467731e+00
2.88230151e-01 3.41236353e-01 -1.12661362e-01 4.49458897e-01
2.97077775e-01 9.72381234e-01 2.80175675e-02 -1.72789693e-01
9.50611413e-01 -2.58429468e-01 -8.57104778e-01 -1.98575631e-01
1.26198798e-01 -3.59799504e-01 6.40577614e-01 3.17203760e-01
-8.92592430e-01 -3.48931700e-01 -8.04869652e-01 3.80655318e-01
-6.00812316e-01 1.27796799e-01 6.72136843e-01 5.13498783e-01
-8.73985291e-01 7.26283908e-01 -1.09068298e+00 -8.46099481e-02
4.48819131e-01 4.19142365e-01 -3.86633426e-01 -2.89715752e-02
-9.55015779e-01 7.68819273e-01 7.39461362e-01 1.74721375e-01
-6.76827669e-01 -6.66833997e-01 -1.05480218e+00 4.62469786e-01
2.49234900e-01 -4.20498639e-01 1.00478208e+00 -5.25732040e-01
-1.21412611e+00 2.31169730e-01 -3.04107428e-01 -5.35086513e-01
1.68975428e-01 -2.72955865e-01 -2.49491021e-01 -2.18600348e-01
-2.38105983e-01 2.84743905e-01 1.05819452e+00 -1.05085731e+00
-4.43401337e-01 -3.41406703e-01 -7.91679919e-01 -2.55662501e-01
-3.65853757e-01 -2.56713361e-01 -2.86736041e-01 -4.72803354e-01
-2.95626134e-01 -5.64309835e-01 -5.18755578e-02 -4.64579225e-01
-3.20060372e-01 -1.99634001e-01 6.16300166e-01 -7.31828392e-01
1.14206529e+00 -2.08311701e+00 1.30188599e-01 5.43635786e-01
1.70719504e-01 1.27253272e-02 -1.61954407e-02 7.50322938e-01
1.39729738e-01 -2.46317893e-01 -4.95845735e-01 -6.79134011e-01
6.27257349e-03 5.42828560e-01 -4.12458517e-02 4.29195136e-01
5.06765068e-01 9.98268127e-01 -6.33694947e-01 -4.30181444e-01
2.58005321e-01 8.88466060e-01 -2.04841211e-01 5.49643755e-01
-3.77404273e-01 3.87500405e-01 -1.90993339e-01 3.65141124e-01
9.54234958e-01 -4.05517310e-01 1.23860568e-01 1.57930002e-01
4.51508816e-03 -1.60503298e-01 -1.03668296e+00 1.27989388e+00
-6.61696255e-01 5.60161293e-01 -7.89508224e-02 -1.33115494e+00
1.25946915e+00 9.62604061e-02 3.68447363e-01 -2.99212307e-01
2.05013603e-01 1.86377078e-01 -2.40303832e-03 -4.07554805e-01
3.93627822e-01 -3.39667082e-01 -2.99891848e-02 4.00431246e-01
4.95686173e-01 8.79286695e-03 2.61418998e-01 -4.61545177e-02
7.91118145e-01 9.73793566e-02 4.86383200e-01 -7.93348327e-02
1.58424005e-01 -5.28812587e-01 2.83796966e-01 7.59522438e-01
2.73044318e-01 5.64322948e-01 8.17078769e-01 -2.51078337e-01
-1.19073641e+00 -1.59768677e+00 -4.44679350e-01 6.69661939e-01
-4.35410291e-01 -1.52277067e-01 -6.26400530e-01 -3.71982336e-01
3.22106421e-01 6.97631538e-01 -7.55733132e-01 -4.63048965e-02
-2.49848604e-01 -1.08291304e+00 1.68624923e-01 6.50034249e-01
1.89640090e-01 -1.17448902e+00 -2.72722244e-02 2.89431959e-01
-6.09635152e-02 -1.14546776e+00 -1.06493182e-01 4.77377683e-01
-9.13831592e-01 -6.80555224e-01 -1.00718260e+00 -3.46871465e-01
4.59157795e-01 -4.71500576e-01 1.28990531e+00 -5.49153805e-01
-4.02572900e-01 1.99364662e-01 -8.24694782e-02 -5.83146036e-01
-4.95321453e-01 1.02926381e-01 4.01346385e-02 1.40957549e-01
7.18184114e-01 -1.20896089e+00 -9.00355875e-02 1.11263670e-01
-9.32612181e-01 -3.80108684e-01 8.46479118e-01 8.42885017e-01
4.74563330e-01 1.76796734e-01 8.94928932e-01 -5.98279476e-01
8.79788101e-01 -1.01652730e+00 -1.03581405e+00 1.76363319e-01
-3.85103792e-01 3.70166451e-01 5.38462400e-01 -3.51660192e-01
-1.35532522e+00 -1.04646143e-02 -2.97332317e-01 -3.40977699e-01
-5.34420729e-01 1.02191329e+00 -3.10507804e-01 4.19626176e-01
3.85923475e-01 3.17893654e-01 3.94928813e-01 -7.10699379e-01
1.41316965e-01 7.12690234e-01 7.00195670e-01 -5.78208148e-01
4.91329521e-01 3.43237162e-01 3.30850005e-01 -9.47341084e-01
-7.24514365e-01 -3.95779222e-01 -8.77143860e-01 -6.23631757e-04
8.70853543e-01 -7.88929701e-01 -8.88966978e-01 6.71344042e-01
-1.15667403e+00 -3.54704708e-01 -4.28335786e-01 7.33224154e-01
-5.75402081e-01 1.76675290e-01 -3.85312200e-01 -1.19233048e+00
-1.46752059e-01 -6.54959142e-01 1.01906192e+00 5.17596483e-01
8.01515132e-02 -1.64828265e+00 6.38390779e-01 -3.01054597e-01
4.58252043e-01 2.62159437e-01 9.18369591e-01 -8.35310400e-01
-4.03245687e-01 -5.53833187e-01 -4.86639500e-01 7.37108648e-01
2.10648939e-01 2.44118590e-02 -9.10661995e-01 1.18446626e-01
-1.04953021e-01 -1.65344238e-01 6.97356880e-01 9.79230106e-01
1.18676889e+00 -1.01597957e-01 -3.05643141e-01 5.89902282e-01
1.38645375e+00 -9.39676389e-02 8.40547860e-01 -2.60994285e-01
5.40068090e-01 6.37133598e-01 5.65524101e-02 7.83984780e-01
4.37934786e-01 4.01070118e-01 3.31638336e-01 2.48739660e-01
5.39761961e-01 -5.44177234e-01 1.53017372e-01 6.35398507e-01
-4.33619134e-02 -3.09038401e-01 -1.00001407e+00 5.45513153e-01
-1.88628364e+00 -1.12990999e+00 -1.63414076e-01 2.43520856e+00
6.93928480e-01 -3.03921495e-02 2.55196720e-01 -3.03925306e-01
7.77735174e-01 -9.28057730e-02 -4.15168822e-01 -2.68636018e-01
-2.09329635e-01 3.65450263e-01 3.29360873e-01 3.63449991e-01
-1.05200553e+00 5.07920921e-01 7.09164667e+00 1.09336495e+00
-8.11518013e-01 -1.43679380e-01 5.96305728e-01 1.44660741e-01
-2.64281649e-02 -2.89711803e-01 -9.23992634e-01 6.78448200e-01
1.37599707e+00 -2.20159501e-01 2.20608741e-01 9.00672376e-01
1.10331640e-01 -4.60278362e-01 -1.03506267e+00 8.83132398e-01
-5.73858209e-02 -1.15808833e+00 -1.75195992e-01 5.60407758e-01
5.01577437e-01 1.02613695e-01 4.23816266e-03 3.08764160e-01
3.93285006e-01 -1.36139584e+00 1.31188422e-01 9.18407977e-01
4.89421248e-01 -1.01408243e+00 1.11351645e+00 7.44457960e-01
-8.79011869e-01 8.41139555e-02 -6.70504630e-01 -3.20621103e-01
5.46251059e-01 1.14395332e+00 -9.80246961e-01 3.56128424e-01
4.13994759e-01 6.29718125e-01 -4.39636469e-01 1.42362332e+00
-8.80801156e-02 7.86144376e-01 -6.78358614e-01 -3.18723410e-01
-1.13698706e-01 -7.12025344e-01 3.11807066e-01 1.06433022e+00
7.23640263e-01 -8.47194791e-01 -1.87529445e-01 1.35434878e+00
3.25815380e-02 -1.52102441e-01 -6.68746829e-01 -2.86336660e-01
4.52496916e-01 1.34095323e+00 -5.07448018e-01 -9.73842889e-02
-3.29817623e-01 6.99584484e-01 7.78087914e-01 3.88453424e-01
-6.38450563e-01 -2.25829571e-01 4.18692470e-01 -1.29266948e-01
5.87355912e-01 -3.94211292e-01 5.28264791e-02 -1.13154924e+00
-4.14987886e-03 -1.74142331e-01 2.61476785e-01 -6.41397595e-01
-1.84779787e+00 2.95954436e-01 5.69125593e-01 -8.83284569e-01
-8.55971098e-01 -7.34356165e-01 -9.75275576e-01 1.17187524e+00
-1.29254317e+00 -9.34686363e-01 -2.13793039e-01 5.56066096e-01
-4.08231765e-02 -2.36523598e-01 9.60014522e-01 2.44654030e-01
-3.85735989e-01 1.55398667e-01 8.09265494e-01 1.32635638e-01
3.63982767e-01 -1.62039328e+00 3.66334260e-01 3.86644959e-01
3.77469569e-01 4.14074272e-01 6.90279186e-01 -7.84780443e-01
-1.03077960e+00 -8.10375333e-01 8.23313296e-01 -3.24724853e-01
8.73610139e-01 -7.02014446e-01 -1.09156311e+00 4.70861107e-01
4.48071733e-02 2.33175661e-02 9.13208485e-01 4.65135068e-01
-7.93628022e-02 -4.38051336e-02 -9.85178769e-01 1.33661881e-01
2.97325820e-01 -3.27016890e-01 -2.85800844e-01 -1.73375979e-02
3.99854928e-01 4.13925461e-02 -8.24310482e-01 2.24684790e-01
5.79964221e-01 -1.09318686e+00 9.02606666e-01 -4.18645173e-01
4.83693153e-01 7.22980127e-02 -8.08431953e-02 -1.64000130e+00
2.84479503e-02 -3.70725155e-01 -6.19723201e-01 1.27957368e+00
5.27928591e-01 -6.14967525e-01 8.39122653e-01 6.38033867e-01
2.25470752e-01 -5.08517265e-01 -1.13267708e+00 -8.96833837e-01
2.74840504e-01 -6.70912206e-01 5.38790643e-01 4.02257323e-01
-7.80462548e-02 3.25953275e-01 -7.18930006e-01 1.18748518e-02
8.84631574e-01 -2.54273772e-01 7.86466718e-01 -1.49619818e+00
-5.74766517e-01 -1.51376367e-01 -4.95173812e-01 -9.58909452e-01
3.91921341e-01 -4.60473686e-01 1.38633311e-01 -1.47740209e+00
2.00774401e-01 -1.97007880e-01 -5.81618510e-02 -1.28215343e-01
-2.06220716e-01 -7.08356872e-02 -1.69063032e-01 -1.22661795e-02
-9.59546790e-02 8.86901975e-01 7.97043562e-01 2.66674459e-01
-1.29447937e-01 4.42975074e-01 -8.72400999e-02 7.21553028e-01
8.05562079e-01 -4.09507394e-01 -3.48359942e-01 -9.97541472e-02
3.76284182e-01 1.12594396e-01 5.10624886e-01 -9.69415605e-01
1.10318534e-01 -9.89746004e-02 6.62244260e-01 -8.34653556e-01
5.19039810e-01 -8.00103545e-01 1.88115999e-01 -1.34378478e-01
1.71257466e-01 -3.71124819e-02 3.55429083e-01 7.77029753e-01
-5.27061760e-01 -5.14753401e-01 4.34234142e-01 5.23535199e-02
-2.47649223e-01 2.08729416e-01 -5.21260500e-01 -4.06276388e-03
7.80755043e-01 2.00823113e-01 -1.56071410e-01 -7.26919472e-01
-8.72036934e-01 -2.60619894e-02 2.64503615e-04 -2.53405739e-02
5.55248737e-01 -1.26590514e+00 -6.94830775e-01 3.48072290e-01
-4.88885716e-02 7.52253383e-02 4.15519238e-01 8.53318393e-01
-3.16326380e-01 2.63106018e-01 -7.56556094e-02 -7.36468315e-01
-6.55723393e-01 4.94907469e-01 2.07296282e-01 -5.43583214e-01
-3.72918814e-01 6.58812642e-01 1.38956189e-01 -5.18940508e-01
1.23037189e-01 -1.11401543e-01 -3.72547805e-01 1.90115437e-01
6.78819656e-01 2.25379854e-01 -2.08813742e-01 -4.18137819e-01
2.25020163e-02 1.82543978e-01 -4.19810601e-02 -2.44701385e-01
1.68572891e+00 -4.61854152e-02 -2.49846414e-01 9.28658068e-01
1.38394356e+00 -2.17922255e-01 -1.32732666e+00 -1.24443173e-01
1.64592460e-01 -3.53761107e-01 -5.81520721e-02 -7.21299589e-01
-6.98734939e-01 1.13405907e+00 2.84753323e-01 6.86387718e-01
8.44564319e-01 3.13079268e-01 3.08203071e-01 3.95066619e-01
7.17996508e-02 -9.16295111e-01 -9.40421373e-02 5.43694258e-01
7.00948119e-01 -1.29277122e+00 -1.46816671e-01 6.23943284e-02
-5.05444705e-01 1.11304009e+00 3.57761323e-01 -2.38879114e-01
1.17767370e+00 6.97450042e-01 -3.15509677e-01 1.81265250e-02
-6.14078164e-01 -2.69995600e-01 4.74651605e-01 1.09758496e+00
3.63030106e-01 -7.82805262e-04 1.20541096e-01 6.46211743e-01
-2.40552723e-01 2.22292230e-01 4.71231133e-01 6.08667672e-01
-3.89798254e-01 -1.26047981e+00 -4.36713099e-01 7.60401368e-01
-3.80794913e-01 -1.67153105e-01 6.34129941e-02 7.84312189e-01
-1.12498298e-01 3.90043139e-01 6.02566063e-01 1.02122584e-02
-1.79793447e-01 2.76564479e-01 3.25213462e-01 -3.90331835e-01
2.04727307e-01 2.24895105e-01 -1.48704544e-01 -1.99775808e-02
-3.13182265e-01 -6.30751550e-01 -8.92247081e-01 -3.30885291e-01
-3.18361968e-01 3.14198345e-01 1.17679024e+00 1.05839503e+00
3.20015788e-01 3.50010931e-01 3.83231729e-01 -1.05076098e+00
-6.01279378e-01 -1.28893960e+00 -1.10778868e+00 2.92761594e-01
3.10184300e-01 -7.80700326e-01 -4.84147877e-01 6.54648021e-02]
|
[7.241539001464844, 3.9494123458862305]
|
cd003731-c2ae-434d-ae8c-d7d858f46769
|
learning-an-optimizer-for-image-deconvolution
|
1804.03368
| null |
https://arxiv.org/abs/1804.03368v2
|
https://arxiv.org/pdf/1804.03368v2.pdf
|
Learning Deep Gradient Descent Optimization for Image Deconvolution
|
As an integral component of blind image deblurring, non-blind deconvolution removes image blur with a given blur kernel, which is essential but difficult due to the ill-posed nature of the inverse problem. The predominant approach is based on optimization subject to regularization functions that are either manually designed, or learned from examples. Existing learning based methods have shown superior restoration quality but are not practical enough due to their restricted and static model design. They solely focus on learning a prior and require to know the noise level for deconvolution. We address the gap between the optimization-based and learning-based approaches by learning a universal gradient descent optimizer. We propose a Recurrent Gradient Descent Network (RGDN) by systematically incorporating deep neural networks into a fully parameterized gradient descent scheme. A hyper-parameter-free update unit shared across steps is used to generate updates from the current estimates, based on a convolutional neural network. By training on diverse examples, the Recurrent Gradient Descent Network learns an implicit image prior and a universal update rule through recursive supervision. The learned optimizer can be repeatedly used to improve the quality of diverse degenerated observations. The proposed method possesses strong interpretability and high generalization. Extensive experiments on synthetic benchmarks and challenging real-world images demonstrate that the proposed deep optimization method is effective and robust to produce favorable results as well as practical for real-world image deblurring applications.
|
['Yanning Zhang', 'Chunhua Shen', 'Anton Van Den Hengel', 'Zhen Zhang', 'Dong Gong', 'Qinfeng Shi']
|
2018-04-10
| null | null | null | null |
['blind-image-deblurring', 'image-deconvolution']
|
['computer-vision', 'computer-vision']
|
[ 1.35033980e-01 -4.52225149e-01 -2.70284489e-02 -3.09923321e-01
-6.59254909e-01 -3.20852667e-01 3.08515370e-01 -9.05470014e-01
-2.77571768e-01 8.02556753e-01 4.03941005e-01 -1.97911680e-01
-1.19247273e-01 -1.82708219e-01 -7.78002381e-01 -9.47751343e-01
1.72512338e-01 1.31851032e-01 -5.90874851e-02 -9.03407186e-02
4.64311063e-01 2.24100485e-01 -1.05409920e+00 -7.46790171e-02
1.26736569e+00 8.40874374e-01 5.84727168e-01 6.90046310e-01
1.45870477e-01 9.42820966e-01 -4.57822919e-01 -4.45549414e-02
4.09238309e-01 -7.53429651e-01 -5.65239668e-01 4.92493391e-01
3.33845884e-01 -6.60322607e-01 -6.03825867e-01 1.32848608e+00
7.16093898e-01 2.47903824e-01 5.64366758e-01 -5.58847189e-01
-1.33387327e+00 2.81629413e-01 -6.65776014e-01 3.82500380e-01
1.12444557e-01 3.93081129e-01 8.01161468e-01 -1.18713737e+00
3.13524872e-01 8.80716860e-01 9.31892037e-01 7.48411417e-01
-1.27435338e+00 -3.37022632e-01 -1.32909149e-03 4.15152192e-01
-1.24062896e+00 -6.63202286e-01 9.15214837e-01 -4.68970001e-01
5.65948904e-01 1.53927788e-01 3.94905627e-01 8.49233568e-01
-4.40971926e-02 6.98948085e-01 1.28449988e+00 -3.24652702e-01
1.77956954e-01 1.74560007e-02 9.50118080e-02 7.36506701e-01
1.68544367e-01 8.78323317e-02 -3.49155486e-01 6.25658222e-03
1.02376163e+00 1.37900382e-01 -1.18019545e+00 -3.62164170e-01
-1.17900312e+00 7.18543410e-01 7.90392160e-01 1.03610553e-01
-4.97925907e-01 2.75448561e-01 1.27591908e-01 3.36563051e-01
5.80563843e-01 5.43463230e-01 -3.97078633e-01 1.54981896e-01
-1.11227345e+00 -2.16152817e-02 7.68804133e-01 4.56917316e-01
9.67857838e-01 3.27170670e-01 -2.06127122e-01 1.15320480e+00
3.69305372e-01 4.67290282e-01 8.44650924e-01 -9.61391866e-01
2.53515631e-01 2.36726448e-01 4.73386914e-01 -9.09582615e-01
-2.48519797e-02 -6.03493869e-01 -1.04754531e+00 4.10438806e-01
2.19815940e-01 -2.96090454e-01 -1.14228988e+00 1.55322707e+00
1.70286357e-01 5.77161789e-01 7.68447341e-03 1.61662209e+00
5.66063881e-01 6.63906872e-01 -5.61697066e-01 -3.00699234e-01
9.43282247e-01 -1.39934289e+00 -8.95235777e-01 -5.49134254e-01
2.30643332e-01 -7.37377644e-01 1.11635554e+00 2.99792528e-01
-1.04707968e+00 -3.70828182e-01 -1.25296974e+00 -1.26603141e-01
8.88799503e-02 5.08300424e-01 4.15033937e-01 5.52132189e-01
-1.15010190e+00 6.50162041e-01 -8.11789930e-01 1.29009411e-03
4.90864217e-01 2.51068562e-01 -1.34437203e-01 -2.49076813e-01
-1.01110375e+00 9.09303904e-01 1.69058025e-01 7.28132546e-01
-1.14915478e+00 -6.71279073e-01 -8.68531048e-01 -4.68227565e-02
2.09239364e-01 -8.84260297e-01 1.09390557e+00 -1.25041294e+00
-2.03486347e+00 5.97924531e-01 -2.01845646e-01 -4.57422316e-01
6.00026965e-01 -6.47583246e-01 -1.65555373e-01 1.79175556e-01
-6.38650805e-02 3.06466725e-02 1.48780227e+00 -1.50776887e+00
-1.79947034e-01 -7.93283992e-03 -7.01012462e-02 3.07280600e-01
-4.17336762e-01 -1.95716083e-01 -5.39675176e-01 -9.12540913e-01
2.46045336e-01 -8.48512590e-01 -3.25031161e-01 3.67310755e-02
-2.75952160e-01 4.25105423e-01 8.88423026e-01 -1.11883545e+00
1.30181050e+00 -1.92747045e+00 3.99158090e-01 -1.60623327e-01
2.42172509e-01 4.50409770e-01 -1.34142399e-01 -4.28190939e-02
-7.43265375e-02 -2.73800790e-01 -7.34581530e-01 -3.91328722e-01
-2.05930784e-01 2.07522810e-01 -4.36263502e-01 9.04495537e-01
2.78463475e-02 8.34076285e-01 -1.01322126e+00 -1.82554834e-02
3.75213951e-01 6.79424942e-01 -6.43940866e-01 7.28061140e-01
-1.40335962e-01 6.95155978e-01 -2.25590810e-01 3.52498233e-01
7.98296750e-01 -6.59735620e-01 -6.71991110e-02 -5.74072659e-01
-7.37304240e-02 9.65865180e-02 -1.20779693e+00 1.71758378e+00
-6.65892363e-01 6.98429883e-01 3.72059137e-01 -1.27384079e+00
6.04905546e-01 2.55644321e-01 5.35103269e-02 -2.87354201e-01
9.83146057e-02 4.42053139e-01 -1.74734846e-01 -8.99916410e-01
2.48757258e-01 -1.67684719e-01 6.11005366e-01 6.49977744e-01
-1.39136612e-03 -2.18088597e-01 -1.89255208e-01 -5.21947406e-02
8.76733959e-01 2.12729946e-01 1.19722813e-01 -2.50813961e-01
8.77142608e-01 -2.18057796e-01 6.06890619e-01 6.84171140e-01
4.23733965e-02 1.23476708e+00 -1.27334250e-02 -4.95152444e-01
-1.00725102e+00 -6.87122047e-01 -1.32761076e-01 7.87728488e-01
4.95525151e-01 1.68870747e-01 -8.75806987e-01 -6.00316226e-01
-2.67331481e-01 4.61765707e-01 -5.27052343e-01 -1.94761887e-01
-7.54394889e-01 -1.29344511e+00 7.17297010e-03 1.20903261e-01
8.65096867e-01 -8.03507686e-01 -3.89236093e-01 2.19433650e-01
-2.66021609e-01 -1.04507887e+00 -1.02523422e+00 -3.54218110e-02
-8.94293189e-01 -9.60555732e-01 -1.36620283e+00 -9.61767733e-01
1.05676579e+00 5.88912547e-01 8.61248553e-01 2.26941362e-01
-1.76455259e-01 3.26880991e-01 -1.78406555e-02 3.42044145e-01
-4.09997284e-01 -1.71431780e-01 7.35519752e-02 4.53343451e-01
-1.36301607e-01 -7.14114547e-01 -1.00963354e+00 3.94870847e-01
-8.10089707e-01 1.06754437e-01 8.15724730e-01 1.50461006e+00
3.61171752e-01 -1.16103545e-01 5.83335698e-01 -7.08173633e-01
8.55435610e-01 -4.36132103e-01 -6.47784173e-01 3.15444350e-01
-9.18182790e-01 4.29403424e-01 6.35154724e-01 -6.46177649e-01
-1.34940684e+00 -9.44012254e-02 2.70617545e-01 -7.99094081e-01
2.80722916e-01 6.24501824e-01 4.43444252e-02 -4.69718516e-01
8.54787230e-01 5.08641839e-01 1.86104178e-01 -6.24529660e-01
4.85218853e-01 6.75819457e-01 8.36780965e-01 -4.75517124e-01
9.06312823e-01 5.01598716e-01 -4.35084879e-01 -6.87514305e-01
-9.30438459e-01 -4.50564653e-01 -3.43027562e-01 -1.21787652e-01
6.36624753e-01 -9.84981358e-01 -5.39588928e-01 1.00379074e+00
-1.18394876e+00 -6.23459935e-01 6.12776317e-02 5.69359720e-01
-4.45198238e-01 7.66712070e-01 -8.27804208e-01 -6.14078343e-01
-5.71308434e-01 -1.25775504e+00 6.56362474e-01 4.22666460e-01
3.26142699e-01 -1.25142777e+00 3.90638858e-02 4.76659238e-01
7.55891979e-01 -1.58253193e-01 5.05345285e-01 -1.40191317e-01
-7.59802401e-01 -7.44177625e-02 -4.85288769e-01 8.06370795e-01
6.05262160e-01 -5.21917880e-01 -9.73164678e-01 -5.59412241e-01
5.60122073e-01 -3.69728416e-01 1.17268825e+00 5.70947289e-01
1.20992100e+00 -6.27178729e-01 -7.34932572e-02 1.11443520e+00
1.50209332e+00 -3.99767429e-01 5.85381985e-01 3.53954196e-01
8.96162212e-01 1.91999108e-01 2.76384391e-02 2.41064131e-01
2.37646282e-01 5.29970109e-01 4.70370471e-01 -1.74520716e-01
-5.39259911e-01 1.44567162e-01 5.27973056e-01 9.36268032e-01
-2.15895116e-01 1.07247189e-01 -7.06668675e-01 6.61286950e-01
-1.94806552e+00 -8.50564241e-01 6.79777935e-02 2.26476169e+00
1.37522316e+00 -1.49258822e-01 -4.42052782e-01 -2.14802399e-01
8.97244751e-01 2.74544746e-01 -7.99155474e-01 9.56631154e-02
-1.97276503e-01 -1.16960339e-01 5.60900390e-01 8.30644190e-01
-1.01372695e+00 8.66444051e-01 6.09317493e+00 5.24638474e-01
-1.42154908e+00 2.00076401e-01 5.11606932e-01 5.54988114e-03
-1.85196057e-01 -4.94942218e-02 -2.86210209e-01 5.86296678e-01
4.59562033e-01 -1.07854180e-01 1.13716185e+00 5.33933580e-01
5.12566328e-01 3.64190564e-02 -9.08108950e-01 1.26059902e+00
3.01513195e-01 -1.54147232e+00 -1.42020956e-01 -3.05924743e-01
1.20342350e+00 1.82682350e-01 1.94299757e-01 -8.85260850e-02
2.73801446e-01 -1.04133701e+00 5.96208572e-01 7.84641445e-01
6.79989934e-01 -2.13539883e-01 7.10098147e-01 2.21136823e-01
-6.32971823e-01 -9.61420462e-02 -2.70704895e-01 3.71731911e-03
2.59699136e-01 8.60364199e-01 -5.29873431e-01 3.99400055e-01
7.56551027e-01 1.06698835e+00 -2.74888068e-01 1.27082205e+00
-5.65505505e-01 8.40037525e-01 -2.07046308e-02 4.02945220e-01
1.60425812e-01 -6.25483990e-01 8.03013206e-01 1.26538455e+00
3.56114149e-01 4.43360023e-02 -1.20772654e-02 1.03000104e+00
-2.21901581e-01 -3.32114577e-01 -1.32697389e-01 1.92017287e-01
1.66747957e-01 1.26604474e+00 -2.16786802e-01 -2.43844539e-01
-4.44589943e-01 1.51805544e+00 3.36865693e-01 9.53707039e-01
-6.98577344e-01 -3.45983684e-01 6.76887035e-01 -1.92329213e-01
5.82818687e-01 -2.88158983e-01 -3.77132833e-01 -1.58313334e+00
1.99291766e-01 -1.00791013e+00 -5.27014211e-03 -9.17146385e-01
-1.57810938e+00 5.99779010e-01 -4.23784316e-01 -1.20580149e+00
-1.58694297e-01 -6.22143745e-01 -8.13835502e-01 1.27001619e+00
-2.03573084e+00 -9.12061036e-01 -6.03655994e-01 5.74547231e-01
6.32598341e-01 -1.46653786e-01 3.94073457e-01 3.28231990e-01
-8.30046535e-01 3.74237120e-01 4.33341473e-01 2.41137352e-02
8.51280689e-01 -1.29450583e+00 6.87767342e-02 1.17665732e+00
-2.24601835e-01 8.39015722e-01 8.90499175e-01 -4.26631927e-01
-1.51705277e+00 -1.03854954e+00 3.24083775e-01 -1.33984059e-01
7.53062844e-01 -2.02985276e-02 -1.31763494e+00 5.16621590e-01
3.88601303e-01 5.85143805e-01 7.98486769e-02 -4.24553812e-01
-1.95955485e-01 -1.78844303e-01 -9.46574390e-01 5.64835131e-01
8.05570543e-01 -6.18132353e-01 -6.54218078e-01 4.31428522e-01
5.80697119e-01 -8.04977775e-01 -4.70901221e-01 2.10627541e-01
2.40620255e-01 -9.42378819e-01 1.06504083e+00 -4.10661429e-01
5.32092690e-01 -6.37941897e-01 1.44025922e-01 -1.52452099e+00
-3.05522263e-01 -1.06692183e+00 -5.04048645e-01 9.57208395e-01
2.72964299e-01 -8.54325712e-01 4.94571120e-01 5.50533831e-01
-3.61089379e-01 -7.71080077e-01 -6.05712891e-01 -6.64634407e-01
-2.27743953e-01 -2.96669211e-02 2.90746719e-01 9.37507033e-01
-4.96561378e-01 2.63702065e-01 -8.82623971e-01 6.31149530e-01
9.87397552e-01 2.14473650e-01 6.08195186e-01 -6.21354163e-01
-4.14144903e-01 -4.41452324e-01 -2.31522266e-02 -1.63253570e+00
1.57184616e-01 -7.05785513e-01 4.87069041e-01 -1.65394664e+00
1.75104111e-01 -4.05626267e-01 -2.46584669e-01 2.50397027e-01
-6.63064480e-01 3.52181107e-01 -1.83975309e-01 7.99913406e-01
-2.55578369e-01 8.23698521e-01 1.40484107e+00 -2.01115742e-01
-2.74464130e-01 3.80108915e-02 -5.70394695e-01 7.73508430e-01
6.00406528e-01 -3.19900692e-01 -4.11210269e-01 -8.89374793e-01
7.90997893e-02 -1.43843427e-01 6.10855281e-01 -6.78787708e-01
5.13635635e-01 -4.29020077e-02 2.36686453e-01 2.93293893e-02
1.40734434e-01 -6.29431188e-01 -2.12099887e-02 3.14445406e-01
-2.69420624e-01 -4.35206562e-01 -1.43252045e-01 6.65441632e-01
-2.35325724e-01 -4.42624092e-01 9.05111730e-01 -1.21125922e-01
-8.31325173e-01 3.26235443e-01 6.44377023e-02 6.35365099e-02
3.85989338e-01 -2.01924548e-01 -2.40150198e-01 -5.89493811e-01
-5.13390899e-01 1.49816275e-01 5.15789807e-01 1.29361972e-01
7.04426348e-01 -1.26754856e+00 -8.48005474e-01 1.71549425e-01
-2.45661542e-01 -5.74769750e-02 1.51617035e-01 1.04851055e+00
-6.16799235e-01 -1.11611642e-01 8.19902346e-02 -5.42998731e-01
-6.48813486e-01 5.09414077e-01 8.94476831e-01 1.29820453e-02
-8.06601822e-01 8.35238159e-01 2.68634945e-01 -3.14375132e-01
2.76523471e-01 -9.99809727e-02 -5.41039631e-02 -3.95471245e-01
6.79720819e-01 3.55481923e-01 -2.64006332e-02 -4.78080124e-01
-7.89684504e-02 7.32617974e-01 3.78214009e-02 -7.44469538e-02
1.46107745e+00 -5.22220075e-01 -4.19973820e-01 4.78265956e-02
1.24633491e+00 5.23147406e-03 -1.83635414e+00 -5.69802105e-01
-1.03616394e-01 -5.87368608e-01 5.13025165e-01 -7.33146489e-01
-1.41086793e+00 6.20553493e-01 6.79955244e-01 -4.49528955e-02
1.25253522e+00 -2.97142327e-01 9.27175820e-01 3.78594339e-01
-8.62497017e-02 -9.07466710e-01 4.04962122e-01 3.94209981e-01
1.19977033e+00 -1.55144083e+00 1.84093505e-01 -1.02113128e-01
-4.14588898e-01 1.26445973e+00 4.77679312e-01 -2.78771102e-01
6.02302074e-01 -7.93934315e-02 3.23832303e-01 -8.62630382e-02
-2.83746600e-01 2.96270400e-02 4.93752390e-01 4.24781650e-01
2.53841758e-01 -4.07839894e-01 -2.15730593e-01 4.57748502e-01
2.82645643e-01 2.89763689e-01 5.33984601e-01 5.59849024e-01
-4.30638701e-01 -6.61996126e-01 -5.30663788e-01 1.83071390e-01
-3.83889765e-01 -2.91165948e-01 2.92609006e-01 2.13307530e-01
-2.05387130e-01 7.73394823e-01 -2.46898249e-01 2.19533220e-02
-8.19599852e-02 -3.28474224e-01 4.93386209e-01 -4.59157199e-01
-3.43967915e-01 2.31409771e-03 -2.76414901e-01 -4.02222991e-01
-5.68157494e-01 -4.79165643e-01 -1.04006195e+00 6.90849684e-03
-5.23495197e-01 8.81396979e-02 4.92038757e-01 9.96246576e-01
3.49856704e-01 4.10277277e-01 9.28461730e-01 -1.21532190e+00
-9.00231421e-01 -9.46145415e-01 -3.02241683e-01 3.96729857e-01
1.05671287e+00 -4.67581064e-01 -8.38184357e-01 4.96812850e-01]
|
[11.621603965759277, -2.655325412750244]
|
0464581d-a27a-48e0-8f33-f7dfc1842d82
|
assessment-of-a-cost-effective-headphone
|
2207.12899
| null |
https://arxiv.org/abs/2207.12899v1
|
https://arxiv.org/pdf/2207.12899v1.pdf
|
Assessment of a cost-effective headphone calibration procedure for soundscape evaluations
|
To increase the availability and adoption of the soundscape standard, a low-cost calibration procedure for reproduction of audio stimuli over headphones was proposed as part of the global ``Soundscape Attributes Translation Project'' (SATP) for validating ISO/TS~12913-2:2018 perceived affective quality (PAQ) attribute translations. A previous preliminary study revealed significant deviations from the intended equivalent continuous A-weighted sound pressure levels ($L_{\text{A,eq}}$) using the open-circuit voltage (OCV) calibration procedure. For a more holistic human-centric perspective, the OCV method is further investigated here in terms of psychoacoustic parameters, including relevant exceedance levels to account for temporal effects on the same 27 stimuli from the SATP. Moreover, a within-subjects experiment with 36 participants was conducted to examine the effects of OCV calibration on the PAQ attributes in ISO/TS~12913-2:2018. Bland-Altman analysis of the objective indicators revealed large biases in the OCV method across all weighted sound level and loudness indicators; and roughness indicators at \SI{5}{\%} and \SI{10}{\%} exceedance levels. Significant perceptual differences due to the OCV method were observed in about \SI{20}{\%} of the stimuli, which did not correspond clearly with the biased acoustic indicators. A cautioned interpretation of the objective and perceptual differences due to small and unpaired samples nevertheless provide grounds for further investigation.
|
['Woon-Seng Gan', 'Trevor Wong', 'Karn N. Watcharasupat', 'Zhen-Ting Ong', 'Kenneth Ooi', 'Bhan Lam']
|
2022-07-24
| null | null | null | null |
['soundscape-evaluation']
|
['audio']
|
[ 2.17765257e-01 -2.34918118e-01 6.87969804e-01 -3.15627605e-01
-1.04455876e+00 -4.61646199e-01 5.30775711e-02 7.49028325e-01
-6.28794134e-01 4.87715572e-01 1.88563690e-01 -4.34555590e-01
-5.38243651e-01 -4.20174450e-01 -5.70177436e-01 -4.39828336e-01
-1.86955556e-01 -2.51579583e-01 2.25405186e-01 1.65830180e-02
2.17802867e-01 2.26174787e-01 -1.99301934e+00 2.48614654e-01
6.89880848e-01 1.28801739e+00 2.23968029e-02 7.23889232e-01
5.27985441e-03 8.63775332e-03 -1.09951937e+00 -5.15691102e-01
5.20252474e-02 -4.78396565e-01 -3.70939314e-01 -3.41431051e-01
5.11942863e-01 -1.19708516e-01 3.57224047e-01 1.01017916e+00
1.06189871e+00 3.82197946e-01 5.59948921e-01 -1.19168198e+00
-4.70087737e-01 4.62931305e-01 -1.73083767e-01 3.06479931e-01
6.57962441e-01 2.43553281e-01 8.98283660e-01 -6.06193542e-01
8.11548755e-02 7.95913041e-01 7.17200696e-01 1.90583780e-01
-1.44155037e+00 -1.22804391e+00 -3.00102048e-02 1.71824813e-01
-1.66989028e+00 -4.95673776e-01 7.48191833e-01 -5.77379763e-01
8.59909117e-01 7.76985407e-01 7.73404717e-01 9.35662448e-01
1.31654784e-01 -1.87332287e-01 1.63264203e+00 -3.82343292e-01
6.17629886e-01 6.18662059e-01 -5.76342978e-02 -1.23232715e-01
4.68838006e-01 9.70482454e-02 -5.50621569e-01 -2.22725227e-01
5.88688374e-01 -9.90486443e-01 -3.10943276e-01 3.31989497e-01
-5.56367695e-01 4.46653336e-01 -6.13462068e-02 5.38948596e-01
-2.09246725e-01 6.87099546e-02 5.98479927e-01 1.84017390e-01
4.24353212e-01 7.72909522e-01 -3.83262128e-01 -7.51587570e-01
-6.67420089e-01 1.06801741e-01 6.48123443e-01 6.84585810e-01
2.75432378e-01 3.18491459e-01 -1.91913307e-01 1.16235065e+00
1.91889375e-01 6.74629152e-01 -1.33017689e-01 -1.15051854e+00
3.08834016e-01 -2.11488828e-01 5.30036390e-01 -1.16797864e+00
-5.79925179e-01 -5.35826087e-01 -1.66721255e-01 2.07194597e-01
3.87127250e-01 -1.16777949e-01 -3.58275354e-01 1.87904775e+00
-8.23482592e-03 -1.93492860e-01 -3.40153664e-01 9.00428414e-01
6.53488278e-01 3.98018211e-01 6.07210159e-01 -6.17648005e-01
1.39697206e+00 1.74922034e-01 -9.42930102e-01 -1.83399677e-01
2.91150689e-01 -1.06877553e+00 1.70637715e+00 6.14178658e-01
-1.30894005e+00 -9.16378319e-01 -1.05548406e+00 4.02797490e-01
-1.19658664e-01 -2.42355004e-01 1.04677014e-01 1.28313684e+00
-9.66818273e-01 4.36418235e-01 -4.59146500e-01 -1.77368835e-01
-1.46987423e-01 1.89294770e-01 -1.13902502e-01 4.23336655e-01
-1.34421217e+00 7.58622408e-01 -2.39806354e-01 4.02892143e-01
-2.67804891e-01 -9.20126677e-01 -8.17604601e-01 -2.93235350e-02
3.28828990e-02 -1.60051271e-01 1.12964678e+00 -6.15462601e-01
-1.59215164e+00 6.20169878e-01 4.54662517e-02 9.86038670e-02
1.26778692e-01 -3.19358259e-01 -1.07010889e+00 2.61070907e-01
1.82894729e-02 3.44506353e-01 4.55077350e-01 -1.23699307e+00
-3.81705731e-01 -1.56642959e-01 -3.59797537e-01 8.50286484e-02
-1.03938825e-01 3.43782365e-01 1.28043205e-01 -5.86890161e-01
7.98825622e-02 -7.85063267e-01 2.55636007e-01 -3.30231547e-01
-1.15970559e-01 -7.09264204e-02 -1.46649927e-01 -7.07976699e-01
1.67562735e+00 -2.39306760e+00 -7.73099184e-01 5.92577815e-01
-1.17301375e-01 1.39128402e-01 -1.91809610e-02 4.79748011e-01
-2.34966725e-01 2.63953507e-01 -6.71356618e-02 -1.08707704e-01
4.15022194e-01 -3.27472985e-01 -6.14490034e-03 5.45377374e-01
-2.15272516e-01 2.05863103e-01 -7.82628536e-01 -2.18095347e-01
3.15546632e-01 5.32403827e-01 -4.82762605e-01 -1.23796966e-02
3.98354441e-01 3.44578832e-01 1.65156707e-01 2.37169266e-01
8.49452615e-01 6.50312424e-01 -3.09224725e-01 -1.87332854e-01
-5.53809941e-01 3.56749326e-01 -1.27694714e+00 1.19151437e+00
-4.94753152e-01 6.22441113e-01 3.82751673e-01 -2.90896624e-01
1.04381919e+00 5.75324237e-01 3.16972464e-01 -1.09525180e+00
2.46421322e-01 6.88363373e-01 3.84059608e-01 -6.24140799e-01
3.90530199e-01 -5.99131942e-01 -8.32734033e-02 3.76824266e-03
-2.01398686e-01 -7.02999353e-01 -1.13628581e-01 -2.90955693e-01
4.57546949e-01 -1.58042103e-01 -1.02110460e-01 -6.23955011e-01
3.99632215e-01 -7.19861150e-01 3.60455394e-01 7.33081222e-01
-5.49253047e-01 6.26479924e-01 4.72676873e-01 4.27833378e-01
-7.04793990e-01 -1.28543603e+00 -6.79604650e-01 9.77762580e-01
-2.08193734e-01 -3.94671768e-01 -8.04781556e-01 2.46712089e-01
-1.59043208e-01 1.37731469e+00 -5.06262183e-01 -2.75599062e-01
1.03022726e-02 -2.51293868e-01 8.21398854e-01 6.11646593e-01
9.84752476e-02 -1.07531655e+00 -8.62163246e-01 2.29523480e-01
-2.51644164e-01 -1.11177480e+00 -2.88097084e-01 1.37747198e-01
-2.80247003e-01 -5.47513187e-01 -4.53231901e-01 -3.83230686e-01
2.24087372e-01 -2.34185666e-01 6.57764018e-01 -3.65859658e-01
-1.29087582e-01 8.19247365e-01 -4.12802279e-01 -1.02640438e+00
-3.11364323e-01 -5.94329536e-01 3.78915340e-01 -2.16227382e-01
4.56392139e-01 -8.06721866e-01 -7.77828932e-01 3.91152650e-01
-7.16670036e-01 -5.03768325e-01 1.45717934e-01 5.11404723e-02
5.21671891e-01 1.00103833e-01 7.84158766e-01 -1.50973514e-01
9.88711238e-01 2.72755455e-02 -2.26017475e-01 -2.50693142e-01
-5.62657833e-01 -6.87523901e-01 3.80729616e-01 -4.14287984e-01
-9.77744460e-01 -9.10563350e-01 -6.70849681e-01 -8.87918770e-02
-6.27382696e-01 3.07705939e-01 -2.18399346e-01 2.04412013e-01
7.02717543e-01 -5.60057461e-02 -1.49097025e-01 -2.22510979e-01
-2.17767864e-01 1.05358827e+00 7.99324036e-01 -7.53548741e-01
5.12887895e-01 2.33932827e-02 -9.36020166e-02 -9.52102125e-01
-4.76184756e-01 -2.53385961e-01 -1.48629397e-01 -5.48309505e-01
8.61191452e-01 -7.83013821e-01 -1.00352359e+00 4.59150732e-01
-6.40736341e-01 -3.77882630e-01 -2.87914991e-01 1.06330466e+00
-5.40440977e-01 2.22635373e-01 -3.60716373e-01 -1.29251432e+00
-2.47226104e-01 -1.12807691e+00 7.82085121e-01 1.19963706e-01
-9.02624130e-01 -6.22944176e-01 7.19745585e-04 4.16002363e-01
5.59696913e-01 3.22028011e-01 9.74959433e-01 -6.10892713e-01
5.72912097e-01 -5.16917944e-01 6.31124824e-02 7.09848046e-01
4.29304153e-01 1.15580626e-01 -1.41375077e+00 -1.00178286e-01
4.18058276e-01 -1.08724952e-01 4.92470264e-02 6.08174562e-01
1.06654561e+00 -1.46404520e-01 4.35828328e-01 5.33470772e-02
1.18780065e+00 6.89648032e-01 8.11950266e-01 1.71845809e-01
1.09740578e-01 9.28449750e-01 5.04966319e-01 5.31365395e-01
-7.37035871e-02 5.80595553e-01 1.72192082e-01 4.83117849e-02
-5.37011027e-02 -1.87697470e-01 4.08133179e-01 9.69496191e-01
4.46306393e-02 -2.36345287e-02 -5.83715916e-01 4.85219657e-01
-4.45444494e-01 -6.78799510e-01 -5.73961020e-01 2.80329871e+00
6.65366888e-01 4.40815747e-01 9.37432572e-02 7.58174241e-01
7.74522305e-01 -2.19379608e-02 1.03777230e-01 -1.24701118e+00
1.82689130e-02 7.69086778e-01 -2.67196260e-02 5.58890581e-01
-4.14995700e-01 1.44346431e-01 5.93186331e+00 7.43281901e-01
-1.25544620e+00 -4.00984995e-02 3.81955922e-01 -2.62296021e-01
-6.03777766e-01 -1.20421804e-01 -2.39633471e-01 6.47061586e-01
1.42751515e+00 -1.52371749e-01 1.42555624e-01 3.07347238e-01
7.34680295e-01 -6.04785740e-01 -9.13046718e-01 8.33106995e-01
-1.06550850e-01 -4.77562547e-01 -3.85568678e-01 8.73611122e-02
2.48619542e-01 -6.39934123e-01 5.20612597e-01 1.57258570e-01
-7.29628682e-01 -8.68488789e-01 1.16022944e+00 3.33197355e-01
1.32592773e+00 -6.95220590e-01 7.29948282e-01 -2.55553216e-01
-1.05178726e+00 4.32737805e-02 -1.32474661e-01 -3.93369049e-01
2.36788899e-01 5.96201658e-01 -6.43218517e-01 4.92993832e-01
9.57045019e-01 -3.91328901e-01 -3.79765272e-01 9.28095341e-01
-1.85091972e-01 1.07997656e+00 -3.37834597e-01 -8.50871205e-02
-5.08093871e-02 -6.74346387e-02 5.90539336e-01 1.22803175e+00
4.11568582e-01 1.60920903e-01 -7.67744482e-01 1.02913105e+00
4.39786285e-01 4.89960492e-01 -1.85613886e-01 -2.88953930e-02
6.91507816e-01 9.32827711e-01 -7.22972453e-01 1.20679937e-01
-5.07205904e-01 4.63615239e-01 -5.06933868e-01 5.83407104e-01
-1.00186598e+00 -8.62139106e-01 4.85224128e-01 4.36523885e-01
-9.64960307e-02 8.07863772e-02 -7.96534002e-01 -4.44379114e-02
1.89970762e-01 -7.69923210e-01 6.96154982e-02 -1.01834822e+00
-1.05075979e+00 5.82294166e-01 5.19281507e-01 -1.33955622e+00
1.40136868e-01 -3.49352062e-01 -4.76638794e-01 1.24303126e+00
-8.10414433e-01 -3.96508098e-01 -3.32611799e-01 3.26650172e-01
7.22038466e-03 5.11912942e-01 8.67639184e-01 3.97060454e-01
-2.35765442e-01 9.01884913e-01 -4.68647480e-02 -3.46182942e-01
9.31309462e-01 -1.16004241e+00 -3.67089771e-02 5.54948390e-01
-4.17364210e-01 9.39132035e-01 1.39030516e+00 -4.18750376e-01
-6.40911579e-01 -6.49532914e-01 8.78888786e-01 -3.16913515e-01
6.81017756e-01 -4.55816746e-01 -9.06686842e-01 1.40896246e-01
2.25739300e-01 -5.25529444e-01 1.17322469e+00 5.99285541e-03
-1.41888082e-01 -4.42838728e-01 -1.21535659e+00 7.23938882e-01
7.07379520e-01 -6.89023972e-01 -3.46790582e-01 -2.54793018e-01
7.09427357e-01 -1.97122484e-01 -1.37729299e+00 5.00095189e-01
8.27236593e-01 -1.10345745e+00 6.00639582e-01 1.24339074e-01
1.28897056e-01 -1.90851897e-01 -2.62032151e-01 -1.26170349e+00
-3.40885907e-01 -7.01328158e-01 7.92253911e-01 1.72989798e+00
4.31383312e-01 -8.98349345e-01 -3.70582156e-02 1.07817960e+00
-5.11127174e-01 -4.51272041e-01 -1.32657015e+00 -6.89040720e-01
2.53997177e-01 -1.23968971e+00 3.41098517e-01 3.92463982e-01
2.47432768e-01 -3.46264318e-02 -7.21688047e-02 2.16702700e-01
2.51698554e-01 -6.55097604e-01 3.10627669e-01 -9.44195271e-01
-2.51489371e-01 -6.12383962e-01 -3.24743420e-01 -2.13626683e-01
-4.38925743e-01 -1.34869188e-01 1.28663376e-01 -1.36499357e+00
-3.08974236e-01 -2.61886239e-01 -6.86766148e-01 1.57478657e-02
-2.12332875e-01 1.97671875e-01 3.26767832e-01 -3.29212844e-01
1.31775722e-01 3.64024848e-01 8.58469844e-01 2.01214880e-01
-5.27404368e-01 2.42153540e-01 -9.17331100e-01 4.63249952e-01
5.96862137e-01 -3.40085626e-01 -5.87552965e-01 -1.81224033e-01
4.16884780e-01 1.09848991e-01 2.88500547e-01 -1.25886643e+00
-1.26757234e-01 -1.45633191e-01 1.43091679e-01 -3.54472607e-01
4.88044769e-01 -8.65275443e-01 6.77958906e-01 2.72243887e-01
-4.76630509e-01 3.54501642e-02 8.06488335e-01 5.52458391e-02
5.52575439e-02 -1.53587565e-01 6.92967236e-01 2.29709595e-01
-3.75663210e-03 -4.19961303e-01 -7.21547544e-01 -8.22856575e-02
7.71647513e-01 -6.17080152e-01 -2.36247465e-01 -7.08521366e-01
-5.99126160e-01 -1.87181160e-01 3.33448827e-01 3.92002970e-01
3.54523957e-01 -1.03348649e+00 -6.11759484e-01 4.87485565e-02
2.76851207e-01 -4.30290014e-01 8.45067203e-01 1.30063403e+00
-3.79916310e-01 4.92340475e-01 -1.62704512e-01 -3.06236207e-01
-1.05543983e+00 3.28912884e-01 3.82005960e-01 4.17906404e-01
-2.17478052e-01 1.07350469e+00 1.92660108e-01 2.49595642e-01
4.04176027e-01 -5.14481068e-01 -1.12450205e-01 7.04124346e-02
4.37252373e-01 9.23141718e-01 3.32689136e-01 -6.45133018e-01
-6.81405425e-01 4.23005849e-01 5.53650200e-01 -8.70277166e-01
7.72624493e-01 -3.35735142e-01 8.46565980e-03 1.05855381e+00
1.21589768e+00 6.63383484e-01 -8.20600808e-01 4.86646861e-01
-4.02055383e-01 -4.90256816e-01 -3.28572989e-01 -1.19328868e+00
-5.30894458e-01 1.02610505e+00 1.00667429e+00 1.92529261e-01
1.28507996e+00 -2.76345015e-01 4.88879889e-01 -4.94066596e-01
2.55580246e-01 -1.60488141e+00 -6.10712692e-02 -8.20349082e-02
1.11244977e+00 -2.81450838e-01 -1.73741952e-01 -3.79520297e-01
-7.30379641e-01 7.32884765e-01 6.08916402e-01 4.30303276e-01
5.39938450e-01 2.07916483e-01 3.61322850e-01 1.77932587e-02
-3.44937563e-01 3.85397673e-02 2.40038559e-01 8.16335499e-01
6.97856188e-01 2.36227617e-01 -9.21321392e-01 1.19341755e+00
-7.40166664e-01 -1.86588541e-01 5.34619570e-01 7.20376611e-01
-3.66878003e-01 -5.86772978e-01 -6.69208586e-01 4.92046773e-01
-8.22834313e-01 -1.11936599e-01 -2.66873121e-01 7.50652134e-01
3.61223876e-01 1.63402665e+00 3.87425065e-01 -7.17170238e-01
9.04508650e-01 3.36636186e-01 3.41375172e-01 -3.61650318e-01
-8.66173148e-01 5.92130125e-01 3.16043079e-01 -3.71488333e-01
-3.66249591e-01 -8.99851024e-01 -1.24171543e+00 -6.06687516e-02
-3.29217225e-01 5.53684711e-01 7.80555010e-01 7.60499060e-01
1.13135338e-01 6.82263792e-01 5.82510233e-01 -3.64504457e-01
-3.25650632e-01 -1.31372750e+00 -9.42671895e-01 6.07058167e-01
-2.24700253e-02 -5.90988696e-01 -7.94668317e-01 -2.64336914e-01]
|
[15.139763832092285, 5.62797737121582]
|
9d41bbfc-606e-49fd-82eb-73964e294df6
|
automatic-quantification-of-settlement-damage
|
2010.05512
| null |
https://arxiv.org/abs/2010.05512v1
|
https://arxiv.org/pdf/2010.05512v1.pdf
|
Automatic Quantification of Settlement Damage using Deep Learning of Satellite Images
|
Humanitarian disasters and political violence cause significant damage to our living space. The reparation cost to homes, infrastructure, and the ecosystem is often difficult to quantify in real-time. Real-time quantification is critical to both informing relief operations, but also planning ahead for rebuilding. Here, we use satellite images before and after major crisis around the world to train a robust baseline Residual Network (ResNet) and a disaster quantification Pyramid Scene Parsing Network (PSPNet). ResNet offers robustness to poor image quality and can identify areas of destruction with high accuracy (92\%), whereas PSPNet offers contextualised quantification of built environment damage with good accuracy (84\%). As there are multiple damage dimensions to consider (e.g. economic loss and fatalities), we fit a multi-linear regression model to quantify the overall damage. To validate our combined system of deep learning and regression modeling, we successfully match our prediction to the ongoing recovery in the 2020 Beirut port explosion. These innovations provide a better quantification of overall disaster magnitude and inform intelligent humanitarian systems of unfolding disasters.
|
['Weisi Guo', 'Lili Lu']
|
2020-10-12
| null | null | null | null |
['scene-parsing']
|
['computer-vision']
|
[ 2.07569331e-01 2.87996829e-01 2.98764795e-01 -7.89202452e-02
-8.08447659e-01 -1.30501091e-01 5.55516183e-01 6.22014225e-01
-9.33291256e-01 8.49173605e-01 1.13294840e+00 -5.53962529e-01
-2.11596370e-01 -1.29336512e+00 -3.61643940e-01 -4.47721511e-01
-3.91014636e-01 2.89686769e-01 -2.70573944e-01 -6.88731849e-01
-5.31503372e-02 8.07630420e-01 -8.97504091e-01 -1.76629797e-02
6.27024651e-01 6.30345821e-01 2.47275531e-01 6.79257691e-01
3.29861671e-01 1.22612846e+00 -6.36877060e-01 6.98747560e-02
2.82957435e-01 1.17724791e-01 -1.06065392e+00 -3.69961292e-01
-1.46241456e-01 -1.06191397e+00 -4.81391042e-01 5.53289711e-01
8.46765041e-01 7.06794187e-02 6.50919437e-01 -7.90788710e-01
-5.37745357e-01 6.10022128e-01 -4.65468287e-01 4.79814708e-01
5.71985841e-01 3.60980839e-01 4.56769556e-01 -6.19320631e-01
6.34019971e-01 1.39181995e+00 1.19556367e+00 -1.12341285e-01
-1.11646461e+00 -5.96637607e-01 -1.78405762e-01 1.42181233e-01
-1.25720227e+00 -3.95946860e-01 3.36746693e-01 -6.94355071e-01
1.49050164e+00 1.83064654e-01 3.98969889e-01 7.71817386e-01
3.95987958e-01 7.92623162e-02 8.40410531e-01 -1.65439993e-01
-1.31398707e-03 -6.20297372e-01 -3.86820376e-01 3.98866504e-01
-1.22440262e-02 2.42076144e-01 2.16642488e-02 6.73136264e-02
7.17808545e-01 4.85084802e-01 -1.90076545e-01 5.70583940e-01
-1.02843201e+00 7.34393656e-01 1.27444923e+00 2.83895046e-01
-8.70830953e-01 3.99574876e-01 3.82202834e-01 8.00418761e-03
6.02971017e-01 3.68574619e-01 -2.95421094e-01 1.32936671e-01
-1.20412683e+00 6.52213916e-02 1.78422883e-01 -1.48960546e-01
6.52728677e-01 3.04999888e-01 6.18408658e-02 7.93096602e-01
-6.03656098e-02 8.44652951e-01 -4.91263494e-02 -8.65934253e-01
7.33456790e-01 6.23101711e-01 1.53208286e-01 -1.69634128e+00
-1.41440666e+00 -3.57450962e-01 -1.43400991e+00 2.96158373e-01
1.36635363e-01 -4.52518016e-01 -9.59570110e-01 1.66031218e+00
3.37223634e-02 -2.90083826e-01 -2.93217272e-01 8.26235235e-01
4.60302830e-01 7.12866902e-01 6.77159607e-01 7.79622644e-02
9.24409330e-01 -6.72434047e-02 -3.31300110e-01 -5.57956398e-01
7.11513162e-01 -1.38106972e-01 8.10655057e-01 -1.08908959e-01
-7.92853177e-01 -1.35646477e-01 -7.35243261e-01 1.45503730e-01
-4.10157114e-01 -3.73748034e-01 3.03389549e-01 1.34966761e-01
-1.13262820e+00 8.33627403e-01 -8.71879518e-01 -7.88275659e-01
6.08973801e-01 1.46515429e-01 -6.69486165e-01 -1.26247555e-01
-1.35377300e+00 1.61497772e+00 3.67059141e-01 2.16489509e-01
-7.98491716e-01 -6.52871847e-01 -9.46264207e-01 1.54023275e-01
-1.67800292e-01 -5.28710365e-01 4.82835561e-01 -2.20678419e-01
-2.61270851e-01 8.37152600e-01 6.63363457e-01 -6.48796260e-01
5.62498093e-01 -1.26392841e-01 -2.36930221e-01 1.19765326e-01
5.09248495e-01 6.78293169e-01 8.21253285e-02 -9.58610058e-01
-7.42000639e-01 -6.35723472e-01 -1.46026444e-02 3.82466048e-01
-3.01606834e-01 5.88163674e-01 5.75446606e-01 -7.04612136e-01
7.58794621e-02 -4.89581555e-01 -6.97954357e-01 -2.67011940e-01
-1.93993092e-01 5.05866587e-01 2.33764112e-01 -1.68389463e+00
1.32889986e+00 -1.83163226e+00 8.78612977e-03 -1.70502320e-01
1.67794928e-01 7.56007731e-02 -1.00955687e-01 6.87304139e-01
-1.27042904e-01 3.95820677e-01 -8.32359493e-01 -1.52062431e-01
-1.86907500e-01 2.19329789e-01 -4.56587583e-01 5.70143044e-01
2.11788088e-01 7.46342778e-01 -9.68072832e-01 -2.25746363e-01
5.52680254e-01 6.31470740e-01 -3.57346654e-01 1.72071695e-01
4.96340215e-01 4.97136265e-01 7.32534379e-02 7.06715226e-01
6.27168238e-01 3.06897432e-01 -3.60885039e-02 5.57082817e-02
-3.01530153e-01 8.75041559e-02 -7.12785542e-01 1.05445790e+00
-5.49049556e-01 4.29332495e-01 2.87510157e-01 -8.46711040e-01
8.40154886e-01 1.77689530e-02 8.17690969e-01 -1.17934453e+00
-5.56286685e-02 -3.34058031e-02 -5.60749948e-01 -4.62824672e-01
7.05128372e-01 -6.06284440e-01 -6.21136129e-01 5.44374228e-01
-1.96857557e-01 -1.09027147e-01 -3.79036009e-01 -2.79015559e-03
1.73106110e+00 -2.16733530e-01 3.25848192e-01 -2.91402102e-01
-1.58063114e-01 2.51952976e-01 6.26352072e-01 4.10140574e-01
-4.28469509e-01 7.08828807e-01 2.17840552e-01 -1.12654245e+00
-1.18637252e+00 -1.14731765e+00 -3.36714610e-02 1.09842587e+00
-2.96871364e-01 1.33002117e-01 -5.38716257e-01 -1.95151702e-01
4.51915376e-02 9.16253984e-01 -4.71421301e-01 -1.60344347e-01
-7.98096359e-01 -1.40460742e+00 1.00378752e+00 6.46870852e-01
6.35318160e-01 -1.46112084e+00 -1.24112713e+00 5.19214690e-01
-6.58343315e-01 -8.03671956e-01 1.97812647e-01 2.88749039e-01
-5.18447816e-01 -1.12104082e+00 -6.77256465e-01 -3.76084328e-01
3.54331166e-01 1.85687929e-01 1.13633561e+00 1.06461942e-01
-6.11727834e-01 4.51546192e-01 -1.61928669e-01 -1.38369530e-01
-1.62444338e-01 -7.67188817e-02 3.14778060e-01 -8.75123024e-01
-8.06238875e-02 -8.01213503e-01 -7.53340840e-01 -1.79265556e-03
-9.43324685e-01 -7.82051310e-02 2.91549206e-01 4.93860632e-01
2.90189117e-01 4.61210936e-01 6.51446164e-01 -1.00535601e-01
6.67918265e-01 -9.31610703e-01 2.46313936e-03 -1.50808610e-03
-4.38493639e-01 -3.07359487e-01 3.11731309e-01 1.93696201e-01
-9.93817210e-01 -1.57724600e-02 -3.82256538e-01 3.58433753e-01
-3.53738636e-01 1.06922305e+00 2.53959477e-01 3.68468225e-01
8.47625732e-01 -3.22540492e-01 -4.82733607e-01 -5.09804189e-01
1.79170758e-01 6.71659350e-01 1.15189731e+00 -2.13227555e-01
7.84258425e-01 5.57141244e-01 -9.10285711e-02 -8.88201952e-01
-5.70526719e-01 -4.83014196e-01 -8.38530838e-01 -5.13160110e-01
1.05681586e+00 -1.14091313e+00 -3.26600194e-01 7.39226818e-01
-1.21946001e+00 -8.07081759e-01 -3.59289140e-01 1.66897893e-01
-3.87661427e-01 1.23918109e-01 -3.33126038e-01 -9.69886839e-01
-6.26256704e-01 -4.60305840e-01 8.50148201e-01 -1.40426397e-01
-1.91161975e-01 -8.68509352e-01 4.08829033e-01 2.93412685e-01
8.05382907e-01 1.35296822e+00 7.73394227e-01 1.39027134e-01
6.51515573e-02 -1.04222104e-01 -6.84463799e-01 5.52222244e-02
3.69199187e-01 -4.59576100e-01 -6.14016771e-01 -2.66910195e-01
-9.38955843e-02 -2.54102856e-01 1.23326337e+00 4.12663996e-01
5.11075795e-01 -7.13953316e-01 1.87692158e-02 6.74661696e-01
1.64628589e+00 -1.38897657e-01 1.24772716e+00 9.23547685e-01
6.64715290e-01 9.47592020e-01 2.93704540e-01 7.98146009e-01
8.71397078e-01 4.49389040e-01 8.26257348e-01 -5.76281846e-01
-1.90156579e-01 -2.02535614e-01 4.45514202e-01 1.63993761e-01
-1.88206509e-01 -7.22557157e-02 -1.88788462e+00 8.53514254e-01
-1.55265427e+00 -1.35711968e+00 -1.39195964e-01 2.12211132e+00
4.16436791e-01 2.73063779e-02 1.85796827e-01 1.56392559e-01
5.69605350e-01 4.13962126e-01 -3.33865821e-01 -3.83397818e-01
-4.55142617e-01 1.45054668e-01 1.13790989e+00 5.56813955e-01
-1.24177468e+00 9.23814476e-01 7.00378132e+00 2.73560822e-01
-9.74524856e-01 1.88354328e-01 8.52370620e-01 -8.52093250e-02
-1.87814951e-01 -7.46927783e-02 -3.23227867e-02 1.65471137e-01
1.28585470e+00 4.81844656e-02 6.11381114e-01 3.49116832e-01
6.74453914e-01 -3.01448733e-01 1.10368328e-02 6.27295554e-01
-2.27954015e-01 -1.09088624e+00 -3.76321167e-01 9.50554758e-02
4.30853218e-01 6.68251038e-01 -2.73289889e-01 2.02043176e-01
7.71126151e-01 -1.30718458e+00 7.00992703e-01 9.18559670e-01
1.07294011e+00 -1.02951801e+00 8.92900586e-01 5.52163899e-01
-1.24357688e+00 -6.49302006e-01 -2.82862812e-01 -5.86078048e-01
6.86187327e-01 5.65317988e-01 -8.25722039e-01 2.12574422e-01
1.01058030e+00 3.76605004e-01 -5.20810843e-01 7.58217752e-01
-3.71055663e-01 4.82854456e-01 -5.91342211e-01 8.24842513e-01
2.68433839e-01 1.29932374e-01 3.86020631e-01 1.35469627e+00
3.89871389e-01 5.24041355e-01 1.80154473e-01 5.04705966e-01
-3.90719622e-02 -1.18965670e-01 -6.49710238e-01 3.19399059e-01
1.67867884e-01 9.62636709e-01 -6.06820822e-01 2.41228491e-01
1.17698342e-01 9.20972705e-01 5.01068652e-01 1.53598279e-01
-6.36681318e-01 -2.16177881e-01 6.82164073e-01 4.97195840e-01
-3.78994763e-01 -5.82544267e-01 -4.30174023e-01 -8.04052949e-01
-4.24539089e-01 -3.40804845e-01 5.03801167e-01 -8.91577125e-01
-9.31286156e-01 5.96667111e-01 9.24391970e-02 -8.61598730e-01
-1.91837102e-01 -1.67377457e-01 -7.98478663e-01 8.21763992e-01
-1.47437036e+00 -1.41611123e+00 -3.99409682e-01 3.43764693e-01
8.37331936e-02 2.90094048e-01 9.87245500e-01 3.07432860e-01
-6.53131008e-01 1.67727843e-01 2.23556068e-02 2.18392491e-01
3.67574751e-01 -1.07471073e+00 7.23547935e-01 9.08459604e-01
-7.71895707e-01 5.79225644e-02 5.95703006e-01 -1.10053599e+00
-7.15666652e-01 -1.45376122e+00 1.02806878e+00 -1.93357721e-01
6.36372507e-01 1.33542687e-01 -6.95147634e-01 5.67517281e-01
-1.95807919e-01 -5.64085722e-01 5.04488766e-01 1.31048605e-01
-2.32929230e-01 4.71947379e-02 -1.73032606e+00 3.66428822e-01
8.74842823e-01 -6.04616940e-01 -4.45626646e-01 4.43201870e-01
6.80801690e-01 1.62836373e-01 -9.89426553e-01 5.96909702e-01
5.56050718e-01 -1.03929400e+00 1.23783481e+00 -4.29243952e-01
5.30660629e-01 -4.31387983e-02 -5.77327788e-01 -1.25475430e+00
-8.19469035e-01 9.40488279e-02 3.43953133e-01 1.17275214e+00
1.27787486e-01 -4.54259962e-01 3.08933437e-01 1.02185726e+00
-1.68442026e-01 -3.06460410e-01 -1.17915356e+00 -7.42518663e-01
4.24538612e-01 -6.93827927e-01 8.05194438e-01 9.24620688e-01
-1.12351306e-01 -5.39338030e-02 -6.16961360e-01 4.73456472e-01
6.33687675e-01 -7.17528760e-01 4.09129649e-01 -1.24976611e+00
4.47441757e-01 -5.22404134e-01 -6.67366505e-01 1.57516748e-01
-2.01191351e-01 -5.22731900e-01 -5.45483604e-02 -2.27923250e+00
3.32593560e-01 -4.93315130e-01 -3.72136950e-01 1.06697571e+00
-7.87768140e-02 3.69619161e-01 2.58784801e-01 2.97676235e-01
-7.43696839e-02 6.73598588e-01 3.82177919e-01 -3.07730317e-01
-2.32835993e-01 -3.83332729e-01 -7.06822872e-01 8.58536184e-01
9.97932255e-01 -5.66220999e-01 3.32669020e-01 -8.95833373e-01
5.06684959e-01 2.87934542e-01 8.41090858e-01 -1.48990953e+00
-1.31877549e-02 -6.05492473e-01 3.93039614e-01 -6.51642263e-01
-2.62477761e-03 -7.20960796e-01 4.88897443e-01 7.77754247e-01
-2.37723161e-03 7.53727183e-02 3.14978123e-01 1.51062489e-01
1.99327454e-01 2.45018542e-01 8.54029953e-01 -1.48330420e-01
-5.69937766e-01 2.77249604e-01 -4.65147972e-01 -2.66881198e-01
4.89464611e-01 -9.06330049e-02 -5.23940980e-01 -5.55546582e-01
-7.59478033e-01 4.58751261e-01 7.58586645e-01 1.54957220e-01
6.48617923e-01 -1.20823753e+00 -1.22564459e+00 -3.20443064e-01
-1.88933179e-01 -1.03085972e-01 5.62265635e-01 5.75434744e-01
-1.09333432e+00 7.17465580e-02 -6.60306811e-01 8.52370188e-02
-6.41793430e-01 4.64068204e-01 3.50671738e-01 -6.30877316e-01
-6.25529110e-01 6.69366360e-01 -1.50624931e-01 -7.78964281e-01
-4.23729327e-03 -1.79265723e-01 -4.08132732e-01 4.08092052e-01
8.23656619e-01 8.92445803e-01 -1.27038449e-01 -1.22154629e+00
-5.31217933e-01 2.80548424e-01 6.30297005e-01 -2.89666146e-01
2.10263133e+00 -5.01066625e-01 -2.45879918e-01 -4.40031588e-02
9.83752549e-01 -4.97513592e-01 -1.28762388e+00 2.63494067e-02
3.61065157e-02 -1.30673051e-01 5.76132774e-01 -1.12256241e+00
-1.02357101e+00 8.32978308e-01 8.90227616e-01 4.50457186e-02
1.58222008e+00 -2.32170448e-01 1.13220131e+00 3.34121436e-01
6.66355073e-01 -1.16170359e+00 -3.28156799e-01 4.93830979e-01
1.48551607e+00 -1.14562082e+00 2.24344581e-01 5.13015270e-01
-4.81281847e-01 7.50325024e-01 2.13864386e-01 -1.01529114e-01
5.67534029e-01 2.18098626e-01 -1.05440006e-01 -2.80325562e-01
-6.51355982e-02 -4.51119989e-01 -3.21903497e-01 9.29892421e-01
-8.86932984e-02 5.80519915e-01 1.04649164e-01 5.44013083e-01
-2.57375985e-01 -2.79300719e-01 4.50460374e-01 7.87317216e-01
-9.37936068e-01 -5.54258764e-01 -7.61429846e-01 5.26166975e-01
-3.21348101e-01 -2.77723730e-01 -3.40078115e-01 5.31992555e-01
2.66113639e-01 1.14237487e+00 -1.38445953e-02 -6.51294112e-01
4.94504362e-01 -3.94504368e-01 -6.41683117e-02 -3.29190105e-01
-8.05914402e-01 -4.12235618e-01 5.20409681e-02 -6.39991403e-01
-2.36959785e-01 -7.99933791e-01 -1.38514483e+00 -7.80492485e-01
2.08009347e-01 -4.93970066e-01 8.17653954e-01 9.26650882e-01
7.81649202e-02 4.34712768e-01 7.52339303e-01 -1.35498047e+00
-1.25106961e-01 -1.19171679e+00 -6.23505235e-01 3.15971345e-01
4.52001244e-01 -4.16520387e-01 -3.82176548e-01 -2.65788764e-01]
|
[9.50593376159668, -1.2932251691818237]
|
14e01fa6-4dae-4da0-9e5d-cca641e5f806
|
distributional-instance-segmentation-modeling
|
2305.01910
| null |
https://arxiv.org/abs/2305.01910v1
|
https://arxiv.org/pdf/2305.01910v1.pdf
|
Distributional Instance Segmentation: Modeling Uncertainty and High Confidence Predictions with Latent-MaskRCNN
|
Object recognition and instance segmentation are fundamental skills in any robotic or autonomous system. Existing state-of-the-art methods are often unable to capture meaningful uncertainty in challenging or ambiguous scenes, and as such can cause critical errors in high-performance applications. In this paper, we explore a class of distributional instance segmentation models using latent codes that can model uncertainty over plausible hypotheses of object masks. For robotic picking applications, we propose a confidence mask method to achieve the high precision necessary in industrial use cases. We show that our method can significantly reduce critical errors in robotic systems, including our newly released dataset of ambiguous scenes in a robotic application. On a real-world apparel-picking robot, our method significantly reduces double pick errors while maintaining high performance.
|
['Xi Chen', 'Pieter Abbeel', 'Nikhil Mishra', 'Yuxuan Liu']
|
2023-05-03
| null | null | null | null |
['object-recognition']
|
['computer-vision']
|
[ 5.03346801e-01 1.70401365e-01 -1.45146132e-01 -5.89251578e-01
-1.14676058e+00 -7.67967880e-01 1.70061246e-01 -6.49244636e-02
-8.96254256e-02 6.91068649e-01 -6.29069448e-01 -2.54566818e-01
-4.66517061e-01 -4.31496531e-01 -1.03639412e+00 -3.59993011e-01
5.09691760e-02 1.09702325e+00 4.18395847e-01 2.72306621e-01
5.59053838e-01 4.82499421e-01 -1.82550538e+00 2.27081656e-01
1.11223447e+00 1.06997216e+00 5.61506867e-01 6.69741392e-01
1.36700302e-01 3.15778255e-01 -5.53253651e-01 -2.21385226e-01
6.39288545e-01 1.60112172e-01 -6.83279216e-01 3.92828166e-01
4.60796177e-01 -2.93733180e-01 1.85174078e-01 1.32263589e+00
3.32943425e-02 5.29064834e-02 8.85340810e-01 -1.59052145e+00
-4.31108594e-01 8.32331359e-01 -5.02102017e-01 -3.93346936e-01
-9.16140750e-02 4.26511794e-01 7.24478304e-01 -9.03732002e-01
3.14696670e-01 1.47649944e+00 7.13026404e-01 2.74398893e-01
-1.38145792e+00 -7.95987248e-01 2.26957008e-01 -1.41412139e-01
-1.37409639e+00 -2.42741585e-01 3.07082832e-01 -7.30267227e-01
7.19000697e-01 2.72574779e-02 2.70683378e-01 7.97927260e-01
5.65369546e-01 9.21629846e-01 1.13039994e+00 -1.02583110e-01
3.82075995e-01 -1.14077982e-02 3.31787199e-01 5.71147919e-01
5.34142673e-01 1.91571593e-01 -2.99926609e-01 -1.54975191e-01
8.74307275e-01 9.70658436e-02 1.83635533e-01 -6.32025659e-01
-1.31208730e+00 6.74819231e-01 2.18655214e-01 -2.90079087e-01
-2.10484356e-01 6.24971807e-01 -5.76882660e-02 -2.14264899e-01
2.83868313e-01 8.70221376e-01 -6.11592889e-01 -3.12607199e-01
-8.22433949e-01 4.68085378e-01 8.97083282e-01 1.69822848e+00
6.26759171e-01 -2.71018833e-01 -3.35035622e-01 7.85031259e-01
3.13448429e-01 3.86570513e-01 -7.70434290e-02 -1.36427844e+00
3.00969630e-01 2.84190446e-01 4.95533258e-01 -5.02270877e-01
-2.92590976e-01 -1.97786167e-01 -2.98182338e-01 4.10651982e-01
3.30996096e-01 -3.13933305e-02 -1.57808554e+00 1.12240314e+00
1.09234424e-02 4.66599222e-03 3.62140834e-02 9.40625012e-01
-2.38842349e-02 3.98474485e-01 -8.55030306e-03 -6.93117380e-02
1.13427365e+00 -8.61905456e-01 -7.98183918e-01 -7.25305438e-01
1.22664243e-01 -1.09685457e+00 9.88747001e-01 9.34041023e-01
-9.70824718e-01 -5.74770808e-01 -1.21939886e+00 -6.24626800e-02
-1.13178127e-01 2.84162372e-01 9.29271996e-01 5.06930411e-01
-5.13246775e-01 6.99863315e-01 -1.08300221e+00 4.85686325e-02
5.93261838e-01 5.96123874e-01 4.11580689e-02 -4.57050949e-01
-5.06082773e-01 9.52825904e-01 5.85557044e-01 2.72712976e-01
-1.26724458e+00 -7.55954742e-01 -1.06615996e+00 -1.39865756e-01
1.04631674e+00 -1.04826115e-01 1.82645559e+00 -3.11903179e-01
-1.53833282e+00 3.96761060e-01 9.72353518e-02 -4.81140643e-01
6.19825125e-01 -7.91341543e-01 1.75737426e-01 -1.49705976e-01
2.06939861e-01 1.06699550e+00 1.04403687e+00 -1.74296272e+00
-7.51000881e-01 -2.15078235e-01 -9.49069113e-02 -4.33133729e-02
4.49576467e-01 -2.81336904e-01 -5.45669913e-01 -4.47652340e-01
6.53890193e-01 -1.42419791e+00 -7.69038498e-01 -2.45764386e-02
-6.45933151e-01 -1.17736630e-01 6.90904200e-01 -3.04914296e-01
4.68365490e-01 -2.14256692e+00 -3.97869432e-03 2.55487442e-01
-2.50264853e-01 -2.26549566e-01 1.90537363e-01 -2.12308645e-01
4.35962200e-01 2.02584699e-01 -8.73060405e-01 -1.38008073e-01
3.24595481e-01 7.04773486e-01 -7.01181173e-01 4.65421826e-01
6.28095627e-01 8.12961876e-01 -1.06281769e+00 -4.65599000e-01
3.70998204e-01 -6.83132932e-02 -6.02009118e-01 -1.23034255e-03
-8.08221281e-01 2.02570453e-01 -3.69176209e-01 9.17342842e-01
7.62830496e-01 1.19996756e-01 8.02607462e-02 1.46302089e-01
-9.91052464e-02 4.20288555e-02 -1.30984831e+00 1.97786570e+00
-4.01413500e-01 4.26591545e-01 2.02225640e-01 -4.72774535e-01
8.90577078e-01 -3.05356234e-01 2.47274280e-01 1.80600256e-01
4.14867587e-02 3.57838243e-01 -5.04465811e-02 -1.45439371e-01
1.12527931e+00 -5.41695096e-02 -5.77699184e-01 -1.50246128e-01
-4.68673259e-02 -1.32796013e+00 1.13134548e-01 1.26048559e-02
8.53818476e-01 3.40047181e-01 -3.16193961e-02 -5.94941318e-01
-3.63721460e-01 4.94125068e-01 8.00691843e-01 1.01350534e+00
-2.47194096e-01 8.62251341e-01 5.49646974e-01 5.35680130e-02
-9.37295675e-01 -1.28391755e+00 -4.76382941e-01 5.84869027e-01
7.74246573e-01 5.47122695e-02 -5.57660520e-01 -6.19954824e-01
5.55487752e-01 1.11292493e+00 -1.01580448e-01 -1.07840104e-02
-1.99675471e-01 -6.08913839e-01 2.65338987e-01 7.35873401e-01
1.54657155e-01 -8.65094304e-01 -6.59579337e-01 3.09204996e-01
1.56418785e-01 -1.50188482e+00 -1.81239754e-01 5.75973034e-01
-8.41696739e-01 -1.14087605e+00 -2.97437608e-01 -6.09910905e-01
7.86160052e-01 1.77010268e-01 1.03421891e+00 -3.17859262e-01
-8.00664842e-01 2.66716301e-01 -1.20930530e-01 -8.85388494e-01
-3.85393769e-01 -2.30133176e-01 3.57562274e-01 -5.33596516e-01
2.57877886e-01 -5.87408803e-02 -3.17141376e-02 6.72242045e-01
-7.68486500e-01 -7.54381865e-02 7.35294044e-01 8.95732880e-01
9.33985353e-01 4.51513439e-01 3.71510476e-01 -8.10753465e-01
5.40803432e-01 -3.56294841e-01 -1.31717181e+00 2.40057409e-01
-6.57489419e-01 1.29896075e-01 -5.28238788e-02 -4.86474544e-01
-1.01655126e+00 7.19957232e-01 3.81311774e-01 -6.30392492e-01
-2.15301007e-01 2.88953573e-01 -4.54876423e-02 1.51787266e-01
6.10226512e-01 -3.83483320e-01 -2.41747528e-01 -3.71552199e-01
4.78556901e-01 7.51327813e-01 9.17215943e-01 -1.05623603e+00
5.88972867e-01 3.04205090e-01 1.31485924e-01 -5.43792188e-01
-7.53553450e-01 -5.37519872e-01 -8.25739741e-01 -9.35168713e-02
7.54765689e-01 -9.68099236e-01 -5.64134836e-01 3.28057051e-01
-1.15601289e+00 -6.42543018e-01 -4.07384694e-01 6.38376594e-01
-1.05365884e+00 2.31854722e-01 -4.95348841e-01 -1.18022513e+00
3.40850621e-01 -1.68465984e+00 1.54916143e+00 1.86052710e-01
-1.50478989e-01 -2.21553184e-02 -6.28901422e-01 2.55314827e-01
-4.92008738e-02 3.16779837e-02 6.36642516e-01 -3.34958762e-01
-1.15894258e+00 -1.33545041e-01 -3.43701988e-01 2.63643444e-01
1.36225879e-01 1.86258271e-01 -8.80296111e-01 3.12856166e-03
-4.69723307e-02 -4.36656713e-01 9.60304737e-01 5.62702596e-01
1.55730176e+00 1.03089266e-01 -4.71084297e-01 2.69682586e-01
1.06533706e+00 1.92012534e-01 5.46844065e-01 -2.50158608e-01
3.93430442e-01 8.08940887e-01 1.49914026e+00 3.41982186e-01
-1.37891963e-01 5.92596352e-01 5.33710897e-01 4.04544860e-01
3.49277884e-01 -1.17089175e-01 2.75748283e-01 1.88394532e-01
4.34418648e-01 -2.31755212e-01 -1.21257842e+00 9.20145631e-01
-2.27951384e+00 -3.94565612e-01 -2.00837493e-01 1.99894118e+00
7.90520847e-01 6.32237136e-01 -5.82679927e-01 -1.13278744e-03
8.00094604e-01 -4.69856113e-01 -8.11026692e-01 -2.33864635e-01
3.74736875e-01 -9.70459729e-03 1.07213438e+00 5.85930049e-01
-1.22054088e+00 1.18042374e+00 7.01950550e+00 8.78946900e-01
-3.93652290e-01 -8.60614888e-03 4.54207808e-01 -1.73820019e-01
-1.56580999e-01 -1.15493044e-01 -1.04621899e+00 1.83070824e-01
6.52306437e-01 1.13761209e-01 3.73000383e-01 1.42580020e+00
-3.19400460e-01 -5.68101585e-01 -1.44940138e+00 1.05213165e+00
-1.30174845e-01 -1.07697845e+00 -3.03906739e-01 -4.03882004e-03
1.01832974e+00 8.68174993e-03 4.05100405e-01 2.74398893e-01
1.09092236e+00 -1.14163935e+00 1.07370007e+00 4.99035656e-01
8.38494897e-01 -8.39837193e-01 7.59519994e-01 4.07887310e-01
-5.33970833e-01 -3.19108874e-01 -6.70790434e-01 4.06095013e-02
3.66592556e-01 1.12325311e+00 -1.14373076e+00 2.48138756e-01
7.29418457e-01 2.72545695e-01 -1.57567695e-01 1.10574245e+00
-3.76658648e-01 3.36652189e-01 -6.29106760e-01 1.28803223e-01
1.83359459e-01 1.20261714e-01 5.64693093e-01 9.83859956e-01
3.76672208e-01 1.66654930e-01 5.58423996e-01 1.30181539e+00
7.85966590e-02 -7.18475163e-01 -7.44917393e-01 -1.24310687e-01
4.65995193e-01 9.52118516e-01 -1.17784417e+00 -2.11708322e-01
2.42600903e-01 9.44933236e-01 1.09186098e-02 1.16194263e-01
-8.78202140e-01 -1.87535435e-01 8.40045571e-01 -3.77798736e-01
3.65532696e-01 -8.92098129e-01 -8.30760419e-01 -6.79639399e-01
1.06637970e-01 -5.28901279e-01 -7.48869032e-02 -8.16185296e-01
-1.45887578e+00 4.93581928e-02 3.60847890e-01 -1.10453224e+00
-2.65214473e-01 -1.13458335e+00 1.73083961e-01 5.88019788e-01
-1.28668094e+00 -8.31099629e-01 -1.55487537e-01 -1.22002829e-02
1.11247504e+00 6.65004849e-02 5.48567891e-01 -2.41127014e-01
-2.28867233e-01 9.66222398e-03 6.48354888e-02 -2.82250524e-01
7.73818731e-01 -1.33853185e+00 4.97881293e-01 8.79847765e-01
4.54423204e-02 7.48189867e-01 8.54765356e-01 -1.09247231e+00
-1.46472824e+00 -1.25193727e+00 9.84406471e-02 -8.62838566e-01
3.80778015e-01 -4.82823163e-01 -5.83079755e-01 8.93877327e-01
-4.49049830e-01 8.21934715e-02 2.50137508e-01 2.70794094e-01
-1.45370379e-01 2.62033045e-01 -1.42626143e+00 5.89199066e-01
1.13878381e+00 -3.66139382e-01 -8.37226868e-01 6.36142015e-01
1.05999815e+00 -8.66315424e-01 -7.26189733e-01 9.59104717e-01
5.39282382e-01 -5.25383770e-01 9.03221309e-01 -2.83086956e-01
5.36270261e-01 -3.98583502e-01 -3.92190009e-01 -1.13032317e+00
-1.46245569e-01 -6.29705012e-01 9.76273883e-03 1.04104877e+00
4.78527963e-01 -3.63091260e-01 6.74525380e-01 1.13493061e+00
-6.31057262e-01 -4.36389893e-01 -9.61312473e-01 -1.14536369e+00
-1.17407314e-01 -1.02119982e+00 4.99541938e-01 4.62443560e-01
-7.79048353e-02 -3.16764086e-01 1.26541881e-02 7.12474167e-01
8.98452044e-01 2.11909354e-01 6.96419179e-01 -1.18651760e+00
-2.49685673e-03 -2.35584542e-01 -5.41836143e-01 -1.06612253e+00
2.93259174e-01 -4.49969679e-01 1.27489889e+00 -1.48673391e+00
9.95807275e-02 -1.02466989e+00 -1.03334546e-01 3.54397386e-01
-1.17782414e-01 2.26002172e-01 2.17394650e-01 1.77405164e-01
-5.77927530e-01 4.25531864e-01 8.84356558e-01 -4.70145822e-01
-2.45988503e-01 4.86213230e-02 -5.09478390e-01 1.16108549e+00
8.15224826e-01 -6.42071903e-01 -3.74979734e-01 -4.97054636e-01
-5.33422120e-02 -2.11585879e-01 4.57176864e-01 -1.17857170e+00
2.07458094e-01 -5.98638296e-01 3.61843139e-01 -9.49000478e-01
4.55391169e-01 -1.08002365e+00 4.30525579e-02 4.50792432e-01
-2.45076567e-01 -3.15807313e-01 5.74008346e-01 8.01417530e-01
7.52716437e-02 -4.85033810e-01 6.08712375e-01 -1.08634822e-01
-8.81428421e-01 9.03864354e-02 -4.05335933e-01 -3.48217696e-01
1.28500533e+00 1.73067614e-01 -1.80577189e-01 4.61495146e-02
-5.11971056e-01 5.78663290e-01 5.63884974e-01 7.61827767e-01
7.03466535e-01 -9.95194376e-01 -5.25544405e-01 1.97947010e-01
2.86978662e-01 7.62201786e-01 -4.01634239e-02 3.24271500e-01
-6.84560180e-01 4.23633754e-01 -1.41094908e-01 -1.20773995e+00
-9.13294315e-01 6.24930561e-01 -7.03938454e-02 -2.53190529e-02
-4.01564658e-01 1.12427390e+00 2.38100037e-01 -5.98346829e-01
2.64952213e-01 -1.11913836e+00 4.92314726e-01 -3.91382366e-01
5.35397641e-02 2.80275851e-01 -9.55077186e-02 -1.45902634e-01
-3.24151814e-01 3.86740476e-01 -6.83617732e-03 -4.46903765e-01
9.36364710e-01 8.33164826e-02 1.71889201e-01 6.22469783e-01
3.85499448e-01 -9.20897871e-02 -1.78187609e+00 1.82032853e-01
3.33232671e-01 -7.42542803e-01 -3.26173119e-02 -8.31861496e-01
-4.51510817e-01 8.20877671e-01 5.42416632e-01 -2.03087509e-01
6.16380930e-01 8.38986859e-02 4.98785317e-01 7.84644127e-01
1.12960064e+00 -1.24257410e+00 -2.20927611e-01 6.14545703e-01
1.00446200e+00 -1.45789707e+00 2.14658618e-01 -9.76700544e-01
-8.04346800e-01 8.54007721e-01 7.60464609e-01 -1.09287072e-02
5.39252996e-01 6.86426878e-01 -1.52256384e-01 -1.21935323e-01
-4.91748691e-01 -1.74973264e-01 1.57013491e-01 6.36556149e-01
-1.79794312e-01 3.94078493e-01 9.25357863e-02 8.03219020e-01
-3.15525025e-01 1.39278695e-01 5.99173129e-01 1.25630248e+00
-5.87346792e-01 -7.40884960e-01 -6.65105879e-01 4.58262861e-01
-4.13504452e-01 1.00534834e-01 -2.02873394e-01 6.20631576e-01
1.96552336e-01 1.14499140e+00 3.17654371e-01 -2.87551016e-01
3.09558094e-01 -3.27381864e-03 7.32866883e-01 -1.02532291e+00
-1.00708403e-01 -1.72487825e-01 2.22545475e-01 -7.58459210e-01
-1.25238255e-01 -8.09781015e-01 -1.62283444e+00 3.18148643e-01
-7.30614841e-01 -1.00572608e-01 1.08662069e+00 8.54963303e-01
2.72945166e-01 6.83413267e-01 1.11194305e-01 -1.17991447e+00
-8.21349382e-01 -9.63508427e-01 -7.06539333e-01 2.02231924e-03
1.92150652e-01 -1.10807443e+00 -9.07801464e-02 1.88508168e-01]
|
[6.800088405609131, -1.066758394241333]
|
6afad2f6-4852-431e-afa4-213388fbb5b7
|
leveraging-declarative-knowledge-in-text-and
|
2004.14201
| null |
https://arxiv.org/abs/2004.14201v2
|
https://arxiv.org/pdf/2004.14201v2.pdf
|
Leveraging Declarative Knowledge in Text and First-Order Logic for Fine-Grained Propaganda Detection
|
We study the detection of propagandistic text fragments in news articles. Instead of merely learning from input-output datapoints in training data, we introduce an approach to inject declarative knowledge of fine-grained propaganda techniques. Specifically, we leverage the declarative knowledge expressed in both first-order logic and natural language. The former refers to the logical consistency between coarse- and fine-grained predictions, which is used to regularize the training process with propositional Boolean expressions. The latter refers to the literal definition of each propaganda technique, which is utilized to get class representations for regularizing the model parameters. We conduct experiments on Propaganda Techniques Corpus, a large manually annotated dataset for fine-grained propaganda detection. Experiments show that our method achieves superior performance, demonstrating that leveraging declarative knowledge can help the model to make more accurate predictions.
|
['Daxin Jiang', 'Ruize Wang', 'Xuanjing Huang', 'Nan Duan', 'Ming Zhou', 'Wanjun Zhong', 'Duyu Tang', 'Zhongyu Wei']
|
2020-04-29
| null |
https://aclanthology.org/2020.emnlp-main.320
|
https://aclanthology.org/2020.emnlp-main.320.pdf
|
emnlp-2020-11
|
['propaganda-detection']
|
['natural-language-processing']
|
[ 1.92343503e-01 2.24280462e-01 -8.63229215e-01 -5.30782759e-01
-7.25750208e-01 -7.11518466e-01 1.03068674e+00 4.73894030e-01
-1.69517528e-02 6.81623518e-01 7.32157350e-01 -6.79302573e-01
3.52679417e-02 -1.30555570e+00 -1.21946490e+00 -8.56512040e-02
-1.61928162e-01 3.40783268e-01 5.07190451e-02 -2.75125533e-01
3.83142740e-01 1.15583897e-01 -9.95286703e-01 1.12497342e+00
1.04598200e+00 9.31609035e-01 -4.97688860e-01 4.98342693e-01
-3.69688570e-01 1.91667509e+00 -7.95531154e-01 -6.62694097e-01
-2.33807899e-02 -3.61079305e-01 -9.59273338e-01 -2.18316957e-01
3.71802956e-01 -3.85654479e-01 -7.23167658e-02 1.15782368e+00
-3.61935675e-01 -2.43234888e-01 9.73581791e-01 -5.71866155e-01
-8.49711835e-01 1.48211014e+00 -4.34204996e-01 3.60717416e-01
6.42588794e-01 -5.94231673e-02 1.39702117e+00 -4.40522492e-01
8.90196502e-01 1.53828120e+00 6.81774318e-01 3.97962332e-01
-1.25706863e+00 -4.78657871e-01 2.66285121e-01 1.39515743e-01
-8.28054488e-01 -1.86029822e-01 7.71159232e-01 -9.11490560e-01
8.97431314e-01 2.40092099e-01 6.81235135e-01 1.35206592e+00
4.42906082e-01 1.04458129e+00 9.79528487e-01 -6.40057385e-01
2.21458167e-01 2.64073461e-01 7.62036264e-01 1.30506337e+00
4.45176274e-01 1.87839314e-01 -5.06801844e-01 -5.32098770e-01
5.34312546e-01 -2.09814191e-01 -1.00231813e-02 3.26874763e-01
-8.34584355e-01 1.37860096e+00 4.09948379e-01 2.71280497e-01
-3.79187435e-01 1.49250746e-01 5.81104279e-01 3.11385393e-01
5.49142599e-01 8.26454103e-01 -4.53197420e-01 -2.10272893e-01
-9.95239973e-01 5.60228705e-01 1.03888476e+00 6.87463939e-01
5.41623652e-01 -2.72981167e-01 -4.69960898e-01 4.09363449e-01
4.28756811e-02 5.18417299e-01 2.17631072e-01 -5.40350914e-01
7.23241448e-01 9.87587631e-01 -6.41086847e-02 -1.18668044e+00
-3.02812427e-01 -2.58060485e-01 -6.10349000e-01 -5.30202031e-01
2.64068514e-01 -2.60516584e-01 -8.64673734e-01 1.50926983e+00
1.34429038e-01 -3.31874728e-01 -1.91721290e-01 5.12364507e-01
6.47544622e-01 9.41621661e-01 1.20233834e-01 -1.47012919e-01
1.44698322e+00 -8.05747688e-01 -6.60678327e-01 -3.04066300e-01
1.00424767e+00 -1.92287162e-01 1.23582757e+00 2.34985217e-01
-7.93286324e-01 6.35838881e-02 -1.02201474e+00 2.88961623e-02
-2.26347536e-01 -9.43461210e-02 8.28473389e-01 3.60211879e-01
4.30452116e-02 4.97439831e-01 -7.72807539e-01 2.43214831e-01
6.41713977e-01 -4.08894897e-01 -3.60030346e-02 8.08375701e-02
-1.55784023e+00 8.85385811e-01 7.05725551e-01 -2.84472615e-01
-8.99244964e-01 -9.96450305e-01 -9.50103581e-01 2.25366965e-01
7.17947900e-01 -5.32500088e-01 1.33680952e+00 -8.40795338e-01
-1.29860306e+00 8.04988742e-01 3.22864484e-03 -8.57171297e-01
2.24424422e-01 -5.53975224e-01 -3.15860629e-01 6.12316169e-02
2.54976898e-01 -1.31948277e-01 8.09280634e-01 -9.02926624e-01
-5.43829858e-01 -6.29365966e-02 5.37701070e-01 -3.64125162e-01
-2.72472411e-01 3.37245405e-01 1.25731438e-01 -7.05784380e-01
-1.42399877e-01 -5.24119914e-01 1.61821783e-01 -5.89877725e-01
-1.04030514e+00 -4.10270035e-01 3.49339873e-01 -6.84798300e-01
1.71638608e+00 -1.91527176e+00 -4.83695529e-02 4.21981186e-01
1.76480323e-01 -8.97941664e-02 1.14503093e-01 3.21270317e-01
1.60636634e-01 6.67117655e-01 2.39486545e-01 3.39447826e-01
1.70575574e-01 2.91224986e-01 -1.01370525e+00 3.16819489e-01
4.31507587e-01 1.02552605e+00 -9.07457471e-01 -7.29257166e-01
-1.97833285e-01 -1.25010580e-01 -9.03811455e-01 1.99813902e-01
-9.19554591e-01 -1.07144140e-01 -9.27228808e-01 6.37439132e-01
2.10876107e-01 -5.01398027e-01 2.61740029e-01 -9.65266302e-02
-1.90885160e-02 1.10238361e+00 -6.36236012e-01 9.85769629e-01
-3.68951410e-01 2.98269749e-01 -1.93377405e-01 -8.04382861e-01
3.14565897e-01 -5.71891591e-02 -1.09666750e-01 -7.34418035e-01
2.35211283e-01 -2.45164663e-01 -3.78972054e-01 -5.90427160e-01
2.53963113e-01 -4.21823323e-01 -6.62967861e-01 8.23288739e-01
-3.25563289e-02 -3.68845575e-02 4.37540203e-01 5.95414102e-01
1.27101910e+00 -8.28842148e-02 7.37591386e-01 -2.54029870e-01
3.77193421e-01 6.52079701e-01 6.33302748e-01 1.20115876e+00
3.71804804e-01 -4.07613665e-01 8.37295175e-01 -8.28778386e-01
-6.78788126e-01 -1.21832788e+00 -1.98284239e-01 1.57183945e+00
-2.55549520e-01 -9.93332148e-01 -7.02786803e-01 -1.19380724e+00
-1.07099758e-02 9.41006303e-01 -1.10424566e+00 -9.23390538e-02
-8.94938529e-01 -4.29594338e-01 8.38948369e-01 5.59261143e-01
4.00917888e-01 -8.80512893e-01 -5.79188943e-01 2.41729885e-01
-2.97551781e-01 -8.69856358e-01 -2.24440038e-01 3.30395490e-01
-6.94209456e-01 -1.22891676e+00 3.67245853e-01 -5.29035449e-01
5.16046822e-01 -3.39808673e-01 1.22057319e+00 2.36485064e-01
-1.87113762e-01 -2.37346247e-01 -4.29180652e-01 -4.52563554e-01
-8.80605757e-01 4.38909046e-02 -4.27459292e-02 -3.99158537e-01
2.94676721e-01 -1.28383875e-01 1.12947740e-01 -1.96541190e-01
-7.51403630e-01 2.49345005e-01 4.31446999e-01 1.13751590e+00
3.10833365e-01 1.78975880e-01 2.05991805e-01 -1.62133276e+00
9.57896471e-01 -5.42133927e-01 -6.69260859e-01 5.07887304e-01
-3.89735490e-01 5.04547000e-01 9.63631868e-01 -4.63477612e-01
-1.19815600e+00 -2.39440590e-01 -2.22265199e-02 2.76737779e-01
-3.37405168e-02 1.18457580e+00 1.66740537e-01 4.71740454e-01
1.26921248e+00 1.74311563e-01 -4.98225331e-01 -3.82679760e-01
7.59195447e-01 5.26416779e-01 6.52859032e-01 -1.17006218e+00
9.24391031e-01 2.47280225e-01 -4.28425670e-01 -4.04747754e-01
-1.75516140e+00 -2.57146627e-01 -2.92191297e-01 2.43311480e-01
6.89744294e-01 -6.36659563e-01 -5.17834604e-01 1.16353966e-02
-1.32208955e+00 -3.37504029e-01 -1.41764462e-01 9.68382955e-02
-5.88031709e-01 3.48657276e-03 -1.12963140e+00 -7.25123823e-01
-3.69878531e-01 -5.24705827e-01 8.20033371e-01 -8.66739824e-02
-4.30037886e-01 -1.16741264e+00 1.78553373e-01 1.63019240e-01
-9.71214473e-03 2.85083890e-01 1.76744199e+00 -9.52040732e-01
-2.90775776e-01 -3.96314383e-01 -3.45769785e-02 1.16778046e-01
3.82576026e-02 -3.68668623e-02 -6.19647086e-01 3.14651191e-01
2.00628832e-01 -8.59210908e-01 7.83819914e-01 1.74153358e-01
1.20908844e+00 -1.38613248e+00 -3.83848578e-01 3.66681248e-01
9.22323287e-01 -2.64025807e-01 3.71857524e-01 6.19406462e-01
3.72452885e-01 4.09767151e-01 7.13936090e-01 4.61499631e-01
1.48427859e-01 3.96782577e-01 3.71991359e-02 2.16776803e-01
2.19004691e-01 -9.10817921e-01 4.24845368e-01 2.92467922e-01
-9.15782899e-02 -1.38541525e-02 -9.03484583e-01 1.55688748e-01
-1.66096723e+00 -1.74983943e+00 2.11858734e-01 1.43697155e+00
1.62004220e+00 4.51613575e-01 -1.53598869e-02 3.14777009e-02
4.29011703e-01 2.88814306e-01 4.98294383e-02 -6.69055223e-01
4.90732864e-02 2.97840595e-01 2.92016026e-02 7.00457215e-01
-1.32258081e+00 1.17826784e+00 6.67765951e+00 9.75319326e-01
-1.07897580e+00 1.62022822e-02 2.71031767e-01 -5.56944907e-02
-5.41189253e-01 -5.76950423e-02 -9.58199739e-01 3.45709503e-01
8.92115176e-01 -3.41919780e-01 4.14853632e-01 1.12731504e+00
-3.25469524e-02 1.78845301e-01 -1.39784324e+00 2.36261919e-01
-2.07454693e-02 -2.14266682e+00 5.45970738e-01 -1.45688713e-01
7.83020437e-01 -2.98521787e-01 -2.65080422e-01 6.89323068e-01
7.78668165e-01 -1.16609895e+00 1.11013651e+00 2.16061756e-01
4.78288442e-01 -4.78727907e-01 5.33227563e-01 1.02212608e+00
-6.99218035e-01 -2.95736611e-01 -2.27825984e-01 -3.78571361e-01
2.05816388e-01 8.59985530e-01 -1.22539341e+00 3.32333207e-01
2.56064743e-01 6.30674124e-01 -3.82739991e-01 3.02284271e-01
-8.89993429e-01 1.16449952e+00 -1.76939026e-01 -3.92017752e-01
3.90156180e-01 4.18323308e-01 2.82395244e-01 1.68913901e+00
-3.33294123e-01 3.00503701e-01 4.64107454e-01 1.14523757e+00
-2.57858366e-01 -1.33712878e-02 -4.51742530e-01 -5.34673870e-01
4.46286887e-01 9.16840732e-01 -2.03288585e-01 -5.23929358e-01
-1.36540771e-01 5.27158797e-01 7.68159091e-01 2.10716799e-01
-1.07219625e+00 -2.15770543e-01 2.99542069e-01 4.37785536e-01
1.63830340e-01 -1.54069215e-01 -4.61910993e-01 -1.30005860e+00
-1.11469366e-01 -8.91710401e-01 7.82723069e-01 -2.81351745e-01
-1.47438002e+00 3.55378747e-01 1.96032584e-01 -5.11575997e-01
-6.90355062e-01 -7.18940854e-01 -8.26973379e-01 5.79682827e-01
-1.09277391e+00 -1.15521026e+00 2.10441053e-01 3.01366359e-01
4.63024378e-01 8.03952739e-02 9.31039989e-01 -2.65087545e-01
-2.85662860e-01 6.03899300e-01 -3.46290797e-01 6.74468815e-01
3.88558090e-01 -9.98453617e-01 1.74315333e-01 8.15477312e-01
2.66353637e-01 1.38999712e+00 7.78893530e-01 -1.09489429e+00
-1.34385943e+00 -1.23931050e+00 1.08409452e+00 -6.79425716e-01
1.28531003e+00 -5.01893938e-01 -8.64887178e-01 1.03303111e+00
-1.63499802e-01 -1.57434925e-01 8.60557675e-01 8.68295610e-01
-1.36136329e+00 3.35127264e-01 -8.41826856e-01 3.53734076e-01
9.06397223e-01 -8.29755485e-01 -1.33994949e+00 5.77687204e-01
6.63513541e-01 -5.05304694e-01 -5.47669232e-01 2.74982512e-01
5.98709166e-01 -3.79676819e-01 7.12311983e-01 -1.40899813e+00
1.15291750e+00 1.09065913e-01 -2.50264138e-01 -1.26171911e+00
-6.95035636e-01 -1.84267864e-01 -6.46007240e-01 9.38897550e-01
7.70123243e-01 -2.26412416e-01 5.28260171e-01 2.04085812e-01
-9.35413763e-02 -1.06113720e+00 -6.28838778e-01 -7.55218446e-01
3.62999529e-01 -5.05811393e-01 3.73840213e-01 1.10113633e+00
6.43211067e-01 6.18259490e-01 -2.94155955e-01 1.01256870e-01
3.67031395e-01 6.11553192e-01 5.75122058e-01 -1.07226801e+00
-6.40921414e-01 -4.46744949e-01 -2.02836484e-01 -1.17786336e+00
4.55655724e-01 -1.07963395e+00 3.11530143e-01 -1.57044041e+00
6.82728171e-01 -3.47288579e-01 1.77981168e-01 7.84129381e-01
-5.65152578e-02 -9.46499556e-02 -7.14317784e-02 8.10398012e-02
-7.19522476e-01 1.97156623e-01 9.52813327e-01 -5.43570518e-01
-1.97870687e-01 -1.09368533e-01 -9.99089658e-01 1.12850416e+00
6.52474880e-01 -7.30841577e-01 -3.00503790e-01 -5.58578730e-01
6.83350325e-01 -1.81316316e-01 6.17346823e-01 -5.23977518e-01
2.35686034e-01 -8.87462258e-01 3.40537339e-01 -4.34026003e-01
-2.91093647e-01 -2.03795269e-01 -3.24297041e-01 5.48890471e-01
-8.58187199e-01 -4.03831363e-01 1.10421732e-01 6.67699516e-01
-1.34205714e-01 -2.63502270e-01 6.55220091e-01 -1.49159998e-01
-3.15615654e-01 -1.35728687e-01 -4.87356901e-01 5.50271749e-01
7.31185853e-01 3.22176427e-01 -1.12539947e+00 -6.79320991e-02
-5.65301895e-01 3.91958468e-02 2.87503153e-01 2.48673201e-01
4.17230129e-01 -1.18200612e+00 -7.15712368e-01 1.61660016e-01
2.26867586e-01 -3.04207236e-01 -2.02543333e-01 5.75179935e-01
-3.96639019e-01 4.95064765e-01 9.53237936e-02 -2.44595394e-01
-9.57058907e-01 6.01558924e-01 1.33346349e-01 -5.99715948e-01
-6.83548748e-01 8.83219421e-01 1.81197211e-01 -1.49803504e-01
2.86026537e-01 -5.90008616e-01 -3.32231112e-02 -2.59132951e-01
9.11733449e-01 -1.10882856e-01 -9.54040587e-02 -5.01132980e-02
-4.99292374e-01 -3.81812453e-02 -5.50286651e-01 9.31065679e-02
1.36223209e+00 4.42654967e-01 -3.39681298e-01 4.74721104e-01
8.65487814e-01 6.94759905e-01 -6.62868440e-01 -2.64754266e-01
2.85384208e-01 -3.08377504e-01 2.91236471e-02 -1.35681808e+00
-1.88983724e-01 6.50498688e-01 -4.38084513e-01 2.97671229e-01
5.67658603e-01 5.23109317e-01 5.64863265e-01 9.28642035e-01
5.04803598e-01 -1.03148448e+00 2.47679636e-01 1.16894281e+00
1.06419313e+00 -6.84977293e-01 1.20800912e-01 -7.19217956e-01
-5.06187499e-01 1.02112472e+00 5.37018955e-01 -2.23917603e-01
4.10186946e-01 6.26570225e-01 -1.81303501e-01 -5.17910242e-01
-8.87721539e-01 4.28916752e-01 5.03119349e-01 1.56644464e-01
6.78343832e-01 2.19218880e-01 -2.86712736e-01 1.37035143e+00
-4.80009198e-01 -7.65949339e-02 1.95092499e-01 1.01446354e+00
-8.22016835e-01 -6.69272244e-01 -2.89901942e-01 8.19240808e-01
-6.16016865e-01 -4.67677563e-01 -8.16974103e-01 5.98330379e-01
1.45789146e-01 7.76030362e-01 -1.98408827e-01 -6.20904565e-01
2.50025690e-01 1.50173917e-01 7.01361954e-01 -1.05311501e+00
-8.65871310e-01 -4.36072588e-01 7.38257945e-01 -7.10721731e-01
1.84785854e-02 -2.45716676e-01 -1.59518552e+00 -6.29257441e-01
-2.93582976e-01 5.47254801e-01 2.93296278e-01 1.18177497e+00
3.55282589e-03 2.80200958e-01 4.16137725e-01 -2.65792847e-01
-1.31064069e+00 -9.18602407e-01 -2.56216407e-01 5.55426359e-01
2.85374820e-01 -6.88035846e-01 -3.96494001e-01 3.12581331e-01]
|
[9.740107536315918, 7.992658615112305]
|
51154fc9-8cb1-48f2-9a43-28f66a60cf24
|
190600575
|
1906.00575
| null |
https://arxiv.org/abs/1906.00575v3
|
https://arxiv.org/pdf/1906.00575v3.pdf
|
Jointly Learning Semantic Parser and Natural Language Generator via Dual Information Maximization
|
Semantic parsing aims to transform natural language (NL) utterances into formal meaning representations (MRs), whereas an NL generator achieves the reverse: producing a NL description for some given MRs. Despite this intrinsic connection, the two tasks are often studied separately in prior work. In this paper, we model the duality of these two tasks via a joint learning framework, and demonstrate its effectiveness of boosting the performance on both tasks. Concretely, we propose a novel method of dual information maximization (DIM) to regularize the learning process, where DIM empirically maximizes the variational lower bounds of expected joint distributions of NL and MRs. We further extend DIM to a semi-supervision setup (SemiDIM), which leverages unlabeled data of both tasks. Experiments on three datasets of dialogue management and code generation (and summarization) show that performance on both semantic parsing and NL generation can be consistently improved by DIM, in both supervised and semi-supervised setups.
|
['Hai Ye', 'Wenjie Li', 'Lu Wang']
|
2019-06-03
|
jointly-learning-semantic-parser-and-natural
|
https://aclanthology.org/P19-1201
|
https://aclanthology.org/P19-1201.pdf
|
acl-2019-7
|
['dialogue-management']
|
['natural-language-processing']
|
[ 4.42734241e-01 1.04469442e+00 -2.38314256e-01 -5.64384878e-01
-1.36234486e+00 -8.00273418e-01 9.60992694e-01 -1.16178170e-02
-4.57421727e-02 8.38993847e-01 5.78975201e-01 -4.55561310e-01
2.83327401e-01 -6.69520378e-01 -7.79212117e-01 -4.12909061e-01
3.04280102e-01 5.51745296e-01 -3.06095660e-01 3.22569944e-02
6.30916432e-02 -9.96632352e-02 -1.27705848e+00 2.98190117e-01
9.74961996e-01 6.28428280e-01 4.38824058e-01 8.17997336e-01
-5.66674113e-01 9.58719969e-01 -5.16996503e-01 -5.23251951e-01
-4.67707403e-02 -7.95173764e-01 -1.36446881e+00 2.97399074e-01
1.24625042e-01 -9.77494866e-02 3.27339396e-02 1.03053761e+00
3.20240974e-01 1.26262739e-01 9.10426855e-01 -1.27764046e+00
-6.73922181e-01 8.97539675e-01 -5.45158327e-01 -2.82485276e-01
6.39059424e-01 -4.83557433e-02 1.57264245e+00 -5.78365564e-01
5.88140011e-01 1.56740940e+00 4.12480742e-01 9.38737631e-01
-1.63825369e+00 -2.70812005e-01 1.09857552e-01 -5.50653696e-01
-7.88741350e-01 -6.17077649e-01 5.19535482e-01 -4.73054707e-01
9.78044629e-01 -1.65683001e-01 1.11277670e-01 1.24140918e+00
-2.40594536e-01 1.22211289e+00 8.37010801e-01 -6.18696034e-01
3.23518813e-01 1.89855158e-01 1.91948295e-01 8.19139600e-01
8.04032758e-02 -3.12192321e-01 -5.81404448e-01 -3.17444652e-01
5.35328269e-01 -4.97949600e-01 -2.11657569e-01 -3.87809306e-01
-1.13144135e+00 1.22952485e+00 -2.31847450e-01 6.27989620e-02
-1.98665276e-01 2.02217177e-01 3.35894436e-01 2.03253612e-01
5.51944911e-01 5.56183517e-01 -5.87053418e-01 -2.99066663e-01
-6.12341046e-01 4.64856654e-01 1.19557559e+00 1.17702651e+00
8.84466588e-01 -1.46591306e-01 -3.18424463e-01 9.99500751e-01
5.22994161e-01 5.27314067e-01 5.55523932e-01 -1.24672186e+00
7.56980062e-01 4.61011767e-01 1.76097199e-01 -3.28628898e-01
-1.81843728e-01 8.10794309e-02 -5.18113375e-01 -3.42235178e-01
5.31348467e-01 -6.51858211e-01 -6.53484344e-01 2.29199481e+00
2.36865729e-01 6.50063157e-02 7.04144657e-01 4.58292961e-01
8.42347145e-01 8.56793106e-01 2.71876872e-01 -2.31637299e-01
1.31901705e+00 -9.32881117e-01 -7.48818398e-01 -4.04233336e-01
9.06745553e-01 -4.43433493e-01 1.13296747e+00 -1.89012997e-02
-1.35574853e+00 -2.85308570e-01 -7.85281539e-01 -2.73265541e-01
1.89765811e-01 1.81524485e-01 8.01989853e-01 5.41290998e-01
-1.16115820e+00 5.19549072e-01 -9.97329295e-01 -1.22767009e-01
3.25512409e-01 7.64117539e-02 -1.82067588e-01 4.64346036e-02
-1.12036586e+00 5.49945712e-01 2.56682456e-01 -2.90646672e-01
-7.77567565e-01 -7.62848556e-01 -1.33273029e+00 4.33982611e-02
3.11609119e-01 -1.02071476e+00 1.91063356e+00 -8.34244668e-01
-1.73583877e+00 1.16319668e+00 -4.22005236e-01 -6.04092300e-01
4.22424734e-01 -2.80187249e-01 2.34882742e-01 9.38961357e-02
5.60285568e-01 9.90471184e-01 6.52255416e-01 -1.17748344e+00
-6.34373844e-01 -2.49809384e-01 4.13193971e-01 2.95639247e-01
-9.25706401e-02 -1.13123655e-01 -3.17349344e-01 -4.01221782e-01
-1.05444200e-01 -7.59545088e-01 -3.61680359e-01 -4.82798338e-01
-7.09658206e-01 -6.00899816e-01 3.23116928e-01 -6.79065824e-01
8.91047716e-01 -1.92747021e+00 5.19006014e-01 -3.79359096e-01
1.01917237e-01 -5.09273484e-02 -1.49583086e-01 4.05633509e-01
1.57570511e-01 2.26112127e-01 -7.15483963e-01 -9.97053027e-01
4.06101435e-01 5.78364551e-01 -5.73536277e-01 1.51574180e-01
5.75567842e-01 1.06013680e+00 -1.05838871e+00 -3.11524153e-01
-5.93360625e-02 1.42777354e-01 -7.77005494e-01 6.54006898e-01
-7.14616418e-01 5.88937402e-01 -6.32276773e-01 2.03378990e-01
4.27122355e-01 -4.66540575e-01 6.40219927e-01 3.25530112e-01
1.94841534e-01 6.77513957e-01 -9.09876108e-01 2.05090761e+00
-8.38432670e-01 3.91975671e-01 3.06934714e-01 -1.16226661e+00
7.45643318e-01 4.50405151e-01 1.33758858e-01 -3.39674860e-01
-1.50341421e-01 1.90944895e-02 -5.08263946e-01 -5.53509533e-01
5.76645613e-01 -4.40067917e-01 -4.22598094e-01 8.78912568e-01
2.90418833e-01 -2.81931192e-01 1.56322032e-01 4.70574349e-01
9.35316920e-01 3.97756010e-01 4.61014926e-01 -1.84448078e-01
3.91419679e-01 -1.45207793e-02 5.27577579e-01 9.50052559e-01
1.01311490e-01 5.72542787e-01 1.27730680e+00 1.59140706e-01
-9.22433913e-01 -1.27784920e+00 -4.82373126e-02 1.13685441e+00
9.01426561e-03 -3.01853150e-01 -1.08826280e+00 -1.21754658e+00
-4.44760919e-03 1.07648289e+00 -3.95555288e-01 1.85663253e-02
-4.75044906e-01 -7.45012581e-01 7.45226324e-01 5.25774956e-01
3.13264042e-01 -1.03155828e+00 -5.43662488e-01 8.52584317e-02
-3.85837406e-01 -1.52953684e+00 -2.16794416e-01 1.12475134e-01
-7.12615311e-01 -1.05031765e+00 -5.31916082e-01 -8.07207227e-01
5.67467988e-01 5.92411272e-02 1.37025845e+00 -6.62306026e-02
-1.64124824e-03 6.50806785e-01 -5.22263229e-01 -1.35158032e-01
-1.15254259e+00 2.17246324e-01 -2.09136397e-01 1.12739194e-03
1.63206875e-01 -5.25706947e-01 -4.16287966e-02 -2.36278698e-01
-1.01583910e+00 5.47172606e-01 4.33549345e-01 9.74011600e-01
2.07201988e-01 -5.29123306e-01 9.61694598e-01 -1.17314684e+00
7.53452539e-01 -7.26907909e-01 -5.92457414e-01 2.51983196e-01
-4.20985907e-01 6.81136549e-01 5.62858164e-01 6.86295256e-02
-1.50696933e+00 7.16952756e-02 -2.46031791e-01 1.19515412e-01
-5.14108002e-01 5.09295583e-01 -3.88005912e-01 6.35488629e-01
2.31600404e-01 1.62592918e-01 6.99604377e-02 -4.95617449e-01
6.94424510e-01 8.37774456e-01 5.20474494e-01 -9.41974401e-01
6.83235109e-01 3.49919468e-01 -1.66358992e-01 -7.15135872e-01
-1.31345046e+00 -2.04414606e-01 -5.59713304e-01 4.15735751e-01
1.22651184e+00 -1.04643881e+00 -3.23334634e-01 5.37879467e-02
-1.47410142e+00 -6.06260777e-01 -3.15032452e-01 2.11753085e-01
-9.22330201e-01 6.46368921e-01 -4.71065819e-01 -1.00352085e+00
-1.70307279e-01 -1.06077588e+00 1.53638244e+00 1.80434227e-01
-2.73780048e-01 -1.46042740e+00 1.50776640e-01 5.22219300e-01
1.29786720e-02 2.72269726e-01 1.19272995e+00 -8.43175590e-01
-3.29133779e-01 1.92603663e-01 -3.12940657e-01 4.86350983e-01
1.79897726e-01 -3.67321759e-01 -1.17790711e+00 4.50846273e-03
8.94850306e-03 -7.16101289e-01 9.67885792e-01 3.29946578e-01
7.00730324e-01 -4.57011789e-01 -1.90725699e-01 4.34972048e-01
1.14755011e+00 -2.49098971e-01 2.54976064e-01 -1.02665432e-01
5.65981209e-01 9.73956466e-01 5.44010639e-01 5.91837168e-01
7.21256018e-01 6.30608201e-01 1.57310754e-01 7.78222010e-02
-2.98936274e-02 -5.24825633e-01 6.49414122e-01 5.41784704e-01
4.83479291e-01 -2.53673106e-01 -9.00435865e-01 6.16476655e-01
-2.07911873e+00 -6.13044620e-01 3.70835029e-02 1.95851743e+00
1.22249174e+00 -2.43772373e-01 3.56500149e-02 -3.62295210e-01
7.04614818e-01 3.34924161e-01 -4.21181321e-01 -4.95089144e-01
-2.45362893e-02 2.45907322e-01 1.91558257e-01 7.54338741e-01
-1.04350030e+00 1.08174825e+00 6.28459930e+00 6.62967265e-01
-5.17150402e-01 1.97474509e-01 6.60688519e-01 1.01081111e-01
-6.47250772e-01 1.17356539e-01 -7.79535830e-01 1.99478850e-01
1.09957719e+00 -4.09340352e-01 3.55848044e-01 8.11594784e-01
1.61439687e-01 -3.20792831e-02 -1.51134121e+00 7.18072951e-01
-6.83445111e-02 -1.30437422e+00 1.28968015e-01 1.72553211e-02
8.88838112e-01 -1.28313541e-01 -1.00754529e-01 5.70352435e-01
9.46263611e-01 -1.00409794e+00 6.35722756e-01 1.01083061e-02
8.50870609e-01 -6.46493077e-01 5.34802258e-01 6.40396953e-01
-9.81443703e-01 5.85287958e-02 -2.44436115e-01 -1.30184799e-01
4.28164542e-01 4.64127481e-01 -8.85168493e-01 7.56686926e-01
6.86732978e-02 7.66643882e-01 -2.08607554e-01 1.25009328e-01
-6.97869837e-01 6.57765806e-01 5.03945686e-02 1.66482165e-01
3.37672442e-01 -2.78875917e-01 5.44204772e-01 1.36579454e+00
-2.15328652e-02 -2.60835979e-02 3.80956650e-01 1.47497141e+00
-3.56806844e-01 -3.58841196e-02 -7.21597075e-01 -3.78260165e-01
3.37654173e-01 1.28277183e+00 -3.71525317e-01 -4.47968423e-01
-4.53379303e-01 1.04592061e+00 4.74080086e-01 3.74426097e-01
-6.76264584e-01 -1.81082949e-01 8.21918726e-01 -4.09952074e-01
2.89123386e-01 -1.36534393e-01 -3.54086429e-01 -1.28585982e+00
2.38619298e-02 -7.16306508e-01 3.28707755e-01 -4.14592773e-01
-1.22822189e+00 2.51019567e-01 2.09529743e-01 -7.40036070e-01
-1.02959347e+00 -4.99184668e-01 -5.83672523e-01 9.18270469e-01
-1.79790068e+00 -1.06609368e+00 8.84976164e-02 1.70132533e-01
8.35704207e-01 -1.68060381e-02 9.14288342e-01 -1.34924307e-01
-5.87956607e-01 5.09270549e-01 4.38586287e-02 1.36140645e-01
3.65796238e-01 -1.85248947e+00 6.99404061e-01 8.00102830e-01
2.48552337e-01 4.64783937e-01 6.33243918e-01 -5.26895344e-01
-1.42944884e+00 -1.13543487e+00 1.01009417e+00 -4.73643988e-01
6.44053400e-01 -5.26679099e-01 -6.79574311e-01 8.42542529e-01
2.81575888e-01 -4.19047266e-01 7.45131075e-01 6.88480660e-02
-4.05609220e-01 7.71613717e-01 -1.12847579e+00 4.69792724e-01
1.08223224e+00 -6.32466257e-01 -7.45938838e-01 6.09363198e-01
1.18192399e+00 -3.95681202e-01 -6.49755180e-01 1.69537142e-01
1.29651651e-01 -1.04633391e+00 6.77961588e-01 -8.57098103e-01
9.37009215e-01 1.35177165e-01 -3.81928831e-01 -1.35770440e+00
3.39699030e-01 -1.01687133e+00 -1.77810594e-01 1.55143130e+00
6.10336006e-01 -5.36148012e-01 7.16065407e-01 6.38401806e-01
-1.72315076e-01 -5.98502159e-01 -7.20678568e-01 -6.51173353e-01
3.32175970e-01 -4.73239303e-01 4.90132838e-01 8.16547751e-01
1.14955217e-01 8.64142060e-01 -2.15723291e-01 1.57465756e-01
5.86359739e-01 3.83733511e-01 8.96058917e-01 -1.00602603e+00
-7.37662971e-01 -3.32773447e-01 3.81706320e-02 -1.56310475e+00
1.00897741e+00 -1.22718918e+00 2.63382971e-01 -1.71121812e+00
3.91495883e-01 -1.61098480e-01 4.35218483e-01 3.71736646e-01
-3.74474972e-01 -3.38262647e-01 5.11246547e-02 3.51844020e-02
-7.92172372e-01 7.23949313e-01 9.34743226e-01 1.64563373e-01
-2.66608119e-01 1.25128344e-01 -1.11362290e+00 7.52235532e-01
5.64386904e-01 -3.39938402e-01 -6.04856312e-01 -5.82929552e-01
1.09940834e-01 4.62119639e-01 2.70042807e-01 -3.79977793e-01
-1.63158998e-01 -9.68524590e-02 -2.32879907e-01 -1.66471049e-01
4.50858772e-02 -5.92821129e-02 -4.92034018e-01 4.25438546e-02
-9.37725842e-01 -3.02259326e-01 5.00881672e-02 7.59213507e-01
-2.29559988e-01 -6.45383179e-01 6.49832726e-01 -2.16353804e-01
-3.89805466e-01 8.78480379e-04 -3.39820296e-01 7.83210397e-01
6.43098712e-01 1.15370937e-01 -2.02500019e-02 -6.97938859e-01
-6.56395495e-01 4.74483341e-01 3.31589639e-01 3.19438219e-01
4.22197014e-01 -1.03117096e+00 -8.60346258e-01 9.24270302e-02
-7.26834238e-02 4.35529262e-01 4.85663638e-02 7.10916162e-01
-8.07390735e-02 5.43976247e-01 4.06082273e-01 -5.27021170e-01
-9.95460212e-01 1.00551583e-01 1.68701082e-01 -7.16865003e-01
-5.48680842e-01 7.77449489e-01 7.04718113e-01 -9.41778898e-01
2.49613509e-01 -1.30283654e-01 -2.24477813e-01 5.27124070e-02
3.79689306e-01 1.56226695e-01 -2.87418306e-01 -5.04467249e-01
1.16942696e-01 1.07505247e-01 -3.67596187e-02 -4.96535838e-01
1.33342516e+00 -3.54841679e-01 -2.10196838e-01 4.08030421e-01
1.29778612e+00 -8.18145126e-02 -1.46183789e+00 -3.91241699e-01
4.98561949e-01 -2.63871867e-02 -2.70513177e-01 -4.84239191e-01
-6.15998387e-01 9.12422001e-01 -1.60494998e-01 3.92399043e-01
7.02138424e-01 4.39478725e-01 9.29592371e-01 4.71238554e-01
1.26554400e-01 -8.08719695e-01 1.58052966e-01 7.19974339e-01
5.79962194e-01 -1.22564411e+00 -4.73921359e-01 -4.33523417e-01
-1.05801666e+00 9.53124464e-01 3.42885971e-01 3.98359261e-02
1.02980152e-01 2.65209407e-01 -1.30189568e-01 -8.97171199e-02
-1.05763662e+00 -2.09316373e-01 -1.23504885e-02 7.12265968e-01
7.39252627e-01 1.20002083e-01 -7.25405058e-03 9.06684816e-01
-3.16910058e-01 -3.60007524e-01 7.54509151e-01 9.89916444e-01
-3.84179175e-01 -1.38472843e+00 2.70841252e-02 2.73132861e-01
-6.96729243e-01 -2.75184423e-01 -5.43959737e-01 4.88852620e-01
-4.50486451e-01 1.13267040e+00 5.01773581e-02 5.88107891e-02
1.37183247e-02 4.04362887e-01 4.33495343e-01 -1.17998302e+00
-3.83996218e-01 -7.15157837e-02 4.52291191e-01 -5.19940794e-01
-3.98143291e-01 -8.53238523e-01 -1.71388984e+00 1.41892850e-01
-1.81077749e-01 3.94627243e-01 5.97411156e-01 1.21320558e+00
3.16527992e-01 3.99949789e-01 7.06727862e-01 -7.22175777e-01
-1.16418862e+00 -9.08279359e-01 -5.58006883e-01 5.69604754e-01
3.46997291e-01 -3.68941218e-01 -4.64675188e-01 2.85006285e-01]
|
[10.672913551330566, 9.103737831115723]
|
98a1a6d7-b9d3-4395-95f2-4f7cba5768e0
|
better-cmos-produces-clearer-images-learning
|
2304.03542
| null |
https://arxiv.org/abs/2304.03542v1
|
https://arxiv.org/pdf/2304.03542v1.pdf
|
Better "CMOS" Produces Clearer Images: Learning Space-Variant Blur Estimation for Blind Image Super-Resolution
|
Most of the existing blind image Super-Resolution (SR) methods assume that the blur kernels are space-invariant. However, the blur involved in real applications are usually space-variant due to object motion, out-of-focus, etc., resulting in severe performance drop of the advanced SR methods. To address this problem, we firstly introduce two new datasets with out-of-focus blur, i.e., NYUv2-BSR and Cityscapes-BSR, to support further researches of blind SR with space-variant blur. Based on the datasets, we design a novel Cross-MOdal fuSion network (CMOS) that estimate both blur and semantics simultaneously, which leads to improved SR results. It involves a feature Grouping Interactive Attention (GIA) module to make the two modalities interact more effectively and avoid inconsistency. GIA can also be used for the interaction of other features because of the universality of its structure. Qualitative and quantitative experiments compared with state-of-the-art methods on above datasets and real-world images demonstrate the superiority of our method, e.g., obtaining PSNR/SSIM by +1.91/+0.0048 on NYUv2-BSR than MANet.
|
['Yong liu', 'Chengjie Wang', 'Yabiao Wang', 'Chao Xu', 'Jiangning Zhang', 'Xuhai Chen']
|
2023-04-07
| null |
http://openaccess.thecvf.com//content/CVPR2023/html/Chen_Better_CMOS_Produces_Clearer_Images_Learning_Space-Variant_Blur_Estimation_for_CVPR_2023_paper.html
|
http://openaccess.thecvf.com//content/CVPR2023/papers/Chen_Better_CMOS_Produces_Clearer_Images_Learning_Space-Variant_Blur_Estimation_for_CVPR_2023_paper.pdf
|
cvpr-2023-1
|
['image-super-resolution']
|
['computer-vision']
|
[-6.29969835e-02 -9.16311383e-01 2.96715558e-01 -2.72287875e-01
-4.07995224e-01 -2.31256843e-01 2.77282834e-01 -6.22053981e-01
-2.61666000e-01 7.54110515e-01 6.02163136e-01 -5.36600538e-02
-4.01977330e-01 -4.17169154e-01 -3.21878225e-01 -8.07654858e-01
1.41266927e-01 -6.07022047e-01 4.40352619e-01 -2.61442989e-01
3.62817138e-01 2.13713542e-01 -1.52266848e+00 2.53088087e-01
1.39407396e+00 7.72836506e-01 6.78179562e-01 3.32094222e-01
1.27327010e-01 8.12205493e-01 -5.49373209e-01 -1.76715523e-01
2.56246269e-01 -5.54976344e-01 -3.39046746e-01 1.24343954e-01
2.76994735e-01 -5.49684405e-01 -8.27440321e-01 1.55896652e+00
7.38419056e-01 1.89768463e-01 3.10311824e-01 -7.76441038e-01
-1.36977911e+00 3.84696424e-01 -9.75328863e-01 5.67814291e-01
4.20430660e-01 3.20967108e-01 5.34969687e-01 -8.27715337e-01
1.64172485e-01 1.44819713e+00 4.69177902e-01 3.84774327e-01
-1.00467825e+00 -8.74615967e-01 -5.57025149e-02 6.15327120e-01
-1.51223564e+00 -5.98532796e-01 5.79869449e-01 -1.82690382e-01
2.75056034e-01 4.94905889e-01 2.85049736e-01 8.31701994e-01
6.86818808e-02 5.61960161e-01 1.47140992e+00 1.01319645e-02
3.78393382e-02 8.31955001e-02 3.95414568e-02 2.88327724e-01
4.78677064e-01 2.72576362e-01 -4.90891337e-01 3.09445441e-01
1.16278231e+00 2.49935478e-01 -1.00107706e+00 -6.20953254e-02
-1.58278263e+00 3.32122862e-01 8.38987708e-01 3.86122108e-01
-1.19671002e-01 -1.54533535e-01 9.37453955e-02 2.42842838e-01
3.34900826e-01 2.30012044e-01 -2.44643331e-01 1.10428564e-01
-8.39687467e-01 3.68160307e-02 3.90832387e-02 1.03945529e+00
5.83951652e-01 -7.26283565e-02 -4.20600176e-01 1.13205481e+00
5.12497425e-01 5.36851406e-01 6.29785776e-01 -8.71124983e-01
3.77217919e-01 2.79092163e-01 4.95380640e-01 -8.42856526e-01
-5.73393749e-03 -6.31074965e-01 -1.27490151e+00 1.33528903e-01
1.85219467e-01 -2.26818044e-02 -9.70607936e-01 1.48503077e+00
5.91885522e-02 4.57106560e-01 2.17955336e-01 1.58677363e+00
9.71919417e-01 5.39622545e-01 -9.51446444e-02 -4.14221078e-01
1.51345384e+00 -1.03764844e+00 -1.04781544e+00 -1.77958369e-01
-1.18287630e-01 -9.70609963e-01 1.14253783e+00 3.11436445e-01
-8.95228565e-01 -9.58073437e-01 -1.04188120e+00 -5.64785264e-02
-1.31538855e-02 4.23718095e-01 4.53261137e-01 4.35624927e-01
-1.04292548e+00 3.26757044e-01 -5.34252405e-01 -2.20015183e-01
4.46603507e-01 2.20929738e-02 -2.25560933e-01 -4.59236741e-01
-1.40927720e+00 8.68532479e-01 2.24324510e-01 4.11166608e-01
-5.06764293e-01 -5.15933275e-01 -6.51054800e-01 -3.87183391e-02
2.23086193e-01 -8.00808609e-01 1.11732376e+00 -8.02435160e-01
-1.20436072e+00 2.91309237e-01 -3.44398022e-01 -5.42975366e-02
6.71770155e-01 -4.53561693e-01 -1.02601790e+00 7.88044706e-02
1.59329087e-01 1.66214809e-01 9.47142959e-01 -1.44361436e+00
-7.08984017e-01 -4.22042578e-01 1.60065144e-01 5.10551214e-01
-2.64375955e-01 3.17284018e-01 -6.60153389e-01 -7.87424624e-01
2.35152259e-01 -4.04585719e-01 -4.26298752e-02 -8.76600966e-02
-3.14374119e-01 1.81945220e-01 8.58161032e-01 -1.05847728e+00
1.49148273e+00 -2.45266676e+00 1.06939010e-01 -3.06458622e-01
3.56668085e-01 3.86062652e-01 -4.08661403e-02 -1.45424549e-02
-7.21507818e-02 -1.29731640e-01 -2.99297750e-01 -3.61704379e-02
-3.75493973e-01 -2.21319258e-01 -8.43277127e-02 6.54570043e-01
-4.07841764e-02 5.66237748e-01 -1.02579391e+00 -4.99400735e-01
4.26002145e-01 5.35704494e-01 -7.99343064e-02 3.37596357e-01
2.61763424e-01 7.22645581e-01 -4.16405708e-01 6.29660606e-01
1.19599473e+00 -1.91147700e-01 -4.05761361e-01 -6.32912278e-01
-4.10539895e-01 -2.56655753e-01 -1.36593437e+00 1.78366542e+00
-5.14764309e-01 5.63175976e-01 1.78630263e-01 -3.63922238e-01
8.17284703e-01 2.26172820e-01 4.29247282e-02 -7.59563982e-01
-2.41299625e-02 3.73978883e-01 2.01242585e-02 -8.09012592e-01
4.37084377e-01 6.34998456e-02 5.46894252e-01 -2.06071995e-02
-3.57037991e-01 2.41569206e-01 -3.18949483e-02 5.69825619e-03
8.64318967e-01 -6.83104172e-02 1.05560549e-01 -2.71528035e-01
9.03277695e-01 -5.33954561e-01 5.30503690e-01 4.87994432e-01
-4.26514924e-01 8.95930469e-01 -1.87429801e-01 -5.78235835e-02
-9.12306607e-01 -1.09234989e+00 -2.70853728e-01 5.01481533e-01
1.03087306e+00 -4.88853119e-02 -7.37630785e-01 -1.93829224e-01
-9.47944075e-02 5.72892964e-01 -4.03622746e-01 -1.59153223e-01
-2.73832768e-01 -8.52063358e-01 1.05062731e-01 4.03009444e-01
1.36881721e+00 -8.30227792e-01 -3.31969172e-01 -1.43075064e-01
-4.02189642e-01 -1.11646545e+00 -8.70303690e-01 -7.37958014e-01
-6.02166533e-01 -9.14692104e-01 -1.24298358e+00 -6.50161326e-01
6.54193342e-01 1.08056307e+00 5.67167878e-01 1.50313571e-01
-7.49734268e-02 -1.08922467e-01 -4.00626600e-01 7.82960802e-02
-1.48546975e-03 -4.18472826e-01 1.21862940e-01 3.44843954e-01
2.51558512e-01 -4.34881598e-01 -1.04304564e+00 5.30732393e-01
-1.01414931e+00 2.60044694e-01 7.88734019e-01 8.36436331e-01
8.29911754e-02 5.72906852e-01 4.84093010e-01 -3.19239289e-01
8.38753343e-01 -3.51230741e-01 -4.95144546e-01 2.99852580e-01
-5.41646898e-01 -2.12049335e-01 4.24473643e-01 -5.21876395e-01
-1.60887313e+00 -3.58669639e-01 2.58754790e-01 -6.48248494e-01
-1.58406898e-01 3.29053998e-01 -4.55881983e-01 -1.19407728e-01
5.34353495e-01 4.63893116e-01 -6.01085834e-02 -7.05761850e-01
2.61730671e-01 1.38276434e+00 7.18405664e-01 1.01549029e-02
9.73139584e-01 3.41402262e-01 -3.48761857e-01 -7.64077127e-01
-6.67893231e-01 -4.93260950e-01 -1.92270055e-01 -6.76356256e-02
8.98829460e-01 -1.30957127e+00 -6.39778972e-01 9.97731924e-01
-1.13682640e+00 4.78367656e-02 1.98330477e-01 7.49298155e-01
-1.92568317e-01 7.05469191e-01 -7.71505296e-01 -8.93405974e-01
-2.76295185e-01 -1.37315035e+00 8.69252384e-01 8.27630281e-01
4.05957252e-01 -5.25187850e-01 -3.66503865e-01 4.67970520e-01
8.29510212e-01 -3.58230352e-01 3.13877821e-01 8.18540826e-02
-8.52650702e-01 2.15012819e-01 -9.81048763e-01 4.72802043e-01
6.05993688e-01 -2.89393634e-01 -1.12217450e+00 -4.36371922e-01
2.40602881e-01 3.03950459e-01 9.40872729e-01 3.59804183e-01
1.23629773e+00 -2.55199909e-01 -8.60449970e-02 7.42564619e-01
1.55012751e+00 3.20053637e-01 9.87041414e-01 5.26246250e-01
6.94848299e-01 3.19791466e-01 8.08028758e-01 2.25212544e-01
2.34298840e-01 7.52705276e-01 3.85444403e-01 -2.67424852e-01
-4.26575273e-01 -4.09086496e-02 3.80757660e-01 5.55586815e-01
-2.35320300e-01 3.17053273e-02 -4.46545720e-01 5.08581519e-01
-1.82960868e+00 -8.70922983e-01 -3.45531791e-01 2.21342301e+00
9.18426037e-01 -1.17738590e-01 -2.08081558e-01 -7.77012482e-02
1.14773202e+00 2.75905907e-01 -5.30788958e-01 3.88695478e-01
-4.02865976e-01 -1.78970635e-01 5.59132397e-01 3.79351467e-01
-1.01892626e+00 5.81561804e-01 4.89555216e+00 9.76419210e-01
-1.02304387e+00 1.78198382e-01 5.57263672e-01 2.11004525e-01
-1.75102234e-01 -9.62935686e-02 -3.07471305e-01 1.03725135e+00
4.90801334e-01 -2.30844393e-01 1.08353186e+00 3.09447080e-01
5.52498102e-01 -2.25622326e-01 -5.30118406e-01 1.42170715e+00
1.35454489e-03 -8.77899289e-01 -1.82654366e-01 -1.48506418e-01
7.01656163e-01 -1.39477536e-01 1.01479970e-01 -9.76358876e-02
9.70425308e-02 -9.13086534e-01 5.82188785e-01 7.82823265e-01
1.17863154e+00 -4.94913429e-01 9.91425514e-01 1.61556974e-01
-1.22098088e+00 -2.28156015e-01 -5.32935739e-01 1.71397880e-01
1.84610605e-01 7.18637645e-01 1.49243549e-01 1.00602591e+00
1.08867824e+00 1.02875245e+00 -7.51474023e-01 1.51483214e+00
-2.40681753e-01 2.27827296e-01 7.63740689e-02 3.69120210e-01
-1.86597005e-01 -3.37602407e-01 6.21172369e-01 1.04766655e+00
6.67291164e-01 2.73448527e-01 -3.06146950e-01 9.81963277e-01
1.44303948e-01 -2.24248052e-01 -2.22664937e-01 2.74641782e-01
6.17478788e-01 1.19466901e+00 -2.04097018e-01 -3.37448001e-01
-5.96097052e-01 1.30103600e+00 -2.62472391e-01 7.84909368e-01
-8.49223137e-01 -6.55585527e-01 8.31738591e-01 -7.86325634e-02
3.45583707e-01 -1.73715442e-01 -3.21533531e-01 -1.57959199e+00
5.78146521e-03 -9.65552568e-01 1.60180535e-02 -1.40094936e+00
-1.50136971e+00 7.03065932e-01 -1.43958405e-01 -1.52713859e+00
5.31248689e-01 -3.13243091e-01 -5.79149723e-01 1.39089417e+00
-1.77305973e+00 -1.03784871e+00 -9.02265191e-01 7.16832101e-01
5.98053157e-01 -1.52479008e-01 1.87535405e-01 5.73716819e-01
-7.34705091e-01 3.78041953e-01 3.14988375e-01 -6.48888946e-02
1.11724448e+00 -1.04916120e+00 2.67054021e-01 1.28648365e+00
-3.15966547e-01 8.21806312e-01 7.69858718e-01 -5.96336782e-01
-1.29802108e+00 -1.03506160e+00 4.62761343e-01 -1.85128346e-01
5.45983613e-01 7.94230774e-02 -1.05676019e+00 3.27031732e-01
3.23483616e-01 1.69441164e-01 -1.69674139e-02 -2.34455004e-01
-3.45235288e-01 -4.96908784e-01 -1.20223629e+00 7.22463191e-01
1.15626669e+00 -4.77926344e-01 -5.56485832e-01 -4.25695218e-02
1.01987457e+00 -5.79835773e-01 -6.71394944e-01 5.63133180e-01
3.51106644e-01 -1.32596111e+00 1.12074089e+00 -6.66627893e-03
3.53833020e-01 -1.00112247e+00 -1.26302540e-01 -1.44901276e+00
-6.69404805e-01 -4.46107805e-01 -2.08346799e-01 1.32180750e+00
-1.77810922e-01 -8.10856879e-01 -1.39124347e-02 2.44132817e-01
-1.01832844e-01 -3.30004930e-01 -6.64878190e-01 -6.06104136e-01
-6.34838641e-01 -5.36077991e-02 8.92046034e-01 9.46604490e-01
-3.72210830e-01 3.61167669e-01 -5.59444368e-01 5.84037244e-01
7.98977792e-01 7.06304610e-02 5.15595138e-01 -9.30544615e-01
-1.07566230e-01 -2.93473959e-01 -2.66814202e-01 -1.19467950e+00
-4.60510731e-01 -1.95593581e-01 8.03326666e-02 -1.71660268e+00
3.31194252e-01 -3.15039605e-01 -5.70198834e-01 9.32295546e-02
-6.30867839e-01 2.62090147e-01 1.98832210e-02 6.75288141e-01
-5.03006816e-01 5.25555789e-01 1.74559236e+00 3.59520456e-03
-1.46020040e-01 3.23801953e-03 -9.51344430e-01 4.19607431e-01
6.45337284e-01 2.57839382e-01 -4.01014596e-01 -7.04752207e-01
-3.06724519e-01 1.27682358e-01 5.25674343e-01 -1.06713831e+00
2.79281974e-01 -3.11818391e-01 5.39625347e-01 -3.35912794e-01
2.04901457e-01 -9.54573393e-01 3.06531578e-01 2.63577789e-01
-8.30511078e-02 -1.14639960e-01 1.15923146e-02 7.52085745e-01
-3.27201426e-01 3.68427560e-02 9.52733338e-01 -1.72629893e-01
-8.47584546e-01 2.34665886e-01 1.85315207e-01 -2.45543048e-01
7.47997940e-01 -2.37400353e-01 -8.02797973e-01 -4.86111045e-01
-1.86477333e-01 1.43804699e-01 6.36128247e-01 5.29831767e-01
7.58117318e-01 -1.31862581e+00 -8.55481803e-01 4.34746504e-01
6.00297190e-02 7.34308437e-02 9.19630408e-01 1.04553473e+00
-5.36067188e-01 1.97468415e-01 -3.00316483e-01 -3.93950284e-01
-1.21701443e+00 7.48579979e-01 3.77114683e-01 1.78806126e-01
-5.49809217e-01 6.67330682e-01 4.92055565e-01 1.65376872e-01
2.55695451e-02 -1.80929810e-01 -3.46642643e-01 -2.88121969e-01
8.13632846e-01 5.73028147e-01 -2.37323150e-01 -6.46145403e-01
-3.01621288e-01 4.82467830e-01 6.81009563e-03 5.21159582e-02
1.17538571e+00 -7.99206734e-01 -4.03735727e-01 2.01460153e-01
8.07825506e-01 3.01742077e-01 -1.28852546e+00 -4.75663602e-01
-3.85532081e-01 -1.17696595e+00 1.95636660e-01 -1.00082195e+00
-1.22940803e+00 7.54666269e-01 9.90592301e-01 1.58948168e-01
1.59126592e+00 -1.54864728e-01 8.10776472e-01 -2.01305270e-01
4.13357317e-01 -6.36425972e-01 -1.49726048e-01 8.54664221e-02
9.49589074e-01 -1.28003240e+00 9.35545266e-02 -4.18374628e-01
-6.07060492e-01 8.66104364e-01 7.16033578e-01 8.58873576e-02
4.21478838e-01 -1.94107145e-01 1.68432564e-01 2.21469447e-01
-2.51568854e-01 -3.42459351e-01 3.74224842e-01 6.11713767e-01
3.58730316e-01 -1.13639146e-01 -3.76597583e-01 5.40723324e-01
8.40036273e-02 2.60838091e-01 5.06232858e-01 5.56661189e-01
-4.56521600e-01 -6.27700090e-01 -6.63916588e-01 3.39838177e-01
-2.82866269e-01 -3.03820074e-01 1.99165761e-01 4.55530435e-01
2.24510938e-01 1.26689899e+00 -2.00270921e-01 -5.38731337e-01
3.27298373e-01 -6.45164132e-01 2.53997147e-01 -1.92770824e-01
-5.36703058e-02 2.38124058e-01 -2.53590375e-01 -3.97614568e-01
-6.99566007e-01 -5.69552004e-01 -8.58320773e-01 -3.87119293e-01
-5.58311999e-01 5.14114015e-02 3.77224207e-01 6.95604384e-01
3.56005222e-01 7.22578287e-01 7.37120032e-01 -6.44054115e-01
-5.29445767e-01 -1.28276622e+00 -9.92405593e-01 4.48344648e-01
7.57383823e-01 -6.72083676e-01 -6.64307237e-01 1.02517679e-01]
|
[11.449225425720215, -2.60996675491333]
|
109ea7a1-4e84-4b6d-a70f-000699897110
|
thifly-research-at-semeval-2023-task-7-a
|
2306.01245
| null |
https://arxiv.org/abs/2306.01245v1
|
https://arxiv.org/pdf/2306.01245v1.pdf
|
THiFLY Research at SemEval-2023 Task 7: A Multi-granularity System for CTR-based Textual Entailment and Evidence Retrieval
|
The NLI4CT task aims to entail hypotheses based on Clinical Trial Reports (CTRs) and retrieve the corresponding evidence supporting the justification. This task poses a significant challenge, as verifying hypotheses in the NLI4CT task requires the integration of multiple pieces of evidence from one or two CTR(s) and the application of diverse levels of reasoning, including textual and numerical. To address these problems, we present a multi-granularity system for CTR-based textual entailment and evidence retrieval in this paper. Specifically, we construct a Multi-granularity Inference Network (MGNet) that exploits sentence-level and token-level encoding to handle both textual entailment and evidence retrieval tasks. Moreover, we enhance the numerical inference capability of the system by leveraging a T5-based model, SciFive, which is pre-trained on the medical corpus. Model ensembling and a joint inference method are further utilized in the system to increase the stability and consistency of inference. The system achieves f1-scores of 0.856 and 0.853 on textual entailment and evidence retrieval tasks, resulting in the best performance on both subtasks. The experimental results corroborate the effectiveness of our proposed method. Our code is publicly available at https://github.com/THUMLP/NLI4CT.
|
['Ji Wu', 'Xinxin You', 'Xien Liu', 'Miao Li', 'Meiwei Li', 'Ziyu Jin', 'Yuxuan Zhou']
|
2023-06-02
| null | null | null | null |
['natural-language-inference']
|
['natural-language-processing']
|
[ 2.45180979e-01 1.58883244e-01 -6.93681538e-01 -3.64786476e-01
-1.32158923e+00 -4.10817742e-01 4.36665714e-01 7.45973110e-01
-4.81305420e-01 9.09613967e-01 2.25547329e-01 -6.77333772e-01
-4.79839623e-01 -6.47714019e-01 -7.51694441e-01 -8.67526233e-02
1.36423573e-01 3.67830187e-01 2.05870140e-02 1.05121575e-01
3.44004005e-01 1.19650356e-01 -1.16219604e+00 8.10761154e-01
1.21556377e+00 1.37836564e+00 -7.43489666e-03 6.25548184e-01
1.08662091e-01 1.01256442e+00 -6.29801154e-01 -7.83576965e-01
-2.46751472e-01 -1.77679375e-01 -1.03645444e+00 -4.43633854e-01
1.68358386e-01 -5.97692132e-01 -1.11805856e-01 1.03089166e+00
4.01839942e-01 -1.22993782e-01 6.78083241e-01 -9.56138968e-01
-4.61684912e-01 8.61147642e-01 -2.60624796e-01 2.00849861e-01
6.90585554e-01 8.22236612e-02 1.18767786e+00 -9.11579609e-01
4.76437241e-01 9.97060299e-01 5.97732008e-01 2.39474565e-01
-6.28363311e-01 -6.61117613e-01 -6.95751533e-02 5.40747941e-01
-1.43673801e+00 -3.26516151e-01 2.96811730e-01 -6.93997592e-02
1.20649767e+00 4.13159519e-01 5.69947660e-01 9.08404708e-01
6.65142894e-01 8.85283828e-01 7.96542943e-01 -3.22103143e-01
3.16741139e-01 -1.01351239e-01 2.11340204e-01 9.09224570e-01
4.46578532e-01 -3.78938556e-01 -4.81915891e-01 -3.58418465e-01
5.19888222e-01 1.07012227e-01 -2.29106158e-01 6.41896605e-01
-1.18695796e+00 7.64214098e-01 4.13923174e-01 2.82806128e-01
-6.40879273e-01 1.75816804e-01 7.15885818e-01 1.42772794e-01
3.48616809e-01 2.47919679e-01 -5.34076154e-01 1.49221579e-02
-1.16646433e+00 3.05003077e-01 7.77851939e-01 1.07510221e+00
9.27002914e-03 -6.34781957e-01 -6.97730958e-01 7.70440578e-01
4.02522534e-01 3.75459403e-01 5.83055198e-01 -8.82905960e-01
7.41799593e-01 7.31203020e-01 -1.76969096e-02 -9.05114830e-01
-2.81765759e-01 -5.53367376e-01 -9.29091632e-01 -5.97103953e-01
6.69926777e-02 3.08643989e-02 -7.66117752e-01 1.59290552e+00
3.24031949e-01 3.14188570e-01 2.20826671e-01 7.61141181e-01
1.22939682e+00 4.03930485e-01 1.89328358e-01 -1.89546481e-01
2.03280568e+00 -8.09473276e-01 -1.23863137e+00 2.64807772e-02
8.73562157e-01 -7.69487500e-01 7.10019648e-01 3.19064260e-01
-1.54576910e+00 -2.39418998e-01 -1.14999330e+00 -3.44180197e-01
-2.60641366e-01 3.85086924e-01 3.56221974e-01 5.39626218e-02
-7.06983626e-01 3.47838551e-01 -6.96103811e-01 1.74269304e-01
4.93759960e-01 1.94512784e-01 -2.11310878e-01 -4.15424734e-01
-1.81637597e+00 1.09606886e+00 6.66138828e-01 5.00589848e-01
-6.05711639e-01 -8.28723967e-01 -1.13259077e+00 2.97280371e-01
5.18187165e-01 -1.04456139e+00 1.32474422e+00 1.53788671e-01
-1.08587873e+00 6.18920803e-01 -1.96073964e-01 -5.21072686e-01
4.37253356e-01 -1.36479437e-02 -3.75455767e-01 5.58056831e-01
2.03085735e-01 2.93129951e-01 3.07312459e-01 -4.69307154e-01
-5.46932280e-01 -2.05951005e-01 2.03854784e-01 7.99251422e-02
-1.20402098e-01 6.41742870e-02 -7.14104354e-01 -6.62306011e-01
-4.89320718e-02 -6.71226501e-01 -1.08001634e-01 -1.36447683e-01
-6.54260576e-01 -5.34159422e-01 1.49328321e-01 -8.20361316e-01
1.64522970e+00 -1.79236817e+00 -6.87466487e-02 2.74576902e-01
2.83371210e-01 2.72286385e-01 8.07397887e-02 3.36638719e-01
1.13328338e-01 2.21899748e-01 -2.74353266e-01 -2.29485512e-01
1.69786289e-01 2.48437852e-01 -6.38904348e-02 9.87177491e-02
5.20894647e-01 1.30448151e+00 -8.94778252e-01 -1.02512419e+00
-1.36451751e-01 1.46679908e-01 -6.33480012e-01 1.15901954e-01
-4.15338010e-01 6.02992326e-02 -7.22650707e-01 7.59715617e-01
5.25868654e-01 -7.35338032e-01 4.40829426e-01 -3.95735711e-01
2.47297540e-01 6.62665188e-01 -8.86298001e-01 1.84568167e+00
-4.08759594e-01 -1.85326219e-01 -2.65448928e-01 -8.80082071e-01
4.19037193e-01 7.09489822e-01 3.22155565e-01 -6.67655706e-01
2.33317181e-01 5.38502812e-01 -3.36923003e-02 -8.46227765e-01
3.95000756e-01 -2.62774855e-01 -3.24730933e-01 5.29374421e-01
-3.00732590e-02 -1.02268137e-01 4.40905154e-01 4.77296352e-01
1.22485960e+00 -2.33780414e-01 6.25223219e-01 8.19607303e-02
7.07261145e-01 -1.21508604e-02 6.11731648e-01 7.89651871e-01
1.44268826e-01 1.92190796e-01 6.31525636e-01 -5.20160608e-02
-5.94877124e-01 -7.38181174e-01 -6.85491741e-01 4.23148930e-01
-1.19716845e-01 -5.59674442e-01 -5.54166317e-01 -7.18328595e-01
1.54194996e-01 7.25107908e-01 -5.89363933e-01 -3.61693889e-01
-3.62696648e-01 -7.35820055e-01 9.81591284e-01 6.42505527e-01
3.78574818e-01 -9.75145280e-01 -4.96287346e-01 2.78210044e-01
-8.78410935e-01 -1.38252640e+00 -5.04880250e-01 4.68999054e-03
-9.01426494e-01 -1.24807084e+00 -5.45924962e-01 -4.84158367e-01
5.91488063e-01 -3.99777710e-01 9.26880956e-01 4.88303751e-01
-3.02041650e-01 1.62851606e-02 -3.97181809e-01 -3.86421293e-01
-4.68552589e-01 -8.79391804e-02 -1.14310861e-01 -5.07233679e-01
3.75466466e-01 -6.20645322e-02 -7.94546485e-01 8.80866721e-02
-1.32983470e+00 6.41623363e-02 8.48844171e-01 9.77663398e-01
6.98378205e-01 -2.92250663e-02 8.00371170e-01 -7.80897319e-01
9.96413052e-01 -6.37600243e-01 -3.24145406e-01 5.81192136e-01
-7.68557847e-01 4.35007885e-02 4.66314405e-01 -2.09524319e-01
-8.44630539e-01 -5.79849124e-01 -4.94991422e-01 -2.08905384e-01
1.30731389e-01 1.38218069e+00 1.10591285e-01 4.89462107e-01
3.11408371e-01 4.65271175e-02 2.76477747e-02 -1.81444988e-01
3.37520421e-01 8.02187800e-01 4.80710715e-01 -6.74962521e-01
1.54604584e-01 -7.29404110e-03 -2.98985839e-02 -1.25034019e-01
-1.32708192e+00 -3.91850889e-01 -1.60560459e-01 -1.55164842e-02
8.06785882e-01 -8.86942208e-01 -1.04554772e+00 1.16278976e-01
-1.36243272e+00 -1.48185343e-01 -5.99104958e-03 7.31550753e-01
-3.09877723e-01 5.19773245e-01 -1.19508541e+00 -4.68384534e-01
-8.64750803e-01 -1.35041785e+00 1.15093017e+00 -2.39019424e-01
-4.61023271e-01 -9.02644634e-01 -2.03799412e-01 6.03729606e-01
8.55097994e-02 1.77111581e-01 1.27943850e+00 -7.84078240e-01
-3.93659651e-01 -6.11846447e-01 -4.69993651e-01 1.47811621e-01
3.56208831e-02 -2.71130770e-01 -5.79293430e-01 7.05079138e-02
1.25198409e-01 -7.42948055e-01 7.96050131e-01 5.28282762e-01
1.42894161e+00 -5.63114226e-01 -2.92420387e-01 6.87639415e-02
1.19681633e+00 -9.53592267e-03 5.53760588e-01 1.07485071e-01
2.69069374e-01 4.08697665e-01 9.56731558e-01 6.40294552e-01
6.25606716e-01 5.89354873e-01 2.10642308e-01 -7.22827464e-02
1.03527196e-02 -1.25694320e-01 -6.87229633e-02 1.03411281e+00
1.14924446e-01 -2.18376294e-01 -8.57010245e-01 3.94766510e-01
-2.05949068e+00 -8.33821177e-01 -1.22753300e-01 1.85821092e+00
1.59450054e+00 3.52321148e-01 -4.43014920e-01 1.93696320e-01
2.57501125e-01 -3.66388053e-01 -5.41686416e-01 -3.89952064e-01
2.66621202e-01 3.79269272e-01 -3.35229523e-02 5.21860540e-01
-5.90007842e-01 5.69969416e-01 6.07537031e+00 1.18227708e+00
-6.70085549e-01 9.88902077e-02 6.39450431e-01 -1.51918009e-01
-3.89298439e-01 -3.63207400e-01 -7.83712864e-01 5.80197930e-01
1.04616106e+00 -2.04922080e-01 -1.24776088e-01 3.40039164e-01
3.01000774e-01 -2.04301998e-01 -1.31290698e+00 4.62954760e-01
-2.67367046e-02 -1.56002510e+00 2.10845575e-01 -4.47917767e-02
4.74264681e-01 -2.02033758e-01 2.76866667e-02 4.16450948e-01
7.88910910e-02 -9.97564316e-01 6.50970340e-01 6.74842596e-01
1.10892749e+00 -4.45523530e-01 1.29684818e+00 3.52019757e-01
-1.15969825e+00 1.75479889e-01 -1.39018863e-01 2.72157282e-01
2.20244169e-01 9.53045726e-01 -1.01423776e+00 1.13214314e+00
3.79812777e-01 6.74961925e-01 -2.36098900e-01 7.75740921e-01
-5.81974089e-01 4.46116537e-01 -2.27119163e-01 -1.25341326e-01
2.31559381e-01 1.74187958e-01 5.57797775e-02 1.49616277e+00
2.13515416e-01 3.47512007e-01 2.09972039e-02 9.39390123e-01
-4.47762668e-01 7.09359860e-03 -2.07535043e-01 -5.82085513e-02
6.43916249e-01 1.12932384e+00 -3.07422459e-01 -9.35316265e-01
-2.21365675e-01 6.54270589e-01 2.95229048e-01 2.28963062e-01
-1.21634209e+00 -4.47950035e-01 3.02895531e-02 -3.78198862e-01
1.91463307e-01 3.23487133e-01 -3.28689277e-01 -1.22138095e+00
3.11159849e-01 -1.10745013e+00 9.17953849e-01 -8.34751308e-01
-1.26390922e+00 5.36103666e-01 1.84338331e-01 -1.09013224e+00
-4.30961460e-01 -5.04199684e-01 -2.99771428e-01 9.76321518e-01
-1.78622758e+00 -9.77844596e-01 -4.11908738e-02 5.40965617e-01
3.68925452e-01 3.26505184e-01 8.17885280e-01 5.48393786e-01
-6.19106710e-01 8.28021944e-01 -3.80023897e-01 3.88929658e-02
6.69377625e-01 -1.00991035e+00 -6.88679367e-02 4.33560282e-01
-5.18189132e-01 1.20216525e+00 2.72804767e-01 -8.66747797e-01
-1.42400157e+00 -1.16761410e+00 1.33228111e+00 -3.30283314e-01
8.31169248e-01 1.22395642e-01 -9.15478170e-01 5.68448126e-01
-2.40045525e-02 -2.33727977e-01 1.00308323e+00 1.00808963e-01
-2.96881974e-01 1.93697527e-01 -1.26479185e+00 5.67958713e-01
6.88353896e-01 -6.57104611e-01 -9.09270346e-01 5.23080826e-01
8.60024512e-01 -6.69970870e-01 -1.55304205e+00 7.10490346e-01
6.45640254e-01 -1.97166592e-01 1.00907159e+00 -7.16347992e-01
1.01612353e+00 -2.27841556e-01 -8.85036141e-02 -7.95545042e-01
-4.41234559e-02 5.99595197e-02 -3.73526335e-01 6.62466049e-01
8.06686997e-01 -5.95472515e-01 1.97230637e-01 6.65437698e-01
-3.42636406e-01 -1.30251229e+00 -8.96239102e-01 -4.58664924e-01
-5.83764613e-02 -6.64876103e-01 6.01280332e-01 9.05201018e-01
7.59962320e-01 2.62438148e-01 -2.88491473e-02 1.51143655e-01
4.36855912e-01 3.90923411e-01 6.19859062e-02 -8.91225815e-01
-2.83762455e-01 -4.79818761e-01 7.61324316e-02 -8.38269055e-01
1.63142592e-01 -1.39788997e+00 1.54118732e-01 -1.85452592e+00
5.99809587e-01 -3.11860293e-01 -5.41679025e-01 8.56056094e-01
-5.14751196e-01 1.66292220e-01 -2.76160806e-01 1.34746805e-01
-8.31599355e-01 4.65931326e-01 1.32234156e+00 -1.29749507e-01
2.67484516e-01 -1.55789331e-01 -7.33630478e-01 4.87657219e-01
7.60138154e-01 -5.23116887e-01 -2.31907219e-01 -3.34206283e-01
5.00657618e-01 5.95039487e-01 3.59596908e-01 -4.51514482e-01
5.12626410e-01 5.03358282e-02 2.85150558e-01 -8.84442806e-01
2.51721144e-01 -6.00767970e-01 -1.45949066e-01 8.05947781e-01
-7.47834206e-01 1.07792981e-01 4.88968521e-01 4.56603944e-01
-3.19472343e-01 -6.05274677e-01 3.07257205e-01 -1.50490552e-01
-2.16644667e-02 3.11342925e-01 -4.18143034e-01 8.00885037e-02
5.22475541e-01 3.65741640e-01 -5.67604303e-01 7.94312079e-03
-5.76029956e-01 6.40314877e-01 -1.89871624e-01 -7.67860375e-03
1.08033657e+00 -1.24924338e+00 -9.34655726e-01 -1.99884102e-01
3.21501553e-01 1.82916328e-01 3.31244648e-01 1.35884666e+00
-3.26827705e-01 8.26196074e-01 1.77881300e-01 -4.73504543e-01
-1.20656526e+00 4.52689826e-01 1.02422349e-01 -8.96578133e-01
-5.98833323e-01 5.38014233e-01 -1.91665873e-01 -3.35166097e-01
2.78102547e-01 -8.57064009e-01 -1.49288401e-01 -1.56251565e-01
6.91917062e-01 1.34521320e-01 4.62788194e-01 1.08985431e-01
-4.88707989e-01 1.20284520e-01 -4.48434561e-01 -1.17127232e-01
1.13298213e+00 2.56819248e-01 -4.98282254e-01 1.03911944e-01
1.01371062e+00 -3.04870933e-01 -2.74653554e-01 -3.56886804e-01
2.75570055e-04 -1.10394070e-02 2.35820934e-01 -1.20176876e+00
-5.65553665e-01 5.82104445e-01 -1.73442915e-01 -5.96175119e-02
1.01904035e+00 1.04890401e-02 1.02873862e+00 4.91633981e-01
5.33534661e-02 -9.22626674e-01 1.16575591e-01 4.28314030e-01
9.81865048e-01 -1.10220075e+00 2.73825675e-01 -3.93229872e-01
-4.04131085e-01 9.18439150e-01 2.10625499e-01 4.27037537e-01
4.10283476e-01 3.64358723e-01 -2.79786110e-01 -5.86592019e-01
-1.13656652e+00 1.48448423e-01 6.80112123e-01 -3.27335954e-01
7.05059886e-01 7.40036219e-02 -7.90796041e-01 8.72263610e-01
1.04321435e-01 4.34913605e-01 1.52066439e-01 1.09089386e+00
-1.18478522e-01 -9.38824177e-01 -2.77811676e-01 7.29538620e-01
-8.20335031e-01 -5.60049474e-01 -1.09967947e-01 4.82474148e-01
-8.22837353e-02 1.14372826e+00 -3.76371861e-01 -1.16869725e-01
3.08416456e-01 1.44003376e-01 5.30202389e-01 -4.35154140e-01
-6.91028357e-01 6.43345192e-02 3.94884706e-01 -5.50264776e-01
-3.00947636e-01 -3.51234436e-01 -1.51793361e+00 -2.92727619e-01
-5.22255003e-01 3.61690849e-01 5.79493284e-01 1.17466390e+00
3.35018724e-01 9.86107230e-01 1.93487361e-01 1.85219962e-02
-7.43589282e-01 -1.15854907e+00 -2.04647839e-01 2.68241376e-01
1.82004049e-01 -4.32978451e-01 -4.62713279e-02 -8.34281668e-02]
|
[8.56010627746582, 8.669601440429688]
|
d43a4a13-f432-4f99-8bd7-c9e9c2a01b89
|
seamless-copy-move-manipulation-in-digital
|
2110.05747
| null |
https://arxiv.org/abs/2110.05747v1
|
https://arxiv.org/pdf/2110.05747v1.pdf
|
Seamless Copy Move Manipulation in Digital Images
|
The importance and relevance of digital image forensics has attracted researchers to establish different techniques for creating as well as detecting forgeries. The core category in passive image forgery is copy-move image forgery that affects the originality of image by applying a different transformation. In this paper frequency domain image manipulation method is being presented.The method exploits the localized nature of discrete wavelet transform (DWT) to get hold of the region of the host image to be manipulated. Both the patch and host image are subjected to DWT at the same level $l$ to get $3l + 1$ sub-bands and each sub-band of the patch is pasted to the identified region in the corresponding sub-band of the host image. The resultant manipulated host sub-bands are then subjected to inverse DWT to get the final manipulated host image. The proposed method shows good resistance against detection by two frequency domain forgery detection methods from the literature. The purpose of this research work is to create the forgery and highlight the need to produce forgery detection methods that are robust against the malicious copy-move forgery.
|
['Khizar Hayat', 'Mushtaq Ali', 'Tanzila Qazi']
|
2021-10-12
| null | null | null | null |
['image-forensics']
|
['computer-vision']
|
[ 6.92491531e-01 -3.72085840e-01 2.61617541e-01 2.49548435e-01
-4.11396474e-01 -5.82967639e-01 1.44160256e-01 -1.73526213e-01
-2.09210664e-01 4.29932445e-01 9.80101712e-03 -2.72217721e-01
-2.00686201e-01 -9.05425370e-01 -4.34119821e-01 -8.56000066e-01
-2.18111109e-02 -6.62134171e-01 3.33418518e-01 -2.56979495e-01
7.88887978e-01 6.55146837e-01 -1.31342816e+00 5.37809134e-01
3.00500453e-01 8.24343383e-01 2.15246707e-01 9.41743612e-01
7.67835826e-02 8.48373830e-01 -1.10436499e+00 -1.69046953e-01
7.60680735e-01 -9.51034486e-01 -7.24967360e-01 5.15149295e-01
2.05025859e-02 -3.88020128e-01 -6.40177786e-01 1.44246352e+00
3.63509327e-01 1.17732532e-01 4.21560258e-01 -1.20256507e+00
-9.59342778e-01 7.62563050e-02 -1.13697851e+00 1.12232459e+00
3.53530705e-01 9.29052457e-02 -8.07466060e-02 -6.28747880e-01
7.09448576e-01 1.16320992e+00 5.78767002e-01 2.18215976e-02
-7.84128785e-01 -7.74403036e-01 -8.45035255e-01 6.24344230e-01
-1.37298417e+00 -3.46308708e-01 1.33202028e+00 2.83466801e-02
8.26376557e-01 5.30853987e-01 2.70411849e-01 3.04873616e-01
7.64108479e-01 4.51028459e-02 1.47721398e+00 -9.56761122e-01
-2.66540021e-01 1.99856564e-01 5.70759252e-02 7.01504350e-01
2.60381401e-01 2.51749277e-01 -4.56710279e-01 -2.33383015e-01
7.74561286e-01 -6.62251050e-03 -6.17831528e-01 5.40432036e-01
-9.07935917e-01 7.06581712e-01 1.50316879e-01 9.25839007e-01
-4.75232422e-01 1.03543907e-01 4.59607869e-01 7.38878489e-01
1.67888820e-01 3.34210731e-02 2.22203299e-01 3.43836397e-02
-1.06829596e+00 1.55219883e-01 4.94596034e-01 4.56150949e-01
4.71168488e-01 2.68887430e-01 3.40181619e-01 4.75239009e-01
1.75457627e-01 2.85533488e-01 7.30543911e-01 -8.21035802e-01
3.38385284e-01 3.52182537e-01 -1.14917703e-01 -1.64876962e+00
2.61978894e-01 -2.21747205e-01 -4.29275513e-01 5.65353513e-01
3.10748667e-01 1.62627742e-01 -7.51527071e-01 1.03376389e+00
3.83020550e-01 1.12716779e-01 9.21946317e-02 6.18811786e-01
4.61022139e-01 9.52561915e-01 -2.62783289e-01 -2.42062271e-01
1.59703362e+00 -4.01799887e-01 -9.07319486e-01 5.31930849e-02
3.73703204e-02 -1.49738920e+00 5.72920799e-01 5.47922909e-01
-1.09053528e+00 -7.27443278e-01 -1.39029050e+00 3.17294210e-01
-3.81665677e-01 -1.07476056e-01 -6.26395941e-02 1.09359765e+00
-8.67889047e-01 4.30448830e-01 -2.59492725e-01 -1.07851960e-01
1.88096806e-01 1.95993215e-01 -4.97756541e-01 -2.49290183e-01
-9.54490066e-01 9.58400488e-01 4.75602686e-01 -1.61150590e-01
-5.18975675e-01 -3.98285925e-01 -4.50251877e-01 -1.22697428e-01
8.57788417e-03 -3.99101451e-02 4.99847263e-01 -1.29821718e+00
-9.06898141e-01 1.14092159e+00 8.81041437e-02 -4.70770657e-01
3.67998630e-01 5.97725093e-01 -1.02907181e+00 8.67775142e-01
2.14270130e-01 -1.90333873e-01 1.83321297e+00 -1.22905445e+00
-6.08114660e-01 -4.70820814e-01 -5.61396956e-01 -2.84430742e-01
-1.20055474e-01 4.47137982e-01 1.03011616e-01 -1.24200892e+00
3.79130900e-01 -3.75459105e-01 4.76495355e-01 -2.26989478e-01
-7.66849592e-02 2.88174629e-01 2.02006149e+00 -1.37594962e+00
1.31420267e+00 -2.65169835e+00 -5.00829279e-01 2.57680058e-01
8.25622901e-02 5.33725083e-01 2.03326009e-02 7.70176828e-01
-5.73364139e-01 8.20775107e-02 -4.28716183e-01 3.54744196e-01
-5.38066208e-01 7.51600116e-02 -3.73532742e-01 1.22568440e+00
9.51113328e-02 3.03647697e-01 -4.16141808e-01 -5.75447381e-01
2.20527276e-01 5.56707621e-01 3.24265435e-02 -4.51444760e-02
4.79667842e-01 1.66983590e-01 -3.57093096e-01 5.67616045e-01
1.32392550e+00 4.70186174e-01 -2.77335316e-01 -4.98757392e-01
-1.66659817e-01 -2.94103861e-01 -1.35677743e+00 8.22454751e-01
1.03409104e-01 9.37224209e-01 4.15083498e-01 -1.41664028e+00
1.22256291e+00 4.98224139e-01 3.80814195e-01 -6.31053150e-01
3.39084178e-01 2.65227199e-01 1.16831772e-02 -1.14757609e+00
6.40637159e-01 -5.64669311e-01 8.96847919e-02 7.59889483e-01
-7.76928663e-02 -4.55569960e-02 -6.85064420e-02 1.74582377e-03
1.09419596e+00 -8.47831666e-02 3.08863640e-01 -1.80402309e-01
8.36335599e-01 1.11251287e-01 2.20085099e-01 4.78230745e-01
-6.37667298e-01 3.70819390e-01 6.04511388e-02 -2.96879411e-01
-1.16373503e+00 -7.43970454e-01 -8.37357640e-02 5.12464523e-01
3.00802767e-01 3.02765965e-01 -9.37895238e-01 -4.39792007e-01
1.61394194e-01 5.53638399e-01 -6.00075722e-01 -3.86414200e-01
-8.28294992e-01 -7.31890082e-01 9.92467523e-01 -2.09263667e-01
1.13926125e+00 -1.14757097e+00 -1.01298499e+00 4.37047482e-01
-2.07909033e-01 -9.28576589e-01 -4.99136925e-01 -2.52169192e-01
-8.88859689e-01 -1.32712340e+00 -5.62500715e-01 -1.08564210e+00
6.67615294e-01 8.12590539e-01 2.85341442e-01 5.35107851e-01
-9.04165924e-01 2.83183813e-01 -5.87458968e-01 -2.09844291e-01
-8.69036555e-01 -8.83802116e-01 -3.66023719e-01 3.68676662e-01
2.43091866e-01 -6.24022603e-01 -5.89026511e-01 1.77719131e-01
-1.47902679e+00 -6.34259343e-01 4.77095664e-01 4.68088031e-01
9.57776010e-02 1.22483981e+00 3.69147480e-01 -4.37237769e-01
9.91098940e-01 -3.41158330e-01 -1.82637051e-01 7.09876698e-03
-2.40280136e-01 -1.78259835e-01 5.56812048e-01 -5.09932518e-01
-1.14825130e+00 -5.45258522e-01 2.03954950e-01 -3.91015351e-01
-7.88334385e-02 1.77188382e-01 6.65308982e-02 -6.57378733e-01
7.35049546e-01 8.81190360e-01 4.82626379e-01 -4.63484496e-01
7.02980533e-02 8.03166449e-01 1.04514003e+00 -9.18159485e-02
1.16787839e+00 8.22070301e-01 4.55262698e-02 -1.10972750e+00
3.43852699e-01 -5.70947766e-01 -1.69955909e-01 -4.51689154e-01
6.86326087e-01 -3.82389694e-01 -3.60378653e-01 7.30489671e-01
-1.12812030e+00 4.13309723e-01 6.76319450e-02 1.55061886e-01
-1.75534531e-01 1.24135590e+00 -7.74676740e-01 -9.73062754e-01
-5.39904475e-01 -9.57064748e-01 4.84280497e-01 1.95777684e-01
1.47125989e-01 -6.19531453e-01 -2.38206774e-01 4.59115118e-01
5.06091237e-01 7.22068489e-01 1.02307057e+00 -1.31816521e-01
-3.58271271e-01 -6.59256816e-01 -2.27263495e-01 5.41793585e-01
5.64119220e-01 -7.03991354e-02 -6.35524392e-01 -4.63526458e-01
1.18957281e+00 2.96150833e-01 5.68343997e-01 1.51497468e-01
6.02360010e-01 -3.85218352e-01 -1.87605634e-01 3.91008735e-01
1.92918205e+00 6.56169116e-01 1.20357883e+00 7.17976332e-01
-3.89524847e-02 5.30830562e-01 5.12359798e-01 4.34204370e-01
-3.92989695e-01 3.95842969e-01 2.50003070e-01 -4.45766114e-02
-4.94314998e-01 2.24417999e-01 3.07730377e-01 4.80709642e-01
-9.99701247e-02 -1.59298256e-01 -4.53309983e-01 5.62553346e-01
-1.15610027e+00 -1.56627679e+00 -4.48369890e-01 1.96783388e+00
4.74957734e-01 -2.78210286e-02 7.55203292e-02 8.93084645e-01
1.09920537e+00 2.62048483e-01 -4.59826551e-02 -6.67548776e-01
-1.61304921e-01 6.60747647e-01 8.33608091e-01 4.71521914e-01
-7.46452630e-01 4.58931327e-01 5.43889189e+00 1.11567545e+00
-1.30055594e+00 3.45908821e-01 2.64400691e-01 3.96971256e-01
1.72080547e-01 2.34300137e-01 -1.29616991e-01 8.55890512e-01
7.40492523e-01 -1.98050037e-01 3.00459057e-01 2.57453114e-01
5.28909206e-01 -7.21959770e-01 -3.63011584e-02 9.77612615e-01
4.00387436e-01 -1.01298368e+00 4.29523587e-02 3.28215033e-01
3.11502308e-01 -9.29173231e-01 1.97096065e-01 -4.15422827e-01
-3.82137865e-01 -8.09521794e-01 5.98072052e-01 2.64882922e-01
5.01019180e-01 -1.14043748e+00 5.12081921e-01 2.94448644e-01
-1.26868844e+00 -5.97565174e-02 -6.08365238e-01 4.98112254e-02
2.34212399e-01 4.69127297e-01 -7.28763640e-01 5.52133024e-01
7.28715181e-01 9.73638445e-02 -3.17315996e-01 8.09617817e-01
3.29500616e-01 5.67836404e-01 -9.90562066e-02 5.14137447e-01
3.70837361e-01 -4.27533001e-01 8.02459836e-01 1.20691729e+00
6.10935926e-01 3.66320103e-01 -4.51443821e-01 8.73396754e-01
6.66987151e-02 -1.25584409e-01 -7.02046335e-01 -1.73143923e-01
4.24517393e-01 7.07883775e-01 -9.96764421e-01 -4.24629182e-01
-2.74646968e-01 1.42270172e+00 -5.26105702e-01 2.03764856e-01
-6.43619776e-01 -9.38352585e-01 2.06814602e-01 2.84190834e-01
5.73923945e-01 -3.51630479e-01 -5.99004030e-02 -5.07966578e-01
1.10528030e-01 -1.22087359e+00 5.00654697e-01 -6.22949660e-01
-1.02159953e+00 4.06184584e-01 3.51639502e-02 -1.34761488e+00
1.63256511e-01 -3.00328642e-01 -8.73222470e-01 1.15451992e+00
-1.28094339e+00 -9.56729710e-01 -1.26306996e-01 9.05530274e-01
6.39374435e-01 -2.18999714e-01 4.03213352e-01 1.71017617e-01
-1.56538084e-01 3.08510989e-01 6.66275397e-02 3.47468764e-01
5.31591952e-01 -4.93543714e-01 2.50574589e-01 1.41844988e+00
-2.88450152e-01 5.76436162e-01 9.38131750e-01 -1.04609752e+00
-1.37310290e+00 -5.33804119e-01 8.24946463e-01 1.21782340e-01
2.95348883e-01 2.40977541e-01 -1.03867805e+00 2.40187570e-01
6.19513750e-01 1.72249856e-03 5.03628492e-01 -1.49666226e+00
-3.20624411e-01 1.75091341e-01 -2.02073288e+00 -4.69311420e-03
2.78241247e-01 -6.99557781e-01 -9.26575363e-01 1.27800088e-02
1.57360435e-01 -2.76172031e-02 -7.34911382e-01 -8.66226554e-02
4.16326910e-01 -1.33757305e+00 1.34607315e+00 -7.43872076e-02
2.23786280e-01 -5.95333278e-01 -1.85448095e-01 -6.11961484e-01
-1.60759374e-01 -8.40536773e-01 2.48547956e-01 1.01296663e+00
-2.59601951e-01 -7.27667332e-01 5.76175869e-01 -1.43341459e-02
2.75688767e-01 -1.43241547e-02 -1.03814101e+00 -7.79986799e-01
-3.44885498e-01 1.20326029e-02 2.60144889e-01 1.25154769e+00
3.04506104e-02 -5.35341144e-01 -3.87279749e-01 3.21125686e-01
1.05891597e+00 -4.92423736e-02 3.44675243e-01 -4.18201536e-01
-4.05056447e-01 -1.65270016e-01 -7.23384261e-01 -1.98116541e-01
-5.42566240e-01 -4.94020700e-01 -3.98177683e-01 -9.95874763e-01
-1.53490216e-01 9.62017919e-04 1.44568622e-01 6.47059977e-02
-2.06095591e-01 7.57620454e-01 3.64452571e-01 6.28078640e-01
5.81205904e-01 -1.70931011e-01 1.12940073e+00 -1.94227081e-02
4.67919745e-02 -2.06502095e-01 -3.62288386e-01 3.42931747e-01
8.17406595e-01 -8.32296610e-01 -2.44181827e-01 6.78443089e-02
-5.51304817e-01 4.23025876e-01 5.46760082e-01 -1.07023752e+00
-6.14836402e-02 -2.14309059e-02 5.14816582e-01 -4.89687502e-01
9.83268619e-02 -9.72036362e-01 4.98412281e-01 1.00027418e+00
2.74240971e-01 5.35505235e-01 2.28652984e-01 6.20522261e-01
-3.76117378e-01 -8.02972615e-01 1.18014622e+00 -5.97990334e-01
-7.70593703e-01 -3.90356153e-01 -6.53973162e-01 -3.96267742e-01
1.34703326e+00 -1.15634692e+00 -2.05265835e-01 -2.85421431e-01
-4.00023460e-01 -8.86432827e-01 5.00763416e-01 5.46929194e-03
9.28466678e-01 -9.55895126e-01 -6.21187210e-01 3.85072887e-01
-4.92555588e-01 -8.49636912e-01 4.09088045e-01 6.30829394e-01
-1.09151983e+00 7.50935869e-03 -7.10057676e-01 5.53770438e-02
-1.79246855e+00 9.52087939e-01 2.38234058e-01 -7.80590391e-03
-8.92179906e-01 5.40943325e-01 -5.28555512e-01 7.38078773e-01
-2.73484170e-01 1.31751612e-01 -8.74344260e-02 -2.01345637e-01
7.88384199e-01 8.40971649e-01 -4.99695763e-02 -1.15044546e+00
-2.34877333e-01 7.53901958e-01 9.01942998e-02 -3.19180101e-01
1.24635828e+00 -3.60590905e-01 -6.53894007e-01 -4.49045658e-01
1.88253772e+00 2.86310375e-01 -8.68506610e-01 -4.02204357e-02
9.11839381e-02 -1.16762877e+00 3.30314159e-01 -4.30979460e-01
-1.19916546e+00 4.00328308e-01 9.14385557e-01 5.56152225e-01
1.56292629e+00 -3.30646604e-01 1.10725379e+00 -4.20675248e-01
1.93359077e-01 -1.07525897e+00 2.57633746e-01 -3.59990537e-01
8.56452584e-01 -5.57139575e-01 2.43797675e-01 -5.76471388e-01
-2.31616423e-01 1.46600544e+00 -1.04884900e-01 -6.38964832e-01
5.98541915e-01 3.88231963e-01 6.97955638e-02 -4.36432660e-01
1.90523654e-01 1.89365968e-01 -1.58475906e-01 9.38795030e-01
1.32802397e-01 -3.42829823e-01 -7.95487106e-01 -7.51388296e-02
-1.79289594e-01 2.72129364e-02 8.32643628e-01 1.60901928e+00
-7.23420084e-01 -1.26118398e+00 -1.54122686e+00 2.01396942e-01
-9.89826083e-01 1.86707646e-01 -4.07744765e-01 9.09838021e-01
3.29229385e-01 1.32004654e+00 -2.53728658e-01 -2.23954663e-01
1.04524568e-01 1.69307604e-01 7.40502715e-01 8.05927068e-02
-8.15149128e-01 1.26383588e-01 -2.37724543e-01 -2.59686887e-01
-4.92126644e-01 -4.76416498e-01 -1.15012956e+00 -6.82952344e-01
-2.56806523e-01 9.83137563e-02 7.64231265e-01 7.02439904e-01
-7.17968941e-02 2.56701946e-01 8.41043234e-01 -5.76583147e-01
-3.88840288e-01 -8.71194601e-01 -8.88423502e-01 8.31702292e-01
6.11179829e-01 -3.15683663e-01 -6.69318557e-01 6.18282259e-01]
|
[12.354524612426758, 0.9405251145362854]
|
b1c241a2-58a2-4e76-add1-ddaa8c818f86
|
reduced-gate-convolutional-lstm-design-using
| null | null |
https://openreview.net/forum?id=rJEyrjRqYX
|
https://openreview.net/pdf?id=rJEyrjRqYX
|
Reduced-Gate Convolutional LSTM Design Using Predictive Coding for Next-Frame Video Prediction
|
Spatiotemporal sequence prediction is an important problem in deep learning. We
study next-frame video prediction using a deep-learning-based predictive coding
framework that uses convolutional, long short-term memory (convLSTM) modules.
We introduce a novel reduced-gate convolutional LSTM architecture. Our
reduced-gate model achieves better next-frame prediction accuracy than the original
convolutional LSTM while using a smaller parameter budget, thereby reducing
training time. We tested our reduced gate modules within a predictive coding architecture
on the moving MNIST and KITTI datasets. We found that our reduced-gate
model has a significant reduction of approximately 40 percent of the total
number of training parameters and training time in comparison with the standard
LSTM model which makes it attractive for hardware implementation especially
on small devices.
|
['Magdy Bayoumi', 'Anthony S. Maida', 'Nelly Elsayed']
|
2018-09-27
| null | null | null | null |
['video-prediction']
|
['computer-vision']
|
[ 2.36311316e-01 -7.92458355e-02 -5.81576049e-01 -5.57097673e-01
-4.60477382e-01 1.92149192e-01 3.23663831e-01 -4.23144937e-01
-3.92553747e-01 6.48717940e-01 1.58951044e-01 -5.14006317e-01
4.85445321e-01 -6.65578246e-01 -1.14521408e+00 -4.96269315e-01
-2.81256646e-01 -1.13011554e-01 7.69241333e-01 8.27033371e-02
1.76519305e-01 1.79908976e-01 -1.32617617e+00 1.13112175e+00
4.10595179e-01 1.41708827e+00 6.85100198e-01 1.02060390e+00
1.64721072e-01 2.07229829e+00 -1.40170872e-01 -2.44671687e-01
9.13302600e-02 -1.84861138e-01 -8.56092215e-01 -5.63082933e-01
5.06500423e-01 -4.71248001e-01 -1.09329593e+00 5.85820198e-01
3.14738601e-01 4.03711170e-01 1.25123948e-01 -8.35038483e-01
-3.67373317e-01 4.83623147e-01 -3.39075997e-02 7.64626741e-01
2.86666723e-03 3.92031893e-02 7.58930504e-01 -7.61829317e-01
4.27666068e-01 9.64268863e-01 1.14989281e+00 7.33637929e-01
-7.52588332e-01 -7.64883280e-01 4.64030318e-02 7.10996389e-01
-1.23411393e+00 -5.08451760e-01 2.89551079e-01 -3.91268790e-01
2.07322478e+00 -1.61030307e-01 9.07243371e-01 1.05313909e+00
1.03143537e+00 9.01550293e-01 2.12226525e-01 -2.35023052e-01
8.75800848e-02 -5.63563168e-01 -9.61490050e-02 1.00979829e+00
-3.10833454e-01 2.53312975e-01 -7.44822860e-01 1.37853190e-01
1.03333247e+00 4.76656228e-01 -6.79032207e-02 4.42664549e-02
-1.17340398e+00 7.14105427e-01 6.40263319e-01 2.78887361e-01
-2.13927388e-01 1.12181735e+00 8.45948815e-01 2.39096835e-01
6.98964357e-01 -3.97474244e-02 -5.90214133e-01 -7.22357452e-01
-1.28262353e+00 1.04804505e-02 5.75350881e-01 1.15451205e+00
4.44468826e-01 7.07322419e-01 -2.35351712e-01 5.39776146e-01
8.70999098e-02 1.40337229e-01 8.31125379e-01 -1.04297066e+00
6.53698504e-01 2.19112784e-01 -2.07603827e-01 -8.92118573e-01
-3.52721483e-01 -2.47994289e-01 -9.38561618e-01 -8.34154934e-02
-2.66352683e-01 -2.11237535e-01 -1.09196329e+00 1.26266956e+00
-5.02076447e-01 8.75261664e-01 6.94820192e-03 6.82608902e-01
8.19535851e-01 1.27005768e+00 5.03822029e-01 -4.04064730e-03
8.33339989e-01 -1.59395778e+00 -5.20681858e-01 -5.43592751e-01
1.28272355e+00 -3.59620869e-01 5.88476181e-01 1.27707765e-01
-9.42647636e-01 -1.07670915e+00 -1.16654563e+00 -4.85780180e-01
-5.05098514e-02 3.86301935e-01 8.30714941e-01 3.89508516e-01
-1.52533174e+00 1.05917788e+00 -1.21985173e+00 -2.38157511e-01
5.34237742e-01 6.38611197e-01 -2.45037526e-01 5.66934906e-02
-9.88440156e-01 8.48268747e-01 7.68723249e-01 -2.06661504e-02
-8.39859009e-01 -6.90638661e-01 -9.52417791e-01 4.22361255e-01
-2.08630815e-01 -6.13630831e-01 1.42111874e+00 -1.18108988e+00
-1.57333982e+00 4.65329319e-01 -7.18482733e-01 -1.32147551e+00
3.79432663e-02 -3.35060179e-01 -6.01103723e-01 7.02500269e-02
-3.38636547e-01 1.06184077e+00 8.49293768e-01 -1.98498920e-01
-8.31493139e-01 1.96383014e-01 -1.68850213e-01 -1.04657523e-01
-1.52036905e-01 -5.84510565e-02 -3.85635585e-01 -9.02032256e-01
-2.06645399e-01 -1.16860139e+00 -4.88194585e-01 -1.95594877e-02
1.82483345e-01 -2.01076970e-01 1.15959585e+00 -6.90199435e-01
1.38903296e+00 -1.95950377e+00 -3.93311292e-01 -2.41397589e-01
2.72348464e-01 7.12451458e-01 3.23704630e-02 1.97742078e-02
-1.63186833e-01 3.01189385e-02 2.24852070e-01 -4.01671767e-01
-5.92966616e-01 4.27701116e-01 -5.60978889e-01 1.51000589e-01
9.08929482e-02 1.19340992e+00 -7.44706810e-01 -5.30856907e-01
4.96923715e-01 5.41087210e-01 -7.95039594e-01 6.96248114e-02
-3.08151871e-01 1.81178808e-01 -3.34512502e-01 5.22638619e-01
1.97121799e-01 -4.98214930e-01 2.02568352e-01 -1.18755288e-01
-1.46753527e-02 5.00637591e-01 -1.29148275e-01 1.91482174e+00
-5.20159364e-01 1.52047479e+00 -9.11404908e-01 -1.11736453e+00
9.59732950e-01 5.40082872e-01 4.54185426e-01 -8.36853027e-01
2.15705723e-01 1.75565034e-01 -5.35780750e-02 -4.28857803e-01
7.19525456e-01 -6.04170524e-02 2.12083727e-01 -3.38649862e-02
3.76319528e-01 6.57777309e-01 -1.07078120e-01 -5.66594042e-02
1.31009603e+00 4.80229914e-01 1.65699005e-01 -3.55044603e-01
1.96372837e-01 -7.87700489e-02 6.34596467e-01 6.47982895e-01
-4.05216187e-01 4.55636799e-01 2.57206380e-01 -1.35136724e+00
-1.30741322e+00 -4.56327975e-01 2.53250122e-01 1.38227439e+00
-5.64304367e-02 -8.18964243e-01 -5.87439656e-01 -3.09057653e-01
-3.89711380e-01 6.26620293e-01 -5.37824154e-01 -3.94179165e-01
-1.09627867e+00 -3.31479818e-01 8.66663396e-01 1.10501707e+00
7.74071753e-01 -1.08531153e+00 -1.23012149e+00 4.84481603e-01
-1.14774249e-01 -1.36545813e+00 -3.51242423e-01 3.11071426e-01
-1.36968780e+00 -5.55872023e-01 -4.20405239e-01 -1.22231889e+00
3.15682441e-02 6.40877634e-02 1.21403301e+00 4.48312387e-02
2.37848446e-01 -3.11279833e-01 -4.39803839e-01 -1.31603643e-01
-2.25186110e-01 2.27434441e-01 -1.73922047e-01 -3.97955626e-01
5.48962474e-01 -3.99682909e-01 -7.24019766e-01 -8.60265195e-02
-3.18659365e-01 6.29761338e-01 3.87884974e-01 9.01150584e-01
5.95727026e-01 -4.20137614e-01 2.18348145e-01 -6.17108762e-01
3.08658965e-02 -5.39780378e-01 -4.58863974e-01 2.37075552e-01
-3.08957607e-01 1.56320825e-01 8.98409843e-01 -3.98058832e-01
-1.10328007e+00 2.71167040e-01 -4.26905572e-01 -9.59337115e-01
2.64075041e-01 4.02993888e-01 6.08984351e-01 -4.86427754e-01
2.47037634e-01 4.99800712e-01 -2.04192728e-01 -3.08578551e-01
-1.14919677e-01 3.46119732e-01 4.58039522e-01 4.88633029e-02
-1.37156874e-01 2.57612407e-01 1.33278236e-01 -7.34046876e-01
-4.45959866e-01 -7.74723105e-03 -5.73681653e-01 -2.71365762e-01
8.71754587e-01 -1.42049098e+00 -8.95283759e-01 3.17343652e-01
-1.32836461e+00 -9.83914495e-01 -7.35905468e-02 6.56664789e-01
-8.12068343e-01 5.15225269e-02 -1.20274889e+00 -4.09586281e-01
-5.82081139e-01 -1.28075862e+00 8.15500557e-01 6.75103441e-02
-1.87247768e-01 -1.38503468e+00 -1.39305532e-01 -2.25324463e-02
7.63021290e-01 1.47272363e-01 7.42185593e-01 -4.11450475e-01
-8.07423592e-01 -2.45310083e-01 -1.74949691e-01 2.15490431e-01
-2.88954377e-01 -9.26376507e-02 -8.64711046e-01 -2.27541000e-01
-4.88681309e-02 -2.76639432e-01 1.37168658e+00 9.13666844e-01
1.81107354e+00 -1.63734749e-01 -6.82537198e-01 1.14710069e+00
1.49451649e+00 6.31450832e-01 1.07839084e+00 3.14677119e-01
9.07361746e-01 -2.45548502e-01 4.63914633e-01 4.70789671e-01
3.42881203e-01 6.69791520e-01 1.50631979e-01 5.88745102e-02
-1.25155628e-01 -5.32263279e-01 5.64502597e-01 9.88872290e-01
-3.02800417e-01 -4.64170188e-01 -9.86153305e-01 4.57168788e-01
-2.11021876e+00 -1.33637977e+00 -1.07158147e-01 1.67964220e+00
3.82166713e-01 2.67520934e-01 -2.77708083e-01 -2.96049207e-01
5.70394337e-01 4.39567566e-01 -4.00097340e-01 -8.17490041e-01
6.84750676e-02 4.95062590e-01 9.46254611e-01 3.00093472e-01
-1.41926396e+00 1.41031659e+00 7.24058247e+00 1.15483415e+00
-1.48822439e+00 2.67333359e-01 1.08300149e+00 -4.14980650e-01
3.04226369e-01 -1.48113389e-02 -1.00213194e+00 6.73692167e-01
1.74445450e+00 1.88420974e-02 -6.61670566e-02 1.18086100e+00
3.43954355e-01 3.58479694e-02 -1.16737449e+00 1.28095341e+00
5.26736341e-02 -2.26416802e+00 2.74847418e-01 -2.46363416e-01
7.42377281e-01 7.00191915e-01 -7.68212005e-02 4.98454779e-01
1.62187427e-01 -1.31937528e+00 8.28189075e-01 5.97383738e-01
1.04965413e+00 -8.28058064e-01 8.30150783e-01 1.69804603e-01
-1.60456252e+00 -3.96441638e-01 -8.23200822e-01 -6.14365876e-01
3.56288970e-01 1.72548756e-01 -4.82104629e-01 -1.51739895e-01
1.00364876e+00 1.34615993e+00 -4.60170805e-01 9.20061052e-01
4.27455485e-01 9.36411023e-01 5.12308963e-02 8.00642744e-03
5.68485796e-01 2.83172458e-01 4.50178534e-02 1.45441914e+00
6.31000221e-01 2.58104831e-01 1.51673228e-01 3.67242694e-01
-2.43432224e-01 -5.23112953e-01 -6.02823198e-01 1.61628410e-01
2.23423228e-01 5.85949481e-01 -5.52275121e-01 -7.39408433e-01
-5.50997376e-01 1.15387166e+00 5.09452045e-01 2.32512638e-01
-1.21245825e+00 -2.74748743e-01 7.48680234e-01 6.87822327e-02
7.89259076e-01 -2.83029556e-01 -3.35733622e-01 -1.18730426e+00
-1.38369232e-01 -2.49114916e-01 1.53437853e-01 -8.19180310e-01
-5.05502045e-01 1.00730944e+00 -3.28127474e-01 -1.54356551e+00
-6.97740734e-01 -6.74857974e-01 -5.70067286e-01 4.67677861e-01
-1.44774091e+00 -1.25031793e+00 -1.22022614e-01 6.22543216e-01
9.98475611e-01 -4.74031746e-01 8.41518939e-01 3.44435453e-01
-5.65267444e-01 6.60192788e-01 2.67286628e-01 2.07730517e-01
1.65231735e-01 -5.52222490e-01 1.01010311e+00 8.93337250e-01
-4.77452725e-02 3.87563080e-01 4.24481213e-01 -6.39000356e-01
-1.18721688e+00 -1.70091271e+00 1.04768741e+00 -4.38586436e-02
5.02223730e-01 -1.77872241e-01 -8.04422796e-01 1.34327233e+00
2.87146986e-01 5.73930323e-01 6.48203015e-01 -4.74533081e-01
-8.30117241e-02 -1.31848276e-01 -6.89193130e-01 3.77844095e-01
1.21267545e+00 -5.80898106e-01 -1.13500789e-01 2.06212416e-01
9.95335221e-01 -4.45233315e-01 -9.55935776e-01 3.71583402e-01
6.87774360e-01 -9.55855310e-01 8.27567637e-01 -4.70947951e-01
8.28579426e-01 1.35247305e-01 -1.97399721e-01 -9.56656575e-01
-7.33095944e-01 -4.34377670e-01 -6.93650305e-01 3.17458481e-01
4.02434349e-01 -3.85903418e-01 1.24704468e+00 6.85751438e-01
-6.15825415e-01 -1.14917421e+00 -1.09901202e+00 -7.54805923e-01
-4.48506400e-02 -6.15853488e-01 1.27574116e-01 5.40670455e-01
1.96098357e-01 1.11847758e-01 -9.66502130e-01 -3.88139993e-01
6.11134991e-02 -1.14216194e-01 2.44757280e-01 -6.72932804e-01
-1.25627056e-01 -3.40210460e-02 -8.97813857e-01 -1.59943783e+00
3.35767478e-01 -6.62962794e-01 2.49558184e-02 -1.31229997e+00
1.34913445e-01 -2.62149990e-01 -3.06856573e-01 6.39668345e-01
2.38952145e-01 4.43185538e-01 2.14050591e-01 3.35633427e-01
-8.68821859e-01 6.74377620e-01 7.47757375e-01 -6.10263459e-02
-5.55861518e-02 -2.47196794e-01 1.30824059e-01 8.81498098e-01
8.06926191e-01 -3.83249849e-01 -3.93172741e-01 -9.88672733e-01
-1.11683406e-01 3.26838672e-01 3.47736269e-01 -1.81232953e+00
5.35388470e-01 4.03787419e-02 7.85434186e-01 -6.34265661e-01
5.84885061e-01 -5.62914014e-01 1.94125757e-01 9.74373877e-01
-5.60349584e-01 3.75713289e-01 4.79966700e-01 4.86960232e-01
-4.23956782e-01 1.81083724e-01 7.46698618e-01 -1.64669245e-01
-1.70296729e+00 4.86595541e-01 -9.79724824e-01 -4.72829521e-01
1.08889556e+00 -6.31370604e-01 -2.29500756e-01 -4.34511095e-01
-6.15900278e-01 -9.04570073e-02 2.26827160e-01 4.56512392e-01
1.10586393e+00 -1.39221895e+00 -3.27010810e-01 1.85024738e-01
-1.11079097e-01 -4.36362714e-01 5.12471557e-01 6.55638874e-01
-1.24909580e+00 1.02753055e+00 -4.61667418e-01 -6.04423285e-01
-1.17658520e+00 4.88113701e-01 6.71216130e-01 -1.93296164e-01
-9.26728666e-01 1.14656448e+00 1.10357545e-01 2.72198915e-01
7.33784065e-02 -7.41879344e-01 -2.61931986e-01 -6.13926649e-01
6.72073960e-01 3.67504358e-01 -2.74277385e-02 -7.76888728e-01
-3.37003946e-01 3.35225046e-01 -4.17382903e-02 5.24464786e-01
1.26598752e+00 6.44090697e-02 1.76509261e-01 4.34261829e-01
1.56641352e+00 -1.05766559e+00 -1.48095548e+00 -1.65804010e-02
1.01254582e-02 -3.84941608e-01 3.22905302e-01 -1.76558495e-02
-1.34791195e+00 1.03621781e+00 8.75157416e-01 -4.45742399e-01
1.02713525e+00 -3.22472990e-01 1.37103772e+00 6.98867500e-01
6.34632587e-01 -9.71263051e-01 -1.79874226e-02 9.51980472e-01
3.73514295e-01 -1.15116715e+00 -1.74707234e-01 -2.14965507e-01
-6.49960816e-01 1.50266385e+00 8.80696774e-01 -6.47756755e-01
9.13856387e-01 3.46643209e-01 -4.76826817e-01 -1.86014865e-02
-1.55826914e+00 1.81915671e-01 1.99656770e-01 4.29922789e-01
7.73103595e-01 8.74937847e-02 1.92837372e-01 4.51713324e-01
-1.27587438e-01 6.59245551e-01 3.59633327e-01 9.50941503e-01
-5.30674279e-01 -6.33142471e-01 3.28687400e-01 8.97302032e-01
-6.23678923e-01 -5.01364648e-01 3.41414809e-01 3.01637173e-01
2.17322469e-01 6.69629872e-01 6.98618174e-01 -1.08321381e+00
-2.93856949e-01 -2.34559346e-02 2.89458990e-01 -3.92244011e-01
-5.52568316e-01 -2.75962532e-01 1.10331848e-01 -1.03280187e+00
-5.79603016e-01 -1.09675303e-01 -1.20517337e+00 -8.34731638e-01
-1.02941208e-01 -2.01250613e-01 1.75224677e-01 1.23493099e+00
8.03991556e-01 6.67261600e-01 2.74096560e-02 -1.29403210e+00
2.21872047e-01 -9.53521729e-01 -2.20314324e-01 -1.61291182e-01
4.36371535e-01 -4.18407112e-01 5.06389022e-01 5.25139689e-01]
|
[8.950385093688965, 0.38918495178222656]
|
937c8cd0-777f-44d5-be3f-c336311266d3
|
combining-word-embeddings-and-n-grams-for
|
2004.14119
| null |
https://arxiv.org/abs/2004.14119v1
|
https://arxiv.org/pdf/2004.14119v1.pdf
|
Combining Word Embeddings and N-grams for Unsupervised Document Summarization
|
Graph-based extractive document summarization relies on the quality of the sentence similarity graph. Bag-of-words or tf-idf based sentence similarity uses exact word matching, but fails to measure the semantic similarity between individual words or to consider the semantic structure of sentences. In order to improve the similarity measure between sentences, we employ off-the-shelf deep embedding features and tf-idf features, and introduce a new text similarity metric. An improved sentence similarity graph is built and used in a submodular objective function for extractive summarization, which consists of a weighted coverage term and a diversity term. A Transformer based compression model is developed for sentence compression to aid in document summarization. Our summarization approach is extractive and unsupervised. Experiments demonstrate that our approach can outperform the tf-idf based approach and achieve state-of-the-art performance on the DUC04 dataset, and comparable performance to the fully supervised learning methods on the CNN/DM and NYT datasets.
|
['Sanjay Krishna', 'Richard Schwartz', 'Manaj Srivastava', 'David Akodes', 'Zhuolin Jiang']
|
2020-04-25
| null | null | null | null |
['sentence-compression', 'extractive-document-summarization']
|
['natural-language-processing', 'natural-language-processing']
|
[ 4.73873466e-01 1.51664108e-01 -3.13294619e-01 -4.32504117e-01
-8.54458451e-01 -3.95305306e-01 5.64910114e-01 1.03278267e+00
-3.88236195e-01 5.52287996e-01 1.24176347e+00 1.09497838e-01
-3.56493622e-01 -9.01870906e-01 -4.84223455e-01 -3.67840827e-01
-8.86112079e-02 4.11936700e-01 -1.19870208e-01 -4.40599889e-01
8.53865206e-01 4.31793332e-02 -1.30187142e+00 4.52063948e-01
1.24279952e+00 9.54442978e-01 2.97161639e-01 9.32476103e-01
-5.87446153e-01 6.51811004e-01 -8.68799329e-01 -5.00039279e-01
8.37300345e-03 -7.30265379e-01 -9.89457250e-01 -2.11631004e-02
8.95331025e-01 -3.21091294e-01 -7.14737833e-01 1.08325410e+00
7.55466402e-01 2.80051529e-01 8.38111818e-01 -6.86056554e-01
-8.55542839e-01 8.98574352e-01 -4.89377618e-01 4.19147581e-01
7.64569819e-01 -3.61333758e-01 1.74651766e+00 -7.24426091e-01
6.71426177e-01 1.26607251e+00 6.26763761e-01 2.80027002e-01
-9.74311888e-01 3.23519111e-02 -1.25897646e-01 3.16864014e-01
-8.42503846e-01 -3.44890326e-01 7.77532756e-01 9.15666148e-02
1.52432466e+00 4.70169604e-01 7.05728292e-01 6.36614263e-01
6.39316738e-01 1.00165546e+00 3.48262310e-01 -5.62043846e-01
2.22198367e-01 -4.00777578e-01 2.69653112e-01 9.36812878e-01
4.93371934e-01 -5.80215693e-01 -7.94182360e-01 -5.88675626e-02
3.36200818e-02 1.89271033e-01 -3.75631541e-01 -1.64140865e-01
-1.23538923e+00 1.09258842e+00 5.68373680e-01 5.84090292e-01
-3.78997952e-01 1.25757709e-01 9.25799370e-01 4.33931559e-01
8.86794508e-01 9.54407513e-01 -6.28633425e-02 -1.91759601e-01
-1.53024077e+00 4.16717350e-01 8.93084705e-01 7.81024456e-01
6.59883082e-01 2.03761496e-02 -8.29137623e-01 9.31161344e-01
-1.60245091e-01 3.57850105e-01 8.19443703e-01 -1.03621316e+00
8.45578611e-01 8.79738927e-01 -5.67688882e-01 -1.30728507e+00
-3.08891982e-01 -5.86385667e-01 -1.15613472e+00 -5.85164309e-01
-4.21820641e-01 2.87311882e-01 -7.74845362e-01 1.21305120e+00
-1.79684877e-01 -9.90761518e-02 1.21416837e-01 5.32808721e-01
1.25724339e+00 7.47054517e-01 -6.36853576e-01 -3.11883390e-01
1.07668483e+00 -1.30425990e+00 -9.23413038e-01 -2.10476488e-01
9.44875717e-01 -6.02655113e-01 7.72757709e-01 -9.38058347e-02
-1.49951816e+00 -3.88943285e-01 -1.28716075e+00 -4.67287004e-01
-3.06476027e-01 -3.68634947e-02 4.48857635e-01 2.17494324e-01
-1.28586483e+00 1.36508358e+00 -5.69501042e-01 -5.71649730e-01
6.32568359e-01 1.86223537e-01 -4.99164850e-01 -2.49812841e-01
-1.08427882e+00 9.07573760e-01 7.98462510e-01 -4.66012985e-01
-2.95192093e-01 -7.00895488e-01 -1.25290978e+00 6.41597629e-01
-3.21453623e-02 -1.15807068e+00 9.65574503e-01 -5.82507074e-01
-1.31236899e+00 6.91443324e-01 -3.94582689e-01 -1.03315282e+00
-7.23749818e-03 -9.17521715e-02 -1.23937398e-01 7.85634696e-01
3.82716119e-01 6.58429265e-01 7.10099638e-01 -7.05899298e-01
-5.79316795e-01 -4.31836605e-01 -6.83470145e-02 5.10352790e-01
-7.85064936e-01 -1.31993771e-01 -9.65173095e-02 -9.06711996e-01
-7.78403739e-03 -4.24671441e-01 -1.07434660e-01 -4.78070974e-01
-6.22650206e-01 -4.00223076e-01 7.25270808e-01 -1.05265999e+00
1.58926523e+00 -1.71319282e+00 4.79679346e-01 -1.06891226e-02
5.43027222e-01 3.31932306e-01 -5.73264897e-01 1.00066519e+00
1.96819797e-01 1.65959463e-01 -6.73223317e-01 -6.53473854e-01
4.54813428e-02 3.67342457e-02 -1.95899352e-01 1.75732061e-01
2.49473795e-01 1.09103549e+00 -1.00845480e+00 -7.22429812e-01
-1.77838560e-02 1.37831151e-01 -5.20810664e-01 4.79424633e-02
-7.81413764e-02 -2.46107236e-01 -3.44614953e-01 2.57688552e-01
4.10602301e-01 -6.06106520e-02 1.02452310e-02 -2.48722494e-01
2.12678388e-01 7.49316156e-01 -3.94008070e-01 2.42670512e+00
-4.43888605e-01 8.34384739e-01 -5.15264153e-01 -1.43448603e+00
1.06287754e+00 3.27146687e-02 5.01246929e-01 -8.13513875e-01
2.41309136e-01 2.85498053e-01 -1.86295599e-01 -2.78463900e-01
1.23113751e+00 2.97609478e-01 -1.19052567e-01 6.10053241e-01
5.25287747e-01 -3.94946933e-01 8.20941031e-01 9.03234899e-01
1.54450405e+00 -3.39080721e-01 4.50276822e-01 -4.26012397e-01
5.69470227e-01 -1.20945096e-01 8.73005241e-02 7.31023192e-01
3.75936687e-01 9.56865668e-01 6.55969143e-01 -1.60906807e-01
-1.27221966e+00 -8.46380532e-01 2.48749912e-01 8.77253592e-01
4.34067212e-02 -1.02967048e+00 -9.39200044e-01 -8.75994682e-01
6.85741380e-02 1.09271324e+00 -4.60949421e-01 -6.83569729e-01
-6.33750439e-01 -3.33678216e-01 4.37016547e-01 5.83703995e-01
5.12454033e-01 -8.22158217e-01 -2.89259136e-01 3.20569307e-01
-5.75878024e-01 -8.47400904e-01 -1.04772496e+00 -3.51677672e-03
-1.26825285e+00 -6.55219436e-01 -7.89653003e-01 -9.13401067e-01
5.01711607e-01 5.09155273e-01 1.22473514e+00 7.34048560e-02
-2.53620803e-01 3.73630941e-01 -7.19894648e-01 -2.30003506e-01
-4.80988950e-01 4.38686192e-01 -1.49605870e-01 -3.14099848e-01
1.97606966e-01 -6.53762162e-01 -7.29287684e-01 -4.84564453e-01
-1.05674613e+00 -2.91101754e-01 5.51513851e-01 9.98436332e-01
4.68847066e-01 -8.71033221e-02 7.95870960e-01 -4.85389531e-01
1.33050942e+00 -5.35681173e-02 3.39737982e-01 2.76178271e-01
-8.12268257e-01 5.62745214e-01 7.29191422e-01 2.53991466e-02
-6.60651326e-01 -5.07122695e-01 -1.20454356e-01 -2.53480703e-01
4.50907946e-01 9.54967380e-01 1.92220539e-01 2.29056418e-01
5.77953398e-01 6.07827187e-01 1.02844477e-01 -4.54660654e-01
4.98560131e-01 7.35948741e-01 6.96209729e-01 1.13988370e-02
4.55712944e-01 4.01689142e-01 1.04602613e-01 -8.82330537e-01
-1.02045310e+00 -7.53577232e-01 -6.31011963e-01 1.36134893e-01
6.10320687e-01 -7.34570324e-01 -1.80801675e-01 3.42780771e-03
-1.43237495e+00 5.03863633e-01 -7.21367121e-01 3.10331225e-01
-6.35823548e-01 1.04547405e+00 -6.47949398e-01 -2.11095974e-01
-1.35084796e+00 -6.08497500e-01 1.37641752e+00 6.27069250e-02
-3.42211366e-01 -1.11883879e+00 2.76332825e-01 3.66853237e-01
3.99955601e-01 1.30177304e-01 1.02284539e+00 -9.76589143e-01
-1.43493310e-01 -6.53358400e-01 -2.33045906e-01 7.25963950e-01
1.96577311e-01 -2.91420579e-01 -4.02171612e-01 -5.70997059e-01
-3.77340168e-02 -2.12764621e-01 1.84866381e+00 5.85485637e-01
1.04769886e+00 -7.61099100e-01 -1.97261661e-01 4.32392389e-01
1.33005953e+00 -3.41572374e-01 8.40696931e-01 2.23580971e-01
7.50102758e-01 4.99963671e-01 3.76326233e-01 4.95910794e-01
3.77004981e-01 2.90303081e-01 2.71618336e-01 2.36668468e-01
-3.09739828e-01 -4.21749681e-01 3.48659545e-01 1.36502600e+00
1.51226774e-01 -5.57358027e-01 -4.69319552e-01 6.05310142e-01
-2.00891161e+00 -1.25110734e+00 1.24410177e-02 2.06681728e+00
7.77211845e-01 1.49728104e-01 -8.62947404e-02 3.09840530e-01
6.35713458e-01 6.71524882e-01 -2.29217201e-01 -9.20324624e-01
-3.44544858e-01 3.74595910e-01 4.43816900e-01 4.12459433e-01
-8.40828300e-01 8.26596379e-01 5.67022324e+00 1.19647956e+00
-8.11352968e-01 -1.38303980e-01 3.38564128e-01 -1.17335983e-01
-6.18651390e-01 -1.21788554e-01 -4.31473494e-01 2.90574253e-01
7.96045065e-01 -6.64459407e-01 2.12238193e-01 5.77556431e-01
7.68207237e-02 -3.92904170e-02 -1.16657019e+00 8.94703567e-01
9.74419773e-01 -1.94610631e+00 3.99936408e-01 -1.21869206e-01
8.81942213e-01 -2.44223251e-04 -2.00866252e-01 2.20292196e-01
-1.65755540e-01 -9.23925698e-01 4.50062454e-01 6.09259367e-01
6.62443936e-01 -1.04692280e+00 1.01123202e+00 1.44017190e-01
-1.15235960e+00 2.09964737e-02 -6.97105169e-01 2.02415176e-02
9.87866148e-02 7.77164102e-01 -8.07471097e-01 1.02111304e+00
4.15969700e-01 1.21135747e+00 -7.41252184e-01 9.19549823e-01
3.80535498e-02 3.32042605e-01 7.44810477e-02 -3.73667955e-01
4.21662420e-01 -9.32939202e-02 1.01638675e+00 1.48477745e+00
5.27154386e-01 -3.71373594e-01 9.11881868e-03 5.36654890e-01
-5.57601273e-01 4.03894395e-01 -6.12288892e-01 -3.84831041e-01
1.95541173e-01 1.27217042e+00 -4.63297784e-01 -4.52977479e-01
-7.42545947e-02 1.41540706e+00 3.39884013e-01 -7.93899894e-02
-3.01912069e-01 -1.00990534e+00 3.15144151e-01 -5.08220941e-02
6.76633239e-01 -1.68358818e-01 -2.85793394e-01 -1.42162728e+00
1.99846417e-01 -7.30725408e-01 2.83680230e-01 -4.89896119e-01
-1.26454890e+00 5.78704476e-01 -1.09564848e-01 -1.21879816e+00
-4.24622029e-01 -9.64152440e-02 -1.10122967e+00 6.13145411e-01
-1.45149755e+00 -1.00146317e+00 -1.64004862e-01 2.15610579e-01
8.04918647e-01 -4.73623902e-01 8.71612251e-01 -1.14987344e-01
-1.79349691e-01 5.92716813e-01 6.21897936e-01 -1.43214851e-03
5.11872470e-01 -1.43358958e+00 7.95811713e-01 7.48606026e-01
1.99032620e-01 3.93847466e-01 8.10954392e-01 -6.48480773e-01
-1.38880575e+00 -1.25259161e+00 1.41821563e+00 -4.02537687e-03
4.77206767e-01 -1.07335135e-01 -8.30007136e-01 2.72138119e-01
7.75586247e-01 -5.03525913e-01 7.37081826e-01 -9.08435807e-02
-3.67361158e-01 -1.76097393e-01 -1.02333164e+00 5.24739683e-01
1.17552686e+00 -3.61933112e-01 -1.14048541e+00 5.72010994e-01
1.17138743e+00 2.71640066e-02 -8.19639981e-01 4.24709976e-01
3.26049954e-01 -9.27638650e-01 8.88170600e-01 -8.00876439e-01
1.12456691e+00 2.75946468e-01 -1.90399542e-01 -1.81061125e+00
-5.14603615e-01 -5.30236363e-01 -4.30689812e-01 1.25735736e+00
3.19511861e-01 -3.43195140e-01 6.90556526e-01 -2.19879135e-01
-5.17963648e-01 -9.43817437e-01 -8.20376396e-01 -9.45556164e-01
1.01852812e-01 1.25933051e-01 5.52099526e-01 5.07474720e-01
4.57495272e-01 8.09492052e-01 -1.47831291e-01 -6.18810654e-01
5.83429754e-01 3.62110555e-01 5.10031223e-01 -1.12425971e+00
-1.81249138e-02 -7.29380310e-01 -6.89559639e-01 -1.02318966e+00
5.06942928e-01 -1.41308773e+00 -1.20078206e-01 -2.41844797e+00
4.64296967e-01 5.93327940e-01 -1.41791478e-01 1.44193694e-02
-1.83394969e-01 -1.28047466e-01 3.04118395e-01 8.62757489e-02
-9.54642713e-01 1.24815047e+00 1.22257137e+00 -7.61429906e-01
-6.57010153e-02 -3.55797529e-01 -9.05639172e-01 3.57750684e-01
7.91868031e-01 -4.95879084e-01 -2.34187424e-01 -5.28499365e-01
1.97283342e-01 9.00482535e-02 -5.87007701e-02 -9.58301961e-01
4.73366767e-01 4.78155017e-01 1.97642356e-01 -9.43955660e-01
1.42635778e-01 -2.92459011e-01 -7.49424219e-01 5.90065956e-01
-7.58700907e-01 2.09880069e-01 -2.30433702e-01 7.43733168e-01
-5.85238457e-01 -5.85338533e-01 4.68415588e-01 -1.35445490e-01
-1.53887898e-01 2.41329208e-01 -1.77400351e-01 2.89986342e-01
4.23995078e-01 -3.27961922e-01 -4.45831299e-01 -6.46610200e-01
-5.56803234e-02 3.56058031e-01 2.47404471e-01 3.81220520e-01
1.15453494e+00 -1.39993417e+00 -1.33226240e+00 -1.91890642e-01
1.82632446e-01 -1.55820116e-01 2.97502965e-01 7.03123927e-01
-6.53058112e-01 6.82733238e-01 -1.77061483e-01 -3.85683358e-01
-1.46361339e+00 3.94351274e-01 -7.11398944e-02 -7.40100682e-01
-7.25796998e-01 5.64249337e-01 -3.30601692e-01 -2.32914463e-01
9.94692072e-02 -3.92954171e-01 -4.26696926e-01 3.33054155e-01
6.01305187e-01 6.80088282e-01 3.13602775e-01 -4.70291853e-01
-1.31091803e-01 5.01142979e-01 -4.17864740e-01 6.01000451e-02
1.43619168e+00 -1.60179585e-01 -5.55294096e-01 8.95889997e-02
1.78386807e+00 -1.15532137e-01 -3.75099778e-01 -4.19991285e-01
8.82703960e-02 -3.96576196e-01 2.21264184e-01 -3.67401421e-01
-8.02238822e-01 8.78024101e-01 -2.87948977e-02 5.06394744e-01
1.26182854e+00 -4.31755595e-02 1.44036627e+00 8.67389739e-01
-1.21263668e-01 -1.34106719e+00 4.80702281e-01 8.43541205e-01
1.32750678e+00 -9.45833325e-01 4.27181661e-01 -2.00053051e-01
-6.19125485e-01 1.51926744e+00 3.76250260e-02 -4.57162827e-01
7.13876113e-02 -1.95874125e-01 -7.07151532e-01 -1.74598902e-01
-6.71336651e-01 -1.61518618e-01 7.33517587e-01 3.51957798e-01
4.93882358e-01 -1.81868598e-01 -8.82224500e-01 2.66899794e-01
-6.19925797e-01 -4.39513594e-01 5.42018294e-01 8.24149251e-01
-8.86051595e-01 -9.15513456e-01 1.77484497e-01 1.13336146e+00
-4.87335294e-01 -4.51784760e-01 -6.34857178e-01 1.66571841e-01
-5.54737449e-01 1.09296989e+00 2.40224510e-01 -5.72071850e-01
2.49737963e-01 -1.29127026e-01 5.65676630e-01 -8.32893133e-01
-7.03028023e-01 -3.09064358e-01 2.98061520e-01 -5.03201425e-01
-3.52645427e-01 -5.23943424e-01 -1.14811265e+00 -5.49083769e-01
-4.27260607e-01 2.01821312e-01 7.38423407e-01 8.96647632e-01
6.14127040e-01 6.82253063e-01 9.38705683e-01 -8.93353522e-01
-7.96737373e-01 -1.28772426e+00 -5.78124523e-01 4.91217673e-01
2.78343797e-01 4.96815406e-02 -4.12685573e-01 -2.71990210e-01]
|
[12.523585319519043, 9.566071510314941]
|
9889fbff-2c46-464b-8850-781f16c99559
|
panocontext-former-panoramic-total-scene
|
2305.12497
| null |
https://arxiv.org/abs/2305.12497v2
|
https://arxiv.org/pdf/2305.12497v2.pdf
|
PanoContext-Former: Panoramic Total Scene Understanding with a Transformer
|
Panoramic image enables deeper understanding and more holistic perception of $360^\circ$ surrounding environment, which can naturally encode enriched scene context information compared to standard perspective image. Previous work has made lots of effort to solve the scene understanding task in a bottom-up form, thus each sub-task is processed separately and few correlations are explored in this procedure. In this paper, we propose a novel method using depth prior for holistic indoor scene understanding which recovers the objects' shapes, oriented bounding boxes and the 3D room layout simultaneously from a single panorama. In order to fully utilize the rich context information, we design a transformer-based context module to predict the representation and relationship among each component of the scene. In addition, we introduce a real-world dataset for scene understanding, including photo-realistic panoramas, high-fidelity depth images, accurately annotated room layouts, and oriented object bounding boxes and shapes. Experiments on the synthetic and real-world datasets demonstrate that our method outperforms previous panoramic scene understanding methods in terms of both layout estimation and 3D object detection.
|
['Zilong Dong', 'Ping Tan', 'Liefeng Bo', 'Chuan Fang', 'Yuan Dong']
|
2023-05-21
| null | null | null | null |
['scene-understanding']
|
['computer-vision']
|
[ 6.28560126e-01 -1.99099436e-01 5.03534377e-01 -7.64853656e-01
-3.47004592e-01 -5.49410284e-01 4.61505979e-01 5.73143363e-02
2.63720214e-01 3.03485900e-01 4.75524694e-01 -1.89671829e-01
-1.91956192e-01 -9.50368583e-01 -8.67632627e-01 -4.92747873e-01
2.35797301e-01 2.01874942e-01 1.58807799e-01 -2.53393143e-01
5.19022048e-01 3.74856353e-01 -1.71965814e+00 2.74944574e-01
1.02369869e+00 9.11760211e-01 1.02426267e+00 7.72252560e-01
9.77234989e-02 7.32239604e-01 -3.03163558e-01 1.75984740e-01
3.84502858e-01 -1.51697204e-01 -5.54807961e-01 6.07411683e-01
1.00382626e+00 -4.96731669e-01 -4.98428822e-01 1.00179267e+00
2.66569853e-01 4.36612993e-01 2.94120342e-01 -8.53055656e-01
-3.55578691e-01 -4.59418166e-03 -5.22035480e-01 6.60208333e-03
7.73351729e-01 -3.92147526e-02 7.38570452e-01 -8.87555897e-01
2.98856765e-01 1.47369885e+00 1.73136026e-01 1.14792146e-01
-8.11148643e-01 -6.06674731e-01 8.30452740e-01 1.65939987e-01
-1.16801703e+00 -2.24795073e-01 1.14155388e+00 -2.61527091e-01
6.76292717e-01 4.24415171e-01 8.12663078e-01 7.82562792e-01
-6.19601943e-02 7.51868188e-01 1.56087911e+00 -4.75636959e-01
2.26016551e-01 1.12148009e-01 1.46333277e-01 1.02532995e+00
2.49967813e-01 2.71604862e-02 -4.24461186e-01 2.60996342e-01
1.17600691e+00 4.75283623e-01 -5.30395150e-01 -8.31458032e-01
-1.28631163e+00 2.50087321e-01 6.53376698e-01 4.39100713e-02
-1.85534805e-02 -8.63354281e-02 -7.13883340e-02 -3.74509096e-01
2.11185038e-01 3.13421100e-01 -3.80119622e-01 1.47725865e-01
-4.25812244e-01 2.04903930e-01 4.96703655e-01 1.41923702e+00
8.56930017e-01 -3.25321704e-01 2.27030590e-01 8.49850714e-01
3.90095741e-01 6.87844276e-01 -7.61175007e-02 -1.07992494e+00
7.80661643e-01 9.56350148e-01 7.56513774e-02 -1.13161802e+00
-4.35733169e-01 -1.46405369e-01 -7.21462846e-01 -1.26074821e-01
3.24021429e-01 2.87340760e-01 -1.11227930e+00 1.46461487e+00
5.42708516e-01 1.74918920e-01 -1.75791532e-01 1.00668168e+00
7.39756584e-01 8.16545904e-01 -3.74584109e-01 6.30302653e-02
1.72925353e+00 -1.28996015e+00 -5.22828281e-01 -8.59554410e-01
-1.35906236e-02 -8.28082800e-01 1.38405526e+00 4.30265307e-01
-7.52771735e-01 -7.97573626e-01 -1.10272813e+00 -4.44287032e-01
-4.35165673e-01 2.04390451e-01 8.44581783e-01 6.83482826e-01
-5.53899288e-01 -6.56622741e-03 -5.84424019e-01 -4.71548885e-01
4.39051777e-01 1.34443313e-01 -4.02441502e-01 -8.81281316e-01
-5.24294138e-01 5.42245209e-01 4.04059470e-01 2.18660921e-01
-1.14602447e+00 -6.66585207e-01 -1.40138006e+00 1.24994390e-01
8.30250382e-01 -8.15460384e-01 9.64158893e-01 -2.11670265e-01
-1.21354282e+00 5.98607302e-01 -4.02460873e-01 2.39657104e-01
-4.05593403e-02 -5.16214728e-01 -2.81348944e-01 4.62100245e-02
8.60849768e-02 5.91332853e-01 5.05236506e-01 -1.84041750e+00
-6.68571949e-01 -8.48521054e-01 4.83888149e-01 6.77193046e-01
1.32649899e-01 -5.37801921e-01 -6.70642316e-01 -3.53067249e-01
8.19117367e-01 -6.83027267e-01 -5.29716671e-01 -4.43234555e-02
-6.23596966e-01 2.95038998e-01 7.86056638e-01 -7.03675270e-01
8.33237410e-01 -1.97299612e+00 4.96732369e-02 1.25044048e-01
-6.09261841e-02 -2.93856680e-01 1.92237884e-01 2.33361259e-01
1.09957583e-01 -3.54886681e-01 -4.71855998e-01 -4.19755340e-01
-4.12105918e-02 3.94378066e-01 -4.97280329e-01 2.68161714e-01
-3.61631989e-01 4.95039642e-01 -1.00773418e+00 -4.07596678e-01
8.44612658e-01 4.89912659e-01 -7.30918705e-01 3.23844641e-01
-3.20100904e-01 6.89666569e-01 -4.98756826e-01 9.06468034e-01
1.09596074e+00 -1.36614665e-01 4.04776782e-01 -1.91693783e-01
-1.35007083e-01 3.17236871e-01 -1.26514268e+00 2.28419089e+00
-8.95188987e-01 6.56645477e-01 -5.31949997e-02 -6.73983693e-01
1.08779085e+00 -2.05205396e-01 -5.71650080e-02 -8.51995945e-01
-8.36081132e-02 -1.06564648e-01 -4.82864022e-01 -5.50049663e-01
8.13553572e-01 1.19184323e-01 -1.98067293e-01 1.90078914e-01
-3.32253486e-01 -7.36589789e-01 -1.21532626e-01 7.55139440e-02
5.97388327e-01 3.77054930e-01 4.48334008e-01 -2.29849175e-01
6.25629187e-01 -1.02673016e-01 4.20632035e-01 6.40098631e-01
7.60633200e-02 7.95689881e-01 -1.02089457e-02 -5.47969699e-01
-1.09846449e+00 -1.19159675e+00 -3.88350487e-02 8.96999776e-01
8.54519010e-01 -5.10483861e-01 -8.17290545e-01 -5.22530258e-01
-4.01254743e-01 7.72443831e-01 -6.20505810e-01 3.14482510e-01
-8.27013910e-01 -5.70530534e-01 -3.09581071e-01 6.85370207e-01
9.61363971e-01 -7.59475112e-01 -9.60063100e-01 -1.32325709e-01
-5.89012742e-01 -1.53302455e+00 -5.51961124e-01 9.67862904e-02
-9.19934571e-01 -1.40755439e+00 -1.40893012e-01 -8.39916587e-01
1.02295709e+00 9.53094482e-01 1.09637332e+00 -2.69608766e-01
-7.21593916e-01 6.52134120e-01 -1.10239424e-01 -3.54877084e-01
3.08227479e-01 -4.57886428e-01 -1.74796954e-01 -2.31528997e-01
1.38645858e-01 -6.41028523e-01 -9.51104820e-01 5.25203645e-01
-5.41717827e-01 6.71887100e-01 5.36657393e-01 4.38570678e-01
7.65044332e-01 2.72437185e-01 -2.29428500e-01 -8.21211159e-01
-5.48504479e-02 -1.26132831e-01 -8.05103660e-01 3.19994271e-01
-2.78346598e-01 -2.41888270e-01 5.81125498e-01 3.30988616e-02
-1.62374997e+00 2.81268209e-01 2.26005107e-01 -1.10741936e-01
-7.09574401e-01 -2.12809995e-01 -9.24429178e-01 2.64188379e-01
2.72632957e-01 4.73986387e-01 -7.10220456e-01 -5.76963961e-01
5.13091266e-01 2.78160930e-01 8.03374827e-01 -8.09031427e-01
8.04581404e-01 9.95648205e-01 7.84059241e-03 -1.03386152e+00
-1.23482180e+00 -6.49597645e-01 -1.05858934e+00 -1.33304119e-01
9.28987920e-01 -1.17505038e+00 -6.73075438e-01 3.74468267e-01
-1.11053669e+00 -2.61250615e-01 -3.42092253e-02 4.28275585e-01
-6.40204191e-01 3.33330452e-01 -1.14600085e-01 -8.96791637e-01
1.67797551e-01 -1.01452446e+00 1.41779852e+00 4.44422066e-01
2.42273346e-01 -7.86634743e-01 -3.34691629e-02 8.83231521e-01
-3.47620994e-02 3.08515459e-01 9.64886725e-01 1.99499696e-01
-1.43866169e+00 1.75451472e-01 -5.14547169e-01 3.57612558e-02
3.32751840e-01 -4.14683193e-01 -1.36518407e+00 3.48257944e-02
2.78805912e-01 7.44934455e-02 6.40628874e-01 4.58571285e-01
1.55999589e+00 -1.42919123e-01 -4.37029600e-01 8.65124524e-01
1.60508454e+00 3.99777979e-01 6.04851067e-01 1.94094941e-01
1.13562369e+00 8.53593767e-01 9.55722272e-01 4.23800141e-01
8.07679892e-01 5.50150096e-01 7.13724971e-01 6.95675015e-02
3.72160114e-02 -7.46149421e-01 8.88127685e-02 8.64825666e-01
-3.82545143e-02 -1.80184379e-01 -9.42179501e-01 4.40183431e-01
-1.61540842e+00 -7.35584378e-01 5.51796407e-02 2.02946234e+00
2.18334958e-01 7.44097978e-02 -3.82155269e-01 -4.62040713e-04
5.03818035e-01 4.65785384e-01 -5.13341248e-01 -6.35650679e-02
2.95168115e-03 -2.13687778e-01 2.47519433e-01 6.27261996e-01
-1.09183168e+00 9.68090296e-01 5.72793531e+00 4.30351943e-01
-6.44440055e-01 -2.75501072e-01 6.90149188e-01 1.76480561e-01
-4.26265001e-01 1.88075155e-01 -8.12972903e-01 1.31351240e-02
1.70984417e-01 2.83207655e-01 5.08362830e-01 1.15405905e+00
1.00431129e-01 -6.26073778e-01 -1.22310066e+00 1.40941143e+00
4.72045302e-01 -1.08161318e+00 -2.92257648e-02 1.41104192e-01
9.02443528e-01 -4.11238194e-01 -8.75599217e-03 1.31142929e-01
2.53949672e-01 -1.07270432e+00 6.13795102e-01 5.30677497e-01
6.46704912e-01 -6.06992006e-01 2.66298324e-01 5.12326121e-01
-1.72253251e+00 -3.82771611e-01 -4.22963887e-01 -3.77491355e-01
1.58879802e-01 4.59779650e-01 -8.27198803e-01 7.44556606e-01
8.71810794e-01 6.50424361e-01 -7.07608223e-01 8.56993914e-01
-3.45008165e-01 2.02389099e-02 -3.75368774e-01 2.25883558e-01
1.22085558e-02 -3.84099633e-01 2.82763064e-01 9.60633814e-01
2.34936967e-01 5.58842123e-01 3.67494494e-01 1.03669310e+00
2.45231107e-01 -1.94675878e-01 -7.48664379e-01 4.60662633e-01
5.10455132e-01 1.25261736e+00 -9.35214043e-01 -3.71738523e-01
-2.54718632e-01 9.35451269e-01 1.10020630e-01 3.70041788e-01
-7.14225888e-01 -2.30685875e-01 4.19125497e-01 3.26179825e-02
2.78622985e-01 -6.17000163e-01 -6.01334453e-01 -1.27792883e+00
7.23809972e-02 -6.67107522e-01 7.06074536e-02 -1.15742493e+00
-7.75128543e-01 3.60838681e-01 2.41471499e-01 -1.02949929e+00
2.79069185e-01 -8.07836294e-01 -5.86631596e-01 5.90107024e-01
-1.55546808e+00 -1.34032786e+00 -1.15830958e+00 5.66095948e-01
1.08647633e+00 4.32175219e-01 7.43843555e-01 1.26719832e-01
-5.77683449e-01 -5.46812192e-02 -1.26861602e-01 -1.95888076e-02
4.76086766e-01 -1.45955896e+00 4.08921868e-01 9.28499103e-01
1.77693203e-01 8.21841896e-01 6.47995532e-01 -3.86814922e-01
-1.40682650e+00 -1.01841664e+00 2.88906008e-01 -7.93961823e-01
-1.18615724e-01 -9.72579002e-01 -5.39435327e-01 6.46340311e-01
-4.58922572e-02 -1.50486335e-01 5.76772451e-01 1.38775036e-01
-3.19034517e-01 -2.48192415e-01 -8.99809957e-01 7.55083680e-01
1.57023168e+00 -6.72328115e-01 -6.43084049e-01 2.51424253e-01
9.14075673e-01 -6.68253005e-01 -5.42064548e-01 5.16346335e-01
5.17878175e-01 -1.23577499e+00 1.44579053e+00 2.25449920e-01
6.16806149e-01 -6.03617489e-01 -7.73652911e-01 -1.01595891e+00
5.10766879e-02 -1.70490444e-01 -9.18067396e-02 9.29334283e-01
-2.33247623e-01 -1.60490826e-01 6.74196005e-01 5.48040807e-01
-4.40201640e-01 -7.59117723e-01 -3.82958204e-01 -3.29579085e-01
-5.75552225e-01 -6.34607017e-01 9.46139514e-01 7.06530035e-01
-2.52103895e-01 4.71197516e-01 -2.65236646e-01 7.49452293e-01
8.28563631e-01 8.59739959e-01 1.19702721e+00 -9.14708436e-01
-1.62341446e-01 1.12020699e-02 -1.56541362e-01 -1.85684347e+00
-2.29925931e-01 -2.74280071e-01 2.84111917e-01 -1.77806377e+00
4.16299999e-01 -5.70330262e-01 9.66646895e-02 1.48719987e-02
-2.71147519e-01 7.00449198e-02 1.19140029e-01 -1.85355574e-01
-7.31640935e-01 7.55675733e-01 1.60169816e+00 -1.23213641e-02
-1.57255024e-01 -1.69966176e-01 -6.73829913e-01 1.11336148e+00
4.64250296e-01 1.59780309e-01 -8.29596043e-01 -6.13755643e-01
-1.37120560e-02 1.08777247e-01 6.53590024e-01 -1.16334474e+00
2.16512486e-01 -3.85037541e-01 7.26935744e-01 -1.34793591e+00
8.76888633e-01 -1.08371210e+00 -3.41298789e-01 1.14307135e-01
8.96968618e-02 -1.39125586e-01 2.81381398e-01 9.09097433e-01
-2.87390146e-02 1.19336002e-01 3.99160594e-01 -4.89957035e-01
-1.17418301e+00 3.21669102e-01 2.28563875e-01 -1.72401577e-01
9.07073796e-01 -4.44524139e-01 -4.61896539e-01 -2.62659818e-01
-4.46998805e-01 3.64412129e-01 7.42406785e-01 5.43792784e-01
1.04875243e+00 -9.69278216e-01 -7.75058791e-02 5.11935651e-01
2.74493277e-01 9.36119854e-01 9.00305748e-01 2.49141246e-01
-7.08182395e-01 6.84714854e-01 -2.04311818e-01 -8.59495819e-01
-1.29113209e+00 7.83192873e-01 3.01999867e-01 2.61666868e-02
-7.92155504e-01 8.89278531e-01 1.27630091e+00 -8.42917502e-01
8.30272734e-02 -8.24747145e-01 -1.23744495e-01 -4.59086686e-01
5.58447063e-01 2.78755844e-01 -4.11027580e-01 -4.87787396e-01
-2.66040474e-01 1.05711126e+00 5.47779277e-02 4.98190522e-02
1.30870342e+00 -8.23872864e-01 -4.55511212e-02 5.84274888e-01
8.47700596e-01 2.15187654e-01 -1.80726147e+00 -3.61615717e-01
-2.96884447e-01 -1.14216578e+00 -2.45764405e-01 -8.29467833e-01
-5.47222793e-01 1.26191318e+00 6.40396655e-01 -3.83247137e-01
1.25210190e+00 1.02073103e-01 6.21841788e-01 3.62555712e-01
7.39057541e-01 -9.29342747e-01 4.77023095e-01 4.72060531e-01
7.69580245e-01 -1.41526330e+00 2.82122761e-01 -1.05193377e+00
-3.86333764e-01 1.08996654e+00 1.11696291e+00 1.00501636e-02
3.41254771e-01 -4.05890159e-02 -2.77429253e-01 -4.75573152e-01
-4.23477560e-01 -1.18433069e-02 3.71101171e-01 6.78382099e-01
-8.70811120e-02 2.53293693e-01 7.97082305e-01 4.42141771e-01
-5.72434545e-01 -5.59668064e-01 4.17915821e-01 8.75505686e-01
-7.03829110e-01 -6.81164503e-01 -7.80138612e-01 -1.06319807e-01
1.83232099e-01 -1.77429095e-01 -9.19817165e-02 8.01515579e-01
3.85114968e-01 7.87364423e-01 1.71833828e-01 -1.50678322e-01
4.59482968e-01 -2.67834246e-01 9.58987653e-01 -8.64654005e-01
2.28131339e-01 3.36780772e-02 -1.42370299e-01 -6.55602336e-01
-2.98576087e-01 -4.81019020e-01 -1.16256011e+00 -1.00059643e-01
-2.74450868e-01 -3.01444948e-01 8.51505816e-01 9.78669047e-01
9.16859880e-03 8.76231074e-01 5.51018834e-01 -1.00095415e+00
3.76007766e-01 -7.26986468e-01 -5.21799743e-01 2.42227808e-01
4.58220333e-01 -6.55991971e-01 -3.58311762e-03 -1.45632038e-02]
|
[8.647789001464844, -2.8411810398101807]
|
c3cd2401-b550-4330-bf96-f276b4eea7d6
|
bayesian-learning-of-effective-chemical
|
2205.06268
| null |
https://arxiv.org/abs/2205.06268v1
|
https://arxiv.org/pdf/2205.06268v1.pdf
|
Bayesian learning of effective chemical master equations in crowded intracellular conditions
|
Biochemical reactions inside living cells often occur in the presence of crowders -- molecules that do not participate in the reactions but influence the reaction rates through excluded volume effects. However the standard approach to modelling stochastic intracellular reaction kinetics is based on the chemical master equation (CME) whose propensities are derived assuming no crowding effects. Here, we propose a machine learning strategy based on Bayesian Optimisation utilising synthetic data obtained from spatial cellular automata (CA) simulations (that explicitly model volume-exclusion effects) to learn effective propensity functions for CMEs. The predictions from a small CA training data set can then be extended to the whole range of parameter space describing physiologically relevant levels of crowding by means of Gaussian Process regression. We demonstrate the method on an enzyme-catalyzed reaction and a genetic feedback loop, showing good agreement between the time-dependent distributions of molecule numbers predicted by the effective CME and CA simulations.
|
['Guido Sanguinetti', 'Ramon Grima', 'Svitlana Braichenko']
|
2022-05-11
| null | null | null | null |
['bayesian-optimisation']
|
['methodology']
|
[ 4.44430113e-01 9.67669189e-02 3.75189722e-01 1.59301847e-01
-1.89667210e-01 -4.72957075e-01 9.92349088e-01 5.64144850e-01
-9.17300463e-01 1.41022670e+00 -2.73851641e-02 -1.36953026e-01
2.72625778e-02 -7.39096642e-01 -6.26882732e-01 -1.52485383e+00
-5.81200160e-02 8.79305840e-01 5.30859590e-01 -1.58542424e-01
3.53175253e-01 7.34160542e-01 -1.11173368e+00 -2.22553238e-01
8.32805753e-01 2.98422068e-01 1.87024593e-01 1.14649987e+00
-2.38648430e-01 7.33605802e-01 -7.67057121e-01 -8.21939204e-03
-3.49416621e-02 -5.09797215e-01 -2.37722993e-01 -9.68030617e-02
-7.17443645e-01 5.70473015e-01 -7.78444931e-02 4.77362156e-01
6.92159593e-01 4.30905223e-01 1.49669838e+00 -6.13619924e-01
-9.71485376e-02 4.00035605e-02 -1.25422984e-01 2.49700636e-01
2.47054949e-01 5.75838625e-01 2.41174772e-01 -5.04044533e-01
8.46190214e-01 1.28493154e+00 4.94157374e-01 5.03805161e-01
-1.80075324e+00 -1.61391973e-01 -3.07485491e-01 -5.20643055e-01
-1.43103111e+00 -2.94891894e-01 2.78601110e-01 -7.38702595e-01
1.08400393e+00 3.01323950e-01 9.78866756e-01 9.24532235e-01
8.50482583e-01 1.61780506e-01 1.24733901e+00 -3.92399997e-01
1.12109303e+00 -1.89225987e-01 -4.48949397e-01 3.55040967e-01
3.43992084e-01 3.46470505e-01 -2.92450577e-01 -5.53014338e-01
9.03332531e-01 -2.24418044e-01 -2.76838094e-01 -5.98608196e-01
-1.26024866e+00 7.81376898e-01 -1.79885015e-01 2.60976881e-01
-4.55290914e-01 3.75940174e-01 3.90796900e-01 -8.02401304e-02
3.94912243e-01 5.45245707e-01 -6.32797420e-01 -1.05989054e-01
-7.57941425e-01 6.33318067e-01 1.32597053e+00 1.69084400e-01
5.91189802e-01 -2.19499275e-01 -9.95932054e-03 3.74966115e-01
5.74801683e-01 6.43629193e-01 3.69144082e-01 -9.48537946e-01
-5.64700216e-02 2.90770173e-01 6.59878612e-01 -4.92370099e-01
-4.24933165e-01 -1.04020521e-01 -8.28050971e-01 3.38096827e-01
1.07096612e+00 -4.76831049e-01 -8.45128417e-01 1.31252563e+00
5.14121175e-01 -5.38279600e-02 4.07213628e-01 5.55772901e-01
2.52005577e-01 8.31198156e-01 4.81616706e-01 -8.99753988e-01
8.15595269e-01 -3.38660806e-01 -7.56173670e-01 3.08488190e-01
1.00730097e+00 -3.94073576e-01 5.20498693e-01 3.42833251e-01
-1.11585093e+00 2.50049774e-02 -8.64542246e-01 4.17555422e-01
-5.61249375e-01 -6.42385244e-01 4.65323597e-01 6.46765947e-01
-8.75091612e-01 1.03008795e+00 -1.08242893e+00 -5.03911614e-01
2.24969804e-01 5.60697079e-01 -2.54425127e-02 2.96858341e-01
-1.01941407e+00 9.36438262e-01 2.04798743e-01 -9.74160805e-03
-9.71379399e-01 -5.30872643e-01 -7.06104755e-01 -2.41328701e-01
4.95625772e-02 -9.26465929e-01 8.02792430e-01 -7.41944849e-01
-1.94159174e+00 6.18882895e-01 -3.83179307e-01 -4.71344918e-01
7.66908944e-01 5.68021417e-01 7.98229873e-02 1.43264070e-01
-2.80119717e-01 5.53425312e-01 2.45929867e-01 -1.37754357e+00
-5.76978968e-03 -1.55129030e-01 -5.61349452e-01 1.28399000e-01
6.02885187e-01 -6.94387108e-02 2.34346300e-01 -2.39911392e-01
-9.28404033e-02 -9.29447412e-01 -8.98242176e-01 -1.60928503e-01
-3.19542468e-01 -1.82214081e-02 4.45558652e-02 -3.09053779e-01
7.04562247e-01 -1.48854136e+00 6.03622913e-01 4.24178839e-01
1.08572416e-01 2.01949403e-01 2.19443917e-01 8.96297872e-01
1.23322651e-01 1.33006215e-01 -6.39243841e-01 1.19518995e-01
-7.99385533e-02 2.44399056e-01 5.62153906e-02 7.25278795e-01
2.09356174e-01 8.01390529e-01 -9.89464164e-01 -3.77742618e-01
1.58652470e-01 7.46890724e-01 -1.35559246e-01 3.43475305e-02
-5.50823808e-01 1.00431669e+00 -3.34652066e-01 1.33505315e-01
5.08695900e-01 -1.15441062e-01 3.13732892e-01 5.08310199e-01
-3.35521936e-01 -2.62581170e-01 -1.05127800e+00 1.14633453e+00
-9.01859626e-02 1.35020390e-01 1.74976006e-01 -7.42607057e-01
1.07680964e+00 4.01662260e-01 5.65840602e-01 -2.15299487e-01
3.87797475e-01 4.01482671e-01 3.97194207e-01 -2.55084753e-01
-2.64333427e-01 -6.30031526e-01 1.72936946e-01 2.58297086e-01
-1.30568817e-01 -4.83971536e-01 4.66714859e-01 3.78493816e-02
1.16288614e+00 3.77563715e-01 4.66059387e-01 -7.68403709e-01
7.39531457e-01 1.35123162e-02 4.31860477e-01 6.59266472e-01
-2.87877500e-01 3.46736580e-01 9.73023295e-01 -4.09342647e-01
-1.39446425e+00 -8.10426056e-01 -1.77692413e-01 3.83877575e-01
2.99848951e-02 -1.26486048e-01 -1.01003838e+00 2.02494357e-02
4.09124121e-02 2.64628619e-01 -8.99501801e-01 7.89351836e-02
-3.98680300e-01 -1.48544264e+00 6.28292739e-01 -3.46943028e-02
-8.34538043e-02 -9.99952257e-01 -4.99892980e-01 5.21758080e-01
3.81338984e-01 -6.65169120e-01 1.50172682e-02 5.80598176e-01
-8.02933395e-01 -1.07929218e+00 -1.09442198e+00 -1.22695059e-01
5.65512717e-01 -6.28367245e-01 5.84816754e-01 -7.70971403e-02
-5.62628806e-01 1.12272523e-01 2.06467092e-01 -6.17283940e-01
-1.01162970e+00 -2.06050456e-01 7.40680099e-02 -3.10705811e-01
2.12296560e-01 -6.57718360e-01 -6.83153629e-01 2.59950221e-01
-9.96110439e-01 -2.80964524e-01 1.93145871e-01 6.48954034e-01
7.27045536e-01 -2.97624230e-01 5.45599639e-01 -7.70457745e-01
5.39299548e-01 -3.67119849e-01 -8.30843449e-01 -8.45175609e-02
-3.08679879e-01 2.76456624e-01 5.55294752e-01 -5.85726023e-01
-1.01338065e+00 3.31408262e-01 -3.69512327e-02 2.08922789e-01
-4.49975252e-01 1.91511586e-01 -2.67257512e-01 -2.21048266e-01
7.41230905e-01 4.52735692e-01 2.59726882e-01 -3.43344000e-04
-1.68609656e-02 2.18408048e-01 -1.09582417e-01 -6.40505016e-01
3.26177329e-01 8.09701145e-01 8.97388101e-01 -1.08887947e+00
8.87754560e-02 -2.85551578e-01 -6.34657204e-01 -1.59197241e-01
7.73607433e-01 -7.08347201e-01 -1.28295076e+00 6.14894450e-01
-1.05762184e+00 -8.50741565e-01 -3.36068720e-01 5.85892916e-01
-1.05278778e+00 5.77309430e-01 -8.42996418e-01 -1.43697822e+00
8.28894824e-02 -1.25422955e+00 8.08995008e-01 2.15696171e-01
-3.08358669e-01 -1.50390959e+00 8.29166651e-01 -8.88696983e-02
4.18414116e-01 7.02457488e-01 8.48113835e-01 -5.63672841e-01
-5.32784879e-01 -1.58133104e-01 4.16951418e-01 -3.02897751e-01
-1.87177688e-01 4.38995272e-01 -8.32188308e-01 -5.34480698e-02
-1.30631343e-01 1.63577110e-01 9.65505004e-01 9.08144772e-01
3.26630026e-01 4.79101054e-02 -8.24679911e-01 4.08545047e-01
1.53026330e+00 3.90050828e-01 8.42785239e-01 1.20416969e-01
2.81632811e-01 8.32798243e-01 2.00479880e-01 4.99706089e-01
-1.41181335e-01 3.33070129e-01 2.47110233e-01 5.82753718e-02
3.88153613e-01 4.93237078e-02 2.72406697e-01 3.95015359e-01
-4.58878815e-01 -4.92042512e-01 -1.00635016e+00 3.15730214e-01
-1.80645478e+00 -7.80216634e-01 -6.49555624e-01 2.38673139e+00
9.70086396e-01 6.75269142e-02 3.53224874e-01 3.75995343e-03
7.40114152e-01 -2.35728309e-01 -4.63281542e-01 -4.65059072e-01
-5.15484869e-01 1.74843490e-01 8.11852574e-01 1.02632678e+00
-6.19314849e-01 5.65246224e-01 7.64543772e+00 7.35887051e-01
-7.84868300e-01 -3.35169584e-02 7.87418425e-01 -1.47207245e-01
-8.66380036e-02 2.37846479e-01 -8.29692006e-01 5.94561934e-01
1.37510073e+00 1.57637745e-02 2.66005099e-01 -9.14597586e-02
7.90016055e-01 -8.50992620e-01 -7.62324691e-01 5.70248425e-01
-4.77546364e-01 -1.54982460e+00 -2.38785461e-01 5.66155910e-01
9.76718605e-01 -1.73004493e-01 -3.33184898e-01 -7.38809258e-02
5.61663866e-01 -1.07113922e+00 2.88855314e-01 9.77605939e-01
3.32370222e-01 -6.45339549e-01 7.93039560e-01 7.84484446e-01
-5.72118104e-01 1.49377123e-01 -4.66381341e-01 -3.07675451e-01
3.62248093e-01 8.31211090e-01 -9.05841291e-01 -9.92118120e-02
-3.09089795e-02 1.68678179e-01 -4.46027033e-02 1.10073721e+00
1.71653837e-01 4.96274114e-01 -6.17252648e-01 -5.35713017e-01
3.68064381e-02 -7.39024699e-01 5.53012729e-01 1.15801907e+00
1.87521860e-01 1.15065150e-01 -5.18097579e-01 9.85629678e-01
4.55579460e-01 4.12174553e-01 -5.88081062e-01 -1.38777941e-01
8.93374532e-03 8.93953800e-01 -1.28710115e+00 -1.23778030e-01
7.96808302e-02 7.47527719e-01 1.43419579e-01 6.75529122e-01
-5.76952100e-01 -2.56967813e-01 5.11445940e-01 3.57209265e-01
4.19098228e-01 -3.15033197e-01 1.34421447e-02 -7.35388935e-01
-5.67215323e-01 -2.10025162e-01 -4.57395017e-02 -5.16803026e-01
-9.71184671e-01 -7.74116740e-02 -6.91603869e-02 -2.09189236e-01
-1.83687642e-01 -7.46526957e-01 -6.46553516e-01 1.21085691e+00
-1.25254560e+00 -4.78023469e-01 3.03944528e-01 9.63552818e-02
4.98391427e-02 1.68012127e-01 8.34904969e-01 -3.49357933e-01
-4.53253657e-01 -3.32323074e-01 7.45607257e-01 -3.09592396e-01
4.68575567e-01 -1.34315848e+00 7.88689479e-02 3.17947388e-01
-6.32370353e-01 5.37503004e-01 1.28959715e+00 -1.05712199e+00
-1.05411875e+00 -9.05282557e-01 7.21197128e-01 -6.63926125e-01
5.42079449e-01 -5.94174087e-01 -7.19028056e-01 8.02944526e-02
3.17314193e-02 2.68260598e-01 8.16378176e-01 -5.43910325e-01
2.68266499e-01 4.81792837e-01 -1.24861324e+00 6.94071651e-01
5.42446136e-01 -8.93247202e-02 -7.07451478e-02 4.94031698e-01
3.16625148e-01 -1.34594992e-01 -1.22186887e+00 2.02385440e-01
2.94017047e-01 -8.01987052e-01 7.53236234e-01 -5.88373959e-01
4.42817025e-02 -3.87007445e-01 4.08997107e-03 -1.27100015e+00
-3.59461196e-02 -1.10163283e+00 -7.02559575e-02 6.54672086e-01
4.96214718e-01 -7.45373309e-01 8.37596476e-01 6.41046762e-01
4.85379875e-01 -8.36564183e-01 -1.16866565e+00 -6.12210095e-01
5.61546087e-01 2.18355879e-01 2.74968296e-01 6.51794195e-01
2.82653153e-01 1.56654656e-01 1.66849211e-01 -9.77884158e-02
6.67840183e-01 -7.35450387e-01 6.58730686e-01 -1.39899588e+00
-4.12067026e-01 -2.37175733e-01 -3.65230381e-01 -5.93745828e-01
1.34558365e-01 -3.46895725e-01 2.88972795e-01 -1.24516881e+00
2.76789576e-01 -2.19105124e-01 2.34248519e-01 -4.70869362e-01
-4.48707975e-02 -1.09018251e-01 -2.43691623e-01 2.15962410e-01
-5.91058612e-01 5.35521984e-01 1.39210379e+00 2.24070549e-01
-4.32679743e-01 -1.81700587e-01 3.04200556e-02 6.04066670e-01
7.09720314e-01 -5.37402391e-01 -2.25107431e-01 7.45074034e-01
5.26402533e-01 2.86624223e-01 2.49565274e-01 -9.57196653e-01
1.51857644e-01 -4.24130738e-01 4.83626842e-01 -1.97715014e-01
3.26872081e-01 -3.88142288e-01 6.91135824e-01 7.97404110e-01
-4.97656673e-01 -4.32688206e-01 1.47046804e-01 1.05182576e+00
2.23034933e-01 -3.49752635e-01 8.95995617e-01 -5.63235879e-01
3.34599882e-01 7.74691328e-02 -1.51455128e+00 -4.70557138e-02
1.27207506e+00 -4.60493088e-01 -1.63024336e-01 -3.10034364e-01
-1.33555663e+00 -1.65546313e-01 1.00877285e+00 -6.54361129e-01
1.60818815e-01 -7.53290653e-01 -5.72171152e-01 -1.39119059e-01
-1.57002255e-01 5.46093360e-02 2.29641989e-01 1.10372972e+00
-1.11659145e+00 6.12677813e-01 7.97327384e-02 -5.44775426e-01
-7.75820076e-01 7.95471609e-01 9.45645154e-01 -4.58865643e-01
3.38779092e-02 5.83229840e-01 3.67731154e-01 -5.68052828e-01
-2.80245841e-01 -3.94968092e-01 -7.46864006e-02 -2.39955783e-01
3.52656215e-01 5.12370348e-01 -6.72223121e-02 -7.73100436e-01
-2.91379113e-02 4.39384848e-01 4.17104423e-01 -3.45013946e-01
1.08137619e+00 -2.18135387e-01 -1.99991047e-01 8.06259155e-01
6.88391089e-01 -1.26207799e-01 -1.86190200e+00 2.25040928e-01
1.95916127e-02 4.11007553e-02 -1.57034114e-01 -6.67427301e-01
-3.56258750e-01 8.37442577e-01 3.79039079e-01 1.58142358e-01
3.06765854e-01 -5.93344010e-02 1.21610157e-01 4.32497174e-01
1.75720692e-01 -1.07200527e+00 -1.96250916e-01 3.67042512e-01
4.30718184e-01 -7.65711665e-01 -1.64302975e-01 -7.02158570e-01
-3.41938585e-01 1.06914711e+00 8.00892264e-02 -2.37360537e-01
6.64984286e-01 6.55255497e-01 -1.01196058e-01 -1.90371513e-01
-1.09852469e+00 3.28591950e-02 -5.64853549e-01 1.10888302e+00
5.76098800e-01 -6.76972717e-02 -7.37456083e-01 1.91233262e-01
6.12612188e-01 1.16828538e-01 8.11125875e-01 8.17464054e-01
-4.92710471e-01 -1.08811224e+00 -5.82266808e-01 1.63961068e-01
-5.07833838e-01 -4.78981202e-03 -4.89223927e-01 7.55305767e-01
8.33105445e-02 8.20716798e-01 6.17253222e-02 5.81524312e-01
-1.05074845e-01 4.70429569e-01 7.50975907e-01 -3.48093778e-01
-5.32451332e-01 3.48846674e-01 -7.60567859e-02 -1.69733495e-01
-5.52450359e-01 -9.15127337e-01 -1.54623580e+00 -2.95609266e-01
-3.93603384e-01 5.59353530e-01 6.94298923e-01 9.87112284e-01
1.86367840e-01 3.39511871e-01 4.87201214e-02 -1.07035398e+00
-2.06295222e-01 -8.57049108e-01 -1.04347992e+00 1.72855124e-01
1.51133329e-01 -4.99512315e-01 -7.70939589e-01 1.67757094e-01]
|
[6.059707164764404, 4.260529041290283]
|
82beab7b-e9c2-4c86-99ea-483ae0f848df
|
dd-tig-at-constraint-acl2022-multimodal
| null | null |
https://aclanthology.org/2022.constraint-1.2
|
https://aclanthology.org/2022.constraint-1.2.pdf
|
DD-TIG at Constraint@ACL2022: Multimodal Understanding and Reasoning for Role Labeling of Entities in Hateful Memes
|
The memes serve as an important tool in online communication, whereas some hateful memes endanger cyberspace by attacking certain people or subjects. Recent studies address hateful memes detection while further understanding of relationships of entities in memes remains unexplored. This paper presents our work at the Constraint@ACL2022 Shared Task: Hero, Villain and Victim: Dissecting harmful memes for semantic role labelling of entities. In particular, we propose our approach utilizing transformer-based multimodal models through a VCR method with data augmentation, continual pretraining, loss re-weighting, and ensemble learning. We describe the models used, the ways of preprocessing and experiments implementation. As a result, our best model achieves the Macro F1-score of 54.707 on the test set of this shared task.
|
['Xiaolong Liu', 'Jun Gao', 'Jingjing Dong', 'Han Zhao', 'Ziming Zhou']
| null | null | null | null |
constraint-acl-2022-5
|
['continual-pretraining']
|
['methodology']
|
[ 3.30966525e-02 1.30853936e-01 1.42968353e-02 -3.54844965e-02
-4.69265848e-01 -9.45856631e-01 1.00349104e+00 4.39502627e-01
-7.78063118e-01 9.17059064e-01 5.86382985e-01 1.83958299e-02
1.47291213e-01 -7.22527802e-01 -4.30878460e-01 -4.42619324e-01
1.31815612e-01 4.05824393e-01 1.56042710e-01 -5.39802730e-01
4.71662879e-01 1.48898482e-01 -9.88894880e-01 6.98815227e-01
8.55632305e-01 6.33919895e-01 -6.78661048e-01 6.95208669e-01
-2.64137894e-01 1.31176090e+00 -9.72459912e-01 -1.32653141e+00
-2.57989943e-01 -3.17855239e-01 -1.06646514e+00 -2.41065055e-01
6.73486590e-01 -1.53536964e-02 -6.29017293e-01 1.01904416e+00
5.95473349e-01 3.80712092e-01 8.62699032e-01 -1.39783645e+00
-9.45687175e-01 8.73195827e-01 -3.78444999e-01 4.92954165e-01
4.50535834e-01 -2.56243438e-01 1.06663501e+00 -1.01169801e+00
8.74617159e-01 1.43966067e+00 7.81683087e-01 7.56695807e-01
-1.26693845e+00 -7.39663124e-01 -2.71456540e-01 3.77554774e-01
-1.44585443e+00 -4.47282702e-01 8.19185913e-01 -4.77861673e-01
9.81402457e-01 3.57557923e-01 2.00705323e-02 1.60208106e+00
9.72813834e-03 8.44133914e-01 8.89765501e-01 -1.29086360e-01
-1.71482295e-01 7.28031278e-01 2.95329422e-01 9.44692910e-01
-3.22218277e-02 -4.47082788e-01 -9.42489445e-01 -6.08818769e-01
2.87510715e-02 -3.24027568e-01 6.36559853e-04 9.30597931e-02
-1.08864367e+00 9.60949659e-01 2.81231701e-01 4.17209685e-01
5.77148609e-02 1.44718811e-01 8.27516854e-01 5.00176191e-01
8.14942122e-01 8.44364345e-01 -1.04495458e-01 1.69853773e-02
-5.78174114e-01 1.33608073e-01 8.89273703e-01 6.26699090e-01
2.84393311e-01 -2.86864609e-01 -2.67110407e-01 1.05328095e+00
1.98363326e-02 4.05742794e-01 6.66366294e-02 -4.79311079e-01
7.19561279e-01 6.79494202e-01 1.03875771e-01 -1.64909875e+00
-4.64690566e-01 -3.32259834e-01 -6.90217316e-01 -4.78753686e-01
2.86881894e-01 -3.18223327e-01 -5.51882505e-01 1.74900746e+00
3.45646560e-01 1.23107217e-01 -1.32630795e-01 6.56807959e-01
1.04074395e+00 7.78255641e-01 6.24166310e-01 1.94615230e-01
1.40524113e+00 -8.67942333e-01 -9.82272804e-01 -5.06404452e-02
8.02131653e-01 -8.28465700e-01 7.65541971e-01 2.07066849e-01
-7.94906914e-01 -9.40857083e-02 -7.24030137e-01 -3.85545284e-01
-9.91914630e-01 6.15683049e-02 5.94440281e-01 8.77953649e-01
-5.49942493e-01 3.48114550e-01 -2.46963143e-01 -5.95011592e-01
4.99895543e-01 4.96935062e-02 -6.33957028e-01 1.12251088e-01
-1.67547739e+00 1.25328588e+00 5.30349255e-01 -2.17993692e-01
-7.92728305e-01 -8.20628643e-01 -7.19805241e-01 -1.89298838e-01
4.26154137e-01 -4.17687356e-01 5.00048339e-01 -6.17757201e-01
-9.41297650e-01 1.45703757e+00 1.55935720e-01 -4.45667684e-01
4.89808232e-01 -3.82193893e-01 -8.20493221e-01 2.55164802e-01
-1.77886914e-02 5.30450881e-01 8.10091853e-01 -1.24116719e+00
-3.61348301e-01 -1.74662516e-01 1.12675861e-01 -3.49548124e-02
-9.79464948e-01 5.27754188e-01 9.60143209e-02 -8.52394462e-01
-4.73821819e-01 -8.02490234e-01 4.31638777e-01 -3.59300047e-01
-7.74492562e-01 -3.73833656e-01 8.74804199e-01 -1.07957971e+00
1.61903131e+00 -2.21724749e+00 3.69830817e-01 8.81899223e-02
6.05083466e-01 3.20514768e-01 -1.86694428e-01 9.16367412e-01
1.14892848e-01 6.10334218e-01 -3.52968395e-01 -5.34350991e-01
1.40630692e-01 -2.34025523e-01 -4.68403816e-01 5.61029553e-01
1.89686179e-01 7.52366364e-01 -9.00769174e-01 -6.04483902e-01
-1.35685280e-01 5.84567249e-01 -3.87425870e-01 1.84226438e-01
-4.51408476e-02 1.80716723e-01 -2.42348894e-01 6.20059311e-01
6.04119182e-01 -9.65467989e-02 2.48028308e-01 -2.50906169e-01
1.03157394e-01 3.07637632e-01 -7.19095409e-01 1.24486411e+00
-2.03014880e-01 9.64209974e-01 2.61238128e-01 -4.86024857e-01
8.55488002e-01 1.45435423e-01 2.53311843e-01 -5.07842362e-01
4.02924895e-01 -8.80028382e-02 -4.16983694e-01 -7.65110672e-01
7.23607063e-01 -2.82385826e-01 -4.14432526e-01 4.67963636e-01
6.54809326e-02 4.06702578e-01 1.99762791e-01 7.82626390e-01
1.25508082e+00 -1.37395069e-01 8.82875323e-02 -5.71607016e-02
6.75933957e-01 1.52571708e-01 2.79928714e-01 4.65323687e-01
-3.65084887e-01 1.43341914e-01 7.43563533e-01 -2.58428037e-01
-9.80845332e-01 -9.79776263e-01 -8.43418017e-02 1.55761170e+00
2.02952161e-01 -5.16115785e-01 -6.24670208e-01 -1.18987215e+00
2.22627684e-01 1.11803293e+00 -8.78372669e-01 -4.71215665e-01
-5.13811707e-01 -9.96372938e-01 1.28986299e+00 4.89210449e-02
5.79860866e-01 -9.31414843e-01 -3.98061462e-02 -6.85452437e-03
-6.28794074e-01 -1.22130561e+00 -4.27938908e-01 -3.28164518e-01
-2.59946674e-01 -1.34239519e+00 -3.12117338e-01 -5.71688950e-01
4.03126478e-01 1.05451256e-01 1.06331384e+00 1.95903569e-01
-1.58599839e-01 3.76236707e-01 -3.48936647e-01 -2.25631326e-01
-4.68200862e-01 2.34745249e-01 -1.64786465e-02 2.85434425e-01
6.16278172e-01 -3.43614250e-01 -1.28127739e-01 3.10100406e-01
-8.36054325e-01 -2.68844545e-01 -2.77998904e-03 7.54654586e-01
-2.44622886e-01 -4.44591753e-02 6.34471297e-01 -1.24856842e+00
7.66177475e-01 -1.17313552e+00 8.99372324e-02 3.91146392e-01
-1.81187987e-01 -2.84384996e-01 5.44359446e-01 -4.57576394e-01
-1.15586698e+00 -5.50194204e-01 1.53900266e-01 -1.29051477e-01
1.09065920e-01 2.54247725e-01 -1.08580068e-01 -9.71214995e-02
8.65727186e-01 -4.18707915e-02 -1.55250579e-01 -5.31951070e-01
1.98048547e-01 7.94763088e-01 1.92358419e-01 -4.70682859e-01
8.77793908e-01 4.28324342e-01 -2.02500612e-01 -1.06330490e+00
-1.02368832e+00 -6.31660283e-01 -3.94928604e-01 -3.34125459e-01
9.22804356e-01 -7.29225636e-01 -9.47151005e-01 5.16668379e-01
-1.38663614e+00 1.83086842e-01 3.96988094e-01 1.15706190e-01
1.71106413e-01 4.94916379e-01 -8.44655871e-01 -9.17950749e-01
-3.15313756e-01 -3.11274618e-01 7.16809630e-01 -1.50561407e-01
-4.10337448e-01 -1.24834120e+00 2.50181705e-01 9.35853124e-01
3.66890103e-01 5.59925556e-01 1.08432806e+00 -1.20967591e+00
-1.12572275e-01 -7.65621662e-02 -2.24543601e-01 1.49220780e-01
-2.94399738e-01 2.65255533e-02 -9.76177335e-01 -1.47614673e-01
-5.28836191e-01 -6.41928911e-01 1.17920518e+00 -4.72780913e-01
8.83467555e-01 -6.13928974e-01 -3.91148895e-01 1.94961950e-01
1.24866080e+00 -2.10145682e-01 6.87205791e-01 3.82228404e-01
9.15296912e-01 1.11050630e+00 2.75815427e-01 5.19165635e-01
3.90119225e-01 4.25851822e-01 3.82868826e-01 3.36216509e-01
-6.15520366e-02 -4.63477641e-01 3.67712200e-01 4.42260653e-01
4.89308089e-02 -6.58627152e-01 -1.09182250e+00 6.27037168e-01
-1.79562211e+00 -1.38560104e+00 -2.85385400e-01 1.71840262e+00
6.52604818e-01 -2.02786669e-01 3.41534913e-01 -1.05663955e-01
1.05532312e+00 3.67585599e-01 -1.70326188e-01 -4.30966973e-01
-3.54915768e-01 -3.85805309e-01 3.93712133e-01 5.74799776e-01
-1.44943297e+00 1.25204718e+00 5.84424496e+00 8.65681887e-01
-5.10652483e-01 6.19611740e-01 3.00954640e-01 -2.19857216e-01
-3.90556693e-01 -5.20926535e-01 -7.45533347e-01 5.83938301e-01
8.04117322e-01 2.30935365e-02 5.25725245e-01 4.71655846e-01
-1.93472460e-01 4.79872897e-02 -7.93502092e-01 9.04258013e-01
6.16515279e-01 -1.38295817e+00 -9.34924930e-02 -1.26019999e-01
5.01773953e-01 -8.74425769e-02 3.79052199e-02 5.70311844e-01
2.52343453e-02 -1.03183579e+00 7.13522077e-01 5.47206759e-01
3.49823326e-01 -8.62868905e-01 7.43987560e-01 2.84091115e-01
-5.51199615e-01 -1.47653401e-01 -1.94898725e-01 3.32290232e-01
5.25017790e-02 4.40134674e-01 -8.48306358e-01 1.72603697e-01
4.42229122e-01 6.64770484e-01 -8.98185372e-01 6.91004634e-01
-2.54993916e-01 8.91043007e-01 -9.11452398e-02 -5.24625957e-01
1.50948659e-01 1.64242268e-01 1.01019645e+00 1.79710281e+00
-2.31732368e-01 5.88216558e-02 -8.86376798e-02 8.09377789e-01
-6.06763959e-01 1.97278500e-01 -6.62331641e-01 -6.25678539e-01
7.26266384e-01 1.39147472e+00 -4.94048059e-01 4.46990691e-03
-8.47362354e-02 1.10999012e+00 5.45280933e-01 2.10119277e-01
-1.10410750e+00 -5.87983608e-01 4.89852726e-01 -2.60462593e-02
-2.67245889e-01 -2.52343774e-01 -1.22766539e-01 -1.20749843e+00
-3.22961092e-01 -7.23545253e-01 7.46188760e-01 -4.88271475e-01
-1.77779424e+00 4.17546451e-01 -3.39767709e-02 -6.76891863e-01
1.94717795e-01 -5.66267192e-01 -3.11113417e-01 4.94038492e-01
-1.04351974e+00 -1.43729150e+00 -4.77421954e-02 4.86085981e-01
2.50242114e-01 -4.46515352e-01 6.72986627e-01 5.78905225e-01
-8.53511453e-01 7.98321128e-01 5.58783766e-03 5.65644860e-01
1.00273848e+00 -1.06357551e+00 9.25272331e-02 6.03377223e-01
-2.97555067e-02 7.96177626e-01 8.40584517e-01 -9.08400595e-01
-1.15508544e+00 -9.83829021e-01 1.29743838e+00 -1.04881322e+00
1.10495710e+00 -7.17914104e-01 -9.80591714e-01 5.70523202e-01
6.00907922e-01 -4.54455972e-01 1.15939236e+00 5.25752664e-01
-9.39297974e-01 3.89939636e-01 -1.56184971e+00 4.86795157e-01
1.11176550e+00 -8.48070264e-01 -8.17663670e-01 5.67054868e-01
6.20676041e-01 5.36677092e-02 -8.24462593e-01 -2.19882786e-01
3.67991745e-01 -6.30392551e-01 8.34644854e-01 -1.35563874e+00
6.53804898e-01 -7.62447715e-02 -2.03758508e-01 -1.16024423e+00
-2.48976752e-01 -6.73715115e-01 -3.45660925e-01 1.67357910e+00
5.43414533e-01 -4.92553294e-01 5.23639560e-01 5.49162030e-01
1.80204675e-01 -2.78316319e-01 -8.77213359e-01 -5.15366077e-01
5.65961599e-02 -2.78376728e-01 1.39790922e-01 1.80649436e+00
2.41394177e-01 8.26762855e-01 -7.65441597e-01 1.67928994e-01
9.23161566e-01 -5.12361228e-01 4.18579876e-01 -1.09540689e+00
1.12261117e-01 -3.02723080e-01 -2.57168949e-01 -1.21633939e-01
6.61940098e-01 -1.04964960e+00 -5.10400891e-01 -1.34849262e+00
6.42179966e-01 -1.62717700e-01 -1.54159278e-01 6.23608887e-01
-2.90041529e-02 4.55744654e-01 3.30173582e-01 1.38903663e-01
-9.02445138e-01 3.62839699e-01 7.81264365e-01 -4.91356790e-01
-3.25165205e-02 -3.90307695e-01 -6.94705844e-01 6.21823072e-01
9.33616102e-01 -7.10364997e-01 -1.50351554e-01 -2.62965262e-01
9.12588954e-01 -4.32004750e-01 6.94894314e-01 -4.59259391e-01
2.09162787e-01 -1.03820756e-01 1.54690132e-01 -3.61673504e-01
5.78490138e-01 -5.64916134e-01 -1.67725563e-01 2.70719230e-01
-7.43768871e-01 -1.33740813e-01 2.00023144e-01 6.92098737e-01
-2.13557389e-02 -3.29748601e-01 8.05754364e-01 7.00716972e-02
-6.65015757e-01 -1.08194649e-01 -5.86365283e-01 4.57845092e-01
9.66590703e-01 1.35878488e-01 -9.61639702e-01 -3.85746092e-01
-8.48979056e-01 2.84690768e-01 1.18425764e-01 6.70164227e-01
6.26471579e-01 -1.27823067e+00 -1.14177632e+00 -4.06914294e-01
2.10802034e-01 -1.11101305e+00 3.58242184e-01 7.45565832e-01
-1.13665558e-01 2.12070346e-01 -2.18417197e-01 1.48604557e-01
-1.57095969e+00 5.05420566e-01 3.05300027e-01 -8.54460299e-02
-1.80816010e-01 1.03137267e+00 -2.99936831e-01 -7.04961121e-01
2.61730045e-01 9.26952004e-01 -5.60608566e-01 6.21798217e-01
6.52659416e-01 1.01350772e+00 -1.17048696e-01 -9.12294865e-01
-6.53726280e-01 4.80649807e-02 -1.47672519e-01 -3.46319340e-02
1.23093486e+00 -1.99950039e-01 -6.26854181e-01 2.85565525e-01
1.61355829e+00 4.35897827e-01 -3.01257372e-01 -1.12264417e-01
2.57771254e-01 -4.42120969e-01 -4.93753292e-02 -1.25610399e+00
-5.65172851e-01 8.15177917e-01 2.46328384e-01 5.16324222e-01
6.07556999e-01 1.12149321e-01 9.51652169e-01 6.01776123e-01
9.63924304e-02 -1.45326364e+00 3.43161404e-01 7.87096679e-01
1.07918799e+00 -1.25489163e+00 -1.92791760e-01 -6.62509501e-01
-9.39029872e-01 8.25445056e-01 6.64602578e-01 1.58568114e-01
2.78458357e-01 -1.35112450e-01 -2.08787471e-01 -4.65038836e-01
-6.97428763e-01 -3.00703701e-02 5.21676898e-01 4.50433105e-01
5.70079446e-01 -5.85634932e-02 -4.62817639e-01 4.75567132e-01
7.62626603e-02 -6.79559052e-01 5.04242480e-01 6.78349555e-01
-4.09504741e-01 -6.02297843e-01 -4.15749639e-01 1.44426703e-01
-8.00038218e-01 -3.53360236e-01 -1.31267953e+00 8.14902723e-01
2.68052518e-01 1.17909896e+00 -9.41573977e-02 -8.09484184e-01
2.05218673e-01 3.41079861e-01 3.42797399e-01 -5.82771778e-01
-1.07795739e+00 -5.43849230e-01 9.55595374e-01 -3.04992527e-01
-3.94577891e-01 -5.30093610e-01 -1.15049827e+00 -8.64273787e-01
-9.73897651e-02 2.36386418e-01 6.81651592e-01 7.89810538e-01
4.54360515e-01 2.94289529e-01 5.99203169e-01 -2.06917793e-01
-2.54031032e-01 -1.09051919e+00 -4.19400066e-01 6.42495394e-01
1.91972956e-01 -6.81073070e-01 -4.98825967e-01 -1.45000204e-01]
|
[8.538400650024414, 10.66923713684082]
|
2eee24a9-3a5e-41bc-ba33-61f5f9fed89c
|
human-action-generation-with-generative
|
1805.10416
| null |
http://arxiv.org/abs/1805.10416v1
|
http://arxiv.org/pdf/1805.10416v1.pdf
|
Human Action Generation with Generative Adversarial Networks
|
Inspired by the recent advances in generative models, we introduce a human
action generation model in order to generate a consecutive sequence of human
motions to formulate novel actions. We propose a framework of an autoencoder
and a generative adversarial network (GAN) to produce multiple and consecutive
human actions conditioned on the initial state and the given class label. The
proposed model is trained in an end-to-end fashion, where the autoencoder is
jointly trained with the GAN. The model is trained on the NTU RGB+D dataset and
we show that the proposed model can generate different styles of actions.
Moreover, the model can successfully generate a sequence of novel actions given
different action labels as conditions. The conventional human action prediction
and generation models lack those features, which are essential for practical
applications.
|
['Mohammad Ahangar Kiasari', 'Minho Lee', 'Dennis Singh Moirangthem']
|
2018-05-26
| null | null | null | null |
['human-action-generation', 'action-generation']
|
['computer-vision', 'computer-vision']
|
[ 5.09699523e-01 3.43835264e-01 7.90750757e-02 -3.54988463e-02
-3.19457442e-01 -4.44670290e-01 8.90408874e-01 -9.63706434e-01
-6.84663327e-03 8.56934547e-01 3.57872933e-01 -2.13192049e-02
4.24209893e-01 -9.47067738e-01 -7.63774157e-01 -8.20560098e-01
4.56640899e-01 3.97811472e-01 3.48427258e-02 -5.35846986e-02
-9.00102556e-02 4.36262995e-01 -1.48143399e+00 2.62485534e-01
7.46233404e-01 7.17356563e-01 1.93327010e-01 1.10170162e+00
3.44145477e-01 1.47007334e+00 -6.74638391e-01 -2.61761069e-01
4.66209978e-01 -1.27240288e+00 -6.71638131e-01 6.42144859e-01
-3.59722190e-02 -6.60354495e-01 -7.00318515e-01 6.87993467e-01
5.96857309e-01 5.56097090e-01 8.90627861e-01 -1.50037026e+00
-9.16307151e-01 2.17266157e-01 -5.65918200e-02 -3.58474016e-01
5.85814416e-01 5.60795724e-01 5.91800332e-01 -3.80248904e-01
7.96029329e-01 1.21841156e+00 2.89330930e-01 1.28181219e+00
-8.51046681e-01 -4.29650515e-01 -7.66243860e-02 1.19399883e-01
-8.44681859e-01 -5.59601523e-02 9.39429879e-01 -4.13277209e-01
6.66499972e-01 -3.14782597e-02 1.05911648e+00 1.65599132e+00
3.67822379e-01 1.09580171e+00 1.00643551e+00 -3.69634032e-01
5.02465963e-01 -5.56228817e-01 -8.38249326e-01 6.23361230e-01
-2.78925806e-01 4.89678830e-01 -1.45607650e-01 2.00200319e-01
1.33743203e+00 4.50194925e-01 -7.68341944e-02 -3.09959292e-01
-1.24703753e+00 8.66210818e-01 3.94227058e-01 2.34411135e-01
-8.41011465e-01 4.83993053e-01 1.16602205e-01 1.65440273e-02
7.13848174e-02 3.52637768e-01 3.40177082e-02 -3.14367473e-01
-6.10664010e-01 5.22377610e-01 5.93036056e-01 1.05019951e+00
5.17980218e-01 5.56585371e-01 -4.98470277e-01 4.15619195e-01
1.63912654e-01 7.49672949e-01 7.56354451e-01 -1.32541239e+00
1.48774847e-01 4.29118216e-01 2.81680018e-01 -8.00671637e-01
-4.89307800e-03 -1.48159415e-01 -9.18498099e-01 3.64533871e-01
2.98629310e-02 -4.33858633e-01 -1.29410231e+00 1.71330929e+00
4.89801407e-01 5.64813018e-01 3.75363797e-01 8.57866287e-01
5.93696833e-01 8.11238229e-01 2.38798663e-01 3.65045331e-02
6.02405190e-01 -1.28227746e+00 -1.00115454e+00 -2.11477503e-01
3.55089813e-01 -5.52370369e-01 8.17582428e-01 2.16616884e-01
-1.17135644e+00 -1.09306347e+00 -8.60473037e-01 5.95116764e-02
-1.03216760e-01 3.33499610e-01 4.37398612e-01 3.74706447e-01
-9.51369286e-01 6.34647131e-01 -1.11025977e+00 -3.07904810e-01
3.18058342e-01 -7.43905781e-03 -2.28657275e-01 -3.97784263e-02
-1.00554264e+00 7.37062275e-01 5.97054482e-01 1.49546430e-01
-1.45474958e+00 7.49687105e-02 -9.39765573e-01 -1.93996087e-01
2.09756330e-01 -1.12427545e+00 1.46011639e+00 -1.23553395e+00
-2.27829957e+00 3.55869740e-01 1.29042208e-01 -4.68565643e-01
7.02395260e-01 -2.61208266e-01 -2.79517710e-01 2.41060898e-01
9.72250775e-02 8.76181722e-01 1.09359121e+00 -1.24497175e+00
-7.17544913e-01 6.54197186e-02 1.18726812e-01 3.08795989e-01
1.88244671e-01 -2.70280570e-01 -3.17396611e-01 -1.00257754e+00
-1.77521184e-01 -1.29865384e+00 -4.93910491e-01 -3.89588803e-01
-5.14344394e-01 1.52979970e-01 7.94331610e-01 -7.16359735e-01
7.87681699e-01 -1.87837708e+00 6.45898819e-01 -1.60440788e-01
-1.09137796e-01 1.77343100e-01 -3.06793690e-01 5.65858424e-01
-1.02419369e-01 -2.15024218e-01 -2.62364239e-01 -3.37022036e-01
2.07889527e-01 5.44147491e-01 -3.40347767e-01 7.56541640e-02
1.81213722e-01 1.22361016e+00 -1.13422537e+00 -2.36390457e-01
4.88896966e-01 5.93361080e-01 -5.64130306e-01 9.16037917e-01
-4.72312212e-01 1.19101989e+00 -7.93352365e-01 4.60141897e-01
1.70663238e-01 -1.28720284e-01 1.12240195e-01 1.44354761e-01
3.45746934e-01 -6.62087798e-02 -9.18020785e-01 1.68141949e+00
-3.91659021e-01 1.70705229e-01 -5.86587310e-01 -7.91715741e-01
9.39202845e-01 5.10629117e-01 4.91923243e-01 -4.33253258e-01
4.07981187e-01 -1.02440782e-01 2.84341723e-02 -6.44500077e-01
1.53332666e-01 -4.11331356e-01 -1.64760426e-01 7.07622051e-01
1.52263954e-01 -3.81648093e-01 1.41147956e-01 -5.25731500e-03
1.25158846e+00 7.80066788e-01 2.86868036e-01 5.53492665e-01
5.29707015e-01 -1.15049936e-01 5.87651670e-01 4.98320997e-01
-7.88133815e-02 8.25592637e-01 3.12877744e-01 -6.30301535e-01
-1.22355318e+00 -1.05092752e+00 7.59073555e-01 8.20556521e-01
-6.09825589e-02 -7.59150237e-02 -9.61277425e-01 -9.93873179e-01
-2.99978226e-01 8.78812194e-01 -8.57420921e-01 -4.83421862e-01
-7.62611508e-01 -2.20806897e-01 3.06761444e-01 9.87409472e-01
7.55918980e-01 -1.85819995e+00 -9.17007744e-01 4.08137530e-01
-2.43657961e-01 -1.27711213e+00 -3.45506668e-01 -1.81260228e-01
-7.20562041e-01 -1.06471002e+00 -1.02880239e+00 -8.58724833e-01
7.84687817e-01 -2.97851950e-01 8.69912207e-01 -2.06887066e-01
8.00403655e-02 5.47670364e-01 -7.99160004e-01 -2.57338762e-01
-1.11017442e+00 -2.58627445e-01 6.77644312e-02 2.81141669e-01
-6.41066134e-02 -7.10416675e-01 -7.13480353e-01 2.55689412e-01
-1.18104184e+00 5.34984827e-01 8.25134456e-01 1.04789197e+00
6.64902568e-01 3.17274034e-03 5.85485876e-01 -5.66535652e-01
4.22825694e-01 -2.56330341e-01 -2.10331559e-01 1.91967845e-01
2.44618673e-02 2.12016981e-02 9.59867537e-01 -6.85346484e-01
-1.26476586e+00 4.13628340e-01 -2.75219023e-01 -8.70026886e-01
-5.13441920e-01 9.27254707e-02 -3.56462598e-01 2.32757047e-01
4.72685218e-01 5.96240342e-01 -5.44444807e-02 -2.79933661e-01
6.84476078e-01 3.85036111e-01 7.83119917e-01 -5.09042919e-01
8.16774964e-01 3.59126151e-01 -2.73499936e-02 -2.79786736e-01
-5.81434488e-01 2.66546875e-01 -9.33295488e-01 -4.70008790e-01
1.39719689e+00 -7.68446922e-01 -4.35100108e-01 8.46500993e-01
-1.23056412e+00 -8.45648050e-01 -5.60090959e-01 4.81148690e-01
-1.36179399e+00 2.14861497e-01 -7.06076264e-01 -6.78422570e-01
-2.87868410e-01 -1.20341718e+00 1.18317950e+00 4.10694182e-01
-5.63839115e-02 -8.57123017e-01 3.03658962e-01 2.21597612e-01
8.55059847e-02 8.54739666e-01 5.20429969e-01 -4.39348340e-01
-6.76383674e-01 -3.68975013e-01 4.95390415e-01 6.50424957e-01
5.11692822e-01 -1.03885353e-01 -5.26421189e-01 -1.75872833e-01
-3.51422396e-03 -4.35310274e-01 4.76626515e-01 3.08263481e-01
1.16241646e+00 -4.19696450e-01 -1.96875155e-01 4.18964028e-01
1.05820000e+00 8.46324682e-01 1.07766199e+00 1.03435047e-01
8.95056307e-01 1.19238921e-01 5.37718475e-01 4.26607370e-01
6.04971275e-02 5.72196543e-01 3.48808527e-01 -4.43255827e-02
-1.95879832e-01 -8.11813831e-01 4.32081968e-01 6.28472745e-01
-4.80439961e-01 -4.63236630e-01 -7.00308621e-01 4.02809560e-01
-2.02605319e+00 -1.26822710e+00 5.12284458e-01 1.69673491e+00
6.07573807e-01 -5.62203974e-02 2.49809980e-01 1.10067956e-01
7.25805938e-01 2.15518475e-01 -6.72919393e-01 -2.84956396e-01
2.12422088e-01 5.55356860e-01 -8.55660159e-03 1.62045941e-01
-1.05211174e+00 1.28754830e+00 6.58467388e+00 5.27146816e-01
-1.03375506e+00 -5.19324318e-02 4.78380889e-01 1.43779546e-01
-5.82698472e-02 -1.13419093e-01 -3.50384980e-01 5.79593599e-01
5.99989355e-01 -2.48053856e-02 4.65534717e-01 1.02365041e+00
1.95000798e-01 1.51329562e-01 -1.00228572e+00 7.88389385e-01
2.22524732e-01 -1.02245808e+00 3.98820043e-01 -5.93616068e-02
1.14947474e+00 -6.71655655e-01 4.10049558e-02 4.47614551e-01
8.61230433e-01 -1.01548624e+00 7.33866930e-01 8.44787598e-01
6.37697339e-01 -8.85842264e-01 5.82609057e-01 5.21789193e-01
-9.22465801e-01 -1.78568423e-01 -1.33948877e-01 -1.86487257e-01
6.51121497e-01 -1.14552025e-03 -8.22362661e-01 5.88123441e-01
2.97587037e-01 9.23197031e-01 -3.21593195e-01 6.21532917e-01
-9.11101997e-01 4.35931653e-01 1.83661997e-01 -7.66047239e-02
3.09479237e-01 -3.24608356e-01 3.06543469e-01 6.70329332e-01
6.78945959e-01 4.36937481e-01 2.87430108e-01 8.49767745e-01
1.30904481e-01 -2.71137953e-01 -7.38230348e-01 -2.94114560e-01
1.34310335e-01 9.98276412e-01 -5.76242864e-01 -5.42149007e-01
-2.28480518e-01 1.56664944e+00 1.10780701e-01 5.98308086e-01
-1.21917307e+00 -1.00454651e-01 2.73542881e-01 -2.41928682e-01
6.34630859e-01 -1.46731809e-01 1.32642999e-01 -1.17680562e+00
-2.45750144e-01 -1.12420022e+00 4.45085727e-02 -1.23948550e+00
-1.20051658e+00 7.44845986e-01 -2.07665563e-01 -1.58868849e+00
-1.04554641e+00 -3.98225456e-01 -7.64132619e-01 6.16177559e-01
-8.08184922e-01 -1.41048467e+00 -4.04158503e-01 6.89787209e-01
6.57811940e-01 -3.63982588e-01 7.96335280e-01 -6.40315861e-02
-4.78006661e-01 2.95680463e-01 -1.98456392e-01 4.06832099e-01
3.22976738e-01 -1.16484070e+00 5.87884963e-01 1.01383257e+00
1.27136931e-01 1.79548874e-01 6.90933704e-01 -9.22933221e-01
-9.19572473e-01 -1.40828431e+00 5.09798288e-01 -4.34661865e-01
2.12395668e-01 -1.18931666e-01 -3.63834232e-01 1.10558712e+00
4.01264250e-01 -2.59743333e-02 5.11545300e-01 -7.93905318e-01
3.27376395e-01 3.16690564e-01 -9.97838080e-01 8.46582353e-01
1.36722434e+00 -3.05120468e-01 -7.80373991e-01 2.25051403e-01
6.63745403e-01 -7.33827651e-01 -7.09955156e-01 5.53172588e-01
5.49402297e-01 -9.27378356e-01 9.47892845e-01 -9.15174842e-01
9.76080537e-01 -3.45854282e-01 -1.21377066e-01 -1.57816017e+00
-4.57213789e-01 -6.44109428e-01 -3.38960409e-01 7.65230834e-01
-1.76197023e-03 -2.79300153e-01 7.21776128e-01 5.38029015e-01
-3.30987930e-01 -6.92332923e-01 -6.83108091e-01 -8.28652024e-01
-8.36351812e-02 -1.56504318e-01 8.03261936e-01 5.86253047e-01
-4.69654977e-01 2.50348926e-01 -9.20488179e-01 -1.79417357e-01
3.03622663e-01 2.38673136e-01 1.26053584e+00 -6.65198684e-01
-6.66461945e-01 -2.96985228e-02 -4.76063520e-01 -1.25333226e+00
3.98204625e-01 -4.78808522e-01 3.00298154e-01 -1.58010352e+00
-1.43781174e-02 3.05521376e-02 -1.67837888e-01 6.29439056e-01
-3.63686532e-01 1.69965416e-01 4.36192632e-01 3.49802561e-02
-4.48657811e-01 1.22093475e+00 1.87207568e+00 -7.15838373e-02
-3.14892381e-01 2.31545791e-01 -2.39870876e-01 6.94299102e-01
7.82268882e-01 -2.76084095e-01 -6.51723504e-01 -2.13096187e-01
-2.04341382e-01 3.84832174e-01 5.31996071e-01 -1.37627411e+00
-1.70756549e-01 -4.37737197e-01 7.52180874e-01 -4.69325989e-01
3.31174016e-01 -7.07128763e-01 5.28711021e-01 6.62830353e-01
-3.93951744e-01 7.30134100e-02 -1.50570422e-01 6.21395826e-01
-2.06666961e-01 2.44594544e-01 6.30316675e-01 -2.89725393e-01
-7.14341581e-01 4.40247267e-01 -3.48899275e-01 -1.95470348e-01
1.28827095e+00 -8.51233006e-02 8.08264688e-02 -7.19889402e-01
-1.14486563e+00 -1.50760695e-01 5.40104389e-01 5.61638892e-01
6.86400771e-01 -1.91649413e+00 -5.97199857e-01 1.76828191e-01
-2.20004529e-01 3.57265323e-02 4.37637329e-01 3.14508080e-01
-5.76269448e-01 1.15681656e-01 -7.49869823e-01 -3.26996386e-01
-7.01371074e-01 8.31946015e-01 4.44092512e-01 -3.92921239e-01
-7.23810554e-01 4.76281881e-01 2.84444153e-01 -4.33694035e-01
-1.37190744e-01 -1.51712567e-01 -1.76313743e-01 -5.43034017e-01
3.78784567e-01 3.64528328e-01 -6.51797533e-01 -7.20082045e-01
1.04237482e-01 3.83900106e-01 4.71716195e-01 -4.06824231e-01
1.26017809e+00 2.27057114e-01 4.11733627e-01 2.41933674e-01
8.16601753e-01 -3.32026780e-01 -1.74353600e+00 1.34325728e-01
-7.15160787e-01 -3.66945952e-01 -4.82376486e-01 -8.03968608e-01
-1.21586180e+00 7.45001733e-01 4.52130109e-01 -4.66119051e-02
1.34482217e+00 -2.19657749e-01 1.18870127e+00 1.60113320e-01
5.46616733e-01 -1.04696584e+00 6.48104370e-01 3.43339503e-01
1.13431072e+00 -9.12253976e-01 -3.78407985e-01 -9.84781757e-02
-1.02661276e+00 9.13070977e-01 7.81137407e-01 -3.84821922e-01
2.63368636e-01 -1.12180198e-02 1.30185783e-01 1.35360926e-01
-5.88472366e-01 -3.74058932e-01 1.96351603e-01 8.90618682e-01
-1.82053132e-03 9.14791152e-02 -3.15915525e-01 5.48370242e-01
-3.23658913e-01 4.96818781e-01 3.62868935e-01 1.00857854e+00
-1.23736158e-01 -1.22616935e+00 -1.43040910e-01 2.28667576e-02
-5.80312498e-02 3.84277135e-01 -4.17147934e-01 7.65350640e-01
3.66552085e-01 7.61371255e-01 6.93586655e-04 -7.68532097e-01
3.77866864e-01 2.36951336e-01 6.30201757e-01 -5.79009354e-01
-2.87820399e-01 -4.25038114e-03 -2.00802281e-01 -6.87565625e-01
-7.43043303e-01 -7.00645387e-01 -1.26071823e+00 -7.07898065e-02
1.38803303e-01 -1.18650377e-01 1.70711745e-02 9.03644860e-01
3.97247791e-01 8.14701557e-01 8.66254807e-01 -1.06736660e+00
-4.09714520e-01 -1.13131690e+00 -4.60355639e-01 8.70671809e-01
4.19496968e-02 -7.19832301e-01 -1.64958641e-01 7.02570319e-01]
|
[7.320693492889404, -0.11622694879770279]
|
252e942c-f6b7-42f1-983c-d9985e722ed1
|
kullback-leibler-maillard-sampling-for-multi
|
2304.14989
| null |
https://arxiv.org/abs/2304.14989v1
|
https://arxiv.org/pdf/2304.14989v1.pdf
|
Kullback-Leibler Maillard Sampling for Multi-armed Bandits with Bounded Rewards
|
We study $K$-armed bandit problems where the reward distributions of the arms are all supported on the $[0,1]$ interval. It has been a challenge to design regret-efficient randomized exploration algorithms in this setting. Maillard sampling~\cite{maillard13apprentissage}, an attractive alternative to Thompson sampling, has recently been shown to achieve competitive regret guarantees in the sub-Gaussian reward setting~\cite{bian2022maillard} while maintaining closed-form action probabilities, which is useful for offline policy evaluation. In this work, we propose the Kullback-Leibler Maillard Sampling (KL-MS) algorithm, a natural extension of Maillard sampling for achieving KL-style gap-dependent regret bound. We show that KL-MS enjoys the asymptotic optimality when the rewards are Bernoulli and has a worst-case regret bound of the form $O(\sqrt{\mu^*(1-\mu^*) K T \ln K} + K \ln T)$, where $\mu^*$ is the expected reward of the optimal arm, and $T$ is the time horizon length.
|
['Chicheng Zhang', 'Kwang-Sung Jun', 'Hao Qin']
|
2023-04-28
| null | null | null | null |
['thompson-sampling', 'multi-armed-bandits']
|
['methodology', 'miscellaneous']
|
[-1.64593056e-01 1.54562980e-01 -8.29576075e-01 -3.42193276e-01
-1.51290107e+00 -8.15432668e-01 -1.48157114e-02 2.37576049e-02
-8.35572898e-01 1.20408237e+00 -6.73218106e-04 -9.78982806e-01
-9.36369002e-01 -7.52024889e-01 -9.51787174e-01 -7.44506538e-01
-5.33300698e-01 5.33414841e-01 -3.30713034e-01 2.07461938e-01
2.84480840e-01 2.56431818e-01 -9.17954028e-01 -3.00813884e-01
8.40128303e-01 1.77397931e+00 1.17151767e-01 7.55539060e-01
3.30015570e-02 6.86961651e-01 -4.18374687e-01 -5.88787258e-01
7.77152300e-01 -6.80891156e-01 -7.88923144e-01 -3.13532978e-01
-2.15107918e-01 -4.83223259e-01 -4.39393550e-01 1.22653878e+00
3.29990417e-01 4.85650629e-01 4.21225399e-01 -8.69860172e-01
-7.56872743e-02 1.10076964e+00 -1.00016499e+00 2.24495798e-01
-2.36843359e-02 -7.76459053e-02 1.22529733e+00 6.25328720e-02
3.26856971e-01 1.29482830e+00 3.57778400e-01 3.48853320e-01
-1.12722087e+00 -8.70970309e-01 4.40465122e-01 -7.39303902e-02
-9.81772780e-01 -4.68936712e-02 4.52737212e-01 -5.82815558e-02
4.93055403e-01 5.34266293e-01 6.73600078e-01 4.37530696e-01
1.19946569e-01 1.19417143e+00 1.27019203e+00 -5.41834295e-01
8.19047272e-01 -9.12080705e-02 -8.67069587e-02 3.84596705e-01
3.74756396e-01 4.90126699e-01 -4.19352114e-01 -3.67175639e-01
7.95848250e-01 9.91971716e-02 -8.82596374e-02 -2.10320115e-01
-7.11230636e-01 1.11713934e+00 3.64222437e-01 -1.51653022e-01
-6.02910101e-01 8.99680853e-01 2.22956002e-01 6.51017964e-01
4.94477451e-01 2.96467066e-01 -4.47061718e-01 -6.57190800e-01
-9.63279366e-01 6.07131958e-01 6.46576226e-01 1.04231930e+00
3.58704895e-01 -1.16405182e-01 -6.64165795e-01 4.38174099e-01
1.14703745e-01 8.33077133e-01 -1.63413715e-02 -1.40013516e+00
1.00660682e+00 -3.51540111e-02 1.17566979e+00 -2.59075761e-01
-4.70524989e-02 -6.41983509e-01 -5.96629977e-01 2.28846475e-01
6.53507948e-01 -6.16879702e-01 -4.29385990e-01 1.84837091e+00
2.25811392e-01 -4.55001146e-01 -7.57043585e-02 8.07084799e-01
-4.18249398e-01 4.19070780e-01 -3.17309916e-01 -8.48501027e-01
7.36215115e-01 -6.13799036e-01 -5.63566029e-01 -2.23225668e-01
5.67449212e-01 -5.37305713e-01 9.01892900e-01 5.50828576e-01
-1.57660234e+00 2.56513268e-01 -6.98677361e-01 6.72384322e-01
2.83493102e-01 -2.77815223e-01 9.16073561e-01 1.10021496e+00
-6.21101379e-01 7.88573503e-01 -6.17499053e-01 3.94848734e-01
7.31879056e-01 2.97371984e-01 3.06951880e-01 -3.90628487e-01
-7.14513063e-01 4.54274148e-01 2.46122852e-01 6.58418685e-02
-1.00178385e+00 -3.85431468e-01 -2.48803690e-01 -9.11771953e-02
8.68852913e-01 -1.34936631e-01 1.70599222e+00 -4.30588007e-01
-1.55722904e+00 2.99537987e-01 -5.29535376e-02 -8.85490358e-01
8.45792770e-01 -1.55381650e-01 1.81265429e-01 -1.99618135e-02
1.31611526e-01 7.06725195e-02 6.29762471e-01 -5.04115999e-01
-8.40453267e-01 -5.66742361e-01 1.70613959e-01 1.44988209e-01
-6.22800514e-02 -4.20620441e-02 8.15551821e-03 -6.71194434e-01
1.56931335e-03 -9.37470734e-01 -7.88064778e-01 -3.67463291e-01
-2.93973833e-01 -1.97920516e-01 -8.17082450e-02 -3.34084898e-01
1.27971733e+00 -1.97876871e+00 -1.83631957e-01 4.84690458e-01
-4.27928865e-01 -1.05941795e-01 8.12518597e-02 3.71780217e-01
3.15884292e-01 6.16066828e-02 3.22957113e-02 -5.64262178e-03
3.74226362e-01 8.48557614e-03 -4.50456828e-01 5.46118319e-01
-7.00997174e-01 8.35837066e-01 -8.76781583e-01 -5.71105182e-02
-7.38332197e-02 -5.34802616e-01 -4.38896835e-01 1.52694553e-01
-6.47771418e-01 5.43554826e-03 -8.76778483e-01 5.62171280e-01
5.96282244e-01 -1.95692435e-01 1.58298805e-01 5.80693364e-01
-8.17427505e-03 1.31219089e-01 -1.37393606e+00 1.43986487e+00
-3.71703804e-01 1.06630556e-01 4.77170199e-01 -1.21023989e+00
7.51518786e-01 9.19954777e-02 5.58482409e-01 -7.60676861e-01
4.33945447e-01 3.97759765e-01 -3.99733782e-01 3.61801195e-03
3.05188447e-01 -5.55865645e-01 -3.61822933e-01 1.03613913e+00
-5.67509830e-01 -1.53642654e-01 1.38448030e-01 3.58734615e-02
1.29610610e+00 -2.20185563e-01 1.97084516e-01 -3.33249062e-01
-3.37593704e-01 -5.47009595e-02 4.70459372e-01 1.46456289e+00
-2.63504773e-01 -5.30088246e-02 7.81062841e-01 -2.68726885e-01
-7.67511010e-01 -1.08963454e+00 6.44845739e-02 1.21112883e+00
1.92246869e-01 1.04091167e-01 -5.23106694e-01 -7.52736509e-01
3.87913495e-01 1.19494069e+00 -8.31221342e-01 2.05358770e-02
-1.47698611e-01 -8.71348083e-01 2.32694745e-01 5.09854615e-01
4.85966116e-01 -7.19254375e-01 -8.72174144e-01 5.51641703e-01
1.01112381e-01 -4.27939922e-01 -7.60078788e-01 5.92756689e-01
-1.01854455e+00 -9.32788908e-01 -9.39827442e-01 -3.19672539e-03
4.69834358e-01 2.69902259e-01 8.36586416e-01 -1.00989997e+00
-1.59471035e-01 4.74421412e-01 -4.40467536e-01 -8.48398626e-01
1.75365135e-01 -3.34585309e-01 -1.03974938e-02 -2.24927813e-01
2.00277358e-01 -3.50572735e-01 -1.09880948e+00 2.76197642e-01
-6.25743687e-01 -4.87434387e-01 6.95608437e-01 8.27176571e-01
8.46258223e-01 1.77995473e-01 6.92158759e-01 -6.29437506e-01
8.39186907e-01 -2.65957385e-01 -1.08733630e+00 3.40020716e-01
-8.29232097e-01 3.51831526e-01 4.59625125e-01 -4.42488670e-01
-8.34150851e-01 -2.05511123e-01 2.26383671e-01 -4.43227381e-01
5.55308044e-01 5.12635827e-01 3.79279971e-01 1.16315991e-01
6.97948277e-01 3.08210015e-01 7.24638179e-02 -6.01366699e-01
4.07032609e-01 7.19094396e-01 1.40654638e-01 -9.63183343e-01
3.11861932e-01 4.02494341e-01 8.51953998e-02 -1.06893055e-01
-1.27515137e+00 -1.48811087e-01 4.71679181e-01 5.28273433e-02
2.06104726e-01 -6.50792480e-01 -1.29037499e+00 -1.92214623e-01
-4.97913629e-01 -5.75430036e-01 -9.54620481e-01 8.61048341e-01
-1.26501107e+00 9.46212113e-02 -2.92103052e-01 -1.75856078e+00
-5.73988616e-01 -7.63648152e-01 4.73808289e-01 3.26208800e-01
2.27936715e-01 -3.73752475e-01 -4.06248495e-02 3.43842387e-01
4.48468179e-01 1.77133128e-01 7.74250865e-01 -2.36603782e-01
-5.57020843e-01 -5.39815664e-01 -2.48805299e-01 2.78068244e-01
-1.44680426e-01 -8.44604611e-01 -1.65021628e-01 -7.34974504e-01
-9.26235411e-03 -5.67567170e-01 7.38609910e-01 1.01122558e+00
1.43368292e+00 -7.34620750e-01 -1.69045299e-01 1.96937829e-01
1.26170480e+00 7.49546051e-01 3.07288080e-01 4.60545748e-01
-5.33340573e-01 4.06384654e-02 1.28790522e+00 1.18822622e+00
-6.09016083e-02 4.48079258e-01 7.28668690e-01 6.99414194e-01
8.30074430e-01 -2.19683334e-01 3.48814636e-01 -1.91029638e-01
1.96473040e-02 8.38895421e-03 -4.18092161e-01 5.93439162e-01
-2.01965690e+00 -9.56771016e-01 4.11287993e-01 2.89037037e+00
1.10914552e+00 3.99627030e-01 5.73616028e-01 -8.33047181e-02
4.81187612e-01 -6.91158995e-02 -1.04382789e+00 -7.56539524e-01
1.55219391e-01 4.57026184e-01 1.21702898e+00 3.58159751e-01
-6.97594464e-01 4.93578345e-01 6.01304770e+00 1.31991279e+00
-7.26633430e-01 1.47582129e-01 9.51109648e-01 -9.29887831e-01
-2.55117059e-01 -2.91275196e-02 -8.47610950e-01 7.34115124e-01
9.77599621e-01 -5.01079857e-01 8.52636218e-01 1.24241197e+00
3.30362856e-01 -5.74103832e-01 -9.30705190e-01 1.02728045e+00
-7.97976851e-01 -1.34196150e+00 -6.72749281e-01 3.32661629e-01
7.89520383e-01 8.30820724e-02 3.07260901e-01 2.98706412e-01
1.14211726e+00 -1.04677844e+00 7.65165329e-01 3.01062018e-01
1.09143698e+00 -1.24199367e+00 6.47380054e-01 6.15767896e-01
-7.94789553e-01 -5.37502825e-01 -5.37400901e-01 -1.26659915e-01
2.33649537e-01 8.24829996e-01 -7.57221162e-01 5.25618136e-01
8.88179541e-01 -6.86285123e-02 4.93158191e-01 1.27755678e+00
1.92022175e-01 5.74528754e-01 -6.67087495e-01 -6.30575836e-01
7.96598613e-01 -3.92178982e-01 4.57364529e-01 7.57671475e-01
6.27595961e-01 3.56529206e-01 1.12663656e-01 5.12775183e-01
-2.45448723e-02 5.76128215e-02 -3.03107291e-01 -3.39298248e-01
6.54035270e-01 4.79899794e-01 -4.79858577e-01 -2.42655557e-02
1.37864903e-01 6.35212064e-01 2.06716120e-01 2.08104327e-01
-7.15559423e-01 -6.74799502e-01 6.49679184e-01 -1.11896574e-01
7.55370200e-01 -1.59683883e-01 -2.61062503e-01 -5.49636126e-01
3.00056159e-01 -3.67227107e-01 7.35927224e-01 -1.78725794e-01
-1.07509100e+00 -1.44581152e-02 1.05649091e-01 -9.44095552e-01
-4.24042284e-01 -3.46620977e-01 -3.21224868e-01 9.56346333e-01
-1.10638213e+00 -3.76559526e-01 5.03890455e-01 5.51331699e-01
2.73974746e-01 -7.93514773e-02 6.24100566e-01 -3.16646516e-01
-3.59844834e-01 8.84350300e-01 1.07766354e+00 -3.33884269e-01
8.11965391e-02 -1.24759459e+00 -1.44540787e-01 3.77592653e-01
-2.62351125e-01 2.90587991e-01 8.56016338e-01 -5.70717394e-01
-1.72374523e+00 -1.01873064e+00 1.27067998e-01 1.67406827e-01
6.33808613e-01 1.04743376e-01 3.26477364e-02 6.97123885e-01
-1.94142923e-01 1.07255243e-01 7.47580469e-01 2.07987875e-01
-1.98655874e-01 -5.32971025e-01 -1.56022215e+00 5.78853130e-01
9.72126365e-01 -2.04269961e-02 -6.70188740e-02 4.14182663e-01
7.31038988e-01 -3.82090896e-01 -9.51451361e-01 1.06711566e-01
6.67277217e-01 -9.29765403e-01 6.06763780e-01 -7.90086091e-01
-1.23428397e-01 3.26564044e-01 -4.96683747e-01 -1.20829892e+00
-4.64802235e-02 -1.41861892e+00 -3.59105825e-01 5.94135046e-01
4.27333653e-01 -5.91922641e-01 1.23231900e+00 7.53235042e-01
2.89127052e-01 -1.07009149e+00 -1.68875241e+00 -1.17447615e+00
3.72374207e-01 -7.44477451e-01 5.76885343e-01 7.09824711e-02
3.39320034e-01 -4.24243182e-01 -5.82203329e-01 -3.47445905e-01
1.03082514e+00 5.76077342e-01 4.52607751e-01 -5.46299577e-01
-8.59483778e-01 -5.69223702e-01 3.55165929e-01 -1.55293477e+00
-3.82807642e-01 -3.76842260e-01 -1.31159304e-02 -1.20419061e+00
2.69172549e-01 -1.02182674e+00 -6.06063128e-01 2.66317248e-01
1.46440297e-01 -3.53458762e-01 2.27848977e-01 -1.06800124e-01
-9.78107035e-01 7.46200383e-01 1.34211326e+00 -1.20988429e-01
-2.89281487e-01 8.48268926e-01 -1.07182634e+00 3.07375014e-01
7.17297435e-01 -5.11496902e-01 -3.53243440e-01 -2.14702874e-01
4.59929615e-01 9.94665921e-01 -1.45949706e-01 -4.76496965e-01
-8.35225806e-02 -8.64318371e-01 8.39414913e-03 -9.58524287e-01
2.63976634e-01 -6.76637650e-01 9.45815444e-02 7.33013332e-01
-7.29630768e-01 -3.05705726e-01 -1.44092709e-01 1.13880718e+00
3.02988946e-01 -5.00874341e-01 6.91149771e-01 -4.72121298e-01
1.90793395e-01 4.21266735e-01 -1.56846404e-01 2.11019576e-01
1.20727873e+00 1.09043606e-01 -1.28905669e-01 -7.76791096e-01
-4.51883942e-01 5.96809983e-01 -7.88383111e-02 -1.70334339e-01
3.88158381e-01 -1.21637595e+00 -4.71586525e-01 -3.67301524e-01
-2.71657407e-01 1.20949797e-01 2.60655850e-01 8.12988937e-01
1.49064874e-02 6.95209920e-01 2.87474483e-01 -1.62078157e-01
-6.76343143e-01 5.68322897e-01 1.73723102e-01 -6.97623312e-01
-3.08880031e-01 1.22928524e+00 -4.06722456e-01 7.80616850e-02
6.98330939e-01 -2.82160342e-01 6.30838096e-01 -1.01443775e-01
8.00345778e-01 7.71434784e-01 -1.82709217e-01 4.18015271e-01
-4.21467908e-02 -4.75946777e-02 -3.03357244e-01 -6.17948771e-01
1.45159638e+00 -1.88995376e-01 1.65575579e-01 1.93200082e-01
9.53025699e-01 -2.53844529e-01 -1.61832595e+00 -4.99908388e-01
-7.36947879e-02 -8.92136931e-01 7.03491420e-02 -1.00850987e+00
-7.94837058e-01 4.32369530e-01 5.94712794e-01 4.16697532e-01
9.98146534e-01 6.21525794e-02 7.25321710e-01 4.47770566e-01
1.14618587e+00 -1.37441373e+00 -1.01076119e-01 2.50054389e-01
7.49457717e-01 -9.58527207e-01 -2.20671017e-02 2.89510339e-01
-5.16485274e-01 6.50793195e-01 5.45029156e-02 -1.57810703e-01
5.16359866e-01 -6.29694387e-02 -5.73601365e-01 2.54460722e-01
-6.53247356e-01 -1.47183672e-01 -1.86175123e-01 9.97975469e-02
-3.56801413e-02 7.03817070e-01 -6.42797649e-01 9.36053753e-01
-3.45896870e-01 2.47406751e-01 1.88604385e-01 1.35602319e+00
-7.87046969e-01 -1.12431872e+00 -5.49547851e-01 1.05755091e+00
-7.65378237e-01 2.06213802e-01 8.50800723e-02 3.59520584e-01
-5.22840559e-01 1.04463708e+00 -1.06901295e-01 -4.94041964e-02
2.43169442e-01 -2.77010351e-01 8.03641021e-01 -9.02467147e-02
-2.58064121e-01 3.00486207e-01 5.65294288e-02 -7.19643056e-01
-5.50156347e-02 -8.27455044e-01 -9.00320470e-01 -5.14793098e-01
-3.04847449e-01 8.27463746e-01 6.86998665e-01 9.49426591e-01
2.37944856e-01 -8.11252892e-02 1.22876310e+00 -4.22206581e-01
-1.62029803e+00 -9.18831348e-01 -1.13215280e+00 -9.33487490e-02
1.92395270e-01 -5.99061668e-01 -2.70757943e-01 -9.61812198e-01]
|
[4.534137725830078, 3.308532238006592]
|
26633a50-c4f6-4eab-ad46-c49f11e7a284
|
a-graph-based-text-similarity-measure-that
| null | null |
https://aclanthology.org/R17-1098
|
https://aclanthology.org/R17-1098.pdf
|
A Graph-based Text Similarity Measure That Employs Named Entity Information
|
Text comparison is an interesting though hard task, with many applications in Natural Language Processing. This work introduces a new text-similarity measure, which employs named-entities{'} information extracted from the texts and the n-gram graphs{'} model for representing documents. Using OpenCalais as a named-entity recognition service and the JINSECT toolkit for constructing and managing n-gram graphs, the text similarity measure is embedded in a text clustering algorithm (k-Means). The evaluation of the produced clusters with various clustering validity metrics shows that the extraction of named entities at a first step can be profitable for the time-performance of similarity measures that are based on the n-gram graph representation without affecting the overall performance of the NLP task.
|
['Iraklis Varlamis', 'Leonidas Tsekouras', 'George Giannakopoulos']
|
2017-09-01
| null | null | null |
ranlp-2017-9
|
['text-clustering']
|
['natural-language-processing']
|
[-1.26941308e-01 2.06446514e-01 1.90779686e-01 -3.45880240e-01
-4.38948274e-01 -7.93147981e-01 7.40779877e-01 1.00402677e+00
-6.49736822e-01 1.66438401e-01 4.37973708e-01 -5.33127427e-01
-5.29618025e-01 -9.38692451e-01 1.43641725e-01 -4.65058684e-01
-1.97815791e-01 7.56599665e-01 4.05022323e-01 -1.76462337e-01
8.05638611e-01 4.65555191e-01 -1.58510852e+00 4.35125709e-01
1.10816622e+00 5.56270480e-01 4.87249084e-02 8.11767220e-01
-9.81182635e-01 4.47507083e-01 -6.98020279e-01 -5.85125685e-01
1.86140433e-01 -4.94915038e-01 -1.27702689e+00 -1.73170179e-01
5.28017245e-02 6.10851765e-01 -1.67456418e-02 1.22539318e+00
4.39821124e-01 4.08695906e-01 9.00173783e-01 -1.18263972e+00
-1.61511093e-01 9.81477439e-01 -8.04791376e-02 9.65053961e-02
7.74842978e-01 -3.27883840e-01 1.14689481e+00 -5.24659693e-01
1.08237314e+00 1.29650843e+00 6.50867164e-01 -1.46599740e-01
-9.38006043e-01 -1.35664031e-01 -4.06033099e-01 2.20571280e-01
-1.57460082e+00 -4.38399352e-02 3.28486711e-01 -5.70774198e-01
1.30914104e+00 6.31989896e-01 2.02432588e-01 3.69809657e-01
8.32005888e-02 1.15963906e-01 7.55700827e-01 -8.95643055e-01
4.88810033e-01 3.58102798e-01 7.41042852e-01 4.39866871e-01
4.88398582e-01 -6.60781980e-01 -5.87257445e-02 -4.88184303e-01
-6.78135231e-02 -2.95735389e-01 -3.46285850e-02 -4.26941365e-01
-1.11876249e+00 8.62253845e-01 3.13098997e-01 1.26468849e+00
-3.44687670e-01 -5.44574916e-01 9.09867346e-01 5.68638742e-01
4.88624483e-01 7.37301111e-01 -1.97266072e-01 -3.85216586e-02
-9.99819100e-01 -1.34731263e-01 1.27381742e+00 1.16258359e+00
5.89110970e-01 -3.69383633e-01 -1.99499816e-01 5.87114394e-01
4.91510719e-01 2.29614004e-01 9.32671666e-01 -2.28682220e-01
4.79771674e-01 1.50832415e+00 -2.33012959e-01 -1.39514685e+00
-5.83866417e-01 1.11058168e-01 -5.79859734e-01 -3.28149796e-01
4.35305923e-01 1.57801360e-02 -5.09465575e-01 9.77403700e-01
7.56305754e-01 -3.70994657e-01 2.04626143e-01 2.46034041e-01
1.01548576e+00 4.77903068e-01 2.39868566e-01 -4.21745062e-01
1.55108261e+00 -5.60300529e-01 -8.44345391e-01 3.85164976e-01
1.46387529e+00 -1.06023467e+00 7.59696066e-01 -8.90022889e-03
-6.83521032e-01 -3.88819903e-01 -5.02258599e-01 2.04360653e-02
-1.19040227e+00 -1.37528375e-01 1.41565934e-01 1.00840282e+00
-1.31546152e+00 7.92454243e-01 -3.76264364e-01 -1.16136670e+00
-2.27980673e-01 2.71408647e-01 -5.29729307e-01 2.04275221e-01
-1.00742948e+00 9.59679127e-01 1.18787324e+00 -4.71848786e-01
2.72334337e-01 -1.93743348e-01 -7.24776030e-01 2.90222555e-01
2.72912920e-01 -4.50828254e-01 6.36651516e-01 -7.00295448e-01
-9.55555558e-01 9.87391651e-01 -5.36601618e-02 -3.22381467e-01
2.94088125e-01 1.96802557e-01 -7.62286723e-01 3.58360201e-01
2.67023116e-01 -1.01719841e-01 4.31265533e-01 -9.82467353e-01
-5.44516981e-01 -6.31859899e-01 -5.78052700e-01 3.90416116e-01
-5.88332236e-01 6.67814493e-01 -3.05905014e-01 -5.12416065e-01
2.11200610e-01 -9.10943508e-01 -1.98246732e-01 -6.29050136e-01
-5.22111177e-01 -6.56348526e-01 6.71866000e-01 -9.42443669e-01
1.80556357e+00 -2.06389856e+00 -6.96457773e-02 1.00171614e+00
3.00311655e-01 3.32863033e-01 1.06270336e-01 1.04661667e+00
-3.57332468e-01 5.29085457e-01 -1.77291170e-01 1.34862021e-01
1.67323589e-01 -1.82395130e-01 2.37466320e-01 2.42399573e-01
-3.07709396e-01 5.76675892e-01 -8.57385755e-01 -9.34638441e-01
2.20765799e-01 1.41952261e-01 3.34483013e-02 1.21498585e-01
2.28059869e-02 -2.28539541e-01 -3.13726276e-01 -3.64953503e-02
3.61164302e-01 -2.29716599e-02 7.65057027e-01 -2.65539493e-02
-2.50393778e-01 3.33256394e-01 -1.46535695e+00 1.30863881e+00
-5.25601096e-02 5.92296720e-01 -3.22446436e-01 -1.09856462e+00
1.04148793e+00 4.32374746e-01 3.83515507e-01 -4.95010883e-01
5.30912220e-01 2.96226591e-01 1.21376589e-01 -7.46375978e-01
7.40428448e-01 4.75574255e-01 -1.27465621e-01 5.51819921e-01
1.58470869e-01 -1.28957734e-01 8.82726431e-01 8.26901317e-01
1.36816585e+00 -3.45273674e-01 7.36536503e-01 -7.80023456e-01
7.83271015e-01 4.39305425e-01 -1.12835415e-01 3.56511831e-01
9.38979909e-02 1.20167255e-01 4.09289032e-01 -5.87321892e-02
-1.17670560e+00 -6.78492606e-01 4.55426946e-02 1.13904333e+00
-2.94771820e-01 -9.90501881e-01 -1.02811432e+00 -8.38187993e-01
-9.92174596e-02 1.09304774e+00 -2.68927097e-01 -2.62531582e-02
-3.03864568e-01 -4.87450391e-01 5.42884886e-01 8.25808942e-02
4.61095348e-02 -9.73179221e-01 -2.28474960e-01 1.46318376e-01
-1.46110311e-01 -9.11143780e-01 -2.70676851e-01 2.46777460e-01
-8.91633093e-01 -1.31786656e+00 -3.37045550e-01 -8.65062058e-01
7.09675252e-01 1.71212718e-01 1.14823699e+00 9.45592523e-02
-2.51205891e-01 6.58999205e-01 -9.38033581e-01 -3.33292156e-01
-9.21968758e-01 2.33782470e-01 -2.02205732e-01 -9.55035910e-02
7.59784400e-01 -3.47996771e-01 -7.53606260e-02 1.52777284e-01
-1.03891766e+00 -4.83469754e-01 1.72864497e-01 1.52523771e-01
1.04327284e-01 4.11706477e-01 2.43304789e-01 -1.10206747e+00
1.03528011e+00 -5.61111927e-01 -2.51485348e-01 6.88835502e-01
-9.20400202e-01 2.40806431e-01 6.75564766e-01 -7.53696114e-02
-1.00723398e+00 -7.05712810e-02 2.00903580e-01 1.37529965e-03
-3.73756647e-01 5.81089079e-01 -1.96115986e-01 9.72593948e-02
9.23588991e-01 7.87825361e-02 -3.11649084e-01 -4.38705832e-01
7.18806922e-01 1.15854251e+00 2.07613379e-01 -2.02634722e-01
6.39514327e-01 9.13526937e-02 -1.00623772e-01 -1.13165641e+00
-3.06811750e-01 -1.49365675e+00 -1.16504443e+00 -2.53331393e-01
9.44732845e-01 -6.05581343e-01 -7.04876125e-01 1.03862705e-02
-1.12258208e+00 2.87299961e-01 -2.07869336e-01 6.20922029e-01
-1.09043181e-01 7.85255671e-01 -3.45999062e-01 -7.31400073e-01
-6.42921627e-01 -5.65107942e-01 6.96207762e-01 1.09491311e-01
-7.01053679e-01 -1.20593882e+00 4.72978681e-01 5.05419075e-01
8.96658823e-02 1.24801859e-01 1.33126712e+00 -1.82811105e+00
1.15387514e-01 -3.87230456e-01 -2.47695521e-01 3.72034535e-02
-1.45187629e-02 4.33975428e-01 -6.07169807e-01 -1.51569188e-01
-1.32924959e-01 2.98600644e-01 2.75258124e-01 -1.00486271e-01
6.09813392e-01 -3.81635487e-01 -5.12314856e-01 -1.30369335e-01
1.45219135e+00 3.83098245e-01 7.04992771e-01 4.77567583e-01
7.99729228e-01 1.01577508e+00 4.74516034e-01 5.13818979e-01
2.95578718e-01 2.39279702e-01 -6.46796599e-02 2.85397649e-01
2.83256173e-01 1.60374761e-01 -6.90326793e-03 1.34335721e+00
6.11334480e-02 -4.56682801e-01 -1.45019209e+00 5.55932581e-01
-1.95419872e+00 -1.12870991e+00 -9.51275587e-01 2.18732595e+00
5.49044549e-01 -2.81376224e-02 2.25540608e-01 4.67602462e-01
1.29550326e+00 -1.71525344e-01 1.46516383e-01 -7.42722034e-01
-1.02692336e-01 2.68958002e-01 3.41540903e-01 2.26115063e-01
-7.64954686e-01 7.69298553e-01 5.47035599e+00 1.06822944e+00
-6.80271685e-01 -1.88937262e-01 9.29054692e-02 4.44216162e-01
-1.62655294e-01 2.84095794e-01 -6.64616406e-01 5.39127827e-01
1.45754671e+00 -7.02067673e-01 4.25484963e-02 8.58493984e-01
1.47226825e-01 -3.23942989e-01 -9.04735386e-01 8.63263011e-01
3.16115975e-01 -1.00643063e+00 2.69576430e-01 -1.20632565e-02
4.31619048e-01 1.28217787e-01 -8.39744031e-01 8.22393522e-02
6.24192953e-01 -4.20422167e-01 3.42438608e-01 3.53112876e-01
3.14313829e-01 -9.87894595e-01 8.04075480e-01 2.41529346e-01
-1.47036183e+00 2.01410174e-01 -3.94349605e-01 3.36887777e-01
-1.90169260e-01 7.20638275e-01 -1.18399096e+00 1.10357225e+00
7.37040520e-01 3.47612977e-01 -1.05098474e+00 1.14740407e+00
1.37376279e-01 4.64414299e-01 -3.81411135e-01 -4.05853629e-01
1.31474987e-01 -6.19260073e-01 5.88750601e-01 1.70030248e+00
3.35899055e-01 -2.10427456e-02 1.15381479e-01 3.76109362e-01
-1.03595063e-01 1.15902030e+00 -7.65291572e-01 -3.18276197e-01
5.95467448e-01 1.51115584e+00 -1.43985152e+00 -7.69717872e-01
-1.30527750e-01 7.93466389e-01 4.03228849e-01 -1.89739048e-01
-2.89342433e-01 -1.12070882e+00 -1.40962705e-01 1.46371931e-01
8.08081254e-02 -1.05903774e-01 -2.84088373e-01 -8.60544264e-01
-2.09796894e-02 -8.33210289e-01 8.54192972e-01 -7.82442093e-01
-1.28942478e+00 8.41038942e-01 6.91312030e-02 -1.07532382e+00
-3.94073039e-01 -4.71567780e-01 -8.05442572e-01 5.54640114e-01
-5.37616193e-01 -6.93763196e-01 -1.61927119e-01 8.49535763e-01
-6.65021967e-03 -1.12067692e-01 9.32834089e-01 6.15739748e-02
-5.23784339e-01 1.97953120e-01 6.54625237e-01 4.07564044e-01
7.84691334e-01 -1.53415298e+00 3.35106552e-01 7.46539772e-01
4.45177019e-01 8.15880597e-01 6.13819599e-01 -9.72238004e-01
-9.40171778e-01 -8.86272252e-01 1.60110497e+00 -4.71653253e-01
1.12845552e+00 -3.20260644e-01 -1.13275802e+00 2.74434298e-01
6.59778476e-01 -5.64307868e-01 1.07046211e+00 -1.18846018e-02
-4.10023898e-01 4.48556930e-01 -1.27321076e+00 5.12561619e-01
8.83522213e-01 -6.32154584e-01 -1.04490495e+00 5.34417868e-01
6.65131032e-01 1.96604192e-01 -1.21154165e+00 -1.74819633e-01
9.48211104e-02 -9.08262491e-01 5.58256090e-01 -6.85543060e-01
-1.48698092e-01 -1.84386641e-01 1.53970532e-02 -1.23858464e+00
-2.09517628e-01 -6.11649036e-01 5.11665344e-01 1.75541222e+00
5.54916501e-01 -4.55849349e-01 5.19260824e-01 7.41679549e-01
1.82252914e-01 1.72655180e-01 -6.64356232e-01 -7.79037118e-01
-1.84679374e-01 -2.67076433e-01 3.94054711e-01 1.52966571e+00
6.82486713e-01 7.04270303e-01 5.28958499e-01 -3.45719792e-02
3.70310962e-01 -9.64853242e-02 7.15813518e-01 -1.73445380e+00
2.41675377e-01 -3.95911336e-01 -7.60958433e-01 2.31385127e-01
1.32727444e-01 -1.22601020e+00 -1.93619445e-01 -1.65128040e+00
1.10467412e-02 -1.49075016e-01 2.86692083e-02 2.73102641e-01
-1.09069765e-01 -5.16619802e-01 4.90709603e-01 4.15057153e-01
-7.46985734e-01 9.66420621e-02 4.32227254e-01 1.61638483e-01
-4.90974426e-01 -2.91905999e-01 -1.08934291e-01 7.53653049e-01
8.03570628e-01 -9.04195547e-01 -2.78916210e-01 2.16494292e-01
4.73455936e-01 -2.37966821e-01 -2.43682012e-01 -9.73680258e-01
6.46218717e-01 6.91083297e-02 -3.83764170e-02 -6.41272426e-01
-4.39325213e-01 -1.06755865e+00 3.63795847e-01 4.98638779e-01
-3.00655544e-01 6.07092381e-01 2.31385287e-02 5.09885132e-01
-2.87384689e-01 -7.31554747e-01 5.87008834e-01 -1.32959574e-01
-6.19815111e-01 -2.52357304e-01 -5.48257649e-01 2.91291684e-01
1.21161723e+00 -3.61677021e-01 -2.22227007e-01 1.20619297e-01
-7.51127839e-01 -1.40631841e-02 5.35240591e-01 3.52038920e-01
6.53812382e-03 -9.30617273e-01 -6.19669259e-01 -3.74365330e-01
3.96052778e-01 -2.84807205e-01 -3.21885161e-02 5.02909184e-01
-7.53319323e-01 3.67036998e-01 -1.48664460e-01 -3.83295387e-01
-1.84435976e+00 7.78331697e-01 -1.59773633e-01 -7.26819277e-01
-2.77153969e-01 1.86916381e-01 -3.71319681e-01 -8.39439213e-01
-5.43722510e-02 5.89137850e-03 -8.63981962e-01 6.53904617e-01
3.84017318e-01 7.98682272e-01 5.37564218e-01 -8.91118646e-01
-3.86416852e-01 2.87690639e-01 -2.20116284e-02 -8.21679458e-02
1.17390776e+00 -3.60441178e-01 -6.30652666e-01 4.60574090e-01
1.27308166e+00 4.65051159e-02 3.53787631e-01 -2.33868971e-01
1.09078360e+00 -2.80752957e-01 -1.19353972e-01 -4.42672819e-01
-6.06307626e-01 4.35105264e-01 3.95141333e-01 8.54997814e-01
7.62868941e-01 -2.39511773e-01 2.26392061e-01 9.68033850e-01
1.23928294e-01 -1.57000756e+00 -2.53598183e-01 6.33560479e-01
4.80902135e-01 -9.89916861e-01 3.52082662e-02 -6.41671538e-01
-6.33998990e-01 1.13173187e+00 1.81993529e-01 2.02407494e-01
7.53907144e-01 2.75266748e-02 5.75014483e-03 -5.71948349e-01
-2.76343644e-01 -4.14100140e-01 5.91366827e-01 5.41476965e-01
7.49926388e-01 3.78948227e-02 -9.55629051e-01 1.28322989e-01
-3.05241883e-01 -5.24494827e-01 3.24237913e-01 9.78408337e-01
-5.86622477e-01 -1.17619777e+00 -5.62327445e-01 7.06799090e-01
-3.99594873e-01 -2.86973804e-01 -1.16317761e+00 8.23911965e-01
-9.90088284e-02 1.25664222e+00 1.02477819e-01 -2.94170707e-01
2.58168191e-01 5.02444386e-01 -1.14374176e-01 -7.80556262e-01
-1.46599722e+00 -5.83124906e-02 3.36951107e-01 -3.56032640e-01
-7.11363256e-01 -5.81033170e-01 -1.52093041e+00 -6.62573516e-01
-7.20861912e-01 8.62669706e-01 1.03625286e+00 9.21483636e-01
5.94603181e-01 1.29091352e-01 6.40251994e-01 -3.12734753e-01
-1.98613778e-01 -1.00784957e+00 -7.78909087e-01 9.19950604e-01
-8.74565363e-01 1.22309178e-01 -4.13916707e-01 2.49058291e-01]
|
[9.979362487792969, 8.745014190673828]
|
486e0887-4aaf-47bf-ab44-873b70c9fbd3
|
an-improved-regret-analysis-for-ucb-n-and-ts
|
2305.04093
| null |
https://arxiv.org/abs/2305.04093v1
|
https://arxiv.org/pdf/2305.04093v1.pdf
|
An improved regret analysis for UCB-N and TS-N
|
In the setting of stochastic online learning with undirected feedback graphs, Lykouris et al. (2020) previously analyzed the pseudo-regret of the upper confidence bound-based algorithm UCB-N and the Thompson Sampling-based algorithm TS-N. In this note, we show how to improve their pseudo-regret analysis. Our improvement involves refining a key lemma of the previous analysis, allowing a $\log(T)$ factor to be replaced by a factor $\log_2(\alpha) + 3$ for $\alpha$ the independence number of the feedback graph.
|
['Nishant A. Mehta']
|
2023-05-06
| null | null | null | null |
['thompson-sampling']
|
['methodology']
|
[-1.27983943e-01 6.51973426e-01 -4.13785398e-01 -3.04779023e-01
-8.61362696e-01 -6.02284014e-01 -3.57210368e-01 3.70089740e-01
-6.25238776e-01 1.23558986e+00 -2.46999145e-01 -7.05131948e-01
-5.85443139e-01 -8.04440916e-01 -1.04598629e+00 -6.34639680e-01
-7.75224388e-01 4.12428349e-01 3.06000024e-01 -5.45691177e-02
6.75585214e-03 2.28628099e-01 -1.17192662e+00 -1.96528226e-01
7.04089224e-01 1.28415167e+00 -2.90391892e-01 9.43554699e-01
-9.70959738e-02 6.41024649e-01 -3.19133073e-01 -7.63151467e-01
7.86437511e-01 -8.77310097e-01 -7.89521873e-01 7.45345280e-02
1.77420542e-01 -2.63393521e-01 -5.58063447e-01 1.36343658e+00
2.95250952e-01 2.88027585e-01 4.21703875e-01 -1.26924884e+00
-6.92235604e-02 1.17096102e+00 -9.41684544e-01 2.92803347e-01
2.58792192e-01 -3.88590008e-01 1.13904381e+00 2.71020383e-02
6.07376039e-01 1.10967588e+00 5.04426301e-01 6.61576748e-01
-1.11337340e+00 -9.41674590e-01 3.08147132e-01 3.41006309e-01
-1.22928405e+00 -6.79347664e-02 5.75841367e-01 -2.28506103e-01
5.59607685e-01 4.03432757e-01 7.51577675e-01 4.40327138e-01
-1.32079676e-01 9.60505009e-01 1.10996258e+00 -8.50284994e-01
4.72297966e-01 1.82074506e-03 1.10465363e-01 1.06369996e+00
7.22480416e-01 2.59217411e-01 -6.92780316e-01 -2.57600814e-01
7.06826985e-01 -2.02500492e-01 -1.71284258e-01 -5.85444987e-01
-2.76111305e-01 1.06262636e+00 4.47975665e-01 -1.75018713e-01
-1.66771915e-02 6.19099915e-01 4.13450450e-01 9.55761909e-01
6.21985614e-01 -2.66172141e-02 -5.45165896e-01 -2.37215027e-01
-7.81687737e-01 -1.67312734e-02 1.18633223e+00 1.38289189e+00
7.89071560e-01 -4.25300822e-02 -1.19691603e-02 3.21149200e-01
2.95868218e-01 3.74545544e-01 -9.58279800e-03 -1.04462278e+00
7.21702337e-01 2.88720042e-01 5.29086292e-01 -5.80825329e-01
-2.28931502e-01 -4.99016345e-01 -5.48698962e-01 4.57765721e-02
9.56806779e-01 -6.57291651e-01 -4.55677301e-01 1.87701344e+00
3.78486186e-01 -5.20036578e-01 -2.76442051e-01 5.96588135e-01
-3.28653693e-01 9.54271704e-02 -5.63973844e-01 -8.09241474e-01
5.52614510e-01 -7.77072668e-01 -6.96532249e-01 4.04445156e-02
7.84321368e-01 -4.09517497e-01 8.14642489e-01 5.45680940e-01
-1.20749867e+00 1.68899462e-01 -1.04664123e+00 5.12328684e-01
1.15891337e-01 -2.77941257e-01 9.34629619e-01 1.25320256e+00
-1.16562593e+00 8.71650875e-01 -6.78346336e-01 -1.79094374e-01
4.70489115e-01 3.17357033e-01 6.82176603e-03 -1.79500848e-01
-9.89077568e-01 3.41200858e-01 1.67490602e-01 -2.48542409e-02
-9.79867399e-01 -5.17410874e-01 -4.59867865e-01 -1.15456041e-02
1.02198505e+00 -4.61110324e-01 1.38793087e+00 -9.71764863e-01
-1.49688542e+00 3.76538754e-01 -1.03396647e-01 -8.59400153e-01
1.07330775e+00 -1.73935845e-01 2.92326212e-01 1.54626131e-01
-1.39743686e-01 -1.62291601e-01 5.73652864e-01 -8.85582685e-01
-7.04156220e-01 -5.71415722e-01 4.16669428e-01 2.27096245e-01
-1.58227146e-01 -1.45121709e-01 -2.38353670e-01 -2.83725768e-01
-6.69048727e-02 -9.15888667e-01 -5.74690938e-01 5.52351214e-02
-1.41123712e-01 -1.40340775e-01 5.36351688e-02 -2.70938784e-01
1.22041380e+00 -1.99062324e+00 -7.03561381e-02 5.83322048e-01
2.48430565e-01 -2.72763938e-01 1.97121337e-01 5.01832724e-01
1.79608330e-01 5.08861244e-01 1.84983775e-01 -3.34335715e-02
-4.05917354e-02 3.53917703e-02 1.02830917e-01 7.58883893e-01
-7.65811503e-01 4.53668952e-01 -1.08798409e+00 -3.48916799e-01
-3.00291628e-01 -4.01703417e-01 -5.34777701e-01 -1.86932012e-01
-3.95595312e-01 -1.99063733e-01 -6.65624321e-01 4.29401845e-01
4.70955282e-01 -5.68468332e-01 6.01304233e-01 4.38065022e-01
1.67083383e-01 2.08818559e-02 -1.51326561e+00 1.36933696e+00
-3.75590801e-01 4.21898067e-01 6.13864660e-01 -9.49312806e-01
4.30349827e-01 2.58782476e-01 6.55428529e-01 -2.27270156e-01
4.78388190e-01 2.40180537e-01 -1.23136215e-01 -1.44233331e-01
9.51438099e-02 -4.88896877e-01 -1.17900334e-02 8.72812629e-01
7.69190788e-02 2.58175284e-01 2.75741577e-01 4.77751940e-01
1.41913521e+00 -2.38788605e-01 2.57417291e-01 -4.59120661e-01
1.68288961e-01 -1.24400131e-01 4.96480495e-01 1.31323802e+00
-5.67959011e-01 4.85553257e-02 1.07404816e+00 -6.76286891e-02
-9.14769411e-01 -7.90560901e-01 2.83366799e-01 1.32441890e+00
-1.58112068e-02 -4.63554144e-01 -6.72898710e-01 -1.03898358e+00
3.33741635e-01 6.67540789e-01 -1.18860734e+00 -1.25369253e-02
1.58780679e-01 -5.53717017e-01 2.37583652e-01 4.79174435e-01
1.76912561e-01 -3.51911068e-01 -4.14789587e-01 8.90375972e-02
1.18157245e-01 -5.38286030e-01 -6.90451682e-01 6.74454749e-01
-1.10382557e+00 -1.24194729e+00 -3.93239886e-01 -1.44877449e-01
7.64030695e-01 2.59712249e-01 6.36944652e-01 -6.84547499e-02
-2.94485483e-02 6.32418871e-01 -5.37073851e-01 -6.57471955e-01
-1.33872837e-01 7.93677494e-02 7.73674250e-02 -2.46576339e-01
3.33188549e-02 -3.62876892e-01 -5.35351932e-01 2.09915355e-01
-5.92922270e-01 -2.89948583e-01 2.59143144e-01 8.26410949e-01
5.29941320e-01 -7.94983208e-02 4.61160690e-01 -1.26084530e+00
4.17547137e-01 -3.97086233e-01 -1.28389800e+00 4.68699545e-01
-1.16214943e+00 4.18243527e-01 8.17564309e-01 -2.32015342e-01
-6.65505052e-01 5.06142303e-02 3.05131376e-01 -4.75075215e-01
7.21119225e-01 5.50390780e-01 1.71035796e-01 -4.22767252e-01
6.96936786e-01 -2.49814019e-02 -1.08503141e-01 -2.22343370e-01
3.74272406e-01 2.85750002e-01 -1.33811891e-01 -6.27160192e-01
6.52355373e-01 3.89852315e-01 3.29656124e-01 -2.00234085e-01
-1.13717794e+00 -3.70989710e-01 -3.59804809e-01 -4.31455344e-01
1.60523593e-01 -6.19996250e-01 -1.23526847e+00 -5.87312505e-02
-5.06059527e-01 -5.87460518e-01 -4.81710881e-01 5.46091020e-01
-9.19396341e-01 4.35151905e-01 -4.26446646e-01 -1.52324843e+00
-8.84282663e-02 -4.32544082e-01 3.66851464e-02 2.94817775e-01
4.58651811e-01 -7.79794395e-01 5.96937835e-02 2.62970775e-01
2.67104894e-01 1.51714608e-01 5.57110071e-01 -6.64758146e-01
-3.82522851e-01 -3.61697078e-01 -2.40540653e-01 2.19190449e-01
-4.58186716e-02 -2.39905998e-01 -5.00971913e-01 -6.63273156e-01
-7.97007829e-02 -4.61328149e-01 8.38847935e-01 4.87042248e-01
1.25026536e+00 -7.51288891e-01 -1.55844510e-01 6.10587597e-01
1.86715078e+00 1.78615317e-01 9.16212574e-02 2.95128793e-01
2.67174393e-01 1.75954476e-01 7.91807652e-01 9.93457973e-01
1.25253173e-02 4.32021283e-02 5.33762336e-01 6.13243580e-01
3.72923166e-01 -4.32600856e-01 3.03290188e-01 3.75684261e-01
9.91905294e-03 -2.13554814e-01 -5.70512235e-01 6.53626621e-01
-2.06004691e+00 -6.80897772e-01 2.56184563e-02 2.75199819e+00
1.11416888e+00 4.49208677e-01 3.62912923e-01 2.47161034e-02
8.77038062e-01 -1.24247685e-01 -8.13224852e-01 -7.55420446e-01
6.31128773e-02 3.03873926e-01 1.49887931e+00 7.95050323e-01
-5.38282275e-01 7.10281730e-01 6.87653589e+00 1.09311461e+00
-3.89621854e-01 1.37325823e-01 6.48723900e-01 -7.51401842e-01
-3.62518072e-01 1.86375126e-01 -6.84730530e-01 4.20307219e-01
1.32347107e+00 -8.65287781e-01 8.15780163e-01 1.09419394e+00
4.22211513e-02 -5.35592377e-01 -1.05959034e+00 5.79280674e-01
-9.47819427e-02 -1.01536572e+00 -3.91967565e-01 2.36935899e-01
9.62117374e-01 1.64186340e-02 -2.31484190e-01 1.78754494e-01
1.18578041e+00 -8.43884706e-01 4.78605896e-01 1.13727920e-01
8.86655986e-01 -1.16907239e+00 6.44118726e-01 3.60483319e-01
-1.06187403e+00 -2.97821909e-01 -4.59180832e-01 -1.37125134e-01
-1.23693116e-01 7.32386291e-01 -8.36847782e-01 9.00728822e-01
7.63212621e-01 3.65023389e-02 -1.62890449e-01 1.07557678e+00
-3.09182853e-01 8.66698086e-01 -8.18321109e-01 -5.04424393e-01
1.70655623e-01 -3.36221486e-01 3.48218203e-01 9.08499181e-01
1.08332165e-01 2.87650943e-01 -1.96746681e-02 2.37905219e-01
-3.85587275e-01 1.87954620e-01 -2.40379572e-01 -1.68884303e-02
5.42780519e-01 9.45358396e-01 -7.87265599e-01 -2.52867281e-01
-2.52521276e-01 7.37063229e-01 6.36441708e-01 4.50391918e-02
-7.59607613e-01 -6.03086233e-01 3.84070963e-01 -4.95045073e-02
9.24419820e-01 -1.82258248e-01 -2.21195072e-01 -8.50539923e-01
1.27738593e-02 -3.77217144e-01 8.44498456e-01 -1.23325855e-01
-1.10406268e+00 -3.58487442e-02 -7.34236836e-02 -8.71228278e-01
-2.18150243e-01 -4.33322966e-01 -3.36313583e-02 7.09204197e-01
-1.21149814e+00 -3.89931470e-01 2.96471298e-01 7.51951396e-01
-3.53971869e-02 2.65762210e-01 5.35413444e-01 -6.98883235e-02
-4.40044463e-01 1.09380794e+00 9.51210022e-01 3.12819518e-02
5.70658207e-01 -1.44899786e+00 -3.50416303e-01 9.50505614e-01
-5.44301122e-02 3.31084490e-01 1.05554628e+00 -4.90970433e-01
-1.52151310e+00 -8.31406415e-01 5.49279630e-01 1.79119855e-01
1.03104484e+00 -3.49501073e-01 -2.75283307e-01 8.45199704e-01
-8.12121779e-02 1.06750228e-01 9.44936395e-01 4.30509418e-01
-4.38902617e-01 -4.59089637e-01 -1.45591247e+00 4.01986480e-01
1.27691197e+00 -2.73536444e-01 6.10457174e-03 3.94361079e-01
6.56484663e-01 -4.40469891e-01 -9.69048142e-01 2.81884205e-02
7.74860203e-01 -1.10448027e+00 1.29011139e-01 -7.92391419e-01
2.77085397e-02 2.31785253e-01 -1.89737663e-01 -1.19861519e+00
-1.41527103e-02 -1.11484468e+00 -4.37391967e-01 6.65553510e-01
6.00285888e-01 -7.56927371e-01 1.23546207e+00 4.06273067e-01
3.51278394e-01 -7.71180630e-01 -1.29903400e+00 -1.03942394e+00
2.59001493e-01 -4.69768286e-01 5.31229153e-02 4.91350383e-01
4.97185439e-01 -1.60141706e-01 -4.56862718e-01 -1.39466271e-01
8.35785806e-01 1.11372219e-02 5.54936886e-01 -9.34889495e-01
-9.00215447e-01 -4.19136792e-01 -2.63460982e-03 -1.00504100e+00
-3.89900416e-01 -6.22951090e-01 -2.07279157e-02 -1.16265273e+00
4.07275647e-01 -5.27805746e-01 -6.83486342e-01 3.66926700e-01
5.21039441e-02 -2.69359559e-01 3.05951208e-01 -2.11613581e-01
-1.19346333e+00 4.49279398e-01 1.20323956e+00 2.92276800e-01
4.19231765e-02 4.51732367e-01 -8.64953458e-01 2.58530945e-01
8.14636827e-01 -7.43882477e-01 -4.52438533e-01 -2.26423204e-01
7.92733252e-01 6.38614297e-01 -2.14267477e-01 -5.98494530e-01
3.25334609e-01 -2.83461154e-01 1.05941787e-01 -5.71043313e-01
-1.90422386e-01 -7.33579934e-01 -1.23286590e-01 8.76356363e-01
-6.91362739e-01 -4.42095846e-02 8.13490376e-02 1.37821591e+00
3.43386531e-01 -5.27114630e-01 6.90904796e-01 -1.81566775e-01
1.58051610e-01 3.25037360e-01 -4.12716031e-01 2.45299831e-01
1.13098025e+00 -2.69934423e-02 -1.54704645e-01 -8.85754168e-01
-6.86489940e-01 5.26476562e-01 3.66869479e-01 -3.09050232e-01
1.61549613e-01 -9.06446993e-01 -3.51626605e-01 -2.33171955e-01
-2.21770391e-01 -3.24993171e-02 3.58525872e-01 9.85474885e-01
-3.51631641e-01 3.20273668e-01 -4.62596752e-02 -3.56172509e-02
-1.18506515e+00 6.61456347e-01 3.70937824e-01 -5.61563075e-01
-6.87531307e-02 1.19588327e+00 -5.53732574e-01 1.67521998e-01
5.71908474e-01 -5.14782332e-02 5.70467949e-01 -1.63096055e-01
3.10187280e-01 8.69558632e-01 -1.12690970e-01 4.18969274e-01
-3.32731664e-01 -6.83955848e-02 -2.76795775e-01 -5.55732906e-01
1.23401701e+00 -4.93533850e-01 1.50863335e-01 5.67823231e-01
1.01074040e+00 1.63089201e-01 -1.41136503e+00 -5.02315462e-01
5.88590801e-02 -6.31166339e-01 -7.11415857e-02 -8.63089740e-01
-1.13285303e+00 3.76300812e-01 5.26944518e-01 6.02300525e-01
1.04365170e+00 -3.46675292e-02 1.72576666e-01 4.57257211e-01
1.08445740e+00 -1.29572701e+00 -2.89147824e-01 4.67083216e-01
3.80553693e-01 -1.10152650e+00 2.11160079e-01 -2.07731709e-01
-3.18789631e-01 1.09437072e+00 3.24374139e-01 -3.13058376e-01
7.72720933e-01 2.25960866e-01 -4.73398566e-01 2.22167313e-01
-1.05035281e+00 -2.65977263e-01 -1.83716282e-01 2.82367140e-01
4.83161420e-01 1.94613799e-01 -8.39675665e-01 6.91941559e-01
-2.39245236e-01 1.27376780e-01 7.43169188e-01 1.07293952e+00
-6.78547204e-01 -1.21102941e+00 -2.24144608e-01 8.27028871e-01
-6.72764361e-01 4.16227169e-02 -4.91065860e-01 7.46676266e-01
-4.64419514e-01 1.11210012e+00 -2.92073131e-01 -4.70601380e-01
-3.66698466e-02 7.63129815e-02 1.00116229e+00 -3.46948802e-01
-4.20016646e-01 2.25249622e-02 9.09819156e-02 -4.75070715e-01
-2.06847876e-01 -8.26310098e-01 -9.94412422e-01 -7.08843470e-01
-6.21082485e-01 6.73196077e-01 5.76044798e-01 7.67654538e-01
1.18249297e-01 2.27189258e-01 1.09289336e+00 -9.69206244e-02
-1.18611217e+00 -1.03273010e+00 -1.14739394e+00 -1.92347676e-01
1.98348880e-01 -6.31387293e-01 -9.50819314e-01 -4.64632124e-01]
|
[4.6529927253723145, 3.4188485145568848]
|
7162aa23-3c42-46e8-97c1-6a7b60b0c31d
|
early-myocardial-infarction-detection-over
|
2111.05790
| null |
https://arxiv.org/abs/2111.05790v3
|
https://arxiv.org/pdf/2111.05790v3.pdf
|
Early Myocardial Infarction Detection over Multi-view Echocardiography
|
Myocardial infarction (MI) is the leading cause of mortality in the world that occurs due to a blockage of the coronary arteries feeding the myocardium. An early diagnosis of MI and its localization can mitigate the extent of myocardial damage by facilitating early therapeutic interventions. Following the blockage of a coronary artery, the regional wall motion abnormality (RWMA) of the ischemic myocardial segments is the earliest change to set in. Echocardiography is the fundamental tool to assess any RWMA. Assessing the motion of the left ventricle (LV) wall only from a single echocardiography view may lead to missing the diagnosis of MI as the RWMA may not be visible on that specific view. Therefore, in this study, we propose to fuse apical 4-chamber (A4C) and apical 2-chamber (A2C) views in which a total of 12 myocardial segments can be analyzed for MI detection. The proposed method first estimates the motion of the LV wall by Active Polynomials (APs), which extract and track the endocardial boundary to compute myocardial segment displacements. The features are extracted from the A4C and A2C view displacements, which are concatenated and fed into the classifiers to detect MI. The main contributions of this study are 1) creation of a new benchmark dataset by including both A4C and A2C views in a total of 260 echocardiography recordings, which is publicly shared with the research community, 2) improving the performance of the prior work of threshold-based APs by a Machine Learning based approach, and 3) a pioneer MI detection approach via multi-view echocardiography by fusing the information of A4C and A2C views. Experimental results show that the proposed method achieves 90.91% sensitivity and 86.36% precision for MI detection over multi-view echocardiography. The software implementation is shared at https://github.com/degerliaysen/MultiEchoAI.
|
['Moncef Gabbouj', 'Rashid Mazhar', 'Tahir Hamid', 'Serkan Kiranyaz', 'Aysen Degerli']
|
2021-11-09
| null | null | null | null |
['myocardial-infarction-detection']
|
['medical']
|
[ 1.67554975e-01 -2.51116186e-01 -1.32080942e-01 2.02927947e-01
-5.35683036e-01 -9.28806663e-01 -3.27742621e-02 -2.81583816e-02
-4.17102575e-02 4.39433783e-01 -2.13929992e-02 -4.85550433e-01
-9.50785726e-02 -5.51437795e-01 -1.36013687e-01 -9.12230968e-01
-3.24151278e-01 2.92587519e-01 2.48184070e-01 4.56813127e-01
1.78917691e-01 7.14757621e-01 -9.19680893e-01 3.60042036e-01
5.41433752e-01 9.76859570e-01 2.95610070e-01 1.26950490e+00
4.63717580e-01 3.90504897e-01 -4.25980002e-01 3.06754589e-01
3.26234221e-01 -8.25952232e-01 -7.04400241e-01 1.51130304e-01
1.80473685e-01 -6.67074859e-01 2.09153160e-01 5.08641362e-01
8.31808925e-01 -4.54239547e-01 4.95057315e-01 -7.13219047e-01
2.73786187e-01 2.51106858e-01 -5.75972676e-01 6.66342795e-01
-8.29496905e-02 -1.36612719e-02 6.60266876e-01 -1.18610823e+00
7.61686265e-01 4.34645444e-01 8.09347212e-01 1.85820907e-01
-9.18772936e-01 -3.46816987e-01 -3.98346364e-01 1.18013285e-01
-1.42604828e+00 -1.09753124e-01 1.00562203e+00 -6.88730419e-01
3.72507840e-01 6.21721745e-01 8.80233824e-01 2.33171985e-01
3.03115129e-01 4.09105390e-01 1.24765050e+00 -3.58912498e-01
-9.02509317e-02 -2.25393161e-01 9.23739374e-02 6.22527957e-01
2.77459651e-01 1.25389770e-01 1.03555724e-01 -5.55976570e-01
9.57833469e-01 3.27397943e-01 -4.62322921e-01 -3.49367440e-01
-1.46839881e+00 4.93198186e-01 -1.89584699e-02 4.30509239e-01
-5.04529715e-01 -2.09672362e-01 6.33671224e-01 2.57682145e-01
2.30944172e-01 1.07764401e-01 -4.73582566e-01 -1.79741904e-01
-1.04473889e+00 1.78053796e-01 4.12214816e-01 1.54494047e-01
2.38943025e-01 -2.22214490e-01 -3.51975560e-02 6.60853386e-01
3.93993855e-01 6.94254220e-01 3.82726133e-01 -1.21272290e+00
2.78253287e-01 7.89194465e-01 3.03193070e-02 -9.99945223e-01
-3.93230140e-01 -6.10444546e-01 -1.12484479e+00 4.34251219e-01
5.95937014e-01 -2.87899524e-01 -5.05785704e-01 1.12790024e+00
4.67160940e-01 1.51707053e-01 -1.50772156e-02 1.17812204e+00
7.84622788e-01 3.71210724e-01 -2.06451923e-01 -6.23945832e-01
1.55976820e+00 -5.62875330e-01 -3.82627279e-01 6.76696226e-02
1.02394998e+00 -8.67925286e-01 3.68872076e-01 3.69217068e-01
-9.88181174e-01 -5.97066402e-01 -1.03112411e+00 5.09978235e-01
2.46019199e-01 7.00187624e-01 6.83915392e-02 6.03947461e-01
-8.13971817e-01 6.64517581e-01 -1.14801311e+00 -1.40229896e-01
2.44353637e-01 1.17709845e-01 -5.08246481e-01 -4.34283726e-02
-9.81242776e-01 7.30626225e-01 8.45443606e-02 5.15684724e-01
-5.03943264e-01 -7.79784501e-01 -5.33358157e-01 -2.00754672e-01
3.13559175e-01 -7.96179473e-01 4.52746630e-01 -6.90943122e-01
-8.98322344e-01 1.10501730e+00 -2.79130787e-01 -1.53474167e-01
6.69094443e-01 -1.73474371e-01 5.16902991e-02 6.67245686e-01
1.30349860e-01 5.87815121e-02 5.22654533e-01 -1.11851943e+00
-6.19706869e-01 -6.01208568e-01 -4.14756328e-01 1.43107042e-01
2.58410186e-01 1.09776340e-01 -1.86725602e-01 -8.82644594e-01
6.86626494e-01 -1.24819589e+00 -1.83760658e-01 -2.53889829e-01
-3.81750882e-01 2.60728091e-01 8.21326733e-01 -1.19815850e+00
1.51753700e+00 -2.13340449e+00 1.74077988e-01 3.51846188e-01
7.23200798e-01 6.11584425e-01 5.73735416e-01 3.29791933e-01
-1.54727578e-01 4.35653299e-01 -3.05071205e-01 1.58581942e-01
-7.30779409e-01 -2.27236860e-02 -6.12291507e-02 6.93737745e-01
-3.05012196e-01 7.81741560e-01 -4.26861048e-01 -8.17629158e-01
2.12617531e-01 1.77341759e-01 -1.86984107e-01 1.56358808e-01
7.06502616e-01 8.39319825e-01 -5.06015480e-01 6.78885698e-01
7.68103600e-01 -2.41343051e-01 6.75281525e-01 -1.27656221e-01
-2.10504442e-01 -7.68072829e-02 -1.28145063e+00 1.30486929e+00
1.32944554e-01 2.42411539e-01 1.04950033e-01 -1.11410856e+00
8.69440019e-01 9.92042601e-01 9.47871745e-01 -3.12876031e-02
1.39726035e-03 5.52430749e-01 4.16559517e-01 -6.90273643e-01
-5.24258077e-01 -2.10912935e-02 9.23143029e-02 4.60162193e-01
-4.82227713e-01 1.83482453e-01 2.29004860e-01 -5.00199646e-02
1.02348256e+00 3.19705576e-01 7.04708517e-01 -2.59563774e-01
9.24385071e-01 -1.64428547e-01 1.21082842e+00 7.45260417e-01
-4.77359891e-01 9.92611110e-01 7.03051388e-01 -8.32644522e-01
-9.45206285e-01 -9.83281851e-01 -4.10593808e-01 1.65571079e-01
-3.22030112e-02 -3.14199030e-01 -5.53456962e-01 -6.42706335e-01
-4.37612608e-02 -2.04663828e-01 -3.17532778e-01 2.09924102e-01
-1.18923819e+00 -8.82006109e-01 4.64318663e-01 7.30583072e-01
4.12918001e-01 -7.19711423e-01 -1.27028215e+00 2.68446892e-01
-5.10259390e-01 -8.92208159e-01 -2.41183013e-01 -4.13791358e-01
-1.71819282e+00 -1.29264462e+00 -1.19052851e+00 -6.93210304e-01
5.51118731e-01 1.12863995e-01 9.61282372e-01 4.85191762e-01
-6.42043769e-01 -3.15185152e-02 -2.13976800e-01 -2.00698733e-01
-5.88973641e-01 -1.50896385e-01 -9.58960280e-02 -2.38350574e-02
-3.81591529e-01 -7.11787283e-01 -1.10486031e+00 4.56273973e-01
-1.98869050e-01 2.31664032e-01 5.87861300e-01 6.98292613e-01
7.87258983e-01 -4.00064021e-01 6.77599311e-01 -9.29768384e-01
1.26138581e-02 -3.46614569e-01 -4.41562682e-01 1.01781674e-01
-8.62615407e-01 -7.79052496e-01 3.09182614e-01 -3.14629823e-02
-6.06511891e-01 1.37541592e-01 -4.52363268e-02 -4.64508682e-01
-3.61245513e-01 5.34413815e-01 7.81275928e-02 1.70120597e-01
3.08186769e-01 1.26705736e-01 1.85725868e-01 -3.56402010e-01
-2.86485642e-01 4.49076921e-01 5.31843483e-01 -2.64030367e-01
4.56295699e-01 5.67173481e-01 4.83130217e-01 -7.90620387e-01
-3.34720105e-01 -8.95629823e-01 -9.81662154e-01 -4.40863639e-01
8.69208395e-01 -6.11822069e-01 -5.31880975e-01 5.59365213e-01
-1.08892655e+00 1.13599293e-01 1.55861154e-02 8.84653449e-01
-3.67849737e-01 8.15640271e-01 -7.59147048e-01 -6.62460744e-01
-7.99053907e-01 -1.02857602e+00 7.65609324e-01 -7.98805058e-02
-3.40352148e-01 -8.74130189e-01 3.15535843e-01 4.10781115e-01
1.04833938e-01 7.40637779e-01 1.03971493e+00 -5.02710402e-01
-5.35485804e-01 -4.00135338e-01 4.52654026e-02 5.24240196e-01
7.90862814e-02 1.05158821e-01 -7.99284816e-01 -2.37755194e-01
1.53865606e-01 4.37170178e-01 6.54788315e-01 1.00366950e+00
5.04145503e-01 1.90640822e-01 -5.10548830e-01 6.14958048e-01
1.31731784e+00 4.85339284e-01 6.61473930e-01 1.93512127e-01
7.11022198e-01 3.60310376e-01 7.48353720e-01 4.01109129e-01
1.13361925e-01 6.23035789e-01 2.96159089e-01 -5.20380855e-01
-2.57107913e-01 4.37358826e-01 3.59494388e-02 1.06678998e+00
-8.45694184e-01 2.52002031e-01 -1.21486306e+00 6.26626790e-01
-1.68835604e+00 -9.17389214e-01 -6.79610789e-01 2.37665868e+00
5.97406983e-01 3.41863222e-02 8.21227655e-02 5.14782429e-01
7.12853670e-01 -6.47373870e-02 -2.56790370e-01 -1.36857361e-01
1.23752750e-01 1.56140747e-02 2.64013827e-01 3.09434593e-01
-1.24815440e+00 -2.41586585e-02 5.47705078e+00 2.33589523e-02
-1.35468388e+00 -1.13830203e-02 7.91004419e-01 3.43113303e-01
2.69230515e-01 1.73335135e-01 -5.94395161e-01 3.44674826e-01
6.42594993e-01 1.23426817e-01 -3.06790441e-01 5.45262516e-01
4.84814644e-01 -1.81960672e-01 -8.73937011e-01 9.61663961e-01
-6.42893985e-02 -1.50267422e+00 -6.29325688e-01 8.12175870e-02
3.60939324e-01 -1.69720098e-01 -2.97492176e-01 -1.20914467e-01
-1.02343416e+00 -6.30092859e-01 3.20890516e-01 7.38303423e-01
1.00098240e+00 -3.94477785e-01 1.18521154e+00 4.68925416e-01
-1.33461607e+00 7.39329681e-02 1.64646804e-01 -3.58140394e-02
1.51170015e-01 6.00039363e-01 -9.20842171e-01 6.12519503e-01
6.33306563e-01 8.17439198e-01 -3.14213932e-01 9.42232728e-01
8.75855088e-02 1.11415303e+00 -7.42130950e-02 7.45911896e-01
-4.22955275e-01 -3.28235000e-01 1.12756228e+00 7.23906398e-01
4.09338236e-01 2.43228465e-01 4.28606033e-01 7.61997402e-01
2.93032914e-01 2.43118495e-01 -6.34632707e-01 4.09888327e-01
3.67651105e-01 1.39720309e+00 -1.08438790e+00 -4.57134485e-01
-5.06057680e-01 5.86135805e-01 -5.10800064e-01 3.33784431e-01
-5.07128716e-01 -2.30269998e-01 6.71945214e-02 4.46418732e-01
6.71220571e-02 4.51148748e-02 -6.54000819e-01 -9.60582435e-01
3.65384996e-01 -8.84569049e-01 5.27251720e-01 -4.23377603e-01
-5.57285905e-01 4.59491163e-01 9.05682705e-03 -1.53415298e+00
-5.31482756e-01 -4.51801062e-01 -8.25467169e-01 1.38950944e+00
-1.11256337e+00 -8.99859846e-01 -1.84375092e-01 -3.33561748e-02
2.88285643e-01 1.79014187e-02 1.15256786e+00 3.96112591e-01
-4.10707563e-01 1.30817860e-01 -3.11855376e-01 4.14659292e-01
7.04911232e-01 -1.40764427e+00 -2.88954732e-04 9.78482187e-01
-4.40403432e-01 5.28605938e-01 4.51553851e-01 -7.51412451e-01
-9.74248886e-01 -7.27569818e-01 1.27353573e+00 -4.73232955e-01
-4.46629450e-02 2.32077494e-01 -7.31479824e-01 4.46145803e-01
-3.46259028e-01 5.65885782e-01 7.76020110e-01 -4.21004266e-01
3.22654635e-01 -1.42987549e-01 -8.17033231e-01 2.05008179e-01
4.23825294e-01 -2.62759805e-01 -3.37572664e-01 -1.89070433e-01
-1.93290547e-01 -4.74066675e-01 -1.45208895e+00 9.21508491e-01
1.00498092e+00 -1.13400805e+00 1.14917779e+00 -1.72327623e-01
3.25912625e-01 -3.40720117e-01 2.09254801e-01 -6.30304813e-01
-1.37049794e-01 -5.68855286e-01 -8.08961242e-02 9.50348377e-01
1.71402931e-01 -6.37331903e-01 6.95976198e-01 4.98043150e-02
-2.48883516e-01 -1.14525640e+00 -8.10622573e-01 -1.55738115e-01
1.07368760e-01 -7.78081417e-02 -2.09778264e-01 8.55852902e-01
-2.11574286e-01 -1.77803919e-01 -1.91565827e-01 2.00521439e-01
5.66634476e-01 5.91031313e-01 4.42651182e-01 -1.45591879e+00
-2.29421586e-01 -1.17758349e-01 -4.68150645e-01 -4.34265047e-01
-5.44287860e-01 -9.92385805e-01 -3.67079437e-01 -1.47488821e+00
4.73484159e-01 -5.95489919e-01 -4.18720931e-01 6.22556321e-02
-4.61037695e-01 5.06457388e-01 1.82961330e-01 6.73835337e-01
1.92834541e-01 -4.54359800e-01 1.48595440e+00 4.31817681e-01
-3.84041280e-01 4.71481740e-01 -1.52715430e-01 1.19691300e+00
8.78734529e-01 -5.86384416e-01 -3.91741246e-01 1.48336887e-01
1.06688768e-01 9.29521203e-01 7.10096955e-01 -8.70511353e-01
-2.75941849e-01 2.67278731e-01 5.92478991e-01 -9.33699489e-01
6.54040575e-02 -4.45583165e-01 1.81147635e-01 8.51431012e-01
-5.88085176e-03 2.39877880e-01 -1.19800046e-01 2.34990314e-01
-2.21081436e-01 -1.81478441e-01 8.19127560e-01 -4.35199916e-01
-3.12753797e-01 2.53473759e-01 -7.07547903e-01 2.75600493e-01
1.07261181e+00 -3.89449209e-01 1.33172991e-02 6.49825111e-02
-1.23091125e+00 -1.29359484e-01 1.24100566e-01 -1.19999178e-01
7.93730497e-01 -9.96974051e-01 -1.06804526e+00 2.03105882e-01
-1.97937876e-01 -1.02728412e-01 5.03983855e-01 1.87884223e+00
-1.03875470e+00 3.67628157e-01 -1.85260370e-01 -1.11732209e+00
-1.81611407e+00 1.98008716e-01 6.88414454e-01 -5.51878572e-01
-1.24045300e+00 4.71235365e-01 2.53675729e-01 -3.82834044e-03
-9.95113477e-02 -3.18732888e-01 -4.99894261e-01 -3.92623013e-03
4.32075053e-01 7.37380147e-01 1.56870410e-02 -7.11245716e-01
-5.97025514e-01 8.98735285e-01 1.98073223e-01 7.40282238e-02
7.47588694e-01 -4.81517166e-01 -3.29495966e-01 6.65076613e-01
9.16019678e-01 2.88480163e-01 -7.16814697e-01 6.38872087e-02
-2.14544177e-01 -3.56697530e-01 1.47075981e-01 -7.16529965e-01
-1.11142492e+00 1.08937478e+00 1.02602136e+00 6.18106537e-02
1.14017808e+00 -1.51316375e-01 8.87599349e-01 -2.58721828e-01
4.16052267e-02 -6.30328417e-01 -3.46966267e-01 -2.41312295e-01
6.77437603e-01 -1.01181793e+00 2.18461677e-01 -5.73384106e-01
-7.32388318e-01 1.45786297e+00 1.59034394e-02 -1.97461084e-01
8.72785151e-01 2.87699491e-01 6.53883100e-01 -7.96459317e-02
-5.63351333e-01 1.36554703e-01 2.59569615e-01 2.71277696e-01
7.63276696e-01 1.91543773e-01 -8.17880213e-01 6.12560570e-01
4.94928539e-01 2.54886389e-01 3.88742983e-01 1.14554548e+00
-3.20576578e-01 -1.09460485e+00 -4.75072563e-01 3.93592507e-01
-1.15984726e+00 -5.25943674e-02 2.36919848e-03 7.12700486e-01
5.61476231e-01 7.57208109e-01 -2.93965757e-01 3.34278494e-02
2.41302967e-01 4.13959742e-01 2.71220297e-01 -4.25474912e-01
-5.75026333e-01 7.02341139e-01 2.94959992e-02 -3.30554634e-01
-4.56904382e-01 -1.04528081e+00 -1.39484918e+00 1.84376329e-01
-9.61941704e-02 1.14284260e-02 3.38815331e-01 8.63646030e-01
3.93759549e-01 6.09382272e-01 7.18338311e-01 -4.69539076e-01
-5.06329060e-01 -8.39606047e-01 -8.41481805e-01 3.78764540e-01
4.46531951e-01 -4.61010426e-01 -4.00831044e-01 6.29158974e-01]
|
[14.247176170349121, -2.3763816356658936]
|
94623665-4cd2-4c48-a46b-7061ea8c9801
|
implicit-and-explicit-aspect-extraction-in
| null | null |
https://aclanthology.org/W18-3108
|
https://aclanthology.org/W18-3108.pdf
|
Implicit and Explicit Aspect Extraction in Financial Microblogs
|
This paper focuses on aspect extraction which is a sub-task of Aspect-based Sentiment Analysis. The goal is to report an extraction method of financial aspects in microblog messages. Our approach uses a stock-investment taxonomy for the identification of explicit and implicit aspects. We compare supervised and unsupervised methods to assign predefined categories at message level. Results on 7 aspect classes show 0.71 accuracy, while the 32 class classification gives 0.82 accuracy for messages containing explicit aspects and 0.35 for implicit aspects.
|
['Gopal Sridhar', 'Thomas Gaillat', 'Manel Zarrouk', 'Bernardo Stearns', 'Ross McDermott', 'Brian Davis']
|
2018-07-01
| null | null | null |
ws-2018-7
|
['aspect-extraction']
|
['natural-language-processing']
|
[-3.35015029e-01 5.89223087e-01 -5.03702700e-01 -6.19626284e-01
-2.92149276e-01 -7.85714984e-01 1.14643764e+00 6.38804138e-01
-5.01883686e-01 6.18421793e-01 6.80297077e-01 -4.61831957e-01
2.50839710e-01 -1.21953118e+00 -9.45362379e-04 -1.81419417e-01
-1.11038581e-01 5.74529409e-01 2.41915107e-01 -4.58772689e-01
9.63997960e-01 1.61304787e-01 -1.41184938e+00 7.30579853e-01
-7.94468299e-02 1.11824548e+00 -6.16803408e-01 7.62196064e-01
-1.09631371e+00 1.38368356e+00 -1.29235041e+00 -8.68099868e-01
3.21913101e-02 -2.48369634e-01 -8.59067559e-01 4.49658602e-01
-4.47139628e-02 1.27607700e-03 4.48102623e-01 8.78919423e-01
-1.38085544e-01 -3.54434967e-01 1.05090725e+00 -1.34698272e+00
-3.57979655e-01 9.81009007e-01 -2.61149645e-01 4.72096682e-01
5.11578977e-01 -4.98793364e-01 1.41357064e+00 -7.88051248e-01
7.42359400e-01 6.87390506e-01 5.96032143e-01 -1.34697124e-01
-6.66535318e-01 -5.31441987e-01 2.30729312e-01 -1.70613676e-01
-8.66314590e-01 -1.32711772e-02 7.07563937e-01 -7.76177883e-01
1.57780445e+00 4.74923640e-01 1.22196686e+00 2.11229593e-01
5.71241200e-01 7.70393193e-01 1.46826828e+00 -1.80895194e-01
4.65352863e-01 1.03125250e+00 1.02202320e+00 4.28211093e-02
8.80988181e-01 -2.82296628e-01 -5.97231984e-01 -5.33609569e-01
2.24259213e-01 -1.31195605e-01 3.58857334e-01 4.57207173e-01
-6.39829516e-01 1.36159790e+00 -3.35778922e-01 7.46491730e-01
-5.81202388e-01 -3.87904525e-01 5.07037699e-01 7.83753991e-01
1.21231115e+00 7.63684511e-01 -9.62276638e-01 -3.87834013e-01
-9.63625312e-01 3.57739985e-01 1.77390182e+00 1.11379313e+00
9.27452087e-01 2.75570691e-01 1.18323430e-01 4.77663517e-01
4.23050940e-01 3.01852763e-01 5.82053602e-01 -3.14880788e-01
2.30455384e-01 1.46478152e+00 -2.15841271e-02 -1.15340614e+00
-3.83769959e-01 -6.87382281e-01 -1.24087362e-02 1.99368283e-01
-3.97037178e-01 -1.59205064e-01 -7.03517616e-01 5.66290379e-01
2.42225230e-01 -5.80674648e-01 4.19241279e-01 3.08044434e-01
1.30530596e+00 5.33448696e-01 2.73029432e-02 -3.07120889e-01
1.78098166e+00 -8.14016819e-01 -1.06158674e+00 -2.38314681e-02
4.80646312e-01 -1.24627674e+00 5.54258585e-01 4.90963280e-01
-8.21051955e-01 9.90711376e-02 -1.18943584e+00 3.82891089e-01
-1.08668172e+00 -2.74201512e-01 1.16637862e+00 1.02526343e+00
-8.05726588e-01 5.98804168e-02 -2.11988613e-01 6.30530640e-02
1.14106402e-01 2.56209493e-01 -1.66926771e-01 7.74228394e-01
-8.75899613e-01 5.61089218e-01 2.23249719e-01 -8.39786172e-01
-1.01810738e-01 -5.30656278e-01 -6.45708203e-01 -1.83879897e-01
9.33866948e-02 -3.16589355e-01 1.51570523e+00 -1.30592048e+00
-1.31324136e+00 1.02748179e+00 -1.06018364e-01 -5.61317384e-01
-2.03394383e-01 -3.95184904e-01 -7.36250162e-01 4.64375228e-01
3.90562356e-01 -3.59573923e-02 7.68997014e-01 -9.33880687e-01
-1.11906552e+00 -1.81589797e-01 4.67743099e-01 -7.23169520e-02
-4.73007739e-01 5.82497239e-01 9.04002860e-02 -1.13046575e+00
-6.53569475e-02 -4.16701347e-01 -7.98749179e-02 -9.04359877e-01
-3.37546319e-01 -3.53961319e-01 8.94452810e-01 -6.29700959e-01
1.43718827e+00 -1.57636130e+00 -7.08059847e-01 2.66030401e-01
3.34639788e-01 -2.45682105e-01 4.61287469e-01 6.77056611e-01
-2.96073053e-02 4.68249977e-01 -9.73803923e-02 -1.94004774e-01
4.34838012e-02 5.25521152e-02 -6.37998343e-01 -4.91170101e-02
1.98900342e-01 7.34960973e-01 -5.00360370e-01 -4.76804584e-01
-4.17372197e-01 1.86367989e-01 -5.60407579e-01 2.36346602e-01
-3.00416768e-01 -5.28533995e-01 -5.70833683e-01 9.49463785e-01
4.13921744e-01 -3.69757861e-02 -2.12753173e-02 -8.91013443e-02
-4.59399432e-01 9.73741710e-01 -9.58713055e-01 5.12430668e-01
-3.03654552e-01 9.46501613e-01 -1.88365668e-01 -8.31234276e-01
1.24346745e+00 5.53850949e-01 5.59296131e-01 -1.96553841e-01
3.94616038e-01 2.18052924e-01 -2.06958130e-01 -3.23032528e-01
1.04093421e+00 -5.37390351e-01 -4.33934003e-01 1.03776860e+00
1.72583044e-01 -5.13001323e-01 4.86021280e-01 6.71749771e-01
9.32475686e-01 -3.23363811e-01 9.08378363e-01 -6.73149168e-01
5.95373929e-01 4.81240392e-01 2.52203584e-01 1.99114010e-01
3.29523869e-02 4.47342843e-01 1.08056748e+00 -6.87771320e-01
-6.18688703e-01 -5.50927103e-01 -2.90203316e-04 7.57539511e-01
-3.26660991e-01 -1.05282891e+00 -5.31860292e-01 -1.03764153e+00
-5.69340214e-02 9.74060953e-01 -8.08261633e-01 4.94225889e-01
-1.39617026e-01 -9.83284235e-01 4.07289416e-02 3.36348414e-02
3.47388476e-01 -1.06593275e+00 -7.01880872e-01 3.29221010e-01
1.22586399e-01 -1.22680998e+00 -7.20258728e-02 1.42865941e-01
-7.48885214e-01 -1.27176774e+00 -2.44645640e-01 -4.18008029e-01
6.46360755e-01 -9.96210352e-02 1.68630683e+00 1.67749166e-01
4.99838367e-02 6.99631333e-01 -7.60957003e-01 -1.04643095e+00
-2.04530925e-01 2.17482731e-01 -3.42957675e-01 7.33310282e-02
1.19857299e+00 -5.21635592e-01 -2.64111400e-01 1.56666383e-01
-1.03029346e+00 -4.90549982e-01 5.98589659e-01 8.31641033e-02
3.81117582e-01 1.00272290e-01 3.94906670e-01 -1.58931088e+00
9.78356361e-01 -7.15909183e-01 -4.38916117e-01 -2.93875486e-01
-1.16585612e+00 -2.15152621e-01 -3.92499827e-02 -2.83114612e-01
-8.45069349e-01 -4.40993667e-01 -1.89329863e-01 8.24450016e-01
-9.97511297e-02 8.97001505e-01 2.48713627e-01 2.06180006e-01
2.86939323e-01 1.36587158e-01 -3.36285412e-01 -3.19670051e-01
-2.20278576e-01 8.54609489e-01 -3.71776819e-01 -6.56686127e-02
4.74866688e-01 4.82383758e-01 -3.14092100e-01 -1.24873364e+00
-1.23297954e+00 -9.29598033e-01 -4.80610669e-01 -3.63252252e-01
6.65091574e-01 -8.87541056e-01 -4.37928364e-02 4.40872788e-01
-9.01610970e-01 3.14510494e-01 -6.58191085e-01 4.23686206e-01
-2.40432218e-01 3.16786654e-02 -7.68901467e-01 -1.07126808e+00
-7.28913546e-01 -6.29158199e-01 7.46618152e-01 3.47150788e-02
-8.32778931e-01 -1.09232271e+00 4.56351787e-01 3.29530329e-01
6.92787230e-01 1.37826756e-01 4.17665660e-01 -1.52078700e+00
-5.15306532e-01 -7.11709738e-01 1.26992673e-01 1.54872477e-01
4.76453900e-01 9.52618793e-02 -8.89174342e-01 3.12566578e-01
5.19126832e-01 1.88517757e-02 6.80335224e-01 1.00839317e-01
-2.42223684e-03 -1.01589870e+00 -2.22960338e-02 7.94554055e-02
1.36716831e+00 2.87102729e-01 5.36338329e-01 1.11876774e+00
1.50176629e-01 9.56158578e-01 9.29674864e-01 5.54048181e-01
4.91782755e-01 2.23252207e-01 -6.82758093e-02 3.19791228e-01
3.05625290e-01 1.99534580e-01 4.83624786e-01 1.21619606e+00
1.38819009e-01 1.46892354e-01 -7.54131615e-01 6.03868306e-01
-1.26963806e+00 -8.99036765e-01 -5.41122198e-01 1.41462243e+00
8.47573161e-01 8.59187603e-01 4.68821108e-01 4.61573571e-01
2.25395918e-01 3.80452216e-01 1.95989460e-01 -8.38637710e-01
-1.22909844e-01 3.22927386e-01 2.39315107e-01 6.13847554e-01
-1.22971511e+00 6.65953755e-01 7.14337206e+00 3.08896452e-01
-8.43708694e-01 1.96279362e-01 6.53159857e-01 1.13675460e-01
-6.67168796e-01 1.98741838e-01 -1.31448722e+00 2.34971926e-01
1.16110682e+00 -2.84278631e-01 -7.21828699e-01 1.14347589e+00
-1.92654401e-01 -4.16943371e-01 -4.92781162e-01 2.38740772e-01
4.12372679e-01 -1.36672223e+00 3.47062588e-01 2.14515880e-01
7.00025380e-01 -1.03794418e-01 -2.32399553e-01 2.73077458e-01
-1.20395254e-02 -5.46144843e-01 7.54586101e-01 4.24955547e-01
6.52634427e-02 -1.04338205e+00 1.30123019e+00 6.16427921e-02
-1.25401437e+00 4.91319805e-01 -2.86792070e-01 -6.49213135e-01
2.70743847e-01 1.23530805e+00 -8.43485832e-01 2.10449889e-01
8.08731854e-01 8.19222569e-01 -4.22202975e-01 7.27465987e-01
-4.35849309e-01 7.98844278e-01 -1.71626821e-01 -6.38829827e-01
2.52655894e-01 -4.56207752e-01 7.52583623e-01 1.82109201e+00
-6.10437281e-02 2.67183840e-01 -7.37625584e-02 2.12093160e-01
3.97942901e-01 6.56180739e-01 -6.68878436e-01 -3.93257231e-01
-1.07085973e-01 1.44213355e+00 -1.33524454e+00 -9.28857327e-01
-6.24557793e-01 2.93320566e-01 -1.74292624e-01 7.61856586e-02
-1.15821987e-01 -7.00219870e-01 7.97828972e-01 3.54976892e-01
5.62584877e-01 -1.25717834e-01 -7.70848274e-01 -1.35895228e+00
-2.68518925e-03 -7.85016835e-01 6.44964099e-01 -2.48948455e-01
-1.13009882e+00 1.11492753e+00 1.63891777e-01 -1.28598201e+00
-5.96827567e-01 -7.48683751e-01 -9.57845151e-01 3.57817471e-01
-1.35048711e+00 -8.67403090e-01 8.16252530e-02 4.34631445e-02
6.70684040e-01 -8.74326825e-01 8.00344527e-01 4.44211662e-02
1.39005780e-01 7.35864788e-02 -7.53020585e-01 1.91971049e-01
2.58071035e-01 -1.42132676e+00 5.02821445e-01 5.63008010e-01
3.89544994e-01 6.22183740e-01 9.79038417e-01 -8.96856725e-01
-7.86172032e-01 -8.53351831e-01 1.91897166e+00 -8.67872477e-01
1.15545869e+00 -1.47847340e-01 -6.55893683e-01 7.15688944e-01
7.85022199e-01 -7.66933203e-01 1.41250896e+00 1.95581481e-01
-5.73939204e-01 1.19542718e-01 -1.13826001e+00 1.70729399e-01
1.89908698e-01 -5.63554704e-01 -1.23691046e+00 1.83082163e-01
8.49999070e-01 3.24393839e-01 -1.02218401e+00 2.17629969e-01
4.51434344e-01 -1.17539692e+00 7.01809883e-01 -4.92752105e-01
5.98927438e-01 -2.16771215e-01 -1.25803366e-01 -8.30774069e-01
2.90812999e-01 -3.66846383e-01 -2.15605944e-01 1.55299497e+00
1.12881184e+00 -1.02330530e+00 9.04325604e-01 5.69106996e-01
2.65625030e-01 -7.04904079e-01 -4.97565657e-01 -5.00267982e-01
-1.42693326e-01 -7.58338332e-01 5.41461468e-01 1.02445376e+00
4.27209228e-01 8.65288675e-01 2.03317553e-01 -3.42222363e-01
4.21850473e-01 5.66452324e-01 6.80074632e-01 -1.07316935e+00
-4.39822584e-01 -7.52433777e-01 -7.53830850e-01 -4.85248655e-01
-5.33411279e-02 -3.13564539e-01 -5.10742545e-01 -1.48383319e+00
2.35210955e-01 -1.06740557e-01 3.91710026e-04 7.03293309e-02
1.59305073e-02 5.11217356e-01 -9.44885090e-02 4.08099830e-01
-5.47743797e-01 2.28756413e-01 4.34867263e-01 -3.62410367e-01
-2.24525094e-01 7.06243753e-01 -1.04769707e+00 9.66984689e-01
1.35338843e+00 -8.72207046e-01 -2.41774261e-01 3.87886584e-01
9.25764024e-01 -6.51969314e-01 -3.49779755e-01 -5.05983233e-01
-2.99157519e-02 -3.83514538e-02 -7.95422569e-02 -1.28560042e+00
2.08077997e-01 -7.42280781e-01 -7.57801607e-02 6.35454834e-01
-1.79904774e-01 4.12293613e-01 -1.50362181e-03 2.47811258e-01
-1.02476776e+00 -7.16085911e-01 1.54908538e-01 -3.98287266e-01
-4.13954169e-01 -1.31593093e-01 -1.15254033e+00 2.19078749e-01
8.62805009e-01 -8.02814364e-02 -1.65286526e-01 -7.47679889e-01
-6.15930796e-01 -2.74196059e-01 2.12253049e-01 2.39659041e-01
6.55072033e-01 -9.68690097e-01 -6.63059473e-01 -5.45973107e-02
1.47329554e-01 -5.91964662e-01 -7.76000082e-01 6.68862224e-01
-6.41705751e-01 6.17621243e-01 1.42777845e-01 -4.09392342e-02
-1.41698587e+00 1.69489726e-01 -6.34358972e-02 -5.51679432e-01
-3.83114457e-01 6.87119186e-01 7.98950419e-02 -4.18236703e-01
-1.17880352e-01 -4.06403750e-01 -1.15922880e+00 8.46610427e-01
8.99411440e-01 2.09220394e-01 1.64781123e-01 -7.05789745e-01
-2.87394881e-01 5.69528878e-01 -1.74792558e-01 -6.08097911e-01
1.65381551e+00 2.36619040e-02 -6.68873668e-01 9.98147011e-01
1.26213682e+00 4.48165506e-01 -1.40347257e-01 -8.42915177e-02
8.39691639e-01 -2.67304003e-01 -1.16137788e-02 -5.35802603e-01
-1.04686630e+00 2.75181949e-01 -2.05713421e-01 1.12351501e+00
9.28406835e-01 2.65386432e-01 4.69101071e-01 2.86371559e-01
1.93837628e-01 -1.24669409e+00 -6.88080341e-02 8.02341819e-01
8.72876108e-01 -1.04863501e+00 7.65806913e-01 -7.37388432e-01
-8.81063581e-01 1.22461736e+00 2.80565590e-01 -4.36110973e-01
1.56328332e+00 6.88836277e-01 4.18893546e-01 -8.96077752e-01
-1.06386614e+00 -6.04281016e-02 4.47912395e-01 2.50989467e-01
8.79548490e-01 1.44416541e-01 -1.04209054e+00 1.06741452e+00
-9.13673162e-01 -4.15670156e-01 9.70550835e-01 1.50222778e+00
-7.11413503e-01 -1.11648309e+00 -1.52815625e-01 9.27663445e-01
-1.45503056e+00 -2.20842153e-01 -1.14034307e+00 1.00537181e+00
-3.56640011e-01 1.11603773e+00 3.70839506e-01 -3.78820956e-01
4.06924069e-01 8.52470770e-02 -3.33626240e-01 -9.37575817e-01
-1.42239773e+00 2.79335588e-01 8.96300197e-01 -2.65517592e-01
-1.10896122e+00 -8.86739314e-01 -1.01428747e+00 -5.45864478e-02
-1.31611183e-01 9.77860272e-01 8.88932765e-01 1.03043842e+00
9.60371941e-02 3.84353518e-01 7.15902567e-01 -2.71758109e-01
2.32514031e-02 -1.06735194e+00 -7.66504765e-01 3.55197228e-02
3.65618914e-01 -3.76993626e-01 -8.01395476e-01 3.07134897e-01]
|
[11.235633850097656, 6.833914756774902]
|
efec6229-ff79-468b-be8f-105894e5e8d3
|
high-accuracy-android-malware-detection-using
|
1608.00835
| null |
http://arxiv.org/abs/1608.00835v1
|
http://arxiv.org/pdf/1608.00835v1.pdf
|
High Accuracy Android Malware Detection Using Ensemble Learning
|
With over 50 billion downloads and more than 1.3 million apps in the Google
official market, Android has continued to gain popularity amongst smartphone
users worldwide. At the same time there has been a rise in malware targeting
the platform, with more recent strains employing highly sophisticated detection
avoidance techniques. As traditional signature based methods become less potent
in detecting unknown malware, alternatives are needed for timely zero-day
discovery. Thus this paper proposes an approach that utilizes ensemble learning
for Android malware detection. It combines advantages of static analysis with
the efficiency and performance of ensemble machine learning to improve Android
malware detection accuracy. The machine learning models are built using a large
repository of malware samples and benign apps from a leading antivirus vendor.
Experimental results and analysis presented shows that the proposed method
which uses a large feature space to leverage the power of ensemble learning is
capable of 97.3 to 99 percent detection accuracy with very low false positive
rates.
|
['Suleiman Y. Yerima', 'Sakir Sezer', 'Igor Muttik']
|
2016-08-02
| null | null | null | null |
['android-malware-detection']
|
['miscellaneous']
|
[ 2.02883050e-01 -2.91375041e-01 -5.20035684e-01 8.94866884e-02
-5.47501206e-01 -7.64758229e-01 7.40548730e-01 -9.62215383e-03
-4.51576896e-02 7.02456772e-01 -3.70568454e-01 -8.07951689e-01
3.15799676e-02 -4.70732003e-01 -2.31696665e-01 -3.84486258e-01
-3.68525803e-01 5.07203974e-02 4.04012889e-01 -1.87417269e-01
6.16316497e-01 2.04650357e-01 -1.83739769e+00 5.30849814e-01
9.22300398e-01 1.01252842e+00 -2.13185340e-01 9.78357434e-01
-2.45377608e-02 4.86631632e-01 -8.35529447e-01 -4.26017314e-01
4.38914567e-01 -2.08535250e-02 -1.71827629e-01 -3.62548709e-01
2.76567228e-02 -4.76202548e-01 1.35311559e-01 8.35730553e-01
3.87574703e-01 -4.65174407e-01 4.21115726e-01 -1.50762391e+00
-2.98999071e-01 1.48324147e-01 -1.11043024e+00 5.25979817e-01
5.81467152e-01 1.79734945e-01 6.11946464e-01 -3.80065918e-01
7.02037811e-02 7.14388013e-01 9.86999989e-01 4.07227486e-01
-9.31029022e-01 -9.80991364e-01 -4.19948339e-01 1.42731786e-01
-1.16443467e+00 -4.73495066e-01 6.39060438e-01 -6.57272041e-01
1.34879363e+00 4.62475717e-01 6.67138159e-01 1.17605829e+00
8.72527182e-01 1.99838728e-01 1.50643587e+00 -1.99857801e-01
1.49037302e-01 7.02042043e-01 2.69182414e-01 7.66229331e-01
7.19232917e-01 1.60148561e-01 -1.19724169e-01 -9.74828482e-01
-3.03935036e-02 3.88225108e-01 2.01526225e-01 -1.17295042e-01
-5.17760217e-01 9.40358996e-01 -4.82129067e-01 3.38054389e-01
-1.89804912e-01 -1.94783390e-01 9.62263763e-01 1.64044797e-01
6.46487057e-01 3.56912255e-01 -8.12183142e-01 -7.97063947e-01
-8.63045216e-01 -3.88543345e-02 9.23173726e-01 3.51279378e-01
3.93253684e-01 2.09774658e-01 6.05913103e-01 7.07738996e-01
5.30122101e-01 4.33432788e-01 9.48545039e-01 -6.08967066e-01
2.40025986e-02 1.06969118e+00 -1.45612136e-01 -1.21385324e+00
1.72076762e-01 -2.95905113e-01 -1.77738503e-01 3.30047786e-01
2.65474953e-02 -1.48222119e-01 -7.43732393e-01 1.22399533e+00
3.20009232e-01 5.42172730e-01 -2.35491335e-01 -3.47878247e-01
8.29057992e-02 5.47692060e-01 1.03503941e-02 -4.85608637e-01
1.22331524e+00 -3.44201922e-01 -3.48244160e-01 9.78883281e-02
6.87181294e-01 -6.58763468e-01 9.98855114e-01 7.96264648e-01
-4.50707763e-01 -2.68376410e-01 -1.24746776e+00 6.35442019e-01
-7.53200531e-01 -1.11843213e-01 6.93877518e-01 1.56927335e+00
-8.31417978e-01 1.85342431e-01 -8.80973279e-01 -4.40985620e-01
7.06287563e-01 8.95997107e-01 -1.82743594e-01 2.59472251e-01
-4.40833330e-01 5.72380841e-01 2.42426574e-01 -6.05378091e-01
-8.84250641e-01 -5.64216018e-01 -2.55763650e-01 -2.28889525e-01
3.88267905e-01 -9.20900106e-02 1.15975487e+00 -1.14811528e+00
-1.47178459e+00 5.55989504e-01 -2.63960809e-02 -5.29137552e-01
1.87464163e-01 -3.66036177e-01 -8.36625159e-01 -2.23557442e-01
1.28929034e-01 -1.25249282e-01 1.26335239e+00 -1.05765629e+00
-8.67136717e-01 -6.44516766e-01 -1.83212399e-01 -5.06417632e-01
-6.45107746e-01 1.48116633e-01 9.88348275e-02 -3.02785099e-01
-4.89047289e-01 -1.22377574e+00 2.67607421e-01 -1.00144160e+00
-1.84963956e-01 -2.60276288e-01 1.92844915e+00 -8.46338153e-01
1.46403790e+00 -1.97226775e+00 -4.25784826e-01 2.98599780e-01
2.27516428e-01 7.98748493e-01 2.47011364e-01 3.81289929e-01
-1.07123191e-02 5.43812752e-01 -2.15417936e-01 1.18489452e-02
-5.99552512e-01 1.10588625e-01 -3.28272372e-01 3.73553634e-01
-5.58502674e-02 4.92015064e-01 -7.88505197e-01 -2.36781299e-01
1.85182869e-01 6.99051797e-01 -4.61558580e-01 -5.27392365e-02
-6.14454560e-02 2.02468127e-01 -5.21557331e-01 1.05620468e+00
4.28102225e-01 -2.40451872e-01 4.05285031e-01 3.45057517e-01
-3.06823757e-02 1.45790726e-01 -6.07371569e-01 7.43797898e-01
-3.78491312e-01 5.47944784e-01 -1.51690781e-01 -6.53483272e-01
6.65540218e-01 1.00664318e-01 8.39589179e-01 -2.59270817e-01
3.47269475e-01 4.52090740e-01 5.21393597e-01 -5.46353698e-01
2.10951418e-01 5.50163209e-01 1.88018784e-01 5.61806738e-01
-3.11923563e-01 4.22937870e-01 -1.38029307e-01 2.21675053e-01
1.47333300e+00 6.41490445e-02 7.55727053e-01 -1.50870189e-01
9.91720915e-01 -1.67153068e-02 3.22769612e-01 4.46114242e-01
-5.20948708e-01 -4.96926010e-01 4.06675577e-01 -2.54340023e-01
-7.11969614e-01 -7.85310864e-01 8.52848217e-02 1.32185841e+00
-3.08245689e-01 -7.60793030e-01 -9.72511232e-01 -1.13188970e+00
1.25278473e-01 3.45564574e-01 -5.78609169e-01 -1.34488210e-01
-5.81868649e-01 -8.16311359e-01 6.85254812e-01 -1.73808075e-02
8.02755952e-01 -6.14357710e-01 -6.75480187e-01 5.68862148e-02
5.15325189e-01 -7.78298974e-01 7.41838440e-02 -1.26757279e-01
-1.02300322e+00 -1.42541337e+00 -1.48759767e-01 -3.35522562e-01
3.21616113e-01 4.66722429e-01 7.66916096e-01 3.29155266e-01
-5.06347895e-01 7.07989633e-01 -4.04703349e-01 -7.60550261e-01
-7.26214409e-01 2.31271595e-01 3.55497926e-01 6.03633299e-02
6.81408167e-01 -6.05780244e-01 -2.72649556e-01 1.87517256e-01
-4.75551426e-01 -7.55941927e-01 6.79988027e-01 4.81349856e-01
2.88179398e-01 4.53718543e-01 9.18130100e-01 -1.08083761e+00
9.25769985e-01 -1.22528446e+00 -4.34054852e-01 -4.02668044e-02
-1.04533505e+00 -4.81311858e-01 5.99354088e-01 -7.68907785e-01
-8.97332668e-01 1.52998447e-01 -1.33481294e-01 2.65182909e-02
3.35358121e-02 9.33930650e-02 2.14599352e-02 -4.89957422e-01
6.88845754e-01 2.35659987e-01 4.21606004e-01 -2.08978832e-01
-8.95326212e-03 1.21764064e+00 -3.55712213e-02 -4.86019515e-02
5.53525925e-01 3.67714792e-01 -2.13433832e-01 -1.29306781e+00
-1.50906578e-01 -5.85077107e-01 -1.74857408e-01 -3.11764002e-01
6.53073490e-01 -5.37352145e-01 -8.69669259e-01 5.72626472e-01
-6.32612705e-01 1.10960811e-01 5.60722530e-01 3.86430919e-02
-5.37331961e-03 5.08295000e-01 -2.81862438e-01 -1.22876549e+00
-4.42552239e-01 -1.18418992e+00 7.98842907e-01 1.42912418e-01
-3.75486195e-01 -8.68453920e-01 3.45608175e-01 4.40605283e-01
6.06547892e-01 3.16528618e-01 5.91481268e-01 -1.18437850e+00
-3.81157577e-01 -7.31832683e-01 -4.72382419e-02 4.69791114e-01
7.23652303e-01 5.60701549e-01 -1.06343615e+00 -4.11522716e-01
3.91850710e-01 1.81320235e-01 3.87033135e-01 1.52342021e-01
1.25187647e+00 -5.86445212e-01 -8.19679558e-01 1.22748904e-01
1.33320856e+00 8.40878785e-01 5.22494197e-01 3.10057223e-01
7.39753604e-01 3.91631901e-01 4.30637717e-01 3.15642953e-01
1.35727987e-01 4.15586889e-01 6.51549935e-01 6.73466504e-01
2.85289705e-01 -2.12431978e-02 7.32891083e-01 7.55319476e-01
-3.34772736e-01 8.29255581e-02 -1.07095385e+00 1.19841635e-01
-1.37518620e+00 -1.14084935e+00 -4.04948890e-02 2.38748789e+00
3.91405463e-01 4.47486669e-01 7.27774262e-01 4.40000176e-01
4.95702118e-01 -7.71035030e-02 -6.37393236e-01 -8.35222602e-01
4.65732932e-01 2.88830847e-01 6.50912344e-01 2.74844080e-01
-1.05613935e+00 4.51213986e-01 6.52005672e+00 7.42986381e-01
-1.35431111e+00 5.55660963e-01 7.54259944e-01 9.08807218e-02
1.83641613e-01 -1.83989182e-01 -8.13815415e-01 1.07065141e+00
1.46661210e+00 -1.46703839e-01 2.97526658e-01 1.34170103e+00
8.14687833e-02 -3.70813161e-01 -4.64575052e-01 1.19830537e+00
2.55824119e-01 -1.24209917e+00 -2.31491387e-01 8.80012333e-01
8.11560631e-01 1.99448258e-01 3.76032174e-01 4.74165231e-01
1.09238029e-02 -8.73554766e-01 -1.90741658e-01 1.51002660e-01
5.66013873e-01 -9.44209695e-01 7.19368696e-01 5.85668862e-01
-9.48560357e-01 -7.57550359e-01 2.55960733e-01 -3.34928751e-01
-3.18854958e-01 3.84566009e-01 -1.38653719e+00 4.15741131e-02
7.64688015e-01 6.67933643e-01 -9.60328817e-01 6.99178159e-01
4.08337027e-01 1.00497520e+00 -2.77713120e-01 -3.03537160e-01
-1.20260581e-01 -2.11365242e-02 5.75856507e-01 1.05459547e+00
4.36083198e-01 -2.30681121e-01 -8.33334681e-03 7.64525458e-02
1.98127761e-01 1.51248917e-01 -1.21174920e+00 -4.63399798e-01
4.70388025e-01 1.48014939e+00 -7.69863844e-01 -3.85054618e-01
-4.20652449e-01 7.23030806e-01 -1.12580255e-01 -2.28687495e-01
-9.82497931e-01 -1.84275061e-01 8.35202634e-01 4.09158438e-01
6.35815635e-02 -4.21047866e-01 -3.02928835e-01 -6.02671742e-01
-2.26646096e-01 -1.42013168e+00 1.49217382e-01 5.60795423e-03
-1.11524677e+00 5.17604828e-01 -5.29646017e-02 -1.09083855e+00
-6.02093995e-01 -8.84147108e-01 -8.96699309e-01 3.64934832e-01
-6.03022933e-01 -1.00914466e+00 -1.24568632e-03 3.31381857e-01
6.24146938e-01 -7.79321730e-01 7.40349591e-01 1.73381880e-01
-6.51570201e-01 4.90666956e-01 2.28073597e-01 -3.57061505e-01
3.41046274e-01 -9.30291176e-01 1.16600461e-01 6.45683467e-01
-1.82923570e-01 1.09081900e+00 5.41273415e-01 -1.26466572e+00
-1.57874966e+00 -8.19015682e-01 3.21695507e-01 -8.92670929e-01
8.24000955e-01 -3.57291460e-01 -7.96655059e-01 5.60164630e-01
3.59374344e-01 -3.98508370e-01 1.09854639e+00 -6.60676807e-02
-5.58866382e-01 -2.34446973e-01 -1.58054876e+00 2.83276528e-01
4.77521271e-01 -6.13774896e-01 -2.82977652e-02 2.64922827e-01
2.46483386e-01 2.13237852e-01 -6.81051314e-01 5.45454144e-01
8.88939500e-01 -1.38272476e+00 6.66411757e-01 -4.64102954e-01
-1.35879964e-01 -1.84882760e-01 -1.57309189e-01 -5.82441270e-01
1.90828130e-01 -8.49117100e-01 -8.17977011e-01 1.14265370e+00
4.62250471e-01 -9.93633568e-01 9.76100504e-01 2.26354048e-01
2.44910583e-01 -1.26997662e+00 -8.28098297e-01 -7.67648578e-01
-3.26050043e-01 -4.70302969e-01 3.37318689e-01 8.54308426e-01
-1.50936857e-01 2.87225693e-01 -4.53782082e-01 -2.13361487e-01
6.11463189e-01 -6.77061260e-01 9.21263576e-01 -1.49344158e+00
-6.55499339e-01 -3.21800053e-01 -6.37439847e-01 -1.86373159e-01
3.45548838e-02 -4.23068523e-01 -6.54879212e-01 -5.98378778e-01
4.00371909e-01 -4.07862335e-01 -1.48641706e-01 3.50566179e-01
9.23279971e-02 6.15283489e-01 -3.09210509e-01 3.50976735e-01
-3.38238358e-01 -3.22547227e-01 3.24807584e-01 5.63513339e-02
-6.34343863e-01 3.66588235e-01 -8.12938392e-01 9.34423923e-01
1.41130853e+00 -2.01158166e-01 -7.98795879e-01 2.90597886e-01
2.63442963e-01 -6.66715801e-01 1.63630724e-01 -1.07123601e+00
-3.56343716e-01 1.17400348e-01 3.97137254e-01 -3.69132251e-01
3.36323977e-01 -7.37968504e-01 4.03822511e-01 8.42720747e-01
3.29874277e-01 4.52509016e-01 3.65884155e-01 9.07298088e-01
3.52552563e-01 1.22643359e-01 7.46781170e-01 -8.68003219e-02
-7.92030871e-01 4.98478813e-03 -5.66851497e-01 -5.40544868e-01
1.82944429e+00 -7.82377422e-01 -3.31684470e-01 -7.65476972e-02
-3.34804446e-01 -5.74082732e-01 5.74654937e-01 7.70882666e-01
4.56859320e-01 -6.76352262e-01 -2.13378325e-01 3.29756618e-01
-4.29925807e-02 -1.13293672e+00 -7.55903423e-02 8.90643835e-01
-3.45260441e-01 3.66737157e-01 -3.49768639e-01 -7.04030335e-01
-1.99679911e+00 5.96839786e-01 2.90433224e-02 -2.28121504e-01
-8.94204229e-02 5.33655047e-01 -5.87305129e-01 -8.63855779e-02
-8.60917717e-02 4.29567486e-01 -3.65981221e-01 -2.08132654e-01
7.87414134e-01 9.41719234e-01 9.04541239e-02 -9.33958948e-01
-6.96636498e-01 1.93881318e-01 -5.84487796e-01 2.10937589e-01
9.18259919e-01 1.42099813e-01 -3.89519781e-01 4.00895596e-01
1.31925809e+00 7.90759921e-01 -6.22687638e-01 8.11949492e-01
4.65495139e-01 -6.85356617e-01 -2.61398315e-01 -1.00683212e+00
-6.65249169e-01 4.28046107e-01 1.23528838e+00 9.14726734e-01
1.08398795e+00 -2.38203824e-01 7.35262513e-01 1.83410451e-01
6.46350980e-01 -7.83246338e-01 2.31385946e-01 2.93254524e-01
2.01049134e-01 -1.43753219e+00 1.84774697e-01 -4.08509344e-01
-2.55151957e-01 7.28344679e-01 6.91161454e-01 -2.50015289e-01
8.01353753e-01 4.32455003e-01 -2.36672431e-01 -1.93764418e-02
-5.09332180e-01 4.62146282e-01 7.65441954e-02 1.04216409e+00
5.78141332e-01 2.69569755e-01 -5.50176382e-01 3.22134793e-01
8.69409293e-02 -2.11870834e-01 4.47243333e-01 1.08771133e+00
-6.58540010e-01 -1.61429083e+00 -3.09032202e-01 1.20949674e+00
-1.20337796e+00 1.16058793e-02 -9.60133851e-01 9.44788694e-01
4.63344872e-01 1.23496079e+00 -2.86680371e-01 -1.04800379e+00
-2.59905994e-01 1.27725959e-01 1.92183569e-01 -6.73747003e-01
-7.84987569e-01 -2.68789768e-01 1.33112803e-01 -6.27340078e-01
-2.31078729e-01 -6.07083380e-01 -8.98064554e-01 -4.99911398e-01
-3.98101479e-01 5.21376133e-02 1.07067752e+00 7.38415182e-01
9.98645723e-01 1.04613423e-01 7.31172740e-01 -6.33021712e-01
-4.75546569e-01 -9.19090152e-01 -1.91499203e-01 -2.65753776e-01
2.75783509e-01 -8.58238041e-01 -4.31429327e-01 -5.29013090e-02]
|
[14.42809009552002, 9.683245658874512]
|
99bd8e84-6a61-4ab2-9c45-a748f845bd4f
|
deepotsu-document-enhancement-and
|
1901.06081
| null |
http://arxiv.org/abs/1901.06081v1
|
http://arxiv.org/pdf/1901.06081v1.pdf
|
DeepOtsu: Document Enhancement and Binarization using Iterative Deep Learning
|
This paper presents a novel iterative deep learning framework and apply it
for document enhancement and binarization. Unlike the traditional methods which
predict the binary label of each pixel on the input image, we train the neural
network to learn the degradations in document images and produce the uniform
images of the degraded input images, which allows the network to refine the
output iteratively. Two different iterative methods have been studied in this
paper: recurrent refinement (RR) which uses the same trained neural network in
each iteration for document enhancement and stacked refinement (SR) which uses
a stack of different neural networks for iterative output refinement. Given the
learned uniform and enhanced image, the binarization map can be easy to obtain
by a global or local threshold. The experimental results on several public
benchmark data sets show that our proposed methods provide a new clean version
of the degraded image which is suitable for visualization and promising results
of binarization using the global Otsu's threshold based on the enhanced images
learned iteratively by the neural network.
|
['Lambert Schomaker', 'Sheng He']
|
2019-01-18
| null | null | null | null |
['document-enhancement']
|
['computer-vision']
|
[ 8.41652036e-01 -1.83095708e-01 1.32482961e-01 -4.47302133e-01
-3.75868648e-01 -2.77058154e-01 4.13388431e-01 -1.49096102e-01
-3.41158897e-01 6.42872214e-01 3.08029771e-01 -9.59713310e-02
-2.23416254e-01 -8.04351032e-01 -5.87540329e-01 -1.17286682e+00
1.07335329e-01 -1.16925575e-02 2.17085615e-01 -2.25516021e-01
5.99196076e-01 5.65055490e-01 -1.59190619e+00 8.58077407e-01
9.91036177e-01 1.16598392e+00 4.44651961e-01 1.11974227e+00
-3.82737443e-03 7.31239676e-01 -1.01324606e+00 -1.05122991e-01
4.02487904e-01 -4.78695154e-01 -6.26748741e-01 5.08127987e-01
4.72925454e-01 -4.85134900e-01 -4.56939459e-01 1.41578996e+00
5.94383895e-01 2.07145080e-01 5.22139311e-01 -8.43351901e-01
-1.05881500e+00 7.44181037e-01 -6.60980821e-01 2.71426499e-01
-1.16895316e-02 -1.37890652e-01 5.08658171e-01 -7.65066981e-01
5.11199951e-01 1.13203180e+00 6.66623533e-01 3.59616876e-01
-1.02532446e+00 -5.55170417e-01 1.08983546e-01 6.93771064e-01
-1.16746938e+00 -2.98859149e-01 7.08152890e-01 -2.32689157e-02
5.36274493e-01 4.29492623e-01 5.00147104e-01 6.60437584e-01
3.49769861e-01 8.43770027e-01 1.44044673e+00 -6.63051903e-01
-6.12851279e-03 5.74436970e-04 4.28701460e-01 6.42163038e-01
8.07422027e-03 6.98609874e-02 -3.31284136e-01 4.46144223e-01
7.92518139e-01 3.37326020e-01 -6.95555866e-01 3.35551016e-02
-1.02202356e+00 3.82079631e-01 7.95283556e-01 4.01966900e-01
-5.92411458e-01 -2.61797637e-01 1.77494258e-01 2.58667797e-01
3.90648037e-01 2.52111971e-01 -1.64355323e-01 4.58626717e-01
-1.49671364e+00 2.65608821e-02 2.48546407e-01 4.50849593e-01
7.46234179e-01 9.50464457e-02 -5.53702474e-01 8.77104938e-01
1.47856027e-01 4.34005409e-01 6.16783798e-01 -9.14776921e-01
4.11661744e-01 5.36361456e-01 9.86230597e-02 -1.09880614e+00
-1.47908241e-01 -5.44878125e-01 -1.49235713e+00 9.39406455e-01
2.05336630e-01 2.85464879e-02 -1.60211062e+00 1.05699003e+00
-9.66108143e-02 1.51306942e-01 2.11358413e-01 9.08475816e-01
7.39150167e-01 1.13505137e+00 -4.30804074e-01 -3.38596702e-01
9.53125000e-01 -1.45426846e+00 -1.23510373e+00 -8.39717612e-02
-6.86177984e-02 -8.45226467e-01 9.14654255e-01 9.38344002e-01
-1.29942966e+00 -1.00580132e+00 -1.56451046e+00 5.39628118e-02
-6.06616437e-01 3.66383791e-01 9.48948488e-02 6.34596646e-01
-1.41342509e+00 9.23602343e-01 -5.93217671e-01 -7.08419159e-02
5.51131010e-01 2.37072423e-01 -2.18186513e-01 -3.02724093e-01
-1.02146637e+00 8.34001660e-01 6.38716221e-01 6.51541650e-01
-1.00139034e+00 -2.06261322e-01 -5.77484488e-01 4.45598394e-01
-4.57338914e-02 -2.13608012e-01 9.53411937e-01 -1.19528568e+00
-1.52848935e+00 4.76109564e-01 -1.81274340e-01 -3.80840003e-01
4.67485011e-01 -2.56743461e-01 -5.06175041e-01 1.82902619e-01
-2.35070258e-01 4.94780093e-01 1.20515764e+00 -1.60560179e+00
-9.42918003e-01 -3.74017149e-01 -1.91542596e-01 3.63552451e-01
-5.38522720e-01 5.63198105e-02 -6.92689836e-01 -7.40552604e-01
4.02766734e-01 -2.96937972e-01 -1.78595275e-01 -1.21233962e-01
-5.19436121e-01 2.26584971e-01 1.18787110e+00 -1.32719743e+00
1.34254050e+00 -2.12423778e+00 3.89434934e-01 4.09626067e-01
1.07455090e-01 3.47393066e-01 -2.56695926e-01 -2.28007138e-01
-2.34177202e-01 5.35529889e-02 -3.53738278e-01 -4.59696829e-01
-2.88918227e-01 -3.73795480e-02 -2.10302085e-01 4.54823345e-01
-4.82445909e-03 5.38626671e-01 -6.03662550e-01 -5.14290035e-01
2.09908009e-01 6.92948163e-01 -1.87488601e-01 4.92883056e-01
1.14272311e-01 2.17560604e-01 2.91761398e-01 5.29959142e-01
1.17210197e+00 -5.78836948e-02 6.04136707e-03 -7.39974618e-01
-1.98949143e-01 -1.55568779e-01 -1.42156613e+00 1.27209175e+00
-9.77026820e-02 9.55718815e-01 1.38682455e-01 -9.00650084e-01
1.05612648e+00 1.95402041e-01 -1.10888958e-01 -8.98148239e-01
1.96711302e-01 -1.78428188e-01 -3.45038205e-01 -4.84960526e-01
8.01188171e-01 2.25593895e-01 4.07144696e-01 6.01954758e-01
1.38391331e-02 5.60357012e-02 3.69911522e-01 3.77005823e-02
7.87718415e-01 1.84061870e-01 -1.87342063e-01 -5.98753861e-04
7.48535216e-01 -1.81927204e-01 4.34874207e-01 1.05774796e+00
-2.29475778e-02 9.73125458e-01 8.30839574e-02 -5.08441746e-01
-1.36046052e+00 -9.87910509e-01 -2.72797216e-02 1.07550490e+00
4.85438257e-01 -3.07235103e-02 -9.96372640e-01 -6.72802687e-01
-6.36425138e-01 5.06852686e-01 -7.54851997e-01 -1.43172607e-01
-6.36172473e-01 -9.71206069e-01 1.79042965e-01 4.59464014e-01
1.07126808e+00 -1.32346117e+00 -1.84902206e-01 -9.74247232e-02
-3.02996457e-01 -5.40306568e-01 -4.08117950e-01 5.42374730e-01
-8.37930918e-01 -7.83721328e-01 -1.03210568e+00 -1.12795687e+00
1.03781056e+00 3.70716572e-01 5.89637518e-01 2.70783693e-01
-6.41732961e-02 -2.50231981e-01 -3.73013556e-01 -1.39638945e-01
-5.32513380e-01 -2.02986881e-01 -1.82428762e-01 2.95731694e-01
-1.77012891e-01 -2.29047626e-01 -6.82561517e-01 4.58333977e-02
-1.23986936e+00 3.17038983e-01 6.97294712e-01 9.77131486e-01
6.96778536e-01 8.38345528e-01 1.45536184e-01 -6.76015735e-01
9.37841892e-01 8.36212188e-02 -4.38647091e-01 6.10621989e-01
-9.01011825e-01 2.70582855e-01 4.74385232e-01 -3.60868096e-01
-1.58448267e+00 -3.05840075e-01 -1.61601752e-02 -1.68188408e-01
-2.21249446e-01 4.78980720e-01 -2.22330987e-01 -1.02808392e-02
6.47123277e-01 5.36248446e-01 -1.22219615e-01 -7.05120027e-01
3.64197999e-01 1.03286958e+00 1.25597465e+00 6.30700141e-02
7.01116145e-01 3.66110086e-01 -4.59444493e-01 -2.94802487e-01
-8.17967117e-01 -1.20925024e-01 -7.77432442e-01 -4.47601020e-01
1.04091036e+00 -5.71093976e-01 -3.24718922e-01 1.05156481e+00
-1.24612296e+00 -3.86335075e-01 -8.45197141e-02 8.24884176e-02
-2.24783301e-01 5.86919785e-01 -8.68413746e-01 -7.21190572e-01
-7.77251124e-01 -1.30647457e+00 8.23981762e-01 7.88407981e-01
2.46288002e-01 -7.81594217e-01 -1.21007301e-01 2.18648016e-01
5.56951523e-01 -3.13056819e-02 1.00660193e+00 -1.07170895e-01
-4.18886542e-01 -1.59720704e-01 -5.93990564e-01 8.84059846e-01
3.83172631e-01 4.71571460e-02 -9.60990667e-01 -4.08162594e-01
1.15996204e-01 2.37585139e-02 1.18018234e+00 6.17320240e-01
1.56002200e+00 -2.68284649e-01 -8.74332041e-02 8.94294977e-01
1.48992515e+00 4.02662635e-01 1.23074877e+00 9.12055910e-01
5.65936029e-01 3.28791477e-02 6.16232514e-01 2.10117668e-01
-1.59837186e-01 2.60630071e-01 5.48599243e-01 -6.51055455e-01
-5.24047077e-01 1.90584242e-01 4.45205569e-01 6.52136683e-01
-3.30934674e-01 -4.57962960e-01 -5.55413604e-01 3.81339639e-01
-1.71797264e+00 -1.08887517e+00 6.77619949e-02 2.01000619e+00
1.00514543e+00 3.39534998e-01 -4.35058147e-01 5.53803682e-01
1.27541792e+00 2.37427816e-01 -3.43476862e-01 -7.38111556e-01
-4.96236622e-01 3.67208451e-01 4.87888396e-01 5.57362437e-01
-1.24031079e+00 7.13203371e-01 6.74980545e+00 7.43866861e-01
-1.32064056e+00 -3.86025682e-02 1.00510788e+00 1.77412048e-01
4.60396111e-02 -4.89077032e-01 -7.58087814e-01 4.12842214e-01
6.90820873e-01 9.56487879e-02 7.08287179e-01 5.22884488e-01
3.20820332e-01 -1.19540267e-01 -6.29610479e-01 9.09467638e-01
3.88516545e-01 -1.43022799e+00 2.35626146e-01 -3.05076718e-01
1.29398596e+00 -3.38955551e-01 4.40835685e-01 1.14019834e-01
2.94921696e-01 -1.05588233e+00 5.47292352e-01 1.04078531e+00
7.65055180e-01 -8.01177680e-01 1.02157259e+00 1.13739960e-01
-7.23887444e-01 -4.02309537e-01 -5.34255207e-01 2.25919694e-01
1.82890862e-01 5.00681162e-01 -6.84535980e-01 4.54409003e-01
1.12184560e+00 6.51714087e-01 -8.76732528e-01 1.41742110e+00
-3.01064223e-01 5.24948299e-01 5.95589876e-02 3.95497561e-01
2.02304497e-01 -3.21009159e-01 3.43801975e-01 1.36361527e+00
5.50051928e-01 -2.98425049e-01 -3.20563942e-01 7.57428050e-01
-1.54421970e-01 -1.73274860e-01 4.93100733e-02 1.04170516e-01
-1.06093716e-02 1.41197157e+00 -9.38755035e-01 -6.85036361e-01
-4.31955904e-02 1.51378953e+00 4.90804985e-02 6.47081494e-01
-7.00947165e-01 -8.73572946e-01 -2.94471476e-02 -2.39950120e-01
7.24793673e-01 1.45195156e-01 -6.52808964e-01 -8.39518309e-01
-1.61848322e-01 -1.07146490e+00 3.49918425e-01 -1.28253889e+00
-8.50366771e-01 1.00461709e+00 -3.07087660e-01 -1.07210112e+00
1.83335729e-02 -7.49846101e-01 -7.65671849e-01 1.10044932e+00
-1.49859631e+00 -8.51049602e-01 -8.51520598e-01 5.90215147e-01
7.22106874e-01 -2.85051405e-01 5.95603704e-01 1.64215162e-01
-7.65222073e-01 3.92586321e-01 6.52294278e-01 -9.07581374e-02
8.44764531e-01 -1.46562207e+00 8.83178413e-02 1.33354783e+00
-1.45812342e-02 2.98367083e-01 8.03288877e-01 -7.03166842e-01
-6.15710974e-01 -1.11023855e+00 5.42895734e-01 6.30218908e-02
1.35571316e-01 1.87868662e-02 -1.08014286e+00 5.21648645e-01
7.68008113e-01 -2.47630745e-01 1.32116899e-01 -3.09787810e-01
-7.35442415e-02 -4.20908689e-01 -1.25756156e+00 6.33116663e-01
5.17668366e-01 -1.34928897e-01 -6.09267890e-01 2.91296065e-01
6.53743863e-01 -4.54628795e-01 -7.18387902e-01 4.97388363e-01
4.53474462e-01 -1.04746199e+00 8.54019761e-01 -3.99654776e-01
6.69716954e-01 -5.90012968e-01 1.60086097e-03 -1.40609241e+00
-7.36766994e-01 -9.71631259e-02 -2.20591336e-01 1.26140404e+00
4.53870147e-01 -7.36786574e-02 6.90118074e-01 1.50703281e-01
-1.73112556e-01 -7.08208084e-01 -3.15161467e-01 -1.63447872e-01
-3.69136930e-01 1.59983896e-02 7.66963601e-01 5.61004877e-01
-4.56062496e-01 -6.28175661e-02 -6.08949184e-01 6.12554312e-01
7.41162658e-01 1.96897149e-01 3.79760742e-01 -9.10422087e-01
-2.71276325e-01 -5.46225727e-01 -2.70350695e-01 -1.09683776e+00
-1.93328664e-01 -5.64774513e-01 3.65033746e-01 -1.82801199e+00
5.00537097e-01 -2.44359821e-02 -7.72627473e-01 4.22095537e-01
-4.81042504e-01 7.64781713e-01 1.64612569e-02 9.44694430e-02
-7.53236890e-01 3.10749918e-01 1.31839478e+00 -7.50791311e-01
-2.68081367e-01 3.48757431e-02 -7.31001794e-01 5.36313474e-01
7.32629061e-01 -2.34077185e-01 -9.21283141e-02 -6.02612019e-01
-2.72358563e-02 -3.99574399e-01 2.05443293e-01 -1.13063872e+00
3.29884440e-01 7.69594759e-02 1.21143794e+00 -9.33166802e-01
9.72413048e-02 -8.19455206e-01 -1.69080585e-01 4.70869809e-01
-5.77413261e-01 -1.19506463e-01 1.47976041e-01 3.34680676e-01
-3.36752295e-01 -5.42698801e-01 1.03940666e+00 -3.38680930e-02
-8.05993438e-01 8.67711455e-02 -2.99924135e-01 -8.11076760e-01
6.25695884e-01 -6.29791081e-01 -5.05401313e-01 -5.64857543e-01
-1.09844708e+00 -7.87774324e-02 2.94067919e-01 2.53495693e-01
9.41772521e-01 -1.31160879e+00 -7.67202020e-01 4.25584406e-01
-4.54013705e-01 -1.04081377e-01 3.80391359e-01 5.16955435e-01
-7.30907559e-01 2.63585895e-02 -6.32445931e-01 -5.67836344e-01
-1.64319587e+00 6.57589436e-01 4.81078595e-01 -4.15842593e-01
-6.43293023e-01 8.87312412e-01 6.50475025e-02 -1.20337740e-01
5.95563948e-01 -1.22904666e-01 -6.55502558e-01 -3.00621688e-01
9.97037947e-01 4.53906685e-01 2.10144579e-01 -5.37502646e-01
2.44336203e-01 3.69265974e-01 -5.08140385e-01 -8.85736942e-02
1.66279066e+00 -4.43455189e-01 -4.65520382e-01 1.38874933e-01
1.08601975e+00 -2.27550656e-01 -1.44187641e+00 -3.61283720e-01
-3.05043906e-01 -5.42924404e-01 5.83315969e-01 -1.13032460e+00
-1.41128874e+00 7.89501131e-01 1.34360087e+00 3.23826343e-01
1.78479767e+00 -5.03247321e-01 6.05857134e-01 2.84372926e-01
-2.55822927e-01 -1.27327096e+00 2.23531619e-01 3.25220019e-01
1.09533095e+00 -9.06689823e-01 8.74220878e-02 6.24880157e-02
-3.89310360e-01 1.38503087e+00 6.06058538e-01 -1.14807136e-01
2.64397115e-01 3.92480582e-01 3.74791354e-01 -2.18900805e-03
-1.93599790e-01 7.83128217e-02 2.53527790e-01 6.37840211e-01
1.96970135e-01 -2.95417219e-01 -1.43636450e-01 4.75256026e-01
-1.18234359e-01 -3.31885926e-02 5.24058223e-01 8.48917186e-01
-7.84969807e-01 -9.47244883e-01 -1.01438236e+00 4.67074811e-01
-3.94251972e-01 -1.90752208e-01 -1.81978196e-01 2.64458925e-01
4.20302629e-01 8.84558082e-01 1.47478998e-01 -4.46523339e-01
1.03986874e-01 6.51853755e-02 5.37913740e-01 -2.17386961e-01
-4.01710004e-01 8.74630585e-02 -3.68365198e-01 -3.31649214e-01
-2.91359097e-01 -4.40734863e-01 -1.01443303e+00 -9.62200388e-02
-4.98824745e-01 4.54361700e-02 7.56394804e-01 9.41792250e-01
6.56263232e-02 9.34437454e-01 7.37232268e-01 -1.28400898e+00
-3.07382315e-01 -1.20343280e+00 -7.28016496e-01 5.53669691e-01
6.60848260e-01 -2.81708688e-01 -3.49684745e-01 7.01225758e-01]
|
[11.35031509399414, -2.080423593521118]
|
49ac2b18-8a4c-45ae-92d1-272a7f7ab221
|
multi-staged-cross-lingual-acoustic-model
| null | null |
https://aclanthology.org/2020.lrec-1.780
|
https://aclanthology.org/2020.lrec-1.780.pdf
|
Multi-Staged Cross-Lingual Acoustic Model Adaption for Robust Speech Recognition in Real-World Applications - A Case Study on German Oral History Interviews
|
While recent automatic speech recognition systems achieve remarkable performance when large amounts of adequate, high quality annotated speech data is used for training, the same systems often only achieve an unsatisfactory result for tasks in domains that greatly deviate from the conditions represented by the training data. For many real-world applications, there is a lack of sufficient data that can be directly used for training robust speech recognition systems. To address this issue, we propose and investigate an approach that performs a robust acoustic model adaption to a target domain in a cross-lingual, multi-staged manner. Our approach enables the exploitation of large-scale training data from other domains in both the same and other languages. We evaluate our approach using the challenging task of German oral history interviews, where we achieve a relative reduction of the word error rate by more than 30{\%} compared to a model trained from scratch only on the target domain, and 6-7{\%} relative compared to a model trained robustly on 1000 hours of same-language out-of-domain training data.
|
['Joachim K{\\"o}hler', 'Oliver Walter', 'Sven Behnke', 'Christoph Schmidt', 'Michael Gref']
|
2020-05-01
| null | null | null |
lrec-2020-5
|
['robust-speech-recognition']
|
['speech']
|
[ 4.49449986e-01 2.32177272e-01 2.26331830e-01 -7.26383448e-01
-1.57080495e+00 -5.22483468e-01 6.26763761e-01 1.64124608e-01
-7.54800439e-01 6.31294191e-01 1.87866241e-01 -2.78290153e-01
2.59183913e-01 -1.90951154e-01 -5.65632761e-01 -5.63673973e-01
2.78459817e-01 8.79266322e-01 4.34788644e-01 -3.25289845e-01
-2.01061994e-01 2.53217906e-01 -1.44238436e+00 2.42274329e-01
6.31566823e-01 6.97883487e-01 5.43104947e-01 5.70611894e-01
-2.13611424e-01 2.71443367e-01 -9.20189679e-01 -3.89049053e-01
5.61180301e-02 -3.96639824e-01 -8.17874670e-01 4.28061962e-01
2.97148556e-01 1.33855566e-02 -1.31725967e-01 9.96195436e-01
6.76958144e-01 2.64381558e-01 3.99141461e-01 -5.00715971e-01
-2.44728118e-01 5.95413208e-01 1.64961070e-01 1.13918550e-01
4.41513121e-01 -2.54910346e-02 6.01236105e-01 -7.10004032e-01
6.57403588e-01 1.04758322e+00 4.86727178e-01 9.01983440e-01
-1.30561900e+00 -4.28740144e-01 1.76751554e-01 -1.48459435e-01
-1.31136930e+00 -1.13304949e+00 7.45970964e-01 -5.70946150e-02
1.12132609e+00 1.07982144e-01 2.38956451e-01 1.45179594e+00
-5.56113422e-01 7.33003795e-01 9.69889283e-01 -7.81610489e-01
4.65742588e-01 4.61971462e-01 -1.70519486e-01 2.52345920e-01
-2.76561230e-01 -4.76854816e-02 -5.86718261e-01 -9.28483158e-02
2.55969852e-01 -6.54075742e-01 -3.19510639e-01 4.75643063e-03
-1.04473197e+00 6.61537826e-01 -2.58442491e-01 8.61237705e-01
-4.31612760e-01 -4.46935862e-01 5.16337276e-01 4.36754674e-01
5.77786088e-01 3.45313966e-01 -7.30361640e-01 -6.09617412e-01
-1.15552771e+00 -4.56065089e-02 1.05034661e+00 9.40027416e-01
4.78906244e-01 4.05151516e-01 4.34516132e-01 1.52478147e+00
2.26212934e-01 6.96483731e-01 8.92684281e-01 -6.14283144e-01
6.10789239e-01 2.81562030e-01 -4.53037806e-02 -2.93917149e-01
-1.06059588e-01 -4.01687771e-01 -4.87212718e-01 -1.39530212e-01
6.65888965e-01 -1.49147928e-01 -1.16645741e+00 1.88779211e+00
4.38740879e-01 -9.18232650e-02 5.56779265e-01 4.35115784e-01
5.24318099e-01 8.84440780e-01 -5.53327017e-02 -5.03167689e-01
1.18869543e+00 -7.92776465e-01 -7.39123344e-01 -7.12981582e-01
6.99482441e-01 -9.79813337e-01 1.25483167e+00 6.87609017e-01
-1.13269377e+00 -6.33313358e-01 -1.05205667e+00 1.51428610e-01
-2.87654638e-01 4.31438014e-02 -6.71736076e-02 9.12034929e-01
-1.10124326e+00 3.44850123e-01 -6.63757622e-01 -6.06845379e-01
-1.21850662e-01 3.66418600e-01 -7.57855654e-01 -3.07502449e-01
-1.01760077e+00 9.46845770e-01 4.78080958e-01 2.26122979e-02
-8.29182506e-01 -3.93301845e-01 -9.24299300e-01 -1.87513813e-01
4.59112942e-01 3.72702256e-02 1.64343631e+00 -1.13829541e+00
-1.77255642e+00 9.01733518e-01 -1.69249102e-01 -4.32487965e-01
3.69730234e-01 -1.11109406e-01 -8.48625779e-01 -7.97695294e-02
-1.50981456e-01 8.77814814e-02 8.26214254e-01 -1.07437778e+00
-5.26059926e-01 -5.02800941e-01 -2.91083097e-01 1.10451229e-01
-5.18153846e-01 2.96331912e-01 -6.68440282e-01 -5.83579063e-01
-9.23035145e-02 -9.67335224e-01 -2.71221548e-01 -5.43577135e-01
-8.29310790e-02 -1.64557621e-01 6.98916435e-01 -1.03054321e+00
1.15690696e+00 -2.34734225e+00 2.80266572e-02 1.56348526e-01
-5.03826022e-01 8.56891096e-01 -2.43386388e-01 5.06149590e-01
7.35735074e-02 -9.39158499e-02 -5.07142782e-01 -7.49154627e-01
-2.20162608e-02 6.38634920e-01 -6.99408948e-02 1.65620103e-01
1.75084338e-01 2.97061265e-01 -8.72528195e-01 -2.37257302e-01
7.96679929e-02 5.62879205e-01 -3.02492827e-01 5.70669115e-01
-8.06421414e-03 3.45605940e-01 -3.21261466e-01 2.94230103e-01
2.37517059e-01 1.60870656e-01 5.56303382e-01 3.14964503e-01
1.35869637e-01 7.48583794e-01 -1.21711719e+00 1.80446899e+00
-8.87381911e-01 4.25379485e-01 4.62646902e-01 -1.20433509e+00
1.19660246e+00 7.86029577e-01 2.65452564e-01 -7.80722558e-01
9.29494295e-03 8.60255301e-01 1.03419684e-01 -4.20957863e-01
2.76100904e-01 -5.66232443e-01 -2.53861219e-01 2.49224231e-01
3.47545147e-01 -5.24822831e-01 3.46676223e-02 -3.10650915e-01
1.34332073e+00 -1.93575278e-01 3.35433662e-01 -1.67115897e-01
7.10605919e-01 -9.39150378e-02 5.14583886e-01 4.72278833e-01
-2.71769851e-01 7.14198470e-01 9.07678623e-03 -1.96758270e-01
-1.19188368e+00 -7.67371058e-01 -1.80725813e-01 1.05976295e+00
-5.28065920e-01 -3.40684921e-01 -1.16463971e+00 -7.47689784e-01
-3.53364438e-01 8.84293914e-01 -2.76472121e-01 -2.04085419e-03
-7.27787614e-01 -4.69944328e-01 9.11627710e-01 3.00057292e-01
2.77297676e-01 -1.28867066e+00 -3.93608771e-02 6.45590961e-01
-1.23523638e-01 -1.56468761e+00 -3.30770701e-01 4.17814523e-01
-5.37502885e-01 -5.82704842e-01 -8.58518779e-01 -9.52090323e-01
3.28722328e-01 -1.49697468e-01 1.01395464e+00 -9.22821462e-02
1.78346902e-01 2.25604489e-01 -5.15776277e-01 -3.93894345e-01
-1.21273267e+00 2.54774600e-01 3.40566218e-01 2.30312161e-02
6.01419747e-01 -4.33176100e-01 1.37813821e-01 3.66975874e-01
-9.22570765e-01 -5.02259731e-01 5.50759137e-01 9.86461341e-01
4.26179022e-01 -7.48409927e-02 9.92671490e-01 -8.56815815e-01
5.25434434e-01 -2.65364856e-01 -4.39430535e-01 1.32881716e-01
-4.81437922e-01 1.50967345e-01 6.72509074e-01 -6.60416663e-01
-1.11501586e+00 2.55198032e-01 -9.56967354e-01 -1.67245939e-01
-5.84825516e-01 4.84466076e-01 -5.21894217e-01 2.74089932e-01
6.08350694e-01 3.86142254e-01 2.30462216e-02 -8.17417026e-01
1.25006154e-01 1.26225114e+00 5.90500474e-01 -4.94763881e-01
6.90807879e-01 -5.22918068e-02 -7.18184829e-01 -1.23015225e+00
-5.65778196e-01 -6.59325600e-01 -7.09403217e-01 6.77394569e-02
6.71265960e-01 -7.73505509e-01 8.63646567e-02 5.17860830e-01
-1.16355884e+00 -5.72784245e-01 -3.86000484e-01 6.05635762e-01
-5.29368579e-01 4.42994505e-01 -2.81517565e-01 -1.07105780e+00
-2.40929991e-01 -1.19938219e+00 1.23758566e+00 -2.38780633e-01
-3.58790100e-01 -1.07073927e+00 2.72358745e-01 5.24759352e-01
3.77967328e-01 -3.46043468e-01 7.26375401e-01 -1.36873472e+00
1.27656102e-01 -4.80169266e-01 3.63394022e-01 1.01962471e+00
3.22782934e-01 -3.06509852e-01 -1.30520511e+00 -3.09632391e-01
2.83594549e-01 -4.91591543e-01 3.60640496e-01 -1.74504891e-02
5.81249654e-01 -4.48097512e-02 -8.58118534e-02 -6.67191204e-03
1.01575565e+00 4.92519915e-01 4.70851719e-01 1.71109244e-01
2.92653471e-01 8.28672051e-01 4.67653900e-01 5.10711931e-02
1.60822183e-01 1.07404506e+00 -2.07066417e-01 -9.49138775e-03
-2.78275877e-01 -2.36728936e-01 6.11097932e-01 1.60851479e+00
3.27030569e-01 -3.30770433e-01 -1.12461662e+00 1.07343674e+00
-1.52068722e+00 -6.80092216e-01 2.41343200e-01 2.47750473e+00
1.13260901e+00 4.40917909e-01 1.68968394e-01 3.92044932e-01
6.91769123e-01 1.39006665e-02 -2.57910937e-01 -6.45678103e-01
-4.34440002e-02 3.45129400e-01 2.23140344e-02 6.22037947e-01
-9.13593829e-01 9.98327255e-01 6.60743237e+00 1.02724242e+00
-1.22599804e+00 3.03011239e-01 4.37046021e-01 6.28757104e-02
-2.07675740e-01 -1.39957175e-01 -6.23793125e-01 4.29827988e-01
1.70228767e+00 1.67143956e-01 3.51322204e-01 9.20949757e-01
2.26524815e-01 8.98888856e-02 -1.11530650e+00 9.54379439e-01
2.11764321e-01 -7.81464815e-01 -3.88773590e-01 2.07850888e-01
4.44547564e-01 2.12821215e-01 -2.28034586e-01 4.70943868e-01
2.71266997e-01 -8.62451017e-01 6.21366322e-01 -1.05603389e-01
7.53507555e-01 -6.15825891e-01 7.95718551e-01 7.44556248e-01
-8.69169414e-01 2.37291500e-01 -2.28737026e-01 2.89390236e-01
3.63484263e-01 5.25244594e-01 -1.34093249e+00 5.16099453e-01
4.72130239e-01 1.69721842e-01 -2.43811920e-01 6.49055362e-01
-8.36348161e-02 9.88311410e-01 -5.35640180e-01 -1.18477315e-01
3.48490000e-01 5.14538214e-02 5.16424179e-01 1.36231101e+00
4.28690344e-01 -1.11499697e-01 1.13958731e-01 1.35376588e-01
-1.69647466e-02 4.98572916e-01 -8.22976172e-01 -2.20485270e-01
4.31015015e-01 7.94774592e-01 -3.90283734e-01 -3.27533722e-01
-5.27949512e-01 9.87896442e-01 3.18951756e-01 2.09347233e-01
-3.33491385e-01 -2.79210955e-01 4.97396022e-01 9.15828049e-02
4.30302382e-01 -4.82489377e-01 1.26852959e-01 -9.79853749e-01
3.41954023e-01 -1.37910903e+00 7.48500451e-02 -3.91606212e-01
-1.28434491e+00 1.28434145e+00 -1.92349404e-01 -9.45131838e-01
-1.01907480e+00 -7.09944010e-01 -2.29946986e-01 9.78723049e-01
-1.38556397e+00 -1.00576401e+00 2.57911801e-01 5.53841412e-01
9.51718509e-01 -6.07837081e-01 1.32589638e+00 6.04522109e-01
-4.58272547e-01 5.63525975e-01 3.39077324e-01 1.11670889e-01
7.91230738e-01 -1.05782199e+00 5.89750230e-01 8.97205472e-01
5.53810060e-01 3.22753489e-01 7.78594136e-01 -3.63709390e-01
-1.20139945e+00 -7.86555946e-01 1.31877375e+00 -5.28195858e-01
6.77550554e-01 -6.79454327e-01 -1.31134200e+00 5.03325999e-01
1.12646475e-01 -1.95915103e-02 7.55006671e-01 3.28172088e-01
-3.02539945e-01 -8.34515020e-02 -1.06707895e+00 2.99655378e-01
9.06491816e-01 -9.32414114e-01 -8.51658404e-01 2.02268571e-01
5.73326170e-01 -1.47780746e-01 -9.55908656e-01 1.31386638e-01
3.22213441e-01 -6.43677115e-01 6.11782730e-01 -4.82275069e-01
-1.61402628e-01 1.11058317e-02 -6.08101010e-01 -1.50185692e+00
2.91853279e-01 -7.95654774e-01 2.94609755e-01 1.53443301e+00
8.06053400e-01 -5.61443150e-01 6.42536461e-01 6.14475846e-01
-4.20661360e-01 -3.05205613e-01 -1.41520786e+00 -1.00466526e+00
1.44416630e-01 -9.23867941e-01 4.56367075e-01 7.88307071e-01
3.85408029e-02 4.77159232e-01 -1.85146913e-01 1.30589560e-01
2.80448347e-01 -4.03013915e-01 8.96373332e-01 -1.07135046e+00
-4.10848349e-01 -1.03260465e-01 -4.93880928e-01 -9.19359922e-01
4.46764201e-01 -5.70808887e-01 6.53144360e-01 -1.23433149e+00
-3.59585643e-01 -6.06920302e-01 -2.11759225e-01 4.95150357e-01
6.00576699e-02 6.63422570e-02 4.74022739e-02 -1.14602141e-01
-1.37966409e-01 4.87004280e-01 6.49642885e-01 -1.27788097e-01
-2.48995528e-01 8.15437958e-02 -3.99014384e-01 6.27591789e-01
4.85724568e-01 -4.92187798e-01 -5.29659748e-01 -5.25461435e-01
-4.13848728e-01 9.49318111e-02 -1.82795957e-01 -1.09771526e+00
1.23852249e-02 -8.12835097e-02 -2.17006400e-01 1.84273403e-02
7.26789474e-01 -8.43937099e-01 1.01037674e-01 1.02896452e-01
-2.98397899e-01 -1.82859406e-01 4.38594848e-01 4.19580370e-01
-5.87896109e-01 -6.31391704e-01 8.00966561e-01 -6.78847358e-02
-7.20045507e-01 -1.25250876e-01 -5.36943793e-01 1.60668075e-01
6.61673725e-01 -1.21206909e-01 4.82300632e-02 -5.91260850e-01
-8.78448606e-01 -3.53587925e-01 3.83409590e-01 6.46491230e-01
3.36874306e-01 -1.14727139e+00 -8.76253068e-01 6.28169775e-01
2.99968839e-01 -1.03128180e-02 1.94822792e-02 5.17352045e-01
-4.24721278e-02 4.20374364e-01 2.22422481e-01 -5.37714303e-01
-1.26159537e+00 3.05073172e-01 3.10483277e-01 -2.47015551e-01
-5.60159385e-01 7.86155045e-01 -3.32533084e-02 -7.76207089e-01
2.47559562e-01 -2.27101803e-01 9.52251479e-02 -7.32778460e-02
5.82157671e-01 -1.72688350e-01 6.79124773e-01 -9.60676670e-01
-4.14836645e-01 2.01193824e-01 -6.00435585e-02 -7.06343472e-01
1.41276503e+00 -9.40096900e-02 5.64808428e-01 7.04828918e-01
1.18722641e+00 3.56242418e-01 -1.17951846e+00 -5.03747821e-01
3.58804792e-01 -2.11961091e-01 7.27884471e-02 -7.81874120e-01
-5.58282673e-01 9.71629977e-01 4.68538135e-01 5.20837069e-01
1.05118632e+00 2.03329921e-01 8.48784685e-01 4.44913626e-01
6.20059848e-01 -1.41408765e+00 -5.17646223e-02 5.45558631e-01
9.31584597e-01 -1.24622953e+00 -5.67365050e-01 -2.49421805e-01
-8.42385709e-01 9.10702288e-01 1.24881275e-01 8.29003230e-02
6.15820706e-01 4.09360379e-01 5.93252540e-01 1.50299102e-01
-6.90427125e-01 -2.21474543e-01 3.33684355e-01 8.64758551e-01
3.99750650e-01 -8.39268193e-02 1.19612060e-01 5.29906213e-01
-2.99229771e-01 -1.98032007e-01 3.42168123e-01 9.26486909e-01
-4.90566552e-01 -1.65806425e+00 -3.14031184e-01 -1.02031790e-03
-4.91207778e-01 -5.22472151e-02 -3.99949878e-01 8.54777396e-01
-1.56300873e-01 1.21674120e+00 -1.12782940e-01 -2.70583004e-01
6.31552577e-01 5.94120681e-01 2.38415748e-01 -9.90150034e-01
-5.11311591e-01 5.47325790e-01 7.49752462e-01 -3.50014299e-01
-4.44757849e-01 -9.38113928e-01 -1.06470501e+00 2.14180171e-01
-2.42917567e-01 4.00707126e-01 1.11325157e+00 1.27957153e+00
6.57345578e-02 2.94416159e-01 6.17246091e-01 -6.10379815e-01
-8.55500042e-01 -1.24228096e+00 -5.74389756e-01 4.53742266e-01
3.92054468e-01 -3.79587322e-01 -3.84337127e-01 2.72200167e-01]
|
[14.418939590454102, 6.665928840637207]
|
86a5623f-7ded-42d8-b429-5a6361fcee1c
|
streamlined-dense-video-captioning
|
1904.03870
| null |
http://arxiv.org/abs/1904.03870v1
|
http://arxiv.org/pdf/1904.03870v1.pdf
|
Streamlined Dense Video Captioning
|
Dense video captioning is an extremely challenging task since accurate and
coherent description of events in a video requires holistic understanding of
video contents as well as contextual reasoning of individual events. Most
existing approaches handle this problem by first detecting event proposals from
a video and then captioning on a subset of the proposals. As a result, the
generated sentences are prone to be redundant or inconsistent since they fail
to consider temporal dependency between events. To tackle this challenge, we
propose a novel dense video captioning framework, which models temporal
dependency across events in a video explicitly and leverages visual and
linguistic context from prior events for coherent storytelling. This objective
is achieved by 1) integrating an event sequence generation network to select a
sequence of event proposals adaptively, and 2) feeding the sequence of event
proposals to our sequential video captioning network, which is trained by
reinforcement learning with two-level rewards at both event and episode levels
for better context modeling. The proposed technique achieves outstanding
performances on ActivityNet Captions dataset in most metrics.
|
['Ning Xu', 'Jonghwan Mun', 'Zhou Ren', 'Linjie Yang', 'Bohyung Han']
|
2019-04-08
|
streamlined-dense-video-captioning-1
|
http://openaccess.thecvf.com/content_CVPR_2019/html/Mun_Streamlined_Dense_Video_Captioning_CVPR_2019_paper.html
|
http://openaccess.thecvf.com/content_CVPR_2019/papers/Mun_Streamlined_Dense_Video_Captioning_CVPR_2019_paper.pdf
|
cvpr-2019-6
|
['dense-video-captioning']
|
['computer-vision']
|
[ 3.97503614e-01 -2.51098555e-02 -2.78232455e-01 -5.33743680e-01
-8.18157017e-01 -5.49834847e-01 8.13890398e-01 3.11823606e-01
-2.94566542e-01 7.89958000e-01 7.58688092e-01 1.31294772e-01
2.78077006e-01 -5.85286856e-01 -1.00108039e+00 -3.36822003e-01
-9.76257473e-02 3.12436372e-01 4.97494817e-01 1.45083806e-02
2.48868957e-01 7.05152843e-03 -1.41753531e+00 6.25438988e-01
5.60088813e-01 9.60235715e-01 5.55779994e-01 6.87892139e-01
-2.52601206e-01 1.51194811e+00 -4.95091766e-01 -1.89016402e-01
-6.26779646e-02 -8.56890380e-01 -7.22370327e-01 4.26195443e-01
2.49200210e-01 -5.15839517e-01 -4.52759773e-01 8.60303223e-01
1.36902168e-01 4.02010500e-01 3.59989256e-01 -1.31730795e+00
-3.79953951e-01 8.44183803e-01 -5.65505266e-01 5.62381387e-01
7.88141012e-01 3.32217664e-01 1.10481119e+00 -7.12305784e-01
7.33295023e-01 9.54807162e-01 1.55645207e-01 6.08985603e-01
-8.63731623e-01 -5.25926113e-01 7.05327272e-01 5.93848705e-01
-1.20417511e+00 -5.62860250e-01 9.43485141e-01 -4.76876080e-01
7.99000084e-01 -4.66785617e-02 7.47537076e-01 1.38954031e+00
-6.98585808e-02 8.61927390e-01 3.82298499e-01 -4.34181020e-02
4.32390928e-01 -1.46811202e-01 -7.76606947e-02 5.06916821e-01
-8.58549923e-02 -3.20516825e-01 -6.46122217e-01 -8.24534670e-02
7.95015574e-01 2.43538007e-01 -2.92420417e-01 -2.30871871e-01
-1.44993186e+00 6.20620787e-01 3.10230076e-01 2.23750025e-01
-9.76618350e-01 4.64774907e-01 5.83830595e-01 -1.76617607e-01
7.11900294e-02 2.06321090e-01 -2.46827435e-02 -1.43377170e-01
-1.10729957e+00 4.21583623e-01 3.47212404e-01 9.71850574e-01
6.22945011e-01 -3.09082940e-02 -8.15678835e-01 6.53995812e-01
3.54562819e-01 8.79767239e-02 3.94121915e-01 -9.53210711e-01
8.72733593e-01 4.80594426e-01 4.14833993e-01 -1.09659612e+00
3.70017774e-02 -1.17434993e-01 -4.38456506e-01 -3.34183931e-01
9.12044942e-02 -1.61196649e-01 -9.32677686e-01 2.04351664e+00
3.12252253e-01 9.40468609e-01 5.38828932e-02 1.16643417e+00
9.02971625e-01 1.27008843e+00 6.99390531e-01 -5.02590358e-01
1.41767657e+00 -1.06885743e+00 -8.93671155e-01 -6.01496160e-01
1.07121967e-01 -5.40489912e-01 6.86740279e-01 -4.19983529e-02
-1.24276042e+00 -5.79110205e-01 -9.03705955e-01 1.37301059e-02
2.42366299e-01 4.12146971e-02 2.41787121e-01 -1.57155573e-01
-8.16097677e-01 1.58696130e-01 -9.11416590e-01 -3.51036340e-01
3.96245837e-01 7.69349560e-02 -2.58411825e-01 -2.56366074e-01
-1.23847139e+00 4.87624556e-01 8.16960156e-01 9.85278264e-02
-1.34743142e+00 -3.09201181e-01 -1.12063527e+00 2.54766166e-01
6.01859450e-01 -6.91173732e-01 1.51192558e+00 -1.43494320e+00
-1.32560658e+00 3.93275887e-01 -3.75033200e-01 -7.08849311e-01
2.85840422e-01 -2.60088056e-01 -3.71156871e-01 6.90061271e-01
2.33947977e-01 9.36770678e-01 9.23164010e-01 -1.19038713e+00
-8.01943660e-01 2.35279463e-02 2.02530250e-01 5.73853076e-01
-2.81041134e-02 1.94262996e-01 -8.12823474e-01 -6.73460126e-01
-1.50995821e-01 -5.87103009e-01 -2.65221536e-01 -1.98730007e-01
-2.00213522e-01 -2.25833088e-01 8.13714862e-01 -7.94911027e-01
1.29693401e+00 -1.95888793e+00 2.44346648e-01 -2.27369741e-01
-5.89641891e-02 1.08315729e-01 -2.52190679e-01 4.51091349e-01
-4.12379466e-02 -2.19256490e-01 -1.58306435e-01 -5.18647492e-01
-2.71957129e-01 2.37102747e-01 -6.10474765e-01 1.33119598e-01
6.41834199e-01 7.42995858e-01 -1.35085678e+00 -7.79442728e-01
3.56084853e-01 5.07655740e-01 -6.56287611e-01 7.17754662e-01
-7.74646342e-01 6.40905321e-01 -6.50055528e-01 4.52193409e-01
2.55103111e-01 -4.91957515e-01 3.07461977e-01 -2.54147559e-01
-6.36584684e-02 2.71134496e-01 -1.12902069e+00 1.96657145e+00
-2.67570019e-01 5.62324405e-01 -2.43616551e-01 -9.58595097e-01
6.37116373e-01 8.04079652e-01 5.86841524e-01 -5.58385313e-01
-3.46509479e-02 -1.93162382e-01 -4.05724943e-01 -8.96824419e-01
6.10168338e-01 -1.95192978e-01 -3.24398398e-01 4.16116238e-01
1.70107141e-01 3.21498126e-01 4.07065779e-01 4.87101316e-01
1.19238019e+00 4.59758162e-01 4.00749326e-01 5.71797550e-01
5.95732927e-01 8.95898268e-02 8.31426859e-01 7.41413772e-01
-2.96345830e-01 9.19300973e-01 6.68930948e-01 -5.24395943e-01
-1.06986737e+00 -1.02287579e+00 5.90186059e-01 1.14153373e+00
3.81800115e-01 -4.41110671e-01 -5.93743563e-01 -7.37792253e-01
-6.06276453e-01 8.07550788e-01 -5.90232909e-01 -9.71027836e-02
-8.09157372e-01 -4.21656430e-01 7.11528957e-02 6.45847499e-01
5.09403110e-01 -1.48210549e+00 -7.00503647e-01 5.94392180e-01
-6.84166193e-01 -1.51984739e+00 -7.49007463e-01 -1.05848029e-01
-5.28848231e-01 -1.00824213e+00 -6.30574763e-01 -9.31386590e-01
5.63277483e-01 1.88692614e-01 1.19058967e+00 -4.23377752e-02
1.33566722e-01 3.59233946e-01 -7.02312112e-01 2.82908063e-02
-4.53557998e-01 -1.36203229e-01 -2.23242685e-01 3.97623390e-01
2.24046215e-01 -3.70687217e-01 -7.79057920e-01 2.19386041e-01
-1.07226384e+00 5.31403184e-01 5.95905662e-01 5.91306210e-01
6.22891188e-01 -3.23569268e-01 7.87125051e-01 -5.59807479e-01
2.59794801e-01 -9.29714024e-01 -4.26527828e-01 2.99483567e-01
7.93052912e-02 -8.60769376e-02 7.90536702e-01 -3.88516039e-01
-1.29530311e+00 4.14703310e-01 9.98207554e-02 -7.37704813e-01
-2.18981132e-01 3.10418308e-01 -5.79017773e-02 6.39514863e-01
3.58417153e-01 4.29097205e-01 -3.89109612e-01 -5.06949686e-02
3.23559970e-01 4.09161687e-01 7.96855390e-01 -4.88155484e-01
5.01324832e-01 4.39071685e-01 -4.74812120e-01 -4.57222253e-01
-1.08271635e+00 -5.37989616e-01 -3.28186452e-01 -5.06052196e-01
1.21309412e+00 -1.24930060e+00 -4.57303613e-01 -3.66222225e-02
-1.53206325e+00 -1.02705382e-01 -1.88632950e-01 6.34108186e-01
-7.56271005e-01 2.64480919e-01 -5.03514171e-01 -6.08804107e-01
-1.32605597e-01 -1.23004413e+00 1.08906949e+00 5.10382891e-01
-2.49970198e-01 -6.58250451e-01 1.49999827e-01 3.69199336e-01
9.64821316e-03 5.48048615e-01 4.37286943e-01 -4.73738074e-01
-8.69647384e-01 -1.90335825e-01 -2.63323158e-01 2.20753485e-03
1.73903331e-01 1.03256265e-02 -7.05524147e-01 3.01326998e-03
-7.49899000e-02 -3.70301336e-01 6.89679563e-01 4.40660477e-01
1.04946446e+00 -5.34179449e-01 -3.41448516e-01 2.13937953e-01
1.37106574e+00 3.00487816e-01 6.03518069e-01 1.68145061e-01
7.27795660e-01 5.29916465e-01 6.68915808e-01 6.60769403e-01
5.30476689e-01 6.95394516e-01 5.66736519e-01 2.52162129e-01
-3.55741270e-02 -6.27883255e-01 4.45883721e-01 4.66986537e-01
6.77213222e-02 -5.37226796e-01 -6.62588060e-01 8.47507179e-01
-2.36571288e+00 -1.45710981e+00 2.64815718e-01 1.95756257e+00
6.73333704e-01 1.92193061e-01 2.20411628e-01 -2.00175911e-01
1.23633265e+00 5.38020432e-01 -5.48213542e-01 -1.45766571e-01
1.52813897e-01 -3.63128036e-01 3.02705448e-02 2.64489263e-01
-1.08036828e+00 9.72781718e-01 5.77963638e+00 3.10700506e-01
-9.93753433e-01 1.40911520e-01 7.31160581e-01 -3.92481118e-01
-1.94354936e-01 1.83506876e-01 -4.81487542e-01 7.07377553e-01
8.76679599e-01 -1.34816781e-01 1.63875118e-01 6.32975996e-01
6.25101805e-01 -1.65442228e-01 -1.23297977e+00 9.21526015e-01
2.01838225e-01 -1.57725871e+00 3.93751651e-01 -5.13792276e-01
7.92003691e-01 -1.22468553e-01 -3.53412271e-01 3.72254312e-01
1.37877122e-01 -8.20462883e-01 9.87805128e-01 5.54770172e-01
3.05170923e-01 -6.38805747e-01 5.16477346e-01 1.64699465e-01
-1.36070657e+00 -1.15727335e-01 -5.69817536e-02 -1.09023280e-01
7.61168718e-01 1.96679562e-01 -1.05753899e+00 4.10233289e-01
3.73201221e-01 9.78144705e-01 -2.36506164e-01 1.19283128e+00
-3.90426844e-01 7.28528500e-01 -4.27834876e-02 8.87421221e-02
4.90607917e-01 2.36267485e-02 5.62391698e-01 1.25698745e+00
2.62770265e-01 3.72007608e-01 5.25304556e-01 7.33034313e-01
-2.12424681e-01 -3.31961140e-02 -5.24457276e-01 -1.05004042e-01
4.95878935e-01 1.19003403e+00 -8.83591235e-01 -6.28479421e-01
-6.05632484e-01 1.08548617e+00 3.62226397e-01 3.95327330e-01
-1.22042537e+00 -2.37774942e-02 4.20543075e-01 7.56849125e-02
5.17795861e-01 -5.41142710e-02 2.19521344e-01 -1.33467400e+00
9.71629769e-02 -5.37835896e-01 7.31864214e-01 -1.22677004e+00
-8.65359902e-01 6.71624720e-01 -5.66746294e-03 -1.41733325e+00
-4.94207621e-01 4.24148366e-02 -8.72034192e-01 4.71933365e-01
-1.51663780e+00 -8.66265774e-01 -4.75383967e-01 5.60782611e-01
1.14479113e+00 2.39324555e-01 3.02009255e-01 3.39383394e-01
-6.85345471e-01 3.60044278e-02 -4.77561802e-01 1.47347242e-01
5.87554634e-01 -1.01429343e+00 3.00671846e-01 1.05928421e+00
2.08618447e-01 9.97832268e-02 9.79403019e-01 -8.54229748e-01
-1.01479471e+00 -1.36122930e+00 9.06620204e-01 -3.32784414e-01
5.22222817e-01 -2.24879101e-01 -9.96655166e-01 8.90926123e-01
4.30072039e-01 1.18997231e-01 4.56141919e-01 -4.63159621e-01
-1.89482033e-01 -3.69846225e-02 -8.32377434e-01 7.00930238e-01
1.01471269e+00 -3.48478138e-01 -9.05437529e-01 4.25991565e-01
8.54110479e-01 -4.91213143e-01 -3.78876686e-01 1.30896121e-01
2.35206693e-01 -7.73669302e-01 8.28543961e-01 -6.19045258e-01
8.89158845e-01 -6.17842615e-01 -3.66929770e-02 -9.23625648e-01
-2.41696268e-01 -6.64745986e-01 -3.22626740e-01 1.26076877e+00
2.27918357e-01 3.23558867e-01 7.50039816e-01 5.22623599e-01
-2.26319313e-01 -3.62140000e-01 -6.81682348e-01 -3.02545428e-01
-7.83592165e-01 -3.44901294e-01 4.65743393e-01 7.70008743e-01
2.89118644e-02 5.25493681e-01 -8.41492295e-01 3.42957437e-01
4.62344170e-01 2.39419654e-01 5.34542203e-01 -6.78892195e-01
-2.05338940e-01 -1.48104623e-01 -3.97088379e-01 -1.02524757e+00
1.53636754e-01 -6.33264780e-01 5.86647451e-01 -1.80522454e+00
3.71426433e-01 1.63316667e-01 -3.36963534e-01 4.29916322e-01
-4.52701390e-01 2.50127524e-01 1.84688181e-01 3.33822489e-01
-1.36501443e+00 6.97425067e-01 1.14296889e+00 -7.15862066e-02
-2.66429275e-01 -2.80258745e-01 -3.86516243e-01 6.45747066e-01
5.46449542e-01 -4.60572630e-01 -6.16183817e-01 -4.96339619e-01
9.42791477e-02 7.99228609e-01 5.35044491e-01 -1.17899108e+00
3.93025458e-01 -5.13538659e-01 4.14405257e-01 -5.65316677e-01
4.62906629e-01 -6.47375286e-01 1.98925942e-01 8.59126970e-02
-6.68424547e-01 1.89273849e-01 4.36962880e-02 9.65408683e-01
-4.61195499e-01 -1.30145222e-01 5.25318146e-01 -3.33820134e-01
-1.16173279e+00 5.26527166e-01 -5.43853045e-01 1.57436296e-01
1.39420438e+00 -1.46323398e-01 -1.03596643e-01 -6.08888030e-01
-8.39046776e-01 5.61793447e-01 3.63987029e-01 6.70866132e-01
6.74395502e-01 -1.39677322e+00 -7.94134915e-01 -1.85895547e-01
2.84266055e-01 1.94245949e-01 5.05325854e-01 4.60587949e-01
-4.41140205e-01 1.96944639e-01 -2.95156658e-01 -6.46601975e-01
-9.70091283e-01 6.99946761e-01 6.36468381e-02 -1.58714265e-01
-6.22362494e-01 6.66641891e-01 3.26155961e-01 4.58650053e-01
3.44091892e-01 -2.61408627e-01 -6.17686629e-01 8.75976086e-02
7.58539081e-01 -1.76037520e-01 -4.02025223e-01 -8.73534501e-01
-3.10065538e-01 3.12561184e-01 -3.30460697e-01 -2.47096106e-01
1.30850732e+00 -3.90129149e-01 3.07133079e-01 4.18864042e-01
1.01754141e+00 -3.68931800e-01 -1.80480421e+00 -2.39215985e-01
-2.45416518e-02 -3.33207786e-01 -1.75655976e-01 -6.44856751e-01
-8.33470762e-01 5.00561893e-01 9.49337706e-02 7.42577091e-02
1.16743755e+00 1.72811791e-01 1.01978660e+00 7.28937387e-02
1.17624737e-01 -9.86474454e-01 5.58194578e-01 2.56662548e-01
8.32212746e-01 -1.27943075e+00 -1.76364481e-01 -2.61665136e-01
-1.06876934e+00 9.47942257e-01 8.40668440e-01 -1.67616934e-01
2.08713897e-02 -2.23203093e-01 -1.32240444e-01 -1.20962597e-01
-1.12230051e+00 -1.23309553e-01 1.20881952e-01 2.18894079e-01
2.92838693e-01 -2.84007043e-01 -2.74060965e-01 5.77321410e-01
4.33073282e-01 2.46184319e-01 6.58225834e-01 1.00675321e+00
-5.05823970e-01 -7.67620385e-01 -3.23752522e-01 2.29331046e-01
-4.37926739e-01 8.47635865e-02 -1.61662668e-01 1.16190724e-01
1.08945213e-01 8.17692757e-01 3.39509398e-01 -1.13482460e-01
1.79054976e-01 9.14307684e-02 1.85987189e-01 -8.31045926e-01
-5.38315415e-01 1.68728754e-01 4.30549234e-02 -5.97466946e-01
-6.18531883e-01 -8.65845919e-01 -1.35558534e+00 3.03440750e-01
8.68940502e-02 4.01796132e-01 3.93947005e-01 9.99521613e-01
4.57925886e-01 6.66515827e-01 7.09173501e-01 -8.94248784e-01
-4.70431969e-02 -7.28847265e-01 -7.49926269e-02 7.78379738e-01
4.16391611e-01 -4.44102973e-01 -9.36516002e-02 5.92695057e-01]
|
[10.39478874206543, 0.6132733225822449]
|
7dcc04c3-b0ed-437e-967d-cbe82c9d6532
|
object-categorization-in-finer-levels
|
1703.09990
| null |
http://arxiv.org/abs/1703.09990v1
|
http://arxiv.org/pdf/1703.09990v1.pdf
|
Object categorization in finer levels requires higher spatial frequencies, and therefore takes longer
|
The human visual system contains a hierarchical sequence of modules that take
part in visual perception at different levels of abstraction, i.e.,
superordinate, basic, and subordinate levels. One important question is to
identify the "entry" level at which the visual representation is commenced in
the process of object recognition. For a long time, it was believed that the
basic level had advantage over two others; a claim that has been challenged
recently. Here we used a series of psychophysics experiments, based on a rapid
presentation paradigm, as well as two computational models, with bandpass
filtered images to study the processing order of the categorization levels. In
these experiments, we investigated the type of visual information required for
categorizing objects in each level by varying the spatial frequency bands of
the input image. The results of our psychophysics experiments and computational
models are consistent. They indicate that the different spatial frequency
information had different effects on object categorization in each level. In
the absence of high frequency information, subordinate and basic level
categorization are performed inaccurately, while superordinate level is
performed well. This means that, low frequency information is sufficient for
superordinate level, but not for the basic and subordinate levels. These finer
levels require high frequency information, which appears to take longer to be
processed, leading to longer reaction times. Finally, to avoid the ceiling
effect, we evaluated the robustness of the results by adding different amounts
of noise to the input images and repeating the experiments. As expected, the
categorization accuracy decreased and the reaction time increased
significantly, but the trends were the same.This shows that our results are not
due to a ceiling effect.
|
['Mohammad Ganjtabesh', 'Timothée Masquelier', 'Matin N. Ashtiani', 'Saeed Reza Kheradpisheh']
|
2017-03-29
| null | null | null | null |
['object-categorization']
|
['computer-vision']
|
[ 8.65874365e-02 -4.78803545e-01 3.62522930e-01 -2.29309544e-01
1.39344379e-01 -7.90688872e-01 4.36102927e-01 6.69346035e-01
-7.37728894e-01 3.15375000e-01 -2.55053081e-02 -4.27358299e-01
-2.53979534e-01 -6.82304204e-01 -4.43934977e-01 -7.31074095e-01
1.03488036e-01 -1.87019691e-01 7.65842497e-01 -2.03639925e-01
7.64311075e-01 6.58799171e-01 -2.30269814e+00 7.58413374e-01
4.86823231e-01 9.06156719e-01 4.22013074e-01 6.20971382e-01
-1.64983645e-01 4.14013684e-01 -8.90537202e-01 1.02259055e-01
4.03656095e-01 -6.27964258e-01 -7.54220545e-01 2.53831625e-01
4.03109998e-01 6.76652491e-02 1.72940701e-01 1.13505065e+00
2.97476500e-01 2.30936930e-01 6.82744861e-01 -9.80598092e-01
-4.73968595e-01 2.34759554e-01 -4.44877148e-01 4.75002706e-01
4.72663909e-01 2.33466983e-01 6.52468562e-01 -6.54457033e-01
3.27918708e-01 1.44478059e+00 1.60149038e-01 1.41695514e-01
-1.73806584e+00 -4.60949957e-01 4.09814090e-01 2.28412360e-01
-1.43956840e+00 -3.07921857e-01 3.94595325e-01 -9.35047150e-01
7.31817782e-01 4.17558521e-01 7.56406784e-01 5.86886048e-01
4.42002863e-01 -2.12660179e-01 1.97429669e+00 -6.73914433e-01
2.81213522e-01 6.52334630e-01 4.45029736e-01 1.02916822e-01
6.56190515e-01 2.87241966e-01 -2.85804182e-01 2.64105760e-03
7.79931188e-01 -2.30242535e-01 -4.59738195e-01 -6.66756406e-02
-9.38715518e-01 5.92955649e-01 4.53822583e-01 9.07051921e-01
-4.16507602e-01 -3.08353961e-01 3.09703350e-01 4.99437034e-01
1.35317538e-02 6.34713173e-01 -1.80466339e-01 1.29327178e-01
-5.53355157e-01 1.13820769e-01 5.60358882e-01 2.34440327e-01
5.66603065e-01 -2.36575678e-01 -2.79851317e-01 7.31548131e-01
1.57728150e-01 -1.81235210e-03 5.41101098e-01 -6.92279577e-01
1.47957325e-01 6.91647470e-01 3.17010134e-01 -1.14550805e+00
-4.89498436e-01 -2.30433553e-01 -4.07077879e-01 9.31654453e-01
1.04229569e+00 3.09113413e-01 -9.45474029e-01 1.74417377e+00
-3.79647277e-02 -6.87654495e-01 -2.54078507e-01 1.20137823e+00
3.01630497e-01 6.83452845e-01 5.17473876e-01 -3.87621194e-01
1.81486118e+00 -1.58862531e-01 -6.53631687e-01 -1.75580382e-01
2.66088903e-01 -9.12471414e-01 1.41224146e+00 5.29562235e-01
-9.35282648e-01 -1.18128705e+00 -1.20005763e+00 1.93600655e-01
-6.01968646e-01 -1.55321583e-01 5.55828989e-01 7.04052031e-01
-1.06569886e+00 6.48275554e-01 -5.11717081e-01 -4.51671720e-01
-2.84071177e-01 1.07132554e-01 -3.54493558e-01 1.80741116e-01
-1.07755506e+00 1.24071944e+00 6.13750815e-01 -1.65321082e-02
-3.31390530e-01 -3.36419314e-01 -5.91785252e-01 3.66521567e-01
-6.84341565e-02 -2.53746629e-01 9.45839107e-01 -1.25524771e+00
-9.76656735e-01 1.00449336e+00 -2.57163554e-01 3.38113196e-02
1.91465586e-01 3.18753451e-01 -4.18755472e-01 1.45429105e-01
9.12311524e-02 4.93744642e-01 7.39675105e-01 -1.49430239e+00
-6.25315845e-01 -6.51660919e-01 1.49849325e-01 1.14548862e-01
1.45449981e-01 1.28494397e-01 2.21635010e-02 -4.80750889e-01
3.52338612e-01 -6.66587830e-01 1.58076081e-02 -1.83017701e-01
3.03037345e-01 -3.67635638e-01 2.70454317e-01 -4.13048565e-01
1.24549651e+00 -2.77477622e+00 -1.88140273e-01 3.62505138e-01
1.82326555e-01 8.55244547e-02 -1.02309082e-02 5.83104849e-01
-5.01944184e-01 2.63615668e-01 3.96911837e-02 4.46267307e-01
-1.11899756e-01 -1.88007966e-01 -7.55342692e-02 3.39751035e-01
-8.39689746e-02 2.86478013e-01 -5.40922046e-01 -1.96198910e-01
1.05280712e-01 2.90785670e-01 -3.55129153e-01 5.91217354e-03
2.57312864e-01 2.50878006e-01 6.64673746e-02 1.72970295e-01
8.26133907e-01 -4.80365306e-02 1.40600905e-01 -3.76509637e-01
-6.36360943e-01 2.66733855e-01 -1.26714003e+00 9.10840631e-01
-1.60212353e-01 8.37123930e-01 2.79493770e-03 -8.67389798e-01
8.16257238e-01 2.95907021e-01 6.06053583e-02 -1.12815666e+00
2.16372401e-01 2.53148139e-01 8.60548496e-01 -4.54804122e-01
3.96273613e-01 -4.30083871e-01 1.44403100e-01 2.87736058e-01
-2.06687197e-01 -8.43640491e-02 5.51160157e-01 -9.81677398e-02
3.41627121e-01 -3.45521897e-01 4.18634981e-01 -4.53573734e-01
4.17482138e-01 -3.21006253e-02 2.97345668e-01 7.93802202e-01
-1.74797386e-01 3.06695253e-01 7.42869556e-01 -3.75803977e-01
-7.81319261e-01 -1.00827050e+00 -4.49420005e-01 1.10146797e+00
4.61716890e-01 -2.47525916e-01 -7.96770990e-01 -4.65382524e-02
-6.37526661e-02 8.33372712e-01 -6.66956246e-01 -2.86916792e-01
-1.65196359e-01 -7.20727742e-01 2.53287535e-02 3.15759897e-01
3.02578479e-01 -1.17330933e+00 -1.26428461e+00 1.23034632e-02
-6.16371110e-02 -9.18062627e-01 -1.39884338e-01 2.42406279e-01
-8.27076495e-01 -1.12347722e+00 -4.73810852e-01 -5.97179592e-01
8.72828662e-01 5.73945999e-01 7.70947218e-01 4.02245939e-01
-4.03803885e-01 2.23116577e-01 -3.37998182e-01 -5.15903652e-01
-4.12219554e-01 -4.74847853e-01 -4.73169424e-02 3.83901456e-03
4.22881484e-01 -3.68125945e-01 -7.10919142e-01 4.99047399e-01
-1.02289402e+00 -9.39513221e-02 8.09642136e-01 4.57711726e-01
1.87100217e-01 3.59180570e-01 3.06803018e-01 -3.02099824e-01
8.93205702e-01 -2.40832176e-02 -6.74003839e-01 -3.91256027e-02
-4.76952970e-01 -1.18608272e-03 6.40890419e-01 -6.26730919e-01
-8.81762922e-01 -2.64785439e-01 4.96433191e-02 1.78315938e-01
-5.64281166e-01 3.80005389e-01 -2.57260771e-03 -2.99179822e-01
1.03401637e+00 1.93810314e-01 -2.25804560e-02 -2.57138520e-01
-2.09996670e-01 6.43686652e-01 1.52937844e-02 -3.55357528e-01
4.84875917e-01 2.89682359e-01 -1.19782485e-01 -9.14389253e-01
-5.40860593e-01 -3.19248915e-01 -6.49396479e-01 -4.28745866e-01
1.00554752e+00 -5.21332145e-01 -8.65213752e-01 3.88490051e-01
-1.06257403e+00 -2.24364862e-01 -1.26951069e-01 7.26148963e-01
-2.58394212e-01 3.43451053e-01 -3.90886009e-01 -8.61829102e-01
3.76212865e-01 -1.20202577e+00 4.45709050e-01 3.10230047e-01
-3.20664644e-01 -5.12394071e-01 -2.73839980e-01 6.41714642e-03
3.40952426e-01 4.54524048e-02 1.20931756e+00 -4.40977067e-01
-3.73408049e-01 -6.74067112e-03 -3.40859234e-01 3.83668333e-01
3.45945150e-01 2.56848950e-02 -9.62567210e-01 -2.79103637e-01
4.37850177e-01 -1.41633362e-01 8.90696943e-01 3.88775229e-01
9.22008693e-01 -8.04848000e-02 -1.46338984e-01 9.07118246e-02
1.62347949e+00 8.10179830e-01 7.45645642e-01 2.91082412e-01
6.66859224e-02 1.06924129e+00 5.40121555e-01 -9.75857899e-02
-8.02858472e-02 6.87456787e-01 -8.08793381e-02 -2.22740501e-01
4.82630432e-02 1.16749860e-01 1.75853029e-01 3.45058143e-01
-2.88332969e-01 -2.85070832e-03 -7.83878386e-01 4.01630700e-01
-1.35827088e+00 -9.90485132e-01 -3.49755138e-01 2.82118177e+00
5.22533476e-01 4.68229920e-01 2.67914832e-01 6.17140651e-01
6.84424520e-01 -1.08291671e-01 -1.13159895e-01 -7.96976447e-01
1.39623061e-01 -1.14983216e-01 2.52442569e-01 4.48085040e-01
-5.57569802e-01 5.06638527e-01 7.13503122e+00 5.26808143e-01
-1.41503239e+00 -2.65183002e-01 4.89309072e-01 2.16563687e-01
-9.18492768e-03 1.74534634e-01 -5.09781003e-01 6.36301577e-01
7.49107897e-01 -2.30840638e-01 2.96229482e-01 4.36771542e-01
4.26884711e-01 -9.23252046e-01 -1.09382832e+00 7.31584251e-01
-2.15688035e-01 -6.07315660e-01 1.80870891e-01 7.70761967e-02
-3.98820490e-02 -6.12553060e-01 -1.40203103e-01 1.34556770e-01
-3.47993523e-01 -9.75457311e-01 7.92693973e-01 3.99050146e-01
3.95042449e-01 -4.42101896e-01 5.43024957e-01 4.41492081e-01
-1.11147404e+00 -1.72092974e-01 -6.11967862e-01 -6.71097696e-01
-1.01190768e-01 3.64267468e-01 -4.40830112e-01 1.11010268e-01
7.50900686e-01 -3.52321744e-01 -8.05161417e-01 1.13507724e+00
-1.27475351e-01 3.39265019e-01 -5.02058528e-02 -4.35462855e-02
-5.83172254e-02 -1.14921637e-01 3.93798381e-01 1.09041131e+00
9.21790078e-02 3.39424849e-01 -1.55756652e-01 9.04481292e-01
6.56687975e-01 1.81730464e-01 -4.00883049e-01 4.72314609e-03
4.73376304e-01 1.13192117e+00 -1.21655822e+00 -4.01396036e-01
-5.55384696e-01 7.32031584e-01 6.32674545e-02 5.99851966e-01
-5.44168591e-01 -5.70108235e-01 3.63332808e-01 4.24351722e-01
4.22483385e-02 -2.79723614e-01 -3.46948802e-01 -8.52860808e-01
-2.14391761e-02 -7.48723269e-01 3.62098426e-01 -7.66871631e-01
-9.79663968e-01 5.39920449e-01 3.50546330e-01 -1.02069163e+00
2.76900064e-02 -8.09552610e-01 -2.85898089e-01 1.40345347e+00
-9.33391273e-01 -3.31596255e-01 -3.94817114e-01 3.97199422e-01
4.15704042e-01 4.23109889e-01 7.23699450e-01 1.02987275e-01
-4.19391319e-02 3.41692746e-01 -4.23078924e-01 6.95106611e-02
6.69954598e-01 -1.18860781e+00 -1.50169849e-01 7.22757101e-01
4.81612980e-02 8.68894756e-01 8.86707306e-01 -3.96360964e-01
-6.48180366e-01 -2.30496868e-01 9.76722002e-01 -2.43766159e-01
2.85528302e-01 -5.39647102e-01 -1.22424316e+00 1.11540340e-01
2.03912735e-01 -5.15510380e-01 6.86903358e-01 2.66543385e-02
-4.24067885e-01 -2.11821288e-01 -1.08831489e+00 6.44247591e-01
6.11562908e-01 -4.20363367e-01 -9.01399851e-01 -1.30069837e-01
4.87574875e-01 4.62288745e-02 -6.58922255e-01 2.98305809e-01
7.44716048e-01 -1.44987142e+00 8.07974160e-01 -3.15381408e-01
7.86177889e-02 -4.77635264e-01 2.31401548e-02 -1.37525821e+00
-9.92791057e-01 3.04163486e-01 7.17493355e-01 9.73511875e-01
4.45704073e-01 -9.58908975e-01 -1.40009290e-02 4.86195773e-01
1.72563717e-01 -3.43788922e-01 -6.05555117e-01 -8.51414800e-01
-1.08498462e-01 -1.49771959e-01 1.15357667e-01 6.50252223e-01
1.77015781e-01 4.16054428e-01 2.60810703e-01 1.90785319e-01
3.62283379e-01 2.72253633e-01 2.33173013e-01 -1.42424953e+00
-2.23189160e-01 -7.95475602e-01 -6.84210479e-01 -5.50725818e-01
-3.33612919e-01 -3.97472411e-01 -3.66727896e-02 -1.47885287e+00
1.03617676e-01 -1.43127948e-01 -4.91216272e-01 2.38316402e-01
-3.43871653e-01 1.93606496e-01 5.73713303e-01 3.24933141e-01
2.34739408e-01 -1.56318158e-01 1.23912036e+00 9.88491327e-02
-4.95116502e-01 2.01309472e-02 -8.97518218e-01 6.93471968e-01
7.78400242e-01 -2.49675229e-01 -4.67876047e-01 -1.31573617e-01
7.43622035e-02 -1.39436005e-02 5.02516985e-01 -1.18725562e+00
1.34431496e-01 -1.57411560e-01 7.43449152e-01 -2.96922714e-01
2.73197442e-01 -8.43292415e-01 1.43589914e-01 7.88219452e-01
-2.88897663e-01 1.93628848e-01 7.13062823e-01 1.47867993e-01
-2.30709508e-01 -4.58799899e-01 1.22273660e+00 -1.87035561e-01
-8.47357273e-01 -4.90605563e-01 -9.33380842e-01 -4.10045981e-01
1.06364739e+00 -6.56336963e-01 -1.07005835e-01 -1.44712359e-01
-1.03029072e+00 -2.95984864e-01 5.01088560e-01 4.35533315e-01
3.23051304e-01 -9.26380992e-01 -5.26468158e-01 2.06778601e-01
1.56914145e-01 -7.40863085e-01 2.95529783e-01 9.09156919e-01
-3.19576979e-01 5.16295731e-01 -6.83366776e-01 -6.70102775e-01
-1.34165931e+00 1.03553951e+00 3.51975948e-01 2.06713855e-01
-2.24119708e-01 4.49679554e-01 7.59491503e-01 3.59033853e-01
1.89020634e-01 -5.20802736e-01 -5.78449845e-01 4.72356409e-01
7.41510808e-01 2.30443031e-01 8.72610584e-02 -5.96104801e-01
-4.17671591e-01 8.15181494e-01 5.59285618e-02 -3.08396399e-01
6.67267740e-01 -2.06096828e-01 -2.21672639e-01 9.92617249e-01
6.77712262e-01 1.77954867e-01 -7.89425790e-01 3.75033736e-01
-1.97734330e-02 -7.77930498e-01 -2.41341442e-01 -8.23516190e-01
-3.59735608e-01 9.65249419e-01 8.38405073e-01 8.12381446e-01
1.36591041e+00 -3.51421051e-02 -2.92137861e-01 -7.26648867e-02
5.24180830e-01 -1.03993797e+00 -1.24544360e-01 1.83054805e-01
1.01148462e+00 -8.45708549e-01 -1.23563714e-01 -6.03972197e-01
-5.08368671e-01 8.78372312e-01 7.40761399e-01 -4.42423932e-02
3.39579970e-01 -7.08950981e-02 1.00017399e-01 -2.72387415e-01
-7.41403282e-01 -3.88232768e-01 3.66573334e-01 3.92055869e-01
7.73216605e-01 -3.74784619e-02 -9.26104665e-01 3.31679165e-01
-1.73113585e-01 -2.17205942e-01 4.42340046e-01 8.00239861e-01
-8.00115466e-01 -8.77371907e-01 -9.15529847e-01 3.72849673e-01
-3.56156766e-01 3.92900147e-02 -5.74978769e-01 1.06473374e+00
4.34291542e-01 1.09270191e+00 4.05561566e-01 -2.76366323e-01
6.88008189e-01 1.68468907e-01 6.98470831e-01 -5.59179664e-01
-5.50597250e-01 1.02587074e-01 -1.89116642e-01 -2.98505694e-01
-2.90843308e-01 -3.78184855e-01 -1.06978643e+00 -4.31722999e-02
-2.61480778e-01 5.57646692e-01 5.69624484e-01 7.08388567e-01
5.08186184e-02 5.22216499e-01 3.81843925e-01 -7.81722128e-01
-3.52060407e-01 -1.12364113e+00 -7.70868659e-01 6.43577635e-01
1.70226499e-01 -7.85573006e-01 -7.74428427e-01 -3.04991379e-02]
|
[10.107155799865723, 2.1731743812561035]
|
6eb3ffef-53df-4b96-aaf9-2236cd7a268f
|
multi-talker-mvdr-beamforming-based-on
|
1910.07753
| null |
http://arxiv.org/abs/1910.07753v1
|
http://arxiv.org/pdf/1910.07753v1.pdf
|
Multi-Talker MVDR Beamforming Based on Extended Complex Gaussian Mixture Model
|
In this letter, we present a novel multi-talker minimum variance
distortionless response (MVDR) beamforming as the front-end of an automatic
speech recognition (ASR) system in a dinner party scenario. The CHiME-5 dataset
is selected to evaluate our proposal for overlapping multi-talker scenario with
severe noise. A detailed study on beamforming is conducted based on the
proposed extended complex Gaussian mixture model (CGMM) integrated with various
speech separation and speech enhancement masks. Three main changes are made to
adopt the original CGMM-based MVDR for the multi-talker scenario. First, the
number of Gaussian distributions is extended to 3 with an additional inference
speaker model. Second, the mixture coefficients are introduced as a supervisor
to generate more elaborate masks and avoid the permutation problems. Moreover,
we reorganize the MVDR and mask-based speech separation to achieve both noise
reduction and target speaker extraction. With the official baseline ASR
back-end, our front-end algorithm gained an absolute WER reduction of 13.87%
compared with the baseline front-end.
|
[]
|
2019-10-17
| null | null | null | null |
['target-speaker-extraction']
|
['audio']
|
[ 1.50191978e-01 -3.02198599e-03 6.53970659e-01 -5.83245635e-01
-1.36989951e+00 -4.21161890e-01 5.38864672e-01 -3.81800860e-01
-4.67501104e-01 1.89653620e-01 6.44557059e-01 -4.25923616e-01
-6.24444187e-02 -1.87318847e-01 -3.06812584e-01 -1.09107077e+00
3.27330440e-01 5.19635417e-02 4.47969176e-02 -3.04327428e-01
-2.03541800e-01 4.84866321e-01 -1.45831943e+00 5.12344182e-01
7.83859611e-01 6.58238351e-01 6.84064031e-01 9.17142153e-01
1.19901806e-01 3.05213511e-01 -1.00770593e+00 -4.40106630e-01
2.00316846e-01 -3.33208919e-01 -2.24107340e-01 1.46092072e-01
1.85268462e-01 -5.94232455e-02 -5.80774844e-01 1.03377628e+00
1.28146005e+00 6.10510409e-01 5.67373037e-01 -6.47524834e-01
-1.89661413e-01 1.06751239e+00 -4.61469740e-01 2.20986351e-01
1.85436279e-01 4.21783850e-02 6.99823797e-01 -1.09397042e+00
-8.35508034e-02 1.54821801e+00 3.16696227e-01 6.34801745e-01
-1.06634820e+00 -6.13523602e-01 1.49382755e-01 2.75717199e-01
-1.74402440e+00 -1.19299424e+00 7.19238281e-01 -2.25617051e-01
9.72462356e-01 7.92873502e-01 -3.00674140e-02 1.05276990e+00
-4.46312904e-01 6.72222614e-01 8.73421907e-01 -5.31373203e-01
2.13029936e-01 1.52764723e-01 2.02985853e-01 -4.76272404e-02
-3.04730814e-02 -6.04700483e-02 -5.54438770e-01 7.63448775e-02
2.70022243e-01 -4.76734459e-01 -4.98764902e-01 3.19322169e-01
-1.09047639e+00 4.10595685e-01 -4.97857518e-02 6.24381125e-01
-3.45832437e-01 -1.12763129e-01 1.04194887e-01 2.14840725e-01
3.97766709e-01 -8.85929260e-03 -2.08344683e-01 8.57451372e-03
-1.21366239e+00 1.18374430e-01 6.87424183e-01 1.05810905e+00
2.26354137e-01 5.30246615e-01 -7.44077504e-01 1.46043456e+00
6.95611775e-01 8.99713039e-01 3.84691566e-01 -5.14503479e-01
9.17375803e-01 -2.60193110e-01 3.10142115e-02 -7.08112180e-01
-1.79962575e-01 -1.22032237e+00 -1.00228775e+00 -1.04165919e-01
1.10843748e-01 -3.65961820e-01 -9.57551122e-01 1.54701209e+00
2.82751024e-01 -5.39845973e-02 3.82423431e-01 1.12088835e+00
8.52398157e-01 1.00297749e+00 -4.73681599e-01 -4.44433659e-01
1.38054621e+00 -1.10492933e+00 -1.26968122e+00 -1.72948658e-01
3.02392334e-01 -1.18015599e+00 7.57229626e-01 7.90989399e-01
-1.09902215e+00 -6.15635335e-01 -1.10489070e+00 1.95385665e-01
3.87981348e-02 5.19930840e-01 -1.44605860e-01 1.36955154e+00
-1.09590447e+00 -9.39251781e-02 -6.91806316e-01 2.18931660e-02
-2.05983922e-01 2.10851446e-01 -1.20540880e-01 -8.41533616e-02
-1.01274896e+00 7.02377081e-01 7.03013539e-02 6.04024649e-01
-8.19867432e-01 -4.16699678e-01 -7.31093109e-01 2.13443875e-01
2.07587436e-01 -4.11039442e-01 1.42975128e+00 -6.32361054e-01
-2.01672220e+00 4.58870769e-01 -6.74350381e-01 -4.04957920e-01
6.30049884e-01 -1.60633221e-01 -8.44998896e-01 -1.71361268e-01
-5.99234641e-01 3.77339758e-02 8.61077845e-01 -1.28531456e+00
-3.48375022e-01 -2.97526211e-01 -5.17149508e-01 6.22661352e-01
-1.32134840e-01 4.80633289e-01 -6.08607411e-01 -1.11065972e+00
5.64701021e-01 -5.46674490e-01 -1.36563510e-01 -1.00014734e+00
-4.99734551e-01 1.47056267e-01 5.51608026e-01 -1.15906537e+00
1.59627998e+00 -2.48776770e+00 1.52045339e-01 3.40496659e-01
-2.14742824e-01 4.34572548e-01 2.57934667e-02 3.34207892e-01
-2.55676657e-01 -3.01648706e-01 -2.28236452e-01 -8.72942924e-01
1.62514493e-01 -1.77214712e-01 -1.50473043e-01 4.71690148e-01
-2.28486642e-01 1.07968345e-01 -3.09581637e-01 -1.07627101e-01
3.91341537e-01 7.44255722e-01 -6.09152198e-01 2.75531173e-01
3.29917878e-01 5.37959814e-01 1.24775045e-01 5.04180849e-01
1.18844295e+00 7.22427487e-01 1.48049906e-01 -2.87969023e-01
-3.33199859e-01 4.71314460e-01 -1.73049128e+00 1.53337622e+00
-5.07396638e-01 4.62281734e-01 8.91809165e-01 -7.08171666e-01
1.17477596e+00 6.38636887e-01 -1.82651274e-03 -3.98962766e-01
1.13967568e-01 3.54655564e-01 4.22976941e-01 -3.87095004e-01
4.60800856e-01 -2.48111397e-01 1.37992963e-01 -1.62303865e-01
1.82455435e-01 -1.00665852e-01 -1.75429389e-01 -5.90684116e-02
8.57699990e-01 -4.46071327e-01 2.27214601e-02 -2.66909480e-01
8.95201266e-01 -7.14279830e-01 5.41720033e-01 6.41217232e-01
-1.03194706e-01 9.83236611e-01 -1.98103234e-01 4.98191774e-01
-6.14231884e-01 -1.19388247e+00 6.26809197e-03 9.74152744e-01
-2.55306929e-01 -2.00705364e-01 -9.34253693e-01 -1.05350651e-01
-5.03634036e-01 1.30177915e+00 1.25246122e-01 2.52739608e-01
-6.14760160e-01 -1.02110612e+00 8.68809640e-01 1.04637705e-01
5.08327127e-01 -4.54449356e-01 2.30170488e-01 1.95449069e-01
-4.73125219e-01 -1.09586549e+00 -8.31074059e-01 2.28006750e-01
-2.85365641e-01 -2.22316161e-01 -9.71736550e-01 -7.44894445e-01
3.73970211e-01 3.55645537e-01 3.59908670e-01 -6.55061543e-01
1.49194136e-01 1.88572675e-01 -5.56593895e-01 -2.17886582e-01
-7.26712167e-01 -2.28977084e-01 2.94052243e-01 6.19281590e-01
5.54436706e-02 -5.39394498e-01 -4.94540185e-01 5.72393477e-01
-6.90656662e-01 -1.95674390e-01 6.85765445e-01 6.65552437e-01
2.31960893e-01 1.90524265e-01 6.18528724e-01 -1.04506709e-01
7.31652677e-01 -2.38379717e-01 -5.26384234e-01 2.78773308e-01
-2.35321373e-01 -1.40082002e-01 3.32248926e-01 -4.71703619e-01
-1.71203232e+00 5.65680414e-02 -9.91141140e-01 -2.29998961e-01
-2.91054070e-01 1.15196109e-01 -1.13008821e+00 3.10805231e-01
4.95891482e-01 4.93609220e-01 -3.78328651e-01 -9.06040430e-01
4.97262031e-01 1.56901753e+00 5.98658741e-01 -1.71106115e-01
6.68748438e-01 -1.15277864e-01 -4.76075083e-01 -1.27144039e+00
-1.14513256e-01 -8.10293376e-01 -2.92986155e-01 -3.81933153e-02
6.50507808e-01 -1.19576168e+00 -6.78101540e-01 7.83259511e-01
-1.36176658e+00 -7.80297443e-02 1.65530160e-01 1.03810418e+00
-2.68595457e-01 4.64313745e-01 -4.40187573e-01 -1.41966343e+00
-5.90515435e-01 -1.41626334e+00 8.48433793e-01 -3.05711888e-02
-1.62681770e-02 -4.10188586e-01 -3.60796563e-02 7.08224058e-01
6.27601802e-01 -5.78988135e-01 3.89093816e-01 -9.47179019e-01
-1.75853431e-01 1.57820895e-01 2.60010540e-01 8.16254854e-01
2.70196766e-01 -3.77154708e-01 -1.48830545e+00 -3.79776239e-01
4.75903064e-01 5.31914055e-01 7.58952916e-01 5.37010014e-01
7.84220815e-01 -4.64941353e-01 -8.58636200e-02 5.94940484e-01
9.71941590e-01 5.59394479e-01 6.60106778e-01 -1.27200261e-01
4.88401830e-01 6.03028119e-01 4.42906201e-01 3.46435398e-01
1.00221239e-01 1.03920174e+00 1.05702415e-01 -9.14144069e-02
-4.49836046e-01 8.08436796e-02 6.97519660e-01 1.46316040e+00
3.09764713e-01 -8.56255651e-01 -7.86523521e-01 4.88637626e-01
-1.42110801e+00 -9.56376314e-01 -4.41113889e-01 2.47912502e+00
6.50517404e-01 -1.36884674e-01 3.43035273e-02 4.44828779e-01
1.09993196e+00 1.68571234e-01 -6.56770542e-02 -5.10795951e-01
-5.02242386e-01 1.24684043e-01 5.18875003e-01 1.02991557e+00
-8.66205215e-01 6.94166780e-01 5.44150591e+00 1.34879446e+00
-1.00538981e+00 4.75567460e-01 4.14176136e-01 -3.71971309e-01
-2.29480758e-01 -4.72120345e-01 -9.99994695e-01 3.36113364e-01
1.14703691e+00 7.70717934e-02 5.43927670e-01 4.00761545e-01
8.32852423e-01 5.05347066e-02 -7.55722702e-01 1.45861232e+00
2.80678123e-01 -4.95608330e-01 -7.19558522e-02 -1.25366107e-01
3.23905498e-01 -7.74811953e-02 1.22359365e-01 1.43932059e-01
4.93518487e-02 -8.49330425e-01 1.01638842e+00 3.24472874e-01
6.46417499e-01 -8.43747914e-01 6.82150483e-01 3.93980026e-01
-1.02849817e+00 -1.58151641e-01 -1.86926182e-02 2.74839520e-01
3.49278450e-01 9.25472319e-01 -1.15540516e+00 9.89814639e-01
3.92136365e-01 -2.89229989e-01 -1.03280611e-01 1.16664445e+00
-2.49574110e-01 9.95069921e-01 -4.33044702e-01 9.49767157e-02
-1.52770668e-01 -6.83496520e-02 1.12983871e+00 1.70687652e+00
6.20765746e-01 4.95125055e-02 -4.10883665e-01 6.85546994e-01
5.61553687e-02 1.67400643e-01 -5.08374833e-02 3.01574439e-01
6.81414604e-01 1.28225541e+00 -2.68620849e-01 -1.62653148e-01
-2.12066963e-01 1.13222051e+00 -3.77518535e-01 6.71566010e-01
-8.53262961e-01 -7.21415222e-01 6.82176054e-01 6.70043156e-02
4.29368049e-01 -3.99122000e-01 -1.65451348e-01 -1.05971992e+00
3.69303785e-02 -1.08654606e+00 -9.52246264e-02 -6.52336299e-01
-8.47455084e-01 9.00324166e-01 -1.28192186e-01 -1.05031884e+00
-1.74235880e-01 -3.17357510e-01 -5.50683320e-01 1.36159515e+00
-1.18696105e+00 -9.16379750e-01 1.85021572e-02 5.15688062e-01
9.43353891e-01 -1.98409483e-01 6.62055075e-01 8.88264954e-01
-7.01364040e-01 1.14530861e+00 4.06829685e-01 5.31930896e-03
7.14514732e-01 -1.02063274e+00 2.78594404e-01 1.46917033e+00
2.27895137e-02 7.47423887e-01 1.11610055e+00 -2.63872534e-01
-1.34372854e+00 -1.11124384e+00 9.04514134e-01 -1.62144229e-02
2.08486184e-01 -9.82619286e-01 -6.77439809e-01 5.17644107e-01
3.91728848e-01 -4.91803408e-01 6.58324838e-01 -6.60410225e-02
-3.16791721e-02 -5.06118655e-01 -1.17847490e+00 5.57603896e-01
9.32914436e-01 -2.78415650e-01 -7.04820633e-01 -6.67141229e-02
8.77447009e-01 -3.92412513e-01 -5.75942695e-01 3.44319046e-01
3.45927119e-01 -7.52413392e-01 8.71852815e-01 2.64633209e-01
-4.34847504e-01 -7.26292729e-01 -8.07296038e-01 -1.60736918e+00
-2.55167037e-01 -1.13742459e+00 8.18671808e-02 1.88696313e+00
6.04026616e-01 -5.63473701e-01 2.11676508e-01 2.05491289e-01
-6.46075666e-01 -1.57258511e-01 -1.15205169e+00 -8.55206490e-01
-3.12592179e-01 -7.10709095e-01 4.66973782e-01 3.72872859e-01
-2.20750108e-01 4.61262196e-01 -5.42632401e-01 7.16692388e-01
5.56756198e-01 -5.18911242e-01 7.21356034e-01 -5.34027934e-01
-8.02708149e-01 -2.77732313e-01 -1.19141592e-02 -1.40049982e+00
-3.03789109e-01 -6.66741192e-01 4.48967636e-01 -1.42753375e+00
-2.92468756e-01 -2.54171714e-02 -2.86150545e-01 -1.66922197e-01
-8.50708112e-02 -3.45235795e-01 2.51576662e-01 -2.12235764e-01
-1.83371350e-01 8.20436299e-01 7.38314390e-01 -1.82708561e-01
-4.35818195e-01 6.73098087e-01 -6.86035693e-01 5.35901308e-01
7.07701087e-01 -3.65638196e-01 -4.06428009e-01 -4.86643821e-01
-4.42613125e-01 2.16179907e-01 -1.57860499e-02 -1.00577474e+00
3.04379523e-01 3.53894502e-01 -1.78087242e-02 -9.15285349e-01
7.21471071e-01 -6.87612653e-01 3.73440325e-01 1.69197097e-01
-3.47722858e-01 -2.97296733e-01 2.38852277e-01 4.25666690e-01
-3.83174539e-01 -8.57148543e-02 9.09393013e-01 1.94216162e-01
-9.04662348e-03 -3.64043504e-01 -7.60927260e-01 -3.80753458e-01
6.01554692e-01 -1.33415610e-01 -1.12892032e-01 -4.94105190e-01
-1.02883482e+00 -8.83719325e-02 -2.87474275e-01 3.34957451e-01
5.78334093e-01 -1.01533270e+00 -1.05781972e+00 1.51353255e-01
-2.15909883e-01 -1.45200402e-01 7.49266803e-01 9.88066137e-01
-1.59734115e-02 5.37264764e-01 3.59792352e-01 -4.00868952e-01
-1.75426495e+00 3.53186488e-01 5.03694654e-01 -8.24286416e-02
-2.63299108e-01 1.13217986e+00 2.05765486e-01 -4.86216784e-01
6.40642762e-01 -4.15914446e-01 -1.50930882e-01 -1.41219094e-01
6.25006258e-01 6.91603959e-01 6.77188814e-01 -7.75281310e-01
-5.06712794e-01 1.44687772e-01 2.35276267e-01 -7.80310631e-01
1.12476110e+00 -6.12270772e-01 1.40586004e-01 3.11489433e-01
1.00053084e+00 7.78891087e-01 -7.50579000e-01 -1.14010207e-01
-2.85532922e-01 -2.23893151e-01 3.10607284e-01 -1.06715739e+00
-7.76377380e-01 1.01230073e+00 8.22132528e-01 -1.08555757e-01
1.29067934e+00 -2.76809037e-01 4.12772655e-01 3.04425806e-01
6.99603334e-02 -9.93408084e-01 -3.71665835e-01 4.87930119e-01
1.18212175e+00 -6.00872040e-01 -4.21932966e-01 -5.11668742e-01
-5.95684171e-01 8.26196253e-01 2.80184835e-01 5.20852625e-01
5.66093385e-01 4.73307431e-01 1.19602256e-01 3.88418049e-01
-4.53817844e-01 -2.16914028e-01 3.30155194e-01 6.13975227e-01
4.53657597e-01 1.50566831e-01 -3.40076745e-01 1.24764359e+00
-4.15162206e-01 -5.08247018e-01 2.72837728e-01 3.66077662e-01
-6.57976389e-01 -9.72197175e-01 -9.89502311e-01 -9.97449309e-02
-5.39000750e-01 -4.30847943e-01 -1.25893056e-01 4.32431921e-02
8.45996849e-03 1.88812363e+00 -1.72067553e-01 -6.12572074e-01
6.78421319e-01 2.22237632e-01 2.59637982e-01 -5.10826170e-01
-7.86751568e-01 8.84767413e-01 3.05672854e-01 -8.41138363e-02
-7.49260560e-02 -6.09854221e-01 -1.02299953e+00 -4.88174222e-02
-5.24973989e-01 3.10895711e-01 1.14569557e+00 8.26931059e-01
4.58245903e-01 1.00829947e+00 7.98333943e-01 -6.14038527e-01
-8.00957382e-01 -1.59583068e+00 -6.74680591e-01 -8.27841386e-02
3.51486623e-01 -2.98122883e-01 -6.74770534e-01 2.91076861e-03]
|
[14.835280418395996, 5.961994647979736]
|
d2608d98-d97f-4f14-98c1-c9afe8912862
|
a-metaheuristic-multi-objective-interaction
|
2211.05423
| null |
https://arxiv.org/abs/2211.05423v1
|
https://arxiv.org/pdf/2211.05423v1.pdf
|
A metaheuristic multi-objective interaction-aware feature selection method
|
Multi-objective feature selection is one of the most significant issues in the field of pattern recognition. It is challenging because it maximizes the classification performance and, at the same time, minimizes the number of selected features, and the mentioned two objectives are usually conflicting. To achieve a better Pareto optimal solution, metaheuristic optimization methods are widely used in many studies. However, the main drawback is the exploration of a large search space. Another problem with multi-objective feature selection approaches is the interaction between features. Selecting correlated features has negative effect on classification performance. To tackle these problems, we present a novel multi-objective feature selection method that has several advantages. Firstly, it considers the interaction between features using an advanced probability scheme. Secondly, it is based on the Pareto Archived Evolution Strategy (PAES) method that has several advantages such as simplicity and its speed in exploring the solution space. However, we improve the structure of PAES in such a way that generates the offsprings, intelligently. Thus, the proposed method utilizes the introduced probability scheme to produce more promising offsprings. Finally, it is equipped with a novel strategy that guides it to find the optimum number of features through the process of evolution. The experimental results show a significant improvement in finding the optimal Pareto front compared to state-of-the-art methods on different real-world datasets.
|
['Mostafa Sabzekar', 'Modjtaba Rouhani', 'Motahare Namakin']
|
2022-11-10
| null | null | null | null |
['metaheuristic-optimization']
|
['methodology']
|
[ 2.59052664e-01 -6.74580753e-01 1.52985200e-01 1.24652851e-02
-1.10646345e-01 4.27736305e-02 9.93724167e-03 3.56378287e-01
-5.22537589e-01 9.61871207e-01 -3.53890419e-01 4.11086559e-01
-8.93119454e-01 -1.10449541e+00 -4.09596376e-02 -1.10979176e+00
-2.40589846e-02 4.63338763e-01 2.81289726e-01 -3.48536730e-01
7.66944170e-01 5.78990579e-01 -2.27663541e+00 -2.26743177e-01
1.36424661e+00 1.01503372e+00 4.81247842e-01 2.26549264e-02
-3.40901762e-01 3.02539021e-02 -7.28270113e-01 -3.04187983e-01
-6.47105575e-02 -5.20546854e-01 -4.31546450e-01 1.27769917e-01
-7.88931191e-01 5.01206458e-01 4.94734496e-01 9.98845935e-01
5.83075404e-01 3.05967718e-01 6.67467356e-01 -1.38408172e+00
-4.25805867e-01 2.81908572e-01 -7.85328567e-01 3.20708334e-01
1.31932735e-01 -6.61304891e-02 8.94852340e-01 -5.21949768e-01
3.27883124e-01 1.19060969e+00 2.58341014e-01 1.54013693e-01
-7.91039765e-01 -5.74701548e-01 -4.03506905e-02 5.98034501e-01
-1.58093905e+00 2.47658119e-02 8.96604300e-01 -1.66098192e-01
5.71961522e-01 6.01828039e-01 1.02657080e+00 2.85396814e-01
5.29087722e-01 6.62382782e-01 9.58706677e-01 -6.13007426e-01
4.26173449e-01 2.92858362e-01 1.12283863e-01 6.18901670e-01
5.60758948e-01 8.87091830e-02 -2.10957870e-01 -3.25245142e-01
5.74374273e-02 5.04948460e-02 -4.73058313e-01 -3.67245793e-01
-7.67411888e-01 9.60435510e-01 1.83227897e-01 7.68241048e-01
-6.06649220e-01 -4.62534338e-01 1.48431644e-01 -5.84158488e-03
2.71338467e-02 7.79499352e-01 -3.62857580e-01 -2.44733974e-01
-6.16130173e-01 1.50418580e-01 5.29906988e-01 1.38027042e-01
7.74239361e-01 -3.21043096e-02 3.73514332e-02 9.08452988e-01
4.15727228e-01 2.90532678e-01 8.01615238e-01 -2.09373862e-01
2.67623514e-01 1.20369053e+00 5.79564683e-02 -1.43475628e+00
-4.68120188e-01 -9.16498005e-01 -6.90008521e-01 4.38140154e-01
-1.05779789e-01 -1.93574309e-01 -4.64995325e-01 1.40220821e+00
4.54938263e-01 -1.55530676e-01 3.12291384e-01 7.58081138e-01
6.47045612e-01 7.09767640e-01 -1.07604405e-02 -7.22200155e-01
1.20249426e+00 -8.04847360e-01 -8.62733126e-01 4.02275287e-02
1.17252082e-01 -8.39975834e-01 6.56987190e-01 5.21626234e-01
-8.39332223e-01 -4.98318791e-01 -1.14420629e+00 8.04850161e-01
-4.53847438e-01 2.12736160e-01 6.46666825e-01 7.77869165e-01
-5.87801218e-01 5.29969811e-01 -5.43052256e-01 -3.00967962e-01
1.33398041e-01 6.55888498e-01 -1.08183905e-01 2.02500835e-01
-1.12524211e+00 1.00018120e+00 9.03399944e-01 2.48231754e-01
8.81807879e-02 -1.58820018e-01 -5.59668064e-01 3.65732580e-01
5.09743989e-01 -5.59525669e-01 4.53162760e-01 -1.22136009e+00
-1.48623562e+00 9.51523185e-02 -1.93673655e-01 8.29733014e-02
4.58788902e-01 1.16046980e-01 -6.84556663e-01 -6.62364066e-02
-1.45898595e-01 1.49461776e-01 6.81536257e-01 -1.07994390e+00
-1.02207744e+00 -3.67719203e-01 -3.55286241e-01 4.45322812e-01
-9.17699814e-01 9.93903354e-02 -2.94016510e-01 -4.07050312e-01
2.28092680e-03 -7.70274103e-01 -3.71136606e-01 -5.56672990e-01
-1.44610158e-03 -2.61814564e-01 8.19166005e-01 -1.56817406e-01
1.45475519e+00 -2.10468340e+00 5.20733416e-01 6.17753386e-01
-2.60594845e-01 5.62301755e-01 -2.30017584e-02 4.52844441e-01
2.77704418e-01 7.52166137e-02 -3.28491807e-01 1.04763851e-01
-3.63014936e-01 9.59924236e-02 4.47190225e-01 2.24927261e-01
2.72332132e-01 3.94991755e-01 -7.75667906e-01 -8.00877869e-01
2.41272196e-01 5.13683617e-01 -5.82140446e-01 2.12005619e-02
7.72419944e-02 3.29684764e-01 -9.01449978e-01 7.56027341e-01
8.44029963e-01 8.36646743e-03 1.00453556e-01 -1.13132007e-01
-6.22667551e-01 -6.45630240e-01 -1.68226683e+00 1.02122033e+00
-2.61168718e-01 7.00576603e-02 -1.62400931e-01 -9.97659028e-01
1.27228022e+00 2.38138482e-01 7.00832784e-01 -5.51156223e-01
6.83287144e-01 5.73449552e-01 2.55280286e-01 -5.95408380e-01
4.99814719e-01 -5.97091578e-02 3.63458395e-01 2.50164658e-01
-1.84247062e-01 1.05146773e-01 5.83467484e-01 -6.03734076e-01
6.33038938e-01 -1.18224509e-01 6.86825812e-01 -1.67351916e-01
1.24278545e+00 -8.11208114e-02 9.92754996e-01 2.17058823e-01
9.24944282e-02 1.48785397e-01 1.25684232e-01 -3.27128768e-01
-4.64713424e-01 -3.95788223e-01 -2.74573058e-01 3.42094690e-01
4.13669795e-01 -1.51051074e-01 -3.03887546e-01 -5.26493192e-01
-1.70005914e-02 7.46767879e-01 -4.20864612e-01 -3.79219294e-01
-4.26598489e-01 -1.26560473e+00 7.33946189e-02 -3.98814119e-02
6.99357033e-01 -1.25058842e+00 -1.12391770e+00 4.65403169e-01
1.07119363e-02 -3.19041967e-01 1.56621471e-01 1.53915621e-02
-9.20160055e-01 -1.02662432e+00 -7.27288663e-01 -7.39777088e-01
7.42302597e-01 2.72057831e-01 6.73285425e-01 6.72952890e-01
-3.85039270e-01 -1.46105379e-01 -7.94243991e-01 -6.67592049e-01
-3.00290346e-01 2.21465304e-01 -1.99737608e-01 3.34050059e-01
3.27446640e-01 -5.33899486e-01 -2.46159270e-01 4.25627828e-01
-9.76513803e-01 -3.67064327e-01 9.41178501e-01 1.05698419e+00
5.67420304e-01 1.16257942e+00 7.45293200e-01 -3.25449765e-01
7.36259222e-01 -4.63359296e-01 -8.27397823e-01 5.77866375e-01
-7.97992051e-01 2.37859666e-01 9.41931784e-01 -2.85747558e-01
-1.14286578e+00 -5.80723137e-02 -1.06017813e-01 -3.96496914e-02
1.26689315e-01 6.59690499e-01 -4.18174326e-01 -4.22662109e-01
1.51547536e-01 5.56386292e-01 2.22098961e-01 -4.19419706e-01
-3.50338995e-01 6.78314090e-01 -1.65437251e-01 -3.48765820e-01
6.20959580e-01 1.26680762e-01 5.24565101e-01 -6.71241105e-01
-1.16594173e-01 -3.45080256e-01 -3.74905348e-01 -2.71890581e-01
5.62511384e-01 -2.26665344e-02 -1.05406463e+00 5.05169392e-01
-8.79459500e-01 8.54200244e-01 7.38502070e-02 7.01796830e-01
-2.70687401e-01 4.03617084e-01 2.73918539e-01 -1.24985814e+00
-5.77167273e-01 -1.49075568e+00 3.86808842e-01 8.62189591e-01
-6.81421310e-02 -6.74625695e-01 -1.33411154e-01 -2.54018530e-02
3.58800024e-01 4.25108075e-01 1.03230131e+00 -5.34440458e-01
-3.10441613e-01 -1.20475993e-01 2.29025379e-01 7.83658102e-02
3.64625692e-01 3.59932452e-01 -1.29390374e-01 -2.88750589e-01
1.32897168e-01 3.08210645e-02 4.51970816e-01 2.71975249e-01
9.10531521e-01 -7.77735114e-02 -4.66222793e-01 4.74837244e-01
1.73081219e+00 9.96816635e-01 5.70570886e-01 9.35807168e-01
3.22203711e-02 6.85056329e-01 1.36787701e+00 8.77979279e-01
-4.64072004e-02 6.59441411e-01 6.51011884e-01 2.30945423e-02
3.95738035e-01 3.56853753e-01 -7.59958029e-02 5.91050625e-01
-2.12958112e-01 -4.65028673e-01 -5.92127502e-01 3.69162500e-01
-1.94482398e+00 -9.60281610e-01 -1.00655705e-01 2.25843644e+00
5.03118336e-01 -4.76469845e-02 8.74316692e-02 7.27673709e-01
1.02268481e+00 -2.38594368e-01 -3.74062985e-01 -5.60494661e-01
-3.00224662e-01 1.88616335e-01 9.76900458e-02 1.68349922e-01
-9.33394790e-01 3.77522528e-01 4.69317389e+00 8.91962945e-01
-1.32742369e+00 -2.72463381e-01 2.13997871e-01 -9.13233757e-02
-2.37929881e-01 -7.99153298e-02 -7.56454349e-01 9.18391764e-01
2.88818598e-01 -4.44795728e-01 4.28060412e-01 5.38748264e-01
1.22337259e-01 -3.56073499e-01 -3.54984403e-01 9.52987671e-01
1.16646945e-01 -8.95925999e-01 1.15263760e-02 1.40489385e-01
7.92495012e-01 -7.00814307e-01 7.86663815e-02 -1.09159306e-01
-2.47694805e-01 -7.34793961e-01 5.82898796e-01 6.16834164e-01
1.18020564e-01 -1.60166132e+00 1.16612220e+00 4.50369567e-01
-1.21344483e+00 -5.10466993e-01 -4.14562017e-01 2.08539650e-01
2.77255028e-01 7.33476281e-01 -4.37668234e-01 1.17054439e+00
6.76779628e-01 3.61301422e-01 -6.31345451e-01 1.78228915e+00
6.35689124e-03 -4.47158664e-02 -3.61789137e-01 -7.11546421e-01
2.47895747e-01 -5.39514363e-01 9.95881915e-01 5.47191978e-01
7.99576581e-01 7.10795224e-02 3.25325392e-02 5.84869921e-01
4.83362079e-01 7.84254670e-01 -2.90395975e-01 -3.52587700e-02
5.13544023e-01 1.23389053e+00 -9.16936874e-01 9.12883431e-02
-2.96295993e-02 6.15598321e-01 6.50750250e-02 -1.73730507e-01
-7.77711391e-01 -8.49736154e-01 1.71423405e-01 -3.96514893e-01
4.43594664e-01 1.73898563e-01 -1.14496648e-01 -6.72536492e-01
7.86955208e-02 -7.94110835e-01 4.79287326e-01 -1.60535783e-01
-9.54885364e-01 9.04135048e-01 -2.10919321e-01 -1.36150789e+00
-9.46496427e-02 -3.05313915e-01 -5.81449270e-01 9.47851300e-01
-1.66693711e+00 -7.80312240e-01 -3.14429522e-01 4.41560775e-01
5.49278080e-01 -4.66089010e-01 8.16828012e-01 3.50664467e-01
-7.76842713e-01 4.52464163e-01 4.82936174e-01 -7.11289883e-01
3.16156089e-01 -7.43765891e-01 -5.56631804e-01 7.60253787e-01
-4.27470326e-01 4.50813413e-01 7.66496420e-01 -6.52604580e-01
-1.37771761e+00 -6.20932341e-01 8.06032956e-01 4.96052027e-01
2.06757531e-01 4.55274343e-01 -5.74654102e-01 -1.91007257e-01
4.54289876e-02 -4.67596322e-01 9.12296653e-01 -7.16710016e-02
7.62938917e-01 -1.50662392e-01 -1.42789328e+00 5.37475467e-01
5.25348663e-01 7.09340215e-01 -4.85957414e-01 -1.30391717e-01
5.58238290e-02 -2.95453425e-02 -7.56713986e-01 7.12421238e-01
6.98559940e-01 -1.12131858e+00 6.61071479e-01 -1.17976859e-01
1.87149927e-01 -7.01577842e-01 8.18416327e-02 -1.59652424e+00
-5.84396362e-01 -3.61096561e-01 1.21128365e-01 1.54363656e+00
2.46482730e-01 -1.13281095e+00 5.67001939e-01 2.63039380e-01
8.72760192e-02 -1.22863102e+00 -8.69240165e-01 -8.89804065e-01
-5.56761146e-01 4.37720299e-01 1.09635770e+00 5.53277850e-01
-2.71366745e-01 5.19171245e-02 -2.26034284e-01 -5.54232448e-02
5.83597302e-01 4.45143610e-01 3.46306652e-01 -1.66956043e+00
-2.99338609e-01 -6.23931110e-01 -5.69575131e-01 -4.00258899e-02
-5.77186001e-03 -5.15682757e-01 -1.63728192e-01 -1.38566816e+00
1.88627154e-01 -7.53102005e-01 -5.03742635e-01 1.67957187e-01
-5.17639279e-01 -9.24792662e-02 3.55396032e-01 1.88274264e-01
-3.43890190e-01 8.84988248e-01 1.22659993e+00 3.45112570e-03
-5.74165821e-01 3.75055373e-01 -7.92311251e-01 5.53342044e-01
8.62510026e-01 -6.10368371e-01 -3.74081373e-01 -1.18402675e-01
4.42159250e-02 -9.24103893e-03 -2.80082136e-01 -1.13958728e+00
1.15135707e-01 -4.10749942e-01 5.02139628e-01 -7.03262985e-01
3.32217544e-01 -1.29583073e+00 5.18738687e-01 8.19002032e-01
1.77975848e-01 2.88946986e-01 2.21765041e-01 3.48108649e-01
-4.52727199e-01 -7.84030795e-01 8.77202272e-01 -1.60622641e-01
-8.45698178e-01 -9.94344279e-02 -3.80529940e-01 -4.77601886e-01
1.59479296e+00 -6.63827658e-01 3.00151538e-02 2.00705037e-01
-2.15044528e-01 5.06777704e-01 5.02464056e-01 4.45519686e-01
5.91581464e-01 -1.16989124e+00 -5.96729815e-01 8.57436508e-02
-4.73997220e-02 -3.81933987e-01 3.67783338e-01 7.82095194e-01
-4.78607297e-01 2.76762992e-01 -6.23179257e-01 -4.02561873e-01
-1.64691162e+00 4.77081805e-01 -1.70155969e-02 -4.26113367e-01
-1.18460432e-01 5.87126911e-01 -4.98114645e-01 -3.04644667e-02
-2.42186841e-02 4.20210391e-01 -1.09332693e+00 4.32839245e-01
4.45055693e-01 8.55737627e-01 1.91028640e-01 -7.29914665e-01
-6.92763746e-01 9.86565888e-01 2.13045135e-01 4.46544029e-02
1.59153843e+00 3.40275057e-02 -5.32918632e-01 -4.16984148e-02
1.05561805e+00 3.43604445e-01 -5.76994896e-01 3.01582783e-01
-1.49503469e-01 -8.90549898e-01 5.77285923e-02 -1.01534045e+00
-1.18185973e+00 3.50302428e-01 6.48944557e-01 1.93174466e-01
1.72208405e+00 -6.82433724e-01 7.32015133e-01 1.63601115e-01
4.47541475e-01 -1.03517258e+00 -9.89130735e-02 2.93937534e-01
8.14127564e-01 -8.64240885e-01 1.46104530e-01 -5.62428474e-01
-6.14554107e-01 1.44459713e+00 6.94453955e-01 2.23823935e-01
4.98029947e-01 2.45562177e-02 -1.48072049e-01 -3.92856300e-02
-4.40287411e-01 -2.96968222e-01 2.56592661e-01 2.79965460e-01
2.32873201e-01 -1.59659192e-01 -1.40455723e+00 5.97929239e-01
-1.79004639e-01 5.82974590e-02 2.12954104e-01 1.19612181e+00
-6.80924535e-01 -1.58360696e+00 -9.60930645e-01 3.46896350e-01
-4.82739389e-01 1.88139051e-01 -2.03209937e-01 7.12412953e-01
5.48792899e-01 1.35755980e+00 -2.92854458e-01 -4.20974582e-01
2.75670052e-01 -2.08150387e-01 2.96362936e-01 -1.47569492e-01
-8.10624540e-01 -6.15272187e-02 -2.67418295e-01 1.04286922e-02
-5.01501203e-01 -7.08976924e-01 -1.32939482e+00 -4.59496565e-02
-9.14530277e-01 9.20783699e-01 9.00069714e-01 8.79962742e-01
4.76523459e-01 6.41745210e-01 1.09060979e+00 -6.02514386e-01
-3.55350763e-01 -5.79784751e-01 -4.87619728e-01 1.16976589e-01
-1.56880349e-01 -1.28281498e+00 -2.34520853e-01 -7.51316428e-01]
|
[5.749634265899658, 3.485461950302124]
|
3ad92053-e4bc-4b84-be68-f2d4c22b6554
|
probability-map-guided-bi-directional
|
1903.00923
| null |
https://arxiv.org/abs/1903.00923v6
|
https://arxiv.org/pdf/1903.00923v6.pdf
|
Pancreas segmentation with probabilistic map guided bi-directional recurrent UNet
|
Pancreas segmentation in medical imaging data is of great significance for clinical pancreas diagnostics and treatment. However, the large population variations in the pancreas shape and volume cause enormous segmentation difficulties, even for state-of-the-art algorithms utilizing fully-convolutional neural networks (FCNs). Specifically, pancreas segmentation suffers from the loss of spatial information in 2D methods, and the high computational cost of 3D methods. To alleviate these problems, we propose a probabilistic-map-guided bi-directional recurrent UNet (PBR-UNet) architecture, which fuses intra-slice information and inter-slice probabilistic maps into a local 3D hybrid regularization scheme, which is followed by bi-directional recurrent network optimization. The PBR-UNet method consists of an initial estimation module for efficiently extracting pixel-level probabilistic maps and a primary segmentation module for propagating hybrid information through a 2.5D U-Net architecture. Specifically, local 3D information is inferred by combining an input image with the probabilistic maps of the adjacent slices into multichannel hybrid data, and then hierarchically aggregating the hybrid information of the entire segmentation network. Besides, a bi-directional recurrent optimization mechanism is developed to update the hybrid information in both the forward and the backward directions. This allows the proposed network to make full and optimal use of the local context information. Quantitative and qualitative evaluation was performed on the NIH Pancreas-CT dataset, and our proposed PBR-UNet method achieved better segmentation results with less computational cost compared to other state-of-the-art methods.
|
['Xiaozhu Lin', 'Xiaohua Qian', 'Jun Li', 'Hui Che', 'Hao Li']
|
2019-03-03
| null | null | null | null |
['pancreas-segmentation']
|
['medical']
|
[-2.60021221e-02 5.88663481e-02 -1.12061255e-01 -2.56304353e-01
-9.01462078e-01 -3.41941416e-01 9.10990406e-03 2.31447458e-01
-3.95346522e-01 5.23204029e-01 2.44750783e-01 -2.18277857e-01
-3.37805957e-01 -8.57331991e-01 -7.29627609e-01 -1.09664738e+00
-3.09448957e-01 5.07724822e-01 5.49364090e-01 3.64316434e-01
-2.40947485e-01 5.59175372e-01 -6.46731794e-01 2.39060864e-01
1.15267003e+00 1.07956374e+00 3.08442712e-01 4.08086121e-01
-4.64870065e-01 4.99948025e-01 8.01399797e-02 1.29272416e-01
2.14047462e-01 -5.54972768e-01 -3.46220434e-01 1.28601402e-01
-2.46260483e-02 -2.89098024e-01 -5.33362627e-01 1.19577110e+00
6.05397403e-01 -4.71723936e-02 5.47551632e-01 -7.49715388e-01
-4.32637125e-01 9.72342312e-01 -4.70254421e-01 1.41541302e-01
-1.24151662e-01 4.19850320e-01 2.15069190e-01 -3.83098781e-01
5.30562401e-01 8.97541881e-01 8.68961990e-01 2.12674022e-01
-1.27661133e+00 -5.40977776e-01 1.65795982e-01 -2.27919117e-01
-1.14472008e+00 3.53562683e-01 9.59871054e-01 -5.30374229e-01
5.07323563e-01 -7.26080909e-02 9.87532258e-01 6.09752297e-01
4.88016039e-01 8.83493423e-01 1.06822002e+00 -7.46623501e-02
-8.53969343e-03 -7.37763569e-02 3.01695377e-01 8.29313517e-01
2.40870997e-01 2.20432371e-01 4.55069682e-03 4.40880880e-02
1.26493645e+00 1.39675766e-01 -5.76656520e-01 -6.46700561e-01
-1.58940911e+00 6.79204166e-01 9.93630588e-01 4.80982780e-01
-8.51495445e-01 -1.01108104e-01 3.35613042e-01 -3.33764911e-01
1.83452219e-01 1.04259685e-01 -4.64913160e-01 2.10605457e-01
-1.06670499e+00 -2.13772297e-01 8.52216125e-01 8.69427085e-01
4.01698530e-01 -9.39438567e-02 -3.26163143e-01 6.63145602e-01
9.01431143e-01 2.79326916e-01 6.50156140e-01 -8.26490402e-01
2.16592446e-01 7.81889141e-01 -3.35719109e-01 -6.83054984e-01
-8.48770976e-01 -6.65621161e-01 -1.31527007e+00 2.10671172e-01
5.75110734e-01 -3.02285820e-01 -1.31561136e+00 1.55632484e+00
4.45096433e-01 1.82782307e-01 -9.44653898e-03 1.29908788e+00
1.14269340e+00 6.63803697e-01 3.15763384e-01 -3.24411750e-01
1.34367108e+00 -9.71362233e-01 -6.00714386e-01 2.32968926e-01
5.48920155e-01 -5.75511992e-01 5.39871216e-01 1.76542252e-01
-1.03554380e+00 -1.47123396e-01 -9.72357929e-01 2.41187826e-01
-1.23396665e-02 1.03100605e-01 4.76144224e-01 4.46052045e-01
-9.40327227e-01 4.47453350e-01 -1.23282576e+00 -2.01556325e-01
6.21391118e-01 3.62055510e-01 -1.62397221e-01 -2.80852973e-01
-9.72240925e-01 7.54291892e-01 5.43571413e-01 4.22625273e-01
-6.24775529e-01 -9.15798008e-01 -1.04161227e+00 2.10158745e-04
2.01764897e-01 -7.56575227e-01 1.00079024e+00 -5.90398848e-01
-1.75166416e+00 4.18994367e-01 6.89791441e-02 -3.63728195e-01
7.88566887e-01 2.60899097e-01 -1.22731537e-01 3.51344377e-01
1.11722071e-02 7.00718403e-01 1.62147686e-01 -1.20711410e+00
-3.30080867e-01 -4.83066261e-01 -4.20676440e-01 2.68645436e-01
2.41058394e-01 -3.31232429e-01 -7.73178816e-01 -8.05956066e-01
9.07918632e-01 -8.76841009e-01 -7.01345325e-01 3.25991720e-01
-4.32810128e-01 5.03969416e-02 7.11601675e-01 -8.94671857e-01
9.18602705e-01 -2.03970480e+00 1.90539002e-01 3.09957027e-01
2.33919621e-01 3.78595144e-02 1.54645309e-01 -5.29035866e-01
-2.36407872e-02 -1.93672836e-01 -6.94226921e-01 6.24642000e-02
-1.41636193e-01 2.51961261e-01 4.28827256e-01 5.89895010e-01
-8.71736407e-02 1.07512653e+00 -9.56354916e-01 -8.76007557e-01
5.50843298e-01 7.36282289e-01 -5.59032857e-01 2.00144678e-01
-1.96530282e-01 8.39407623e-01 -5.84525585e-01 7.28027046e-01
8.22635353e-01 -4.48370636e-01 2.62751222e-01 -6.88102543e-01
-4.37649280e-01 8.70959014e-02 -1.02377582e+00 1.99384952e+00
-1.63954303e-01 1.31543085e-01 2.89973676e-01 -8.83738458e-01
6.57286704e-01 5.06483674e-01 1.12002885e+00 -7.34301209e-01
2.12645307e-01 4.53338325e-01 -3.72794345e-02 -6.02469146e-01
-1.26980186e-01 -3.38459134e-01 -2.55683493e-02 2.55710632e-01
1.11413583e-01 1.22711658e-02 1.67576164e-01 -8.09035525e-02
7.03795135e-01 4.07002211e-01 1.71527695e-02 -5.02481937e-01
4.54870015e-01 1.70106187e-01 8.91594470e-01 5.54228365e-01
-4.61987883e-01 7.61846781e-01 5.22111297e-01 -4.10767972e-01
-7.46760726e-01 -1.33783090e+00 -4.44104970e-01 3.41679245e-01
5.47239542e-01 2.26818457e-01 -6.33077681e-01 -8.80990803e-01
1.08096637e-01 5.96294582e-01 -4.14252043e-01 6.89107180e-02
-8.24288547e-01 -1.26288974e+00 4.09320235e-01 6.34197772e-01
7.64316380e-01 -9.52843785e-01 -4.70903695e-01 6.80060923e-01
-2.41437703e-01 -7.86377549e-01 -7.15457141e-01 5.21784961e-01
-1.44577312e+00 -9.15792048e-01 -1.13487649e+00 -8.91897976e-01
8.18724692e-01 -1.23270929e-01 9.26368177e-01 -3.16559911e-01
-2.35679209e-01 3.64497514e-03 -6.97833486e-03 2.44409680e-01
-3.78111601e-01 -8.17746744e-02 -4.27906275e-01 -3.33483696e-01
9.17598903e-02 -6.75423384e-01 -8.24449718e-01 3.73402804e-01
-9.24798548e-01 1.65870368e-01 9.37399507e-01 9.20538247e-01
8.85126948e-01 -6.68964237e-02 4.18187648e-01 -6.39540374e-01
1.61663190e-01 -5.24847746e-01 -8.42857897e-01 2.38703862e-01
-4.94306237e-01 1.11003265e-01 2.50322342e-01 -4.16010261e-01
-1.10467803e+00 3.62523854e-01 -2.89669573e-01 -4.70263213e-01
-2.36714229e-01 7.37448215e-01 -1.14444584e-01 -2.35518143e-01
3.12022597e-01 4.67718363e-01 1.37787357e-01 -4.09722716e-01
2.28793502e-01 2.66851366e-01 5.28537095e-01 -4.49520946e-01
1.55294955e-01 2.47813672e-01 -1.30846128e-02 -5.13728738e-01
-3.88520032e-01 -3.37117851e-01 -8.65951955e-01 -3.21132958e-01
1.26590323e+00 -6.24922633e-01 -6.74719870e-01 7.03557193e-01
-1.02090859e+00 -4.06064481e-01 -3.43435466e-01 1.12885857e+00
-3.19763899e-01 4.90086704e-01 -1.03818369e+00 -2.74339825e-01
-4.25690860e-01 -1.88732016e+00 6.30256355e-01 5.34130573e-01
2.52127051e-01 -1.09550714e+00 -1.35713086e-01 1.53172910e-01
5.13217151e-01 2.85350978e-01 1.05421638e+00 -4.85139638e-01
-1.03227293e+00 9.19958390e-03 -7.07685351e-01 2.18824908e-01
1.27446085e-01 -3.63460422e-01 -4.18617606e-01 1.25549454e-02
7.12811365e-04 1.69809610e-01 8.56725037e-01 1.27506709e+00
1.04578352e+00 -1.13146797e-01 -3.99852306e-01 9.53498244e-01
1.57171452e+00 3.87988508e-01 2.84475327e-01 1.43846959e-01
8.92956913e-01 7.84879699e-02 2.41894685e-02 2.10733682e-01
5.17560720e-01 3.90207499e-01 3.88757646e-01 -3.23954880e-01
-3.61986995e-01 -1.83792934e-02 7.55028054e-03 1.05968595e+00
1.40821651e-01 -6.39117658e-02 -9.38198030e-01 5.62663496e-01
-1.85323167e+00 -5.20637393e-01 -1.56128228e-01 2.12432265e+00
9.43384528e-01 8.01678095e-03 -1.62392348e-01 -4.63125557e-01
8.67152512e-01 -1.19925113e-02 -6.65480256e-01 2.07515195e-01
-3.90565884e-03 -2.86888868e-01 7.55879581e-01 4.21833158e-01
-1.30502391e+00 3.27246755e-01 5.56572723e+00 5.17841935e-01
-1.27116776e+00 1.01919778e-01 6.14834011e-01 1.50962800e-01
-2.80285001e-01 -2.89316475e-01 -4.82251346e-01 5.80543756e-01
3.68427753e-01 2.12778792e-01 1.41726851e-01 6.03869379e-01
1.83107570e-01 -3.15582931e-01 -8.42265666e-01 8.09595048e-01
-5.59372231e-02 -1.37094283e+00 -1.46172028e-02 -6.43878952e-02
6.68059647e-01 5.14911115e-01 -2.36909613e-01 1.40577793e-01
2.99911290e-01 -7.67305791e-01 4.25276011e-01 7.09885359e-01
6.07032597e-01 -5.51604152e-01 1.06050014e+00 2.87872165e-01
-1.23505867e+00 1.79793447e-01 7.05066621e-02 7.34645665e-01
6.08591020e-01 1.03536904e+00 -6.57952666e-01 7.55788863e-01
7.28818774e-01 6.84903383e-01 9.14056436e-04 1.62825644e+00
-2.11968660e-01 4.12453949e-01 -8.74013782e-01 1.74089313e-01
4.87060636e-01 -6.29811049e-01 8.76826465e-01 1.45365489e+00
5.21364927e-01 2.42051244e-01 4.13203716e-01 1.15916502e+00
-1.39527336e-01 -7.27622351e-03 -1.28161818e-01 2.23625839e-01
-6.23915084e-02 1.25541723e+00 -1.16013110e+00 -4.76005077e-01
-3.72739941e-01 6.69165015e-01 -1.56456560e-01 4.20270056e-01
-8.05906892e-01 -1.02228910e-01 1.37768835e-01 -1.13675952e-01
2.12061584e-01 -1.59459785e-01 -6.67901754e-01 -1.26678872e+00
-9.76763591e-02 -2.34376505e-01 6.15376949e-01 -5.43862224e-01
-1.43002093e+00 5.76312006e-01 -1.89873889e-01 -1.21819305e+00
1.02156170e-01 -5.12430727e-01 -5.47721982e-01 1.07270277e+00
-1.73708224e+00 -1.04243135e+00 -4.92707282e-01 4.40597326e-01
3.86543661e-01 3.90876859e-01 5.99624276e-01 5.36451042e-01
-5.08608341e-01 3.26988161e-01 1.12999000e-01 3.28255564e-01
4.67130899e-01 -1.30346787e+00 -3.11171204e-01 7.82699168e-01
-5.33220768e-01 2.51554847e-01 2.07060039e-01 -9.13641691e-01
-1.23476553e+00 -1.32181692e+00 4.70554471e-01 1.75485075e-01
4.86087680e-01 3.05345356e-01 -1.02301109e+00 6.75308347e-01
3.34642194e-02 4.75605339e-01 4.93984044e-01 -4.79119748e-01
1.60968304e-01 4.89053130e-02 -1.43479228e+00 4.76096541e-01
6.03142083e-01 -3.35945636e-02 -6.04619324e-01 1.97168201e-01
7.95651793e-01 -8.64567697e-01 -1.27952564e+00 8.22457135e-01
6.68485880e-01 -7.58231640e-01 9.86190140e-01 1.52202800e-01
2.92285651e-01 -4.69146937e-01 -2.80092340e-02 -1.27855933e+00
-3.88515890e-01 -2.98883498e-01 1.34825975e-01 7.07999349e-01
5.14330685e-01 -6.87749505e-01 8.08846831e-01 6.51591897e-01
-7.21183538e-01 -8.29789162e-01 -1.06353545e+00 -2.95304120e-01
1.78067267e-01 -3.87057066e-01 2.09426150e-01 9.28180754e-01
-1.87111180e-02 -2.91489780e-01 2.91040719e-01 5.25772512e-01
1.13221395e+00 2.95195580e-01 -3.37206312e-02 -9.95755315e-01
-6.24387339e-02 -7.58534908e-01 -2.58652896e-01 -1.20592070e+00
-2.80479491e-01 -1.18546855e+00 3.68462205e-01 -1.93064964e+00
2.17936948e-01 -6.22361541e-01 -5.97056150e-01 3.09876174e-01
-1.16623253e-01 2.26158559e-01 1.05384074e-01 2.43847147e-01
-4.06372011e-01 5.46146870e-01 1.87395513e+00 -1.72916159e-01
-5.52029848e-01 8.28356743e-02 -2.29367197e-01 7.69517004e-01
5.12751937e-01 -5.30568302e-01 -1.14692627e-02 -3.26786727e-01
-4.02098209e-01 6.32728815e-01 5.56270540e-01 -9.97286856e-01
5.98033309e-01 1.38467073e-01 7.78743446e-01 -9.98984039e-01
-7.07521141e-02 -1.04340208e+00 1.88809320e-01 6.36754692e-01
-1.60966739e-01 -3.20495605e-01 2.73109615e-01 3.89928162e-01
-3.58575195e-01 -5.69988713e-02 1.14441788e+00 -4.79437768e-01
-2.29786426e-01 6.29562676e-01 -4.12006766e-01 -2.51670122e-01
9.69762027e-01 -2.20506176e-01 1.38225347e-01 2.33313665e-01
-1.09392691e+00 4.31012809e-01 1.97639093e-01 -1.11946836e-01
5.27032554e-01 -1.38010430e+00 -5.98558724e-01 3.84600967e-01
-3.18629831e-01 7.11619437e-01 8.29972625e-01 1.57987392e+00
-8.58176231e-01 4.44593519e-01 -1.74174830e-01 -1.16262221e+00
-5.95349133e-01 3.28059226e-01 7.88930297e-01 -5.49942017e-01
-1.17778528e+00 7.26887584e-01 3.60557914e-01 -7.87986338e-01
2.50547260e-01 -8.20296288e-01 -2.13438213e-01 -3.61754000e-02
3.48778628e-02 1.39521584e-01 2.06758659e-02 -5.08219063e-01
-2.95894355e-01 6.62397861e-01 2.39209700e-02 6.96848556e-02
1.39431536e+00 -7.06042424e-02 -2.49900743e-01 1.09596863e-01
1.09172618e+00 -3.79394799e-01 -1.49217474e+00 -3.75895441e-01
-1.78411081e-01 -1.83504168e-02 5.84267497e-01 -1.19062328e+00
-1.60452700e+00 7.28321970e-01 7.32066691e-01 -1.37543231e-01
1.18473458e+00 -4.95619774e-02 1.04875517e+00 -2.40033239e-01
2.18468219e-01 -5.11939108e-01 -4.34457213e-01 4.12533641e-01
4.85710859e-01 -1.23290801e+00 -6.91202432e-02 -4.98938233e-01
-4.80666190e-01 1.35282147e+00 4.45418179e-01 -1.34791046e-01
1.02887368e+00 5.01086473e-01 2.82228440e-01 -1.60361856e-01
-9.14425850e-02 1.27110571e-01 3.29854071e-01 3.65068436e-01
4.46724594e-01 1.93693265e-01 -2.67525792e-01 8.68462026e-01
4.63762760e-01 3.87738824e-01 5.53922504e-02 8.22482467e-01
-2.09079638e-01 -6.39966547e-01 -4.29055989e-01 4.38922524e-01
-3.21228087e-01 -1.21856174e-02 5.14111698e-01 7.89069414e-01
1.48279324e-01 2.39002854e-01 -3.55879031e-02 3.60709429e-02
3.05764377e-01 1.12300679e-01 4.57674950e-01 -2.58852452e-01
-6.59829259e-01 7.27750540e-01 -3.70128363e-01 -5.74164569e-01
-4.03087199e-01 -7.70076811e-01 -1.68007827e+00 2.50558078e-01
-1.69884831e-01 6.68755174e-02 7.71140933e-01 9.38142836e-01
1.73548907e-01 9.89049077e-01 3.48773181e-01 -1.15588880e+00
-3.36892068e-01 -9.77397501e-01 -5.18310249e-01 8.19458663e-02
2.19295979e-01 -5.06905735e-01 -3.51080984e-01 -9.60986093e-02]
|
[14.5484037399292, -2.724985361099243]
|
d5948d03-ab2c-4e5e-9fea-c06b2a9f0793
|
multiscale-analysis-for-improving-texture
|
2204.09841
| null |
https://arxiv.org/abs/2204.09841v1
|
https://arxiv.org/pdf/2204.09841v1.pdf
|
Multiscale Analysis for Improving Texture Classification
|
Information from an image occurs over multiple and distinct spatial scales. Image pyramid multiresolution representations are a useful data structure for image analysis and manipulation over a spectrum of spatial scales. This paper employs the Gaussian-Laplacian pyramid to treat different spatial frequency bands of a texture separately. First, we generate three images corresponding to three levels of the Gaussian-Laplacian pyramid for an input image to capture intrinsic details. Then we aggregate features extracted from gray and color texture images using bio-inspired texture descriptors, information-theoretic measures, gray-level co-occurrence matrix features, and Haralick statistical features into a single feature vector. Such an aggregation aims at producing features that characterize textures to their maximum extent, unlike employing each descriptor separately, which may lose some relevant textural information and reduce the classification performance. The experimental results on texture and histopathologic image datasets have shown the advantages of the proposed method compared to state-of-the-art approaches. Such findings emphasize the importance of multiscale image analysis and corroborate that the descriptors mentioned above are complementary.
|
['Alessandro L. Koerich', 'Alceu S. Britto Jr.', 'Jonathan de Matos', 'Diego Saqui', 'Steve T. M. Ataky']
|
2022-04-21
| null | null | null | null |
['texture-classification']
|
['computer-vision']
|
[ 6.34025156e-01 -4.33284074e-01 -1.66305363e-01 2.56243232e-03
-9.68984067e-01 -3.20482731e-01 2.98030972e-01 6.27525091e-01
-4.09764349e-01 5.44192135e-01 1.42543316e-01 2.85625994e-01
-5.93254328e-01 -9.47307706e-01 -4.23480533e-02 -1.18865311e+00
-3.05949479e-01 -1.64370328e-01 5.98159611e-01 -6.82245642e-02
6.09925866e-01 8.50550890e-01 -1.95451713e+00 7.29208887e-01
7.91004896e-01 1.20301008e+00 1.85215101e-01 5.68985999e-01
-9.97928008e-02 6.40560389e-01 -4.53502297e-01 -2.11748984e-02
-1.66932032e-01 -1.72478035e-01 -6.94422781e-01 4.31936085e-01
1.37284145e-01 1.06320806e-01 4.84782234e-02 1.29451811e+00
3.01669240e-01 -4.39381525e-02 1.01319051e+00 -6.18519247e-01
-9.85919952e-01 -2.11375937e-01 -1.01545548e+00 5.12052953e-01
2.98322558e-01 -3.46577376e-01 6.00779593e-01 -9.20670569e-01
6.32053792e-01 1.29556215e+00 5.49661756e-01 -2.96753556e-01
-1.45981717e+00 -1.96683064e-01 -4.51244652e-01 3.86544555e-01
-1.62847579e+00 -5.63284196e-02 9.56940353e-01 -5.18129230e-01
3.03320289e-01 4.44567084e-01 3.47278118e-01 3.73991221e-01
9.66655672e-01 3.31130773e-01 1.82962799e+00 -5.41998148e-01
3.36623713e-02 -1.12049766e-01 2.23951470e-02 9.89609957e-01
1.82633221e-01 -1.65662140e-01 -4.76552159e-01 -3.91128272e-01
9.78013575e-01 2.70395935e-01 -2.82923542e-02 -1.94508821e-01
-1.26957238e+00 6.94013894e-01 2.58305132e-01 1.13851488e+00
-7.25116968e-01 -3.88186514e-01 3.47705781e-01 2.45089516e-01
3.57562572e-01 -1.71741784e-01 2.42413521e-01 2.32729614e-01
-8.33722770e-01 -1.88646346e-01 1.42003059e-01 3.62001002e-01
1.16270840e+00 -2.41501108e-01 -2.31234923e-01 9.89201605e-01
-5.56445271e-02 5.33160090e-01 7.01185644e-01 -7.89491653e-01
-2.26790354e-01 7.72182047e-01 -2.15679944e-01 -1.67420626e+00
-4.35440719e-01 -1.08800046e-01 -1.29893994e+00 3.68305415e-01
3.30235332e-01 5.85721672e-01 -7.05526710e-01 1.13534606e+00
2.46118575e-01 -3.92375663e-02 -9.93324071e-02 6.41378999e-01
6.62013412e-01 3.68619859e-01 1.38764650e-01 -1.99626595e-01
1.81317544e+00 -2.20025092e-01 -8.28329682e-01 4.34991300e-01
5.40501997e-02 -1.18365717e+00 9.61544394e-01 3.70742857e-01
-8.98296058e-01 -8.15302849e-01 -1.05602598e+00 9.96388122e-02
-5.94942033e-01 3.52676392e-01 4.64788944e-01 6.35695755e-01
-1.03351164e+00 6.10828102e-01 -8.23000610e-01 -2.86346257e-01
8.09035823e-02 2.01849863e-01 -8.08181882e-01 -3.55672091e-02
-8.74147713e-01 6.53830111e-01 6.72694966e-02 -9.16354954e-02
-2.50669997e-02 -2.85996139e-01 -7.31980860e-01 -7.91023821e-02
-2.65831828e-01 -1.81771532e-01 2.87387013e-01 -5.24130762e-01
-1.20366061e+00 1.01564741e+00 -4.55632955e-01 1.16201080e-01
-2.94324160e-02 5.29568255e-01 -5.87358177e-01 9.18588519e-01
3.07308346e-01 7.28649646e-02 8.80851626e-01 -1.25418365e+00
-7.38154769e-01 -6.00741327e-01 -3.04612488e-01 -1.11437924e-02
-3.04458022e-01 5.07985353e-02 -2.93626845e-01 -8.52194786e-01
5.95105469e-01 -5.84843695e-01 -6.78801686e-02 -2.00580493e-01
-1.91597089e-01 5.67559749e-02 7.05460846e-01 -6.47444189e-01
1.17453980e+00 -2.34058523e+00 8.80475249e-03 5.30207336e-01
1.87794507e-01 -2.59501606e-01 -5.41573428e-02 4.37512338e-01
5.75288534e-02 2.17863336e-01 2.00501550e-02 2.41216674e-01
-4.14081156e-01 -7.65627176e-02 3.30210030e-01 9.05120015e-01
3.28060210e-01 3.48127097e-01 -6.19806170e-01 -1.01751268e+00
4.94281411e-01 7.35951424e-01 -2.42049918e-01 -3.83284360e-01
6.21803999e-01 4.86070037e-01 -9.49986219e-01 8.92608047e-01
8.99065852e-01 -2.31151521e-01 5.97191788e-02 -7.13959396e-01
-1.93080828e-01 -6.78710520e-01 -1.03203380e+00 1.28797269e+00
-4.44031924e-01 3.29959273e-01 3.61048169e-02 -1.02334678e+00
1.07044244e+00 2.44500324e-01 8.80778432e-01 -7.47118592e-01
1.45246908e-01 1.81787908e-01 -2.01506108e-01 -6.30222738e-01
4.76968467e-01 -3.73550117e-01 -1.67644247e-01 -1.37270112e-02
-7.90025480e-03 5.72989974e-03 4.47587222e-01 -3.55127484e-01
8.26157153e-01 -3.31060201e-01 4.93605316e-01 -7.36088216e-01
1.08216691e+00 -1.59695327e-01 3.40063423e-01 6.21812522e-01
-2.62996703e-01 5.67035317e-01 4.62820798e-01 -4.12002772e-01
-8.77949178e-01 -1.17005587e+00 -6.43685579e-01 9.78329420e-01
3.39506984e-01 6.08272739e-02 -6.17176354e-01 -1.22733556e-01
9.04183686e-02 -7.75624588e-02 -1.01566970e+00 1.74851805e-01
-2.74044991e-01 -1.00735033e+00 2.85033166e-01 -3.73822674e-02
5.65231085e-01 -7.47135878e-01 -6.42407656e-01 1.57508925e-01
-2.11763397e-01 -9.15274382e-01 -3.16085458e-01 -8.06373805e-02
-8.55339170e-01 -1.18607295e+00 -8.22251678e-01 -1.02996838e+00
8.21907401e-01 4.01219398e-01 6.63521349e-01 -2.18262170e-02
-8.66812408e-01 3.69173557e-01 -4.73866880e-01 5.71058728e-02
-4.43914711e-01 -2.75450051e-01 -6.63974211e-02 5.53496003e-01
3.78451794e-01 -7.96170115e-01 -7.87392855e-01 3.23684365e-01
-1.21666861e+00 -2.20037356e-01 1.03052831e+00 8.71294320e-01
1.18279731e+00 5.63657165e-01 4.52159196e-01 -6.32524490e-01
7.14220941e-01 -1.46364599e-01 -2.64659286e-01 4.24574405e-01
-1.32772788e-01 6.27297312e-02 6.49186552e-01 -1.95502922e-01
-1.03084743e+00 -2.65611589e-01 2.91956633e-01 -6.26607686e-02
-2.98221499e-01 5.26975989e-01 2.96770334e-01 -5.06353378e-01
7.17172265e-01 8.11681390e-01 3.04415762e-01 -4.78595436e-01
3.64996158e-02 7.29609191e-01 5.73050559e-01 -6.59756303e-01
3.68179172e-01 1.02401817e+00 3.83647710e-01 -1.34106994e+00
-4.33596969e-01 -6.84040189e-01 -8.24881792e-01 -9.51260701e-02
1.07648027e+00 -5.03053129e-01 -8.62021744e-01 4.95706350e-01
-7.48021424e-01 4.93601918e-01 2.14887150e-02 4.12634373e-01
-6.15544081e-01 8.22027922e-01 -7.32409835e-01 -6.36681616e-01
-1.43205315e-01 -1.21475768e+00 1.24911106e+00 2.85895169e-01
1.75738230e-01 -1.06897545e+00 -1.18571959e-01 1.53458178e-01
6.47338748e-01 7.76006818e-01 1.29864168e+00 4.08513173e-02
-1.46861196e-01 -4.14645374e-01 -4.41397995e-01 1.60734072e-01
7.55886793e-01 7.22427815e-02 -6.54215097e-01 -1.78159267e-01
2.62905568e-01 -1.67640135e-01 7.11896181e-01 4.81044978e-01
1.35054600e+00 -9.96114314e-02 -2.49255836e-01 3.22887450e-01
1.83192742e+00 2.93739498e-01 5.60876012e-01 3.90738785e-01
9.00698826e-02 6.11055911e-01 5.59658885e-01 4.68795180e-01
1.10504739e-01 5.44231653e-01 -9.57833976e-02 -5.49045920e-01
-1.35885298e-01 2.89921790e-01 -1.47451714e-01 8.80063713e-01
-4.30618286e-01 4.99783874e-01 -6.52561426e-01 4.81021136e-01
-1.36122274e+00 -1.04323125e+00 1.25138927e-02 2.11919308e+00
8.02522123e-01 -1.22824304e-01 -9.27938521e-02 2.93415070e-01
1.06211269e+00 1.89856276e-01 -1.08703338e-01 -3.20619583e-01
-4.56229150e-01 3.17122489e-01 3.95113975e-01 3.13945800e-01
-1.35155976e+00 4.65422362e-01 6.41884804e+00 1.34581006e+00
-1.25931668e+00 -1.81756504e-02 7.51881540e-01 5.43634772e-01
-9.05280858e-02 -3.76289159e-01 -1.55046791e-01 3.24484169e-01
5.70829630e-01 -4.27401990e-01 1.41275585e-01 4.41854686e-01
2.73004502e-01 -6.05314970e-01 -4.28363651e-01 1.03651631e+00
-9.17568579e-02 -1.16379106e+00 2.90837526e-01 2.27619559e-01
6.95712388e-01 -4.00952041e-01 3.87995362e-01 -3.55717719e-01
-1.71415552e-01 -8.98661256e-01 2.70617306e-01 9.01379466e-01
9.80682611e-01 -7.71087646e-01 8.86764228e-01 -1.36801854e-01
-1.58303964e+00 1.24982409e-01 -6.56383514e-01 2.67870754e-01
-2.65049875e-01 6.94134057e-01 -1.74864292e-01 8.49136651e-01
6.33569062e-01 3.61855656e-01 -8.04165542e-01 8.94684374e-01
4.00601447e-01 1.51257720e-02 -1.09655231e-01 7.72510841e-02
2.36124888e-01 -4.48471427e-01 2.71371245e-01 1.32717574e+00
4.09893185e-01 1.39805987e-01 2.29439288e-01 4.75332111e-01
4.29725051e-01 6.50336623e-01 -5.78961909e-01 8.79761800e-02
3.59323233e-01 1.48179448e+00 -1.23704481e+00 -4.19486642e-01
-7.08848000e-01 9.38769698e-01 -7.86036700e-02 2.78080761e-01
-4.04237628e-01 -7.17259526e-01 3.07094365e-01 5.79882599e-02
2.11561784e-01 -1.59636945e-01 -3.09429854e-01 -8.12323272e-01
-3.39136384e-02 -7.53706276e-01 3.01966310e-01 -3.62127662e-01
-1.31378233e+00 7.25571692e-01 4.90280017e-02 -1.54440057e+00
1.19016565e-01 -6.33008063e-01 -3.00001591e-01 1.04875588e+00
-1.43272614e+00 -1.23940837e+00 -3.39895993e-01 7.25878119e-01
1.41070381e-01 -4.12452631e-02 1.05224812e+00 3.39315273e-02
-8.74228999e-02 2.28224054e-01 5.90948939e-01 -2.78019495e-02
5.50965726e-01 -1.18459189e+00 -3.95116508e-01 5.15842617e-01
-3.59075546e-01 6.45433128e-01 4.76435035e-01 -5.07035494e-01
-1.32937121e+00 -6.70307696e-01 5.94705939e-01 1.65486727e-02
8.10723662e-01 1.89994797e-01 -8.50510478e-01 -6.57812059e-02
-1.05236666e-02 1.78872168e-01 1.04678953e+00 -2.41068110e-01
-6.90763146e-02 -1.73425853e-01 -1.40225852e+00 3.97266448e-01
3.78406018e-01 -7.42887139e-01 -5.08327842e-01 1.60467297e-01
-1.00965433e-01 6.91166967e-02 -1.72678864e+00 4.20262009e-01
8.00965846e-01 -1.40073705e+00 1.01934731e+00 -5.56231476e-02
3.47931653e-01 -4.37494487e-01 -3.59643131e-01 -9.32057142e-01
-7.87176192e-01 -9.17574167e-02 7.30193019e-01 9.29197907e-01
6.62043840e-02 -5.63045859e-01 5.51623404e-01 -6.58884570e-02
2.80745357e-01 -7.31267333e-01 -1.17777455e+00 -6.71800673e-01
5.73198833e-02 1.54050812e-01 3.10174137e-01 7.90931404e-01
2.67782837e-01 -3.11934292e-01 -4.94591370e-02 1.68369710e-01
1.08649909e+00 4.25440431e-01 1.97138712e-01 -1.28613162e+00
9.98554155e-02 -7.06779540e-01 -9.88138795e-01 -1.14309072e-01
-2.11717755e-01 -7.85262525e-01 -2.80729085e-01 -1.38390362e+00
5.21818221e-01 -3.15034807e-01 -6.27713859e-01 2.18843311e-01
-1.98739082e-01 7.40333974e-01 -1.02638006e-01 3.42021674e-01
-2.80476153e-01 2.23536506e-01 1.54565382e+00 -2.50600845e-01
1.68938786e-02 -2.89803714e-01 -6.68088317e-01 7.33546793e-01
5.42712629e-01 -1.32526904e-01 -8.84472206e-02 9.68507007e-02
-5.09240687e-01 1.95070490e-01 1.96886763e-01 -1.16362429e+00
1.84023455e-01 -2.84696639e-01 6.16820455e-01 -2.52061993e-01
2.85286486e-01 -6.87125742e-01 2.56884366e-01 6.42322421e-01
-5.71886152e-02 -6.88977391e-02 2.30734572e-01 5.34021497e-01
-8.12661827e-01 7.47758672e-02 1.17535913e+00 -1.23706825e-01
-8.82708132e-01 2.16683149e-02 -7.18999922e-01 -5.45544207e-01
9.85282183e-01 -5.94939828e-01 -3.56549323e-01 9.33127664e-03
-1.00584030e+00 -5.50312757e-01 6.33728147e-01 2.02685371e-01
7.12280810e-01 -1.48622727e+00 -7.94916868e-01 3.64926398e-01
3.13264787e-01 -8.04447591e-01 8.41119468e-01 1.16133893e+00
-7.36513436e-01 3.89618367e-01 -8.40894043e-01 -7.92082667e-01
-1.38997090e+00 5.56213439e-01 1.16808087e-01 -4.89596188e-01
-5.45410037e-01 2.37595722e-01 5.41067004e-01 1.23709001e-01
-3.74434918e-01 -4.18268204e-01 -6.12840176e-01 1.81871101e-01
6.69655204e-01 4.34610486e-01 1.97673291e-01 -9.65600908e-01
-3.76474440e-01 1.45013130e+00 1.04691088e-02 5.42310737e-02
1.18187511e+00 -3.83295804e-01 -5.30330598e-01 4.38511431e-01
1.47817910e+00 3.14758003e-01 -6.77025795e-01 -3.59169900e-01
5.40155545e-02 -5.95239937e-01 1.66117296e-01 -3.25766027e-01
-8.00683022e-01 7.78852820e-01 9.49557245e-01 6.55579627e-01
1.55696607e+00 -3.16317491e-02 2.01528877e-01 -3.81683558e-02
5.50546706e-01 -1.04727030e+00 5.88849820e-02 -3.09558194e-02
9.02592897e-01 -9.14398909e-01 1.28901005e-01 -6.91095471e-01
-4.25479621e-01 1.40441072e+00 -5.41263260e-02 -2.71508366e-01
8.38528097e-01 3.03248346e-01 3.92502621e-02 -2.01370895e-01
-2.13947922e-01 -4.35405225e-01 4.74049777e-01 5.74388206e-01
6.65312529e-01 8.12343657e-02 -6.20798528e-01 3.40423793e-01
9.01378915e-02 9.74468663e-02 3.44537199e-01 1.02521539e+00
-8.32095981e-01 -1.00894094e+00 -8.06480706e-01 5.95545888e-01
-1.00353038e+00 1.57640830e-01 1.29093021e-01 6.97154999e-01
1.06943421e-01 9.54003274e-01 -9.18358490e-02 -3.92938197e-01
2.83231378e-01 -1.78290591e-01 5.47698975e-01 -1.11099355e-01
-3.46871167e-02 4.41845596e-01 -4.04158294e-01 -5.78064620e-01
-7.62987256e-01 -5.43857396e-01 -9.73007500e-01 -2.26296112e-01
5.17788939e-02 1.71678305e-01 5.08097291e-01 5.94769835e-01
3.20460618e-01 5.60287952e-01 7.16097295e-01 -1.06243360e+00
-2.09305987e-01 -8.71118426e-01 -1.24978995e+00 6.73222542e-01
4.27847862e-01 -8.75444412e-01 -3.23855758e-01 3.83150846e-01]
|
[10.431540489196777, -0.3835970163345337]
|
d29af8b2-0253-4626-848c-c357184a488b
|
hiding-image-in-image-by-five-modulus-method
|
1304.1571
| null |
http://arxiv.org/abs/1304.1571v1
|
http://arxiv.org/pdf/1304.1571v1.pdf
|
Hiding Image in Image by Five Modulus Method for Image Steganography
|
This paper is to create a practical steganographic implementation to hide
color image (stego) inside another color image (cover). The proposed technique
uses Five Modulus Method to convert the whole pixels within both the cover and
the stego images into multiples of five. Since each pixels inside the stego
image is divisible by five then the whole stego image could be divided by five
to get new range of pixels 0..51. Basically, the reminder of each number that
is not divisible by five is either 1,2,3 or 4 when divided by 5. Subsequently,
then a 4-by-4 window size has been implemented to accommodate the proposed
technique. For each 4-by-4 window inside the cover image, a number from 1 to 4
could be embedded secretly from the stego image. The previous discussion must
be applied separately for each of the R, G, and B arrays. Moreover, a stego-key
could be combined with the proposed algorithm to make it difficult for any
adversary to extract the secret image from the cover image. Based on the PSNR
value, the extracted stego image has high PSNR value. Hence this new
steganography algorithm is very efficient to hide color images.
|
['Firas A. Jassim']
|
2013-04-04
| null | null | null | null |
['image-steganography']
|
['computer-vision']
|
[ 7.62420356e-01 9.16947871e-02 1.97115317e-01 1.83330551e-01
-1.82479158e-01 -6.69552624e-01 -1.15852594e-01 -3.56329173e-01
-3.28764647e-01 6.64734006e-01 -4.55240548e-01 -6.94929540e-01
4.19323325e-01 -1.10101604e+00 -4.55269486e-01 -1.05264640e+00
-9.21783522e-02 -2.70686537e-01 5.64006150e-01 -3.69837970e-01
5.17452538e-01 2.96034485e-01 -1.26391673e+00 2.14826763e-01
5.93555808e-01 9.29641664e-01 1.39000669e-01 8.48424017e-01
-1.24610744e-01 3.49388748e-01 -8.03509951e-01 -1.36197492e-01
7.23442316e-01 -1.03720224e+00 -4.00414884e-01 4.93653268e-01
-2.95965850e-01 -6.54446542e-01 -2.23756991e-02 1.36522448e+00
1.42687876e-02 -4.33676213e-01 5.02879620e-01 -1.19471931e+00
-5.23375511e-01 4.90035027e-01 -1.16140497e+00 -2.64318466e-01
2.01365471e-01 7.88166523e-02 4.68602955e-01 -3.99778455e-01
3.88002127e-01 8.62259984e-01 1.88747689e-01 6.97618946e-02
-6.68488026e-01 -9.94157255e-01 -4.15146828e-01 9.71975103e-02
-1.27083385e+00 -8.70092064e-02 7.58152306e-01 1.18725903e-01
4.74374652e-01 7.48436213e-01 1.05603695e+00 -2.43222956e-02
6.05016470e-01 2.41762511e-02 1.59254110e+00 -9.75571930e-01
-1.05040774e-01 4.49852437e-01 -4.33286071e-01 6.38020933e-01
7.97966361e-01 -6.47379681e-02 3.20036352e-01 1.75297037e-01
8.05256009e-01 1.24701433e-01 -3.91796947e-01 -8.81517492e-03
-1.07778001e+00 8.12040269e-01 3.39894235e-01 6.16527140e-01
-3.47442925e-03 4.68126833e-01 6.14541322e-02 5.60235441e-01
-2.54160762e-01 1.50534948e-02 3.43254022e-02 3.70060712e-01
-8.75656903e-01 -2.70107180e-01 8.46764088e-01 7.14341998e-01
1.00643981e+00 2.82905251e-01 8.16993177e-01 1.46222129e-01
5.34407675e-01 7.44526923e-01 1.82340652e-01 -6.85174286e-01
4.46460873e-01 4.98961449e-01 8.03748518e-02 -1.36957264e+00
5.52003048e-02 1.67366043e-02 -9.22503769e-01 6.47792637e-01
5.22222221e-01 -1.85200870e-01 -9.08445656e-01 1.07721698e+00
1.82754189e-01 -7.35674277e-02 2.94318557e-01 8.06230485e-01
4.60689396e-01 1.21219885e+00 -4.04505581e-01 -2.60847121e-01
1.72738004e+00 -6.54402435e-01 -6.28544092e-01 -3.97627860e-01
2.31201589e-01 -1.18355608e+00 3.15244704e-01 3.54963452e-01
-1.02606928e+00 -3.92706156e-01 -1.51115584e+00 4.83719736e-01
-3.85612935e-01 -1.23866880e-02 2.19028950e-01 1.18153167e+00
-9.03241634e-01 1.27598057e-02 -2.93954372e-01 9.67812538e-02
-2.34120652e-01 5.14520466e-01 -3.15918565e-01 -8.80078301e-02
-1.39805031e+00 5.54394007e-01 8.05866838e-01 1.38010949e-01
-2.85379469e-01 2.65085787e-01 -7.22080112e-01 1.78311110e-01
2.32555106e-01 -2.74250478e-01 2.51598179e-01 -1.38574910e+00
-1.01670194e+00 9.60398912e-01 -2.27151196e-02 -4.51285630e-01
4.68243122e-01 8.42786968e-01 -6.73057854e-01 5.39751291e-01
-2.10895643e-01 3.86936963e-01 1.22391367e+00 -1.50139678e+00
-7.73808002e-01 -1.21456787e-01 6.20818883e-02 7.88215697e-02
3.97648662e-02 4.69324291e-02 -4.18988168e-01 -5.40175259e-01
5.77105403e-01 -1.25729358e+00 -1.43639863e-01 -1.63220108e-01
-4.56748217e-01 8.11907411e-01 8.49963069e-01 -8.82099569e-01
1.43098593e+00 -2.48792005e+00 -3.93892407e-01 8.88203502e-01
1.59502849e-01 3.09787095e-01 2.85826921e-01 5.49081266e-01
-2.13805601e-01 3.32176059e-01 -3.38589549e-01 5.12594283e-01
-3.51469249e-01 -1.61588360e-02 1.18406922e-01 8.33655834e-01
-3.17908138e-01 3.68581891e-01 -3.66113782e-01 -4.41347897e-01
3.27160805e-02 5.90344727e-01 -2.56223351e-01 -3.81697446e-01
3.46988112e-01 2.56758481e-01 -4.21542287e-01 4.51311737e-01
1.50540304e+00 -7.89652094e-02 4.49531525e-01 -1.27083853e-01
-3.82331163e-01 -3.48370612e-01 -1.55232370e+00 3.45595032e-01
-1.26079042e-02 5.28091848e-01 1.58680499e-01 -7.55156994e-01
1.08810782e+00 5.20643532e-01 3.97390604e-01 -4.90555793e-01
5.09596467e-01 4.29925740e-01 1.55742079e-01 -4.48367953e-01
5.51123261e-01 -2.86878586e-01 -1.10068381e-01 8.14759016e-01
-8.87650311e-01 -5.54667935e-02 3.41668315e-02 -1.54429212e-01
5.31637490e-01 -3.14987838e-01 3.68051052e-01 2.65035294e-02
9.72359955e-01 3.64279486e-02 4.87218142e-01 4.17142689e-01
-2.86592603e-01 2.77251393e-01 4.53020066e-01 -1.06103338e-01
-1.29075694e+00 -3.84990364e-01 3.46416086e-01 4.14656371e-01
7.79691756e-01 1.50624290e-01 -8.91717434e-01 -1.19182192e-01
-2.23092258e-01 2.13796660e-01 -3.09285820e-01 -2.90876329e-02
-6.32385015e-01 -3.66531253e-01 4.47485626e-01 -2.03357324e-01
1.18641233e+00 -1.13771260e+00 -9.09502387e-01 3.61389011e-01
-2.26790905e-01 -7.72679269e-01 -4.55240935e-01 -8.14140737e-02
-7.07117081e-01 -1.21334326e+00 -6.63634300e-01 -1.25833297e+00
1.07075405e+00 9.48457658e-01 2.14277655e-01 7.01644480e-01
-1.09563060e-01 -1.99087545e-01 -5.69790125e-01 -1.43225878e-01
-7.08440125e-01 -4.29917604e-01 -4.71805274e-01 7.25302994e-02
2.45280087e-01 -2.39014894e-01 -9.81826663e-01 4.16606486e-01
-1.22434509e+00 1.64015412e-01 6.44328058e-01 6.53096855e-01
2.52178222e-01 9.02592361e-01 7.17681646e-02 -5.93643844e-01
1.69186577e-01 -2.15542629e-01 -7.29630172e-01 1.99066281e-01
-4.62269038e-01 -1.70592442e-01 5.11591434e-01 -2.17248097e-01
-6.68505669e-01 -8.32602084e-02 1.28942996e-01 2.67428935e-01
8.88010263e-02 2.46348009e-01 -2.92934865e-01 -5.09911001e-01
-9.12406594e-02 5.36774278e-01 5.11936128e-01 1.41186133e-01
2.38592386e-01 8.08248103e-01 4.40664589e-01 5.47436118e-01
1.09764171e+00 7.26567149e-01 2.22574428e-01 -5.93341470e-01
4.31497484e-01 -1.72758654e-01 1.08697349e-02 -9.54284966e-02
9.74323452e-01 -8.06294918e-01 -8.61560464e-01 8.41901422e-01
-9.82183456e-01 3.58543664e-01 4.25493449e-01 4.03947771e-01
-8.84308144e-02 6.71851039e-01 -5.56979120e-01 -8.06248069e-01
-4.18686032e-01 -1.50728512e+00 1.09616391e-01 2.72947580e-01
7.02310801e-02 -8.57769847e-01 -6.73072159e-01 2.39009738e-01
4.58916098e-01 7.02296793e-01 7.72919714e-01 -1.93033576e-01
-9.79324639e-01 -5.16403854e-01 -3.79122555e-01 2.95144618e-01
3.75639349e-01 2.42041320e-01 -1.45546719e-01 -5.77370048e-01
4.73543704e-01 3.72563094e-01 9.39017832e-01 1.55026197e-01
5.98125398e-01 -6.66217923e-01 -3.60631406e-01 6.27829254e-01
2.01511145e+00 8.64120066e-01 1.27021420e+00 9.58198845e-01
4.24205095e-01 3.48035604e-01 6.13557518e-01 3.84237289e-01
2.51466841e-01 1.75261185e-01 6.79848492e-01 -5.22811949e-01
2.63747066e-01 2.64342606e-01 3.68256032e-01 7.02852249e-01
-8.92958604e-03 -4.48609859e-01 -3.14265817e-01 3.62302423e-01
-9.08987463e-01 -1.04265749e+00 -6.44601464e-01 1.94981039e+00
6.92994475e-01 2.04688907e-01 -3.06381166e-01 7.81984091e-01
1.20693171e+00 3.30731183e-01 -1.34396866e-01 -8.53427231e-01
-1.31606087e-01 4.70814407e-02 1.31180334e+00 5.99361181e-01
-6.76159143e-01 5.21872818e-01 5.30215263e+00 7.91250765e-01
-1.41615009e+00 -2.44853660e-01 6.08304083e-01 6.07793748e-01
-6.62187636e-01 5.71506560e-01 -7.26819038e-01 7.76992381e-01
4.06589478e-01 3.70822195e-03 4.91234779e-01 2.81371593e-01
2.52926230e-01 -8.32058370e-01 -1.84148327e-01 6.85923696e-01
1.63031518e-01 -8.44331920e-01 2.00917497e-02 4.43620175e-01
9.08997059e-01 -1.02516210e+00 4.76564080e-01 -4.72665101e-01
-2.28172079e-01 -6.95940316e-01 5.60703635e-01 -1.30503783e-02
9.61945057e-01 -9.24721539e-01 7.34865427e-01 2.13043898e-01
-1.17603564e+00 -5.68981171e-02 -3.95192236e-01 1.84394255e-01
1.75502583e-01 3.67855698e-01 -7.95379877e-01 4.75255400e-01
5.13852477e-01 -1.73775479e-01 -1.44171447e-01 7.24151790e-01
-6.50847256e-02 4.51335430e-01 -2.62594640e-01 7.23980218e-02
6.59794390e-01 -4.63590056e-01 5.46712518e-01 7.28951871e-01
8.95384312e-01 2.75877625e-01 -3.31158280e-01 4.57824647e-01
9.31105167e-02 1.71424866e-01 -5.01579463e-01 1.69382785e-02
6.41180217e-01 8.01651895e-01 -1.19506478e+00 -6.25904620e-01
-5.47902882e-01 1.04569614e+00 -8.73851359e-01 2.48455927e-01
-7.86077976e-01 -1.10228074e+00 9.48747024e-02 4.06088054e-01
1.01010489e+00 -1.94620356e-01 -2.86672711e-01 -5.22682846e-01
-2.64442265e-01 -1.35923207e+00 1.32026404e-01 -5.07215559e-01
-2.84166932e-01 6.33391678e-01 -2.52540499e-01 -1.56245315e+00
7.16026407e-03 -2.06850469e-01 -6.99302495e-01 1.07558918e+00
-1.45057774e+00 -1.00197470e+00 -2.32362419e-01 5.73255658e-01
-1.24354526e-01 -1.49066910e-01 5.47499955e-01 -6.85357973e-02
-1.16246529e-01 4.95831132e-01 4.35885251e-01 1.19889006e-01
6.04446709e-01 -4.84347463e-01 2.91280806e-01 1.08036411e+00
-5.97163498e-01 6.09953105e-01 7.31026590e-01 -7.99220860e-01
-1.19394565e+00 -4.18997109e-01 8.17383945e-01 4.40679103e-01
2.26673916e-01 5.47673106e-02 -7.28702307e-01 6.09081626e-01
6.12100065e-01 -4.94300574e-01 6.39620185e-01 -1.29930890e+00
-1.19986072e-01 -1.20603859e-01 -1.62838447e+00 4.59581941e-01
9.09848586e-02 -1.28654808e-01 -3.59597392e-02 -2.66226709e-01
3.87049645e-01 -4.86770779e-01 -6.13137484e-01 1.02516934e-01
7.71942139e-01 -1.26294434e+00 7.99132049e-01 3.46431434e-01
3.83404791e-01 -8.10956001e-01 -2.02804375e-02 -7.55936503e-01
-1.06697410e-01 -5.85257947e-01 5.28894007e-01 8.93534958e-01
4.00523067e-01 -1.05073869e+00 6.70204759e-01 2.04479873e-01
5.35716951e-01 -3.39497626e-01 -6.97556198e-01 -3.37164074e-01
-2.41813391e-01 2.91686952e-01 1.01808143e+00 7.91836798e-01
-1.28070131e-01 -4.04345810e-01 -8.57553780e-01 3.28109562e-01
7.48637319e-01 2.41834670e-01 7.12202728e-01 -6.95615470e-01
-2.84053534e-01 -2.23726705e-01 -4.91298139e-01 -8.80802810e-01
-6.16749048e-01 -5.11998415e-01 -3.57183546e-01 -1.48525023e+00
-6.64397031e-02 -6.27158582e-01 -9.52050835e-02 3.09827477e-01
-4.35003154e-02 1.01194096e+00 4.19807583e-01 3.75556529e-01
1.55453250e-01 -3.88131142e-01 1.70125532e+00 -1.05984010e-01
-2.09441945e-01 -7.31885433e-02 -8.32020223e-01 4.18895394e-01
1.02734590e+00 -5.85599303e-01 -1.46232456e-01 -6.60414025e-02
2.53780007e-01 5.01919687e-01 1.77803278e-01 -6.40534878e-01
8.24588463e-02 -1.29027069e-01 4.97980058e-01 -6.36821270e-01
1.11621737e-01 -1.49438214e+00 7.95309126e-01 1.17922401e+00
2.98296958e-01 -4.64557745e-02 -6.64401576e-02 3.19057584e-01
-1.95071310e-01 -5.50959229e-01 8.09283555e-01 -4.81207848e-01
-8.54419351e-01 -2.86370009e-01 -6.57595575e-01 -7.64692962e-01
1.54458284e+00 -1.30976892e+00 -2.40448028e-01 -5.07548928e-01
-7.92349875e-01 -8.76698866e-02 7.03049421e-01 -1.85332000e-01
7.80579269e-01 -1.28981256e+00 -6.55297160e-01 6.31838083e-01
-1.22237094e-01 -5.58767855e-01 3.93206120e-01 7.76984751e-01
-1.45183718e+00 1.19387358e-01 -6.00368142e-01 -1.35427862e-01
-1.74327922e+00 5.20809114e-01 1.50703490e-01 -2.43060261e-01
-4.41150725e-01 5.58974683e-01 -2.24183246e-01 5.79423010e-01
-2.36924261e-01 -1.86812282e-01 -2.64517754e-01 -3.18235010e-02
5.75975180e-01 3.99464518e-01 -5.56697190e-01 -9.05990899e-01
-1.84390649e-01 9.64354575e-01 5.27821109e-02 -3.84733975e-01
1.01210546e+00 -6.54713154e-01 -8.26268613e-01 -2.98904479e-01
1.45031905e+00 5.44874191e-01 -8.05076420e-01 2.70964116e-01
-4.96482939e-01 -8.59820724e-01 -2.96793669e-01 -2.87488610e-01
-1.22941339e+00 2.90052831e-01 6.22853637e-01 6.00963891e-01
1.47859442e+00 -5.02159655e-01 1.12762797e+00 -2.76980132e-01
4.99969780e-01 -8.74608994e-01 -1.24421060e-01 3.44257392e-02
3.25785667e-01 -1.05711925e+00 1.16246574e-01 -6.20302498e-01
-7.09277391e-01 1.39981019e+00 5.04542850e-02 -2.95453161e-01
5.32918453e-01 2.72376567e-01 1.12616949e-01 2.18974873e-02
6.79811984e-02 1.85651500e-02 -1.59390852e-01 3.91684413e-01
-6.84717521e-02 -1.46868629e-02 -9.02138412e-01 -8.43949020e-02
-1.26517355e-01 -6.80586547e-02 1.09192717e+00 1.15796506e+00
-1.13485873e+00 -1.31847250e+00 -1.13377619e+00 -9.85896215e-02
-8.72686327e-01 2.09585950e-02 6.14085682e-02 8.12437654e-01
4.52583939e-01 1.10367632e+00 -1.01025350e-01 -3.31573009e-01
-2.42907941e-01 -3.15512538e-01 3.53350639e-01 6.90382421e-02
-4.59849238e-01 4.38365281e-01 -1.63543284e-01 3.55611630e-02
-4.57209736e-01 -7.14764521e-02 -1.59784067e+00 -8.79630029e-01
-5.52014470e-01 2.40777373e-01 9.73511517e-01 4.61810052e-01
-3.59638602e-01 3.70247394e-01 1.07781971e+00 -1.14389472e-01
-3.66794765e-02 -5.33594310e-01 -8.67528439e-01 -4.30234857e-02
5.02010345e-01 6.91562146e-02 -7.77756035e-01 2.81753331e-01]
|
[4.300374507904053, 8.05219554901123]
|
2660fc86-37cc-4234-8ebe-51bedd7d19d8
|
a-symbolic-framework-for-systematic
|
2305.12563
| null |
https://arxiv.org/abs/2305.12563v1
|
https://arxiv.org/pdf/2305.12563v1.pdf
|
A Symbolic Framework for Systematic Evaluation of Mathematical Reasoning with Transformers
|
Whether Transformers can learn to apply symbolic rules and generalise to out-of-distribution examples is an open research question. In this paper, we devise a data generation method for producing intricate mathematical derivations, and systematically perturb them with respect to syntax, structure, and semantics. Our task-agnostic approach generates equations, annotations, and inter-equation dependencies, employing symbolic algebra for scalable data production and augmentation. We then instantiate a general experimental framework on next-equation prediction, assessing systematic mathematical reasoning and generalisation of Transformer encoders on a total of 200K examples. The experiments reveal that perturbations heavily affect performance and can reduce F1 scores of $97\%$ to below $17\%$, suggesting that inference is dominated by surface-level patterns unrelated to a deeper understanding of mathematical operators. These findings underscore the importance of rigorous, large-scale evaluation frameworks for revealing fundamental limitations of existing models.
|
['Andre Freitas', 'Damien Teney', 'Marco Valentino', 'Jordan Meadows']
|
2023-05-21
| null | null | null | null |
['mathematical-reasoning']
|
['natural-language-processing']
|
[ 4.28973347e-01 4.61389542e-01 3.67637463e-02 -3.75458628e-01
-4.84211832e-01 -8.87280822e-01 5.93086600e-01 2.31579412e-02
1.79096445e-01 6.08228207e-01 6.99211583e-02 -1.09845734e+00
-1.46121532e-01 -8.36826861e-01 -1.06217885e+00 1.23681709e-01
-2.01428160e-01 3.01476359e-01 1.28019840e-01 -5.25284290e-01
2.19364047e-01 3.48403394e-01 -1.52259409e+00 5.01002729e-01
1.08426559e+00 7.75089204e-01 -2.73852050e-01 8.03082645e-01
-1.36190295e-01 1.27638984e+00 -6.88187897e-01 -1.07569432e+00
8.83540735e-02 -4.50129449e-01 -8.39470983e-01 -5.83916664e-01
7.51729131e-01 -4.54579473e-01 -1.01882346e-01 1.07695687e+00
6.50925785e-02 -4.23505455e-02 6.83986425e-01 -1.40303075e+00
-9.97946620e-01 1.30154335e+00 2.78979719e-01 2.70575464e-01
6.35152698e-01 4.37296867e-01 1.43077099e+00 -4.67649728e-01
5.94394624e-01 1.10576379e+00 1.09144783e+00 5.35176098e-01
-1.66171920e+00 -8.80932152e-01 1.67668596e-01 2.44343117e-01
-1.33094001e+00 -7.82787085e-01 6.69441402e-01 -4.91379678e-01
1.53801477e+00 3.84417623e-01 6.45821631e-01 1.24037206e+00
1.02246217e-01 5.44520140e-01 9.09483492e-01 -5.25529325e-01
-1.48306647e-02 1.69471696e-01 1.80756629e-01 1.09137845e+00
3.90148401e-01 1.69548601e-01 -7.31243670e-01 -7.05005303e-02
7.62265444e-01 -6.07314050e-01 -2.00356841e-01 -3.12097728e-01
-1.01086557e+00 6.84921503e-01 2.80417968e-02 1.08374365e-01
8.87878761e-02 3.45553219e-01 4.77677524e-01 5.90602279e-01
-2.37032861e-01 1.30326653e+00 -8.15091729e-01 -5.60653687e-01
-7.38029599e-01 7.85784662e-01 9.97267306e-01 1.43878806e+00
4.64015901e-01 2.12173119e-01 -1.12917364e-01 3.66519660e-01
2.28954237e-02 4.17521954e-01 3.23479831e-01 -1.13200104e+00
5.95172226e-01 7.02645540e-01 -1.56840548e-01 -9.09330785e-01
-1.66876942e-01 -3.53167892e-01 -4.38009411e-01 -1.30181342e-01
5.97681880e-01 -1.48102790e-01 -4.68605489e-01 2.05854130e+00
-2.24623352e-01 -2.79585458e-02 1.77629650e-01 3.79679382e-01
6.12189651e-01 3.75099421e-01 1.18967883e-01 7.61547387e-02
1.11526942e+00 -5.63785017e-01 -4.96594936e-01 -2.61165798e-01
9.83745396e-01 -3.53872567e-01 1.51828897e+00 5.57661653e-01
-1.57191432e+00 -4.58145797e-01 -1.08389938e+00 -3.62345904e-01
-3.33764493e-01 -1.38752028e-01 1.20944953e+00 6.39274895e-01
-9.56911802e-01 7.80952990e-01 -7.03059077e-01 1.45641729e-01
4.60421473e-01 3.26119721e-01 4.32386110e-03 5.78527078e-02
-1.37792253e+00 1.15383863e+00 3.81448179e-01 -1.66818708e-01
-6.90313220e-01 -1.47851658e+00 -1.14187264e+00 2.80435085e-01
2.55226433e-01 -7.15704620e-01 1.68284893e+00 -6.31863952e-01
-1.53039610e+00 4.80771750e-01 -2.34032929e-01 -7.35015571e-01
2.03450009e-01 -2.27852967e-02 -5.68114221e-01 -2.53467917e-01
-5.58626801e-02 2.91503936e-01 5.21888137e-01 -7.38253117e-01
-4.32578295e-01 -5.11224754e-02 5.21478057e-01 -1.51939809e-01
-1.60725489e-01 6.54926747e-02 1.04530990e-01 -6.75482035e-01
-1.34126171e-01 -6.91228688e-01 1.30763769e-01 -3.57896745e-01
-4.07530099e-01 -3.04925859e-01 3.60335521e-02 -6.61870360e-01
1.53442872e+00 -1.87942243e+00 2.68142998e-01 3.04906785e-01
2.67626107e-01 3.25651355e-02 1.36832267e-01 3.47318918e-01
-2.47983426e-01 4.93952066e-01 -5.32433204e-02 -1.17940463e-01
6.40181124e-01 1.79380458e-02 -7.36606121e-01 -1.09999098e-01
6.25355005e-01 1.34717941e+00 -7.96044767e-01 -5.26639163e-01
1.08303741e-01 3.80280539e-02 -1.09411788e+00 1.60092011e-01
-7.03634143e-01 -2.39833549e-01 -1.36785358e-01 6.00056589e-01
2.19956547e-01 -3.96293521e-01 3.34719270e-01 -1.35168016e-01
1.31421611e-01 1.02305984e+00 -9.90072250e-01 1.79447043e+00
-4.68331069e-01 7.53522336e-01 -4.06966269e-01 -8.12168241e-01
5.85640550e-01 -1.14014987e-02 -2.37334386e-01 -8.21924150e-01
-6.49988353e-02 1.33505449e-01 5.09863734e-01 -4.19114828e-01
4.39763814e-01 -3.07771027e-01 -4.08914089e-01 3.80300015e-01
9.24335122e-02 -4.64723676e-01 3.98677021e-01 4.10598457e-01
1.32954526e+00 3.26639473e-01 2.09176883e-01 -1.67449147e-01
3.26245606e-01 2.51335889e-01 3.28085005e-01 8.19418192e-01
3.07437360e-01 -6.28383309e-02 8.24226320e-01 -4.88162398e-01
-1.06032670e+00 -1.09481466e+00 -1.26132503e-01 1.26630998e+00
-3.49387705e-01 -1.02271497e+00 -6.76983595e-01 -3.62762451e-01
2.50238299e-01 1.35391319e+00 -5.24992287e-01 -5.41522384e-01
-7.31236279e-01 -3.01952779e-01 1.34369779e+00 7.67898262e-01
1.99196100e-01 -8.86735082e-01 -5.91521859e-01 3.42402346e-02
-1.02490306e-01 -1.29697311e+00 7.34071955e-02 3.24314624e-01
-7.58353651e-01 -1.08666027e+00 9.62371007e-02 -4.87651795e-01
5.27651429e-01 -7.10961223e-01 1.57390702e+00 1.98594704e-01
-4.91052940e-02 -6.13444224e-02 -3.78962569e-02 -3.47998947e-01
-8.47697556e-01 3.15708846e-01 -8.10345560e-02 -7.59045184e-01
4.22871917e-01 -9.47527289e-01 6.58635572e-02 -1.48346558e-01
-4.22179729e-01 3.31834286e-01 6.09055400e-01 5.48305750e-01
4.37209085e-02 6.37091249e-02 1.11147679e-01 -1.01425195e+00
8.99305761e-01 -6.02544174e-02 -9.45196986e-01 3.81034881e-01
-9.55309510e-01 5.75040162e-01 1.10040855e+00 -4.22700286e-01
-8.43773186e-01 -2.36403435e-01 3.07698935e-01 -4.41058576e-01
-1.01922140e-01 4.75811303e-01 -1.59538776e-01 1.49860322e-01
8.54038119e-01 3.36704522e-01 -2.09147573e-01 -1.44574195e-01
6.67003691e-01 1.41066164e-01 7.19223022e-01 -1.30621016e+00
1.17416394e+00 -3.05169731e-01 1.20257884e-01 -3.12692761e-01
-6.49134636e-01 4.34756726e-01 -2.94805825e-01 5.23458302e-01
2.89936811e-01 -8.43249142e-01 -1.06264782e+00 1.59362644e-01
-1.22289550e+00 -9.34280574e-01 -3.76121938e-01 1.03695549e-01
-4.65667576e-01 2.36658335e-01 -7.04373300e-01 -5.63218534e-01
-2.01105535e-01 -1.32757545e+00 7.87963212e-01 -2.41708681e-01
-1.02833521e+00 -9.44316328e-01 -1.58871025e-01 2.36104146e-01
3.79373193e-01 -5.82026429e-02 1.58228862e+00 -7.03244448e-01
-8.16044688e-01 1.56601802e-01 -1.94365799e-01 4.37599659e-01
-1.16636671e-01 1.44283757e-01 -8.32651854e-01 3.17836851e-01
-4.65572000e-01 -4.28822190e-01 1.88124716e-01 -3.09180409e-01
1.32338142e+00 -6.27156734e-01 -4.46069688e-02 8.67071867e-01
1.00678444e+00 -8.01706165e-02 6.06483519e-01 6.95385933e-02
5.24520934e-01 1.52125984e-01 -1.09316535e-01 1.70934498e-01
6.19949818e-01 5.93729198e-01 -3.28944139e-02 5.61053753e-01
-4.59462442e-02 -7.25534856e-01 5.17525196e-01 4.17068928e-01
-5.43310866e-02 1.43848896e-01 -1.11108720e+00 4.85404044e-01
-1.29862690e+00 -9.75214660e-01 2.87151728e-02 1.79498899e+00
1.58598495e+00 4.75864768e-01 -1.60139669e-02 3.83290410e-01
2.96049323e-02 -2.33357593e-01 -3.34337085e-01 -8.56698930e-01
2.49210857e-02 1.01014256e+00 3.16630006e-01 6.93972230e-01
-5.55674493e-01 1.11530471e+00 7.31726265e+00 5.83244383e-01
-9.64692950e-01 -5.65215528e-01 1.27340615e-01 -4.73002046e-02
-7.78940856e-01 1.12494752e-01 -8.31150293e-01 5.47287047e-01
1.34678662e+00 -3.06418568e-01 1.09429514e+00 7.90722013e-01
-2.50573665e-01 3.66000593e-01 -1.84405422e+00 6.33867085e-01
-5.28765023e-02 -1.56774545e+00 1.81581140e-01 -1.69913784e-01
4.45939004e-01 -4.43823576e-01 1.58217460e-01 7.77364790e-01
7.84065962e-01 -1.51411867e+00 9.27744627e-01 4.35631096e-01
8.11111391e-01 -5.57017922e-01 2.80481607e-01 2.35235333e-01
-9.71340001e-01 -9.65720192e-02 1.16749771e-01 -7.55847454e-01
-1.51023462e-01 1.31170407e-01 -9.77803588e-01 2.23635018e-01
4.62003320e-01 6.13676786e-01 -9.80639994e-01 1.67880788e-01
-8.26480210e-01 7.06698060e-01 -3.58792722e-01 -1.92524567e-01
-1.07735373e-01 1.98017195e-01 1.15139395e-01 1.32196259e+00
1.06069185e-01 1.76316619e-01 -3.04693878e-01 1.55875599e+00
-5.73912375e-02 -3.39406133e-01 -5.93557298e-01 -3.21725577e-01
7.34611332e-01 7.47474194e-01 -8.89796019e-02 -6.14088893e-01
-2.45292842e-01 7.73091674e-01 6.07229233e-01 3.63438189e-01
-1.19881773e+00 -4.43333685e-01 7.48498917e-01 6.46841750e-02
4.95873421e-01 -1.61346167e-01 -7.87748277e-01 -1.28182173e+00
2.79967457e-01 -1.29136133e+00 1.04614198e-01 -9.94242191e-01
-8.72402251e-01 1.07294075e-01 5.03453851e-01 -5.00967741e-01
-7.65040755e-01 -8.30797136e-01 -5.35143852e-01 9.84234989e-01
-1.29596388e+00 -9.91995096e-01 -1.44380942e-01 4.20940936e-01
8.88793841e-02 -1.18822493e-01 1.13234675e+00 1.01529412e-01
-4.76987213e-01 1.26752520e+00 -5.89850545e-01 4.07994956e-01
8.96985084e-02 -1.46992767e+00 6.65547073e-01 9.29149747e-01
1.69137776e-01 1.36542571e+00 8.51141274e-01 -3.59969020e-01
-1.73597658e+00 -8.29328716e-01 1.19903028e+00 -9.33386385e-01
1.01630259e+00 -6.03662670e-01 -6.37927294e-01 1.26957214e+00
1.53211541e-02 -9.99198556e-02 6.33805692e-01 5.25541425e-01
-9.23843324e-01 -2.46712342e-02 -9.60103989e-01 9.84288990e-01
1.59666061e+00 -8.42024028e-01 -9.33351099e-01 2.29375400e-02
9.96170759e-01 -6.93121850e-01 -1.29788482e+00 3.85821998e-01
7.81942368e-01 -9.35354233e-01 9.64920461e-01 -1.22759664e+00
1.10298502e+00 -3.61539200e-02 -3.15919697e-01 -1.21238697e+00
-3.68290454e-01 -1.04749835e+00 -6.82647109e-01 1.25248027e+00
8.20391297e-01 -6.29648626e-01 6.23704493e-01 9.37057257e-01
-3.97807568e-01 -7.57102489e-01 -4.70078319e-01 -6.41557157e-01
5.11608779e-01 -9.16236758e-01 1.09705710e+00 9.24802661e-01
5.25876641e-01 5.95684230e-01 2.67021954e-01 1.74272940e-01
3.66279244e-01 2.10344717e-01 9.58724082e-01 -9.30131376e-01
-6.23733222e-01 -7.22676098e-01 -4.46508765e-01 -8.75860035e-01
6.15056932e-01 -1.21243060e+00 -2.53402084e-01 -9.94602621e-01
-1.55372933e-01 -4.20255780e-01 -2.95399800e-02 5.93348742e-01
-1.38151124e-02 -8.90553147e-02 5.57835475e-02 -2.32230484e-01
-3.79347205e-01 2.33774617e-01 9.57913995e-01 -1.30876601e-02
2.23344117e-01 -4.00954664e-01 -1.18603086e+00 7.50293732e-01
6.79218411e-01 1.33133484e-02 -7.06396520e-01 -7.04801261e-01
7.44701207e-01 -2.07816258e-01 6.28138900e-01 -9.59043801e-01
9.79466885e-02 -3.30700904e-01 2.56301969e-01 -8.44874904e-02
1.09633826e-01 -6.70433640e-01 1.15910202e-01 2.28390291e-01
-8.38776767e-01 4.56671655e-01 5.15543103e-01 -1.85428711e-03
2.56767631e-01 1.15157571e-03 3.22844028e-01 -1.84028476e-01
-5.97529829e-01 -3.29900414e-01 -9.82731581e-02 6.34658992e-01
5.97794950e-01 -1.69415906e-01 -3.52087736e-01 -1.47592589e-01
-4.13110495e-01 -7.45454058e-02 5.33419669e-01 1.21453665e-01
3.34153473e-01 -1.12365067e+00 -3.97206843e-01 3.98264825e-01
1.91848800e-01 2.33822301e-01 -3.71182024e-01 5.52483797e-01
-4.93113309e-01 6.53595150e-01 1.27440160e-02 -2.21654147e-01
-9.04748082e-01 5.03343463e-01 5.51811397e-01 -3.98657650e-01
-4.95636076e-01 1.01707220e+00 -2.18709767e-01 -6.16126001e-01
2.34756693e-01 -1.43650496e+00 5.25471210e-01 -5.42539597e-01
2.32509211e-01 2.00281322e-01 1.55805543e-01 3.99206094e-02
-3.28139693e-01 2.15250701e-01 -8.99418741e-02 6.92698956e-02
1.15068877e+00 4.64489043e-01 -1.29776627e-01 3.67075264e-01
8.76637399e-01 5.93736768e-02 -7.46748567e-01 -1.22765630e-01
-3.22062075e-02 1.63824875e-02 -3.79942954e-01 -1.16310668e+00
-5.08935869e-01 8.03541481e-01 -2.27715373e-01 1.74397022e-01
8.13943028e-01 1.10974520e-01 6.19582176e-01 8.19738209e-01
2.21223399e-01 -7.80864537e-01 -1.10553026e-01 7.34698057e-01
7.09950984e-01 -7.36974299e-01 -1.89621598e-02 -5.00045776e-01
-4.03446704e-01 9.24596608e-01 7.01053500e-01 -1.13076694e-01
1.92542985e-01 7.05273688e-01 -5.25721967e-01 -1.64894879e-01
-1.20908737e+00 1.84566721e-01 2.24810764e-01 6.17512763e-01
6.31935954e-01 5.77014638e-03 3.62795144e-01 1.09176183e+00
-1.27991045e+00 3.15542132e-01 2.69658178e-01 7.67669201e-01
1.46230280e-01 -1.07406294e+00 -8.90523419e-02 5.06530523e-01
-3.75457615e-01 -6.94024086e-01 -5.40892422e-01 1.09448195e+00
2.24885643e-01 4.81501222e-01 8.51406716e-03 -4.44233447e-01
4.26751763e-01 6.75276816e-01 1.06971872e+00 -5.79096317e-01
-5.97241223e-01 -7.90382743e-01 5.36447465e-01 -4.25195307e-01
2.06475407e-01 -7.80341983e-01 -1.33840287e+00 -5.72085798e-01
6.66710362e-02 2.35778298e-02 2.05984935e-01 1.00264382e+00
5.45315742e-01 7.32767820e-01 7.19777774e-03 -1.67736530e-01
-1.06459737e+00 -8.78769338e-01 -3.63741741e-02 5.67357600e-01
2.72472411e-01 -4.21780914e-01 -4.06475008e-01 2.97163904e-01]
|
[9.336137771606445, 7.243639945983887]
|
dd8f903a-2e39-4265-bdf0-6de49c401915
|
low-rank-optimization-for-efficient-deep
|
2303.13635
| null |
https://arxiv.org/abs/2303.13635v1
|
https://arxiv.org/pdf/2303.13635v1.pdf
|
Low Rank Optimization for Efficient Deep Learning: Making A Balance between Compact Architecture and Fast Training
|
Deep neural networks have achieved great success in many data processing applications. However, the high computational complexity and storage cost makes deep learning hard to be used on resource-constrained devices, and it is not environmental-friendly with much power cost. In this paper, we focus on low-rank optimization for efficient deep learning techniques. In the space domain, deep neural networks are compressed by low rank approximation of the network parameters, which directly reduces the storage requirement with a smaller number of network parameters. In the time domain, the network parameters can be trained in a few subspaces, which enables efficient training for fast convergence. The model compression in the spatial domain is summarized into three categories as pre-train, pre-set, and compression-aware methods, respectively. With a series of integrable techniques discussed, such as sparse pruning, quantization, and entropy coding, we can ensemble them in an integration framework with lower computational complexity and storage. Besides of summary of recent technical advances, we have two findings for motivating future works: one is that the effective rank outperforms other sparse measures for network compression. The other is a spatial and temporal balance for tensorized neural networks.
|
['Yipeng Liu', 'Ce Zhu', 'Zhangxin Chen', 'Xinwei Ou']
|
2023-03-22
| null | null | null | null |
['model-compression']
|
['methodology']
|
[ 1.58497900e-01 -1.43903628e-01 -4.32320058e-01 -3.71294826e-01
-1.86182842e-01 1.02403481e-02 1.95643172e-01 -6.33418187e-02
-5.42457819e-01 5.58375478e-01 1.84229597e-01 -1.61182228e-02
-5.86815238e-01 -8.04078996e-01 -7.37117469e-01 -9.37421679e-01
-1.13477679e-02 1.74307436e-01 7.77169243e-02 8.75685364e-02
4.28602621e-02 5.69043636e-01 -1.67378092e+00 2.65028447e-01
7.34694660e-01 1.57083392e+00 4.40006286e-01 5.07459417e-02
-3.17904539e-02 9.27905500e-01 -3.03476572e-01 -3.65103453e-01
3.84584278e-01 -2.12669864e-01 -5.28922677e-01 -3.03278744e-01
4.58899289e-01 -6.96677387e-01 -9.86948371e-01 1.23018920e+00
5.93429506e-01 2.50310004e-01 3.40919524e-01 -1.00303316e+00
-3.87334257e-01 7.39977896e-01 -1.73183426e-01 9.67766494e-02
-5.20748436e-01 -2.25170270e-01 8.93575788e-01 -8.16035092e-01
4.03425992e-01 9.98542547e-01 6.63447022e-01 4.63143140e-01
-7.55666971e-01 -7.88057089e-01 1.70044839e-01 6.64950430e-01
-1.44154513e+00 -7.12515175e-01 8.62298846e-01 -2.58280247e-01
1.06687331e+00 2.19888061e-01 8.73350203e-01 8.22167456e-01
4.59124111e-02 8.35647643e-01 5.61207652e-01 -2.62708664e-01
5.78836024e-01 -5.56377321e-02 3.82494666e-02 6.97533965e-01
5.00168622e-01 9.22259241e-02 -8.74568343e-01 1.82499081e-01
8.53535593e-01 4.92154181e-01 -2.09232137e-01 -3.90014529e-01
-9.21199918e-01 1.01054645e+00 5.23602307e-01 5.31943858e-01
-3.09822023e-01 3.45327705e-01 8.02412033e-01 2.58746177e-01
4.64640081e-01 2.73634672e-01 -4.73510444e-01 -3.32661837e-01
-1.30647290e+00 6.09554425e-02 6.41161919e-01 9.73439455e-01
5.50049424e-01 5.44898510e-01 -3.21219414e-02 1.19397557e+00
1.60399824e-01 2.83629894e-01 8.32823634e-01 -1.15848482e+00
6.27683818e-01 3.81891519e-01 -5.46044946e-01 -1.29145002e+00
-3.46149594e-01 -8.87812912e-01 -1.67627323e+00 -2.37728834e-01
-6.38279840e-02 3.43822241e-02 -6.34142339e-01 1.60135961e+00
8.41766149e-02 2.37924159e-01 1.14906289e-01 9.47740674e-01
8.44876707e-01 6.76450908e-01 -2.77559549e-01 -3.62431437e-01
1.26305771e+00 -1.21683300e+00 -9.47405636e-01 -2.53155768e-01
7.85876811e-01 -3.19583118e-01 8.82683516e-01 5.77850938e-01
-1.13823962e+00 -3.13918382e-01 -1.32757354e+00 -3.69568378e-01
-1.91527694e-01 3.93277645e-01 9.26304519e-01 4.21785951e-01
-1.05693841e+00 1.17523122e+00 -1.09179366e+00 -1.61218703e-01
8.44224870e-01 4.51928943e-01 -1.85652301e-01 -2.26353467e-01
-1.13090575e+00 6.09013915e-01 6.99803472e-01 1.39239386e-01
-7.00126886e-01 -6.52319908e-01 -7.47871220e-01 4.91534352e-01
8.27922970e-02 -5.46840727e-01 9.61529195e-01 -4.54882830e-01
-1.58993280e+00 3.14276129e-01 -1.57123595e-01 -8.28360975e-01
8.43997821e-02 -4.88553792e-01 -2.91006714e-01 4.33301181e-01
-2.56940365e-01 5.14152110e-01 8.25370133e-01 -7.03770757e-01
-4.84770775e-01 -4.74715799e-01 -2.17408150e-01 3.20915282e-01
-1.35059202e+00 -2.07591549e-01 -6.15155518e-01 -8.01359475e-01
5.00626147e-01 -6.93930924e-01 -3.14953625e-01 3.03932697e-01
-2.02532753e-01 -4.95275781e-02 1.04667103e+00 -7.61454821e-01
1.54244709e+00 -2.32508874e+00 2.74531633e-01 3.83311436e-02
4.15840745e-01 4.77660120e-01 -1.54302001e-01 9.38124880e-02
-2.01655310e-02 1.29329130e-01 -5.93614168e-02 -4.38523918e-01
-1.59381494e-01 4.26204860e-01 -4.99340564e-01 3.68953228e-01
-1.41679168e-01 6.60093307e-01 -5.20301640e-01 -5.85341334e-01
8.61762241e-02 6.42697155e-01 -7.92162478e-01 -2.39660386e-02
1.90470770e-01 -1.52618438e-01 -3.94018918e-01 4.12290454e-01
6.00803375e-01 -4.04062390e-01 8.43649507e-02 -6.74528539e-01
2.02815514e-02 5.61969459e-01 -1.13439417e+00 1.87790322e+00
-4.79540348e-01 6.73130572e-01 1.47589311e-01 -1.47453880e+00
8.32468748e-01 1.26306668e-01 6.47479236e-01 -9.26347077e-01
2.19349191e-01 3.07444096e-01 -1.66740984e-01 -4.52012628e-01
5.10504007e-01 1.63246304e-01 3.82318467e-01 3.29890341e-01
2.89310336e-01 1.89250216e-01 3.66019458e-01 1.60918787e-01
9.68765795e-01 -3.50319415e-01 7.84096718e-02 -1.06058717e-01
1.42122224e-01 -2.79063821e-01 5.95383525e-01 4.13860112e-01
-2.87952982e-02 4.69178706e-01 3.58331978e-01 -5.09761989e-01
-1.28976786e+00 -4.86313403e-01 -2.64815539e-01 1.16086912e+00
1.31216375e-02 -6.52351677e-01 -6.84006512e-01 -1.53658941e-01
-1.93352416e-01 3.55249524e-01 -1.94248080e-01 -4.53015864e-01
-8.23110163e-01 -6.70554399e-01 4.51279074e-01 5.90033352e-01
8.29992533e-01 -6.84036613e-01 -5.96272409e-01 3.83494422e-02
-3.09443865e-02 -1.07706857e+00 -2.45979071e-01 4.76535261e-01
-1.60684228e+00 -5.09582460e-01 -6.90122426e-01 -9.46818650e-01
4.43690181e-01 4.52401638e-01 7.68590093e-01 1.49745986e-01
7.29808807e-02 -2.22323954e-01 -3.87100339e-01 -2.91787028e-01
1.78298250e-01 2.69943058e-01 4.06723410e-01 -1.68084502e-01
2.16582179e-01 -9.67337728e-01 -6.99825168e-01 6.46509230e-02
-9.36826646e-01 2.60449022e-01 9.41850841e-01 9.58289862e-01
9.41097677e-01 4.19395089e-01 2.02900380e-01 -5.20755947e-01
4.88537073e-01 -3.07844669e-01 -3.61202478e-01 2.07238439e-02
-7.35767901e-01 2.68883884e-01 9.07793581e-01 -5.61771512e-01
-6.08443201e-01 8.58977214e-02 -1.96759790e-01 -9.57096636e-01
2.68305242e-01 7.89691210e-01 -5.90990558e-02 -1.38393074e-01
4.69669729e-01 5.86032331e-01 1.75499573e-01 -8.61609042e-01
8.13688114e-02 4.72485870e-01 2.96452373e-01 -3.99216682e-01
7.49511957e-01 3.77535015e-01 1.54740348e-01 -9.51676667e-01
-1.00105751e+00 -1.09426983e-01 -4.33031499e-01 1.43060893e-01
4.68663663e-01 -1.07813513e+00 -3.80553484e-01 3.76890957e-01
-1.02883232e+00 -1.60817593e-01 -6.35300756e-01 7.53872454e-01
-3.00927430e-01 4.14664239e-01 -8.51342142e-01 -4.78481680e-01
-6.38517439e-01 -1.07613885e+00 7.20732868e-01 1.09795637e-01
2.33299419e-01 -7.32783735e-01 -3.62653941e-01 2.47568503e-01
6.66179419e-01 -1.82971552e-01 9.48769331e-01 -6.57179594e-01
-6.23075426e-01 -2.35620931e-01 -4.13708299e-01 6.75745726e-01
-3.01573873e-01 -4.04479384e-01 -7.97231853e-01 -5.74092329e-01
4.84462261e-01 -4.27496433e-01 1.05992055e+00 5.40081382e-01
1.94873631e+00 -8.22599113e-01 -1.67543545e-01 1.15237808e+00
1.32544601e+00 1.81646600e-01 4.71233904e-01 1.91163808e-01
8.70352328e-01 3.11309546e-01 3.35558981e-01 4.51257885e-01
1.95681695e-02 5.49207985e-01 5.11965394e-01 -1.53050898e-03
-6.10760748e-02 -1.53324038e-01 1.59691051e-01 1.60043323e+00
-2.34256864e-01 4.33475040e-02 -6.11039042e-01 3.23379993e-01
-1.84793591e+00 -1.03714561e+00 2.97200710e-01 2.06358004e+00
8.46077025e-01 -2.50237044e-02 -1.62498578e-01 5.08090973e-01
4.38672721e-01 3.60387623e-01 -9.49761569e-01 -1.35508478e-01
-1.03781827e-01 1.56467319e-01 6.29820466e-01 3.60957049e-02
-1.04310596e+00 5.44861794e-01 6.19083548e+00 1.35566509e+00
-1.32668197e+00 4.08841312e-01 8.89546454e-01 -7.51016974e-01
-1.57247171e-01 -3.35139275e-01 -9.26771581e-01 4.47758108e-01
1.15408909e+00 -1.36125877e-01 7.26156890e-01 1.36041141e+00
1.35953784e-01 3.74852300e-01 -1.24543619e+00 1.57138574e+00
2.15137042e-02 -1.72536373e+00 2.71313667e-01 1.01326056e-01
6.02111697e-01 2.99464822e-01 2.50211149e-01 3.73963386e-01
-3.42410803e-01 -1.08131552e+00 7.93407321e-01 2.49897569e-01
1.06695199e+00 -8.66567612e-01 7.91450202e-01 4.06783372e-01
-1.16380477e+00 -3.97842109e-01 -9.51090753e-01 -8.84942710e-02
-5.78307360e-03 1.01630569e+00 -3.81363928e-02 2.49753579e-01
1.02585661e+00 1.02127850e+00 -3.07458222e-01 9.44210410e-01
3.16267073e-01 6.18053496e-01 -4.43466306e-01 -1.65660769e-01
1.32868737e-01 -4.97142047e-01 2.77965605e-01 1.00185859e+00
6.53014123e-01 1.16800480e-01 -1.11940540e-01 6.12025142e-01
-2.20288798e-01 7.83842895e-03 -4.70381111e-01 -3.25216055e-01
6.65759206e-01 1.12698901e+00 -3.85504156e-01 -2.66395360e-01
-2.89248258e-01 7.50107467e-01 4.37937379e-01 3.01373303e-01
-6.57497108e-01 -5.58649838e-01 5.67544341e-01 2.31726896e-02
3.16809863e-01 -4.38291937e-01 -6.30596220e-01 -1.27668893e+00
3.29844326e-01 -7.63341486e-01 1.89814150e-01 -3.76497805e-01
-8.35993528e-01 6.37373865e-01 -1.76568013e-02 -1.38797021e+00
-6.14205636e-02 -6.64789557e-01 -1.66568741e-01 4.71646965e-01
-1.45453858e+00 -6.27412021e-01 -2.31367663e-01 4.85726386e-01
5.01085341e-01 -6.17497146e-01 6.47175670e-01 8.96333277e-01
-9.04590428e-01 8.81701827e-01 5.84666729e-01 2.17787549e-02
1.20035030e-01 -5.92691243e-01 -2.41022930e-02 7.41864204e-01
4.91266251e-02 7.19925463e-01 3.51662606e-01 -2.04004258e-01
-1.63755894e+00 -1.37012589e+00 7.78029382e-01 4.50863093e-01
5.31018674e-01 -3.74960452e-01 -1.08061671e+00 1.75544471e-01
-2.00269222e-01 1.68973327e-01 6.11140132e-01 2.20736414e-01
-2.49711752e-01 -6.31775916e-01 -8.99578512e-01 5.54338038e-01
1.22236204e+00 -6.27623856e-01 -5.00750057e-02 7.04419017e-01
1.11648810e+00 -4.14276779e-01 -1.09165680e+00 4.01569575e-01
5.23091257e-01 -9.36027110e-01 9.87556159e-01 -4.78676736e-01
7.77431369e-01 9.92827490e-02 -4.42952096e-01 -9.48120534e-01
-5.46898425e-01 -4.61475432e-01 -9.20632839e-01 1.00677490e+00
2.06121013e-01 -4.74401563e-01 1.14428151e+00 5.18092334e-01
-3.70666295e-01 -1.44228625e+00 -1.22625864e+00 -9.66356397e-01
-6.88474253e-02 -4.31454271e-01 5.28125584e-01 7.75195897e-01
-2.27453515e-01 2.92633921e-01 -5.30139625e-01 -1.41470283e-01
4.50351596e-01 -1.04464211e-01 1.96948841e-01 -1.15744221e+00
-3.53246510e-01 -5.33088744e-01 -3.68763834e-01 -1.44124269e+00
-2.09550727e-02 -8.92434955e-01 -2.86779374e-01 -1.48958039e+00
3.39252442e-01 -5.33015072e-01 -4.83293951e-01 6.07846916e-01
4.94650930e-01 1.84363708e-01 1.57987997e-01 6.95234954e-01
-6.01008773e-01 1.05926740e+00 1.18551660e+00 -2.92575508e-01
-1.42867282e-01 -1.82011902e-01 -6.30967259e-01 7.54918218e-01
8.23914230e-01 -4.43742901e-01 -5.61952472e-01 -9.13376629e-01
2.85700619e-01 -7.02321669e-03 -1.40862269e-02 -1.47323680e+00
4.39689815e-01 -9.13422182e-02 3.03668141e-01 -5.86134493e-01
6.22416973e-01 -9.71322000e-01 -9.55303758e-02 6.59196138e-01
-4.06136841e-01 -7.32933804e-02 6.37594983e-02 5.40253222e-01
-4.29202527e-01 -5.05221963e-01 9.20465708e-01 7.63474107e-02
-4.68542308e-01 7.65683711e-01 -4.39314917e-02 -1.95302591e-01
6.35068595e-01 -3.27374935e-01 -2.82845795e-01 -3.26753259e-01
-4.01577562e-01 -7.94214010e-02 4.14427146e-02 2.25854203e-01
7.64650106e-01 -1.60167933e+00 -3.71354073e-01 3.33483189e-01
-4.03963149e-01 3.58195752e-01 5.99496961e-01 7.43184805e-01
-6.51513278e-01 6.56440198e-01 -2.20236689e-01 -5.57873905e-01
-1.10193408e+00 4.93873596e-01 2.01878235e-01 -3.50466102e-01
-6.69644356e-01 9.16903913e-01 2.61441823e-02 1.33174844e-03
7.80266464e-01 -3.26157928e-01 -3.46791416e-01 1.33400438e-02
7.81172454e-01 5.41286290e-01 2.74010509e-01 -4.93169010e-01
-1.42453909e-01 5.09519458e-01 -1.97036937e-01 2.51908779e-01
1.67632604e+00 1.57003731e-01 -2.76116580e-01 2.69105166e-01
1.51774347e+00 -5.64738095e-01 -1.16910994e+00 -4.21958297e-01
-3.83392543e-01 -2.43777141e-01 6.76360548e-01 -1.92076817e-01
-1.73145795e+00 1.06641269e+00 7.60197639e-01 7.47846290e-02
1.40502059e+00 -2.76876718e-01 1.10660124e+00 9.75132942e-01
2.31509373e-01 -1.31098652e+00 2.01427802e-01 8.51261139e-01
9.61695194e-01 -8.60920846e-01 3.40551794e-01 -2.04054117e-01
-2.60268778e-01 1.09753704e+00 6.03011489e-01 -7.14094639e-02
8.44107449e-01 2.73872912e-01 -5.85597157e-01 -1.47441208e-01
-7.62811840e-01 3.47997218e-01 3.03327411e-01 3.90506417e-01
3.53247494e-01 -2.15627357e-01 -3.06687206e-01 6.78588390e-01
-5.30416012e-01 -9.72398594e-02 2.61986349e-02 6.81675434e-01
-6.61579669e-01 -7.88807154e-01 5.52306809e-02 1.07614088e+00
-3.56512755e-01 -2.84010917e-01 2.07848296e-01 3.04921269e-01
7.13349283e-02 6.03892863e-01 2.41731659e-01 -7.99139559e-01
1.71611682e-01 -1.25700668e-01 2.41149336e-01 -4.19835627e-01
-8.31437260e-02 -1.38154954e-01 -9.19755027e-02 -7.63336718e-01
-2.91496634e-01 -4.32743162e-01 -9.15323138e-01 -5.32784998e-01
-3.34648758e-01 3.77728343e-02 8.83273959e-01 8.76640737e-01
6.46127164e-01 6.09668612e-01 3.98334384e-01 -1.03089988e+00
-9.09342527e-01 -8.75282049e-01 -6.95236385e-01 1.95029303e-01
1.12640105e-01 -4.94487882e-01 -3.50800157e-01 -1.19864836e-01]
|
[8.485584259033203, 3.0683698654174805]
|
1b405a27-bf66-4c8e-9fe5-abd39026c768
|
conmae-contour-guided-mae-for-unsupervised
|
2302.05673
| null |
https://arxiv.org/abs/2302.05673v1
|
https://arxiv.org/pdf/2302.05673v1.pdf
|
ConMAE: Contour Guided MAE for Unsupervised Vehicle Re-Identification
|
Vehicle re-identification is a cross-view search task by matching the same target vehicle from different perspectives. It serves an important role in road-vehicle collaboration and intelligent road control. With the large-scale and dynamic road environment, the paradigm of supervised vehicle re-identification shows limited scalability because of the heavy reliance on large-scale annotated datasets. Therefore, the unsupervised vehicle re-identification with stronger cross-scene generalization ability has attracted more attention. Considering that Masked Autoencoder (MAE) has shown excellent performance in self-supervised learning, this work designs a Contour Guided Masked Autoencoder for Unsupervised Vehicle Re-Identification (ConMAE), which is inspired by extracting the informative contour clue to highlight the key regions for cross-view correlation. ConMAE is implemented by preserving the image blocks with contour pixels and randomly masking the blocks with smooth textures. In addition, to improve the quality of pseudo labels of vehicles for unsupervised re-identification, we design a label softening strategy and adaptively update the label with the increase of training steps. We carry out experiments on VeRi-776 and VehicleID datasets, and a significant performance improvement is obtained by the comparison with the state-of-the-art unsupervised vehicle re-identification methods. The code is available on the website of https://github.com/2020132075/ConMAE.
|
['Hongke Xu', 'Jianwu Fang', 'Jing Yang']
|
2023-02-11
| null | null | null | null |
['vehicle-re-identification']
|
['computer-vision']
|
[-9.12627578e-02 -2.75363535e-01 -3.27964693e-01 -4.26191121e-01
-4.77704853e-01 -2.85504371e-01 6.85173631e-01 -1.60623372e-01
-3.48728031e-01 2.96648145e-01 -1.95847563e-02 -2.05433160e-01
-1.40098843e-03 -7.94346631e-01 -5.85515797e-01 -9.26470935e-01
1.92372695e-01 3.79337370e-01 3.68238777e-01 -9.20358375e-02
1.93732575e-01 5.50015569e-01 -1.95835304e+00 -8.48674960e-03
9.29733276e-01 8.03666592e-01 3.36962610e-01 2.02361211e-01
-1.41116872e-01 3.81149679e-01 -1.40123919e-01 -3.98281306e-01
3.11109960e-01 -1.60788044e-01 -4.09469754e-01 3.16942811e-01
5.33986568e-01 -3.59743059e-01 -5.49662471e-01 1.43201208e+00
2.92738289e-01 3.24469805e-01 7.04033256e-01 -1.60401797e+00
-4.69368190e-01 3.26195747e-01 -6.11512125e-01 1.91304773e-01
-3.84166181e-01 7.12047741e-02 6.01484239e-01 -1.03934169e+00
4.16577965e-01 1.11708856e+00 6.75573945e-01 5.72157264e-01
-1.08529139e+00 -1.03308427e+00 6.28978387e-02 8.67617428e-01
-1.79761457e+00 -5.51324368e-01 1.07421088e+00 -7.31697917e-01
4.32165653e-01 1.63353026e-01 4.19497788e-01 8.01390052e-01
-1.37527004e-01 6.56780601e-01 9.78661001e-01 -1.02648698e-01
-8.93785581e-02 5.05197823e-01 1.86236620e-01 5.89254081e-01
2.72306621e-01 5.81980586e-01 -4.38023992e-02 2.54811972e-01
3.40100050e-01 3.52504671e-01 -2.60680951e-02 -3.48197073e-01
-1.12828481e+00 8.08136761e-01 3.16788107e-01 2.41327286e-01
-1.59300938e-01 -3.66881453e-02 5.49526453e-01 1.77315116e-01
3.40701193e-01 -2.43537903e-01 -1.12497725e-01 3.19480211e-01
-6.92610204e-01 -1.82835624e-01 1.04517616e-01 8.83631051e-01
1.35862792e+00 3.92127931e-01 1.98686481e-01 1.02948558e+00
3.59006345e-01 7.78655052e-01 6.13701761e-01 -6.81961834e-01
2.22466245e-01 7.07319498e-01 -1.28552377e-01 -1.31974423e+00
-2.45190591e-01 -3.94192278e-01 -1.09371507e+00 5.84910691e-01
2.54029576e-02 7.50082508e-02 -9.38299775e-01 1.52645290e+00
4.94540572e-01 4.93515551e-01 4.09924865e-01 8.34196866e-01
1.12406313e+00 7.40680158e-01 2.31002316e-01 -2.10260414e-03
1.31071007e+00 -9.87916470e-01 -7.10090637e-01 -1.97072104e-01
5.20795643e-01 -7.17475176e-01 5.85562527e-01 -9.57322419e-02
-2.96738178e-01 -1.13468719e+00 -1.15550768e+00 3.98136050e-01
-7.78636336e-01 3.45659822e-01 2.11035892e-01 6.05898857e-01
-8.31409872e-01 1.15446374e-01 -3.81723076e-01 -3.83795261e-01
2.64373928e-01 3.04125041e-01 -6.63719058e-01 3.21466662e-02
-1.27951252e+00 6.85108364e-01 7.02290893e-01 2.56487429e-01
-8.20067465e-01 -5.18453956e-01 -9.04660702e-01 -1.95385829e-01
3.11932862e-01 -7.57555710e-03 7.12619007e-01 -1.11447942e+00
-1.32862031e+00 9.82008457e-01 -1.75778970e-01 -3.88445705e-01
4.19931829e-01 3.46298754e-01 -1.01235533e+00 2.18191314e-02
2.78457552e-01 8.60116959e-01 1.08676720e+00 -1.64895105e+00
-9.64518964e-01 -1.92340821e-01 -3.84011090e-01 4.74782735e-02
-3.84757519e-01 -1.91405844e-02 -5.77065289e-01 -7.70365596e-01
1.77731141e-02 -1.16602218e+00 1.46454707e-01 -5.55340171e-01
-3.03847998e-01 -2.74537653e-01 1.34540868e+00 -8.35033715e-01
1.13560295e+00 -2.44347262e+00 -3.42632502e-01 3.11391294e-01
2.49401748e-01 3.95759463e-01 -1.07852623e-01 1.35408223e-01
-3.99647623e-01 1.10381983e-01 -2.36612707e-01 -2.10313141e-01
-5.85050024e-02 4.72072065e-02 -4.54626679e-02 5.43392360e-01
1.58538949e-02 8.24097812e-01 -5.61818719e-01 -8.12894344e-01
5.49524605e-01 2.29715347e-01 -6.30965233e-02 9.55268741e-02
2.25145459e-01 3.28303635e-01 -3.69113982e-01 7.55494535e-01
1.18386972e+00 -7.32652694e-02 -7.80224800e-02 -5.97717106e-01
-4.30529952e-01 -6.12055540e-01 -1.29950762e+00 7.47989297e-01
-1.81802854e-01 9.70388114e-01 7.40446225e-02 -1.10347223e+00
1.08773518e+00 1.47814557e-01 6.20274484e-01 -1.05615199e+00
2.17558503e-01 1.71296149e-01 -2.19983727e-01 -5.78306913e-01
5.45995235e-01 2.53352344e-01 1.39993830e-02 1.86600983e-01
-2.59294778e-01 4.03078079e-01 1.80610642e-01 -6.11578021e-03
3.42482984e-01 -4.28456604e-01 -1.08272910e-01 -3.87098819e-01
1.06877553e+00 1.79067880e-01 5.73027968e-01 2.21004471e-01
-3.53318125e-01 2.11295471e-01 -3.78413230e-01 -5.47469795e-01
-1.15061724e+00 -6.65153623e-01 -4.17307377e-01 8.33031237e-01
7.39122093e-01 1.14111096e-01 -7.75269508e-01 -6.42868161e-01
2.73341686e-01 4.98264760e-01 -6.92776799e-01 -2.62752056e-01
-6.67304933e-01 -6.11155033e-01 4.63603526e-01 5.41533530e-01
1.04245567e+00 -9.70040798e-01 -6.83497712e-02 -2.82300636e-02
-1.32472113e-01 -1.19216728e+00 -6.15708113e-01 -2.46193856e-01
-5.28241515e-01 -1.06428134e+00 -6.45390868e-01 -1.38897657e+00
8.41931462e-01 7.58730233e-01 6.34668410e-01 2.60179132e-01
-1.73234642e-02 1.79427639e-01 -3.05453390e-01 -9.77728963e-02
-6.85054898e-01 4.40108962e-02 3.36555153e-01 5.80103338e-01
6.24866903e-01 -3.00274044e-01 -5.27531445e-01 8.58892143e-01
-6.89813733e-01 2.28663117e-01 5.46631515e-01 8.87172043e-01
8.21590364e-01 5.19059002e-01 5.85446060e-01 -5.10812938e-01
1.25460193e-01 -4.52833027e-01 -8.85586083e-01 8.89794454e-02
-9.09789920e-01 -1.24357454e-01 6.18340075e-01 -4.56118584e-01
-1.19800246e+00 5.66130131e-02 -1.50839671e-01 -5.97856879e-01
-4.94042307e-01 2.49967963e-01 -3.90475631e-01 -3.89682829e-01
2.16333464e-01 4.58630055e-01 2.32879281e-01 -4.52843428e-01
3.10661703e-01 9.64787066e-01 6.39685392e-01 -5.07836826e-02
1.12675846e+00 5.01757383e-01 -2.63825148e-01 -8.69066477e-01
-1.51821254e-02 -6.63852990e-01 -4.49299723e-01 -6.66655958e-01
1.11860716e+00 -1.01597989e+00 -8.25026751e-01 7.09657669e-01
-8.37440491e-01 -2.99626052e-01 7.59663433e-02 6.37605071e-01
-2.04901338e-01 5.27765572e-01 -1.69137999e-01 -4.86954212e-01
-2.64169306e-01 -1.23196375e+00 6.99200630e-01 6.00925148e-01
3.75644207e-01 -7.12671459e-01 8.14940035e-02 4.56617266e-01
3.55051607e-01 -5.45110404e-02 6.73447669e-01 -7.15099037e-01
-7.38743901e-01 -2.28942797e-01 -5.05815327e-01 4.34827268e-01
1.32645682e-01 -4.02685627e-02 -8.31441462e-01 -2.90159017e-01
-3.98306936e-01 9.99059677e-02 9.83578920e-01 1.26743674e-01
9.67917502e-01 -2.36354336e-01 -6.10216200e-01 7.25711524e-01
1.37661982e+00 5.34412384e-01 5.68806469e-01 6.42134070e-01
1.03678977e+00 7.88702309e-01 7.06161737e-01 1.20555364e-01
5.85828245e-01 6.95405543e-01 3.91086191e-01 -2.50603855e-01
-2.51491547e-01 -2.24924192e-01 1.87952682e-01 8.60300422e-01
1.07464097e-01 -8.82084444e-02 -9.47490573e-01 8.24940383e-01
-1.78744805e+00 -1.22406530e+00 -1.71314210e-01 1.98416483e+00
4.13209707e-01 5.66712022e-02 8.57948288e-02 1.43841162e-01
1.25021660e+00 1.73814684e-01 -5.88250399e-01 -8.44797269e-02
-3.25550705e-01 -4.99738544e-01 9.37945008e-01 5.57771742e-01
-1.44060194e+00 9.77776229e-01 4.94370222e+00 1.39268684e+00
-1.21728706e+00 1.61013335e-01 6.23539209e-01 8.76994789e-01
-1.72592387e-01 -1.51078328e-01 -9.22097981e-01 7.28681326e-01
5.92733443e-01 -1.56328022e-01 3.72058630e-01 9.67283607e-01
2.35729784e-01 1.00654006e-01 -6.07538640e-01 1.18897700e+00
1.44216478e-01 -1.29168904e+00 -9.66947153e-02 3.91004458e-02
9.04091120e-01 2.25865856e-01 9.53384563e-02 1.08282626e-01
1.76568091e-01 -7.89865732e-01 6.87967479e-01 5.55077732e-01
9.20372903e-01 -8.23263824e-01 9.07049716e-01 8.77022445e-02
-1.69874787e+00 -9.38213468e-02 -1.28821135e-01 6.30106270e-01
2.64119450e-02 2.27678597e-01 -6.44962192e-01 4.06248212e-01
8.01101267e-01 8.22781205e-01 -7.74095595e-01 9.71674800e-01
1.46763831e-01 5.95416605e-01 -1.89417183e-01 1.29375488e-01
-5.07315658e-02 -4.16817278e-01 6.50693417e-01 1.31713963e+00
1.57078564e-01 -1.62972286e-01 2.33778194e-01 6.40720546e-01
9.96691883e-02 1.45371303e-01 -4.97542679e-01 5.36995754e-02
6.73229635e-01 1.18310249e+00 -6.81571484e-01 -5.24529696e-01
-4.18612093e-01 7.03255951e-01 -2.09910676e-01 5.70744812e-01
-9.58472788e-01 -3.33561093e-01 6.48692966e-01 8.55035111e-02
5.18311501e-01 -1.52014539e-01 -5.50495088e-02 -7.21963704e-01
-2.06196710e-01 -7.71590412e-01 3.09227198e-01 -5.66411614e-01
-1.20731926e+00 7.97837257e-01 1.41697228e-01 -1.67111647e+00
-5.16330935e-02 -4.62573797e-01 -5.93030691e-01 5.23777664e-01
-1.65893734e+00 -1.31725419e+00 -7.73999631e-01 6.21579230e-01
6.04689002e-01 -6.85990930e-01 4.12005007e-01 7.32856333e-01
-7.87283480e-01 8.91521931e-01 5.13666868e-01 2.69939929e-01
6.13333464e-01 -7.00794101e-01 3.12785715e-01 8.75115633e-01
-1.97414964e-01 2.08268568e-01 5.09493530e-01 -7.61749387e-01
-1.15857279e+00 -1.43855309e+00 6.96070015e-01 9.17677581e-02
5.49988389e-01 -8.49125758e-02 -9.24069643e-01 4.72723067e-01
2.32267514e-01 -6.26817048e-02 3.28304619e-01 -4.78405446e-01
-2.66072690e-01 -6.85382545e-01 -9.31059837e-01 4.96406436e-01
7.79397011e-01 -6.08244359e-01 -2.08472475e-01 2.15270724e-02
5.74516416e-01 -2.97076013e-02 -6.79819703e-01 7.52710283e-01
5.57881653e-01 -7.66972601e-01 1.01018083e+00 2.43668724e-02
-3.30440849e-02 -8.45512629e-01 -1.08491741e-01 -9.28393781e-01
-4.70970958e-01 -1.95335954e-01 3.06698382e-01 1.52651668e+00
1.49772897e-01 -9.15504038e-01 9.74037588e-01 2.08035305e-01
-1.53886512e-01 -3.53146136e-01 -8.35887611e-01 -6.81666195e-01
-2.56342709e-01 -2.30839998e-01 7.87236750e-01 9.77583945e-01
-5.14650285e-01 -4.41863202e-02 -5.27844906e-01 3.40427667e-01
8.75101268e-01 2.30791807e-01 1.01449013e+00 -1.24764621e+00
2.57554591e-01 -5.94482839e-01 -6.57206416e-01 -1.01258516e+00
4.71871734e-01 -8.98710072e-01 1.71662837e-01 -9.87060547e-01
3.09694141e-01 -6.70299411e-01 -3.13265294e-01 2.57277399e-01
-5.02050892e-02 3.70652199e-01 6.21477962e-02 4.80617523e-01
-6.25660658e-01 7.35186338e-01 8.73473287e-01 -6.46333098e-01
-4.20806222e-02 1.93662979e-02 -3.43494534e-01 6.84233129e-01
1.19481874e+00 -4.38516617e-01 -2.78860420e-01 -3.13644201e-01
-3.32187623e-01 -2.98124850e-01 4.64805752e-01 -1.09964514e+00
5.11881888e-01 -1.91365302e-01 1.23380296e-01 -1.12629175e+00
6.35967702e-02 -1.21447122e+00 6.17648840e-01 3.30042541e-01
1.50805572e-02 2.18901694e-01 3.54384452e-01 7.39388406e-01
-5.13248682e-01 -2.52368391e-01 8.33010316e-01 1.27245292e-01
-1.54854214e+00 4.20467764e-01 -5.32683849e-01 -2.68092722e-01
1.20205402e+00 -6.98315382e-01 -2.02872753e-01 -5.03743105e-02
-3.60869676e-01 4.51150954e-01 7.11599946e-01 6.34410381e-01
7.34244525e-01 -1.66921473e+00 -7.40229905e-01 5.12297630e-01
5.03790200e-01 -3.36854130e-01 8.59136939e-01 6.03804946e-01
-4.45513308e-01 3.55984002e-01 -2.47117773e-01 -6.47608161e-01
-1.46361566e+00 7.52560616e-01 3.30873132e-01 1.05438136e-01
-5.72025716e-01 4.42331374e-01 4.49675798e-01 -3.23955357e-01
2.47433364e-01 3.38270366e-01 -6.92128956e-01 1.61283121e-01
3.75727057e-01 5.77474415e-01 -5.83970398e-02 -1.46117401e+00
-5.29530585e-01 1.05184162e+00 7.63916643e-03 2.67798215e-01
8.96222830e-01 -5.58213353e-01 3.00430767e-02 3.19732609e-03
1.49716747e+00 1.30446419e-01 -1.18316233e+00 -3.70514572e-01
-1.67989895e-01 -4.25159812e-01 2.52517909e-01 -3.05379480e-01
-1.43308628e+00 5.35373449e-01 1.18774474e+00 9.31308139e-03
1.04639220e+00 -5.77206537e-02 8.86916935e-01 2.66740859e-01
9.77993831e-02 -1.18098795e+00 -2.62993369e-02 3.19409221e-01
6.68000042e-01 -1.73248351e+00 -2.18315572e-01 -3.57977629e-01
-7.19765782e-01 9.38293874e-01 7.22289622e-01 6.57938272e-02
8.03502083e-01 8.37258697e-02 2.58499414e-01 -9.85077024e-02
-1.02522656e-01 -4.78662968e-01 2.87153184e-01 6.34210765e-01
-3.81673872e-01 2.63931930e-01 -4.39352505e-02 3.97806048e-01
4.21900786e-02 -3.11454713e-01 -4.44521457e-02 3.86447877e-01
-5.00088155e-01 -9.58392978e-01 -4.80162799e-01 1.53709844e-01
1.83326248e-02 2.67781205e-02 -6.23215083e-03 7.93012321e-01
5.12180865e-01 1.06703472e+00 2.19570532e-01 -1.02693212e+00
9.31573957e-02 -1.67075574e-01 -2.40125135e-01 4.03326340e-02
-1.31404772e-01 -4.51549701e-02 -5.35184816e-02 -1.60295784e-01
-6.73432171e-01 -6.22286141e-01 -1.33157420e+00 -5.78280985e-01
-4.08519745e-01 2.16651887e-01 6.17289960e-01 6.79470778e-01
4.93442535e-01 2.54690975e-01 1.08738267e+00 -7.20703006e-01
-1.47135668e-02 -7.43484378e-01 -3.66339296e-01 5.81469297e-01
4.95745838e-01 -8.83079529e-01 -4.55862820e-01 3.12793523e-01]
|
[8.144396781921387, -0.9949756860733032]
|
2b912f4a-1d62-4eb0-921f-6bf25a601c94
|
sapi-surroundings-aware-vehicle-trajectory
|
2306.01812
| null |
https://arxiv.org/abs/2306.01812v1
|
https://arxiv.org/pdf/2306.01812v1.pdf
|
SAPI: Surroundings-Aware Vehicle Trajectory Prediction at Intersections
|
In this work we propose a deep learning model, i.e., SAPI, to predict vehicle trajectories at intersections. SAPI uses an abstract way to represent and encode surrounding environment by utilizing information from real-time map, right-of-way, and surrounding traffic. The proposed model consists of two convolutional network (CNN) and recurrent neural network (RNN)-based encoders and one decoder. A refiner is proposed to conduct a look-back operation inside the model, in order to make full use of raw history trajectory information. We evaluate SAPI on a proprietary dataset collected in real-world intersections through autonomous vehicles. It is demonstrated that SAPI shows promising performance when predicting vehicle trajectories at intersection, and outperforms benchmark methods. The average displacement error(ADE) and final displacement error(FDE) for 6-second prediction are 1.84m and 4.32m respectively. We also show that the proposed model can accurately predict vehicle trajectories in different scenarios.
|
['Lei Wang', 'Yiqian Gan', 'Hao Xiao', 'Ethan Zhang']
|
2023-06-02
| null | null | null | null |
['trajectory-prediction', 'autonomous-vehicles']
|
['computer-vision', 'computer-vision']
|
[-1.88308984e-01 1.61286891e-02 -1.58728153e-01 -6.91768825e-01
-6.53444886e-01 -1.20286651e-01 6.11190736e-01 -6.46345643e-03
-4.19814259e-01 6.91533327e-01 3.07129711e-01 -7.48951375e-01
1.34172529e-01 -1.17570627e+00 -1.14637995e+00 -4.09508139e-01
-4.58760470e-01 2.25550979e-01 5.00895739e-01 -3.01491946e-01
6.58152625e-02 6.14021242e-01 -1.55653965e+00 1.38616428e-01
6.44651949e-01 9.45749938e-01 2.31577173e-01 6.15755856e-01
3.33654433e-02 9.92284656e-01 -3.32312047e-01 -1.79334626e-01
1.93135202e-01 2.38201767e-01 -5.28823853e-01 -6.45025313e-01
-1.37931341e-02 -8.66790235e-01 -1.23136342e+00 5.64731061e-01
2.46240109e-01 3.20672333e-01 4.96198058e-01 -1.47100222e+00
-2.72559285e-01 4.30308253e-01 -8.91505480e-02 4.04826760e-01
-1.37496188e-01 3.03234309e-01 2.87434071e-01 -8.05563271e-01
4.51007307e-01 1.08170283e+00 9.82822835e-01 3.37000728e-01
-7.44029939e-01 -1.04782438e+00 1.10665135e-01 5.10985434e-01
-1.66315937e+00 -4.66324538e-01 5.05158544e-01 -4.89982337e-01
1.31919742e+00 2.21574511e-02 4.42358375e-01 9.55741167e-01
7.19188273e-01 8.83052766e-01 2.30452508e-01 3.05174351e-01
6.53830543e-02 7.21828267e-03 1.66083455e-01 3.66007626e-01
9.95678604e-02 5.81836820e-01 -8.09217840e-02 3.46277863e-01
3.68885338e-01 3.13366532e-01 1.48595572e-01 4.16067481e-01
-9.57320392e-01 5.38045406e-01 7.74688721e-01 -1.16827875e-01
-6.67245448e-01 5.07136345e-01 4.83425260e-01 2.69233417e-02
3.72332364e-01 -3.31228137e-01 -2.06602991e-01 -5.99395275e-01
-8.61662805e-01 4.66330111e-01 3.37641031e-01 1.30448198e+00
9.63188589e-01 3.47365916e-01 -2.52753764e-01 5.06561220e-01
4.19127136e-01 8.37285817e-01 2.77929008e-01 -8.80101323e-01
8.83756518e-01 3.41719210e-01 3.06268811e-01 -1.21906197e+00
-6.61796093e-01 -3.05501580e-01 -9.91650760e-01 1.25250220e-01
-7.03188181e-02 -5.39114475e-01 -1.03830338e+00 1.62431753e+00
5.39894821e-03 1.01349616e+00 2.52495974e-01 7.95495808e-01
9.93513703e-01 1.23953032e+00 1.75227910e-01 3.44304442e-01
5.18401861e-01 -1.05449891e+00 -6.15716636e-01 -1.06093414e-01
9.46839154e-01 -1.18280664e-01 3.29575330e-01 -6.11478128e-02
-1.02253032e+00 -9.01081622e-01 -1.11415339e+00 1.10467978e-01
-6.90585971e-01 1.72219351e-01 -6.74086250e-03 2.18752638e-01
-1.08431101e+00 5.56984723e-01 -8.51574123e-01 -2.37216681e-01
2.85639077e-01 4.61436063e-01 -2.73431987e-01 1.16314843e-01
-1.56587648e+00 9.52720046e-01 3.25766981e-01 5.04970551e-01
-1.21639693e+00 -8.19610775e-01 -1.06318343e+00 1.53563485e-01
-4.75563342e-03 -1.67394653e-01 1.19315577e+00 -1.64957866e-01
-1.44048250e+00 1.47248909e-01 -4.39013988e-01 -1.09568775e+00
2.71587878e-01 -3.15302461e-01 -9.61197257e-01 -3.55499297e-01
1.18896596e-01 7.76481748e-01 1.98554412e-01 -1.14930701e+00
-1.07502735e+00 -7.39944950e-02 -1.40320644e-01 -1.62244007e-01
1.83841273e-01 -2.23131359e-01 -6.06509864e-01 -2.42688075e-01
-2.02024370e-01 -1.13709509e+00 -5.62994599e-01 -5.89435339e-01
-4.61936921e-01 -5.77803664e-02 1.12010884e+00 -9.29473042e-01
1.58995140e+00 -2.01101232e+00 -6.26978993e-01 4.28689450e-01
1.54481456e-01 7.67906368e-01 -2.47801557e-01 5.69328785e-01
1.96359828e-01 -7.88198784e-02 2.24019945e-01 -3.57090533e-01
-2.88784634e-02 4.61975008e-01 -4.34100330e-01 3.60714585e-01
2.44544983e-01 1.30198300e+00 -8.03363681e-01 -6.90100119e-02
5.40606499e-01 6.54213130e-01 -3.36981177e-01 1.67095333e-01
1.71904087e-01 4.00164366e-01 -4.95206773e-01 2.16025472e-01
9.99899507e-01 1.24154076e-01 -3.04821342e-01 1.04430497e-01
-6.10968351e-01 5.02912700e-01 -9.18368042e-01 1.21322727e+00
-5.67535639e-01 1.27384830e+00 -3.26999635e-01 -6.84078693e-01
1.13010859e+00 2.56903857e-01 3.49133849e-01 -1.45101595e+00
1.19161710e-01 1.39652759e-01 -3.16359401e-01 -4.81243163e-01
9.37344968e-01 5.81877112e-01 -1.11700140e-01 6.35112152e-02
-3.63869131e-01 8.50728750e-01 -5.05745634e-02 4.72294874e-02
1.15169048e+00 -1.44083053e-01 -2.62294739e-01 4.71094474e-02
6.13065362e-01 4.68399934e-02 8.08331192e-01 5.25583804e-01
-1.62613526e-01 3.84100765e-01 1.69599354e-01 -8.64962578e-01
-1.20204043e+00 -9.29322183e-01 4.77657504e-02 5.69575608e-01
4.71911013e-01 -3.20751965e-01 -5.27082920e-01 -3.13945472e-01
6.40992150e-02 9.55867946e-01 -5.23851454e-01 -3.19312125e-01
-1.01586413e+00 -2.51072526e-01 7.40380406e-01 1.03268802e+00
7.82118976e-01 -7.73911417e-01 -6.79173231e-01 5.68016469e-01
-2.64195595e-02 -1.24632716e+00 -1.85202032e-01 -2.36238375e-01
-4.38101560e-01 -6.39479816e-01 -3.48747998e-01 -6.38028324e-01
3.14447820e-01 4.96321917e-01 7.58880973e-01 -1.24548033e-01
3.73397529e-01 -1.94755569e-01 -1.79684013e-01 -3.72029662e-01
-4.57758725e-01 2.89515853e-01 5.22029512e-02 2.35347338e-02
5.68254471e-01 -5.44375777e-01 -7.36230373e-01 4.76243228e-01
-4.16481733e-01 2.54176378e-01 5.03948212e-01 4.69722778e-01
5.74069142e-01 2.18780532e-01 7.11715221e-01 -3.59133899e-01
3.13285083e-01 -9.78502691e-01 -6.92660630e-01 -1.85632586e-01
-6.46984935e-01 -7.21020177e-02 6.28161728e-01 2.38545537e-02
-9.60145116e-01 -5.07198833e-02 -6.93420351e-01 -5.65447092e-01
-3.23576212e-01 5.95902026e-01 -5.90915270e-02 2.90080518e-01
1.88855737e-01 3.24204355e-01 -1.38154864e-01 -3.82957160e-01
2.30733573e-01 9.30521369e-01 8.75598967e-01 9.00695920e-02
6.19941413e-01 2.96684384e-01 -4.97492217e-02 -6.87624395e-01
-9.25866291e-02 -3.45857292e-01 -4.68491465e-01 -5.49472153e-01
7.82659888e-01 -1.27579296e+00 -1.04080522e+00 4.79928493e-01
-1.38759530e+00 -6.50985658e-01 6.91105351e-02 6.55406237e-01
-4.66340631e-01 -2.07106993e-01 -6.38049245e-01 -7.41324782e-01
-3.60166103e-01 -1.24908793e+00 9.76088822e-01 2.71219730e-01
8.41149464e-02 -8.08508098e-01 1.17702084e-02 -9.62125659e-02
7.18516350e-01 4.33019906e-01 5.25181770e-01 -5.74680567e-01
-1.00479007e+00 -4.42034304e-01 -4.84717786e-01 1.44530237e-01
-2.19861954e-01 -9.97161865e-02 -7.69574225e-01 -1.14078455e-01
-7.10687637e-01 2.89003342e-01 8.69332135e-01 5.19558489e-01
1.12212265e+00 -5.41631520e-01 -8.51732075e-01 6.41815424e-01
1.44789541e+00 7.98068821e-01 1.11107695e+00 3.31102818e-01
7.39551604e-01 2.61252671e-01 6.45380676e-01 4.05914485e-01
1.00445282e+00 8.33743751e-01 6.48649633e-01 2.99217235e-02
-1.99807525e-01 -6.43268466e-01 3.31917256e-01 7.89907277e-01
8.50530863e-02 -5.15441358e-01 -1.36390996e+00 9.16477919e-01
-2.10753632e+00 -1.17901933e+00 -5.32744586e-01 1.97068489e+00
2.05660667e-02 3.39082479e-01 1.30317524e-01 -6.43247217e-02
4.75996286e-01 9.45304185e-02 -5.50604045e-01 -5.97455978e-01
2.85874933e-01 -2.25957721e-01 1.09505391e+00 6.87762737e-01
-1.09160519e+00 1.02479839e+00 6.04522848e+00 5.39058149e-01
-1.36840308e+00 3.96800712e-02 6.32939100e-01 -5.20418361e-02
-9.94516835e-02 -3.67009938e-01 -9.95453119e-01 9.28425372e-01
1.85382485e+00 -4.80093844e-02 1.62309334e-01 8.63369644e-01
6.64139986e-01 2.62139559e-01 -8.38191569e-01 8.34070623e-01
-4.42985862e-01 -1.79454124e+00 -1.23998746e-01 7.53095597e-02
5.69804311e-01 7.42685318e-01 1.82119429e-01 6.95681572e-01
2.03768805e-01 -1.23563695e+00 6.72930419e-01 9.92710233e-01
7.54307926e-01 -1.14597702e+00 1.04591823e+00 5.05507827e-01
-1.55880415e+00 -2.36736462e-01 -1.41455457e-01 -1.45167708e-01
6.24975204e-01 1.11828268e-01 -9.69748259e-01 4.99279737e-01
7.07669079e-01 1.12890244e+00 -3.12235087e-01 1.07525027e+00
2.95978099e-01 8.81782591e-01 -2.96788573e-01 4.85438667e-02
7.92228341e-01 1.29438443e-02 3.61754656e-01 1.55891335e+00
5.59443831e-01 1.22167647e-01 2.30568573e-02 6.98780358e-01
1.46298468e-01 -4.52388167e-01 -9.25849140e-01 5.70652008e-01
7.90396392e-01 7.10429192e-01 -1.09599670e-02 -3.43237907e-01
-4.71144140e-01 4.63614970e-01 6.27213344e-02 7.15269804e-01
-1.33527982e+00 -6.64442778e-01 1.10092175e+00 2.59788394e-01
5.58311999e-01 -4.31485593e-01 -1.35250494e-01 -4.44028139e-01
-3.60011235e-02 -1.40716702e-01 -3.19521308e-01 -8.56728017e-01
-5.00993311e-01 7.21293449e-01 3.73080932e-02 -1.45539534e+00
-4.99350935e-01 -4.27864492e-01 -9.38953161e-01 9.03869152e-01
-1.68805432e+00 -1.09331024e+00 -3.94272178e-01 3.25770736e-01
5.61746955e-01 -3.68917555e-01 3.27295423e-01 9.55598772e-01
-1.05699444e+00 7.48891175e-01 4.64406431e-01 4.27197456e-01
4.21086233e-03 -5.91908634e-01 1.20438111e+00 7.97309160e-01
-3.11634064e-01 2.94175208e-01 5.92101514e-01 -6.47447169e-01
-1.30479383e+00 -2.06932425e+00 1.11790562e+00 -2.31992885e-01
3.57987881e-01 -8.81721824e-02 -8.99698973e-01 1.04899025e+00
1.41360253e-01 2.67807161e-03 9.97390002e-02 -3.47900391e-01
1.31884694e-01 -3.69996130e-01 -8.18393409e-01 6.03851676e-01
9.46972191e-01 -3.03419679e-01 -1.42049612e-02 -1.36713564e-01
9.83531833e-01 -5.80206871e-01 -7.57990003e-01 5.47333956e-01
6.80086792e-01 -8.17756176e-01 9.50199127e-01 -4.61354494e-01
2.73930311e-01 -3.97780508e-01 -2.52292484e-01 -1.19831526e+00
-4.75368381e-01 -3.76734883e-01 -3.08105081e-01 9.55498397e-01
7.12103605e-01 -7.21715152e-01 8.30329835e-01 7.36800790e-01
-7.05921233e-01 -1.13095808e+00 -1.00683355e+00 -6.86788619e-01
4.13039811e-02 -1.00922894e+00 1.36622822e+00 2.29750693e-01
-3.49866927e-01 9.27060544e-02 -6.22345924e-01 5.97357452e-01
1.22673377e-01 -3.19061249e-01 1.01333606e+00 -7.86651611e-01
5.65565467e-01 -3.41947019e-01 -9.39151347e-01 -1.60195053e+00
2.71589130e-01 -6.27729774e-01 9.79707614e-02 -1.69326293e+00
-2.79701203e-01 -6.16286099e-01 -4.84794885e-01 2.73146749e-01
3.45774859e-01 5.33634685e-02 7.39905378e-03 -3.94987036e-03
-6.76926196e-01 7.18723178e-01 6.71041608e-01 -3.51463437e-01
-3.02931249e-01 2.93752730e-01 -1.32371873e-01 5.17384529e-01
1.14129424e+00 -3.46571237e-01 -5.40509045e-01 -8.03700328e-01
-2.16048136e-02 3.32259595e-01 5.98722816e-01 -1.48641455e+00
7.06633389e-01 4.11003921e-03 1.87515303e-01 -1.53892195e+00
4.01856989e-01 -8.18799198e-01 4.04808849e-01 3.31712216e-01
-2.91303307e-01 4.12178636e-01 5.77561438e-01 7.86011636e-01
-4.03840184e-01 3.65259379e-01 2.08285704e-01 4.80335563e-01
-1.21101439e+00 7.09660649e-01 -7.03222632e-01 -4.02281135e-01
1.12615967e+00 -3.28917623e-01 -2.37883121e-01 -6.50736392e-01
-3.90846074e-01 7.88491189e-01 1.87363345e-02 7.41057515e-01
9.51708078e-01 -1.82390046e+00 -6.83342099e-01 4.85707104e-01
6.27565309e-02 -9.69491675e-02 5.98026693e-01 6.41573489e-01
-6.76003397e-01 1.07254982e+00 -1.10801466e-01 -5.51195204e-01
-8.14118445e-01 4.52874601e-01 4.17089462e-01 -9.60083865e-03
-9.01772738e-01 4.41110343e-01 -1.34889141e-01 -3.74226749e-01
3.20877999e-01 -7.17573762e-01 -2.72853076e-01 -4.89679635e-01
7.92673767e-01 7.53688157e-01 1.02169931e-01 -1.25961423e+00
-3.26665163e-01 3.37548941e-01 7.67790228e-02 1.57505106e-02
1.11724150e+00 -4.09322798e-01 4.52751577e-01 2.07978457e-01
1.55930173e+00 -5.75200498e-01 -1.57770479e+00 -3.21872175e-01
-8.64297003e-02 -2.57956088e-01 1.30228177e-01 -6.38401151e-01
-1.11177647e+00 1.01374352e+00 7.45639324e-01 1.02956891e-02
9.51205194e-01 -3.58681858e-01 1.69461119e+00 4.30310041e-01
4.18060601e-01 -9.87314045e-01 -6.56341612e-01 8.98393333e-01
6.13624096e-01 -1.25278676e+00 -6.64121330e-01 8.38402510e-02
-6.79718912e-01 7.84197628e-01 7.38367498e-01 -9.65118334e-02
1.00326777e+00 1.97971761e-01 -1.67999119e-02 -7.54274279e-02
-1.00855970e+00 -1.63561136e-01 1.76745147e-01 4.77001727e-01
-5.97827025e-02 1.91361174e-01 3.09715867e-01 6.25406444e-01
-2.59194583e-01 1.23231724e-01 2.59041965e-01 6.25393152e-01
-5.30832648e-01 -5.64665318e-01 9.75893289e-02 4.43436742e-01
2.47042831e-02 3.17221671e-01 3.71400028e-01 7.84278929e-01
2.94244111e-01 1.09970725e+00 5.83742559e-01 -1.16089380e+00
4.58689213e-01 -1.59110978e-01 -3.61647069e-01 -3.99318617e-03
-2.62261450e-01 -7.38048851e-01 2.82317579e-01 -1.01502490e+00
1.43276509e-02 -4.55876559e-01 -1.58154786e+00 -8.26052248e-01
1.88713834e-01 2.56821383e-02 7.51220942e-01 8.68218541e-01
9.54111934e-01 7.26074398e-01 7.89095163e-01 -9.01441276e-01
1.86592266e-02 -7.18262792e-01 -2.97451615e-01 4.12504151e-02
8.78188431e-01 -5.86261868e-01 4.03164238e-01 -2.23981753e-01]
|
[6.0944976806640625, 1.2894493341445923]
|
e3313ec2-9da4-40d5-a14a-ff5ada42feac
|
the-ethics-of-ai-in-games
|
2305.07392
| null |
https://arxiv.org/abs/2305.07392v1
|
https://arxiv.org/pdf/2305.07392v1.pdf
|
The Ethics of AI in Games
|
Video games are one of the richest and most popular forms of human-computer interaction and, hence, their role is critical for our understanding of human behaviour and affect at a large scale. As artificial intelligence (AI) tools are gradually adopted by the game industry a series of ethical concerns arise. Such concerns, however, have so far not been extensively discussed in a video game context. Motivated by the lack of a comprehensive review of the ethics of AI as applied to games, we survey the current state of the art in this area and discuss ethical considerations of these systems from the holistic perspective of the affective loop. Through the components of this loop, we study the ethical challenges that AI faces in video game development. Elicitation highlights the ethical boundaries of artificially induced emotions; sensing showcases the trade-off between privacy and safe gaming spaces; and detection, as utilised during in-game adaptation, poses challenges to transparency and ownership. This paper calls for an open dialogue and action for the games of today and the virtual spaces of the future. By setting an appropriate framework we aim to protect users and to guide developers towards safer and better experiences for their customers.
|
['Georgios N. Yannakakis', 'Christoffer Holmgård', 'Benedikte Mikkelsen', 'Julian Togelius', 'David Melhart']
|
2023-05-12
| null | null | null | null |
['ethics']
|
['miscellaneous']
|
[ 3.22869211e-01 6.60552382e-01 4.41170335e-01 1.62763931e-02
6.41854405e-02 -7.73963094e-01 4.62827951e-01 -3.42928953e-02
-5.23935258e-01 5.27866721e-01 1.76040217e-01 -3.11622322e-01
-4.79116850e-02 -5.92826962e-01 -2.77234256e-01 -3.13839883e-01
2.02928185e-01 -9.92813855e-02 1.19957618e-01 -5.68218946e-01
2.24564523e-01 1.84228048e-01 -1.95912516e+00 1.81143299e-01
5.77693105e-01 9.13432181e-01 -4.04192924e-01 6.23168409e-01
1.63468301e-01 1.13128555e+00 -6.03224277e-01 -1.15081608e+00
3.00248325e-01 -6.72181249e-01 -8.28514993e-01 -4.60470989e-02
-3.27798903e-01 -4.57422435e-01 3.62963051e-01 1.13383269e+00
6.37774527e-01 -9.71153229e-02 1.11227818e-01 -1.47620845e+00
-3.96591902e-01 1.67854905e-01 -1.65522143e-01 -1.51070178e-01
6.03078127e-01 2.39590794e-01 6.69588804e-01 -1.09116815e-01
8.28830063e-01 6.33114576e-01 4.79129523e-01 8.18963766e-01
-8.40179026e-01 -6.65304899e-01 2.07552202e-02 -3.68246958e-02
-1.17178690e+00 -6.15622580e-01 5.22980154e-01 -6.46998286e-01
9.28267002e-01 7.20585763e-01 1.23423398e+00 1.11911798e+00
1.89060494e-01 3.48003060e-01 8.30970824e-01 -5.96382558e-01
5.82106650e-01 8.61698270e-01 -5.23060501e-01 -9.27922409e-03
3.05256337e-01 3.04399412e-02 -3.75547349e-01 -3.23848665e-01
7.47688055e-01 -6.09856248e-01 -7.80752301e-02 -4.89096254e-01
-5.44153094e-01 8.26086879e-01 -4.04773891e-01 4.94460374e-01
-5.05434155e-01 1.80236027e-01 8.90700519e-01 2.60605633e-01
4.39895451e-01 6.32367373e-01 -2.00940132e-01 -1.16188443e+00
-3.43201876e-01 2.50149071e-01 9.53593850e-01 5.74185729e-01
4.17890064e-02 -2.19321772e-01 4.89868820e-01 6.85245752e-01
5.38321853e-01 -7.97000229e-02 9.26598534e-03 -1.05185294e+00
-2.78907180e-01 5.73496044e-01 3.05477113e-01 -1.37148154e+00
2.25290004e-02 -5.44509739e-02 -1.93462729e-01 7.53762007e-01
2.33142227e-01 -3.48518312e-01 9.86356959e-02 1.65995944e+00
5.17518818e-01 -1.94809586e-01 -2.07232963e-02 7.77143836e-01
5.71521997e-01 2.71496505e-01 3.69085103e-01 -2.78694153e-01
1.49663258e+00 2.63298508e-02 -8.62296224e-01 -4.12729472e-01
5.98535657e-01 -5.78515828e-01 9.50687826e-01 7.70249963e-01
-1.34386289e+00 1.81216419e-01 -1.07655811e+00 2.53345340e-01
-3.20538223e-01 -5.60513973e-01 6.27011657e-01 1.65483332e+00
-1.07631600e+00 2.53306419e-01 -5.17568946e-01 -8.43071342e-01
3.62022787e-01 3.96660626e-01 -3.17146510e-01 4.13622111e-01
-1.32230413e+00 1.05541468e+00 6.67705983e-02 1.29335240e-01
1.98888734e-01 -4.52284247e-01 -6.88296616e-01 -3.53351980e-01
4.19829607e-01 -2.89914757e-01 1.20042777e+00 -1.60184836e+00
-1.76497316e+00 1.39400494e+00 8.56372178e-01 -1.70891240e-01
7.47672319e-01 -3.03856581e-01 -5.04113793e-01 -1.18833929e-01
5.74371554e-02 2.51472741e-01 2.48407900e-01 -1.07800639e+00
-6.44633830e-01 -1.48014098e-01 2.88939297e-01 3.31218898e-01
-2.92789906e-01 8.79330695e-01 -1.69187531e-01 -1.33330390e-01
-7.15410173e-01 -9.00179565e-01 -2.06549019e-01 8.19549616e-03
2.52110988e-01 2.58351892e-01 4.65610832e-01 -5.61112106e-01
1.60461950e+00 -2.41969562e+00 -2.46213719e-01 2.07543463e-01
3.26024890e-01 5.84409714e-01 4.40788478e-01 7.49908984e-01
1.14396936e-03 3.88528526e-01 2.70519316e-01 2.58913130e-01
4.07233179e-01 -8.68418515e-02 4.42675650e-02 4.25654382e-01
-1.15298554e-01 5.57461202e-01 -9.13454890e-01 -2.89679617e-01
3.33597392e-01 5.73862731e-01 -7.33756781e-01 2.93608725e-01
2.01631978e-01 2.76052892e-01 -3.30764830e-01 2.16816574e-01
4.74157721e-01 4.59038645e-01 4.63121861e-01 4.92982149e-01
-4.04144913e-01 1.16659872e-01 -1.03691089e+00 1.22433054e+00
-2.34848812e-01 5.54394662e-01 4.98218596e-01 -5.01663268e-01
7.70157456e-01 6.36016846e-01 5.13945639e-01 -1.02127099e+00
5.89698911e-01 2.18870655e-01 3.31042200e-01 -7.70735264e-01
4.08637881e-01 -3.21184814e-01 -3.60998392e-01 7.39189029e-01
-4.26241428e-01 -4.36796784e-01 -2.69231021e-01 -1.22011062e-02
1.09446061e+00 3.15649867e-01 6.76269174e-01 -1.33288622e-01
3.23097467e-01 -2.73875594e-01 4.35327142e-01 2.45143786e-01
-8.59277606e-01 4.63002235e-01 9.29711282e-01 -5.74160457e-01
-7.34028280e-01 -4.38560873e-01 1.37705639e-01 1.17878246e+00
-1.06471088e-02 -6.55402064e-01 -1.33085561e+00 -3.13519299e-01
-3.40206146e-01 8.83659780e-01 -7.12728858e-01 -3.70658606e-01
-2.15392355e-02 -3.37839514e-01 5.15569746e-01 -2.99653262e-02
1.91705227e-01 -1.41329539e+00 -1.77673340e+00 3.80125165e-01
-1.88341085e-02 -1.13490903e+00 2.05523849e-01 -2.64096022e-01
-1.94034740e-01 -1.09082496e+00 -8.59207436e-02 -3.52236450e-01
4.78286296e-02 -2.60573953e-01 1.18747723e+00 2.36391604e-01
-3.10619563e-01 8.64901960e-01 -6.74159765e-01 -1.02660143e+00
-7.57326722e-01 -4.97063011e-01 -1.71128154e-01 -4.34055030e-02
7.73924291e-01 -7.68414557e-01 -5.51152766e-01 2.45272666e-01
-1.08857775e+00 1.37621418e-01 7.54807740e-02 1.77272379e-01
-2.02609167e-01 -9.49513689e-02 4.66637284e-01 -1.08881724e+00
9.86034155e-01 -5.74167967e-01 -2.62765765e-01 -3.46660130e-02
-2.89709866e-01 -8.13263535e-01 2.51813531e-01 -1.80401549e-01
-9.01772738e-01 -3.00810724e-01 -2.88163483e-01 3.92528504e-01
-5.08206785e-01 2.34756202e-01 -4.27904189e-01 -2.68837810e-01
8.22444141e-01 -5.03150165e-01 3.18850815e-01 2.03134045e-01
2.81004488e-01 8.85242999e-01 4.68337461e-02 -3.18859428e-01
1.85485080e-01 4.22146440e-01 -4.25202608e-01 -1.02444017e+00
1.26303032e-01 -1.85815141e-01 -3.33709002e-01 -9.57731426e-01
9.09649968e-01 -6.53476059e-01 -9.07175660e-01 4.80124503e-01
-9.49755311e-01 -3.14529419e-01 -5.19409359e-01 1.94204107e-01
-7.01275587e-01 1.77571625e-01 -2.12898552e-01 -1.36987352e+00
-3.01784903e-01 -1.00191891e+00 5.10830164e-01 2.16849521e-01
-1.09251952e+00 -7.52699077e-01 2.73721665e-01 5.40536225e-01
5.48765182e-01 8.37910235e-01 6.08525634e-01 -4.24422354e-01
-2.88640056e-02 -5.16140699e-01 2.12072626e-01 2.77138889e-01
4.64390218e-02 2.70496041e-01 -1.08115304e+00 1.77458197e-01
3.44485730e-01 -3.45454901e-01 -4.95862424e-01 8.06664303e-02
2.64857709e-01 -4.23764884e-01 1.32775739e-01 -1.69648584e-02
1.40300465e+00 8.14940095e-01 1.16442037e+00 8.22364390e-01
2.93103844e-01 1.08846045e+00 1.03398883e+00 9.40686285e-01
1.76094085e-01 4.86083925e-01 7.04014361e-01 1.19512558e-01
6.76535368e-01 2.18231827e-01 2.94022024e-01 3.59804362e-01
-3.10500503e-01 -3.02945785e-02 -1.01743901e+00 5.13774574e-01
-1.88327265e+00 -8.54307353e-01 -2.16485813e-01 2.33466887e+00
4.54084307e-01 2.53475606e-01 5.18047154e-01 2.18297258e-01
6.12257361e-01 -3.26928608e-02 1.50014572e-02 -1.40301561e+00
4.33478087e-01 2.95070380e-01 7.43628889e-02 1.49542704e-01
-9.01229203e-01 8.13477039e-01 6.20182085e+00 2.90536702e-01
-8.18087339e-01 7.56638944e-02 7.48693287e-01 -3.29896629e-01
-4.02569860e-01 -5.52970842e-02 2.54691154e-01 2.90562183e-01
8.28898668e-01 -3.17713916e-01 4.44710761e-01 9.04167175e-01
4.17603731e-01 -4.51638222e-01 -8.60276580e-01 9.55070317e-01
1.79412141e-02 -7.47867107e-01 -7.82321095e-01 4.25483257e-01
2.35080212e-01 -3.50558192e-01 1.47768185e-01 -3.86983790e-02
1.73922271e-01 -1.17073905e+00 1.02596045e+00 2.94970199e-02
6.66332841e-01 -1.16490126e+00 7.74524808e-01 1.41280517e-01
-6.23447776e-01 4.76970598e-02 3.32356319e-02 -8.93370152e-01
1.69654489e-01 7.80861303e-02 -2.90721774e-01 -1.63730551e-02
8.86114836e-01 4.92635928e-02 -1.47717461e-01 7.72794783e-01
-7.87340943e-03 3.09503615e-01 -2.03469738e-01 -3.83786917e-01
9.94356349e-02 -5.30719936e-01 4.64006215e-01 9.63716686e-01
4.68035638e-02 5.18683195e-01 -7.02897668e-01 7.02908099e-01
5.64669013e-01 4.76544470e-01 -1.02594709e+00 -4.25865561e-01
3.52327555e-01 1.15948141e+00 -8.49030733e-01 4.34356809e-01
-6.29446089e-01 1.04452896e+00 -1.53946742e-01 -8.69019926e-02
-8.01170409e-01 -2.82204509e-01 1.24947500e+00 3.47435057e-01
-1.72853842e-02 3.82551640e-01 -4.18247342e-01 -5.32654345e-01
1.26643598e-01 -1.34054768e+00 1.24332830e-01 -5.62710583e-01
-7.93768644e-01 4.55741316e-01 -1.77046940e-01 -1.12311852e+00
-5.90139747e-01 -3.98761958e-01 -5.97334504e-01 4.31002796e-01
-4.50417966e-01 -1.10160625e+00 -1.05629772e-01 4.78260666e-02
-1.37402266e-01 2.18779743e-01 1.13674617e+00 1.80018485e-01
-3.01280260e-01 4.41805094e-01 -9.20157954e-02 -1.87299132e-01
4.88076419e-01 -7.74195194e-01 6.13179266e-01 7.94586182e-01
-3.85113984e-01 5.40817499e-01 9.64490294e-01 -4.65014726e-01
-1.28526759e+00 -2.77989030e-01 7.07887292e-01 -6.45224154e-01
6.47280872e-01 -6.63551986e-01 -4.44602668e-01 4.46296632e-01
5.10147572e-01 -4.82425362e-01 1.34546447e+00 -1.40327606e-02
1.30984873e-01 2.34002873e-01 -1.54136610e+00 9.16199803e-01
8.73962641e-01 -6.45974457e-01 -1.19608931e-01 -2.57048100e-01
3.47936511e-01 -2.75037050e-01 -8.06664586e-01 -7.42752999e-02
9.73170578e-01 -1.80180407e+00 5.85685194e-01 -2.92391747e-01
3.98508608e-01 2.35058982e-02 1.07163571e-01 -7.94190347e-01
-2.10504621e-01 -1.22751534e+00 6.11368001e-01 1.68156219e+00
1.33897215e-01 -7.59293795e-01 9.20191824e-01 1.73637879e+00
3.50615144e-01 -6.80594206e-01 -8.63285959e-01 -1.34825110e-01
1.27402246e-01 -9.63983595e-01 6.11158133e-01 1.05736721e+00
9.21122193e-01 -1.58908740e-01 -4.80957687e-01 -2.06702963e-01
1.64187133e-01 -7.42958844e-01 8.65889072e-01 -1.17626882e+00
-2.36106098e-01 -5.74240983e-01 -1.05004168e+00 1.74197569e-01
-4.03701156e-01 1.44562107e-02 -2.40134913e-02 -1.31809306e+00
1.60727762e-02 2.71085873e-02 1.49574772e-01 -1.36037245e-01
2.06200063e-01 4.43624049e-01 4.06077474e-01 -3.00784856e-01
-5.66520154e-01 -7.93153793e-02 6.11355186e-01 7.28872538e-01
-2.79697418e-01 -1.31112128e-01 -1.50553679e+00 8.17807436e-01
7.33212769e-01 -1.73920989e-01 -6.82832003e-01 1.10139765e-01
1.03709280e+00 -2.82864481e-01 8.42904150e-02 -9.44011211e-01
4.26433496e-02 -3.72453868e-01 -2.22069412e-01 2.94586778e-01
2.21959472e-01 -1.28599131e+00 8.64174485e-01 2.91180789e-01
2.44287904e-02 -2.92750180e-01 5.12537837e-01 1.72433913e-01
6.38915673e-02 -2.89605319e-01 7.06099570e-01 -2.54946113e-01
-5.21434247e-01 -2.35568792e-01 -1.18670309e+00 2.27695674e-01
1.66816616e+00 -9.62385893e-01 1.75876260e-01 -7.89414585e-01
-4.57056254e-01 4.66834530e-02 1.17431223e+00 4.67178762e-01
2.88508236e-01 -9.46112156e-01 -2.73159355e-01 1.69623628e-01
2.54854292e-01 -3.21352273e-01 4.64668930e-01 4.69445407e-01
-8.93824756e-01 -1.01430133e-01 -5.61698794e-01 4.67751548e-02
-1.60268152e+00 3.68531495e-01 3.64951432e-01 8.50905851e-02
-3.34641367e-01 9.70147491e-01 3.88464242e-01 -3.10263503e-02
1.27821356e-01 2.28460088e-01 -4.85341966e-01 1.25795543e-01
7.18070328e-01 3.88207316e-01 -2.88735665e-02 -9.16499376e-01
-5.51436901e-01 7.68842027e-02 1.48554116e-01 -6.40194058e-01
1.35761714e+00 -3.57611030e-01 -4.44583386e-01 4.68279004e-01
7.00871289e-01 1.03468224e-01 -9.40052152e-01 6.94272161e-01
-9.17115510e-02 -8.81106853e-01 -3.19764733e-01 -7.96398818e-01
-7.66935706e-01 5.89364409e-01 5.77367902e-01 7.94722080e-01
1.27308941e+00 -3.14299494e-01 4.13706392e-01 -5.09767413e-01
5.28667867e-01 -1.53821957e+00 -1.95048988e-01 1.14434175e-01
7.77471483e-01 -7.01983988e-01 -1.11452512e-01 -7.22397983e-01
-1.21557724e+00 8.14579666e-01 6.26993537e-01 2.32112378e-01
6.24602258e-01 5.23751974e-01 4.53955114e-01 -2.93736041e-01
-6.51436687e-01 -5.67502007e-02 -3.59130800e-01 1.10532534e+00
6.92142248e-01 1.66943237e-01 -7.23828852e-01 9.97979939e-01
-2.87072688e-01 4.02251929e-01 8.51585746e-01 1.24168050e+00
-1.02155037e-01 -1.31899083e+00 -4.60349649e-01 3.05524409e-01
-1.04439998e+00 1.14721008e-01 -1.05363941e+00 8.96694601e-01
3.86405557e-01 1.28939950e+00 -2.21313253e-01 -7.04677045e-01
5.14830291e-01 -5.42839542e-02 2.95150578e-01 -6.08261108e-01
-1.37763262e+00 2.42211651e-02 6.23113632e-01 -7.13996768e-01
-3.06783468e-01 -9.87947643e-01 -8.04057717e-01 -5.74364960e-01
-3.15424889e-01 2.32604280e-01 6.85249448e-01 7.47925401e-01
5.22356868e-01 2.14863211e-01 2.20428824e-01 -6.12727582e-01
5.24704605e-02 -2.78542578e-01 -8.58353853e-01 4.75245804e-01
-3.08083713e-01 -2.76097596e-01 -2.26472512e-01 -1.40631840e-01]
|
[9.135711669921875, 6.445724010467529]
|
1e0cfd09-1e61-4842-b564-afe367b20619
|
self-explaining-neural-network-with-plausible
|
2110.04598
| null |
https://arxiv.org/abs/2110.04598v3
|
https://arxiv.org/pdf/2110.04598v3.pdf
|
Self-explaining Neural Network with Concept-based Explanations for ICU Mortality Prediction
|
Complex deep learning models show high prediction tasks in various clinical prediction tasks but their inherent complexity makes it more challenging to explain model predictions for clinicians and healthcare providers. Existing research on explainability of deep learning models in healthcare have two major limitations: using post-hoc explanations and using raw clinical variables as units of explanation, both of which are often difficult for human interpretation. In this work, we designed a self-explaining deep learning framework using the expert-knowledge driven clinical concepts or intermediate features as units of explanation. The self-explaining nature of our proposed model comes from generating both explanations and predictions within the same architectural framework via joint training. We tested our proposed approach on a publicly available Electronic Health Records (EHR) dataset for predicting patient mortality in the ICU. In order to analyze the performance-interpretability trade-off, we compared our proposed model with a baseline having the same set-up but without the explanation components. Experimental results suggest that adding explainability components to a deep learning framework does not impact prediction performance and the explanations generated by the model can provide insights to the clinicians to understand the possible reasons behind patient mortality.
|
['Andrew Michelson', 'Zachary Abrams', 'Thomas Kannampallil', 'Philip R. O. Payne', 'Sean C. Yu', 'Sayantan Kumar']
|
2021-10-09
| null | null | null | null |
['explainable-models', 'icu-mortality']
|
['computer-vision', 'medical']
|
[-1.35057494e-02 9.94319022e-01 -9.26627144e-02 -6.54895902e-01
-1.70448110e-01 -2.78899577e-02 8.51191133e-02 4.97109920e-01
2.03973383e-01 7.62506843e-01 5.41718900e-01 -7.53618717e-01
-5.46565592e-01 -5.08952200e-01 -6.21873617e-01 -2.14076906e-01
-1.53980508e-01 7.87548840e-01 -3.94614458e-01 2.12386344e-02
1.41226929e-02 3.03778142e-01 -1.23958266e+00 1.04088247e+00
1.16022897e+00 8.99408162e-01 -1.06943242e-01 7.23959148e-01
-6.34950101e-02 1.29347062e+00 -3.85395855e-01 -3.37318242e-01
2.92073011e-01 -7.44021118e-01 -8.77192318e-01 1.04749128e-02
1.17810350e-03 -5.22487402e-01 -1.46325409e-01 2.13795796e-01
3.00590307e-01 -2.75579989e-01 6.76105678e-01 -1.44774818e+00
-1.04701936e+00 9.89386141e-01 1.94686353e-01 1.87933091e-02
2.28307530e-01 2.89571166e-01 1.02152753e+00 -5.80916584e-01
1.41032711e-01 9.88010705e-01 1.09077382e+00 1.01606119e+00
-1.16015506e+00 -4.53929782e-01 1.58107504e-01 4.56272244e-01
-8.56941819e-01 -4.88944426e-02 3.77925515e-01 -6.54505074e-01
1.26571321e+00 4.29678470e-01 7.40849674e-01 1.06964290e+00
6.62549555e-01 4.18225139e-01 7.38278925e-01 -2.22569078e-01
3.04870009e-01 3.26665103e-01 5.21777689e-01 7.29844451e-01
5.87500274e-01 2.28382498e-01 -2.70690709e-01 -2.92218685e-01
6.98276758e-01 8.64310324e-01 -4.49752212e-01 -1.88753828e-01
-1.22382402e+00 1.07201135e+00 7.15103626e-01 2.27106199e-01
-7.20215201e-01 2.18059286e-01 3.06865484e-01 1.80211693e-01
8.66717547e-02 7.94120133e-01 -1.08397388e+00 9.28682908e-02
-6.88838661e-01 1.03387929e-01 9.32427227e-01 6.92577302e-01
4.34430420e-01 1.02083638e-01 -3.69575262e-01 2.26959363e-01
1.95503309e-01 -2.53522638e-02 8.00315320e-01 -7.41804063e-01
2.27580816e-01 1.02682042e+00 2.16428787e-02 -8.07231903e-01
-7.89905787e-01 -7.53272772e-01 -1.08666861e+00 9.58635136e-02
1.47391856e-01 -3.79601568e-01 -9.72426236e-01 1.53708732e+00
-1.72455519e-01 2.31794745e-01 4.34396088e-01 9.76289988e-01
1.10176718e+00 2.43707210e-01 3.84345174e-01 -7.59686455e-02
1.27536523e+00 -1.30629337e+00 -8.43905866e-01 -1.75766945e-01
1.05327082e+00 -2.39072949e-01 9.13412690e-01 3.05254638e-01
-9.75009739e-01 -6.27813876e-01 -8.33655179e-01 -3.33974101e-02
-1.30252585e-01 1.25806242e-01 6.98613822e-01 1.67398915e-01
-9.91870046e-01 8.70377123e-01 -9.20006514e-01 -4.15446162e-01
4.32358593e-01 6.03175163e-01 -2.84959167e-01 2.15163037e-01
-1.05377090e+00 9.71064448e-01 4.08216566e-01 -1.88802555e-01
-7.02400744e-01 -9.55333412e-01 -8.11247945e-01 7.32832313e-01
-2.59500425e-02 -1.50367820e+00 1.21487629e+00 -1.12568736e+00
-9.89955664e-01 2.89027512e-01 -2.52099365e-01 -8.68011892e-01
6.11473143e-01 -5.94559848e-01 -2.42428154e-01 -1.41191274e-01
-1.03015386e-01 4.92061704e-01 2.48962909e-01 -1.23045528e+00
-3.35097939e-01 -2.83216298e-01 9.97533128e-02 -1.50562376e-01
-1.32272243e-01 -6.56179070e-01 2.61926800e-01 -4.89548326e-01
1.53842315e-01 -9.26892102e-01 -5.59413671e-01 -9.53813121e-02
-7.13164866e-01 1.10910654e-01 5.49939632e-01 -7.17870057e-01
1.33044016e+00 -1.78309202e+00 7.79365003e-02 -6.60696775e-02
8.62032413e-01 1.62981823e-01 4.23415527e-02 4.69501168e-01
-6.01886690e-01 4.97082412e-01 -2.63190448e-01 -3.36361289e-01
-1.46202087e-01 5.30842304e-01 -2.79916793e-01 -1.63584322e-01
4.07810479e-01 1.14348054e+00 -6.57401085e-01 -3.46872002e-01
2.90437818e-01 7.23528802e-01 -1.09103751e+00 7.90069163e-01
-1.70032114e-01 7.33886182e-01 -5.49522400e-01 3.42439890e-01
2.38685265e-01 -8.84569585e-01 2.10885614e-01 -3.72943804e-02
3.59037638e-01 2.55015463e-01 -6.31389201e-01 1.14815593e+00
-3.45697641e-01 4.47420031e-01 -6.67130709e-01 -1.03330052e+00
6.72183812e-01 6.66928172e-01 5.33499479e-01 -2.28159145e-01
1.57817140e-01 7.45711923e-02 4.42348152e-01 -8.05977225e-01
-1.00411028e-01 -4.57488120e-01 2.98043609e-01 5.16971469e-01
-3.21872264e-01 4.54726279e-01 -4.65337306e-01 7.85691291e-02
1.35094953e+00 -1.54406682e-01 8.65141034e-01 -4.53209318e-02
4.12199914e-01 3.20799232e-01 7.39520550e-01 8.12946796e-01
-2.85430197e-02 9.20337200e-01 5.31862199e-01 -1.40793824e+00
-9.00060236e-01 -7.60518909e-01 -8.12431891e-03 5.30023098e-01
-2.45486557e-01 -4.99859422e-01 -8.07514250e-01 -9.83485639e-01
1.02916524e-01 1.00701845e+00 -1.11202025e+00 -3.57112795e-01
-3.61686498e-01 -6.07023716e-01 1.99587792e-01 9.95905221e-01
5.08272611e-02 -1.20463455e+00 -1.24589550e+00 3.63900155e-01
-1.74447477e-01 -9.16141808e-01 8.76996783e-04 5.12486935e-01
-1.30507934e+00 -1.39517438e+00 -2.77412236e-01 -3.91034782e-01
7.78392494e-01 -1.02390401e-01 1.37628376e+00 7.72458613e-01
-2.58131653e-01 1.51191548e-01 -4.62467164e-01 -8.08635414e-01
-6.26382113e-01 8.86017010e-02 -2.06630826e-02 -3.36484820e-01
4.25303757e-01 -3.63892138e-01 -1.03925896e+00 2.15395913e-01
-9.48953748e-01 4.93190318e-01 7.53962696e-01 1.07525110e+00
4.54799473e-01 -3.73907983e-01 7.29523361e-01 -1.31403410e+00
6.99897647e-01 -8.73558223e-01 1.79807559e-01 1.94600597e-01
-1.24196231e+00 4.59179491e-01 9.40245092e-01 -1.80116698e-01
-6.93298757e-01 1.27621382e-01 -1.30804420e-01 -3.54414225e-01
-4.62073982e-01 5.07530034e-01 1.57814935e-01 5.10806501e-01
6.65495038e-01 -3.19742300e-02 3.82904150e-02 -5.01472712e-01
1.18229732e-01 4.30006564e-01 1.34253114e-01 5.06704636e-02
5.47270596e-01 1.75891250e-01 2.61776838e-02 5.49825206e-02
-1.04841495e+00 -2.89310180e-02 -6.20418727e-01 2.40458578e-01
1.15936542e+00 -7.39169121e-01 -8.95630419e-01 -4.63765115e-01
-1.32757163e+00 -1.87882721e-01 -3.77493501e-01 4.74315256e-01
-5.95212281e-01 1.60803273e-01 -5.98706782e-01 -5.26059985e-01
-6.90579534e-01 -1.28456068e+00 8.79178047e-01 3.20801139e-02
-9.45174515e-01 -1.34986448e+00 -6.44779429e-02 4.35450375e-01
6.33017302e-01 6.21741831e-01 1.45004535e+00 -1.31363773e+00
-4.70505655e-01 -6.85181543e-02 -2.32831076e-01 1.21555761e-01
4.02500570e-01 -3.74585509e-01 -8.58245850e-01 1.07092023e-01
1.67199373e-02 -1.09254876e-02 6.50446296e-01 5.42576730e-01
1.59162319e+00 -8.34700227e-01 -4.03305709e-01 7.78700173e-01
1.29312110e+00 2.70136833e-01 5.12938976e-01 5.00057697e-01
7.56867647e-01 5.39878964e-01 1.66677207e-01 6.93033874e-01
6.34265423e-01 3.44478577e-01 6.72071815e-01 -6.31805778e-01
4.44729887e-02 -5.30651845e-02 -8.04839209e-02 6.58577323e-01
-1.55181497e-01 -1.92747623e-01 -1.28441143e+00 3.54203552e-01
-2.30811739e+00 -7.33200133e-01 -3.84711981e-01 1.64195967e+00
4.77653533e-01 -2.16120914e-01 -2.66173154e-01 2.09249571e-01
2.47858703e-01 -6.47221863e-01 -6.94272995e-01 -8.48311484e-01
1.40943006e-01 2.53420826e-02 9.27089080e-02 4.35267597e-01
-6.76597774e-01 3.28866690e-01 6.63559484e+00 -3.32866609e-01
-1.05462015e+00 5.67381978e-02 9.52781558e-01 -2.72047240e-02
-3.53351623e-01 -1.01776905e-01 -2.10743889e-01 1.60222352e-01
1.14948130e+00 -4.35022898e-02 5.79485931e-02 1.10103953e+00
5.96401393e-01 5.22114098e-01 -1.84847105e+00 7.87900269e-01
-9.55353826e-02 -1.57641697e+00 4.80147660e-01 7.70984739e-02
6.21581376e-01 -3.08307528e-01 -2.20902022e-02 1.99516639e-01
3.38362426e-01 -1.55034590e+00 3.76048118e-01 8.47452521e-01
3.91574830e-01 -3.58640105e-01 1.36587954e+00 4.20845747e-01
-6.45299256e-01 -5.51471293e-01 -1.45667091e-01 -5.09003043e-01
-2.16645887e-03 3.54260653e-01 -1.49045110e+00 4.33130026e-01
6.53429866e-01 7.24208295e-01 -5.59554219e-01 7.07176924e-01
-1.03557765e-01 7.08028972e-01 3.90099347e-01 4.12841052e-01
1.36251926e-01 3.67052078e-01 7.90553316e-02 1.18948293e+00
4.91700768e-01 4.61241305e-01 8.06370899e-02 1.09064269e+00
1.66256323e-01 2.14308202e-02 -5.04420340e-01 1.67391226e-01
3.64123024e-02 8.98138881e-01 -3.49253744e-01 -6.98813677e-01
-3.19785893e-01 8.79908502e-01 2.94067144e-01 2.04442069e-01
-9.99298513e-01 1.52120501e-01 6.98577225e-01 5.72295785e-01
5.18445931e-02 3.80664885e-01 -1.05344987e+00 -1.27452278e+00
-3.24508607e-01 -1.02185571e+00 6.40110672e-01 -8.88021350e-01
-1.33331966e+00 1.14910293e+00 -2.21698716e-01 -1.17037463e+00
-6.48952246e-01 -5.17506838e-01 -7.13366032e-01 9.47904408e-01
-1.54251850e+00 -1.10859036e+00 -8.12229693e-01 4.42458332e-01
6.72909677e-01 -2.48204485e-01 1.23009634e+00 -5.33556268e-02
-5.04605889e-01 5.45063317e-01 -2.50812322e-01 1.58422515e-02
3.26476812e-01 -1.27478385e+00 2.72783428e-01 3.02906305e-01
-2.66983479e-01 9.41852331e-01 8.78425837e-01 -6.77258849e-01
-1.02789140e+00 -1.35149527e+00 1.16511226e+00 -7.72455871e-01
1.37984872e-01 2.36993834e-01 -1.34852636e+00 1.04257333e+00
2.87306011e-01 -6.29784837e-02 1.34735215e+00 1.51078418e-01
-3.93205062e-02 1.99461699e-01 -1.22841394e+00 2.84059078e-01
8.87758791e-01 6.42692149e-02 -9.59064782e-01 2.61427909e-01
9.58376586e-01 -1.25172198e-01 -8.89506698e-01 6.23036683e-01
6.27006173e-01 -1.23407042e+00 7.17236400e-01 -1.34041786e+00
1.19258356e+00 -7.18610138e-02 4.21396270e-02 -1.34507275e+00
-6.30709589e-01 -1.92153156e-01 -2.96619534e-01 5.76172411e-01
7.29467750e-01 -7.23509371e-01 8.22609305e-01 1.22936594e+00
-3.17052037e-01 -1.29724610e+00 -3.68396670e-01 -2.47988895e-01
-5.23003116e-02 -2.16883615e-01 1.09479833e+00 1.00057399e+00
1.79834694e-01 3.58173162e-01 -4.82260257e-01 4.19684529e-01
3.58122855e-01 1.72258243e-01 7.05338120e-01 -1.48415148e+00
-6.19513214e-01 -1.77582920e-01 -4.31526095e-01 -3.71495664e-01
-4.02775109e-02 -8.42663229e-01 -3.00108731e-01 -2.08943963e+00
6.16953313e-01 -2.63986379e-01 -6.09023750e-01 9.62653875e-01
-5.13524413e-01 -2.49020845e-01 2.56617695e-01 4.62114811e-01
-3.64117801e-01 4.26100612e-01 1.00171947e+00 1.13535568e-03
-2.31331348e-01 -1.63561031e-01 -1.09492123e+00 8.55380833e-01
8.42898071e-01 -6.44213140e-01 -4.67595696e-01 -8.11403334e-01
-1.79718826e-02 4.66632277e-01 4.46624219e-01 -9.98185277e-01
-8.16744640e-02 -1.91126749e-01 6.94481075e-01 -1.63817778e-01
-6.74294606e-02 -1.12751889e+00 5.99178553e-01 1.16531491e+00
-7.64990687e-01 4.74047363e-01 2.52945632e-01 5.22016287e-01
-1.55964389e-01 2.38491103e-01 5.03681123e-01 -1.92201808e-01
-2.68540710e-01 1.65639654e-01 -3.50834608e-01 -3.69504690e-01
1.10623109e+00 -3.10788006e-01 -1.78690508e-01 -5.34987807e-01
-1.11117375e+00 1.34152427e-01 2.73882121e-01 4.07839477e-01
8.88974190e-01 -1.07960689e+00 -7.32536554e-01 1.81680337e-01
2.47143194e-01 -1.25176534e-01 2.38350213e-01 6.41863704e-01
-7.10714877e-01 8.25216889e-01 -4.22252446e-01 -5.50000131e-01
-1.10870278e+00 7.59673655e-01 5.29937565e-01 -4.96637583e-01
-8.43163908e-01 4.15386766e-01 7.75621295e-01 -4.89859700e-01
1.33191839e-01 -8.05018961e-01 -3.84874672e-01 -5.76219618e-01
4.25389677e-01 1.50061965e-01 -4.70247194e-02 -2.36323372e-01
-4.36875612e-01 2.00217590e-01 -3.87378130e-03 5.69255292e-01
1.70088542e+00 2.29787193e-02 2.22958684e-01 2.96990752e-01
7.06796408e-01 -5.81377387e-01 -8.94693851e-01 1.06198847e-01
8.57849494e-02 -2.10988775e-01 -2.53373116e-01 -1.26687467e+00
-1.11725378e+00 1.18629885e+00 5.98254681e-01 1.88141257e-01
1.13111520e+00 8.18566233e-03 8.32291961e-01 4.06900495e-01
-9.42273363e-02 -3.17449272e-01 1.12353072e-01 1.81177422e-01
1.07532644e+00 -1.53104186e+00 -1.46726876e-01 -3.81771773e-01
-1.01215231e+00 1.26121879e+00 8.62949908e-01 4.71980385e-02
5.50709367e-01 1.15024686e-01 4.52412754e-01 -4.00276572e-01
-1.26120365e+00 2.55220622e-01 5.15350521e-01 4.58331972e-01
9.79159534e-01 3.08304518e-01 -2.73284465e-01 1.39677930e+00
-2.93251753e-01 2.94325143e-01 6.31285667e-01 5.43327034e-01
-1.90302834e-01 -9.82870579e-01 -2.58427143e-01 7.53591597e-01
-6.41851962e-01 -2.67297029e-01 -5.43169975e-01 6.65246964e-01
3.42487931e-01 9.34294224e-01 -1.60333067e-01 -6.72605813e-01
4.42448974e-01 2.83095986e-01 -2.84082770e-01 -9.01589394e-01
-9.44081604e-01 -3.81067783e-01 -1.36858314e-01 -6.73727512e-01
3.61520355e-03 -2.29081213e-01 -1.77934706e+00 -1.35824189e-01
-1.37428902e-02 3.06913048e-01 3.97639155e-01 1.09399176e+00
1.01579368e+00 1.20675325e+00 2.07638875e-01 -2.78868943e-01
-6.03943527e-01 -9.75563705e-01 -1.00105517e-01 8.70162904e-01
7.20228672e-01 -4.31894124e-01 -2.90693909e-01 3.70621800e-01]
|
[8.341941833496094, 5.987630844116211]
|
3377db2d-34b1-4347-bf4a-9f0dbfa77087
|
the-past-and-the-present-of-the-color-checker
|
1903.04473
| null |
http://arxiv.org/abs/1903.04473v1
|
http://arxiv.org/pdf/1903.04473v1.pdf
|
The Past and the Present of the Color Checker Dataset Misuse
|
The pipelines of digital cameras contain a part for computational color
constancy, which aims to remove the influence of the illumination on the scene
colors. One of the best known and most widely used benchmark datasets for this
problem is the Color Checker dataset. However, due to the improper handling of
the black level in its images, this dataset has been widely misused and while
some recent publications tried to alleviate the problem, they nevertheless
erred and created additional wrong data. This paper gives a history of the
Color Checker dataset usage, it describes the origins and reasons for its
misuses, and it explains the old and new mistakes introduced in the most recent
publications that tried to handle the issue. This should, hopefully, help to
prevent similar future misuses.
|
['Sven Lon{č}arić', 'Marko Subašić', 'Karlo Koš{č}ević', 'Nikola Banić']
|
2019-03-11
| null | null | null | null |
['color-constancy']
|
['computer-vision']
|
[ 5.43972179e-02 -4.11008239e-01 3.62681836e-01 -1.64261848e-01
1.49219513e-01 -6.50737941e-01 6.17714167e-01 -1.24512672e-01
-3.86221647e-01 7.20906496e-01 -2.10422993e-01 -2.93947995e-01
-5.45045696e-02 -5.06800175e-01 -5.45013905e-01 -9.36846912e-01
2.46059895e-01 -5.85906878e-02 6.41147196e-01 -1.50417298e-01
6.67095184e-01 3.97356778e-01 -1.91118205e+00 7.98966736e-02
6.96760774e-01 6.67474031e-01 5.04348017e-02 4.36400980e-01
-2.56151646e-01 9.84270811e-01 -8.31464291e-01 -7.33133435e-01
5.65048218e-01 -8.21182966e-01 -4.45408762e-01 1.90691546e-01
5.01139462e-01 -2.18688965e-01 -1.07125632e-01 1.23234308e+00
3.57157946e-01 -5.35348840e-02 3.37059647e-01 -1.45136297e+00
-5.65052211e-01 -3.45729105e-02 -7.77701974e-01 3.12542021e-02
2.35522315e-01 3.41605991e-01 4.16182905e-01 -4.19367075e-01
7.94598401e-01 7.85509408e-01 6.00503922e-01 5.87220669e-01
-9.57664490e-01 -6.54816151e-01 -1.98541239e-01 1.77276716e-01
-1.20007968e+00 -3.17224890e-01 7.11280346e-01 -3.53842378e-01
5.88102341e-01 3.84075731e-01 9.95968282e-01 9.77390289e-01
1.91774637e-01 2.79503614e-01 1.71819568e+00 -7.43643701e-01
2.15094760e-01 3.27910244e-01 1.60268098e-02 5.38088143e-01
6.23878360e-01 2.01449111e-01 -3.29768926e-01 5.74255455e-03
5.65274179e-01 -2.20848113e-01 -1.93376824e-01 -3.43565047e-01
-9.08234239e-01 4.49326664e-01 3.04980338e-01 5.14678180e-01
1.46651074e-01 -1.89889759e-01 2.35690415e-01 2.69938111e-01
1.99922726e-01 5.06433070e-01 -4.56221730e-01 -4.27707016e-01
-7.18325436e-01 -2.88712475e-02 5.31092584e-01 8.00813496e-01
7.14260042e-01 1.58671528e-01 3.29876810e-01 7.17767656e-01
1.68792471e-01 3.49396795e-01 1.50224566e-01 -6.79254353e-01
-8.02347809e-02 9.21396971e-01 1.16990902e-01 -1.10359728e+00
-4.28186625e-01 -2.00153947e-01 -6.26333892e-01 8.91629577e-01
1.00997829e+00 -3.43249626e-02 -1.02414727e+00 1.08101976e+00
1.94269374e-01 -4.31234948e-02 -2.87185669e-01 1.09200394e+00
4.62092847e-01 1.28035262e-01 -8.69633853e-02 7.24739805e-02
1.23507082e+00 -6.19051337e-01 -7.33112335e-01 9.50814225e-03
7.19986930e-02 -1.28106868e+00 9.62291420e-01 9.95105445e-01
-8.27270210e-01 -4.19557303e-01 -1.30165863e+00 -3.15702595e-02
-8.00933003e-01 -3.71397026e-02 7.57291734e-01 1.20656800e+00
-9.37315881e-01 5.07977009e-01 -4.48157966e-01 -4.10754979e-01
2.60112464e-01 1.12501415e-03 -2.07974091e-01 -2.16643780e-01
-8.06632698e-01 1.09941876e+00 1.97520882e-01 2.59536207e-01
-1.97209120e-01 -3.57924074e-01 -3.07510465e-01 -3.79360050e-01
5.14579356e-01 -1.07645713e-01 7.06310928e-01 -1.28187740e+00
-1.45114863e+00 1.13878477e+00 2.11371228e-01 -3.10360461e-01
9.19876993e-01 -2.05081791e-01 -6.29745185e-01 -1.62610695e-01
-3.07338566e-01 2.52031505e-01 8.75459492e-01 -1.18885720e+00
-5.48032165e-01 -1.52805254e-01 5.07190749e-02 -2.58360535e-01
-4.73130308e-02 1.40001893e-01 -8.89179230e-01 -5.68566978e-01
-1.59877650e-02 -8.75034511e-01 5.21197431e-02 -1.94944814e-01
-3.55822414e-01 2.87031144e-01 6.95467651e-01 -5.04516542e-01
9.86024976e-01 -2.58695936e+00 -3.70467424e-01 1.19369179e-01
3.54937464e-02 4.84577090e-01 1.34356990e-01 4.93191570e-01
-1.84796765e-01 1.32004440e-01 -2.18483850e-01 -1.72070175e-01
-2.50614196e-01 1.81040883e-01 -1.05802462e-01 8.94337118e-01
1.96701914e-01 4.18599039e-01 -8.15247536e-01 -2.53183872e-01
6.71957254e-01 6.26454890e-01 -1.40797138e-01 5.23593538e-02
-1.50544047e-01 4.27228838e-01 3.13892454e-01 7.01459408e-01
1.08335721e+00 3.91204059e-01 1.94119606e-02 -2.07491592e-01
-6.15381360e-01 2.28304286e-02 -1.49579537e+00 1.13744020e+00
9.48423371e-02 9.78934765e-01 -1.82215378e-01 -4.70862299e-01
8.30886066e-01 5.64725026e-02 5.47571540e-01 -1.10359859e+00
1.60892949e-01 3.98781419e-01 1.29746899e-01 -4.96726394e-01
7.21787810e-01 -1.77645329e-02 4.54705387e-01 1.42872676e-01
-4.65478808e-01 -3.93409610e-01 4.24921691e-01 2.43249238e-02
7.26645410e-01 5.98583341e-01 2.31963485e-01 -8.74074101e-02
6.20455682e-01 2.93840498e-01 4.63498861e-01 5.59465826e-01
-3.28175753e-01 1.16101789e+00 6.54123664e-01 -7.38796353e-01
-1.02972543e+00 -6.21905565e-01 -1.42506644e-01 6.34611607e-01
2.80153126e-01 -3.05726916e-01 -6.71632886e-01 -3.28926533e-01
-1.10321343e-01 5.81630528e-01 -6.68738365e-01 -7.46589079e-02
-3.13715488e-01 -8.95288944e-01 5.38306952e-01 6.48024827e-02
6.26010895e-01 -9.64190722e-01 -1.14853561e+00 -2.14749590e-01
3.02489161e-01 -1.04386270e+00 1.77138984e-01 1.61672562e-01
-7.13745832e-01 -1.71853161e+00 -6.84541106e-01 -2.30720520e-01
7.81625152e-01 4.73856837e-01 9.80790675e-01 5.96483707e-01
-5.50066173e-01 2.72365212e-01 -6.15302861e-01 -5.39723098e-01
-4.81573254e-01 -3.46402794e-01 -4.62990761e-01 -6.72841296e-02
4.37644750e-01 4.22925875e-02 -5.18009007e-01 3.45991671e-01
-1.25216591e+00 8.45503956e-02 3.46630186e-01 5.25386691e-01
2.00233579e-01 2.58271933e-01 -1.52685776e-01 -1.22386086e+00
3.29068124e-01 1.00417756e-01 -1.15572238e+00 2.07910284e-01
-8.31320584e-01 2.96124779e-02 4.41170931e-01 -1.93065964e-02
-1.05035520e+00 1.69910148e-01 1.42107025e-01 -2.13105138e-02
-3.66626173e-01 -1.80906475e-01 6.27434719e-03 -4.57351834e-01
4.77815837e-01 9.83823091e-02 -1.80189893e-01 -5.14201701e-01
3.84858474e-02 4.27430660e-01 4.61003065e-01 -1.51532874e-01
7.69936562e-01 6.05992615e-01 1.83106199e-01 -9.94903743e-01
-3.96133065e-01 -4.42119002e-01 -4.83994484e-01 -5.03746212e-01
7.93861568e-01 -4.73144919e-01 -5.92522979e-01 9.87198055e-01
-9.77508128e-01 -1.24538168e-01 -1.05266795e-01 7.34459385e-02
-3.07922605e-02 4.69614625e-01 -2.04505518e-01 -1.03326011e+00
2.00534642e-01 -1.19374609e+00 4.65351701e-01 6.28162861e-01
1.59301326e-01 -9.06639218e-01 5.79776391e-02 2.13711709e-01
9.15923893e-01 5.94309270e-01 8.35216343e-01 4.31979522e-02
-4.76356000e-01 -3.01018357e-01 -3.88944834e-01 5.43944120e-01
3.46469581e-01 8.47000301e-01 -1.22756469e+00 -5.14728650e-02
-5.33943065e-02 1.42305732e-01 9.19189274e-01 6.04739450e-02
7.73810208e-01 2.99221694e-01 1.49496585e-01 4.78202909e-01
1.74767065e+00 4.02011007e-01 1.08232522e+00 8.26728106e-01
4.54185575e-01 5.52778363e-01 5.09670436e-01 3.12217474e-01
-7.97508061e-02 6.83998883e-01 8.96844566e-01 -4.67406124e-01
-3.98320943e-01 8.03742856e-02 2.13530570e-01 2.92060643e-01
-5.46060264e-01 5.99727500e-03 -7.75934100e-01 1.65738866e-01
-1.35574591e+00 -5.64790845e-01 -9.37021732e-01 2.49050236e+00
5.26540756e-01 2.33248040e-01 1.87566131e-01 3.89912009e-01
6.33042693e-01 -3.88931111e-02 -1.97647423e-01 -4.66015071e-01
-7.34847844e-01 2.10467249e-01 7.45817542e-01 7.40842298e-02
-1.00100362e+00 6.52852833e-01 6.64669085e+00 9.60550159e-02
-1.45015240e+00 -1.52845785e-01 4.96018469e-01 1.87490582e-01
1.86933085e-01 2.54826903e-01 -3.52433801e-01 6.42575681e-01
4.55887645e-01 2.31644630e-01 5.16527534e-01 5.83504081e-01
6.70572966e-02 -9.03859615e-01 -5.55300772e-01 1.01896954e+00
2.44184792e-01 -8.59535694e-01 -2.27501065e-01 -4.88512330e-02
7.42275238e-01 -1.00130416e-01 3.58660482e-02 -1.40323356e-01
-3.06150001e-02 -7.20530570e-01 8.84233356e-01 6.69220090e-01
5.99510550e-01 -6.64218426e-01 9.45819080e-01 -9.99750718e-02
-6.11737728e-01 2.54690484e-03 -4.28882539e-01 -4.13674951e-01
-2.00964883e-01 7.82597005e-01 -5.78889728e-01 5.26525199e-01
7.86625624e-01 5.06957054e-01 -1.28046012e+00 1.65538180e+00
-5.03424466e-01 3.24681669e-01 -4.14016843e-02 6.20892793e-02
1.46287516e-01 -5.40139854e-01 3.20335478e-01 1.12635505e+00
-1.94283724e-02 -4.55058694e-01 -4.09798533e-01 5.86947501e-01
2.16025934e-01 3.81951332e-02 -3.95075113e-01 -4.57882397e-02
5.38220331e-02 1.14716029e+00 -1.04597831e+00 -6.31323531e-02
-7.39006400e-01 1.16272199e+00 -7.32590184e-02 3.28856081e-01
-1.00380421e+00 -3.39871109e-01 5.01977921e-01 2.57286191e-01
3.49096097e-02 -2.91968822e-01 -4.46465731e-01 -9.06601965e-01
-1.08980365e-01 -1.04869640e+00 3.46471936e-01 -1.04612315e+00
-1.09283650e+00 3.10480028e-01 -1.39293581e-01 -1.36675668e+00
2.27119565e-01 -8.75322223e-01 -3.03167343e-01 6.95524096e-01
-1.51432312e+00 -8.81474733e-01 -4.68985319e-01 4.83226866e-01
3.71697515e-01 -7.44528174e-02 6.91054225e-01 2.79029787e-01
-4.95145500e-01 1.54113293e-01 1.95788190e-01 2.00845197e-01
1.29795039e+00 -1.36962509e+00 1.93230867e-01 9.73043740e-01
1.44497201e-01 5.75605392e-01 9.89613593e-01 -3.82063359e-01
-1.47568977e+00 -3.75462323e-01 4.91866678e-01 -3.24002773e-01
4.13650662e-01 -1.81838602e-01 -7.97339737e-01 2.13784263e-01
6.71847820e-01 -2.19732478e-01 3.93131167e-01 -2.63293356e-01
-4.63149518e-01 -3.41626555e-01 -1.04499698e+00 5.21834195e-01
4.24727440e-01 -1.48761421e-01 -3.38900983e-01 -3.70957479e-02
-9.77660492e-02 -6.43563747e-01 -2.35480890e-01 1.60502195e-02
7.06106007e-01 -1.89109123e+00 5.60274124e-01 -2.35157937e-01
4.51144278e-01 -5.86789310e-01 2.53379047e-01 -1.12414622e+00
-6.24848530e-02 -4.29225475e-01 4.32509869e-01 1.31289172e+00
9.45223346e-02 -8.01625311e-01 4.55036253e-01 8.57132077e-01
1.27419844e-01 -1.47848904e-01 -9.44293380e-01 -5.63256800e-01
-1.40732571e-01 -3.93018782e-01 4.93877321e-01 1.02675724e+00
-4.41363096e-01 -3.32052380e-01 -7.19080627e-01 -1.48663089e-01
6.29392684e-01 -3.71257849e-02 9.08723950e-01 -1.27687216e+00
9.73031148e-02 -6.83218002e-01 -5.69683731e-01 -1.02147400e-01
-5.88143349e-01 -1.36288062e-01 -2.11214215e-01 -1.27913463e+00
1.43277884e-01 -3.52478683e-01 -1.46633044e-01 4.53192353e-01
-3.27040911e-01 8.75410676e-01 4.42599952e-01 2.89384322e-03
-4.32310313e-01 -2.51999617e-01 1.18078232e+00 9.70669165e-02
-7.70482272e-02 -9.23119336e-02 -5.76791763e-01 6.47642076e-01
8.78450811e-01 -5.16370058e-01 5.23037016e-02 -4.01786238e-01
7.71632671e-01 -7.37400115e-01 4.85753626e-01 -1.29099536e+00
1.90345168e-01 -2.43362203e-01 7.13153422e-01 -4.39908177e-01
1.53495491e-01 -1.33448994e+00 5.55312574e-01 6.56472623e-01
1.06215216e-01 1.05217315e-01 3.96557897e-01 1.21088415e-01
-1.18165195e-01 -4.04932827e-01 1.12921369e+00 -3.59332174e-01
-9.34263527e-01 -3.21425498e-01 -3.34024608e-01 -8.27724785e-02
1.16758215e+00 -5.20212054e-01 -5.71024120e-01 -1.49873912e-01
-1.24071814e-01 -2.39005789e-01 1.20416677e+00 5.08540094e-01
2.72569567e-01 -8.07804704e-01 -3.14455420e-01 4.25582677e-01
1.02068879e-01 -3.70009780e-01 2.18933359e-01 1.03197134e+00
-9.33791339e-01 1.52416855e-01 -6.28535450e-01 -3.53352636e-01
-1.43199193e+00 5.76878726e-01 5.05739331e-01 2.25684717e-01
-6.53522849e-01 4.31479663e-01 -4.56087321e-01 3.40164989e-01
1.16534598e-01 -2.03741759e-01 -1.85779750e-01 3.69442478e-02
6.53634846e-01 7.32791722e-01 4.70870167e-01 -6.26344800e-01
-3.93750578e-01 5.84126472e-01 2.11539313e-01 1.21781565e-01
1.40626550e+00 -1.91304877e-01 -4.29936767e-01 6.27110839e-01
5.42626202e-01 4.34743285e-01 -1.21755123e+00 3.34311426e-01
6.61239177e-02 -1.03038502e+00 -1.30278662e-01 -1.13183773e+00
-1.40644014e+00 7.23894775e-01 7.20598698e-01 6.37373388e-01
1.42884588e+00 -5.28375328e-01 2.69638389e-01 -6.36772364e-02
4.09734696e-01 -1.24667835e+00 -2.96105921e-01 4.95680809e-01
5.25710881e-01 -1.19105852e+00 4.31292772e-01 -4.18648243e-01
-6.81665480e-01 1.63637781e+00 4.23325032e-01 1.46753490e-02
2.13837266e-01 3.43358338e-01 4.10186291e-01 -8.27145576e-02
-2.45292813e-01 -3.10961008e-01 1.18424417e-02 8.85990024e-01
8.20096910e-01 -1.24786325e-01 -6.51343584e-01 1.14128925e-01
5.53345196e-02 8.48364308e-02 8.40741634e-01 9.34160948e-01
-1.32066444e-01 -1.31255937e+00 -9.27260578e-01 2.66248077e-01
-4.96848315e-01 8.55248347e-02 -8.40313256e-01 1.07078934e+00
6.58570468e-01 8.07464182e-01 -2.39395261e-01 -1.75454333e-01
5.88389039e-01 1.61916599e-01 5.36017358e-01 -2.53464244e-02
-9.29410934e-01 1.54351573e-02 -2.08850101e-01 -6.30913138e-01
-7.22024560e-01 -5.55248857e-01 -8.92690778e-01 -3.96721035e-01
-3.08592230e-01 -5.92030026e-02 9.45848286e-01 5.95192730e-01
-3.06516830e-02 5.21241724e-01 4.26277071e-01 -5.90959072e-01
-9.53778625e-02 -8.42890084e-01 -8.08935583e-01 8.66531849e-01
2.58019716e-01 -7.60318041e-01 -6.03153110e-01 2.75343746e-01]
|
[10.398168563842773, -2.5140202045440674]
|
1f3ec439-6587-4d62-8b35-473cc864ca39
|
convolutional-recurrent-neural-networks-for-2
|
1703.05390
| null |
http://arxiv.org/abs/1703.05390v3
|
http://arxiv.org/pdf/1703.05390v3.pdf
|
Convolutional Recurrent Neural Networks for Small-Footprint Keyword Spotting
|
Keyword spotting (KWS) constitutes a major component of human-technology
interfaces. Maximizing the detection accuracy at a low false alarm (FA) rate,
while minimizing the footprint size, latency and complexity are the goals for
KWS. Towards achieving them, we study Convolutional Recurrent Neural Networks
(CRNNs). Inspired by large-scale state-of-the-art speech recognition systems,
we combine the strengths of convolutional layers and recurrent layers to
exploit local structure and long-range context. We analyze the effect of
architecture parameters, and propose training strategies to improve
performance. With only ~230k parameters, our CRNN model yields acceptably low
latency, and achieves 97.71% accuracy at 0.5 FA/hour for 5 dB signal-to-noise
ratio.
|
['Sercan O. Arik', 'Markus Kliegl', 'Rewon Child', 'Adam Coates', 'Andrew Gibiansky', 'Ryan Prenger', 'Joel Hestness', 'Chris Fougner']
|
2017-03-15
| null | null | null | null |
['small-footprint-keyword-spotting']
|
['speech']
|
[ 6.87203333e-02 -2.43731529e-01 -2.78616279e-01 -2.80651003e-01
-7.76672244e-01 -1.94570005e-01 1.75641298e-01 -8.29265043e-02
-5.37155509e-01 3.27730924e-01 1.21701941e-01 -9.93844748e-01
-4.73989733e-03 -3.84160191e-01 -3.93711865e-01 -2.88635433e-01
-1.14750996e-01 -3.90030354e-01 4.87292528e-01 -1.21783182e-01
-2.30141990e-02 7.51313388e-01 -1.27644682e+00 4.80210692e-01
3.41046959e-01 1.27897775e+00 1.83509082e-01 1.15479290e+00
4.90425937e-02 7.78569400e-01 -9.14116800e-01 -9.66628417e-02
-1.97439879e-01 -6.39834851e-02 -5.33708692e-01 -4.78777617e-01
8.47754329e-02 -3.88422459e-01 -8.49844813e-01 7.16799080e-01
8.94665480e-01 -1.14117330e-02 2.05646038e-01 -9.87206638e-01
-4.41488475e-01 6.97777927e-01 -2.31002033e-01 6.22747183e-01
1.20708004e-01 8.21990669e-02 9.38330710e-01 -1.10148525e+00
-1.19896516e-01 9.50540662e-01 7.15365052e-01 4.70292300e-01
-1.16106641e+00 -9.37073231e-01 1.83729067e-01 1.52075291e-01
-1.63078213e+00 -1.05477357e+00 4.84384269e-01 1.72054708e-01
1.69352627e+00 4.08482999e-01 1.31889224e-01 1.21537495e+00
1.30719021e-01 9.13851976e-01 4.70030785e-01 -4.92045820e-01
1.32521376e-01 5.52549958e-02 4.17390645e-01 5.89898229e-01
8.01408291e-02 -6.58375621e-02 -9.50352907e-01 -2.75410265e-01
5.26053309e-01 7.12007657e-02 -1.48535743e-01 3.75662267e-01
-7.40641713e-01 4.37535256e-01 3.50923270e-01 3.40729296e-01
-1.31963328e-01 5.10057449e-01 3.86955798e-01 3.49683732e-01
1.13883950e-01 3.88026714e-01 -4.75854456e-01 -5.93681991e-01
-1.02162278e+00 -4.32386510e-02 7.90601254e-01 1.04438221e+00
2.04964221e-01 5.98649204e-01 -2.87361324e-01 1.11395228e+00
2.10337922e-01 6.68039858e-01 4.08447444e-01 -3.62817287e-01
5.14853537e-01 2.59597570e-01 -1.35191396e-01 -7.72195935e-01
-5.36974847e-01 -8.09887767e-01 -6.69877231e-01 -3.62105489e-01
-7.56662190e-02 -2.37817541e-01 -1.08345687e+00 1.52308953e+00
-2.30576888e-01 9.92687419e-02 -7.94758648e-02 5.71631551e-01
3.80790561e-01 7.60394454e-01 -1.47666112e-01 -4.26248491e-01
1.34201086e+00 -7.23322272e-01 -8.21985006e-01 -4.30335313e-01
6.33439004e-01 -8.80268216e-01 1.11348903e+00 3.36617142e-01
-8.66929948e-01 -3.51622313e-01 -1.28385794e+00 9.79230702e-02
-3.15189034e-01 5.34839869e-01 3.70632946e-01 1.01833797e+00
-1.06626117e+00 3.15824628e-01 -7.95490086e-01 -7.04666302e-02
1.52602866e-01 5.36171615e-01 2.15819106e-01 3.46041411e-01
-1.31879222e+00 3.89299959e-01 -1.72618687e-01 2.96072036e-01
-7.19454587e-01 -5.74315786e-01 -2.60997087e-01 3.95736516e-01
4.61500049e-01 -8.51993188e-02 1.68135595e+00 -3.52600187e-01
-1.60620916e+00 2.56538630e-01 -4.20854598e-01 -7.10450411e-01
2.16870289e-02 -5.35967588e-01 -1.24668336e+00 2.92489193e-02
-3.61554384e-01 1.68520007e-02 9.01025593e-01 -5.86645782e-01
-6.31964624e-01 -9.70953107e-02 -3.07340711e-01 -3.08976561e-01
-9.15621102e-01 3.92726213e-01 -6.48677468e-01 -8.63653839e-01
7.65801072e-02 -8.03698301e-01 -1.34437010e-01 -3.88717622e-01
-4.60050553e-01 -8.99243280e-02 1.12368727e+00 -5.34931600e-01
1.68714368e+00 -2.35350919e+00 -6.71099722e-01 4.49118167e-01
2.44010642e-01 8.78367007e-01 -6.43996447e-02 2.34850287e-01
1.20413065e-01 2.21406013e-01 2.87023097e-01 -3.43975753e-01
-2.10155100e-01 -1.10751070e-01 -4.72694695e-01 2.29032636e-01
2.32871145e-01 1.02524948e+00 -6.44918680e-01 1.59147546e-01
1.66169003e-01 6.01817429e-01 -3.56453151e-01 3.43800277e-01
4.68917750e-02 -4.73975569e-01 -3.69248658e-01 7.14029431e-01
1.47618309e-01 -3.48008931e-01 2.27900788e-01 -2.61152536e-01
-1.73383325e-01 8.80297005e-01 -9.61989462e-01 1.16216695e+00
-7.94406235e-01 9.83901381e-01 -4.84991185e-02 -7.41327345e-01
9.61989760e-01 3.60397130e-01 -3.53395157e-02 -1.16043043e+00
1.60894737e-01 3.63201082e-01 1.80095043e-02 -5.42858690e-02
7.05460250e-01 4.53400105e-01 -1.05137706e-01 3.98032606e-01
-3.87430266e-02 5.68466663e-01 -5.13167679e-01 3.44793618e-01
1.35298586e+00 -8.33582997e-01 8.34115520e-02 -1.79940015e-01
3.17192450e-02 -7.73469567e-01 3.23926270e-01 1.25866389e+00
-1.74258485e-01 5.30785501e-01 4.46094573e-01 -4.05287236e-01
-8.26114774e-01 -1.10140300e+00 -3.59776914e-02 1.22309208e+00
-2.31422231e-01 -6.07118070e-01 -6.87343717e-01 -1.23567246e-01
-2.33820885e-01 4.74537045e-01 -1.70367971e-01 -4.02111620e-01
-5.92676938e-01 -5.40009677e-01 1.34751642e+00 7.67850757e-01
4.48707730e-01 -7.44542241e-01 -7.46710062e-01 3.97230685e-01
7.11568743e-02 -1.53465378e+00 -6.80120170e-01 5.06411493e-01
-2.89425075e-01 -4.79564190e-01 -6.02358401e-01 -6.13224685e-01
1.43444687e-01 5.87189615e-01 8.24135900e-01 -7.93580711e-02
-5.66455960e-01 1.18482828e-01 -4.10900950e-01 -3.64829451e-01
-1.87384456e-01 3.87255132e-01 5.34776628e-01 -3.65283266e-02
3.40855271e-01 -6.35402143e-01 -5.60574472e-01 4.07064021e-01
-5.02243996e-01 -3.26771110e-01 8.74895513e-01 7.50483453e-01
2.19463781e-01 -6.48607388e-02 6.45259202e-01 -5.23528814e-01
9.18999195e-01 -1.03050910e-01 -6.44293249e-01 3.94064069e-01
-8.32377017e-01 1.75105762e-02 5.46165109e-01 -8.21147859e-01
-6.51235282e-01 -2.19285533e-01 -4.29360926e-01 -5.49605370e-01
1.47855058e-01 2.52767831e-01 -4.69697453e-02 -1.91534922e-01
7.16721654e-01 3.39468181e-01 -2.06757337e-01 -4.58198220e-01
1.23073049e-01 1.29789698e+00 4.16398972e-01 -2.24729896e-01
3.18161935e-01 1.16929738e-02 -5.14188170e-01 -1.37290609e+00
-3.10817689e-01 -6.57218039e-01 1.48194414e-02 -7.33275563e-02
1.84370518e-01 -1.03331780e+00 -9.73271310e-01 4.20184553e-01
-1.18135846e+00 -1.57019079e-01 7.27023259e-02 3.03808391e-01
-2.95629017e-02 -7.24169910e-02 -7.80414045e-01 -1.42888665e+00
-8.96008074e-01 -1.12487102e+00 1.01066816e+00 9.88701954e-02
-3.76310676e-01 -2.09904581e-01 -7.23606229e-01 -3.19573400e-03
1.08262157e+00 -6.01319551e-01 6.74499571e-01 -8.67461026e-01
-4.84475732e-01 -4.00968283e-01 -3.96438926e-01 3.40692699e-01
1.56713262e-01 -2.80560166e-01 -1.33058500e+00 -2.22279072e-01
-8.76281634e-02 -5.36401607e-02 7.01207340e-01 3.40390235e-01
1.45936775e+00 -5.06200016e-01 -3.79294157e-01 4.00232881e-01
9.11101520e-01 5.12063205e-01 6.14052892e-01 -2.24150628e-01
6.66293740e-01 2.89250631e-02 1.19467653e-01 6.41459882e-01
-1.46934152e-01 9.03696835e-01 -1.36121273e-01 8.23590532e-02
-9.92833599e-02 -2.49475390e-01 5.14118373e-01 1.08048391e+00
5.10733604e-01 -4.97669667e-01 -1.12973404e+00 6.37730539e-01
-1.64920318e+00 -7.31696010e-01 3.10291231e-01 2.18810940e+00
6.48093581e-01 7.57092834e-01 2.55914271e-01 4.03971404e-01
5.82542121e-01 2.79862255e-01 -6.06098890e-01 -7.04415262e-01
-6.18334971e-02 4.32218134e-01 8.16087544e-01 5.43841898e-01
-8.85733664e-01 9.97877121e-01 6.88856363e+00 1.22859454e+00
-1.33035743e+00 -8.88015181e-02 9.33021903e-01 -5.23596525e-01
-4.62235510e-02 -5.19765854e-01 -1.21974885e+00 3.36976409e-01
1.87929237e+00 1.62725940e-01 5.77227712e-01 9.57167804e-01
2.52675325e-01 3.87671530e-01 -7.87627280e-01 1.06982958e+00
-6.70290366e-02 -1.57374287e+00 -4.38854545e-02 7.34586641e-02
9.77772549e-02 2.53100574e-01 3.00309837e-01 2.42590219e-01
7.16762319e-02 -1.18231356e+00 6.07688665e-01 2.63815373e-01
1.22323549e+00 -1.04129922e+00 5.44989705e-01 6.73269704e-02
-1.39079976e+00 -2.75757015e-01 -5.75304516e-02 -4.28986177e-02
-8.91475764e-04 9.10767078e-01 -1.03304827e+00 -9.82044265e-03
9.43610251e-01 5.35505079e-02 -1.93106875e-01 6.78467870e-01
3.56624015e-02 9.39733863e-01 -7.27916837e-01 -6.43589854e-01
2.60965735e-01 5.70012510e-01 3.32630098e-01 1.65029478e+00
1.21378526e-01 1.74331188e-01 -7.54916966e-02 5.78541815e-01
-4.11254019e-01 -2.59367168e-01 -3.53215784e-01 -2.98322737e-01
1.18081200e+00 8.68361831e-01 -4.72586721e-01 -7.06753880e-02
-2.98965156e-01 1.01425827e+00 1.98294505e-01 5.02760470e-01
-6.37337565e-01 -9.31490481e-01 9.74751055e-01 1.15536012e-01
3.20032895e-01 -4.25969124e-01 -4.05325592e-01 -6.45540774e-01
3.52611303e-01 -8.77699375e-01 -1.56847730e-01 -3.19219410e-01
-7.12159872e-01 7.77797937e-01 -5.12875736e-01 -9.16734636e-01
-2.37720102e-01 -6.04782641e-01 -4.43852484e-01 9.32615042e-01
-1.25476623e+00 -8.81304383e-01 1.78477466e-01 3.15358698e-01
6.15874350e-01 -1.94360703e-01 8.80590320e-01 4.98467326e-01
-6.95465684e-01 1.53728330e+00 4.88841310e-02 3.59036833e-01
2.06768587e-01 -7.35241830e-01 1.28860068e+00 1.06734431e+00
1.88248858e-01 1.09922886e+00 4.54404801e-01 -3.79342705e-01
-1.61216176e+00 -1.09226942e+00 9.74716961e-01 1.23324003e-02
7.22607851e-01 -8.78425896e-01 -8.22571874e-01 8.79356042e-02
-2.17043236e-01 2.28831857e-01 5.33746481e-01 5.21899045e-01
-7.31771648e-01 -2.66591460e-01 -6.18750155e-01 9.62293088e-01
1.22871459e+00 -1.28349435e+00 9.97011587e-02 2.73753721e-02
1.40237403e+00 -3.40471953e-01 -3.96926790e-01 3.81819248e-01
9.61232722e-01 -5.77035546e-01 1.17539668e+00 -5.46274841e-01
-3.77268821e-01 -2.15815976e-02 -2.89292634e-01 -8.75287473e-01
-2.66637683e-01 -1.13805223e+00 -4.80793059e-01 1.02651572e+00
9.52880263e-01 -5.00212729e-01 9.03511703e-01 6.64185941e-01
-2.81436861e-01 -8.88814211e-01 -1.16358101e+00 -1.03414857e+00
-4.63868588e-01 -8.09513211e-01 4.32119697e-01 3.60420316e-01
2.32823845e-02 3.31956506e-01 -7.76334524e-01 4.48604822e-01
1.50605500e-01 -4.00196135e-01 3.46893340e-01 -6.20284557e-01
-2.99210936e-01 -5.36549509e-01 -3.52008969e-01 -1.53979111e+00
-2.34109372e-01 -6.59535229e-02 1.14639491e-01 -7.66312480e-01
-2.99618840e-01 -5.07441461e-01 -6.94920123e-01 6.66415751e-01
-1.52059961e-02 5.95050417e-02 -1.71883088e-02 -2.55656354e-02
-6.44738674e-01 3.62713069e-01 1.18499115e-01 -1.00209214e-01
-3.98684800e-01 5.82735278e-02 -6.96721852e-01 3.45473796e-01
9.93548989e-01 -2.92909503e-01 -4.37959492e-01 -4.26470876e-01
-2.75120558e-03 -5.36497384e-02 -1.32595524e-01 -1.24611747e+00
7.21260548e-01 6.57052547e-03 1.75504729e-01 -5.44032395e-01
6.16970479e-01 -5.58097959e-01 -1.31243870e-01 5.00298858e-01
-5.77895463e-01 1.51749983e-01 3.94023806e-01 6.41206801e-01
2.14598011e-02 2.73779482e-01 6.94473088e-01 3.70571971e-01
-4.92735177e-01 9.20918509e-02 -6.92687094e-01 -1.90775543e-01
4.59953129e-01 7.17295855e-02 -4.47533995e-01 -4.28234160e-01
-1.40087947e-01 -1.82041451e-02 -3.03326666e-01 8.06729257e-01
1.01523232e+00 -9.98441935e-01 -1.76016644e-01 4.43148255e-01
2.80863464e-01 -4.95805949e-01 5.90637922e-02 3.86215895e-01
-1.84438661e-01 8.20594430e-01 4.43508625e-01 -4.20899153e-01
-1.35639751e+00 8.97288471e-02 5.40625870e-01 7.35441893e-02
-5.34345210e-01 1.15233898e+00 -3.38342220e-01 4.15804796e-02
8.45200837e-01 -4.35073286e-01 1.84655160e-01 -4.01080042e-01
1.16803718e+00 3.57313246e-01 6.53143167e-01 -9.20198262e-02
-6.41105056e-01 -1.09136723e-01 -6.98283911e-01 -1.27675489e-01
8.35356295e-01 3.60949263e-02 5.24501920e-01 5.52279115e-01
1.42098665e+00 -5.61044924e-02 -7.13088632e-01 -5.02109110e-01
2.89948076e-01 -1.12035319e-01 4.83873576e-01 -7.88675129e-01
-6.69364929e-01 9.52384055e-01 9.68418539e-01 1.30842999e-01
9.45425689e-01 -9.09330770e-02 1.15488851e+00 6.46998763e-01
2.15303689e-01 -1.26403582e+00 1.63430378e-01 7.08217800e-01
5.37291408e-01 -7.60209560e-01 -2.72111833e-01 -1.52374536e-01
-3.01476151e-01 9.20037389e-01 4.81615722e-01 2.03558281e-01
7.65674174e-01 8.78625810e-01 -9.60120931e-03 -1.24013247e-02
-9.70018029e-01 -1.35468781e-01 1.24765567e-01 2.69006938e-01
3.82830799e-01 3.38009953e-01 2.83232301e-01 6.68644190e-01
-3.72073092e-02 -2.50859618e-01 1.61031976e-01 8.56463253e-01
-6.57698631e-01 -9.23665881e-01 -5.20101786e-02 9.10747111e-01
-6.23761773e-01 -5.89996636e-01 -4.45986986e-01 8.38922411e-02
-6.70874774e-01 1.31123734e+00 1.01032510e-01 -1.09382844e+00
6.08472049e-01 -5.64953350e-02 -2.83158362e-01 -3.35059464e-01
-6.18550479e-01 1.83725625e-01 3.57180476e-01 -7.91787863e-01
1.33788154e-01 -2.43838146e-01 -1.06009376e+00 -5.66957295e-01
-6.55428648e-01 1.52595237e-01 8.28558147e-01 9.40683305e-01
9.71638143e-01 9.06162143e-01 7.32904315e-01 -2.96776384e-01
-4.78146464e-01 -9.31426704e-01 -4.03733552e-01 -3.84387523e-01
7.24396110e-01 -2.72302002e-01 -2.82220006e-01 -4.60858494e-01]
|
[14.354581832885742, 6.223809719085693]
|
979ee4ef-278a-4e59-9244-a15590898bd7
|
lider-an-efficient-high-dimensional-learned
|
2205.00970
| null |
https://arxiv.org/abs/2205.00970v3
|
https://arxiv.org/pdf/2205.00970v3.pdf
|
LIDER: An Efficient High-dimensional Learned Index for Large-scale Dense Passage Retrieval
|
Many recent approaches of passage retrieval are using dense embeddings generated from deep neural models, called "dense passage retrieval". The state-of-the-art end-to-end dense passage retrieval systems normally deploy a deep neural model followed by an approximate nearest neighbor (ANN) search module. The model generates embeddings of the corpus and queries, which are then indexed and searched by the high-performance ANN module. With the increasing data scale, the ANN module unavoidably becomes the bottleneck on efficiency. An alternative is the learned index, which achieves significantly high search efficiency by learning the data distribution and predicting the target data location. But most of the existing learned indexes are designed for low dimensional data, which are not suitable for dense passage retrieval with high-dimensional dense embeddings. In this paper, we propose LIDER, an efficient high-dimensional Learned Index for large-scale DEnse passage Retrieval. LIDER has a clustering-based hierarchical architecture formed by two layers of core models. As the basic unit of LIDER to index and search data, a core model includes an adapted recursive model index (RMI) and a dimension reduction component which consists of an extended SortingKeys-LSH (SK-LSH) and a key re-scaling module. The dimension reduction component reduces the high-dimensional dense embeddings into one-dimensional keys and sorts them in a specific order, which are then used by the RMI to make fast prediction. Experiments show that LIDER has a higher search speed with high retrieval quality comparing to the state-of-the-art ANN indexes on passage retrieval tasks, e.g., on large-scale data it achieves 1.2x search speed and significantly higher retrieval quality than the fastest baseline in our evaluation. Furthermore, LIDER has a better capability of speed-quality trade-off.
|
['Daisy Zhe Wang', 'Haodi Ma', 'Yifan Wang']
|
2022-05-02
| null | null | null | null |
['passage-retrieval']
|
['natural-language-processing']
|
[-8.90344679e-01 -7.55570710e-01 -3.34721893e-01 1.26381069e-01
-1.22179031e+00 -5.10928452e-01 5.62969387e-01 3.84490043e-01
-7.98149586e-01 3.78173202e-01 6.59706533e-01 -3.75684798e-02
-5.22480428e-01 -1.19976234e+00 -5.06233454e-01 -5.16093373e-01
-8.33107680e-02 1.10479069e+00 4.27056998e-01 -3.79319578e-01
2.84530967e-01 3.33353609e-01 -1.57819927e+00 2.94219553e-01
7.03334630e-01 1.12290251e+00 2.00769097e-01 7.75167346e-01
-3.75736088e-01 4.17064071e-01 -5.18444300e-01 5.23793362e-02
3.25912774e-01 -2.03848228e-01 -9.16571796e-01 -1.00680983e+00
2.08369479e-01 -7.25443840e-01 -9.75423038e-01 5.89327931e-01
1.05285728e+00 5.29806077e-01 7.77751386e-01 -9.13283885e-01
-9.92455959e-01 5.67262053e-01 -5.17427512e-02 4.68294501e-01
4.10423398e-01 -2.30422303e-01 1.42345905e+00 -1.21385849e+00
4.34183300e-01 1.03190231e+00 5.05652666e-01 3.81168753e-01
-9.54311848e-01 -3.59288871e-01 -2.54355013e-01 3.65039170e-01
-1.82085669e+00 -1.96701452e-01 3.89738947e-01 9.76893958e-03
1.39612651e+00 4.40209061e-01 7.06513882e-01 6.89982295e-01
-6.12647980e-02 1.14284956e+00 1.88118353e-01 -3.72485220e-01
3.39867741e-01 -1.99561790e-02 4.04712439e-01 3.81343991e-01
1.56100497e-01 4.08021808e-02 -3.18181872e-01 -4.86718088e-01
4.81065184e-01 5.18904388e-01 -3.78164351e-01 -7.42076710e-02
-9.97507215e-01 9.85187471e-01 9.97556746e-01 4.18936104e-01
-5.19899726e-01 3.83755974e-02 6.56779766e-01 4.67476308e-01
2.76637763e-01 6.69200182e-01 -4.67180818e-01 -2.50311822e-01
-1.24395716e+00 6.09290183e-01 1.02869272e+00 9.10974383e-01
6.84355676e-01 -5.99046052e-01 -6.51174605e-01 1.22432375e+00
3.23187709e-01 5.30295432e-01 9.23627853e-01 -4.76222903e-01
4.52702016e-01 7.26804614e-01 -3.70335691e-02 -1.05169392e+00
-5.05926192e-01 -4.69813198e-01 -9.58060503e-01 -5.39587379e-01
-1.19920336e-01 2.81569362e-01 -7.47638822e-01 1.37913549e+00
1.73768252e-01 -1.44121414e-02 1.33822754e-01 1.01215589e+00
8.73416603e-01 1.18808794e+00 -1.85951561e-01 9.28022861e-02
1.28532004e+00 -1.44130313e+00 -4.28907931e-01 2.39277244e-01
1.07611096e+00 -7.56137311e-01 1.38529992e+00 -4.67850417e-02
-1.21566021e+00 -5.56517541e-01 -9.06732857e-01 -7.14763224e-01
-7.54806459e-01 5.61617501e-02 2.93277651e-01 -3.37894820e-02
-1.31300008e+00 4.74372417e-01 -6.70062304e-01 -4.24220651e-01
3.43535841e-02 2.85119742e-01 -1.07348956e-01 -3.18458050e-01
-1.63191628e+00 6.27566993e-01 6.33977532e-01 -7.26246461e-02
-6.35727286e-01 -9.74273324e-01 -7.23975241e-01 6.14717245e-01
-1.33543372e-01 -9.44161355e-01 1.12574422e+00 1.51544601e-01
-1.23286450e+00 5.60985565e-01 8.39615765e-04 -4.62521613e-01
7.06294551e-02 -5.28577626e-01 -3.83842021e-01 3.37166339e-01
5.61756007e-02 5.99248230e-01 3.95400912e-01 -8.00770998e-01
-6.42496884e-01 -1.87705740e-01 -4.28190129e-03 3.85787040e-01
-9.79728758e-01 -1.20833747e-01 -1.21723008e+00 -5.52500665e-01
-1.72378272e-01 -8.19827080e-01 -4.49292809e-02 -8.68130922e-02
-1.01715967e-01 -7.24352717e-01 7.45258451e-01 -4.74964827e-01
2.23303962e+00 -2.15502405e+00 2.66708851e-01 3.61594766e-01
3.76491159e-01 7.14803398e-01 -4.93282765e-01 9.09512043e-01
3.91875982e-01 6.00653887e-02 3.59311283e-01 -5.42958260e-01
5.26664674e-01 7.35621676e-02 -3.92761856e-01 1.66316137e-01
-4.87862766e-01 1.23739529e+00 -8.94744098e-01 -6.22226596e-01
1.11532370e-02 4.71475810e-01 -8.91849220e-01 6.77151561e-01
-2.65023917e-01 -6.13528252e-01 -5.11156142e-01 4.25378442e-01
5.16104162e-01 -4.70869064e-01 -3.60939562e-01 -2.22384050e-01
-1.28493058e-02 5.64216554e-01 -7.54483938e-01 2.05157709e+00
-5.52849948e-01 4.64286506e-01 -3.64341348e-01 -5.84533334e-01
9.01083887e-01 1.28413886e-01 3.85430038e-01 -1.19442546e+00
-1.04581214e-01 4.08874184e-01 -4.29271519e-01 -4.23374206e-01
1.07115328e+00 4.17116165e-01 -3.94660860e-01 7.95194685e-01
-1.24559082e-01 1.84595346e-01 4.01832879e-01 6.36577308e-01
1.36053777e+00 -3.18133235e-01 -2.59129256e-01 -1.40832826e-01
5.45695305e-01 1.27346665e-01 2.43454814e-01 9.49732363e-01
2.33638614e-01 7.44355857e-01 6.80301413e-02 -5.88916659e-01
-1.30634749e+00 -1.00988853e+00 -4.49419655e-02 1.29323626e+00
1.39881700e-01 -8.29764783e-01 -6.10613823e-01 -2.79151708e-01
2.44890600e-01 2.26476848e-01 -4.52704310e-01 -5.94612420e-01
-6.82807982e-01 -4.84784096e-01 6.82041585e-01 5.62054694e-01
5.73859990e-01 -1.23015022e+00 -1.76643685e-01 3.30842495e-01
-2.58150160e-01 -7.00179160e-01 -9.13639784e-01 -1.04793943e-01
-6.34601951e-01 -7.37133861e-01 -9.54933643e-01 -1.06817079e+00
4.27140981e-01 3.37019652e-01 1.29671240e+00 6.09410048e-01
-3.25546861e-01 3.32685828e-01 -7.60646820e-01 7.76690915e-02
3.49694639e-02 8.37876618e-01 3.00811380e-01 -5.38874447e-01
8.28598499e-01 -3.66694421e-01 -1.34225690e+00 2.35832259e-01
-1.25876653e+00 -5.10266185e-01 6.49740934e-01 9.54216301e-01
7.27292895e-01 -2.31307652e-02 4.69474494e-01 -2.99645483e-01
1.18314517e+00 -7.60556936e-01 -4.45404172e-01 4.59475905e-01
-8.06705236e-01 2.12560147e-01 8.91409099e-01 -3.55103552e-01
-3.50479126e-01 -6.96628690e-01 -3.86995316e-01 -7.01238096e-01
2.56248891e-01 7.33108938e-01 2.95082062e-01 3.46843630e-01
7.30993569e-01 5.24516404e-01 -6.49916232e-02 -9.23501432e-01
4.12057102e-01 1.09379637e+00 1.48472369e-01 -5.54878294e-01
7.32042551e-01 1.32194813e-02 -4.74359065e-01 -6.40542746e-01
-6.73689544e-01 -9.52055454e-01 -4.35524374e-01 4.01830465e-01
6.43584430e-01 -8.42235804e-01 -4.64468360e-01 3.47391129e-01
-9.62547243e-01 -3.85877222e-01 -4.11205262e-01 4.99488592e-01
-2.83198625e-01 2.31070057e-01 -1.19949389e+00 -1.99486986e-01
-1.12198877e+00 -8.76375437e-01 1.11686635e+00 3.25128734e-01
-9.45031419e-02 -1.00803399e+00 6.38533115e-01 8.90496895e-02
1.01069558e+00 -6.58508778e-01 1.19813478e+00 -1.06849682e+00
-5.91504037e-01 -5.76707959e-01 -3.65256608e-01 2.23748207e-01
-2.06223160e-01 -2.85470963e-01 -5.60112476e-01 -6.31554961e-01
-4.85605389e-01 -5.42189360e-01 9.09451485e-01 1.47297159e-01
1.21198070e+00 -3.45285833e-01 -4.72410202e-01 7.61035681e-01
1.57823670e+00 -3.16468291e-02 7.12335110e-01 5.85782647e-01
5.37890673e-01 -9.03797075e-02 5.84732473e-01 3.28431338e-01
6.10354662e-01 7.70820975e-01 -1.14959208e-02 -1.88258722e-01
-1.89624876e-01 -6.52999401e-01 9.41558555e-02 1.44121385e+00
5.31258106e-01 -5.08708835e-01 -9.17681754e-01 7.03018546e-01
-1.97191215e+00 -8.36234570e-01 4.80476052e-01 2.20944023e+00
9.07361984e-01 -1.14739858e-01 2.28058025e-01 6.74494877e-02
1.72197700e-01 1.44295782e-01 -4.26731706e-01 -3.61713052e-01
2.07475752e-01 2.47894332e-01 3.51664759e-02 4.50216740e-01
-8.24100018e-01 9.62410331e-01 5.83793449e+00 1.08864343e+00
-7.88842380e-01 -1.50065824e-01 1.86367661e-01 -4.49445218e-01
-3.49426091e-01 -3.15870762e-01 -1.21626461e+00 5.96120298e-01
1.06398189e+00 -3.21668386e-01 4.37806338e-01 8.81818950e-01
-3.99039164e-02 2.63886362e-01 -1.22251713e+00 1.11394763e+00
9.85438973e-02 -1.58235705e+00 4.51944411e-01 9.84813720e-02
4.39589918e-01 4.63900030e-01 -3.58866043e-02 8.86961699e-01
1.04072563e-01 -7.63879180e-01 9.52263698e-02 8.00541043e-01
6.49845362e-01 -8.01919222e-01 7.67430902e-01 4.94478315e-01
-1.32681084e+00 -2.36633107e-01 -8.25769246e-01 2.22671807e-01
1.06338523e-01 3.36654127e-01 -3.33559453e-01 3.18094105e-01
8.62744212e-01 6.53060317e-01 -3.15643430e-01 1.53445721e+00
2.18812287e-01 2.25988165e-01 -7.18931913e-01 -3.43889177e-01
5.42052686e-01 -1.52865916e-01 2.93065220e-01 1.35369754e+00
3.87894750e-01 1.14002384e-01 2.22061485e-01 5.68788648e-01
-4.17017579e-01 2.98830122e-01 -5.24520278e-01 1.95401330e-02
9.12097037e-01 1.11995363e+00 -1.11970410e-01 -5.77242553e-01
-3.32527548e-01 1.06273258e+00 6.57396972e-01 4.71904486e-01
-5.53917885e-01 -8.95101309e-01 7.84515977e-01 8.52082148e-02
5.15090883e-01 -8.74729455e-02 4.31853831e-01 -1.17069411e+00
2.45034218e-01 -8.18631291e-01 8.33607435e-01 -4.79174793e-01
-1.48115730e+00 8.12123775e-01 -8.31234679e-02 -1.25574684e+00
-4.45596963e-01 -1.91436499e-01 -4.36901540e-01 9.89724159e-01
-1.75524664e+00 -8.07471931e-01 -2.17825279e-01 7.49620438e-01
5.04567504e-01 -2.83564806e-01 1.27881944e+00 8.33448648e-01
-5.58762610e-01 1.08794892e+00 7.46270657e-01 5.01333296e-01
6.13649607e-01 -1.11105883e+00 2.52278298e-01 3.30730051e-01
2.63223588e-01 9.58860278e-01 -2.81236153e-02 -2.16732800e-01
-1.69250989e+00 -9.47327852e-01 1.15782571e+00 -2.87616998e-01
7.54863322e-01 -4.11800891e-01 -1.11287177e+00 2.49145940e-01
5.25149368e-02 2.38495409e-01 9.35210526e-01 2.12029591e-01
-3.45011145e-01 -4.11764681e-01 -8.21702659e-01 6.57607615e-01
8.14863026e-01 -8.42696130e-01 -7.60709524e-01 5.24521053e-01
1.10017371e+00 -3.54590327e-01 -1.30544484e+00 1.24797694e-01
6.13968611e-01 -7.11802602e-01 1.21294999e+00 -7.25854993e-01
2.31038287e-01 -1.76250771e-01 -2.25533307e-01 -1.22012877e+00
-6.82250679e-01 -3.20057064e-01 -6.33017600e-01 1.05980957e+00
4.80266482e-01 -3.76565784e-01 6.79865658e-01 6.95068836e-01
-7.34672472e-02 -1.21257567e+00 -6.68307960e-01 -6.67228162e-01
3.61648142e-01 -4.86065075e-02 9.52571988e-01 4.28223491e-01
1.06392927e-01 4.61199403e-01 1.26134465e-02 6.73795193e-02
3.95160407e-01 5.15343070e-01 8.68119180e-01 -9.99573231e-01
-1.82910487e-01 -6.25532150e-01 -3.66762340e-01 -1.83033991e+00
-2.90699974e-02 -9.65874314e-01 -1.58560917e-01 -1.86038947e+00
2.82009572e-01 -8.04798305e-01 -7.20784009e-01 2.22275868e-01
-1.23701729e-01 9.20542106e-02 4.25425619e-02 6.79844558e-01
-1.11189783e+00 9.39885736e-01 8.14371049e-01 -1.67463601e-01
-4.79902595e-01 -1.63489118e-01 -4.65126336e-01 1.75052628e-01
5.67328751e-01 -5.03900588e-01 -5.08019030e-01 -8.71136844e-01
3.46199483e-01 -1.41648397e-01 -1.12453565e-01 -9.52707231e-01
8.65359366e-01 5.15998065e-01 3.38792890e-01 -8.76689732e-01
2.89599389e-01 -8.25991452e-01 -3.99761349e-01 2.44113252e-01
-5.25914431e-01 3.94059718e-01 5.98531216e-02 4.48670685e-01
-6.04795754e-01 -1.06036507e-01 3.28722626e-01 1.64212883e-02
-5.43191671e-01 8.40080321e-01 -3.63996401e-02 3.01951498e-01
5.81938386e-01 2.01455414e-01 -3.54614884e-01 -2.92092294e-01
-4.91396785e-01 6.96302831e-01 3.01268876e-01 6.65693521e-01
9.18894768e-01 -1.73131573e+00 -6.51081562e-01 3.30490530e-01
2.23909333e-01 2.75542200e-01 1.42286763e-01 4.58099604e-01
-6.64307654e-01 7.92633891e-01 3.33380193e-01 -4.35749829e-01
-8.94714713e-01 7.97225296e-01 2.03162581e-01 -7.62747169e-01
-8.79787743e-01 7.84651101e-01 -2.73161203e-01 -3.91369522e-01
5.92478871e-01 -2.44160011e-01 -1.85115620e-01 1.82489678e-01
1.04449940e+00 3.46385270e-01 1.29101440e-01 -2.30351940e-01
-2.04825729e-01 7.04672754e-01 -6.70451462e-01 1.15388185e-01
1.41159546e+00 -2.00732306e-01 -3.10483038e-01 1.16894126e-01
1.96474433e+00 -4.12845790e-01 -5.81069767e-01 -7.64550209e-01
-9.58212242e-02 -4.03724998e-01 2.72650123e-01 -5.10309458e-01
-9.69721079e-01 7.53725827e-01 4.00780380e-01 3.33334118e-01
1.12817562e+00 -4.32828069e-02 1.64655113e+00 8.59054089e-01
3.47734421e-01 -1.25323117e+00 1.11035809e-01 9.01523173e-01
8.40998054e-01 -1.00492632e+00 -8.42534453e-02 5.81545830e-01
-1.11399382e-01 9.35355544e-01 4.85140979e-01 -3.38219225e-01
9.70112562e-01 1.07183300e-01 -3.39177065e-02 -3.66149724e-01
-8.68632674e-01 -1.32235497e-01 6.16867721e-01 1.01784013e-01
3.60580459e-02 -2.22402230e-01 -2.31413350e-01 4.96171802e-01
-1.68811873e-01 3.87344602e-03 -3.48160148e-01 8.16894948e-01
-5.53692997e-01 -1.19568658e+00 -1.68070029e-02 6.90942585e-01
-2.61382341e-01 -5.02323151e-01 2.69024782e-02 6.74995482e-01
-5.11339605e-01 6.64512992e-01 3.39123756e-01 -5.90988815e-01
4.76276964e-01 -8.71692505e-03 -2.88260076e-02 -4.68703777e-01
-7.53422320e-01 -1.89475477e-01 -2.41878361e-01 -8.11101615e-01
2.27109060e-01 -1.58740625e-01 -1.30885899e+00 -5.91136158e-01
-3.51138085e-01 8.27067375e-01 4.86581892e-01 6.57494664e-01
8.41979325e-01 7.76909664e-02 7.57463038e-01 -7.01125205e-01
-7.87661552e-01 -1.01062548e+00 -5.93657553e-01 4.85975504e-01
4.51283604e-01 -1.56084836e-01 -5.65094411e-01 -7.62111783e-01]
|
[11.415264129638672, 7.6404500007629395]
|
bbd54393-bedb-48ec-bffc-7d36cdae797a
|
fba-net-foreground-and-background-aware
|
2306.15189
| null |
https://arxiv.org/abs/2306.15189v1
|
https://arxiv.org/pdf/2306.15189v1.pdf
|
FBA-Net: Foreground and Background Aware Contrastive Learning for Semi-Supervised Atrium Segmentation
|
Medical image segmentation of gadolinium enhancement magnetic resonance imaging (GE MRI) is an important task in clinical applications. However, manual annotation is time-consuming and requires specialized expertise. Semi-supervised segmentation methods that leverage both labeled and unlabeled data have shown promise, with contrastive learning emerging as a particularly effective approach. In this paper, we propose a contrastive learning strategy of foreground and background representations for semi-supervised 3D medical image segmentation (FBA-Net). Specifically, we leverage the contrastive loss to learn representations of both the foreground and background regions in the images. By training the network to distinguish between foreground-background pairs, we aim to learn a representation that can effectively capture the anatomical structures of interest. Experiments on three medical segmentation datasets demonstrate state-of-the-art performance. Notably, our method achieves a Dice score of 91.31% with only 20% labeled data, which is remarkably close to the 91.62% score of the fully supervised method that uses 100% labeled data on the left atrium dataset. Our framework has the potential to advance the field of semi-supervised 3D medical image segmentation and enable more efficient and accurate analysis of medical images with a limited amount of annotated labels.
|
['Jihun Hamm', 'Nassir Marrouche', 'Chao Huang', 'Chanho Lim', 'Yunsung Chung']
|
2023-06-27
| null | null | null | null |
['contrastive-learning', 'medical-image-segmentation', 'contrastive-learning']
|
['computer-vision', 'medical', 'methodology']
|
[ 4.96464759e-01 4.24010932e-01 -2.81149536e-01 -6.19482458e-01
-1.21907926e+00 -4.74354774e-01 2.08886206e-01 2.05434784e-01
-6.45512760e-01 5.01694083e-01 -1.08688198e-01 -4.59190428e-01
1.87428251e-01 -2.66037583e-01 -2.65525669e-01 -9.56125200e-01
-2.39253968e-01 7.99548447e-01 2.11400703e-01 2.41335541e-01
-1.55526996e-01 5.18218994e-01 -8.48781228e-01 3.29483330e-01
9.48071539e-01 1.06975651e+00 3.27298313e-01 5.22534668e-01
-2.38114670e-01 9.52505171e-01 -4.27277714e-01 -6.52282638e-03
3.84912789e-01 -5.53282678e-01 -1.22274041e+00 4.27348435e-01
2.92846560e-01 -3.01383168e-01 -1.20039970e-01 1.04566109e+00
7.11223006e-01 -2.18125373e-01 8.06438088e-01 -7.34949291e-01
-3.69600505e-02 5.26817620e-01 -8.67550373e-01 7.55701482e-01
-3.18525434e-01 4.77021188e-02 5.94875336e-01 -7.08748996e-01
5.83730996e-01 5.14433205e-01 5.83780289e-01 7.10673928e-01
-1.23826087e+00 -6.52872145e-01 -5.66916317e-02 -2.73747653e-01
-1.21722662e+00 -1.22392833e-01 7.07572341e-01 -7.27427959e-01
4.37467366e-01 5.54747693e-02 6.76610231e-01 4.59782869e-01
1.26596555e-01 1.18842876e+00 1.46805120e+00 -3.10892105e-01
1.14767797e-01 -1.88266054e-01 2.73253590e-01 9.18507516e-01
2.55476415e-01 -2.03738153e-01 9.23821777e-02 -1.02725580e-01
9.96003509e-01 1.82375655e-01 -2.27789298e-01 -8.07481706e-01
-1.31057966e+00 7.70102441e-01 4.83663946e-01 4.76291895e-01
-5.49201965e-01 -3.28890160e-02 4.53792989e-01 -3.00528318e-01
7.88268030e-01 3.06728870e-01 -2.99523205e-01 1.56989738e-01
-1.46309662e+00 -3.02144974e-01 5.62515259e-01 4.87887740e-01
4.21770483e-01 7.48356879e-02 -1.67663783e-01 8.02222192e-01
2.56303012e-01 3.11238647e-01 5.55239320e-01 -8.36030900e-01
1.28390148e-01 6.99297845e-01 -1.61144599e-01 -3.78619581e-01
-6.41350627e-01 -5.71203053e-01 -8.31246555e-01 3.50556970e-01
6.72870100e-01 -2.48936579e-01 -1.55730295e+00 1.22325909e+00
4.92910087e-01 9.58094224e-02 -1.47444636e-01 1.00572824e+00
1.10706723e+00 2.02609152e-01 2.93195367e-01 -3.03555042e-01
1.26257288e+00 -1.02253938e+00 -7.21631467e-01 -3.07147056e-01
8.50841701e-01 -5.59298098e-01 7.62127697e-01 1.14570782e-01
-1.27934563e+00 -9.69841853e-02 -1.05884480e+00 2.39469841e-01
6.21643923e-02 1.36362910e-01 7.34408081e-01 8.79371405e-01
-8.52085829e-01 4.82934535e-01 -1.47516322e+00 3.23813945e-01
1.17100835e+00 6.35226071e-01 -3.36217493e-01 -2.81752020e-01
-8.09220076e-01 9.12139058e-01 4.75664437e-01 1.61266044e-01
-1.00893915e+00 -9.41318274e-01 -8.55852544e-01 -4.10297334e-01
4.53552425e-01 -3.12714130e-01 1.20878875e+00 -9.14554477e-01
-1.13728297e+00 1.61532080e+00 1.62448853e-01 -4.77002859e-01
7.91564584e-01 1.62363753e-01 -7.86741674e-02 8.00289512e-01
2.90528387e-01 7.67531455e-01 7.15898216e-01 -1.33085263e+00
-4.79514629e-01 -5.18024445e-01 -2.72769362e-01 1.07589662e-01
2.51312941e-01 1.32838681e-01 -3.46972555e-01 -6.40416324e-01
5.16135395e-01 -9.47953463e-01 -6.84702277e-01 9.49747488e-02
-3.40645939e-01 1.07498996e-01 6.11382365e-01 -9.30744588e-01
8.93263042e-01 -1.99256873e+00 -2.12938502e-01 2.74844557e-01
8.25349867e-01 5.20842314e-01 3.25043231e-01 -6.54289246e-01
-1.79306701e-01 6.40221238e-02 -7.94049501e-01 -1.45986214e-01
-4.62596476e-01 1.72860369e-01 2.77778536e-01 7.27250576e-01
2.20782563e-01 1.14424503e+00 -1.13451612e+00 -9.57602739e-01
3.17570955e-01 4.92703289e-01 -4.15538222e-01 1.55266836e-01
1.95749968e-01 1.02368402e+00 -4.84559268e-01 7.82450855e-01
6.61022663e-01 -7.21927941e-01 4.67127293e-01 -1.46005914e-01
3.08699220e-01 1.22618735e-01 -8.51419032e-01 1.81608212e+00
-2.29617491e-01 4.10947740e-01 3.30751151e-01 -1.31909478e+00
7.11099565e-01 3.93620819e-01 1.16025889e+00 -5.87620020e-01
3.73262703e-01 4.78454083e-01 2.07806066e-01 -4.28860515e-01
-1.78611860e-01 -5.47261775e-01 1.23760931e-01 6.90216660e-01
1.99297383e-01 -3.95373136e-01 1.14525631e-01 1.63105413e-01
9.10808563e-01 1.50690421e-01 1.93830803e-01 -4.57809538e-01
3.80566388e-01 8.04733709e-02 5.34323812e-01 7.45899975e-01
-7.15896726e-01 8.66821289e-01 3.92212033e-01 -6.20509267e-01
-9.05629218e-01 -1.06137896e+00 -3.96329910e-01 6.12251639e-01
1.02041371e-01 1.15451999e-02 -1.04364729e+00 -1.27413762e+00
-3.67058188e-01 1.44259170e-01 -6.39560223e-01 6.62746727e-02
-8.20573866e-01 -1.12113702e+00 4.29890186e-01 7.15999603e-01
4.48439956e-01 -9.95153546e-01 -9.69668984e-01 1.80669039e-01
-4.28265631e-01 -1.33412862e+00 -2.90913790e-01 4.86385167e-01
-1.28981435e+00 -1.29543209e+00 -1.19901836e+00 -1.01461530e+00
1.14221561e+00 1.38566822e-01 1.32271528e+00 1.76280990e-01
-7.25657523e-01 3.04446131e-01 -7.93432221e-02 -2.45190278e-01
-4.78351891e-01 -2.74046920e-02 -3.77696753e-01 -2.42449462e-01
9.42763388e-02 -2.37952724e-01 -9.53857660e-01 3.44525546e-01
-9.01297808e-01 1.10841490e-01 6.05544209e-01 1.00574410e+00
1.09674668e+00 -2.50223875e-01 3.46933872e-01 -1.37017202e+00
7.55159706e-02 -3.16123366e-01 -3.68632615e-01 2.82337755e-01
-4.79131162e-01 -1.01009443e-01 1.28697366e-01 -3.04966092e-01
-1.06261706e+00 4.46928084e-01 -1.13738440e-01 -3.77484709e-01
-3.20096225e-01 3.23806942e-01 2.50815988e-01 -2.92724729e-01
5.02197087e-01 -1.60771683e-02 2.82204837e-01 -2.81323731e-01
1.96429670e-01 5.76609731e-01 6.32549465e-01 -3.90925556e-01
3.54759157e-01 9.25315797e-01 1.95812918e-02 -4.89881188e-01
-1.24260771e+00 -7.09088862e-01 -1.03935683e+00 -2.35724166e-01
1.16040635e+00 -7.28851676e-01 -1.55812070e-01 2.80655533e-01
-6.26169920e-01 -6.41019702e-01 -4.55531210e-01 6.63805187e-01
-6.40940189e-01 5.31180143e-01 -7.67385304e-01 -4.50530380e-01
-5.13783097e-01 -1.58986890e+00 9.40964997e-01 7.02416003e-02
-6.49359226e-02 -1.19554806e+00 -1.86532870e-01 6.85152292e-01
2.55995810e-01 5.57021022e-01 7.91497707e-01 -9.59019542e-01
-3.08796346e-01 -2.55061090e-01 -2.86040515e-01 5.21100581e-01
4.24272507e-01 -5.30231476e-01 -9.16266978e-01 -2.53187448e-01
1.90571845e-01 -4.20048743e-01 9.69234705e-01 9.50864315e-01
1.35007894e+00 2.55211890e-01 -2.48931229e-01 7.68604875e-01
1.06447947e+00 2.65756935e-01 3.33482742e-01 5.78439608e-02
7.93784916e-01 6.32419527e-01 5.94991565e-01 2.40232438e-01
1.02387451e-01 1.83650061e-01 3.50603402e-01 -8.23046625e-01
-3.14743221e-01 1.11399405e-01 -3.94508928e-01 7.28761673e-01
-1.82888508e-02 3.21011692e-01 -1.36082506e+00 5.30361354e-01
-1.42415380e+00 -4.73165929e-01 2.43996605e-02 1.91714334e+00
1.03340590e+00 2.61060834e-01 2.05427423e-01 1.55156076e-01
8.69804859e-01 3.39766853e-02 -6.53675854e-01 1.95052743e-01
1.16711818e-01 5.96871436e-01 6.47662878e-01 3.76269937e-01
-1.58912671e+00 7.33102143e-01 6.92664385e+00 6.77444041e-01
-1.17641222e+00 3.67767483e-01 1.20111036e+00 5.30076511e-02
7.56862536e-02 -3.27437192e-01 -4.84379232e-01 3.66123021e-01
6.57497704e-01 3.04822683e-01 1.10380230e-02 6.73783958e-01
8.35896134e-02 -2.00848117e-01 -8.14448476e-01 1.06683719e+00
1.28656834e-01 -1.39882338e+00 -2.02570915e-01 7.06660897e-02
1.03300428e+00 6.02030829e-02 -1.19414059e-02 -5.07656336e-02
1.46836087e-01 -1.05516601e+00 2.45806202e-01 1.73668429e-01
1.04977727e+00 -5.49111784e-01 8.79214048e-01 2.47169927e-01
-8.18244815e-01 2.65919805e-01 5.08386716e-02 4.18771982e-01
3.14339250e-01 5.63296437e-01 -1.00867319e+00 2.35466897e-01
6.03658974e-01 5.85670233e-01 -3.92201692e-01 1.16103709e+00
-2.45329306e-01 7.64958560e-01 -1.00575492e-01 4.98487383e-01
5.00112116e-01 -1.53420448e-01 2.44507730e-01 1.19158864e+00
-2.41360247e-01 3.50060254e-01 5.51891327e-01 7.79226005e-01
-9.61890295e-02 1.91773146e-01 -2.02003136e-01 -3.97988856e-02
-1.00592315e-01 1.38473046e+00 -1.53660941e+00 -6.06392920e-01
-3.04121345e-01 7.86995053e-01 8.36462900e-02 1.09191820e-01
-6.72715247e-01 5.51294275e-02 -1.50622278e-01 2.97846884e-01
1.88120902e-01 -1.16178945e-01 -4.23262686e-01 -9.40931499e-01
-3.73266935e-01 -5.91261625e-01 5.94461799e-01 -3.34613502e-01
-1.15188873e+00 6.98371887e-01 1.43984005e-01 -1.13486826e+00
-6.61585033e-02 -6.13694489e-01 -3.24810356e-01 6.25933766e-01
-1.76849055e+00 -1.06928694e+00 -3.27723354e-01 4.63694900e-01
3.90543282e-01 5.33056632e-02 6.55025423e-01 4.88155067e-01
-3.50334972e-01 3.55687171e-01 -1.79706477e-02 5.50618231e-01
5.54535806e-01 -1.48916376e+00 2.97803525e-02 6.48892522e-01
4.10087518e-02 4.03405458e-01 1.68228745e-01 -5.95670938e-01
-7.56934285e-01 -8.97959828e-01 3.84762973e-01 -2.63951689e-01
3.89652401e-01 1.50642004e-02 -8.88499022e-01 5.86597919e-01
-1.82874829e-01 8.70865464e-01 1.03767538e+00 -4.57806498e-01
1.04803391e-01 2.60199130e-01 -1.40684366e+00 2.95062959e-01
6.75076663e-01 -3.80942732e-01 -5.36982894e-01 5.55869460e-01
3.77955973e-01 -9.15716946e-01 -9.93453145e-01 5.49465775e-01
2.87605107e-01 -7.36346066e-01 1.04169834e+00 -5.64897120e-01
3.26475501e-01 -1.10684976e-01 2.45241657e-01 -8.85186255e-01
2.93951426e-02 -4.54432935e-01 -4.08289023e-02 5.95001638e-01
3.51814449e-01 -4.04270053e-01 1.14937115e+00 7.90620029e-01
-2.92324364e-01 -1.06794691e+00 -8.29112470e-01 -4.34499979e-01
3.57593417e-01 -2.50037432e-01 -1.38654718e-02 1.02725565e+00
-2.84101248e-01 -1.30724907e-02 4.81091328e-02 -8.92159864e-02
9.84271705e-01 1.05981752e-01 1.78246289e-01 -1.24438357e+00
2.34716032e-02 -4.33976084e-01 -3.57385606e-01 -9.19543147e-01
2.72304326e-01 -1.31093562e+00 1.51863083e-01 -1.77058566e+00
5.06899238e-01 -7.58718133e-01 -5.28130233e-01 5.26590407e-01
-3.39817196e-01 8.24187517e-01 -1.17388561e-01 3.36030185e-01
-7.57830918e-01 9.70665812e-02 1.77693069e+00 -3.56321394e-01
-1.71727881e-01 1.84212938e-01 -5.65380812e-01 1.00633740e+00
7.83292353e-01 -5.66145241e-01 -2.12243706e-01 -2.46985301e-01
-3.98960620e-01 1.41850039e-01 3.11530948e-01 -6.94535792e-01
4.09168154e-02 2.77030438e-01 5.60957491e-01 -6.13184750e-01
2.31364910e-02 -7.11568832e-01 -3.91304821e-01 6.98776007e-01
-3.99506122e-01 -2.14482889e-01 6.93676323e-02 3.44227165e-01
-2.58926392e-01 -3.31077337e-01 1.18889272e+00 -6.36466980e-01
-5.93938053e-01 4.95518476e-01 -3.73758376e-01 5.86050272e-01
1.04583371e+00 -3.44095230e-01 2.12709576e-01 -5.55626117e-02
-1.23430407e+00 1.58662513e-01 1.04188859e-01 -1.35896683e-01
6.78624451e-01 -9.21606004e-01 -5.70005059e-01 1.06171735e-01
-1.15279384e-01 3.14215451e-01 3.85426521e-01 1.27300096e+00
-8.36888075e-01 2.35573128e-01 -2.55645484e-01 -1.03743660e+00
-1.06265545e+00 1.70090988e-01 6.46158576e-01 -6.02916956e-01
-8.66381586e-01 7.54903316e-01 3.04085195e-01 -5.30456662e-01
1.02071196e-01 -3.92730027e-01 -2.23933324e-01 -4.49421592e-02
4.50966746e-01 2.02705979e-01 1.98591664e-01 -8.31077814e-01
-4.93981540e-01 5.01122534e-01 -3.09585214e-01 7.53691345e-02
1.29263389e+00 3.02531961e-02 2.39461046e-02 2.25685105e-01
1.19010711e+00 -3.49054217e-01 -1.47714233e+00 -3.73853385e-01
8.25422332e-02 -3.06240737e-01 4.85509932e-01 -8.98332357e-01
-1.40920222e+00 1.15988684e+00 1.00628126e+00 -1.61286518e-01
8.77419531e-01 1.85202718e-01 7.55482614e-01 -1.47502616e-01
1.59502134e-01 -8.95707905e-01 2.11637318e-01 9.71857756e-02
2.17285171e-01 -1.65986741e+00 1.63503841e-01 -5.13010621e-01
-9.76656854e-01 8.56735706e-01 4.10905570e-01 -1.85891613e-01
7.39169002e-01 4.69931841e-01 4.69435871e-01 -5.18632472e-01
-1.09837418e-02 -2.80056864e-01 5.08256257e-01 6.58512712e-01
5.80015600e-01 8.48204270e-02 -2.72473186e-01 5.61361849e-01
3.61078382e-01 -7.81564787e-02 2.58001834e-01 1.23728096e+00
-3.26657206e-01 -8.97003293e-01 -1.22273676e-01 6.79614663e-01
-1.13581181e+00 -1.00647032e-01 -7.65177310e-02 7.60474563e-01
-6.00617155e-02 6.42383695e-01 -9.29794535e-02 3.35317343e-01
2.23865323e-02 -1.60657316e-02 6.65102899e-01 -7.77088106e-01
-7.07283854e-01 5.05973518e-01 -2.90317684e-01 -4.28188950e-01
-8.05462480e-01 -4.90371704e-01 -1.72103226e+00 4.15663928e-01
-3.76141280e-01 1.06018960e-01 6.46497965e-01 1.01680803e+00
-4.29422827e-03 5.68913758e-01 5.86948037e-01 -9.73571122e-01
-3.97739768e-01 -6.25799835e-01 -6.97628379e-01 5.65093279e-01
3.22677642e-01 -7.47058272e-01 -2.08677739e-01 1.42612025e-01]
|
[14.614500045776367, -2.249267101287842]
|
77545e2b-c969-4c55-9f5b-f1c4a7c72f61
|
trove-ontology-driven-weak-supervision-for
|
2008.01972
| null |
https://arxiv.org/abs/2008.01972v2
|
https://arxiv.org/pdf/2008.01972v2.pdf
|
Ontology-driven weak supervision for clinical entity classification in electronic health records
|
In the electronic health record, using clinical notes to identify entities such as disorders and their temporality (e.g. the order of an event relative to a time index) can inform many important analyses. However, creating training data for clinical entity tasks is time consuming and sharing labeled data is challenging due to privacy concerns. The information needs of the COVID-19 pandemic highlight the need for agile methods of training machine learning models for clinical notes. We present Trove, a framework for weakly supervised entity classification using medical ontologies and expert-generated rules. Our approach, unlike hand-labeled notes, is easy to share and modify, while offering performance comparable to learning from manually labeled training data. In this work, we validate our framework on six benchmark tasks and demonstrate Trove's ability to analyze the records of patients visiting the emergency department at Stanford Health Care for COVID-19 presenting symptoms and risk factors.
|
['Saelig Khattar', 'Scott L. Fleming', 'Jose Posada', 'Jason A. Fries', 'Ethan Steinberg', 'Nigam H. Shah', 'Alison Callahan']
|
2020-08-05
| null | null | null | null |
['temporal-information-extraction']
|
['natural-language-processing']
|
[-6.55195788e-02 3.41444999e-01 -5.39343476e-01 -5.27915776e-01
-7.77640522e-01 -8.46877694e-01 5.14771529e-02 1.28723609e+00
-6.24334753e-01 9.08149242e-01 5.29282510e-01 -6.18981361e-01
-6.61975324e-01 -5.67323029e-01 -4.53238934e-01 -2.02116862e-01
-6.35573447e-01 9.95061636e-01 -1.75742298e-01 2.52425849e-01
-3.57208580e-01 3.16468239e-01 -7.22717941e-01 6.79126561e-01
6.93787515e-01 7.33695626e-01 -2.60873586e-01 4.45588231e-01
1.22630283e-01 1.19784939e+00 -4.04210865e-01 -4.30757731e-01
9.60188508e-02 -2.04377130e-01 -1.04090321e+00 -4.26808447e-01
6.56455383e-02 -1.42668143e-01 -5.50365448e-02 5.03434539e-01
6.35809422e-01 -1.53701693e-01 6.23226106e-01 -1.36859083e+00
-3.52270663e-01 7.47338593e-01 3.20552349e-01 3.48052323e-01
3.54347557e-01 -2.59379297e-01 1.13630879e+00 -3.05019706e-01
1.39563251e+00 4.32169050e-01 1.18883407e+00 7.15921581e-01
-1.14758980e+00 -5.28226852e-01 -2.46661395e-01 -4.35437635e-02
-1.28739619e+00 -4.46994036e-01 -2.68935949e-01 -7.56008804e-01
1.17299616e+00 4.48599428e-01 4.76290405e-01 8.10310304e-01
1.35642335e-01 4.98106837e-01 6.96975350e-01 -1.05127178e-01
3.79515767e-01 2.91793644e-01 2.79219329e-01 7.23408401e-01
3.62413853e-01 -2.87287116e-01 -2.15983778e-01 -1.01667237e+00
2.38035053e-01 4.96489733e-01 -2.26530775e-01 -2.10819662e-01
-1.34414327e+00 4.05573189e-01 1.77667707e-01 2.03882977e-01
-5.92052042e-01 -2.93040454e-01 6.77316546e-01 1.31253913e-01
3.41760576e-01 7.95861185e-01 -1.17730319e+00 -8.18262920e-02
-8.58049631e-01 2.31906876e-01 1.19087040e+00 1.09769273e+00
1.13811925e-01 -7.68969059e-01 -1.27034873e-01 5.80036342e-01
2.06367165e-01 4.81407084e-02 4.66627866e-01 -7.92834103e-01
4.25709486e-01 7.72080958e-01 4.23203677e-01 -6.26463592e-01
-9.67473388e-01 -1.08926460e-01 -6.02552235e-01 -5.53388298e-01
4.84689057e-01 -5.28655827e-01 -8.04968297e-01 1.58886123e+00
4.29727376e-01 2.25012898e-01 5.36411822e-01 4.25049901e-01
1.04777980e+00 2.10905597e-01 6.88986421e-01 -3.96242768e-01
1.66780841e+00 -3.18919569e-01 -9.97202754e-01 1.59143299e-01
1.37416208e+00 -3.65430266e-01 9.06829089e-02 1.10362031e-01
-9.53461111e-01 4.30694073e-01 -3.22008520e-01 8.59135464e-02
-6.53419673e-01 -8.06430429e-02 6.85346603e-01 3.11107814e-01
-7.21788645e-01 6.28417671e-01 -1.14625537e+00 -7.65588582e-01
6.79846942e-01 3.57640713e-01 -7.09338307e-01 1.55158997e-01
-1.30414081e+00 1.04247212e+00 5.87808251e-01 -5.13786614e-01
-3.62444252e-01 -1.40632725e+00 -8.72483134e-01 1.27554923e-01
2.76790321e-01 -9.08286989e-01 1.24178016e+00 -2.35578604e-02
-5.48728228e-01 1.17670763e+00 -1.29998565e-01 -6.06348455e-01
3.26576322e-01 -5.16639687e-02 -8.93947601e-01 1.02392964e-01
2.67276675e-01 1.68552354e-01 -2.09543016e-02 -6.13968670e-01
-8.88323069e-01 -3.70595932e-01 -2.80126840e-01 -2.12904304e-01
-3.01792353e-01 3.38957071e-01 -8.21176618e-02 -4.99251246e-01
-2.06966296e-01 -8.71154130e-01 -6.10629678e-01 -1.56200364e-01
-2.96428472e-01 -1.65340617e-01 3.72343630e-01 -8.21756721e-01
1.72867632e+00 -2.16709661e+00 -3.92567635e-01 2.37441868e-01
6.36037946e-01 2.28894651e-01 4.95537788e-01 7.80606449e-01
-2.91374505e-01 4.94297534e-01 -1.20933183e-01 6.23764507e-02
-1.40134364e-01 3.25432003e-01 -9.78092328e-02 2.35299975e-01
1.73600346e-01 1.00754368e+00 -1.08781099e+00 -7.60839403e-01
-2.55565763e-01 3.59100163e-01 -8.71588588e-01 2.03336045e-01
-2.75266677e-01 3.99715930e-01 -6.53326690e-01 7.23479152e-01
7.50872865e-02 -1.04705739e+00 7.45603919e-01 -2.05423638e-01
-2.64326409e-02 5.98026097e-01 -9.35622334e-01 1.39128757e+00
-1.25720710e-01 8.95414278e-02 5.78495115e-02 -5.43546736e-01
3.60099256e-01 1.00055110e+00 1.21034908e+00 -1.49981171e-01
-1.55404685e-02 2.16040418e-01 -1.63376242e-01 -9.44045007e-01
-5.72268181e-02 -1.44852594e-01 -1.69056579e-01 6.65572643e-01
-9.51753706e-02 4.38149840e-01 2.29896486e-01 3.35947812e-01
1.64454901e+00 -3.43772620e-01 9.49813843e-01 -1.24844305e-01
4.83864136e-02 3.99699926e-01 1.00657308e+00 5.03348172e-01
-1.81873605e-01 1.85127303e-01 3.53727907e-01 -7.86108613e-01
-7.58956909e-01 -9.55486119e-01 -8.13184917e-01 7.34645844e-01
-3.50568950e-01 -9.35031056e-01 -2.85384893e-01 -8.93768370e-01
3.22797686e-01 5.11814892e-01 -6.31082773e-01 5.82964309e-02
-3.82020086e-01 -7.23851085e-01 8.60243678e-01 6.36267543e-01
-2.15936616e-01 -9.90074337e-01 -8.95375311e-01 5.59982896e-01
-4.30101484e-01 -1.30527484e+00 -5.54087102e-01 2.06786647e-01
-8.70097756e-01 -1.56884563e+00 -3.80143464e-01 -8.15472305e-01
7.31841564e-01 -6.88284755e-01 1.35216820e+00 6.79732114e-02
-6.68469787e-01 5.24192214e-01 -2.39457592e-01 -7.83668578e-01
-4.58833188e-01 1.27751783e-01 8.47948045e-02 -2.75655866e-01
8.10382485e-01 -3.21781993e-01 -6.66518450e-01 1.87328696e-01
-9.60828006e-01 -2.16291845e-01 3.08705330e-01 7.31284022e-01
5.95184088e-01 -1.74026713e-01 7.32811451e-01 -1.69615173e+00
5.40001273e-01 -9.62734222e-01 -3.24152976e-01 5.93640506e-01
-1.15588498e+00 -6.68876544e-02 4.19526815e-01 -8.31209645e-02
-8.57161880e-01 2.10600480e-01 2.15263411e-01 -7.43703498e-03
-3.04017514e-01 7.80869603e-01 3.18850189e-01 6.11128867e-01
7.91678905e-01 -2.73145497e-01 -1.93652779e-01 -4.71627980e-01
9.90268663e-02 9.11859095e-01 3.99057388e-01 -5.26953757e-01
3.60392034e-01 4.87677962e-01 -1.15340635e-01 -3.05451989e-01
-9.11697030e-01 -8.69653761e-01 -5.19288659e-01 3.94350171e-01
1.13853157e+00 -1.04980814e+00 -9.88173842e-01 -1.25498891e-01
-9.04455960e-01 -2.04306275e-01 -4.15426582e-01 7.38205791e-01
-3.91371220e-01 -9.64577645e-02 -8.02127779e-01 -3.79618198e-01
-4.90255117e-01 -5.71803153e-01 8.69404733e-01 -1.31970525e-01
-8.03746104e-01 -1.38897026e+00 4.48211908e-01 3.36448967e-01
2.56175131e-01 5.78070462e-01 1.30287206e+00 -1.34554088e+00
-3.21402639e-01 -4.69049692e-01 -6.11666124e-03 -4.07658786e-01
5.84487319e-01 -1.41048506e-01 -5.99103868e-01 -9.71065164e-02
-3.73110801e-01 -1.86416253e-01 3.02990913e-01 1.35180444e-01
1.04947758e+00 -7.29124427e-01 -9.32607412e-01 6.64697289e-01
1.27352536e+00 4.28158581e-01 1.97521061e-01 2.90666014e-01
4.27674681e-01 4.08287793e-01 4.75968808e-01 8.85286391e-01
7.27109134e-01 2.80938655e-01 -1.67683616e-01 -1.74103484e-01
5.76985061e-01 -1.07614763e-01 -3.56335193e-01 3.82132232e-01
-2.02956140e-01 -1.65520296e-01 -1.33192408e+00 7.38552749e-01
-1.88327694e+00 -1.01511669e+00 -2.22606938e-02 2.12899876e+00
1.31411672e+00 -3.27304184e-01 2.30410360e-02 -3.51484001e-01
4.83911902e-01 -4.50040251e-01 -6.09225333e-01 -1.66661009e-01
2.08212435e-01 3.72618228e-01 6.82217419e-01 1.60119846e-01
-1.09113550e+00 4.97262836e-01 6.82058048e+00 -6.75462186e-02
-8.79638851e-01 1.88101098e-01 4.54266936e-01 -2.48801604e-01
-9.71368402e-02 -1.59690738e-01 -7.22192526e-01 3.70546848e-01
1.12565076e+00 -3.42927724e-01 1.61374107e-01 7.03424394e-01
2.05874234e-01 3.71282965e-01 -1.60747159e+00 8.10932219e-01
-2.66203076e-01 -1.65704846e+00 -3.31797928e-01 1.63498357e-01
6.92659557e-01 1.98843479e-01 -2.81074703e-01 1.12614796e-01
7.36835659e-01 -1.11050069e+00 7.79719949e-02 6.66754842e-01
9.18188274e-01 -1.72859922e-01 6.18874729e-01 2.31504083e-01
-9.62629378e-01 -1.79254040e-01 1.33132264e-01 3.87178183e-01
2.63246447e-01 5.86562812e-01 -1.63176060e+00 5.49634814e-01
8.63002717e-01 9.62730348e-01 -2.24014848e-01 1.08402157e+00
2.44707167e-01 6.74269199e-01 -3.55936289e-01 3.86167943e-01
-9.35510248e-02 2.92911828e-01 1.87169567e-01 1.35521996e+00
2.89298624e-01 5.87768316e-01 2.17713550e-01 3.55506152e-01
-3.42350870e-01 4.60273236e-01 -8.65622342e-01 -4.97486562e-01
5.75167537e-01 1.26854908e+00 -4.73739088e-01 -6.26927733e-01
-3.57280731e-01 3.55272084e-01 2.44982094e-01 1.68487757e-01
-6.85568452e-01 -2.43522540e-01 6.69180930e-01 3.85582298e-01
1.89444229e-01 3.45053583e-01 -3.81752253e-02 -1.19837236e+00
-1.69736907e-01 -1.01260602e+00 1.30387807e+00 -6.22466445e-01
-1.38807905e+00 6.66016400e-01 -1.54750064e-01 -1.32332313e+00
-5.68947256e-01 -4.47673500e-01 -3.23689245e-02 5.71212113e-01
-1.27153027e+00 -9.10557210e-01 -6.01195469e-02 6.06607616e-01
-2.15849593e-01 -6.99012205e-02 1.47200871e+00 6.72805130e-01
-3.83025616e-01 6.22969627e-01 1.42523572e-01 5.25339186e-01
1.12707305e+00 -1.18853152e+00 1.15041025e-01 2.95495596e-02
-2.00755820e-01 9.85177994e-01 3.18895876e-01 -8.87018979e-01
-9.76693332e-01 -1.35870540e+00 1.54150641e+00 -1.14156854e+00
6.39082730e-01 -4.15034778e-02 -9.89544749e-01 1.32673299e+00
2.64462754e-02 1.21736571e-01 1.61894691e+00 3.70459437e-01
-5.26568472e-01 1.09945007e-01 -1.51116061e+00 3.07110071e-01
1.16726351e+00 -6.09525502e-01 -9.17824924e-01 7.83437967e-01
5.58500469e-01 -3.55199397e-01 -1.74600208e+00 5.06583333e-01
5.63605487e-01 -2.74901479e-01 9.30435598e-01 -1.70674145e+00
2.03850329e-01 -1.82078928e-01 1.75896630e-01 -1.16553843e+00
-1.78438351e-01 -8.06340933e-01 -1.27548501e-01 1.01084781e+00
8.53670537e-01 -8.65946054e-01 5.62616408e-01 1.36186171e+00
2.85302192e-01 -5.80894113e-01 -5.02837837e-01 -3.99161458e-01
-3.50273341e-01 -3.59747969e-02 7.71487355e-01 1.89208472e+00
6.68639064e-01 1.31903365e-01 -4.57338579e-02 5.56171775e-01
4.47793126e-01 1.32632792e-01 2.80310988e-01 -1.53365326e+00
-1.98431715e-01 8.30315500e-02 -4.10014153e-01 -4.86728624e-02
-1.41920999e-01 -1.20192266e+00 -3.80470276e-01 -1.85869062e+00
3.24751437e-01 -1.04921138e+00 -6.43657267e-01 1.08240390e+00
-2.36314356e-01 -1.80471331e-01 -1.75365329e-01 3.14519942e-01
-7.74713933e-01 -3.85526747e-01 7.17042506e-01 -2.22621355e-02
-3.26731592e-01 -5.08382358e-02 -6.32101238e-01 7.13501036e-01
7.60813415e-01 -1.08543134e+00 -2.98049062e-01 -3.27734530e-01
4.95877773e-01 4.18737173e-01 -1.39076114e-02 -6.96325481e-01
5.80908835e-01 -2.58424491e-01 1.52901337e-01 -3.60354722e-01
-5.86239807e-02 -1.08245373e+00 4.92535770e-01 5.61606348e-01
-7.23872662e-01 2.62217551e-01 2.47157261e-01 5.31670392e-01
-1.23032883e-01 -6.68221787e-02 2.69602954e-01 -3.69369596e-01
-3.60343516e-01 4.97969419e-01 -5.10234714e-01 6.64323032e-01
1.17343450e+00 3.77610594e-01 -3.72155607e-01 -7.40029737e-02
-1.20557976e+00 5.49248636e-01 2.43810117e-01 3.03041756e-01
2.00494811e-01 -9.27552342e-01 -6.99078441e-01 -1.27531171e-01
5.93492568e-01 -2.26545855e-02 1.82845101e-01 9.46896732e-01
-4.89056379e-01 6.01344824e-01 6.20841328e-03 -3.53713483e-01
-1.49084640e+00 8.19797158e-01 2.82307208e-01 -6.22512579e-01
-6.19111001e-01 3.95893008e-01 -4.60555553e-02 -6.91458821e-01
2.92708993e-01 -5.22522390e-01 -3.07553381e-01 1.33428186e-01
6.13993764e-01 5.57332151e-02 2.75956631e-01 -2.06015155e-01
-7.71160066e-01 4.20169346e-02 -3.27558756e-01 2.67150491e-01
1.72001576e+00 3.25771958e-01 -3.57807904e-01 1.93856120e-01
1.03949583e+00 1.63789630e-01 -3.61323237e-01 -3.01478624e-01
5.21924496e-01 2.29516346e-02 -4.03579801e-01 -1.16134834e+00
-7.10429490e-01 3.07394922e-01 3.65026176e-01 1.83923021e-01
9.18177903e-01 1.63572401e-01 7.49444723e-01 8.02648485e-01
2.27573127e-01 -8.04135501e-01 -6.46420836e-01 2.60610729e-01
2.30596557e-01 -1.17331767e+00 1.90131087e-02 -3.94112289e-01
-7.53503203e-01 7.37100184e-01 2.34116718e-01 4.82404798e-01
1.03390133e+00 5.80838978e-01 5.93092978e-01 -5.76190472e-01
-1.28404665e+00 1.17749214e-01 8.39432925e-02 4.20644879e-01
6.89451277e-01 2.73222804e-01 -2.33549789e-01 7.02332675e-01
8.86017233e-02 6.83658361e-01 3.14090520e-01 1.28516889e+00
2.04873726e-01 -1.30948412e+00 -1.04296684e-01 1.02028477e+00
-1.09137118e+00 -4.54953283e-01 -2.45677590e-01 5.60828149e-01
2.60237455e-01 6.31849885e-01 1.73638836e-01 3.91391777e-02
4.81610417e-01 6.34818554e-01 3.11690923e-02 -9.03163731e-01
-9.85922933e-01 -3.62648338e-01 5.45575500e-01 -4.06175852e-01
-4.72323000e-01 -8.13633144e-01 -1.80850911e+00 1.60174906e-01
7.30503649e-02 6.59244597e-01 3.62501562e-01 8.29246044e-01
1.04661632e+00 4.68547076e-01 1.41226619e-01 5.24538159e-01
-3.43282640e-01 -4.46068794e-01 -3.33220989e-01 9.06551540e-01
4.31431800e-01 -3.03370714e-01 -1.89977158e-02 6.47861600e-01]
|
[8.409740447998047, 8.563154220581055]
|
e90869d8-00d8-4981-ab49-fca2014acc4c
|
explan-explaining-black-box-classifiers-using
| null | null |
https://ieeexplore.ieee.org/document/9206710
|
https://ieeexplore.ieee.org/document/9206710
|
EXPLAN: Explaining Black-box Classifiers using Adaptive Neighborhood Generation
|
Defining a representative locality is an urgent challenge in perturbation-based explanation methods, which influences the fidelity and soundness of explanations. We address this issue by proposing a robust and intuitive approach for EXPLaining black-box classifiers using Adaptive Neighborhood generation (EXPLAN). EXPLAN is a module-based algorithm consisted of dense data generation, representative data selection, data balancing, and rule-based interpretable model. It takes into account the adjacency information derived from the black-box decision function and the structure of the data for creating a representative neighborhood for the instance being explained. As a local model-agnostic explanation method, EXPLAN generates explanations in the form of logical rules that are highly interpretable and well-suited for qualitative analysis of the model's behavior. We discuss fidelity-interpretability trade-offs and demonstrate the performance of the proposed algorithm by a comprehensive comparison with state-of-the-art explanation methods LIME, LORE, and Anchor. The conducted experiments on real-world data sets show our method achieves solid empirical results in terms of fidelity, precision, and stability of explanations.
|
['Ingrid Chieh Yu', 'Peyman Rasouli']
|
2020-07-19
| null | null | null |
2020-international-joint-conference-on-neural-1
|
['explainable-models', 'explanation-fidelity-evaluation']
|
['computer-vision', 'methodology']
|
[ 3.99766937e-02 5.60753226e-01 -4.07269686e-01 -6.54498637e-01
-2.38430440e-01 -1.63659588e-01 6.80627465e-01 5.58351994e-01
3.65174919e-01 6.88435435e-01 5.21484613e-01 -5.91134250e-01
-8.28479230e-01 -7.05518305e-01 -6.21188223e-01 -5.37785590e-01
6.40200749e-02 4.82892185e-01 -1.04857482e-01 -2.03862697e-01
6.76313818e-01 3.97010356e-01 -1.82819045e+00 3.63361031e-01
1.44627595e+00 7.24304974e-01 -2.31380939e-01 3.78254175e-01
-1.87537238e-01 5.86668134e-01 -4.41973954e-01 -1.33750558e-01
7.15226773e-03 -8.23311150e-01 -7.03344107e-01 4.48273085e-02
1.11342631e-01 2.28547141e-01 -4.60798591e-02 6.98615134e-01
1.88490570e-01 1.92059904e-01 9.24239695e-01 -1.65526092e+00
-9.71539497e-01 9.88058567e-01 -2.92724460e-01 5.48619367e-02
4.04770732e-01 1.72787860e-01 8.82233799e-01 -8.33425462e-01
4.86806422e-01 1.34253323e+00 5.43387830e-01 5.99362075e-01
-1.29121709e+00 -5.13764083e-01 4.35525358e-01 5.15630662e-01
-1.14717221e+00 -2.97145814e-01 7.85788596e-01 -3.37242603e-01
7.92635143e-01 6.97126269e-01 7.17929006e-01 8.20536733e-01
5.20589113e-01 3.53328258e-01 7.98978627e-01 -5.16720355e-01
8.06183577e-01 2.96008706e-01 5.20587265e-01 8.15097749e-01
5.70778430e-01 2.29400575e-01 -9.10976052e-01 -4.16091919e-01
4.94412839e-01 3.48046347e-02 -3.51991683e-01 -5.98332345e-01
-1.07141590e+00 7.97173202e-01 8.69926989e-01 -4.77721170e-02
-3.92708361e-01 9.42050442e-02 1.92759648e-01 6.16225377e-02
1.62705481e-01 5.09655178e-01 -2.46159643e-01 1.94952741e-01
-5.15241444e-01 1.91915706e-01 6.26521766e-01 9.52207983e-01
6.68269634e-01 8.10216293e-02 -3.35745543e-01 3.62242192e-01
5.57320833e-01 1.93345934e-01 5.23948491e-01 -6.07574880e-01
2.49573559e-01 1.37292337e+00 -1.17545381e-01 -1.17722070e+00
-3.28065544e-01 -6.05028629e-01 -1.02156866e+00 3.15575212e-01
-6.57569692e-02 8.73811767e-02 -8.27842593e-01 1.69376349e+00
5.77587724e-01 9.27282199e-02 3.22190642e-01 9.59197640e-01
8.54859531e-01 4.08587247e-01 8.26347917e-02 -2.59593934e-01
9.83381212e-01 -1.02158415e+00 -6.89341366e-01 -1.60725638e-01
4.23542351e-01 -1.85918167e-01 1.16183758e+00 8.74788612e-02
-6.35823011e-01 -6.01157129e-01 -1.18572938e+00 2.09566519e-01
-3.78058851e-01 -7.78788747e-03 9.34297979e-01 3.29387993e-01
-6.13509357e-01 6.85229599e-01 -5.55994868e-01 -5.26470780e-01
1.97469115e-01 3.64746124e-01 -5.06151438e-01 2.92696327e-01
-1.02606595e+00 7.51430571e-01 4.39941227e-01 9.77468789e-02
-7.18177199e-01 -7.71961391e-01 -8.17428529e-01 4.04523313e-01
3.78627442e-02 -1.11138844e+00 7.02357471e-01 -8.92319262e-01
-1.08713067e+00 1.99585393e-01 -3.93679202e-01 -6.14575922e-01
4.09900069e-01 8.81397873e-02 -4.92205054e-01 -2.43231714e-01
9.30813774e-02 5.92270434e-01 5.48558593e-01 -1.41315103e+00
-4.54622835e-01 -3.06830347e-01 -2.37193003e-01 6.91350847e-02
-1.59609824e-01 -7.25829601e-01 -3.61903273e-02 -3.92807394e-01
4.81757373e-01 -6.40603244e-01 -3.86011869e-01 4.24059294e-02
-8.05006266e-01 -1.32166594e-01 8.22870553e-01 -1.84608370e-01
1.62478900e+00 -2.08535480e+00 6.49436936e-02 6.40440643e-01
3.24050158e-01 -1.02047995e-01 3.58071662e-02 5.14522910e-01
-3.37704718e-01 4.57714349e-01 -3.51777852e-01 5.24730720e-02
-4.55812784e-04 1.44938290e-01 -3.33521932e-01 2.68352866e-01
2.22626850e-01 6.19226813e-01 -7.98439741e-01 -5.05352497e-01
3.39401603e-01 2.98363805e-01 -5.73596716e-01 3.46638501e-01
-2.37139761e-01 4.23835307e-01 -5.10247052e-01 5.54285884e-01
3.83049965e-01 -2.64549375e-01 1.08562894e-01 -1.81282908e-01
-2.31637601e-02 2.23905444e-01 -1.35108924e+00 1.26658249e+00
-3.52182388e-01 5.61439812e-01 -6.67282879e-01 -5.59418678e-01
1.34592056e+00 7.45752081e-02 -9.29120034e-02 -3.17322463e-01
4.03596945e-02 1.28593743e-01 1.91546753e-01 -5.93172669e-01
1.76633790e-01 9.76219699e-02 6.92514926e-02 5.05233705e-01
-4.85135734e-01 3.85425687e-02 1.66654468e-01 2.51938522e-01
1.24394238e+00 -9.82716680e-02 8.44806254e-01 -5.17466605e-01
5.57666421e-01 3.82043362e-01 6.07872248e-01 6.92024529e-01
1.31819576e-01 6.10437930e-01 5.37967980e-01 -7.30713248e-01
-8.80207300e-01 -8.89726877e-01 -2.14624420e-01 2.60055512e-01
4.25376475e-01 -5.55618048e-01 -7.36575484e-01 -7.59519458e-01
2.63688326e-01 1.33644366e+00 -1.15476608e+00 -6.18045628e-01
-2.28730328e-02 -3.64895761e-01 1.26581490e-01 5.26073694e-01
6.18413925e-01 -1.00179458e+00 -6.37931108e-01 1.23254983e-02
-3.53991706e-03 -4.72241670e-01 -1.69304416e-01 3.26274782e-01
-1.05843699e+00 -1.21670878e+00 2.85637289e-01 -1.40155330e-01
1.26543856e+00 1.98012590e-01 1.00621378e+00 5.29551744e-01
-1.08625598e-01 -9.19273570e-02 -2.80867726e-01 -3.68300170e-01
-5.13860464e-01 8.61311704e-03 5.53520136e-02 1.07635735e-02
2.70683527e-01 -6.84580505e-01 -6.90151989e-01 4.81544524e-01
-8.02847624e-01 2.93033630e-01 7.93891549e-01 9.61134255e-01
6.90315545e-01 3.58243547e-02 6.40179574e-01 -1.11565804e+00
1.01504052e+00 -6.78763449e-01 -3.36039841e-01 5.50261915e-01
-1.25205564e+00 6.51038349e-01 9.14547563e-01 -3.67217183e-01
-1.00537491e+00 1.28411412e-01 4.76139635e-01 -3.75474870e-01
-2.49468610e-01 4.36258703e-01 -2.13350877e-01 1.39888376e-01
1.22774971e+00 4.84101549e-02 -2.85278820e-02 -1.86801746e-01
6.46470070e-01 6.78363979e-01 6.38049543e-01 -3.57313812e-01
8.25364649e-01 3.71599764e-01 1.86581075e-01 -3.29852194e-01
-5.30951381e-01 -1.44887373e-01 -6.40587211e-01 -2.14856327e-01
5.30845821e-01 -3.72791797e-01 -8.09812665e-01 -4.69815612e-01
-1.08622062e+00 1.81258410e-01 -5.16390085e-01 2.86106229e-01
-4.77230251e-01 -2.22938019e-03 1.71864659e-01 -8.57378662e-01
-4.41698104e-01 -1.06080067e+00 6.74165487e-01 3.36697638e-01
-8.97488117e-01 -1.04157495e+00 2.69391872e-02 2.05409214e-01
2.48302743e-01 6.39469147e-01 1.47924376e+00 -8.60741377e-01
-6.17365718e-01 -2.12620318e-01 -3.95620018e-02 -1.94220558e-01
1.20809637e-01 2.94540435e-01 -8.57168615e-01 2.69482553e-01
-3.00351948e-01 1.99574023e-01 6.81733847e-01 2.13105962e-01
1.07392085e+00 -6.75601900e-01 -6.34420335e-01 4.87044185e-01
1.29590178e+00 -9.99982562e-03 3.96722883e-01 3.15930933e-01
5.70429087e-01 5.06706834e-01 8.79337907e-01 5.48090458e-01
1.59661144e-01 4.22091991e-01 7.92936563e-01 -1.61230981e-01
-1.59339309e-01 -6.29392087e-01 2.53140107e-02 3.41455042e-01
2.15494543e-01 -3.77170295e-01 -9.99713659e-01 5.66217721e-01
-2.51005816e+00 -9.81306851e-01 -4.70245153e-01 2.13755536e+00
2.92965680e-01 8.74002874e-02 -1.67538747e-01 3.92340660e-01
7.16846406e-01 -2.46844396e-01 -8.74266028e-01 -8.58580351e-01
7.61096925e-02 -4.99359369e-01 5.39554358e-02 6.49511814e-01
-4.42498147e-01 6.22079790e-01 6.62559795e+00 4.32582885e-01
-8.14001381e-01 -3.02434981e-01 6.50390565e-01 1.10904396e-01
-9.13346112e-01 2.99107313e-01 -5.27686417e-01 2.30659246e-01
8.49867821e-01 -4.55340773e-01 3.14304650e-01 1.13092685e+00
6.74189866e-01 -9.62705761e-02 -1.48645043e+00 5.74323177e-01
-5.04180081e-02 -1.60728812e+00 5.86695194e-01 -6.00420982e-02
8.07899356e-01 -6.58731520e-01 -2.79006828e-02 -1.41103253e-01
1.94659919e-01 -1.19018555e+00 8.83264124e-01 7.97128201e-01
3.82716268e-01 -7.99255908e-01 8.37004781e-01 4.31419730e-01
-1.08090854e+00 -3.94221932e-01 -4.25112456e-01 -2.28076339e-01
-1.70660883e-01 8.55673850e-01 -9.40581143e-01 7.07115948e-01
5.13573110e-01 5.77080488e-01 -6.97576225e-01 8.33428621e-01
-6.31861448e-01 5.25864005e-01 -1.03323504e-01 -2.90434837e-01
-4.94353613e-03 -6.55135214e-02 5.06970823e-01 9.10230935e-01
2.37060726e-01 2.58892924e-01 -2.29733884e-01 1.30910122e+00
1.69686437e-01 1.26304165e-01 -7.73531854e-01 2.34820813e-01
8.48204792e-01 1.06589329e+00 -5.24205744e-01 -2.77793646e-01
1.67912006e-01 5.18957138e-01 4.33669239e-01 1.74217641e-01
-9.18143213e-01 -6.62979856e-02 5.44492304e-01 2.69817233e-01
-1.61928549e-01 2.93919563e-01 -1.04967880e+00 -7.70808935e-01
3.19803171e-02 -1.06532896e+00 4.60167080e-01 -9.02696729e-01
-1.15507209e+00 9.07296181e-01 2.04052582e-01 -1.40160227e+00
-4.38247263e-01 -1.88840166e-01 -1.16770947e+00 7.53938854e-01
-9.00140882e-01 -9.01898682e-01 -8.32619429e-01 2.92482972e-01
4.68516618e-01 -2.82884181e-01 8.87554526e-01 -3.86572778e-01
-7.75101900e-01 4.88812864e-01 -2.41356492e-02 -5.29074430e-01
1.73690453e-01 -1.16810870e+00 4.27258134e-01 9.33106303e-01
1.73532322e-01 9.42151606e-01 1.07106304e+00 -6.93365514e-01
-1.18294859e+00 -9.99880075e-01 9.13574457e-01 -4.00759429e-01
2.85161018e-01 -1.82037845e-01 -9.17197883e-01 4.50372458e-01
6.07932620e-02 -1.55283436e-01 6.60160720e-01 2.58264422e-01
-2.70422667e-01 -3.19348067e-01 -1.32108808e+00 9.20298159e-01
1.10921645e+00 6.11760132e-02 -5.98478615e-01 2.14249417e-01
5.65408707e-01 -1.52641103e-01 -4.27013457e-01 3.98455054e-01
4.05679524e-01 -1.45725286e+00 5.54674804e-01 -6.64436400e-01
4.89336759e-01 -6.34779990e-01 1.03937447e-01 -1.38127279e+00
-5.90377867e-01 -5.80893457e-01 -8.38029087e-02 1.23376906e+00
8.71836483e-01 -6.55398905e-01 5.98840654e-01 9.44130063e-01
-1.26921952e-01 -1.10310125e+00 -6.34829879e-01 -4.88922298e-01
-4.73657280e-01 -1.53381005e-01 1.22875416e+00 7.66749918e-01
5.21599293e-01 4.54901278e-01 6.23130612e-02 4.64623064e-01
5.95446706e-01 3.35115016e-01 1.01555240e+00 -1.20334125e+00
-1.13047220e-01 -3.67776662e-01 -5.67325532e-01 -4.79821682e-01
2.01292798e-01 -7.74988532e-01 -1.80269033e-02 -1.82954919e+00
2.94346571e-01 -2.79400051e-01 -6.83143511e-02 6.13450706e-01
-3.38720858e-01 -4.31445777e-01 -6.07222468e-02 3.77295345e-01
-3.55376124e-01 7.09134579e-01 9.96266127e-01 -4.14800234e-02
-3.90690118e-01 -9.35852453e-02 -9.64809895e-01 6.16938651e-01
8.56610298e-01 -5.97404003e-01 -8.57240140e-01 -9.70861763e-02
-4.37370548e-03 -8.17246959e-02 4.04566586e-01 -1.05902123e+00
3.69998455e-01 -5.03046095e-01 3.93482774e-01 -3.21236432e-01
-1.67201366e-02 -9.90293205e-01 6.14386559e-01 6.67908669e-01
-8.31394136e-01 3.52662146e-01 1.60938144e-01 8.25604856e-01
-1.82409868e-01 -5.91604188e-02 6.60866857e-01 2.70405143e-01
-5.85431755e-01 3.37676378e-03 5.01409806e-02 -2.63103217e-01
1.17640591e+00 -4.79120046e-01 -5.94462335e-01 -4.61685866e-01
-4.89285678e-01 3.06532949e-01 3.90324622e-01 3.74718308e-01
8.49962771e-01 -1.38091612e+00 -6.46426976e-01 4.37116504e-01
4.09736812e-01 -5.89009076e-02 5.35521656e-02 5.28594971e-01
-3.67033929e-01 3.56766969e-01 -1.74003005e-01 -6.45448446e-01
-1.13455307e+00 5.57526290e-01 3.91026407e-01 -1.48858657e-04
-5.90647042e-01 3.62112463e-01 7.09757954e-02 -4.28634226e-01
1.21020757e-01 -4.57981616e-01 -1.55617818e-01 -5.26194453e-01
2.89098769e-01 4.99535650e-01 -6.27440289e-02 -1.45062730e-01
-5.89439929e-01 3.55758607e-01 2.88796932e-01 3.96745056e-02
1.10173869e+00 -5.30468412e-02 -8.12240839e-02 5.63736498e-01
5.24616003e-01 -1.70807615e-01 -1.09463573e+00 2.39252761e-01
-6.59696385e-02 -4.58224565e-01 -1.12855636e-01 -1.05633986e+00
-6.83496296e-01 6.68467402e-01 4.41881597e-01 3.72260302e-01
1.01668060e+00 -6.97130114e-02 3.79013158e-02 2.66644120e-01
-1.69054084e-02 -6.90340936e-01 -2.76946388e-02 -7.57834986e-02
1.26274180e+00 -1.23704326e+00 1.72440588e-01 -5.66397071e-01
-6.45975173e-01 1.14562333e+00 9.09818411e-01 1.19343206e-01
5.55189013e-01 -4.33585122e-02 1.75464407e-01 -3.86379093e-01
-1.22996747e+00 2.07875803e-01 6.55857444e-01 5.99279821e-01
2.46504694e-01 5.34163751e-02 -3.46705586e-01 7.54971623e-01
-4.09015387e-01 -2.89784253e-01 2.44705334e-01 4.81189400e-01
-5.57765961e-01 -6.54049516e-01 -5.09776413e-01 3.36149186e-01
4.80376571e-01 6.89698085e-02 -8.66240501e-01 1.08833814e+00
1.27170905e-01 1.25863016e+00 -8.95810947e-02 -5.36495626e-01
4.28880900e-01 1.63756832e-02 -1.86808780e-01 -5.36325693e-01
-7.34064341e-01 -6.03393614e-01 -1.54548921e-02 -5.15534759e-01
-8.75226129e-03 -3.36880535e-01 -1.77747953e+00 -7.05574512e-01
-5.51833630e-01 6.23920858e-01 7.41677940e-01 1.15142226e+00
8.70696485e-01 8.25805664e-01 6.84729874e-01 -2.80580223e-01
-4.50397551e-01 -7.79959321e-01 -2.16905639e-01 4.97080266e-01
1.65997446e-01 -6.46563828e-01 -7.62130678e-01 -1.59590498e-01]
|
[8.76427173614502, 5.6784796714782715]
|
6a299f9a-15ba-46c6-a0f0-3085ef747cdc
|
fusing-motion-patterns-and-key-visual
|
2007.06288
| null |
https://arxiv.org/abs/2007.06288v1
|
https://arxiv.org/pdf/2007.06288v1.pdf
|
Fusing Motion Patterns and Key Visual Information for Semantic Event Recognition in Basketball Videos
|
Many semantic events in team sport activities e.g. basketball often involve both group activities and the outcome (score or not). Motion patterns can be an effective means to identify different activities. Global and local motions have their respective emphasis on different activities, which are difficult to capture from the optical flow due to the mixture of global and local motions. Hence it calls for a more effective way to separate the global and local motions. When it comes to the specific case for basketball game analysis, the successful score for each round can be reliably detected by the appearance variation around the basket. Based on the observations, we propose a scheme to fuse global and local motion patterns (MPs) and key visual information (KVI) for semantic event recognition in basketball videos. Firstly, an algorithm is proposed to estimate the global motions from the mixed motions based on the intrinsic property of camera adjustments. And the local motions could be obtained from the mixed and global motions. Secondly, a two-stream 3D CNN framework is utilized for group activity recognition over the separated global and local motion patterns. Thirdly, the basket is detected and its appearance features are extracted through a CNN structure. The features are utilized to predict the success or failure. Finally, the group activity recognition and success/failure prediction results are integrated using the kronecker product for event recognition. Experiments on NCAA dataset demonstrate that the proposed method obtains state-of-the-art performance.
|
['Qi. Wang', 'Junchi Yan', 'Lifang Wu', 'Zhou Yang', 'Chang Wen Chen', 'Boxuan Zhao', 'Meng Jian']
|
2020-07-13
| null | null | null | null |
['group-activity-recognition']
|
['computer-vision']
|
[-1.46201625e-01 -9.10232604e-01 -3.46746266e-01 -6.84438646e-02
-4.19412106e-01 -4.50420618e-01 4.28507298e-01 2.76019778e-02
-4.60807413e-01 3.69911373e-01 2.97654837e-01 4.12215084e-01
-3.33907336e-01 -7.75957584e-01 -3.92920375e-01 -1.03126729e+00
-1.65144980e-01 -9.25834775e-02 3.85586739e-01 2.90546869e-03
3.60438824e-01 5.59504867e-01 -1.74850810e+00 3.57862741e-01
3.34367335e-01 1.28701150e+00 3.07449341e-01 8.45978737e-01
-6.35640323e-02 1.27837598e+00 -6.96855068e-01 3.81947681e-02
3.05854559e-01 -6.53476357e-01 -4.86365139e-01 3.24969500e-01
8.61717388e-02 -3.30137342e-01 -5.87430894e-01 9.02075350e-01
5.01887977e-01 4.64081645e-01 4.57292378e-01 -1.19508219e+00
6.47512004e-02 3.43948416e-02 -4.62651759e-01 7.80904353e-01
4.68617469e-01 1.61679879e-01 8.15893233e-01 -8.79393339e-01
6.47782922e-01 9.33194160e-01 3.31249356e-01 -1.52789745e-02
-7.80292034e-01 -6.47503495e-01 9.70174000e-02 8.82732272e-01
-1.31528568e+00 -2.03621998e-01 1.01308787e+00 -5.53023577e-01
6.07234418e-01 1.32028341e-01 1.23530746e+00 9.74919379e-01
3.77865553e-01 1.17390883e+00 6.62382185e-01 -8.59130546e-02
1.09445445e-01 -4.79876250e-01 3.65672819e-02 6.01081848e-01
1.56432148e-02 7.30437338e-02 -9.50741649e-01 2.46396020e-01
9.57815826e-01 2.67325759e-01 -3.14055622e-01 -2.49190077e-01
-1.42946255e+00 5.50704181e-01 1.29947677e-01 2.48486787e-01
-5.65115511e-01 -1.72998197e-02 5.66517234e-01 9.63770822e-02
2.19757855e-01 1.09832305e-02 -1.06001221e-01 -6.65143967e-01
-9.17179108e-01 3.60097736e-01 5.82661152e-01 5.01281023e-01
6.55152321e-01 8.28014463e-02 -2.73748845e-01 7.71477699e-01
2.48222515e-01 3.84219795e-01 4.85994637e-01 -8.71441603e-01
5.87829590e-01 6.70491397e-01 2.00686127e-01 -1.59692574e+00
-4.49724555e-01 -3.74578983e-01 -6.83539867e-01 1.52639046e-01
7.99073815e-01 3.56100202e-02 -5.95703840e-01 1.30211854e+00
3.96801919e-01 5.95039785e-01 -4.10374533e-03 1.22697592e+00
9.19141591e-01 7.82103896e-01 -6.25164658e-02 -3.61503512e-01
1.41715145e+00 -7.68372715e-01 -9.49971855e-01 -1.64005384e-01
4.30886775e-01 -8.69132102e-01 4.09046382e-01 5.01204312e-01
-1.14551628e+00 -9.23359156e-01 -1.10233188e+00 3.10206294e-01
3.79743464e-02 4.42655861e-01 3.33665907e-01 2.91849077e-01
-3.64216328e-01 4.75877374e-01 -1.22982752e+00 -1.68416470e-01
2.04982236e-03 1.34333715e-01 -3.85629028e-01 1.66151747e-02
-1.00727773e+00 5.97873032e-01 5.99399447e-01 5.89426279e-01
-8.40692639e-01 -2.62355387e-01 -9.65994000e-01 -1.25444666e-01
3.19219291e-01 -1.53692633e-01 8.65685105e-01 -9.54061627e-01
-1.38115919e+00 4.59881574e-01 -1.91056505e-01 -1.78420275e-01
5.62481284e-01 -1.50146171e-01 -5.33025384e-01 5.33071101e-01
2.89192140e-01 7.73825571e-02 5.19269764e-01 -7.18707621e-01
-1.11022592e+00 -1.58255547e-01 -1.28916398e-01 6.19552195e-01
-1.42728344e-01 2.18779966e-01 -6.90566719e-01 -7.88529217e-01
5.03402352e-01 -7.42607594e-01 1.21257983e-01 -1.46376580e-01
-2.90660169e-02 -7.07283542e-02 8.65917563e-01 -8.52591097e-01
1.25712776e+00 -2.27156472e+00 2.74251074e-01 2.70430624e-01
-3.18801613e-03 2.27781042e-01 1.34685054e-01 2.76032448e-01
-4.33461666e-02 -5.10421693e-01 2.30110720e-01 7.72270411e-02
-3.97587061e-01 2.36045673e-01 1.05801582e-01 6.37630463e-01
1.02781661e-01 6.48889184e-01 -1.02405572e+00 -7.23306596e-01
5.81547379e-01 2.57161349e-01 -1.89994738e-01 2.40104705e-01
5.26080608e-01 7.82854855e-01 -6.60127640e-01 6.50814414e-01
5.01085877e-01 1.64928421e-01 7.06192106e-02 -3.66829485e-01
-2.68388540e-01 7.67789632e-02 -1.84774852e+00 1.64017069e+00
-9.69407782e-02 5.97969413e-01 -8.47014338e-02 -1.42504632e+00
1.01735413e+00 5.09049058e-01 9.46731627e-01 -7.24336565e-01
2.34886572e-01 1.59538329e-01 3.28023061e-02 -7.93274462e-01
4.21397120e-01 -5.15098572e-02 -4.91413437e-02 1.40065849e-01
2.03750655e-01 4.91516858e-01 4.85234708e-01 -1.49903521e-01
1.04512298e+00 3.06388974e-01 2.86011696e-01 5.08654490e-02
5.92518091e-01 4.35078554e-02 1.13905632e+00 6.67574644e-01
-4.95595336e-01 6.87229693e-01 2.75940567e-01 -8.08225572e-01
-6.05913997e-01 -1.11273563e+00 3.04231644e-01 6.54764771e-01
7.86567688e-01 -3.03419709e-01 -3.65259290e-01 -4.83796656e-01
-4.14643407e-01 -1.36039406e-01 -3.60309064e-01 -3.66797894e-01
-9.22497511e-01 -8.17655563e-01 3.63297522e-01 8.69485319e-01
9.57149386e-01 -1.26558065e+00 -6.72008395e-01 6.07324064e-01
-6.83036327e-01 -1.20883727e+00 -5.21657050e-01 -1.56704381e-01
-8.63849163e-01 -1.23357630e+00 -6.56734049e-01 -8.07681203e-01
1.98294461e-01 5.12606740e-01 6.55343235e-01 -5.41203953e-02
-3.65681320e-01 3.62015128e-01 -4.79525566e-01 -7.25812465e-02
6.91788644e-02 -4.38361257e-01 1.18422858e-01 7.09407628e-01
4.57507908e-01 -4.33994383e-01 -9.55809414e-01 5.71441114e-01
-7.31717706e-01 -4.08256277e-02 4.45252091e-01 7.59649992e-01
5.53018034e-01 1.62094370e-01 1.92790061e-01 1.22625887e-01
2.36640722e-01 -2.81236321e-01 -2.22618133e-01 1.31951779e-01
1.96158007e-01 -3.93868417e-01 2.84846991e-01 -6.40430987e-01
-1.23403203e+00 1.38868734e-01 1.34747416e-01 -5.49181044e-01
-2.94057339e-01 5.22452533e-01 -2.03815430e-01 1.66597977e-01
3.03836405e-01 4.16798741e-01 -1.21104695e-01 -3.38140726e-01
-1.38249779e-02 5.29854774e-01 5.82285702e-01 -4.55789208e-01
4.23931360e-01 7.37075984e-01 4.41873670e-02 -1.00114393e+00
-6.95494235e-01 -1.05926049e+00 -6.44137442e-01 -9.24193919e-01
1.45005524e+00 -1.10699975e+00 -8.60871255e-01 9.51576710e-01
-1.04595876e+00 1.37106791e-01 -2.96445787e-01 1.30232239e+00
-4.97661918e-01 6.16385400e-01 -8.72463465e-01 -8.13088953e-01
4.47912738e-02 -1.28858149e+00 8.64436984e-01 5.28108358e-01
-1.28223807e-01 -1.00132930e+00 5.49696609e-02 5.00663996e-01
-2.57817626e-01 3.35501224e-01 2.10193038e-01 -2.81748921e-01
-6.79037869e-01 -4.19083744e-01 -1.83416549e-02 4.01554227e-01
2.81276882e-01 6.16949387e-02 -5.44396758e-01 -3.39017287e-02
1.43712983e-01 7.96506777e-02 6.65653765e-01 8.03899109e-01
7.75458395e-01 1.58591121e-01 -1.34306476e-01 6.34857655e-01
1.15708911e+00 5.23131371e-01 7.18758643e-01 5.86612046e-01
9.06899214e-01 4.29312646e-01 1.10377395e+00 6.41863644e-01
4.63958718e-02 9.19508159e-01 2.22762510e-01 1.13997117e-01
-1.39771000e-01 -9.54195336e-02 6.91649973e-01 1.07318604e+00
-7.45989740e-01 1.61292609e-02 -5.87012768e-01 6.80370688e-01
-2.23211408e+00 -1.51427805e+00 -3.64365667e-01 2.20704794e+00
3.17513555e-01 1.02486379e-01 1.86629638e-01 4.31787878e-01
1.08693087e+00 2.82491088e-01 -1.63969964e-01 -4.75089327e-02
-1.74452230e-01 1.69797465e-01 4.94796395e-01 1.74621008e-02
-1.32180095e+00 5.14169991e-01 5.17799282e+00 1.21159005e+00
-1.07646275e+00 1.74591446e-03 2.86217570e-01 -2.92609125e-01
4.28860396e-01 -8.03776667e-04 -6.84051275e-01 4.92650658e-01
3.80600035e-01 2.71532208e-01 1.12394895e-02 4.97254640e-01
6.00302219e-01 -3.53863180e-01 -5.63470006e-01 1.19875479e+00
5.34710772e-02 -1.15799022e+00 -2.35190660e-01 -6.15461543e-02
4.35620010e-01 -4.00470883e-01 -3.67021739e-01 9.59073901e-02
-1.03842959e-01 -5.37067294e-01 1.02395582e+00 8.15191507e-01
2.36720890e-01 -8.01683247e-01 7.23015249e-01 5.11985958e-01
-1.85038257e+00 -3.25372279e-01 -2.20431164e-01 -4.32358027e-01
5.08578002e-01 3.72232884e-01 -2.63313651e-01 8.39636326e-01
9.61127579e-01 1.26534009e+00 -2.00402975e-01 1.26667249e+00
-1.66633055e-01 7.03172684e-01 -2.29509458e-01 3.94783206e-02
1.40222386e-01 -5.77337027e-01 9.03240383e-01 1.06341147e+00
5.44086933e-01 1.10260114e-01 5.08717179e-01 4.88458812e-01
5.57377934e-01 1.16424508e-01 -3.65894616e-01 1.02072433e-01
7.71608800e-02 1.31311309e+00 -9.92401838e-01 -3.58304709e-01
-6.06461883e-01 8.76090407e-01 -5.09097837e-02 3.86203170e-01
-8.33319008e-01 -3.58431309e-01 6.91745520e-01 -1.19971909e-01
3.20414245e-01 -3.74656975e-01 6.61102869e-03 -1.52521753e+00
3.21200877e-01 -5.47432959e-01 6.35181069e-01 -7.93909848e-01
-1.04870474e+00 1.21996619e-01 3.18307579e-02 -1.93534291e+00
-6.12295866e-02 -4.82977957e-01 -7.81344533e-01 5.59217751e-01
-1.06453586e+00 -7.87854612e-01 -5.79449594e-01 7.42283046e-01
8.10387611e-01 -1.18699320e-01 3.24582964e-01 5.21174729e-01
-6.64876997e-01 1.46730796e-01 -1.08737528e-01 6.19743049e-01
5.39413095e-01 -9.69706297e-01 -3.34991634e-01 1.18020916e+00
2.22895056e-01 6.17797934e-02 5.14955819e-01 -6.94959939e-01
-1.19729185e+00 -8.43271196e-01 6.66331768e-01 -1.77468300e-01
5.20153999e-01 3.74953300e-02 -6.08654499e-01 4.67400044e-01
-2.07386374e-01 1.83906212e-01 4.29388791e-01 -2.93097794e-01
2.69970477e-01 -3.04530621e-01 -5.55372179e-01 4.17617083e-01
9.92763340e-01 -3.52169067e-01 -7.20662892e-01 -6.71975538e-02
-1.40840188e-02 -5.44251919e-01 -8.86610210e-01 4.38546270e-01
6.52745426e-01 -1.00968158e+00 1.13214171e+00 -3.15180421e-01
4.31139112e-01 -7.18363702e-01 -6.19885325e-02 -9.30063188e-01
-3.54929298e-01 -2.26182088e-01 -9.27533209e-02 1.07074928e+00
-2.87262380e-01 -1.50686160e-01 7.97395170e-01 1.51027977e-01
-1.29028335e-01 -3.46306652e-01 -1.03012228e+00 -7.45992422e-01
-6.49015963e-01 -6.99352741e-01 1.85239926e-01 9.32946444e-01
3.04593612e-02 5.65273240e-02 -6.51552975e-01 2.32832327e-01
4.07187849e-01 1.33037850e-01 7.19025493e-01 -1.06430399e+00
-3.55106711e-01 -3.27660322e-01 -1.14005435e+00 -1.20830834e+00
-2.37203479e-01 -6.60102427e-01 1.02512620e-01 -1.49253666e+00
2.87306547e-01 4.17798497e-02 -4.95065451e-01 1.35206744e-01
-2.14934304e-01 4.68275875e-01 3.64750624e-01 3.89708906e-01
-6.25852883e-01 4.49791551e-01 1.44460833e+00 -1.28903925e-01
-4.52051222e-01 1.61207020e-01 5.64266257e-02 7.70229757e-01
5.50837934e-01 -2.38800913e-01 -1.62969410e-01 -6.09276444e-02
2.25691959e-01 5.82357347e-01 5.29463708e-01 -1.33274531e+00
2.74247915e-01 -3.14331144e-01 4.58260864e-01 -6.43985331e-01
4.92886126e-01 -5.25144458e-01 1.40854612e-01 4.05878574e-01
-6.53885454e-02 2.60425005e-02 -1.40532240e-01 8.12904596e-01
-8.32026422e-01 -1.70728296e-01 4.99082983e-01 -2.48637483e-01
-1.05129063e+00 3.85058820e-01 -7.85253584e-01 -1.36215240e-01
1.13508451e+00 -7.55517125e-01 5.19186035e-02 -4.12299246e-01
-1.21859729e+00 1.28967643e-01 -1.16493583e-01 5.79706609e-01
8.04164767e-01 -1.66497111e+00 -6.42566800e-01 1.54669315e-01
1.91306576e-01 -1.54472619e-01 6.78077519e-01 1.29776180e+00
-8.27492177e-01 -1.39412796e-02 -4.34634119e-01 -8.99266779e-01
-1.43461740e+00 2.08429858e-01 3.93079549e-01 -2.96978325e-01
-7.30204880e-01 4.94631588e-01 3.22854698e-01 1.08389534e-01
1.15455333e-02 -3.08444500e-01 -7.15095520e-01 3.19044292e-01
7.11060286e-01 6.66277409e-01 -1.76936071e-02 -1.11608303e+00
-1.73128217e-01 8.36778164e-01 3.44418198e-01 -8.31348896e-02
1.13949430e+00 -3.06852013e-01 1.38911426e-01 5.84014595e-01
1.11757696e+00 -1.19460560e-01 -1.37897789e+00 -1.53533667e-01
-2.49401480e-01 -7.15730786e-01 1.37040094e-02 -1.07528478e-01
-1.39513886e+00 1.13895023e+00 7.45330334e-01 1.19556390e-01
1.23285317e+00 -1.39948145e-01 8.83808732e-01 -6.49048835e-02
2.74835020e-01 -1.33562446e+00 3.79390955e-01 3.56576413e-01
3.87703985e-01 -8.96543503e-01 -1.63744599e-01 -3.57567519e-01
-8.28442872e-01 1.39789236e+00 3.72085989e-01 -3.18460375e-01
6.29839599e-01 -1.23916365e-01 1.33889243e-01 -2.65653133e-01
-3.12624931e-01 -3.67834538e-01 5.04126012e-01 4.21222955e-01
5.81978671e-02 3.21896374e-02 -3.99768561e-01 5.96812189e-01
2.06474155e-01 -4.56489362e-02 5.32477200e-01 1.08460891e+00
-5.05929470e-01 -8.54912281e-01 -6.61769092e-01 1.68405056e-01
-6.60858274e-01 4.46492344e-01 1.22177996e-01 7.36461520e-01
5.39099991e-01 1.18620026e+00 2.80736715e-01 -4.66491908e-01
4.76706803e-01 -1.56765077e-02 5.47996223e-01 -3.29963148e-01
-4.79020268e-01 4.65242654e-01 4.61004302e-02 -8.45407009e-01
-8.84533942e-01 -9.01148081e-01 -1.11931646e+00 -3.53064537e-02
-3.04808140e-01 2.45793127e-02 2.75987417e-01 1.17900586e+00
-6.21862300e-02 4.84199166e-01 5.07217228e-01 -9.85616922e-01
2.48443689e-02 -6.95130110e-01 -1.03115666e+00 8.54438543e-01
1.24570973e-01 -9.78997350e-01 -3.61453533e-01 2.56712854e-01]
|
[8.17525577545166, 0.4657060503959656]
|
ce41ab97-ba15-4d58-ae07-89cccdc557d7
|
automated-fake-news-detection-using-cross
|
2201.00083
| null |
https://arxiv.org/abs/2201.00083v1
|
https://arxiv.org/pdf/2201.00083v1.pdf
|
Automated Fake News Detection using cross-checking with reliable sources
|
Over the past decade, fake news and misinformation have turned into a major problem that has impacted different aspects of our lives, including politics and public health. Inspired by natural human behavior, we present an approach that automates the detection of fake news. Natural human behavior is to cross-check new information with reliable sources. We use Natural Language Processing (NLP) and build a machine learning (ML) model that automates the process of cross-checking new information with a set of predefined reliable sources. We implement this for Twitter and build a model that flags fake tweets. Specifically, for a given tweet, we use its text to find relevant news from reliable news agencies. We then train a Random Forest model that checks if the textual content of the tweet is aligned with the trusted news. If it is not, the tweet is classified as fake. This approach can be generally applied to any kind of information and is not limited to a specific news story or a category of information. Our implementation of this approach gives a $70\%$ accuracy which outperforms other generic fake-news classification models. These results pave the way towards a more sensible and natural approach to fake news detection.
|
['Sadegh Raeisi', 'Fakhteh Ghanbarnejad', 'Milad Ranjbar', 'Zahra Ghadiri']
|
2022-01-01
| null | null | null | null |
['news-classification']
|
['natural-language-processing']
|
[-6.79740310e-02 3.85668397e-01 -5.34451485e-01 -3.48232001e-01
-7.53283620e-01 -6.22599661e-01 1.17656016e+00 9.23475266e-01
-3.74253511e-01 8.76538634e-01 2.07801253e-01 -5.47442079e-01
5.83167732e-01 -1.25911856e+00 -8.33481848e-01 -2.76900202e-01
3.39008600e-01 5.75071275e-01 7.08072662e-01 -5.06020784e-01
4.19262975e-01 4.58683997e-01 -1.12970841e+00 8.27806175e-01
5.68871498e-01 6.51290834e-01 -6.65132523e-01 4.08787668e-01
-2.55747259e-01 1.20330858e+00 -9.06437278e-01 -7.69057274e-01
5.89775387e-03 -5.44320345e-01 -1.01665199e+00 -1.78660601e-01
2.35997543e-01 -2.19027717e-02 -1.10859852e-04 1.28605342e+00
-6.11019135e-02 -4.90832835e-01 4.35848743e-01 -1.18179584e+00
-9.81562659e-02 7.45520353e-01 -3.45554709e-01 3.21914285e-01
5.30918658e-01 -1.33010104e-01 5.47848761e-01 -6.53924286e-01
1.02216649e+00 1.22084475e+00 7.73870707e-01 2.32820138e-01
-9.52045083e-01 -7.19607294e-01 -3.60039085e-01 -5.86951822e-02
-9.16682363e-01 -2.96177179e-01 4.19612586e-01 -6.92153633e-01
3.40471655e-01 4.04790431e-01 6.64184391e-01 1.32698667e+00
5.85356295e-01 5.46388328e-01 1.60773623e+00 -5.31592309e-01
2.04780504e-01 9.86694694e-01 4.03255492e-01 6.74303532e-01
5.65262794e-01 7.30825588e-02 -5.60238242e-01 -7.44908452e-01
-8.96239206e-02 -1.15936048e-01 -5.62575795e-02 2.59680212e-01
-1.17782116e+00 1.27138388e+00 4.38848495e-01 7.26608455e-01
-3.06703657e-01 -1.22509234e-01 6.37107015e-01 6.62464499e-01
9.33111548e-01 7.26117015e-01 -1.99919030e-01 1.21329494e-01
-1.11915779e+00 3.60112876e-01 1.22553837e+00 4.24844146e-01
6.07913136e-01 -4.12472397e-01 7.50803575e-03 4.41460162e-01
2.60409802e-01 5.29421985e-01 6.13570213e-01 -2.50465035e-01
3.59506577e-01 7.44784176e-01 5.41392922e-01 -1.86121583e+00
-3.18252593e-01 -3.85390282e-01 -5.94204962e-01 -5.44800051e-02
4.99344558e-01 4.46721166e-02 -4.61026996e-01 9.77851391e-01
6.74619436e-01 -6.46663979e-02 -7.36751035e-02 7.95813560e-01
7.96437979e-01 8.41461480e-01 -1.17911875e-01 -1.99205682e-01
1.40163195e+00 -6.76601171e-01 -8.34335208e-01 -1.83463350e-01
6.62601829e-01 -1.01293027e+00 5.96988022e-01 4.48727459e-01
-4.71352637e-01 4.31847060e-03 -9.56098139e-01 1.07017472e-01
-1.08336711e+00 -2.28081673e-01 3.04516733e-01 8.30872715e-01
-5.41036725e-01 7.26965725e-01 -4.60735679e-01 -4.02506918e-01
2.57662445e-01 -1.10428736e-01 -4.20422524e-01 2.39528175e-02
-1.57299197e+00 1.18707991e+00 4.86788332e-01 -4.18005943e-01
-5.99334836e-01 2.66799498e-02 -5.72836459e-01 -3.55892658e-01
4.65697259e-01 -2.83805519e-01 1.14701462e+00 -1.23772871e+00
-1.15892601e+00 1.14954984e+00 -2.82558762e-02 -8.11284482e-01
9.79584754e-01 1.35725588e-01 -9.41210628e-01 3.90147157e-02
5.20627141e-01 -1.01912104e-01 1.29247046e+00 -1.29115975e+00
-5.29656589e-01 -3.09016615e-01 -3.04317683e-01 -6.70144141e-01
-1.07456416e-01 5.02269149e-01 4.84997444e-02 -7.26342976e-01
1.55708730e-01 -9.56358373e-01 9.35698375e-02 -4.52660561e-01
-7.59197950e-01 -2.56339833e-02 8.38296533e-01 -6.96069956e-01
1.33840585e+00 -1.88480866e+00 -6.12407386e-01 5.82918048e-01
2.72356629e-01 3.84774745e-01 4.18278724e-01 5.99101663e-01
1.48855895e-01 4.58578259e-01 5.50981006e-03 -1.02338150e-01
-2.83705682e-01 3.55522782e-02 -6.43791974e-01 7.75390625e-01
6.55393526e-02 6.20933831e-01 -1.25311017e+00 -4.07037795e-01
-2.54279077e-01 1.51348636e-01 -1.41837239e-01 -1.86144412e-01
-3.10394496e-01 4.42779601e-01 -5.15596390e-01 4.74259645e-01
5.48017621e-01 -2.79387087e-01 1.55548826e-01 5.93282953e-02
-2.88441241e-01 8.12292755e-01 -8.40231955e-01 5.57075441e-01
-2.37486660e-01 7.18761265e-01 -1.92727223e-01 -7.26958752e-01
9.91537452e-01 2.43767247e-01 7.49484673e-02 -4.17594433e-01
5.12092292e-01 5.93927383e-01 -4.22479868e-01 -5.64001679e-01
6.61475241e-01 -2.82238036e-01 -3.95989209e-01 8.21607471e-01
-2.00887218e-01 -2.52737939e-01 -1.50214031e-01 2.81166971e-01
1.09946084e+00 -3.71347010e-01 6.21720910e-01 -4.19416539e-02
6.55651748e-01 5.90824306e-01 1.03902332e-01 8.59414577e-01
-1.65759504e-01 8.19554776e-02 4.98941720e-01 -7.85647571e-01
-1.01139510e+00 -3.03451747e-01 1.17903436e-02 9.28783715e-01
7.92637691e-02 -3.47874463e-01 -5.55458307e-01 -1.04182827e+00
9.94997397e-02 8.38987410e-01 -7.30932891e-01 -1.14113010e-01
-3.45770597e-01 -7.93723226e-01 6.09993219e-01 -5.30650556e-01
5.55091679e-01 -8.83006990e-01 -3.37860852e-01 4.27818090e-01
-5.81422985e-01 -1.14859629e+00 -9.72765908e-02 -1.86002612e-01
-4.91112858e-01 -1.18050945e+00 -1.00492239e-01 -3.41576904e-01
5.85557401e-01 3.11748207e-01 8.93221200e-01 2.88750917e-01
2.20707476e-01 -2.90058374e-01 -5.96805632e-01 -6.08495533e-01
-1.48637760e+00 -8.08294937e-02 5.67058101e-02 4.27185625e-01
4.16017532e-01 7.33590871e-02 -3.07154283e-02 4.27546442e-01
-1.10145974e+00 -2.47376785e-01 2.89943725e-01 6.27781510e-01
8.78856331e-02 1.94447994e-01 5.22839427e-01 -1.49760175e+00
7.25179195e-01 -9.58003163e-01 -6.44467533e-01 1.01597071e-01
-5.10938346e-01 2.95506939e-02 6.32774115e-01 -3.75773519e-01
-5.75277090e-01 -3.70689258e-02 -1.99003831e-01 2.57053673e-01
-7.88637772e-02 6.77261889e-01 3.79125327e-01 -1.46793261e-01
1.14034557e+00 -1.50477455e-03 -9.53918546e-02 -3.16898555e-01
1.49715543e-01 1.16787410e+00 2.06502989e-01 6.27171993e-02
1.00887668e+00 8.21962953e-01 -3.12812954e-01 -8.73182893e-01
-1.25285470e+00 -7.25378811e-01 -2.37735420e-01 -3.61058980e-01
5.16318381e-01 -6.07451439e-01 -4.11405742e-01 4.48948920e-01
-1.42261004e+00 2.47074023e-01 1.49022952e-01 4.14426357e-01
5.35199828e-02 3.66694927e-01 -5.77204466e-01 -8.15762401e-01
-1.29471883e-01 -7.75358021e-01 7.33534753e-01 -3.49205673e-01
-4.11835402e-01 -7.86547542e-01 3.41110945e-01 6.11771464e-01
4.09147292e-01 5.25948644e-01 5.33499777e-01 -1.21583796e+00
-1.53850675e-01 -9.79244232e-01 -2.83484221e-01 1.26619846e-01
-1.05762528e-02 9.16238129e-02 -8.97322357e-01 -5.09787574e-02
2.97862560e-01 -2.43086278e-01 6.80835843e-01 -3.16209674e-01
3.84485483e-01 -1.28583562e+00 -5.63584328e-01 -1.76538751e-01
1.19401002e+00 -2.35008150e-01 4.81303334e-01 8.38325858e-01
2.92457551e-01 8.00895989e-01 7.09448397e-01 2.54194736e-01
2.76725590e-01 7.01852858e-01 3.72116834e-01 3.20608122e-03
3.16303551e-01 -5.49327195e-01 4.09196794e-01 4.97552097e-01
2.87281513e-01 -2.43741348e-01 -8.83541346e-01 5.11900485e-01
-1.72430885e+00 -1.32699823e+00 -6.80808783e-01 2.16745496e+00
9.62099373e-01 4.49722648e-01 3.14030856e-01 2.02182844e-01
8.85771155e-01 2.25549433e-02 1.60780773e-01 -7.02923417e-01
-4.90555307e-03 -5.97586334e-02 8.98707867e-01 6.63831115e-01
-1.27737546e+00 1.11978590e+00 5.67893982e+00 6.53909564e-01
-1.41884434e+00 5.89378476e-01 5.64036310e-01 4.75871086e-01
-2.54321784e-01 -7.58122131e-02 -8.41017783e-01 6.35430396e-01
1.13125026e+00 1.07842691e-01 8.88208300e-02 9.38661754e-01
5.74475110e-01 -2.97489077e-01 -5.34052074e-01 4.45297539e-01
2.68778235e-01 -1.63394988e+00 5.79781132e-03 5.82737066e-02
6.80862844e-01 2.19217941e-01 -3.75004113e-01 1.95304930e-01
5.24512947e-01 -7.43798077e-01 9.37041998e-01 3.88352305e-01
2.83739388e-01 -5.04948080e-01 1.19900274e+00 9.15410519e-01
-2.18856737e-01 1.32762283e-01 -1.10291682e-01 -1.07579185e-02
1.22634768e-01 1.13727272e+00 -1.48548460e+00 2.57799655e-01
6.41539276e-01 5.32451630e-01 -6.56532109e-01 9.31364715e-01
-5.38244903e-01 6.35977328e-01 -3.66415024e-01 -3.94942075e-01
3.40974301e-01 2.34406799e-01 7.51903594e-01 1.48239088e+00
1.22694701e-01 -2.53441364e-01 8.35314021e-02 6.30520523e-01
-1.99037522e-01 3.18255454e-01 -9.40682352e-01 -2.52767056e-01
2.93527633e-01 8.84344697e-01 -8.91680002e-01 -7.89069831e-01
-1.40860528e-01 9.32974637e-01 8.65264535e-02 -2.75787711e-01
-7.07599938e-01 -3.66763249e-02 -1.23911472e-02 6.37887716e-01
-5.80293052e-02 1.30632132e-01 -1.43434137e-01 -1.41547537e+00
-1.89219847e-01 -1.18939567e+00 4.08561736e-01 -4.68436688e-01
-1.40251803e+00 9.20986414e-01 -2.22090453e-01 -1.24678731e+00
-9.60892662e-02 -2.19591975e-01 -3.69133979e-01 3.62834781e-01
-1.41557419e+00 -9.91043270e-01 -7.57746845e-02 3.84779781e-01
6.32565692e-02 1.14946784e-02 7.38420308e-01 1.63009003e-01
-6.26659170e-02 2.43979454e-01 -4.21210900e-02 4.65840667e-01
9.09077346e-01 -7.67446518e-01 5.16603887e-01 8.95148158e-01
2.19191045e-01 5.50577819e-01 1.15198958e+00 -1.07549942e+00
-7.11850047e-01 -1.30079174e+00 1.84408927e+00 -7.97031641e-01
1.25082278e+00 -2.54284233e-01 -8.12647462e-01 5.14815986e-01
-7.70583749e-02 -1.23708598e-01 6.25583291e-01 -3.21433663e-01
-8.08907747e-01 2.53039271e-01 -1.57097054e+00 1.46077946e-01
1.93193361e-01 -5.76797068e-01 -7.94671834e-01 9.34830129e-01
5.92366934e-01 -2.54200608e-01 -3.71830761e-01 -1.00120805e-01
4.60384876e-01 -9.35828149e-01 5.13689816e-01 -8.87180924e-01
4.26613450e-01 -4.62459743e-01 2.97669500e-01 -1.29236221e+00
6.44718036e-02 -6.53807342e-01 2.01972201e-01 8.93936336e-01
6.61797523e-01 -9.58393753e-01 4.49216545e-01 2.19480664e-01
3.02862287e-01 -1.02453835e-01 -8.28601241e-01 -6.96386874e-01
-1.86224118e-01 -4.49048370e-01 4.21931565e-01 1.57670379e+00
5.31608105e-01 2.44055539e-01 -6.15068674e-01 3.45590800e-01
3.49127471e-01 1.07838832e-01 8.69845092e-01 -1.30475986e+00
-2.16072828e-01 -1.09192856e-01 -4.41140711e-01 -4.69103634e-01
-8.48113298e-02 -6.74506724e-01 3.34993005e-02 -1.13970685e+00
-6.85169026e-02 -4.99959618e-01 3.97875905e-01 5.39953649e-01
1.99495360e-01 7.06599116e-01 -9.43520814e-02 5.58795929e-01
-3.72907639e-01 -1.52319416e-01 8.40840161e-01 -2.88339674e-01
-2.13818297e-01 5.31036019e-01 -4.79763001e-01 9.13902044e-01
8.66765082e-01 -1.25991774e+00 5.15241802e-01 1.68686837e-01
8.86012316e-01 -4.69502695e-02 6.16181016e-01 -7.36523271e-01
2.06102654e-01 -2.42341191e-01 -1.03479899e-01 -3.86724770e-01
-1.23740733e-01 -9.32100236e-01 1.35806784e-01 9.01662707e-01
-2.96823680e-01 -8.15685764e-02 -2.09114090e-01 7.15193689e-01
-3.78913283e-01 -3.67618114e-01 1.02604580e+00 -5.00893772e-01
-3.65746878e-02 -2.14811802e-01 -8.84759426e-01 -1.94686264e-01
9.78600681e-01 1.51109278e-01 -7.41803586e-01 -5.47433257e-01
-7.20502973e-01 -3.47333938e-01 5.71308672e-01 3.72818649e-01
2.50213802e-01 -9.83379245e-01 -8.29126060e-01 4.74732816e-02
4.17918414e-01 -6.44954085e-01 -5.64723611e-01 9.36730683e-01
-7.29207218e-01 3.46980959e-01 1.17762104e-01 -1.62074044e-01
-1.22343302e+00 6.20431662e-01 1.81763858e-01 -3.85980725e-01
-2.85176218e-01 4.60777283e-01 -8.89523149e-01 -4.52275246e-01
-2.70612061e-01 -2.84985632e-01 -5.13023138e-01 2.33377531e-01
8.85692775e-01 3.26198041e-01 4.52343374e-01 -1.21549642e+00
-3.56850237e-01 -9.10786986e-02 -9.32474881e-02 -1.48883954e-01
1.41894758e+00 1.21996820e-01 -5.97381711e-01 5.12294710e-01
1.13561761e+00 7.13735223e-01 -8.72863680e-02 -2.60032892e-01
3.34810734e-01 -7.03780830e-01 4.98534441e-02 -9.46554184e-01
-6.62832797e-01 2.19843373e-01 7.89163038e-02 9.68375027e-01
3.13541919e-01 1.22409396e-01 7.79814839e-01 2.43199423e-01
6.39105797e-01 -9.35333133e-01 -1.95894372e-02 5.61815560e-01
7.01190174e-01 -1.51116574e+00 1.79622397e-01 -6.84925973e-01
-5.11075258e-01 1.16583872e+00 -3.67889255e-02 -1.36959627e-01
7.42396533e-01 -1.77074999e-01 1.31375358e-01 -6.13221169e-01
-3.11285436e-01 1.06000371e-01 2.37746641e-01 1.60971001e-01
2.47830898e-01 1.53615966e-01 -8.21157336e-01 2.62541443e-01
-4.23852235e-01 1.27916366e-01 9.91190016e-01 8.41919184e-01
-7.52304494e-01 -1.03549230e+00 -7.91805506e-01 4.27785009e-01
-8.62417638e-01 -8.24478939e-02 -1.03627706e+00 6.02254391e-01
3.53934735e-01 1.30182767e+00 -4.03536469e-01 -4.94184852e-01
3.88111435e-02 5.33393472e-02 -2.18362272e-01 -8.04526448e-01
-1.09469414e+00 -3.00016344e-01 5.54640412e-01 -5.24721861e-01
-6.82214081e-01 -4.70310956e-01 -8.78414035e-01 -7.02466369e-01
-6.55322313e-01 5.68798721e-01 1.00179362e+00 1.25976264e+00
1.86265379e-01 -2.48782203e-01 7.87550390e-01 -2.60802120e-01
-3.43317330e-01 -8.91048491e-01 -1.65281519e-01 5.07222235e-01
5.96484721e-01 -4.32076871e-01 -6.22820079e-01 1.05379194e-01]
|
[8.16787338256836, 10.234122276306152]
|
5a04e966-9f14-49f5-a40a-882d73d6fdfd
|
gowfed-a-novel-federated-network-intrusion
|
2210.16441
| null |
https://arxiv.org/abs/2210.16441v2
|
https://arxiv.org/pdf/2210.16441v2.pdf
|
GowFed -- A novel Federated Network Intrusion Detection System
|
Network intrusion detection systems are evolving into intelligent systems that perform data analysis while searching for anomalies in their environment. Indeed, the development of deep learning techniques paved the way to build more complex and effective threat detection models. However, training those models may be computationally infeasible in most Edge or IoT devices. Current approaches rely on powerful centralized servers that receive data from all their parties - violating basic privacy constraints and substantially affecting response times and operational costs due to the huge communication overheads. To mitigate these issues, Federated Learning emerged as a promising approach, where different agents collaboratively train a shared model, without exposing training data to others or requiring a compute-intensive centralized infrastructure. This work presents GowFed, a novel network threat detection system that combines the usage of Gower Dissimilarity matrices and Federated averaging. Different approaches of GowFed have been developed based on state-of the-art knowledge: (1) a vanilla version; and (2) a version instrumented with an attention mechanism. Furthermore, each variant has been tested using simulation oriented tools provided by TensorFlow Federated framework. In the same way, a centralized analogous development of the Federated systems is carried out to explore their differences in terms of scalability and performance - across a set of designed experiments/scenarios. Overall, GowFed intends to be the first stepping stone towards the combined usage of Federated Learning and Gower Dissimilarity matrices to detect network threats in industrial-level networks.
|
['Javier Navaridas', 'Jose A. Pascual', 'Aitor Belenguer']
|
2022-10-28
| null | null | null | null |
['network-intrusion-detection']
|
['miscellaneous']
|
[-1.43881336e-01 -4.67581004e-02 -4.16209586e-02 -1.62505403e-01
-2.14404181e-01 -7.20463991e-01 7.99119651e-01 3.05502325e-01
-2.48545349e-01 6.08553052e-01 -3.54198277e-01 -5.18834710e-01
-7.06872880e-01 -1.07767808e+00 -3.02983463e-01 -7.24068344e-01
-6.48115277e-01 4.86939818e-01 1.42038718e-01 -1.58213779e-01
1.73842847e-01 8.84355366e-01 -1.53470767e+00 2.06198052e-01
4.10831541e-01 1.25082517e+00 -5.74964106e-01 6.00172400e-01
-5.27741499e-02 6.79178953e-01 -6.69198990e-01 -6.24577343e-01
7.99129784e-01 1.17577054e-01 -7.03158855e-01 -3.48713458e-01
2.47939080e-01 -2.70557046e-01 -2.30348974e-01 1.07272828e+00
4.13835466e-01 -6.68452233e-02 -4.14274260e-03 -1.79218709e+00
-1.87709853e-01 5.84891200e-01 -1.19798884e-01 1.48481429e-01
2.94749469e-01 5.78704655e-01 5.99663377e-01 -1.49974272e-01
6.13171995e-01 8.86828601e-01 4.81364608e-01 2.97643423e-01
-1.11591887e+00 -7.88213074e-01 1.05633274e-01 3.22267443e-01
-1.02115023e+00 -1.56688213e-01 8.44848633e-01 -3.76539737e-01
9.82681751e-01 6.19412482e-01 1.91453636e-01 1.37557006e+00
2.78567463e-01 1.58920497e-01 1.14590991e+00 -2.22841352e-01
7.30196178e-01 4.34255928e-01 3.06592137e-01 4.40144271e-01
6.68558598e-01 4.59781855e-01 -2.07617477e-01 -5.41679680e-01
2.36699224e-01 1.83673710e-01 -4.29167561e-02 -5.12925923e-01
-7.21866846e-01 6.65837288e-01 3.56251299e-01 7.14882731e-01
-6.42378986e-01 -2.12655619e-01 1.06130707e+00 6.23036325e-01
3.47459853e-01 3.86177957e-01 -6.05553687e-01 5.83537593e-02
-7.30927110e-01 9.80789214e-02 1.19548965e+00 4.70736951e-01
7.87096441e-01 1.67158023e-01 1.30477935e-01 -6.88715354e-02
2.13491142e-01 3.07008587e-02 2.43593842e-01 -5.18913925e-01
2.62615740e-01 9.54471111e-01 -1.24160767e-01 -1.22361147e+00
-4.65373814e-01 -5.94019830e-01 -7.19930530e-01 7.26418495e-01
2.67179370e-01 -5.23309886e-01 -3.28750283e-01 1.54258072e+00
6.82091534e-01 3.99194479e-01 8.64517763e-02 8.25136840e-01
-1.07563399e-01 3.02648365e-01 1.21001370e-01 7.44828284e-02
1.04460347e+00 -6.20451331e-01 -7.10909963e-01 4.14486259e-01
6.25214577e-01 -5.30607343e-01 4.65200961e-01 8.24386835e-01
-5.72006941e-01 -2.71838576e-01 -1.31379175e+00 6.18507445e-01
-1.18727326e+00 -4.52392995e-01 8.59269261e-01 1.45758355e+00
-9.14142370e-01 8.43227029e-01 -6.21032596e-01 -3.64758104e-01
5.74884415e-01 6.12558484e-01 -4.10008848e-01 1.87581498e-02
-1.02063346e+00 7.91784346e-01 4.14954901e-01 -3.08784228e-02
-1.23459947e+00 -5.74316800e-01 -4.07574326e-01 1.62923429e-02
4.10296023e-01 -6.10252380e-01 8.14094961e-01 -8.52712333e-01
-1.36324847e+00 4.87997591e-01 7.80091584e-01 -9.08485115e-01
6.54862046e-01 -1.78996310e-01 -9.21894252e-01 4.43881229e-02
-2.66938597e-01 -3.86308104e-01 7.43201971e-01 -1.05551779e+00
-6.28239453e-01 -6.57419264e-01 2.66812354e-01 -5.21688223e-01
-7.31223643e-01 2.96671361e-01 4.18469548e-01 -2.08963603e-01
-5.59057593e-01 -5.69653332e-01 -3.81591469e-01 -6.90837950e-02
-3.95754904e-01 3.38282846e-02 1.53627610e+00 -4.51345921e-01
9.58901823e-01 -1.97395349e+00 -3.42921227e-01 6.41555607e-01
2.60761976e-01 8.26725781e-01 -4.50620875e-02 8.15534413e-01
-2.51022279e-01 2.18001857e-01 8.46018493e-02 -2.69244313e-01
3.12750310e-01 2.16181017e-02 -2.71944821e-01 7.04750896e-01
-6.41347170e-02 3.29588294e-01 -6.63854957e-01 -1.80968881e-01
6.80852950e-01 4.43262756e-01 -2.98462093e-01 3.27414349e-02
-2.70152271e-01 3.74705166e-01 -6.99521840e-01 7.17419207e-01
7.85753787e-01 1.79882377e-01 3.58523488e-01 -9.16445106e-02
-3.67488116e-01 -7.82458037e-02 -1.39727271e+00 1.38889217e+00
-5.42589426e-01 1.42644361e-01 6.49996996e-01 -1.18550658e+00
1.03846681e+00 6.14252508e-01 9.52568471e-01 -6.79890573e-01
6.80548429e-01 1.62197888e-01 -2.00670063e-01 -4.18653488e-01
-1.48175145e-02 1.80266500e-01 -4.75750491e-02 6.22690558e-01
2.97637612e-01 4.97931540e-01 -9.81363654e-03 9.10611916e-03
1.63411736e+00 -1.29761547e-01 1.53455451e-01 -1.20597631e-01
1.00307798e+00 -3.09340786e-02 5.16317308e-01 5.74245155e-01
-4.66745317e-01 -3.19373399e-01 2.61608511e-01 -9.15782452e-01
-6.80192828e-01 -9.99399602e-01 1.25583991e-01 5.97975373e-01
-6.33114055e-02 -3.38039190e-01 -7.30673015e-01 -1.08641732e+00
-3.19366679e-02 7.72166908e-01 -3.29339355e-01 -3.72472793e-01
-3.86804193e-01 -4.28036869e-01 9.05954242e-01 7.88681433e-02
6.22707069e-01 -8.79081607e-01 -9.15248752e-01 2.41814584e-01
5.50654709e-01 -9.98918355e-01 2.59737611e-01 3.87984097e-01
-5.15750706e-01 -1.48871744e+00 9.19717327e-02 -2.08987296e-01
2.03212425e-01 -1.73482951e-02 8.66058528e-01 -1.77893918e-02
-7.69240379e-01 5.19007146e-01 -4.30492789e-01 -7.20873117e-01
-4.16811109e-01 -1.56591728e-01 2.96163619e-01 4.64607000e-01
4.65089500e-01 -9.03721929e-01 -4.68666822e-01 4.39867437e-01
-1.05410278e+00 -9.06464994e-01 4.96871799e-01 2.78797209e-01
-1.13800906e-01 5.09272695e-01 6.67355299e-01 -8.64330947e-01
6.85223877e-01 -7.10676312e-01 -1.11125219e+00 1.37203068e-01
-7.31418729e-01 -2.20573768e-01 1.13003838e+00 -2.71321833e-01
-9.83481944e-01 -1.60850570e-01 5.76913729e-02 -5.95329821e-01
-6.56028926e-01 1.04346767e-01 -2.94089824e-01 -4.93184656e-01
6.93960905e-01 -2.54305720e-01 2.72774771e-02 -4.14162636e-01
2.66462833e-01 8.15924048e-01 2.70604074e-01 -4.99754727e-01
1.13007498e+00 6.45927787e-01 2.44874209e-01 -6.67344928e-01
-1.79603219e-01 -3.72846842e-01 -1.31037876e-01 -2.58530378e-01
5.55170178e-01 -5.12980402e-01 -1.11097205e+00 3.96916866e-01
-9.28346515e-01 1.29467607e-01 -4.01119590e-01 4.60610867e-01
-1.57104447e-01 3.54589522e-01 -4.95562583e-01 -8.93602669e-01
-6.92346215e-01 -9.30982709e-01 4.38400000e-01 4.94914874e-02
-4.75097597e-02 -9.15824473e-01 3.90310377e-01 2.87654579e-01
1.00387573e+00 6.55441642e-01 7.63219297e-01 -1.35926592e+00
-5.16400278e-01 -8.14219177e-01 1.40774455e-02 4.90723580e-01
1.65444225e-01 -7.65579846e-03 -1.06914878e+00 -4.70907122e-01
2.87333846e-01 -1.94965124e-01 -3.33236866e-02 -2.08838239e-01
8.84692788e-01 -4.33629274e-01 -2.62847841e-01 4.41357017e-01
1.54708791e+00 2.48705819e-01 4.50379759e-01 5.06404936e-01
2.00845137e-01 6.31368339e-01 4.33635741e-01 6.90686524e-01
-1.79347098e-01 5.22986531e-01 1.19314754e+00 3.82420607e-02
3.18118840e-01 1.19806878e-01 3.18982840e-01 1.72132373e-01
4.18813378e-02 -1.98347807e-01 -6.99765146e-01 2.82226264e-01
-1.57387412e+00 -1.00610805e+00 -3.19627747e-02 2.27598548e+00
-1.27536133e-01 3.42608035e-01 2.14535385e-01 3.89052451e-01
7.21003413e-01 1.09253265e-01 -4.68880653e-01 -8.11821878e-01
1.37793571e-01 5.16531050e-01 4.99309003e-01 3.93730141e-02
-1.09840262e+00 3.70698690e-01 4.94303608e+00 4.16419715e-01
-1.28649378e+00 2.94470996e-01 1.31581649e-01 -9.39682219e-03
2.25193813e-01 2.04770938e-01 -3.66903603e-01 3.05567443e-01
1.34910572e+00 -2.32243836e-01 4.30482507e-01 1.23016286e+00
8.63388479e-02 3.78737479e-01 -9.68992293e-01 7.74697602e-01
-2.06726521e-01 -1.25882471e+00 -2.08338752e-01 4.30518925e-01
3.12429041e-01 3.68717313e-01 -1.21546715e-01 2.60737211e-01
5.88311613e-01 -7.46092081e-01 3.23616713e-01 4.28573132e-01
1.65703773e-01 -8.91049743e-01 9.59563255e-01 3.54595780e-01
-9.72767889e-01 -5.69127977e-01 -2.37390362e-02 -1.10315666e-01
5.77728860e-02 6.40215278e-01 -7.73003817e-01 1.18756855e+00
6.83905423e-01 7.93337524e-02 -4.45211738e-01 1.07606125e+00
2.62139171e-01 6.25980020e-01 -4.85645920e-01 -2.30243523e-03
3.47392708e-01 -1.31798804e-01 8.65202606e-01 8.87535989e-01
1.42262176e-01 -5.65937281e-01 3.37976038e-01 8.45275044e-01
1.63456902e-01 -5.75715303e-03 -8.77178848e-01 -4.35085222e-02
4.35231507e-01 1.76566219e+00 -5.53146005e-01 1.52567908e-01
-3.00335526e-01 6.33631587e-01 -5.76777458e-02 -6.26432002e-02
-8.30909014e-01 -3.61984015e-01 1.01485634e+00 4.38725278e-02
1.81804597e-01 -1.17327817e-01 3.92845422e-02 -7.70223439e-01
-8.06683972e-02 -1.09215188e+00 7.46215224e-01 -9.93514881e-02
-1.55416334e+00 9.41224813e-01 -1.77440405e-01 -1.21166325e+00
-2.10417598e-01 -7.01497614e-01 -8.23770344e-01 6.05229497e-01
-1.23628998e+00 -1.20361328e+00 -1.61534876e-01 1.16818810e+00
1.60912648e-01 -4.72726911e-01 1.28435123e+00 5.19298255e-01
-8.92130792e-01 4.60352659e-01 -1.53068155e-01 2.29907893e-02
2.53027052e-01 -7.87164271e-01 1.35655984e-01 1.03682554e+00
2.18844533e-01 5.52671492e-01 5.98636150e-01 -5.97567618e-01
-1.51961255e+00 -1.02456677e+00 2.40558386e-01 -5.19451760e-02
8.71957123e-01 -4.31141287e-01 -5.71559191e-01 6.89386487e-01
5.10234296e-01 4.00866747e-01 7.14688897e-01 1.69906080e-01
-4.23809469e-01 -3.94675314e-01 -1.74732816e+00 4.25794125e-01
5.46520829e-01 -3.62130046e-01 -1.60248607e-01 4.34563130e-01
6.46234751e-01 2.19727963e-01 -9.49096978e-01 1.64948672e-01
8.55729431e-02 -1.41341233e+00 6.55726314e-01 -7.15091348e-01
-4.86347675e-01 -2.89465427e-01 -2.22366750e-01 -9.42884445e-01
-1.63730755e-01 -1.03174543e+00 -2.64316261e-01 1.47484851e+00
7.67890289e-02 -1.15806890e+00 9.56535935e-01 6.98744953e-01
4.03498039e-02 -5.25206923e-01 -1.26647413e+00 -8.81569564e-01
-3.30985427e-01 -6.62630439e-01 1.02873838e+00 1.10867739e+00
-5.00811003e-02 -1.56446069e-01 -9.56263244e-02 5.88818550e-01
8.23675096e-01 -2.52901137e-01 8.60071838e-01 -1.37557471e+00
-2.70839423e-01 -2.02076554e-01 -8.96271169e-01 3.58986646e-01
4.20999825e-02 -6.94105327e-01 -8.35557222e-01 -7.26238847e-01
-5.93949914e-01 -5.37846684e-01 -7.06092238e-01 6.38071239e-01
8.04418325e-01 -1.27694577e-01 3.14926475e-01 -2.78993368e-01
-5.36068380e-01 2.36180454e-01 2.92972624e-01 -1.56643003e-01
1.18595913e-01 1.35201454e-01 -4.17927384e-01 6.95541978e-01
1.15636802e+00 -6.27221704e-01 -4.99164760e-01 2.88659818e-02
-5.38279628e-03 -3.13595384e-02 8.05739880e-01 -1.52960300e+00
4.35199946e-01 1.09806128e-01 5.92656359e-02 -4.10531253e-01
1.44616142e-01 -1.61103916e+00 4.48449612e-01 7.19742715e-01
1.49291769e-01 1.17529429e-01 1.80524871e-01 5.73571324e-01
-1.50400503e-02 -8.63582790e-02 4.59589183e-01 -1.56831726e-01
-5.67672372e-01 2.84474105e-01 -2.86512434e-01 -5.73352337e-01
1.69816256e+00 -2.72605777e-01 -4.71878409e-01 -3.03801429e-02
-5.71133375e-01 1.59630179e-01 3.65565568e-01 4.99903679e-01
2.81375855e-01 -7.12593496e-01 -3.13303292e-01 3.50993663e-01
-2.18861207e-01 -4.82696027e-01 3.06770772e-01 5.97330809e-01
-2.68071026e-01 5.05711079e-01 -5.22796571e-01 -1.98997885e-01
-1.21305490e+00 1.18210864e+00 4.13740724e-01 -6.96626723e-01
-5.91614902e-01 3.49724203e-01 -6.23441219e-01 -7.02942014e-01
3.67865622e-01 1.71972916e-01 -5.17188921e-04 -5.77882417e-02
5.98266006e-01 8.50508392e-01 6.79237843e-01 -2.57973701e-01
-6.27622366e-01 4.14437428e-02 -7.50075653e-02 1.48449674e-01
1.39743090e+00 8.61869007e-02 -2.15036869e-01 -1.58710510e-01
1.05164158e+00 3.59239010e-03 -7.85392165e-01 6.19975962e-02
3.21537256e-01 -4.22655791e-01 4.26184386e-01 -1.08040488e+00
-1.34304512e+00 4.11817223e-01 9.54015076e-01 4.80517536e-01
1.23189795e+00 -6.33200049e-01 5.48960567e-01 3.82184863e-01
9.37136710e-01 -8.07263196e-01 -1.99266180e-01 9.30714328e-03
2.16423377e-01 -8.32721233e-01 -6.41579852e-02 -3.45088959e-01
-2.62731075e-01 1.18385530e+00 5.31585157e-01 -1.69333637e-01
8.21105182e-01 5.99687040e-01 1.33421961e-02 -5.47090054e-01
-7.70312965e-01 1.55321673e-01 -3.72809708e-01 8.15712750e-01
-1.13331206e-01 -9.37889665e-02 -1.96028113e-01 5.72923303e-01
2.67215848e-01 1.45357370e-01 2.56752580e-01 1.11220348e+00
8.23845938e-02 -1.48217499e+00 -6.42931163e-01 2.55060524e-01
-6.06909215e-01 5.01743734e-01 -3.98191452e-01 8.44407499e-01
4.59791899e-01 1.24979091e+00 -3.92783791e-01 -6.35685802e-01
3.67430717e-01 9.34205353e-02 -4.83293906e-02 -1.49425462e-01
-1.33128452e+00 -5.35823703e-01 1.82979152e-01 -1.10875249e+00
-1.77679732e-01 -5.01410663e-01 -8.24008465e-01 -4.40709114e-01
-4.01192576e-01 1.90168574e-01 1.17811716e+00 6.84204698e-01
5.94314694e-01 6.25693917e-01 9.90873277e-01 -7.61181474e-01
-9.96159554e-01 -5.85229576e-01 -6.01803243e-01 2.78859019e-01
1.21824399e-01 -5.19569218e-01 -4.90453273e-01 -7.52882838e-01]
|
[5.362105369567871, 7.116565227508545]
|
a939ab6c-62dc-4170-9389-91f3ca257b80
|
snowman-a-million-scale-chinese-commonsense
|
2306.10241
| null |
https://arxiv.org/abs/2306.10241v1
|
https://arxiv.org/pdf/2306.10241v1.pdf
|
Snowman: A Million-scale Chinese Commonsense Knowledge Graph Distilled from Foundation Model
|
Constructing commonsense knowledge graphs (CKGs) has attracted wide research attention due to its significant importance in cognitive intelligence. Nevertheless, existing CKGs are typically oriented to English, limiting the research in non-English languages. Meanwhile, the emergence of foundation models like ChatGPT and GPT-4 has shown promising intelligence with the help of reinforcement learning from human feedback. Under the background, in this paper, we utilize foundation models to construct a Chinese CKG, named Snowman. Specifically, we distill different types of commonsense head items from ChatGPT, and continue to use it to collect tail items with respect to the head items and pre-defined relations. Based on the preliminary analysis, we find the negative commonsense knowledge distilled by ChatGPT achieves lower human acceptance compared to other knowledge. Therefore, we design a simple yet effective self-instruct filtering strategy to filter out invalid negative commonsense. Overall, the constructed Snowman covers more than ten million Chinese commonsense triples, making it the largest Chinese CKG. Moreover, human studies show the acceptance of Snowman achieves 90.6\%, indicating the high-quality triples distilled by the cutting-edge foundation model. We also conduct experiments on commonsense knowledge models to show the usability and effectiveness of our Snowman.
|
['Xin Zheng', 'Guanfeng Liu', 'An Liu', 'Zhixu Li', 'Yunlong Liang', 'Jianfeng Qu', 'Jiaan Wang']
|
2023-06-17
| null | null | null | null |
['knowledge-graphs']
|
['knowledge-base']
|
[-6.22360855e-02 2.06876859e-01 -2.31368601e-01 -1.34931847e-01
1.44736394e-01 -2.48248458e-01 2.87086397e-01 5.72933070e-02
-4.85175788e-01 8.54817748e-01 2.74048388e-01 -2.90512025e-01
-2.49204621e-01 -1.11988831e+00 -3.82984966e-01 -2.03806728e-01
3.33919019e-01 3.15540254e-01 4.23137486e-01 -7.34257281e-01
4.00271714e-01 -1.85532585e-01 -1.37619424e+00 7.75576159e-02
1.80693460e+00 8.20375264e-01 2.54641294e-01 -6.59119040e-02
-2.17791662e-01 1.21730626e+00 -8.00067365e-01 -9.65242743e-01
-1.93074197e-01 -6.82095110e-01 -1.12061715e+00 -2.63681859e-01
-1.33931711e-01 -2.31309831e-01 -2.89035022e-01 1.30615354e+00
2.82209486e-01 1.79430366e-01 2.09785506e-01 -1.30456674e+00
-1.54993320e+00 1.38555622e+00 -1.28536940e-01 1.09238878e-01
6.20542645e-01 3.71256210e-02 1.30748618e+00 -6.49472952e-01
7.20096231e-01 1.09601390e+00 5.95121264e-01 5.43187141e-01
-7.32174695e-01 -9.95461762e-01 2.50489801e-01 8.74970555e-01
-1.27216685e+00 2.69815903e-02 8.32634985e-01 -1.26693487e-01
1.09980953e+00 3.48565161e-01 1.31309795e+00 9.76468384e-01
-6.36029508e-05 1.00908470e+00 1.40541351e+00 -5.75235844e-01
3.25909972e-01 1.21447802e-01 5.67945480e-01 7.31868923e-01
6.52327716e-01 -1.20702557e-01 -8.17785144e-01 1.35807663e-01
4.66191858e-01 -1.53586552e-01 -2.51880139e-01 3.93291526e-02
-1.12364721e+00 7.68266499e-01 6.59865081e-01 4.51694071e-01
-1.96161598e-01 -1.53039753e-01 4.05370027e-01 2.82705486e-01
1.28243253e-01 6.78990841e-01 -3.46843719e-01 -3.70909870e-01
-5.08629978e-01 1.10487677e-01 7.19897151e-01 1.28993356e+00
5.83033741e-01 4.21343967e-02 -1.39713004e-01 7.81197786e-01
-3.08294035e-03 5.71588755e-01 7.18378901e-01 -8.14056277e-01
2.65590668e-01 1.16538417e+00 -1.86422229e-01 -1.23054790e+00
-3.76267314e-01 -4.88057792e-01 -8.77329588e-01 -5.22266805e-01
-7.47379437e-02 -4.33262400e-02 -6.36490881e-01 1.78131950e+00
5.54715609e-03 -4.07781266e-03 2.18549207e-01 6.93100870e-01
1.06099021e+00 3.45508009e-01 8.62481743e-02 -3.43918085e-01
1.31566453e+00 -6.66065574e-01 -9.18218017e-01 -3.24898392e-01
7.39327013e-01 -1.25223100e-01 1.67638433e+00 5.77260554e-01
-6.73577070e-01 -2.62692004e-01 -1.03443038e+00 -4.71556745e-02
-5.88809371e-01 4.45992611e-02 1.06630337e+00 7.94002593e-01
-7.24566042e-01 3.89341682e-01 -4.75601226e-01 -3.41152310e-01
3.55803668e-01 4.44224030e-02 5.74131086e-02 -2.49829873e-01
-2.00950265e+00 1.25469124e+00 1.13572502e+00 8.42074081e-02
-4.94010597e-01 -3.84278893e-01 -7.36098945e-01 7.47104660e-02
9.05822635e-01 -6.75963581e-01 8.86532009e-01 -7.55726576e-01
-1.46098042e+00 6.71897709e-01 8.03127326e-03 -2.85257727e-01
2.00622290e-01 -3.53430659e-01 -6.65575862e-01 8.65946338e-03
3.10583025e-01 2.33886331e-01 2.66253173e-01 -9.53934908e-01
-6.89280748e-01 -1.91300556e-01 4.74402279e-01 2.18911424e-01
-8.96374464e-01 -1.79974884e-01 -3.59092146e-01 -7.12795675e-01
1.49604619e-01 -8.21448803e-01 1.99948430e-01 -7.15256631e-01
-6.72344625e-01 -5.50556600e-01 4.56893504e-01 -7.18431473e-01
1.76378477e+00 -1.98002863e+00 1.20373726e-01 3.73291671e-01
5.24760664e-01 1.43899024e-01 1.68233231e-01 1.50421143e-01
1.13034248e-01 3.86739284e-01 -1.76824257e-01 3.86906832e-01
1.32136405e-01 2.93245882e-01 -1.50701970e-01 -3.41876715e-01
-5.57633266e-02 1.21472967e+00 -1.38232517e+00 -6.67866409e-01
5.13134487e-02 -2.50396907e-01 -6.43399298e-01 -1.00101821e-01
-1.73058927e-01 -3.78278524e-01 -5.94024420e-01 7.79747605e-01
3.39283913e-01 -4.07955945e-01 2.83394605e-01 -1.40334025e-01
1.92443699e-01 3.93460989e-01 -8.43339086e-01 1.26548767e+00
-1.36728615e-01 2.61546284e-01 -7.07669258e-01 -6.46522582e-01
1.06000400e+00 -1.69494659e-01 -2.86066718e-03 -9.16895568e-01
2.47311234e-01 3.40902507e-01 3.30013067e-01 -5.25448799e-01
7.65366614e-01 -4.78291422e-01 -2.18353301e-01 2.11641014e-01
7.26183355e-02 -3.91300023e-01 6.93493485e-01 6.92469239e-01
1.11513782e+00 -2.79812720e-02 6.70606196e-01 -2.35980436e-01
3.37582052e-01 1.77611426e-01 7.63033211e-01 6.39861286e-01
-3.37632000e-01 -6.41067028e-02 6.60072386e-01 6.18963577e-02
-4.54801530e-01 -8.88327956e-01 2.04941601e-01 1.03460693e+00
5.36074638e-01 -8.17490578e-01 -7.54323006e-01 -5.95664084e-01
-1.61188737e-01 1.39747870e+00 -6.85014188e-01 -7.86392272e-01
-2.12849081e-01 -7.22029746e-01 7.66848981e-01 6.01879478e-01
1.19529772e+00 -1.54345024e+00 -5.05925179e-01 1.45641804e-01
-6.10907674e-01 -1.06688440e+00 -2.13580072e-01 1.09015755e-01
-5.63647807e-01 -1.24250758e+00 -2.54860759e-01 -7.67952323e-01
3.34653944e-01 2.84228146e-01 1.17318642e+00 3.25131059e-01
1.90534204e-01 8.15088749e-02 -9.39297199e-01 -5.20269275e-01
-2.17906073e-01 1.32531866e-01 8.47632065e-02 -5.19316673e-01
9.01985824e-01 -5.45637667e-01 -1.56082079e-01 9.92445201e-02
-6.94121003e-01 2.27289826e-01 5.41119695e-01 8.14449191e-01
1.51319176e-01 5.32443345e-01 8.51145267e-01 -8.75073791e-01
1.28468299e+00 -3.56839597e-01 -1.97085440e-01 5.78945339e-01
-9.39444065e-01 -1.68245569e-01 7.07470894e-01 -2.85665005e-01
-1.06397879e+00 -6.21360004e-01 2.13007301e-01 -7.73041546e-02
3.66636485e-01 1.00843406e+00 -3.90462101e-01 2.38688320e-01
7.68957376e-01 5.56420505e-01 -3.45641106e-01 9.73396897e-02
4.11082506e-01 6.48235619e-01 4.56124097e-01 -8.17708850e-01
5.11332154e-01 -1.87503695e-02 -5.13604581e-01 -6.79901481e-01
-1.24961317e+00 -4.28700335e-02 -1.63575903e-01 -1.62122577e-01
7.64559448e-01 -5.70666611e-01 -1.06029773e+00 4.26059276e-01
-9.21357930e-01 -1.33256987e-01 -3.20346236e-01 4.25948322e-01
-3.33699197e-01 4.17734891e-01 -7.84611940e-01 -6.83242798e-01
-4.84988123e-01 -7.24422336e-01 3.13922286e-01 3.51337790e-01
-5.03109515e-01 -8.35325599e-01 -2.33249888e-01 3.89181197e-01
2.30626777e-01 -9.84630212e-02 1.32890797e+00 -7.43477881e-01
-2.75324434e-01 3.89911085e-02 -3.59524190e-01 5.31697035e-01
1.93129405e-01 -1.01241939e-01 -5.60461998e-01 2.66794533e-01
-3.74764614e-02 -6.45740271e-01 8.27554286e-01 1.00418940e-01
1.07913780e+00 -1.22552775e-01 -1.26076199e-03 1.92741305e-01
1.10799527e+00 2.93567449e-01 7.01065123e-01 7.76481211e-01
6.60077333e-01 2.19392121e-01 6.89058661e-01 5.38118780e-01
7.17163563e-01 6.31188378e-02 1.81236327e-01 3.84168714e-01
1.58707246e-01 -5.26351273e-01 4.52414483e-01 1.32167339e+00
-5.29694557e-01 -5.98386861e-03 -9.12493527e-01 4.93342191e-01
-1.68794179e+00 -1.18489027e+00 -3.13663296e-02 1.74441493e+00
1.44699657e+00 5.99547684e-01 -9.48613808e-02 3.63592148e-01
6.69006288e-01 -1.51707679e-01 -5.83415985e-01 -1.30484760e-01
-5.32734394e-01 2.06244498e-01 -9.93206725e-03 2.60647058e-01
-5.06530583e-01 1.50202167e+00 5.96620846e+00 1.10666287e+00
-6.34845257e-01 -1.24077193e-01 1.28808588e-01 2.30807424e-01
-6.17442310e-01 1.64789483e-01 -6.58315241e-01 4.87561584e-01
4.13237542e-01 -8.11649382e-01 5.95275998e-01 8.51110339e-01
-1.88356079e-02 -2.16573671e-01 -6.12681329e-01 9.33190465e-01
1.38999641e-01 -1.04489803e+00 4.11725789e-01 -2.94923693e-01
7.02074587e-01 -4.26699579e-01 -2.74793118e-01 9.44564283e-01
7.54420459e-01 -6.81760311e-01 8.43201160e-01 3.86206806e-01
4.39949811e-01 -9.77500260e-01 8.41807604e-01 4.74918514e-01
-9.48012412e-01 -1.08411975e-01 -5.61819613e-01 -2.72876948e-01
-3.27994190e-02 9.61946547e-01 -7.04599857e-01 7.60411978e-01
6.34098530e-01 8.87227714e-01 -9.22316849e-01 6.28453970e-01
-8.81340504e-01 8.62902582e-01 -1.14362024e-01 -6.08845472e-01
-1.99905299e-02 -2.11689681e-01 1.63990170e-01 1.01414847e+00
2.61323482e-01 4.95339930e-01 3.12837034e-01 1.08780277e+00
-3.48796546e-01 2.40384936e-01 -2.96304286e-01 -4.37145978e-01
8.27522099e-01 9.52464819e-01 -7.55417049e-01 -8.02687287e-01
-1.55666664e-01 8.24334860e-01 6.20251536e-01 2.62674928e-01
-9.76071179e-01 -8.15985978e-01 1.68695554e-01 -2.24014103e-01
-4.24668342e-02 1.94184724e-02 -6.24069929e-01 -1.31686628e+00
1.42280295e-01 -9.19051111e-01 4.73053783e-01 -1.14380288e+00
-1.46198845e+00 6.35063171e-01 1.49332151e-01 -8.48004341e-01
7.01548606e-02 -4.85494822e-01 -3.94163489e-01 5.73779106e-01
-1.10444689e+00 -8.61718953e-01 -5.91564596e-01 7.88355529e-01
2.35317573e-01 -8.22459087e-02 6.22012556e-01 -2.68057406e-01
-6.91627145e-01 5.33799052e-01 -3.18733126e-01 2.46605739e-01
3.88493508e-01 -1.26242316e+00 2.09097385e-01 7.85091519e-01
-1.20080635e-01 1.22802830e+00 6.40861630e-01 -1.29560637e+00
-1.02179146e+00 -6.71768010e-01 1.05736995e+00 -2.99155742e-01
9.77762103e-01 5.95394447e-02 -1.06090212e+00 6.97275937e-01
2.29863748e-01 -4.25319314e-01 7.72086561e-01 5.86860061e-01
-4.78491753e-01 -9.40146018e-03 -9.24626946e-01 9.12645280e-01
1.48017061e+00 -4.88982856e-01 -1.34927905e+00 1.37922287e-01
1.01203954e+00 -1.26995459e-01 -6.77465975e-01 2.69793540e-01
2.12158471e-01 -1.03577554e+00 4.54370260e-01 -4.66757238e-01
7.74764895e-01 -3.36795300e-01 -3.54541764e-02 -1.66970015e+00
-6.34759843e-01 -3.77731830e-01 -3.02468181e-01 1.18897760e+00
3.86855274e-01 -7.33738244e-01 6.61532462e-01 6.32890522e-01
-2.10730776e-01 -6.20854259e-01 -6.14817619e-01 -1.04689252e+00
-1.65101022e-01 -5.14604151e-01 6.99782193e-01 1.29733932e+00
8.43921125e-01 7.40982056e-01 -3.05814356e-01 -3.94396961e-01
3.54337931e-01 1.99722677e-01 2.74234653e-01 -1.18392217e+00
-1.58706427e-01 -4.57795769e-01 -1.30767331e-01 -8.84365022e-01
2.56470114e-01 -1.16117823e+00 9.65517201e-03 -1.68703270e+00
5.72108448e-01 -4.89441961e-01 -3.93348634e-01 8.11983883e-01
-8.18919659e-01 -7.67235756e-02 1.32096648e-01 -5.29850721e-02
-7.31173992e-01 7.65492260e-01 1.56202829e+00 -1.09340832e-01
-2.34610170e-01 -4.82165664e-01 -1.25526428e+00 7.41207778e-01
1.03517604e+00 -1.91008940e-01 -7.68336356e-01 -1.68857560e-01
7.38751888e-01 -5.08053303e-01 2.55174309e-01 -8.57450128e-01
1.17556371e-01 -5.81864297e-01 1.86968714e-01 -3.31302762e-01
9.98663343e-03 -4.96924609e-01 -2.24547029e-01 4.69068587e-01
-2.78420478e-01 5.70158567e-03 6.90046772e-02 3.14704686e-01
-1.90381959e-01 -2.50120878e-01 2.96405941e-01 -3.15066218e-01
-1.14148676e+00 -1.87574059e-01 5.46978973e-03 6.83994889e-01
8.55437458e-01 -4.38973397e-01 -4.36278015e-01 -1.96162552e-01
-6.09103262e-01 2.15648904e-01 2.29171365e-01 4.15468037e-01
8.79369199e-01 -1.35002184e+00 -5.99999309e-01 -1.20369643e-01
3.54061097e-01 -1.53679505e-01 1.13519870e-01 7.39551961e-01
-3.39911133e-01 2.90815860e-01 -3.04326385e-01 -3.19546685e-02
-1.00994003e+00 7.52462685e-01 -2.29956228e-02 -1.51230887e-01
-6.95361376e-01 7.37491131e-01 -2.53426164e-01 -3.18028241e-01
-1.80342235e-02 -6.21432364e-01 -3.43071789e-01 -2.70125316e-03
4.35737371e-01 5.54110050e-01 2.48794630e-02 -9.95447412e-02
-4.35660928e-01 4.66394089e-02 -1.69811293e-01 7.09684193e-02
1.12418580e+00 1.16583221e-01 -4.60311472e-01 5.29934943e-01
4.71033186e-01 1.24885380e-01 -3.41089159e-01 -1.27468497e-01
2.85282910e-01 -4.06522274e-01 -2.36146137e-01 -1.13235033e+00
-8.48248124e-01 4.72284853e-01 -1.70048788e-01 2.91749328e-01
1.19151855e+00 -1.21989967e-02 7.73645878e-01 8.48379433e-01
7.33543515e-01 -1.43016946e+00 1.01479307e-01 9.93528068e-01
8.65401864e-01 -8.43952298e-01 1.21338710e-01 -6.80820644e-01
-8.81518066e-01 7.87188411e-01 1.05042017e+00 1.31480936e-02
3.80720049e-01 -1.13519505e-01 -2.86935925e-01 -2.23393202e-01
-7.27485180e-01 -4.09145921e-01 1.83518250e-02 7.05379426e-01
4.81205046e-01 2.98187375e-01 -7.00498939e-01 1.11014068e+00
-9.13149059e-01 4.49600041e-01 6.81949794e-01 9.21761513e-01
-7.42489278e-01 -6.84355259e-01 -1.59343287e-01 6.74232125e-01
1.85705423e-01 -6.09538317e-01 -8.39728236e-01 9.39565957e-01
3.10566694e-01 1.00155139e+00 -6.26470447e-01 -8.77828181e-01
7.01623321e-01 7.59236738e-02 4.61398900e-01 -5.05217671e-01
-6.45305037e-01 -5.31060338e-01 2.40480438e-01 -3.33048493e-01
-3.90947133e-01 -3.10833037e-01 -1.68781590e+00 -6.54910982e-01
-6.18076861e-01 3.10374111e-01 -1.41909719e-01 1.23219192e+00
8.59401748e-02 6.53345585e-01 1.45013258e-01 7.46066794e-02
-4.87811446e-01 -1.14314091e+00 -6.74995482e-01 4.93734539e-01
-4.07212466e-01 -8.78882587e-01 -2.83555984e-01 -7.35932663e-02]
|
[9.999024391174316, 8.053938865661621]
|
da12f079-9725-4730-88b7-2d858483c3cf
|
three-dimensional-bin-packing-and-mixed-case
| null | null |
https://pubsonline.informs.org/doi/10.1287/ijoo.2019.0013
|
https://pubsonline.informs.org/doi/pdf/10.1287/ijoo.2019.0013
|
Three-Dimensional Bin Packing and Mixed-Case Palletization
|
Despite its wide range of applications, the three-dimensional bin-packing problem is still one of the most difficult optimization problems to solve. Currently, medium- to large-size instances are only solved heuristically and remain out of reach of exact methods. This is particularly true for its practical variant, the mixed-case palletization problem, where item support is needed. This and the lack of a realistic benchmark data set are identified as major research gaps by a recent survey. In this work, we propose a novel formulation and a column-generation solution approach, where the pricing subproblem is a two-dimensional layer-generation problem. Layers are highly desirable in practical packings as they are easily packable and can accommodate important practical constraints such as item support, family groupings, isle friendliness, and load bearing. Being key to the success of the column-generation approach, the pricing subproblem is solved optimally as well as heuristically and is enhanced by using item grouping, item replacement, layer reorganization, and layer spacing. We conduct extensive computational experiments and compare against existing approaches. We also use industrial data to train and propose a realistic data set. The proposed approach outperforms the best-performing algorithm in the literature on most instances and succeeds to solve practical size instances in very reasonable computational times.
|
['Burak Yildiz', 'Fatma Gzara', 'Samir Elhedhli']
|
2019-07-08
| null | null | null |
informs-2019-7
|
['3d-bin-packing']
|
['miscellaneous']
|
[ 2.90132291e-03 1.92067574e-03 -6.34057820e-01 -1.60027221e-01
-3.67269307e-01 -5.63762546e-01 -3.80741626e-01 5.36206663e-01
-2.60449618e-01 1.35952961e+00 -2.84425139e-01 -3.94203246e-01
-8.78465295e-01 -8.94815028e-01 -8.19292724e-01 -6.54822946e-01
-5.26109695e-01 1.06822181e+00 1.32583991e-01 -1.98755026e-01
2.28927985e-01 6.47836745e-01 -1.64126694e+00 1.07669704e-01
1.26305890e+00 1.19601119e+00 5.19609988e-01 2.34004915e-01
-2.75663614e-01 -1.27121359e-01 -7.67028391e-01 -4.25001591e-01
5.91225863e-01 -1.20169697e-02 -6.54012978e-01 7.36851692e-01
-3.68136093e-02 -2.18952730e-01 3.86448860e-01 5.79522848e-01
4.07854825e-01 3.03262293e-01 2.70864666e-01 -1.50395107e+00
-4.19136167e-01 4.93591130e-01 -9.90110695e-01 2.79854178e-01
4.12467450e-01 -3.96794289e-01 1.16055405e+00 -4.59512860e-01
2.60928094e-01 8.65565956e-01 2.32943356e-01 1.35906965e-01
-1.35742974e+00 -3.18666041e-01 4.94088709e-01 1.94964930e-01
-1.21142280e+00 7.27450773e-02 4.16160762e-01 -4.83580306e-02
1.19421434e+00 6.37382925e-01 8.12697768e-01 2.67822295e-01
4.79247347e-02 8.69572043e-01 8.73088241e-01 -4.76245314e-01
5.53341866e-01 1.57930136e-01 2.01538920e-01 2.76619375e-01
1.11047554e+00 -2.61379004e-01 -3.85066145e-03 -1.67058349e-01
5.27061522e-01 1.93525404e-01 -3.18721771e-01 -7.22143233e-01
-6.85055077e-01 9.00764823e-01 2.66988903e-01 -7.96453655e-02
-5.09917498e-01 -4.58997846e-01 1.95656255e-01 1.80827752e-01
4.20191079e-01 7.23137856e-01 -5.64861178e-01 -1.39518797e-01
-1.08747470e+00 7.96381354e-01 1.29503334e+00 1.39686930e+00
2.25150168e-01 -1.41658783e-01 -1.24907523e-01 9.68543470e-01
-8.08406621e-03 3.20616513e-01 -3.36199105e-02 -3.35525036e-01
1.07897222e+00 5.51403284e-01 6.46039009e-01 -7.76924372e-01
-6.22091770e-01 -6.14305377e-01 -6.48517430e-01 -2.12532252e-01
3.46966326e-01 -1.44503564e-01 -7.75207162e-01 1.17194617e+00
3.66527736e-01 -1.16890639e-01 5.51727088e-03 9.09126997e-01
3.09910387e-01 8.64924014e-01 -1.98594913e-01 -9.10671592e-01
1.25570059e+00 -1.24128926e+00 -7.63704658e-01 -1.67765856e-01
5.47567785e-01 -8.34430099e-01 4.36224699e-01 6.60180271e-01
-1.62965858e+00 -1.91172361e-01 -1.15228355e+00 5.84537566e-01
-4.41121578e-01 -1.32834375e-01 1.02457607e+00 7.46080101e-01
-5.34956932e-01 4.50736135e-01 -5.45948446e-01 -2.48415887e-01
1.76901072e-01 7.78979301e-01 -2.97779381e-01 -5.62166393e-01
-8.18569183e-01 8.73897552e-01 7.03234196e-01 3.62117708e-01
2.82756332e-02 -5.94664574e-01 -8.84865940e-01 3.30558091e-01
9.10591006e-01 -4.89399701e-01 1.07266438e+00 -1.81435227e-01
-1.03929317e+00 2.06422806e-01 7.74685368e-02 -4.30103779e-01
1.33394048e-01 9.99969244e-02 -4.52152878e-01 -1.82275280e-01
-5.68295605e-02 4.73717414e-02 1.23250715e-01 -1.32448363e+00
-1.18248701e+00 -1.49380341e-01 4.24487263e-01 3.08015257e-01
-2.27720037e-01 -9.91682261e-02 -4.99831229e-01 -4.83613610e-01
1.27221376e-01 -7.60011077e-01 -6.01204276e-01 -7.07709849e-01
-3.69898140e-01 -1.02444522e-01 2.44709358e-01 -3.73571575e-01
1.76923573e+00 -1.82089925e+00 2.63694406e-01 4.93270636e-01
-3.03008795e-01 3.76532584e-01 8.29569846e-02 9.11903322e-01
-1.33797646e-01 4.82092388e-02 7.30736479e-02 -2.83430576e-01
4.19873387e-01 7.66943753e-01 -2.08011530e-02 3.79580140e-01
8.68602768e-02 4.16408062e-01 -6.49858415e-01 -2.25038812e-01
2.02402800e-01 -3.01326126e-01 -9.08549488e-01 1.83882505e-01
-2.47142211e-01 -3.09276015e-01 -3.14621449e-01 9.88124251e-01
1.10268033e+00 -1.34800291e-02 2.08315447e-01 -2.58629508e-02
-3.86683613e-01 3.01677119e-02 -1.78340840e+00 1.07324123e+00
-3.70131582e-01 -3.10044140e-01 3.28116566e-01 -1.45375741e+00
8.93103063e-01 -1.18299329e-03 5.96924067e-01 -7.33150244e-01
-5.86035475e-02 3.97939980e-01 2.00470462e-01 -5.12597501e-01
8.15922678e-01 -1.55261178e-02 -3.08154345e-01 1.73883826e-01
-4.52220827e-01 -7.73106068e-02 1.13997364e+00 -1.50610685e-01
8.72102082e-01 -3.15093488e-01 3.86836052e-01 -4.90493089e-01
2.13407457e-01 2.11782247e-01 9.20694828e-01 4.41347808e-01
3.34456354e-01 4.76481706e-01 6.56824410e-01 -3.70360345e-01
-9.93966877e-01 -7.42563069e-01 -3.70806366e-01 8.39881778e-01
5.84455967e-01 -1.90325156e-01 -5.30997336e-01 -3.27350050e-01
6.03733420e-01 6.88898742e-01 -3.00571024e-01 3.85447025e-01
-5.31116307e-01 -1.17630219e+00 -5.18708766e-01 5.78746617e-01
-7.01693352e-03 -6.36821806e-01 -3.86866897e-01 8.37388396e-01
1.90930352e-01 -1.17891228e+00 -2.28850394e-01 6.42619669e-01
-8.11975777e-01 -1.11548018e+00 -6.63076699e-01 -9.18758690e-01
9.14723277e-01 5.55421293e-01 1.01086450e+00 2.09914386e-01
-4.42691892e-01 -2.87480623e-01 -7.25252569e-01 -5.08189917e-01
2.13003740e-01 2.99155563e-01 1.57019183e-01 -3.26229364e-01
1.15729563e-01 -3.05608988e-01 -4.26447213e-01 8.41087282e-01
-1.05505681e+00 1.26231074e-01 6.28254533e-01 9.90693748e-01
7.95856833e-01 7.63849795e-01 7.39662051e-01 -9.55515563e-01
7.77372181e-01 -6.78843737e-01 -9.77807462e-01 4.40211743e-01
-5.54789007e-01 -1.65531605e-01 8.39259863e-01 -2.13743180e-01
-6.35715783e-01 -2.73431659e-01 -1.04627505e-01 1.69934649e-02
-2.45994687e-01 7.56667614e-01 -3.83716166e-01 9.37335193e-02
-2.43735656e-01 -1.56197295e-01 -2.49381498e-01 -6.40530646e-01
-3.37994546e-01 4.69716191e-01 1.47800952e-01 -8.95566881e-01
6.80495381e-01 -8.97903927e-03 1.37644947e-01 -5.51854849e-01
-9.48787093e-01 -7.23403096e-01 -2.94378251e-01 2.60823011e-01
2.44938567e-01 -4.54913616e-01 -8.50640953e-01 -5.47385812e-02
-7.44548559e-01 -8.88188407e-02 -2.82686114e-01 4.09806192e-01
-5.26637197e-01 2.56411642e-01 -3.86329591e-01 -8.91074479e-01
-1.31555766e-01 -1.15917897e+00 6.47976816e-01 3.36175442e-01
1.04910880e-01 -7.89891243e-01 -2.44780511e-01 4.85112667e-01
2.98235327e-01 5.28374970e-01 1.06526971e+00 -6.65165603e-01
-4.26907331e-01 -4.30271715e-01 -1.95901200e-01 -2.66950522e-02
1.63777843e-01 -6.30252510e-02 1.02968484e-01 -7.30030715e-01
-3.97423208e-01 -2.99848821e-02 2.64338523e-01 4.34377462e-01
1.23063040e+00 -2.51997262e-01 -5.61966062e-01 4.88833576e-01
1.62704611e+00 5.06807864e-01 4.21240956e-01 5.74082553e-01
8.29846188e-02 5.15146732e-01 1.38067627e+00 9.56719220e-01
3.85388345e-01 9.93944943e-01 7.15160310e-01 -3.23675543e-01
5.41011989e-01 9.03970003e-02 -3.08892041e-01 6.61031187e-01
-3.18540514e-01 -9.20979679e-01 -6.99458480e-01 5.98298013e-01
-2.10919166e+00 -9.11520600e-01 -1.91809461e-01 2.42528796e+00
4.64633554e-01 4.71650630e-01 3.69412571e-01 6.44005120e-01
6.40198231e-01 -3.41570199e-01 -2.41664648e-01 -1.15238404e+00
-1.07283425e-02 2.58764625e-01 9.18647408e-01 2.39335239e-01
-1.04456329e+00 5.09836733e-01 6.08326244e+00 7.70838320e-01
-4.95658070e-01 -3.43977660e-01 4.85114753e-01 -4.35048759e-01
-2.18065511e-02 -2.09206566e-01 -1.38311160e+00 7.60190606e-01
7.01915383e-01 -2.58332282e-01 5.28098047e-01 8.58594000e-01
3.09533596e-01 -5.74188530e-01 -1.23854017e+00 7.58404911e-01
2.44424343e-02 -1.28983426e+00 -3.07845861e-01 5.28646529e-01
9.83067513e-01 -7.62559354e-01 -1.26740471e-01 2.63056785e-01
2.10671723e-02 -8.56265485e-01 4.66967285e-01 9.40534472e-02
4.04419541e-01 -1.29930675e+00 1.14893711e+00 3.93143386e-01
-1.25727916e+00 -4.67529178e-01 -6.78995073e-01 -2.92452842e-01
7.48663604e-01 7.26072609e-01 -5.57096362e-01 9.56695020e-01
6.98301017e-01 1.16276734e-01 1.34026438e-01 1.99845266e+00
2.62167037e-01 2.57499814e-01 -6.33027971e-01 -2.21584067e-01
4.93903309e-01 -2.56330252e-01 1.04451522e-01 1.06538844e+00
3.68224978e-01 3.73369306e-01 7.51687288e-01 2.87581295e-01
3.07618946e-01 2.75800824e-01 -2.21478462e-01 7.70874321e-02
5.25065958e-01 1.11314559e+00 -8.74717355e-01 3.43138464e-02
-5.27158082e-01 4.72619444e-01 3.15532267e-01 2.29943424e-01
-1.03070641e+00 -4.65610087e-01 6.89889491e-01 2.75453985e-01
8.39285314e-01 -2.76365280e-01 -1.52499378e-01 -6.49943888e-01
2.43183017e-01 -6.18135035e-01 4.57639545e-01 -1.14943184e-01
-1.34484005e+00 3.13505411e-01 5.24798751e-01 -1.12503183e+00
-1.82085231e-01 -8.58623445e-01 -4.11958694e-01 8.07885170e-01
-1.80358958e+00 -5.54426134e-01 -6.09507561e-02 3.09464723e-01
7.30249763e-01 6.98580267e-03 6.77587509e-01 6.43190920e-01
-9.53104198e-01 7.08872497e-01 4.56245363e-01 -5.77961981e-01
1.80303037e-01 -1.23484552e+00 -5.87968640e-02 5.95689237e-01
-5.13507724e-01 4.32679653e-01 9.76548970e-01 -4.69669610e-01
-1.82378733e+00 -7.96842158e-01 6.73369765e-01 3.89520705e-01
6.75457954e-01 -5.38909912e-01 -6.88646674e-01 1.86355993e-01
8.70999042e-03 -9.49495658e-02 8.61603022e-01 2.28187010e-01
4.58075017e-01 -2.94646800e-01 -1.32356644e+00 2.67813116e-01
9.96049345e-01 7.18018055e-01 -3.50387394e-01 6.55889452e-01
4.67382729e-01 -8.25078607e-01 -1.02945518e+00 4.68668938e-01
3.62229437e-01 -8.08760583e-01 7.27090776e-01 -6.21214867e-01
1.77842498e-01 -4.92318124e-02 -2.41615355e-01 -1.36296773e+00
-4.69238728e-01 -5.33855975e-01 -1.96622655e-01 1.24004090e+00
7.17737079e-01 -6.25960827e-01 9.27114308e-01 1.04130745e+00
-3.30837786e-01 -1.43897426e+00 -8.00181627e-01 -1.17404997e+00
-2.88586795e-01 8.80285874e-02 1.12694514e+00 5.99843204e-01
1.98901877e-01 -2.05582142e-01 -4.14869368e-01 3.04226041e-01
4.81320471e-01 8.48711193e-01 7.27114916e-01 -1.10379231e+00
-6.67670369e-01 -3.42496365e-01 -2.82482773e-01 -1.10608399e+00
-2.09870458e-01 -5.55188000e-01 -2.24188417e-02 -2.01239586e+00
4.04750183e-02 -1.01775134e+00 -1.79883927e-01 4.29221749e-01
1.65771574e-01 -1.31453589e-01 2.41726175e-01 -2.49255925e-01
-7.54852772e-01 1.40235364e-01 1.26865780e+00 1.09752424e-01
-3.63725483e-01 4.62755382e-01 -7.64907479e-01 3.35904002e-01
8.24485958e-01 -2.26796135e-01 -4.08635676e-01 -3.99640352e-01
3.53437483e-01 4.38371420e-01 -4.37386721e-01 -6.88361108e-01
-2.26881765e-02 -6.75285220e-01 7.99817294e-02 -9.44706857e-01
3.83719414e-01 -1.35218847e+00 3.68862122e-01 3.13137919e-01
2.29039803e-01 5.59975922e-01 3.66521209e-01 4.57926303e-01
-1.14843346e-01 -6.99555159e-01 2.60804683e-01 1.42633468e-01
-5.49561679e-01 4.49803382e-01 2.97570117e-02 -5.55698276e-02
1.64059424e+00 -8.32149804e-01 -4.27116960e-01 1.88767299e-01
-6.67326868e-01 9.07845318e-01 3.89247715e-01 3.17784160e-01
3.81342441e-01 -1.06046474e+00 -4.90270108e-01 2.36528218e-01
-5.53228660e-03 3.16866100e-01 2.79289931e-01 7.98135519e-01
-6.86414421e-01 7.23768890e-01 -2.98441797e-01 -2.94791132e-01
-1.14795709e+00 9.35629189e-01 -3.40330809e-01 -7.50117958e-01
-3.67877454e-01 6.10753059e-01 -3.04784328e-01 2.08803639e-01
4.61887032e-01 -4.75018889e-01 -1.55327514e-01 1.06552482e-01
4.63234991e-01 4.64573890e-01 3.72641176e-01 -1.90336376e-01
-1.77514136e-01 2.24045798e-01 -3.87180954e-01 6.32215858e-01
1.71359110e+00 6.36001397e-03 -1.63765877e-01 -7.77583942e-02
7.20663369e-01 -5.24993613e-02 -7.91315019e-01 3.50396521e-02
-4.56609158e-03 -8.38957727e-01 -3.51193428e-01 -6.74753070e-01
-1.18694830e+00 2.67551005e-01 2.45681405e-02 9.16729271e-01
1.29903853e+00 -1.59820139e-01 8.27111423e-01 2.29308471e-01
8.64771485e-01 -1.29481077e+00 -4.84307408e-01 3.35082650e-01
7.99276352e-01 -9.56931174e-01 4.29017246e-01 -8.76734436e-01
-4.25374240e-01 9.08654392e-01 8.17578793e-01 -2.20572844e-01
4.21775371e-01 7.01970339e-01 -8.02622259e-01 1.31396979e-01
-8.90394628e-01 -2.99135387e-01 8.23519006e-02 2.57222027e-01
2.70013452e-01 4.47481006e-01 -9.13726091e-01 8.15521002e-01
-2.64912903e-01 -2.64626980e-01 4.89176482e-01 1.26200271e+00
-6.28495455e-01 -1.36467206e+00 -5.70132732e-01 8.88931513e-01
-4.71987635e-01 1.39095202e-01 3.62898886e-01 9.96617615e-01
2.62937218e-01 1.12375247e+00 1.95131838e-01 1.36552900e-01
7.56858468e-01 -4.55778569e-01 5.96008837e-01 -8.54087710e-01
-5.13494313e-01 1.60764366e-01 5.19808292e-01 -3.37999344e-01
-2.18825147e-01 -4.95110899e-01 -1.06810188e+00 -6.19358480e-01
-9.01300251e-01 6.49097323e-01 7.36196995e-01 9.11764979e-01
1.23112619e-01 5.04266024e-01 7.39282012e-01 -9.62034225e-01
-4.79524046e-01 -4.84572560e-01 -1.11360598e+00 2.20324546e-01
-1.05646387e-01 -1.18507326e+00 4.89442982e-02 -5.35053492e-01]
|
[5.091279029846191, 2.81113338470459]
|
2062f839-fbcd-4395-9ac2-d1f0ca46ab20
|
enhancing-the-spatial-resolution-of-stereo
| null | null |
http://openaccess.thecvf.com/content_cvpr_2018/html/Jeon_Enhancing_the_Spatial_CVPR_2018_paper.html
|
http://openaccess.thecvf.com/content_cvpr_2018/papers/Jeon_Enhancing_the_Spatial_CVPR_2018_paper.pdf
|
Enhancing the Spatial Resolution of Stereo Images Using a Parallax Prior
|
We present a novel method that can enhance the spatial resolution of stereo images using a parallax prior. While traditional stereo imaging has focused on estimating depth from stereo images, our method utilizes stereo images to enhance spatial resolution instead of estimating disparity. The critical challenge for enhancing spatial resolution from stereo images: how to register corresponding pixels with subpixel accuracy. Since disparity in traditional stereo imaging is calculated per pixel, it is directly inappropriate for enhancing spatial resolution. We, therefore, learn a parallax prior from stereo image datasets by jointly training two-stage networks. The first network learns how to enhance the spatial resolution of stereo images in luminance, and the second network learns how to reconstruct a high-resolution color image from high-resolution luminance and chrominance of the input image. Our two-stage joint network enhances the spatial resolution of stereo images significantly more than single-image super-resolution methods. The proposed method is directly applicable to any stereo depth imaging methods, enabling us to enhance the spatial resolution of stereo images.
|
['Seung-Hwan Baek', 'Min H. Kim', 'Inchang Choi', 'Daniel S. Jeon']
|
2018-06-01
| null | null | null |
cvpr-2018-6
|
['stereo-image-super-resolution']
|
['computer-vision']
|
[ 9.34372246e-01 -3.68201323e-02 -1.19354212e-04 -4.10835505e-01
-7.92108417e-01 -1.76868558e-01 1.97264016e-01 -5.74904442e-01
-8.04971576e-01 1.08332193e+00 2.94197530e-01 4.03435081e-02
1.86738595e-01 -1.04744434e+00 -8.59835804e-01 -7.80807793e-01
3.89249027e-01 -1.94674045e-01 6.43192530e-01 -2.37238668e-02
5.28424561e-01 4.99930650e-01 -1.85461843e+00 3.98703188e-01
8.85576069e-01 7.92472661e-01 8.12796772e-01 7.71034837e-01
1.03987595e-02 9.46190178e-01 -3.02445143e-01 2.62006223e-01
4.38307375e-01 -3.31568927e-01 -7.93555379e-01 -7.16778412e-02
9.14009094e-01 -1.05387914e+00 -4.79534000e-01 1.28578401e+00
3.94175500e-01 -1.01479024e-01 3.13547015e-01 -5.69746852e-01
-3.84245306e-01 2.81328082e-01 -9.37141895e-01 5.76487958e-01
3.49625498e-01 -1.44275412e-01 3.63634676e-01 -6.68012857e-01
5.74772537e-01 1.17866707e+00 3.17128360e-01 5.06907046e-01
-1.36879623e+00 -7.69183278e-01 -1.50466397e-01 8.65319371e-02
-1.43512845e+00 -4.74459469e-01 7.54822433e-01 -2.73333222e-01
6.62575126e-01 -2.32998058e-02 5.35764873e-01 5.32701552e-01
1.85647845e-01 3.95978600e-01 1.66944122e+00 -4.12788242e-01
-6.97143550e-05 -1.22091442e-01 -4.24548239e-01 3.66387248e-01
1.14965975e-01 7.44446039e-01 -8.39655876e-01 3.77586365e-01
1.97324657e+00 5.49725071e-02 -6.20928645e-01 -2.18798175e-01
-1.11421072e+00 2.61818171e-01 4.81401831e-01 3.48215699e-01
-2.55079478e-01 2.63547808e-01 -2.62281597e-01 -7.64607042e-02
3.66291374e-01 3.42806131e-01 -1.89364627e-01 -9.59547423e-03
-9.83643830e-01 -1.53547168e-01 -7.00148940e-02 7.56079733e-01
1.08528769e+00 3.40299271e-02 2.54119158e-01 7.33601391e-01
6.00910075e-02 3.92019004e-01 3.46264631e-01 -1.74257648e+00
4.43235427e-01 1.94940940e-01 2.68368542e-01 -6.24701738e-01
-1.35672212e-01 -1.05588652e-01 -9.94830549e-01 9.61185217e-01
3.84933084e-01 -7.17588514e-02 -7.49381363e-01 1.63701153e+00
2.45218799e-01 1.02145493e-01 3.17572281e-02 1.19292164e+00
6.06855929e-01 6.33083999e-01 -4.03017789e-01 -1.55467659e-01
9.01380658e-01 -5.57040036e-01 -7.28148460e-01 -4.23208356e-01
-1.75873384e-01 -7.48014092e-01 8.14735711e-01 3.71162653e-01
-1.73289084e+00 -8.24532032e-01 -1.23389530e+00 -6.53185129e-01
2.25857142e-02 -1.46012545e-01 4.16600883e-01 2.90937513e-01
-1.46596837e+00 5.38708091e-01 -7.01403141e-01 2.13480666e-01
2.92937607e-01 5.19553423e-01 -4.80809748e-01 -2.15729058e-01
-1.13994944e+00 7.86718488e-01 4.43811148e-01 -2.20890164e-01
-6.32649124e-01 -8.64963472e-01 -1.00348306e+00 -9.41192880e-02
-1.58631336e-03 -8.01811874e-01 1.01873207e+00 -1.04327643e+00
-1.58398902e+00 1.00153077e+00 -5.86306155e-01 -1.85227633e-01
4.38557804e-01 -1.19354270e-01 3.90713587e-02 4.77259457e-01
1.90031618e-01 1.11385441e+00 9.70913172e-01 -1.46021283e+00
-1.29086483e+00 -5.63240469e-01 2.39643827e-01 6.48614228e-01
9.01237950e-02 -2.77587980e-01 -6.76933467e-01 -2.91656405e-01
2.76637405e-01 -1.20882951e-01 -1.78554118e-01 3.13786685e-01
-1.13511741e-01 4.43017095e-01 6.08200312e-01 -5.48513234e-01
7.75533915e-01 -2.16996145e+00 3.69136482e-02 -2.25432456e-01
4.47090745e-01 -6.33806437e-02 -1.77824244e-01 -2.87740529e-01
-9.28813890e-02 -2.01793239e-01 -1.71718150e-01 -1.59428090e-01
-7.06539094e-01 7.54986107e-02 -2.30123639e-01 4.74912077e-01
1.88306104e-02 5.99488437e-01 -9.80127394e-01 -5.90139806e-01
7.33641088e-01 1.11986136e+00 -5.21667421e-01 1.58413962e-01
2.94369400e-01 9.31507349e-01 -1.43752754e-01 3.72729182e-01
1.16715658e+00 -1.78250611e-01 3.14753205e-02 -3.91654581e-01
-6.54851079e-01 2.02143535e-01 -1.07143271e+00 1.79193246e+00
-6.58693969e-01 9.35904503e-01 1.75408632e-01 -3.44837934e-01
7.10505486e-01 -1.95098780e-02 4.97897714e-01 -1.44473302e+00
-5.40660545e-02 1.49216115e-01 -4.36892599e-01 -1.24327600e-01
6.35710120e-01 -3.03316027e-01 4.14082766e-01 2.95188814e-01
-3.37823689e-01 -4.54061240e-01 -3.83565053e-02 -2.20968187e-01
5.68545401e-01 1.86989367e-01 3.11302602e-01 -6.61278740e-02
5.64994335e-01 -4.19780642e-01 5.50259292e-01 5.38939595e-01
6.91100881e-02 1.09248221e+00 9.66588557e-02 -4.34651256e-01
-1.53645992e+00 -1.40580928e+00 -4.43762988e-01 5.76529384e-01
6.73743069e-01 1.55185670e-01 -6.77136600e-01 2.85050496e-02
-2.59276330e-01 1.67716041e-01 -5.37087560e-01 2.98375010e-01
-6.83107197e-01 -2.95625061e-01 1.57157332e-01 6.86548769e-01
1.24551320e+00 -6.33971810e-01 -7.88827658e-01 5.03462367e-03
-5.37874401e-01 -1.35556328e+00 -5.62591016e-01 1.40356660e-01
-1.10982347e+00 -1.08741832e+00 -9.86806512e-01 -8.88170004e-01
8.50803912e-01 7.48795331e-01 1.03345096e+00 -1.85964078e-01
-2.94909984e-01 -5.11905812e-02 2.55967259e-01 5.52520715e-02
-5.39674088e-02 -2.95003980e-01 -1.74586102e-01 -1.77906647e-01
1.07125223e-01 -9.41927910e-01 -1.01912212e+00 3.86321247e-01
-1.07923651e+00 7.09015846e-01 6.53007507e-01 7.53689587e-01
8.40218782e-01 4.40297663e-01 -1.52327716e-01 -5.82015991e-01
5.55547811e-02 2.51883864e-01 -1.07127750e+00 -2.56311089e-01
-3.77359122e-01 1.12194695e-01 4.11860079e-01 -8.13468769e-02
-1.50534689e+00 1.52765989e-01 -9.22731534e-02 -1.31877467e-01
-2.90109098e-01 -8.71790051e-02 -7.11080581e-02 -3.36394936e-01
6.09951794e-01 4.02441651e-01 1.14431627e-01 -2.22870246e-01
5.64150214e-02 4.98595059e-01 9.46287990e-01 -2.59864509e-01
6.08038783e-01 1.20846164e+00 1.94704548e-01 -8.44222188e-01
-8.15431535e-01 -4.55803990e-01 -9.15761471e-01 -7.01304972e-02
1.10364997e+00 -1.34088373e+00 -7.93344438e-01 6.14778936e-01
-1.05804920e+00 -5.24921715e-01 -1.13013096e-01 6.02971554e-01
-6.46805704e-01 5.20767152e-01 -7.55015075e-01 -3.78706068e-01
1.69802830e-01 -1.18870378e+00 1.32867134e+00 4.00913268e-01
2.07617342e-01 -9.68059897e-01 2.77029071e-02 6.01392627e-01
4.35374916e-01 5.76074095e-03 7.24391282e-01 9.22410786e-01
-1.11516571e+00 3.28794867e-01 -8.74081194e-01 4.24605280e-01
3.60916972e-01 -3.48749012e-01 -1.25283921e+00 -1.25055030e-01
1.94707140e-01 -1.98240623e-01 8.36755037e-01 9.71301973e-01
1.39456403e+00 7.42164254e-02 6.57342467e-03 1.17086542e+00
1.85873413e+00 1.57241061e-01 1.17112589e+00 5.03696978e-01
6.57742620e-01 5.72702885e-01 5.12435079e-01 1.15549996e-01
4.16328043e-01 8.40340555e-01 3.21091980e-01 -6.19227290e-01
-5.78807652e-01 -2.64797777e-01 1.87820926e-01 2.90284842e-01
-4.20772642e-01 3.96883726e-01 -6.00431621e-01 3.72920156e-01
-1.32858443e+00 -9.86374497e-01 9.54731777e-02 2.36104655e+00
1.16582882e+00 -1.16824873e-01 -3.00143510e-01 1.29085332e-01
8.53420079e-01 1.69252694e-01 -7.28619397e-01 -1.04129314e-03
-4.98867810e-01 1.99240372e-01 8.21566105e-01 1.08441341e+00
-8.69732738e-01 8.40574503e-01 7.23368216e+00 6.40729427e-01
-1.23167002e+00 -1.24949299e-01 6.87343001e-01 -3.87102395e-01
-3.69150430e-01 -2.24233642e-01 -7.96202719e-01 3.83203804e-01
3.02619457e-01 -1.32331669e-01 7.06098080e-01 3.83876085e-01
4.82228011e-01 -6.68399751e-01 -1.11347437e+00 1.55084348e+00
6.97367862e-02 -1.27798522e+00 1.76734477e-01 1.88997537e-01
1.21392345e+00 -1.66025534e-02 4.51833755e-01 -4.00362790e-01
2.02503249e-01 -1.27948153e+00 2.36087412e-01 3.65445763e-01
1.32554340e+00 -8.55277717e-01 5.03208220e-01 2.83573329e-01
-1.11615407e+00 -2.70354562e-03 -5.34066439e-01 -2.17641383e-01
2.56318331e-01 5.97038865e-01 -3.11268240e-01 4.54286002e-02
1.03130066e+00 7.36142457e-01 -2.58382648e-01 8.75649035e-01
-2.90475607e-01 -4.20755416e-01 -8.64928588e-02 7.32923925e-01
-7.03083202e-02 -2.57489175e-01 1.09511971e-01 7.73612976e-01
2.95068771e-01 2.36112386e-01 -3.04875553e-01 1.02036572e+00
-5.16646989e-02 -4.45107818e-01 -7.52007902e-01 4.66938764e-01
5.66362917e-01 9.36439693e-01 -2.29427144e-01 -3.02116960e-01
-5.80324948e-01 1.15687943e+00 1.89616591e-01 7.23406672e-01
-5.07751405e-01 -2.19049796e-01 6.88422799e-01 4.27337259e-01
1.01750873e-01 -1.47612929e-01 -6.17177248e-01 -1.11847532e+00
-1.35071725e-02 -6.82460487e-01 -5.48011027e-02 -1.41775513e+00
-6.80802584e-01 5.20595074e-01 -1.51501611e-01 -1.31185436e+00
-3.25595319e-01 -5.88295639e-01 -4.46604006e-02 1.31139743e+00
-2.24560833e+00 -8.24281752e-01 -6.57741785e-01 8.83271337e-01
5.39509118e-01 3.64930540e-01 3.84002835e-01 2.51509577e-01
7.64793828e-02 7.85991549e-02 1.24590658e-01 3.78363729e-02
8.47070634e-01 -1.31173408e+00 1.75567821e-01 8.35612953e-01
-4.88772362e-01 5.13220012e-01 4.20856237e-01 -4.02369857e-01
-9.89634931e-01 -8.17270815e-01 6.91804111e-01 -2.45253846e-01
2.84983069e-02 -6.59670606e-02 -7.57381499e-01 3.38520914e-01
2.22326592e-01 -7.85544887e-02 2.57940650e-01 -3.06624442e-01
-2.69849896e-01 -3.50931883e-01 -1.16287184e+00 6.88453674e-01
9.41656768e-01 -1.01694787e+00 -4.20535773e-01 -3.22281480e-01
4.82297122e-01 -6.82405710e-01 -7.31824994e-01 5.02560019e-01
7.34203279e-01 -1.65048456e+00 1.31567323e+00 2.91877270e-01
8.27915847e-01 -6.53341711e-01 -1.58719003e-01 -1.12709391e+00
-3.43074918e-01 -1.35699362e-01 4.23960745e-01 6.68310404e-01
1.61297783e-01 -5.57344377e-01 9.27304208e-01 5.02795994e-01
2.40827560e-01 -2.47514918e-01 -9.32280898e-01 -3.57152373e-01
-3.28720249e-02 3.74964960e-02 3.47873151e-01 7.62914836e-01
-2.40967244e-01 1.06831670e-01 -2.79406667e-01 4.13314283e-01
1.22767746e+00 2.13635579e-01 6.30685687e-01 -9.28717613e-01
-2.83122748e-01 -4.76478726e-01 -2.90039361e-01 -1.56835544e+00
-1.30110353e-01 -2.84722865e-01 7.77177885e-02 -1.45161974e+00
6.71110272e-01 2.77228244e-02 -5.12300804e-02 -9.61040407e-02
-1.82438612e-01 7.41892695e-01 -9.15958583e-02 1.95346236e-01
-1.49272710e-01 1.93993106e-01 1.84298611e+00 1.08854793e-01
-2.20944196e-01 -3.55962336e-01 -6.33954525e-01 8.49410713e-01
6.56404257e-01 8.85415077e-03 -6.10976815e-01 -9.18350816e-01
3.74353528e-01 4.29718554e-01 6.06141865e-01 -9.34435129e-01
4.66797769e-01 -1.73945308e-01 9.35501873e-01 -7.94937432e-01
5.58426678e-01 -7.71338344e-01 7.61519326e-03 1.29772946e-01
-3.07588726e-01 -2.45072752e-01 1.60346389e-01 3.12507927e-01
-5.46314716e-01 1.19114913e-01 1.29378903e+00 -3.89941841e-01
-1.07198107e+00 3.54624510e-01 -3.68788242e-01 -1.34390816e-01
5.64850628e-01 -7.09129035e-01 -4.61745739e-01 -4.53233689e-01
-5.11147022e-01 -8.83657262e-02 1.00542212e+00 6.53912053e-02
1.19532716e+00 -1.25120962e+00 -4.47402924e-01 4.86566693e-01
-1.54518008e-01 3.57673883e-01 5.59313953e-01 5.61888814e-01
-8.47545564e-01 4.04297024e-01 -7.24756777e-01 -8.72585773e-01
-1.31816435e+00 3.58143717e-01 7.15611756e-01 9.80080962e-02
-6.41395628e-01 6.24284089e-01 9.31428671e-01 9.04023200e-02
2.87011385e-01 -3.20618123e-01 -2.36932904e-01 -6.28398955e-01
9.77340221e-01 2.95946568e-01 -4.33534652e-01 -4.04404283e-01
5.09216189e-02 1.31185019e+00 -1.57769769e-02 -5.77271521e-01
1.22676361e+00 -8.32474232e-01 -2.11193174e-01 4.55131441e-01
1.17560863e+00 3.42857204e-02 -1.88612521e+00 -2.90190458e-01
-7.38461971e-01 -1.05784047e+00 5.69824696e-01 -5.73225796e-01
-1.07826233e+00 1.18140996e+00 7.49999106e-01 -3.17154109e-01
1.59440219e+00 -3.35115463e-01 6.15463793e-01 4.63878028e-02
6.77639365e-01 -1.23260069e+00 2.19696984e-01 4.42074090e-01
6.59331024e-01 -1.50711167e+00 -4.99253646e-02 -5.77094495e-01
-2.95389414e-01 1.23493910e+00 7.83156395e-01 7.69488439e-02
3.91207665e-01 7.47677505e-01 2.03368321e-01 1.24895282e-01
-4.14974958e-01 -4.56942528e-01 1.31729215e-01 8.34293902e-01
5.14393985e-01 -3.96465212e-01 7.12246001e-02 -3.00207525e-01
-9.94956773e-03 2.78575212e-01 5.88113844e-01 5.75560451e-01
-6.56887233e-01 -9.15986836e-01 -4.89152759e-01 -1.35178477e-01
-4.82192099e-01 -2.96594292e-01 -8.90676901e-02 6.34968162e-01
2.35187098e-01 7.55453229e-01 4.46397513e-01 -1.26553968e-01
8.72012135e-03 -7.17258394e-01 9.04096127e-01 -4.98358548e-01
2.34360605e-01 2.09355652e-01 -2.47535363e-01 -9.01900053e-01
-7.06099272e-01 -3.42299402e-01 -1.22715056e+00 -3.33587021e-01
1.68623358e-01 -3.06979001e-01 7.11663187e-01 5.90307057e-01
1.28860697e-01 3.80843222e-01 6.86488152e-01 -1.07103503e+00
3.38717192e-01 -5.43868124e-01 -8.87279451e-01 1.42894283e-01
8.42443109e-01 -3.53277892e-01 -4.79057342e-01 1.88608810e-01]
|
[10.008146286010742, -2.398576498031616]
|
b31ceddc-8fff-4d67-a159-5861c08e5c74
|
learning-normal-form-autoencoders-for-data
|
2106.05102
| null |
https://arxiv.org/abs/2106.05102v1
|
https://arxiv.org/pdf/2106.05102v1.pdf
|
Learning normal form autoencoders for data-driven discovery of universal,parameter-dependent governing equations
|
Complex systems manifest a small number of instabilities and bifurcations that are canonical in nature, resulting in universal pattern forming characteristics as a function of some parametric dependence. Such parametric instabilities are mathematically characterized by their universal un-foldings, or normal form dynamics, whereby a parsimonious model can be used to represent the dynamics. Although center manifold theory guarantees the existence of such low-dimensional normal forms, finding them has remained a long standing challenge. In this work, we introduce deep learning autoencoders to discover coordinate transformations that capture the underlying parametric dependence of a dynamical system in terms of its canonical normal form, allowing for a simple representation of the parametric dependence and bifurcation structure. The autoencoder constrains the latent variable to adhere to a given normal form, thus allowing it to learn the appropriate coordinate transformation. We demonstrate the method on a number of example problems, showing that it can capture a diverse set of normal forms associated with Hopf, pitchfork, transcritical and/or saddle node bifurcations. This method shows how normal forms can be leveraged as canonical and universal building blocks in deep learning approaches for model discovery and reduced-order modeling.
|
['J. Nathan Kutz', 'Christoph Brune', 'Hil G. E. Meijer', 'Steven L. Brunton', 'Manu Kalia']
|
2021-06-09
| null | null | null | null |
['model-discovery']
|
['miscellaneous']
|
[-4.48745996e-01 6.28315564e-03 -3.39096576e-01 6.07203096e-02
1.23272121e-01 -9.23465490e-01 1.02813387e+00 -2.43472531e-01
4.49863493e-01 3.34280074e-01 3.15825135e-01 -4.66188818e-01
-4.11454409e-01 -4.67241108e-01 -7.52214849e-01 -1.08121407e+00
-5.84475040e-01 7.28122711e-01 -2.07177460e-01 -7.40313709e-01
-6.32977579e-03 9.04229462e-01 -1.35693717e+00 -2.85907060e-01
5.12530684e-01 6.43540263e-01 -2.90024251e-01 6.93563819e-01
6.30489886e-02 3.27444851e-01 -4.37971324e-01 1.11469030e-01
4.89394277e-01 -5.50723255e-01 -7.25624025e-01 -4.18710858e-02
2.34641895e-01 -5.21640927e-02 -5.58089137e-01 7.89483428e-01
1.10757284e-01 1.19467750e-01 1.27341557e+00 -1.41600978e+00
-6.14192963e-01 1.98403671e-01 -6.29727393e-02 2.66752601e-01
-7.46828504e-03 3.21052849e-01 1.46598291e+00 -4.55957681e-01
6.91397190e-01 1.34101200e+00 8.01008642e-01 5.54421127e-01
-1.75728369e+00 -4.75278348e-01 -3.23859274e-01 2.24750750e-02
-1.20854139e+00 -3.74995053e-01 1.07298040e+00 -9.69276309e-01
7.20617354e-01 9.98657569e-02 1.19236124e+00 1.06254530e+00
7.17352986e-01 3.48111331e-01 8.35957587e-01 -3.84370357e-01
2.66296923e-01 -2.95711011e-01 1.42957300e-01 7.88174272e-01
1.56720325e-01 3.50019455e-01 -1.05293177e-01 -2.49669373e-01
1.26380885e+00 -8.76039192e-02 -3.58980060e-01 -7.07604766e-01
-1.10248232e+00 1.03518558e+00 5.45539975e-01 4.61390793e-01
-1.50273681e-01 2.03182980e-01 7.51825646e-02 5.11947155e-01
-1.38859555e-01 9.57986474e-01 -5.80605209e-01 -1.05080500e-01
-4.77586001e-01 3.81094724e-01 1.02044892e+00 5.58813572e-01
1.09954357e+00 2.40520105e-01 3.87597650e-01 3.76875162e-01
1.96377963e-01 1.93934962e-01 7.90071487e-01 -1.14713132e+00
-4.87745255e-01 8.81854892e-01 -3.96733992e-02 -1.03845441e+00
-7.21885741e-01 -4.55131799e-01 -1.17900288e+00 2.99821258e-01
5.11066020e-01 4.69185226e-02 -8.23955894e-01 2.06428194e+00
2.35065326e-01 1.34207845e-01 8.83962587e-02 6.43951595e-01
2.72660494e-01 9.50598240e-01 -4.00247812e-01 -9.88402441e-02
1.23081136e+00 -3.14033389e-01 -6.09029055e-01 2.11146474e-01
6.18025064e-01 -4.17551607e-01 1.22803450e+00 -5.95160015e-02
-8.37654889e-01 -1.37195453e-01 -1.08200526e+00 5.16469143e-02
-3.14817846e-01 -4.13270235e-01 5.88276565e-01 1.63365528e-01
-1.27214551e+00 8.92294765e-01 -1.14831078e+00 -4.60022390e-01
-1.27999201e-01 3.52424741e-01 -5.08090556e-01 7.52577722e-01
-1.23487723e+00 8.97215605e-01 1.81831166e-01 -1.61695734e-01
-8.88454676e-01 -1.04044759e+00 -6.23470545e-01 1.35837272e-01
-3.90440255e-01 -8.81313860e-01 1.23080158e+00 -8.09197068e-01
-1.51031148e+00 7.30693698e-01 -1.70895755e-01 -3.87349695e-01
1.23287119e-01 1.96887836e-01 -2.64304787e-01 3.08758896e-02
-1.76064432e-01 6.62656903e-01 1.22185278e+00 -1.01813805e+00
1.80555314e-01 -6.60429001e-02 7.73832351e-02 -2.21088212e-02
-4.61682498e-01 -2.94881642e-01 2.45164767e-01 -5.28632522e-01
3.98890525e-01 -1.09000313e+00 -2.73071498e-01 1.04196660e-01
-5.65741837e-01 -4.14693266e-01 1.07993793e+00 -2.25647137e-01
1.28134632e+00 -1.84974253e+00 7.46848166e-01 2.15996325e-01
6.10167563e-01 -1.23422883e-01 2.04262272e-01 7.81722486e-01
-4.79450583e-01 5.00156939e-01 -3.63512427e-01 5.60955927e-02
1.06059119e-01 5.40351868e-01 -5.50803900e-01 6.49199605e-01
4.16812509e-01 8.51044595e-01 -4.56109703e-01 2.33571120e-02
2.53019538e-02 5.75282037e-01 -6.98013306e-01 3.67725998e-01
-1.93301290e-01 7.31499493e-01 -1.56806424e-01 1.91299707e-01
1.53535530e-01 -2.77509093e-01 -2.09189445e-01 -2.49191090e-01
-1.23281851e-01 1.82466075e-01 -9.69313920e-01 9.65071857e-01
-3.35554779e-01 9.65763032e-01 2.16358751e-01 -1.29994881e+00
9.50165927e-01 2.07733035e-01 7.12566257e-01 2.85449736e-02
2.31708676e-01 1.17751725e-01 5.16901851e-01 -4.54560727e-01
-1.44579494e-02 -4.38875705e-01 -2.02429250e-01 5.56227267e-01
4.01031822e-01 -2.67694116e-01 1.02769442e-01 9.50712711e-02
9.32993114e-01 -4.05078828e-01 4.34098035e-01 -9.25729752e-01
3.99264276e-01 -1.28038362e-01 5.12109280e-01 3.99634182e-01
-1.62884980e-01 3.52509409e-01 9.51279223e-01 -1.00962186e+00
-1.56699264e+00 -1.11575842e+00 -5.39654315e-01 5.41246176e-01
-6.96896762e-02 -5.44663191e-01 -7.58245945e-01 2.78903425e-01
-5.51457331e-02 1.29144996e-01 -9.71964061e-01 -7.55138993e-01
-8.15057516e-01 -6.32690966e-01 6.57922029e-01 1.77779883e-01
5.25981076e-02 -1.00165737e+00 -8.29599440e-01 6.99033029e-03
6.20319359e-02 -7.87805915e-01 -3.95762831e-01 4.11070526e-01
-1.00100839e+00 -1.25802243e+00 -4.39504176e-01 -6.86521053e-01
4.85022843e-01 -3.11823726e-01 1.07212579e+00 4.71293926e-03
-5.41249871e-01 2.65284777e-01 1.46735460e-01 1.43233106e-01
-9.38684225e-01 1.64911434e-01 7.58794129e-01 -6.85541779e-02
-6.97776151e-04 -1.15361333e+00 -4.73882735e-01 2.08595425e-01
-9.57382202e-01 -8.54383633e-02 2.70241797e-01 9.66846704e-01
3.13771844e-01 -9.88468304e-02 4.18408751e-01 -2.10699514e-02
8.80683661e-01 -5.92643499e-01 -6.02176487e-01 -1.99786663e-01
-4.61038381e-01 7.30306387e-01 9.57461953e-01 -6.18227601e-01
-2.75721669e-01 5.97358309e-02 -3.66034247e-02 -7.54674375e-01
-3.36959243e-01 4.81659055e-01 -4.85910624e-02 -4.88330126e-02
8.57262850e-01 4.36263502e-01 4.99508828e-01 -5.00317693e-01
5.17546356e-01 2.31967747e-01 6.13184631e-01 -7.10104108e-01
1.07953727e+00 3.68869185e-01 5.37228405e-01 -1.13797760e+00
-3.22131932e-01 -3.11197210e-02 -1.00146878e+00 -2.20038503e-01
8.38358402e-01 -4.95318323e-01 -9.19720888e-01 3.76024902e-01
-1.05011594e+00 -1.85064256e-01 -5.41546583e-01 2.39365827e-02
-8.84120762e-01 7.85202459e-02 -5.97937346e-01 -5.14267981e-01
-2.15893183e-02 -1.23670936e+00 7.42573321e-01 1.36388078e-01
-4.99385655e-01 -1.28113735e+00 5.24163723e-01 -6.68817163e-01
5.55685759e-01 5.94743550e-01 1.54888999e+00 -5.74474096e-01
-3.94289047e-01 -3.08133870e-01 5.45652449e-01 -1.39372339e-02
1.67500556e-01 5.29579222e-01 -5.83944798e-01 -2.90039986e-01
1.28312176e-02 1.88866496e-01 6.11450016e-01 5.39810479e-01
6.38516843e-01 -7.76156604e-01 -1.53887480e-01 1.04093337e+00
9.43447232e-01 1.49413213e-01 9.22197998e-02 1.24454387e-01
6.19540274e-01 4.49271768e-01 -8.01013291e-01 2.16847867e-01
2.07286343e-01 5.72524190e-01 4.05337572e-01 1.59910932e-01
2.58451283e-01 -2.41080418e-01 5.08519232e-01 1.15667212e+00
1.16489328e-01 5.88636458e-01 -1.21793628e+00 3.96119207e-01
-1.50564754e+00 -9.16289210e-01 6.49603456e-02 1.94808698e+00
9.13583994e-01 -4.92852144e-02 4.89662111e-01 1.11952916e-01
6.43208325e-01 1.29536048e-01 -9.03850436e-01 -5.92125595e-01
-3.84814888e-01 2.83037387e-02 -1.07934894e-02 6.19730175e-01
-1.06358349e+00 8.81933212e-01 7.41045809e+00 2.51947969e-01
-1.51977563e+00 -2.59010136e-01 3.13403308e-01 3.90338480e-01
-2.97329068e-01 2.61450022e-01 -5.73600531e-01 2.01259613e-01
1.13593805e+00 -5.87592900e-01 6.62940562e-01 6.71079516e-01
2.35441387e-01 5.99852622e-01 -1.20524073e+00 8.23216081e-01
-4.76316214e-01 -1.64579773e+00 2.23770276e-01 2.42527649e-01
7.19749331e-01 -1.47090778e-01 2.21131012e-01 1.16820425e-01
3.33254457e-01 -1.18829107e+00 4.64410245e-01 5.68583548e-01
7.73590684e-01 -5.46710551e-01 -6.91670030e-02 4.69199687e-01
-1.07420683e+00 -3.81199062e-01 -2.85202026e-01 -3.59100938e-01
-1.00618294e-04 3.46265852e-01 -5.44660330e-01 -9.39430445e-02
5.67740738e-01 9.23006713e-01 -3.61234218e-01 8.49028647e-01
3.45119908e-02 7.77098715e-01 -7.32560694e-01 6.52305931e-02
2.75044560e-01 -5.60833454e-01 1.03074217e+00 9.15692031e-01
2.76006401e-01 2.29020603e-02 -3.67980972e-02 1.31906235e+00
-4.40916903e-02 -2.06556261e-01 -9.17943597e-01 -2.45252654e-01
4.76821482e-01 1.20010245e+00 -6.20977640e-01 1.02926120e-02
1.40865758e-01 3.21142524e-01 1.76917359e-01 5.52219748e-01
-3.78251404e-01 -2.61412650e-01 1.38221014e+00 4.02922183e-01
1.60486087e-01 -6.49147749e-01 -2.16609538e-01 -1.47788024e+00
-1.94655791e-01 -8.06272984e-01 -9.55434982e-03 -4.26745355e-01
-1.14237344e+00 6.64115667e-01 2.48815730e-01 -1.30848169e+00
-8.89619648e-01 -6.72913015e-01 -1.16295278e+00 7.10006654e-01
-7.77366579e-01 -9.75572705e-01 -3.46658193e-02 7.59323359e-01
7.31185377e-02 -3.71342123e-01 1.04574430e+00 -2.07703963e-01
-6.90495431e-01 3.51527452e-01 4.00112391e-01 2.02931613e-02
1.94168448e-01 -1.56734550e+00 3.97538185e-01 6.27316892e-01
1.99737266e-01 9.60390449e-01 1.22440290e+00 -1.19022876e-01
-1.62026739e+00 -7.96985149e-01 4.54709649e-01 -4.19664085e-01
1.17908382e+00 -4.43896800e-01 -1.02860463e+00 7.79177964e-01
1.00991234e-01 4.53764275e-02 4.48025078e-01 -4.24091741e-02
-4.58141595e-01 -7.27943778e-02 -6.74264729e-01 9.26309943e-01
7.66845047e-01 -6.37820601e-01 -6.39962256e-01 3.18433225e-01
7.44897902e-01 5.45015670e-02 -8.41109216e-01 2.34983549e-01
6.69798374e-01 -8.60192120e-01 9.29245651e-01 -1.12669325e+00
5.94475448e-01 -2.20161930e-01 1.03896588e-01 -1.62971282e+00
-5.64949811e-01 -1.29922390e+00 -6.86478555e-01 6.41924441e-01
2.39779368e-01 -6.45389020e-01 4.10322130e-01 3.33473951e-01
-2.15879008e-01 -9.23432529e-01 -1.27027893e+00 -7.71996856e-01
1.02603459e+00 -6.83477521e-03 5.90891540e-01 1.03511596e+00
1.99093029e-01 3.10281038e-01 -1.45390451e-01 6.02841116e-02
3.99177343e-01 5.95079102e-02 7.62786865e-01 -1.76620936e+00
-1.75717026e-01 -1.02989769e+00 -7.78229296e-01 -1.01390231e+00
5.22399127e-01 -9.76959884e-01 -2.42516458e-01 -7.85597205e-01
-3.00729036e-01 -7.24673718e-02 -3.48795541e-02 2.17001125e-01
3.66359740e-01 -2.09782511e-01 1.96191426e-02 7.64234543e-01
2.58590460e-01 7.44566500e-01 1.03977406e+00 8.74065235e-02
-4.49734598e-01 -1.62304677e-02 -6.37045383e-01 8.16209674e-01
8.58362317e-01 -1.54205099e-01 -1.56549275e-01 4.00396064e-02
3.43982428e-01 6.91493973e-02 4.90223914e-01 -1.04463089e+00
2.91376442e-01 -5.89904934e-02 1.32297024e-01 -1.28560036e-01
2.85302579e-01 -7.02368915e-01 1.67779863e-01 6.17931187e-01
-4.82598275e-01 6.52297318e-01 1.55652389e-01 5.39653420e-01
-1.46972999e-01 1.47967726e-01 1.01965463e+00 1.19116090e-01
-2.02387974e-01 3.24173748e-01 -8.09332073e-01 9.48147774e-02
8.15185428e-01 -1.55943543e-01 3.36894132e-02 -7.30796456e-01
-9.18534219e-01 7.32999109e-03 5.56803882e-01 4.05678332e-01
1.48604035e-01 -1.44419611e+00 -4.15544957e-01 6.42544866e-01
-8.37105662e-02 -1.19080774e-01 -1.68792456e-01 7.94486105e-01
-5.87708294e-01 4.08910751e-01 -6.14753723e-01 -1.03554261e+00
-6.02805197e-01 4.49871391e-01 1.07032347e+00 -9.88093913e-02
-8.05425763e-01 3.38502973e-01 2.88272917e-01 -5.95717609e-01
-1.12605378e-01 -7.07954526e-01 -5.72989434e-02 7.90661480e-03
4.84920666e-02 1.96234435e-02 -2.66525894e-01 -1.06259489e+00
-1.46803468e-01 9.30784702e-01 3.67191494e-01 -6.91716745e-02
1.20151603e+00 5.59179299e-02 -3.83783400e-01 7.09684193e-01
1.45971894e+00 -4.50001746e-01 -1.32278728e+00 1.83377732e-02
4.30329107e-02 2.77939528e-01 -1.51135311e-01 -2.02265844e-01
-8.74001861e-01 1.07892847e+00 2.76977926e-01 6.85229599e-01
8.68259966e-01 2.25943103e-01 5.97222567e-01 7.69048095e-01
-1.92903817e-01 -3.84219617e-01 -9.57845300e-02 8.94126773e-01
1.12159777e+00 -7.10515261e-01 -5.57326734e-01 2.15765625e-01
-4.76391427e-02 1.66760731e+00 2.51400679e-01 -8.73296201e-01
1.29303825e+00 3.96544546e-01 -6.55352324e-02 -1.75301388e-01
-9.81563449e-01 2.37433121e-01 4.48865354e-01 5.60402811e-01
2.38063514e-01 2.19096512e-01 -1.23250792e-02 4.83120173e-01
-8.43625188e-01 -7.51556277e-01 8.09637785e-01 3.66083562e-01
-5.24670839e-01 -8.34568143e-01 -2.42230684e-01 3.50628197e-01
-1.87272429e-01 2.30939642e-01 -5.45237005e-01 1.05762780e+00
-2.71057546e-01 3.50626707e-01 3.37076187e-01 -4.98729885e-01
7.61400387e-02 4.36709970e-01 2.07811981e-01 -1.30707309e-01
-2.72383720e-01 8.96300748e-02 -6.61965013e-01 -4.18368697e-01
-1.21061399e-03 -7.50637710e-01 -9.94505525e-01 -4.89584178e-01
5.88398054e-02 1.84985593e-01 3.19558889e-01 1.04057193e+00
5.28959930e-01 2.02878192e-01 7.01928616e-01 -1.10654104e+00
-8.36149931e-01 -1.01876092e+00 -4.50896829e-01 6.31908774e-01
1.15708506e+00 -9.37063754e-01 -8.20919454e-01 3.27984244e-01]
|
[6.585069179534912, 3.6192548274993896]
|
70e48a75-bfff-4be0-adcd-beba8d793777
|
why-does-my-medical-ai-look-at-pictures-of
|
2306.17555
| null |
https://arxiv.org/abs/2306.17555v1
|
https://arxiv.org/pdf/2306.17555v1.pdf
|
Why does my medical AI look at pictures of birds? Exploring the efficacy of transfer learning across domain boundaries
|
It is an open secret that ImageNet is treated as the panacea of pretraining. Particularly in medical machine learning, models not trained from scratch are often finetuned based on ImageNet-pretrained models. We posit that pretraining on data from the domain of the downstream task should almost always be preferred instead. We leverage RadNet-12M, a dataset containing more than 12 million computed tomography (CT) image slices, to explore the efficacy of self-supervised pretraining on medical and natural images. Our experiments cover intra- and cross-domain transfer scenarios, varying data scales, finetuning vs. linear evaluation, and feature space analysis. We observe that intra-domain transfer compares favorably to cross-domain transfer, achieving comparable or improved performance (0.44% - 2.07% performance increase using RadNet pretraining, depending on the experiment) and demonstrate the existence of a domain boundary-related generalization gap and domain-specific learned features.
|
['Jens Kleesiek', 'Jan Egger', 'Constantin Seibold', 'Michael Kamp', 'Felix Nensa', 'René Hosch', 'Johannes Haubold', 'Janis Evers', 'Enrico Nasca', 'Moon Kim', 'Frederic Jonske']
|
2023-06-30
| null | null | null | null |
['computed-tomography-ct', 'transfer-learning']
|
['methodology', 'miscellaneous']
|
[ 4.91807699e-01 4.89374548e-01 -3.62927884e-01 -5.68927526e-01
-1.06116462e+00 -6.63137674e-01 6.52741075e-01 2.13665783e-01
-9.30080175e-01 7.19541311e-01 4.58196253e-01 -3.84156585e-01
-1.03490464e-01 -5.32664895e-01 -6.53899968e-01 -4.62653935e-01
-1.86190769e-01 6.25994325e-01 3.31134260e-01 -2.11365670e-01
-7.22177997e-02 3.52714807e-01 -8.61825228e-01 6.69224143e-01
5.73091209e-01 9.33634639e-01 5.17662540e-02 6.45362973e-01
2.17824504e-01 5.36368489e-01 -4.07669753e-01 -1.35827720e-01
4.01701838e-01 -2.17966005e-01 -1.14820206e+00 1.05211250e-01
5.93952954e-01 -3.33566666e-01 -3.08816701e-01 8.38155508e-01
5.65373600e-01 -1.06551774e-01 7.09534705e-01 -6.84879124e-01
-6.93410933e-01 4.65437442e-01 -3.57520103e-01 6.79195404e-01
-1.18964493e-01 4.42129076e-01 7.87124157e-01 -6.07687652e-01
1.02797365e+00 9.00207043e-01 8.87654901e-01 7.89160430e-01
-1.54404223e+00 -5.41902781e-01 -5.42760454e-02 -9.43905339e-02
-1.13108420e+00 -1.61134958e-01 2.76707053e-01 -5.58124661e-01
8.83487046e-01 -3.08013558e-02 4.37181354e-01 1.33408999e+00
5.50508916e-01 4.81952637e-01 1.42630661e+00 -1.77920237e-01
2.74935341e-03 2.08766565e-01 -2.78788731e-02 5.68529963e-01
1.40027463e-01 4.58527505e-01 -2.23520264e-01 -1.00813195e-01
1.09114218e+00 -1.88967481e-01 -3.05122614e-01 -3.95073205e-01
-1.33148038e+00 9.41210508e-01 8.55232954e-01 4.20008361e-01
-3.89984757e-01 -8.92017037e-03 7.50805080e-01 6.09727085e-01
4.50669408e-01 8.11300695e-01 -7.39280522e-01 -2.18221750e-02
-8.64468277e-01 4.34206352e-02 3.94131184e-01 6.67664051e-01
6.93599820e-01 -1.58876270e-01 3.53615056e-03 7.20872283e-01
-2.61056393e-01 4.71433662e-02 8.71585071e-01 -7.46221006e-01
2.91841716e-01 4.47007865e-01 -2.68612623e-01 -3.48142177e-01
-7.06171334e-01 -7.42894113e-01 -8.28710854e-01 2.18783110e-01
5.56068182e-01 -3.59956682e-01 -1.50234246e+00 1.81384861e+00
9.62311253e-02 1.33903623e-01 7.26902857e-02 8.10899198e-01
7.97805846e-01 2.04706505e-01 5.11482954e-01 1.68433607e-01
1.12489593e+00 -8.26855361e-01 -1.13365561e-01 -3.02624136e-01
8.33949506e-01 -5.52120090e-01 1.27971065e+00 4.17526931e-01
-8.92542958e-01 -5.11280656e-01 -9.10857379e-01 -5.81959374e-02
-4.72761661e-01 -2.32952058e-01 3.79132926e-01 4.20391887e-01
-1.19045961e+00 8.11691523e-01 -8.91992390e-01 -5.14730453e-01
6.60502553e-01 6.26875818e-01 -6.76460683e-01 -2.65823603e-01
-1.03221297e+00 1.07895374e+00 5.99650681e-01 -4.50780481e-01
-1.03857112e+00 -1.32292426e+00 -6.23610675e-01 -1.18308082e-01
-1.11319162e-01 -9.27021861e-01 1.13294339e+00 -1.27636838e+00
-1.06471848e+00 1.41897535e+00 5.84146678e-01 -7.40443051e-01
7.04274297e-01 -1.14248626e-01 -4.55509663e-01 3.47000867e-01
1.84478834e-01 1.16853654e+00 8.20185184e-01 -1.04699695e+00
-6.06912434e-01 -2.87014484e-01 2.31236350e-02 2.99917102e-01
-2.97245830e-01 -2.58576214e-01 -4.29753587e-02 -7.14291930e-01
-2.10294351e-01 -1.01048136e+00 -5.11905491e-01 -1.22300945e-01
-5.48296683e-02 -6.23001046e-02 5.86392879e-01 -5.63766539e-01
7.76840746e-01 -2.23835206e+00 -2.05729842e-01 1.71373740e-01
3.91509384e-01 2.06506446e-01 -3.46564293e-01 2.00621262e-02
-4.08512294e-01 2.82544419e-02 -4.28177804e-01 6.32328093e-02
-4.50835109e-01 2.77311653e-01 -3.30996290e-02 6.02644801e-01
3.28158379e-01 1.04189479e+00 -8.42204034e-01 -4.34478909e-01
7.53122941e-02 2.86213547e-01 -6.56057894e-01 -8.16678852e-02
-6.30319864e-02 9.01924968e-01 -5.23394406e-01 4.37003344e-01
4.84699219e-01 -8.17094982e-01 1.92859277e-01 -1.94546282e-01
1.80805832e-01 2.60957003e-01 -5.20696580e-01 1.94153512e+00
-4.58310813e-01 4.87571806e-01 1.26821697e-01 -1.11309719e+00
5.79650044e-01 3.82471293e-01 7.85390735e-01 -9.04159904e-01
6.83485866e-02 2.48292759e-01 4.72771972e-01 -2.95970380e-01
1.02161758e-01 -6.49490118e-01 -9.70104411e-02 3.56188208e-01
5.27978837e-01 -5.93239553e-02 -1.13681011e-01 -9.88038480e-02
1.49957633e+00 -1.76299512e-01 3.48908335e-01 -7.94666588e-01
2.82164484e-01 4.32351947e-01 3.02033424e-01 6.83646739e-01
-3.29831272e-01 8.03268492e-01 2.71609575e-01 -6.88412726e-01
-1.15187752e+00 -1.26216185e+00 -6.37454748e-01 1.26682866e+00
2.58555985e-03 -1.43578023e-01 -7.08247244e-01 -1.12541640e+00
1.64791554e-01 4.21549499e-01 -1.11629224e+00 -3.24956596e-01
-6.93662226e-01 -8.15256476e-01 6.57796025e-01 7.46389866e-01
3.96383077e-01 -1.10611677e+00 -6.60582662e-01 2.15578556e-01
2.46769026e-01 -1.22627532e+00 -3.00229013e-01 5.46952486e-01
-1.27624083e+00 -9.36248720e-01 -9.94060218e-01 -7.39862978e-01
8.59983802e-01 -9.39759761e-02 1.46257114e+00 1.03497908e-01
-3.34154397e-01 6.05675519e-01 -2.32306406e-01 -1.77980080e-01
-6.49989963e-01 5.83760321e-01 -1.27028421e-01 -5.21367550e-01
1.71777546e-01 -6.53035223e-01 -9.43320394e-01 3.29083890e-01
-9.76197243e-01 -1.96718425e-02 1.04109585e+00 1.16115308e+00
5.33801079e-01 -3.11383367e-01 6.35368228e-01 -1.32487524e+00
5.77787399e-01 -5.71745515e-01 -2.27024898e-01 1.45790607e-01
-8.19569170e-01 7.78915659e-02 5.82053781e-01 -4.25706506e-01
-9.63194549e-01 5.06727397e-02 -1.08755954e-01 -4.52730507e-01
-5.18511176e-01 5.32519042e-01 3.76557469e-01 -3.08603168e-01
1.17571437e+00 7.99398050e-02 1.50390640e-01 -3.47488493e-01
3.13025236e-01 2.18842283e-01 5.84676385e-01 -7.49875665e-01
6.68088019e-01 6.85269117e-01 -2.71317273e-01 -5.32918692e-01
-8.00597131e-01 -3.75829250e-01 -7.87423909e-01 1.30853236e-01
1.13898730e+00 -9.45628345e-01 -1.35724515e-01 -7.42677301e-02
-7.20542729e-01 -8.34439278e-01 -5.38038433e-01 4.77781028e-01
-5.39453030e-01 -4.78918105e-02 -7.69469440e-01 3.55147392e-01
-4.02905315e-01 -1.24871778e+00 1.02354145e+00 1.65058263e-02
-3.12567264e-01 -1.33567882e+00 2.49451712e-01 2.73771495e-01
5.22482395e-01 3.12627226e-01 1.12583232e+00 -8.63536596e-01
-2.31688887e-01 -6.70963153e-02 -4.30670679e-01 3.61303896e-01
3.44755054e-01 -5.53992450e-01 -1.00158596e+00 -4.98575896e-01
-2.49912649e-01 -7.16861308e-01 1.02084959e+00 4.56748426e-01
1.22163010e+00 -7.55144656e-02 -3.81467044e-01 7.89714813e-01
1.42356598e+00 -1.23582192e-01 5.60765922e-01 4.98057246e-01
4.38481838e-01 4.30367410e-01 3.42405528e-01 1.26370534e-01
1.19936660e-01 3.23639214e-01 2.31814623e-01 -4.83039230e-01
-3.05996627e-01 -1.66858554e-01 4.42525260e-02 4.16838974e-01
8.98239017e-02 1.46027863e-01 -1.26459348e+00 6.15103424e-01
-1.30910313e+00 -5.02163112e-01 3.88154119e-01 1.98134303e+00
1.04583955e+00 4.10155863e-01 2.76848346e-01 -3.46895009e-01
4.97088760e-01 -1.39309585e-01 -6.25929177e-01 -1.66072145e-01
1.40639439e-01 6.43893063e-01 8.37851107e-01 2.49376759e-01
-1.28956735e+00 9.38479841e-01 7.14352608e+00 7.44975626e-01
-1.74654067e+00 3.68979394e-01 1.04092693e+00 1.17135085e-01
-9.85946283e-02 -3.35637063e-01 -3.60710919e-01 2.46921420e-01
1.04705405e+00 -2.91114748e-02 -3.97369452e-02 8.13913465e-01
-2.34682739e-01 6.81937337e-02 -1.30654848e+00 7.12161839e-01
-2.47238293e-01 -1.57925379e+00 4.33872566e-02 2.63059139e-01
1.00514209e+00 5.94858170e-01 3.49187702e-01 4.90661860e-01
4.44704086e-01 -1.39769578e+00 2.14193881e-01 -3.91192660e-02
1.15744603e+00 -5.06152987e-01 6.51270032e-01 1.77189305e-01
-7.27402449e-01 8.58057290e-02 -1.45895630e-01 3.11399519e-01
-8.27557221e-02 3.03491026e-01 -1.32798743e+00 3.81745011e-01
8.92884552e-01 7.35687733e-01 -7.15097249e-01 8.18301439e-01
2.78453920e-02 7.31707871e-01 -2.25834370e-01 5.31379700e-01
6.48934126e-01 2.86625594e-01 1.61396623e-01 1.46627545e+00
1.97378267e-02 1.67329550e-01 2.58381516e-01 5.83375514e-01
-1.85338408e-01 -1.06392354e-02 -5.43581069e-01 1.06449522e-01
-7.75044113e-02 1.25818372e+00 -9.13339555e-01 -4.99891490e-01
-4.99101311e-01 8.42326820e-01 2.51603723e-01 2.35885471e-01
-8.01361680e-01 6.41146600e-02 5.76129973e-01 4.30406213e-01
4.63155597e-01 -1.51064247e-02 -6.21410131e-01 -9.62315738e-01
-4.48215544e-01 -1.05092382e+00 7.73790359e-01 -3.36133927e-01
-1.56431031e+00 8.60644400e-01 1.25220660e-02 -1.26322150e+00
-2.45965153e-01 -9.01919127e-01 -4.50548351e-01 5.46646714e-01
-1.53389037e+00 -1.07910049e+00 -2.93192357e-01 7.91968346e-01
3.20817441e-01 -4.51396257e-02 8.17496121e-01 4.20490623e-01
-2.16720216e-02 7.57429659e-01 6.94440529e-02 3.07030708e-01
9.57172513e-01 -1.32762992e+00 3.36137116e-01 3.13754350e-01
-3.99398282e-02 6.03779972e-01 5.59155822e-01 -5.07682443e-01
-8.27366889e-01 -1.22850072e+00 2.52048522e-01 -5.87117136e-01
7.70456314e-01 -2.34392986e-01 -1.01261818e+00 8.69945526e-01
3.48785490e-01 3.69933605e-01 7.81702816e-01 2.03145251e-01
-4.37991619e-01 9.67130065e-03 -1.37431097e+00 3.15933347e-01
1.04488492e+00 -5.00464559e-01 -6.47164464e-01 4.05043691e-01
6.83047235e-01 -5.94887197e-01 -1.36698234e+00 6.96972549e-01
3.14166456e-01 -7.98505127e-01 1.09466767e+00 -8.87665749e-01
6.47496998e-01 2.75916874e-01 -2.71261856e-02 -1.32881451e+00
-4.88669991e-01 -2.53214955e-01 4.72870946e-01 5.86714983e-01
6.83921397e-01 -6.59871817e-01 1.06789613e+00 3.86148125e-01
-3.66915464e-01 -8.03260624e-01 -9.40603137e-01 -7.65684962e-01
8.03030312e-01 -2.20312431e-01 2.09597453e-01 1.24173999e+00
-6.86596856e-02 3.22001040e-01 1.17162764e-01 -6.72210380e-02
4.18747216e-01 3.37805934e-02 5.74572563e-01 -1.06936586e+00
-5.27393341e-01 -4.98609811e-01 -4.20433313e-01 -7.53973305e-01
9.21506956e-02 -1.12964869e+00 -3.75447385e-02 -1.13489306e+00
2.32122630e-01 -6.09804213e-01 -7.42241025e-01 5.75729609e-01
-3.34585346e-02 4.65122014e-01 -2.30063181e-02 2.12848276e-01
-4.31887329e-01 1.77183509e-01 1.60807347e+00 -9.13366601e-02
-5.65658696e-02 -1.38358667e-01 -7.61976242e-01 6.12385035e-01
8.22317541e-01 -4.73928243e-01 -5.01900434e-01 -5.16157925e-01
-2.99506813e-01 5.87635376e-02 4.27910507e-01 -1.05259955e+00
-5.76044358e-02 -1.13465033e-01 6.22190356e-01 -1.39412716e-01
1.16848178e-01 -6.61013007e-01 -1.37664884e-01 6.66104257e-01
-6.29877090e-01 -2.42003687e-02 6.32143021e-01 4.97851133e-01
-2.26190448e-01 1.24341629e-01 1.15407455e+00 -4.62048918e-01
-8.48291218e-01 4.93297905e-01 -2.36648411e-01 4.56650943e-01
7.22909868e-01 -2.38870129e-01 -3.19124937e-01 -1.16934404e-01
-1.14636421e+00 5.37816063e-02 4.03422385e-01 2.44512483e-01
4.54421580e-01 -1.03152478e+00 -7.28246391e-01 8.14181268e-02
3.42723250e-01 -6.23050742e-02 4.01857525e-01 8.95067871e-01
-5.03570914e-01 5.20254195e-01 -5.50869405e-01 -8.63903463e-01
-1.01013231e+00 4.22547847e-01 4.34290111e-01 -7.74304092e-01
-7.36451268e-01 9.56599116e-01 8.05306375e-01 -7.46044159e-01
-8.33429992e-02 -4.93677735e-01 2.48867989e-01 -3.32836509e-01
2.07500666e-01 -1.56698879e-02 3.49032789e-01 -2.36065194e-01
-4.39954877e-01 2.99187154e-01 -4.83844966e-01 -1.58827886e-01
1.45203161e+00 2.60668367e-01 3.52396071e-01 1.95171237e-02
1.41610408e+00 -4.67417955e-01 -1.38081527e+00 -4.47118074e-01
1.02118067e-01 -5.86851947e-02 1.44366384e-01 -1.13216472e+00
-1.07570446e+00 8.17866743e-01 8.48021984e-01 -2.36317322e-01
1.13084674e+00 2.37121433e-01 7.16521263e-01 2.81076103e-01
3.44292045e-01 -9.65745509e-01 3.18954110e-01 3.60809952e-01
8.61057818e-01 -1.36919022e+00 4.02650945e-02 -4.57286052e-02
-7.42811263e-01 9.77108240e-01 8.36360574e-01 -5.92357397e-01
8.77146900e-01 2.55019695e-01 2.25702658e-01 -3.75645280e-01
-6.68220401e-01 -2.41661239e-02 3.85194391e-01 5.85809052e-01
4.66932684e-01 1.18659467e-01 1.58036258e-02 4.50273693e-01
-2.32044458e-01 1.45466119e-01 1.10001251e-01 9.55220461e-01
-2.99801350e-01 -1.21465325e+00 -2.10861653e-01 6.13967061e-01
-5.20893812e-01 -2.12854400e-01 -2.67347127e-01 1.26468682e+00
2.56989479e-01 2.97597051e-01 1.36949554e-01 -3.34139556e-01
1.51452333e-01 -4.92359437e-02 7.25765467e-01 -9.50717986e-01
-8.81260991e-01 2.57045794e-02 -9.02133957e-02 -5.38760602e-01
-2.54392892e-01 -5.85111022e-01 -1.36178815e+00 -6.84313402e-02
5.20835333e-02 -4.49373871e-02 3.46863657e-01 8.98968875e-01
4.43903267e-01 5.69969118e-01 2.69633293e-01 -5.68625450e-01
-7.27680445e-01 -9.67329323e-01 -2.90350854e-01 6.60043478e-01
4.46039796e-01 -6.27355337e-01 -1.01944625e-01 1.48586079e-01]
|
[14.851841926574707, -2.327594757080078]
|
1a06373e-f073-4b58-a68c-a31d8d9c9553
|
deepkspd-learning-kernel-matrix-based-spd
|
1711.04047
| null |
http://arxiv.org/abs/1711.04047v1
|
http://arxiv.org/pdf/1711.04047v1.pdf
|
DeepKSPD: Learning Kernel-matrix-based SPD Representation for Fine-grained Image Recognition
|
Being symmetric positive-definite (SPD), covariance matrix has traditionally
been used to represent a set of local descriptors in visual recognition. Recent
study shows that kernel matrix can give considerably better representation by
modelling the nonlinearity in the local descriptor set. Nevertheless, neither
the descriptors nor the kernel matrix is deeply learned. Worse, they are
considered separately, hindering the pursuit of an optimal SPD representation.
This work proposes a deep network that jointly learns local descriptors,
kernel-matrix-based SPD representation, and the classifier via an end-to-end
training process. We derive the derivatives for the mapping from a local
descriptor set to the SPD representation to carry out backpropagation. Also, we
exploit the Daleckii-Krein formula in operator theory to give a concise and
unified result on differentiating SPD matrix functions, including the matrix
logarithm to handle the Riemannian geometry of kernel matrix. Experiments not
only show the superiority of kernel-matrix-based SPD representation with deep
local descriptors, but also verify the advantage of the proposed deep network
in pursuing better SPD representations for fine-grained image recognition
tasks.
|
['Lei Wang', 'Xinwang Liu', 'Melih Engin', 'Luping Zhou']
|
2017-11-11
|
deepkspd-learning-kernel-matrix-based-spd-1
|
http://openaccess.thecvf.com/content_ECCV_2018/html/Melih_Engin_DeepKSPD_Learning_Kernel-matrix-based_ECCV_2018_paper.html
|
http://openaccess.thecvf.com/content_ECCV_2018/papers/Melih_Engin_DeepKSPD_Learning_Kernel-matrix-based_ECCV_2018_paper.pdf
|
eccv-2018-9
|
['fine-grained-image-recognition']
|
['computer-vision']
|
[-1.20658010e-01 -3.41235965e-01 -1.31594375e-01 -5.45159757e-01
-5.70268035e-01 -4.30505604e-01 7.31671035e-01 -2.14953303e-01
-3.37972581e-01 2.36707464e-01 3.82502750e-02 -1.40421987e-01
-7.30779231e-01 -5.67649662e-01 -5.75410903e-01 -1.06440353e+00
-2.87344486e-01 -5.76722734e-02 -2.27126092e-01 -1.05235070e-01
3.61247689e-01 1.09960389e+00 -1.33355772e+00 -1.82743110e-02
7.14865267e-01 1.29124415e+00 9.62274969e-02 3.38541538e-01
2.12372512e-01 6.94616079e-01 -4.30835426e-01 -2.29679838e-01
5.72918952e-01 -1.52731180e-01 -6.82082057e-01 5.24713583e-02
5.47029555e-01 -1.92620769e-01 -6.31853640e-01 1.17998648e+00
4.37687933e-01 3.19959164e-01 1.19958580e+00 -1.17168677e+00
-1.25937521e+00 -3.75291109e-02 -3.94601822e-01 2.03510135e-01
2.04771727e-01 -1.44575402e-01 1.14077687e+00 -1.17315912e+00
5.13479352e-01 1.08296454e+00 6.54923081e-01 1.40204206e-01
-1.19523358e+00 -3.03800613e-01 -1.45777419e-01 3.20249081e-01
-1.86890495e+00 -7.82541856e-02 9.05566156e-01 -5.56647837e-01
9.93915021e-01 1.94419131e-01 4.91788656e-01 8.53262722e-01
3.80473286e-01 6.43178284e-01 8.30332100e-01 -3.32292199e-01
-1.94711328e-01 8.36291537e-02 1.37047514e-01 9.67719972e-01
1.62620395e-01 5.68394996e-02 -3.11678380e-01 -9.94952917e-02
1.14251864e+00 2.37847596e-01 -2.54132718e-01 -9.04219449e-01
-1.22700274e+00 9.41199124e-01 7.76477098e-01 4.38858360e-01
-5.00992119e-01 -9.29788034e-03 2.85278380e-01 5.47017932e-01
3.34998697e-01 2.42614180e-01 -1.41183451e-01 -4.71866131e-02
-4.63564068e-01 -4.17417400e-02 6.24993622e-01 6.34133160e-01
8.56441498e-01 -5.83485402e-02 -2.14125127e-01 9.60355759e-01
3.11814308e-01 5.96788943e-01 4.40410554e-01 -7.74323106e-01
4.83971715e-01 6.61803901e-01 -7.16095567e-02 -1.63698864e+00
-4.54409242e-01 -5.06490529e-01 -1.15602589e+00 2.14267880e-01
4.12409395e-01 2.97910213e-01 -4.06802237e-01 1.71871841e+00
1.14764534e-01 -1.43853901e-02 -1.21709168e-01 1.28421772e+00
3.58189076e-01 6.24810517e-01 -2.69499779e-01 3.71634722e-01
9.38359797e-01 -7.09888935e-01 -3.48749131e-01 1.55922472e-01
8.92578125e-01 -4.75221157e-01 8.98044348e-01 1.16135113e-01
-7.21103728e-01 -5.28448522e-01 -1.23253834e+00 -4.32256132e-01
-5.19228816e-01 5.69251418e-01 6.23257577e-01 1.98791951e-01
-1.27281952e+00 8.59803319e-01 -6.59713924e-01 -2.81746745e-01
1.48458824e-01 3.95457327e-01 -8.17572296e-01 2.26328019e-02
-1.17546308e+00 1.08616316e+00 2.35423535e-01 5.30723512e-01
-6.01776659e-01 -6.03557646e-01 -9.28029776e-01 1.19499274e-01
-2.00759768e-01 -5.26161253e-01 5.78534842e-01 -5.95315933e-01
-1.62758768e+00 1.05593419e+00 1.33929029e-01 -1.75249264e-01
2.42235422e-01 8.63982812e-02 -1.91685855e-01 2.28787959e-01
-8.26809704e-02 3.05248946e-01 1.05550098e+00 -5.58266759e-01
-1.06325388e-01 -4.68470842e-01 2.97571778e-01 2.91319162e-01
-3.91877711e-01 -2.31633589e-01 -1.19855024e-01 -5.01758814e-01
1.83638334e-01 -6.92431092e-01 1.51711658e-01 2.58889884e-01
-2.67746598e-01 -5.50934494e-01 4.50638205e-01 -6.14092588e-01
1.03700173e+00 -2.60961366e+00 4.26901996e-01 5.04745126e-01
2.31508657e-01 2.04718560e-01 -5.67255914e-01 5.18768668e-01
-3.41073334e-01 -9.23946649e-02 -5.10425016e-04 -1.06449634e-01
3.21802050e-01 5.81498481e-02 -2.15085655e-01 9.31980371e-01
6.62648797e-01 9.22350109e-01 -6.57070518e-01 -1.01066835e-01
1.99139014e-01 8.90211344e-01 -2.84338713e-01 2.29731807e-03
5.72698593e-01 -3.98414629e-03 -5.76269448e-01 1.90145805e-01
9.08475757e-01 -2.40564585e-01 -1.96097240e-01 -5.29869437e-01
-1.38738841e-01 2.29531854e-01 -1.21196234e+00 1.63942409e+00
-5.88049114e-01 6.55269921e-01 1.53715089e-01 -1.43373263e+00
1.16176069e+00 -1.36459127e-01 2.65978009e-01 -7.91695416e-01
2.27287382e-01 3.33391011e-01 -1.02054454e-01 -2.32500643e-01
7.57242516e-02 -9.83829275e-02 1.38294548e-01 2.57581919e-01
2.14116991e-01 2.61899203e-01 8.45290944e-02 -4.89569269e-04
9.35435772e-01 -1.46475315e-01 1.75186425e-01 -6.57034099e-01
1.06165302e+00 -5.96070647e-01 1.45999923e-01 5.47465563e-01
-1.51686668e-01 5.83047688e-01 7.18717933e-01 -4.27521795e-01
-7.37982035e-01 -1.14373982e+00 -6.34232700e-01 8.48594546e-01
2.00525343e-01 -1.43884793e-01 -4.97997224e-01 -5.67682505e-01
1.99690983e-01 1.92890257e-01 -8.34817410e-01 -7.10417390e-01
-2.50322610e-01 -4.13642049e-01 5.03578842e-01 4.83892828e-01
6.19842172e-01 -5.37019491e-01 -1.24052972e-01 -1.99444205e-01
3.13713729e-01 -7.90001631e-01 -8.66782963e-01 2.85883188e-01
-7.21709847e-01 -1.06750500e+00 -9.89431977e-01 -9.84810054e-01
8.29501629e-01 3.89253736e-01 4.42429900e-01 -1.11409388e-01
-2.51356751e-01 6.91245139e-01 -6.48215190e-02 2.13983625e-01
-1.29367501e-01 3.19324583e-02 6.19388223e-02 5.20887911e-01
4.44631845e-01 -4.31996703e-01 -6.67506099e-01 4.67090696e-01
-8.44476700e-01 -3.02835703e-01 7.82844007e-01 1.04664886e+00
4.54518408e-01 -1.86279222e-01 3.32336843e-01 -1.54213443e-01
1.02548051e+00 -2.75077432e-01 -6.15539134e-01 2.84955442e-01
-3.90924543e-01 5.34561813e-01 7.83991992e-01 -3.71585965e-01
-6.33628070e-01 -1.17008887e-01 5.84737919e-02 -5.18160760e-01
5.99187836e-02 7.03182220e-01 -5.12877554e-02 -5.30078828e-01
6.08522415e-01 5.42346656e-01 3.83348972e-01 -6.22489810e-01
2.46457055e-01 5.99019706e-01 2.42575452e-01 -4.42595452e-01
9.42931771e-01 3.85786265e-01 3.26442510e-01 -9.13344860e-01
-8.81602526e-01 -6.47506416e-01 -1.09905636e+00 1.78369001e-01
7.61233211e-01 -8.24849784e-01 -9.19519186e-01 5.72903454e-01
-1.06325436e+00 4.30677906e-02 -3.28811854e-01 6.28625512e-01
-5.21522224e-01 4.74323153e-01 -5.61560273e-01 -4.69006211e-01
-7.91619718e-02 -9.40690219e-01 1.25117481e+00 -4.07476425e-02
2.58274108e-01 -1.38851857e+00 6.37079775e-02 2.43155751e-02
4.83846784e-01 1.41626537e-01 8.81188393e-01 -4.94262636e-01
-4.68223184e-01 -7.38803923e-01 -6.82254553e-01 8.52457047e-01
5.50594330e-02 -1.81986630e-01 -7.05363631e-01 -3.85715336e-01
2.00261444e-01 -1.06438175e-01 8.04028869e-01 1.08742289e-01
1.00391281e+00 -1.33458927e-01 8.19274038e-02 9.64060962e-01
1.32970273e+00 1.43597052e-02 3.90105218e-01 3.54169279e-01
7.90824115e-01 3.78297657e-01 3.66105944e-01 3.90061885e-01
3.38902593e-01 7.20810115e-01 1.66831195e-01 -9.53346267e-02
-2.64197830e-02 -1.85132414e-01 5.98110557e-01 1.10006189e+00
-9.77811664e-02 3.78447682e-01 -7.56787121e-01 2.74006456e-01
-1.63132191e+00 -7.15089738e-01 1.51091769e-01 2.37052274e+00
5.08756876e-01 -1.41144380e-01 -1.35925636e-01 -1.31559772e-02
6.38175309e-01 2.46865228e-01 -6.01628125e-01 -5.57031751e-01
-3.51508766e-01 7.90695921e-02 3.61506164e-01 6.14310801e-01
-1.18185985e+00 7.15462863e-01 5.69345808e+00 9.08584118e-01
-1.37128365e+00 -6.93404153e-02 3.87627743e-02 4.49041784e-01
-1.01052128e-01 -3.45339954e-01 -6.18985236e-01 6.54997677e-02
7.32568383e-01 -1.51333883e-01 5.80734432e-01 8.03070068e-01
-6.94487020e-02 2.36165434e-01 -1.27457857e+00 1.46252203e+00
2.35093072e-01 -1.11087966e+00 3.27479869e-01 4.07175034e-01
5.19203782e-01 5.48732979e-03 3.66675526e-01 3.19769889e-01
-4.23529655e-01 -1.03084230e+00 5.58579683e-01 8.28121901e-01
6.53446496e-01 -8.93195689e-01 7.26396620e-01 3.03477168e-01
-1.23310471e+00 -4.21556681e-02 -9.94275987e-01 -1.37244314e-01
-4.20039356e-01 5.65605521e-01 -6.27214968e-01 5.03245890e-01
2.35658541e-01 1.17292809e+00 -8.23921025e-01 1.02202332e+00
-8.67573395e-02 -5.52542843e-02 -3.59853506e-01 -8.43255296e-02
6.10677063e-01 -8.84828508e-01 4.54742104e-01 1.16461551e+00
3.32917541e-01 -2.51366556e-01 -1.50727749e-01 1.18483627e+00
7.16543710e-03 2.92146415e-01 -1.01304901e+00 -3.28608394e-01
-2.13275924e-02 1.27679479e+00 -2.89262891e-01 1.25815049e-01
-4.14544404e-01 1.18256021e+00 6.09699368e-01 5.61373293e-01
-6.07182503e-01 -9.57951903e-01 1.07224274e+00 -1.05029076e-01
5.52457333e-01 -6.91339970e-01 3.97348106e-02 -1.45195413e+00
2.81691074e-01 -5.33917487e-01 9.21342224e-02 -5.20569444e-01
-1.53109324e+00 4.51588333e-01 -6.54702261e-02 -1.33495736e+00
-1.72941685e-01 -1.22840166e+00 -4.95945573e-01 1.16254330e+00
-1.46772516e+00 -1.06650853e+00 -1.13763273e-01 9.12803769e-01
-1.56786397e-01 -1.50286123e-01 7.61854231e-01 5.88130713e-01
-5.01408339e-01 8.24622691e-01 5.58008432e-01 4.64203119e-01
5.77674448e-01 -1.27474844e+00 1.19122490e-01 6.23166084e-01
5.38283214e-02 9.58841980e-01 1.29067913e-01 -9.82535183e-02
-1.65521264e+00 -8.30014944e-01 7.46914744e-01 -3.75056028e-01
1.01972210e+00 -4.70655859e-01 -8.30570877e-01 3.73696029e-01
-2.39138007e-01 3.52887779e-01 4.13727492e-01 -1.17626162e-02
-5.42558312e-01 -5.56733370e-01 -7.74765193e-01 4.09930825e-01
1.08061445e+00 -1.27134061e+00 -3.40101033e-01 3.90411228e-01
2.31277570e-01 -1.39281094e-01 -1.09676552e+00 8.89757797e-02
5.21245062e-01 -9.09127176e-01 1.10632777e+00 -6.42140627e-01
9.01347175e-02 -3.70694041e-01 -3.43155921e-01 -1.32806158e+00
-6.67839348e-01 -2.61424243e-01 1.88020952e-02 9.79596436e-01
2.50964940e-01 -9.62002039e-01 2.86354363e-01 4.46019292e-01
-1.20023742e-01 -9.61853325e-01 -1.18718815e+00 -1.18517685e+00
2.52780586e-01 -2.48346865e-01 2.57196754e-01 8.41928244e-01
2.59289198e-04 3.23011011e-01 -2.58675516e-01 2.04366624e-01
6.17880702e-01 5.13374582e-02 5.38271844e-01 -1.14060318e+00
-9.88738537e-02 -6.94218755e-01 -1.30018985e+00 -1.44385469e+00
4.08621341e-01 -1.57141674e+00 -2.03961879e-01 -1.33553553e+00
1.12216741e-01 -1.28585964e-01 -6.00529492e-01 2.58272707e-01
3.91219199e-01 7.67424032e-02 2.29485944e-01 1.64664447e-01
-4.78506327e-01 8.38698804e-01 1.54383838e+00 -3.10965478e-01
7.30107948e-02 -1.25507843e-02 -5.46962202e-01 4.97505099e-01
4.93438870e-01 -1.80454552e-01 -4.35141325e-01 -3.65609407e-01
2.27311328e-01 -3.09863687e-01 5.83701253e-01 -8.60989511e-01
2.49776855e-01 9.27941874e-02 3.86668295e-01 -1.94016829e-01
4.51839238e-01 -7.61916637e-01 -3.63178939e-01 3.38067412e-01
-3.56165469e-01 -1.18642814e-01 -1.44900754e-01 5.66518545e-01
-5.43411553e-01 -3.69921744e-01 8.28702271e-01 3.21899474e-01
-5.27770996e-01 4.78363454e-01 -2.07133412e-01 -5.61941788e-02
8.45049560e-01 -3.83158922e-01 -1.01171002e-01 -2.19381675e-01
-7.90257275e-01 -6.02335855e-02 2.27321163e-01 4.17814702e-01
8.33953023e-01 -1.78865814e+00 -5.51092207e-01 7.15910792e-01
4.70400415e-02 -3.46822590e-01 2.66588241e-01 1.33952212e+00
-7.57957578e-01 6.93761230e-01 -3.52475643e-01 -7.46362388e-01
-8.53439629e-01 5.14125884e-01 6.87146664e-01 -8.89270306e-02
-7.28657246e-01 8.68185341e-01 4.67041492e-01 -6.50387645e-01
4.17682230e-01 -5.88457167e-01 -2.82584261e-02 1.37400523e-01
4.71723914e-01 3.68443370e-01 8.46901760e-02 -7.55978465e-01
-5.72427571e-01 1.06684327e+00 -3.86178903e-02 2.66389102e-01
1.37424040e+00 -1.86419547e-01 -3.52827907e-01 4.86321807e-01
2.20002174e+00 -4.13838118e-01 -1.20303226e+00 -3.35554659e-01
1.09657668e-01 -4.97777641e-01 1.77201107e-01 -3.33236188e-01
-1.04998744e+00 1.12615991e+00 6.71567082e-01 2.52991676e-01
8.90051723e-01 -3.47734690e-02 4.97844309e-01 8.44318569e-01
2.27145493e-01 -8.89899075e-01 8.62665381e-03 9.00377333e-01
1.45456815e+00 -1.11071718e+00 -1.44792855e-01 -9.10094306e-02
-5.28457224e-01 1.47567272e+00 2.39621922e-01 -5.73159635e-01
1.14052677e+00 -3.49681407e-01 -5.94770769e-03 -2.38214523e-01
-3.67268801e-01 3.62084620e-02 9.24736440e-01 5.89392245e-01
3.59597147e-01 -1.24340542e-02 -1.40690535e-01 3.02615285e-01
-1.23685412e-02 -4.47126329e-01 1.91956311e-01 4.93824273e-01
-2.21561030e-01 -9.53243792e-01 -1.86260968e-01 2.72340149e-01
9.68659222e-02 6.92668781e-02 -4.51367527e-01 7.92488694e-01
-2.04089228e-02 4.53869700e-01 3.18863355e-02 -3.79031122e-01
4.23737377e-01 1.24644287e-01 6.56286061e-01 -3.60094279e-01
-1.68946773e-01 -2.56701469e-01 -4.11934733e-01 -6.33624375e-01
-2.10138038e-01 -5.50359547e-01 -9.74799275e-01 -3.94463062e-01
-1.81657776e-01 1.35488927e-01 7.24643528e-01 8.36609066e-01
5.98421216e-01 2.90047340e-02 9.23840046e-01 -9.41811800e-01
-9.48534250e-01 -7.75698960e-01 -1.17258537e+00 5.80952108e-01
4.76971835e-01 -8.89382720e-01 -7.19944715e-01 -4.09961551e-01]
|
[8.942140579223633, 2.1675868034362793]
|
f56f1ecd-18d1-4e02-ba85-f67358867113
|
a-semi-trailer-truck-right-hook-turn-blind
|
2303.11223
| null |
https://arxiv.org/abs/2303.11223v1
|
https://arxiv.org/pdf/2303.11223v1.pdf
|
A semi-trailer truck right-hook turn blind spot alert system for detecting vulnerable road users using transfer learning
|
Cycling is an increasingly popular method of transportation for sustainability and health benefits. However, cyclists face growing risks, especially when encountering semi-trailer trucks. This study aims to reduce the number of truck-cyclist collisions, which are often caused by semi-trailer trucks making right-hook turns and poor driver attention to blind spots. To achieve this, we designed a visual-based blind spot warning system that can detect cyclists for semi-trailer truck drivers using deep learning. First, several greater than 90% mAP cyclist detection models, such as the EfficientDet Lite 1 and SSD MobileNetV2, were created using state-of-the-art lightweight deep learning architectures fine-tuned on a newly proposed cyclist image dataset composed of a diverse set of over 20,000 images. Next, the object detection model was deployed onto a Google Coral Dev Board mini-computer with a camera module and analyzed for speed, reaching inference times as low as 15 milliseconds. Lastly, the end-to-end blind spot cyclist detection device was tested in real-time to model traffic scenarios and analyzed further for performance and feasibility. We concluded that this portable blind spot alert device can accurately and quickly detect cyclists and have the potential to significantly improve cyclist safety. Future studies could determine the feasibility of the proposed device in the trucking industry and improvements to cyclist safety over time.
|
['Charles Tang']
|
2023-01-16
| null | null | null | null |
['2d-cyclist-detection']
|
['computer-vision']
|
[-1.16684824e-01 -2.75862962e-01 -1.52282789e-01 -9.28712860e-02
-6.71577811e-01 -4.99748975e-01 2.99917042e-01 -1.87491789e-01
-7.07368314e-01 1.34406298e-01 -1.36979565e-01 -9.41225469e-01
2.09613871e-02 -5.62328219e-01 -6.62102580e-01 -3.46139014e-01
1.36546522e-01 3.57358605e-01 7.69073784e-01 -2.93271214e-01
2.18478456e-01 8.75755548e-01 -1.92541885e+00 1.18628561e-01
7.04099357e-01 9.16016042e-01 3.53266060e-01 9.20236230e-01
4.31416094e-01 4.47014540e-01 -3.79586011e-01 -4.30082381e-01
3.62478435e-01 2.94012100e-01 -2.51919534e-02 -3.58485997e-01
6.70760334e-01 -6.31076753e-01 -7.13838875e-01 5.29287279e-01
9.03489769e-01 7.57939890e-02 5.45091569e-01 -1.77121353e+00
-2.51750171e-01 -1.44225642e-01 -5.24421155e-01 6.54939473e-01
7.64244050e-02 9.80070770e-01 2.55340099e-01 -8.11721861e-01
-1.51812494e-01 1.22283053e+00 9.99466240e-01 3.05582911e-01
-6.11751020e-01 -1.20943403e+00 -1.78569198e-01 8.87898147e-01
-1.34031057e+00 -6.08459175e-01 2.11402699e-01 -4.82648313e-01
1.35144126e+00 3.33945096e-01 5.57299554e-01 8.84881437e-01
3.72307032e-01 6.96417809e-01 7.81763732e-01 -4.57965732e-02
-6.35119230e-02 1.38383105e-01 2.92409629e-01 5.73234320e-01
5.79659224e-01 7.43396103e-01 -4.22080189e-01 5.21263659e-01
2.21981853e-01 -9.44862738e-02 3.33710581e-01 1.33036673e-01
-8.56937289e-01 9.13051605e-01 7.82758415e-01 -2.65892088e-01
-3.04805309e-01 2.98217416e-01 5.88370025e-01 -1.06689401e-01
1.82090327e-01 -2.39502653e-01 1.41755603e-02 -1.81810498e-01
-6.79801762e-01 4.56784852e-02 3.74351203e-01 1.06073761e+00
7.10831761e-01 1.11720406e-01 -5.09083152e-01 4.06327456e-01
4.53192979e-01 1.04071319e+00 -2.70266794e-02 -7.08261847e-01
6.08409941e-01 5.23918509e-01 2.23362312e-01 -9.58646774e-01
-8.38272810e-01 -2.68933892e-01 -4.27196622e-01 6.53456569e-01
4.01781738e-01 -2.40160510e-01 -1.20946240e+00 8.53493512e-01
2.57233202e-01 2.65675932e-01 -2.80217111e-01 1.17101109e+00
9.84906375e-01 3.10521841e-01 4.24459547e-01 3.22344422e-01
1.62089002e+00 -9.51232672e-01 -5.68113089e-01 -8.32958281e-01
4.16859329e-01 -7.33904839e-01 1.01744688e+00 -2.09182967e-02
-6.57701731e-01 -8.56039524e-01 -1.14663136e+00 -1.71515673e-01
-7.94287086e-01 4.10764813e-01 4.16435957e-01 1.33856225e+00
-1.01325321e+00 -2.26903722e-01 -8.68886113e-01 -8.02280605e-01
5.70388317e-01 3.76063257e-01 9.99462456e-02 -1.91299602e-01
-1.04912090e+00 1.39389741e+00 -4.59289700e-02 5.34459591e-01
-1.30083776e+00 -8.22406352e-01 -7.35227764e-01 6.75580290e-04
3.31276268e-01 -5.17932475e-01 1.40285039e+00 -2.09295332e-01
-8.88963878e-01 8.60161960e-01 -3.60562265e-01 -6.52548194e-01
7.28572547e-01 -2.17865363e-01 -9.71323311e-01 1.67211089e-02
4.10802037e-01 7.10236549e-01 8.83014143e-01 -1.10572231e+00
-1.13618827e+00 -4.63081837e-01 -1.81563944e-02 2.28666246e-01
5.91147505e-02 3.34245324e-01 -5.05824924e-01 9.66213122e-02
-5.27457952e-01 -1.15240693e+00 1.92582142e-02 2.40702406e-01
-4.25714135e-01 -2.33715728e-01 1.16182578e+00 -6.12652063e-01
1.00768995e+00 -2.14521575e+00 -1.06798744e+00 1.43561810e-01
2.11098686e-01 9.86836910e-01 -2.24473968e-01 1.93681434e-01
2.36057192e-02 -3.98307770e-01 1.21658243e-01 -2.33909905e-01
-5.99435270e-02 -9.85380728e-03 1.95091404e-02 7.27612793e-01
3.50416079e-02 1.15654767e+00 -7.12145686e-01 -3.43348920e-01
7.65795112e-01 4.48906958e-01 8.86424333e-02 7.62705207e-02
3.86949539e-01 3.01922858e-02 -1.24115393e-01 8.03225636e-01
9.95237350e-01 3.88176709e-01 -4.51861501e-01 -2.15978205e-01
-5.89265406e-01 8.14450160e-02 -1.00595856e+00 8.45409691e-01
-5.43258071e-01 1.23002148e+00 1.82234332e-01 -7.31837928e-01
7.33717322e-01 -9.84306037e-02 -4.13207710e-02 -1.40665841e+00
1.85692400e-01 -1.05584217e-02 -1.82937711e-01 -9.54369843e-01
7.32101798e-01 2.48051256e-01 -7.78821036e-02 3.67335111e-01
-6.92949295e-01 3.64638686e-01 1.37702763e-01 9.62845832e-02
1.01769996e+00 -3.88769716e-01 -3.47156376e-01 6.39902726e-02
2.81633973e-01 4.57569152e-01 4.76626419e-02 8.26345801e-01
-7.50441194e-01 3.89071226e-01 5.95326573e-02 -5.22669852e-01
-8.03350210e-01 -1.20225859e+00 -5.37053160e-02 1.26599252e+00
4.76198196e-01 1.89053267e-01 -8.27500045e-01 -5.40997505e-01
3.93424332e-01 9.45560277e-01 -4.82547253e-01 -4.66039121e-01
-5.41129291e-01 -7.31175482e-01 9.14565802e-01 1.01529181e+00
7.69474268e-01 -9.15841401e-01 -9.50348735e-01 8.85313377e-02
-2.52704114e-01 -1.18719506e+00 -7.89219677e-01 -5.71037643e-02
-1.17271528e-01 -1.35643077e+00 -5.03422260e-01 -8.16751063e-01
3.06623966e-01 1.18566346e+00 7.36757636e-01 2.58640908e-02
-6.83518887e-01 3.95968199e-01 6.40997067e-02 -8.29845011e-01
-2.63799161e-01 -9.66792777e-02 1.39642775e-01 9.60103944e-02
1.12560618e+00 2.84218252e-01 -1.16972184e+00 9.29679811e-01
-2.49902636e-01 -3.52037638e-01 6.34838104e-01 3.59859727e-02
7.61418492e-02 1.15719162e-01 6.35980129e-01 -9.58737582e-02
5.72414815e-01 -3.61966223e-01 -7.67476439e-01 2.76489314e-02
-7.27189183e-01 -4.52332616e-01 2.32308451e-02 -2.22316191e-01
-1.03930759e+00 2.93132007e-01 -7.91050047e-02 -7.23610595e-02
-3.18335801e-01 -5.56085110e-02 -2.45920587e-02 -3.51908237e-01
8.39302838e-01 7.54372776e-02 2.22062111e-01 -1.28649473e-01
4.52588379e-01 1.24212766e+00 7.83966482e-01 3.67581159e-01
6.66299820e-01 6.60392821e-01 -1.41199395e-01 -1.02595484e+00
-3.73619124e-02 -9.80779290e-01 -2.79612720e-01 -8.40530455e-01
1.05556250e+00 -1.27335536e+00 -1.42301130e+00 7.82209992e-01
-1.14790821e+00 -3.63481700e-01 3.16549659e-01 4.81102198e-01
-1.72417894e-01 2.42598280e-01 -1.27906576e-01 -1.01816881e+00
-3.96836221e-01 -1.09000051e+00 1.17043817e+00 3.31512511e-01
-6.50281310e-02 -3.34351718e-01 -1.80221796e-01 8.14786553e-01
5.77089787e-01 -1.71333417e-01 3.40760857e-01 -1.32380232e-01
-7.48779893e-01 -6.12202108e-01 -9.14292574e-01 1.88132629e-01
-1.82923585e-01 -2.48413891e-01 -1.19418359e+00 -3.21091235e-01
-6.41038299e-01 1.41786560e-01 9.66468990e-01 6.04085207e-01
6.87817574e-01 1.02058820e-01 -9.58509862e-01 3.18008661e-01
1.11128700e+00 5.02980292e-01 9.68672931e-01 4.42880124e-01
7.72750735e-01 4.46555376e-01 7.48322010e-01 3.80333234e-03
7.18654394e-01 6.47782028e-01 7.53494501e-01 -3.30550671e-01
-6.60257101e-01 -1.45785302e-01 3.06057155e-01 -1.28737897e-01
-9.62967724e-02 -3.86376858e-01 -1.01616716e+00 9.06855762e-01
-1.62727308e+00 -9.28955853e-01 -8.23478460e-01 2.20140076e+00
-1.01289257e-01 3.06809515e-01 7.70833075e-01 4.38437015e-02
1.08444858e+00 -2.46316224e-01 -7.96496093e-01 -4.98249084e-01
3.25824112e-01 -2.95739114e-01 1.32668519e+00 4.10773396e-01
-1.01971090e+00 8.25015783e-01 6.49489546e+00 7.09487140e-01
-1.08770394e+00 4.73947644e-01 2.68857151e-01 -1.34120971e-01
3.14605564e-01 -3.61734420e-01 -1.19726562e+00 6.69933856e-01
1.26997328e+00 1.16572984e-01 3.77375662e-01 8.22765887e-01
7.44296134e-01 -4.19241279e-01 -5.79625607e-01 8.00570786e-01
2.72205118e-02 -1.13305783e+00 -7.41625726e-01 -1.50103802e-02
1.12454087e-01 5.79275310e-01 2.40968928e-01 4.20221418e-01
1.45613343e-01 -9.67306256e-01 8.13596606e-01 3.31695765e-01
1.11318815e+00 -1.00145233e+00 6.25273824e-01 2.61663586e-01
-1.50305879e+00 -5.62315106e-01 -2.09871843e-01 -6.05600141e-02
3.56854737e-01 1.10639021e-01 -1.21732914e+00 -1.05839796e-01
9.72052515e-01 4.49502856e-01 -1.01553762e+00 1.54039705e+00
7.72555405e-03 5.36409616e-01 -5.06057978e-01 -3.97261620e-01
4.74969298e-01 3.53112549e-01 4.19130713e-01 1.54335141e+00
2.38665193e-01 -2.66016126e-01 -2.88070202e-01 6.29109621e-01
3.69175076e-01 -6.80119216e-01 -8.46300244e-01 4.79292125e-01
3.87447476e-01 1.27451539e+00 -8.19036901e-01 -7.61594847e-02
-6.68067694e-01 5.39873242e-01 -2.59999901e-01 6.08027101e-01
-1.36259043e+00 -5.56896567e-01 9.28275287e-01 6.72869802e-01
4.04913217e-01 -2.05861628e-01 -5.26012421e-01 -1.28470898e-01
2.90810484e-02 -3.54563415e-01 1.94198593e-01 -1.10941517e+00
-8.96180689e-01 4.62327540e-01 -1.29681854e-02 -1.34047604e+00
3.46668631e-01 -8.77590060e-01 -8.89316857e-01 8.09245765e-01
-1.74799120e+00 -1.34024906e+00 -8.09878588e-01 7.59828925e-01
6.13135397e-01 -2.91324615e-01 2.60640025e-01 6.85871124e-01
-8.18588078e-01 1.03529370e+00 5.85788786e-02 -8.65322351e-02
5.81424713e-01 -7.56664395e-01 9.40670431e-01 9.89998579e-01
-5.06345093e-01 1.48071557e-01 5.29089808e-01 -8.20883453e-01
-1.44258857e+00 -1.55977774e+00 8.85174811e-01 -8.16092968e-01
3.60768944e-01 -4.54822987e-01 -4.76577759e-01 6.43463671e-01
1.03534892e-01 -1.44769922e-01 1.31381452e-01 -2.85914570e-01
-4.08260785e-02 -5.54184258e-01 -1.22701192e+00 5.55288434e-01
9.48470950e-01 -5.67371309e-01 -2.96847433e-01 5.36202967e-01
4.12806332e-01 -3.61830771e-01 1.10164903e-01 4.43383530e-02
7.54460990e-01 -8.31874907e-01 1.19430530e+00 -2.23529875e-01
-3.88459682e-01 -4.88469273e-01 3.33734691e-01 -1.12088287e+00
-4.42411005e-01 -6.22446716e-01 7.56390691e-02 8.02832246e-01
4.01977330e-01 -7.21543789e-01 8.92396808e-01 7.10477531e-01
-6.61567748e-01 5.09314574e-02 -1.33565497e+00 -9.74858940e-01
-4.07750964e-01 -1.09355819e+00 5.83486736e-01 4.22372520e-02
-2.72684872e-01 2.39656940e-01 -2.98936576e-01 5.62572241e-01
6.31831110e-01 -5.56866646e-01 7.74172604e-01 -8.94524515e-01
6.29745901e-01 -4.36149597e-01 -7.06807256e-01 -7.72524714e-01
-2.41921961e-01 -7.28004217e-01 4.21627373e-01 -1.72018862e+00
-1.64959520e-01 -3.19377989e-01 1.79348513e-02 4.81627077e-01
-9.50438529e-02 6.60744488e-01 -7.13893846e-02 -2.68757403e-01
-4.74852890e-01 5.31474985e-02 1.02332461e+00 -4.24120873e-01
-9.42559987e-02 4.83551353e-01 -6.36962712e-01 4.14046496e-01
5.97398102e-01 -4.96749997e-01 -2.02251464e-01 -5.47013938e-01
1.03029393e-01 -3.24895740e-01 1.02036691e+00 -1.25029087e+00
8.29199016e-01 1.48291424e-01 1.55613139e-01 -9.76309478e-01
4.28427011e-01 -9.96227980e-01 -3.58588882e-02 6.80548370e-01
2.32198566e-01 1.11331688e-02 8.50186765e-01 8.36439788e-01
5.19988835e-01 1.77584559e-01 7.06551731e-01 3.78190219e-01
-1.28564668e+00 6.95736259e-02 -1.11903286e+00 -3.04987222e-01
1.51582778e+00 -6.44789636e-01 -7.27243781e-01 -2.62566358e-01
-3.77652258e-01 6.17434621e-01 -1.40147433e-01 9.06075239e-01
6.78009272e-01 -1.18099964e+00 -8.04583430e-01 5.55833101e-01
5.37592947e-01 -4.49634910e-01 5.35775900e-01 9.32381451e-01
-7.30372369e-01 9.90182638e-01 -3.04575145e-01 -8.13195407e-01
-1.12853754e+00 7.97363818e-01 4.33748364e-01 5.36797822e-01
-7.60974944e-01 5.72383583e-01 -7.11371973e-02 -2.40920201e-01
4.93665367e-01 -4.04826164e-01 -2.09362328e-01 1.30945072e-01
7.51488984e-01 1.18286288e+00 5.45180500e-01 -9.29163873e-01
-8.94720852e-01 7.15689361e-01 1.31898358e-01 3.44556272e-01
6.05979085e-01 -5.23106158e-01 6.79135025e-01 -3.50659877e-01
1.00859833e+00 -6.58787549e-01 -1.49459279e+00 2.30441570e-01
-2.78349459e-01 -3.62996906e-01 3.67858887e-01 -1.10738027e+00
-1.11238861e+00 9.00540948e-01 1.57265079e+00 9.67888758e-02
8.77381265e-01 4.79684919e-02 1.13055170e+00 3.39616239e-01
1.09756872e-01 -1.07918227e+00 -1.88834965e-01 2.13155076e-01
6.31157815e-01 -1.75682569e+00 -3.43940437e-01 -1.27720922e-01
-8.31744611e-01 7.11106896e-01 7.34811127e-01 1.63350746e-01
4.89603966e-01 2.58797765e-01 3.57909828e-01 -3.96612853e-01
-1.06087461e-01 -7.30847299e-01 3.11825722e-01 1.31556523e+00
-3.38956267e-01 3.99612814e-01 1.50964692e-01 3.06597143e-01
9.92169157e-02 1.50268450e-01 2.82614052e-01 6.71459198e-01
-5.72568536e-01 -1.66488722e-01 -5.31981170e-01 5.19360542e-01
1.58843219e-01 1.37152234e-02 -7.51905069e-02 9.11153615e-01
3.45546424e-01 1.74810243e+00 2.76789933e-01 -5.24211109e-01
8.81702304e-01 -2.41715625e-01 -1.26394242e-01 -2.31690675e-01
-4.36518729e-01 -4.40282673e-01 3.34764749e-01 -7.53704488e-01
1.19185699e-02 -5.01511455e-01 -1.00408399e+00 -7.06106961e-01
-2.01488048e-01 -4.24041599e-01 1.10039628e+00 9.73930955e-01
5.04119098e-01 5.89098334e-01 4.63278830e-01 -1.09321237e+00
-1.19416498e-01 -8.38208437e-01 -4.56765503e-01 -1.88694954e-01
5.93586981e-01 -7.83656478e-01 -3.04754138e-01 -1.35818690e-01]
|
[7.918096542358398, -0.9056409001350403]
|
beecf4e9-7000-4cde-a3c7-bcc71b74d946
|
interpretable-anomaly-detection-in-cellular
|
2306.15938
| null |
https://arxiv.org/abs/2306.15938v1
|
https://arxiv.org/pdf/2306.15938v1.pdf
|
Interpretable Anomaly Detection in Cellular Networks by Learning Concepts in Variational Autoencoders
|
This paper addresses the challenges of detecting anomalies in cellular networks in an interpretable way and proposes a new approach using variational autoencoders (VAEs) that learn interpretable representations of the latent space for each Key Performance Indicator (KPI) in the dataset. This enables the detection of anomalies based on reconstruction loss and Z-scores. We ensure the interpretability of the anomalies via additional information centroids (c) using the K-means algorithm to enhance representation learning. We evaluate the performance of the model by analyzing patterns in the latent dimension for specific KPIs and thereby demonstrate the interpretability and anomalies. The proposed framework offers a faster and autonomous solution for detecting anomalies in cellular networks and showcases the potential of deep learning-based algorithms in handling big data.
|
['Markus Lange-Hegermann', 'Michael Weber', 'Amandeep Singh']
|
2023-06-28
| null | null | null | null |
['anomaly-detection']
|
['methodology']
|
[-1.12265734e-04 3.48366201e-01 3.03540409e-01 -1.38389692e-01
-6.76046789e-01 -4.29234385e-01 6.48833752e-01 3.83087814e-01
1.09288126e-01 5.70457458e-01 2.84199268e-01 -1.86339572e-01
-8.82432342e-01 -7.96146154e-01 -5.25609910e-01 -1.15147853e+00
-3.59273851e-01 8.62164557e-01 -2.62941808e-01 -2.65253056e-03
2.62532830e-01 9.45673466e-01 -1.34890771e+00 4.11685884e-01
4.75302517e-01 1.24871671e+00 -5.06937385e-01 1.04777348e+00
6.84830360e-03 6.39020264e-01 -7.04279125e-01 -2.92284817e-01
8.05425048e-02 -9.28818360e-02 -5.76440632e-01 3.23287725e-01
5.12117706e-02 -1.74417466e-01 -3.80409926e-01 6.87918484e-01
1.48215458e-01 -2.61274993e-01 1.14592993e+00 -1.66824937e+00
-5.88632166e-01 2.90764898e-01 -4.11520064e-01 7.12091029e-01
-2.42706597e-01 -5.55895567e-02 1.24464214e+00 -6.29000962e-01
5.05075753e-01 1.12977147e+00 8.02849412e-01 3.81182015e-01
-1.42477083e+00 -2.17596084e-01 -1.91007983e-02 3.32294673e-01
-1.22638178e+00 -4.08404917e-01 5.73397577e-01 -7.19375074e-01
8.99706066e-01 4.57985461e-01 3.17254484e-01 1.16298246e+00
4.51486379e-01 5.33281028e-01 5.27641594e-01 -2.09957615e-01
5.15456438e-01 1.19672427e-02 3.45745057e-01 7.25191593e-01
2.91858703e-01 5.74198328e-02 -3.21170270e-01 -5.59609354e-01
6.79720879e-01 4.98914450e-01 1.41612157e-01 -2.30556384e-01
-1.13967633e+00 1.03413010e+00 7.80578330e-02 2.42289633e-01
-6.82272673e-01 6.49937212e-01 6.74836695e-01 4.01978314e-01
6.32048130e-01 4.65661794e-01 -5.75185478e-01 5.10605574e-02
-1.05507731e+00 -7.39890262e-02 3.39806378e-01 3.71231586e-01
4.51551318e-01 3.05060416e-01 -5.82506418e-01 3.06121439e-01
6.26666546e-01 1.99194655e-01 3.75188649e-01 -1.31464958e+00
3.92525084e-03 8.31705451e-01 -2.16599107e-01 -1.20586145e+00
-6.47299111e-01 -7.53098905e-01 -1.06444311e+00 3.55341345e-01
2.89954633e-01 -6.33130595e-02 -7.85937965e-01 1.55249047e+00
2.03589901e-01 6.80681705e-01 1.53802574e-01 3.09880376e-01
4.20490712e-01 6.82700276e-01 -5.58823310e-02 -3.52171659e-01
1.18594909e+00 -4.46153700e-01 -8.68507385e-01 6.38381481e-01
6.54567838e-01 6.34083971e-02 4.39910442e-01 4.16733593e-01
-7.57348597e-01 -2.16918096e-01 -7.80866325e-01 2.38210037e-01
-5.95875323e-01 -5.54819405e-03 5.29070258e-01 6.93650901e-01
-1.22124076e+00 5.60690939e-01 -1.11669540e+00 -3.78756583e-01
7.59043813e-01 4.92168963e-01 -2.80859500e-01 4.83928025e-01
-8.02219510e-01 1.65951312e-01 1.99924365e-01 5.00347205e-02
-1.30152512e+00 -7.99773455e-01 -6.35738015e-01 6.35304213e-01
-3.10489852e-02 -5.83655477e-01 3.88735682e-01 -7.89683580e-01
-1.11106181e+00 7.39558458e-01 -3.29880714e-01 -8.67649794e-01
5.29202223e-01 1.46799132e-01 -5.34034133e-01 3.55455279e-01
-2.23059192e-01 2.39826053e-01 1.07362032e+00 -1.32163489e+00
-5.92787802e-01 -5.16670108e-01 -3.58792484e-01 -6.09873772e-01
-6.84820473e-01 -2.35696539e-01 -2.33113274e-01 -6.78197503e-01
2.88077652e-01 -8.39673102e-01 -3.17104995e-01 -1.29771143e-01
-6.94382429e-01 -2.63367504e-01 1.08510566e+00 -8.26056957e-01
1.29019225e+00 -2.05143476e+00 1.48732260e-01 9.00362492e-01
1.05933976e+00 -1.87362373e-01 -1.26019686e-01 3.25860858e-01
-2.30187550e-01 1.65121749e-01 -2.24788621e-01 -6.01195872e-01
-5.47767617e-02 6.63295507e-01 -5.78281462e-01 6.14820004e-01
5.13479412e-01 9.33524907e-01 -6.16534293e-01 -1.63519964e-01
1.74108237e-01 4.90929037e-01 -7.64294088e-01 1.11446688e-02
-3.59871760e-02 5.73358238e-01 -2.86018759e-01 7.08627582e-01
4.93249774e-01 -5.69423199e-01 8.33013356e-02 1.81003615e-01
1.32074982e-01 -5.13017118e-01 -1.00598514e+00 9.38639641e-01
-1.89579532e-01 1.06045735e+00 -2.10470617e-01 -1.37776685e+00
7.94931710e-01 4.92111713e-01 9.63943183e-01 -3.44793409e-01
-5.28571010e-02 -1.51560113e-01 -3.12197149e-01 -3.72696072e-01
-8.09483696e-03 2.68008441e-01 -5.65733528e-03 5.46559513e-01
2.39912897e-01 8.33475590e-01 -1.33096069e-01 3.49422961e-01
1.40744805e+00 -5.96739948e-01 7.38060474e-03 -3.63429755e-01
7.36174822e-01 -5.21652639e-01 5.36940753e-01 8.72947395e-01
-3.25097084e-01 4.70803887e-01 1.20172346e+00 -7.79758275e-01
-1.34263635e+00 -1.38769269e+00 -3.45193624e-01 9.12345409e-01
-6.17822945e-01 -2.57416755e-01 -6.92708373e-01 -7.87008584e-01
2.68410712e-01 6.44674122e-01 -1.25191987e+00 -2.36888587e-01
3.43907028e-02 -1.13000751e+00 6.13056540e-01 5.40183663e-01
-1.53092071e-01 -7.51798987e-01 -3.16112250e-01 1.07077265e-03
-9.24121514e-02 -9.16127205e-01 3.58412713e-01 1.36247441e-01
-9.79176283e-01 -8.60597014e-01 -2.52342731e-01 -2.10865349e-01
9.98092353e-01 -3.20108831e-01 1.04773915e+00 1.36276618e-01
-3.42872620e-01 7.35088110e-01 -9.19148512e-03 -5.62024713e-01
-5.91963828e-01 -1.57412082e-01 4.57483649e-01 4.73474771e-01
3.18745315e-01 -6.53025687e-01 -5.27770817e-01 1.20708667e-01
-9.40850019e-01 -6.63564980e-01 2.35548124e-01 9.21136320e-01
8.34215641e-01 4.61898506e-01 8.39197516e-01 -8.08098018e-01
6.90063953e-01 -9.65409338e-01 -7.51293302e-01 9.93112400e-02
-9.87440348e-01 3.33943218e-01 4.35312152e-01 6.08058050e-02
-6.15773380e-01 -2.49460995e-01 -9.59843695e-02 -7.65045524e-01
-3.60310107e-01 2.18212157e-01 2.33422905e-01 1.55502319e-01
5.69137514e-01 2.94554323e-01 2.41087899e-02 -2.55426705e-01
6.29402548e-02 5.51742911e-01 3.08281213e-01 -1.73862308e-01
7.60000765e-01 1.01183808e+00 5.37212610e-01 -8.20610940e-01
-6.28261864e-01 -5.22240400e-01 -6.63705707e-01 -2.53430009e-01
8.70898545e-01 -8.40904534e-01 -9.02943194e-01 2.30075642e-01
-8.87422025e-01 1.93865389e-01 -5.79883754e-01 1.27956510e-01
-6.63243949e-01 3.87768418e-01 -8.18850935e-01 -9.79252696e-01
-3.96406919e-01 -1.04647505e+00 1.20831990e+00 -1.74421415e-01
-3.32563072e-01 -1.36466718e+00 4.30569679e-01 3.10610205e-01
3.37080538e-01 6.27527297e-01 1.20379877e+00 -1.09352791e+00
-6.99567199e-01 -5.04269361e-01 -3.00939918e-01 1.78012714e-01
-2.34382302e-01 5.21290362e-01 -1.49278760e+00 -3.37995023e-01
-3.26991051e-01 2.75899410e-01 1.07146478e+00 9.08024073e-01
1.72525895e+00 -4.64329302e-01 -4.48082775e-01 6.70274138e-01
1.33592856e+00 -2.23625973e-01 7.18699396e-01 1.91927701e-01
5.63789189e-01 6.28882468e-01 -1.55728698e-01 9.73444939e-01
1.24460429e-01 4.37757760e-01 8.06743622e-01 -1.72336996e-01
4.43601936e-01 6.69611618e-02 2.13042989e-01 6.48757756e-01
-1.12211056e-01 -3.35746557e-01 -9.41117942e-01 6.13315642e-01
-1.98300683e+00 -1.17803514e+00 -2.06192121e-01 1.82813454e+00
-5.75136393e-02 6.34170100e-02 5.01904935e-02 4.60663825e-01
4.55376029e-01 -1.97571427e-01 -5.13585448e-01 -5.73494732e-01
-1.73417300e-01 -2.21691623e-01 2.49393314e-01 3.82574558e-01
-9.63231683e-01 3.77456039e-01 7.43686914e+00 6.15909219e-01
-7.38676190e-01 1.38884364e-02 1.18320966e+00 -4.43241186e-02
-4.63811696e-01 -3.82050693e-01 -5.29244363e-01 5.10547042e-01
1.47198236e+00 2.21263856e-01 3.25236768e-01 6.93435788e-01
4.72311527e-01 4.20706004e-01 -1.06036615e+00 9.25209343e-01
-2.41381556e-01 -1.76385522e+00 2.66172737e-01 6.91633046e-01
8.30012798e-01 1.25388011e-01 5.92364669e-01 1.50788605e-01
1.19889893e-01 -1.12994409e+00 1.81691706e-01 1.08306444e+00
5.06873488e-01 -1.12396216e+00 9.16631103e-01 -1.60622634e-02
-5.74998856e-01 -7.14396000e-01 -5.76179862e-01 8.29689279e-02
-2.20822915e-01 9.86576915e-01 -1.05759931e+00 1.99860752e-01
6.87381387e-01 7.15804636e-01 -6.88931048e-01 7.46695459e-01
3.80950332e-01 8.27276886e-01 -3.34960192e-01 4.79607821e-01
2.37844244e-01 -3.13983887e-01 8.54469597e-01 1.07421112e+00
6.05040014e-01 -4.48076516e-01 -3.93287271e-01 9.63007033e-01
-1.47509640e-02 -1.03736356e-01 -6.85503244e-01 -1.96544409e-01
7.69942030e-02 1.42185915e+00 -8.49060059e-01 -3.15854371e-01
-9.95908678e-02 9.00028288e-01 1.31244242e-01 4.06768203e-01
-5.59270680e-01 1.02645643e-01 1.12124777e+00 4.35493402e-02
4.47136223e-01 9.47616398e-02 -5.52376807e-01 -9.76203322e-01
-3.48338425e-01 -5.41997969e-01 9.63953972e-01 -6.26520038e-01
-1.31196332e+00 4.08965230e-01 -4.36870188e-01 -1.15605319e+00
-4.57984686e-01 -8.88119578e-01 -8.03490996e-01 3.87461960e-01
-1.35502303e+00 -9.75423455e-01 -2.07732171e-01 5.07527411e-01
3.02859783e-01 -6.56226814e-01 1.03097844e+00 7.54131451e-02
-1.04078805e+00 7.44842887e-01 9.48726714e-01 1.37267217e-01
1.27537981e-01 -1.56742072e+00 2.23639652e-01 7.33239591e-01
4.78926718e-01 3.09251994e-01 9.14695144e-01 -2.28147238e-01
-7.30163336e-01 -1.05718613e+00 6.61092758e-01 -1.13723302e+00
6.38781667e-01 -1.96392402e-01 -7.92536557e-01 7.83427536e-01
-1.00837916e-01 1.41078860e-01 1.39057076e+00 3.10013026e-01
-2.08699763e-01 -1.25274166e-01 -1.29177618e+00 3.79331887e-01
5.23483872e-01 -5.73303103e-01 -3.21533298e-03 4.34489965e-01
4.85241920e-01 2.74313033e-01 -9.87757504e-01 2.12462485e-01
3.30944926e-01 -1.04891372e+00 1.03981435e+00 -1.23652315e+00
4.34430510e-01 -2.66970605e-01 -2.53975391e-01 -1.20714366e+00
-6.04377985e-01 -2.86553413e-01 -9.07770753e-01 1.04066145e+00
4.03842628e-01 -5.61587393e-01 9.63525534e-01 3.40413660e-01
1.15345202e-01 -8.32633674e-01 -1.19467187e+00 -2.51441747e-01
-5.63470796e-02 -6.37858212e-01 7.17657089e-01 9.95036542e-01
-2.89838631e-02 -1.90096542e-01 -2.62223035e-01 7.81916916e-01
8.95817339e-01 -1.83845624e-01 3.88571709e-01 -1.79867160e+00
-2.83496171e-01 -4.36043531e-01 -1.17101586e+00 -1.45374715e-01
3.06647778e-01 -8.89201880e-01 -7.19010472e-01 -1.09736776e+00
1.84185177e-01 -6.90579042e-02 -1.05389261e+00 2.44860664e-01
1.83234230e-01 3.67266178e-01 -2.96262145e-01 2.46580750e-01
-8.94352496e-01 4.13102001e-01 4.65521872e-01 -1.72031641e-01
-4.93980832e-02 1.75733462e-01 -7.18214154e-01 6.43559635e-01
8.14100683e-01 -3.51251185e-01 -2.79722095e-01 -9.73854512e-02
4.76537347e-01 -3.14447656e-02 5.96859694e-01 -1.09711218e+00
-7.49745071e-02 2.68482268e-01 9.81832504e-01 -5.82712650e-01
1.16804145e-01 -9.15075600e-01 -8.42394866e-03 5.25831282e-01
-5.05194783e-01 1.69402316e-01 1.42667517e-01 1.27609825e+00
-4.47362736e-02 1.48031279e-01 5.80758333e-01 2.66534001e-01
-4.75839257e-01 4.63864207e-01 -9.03058112e-01 -2.79313684e-01
1.24996924e+00 -4.46497984e-02 -1.25093432e-02 -6.08304024e-01
-1.10773730e+00 3.66532147e-01 1.39923453e-01 1.45915300e-01
7.80577183e-01 -1.36022854e+00 -8.68311882e-01 5.62702894e-01
4.60886270e-01 -4.31564718e-01 4.77727264e-01 9.29781437e-01
-6.87478721e-01 4.52390790e-01 -3.88611078e-01 -8.94425213e-01
-1.15984583e+00 7.09130228e-01 6.63870156e-01 -6.41697645e-01
-4.70935762e-01 7.77478397e-01 2.03393400e-01 -2.46977717e-01
1.45609066e-01 -1.33570194e-01 -6.34583950e-01 3.38541538e-01
5.88312089e-01 6.98788881e-01 -2.31606457e-02 -5.74327171e-01
-1.79493412e-01 1.28679767e-01 -9.57116038e-02 1.60403654e-01
1.55698228e+00 -3.79942417e-01 -2.00420856e-01 3.61944765e-01
1.30837917e+00 -2.23802194e-01 -1.33495879e+00 -1.61099866e-01
3.03323805e-01 -6.17334694e-02 3.19344163e-01 -4.31159019e-01
-1.30857873e+00 9.73581076e-01 1.01719010e+00 4.80738491e-01
9.64128196e-01 2.71711230e-01 4.15966392e-01 5.39940357e-01
-3.66658211e-01 -1.15702331e+00 1.56854734e-01 2.90555120e-01
4.44472015e-01 -1.14446282e+00 -2.50032336e-01 1.15804128e-01
-4.02974457e-01 1.37430990e+00 7.24973977e-02 -8.68359059e-02
8.71456206e-01 9.82561633e-02 2.80728430e-01 -7.23276496e-01
-1.01007473e+00 1.21539429e-01 2.00740501e-01 7.72012651e-01
6.77831098e-02 7.09034950e-02 3.73725951e-01 4.91908193e-01
1.76184624e-01 -6.22843266e-01 6.31522655e-01 1.67680740e-01
-3.03382546e-01 -5.44099391e-01 -4.29621756e-01 9.12967026e-01
-6.15616977e-01 2.27185607e-01 -2.38887593e-01 1.83038432e-02
-4.04159799e-02 7.85888970e-01 5.81986129e-01 -2.08577886e-01
-9.84536633e-02 3.46956223e-01 -2.17541829e-01 5.48128523e-02
-4.75204550e-02 -4.22454998e-02 -3.52665007e-01 -7.52885997e-01
2.37300806e-02 -7.98306942e-01 -1.17423201e+00 -3.24171633e-01
-6.26608133e-02 5.66171035e-02 4.04567391e-01 1.06695747e+00
7.33213723e-01 1.01468801e+00 9.57448065e-01 -4.08711135e-01
-3.55556846e-01 -5.21411240e-01 -8.74231935e-01 6.81176186e-01
8.16592455e-01 -5.48721433e-01 -5.58389366e-01 2.28237212e-02]
|
[7.6081743240356445, 2.446704149246216]
|
ad5ea8cd-8988-49c3-9ff6-d2ead7d95df7
|
joint-texture-and-geometry-optimization-for
| null | null |
http://openaccess.thecvf.com/content_CVPR_2020/html/Fu_Joint_Texture_and_Geometry_Optimization_for_RGB-D_Reconstruction_CVPR_2020_paper.html
|
http://openaccess.thecvf.com/content_CVPR_2020/papers/Fu_Joint_Texture_and_Geometry_Optimization_for_RGB-D_Reconstruction_CVPR_2020_paper.pdf
|
Joint Texture and Geometry Optimization for RGB-D Reconstruction
|
Due to inevitable noises and quantization error, the reconstructed 3D models via RGB-D sensors always accompany geometric error and camera drifting, which consequently lead to blurring and unnatural texture mapping results. Most of the 3D reconstruction methods focus on either geometry refinement or texture improvement respectively, which subjectively decouples the inter-relationship between geometry and texture. In this paper, we propose a novel approach that can jointly optimize the camera poses, texture and geometry of the reconstructed model, and color consistency between the key-frames. Instead of computing Shape-From-Shading (SFS) expensively, our method directly optimizes the reconstructed mesh according to color and geometric consistency and high-boost normal cues, which can effectively overcome the texture-copy problem generated by SFS and achieve more detailed shape reconstruction. As the joint optimization involves multiple correlated terms, therefore, we further introduce an iterative framework to interleave the optimal state. The experiments demonstrate that our method can recover not only fine-scale geometry but also high-fidelity texture.
|
[' Chunxia Xiao', ' Jie Liao', ' Qingan Yan', 'Yanping Fu']
|
2020-06-01
| null | null | null |
cvpr-2020-6
|
['rgb-d-reconstruction']
|
['computer-vision']
|
[ 2.33949184e-01 -2.67600983e-01 3.85545880e-01 -2.88251191e-01
-6.40108049e-01 -3.78405809e-01 2.63806611e-01 -2.03839064e-01
6.78116679e-02 3.18628132e-01 1.06971487e-01 1.17344238e-01
5.25238924e-02 -8.31888318e-01 -5.85333109e-01 -8.68494987e-01
7.16905415e-01 2.80611575e-01 4.73231912e-01 8.06234106e-02
2.85368085e-01 5.30342460e-01 -1.47884500e+00 -1.04588538e-01
9.74342704e-01 1.20393813e+00 3.37074816e-01 4.82996225e-01
-1.43100485e-01 4.42983121e-01 -1.77458256e-01 -1.65825292e-01
3.06418300e-01 -1.85010642e-01 -2.99523681e-01 4.71787602e-01
4.06006873e-01 -6.59652770e-01 -4.09814686e-01 1.35991788e+00
4.11140978e-01 -4.69425470e-02 3.16586792e-01 -7.79623151e-01
-4.83807445e-01 -4.56972629e-01 -8.35640073e-01 -5.34500420e-01
5.48327863e-01 1.79239556e-01 3.88709575e-01 -1.04029238e+00
5.72608531e-01 1.35699975e+00 6.41755879e-01 2.33724371e-01
-1.20268917e+00 -6.28789127e-01 9.53916535e-02 5.86905740e-02
-1.66646504e+00 -5.53358734e-01 1.18090129e+00 -2.74599880e-01
2.63359219e-01 4.70056742e-01 8.24463248e-01 7.22824335e-01
9.95850563e-03 4.19119298e-01 1.23375058e+00 -2.96899974e-01
1.70229718e-01 -1.11214995e-01 -6.11830592e-01 7.22754061e-01
1.38588458e-01 2.97237456e-01 -5.44787765e-01 -8.97087231e-02
1.50564420e+00 2.39326298e-01 -7.53910363e-01 -6.05276644e-01
-1.42101192e+00 1.66568920e-01 1.59267694e-01 -1.66466236e-01
-2.18609363e-01 1.10735171e-01 -6.21159673e-02 -5.67416660e-02
6.36271775e-01 -6.22606277e-03 -3.88266593e-01 -3.08250070e-01
-6.95732713e-01 7.25179911e-02 3.79660457e-01 1.09944999e+00
1.15654397e+00 -8.69325772e-02 6.65442925e-03 7.97014892e-01
8.12116206e-01 9.62519884e-01 -5.02639189e-02 -1.23896301e+00
4.12457019e-01 6.39079392e-01 4.51322794e-01 -1.49282587e+00
-1.55705484e-02 -6.88394532e-02 -1.00274360e+00 2.04622447e-01
2.29707137e-01 2.50168353e-01 -8.21174085e-01 1.33746541e+00
8.33849311e-01 3.51159424e-01 -4.67901170e-01 1.21228135e+00
7.03257143e-01 4.47386384e-01 -5.34469724e-01 -4.14407104e-01
9.39862072e-01 -7.24019229e-01 -9.48060811e-01 -1.33437291e-02
-2.50015631e-02 -1.28350890e+00 1.04067492e+00 3.10624570e-01
-1.13854456e+00 -3.93488020e-01 -1.03585792e+00 -3.82429212e-01
4.45927382e-01 4.43240553e-02 2.28735760e-01 3.65467131e-01
-8.74516249e-01 5.42730510e-01 -1.02311420e+00 7.46654943e-02
1.15561269e-01 1.48860976e-01 -1.54531851e-01 -4.22357887e-01
-7.74287701e-01 6.63039625e-01 -2.80962256e-03 4.43975836e-01
-4.95875955e-01 -5.42750776e-01 -6.77256584e-01 -2.90508866e-01
6.09054506e-01 -6.99796259e-01 9.37505305e-01 -7.33239114e-01
-2.00237346e+00 7.67182291e-01 -3.58253598e-01 7.01188207e-01
5.62675118e-01 -1.47165403e-01 -2.68040091e-01 3.25038508e-02
2.61326674e-02 1.42950162e-01 8.91214490e-01 -1.66696966e+00
-3.50433975e-01 -4.80702609e-01 -2.26598263e-01 5.21832168e-01
-3.97131518e-02 -3.67783725e-01 -1.12162054e+00 -6.63536966e-01
8.48062456e-01 -5.41260421e-01 -2.31630638e-01 5.19252419e-01
-3.34316492e-01 4.68462944e-01 8.59268546e-01 -6.94886863e-01
1.12250841e+00 -2.31984019e+00 2.76568651e-01 2.12860018e-01
2.52269059e-01 -2.03694388e-01 7.48195574e-02 3.70001905e-02
3.20582420e-01 -1.26141712e-01 -3.56759936e-01 -4.73806977e-01
-1.46957591e-01 5.02215862e-01 -9.37859192e-02 6.37814820e-01
-3.59419435e-02 7.27475047e-01 -8.79009545e-01 -5.84494174e-01
5.70870161e-01 7.72875428e-01 -4.74171758e-01 4.19156462e-01
-1.38122335e-01 9.33638096e-01 -7.98609853e-01 8.50488961e-01
1.30182648e+00 -1.79849818e-01 -7.22805560e-02 -5.87644577e-01
-3.11140865e-01 4.23339717e-02 -1.53478825e+00 2.10850620e+00
-3.21277380e-01 -1.21554276e-02 5.41904032e-01 -4.46295857e-01
1.00814259e+00 1.38608202e-01 5.24963975e-01 -9.06815112e-01
1.36279285e-01 3.90677631e-01 -6.34741843e-01 -4.38716620e-01
4.59112793e-01 -2.10630789e-01 2.91454762e-01 1.91251546e-01
-5.48218668e-01 -5.50366402e-01 -6.40597641e-01 -2.03812703e-01
6.24700546e-01 6.93898082e-01 -8.87765810e-02 -1.87896028e-01
6.27095699e-01 -2.99922049e-01 8.51360440e-01 2.54270554e-01
1.44286722e-01 1.12801194e+00 1.65546894e-01 -3.40769351e-01
-1.26099646e+00 -1.06241643e+00 -1.85261279e-01 2.54798770e-01
9.57121491e-01 -3.72590214e-01 -6.66223943e-01 -1.72167867e-01
-2.14723676e-01 2.35709026e-01 -3.61129105e-01 -1.15139917e-01
-7.09032953e-01 -5.63945711e-01 -1.43382192e-01 1.37061356e-02
6.13997638e-01 -4.31519240e-01 -3.36576253e-01 2.35370696e-01
-2.81319618e-01 -1.10841942e+00 -8.13664198e-01 -4.46059793e-01
-1.10411644e+00 -1.08260214e+00 -7.32616663e-01 -4.00494456e-01
8.35341036e-01 5.45907557e-01 1.05173492e+00 4.06755060e-01
-6.81295171e-02 2.23171830e-01 -3.16128522e-01 1.34130523e-01
-1.73861489e-01 -4.22903776e-01 -1.08674243e-01 4.73577499e-01
-1.95163503e-01 -7.96618283e-01 -9.36528563e-01 7.38504171e-01
-9.68210161e-01 7.98848629e-01 3.59604239e-01 6.20791554e-01
1.08718765e+00 4.04851846e-02 -4.05189723e-01 -3.97424281e-01
-9.95963928e-04 -7.79635506e-03 -8.78632009e-01 2.72867769e-01
-6.39981449e-01 -4.57868613e-02 3.48420650e-01 -2.04278395e-01
-1.24969816e+00 2.59978622e-01 -9.73427203e-03 -8.91583741e-01
-6.65513426e-02 1.51419729e-01 -5.19032896e-01 -3.40349674e-01
7.43555874e-02 4.61709738e-01 -4.51483019e-02 -8.20034206e-01
1.02828801e-01 4.52473223e-01 5.28376877e-01 -7.15340018e-01
1.04400516e+00 7.36126721e-01 1.52161151e-01 -6.26838326e-01
-8.01394820e-01 -1.52935773e-01 -6.07459843e-01 -4.49021906e-01
7.84376204e-01 -9.63739216e-01 -7.85338461e-01 8.53214681e-01
-1.37201393e+00 -2.28354767e-01 2.52249371e-02 5.29510319e-01
-3.50800663e-01 8.40738595e-01 -6.58102274e-01 -7.22013116e-01
-1.35747734e-02 -1.34424686e+00 1.61488521e+00 2.48389274e-01
3.74525309e-01 -6.02074802e-01 -6.50999993e-02 2.12364390e-01
3.48388612e-01 3.33942622e-01 6.91940069e-01 8.74372363e-01
-1.27721083e+00 9.59000289e-02 -5.71237981e-01 6.47241846e-02
3.48258018e-01 4.82456665e-03 -8.99143636e-01 -2.01891080e-01
3.79239082e-01 8.27284753e-02 1.72239050e-01 1.93014160e-01
1.24760234e+00 -4.72059101e-02 -2.79278271e-02 1.14526188e+00
1.64345574e+00 1.19944438e-02 6.94692969e-01 3.01075757e-01
1.35409105e+00 4.05662805e-01 8.66167426e-01 5.79733551e-01
6.18986011e-01 1.08076406e+00 5.54016054e-01 -2.57924318e-01
-2.18516842e-01 -4.26916242e-01 1.06018782e-01 1.30222380e+00
-3.50380778e-01 8.26302022e-02 -6.78752542e-01 -3.38520668e-02
-1.83756399e+00 -4.15278614e-01 -4.93030250e-01 2.53753400e+00
1.00608003e+00 -2.07525864e-01 -5.52442789e-01 -1.24567404e-01
7.57574320e-01 1.03209689e-01 -6.51531160e-01 1.28483057e-01
-1.77529290e-01 2.42403001e-02 6.42499685e-01 7.96107531e-01
-6.03467166e-01 7.46295094e-01 6.02581930e+00 9.96613562e-01
-1.17150605e+00 2.39608269e-02 6.34561360e-01 -6.39256686e-02
-8.56671035e-01 1.85901776e-01 -4.61340427e-01 6.62950873e-01
1.73479542e-01 2.00236008e-01 7.86013722e-01 4.95365769e-01
3.75115842e-01 -2.75785595e-01 -7.61261284e-01 1.49541068e+00
-9.62926149e-02 -1.01575267e+00 -2.80964542e-02 2.15222478e-01
9.07408595e-01 -2.52626657e-01 -2.09272608e-01 -3.84996682e-01
7.16689900e-02 -8.39218438e-01 1.08378959e+00 8.72748971e-01
1.15613353e+00 -5.33037961e-01 4.66847003e-01 2.72958755e-01
-1.36946762e+00 4.88263428e-01 -3.89866084e-01 7.64304772e-02
4.57416981e-01 9.03485775e-01 2.81801187e-02 8.79178047e-01
7.64439762e-01 9.07247305e-01 -3.45793217e-01 8.01090717e-01
-3.14804018e-01 3.32826748e-02 -4.78398412e-01 3.52788746e-01
-2.49632105e-01 -7.43523538e-01 4.98538941e-01 4.85068679e-01
5.55294931e-01 4.70175922e-01 4.13868517e-01 1.05035675e+00
2.71600097e-01 4.61863680e-03 -7.34159127e-02 3.47284138e-01
5.63235879e-01 1.15933740e+00 -6.58225596e-01 -3.04024100e-01
-4.60531324e-01 1.29234934e+00 1.22698329e-01 5.69305301e-01
-8.54537904e-01 2.82964204e-02 6.28104389e-01 2.60384530e-01
2.05631658e-01 -4.83156115e-01 -7.22151756e-01 -1.55566931e+00
4.58999842e-01 -6.77777112e-01 -3.35454613e-01 -1.06618047e+00
-1.05500710e+00 3.41141611e-01 -3.21839541e-01 -1.54003453e+00
3.98579717e-01 -2.55979687e-01 -2.13857949e-01 1.09636891e+00
-1.65909386e+00 -9.28918004e-01 -6.61126971e-01 7.14427769e-01
2.33272016e-01 5.48434377e-01 6.11994326e-01 5.30287981e-01
-5.27610540e-01 1.53399780e-01 1.94333091e-01 -3.95827740e-01
8.03030133e-01 -8.89484823e-01 1.10228166e-01 6.95268869e-01
-3.85838568e-01 2.89831251e-01 5.24414122e-01 -7.44436920e-01
-1.83589458e+00 -8.62352133e-01 4.74251062e-01 -3.57122034e-01
1.81845948e-01 -2.86875665e-01 -9.31185067e-01 2.23499164e-01
-3.01457465e-01 1.07973956e-01 3.90092023e-02 -4.00440633e-01
-1.64258376e-01 -2.42745146e-01 -1.06733477e+00 6.06379271e-01
1.21732926e+00 -7.14540005e-01 -1.30330190e-01 1.60159115e-02
6.96207881e-01 -1.07617950e+00 -9.42097366e-01 5.99027157e-01
5.92350781e-01 -1.32139826e+00 1.01477039e+00 3.44467610e-01
5.26333034e-01 -9.00752187e-01 -2.79284805e-01 -9.39492524e-01
-3.59730929e-01 -6.76633418e-01 -1.56529590e-01 1.29560030e+00
-9.97419953e-02 -4.75334555e-01 7.42001772e-01 8.07200313e-01
-2.69098580e-01 -6.23392045e-01 -9.80295479e-01 -3.01197737e-01
-5.20801663e-01 -3.69458914e-01 9.48853850e-01 1.07753539e+00
-6.04241610e-01 1.22962788e-01 -5.50673842e-01 4.81766254e-01
9.17638004e-01 4.90569055e-01 8.13674390e-01 -1.04090953e+00
-3.41546386e-01 -3.07325363e-01 -1.92321450e-01 -1.66548502e+00
-5.65804303e-01 -3.00427943e-01 3.62744272e-01 -1.22586441e+00
1.41142309e-01 -7.93541133e-01 -2.32506711e-02 5.10444641e-02
-2.64697254e-01 3.05776507e-01 -1.89298972e-01 3.73687208e-01
-4.96620297e-01 8.87833595e-01 1.80049801e+00 2.24791497e-01
-1.40532970e-01 -4.40593362e-01 -4.39553589e-01 7.72335410e-01
2.64997452e-01 -2.50751108e-01 -2.86091805e-01 -1.06962895e+00
2.90277064e-01 4.49970961e-01 4.08587396e-01 -6.41528964e-01
1.27493113e-01 -5.58329880e-01 5.69594026e-01 -6.51866257e-01
4.37340051e-01 -1.17618978e+00 5.36436677e-01 3.44596207e-02
2.64030159e-01 -1.43743485e-01 -1.18707538e-01 7.88175762e-01
-2.14020029e-01 1.47731349e-01 7.65056908e-01 -1.25236243e-01
-3.44540536e-01 8.89730632e-01 1.33948177e-01 -2.12512985e-01
6.24907136e-01 -5.05321085e-01 1.01147152e-01 -2.84168303e-01
-3.75430971e-01 1.00749776e-01 1.34947646e+00 2.79122353e-01
7.15666175e-01 -1.77360070e+00 -4.34289306e-01 5.74434638e-01
6.89198449e-02 7.63518870e-01 5.66870093e-01 9.01390314e-01
-8.70994508e-01 -1.50085032e-01 1.76318079e-01 -9.42384124e-01
-1.06247544e+00 2.42060259e-01 3.33797991e-01 2.75225192e-02
-7.62053072e-01 7.66923666e-01 3.74026567e-01 -4.79733407e-01
1.33990809e-01 -1.77375227e-01 4.26490396e-01 -5.50416291e-01
4.00853872e-01 3.34875882e-01 6.09305687e-02 -7.14398563e-01
-2.79591262e-01 1.27419770e+00 2.85968155e-01 -1.53459534e-01
1.21368778e+00 -7.60175109e-01 -4.00864244e-01 3.36831450e-01
1.19141543e+00 1.56941205e-01 -1.72421205e+00 -2.79315710e-01
-4.65693206e-01 -1.18000960e+00 3.80949974e-01 -3.62246186e-01
-1.26073778e+00 9.43758368e-01 5.39146543e-01 -1.51843444e-01
1.26045370e+00 -3.07734549e-01 1.10707879e+00 -1.03497863e-01
5.78770518e-01 -1.16061246e+00 -1.18236266e-01 4.26449329e-01
8.15089345e-01 -1.01840854e+00 3.51056576e-01 -8.02229583e-01
-1.18389569e-01 1.19822526e+00 5.75856984e-01 5.44149242e-02
5.72052836e-01 1.15271159e-01 6.39436841e-02 -2.86499500e-01
-3.18540484e-01 1.51391685e-01 3.48885357e-01 4.20736432e-01
1.82225794e-01 -5.01729958e-02 -4.59734276e-02 2.26023152e-01
7.88540691e-02 5.39320614e-03 1.98101938e-01 7.64844239e-01
-2.29639083e-01 -1.23815739e+00 -7.30894983e-01 7.77401552e-02
5.66101912e-03 -7.06484020e-02 -5.14688492e-02 2.92061806e-01
1.77907810e-01 6.99049354e-01 1.68768223e-02 -4.64236617e-01
5.36634684e-01 -4.44746912e-01 6.72748744e-01 -3.31703663e-01
1.68504894e-01 7.41267025e-01 -1.35454267e-01 -9.52852011e-01
-5.33723950e-01 -5.41406572e-01 -9.78538096e-01 -4.50176328e-01
-4.50409293e-01 -2.98559248e-01 5.28904915e-01 7.52323449e-01
4.06531364e-01 2.29376063e-01 1.00364184e+00 -1.01712942e+00
-1.67731047e-01 -6.64740801e-01 -8.23815167e-01 4.99039799e-01
2.38443911e-01 -7.30037868e-01 -4.22727376e-01 -6.71122894e-02]
|
[9.299057960510254, -2.8902955055236816]
|
daf9e9af-8301-4c3d-9bac-5d4a48132c04
|
libris2s-a-german-english-speech-to-speech
|
2204.10593
| null |
https://arxiv.org/abs/2204.10593v1
|
https://arxiv.org/pdf/2204.10593v1.pdf
|
LibriS2S: A German-English Speech-to-Speech Translation Corpus
|
Recently, we have seen an increasing interest in the area of speech-to-text translation. This has led to astonishing improvements in this area. In contrast, the activities in the area of speech-to-speech translation is still limited, although it is essential to overcome the language barrier. We believe that one of the limiting factors is the availability of appropriate training data. We address this issue by creating LibriS2S, to our knowledge the first publicly available speech-to-speech training corpus between German and English. For this corpus, we used independently created audio for German and English leading to an unbiased pronunciation of the text in both languages. This allows the creation of a new text-to-speech and speech-to-speech translation model that directly learns to generate the speech signal based on the pronunciation of the source language. Using this created corpus, we propose Text-to-Speech models based on the example of the recently proposed FastSpeech 2 model that integrates source language information. We do this by adapting the model to take information such as the pitch, energy or transcript from the source speech as additional input.
|
['Jan Niehues', 'Pedro Jeuris']
|
2022-04-22
| null |
https://aclanthology.org/2022.lrec-1.98
|
https://aclanthology.org/2022.lrec-1.98.pdf
|
lrec-2022-6
|
['speech-to-text-translation', 'speech-to-speech-translation']
|
['natural-language-processing', 'speech']
|
[ 3.37709427e-01 3.98096412e-01 1.05555296e-01 -3.82392555e-01
-1.21900582e+00 -4.81074005e-01 8.36255014e-01 1.28534392e-01
-3.68509024e-01 7.36600697e-01 5.10935426e-01 -4.66204077e-01
3.28312576e-01 -4.81191337e-01 -6.81063771e-01 -4.54749852e-01
3.68519992e-01 5.40673077e-01 2.30814472e-01 -3.74399453e-01
-5.01170009e-02 2.48220921e-01 -1.55716622e+00 4.30621207e-01
7.79366612e-01 4.90333885e-01 5.36806285e-01 8.60847712e-01
-2.53670067e-01 5.79496980e-01 -8.22380304e-01 -3.71998340e-01
9.68336537e-02 -9.78900194e-01 -7.58497059e-01 -8.96175802e-02
4.98685576e-02 -8.37878883e-02 -1.05658628e-01 7.84324288e-01
6.92120850e-01 -3.84737961e-02 4.65244681e-01 -7.74397910e-01
-3.23861867e-01 8.16485167e-01 2.50724226e-01 9.45988372e-02
5.48841894e-01 -1.51697010e-01 8.60998631e-01 -9.48395967e-01
7.01662421e-01 1.23506188e+00 2.01360255e-01 4.61767793e-01
-1.09129131e+00 -2.87644625e-01 -2.10572839e-01 2.96142474e-02
-1.37293673e+00 -1.01411140e+00 6.07365906e-01 -1.90121531e-01
1.16801655e+00 2.42452681e-01 6.74233615e-01 1.42237580e+00
2.96182446e-02 7.36116052e-01 1.12839794e+00 -1.19240940e+00
2.01219037e-01 4.94546831e-01 -6.08900249e-01 1.23587035e-01
-4.30186301e-01 3.05696219e-01 -7.05947876e-01 1.83507189e-01
4.27015185e-01 -7.19666839e-01 -2.14969009e-01 1.67969927e-01
-1.37265420e+00 6.32959783e-01 -1.56529516e-01 7.43714213e-01
-3.27985018e-01 -7.80478725e-03 4.75694597e-01 5.88538706e-01
6.29131854e-01 2.15860635e-01 -3.76271695e-01 -6.16361022e-01
-1.21874273e+00 1.75262943e-01 1.11375856e+00 8.39250207e-01
2.73213565e-01 3.66768777e-01 3.95780131e-02 9.63585138e-01
4.21815515e-01 8.79190087e-01 5.82145274e-01 -6.63993239e-01
7.20279157e-01 -4.86725532e-02 6.73672408e-02 -5.47375798e-01
-4.89499718e-02 -3.93411934e-01 -3.99834812e-01 6.91391528e-02
4.36617255e-01 -3.07321221e-01 -7.51837909e-01 1.71135092e+00
2.23760039e-01 -2.06565097e-01 4.56069827e-01 7.47327209e-01
5.81808627e-01 1.05138218e+00 -2.62619704e-01 -5.01084805e-01
1.19442320e+00 -8.28240097e-01 -1.02894711e+00 -1.68145180e-01
4.58547831e-01 -1.34516931e+00 1.01252091e+00 3.99247468e-01
-1.35199583e+00 -6.51659429e-01 -8.97730231e-01 4.37657870e-02
-3.69202644e-01 2.69769281e-01 1.76582802e-02 7.84408033e-01
-1.16046524e+00 4.70488489e-01 -8.10124815e-01 -6.67631686e-01
-5.21992624e-01 1.75733268e-01 -3.25183898e-01 3.10643613e-01
-1.55534256e+00 1.12533820e+00 3.26675147e-01 -2.05296934e-01
-5.50043166e-01 -2.25376785e-01 -7.56463289e-01 -3.78781892e-02
2.69048005e-01 -3.86336893e-01 1.64619827e+00 -1.24032831e+00
-2.16336966e+00 6.15020037e-01 -2.76351184e-01 -4.46066141e-01
5.70556760e-01 5.55453561e-02 -6.89103186e-01 2.19166815e-01
-2.76436098e-02 5.88483632e-01 8.05222631e-01 -9.22972918e-01
-5.77106893e-01 -4.05292511e-02 -4.50891882e-01 2.84903079e-01
-1.23580649e-01 4.50225711e-01 -3.94765019e-01 -8.65698099e-01
-8.26864094e-02 -1.01048160e+00 1.88673660e-01 -3.68512243e-01
-2.60210097e-01 3.69015262e-02 5.53737760e-01 -9.88208950e-01
1.20311809e+00 -2.16521025e+00 2.12509543e-01 -4.62295339e-02
-6.17373228e-01 3.59148026e-01 -1.43087327e-01 1.06386244e+00
-5.13918735e-02 -2.30997689e-02 -1.62008613e-01 -6.95921957e-01
-5.37520684e-02 3.08165520e-01 -4.84522849e-01 1.54287934e-01
3.57712507e-01 7.22707510e-01 -9.02793765e-01 -4.23806608e-01
2.78532624e-01 7.51056314e-01 -3.35013092e-01 4.13615465e-01
-8.04373249e-02 6.98848069e-01 -1.59110010e-01 1.71626776e-01
1.20367393e-01 5.23341060e-01 3.47194523e-02 2.30840251e-01
-4.64977801e-01 1.10702360e+00 -1.06087208e+00 1.52519333e+00
-7.46190786e-01 6.90605879e-01 2.15333968e-01 -8.23676467e-01
9.95499313e-01 1.06250453e+00 2.38822550e-01 -7.88638234e-01
2.06218034e-01 8.45954776e-01 3.29837680e-01 -5.52392840e-01
4.85485613e-01 -4.45984095e-01 -3.34146060e-03 2.88810015e-01
3.20759177e-01 -7.92666018e-01 2.09651113e-01 -8.13720003e-02
5.81624389e-01 2.26992175e-01 3.10851455e-01 -1.16143875e-01
4.73405719e-01 -4.68658358e-02 8.66124686e-03 2.33667985e-01
1.21160693e-01 9.31630015e-01 1.51500836e-01 1.81009918e-01
-1.16800129e+00 -9.31441188e-01 7.49373958e-02 8.08570921e-01
-5.40307701e-01 -5.68343580e-01 -1.12014973e+00 -2.65222222e-01
-5.67577422e-01 1.08354080e+00 -1.04125485e-01 8.68530348e-02
-6.56085610e-01 -2.86099344e-01 7.73066163e-01 1.19410433e-01
8.40752721e-02 -1.19361675e+00 -1.78774819e-01 4.94831294e-01
-4.64645028e-01 -1.31765163e+00 -6.27045155e-01 1.82203740e-01
-7.99915671e-01 -4.77122396e-01 -9.50073779e-01 -7.40809619e-01
2.28323504e-01 -9.23726335e-02 8.36189032e-01 -3.85192782e-01
1.89383671e-01 2.42327422e-01 -5.58475494e-01 -6.54042602e-01
-1.47933364e+00 2.91724861e-01 1.45900711e-01 6.36257827e-02
4.67772856e-02 -4.67267662e-01 -5.91087751e-02 3.14017355e-01
-9.40613031e-01 1.84036374e-01 6.00767732e-01 4.19618011e-01
3.30388248e-01 -3.07855904e-01 8.42338681e-01 -5.10720611e-01
8.07669401e-01 -2.60498703e-01 -4.56222534e-01 -8.11851472e-02
-3.45218003e-01 9.61928293e-02 9.65965688e-01 -4.63812321e-01
-9.32966292e-01 9.92298052e-02 -8.27778935e-01 -1.86315760e-01
-3.29632133e-01 7.06708848e-01 -2.75194496e-01 2.82064199e-01
6.46954477e-01 4.26727563e-01 4.34330404e-02 -4.64207292e-01
3.01908612e-01 1.32313395e+00 6.07037127e-01 -3.06810915e-01
6.91914856e-01 -1.00267217e-01 -1.92393035e-01 -1.29798996e+00
-2.82448471e-01 -3.84187162e-01 -5.34060299e-01 -7.93180019e-02
6.66798651e-01 -8.32641006e-01 -1.85891151e-01 4.71568376e-01
-1.36612821e+00 -2.65435129e-01 -5.01690805e-01 8.45570445e-01
-9.01258290e-01 3.33986074e-01 -5.06852210e-01 -1.04731607e+00
-2.56364703e-01 -1.26600742e+00 1.10087311e+00 -2.28385702e-01
-4.49588239e-01 -9.04387772e-01 1.66751862e-01 2.41600677e-01
6.64236605e-01 -1.80160984e-01 6.66904449e-01 -8.20158899e-01
-3.55896652e-01 -1.38776556e-01 2.73475766e-01 7.15473831e-01
3.46368462e-01 1.09389603e-01 -1.02551639e+00 -1.07641846e-01
2.82387078e-01 -1.33089051e-01 3.91488165e-01 2.14721352e-01
1.59248516e-01 -3.10953349e-01 5.64438216e-02 9.55982879e-02
8.90294671e-01 3.46839577e-01 6.32622302e-01 -4.88984473e-02
2.43666321e-01 8.31298709e-01 4.50880706e-01 1.56096563e-01
3.33195031e-01 1.16159260e+00 -7.36157075e-02 -1.96508706e-01
-5.84801912e-01 -5.47796726e-01 8.33512664e-01 1.56860840e+00
1.12205550e-01 -4.31905866e-01 -7.38410711e-01 7.81506658e-01
-1.43162227e+00 -8.75128627e-01 -1.53267175e-01 2.45476866e+00
1.01138520e+00 1.80325300e-01 2.12017521e-01 3.38682830e-01
5.32881796e-01 5.99589571e-02 1.14576660e-01 -7.92661011e-01
6.20466359e-02 2.95419693e-01 1.67278886e-01 1.00090373e+00
-6.93089306e-01 1.06139600e+00 6.40511227e+00 8.98837924e-01
-1.48649943e+00 1.17882393e-01 2.69076556e-01 1.07579038e-01
-3.13075751e-01 1.71965174e-02 -6.90048456e-01 4.68154728e-01
1.77993464e+00 -2.76289046e-01 6.59339011e-01 4.61010247e-01
8.53562713e-01 -2.94884872e-02 -1.07138550e+00 7.35009491e-01
2.26338804e-01 -8.98249388e-01 6.48406614e-03 -1.06946848e-01
3.94562662e-01 1.77652299e-01 -2.51044154e-01 1.96433991e-01
-2.59613216e-01 -7.83475816e-01 1.17903471e+00 1.59144774e-01
9.86806452e-01 -5.64130962e-01 5.72146177e-01 6.21117175e-01
-1.04709637e+00 3.75963449e-01 2.85161119e-02 -6.41818568e-02
4.76852179e-01 5.14959872e-01 -1.30890167e+00 6.52291596e-01
8.33042935e-02 3.35471123e-01 -1.23935238e-01 7.96591699e-01
-4.62494761e-01 1.07993960e+00 -4.28589880e-01 -1.22457527e-01
1.15951352e-01 -1.12887383e-01 6.97901785e-01 1.33761311e+00
8.26337457e-01 -2.88855642e-01 1.25468582e-01 6.21590376e-01
4.47706617e-02 6.40481830e-01 -7.11260498e-01 -4.95919555e-01
3.48583937e-01 8.38927805e-01 -6.07090771e-01 -3.50335568e-01
-5.10501921e-01 1.13026536e+00 -1.90010235e-01 3.31600457e-01
-4.12443697e-01 -4.65384722e-01 3.12439948e-01 3.86017174e-01
1.93410411e-01 -4.50977683e-01 8.68314728e-02 -1.00049543e+00
1.62156746e-02 -1.13040280e+00 -2.19863325e-01 -8.13386083e-01
-8.77205431e-01 9.80399847e-01 2.00663239e-01 -1.11360300e+00
-1.00079036e+00 -3.76926929e-01 -3.42006892e-01 1.39686656e+00
-1.54339445e+00 -1.05462158e+00 3.82832438e-01 2.13095695e-01
8.67091894e-01 -2.27713794e-01 1.07795608e+00 3.75307351e-01
5.97400516e-02 3.22644949e-01 4.93209660e-02 -3.32518891e-02
9.72653210e-01 -1.00215888e+00 9.41114366e-01 8.32156599e-01
4.60286885e-01 4.27848876e-01 1.10704422e+00 -4.94092822e-01
-1.17020428e+00 -8.27789545e-01 1.71792841e+00 -4.29166436e-01
6.85994744e-01 -6.35734975e-01 -8.07984293e-01 4.50316310e-01
5.14804244e-01 -4.58944708e-01 5.59160590e-01 -3.06165695e-01
1.10889740e-01 1.12333335e-02 -7.91609824e-01 6.39854848e-01
6.97387576e-01 -7.83101201e-01 -7.29336500e-01 2.70249128e-01
8.36754620e-01 -5.33054113e-01 -6.21705234e-01 -9.49811935e-03
2.42316604e-01 -7.93069243e-01 5.82625031e-01 -1.35246783e-01
2.47980297e-01 -3.51866692e-01 -3.73522878e-01 -1.73516858e+00
2.76569635e-01 -1.03775084e+00 2.08817348e-01 1.35428202e+00
6.58123195e-01 -7.01195419e-01 3.17698389e-01 1.38813615e-01
-4.64724839e-01 -3.16179454e-01 -1.22452843e+00 -8.68600190e-01
2.32224882e-01 -6.46480381e-01 5.25989413e-01 5.02270937e-01
2.36227825e-01 6.42571867e-01 -5.00806868e-01 5.12443259e-02
-2.74187326e-02 -1.79051533e-01 7.55530119e-01 -1.09472752e+00
-5.82747817e-01 -1.99963674e-01 -5.14354557e-02 -1.06817210e+00
1.09366467e-02 -8.94567132e-01 3.14431101e-01 -1.57507145e+00
-2.95095772e-01 -8.44605416e-02 1.78137928e-01 3.10885787e-01
1.13913333e-02 1.12156883e-01 4.52948809e-01 4.83282916e-02
1.40429541e-01 5.94919920e-01 1.20161712e+00 1.75625935e-01
-3.23951840e-01 2.71467566e-01 -4.64656174e-01 2.67970622e-01
7.58673966e-01 -6.26093626e-01 -4.30051208e-01 -2.46137336e-01
1.05778165e-01 3.48066330e-01 -4.34817448e-02 -9.66464758e-01
-3.95242088e-02 1.41666699e-02 -2.07739860e-01 -2.83317953e-01
5.69529116e-01 -7.88279474e-01 2.38980561e-01 3.14476192e-01
-3.98701459e-01 6.62838072e-02 3.60953867e-01 4.44768704e-02
-6.25729620e-01 -4.38434064e-01 6.26286626e-01 2.53464840e-02
-1.13717332e-01 -3.37371558e-01 -8.73465836e-01 -5.14146574e-02
6.44417703e-01 -1.46975517e-01 8.67826119e-02 -9.01289344e-01
-6.48596287e-01 -3.79172534e-01 4.13339376e-01 6.56737089e-01
3.24666321e-01 -1.06365359e+00 -1.07443988e+00 5.40525436e-01
9.10034701e-02 -4.11913842e-01 -1.81069314e-01 7.70764410e-01
-2.15533584e-01 8.64880025e-01 9.86155942e-02 -4.34997946e-01
-1.34797192e+00 3.56525809e-01 3.32569569e-01 3.15368697e-02
-2.89403528e-01 2.74463654e-01 -2.36338049e-01 -5.72246969e-01
2.13472694e-02 -5.24139166e-01 6.90306723e-02 -4.11379002e-02
4.42159653e-01 1.17415674e-01 4.39164251e-01 -9.60698485e-01
-1.83927119e-01 2.68616170e-01 2.05335185e-01 -9.07980561e-01
1.04244781e+00 -2.28808045e-01 1.57603040e-01 7.90587723e-01
1.01706159e+00 5.90379000e-01 -8.29541504e-01 8.65743868e-03
-4.19587567e-02 -2.89054155e-01 8.24539810e-02 -8.48170042e-01
-4.95346993e-01 1.17478812e+00 3.40526044e-01 3.43682379e-01
1.00042772e+00 -1.29198015e-01 8.75398219e-01 1.78007782e-01
4.73764747e-01 -1.33274817e+00 -1.65350318e-01 9.01098013e-01
9.27128494e-01 -9.54029083e-01 -5.06626904e-01 -4.25325572e-01
-6.40552640e-01 1.17857873e+00 -1.37245610e-01 3.76135528e-01
3.69242847e-01 4.82525587e-01 5.56287944e-01 3.93893331e-01
-7.23557591e-01 -2.75161088e-01 2.32597619e-01 6.63031340e-01
9.61944163e-01 -2.70005018e-02 -5.43798566e-01 3.24889004e-01
-4.83137250e-01 2.32432142e-01 6.23172641e-01 7.10747302e-01
-4.05187875e-01 -1.80928326e+00 -5.89761496e-01 -1.72505096e-01
-6.76839173e-01 -3.56138349e-01 -6.21826828e-01 5.96005917e-01
-2.35068619e-01 1.32520163e+00 -2.03786120e-01 -1.02492884e-01
4.95630175e-01 4.95782733e-01 4.33897972e-01 -9.05679882e-01
-6.46231830e-01 5.35311520e-01 4.24715370e-01 -1.86271176e-01
-2.99189270e-01 -7.52880991e-01 -1.24061632e+00 -3.99186611e-02
-2.94304878e-01 4.50354815e-01 1.27334237e+00 1.04412115e+00
3.03018034e-01 6.72731042e-01 6.06178403e-01 -1.03914917e+00
-5.65940320e-01 -1.21783996e+00 -3.93110782e-01 1.64536968e-01
3.28916669e-01 -1.35792300e-01 -3.23488444e-01 2.22918212e-01]
|
[14.520392417907715, 6.984955787658691]
|
4d5da945-eb18-4c90-adc8-f3290b266827
|
extended-u-net-for-speaker-verification-in
|
2206.13044
| null |
https://arxiv.org/abs/2206.13044v1
|
https://arxiv.org/pdf/2206.13044v1.pdf
|
Extended U-Net for Speaker Verification in Noisy Environments
|
Background noise is a well-known factor that deteriorates the accuracy and reliability of speaker verification (SV) systems by blurring speech intelligibility. Various studies have used separate pretrained enhancement models as the front-end module of the SV system in noisy environments, and these methods effectively remove noises. However, the denoising process of independent enhancement models not tailored to the SV task can also distort the speaker information included in utterances. We argue that the enhancement network and speaker embedding extractor should be fully jointly trained for SV tasks under noisy conditions to alleviate this issue. Therefore, we proposed a U-Net-based integrated framework that simultaneously optimizes speaker identification and feature enhancement losses. Moreover, we analyzed the structural limitations of using U-Net directly for noise SV tasks and further proposed Extended U-Net to reduce these drawbacks. We evaluated the models on the noise-synthesized VoxCeleb1 test set and VOiCES development set recorded in various noisy scenarios. The experimental results demonstrate that the U-Net-based fully joint training framework is more effective than the baseline, and the extended U-Net exhibited state-of-the-art performance versus the recently proposed compensation systems.
|
['Ha-Jin Yu', 'Hye-jin Shim', 'Jungwoo Heo', 'Ju-ho Kim']
|
2022-06-27
| null | null | null | null |
['speaker-identification']
|
['speech']
|
[ 1.06800549e-01 3.42652127e-02 3.93052965e-01 -5.88362932e-01
-9.99565065e-01 -2.27591693e-01 2.10938588e-01 -6.39037132e-01
-4.54965025e-01 4.74077195e-01 5.52316606e-01 -2.89511621e-01
1.60518289e-01 -1.37133300e-01 -4.15761083e-01 -1.00152743e+00
3.19428325e-01 -3.75206470e-01 -1.35492399e-01 -2.85205543e-01
-2.17078686e-01 3.31998795e-01 -1.66124082e+00 1.83125660e-01
1.15096796e+00 1.02178168e+00 4.53057587e-01 6.99630201e-01
1.32504433e-01 4.69027847e-01 -1.00935280e+00 -4.62830365e-01
2.14788973e-01 -3.19624096e-01 -2.92256996e-02 1.15712315e-01
4.78738815e-01 -4.72541064e-01 -6.75822556e-01 1.38008058e+00
1.23559320e+00 2.43727013e-01 3.74564081e-01 -9.63536620e-01
-8.51196766e-01 8.08882892e-01 -2.76637614e-01 3.25023562e-01
1.22335255e-02 2.57811725e-01 6.49242043e-01 -1.11280406e+00
1.61102876e-01 1.36165380e+00 8.58408809e-01 9.74690199e-01
-1.02479458e+00 -9.38917816e-01 1.17744707e-01 3.29535067e-01
-1.33804858e+00 -1.35660017e+00 1.03577828e+00 -3.78331691e-02
9.34760451e-01 5.47784686e-01 1.41157568e-01 1.37992692e+00
-2.33205393e-01 7.89856195e-01 1.02017915e+00 -3.67852092e-01
-9.99384448e-02 5.35459459e-01 3.71799082e-01 1.93324283e-01
-4.20892499e-02 5.25795341e-01 -6.04532063e-01 -5.59148416e-02
4.05484319e-01 -5.48448801e-01 -7.70162523e-01 3.07809889e-01
-7.91667163e-01 4.37668413e-01 1.34869799e-01 4.43860739e-01
-3.40169251e-01 -1.87927276e-01 5.16181052e-01 4.13038284e-01
7.23837256e-01 1.14916600e-01 -4.74718243e-01 1.11325234e-02
-1.08071661e+00 -4.76753078e-02 4.96400416e-01 7.17748940e-01
1.09179899e-01 7.66895831e-01 -5.81816912e-01 1.32950532e+00
4.74974006e-01 5.51948190e-01 6.23793304e-01 -4.76168454e-01
6.01722181e-01 -4.81661409e-02 -8.16047862e-02 -6.29947484e-01
-8.37466568e-02 -1.08232152e+00 -9.35542583e-01 1.14993393e-01
6.96936101e-02 -4.23842400e-01 -9.66632903e-01 2.07558298e+00
3.68003041e-01 2.65126973e-01 4.64111120e-01 1.00778389e+00
1.27507663e+00 6.66782498e-01 2.65588369e-02 -4.60760087e-01
1.16595662e+00 -1.13836098e+00 -1.60192752e+00 -1.89033806e-01
3.51779275e-02 -1.09997499e+00 9.29806054e-01 3.68808955e-01
-1.13703871e+00 -1.07350004e+00 -1.28231096e+00 4.61910432e-03
-6.98227286e-02 4.44122970e-01 -3.65251079e-02 1.37202322e+00
-1.23106289e+00 4.72669065e-01 -6.09778881e-01 8.59625116e-02
2.28193209e-01 3.54490787e-01 -2.93751240e-01 1.60987273e-01
-1.45755064e+00 9.88041699e-01 -7.68134072e-02 7.63754964e-01
-1.01316917e+00 -4.06421870e-01 -9.96621132e-01 2.42082998e-01
1.10425010e-01 -3.67678702e-01 1.36525536e+00 -7.26132214e-01
-1.83995175e+00 3.53350908e-01 -4.99355763e-01 -3.70713711e-01
5.76031625e-01 -3.63301903e-01 -1.12514627e+00 -1.75669089e-01
-3.60991180e-01 1.67602345e-01 1.29110622e+00 -1.32234955e+00
-2.55190611e-01 -4.52248931e-01 -5.94717979e-01 3.19863290e-01
-6.80539727e-01 3.75440687e-01 -2.80527502e-01 -8.77835453e-01
1.87849075e-01 -3.66753042e-01 8.20595920e-02 -3.10961992e-01
-3.52712005e-01 -1.13172298e-02 1.28053653e+00 -1.51660109e+00
1.34436798e+00 -2.48918533e+00 -1.49117365e-01 -2.23319959e-02
-2.04167962e-02 8.76022935e-01 -3.11753929e-01 -2.07280591e-02
-1.62442401e-01 -2.19910350e-02 -1.94079518e-01 -9.25511122e-01
1.08419910e-01 -2.50028279e-02 -2.07766458e-01 5.22204757e-01
4.35497202e-02 4.66521561e-01 -4.38672692e-01 -2.90255427e-01
2.62422711e-01 1.11551523e+00 -3.32059503e-01 5.18898547e-01
3.99165034e-01 4.64282066e-01 1.13372236e-01 5.71116328e-01
1.23377633e+00 6.27232373e-01 -1.32670924e-01 -3.36997032e-01
2.81675514e-02 4.62737113e-01 -1.39282906e+00 1.37238920e+00
-3.69885325e-01 7.98869669e-01 9.02284265e-01 -5.69915533e-01
1.07380641e+00 9.35011208e-01 -1.18925020e-01 -4.02296990e-01
2.41606623e-01 3.07170391e-01 2.05474198e-01 -7.15208352e-01
4.85013455e-01 -3.85249466e-01 4.78054732e-01 -1.15329891e-01
2.65369058e-01 9.70546529e-03 -4.68247086e-01 -2.22917691e-01
7.15187967e-01 -2.69667178e-01 1.02955282e-01 -6.88775107e-02
9.47993100e-01 -9.68713522e-01 9.20474291e-01 6.54994249e-01
-7.16015279e-01 7.69981325e-01 -1.30738929e-01 3.62493455e-01
-6.87255800e-01 -9.92384732e-01 -3.36127251e-01 9.18259621e-01
-3.64375189e-02 -2.36827999e-01 -1.07351744e+00 -5.18427253e-01
-2.85966277e-01 9.34255838e-01 -2.46642455e-01 -2.21088529e-01
-4.75033492e-01 -6.96879864e-01 1.03533173e+00 4.49449182e-01
7.86620200e-01 -8.70230794e-01 2.66993761e-01 3.06364805e-01
-5.20801663e-01 -1.15589833e+00 -9.22776103e-01 1.38787553e-01
-4.64611024e-01 -5.31513870e-01 -8.95159900e-01 -8.94014299e-01
4.36088443e-01 3.30915451e-01 3.93046439e-01 -5.93289547e-02
2.76885778e-01 1.34980291e-01 -1.03999898e-01 -5.25552988e-01
-8.42979670e-01 -1.90668926e-01 4.59342211e-01 4.23020929e-01
3.22071016e-01 -4.82529789e-01 -5.22739232e-01 4.71233100e-01
-6.82401478e-01 -3.85450125e-01 4.07884419e-01 1.12727010e+00
5.64727932e-02 2.06328794e-01 1.03709972e+00 -3.25755268e-01
1.00679564e+00 6.14663064e-02 -4.87984389e-01 2.19679564e-01
-3.97811681e-01 -1.69567481e-01 3.90487343e-01 -5.71958244e-01
-1.85651374e+00 -1.57963887e-01 -8.94720972e-01 -4.35168952e-01
-2.89669693e-01 1.31305829e-01 -9.92635548e-01 -1.13109604e-01
5.22627473e-01 4.09928411e-01 -1.31747415e-02 -8.19723368e-01
1.08467884e-01 1.49966466e+00 6.58209205e-01 -1.25631273e-01
7.53815472e-01 -2.08431687e-02 -6.64246619e-01 -1.27224898e+00
-4.02063131e-01 -5.92484891e-01 -2.11501494e-01 -2.05502763e-01
5.70563436e-01 -1.19540727e+00 -5.32071114e-01 8.76925468e-01
-1.49790525e+00 4.24860194e-02 -3.82004492e-02 6.48956418e-01
1.42938113e-02 7.41891682e-01 -8.51669848e-01 -1.36061060e+00
-7.19402075e-01 -1.47438681e+00 7.53383577e-01 1.83414638e-01
8.83152559e-02 -6.53658032e-01 -3.14389020e-02 6.42308950e-01
9.36217129e-01 -4.98544514e-01 3.32893789e-01 -6.97804451e-01
-2.08011344e-02 -9.18816701e-02 -2.40416545e-02 1.27001929e+00
2.89815068e-01 -3.30618024e-01 -2.02590346e+00 -4.12851483e-01
8.07140231e-01 1.99806437e-01 9.66316998e-01 5.14125228e-01
9.61132050e-01 -4.98716474e-01 6.28262982e-02 7.33780742e-01
1.06200218e+00 3.11462909e-01 8.59407365e-01 -1.16935045e-01
4.64215994e-01 5.94784379e-01 2.02316508e-01 7.04659969e-02
-4.81019169e-02 7.14645088e-01 1.61608443e-01 -2.58810699e-01
-5.91729879e-01 -1.16392985e-01 7.65314758e-01 1.26144373e+00
3.26064117e-02 -3.21281850e-01 -3.41367662e-01 5.85909069e-01
-1.38098955e+00 -1.04590631e+00 -4.41834331e-02 2.12726355e+00
8.39159250e-01 -4.35188934e-02 -1.49827823e-01 4.37892228e-01
1.14962220e+00 2.48625994e-01 -4.11779076e-01 -2.76232660e-01
-5.13810217e-01 9.88064185e-02 1.50818557e-01 6.82563961e-01
-1.04178667e+00 6.46199822e-01 6.31492901e+00 1.00699925e+00
-1.15583444e+00 5.32620370e-01 4.78566796e-01 -1.00341596e-01
-4.51781340e-02 -6.06136382e-01 -9.18615341e-01 3.96797091e-01
1.09099698e+00 8.29778761e-02 2.85415381e-01 8.80923033e-01
5.38462341e-01 5.00031531e-01 -8.38448703e-01 1.13077700e+00
2.78550416e-01 -5.74489117e-01 -3.43549520e-01 2.26712320e-02
6.04662836e-01 -7.27704167e-02 3.80894035e-01 5.07107198e-01
-2.45915160e-01 -8.13827395e-01 8.45641851e-01 1.69694424e-01
7.21372604e-01 -5.65096080e-01 1.03553689e+00 3.00282180e-01
-1.07975197e+00 -7.61301890e-02 -4.08753872e-01 1.51614651e-01
3.97359520e-01 8.08613598e-01 -9.06623602e-01 5.56984603e-01
5.68752050e-01 1.73897997e-01 -3.60301197e-01 1.05008185e+00
-4.48841363e-01 8.51810694e-01 -1.30474880e-01 2.66877890e-01
-3.56933802e-01 5.15034683e-02 1.09255373e+00 1.43880212e+00
4.03498024e-01 -4.12564054e-02 -5.14965355e-01 8.21932673e-01
-1.76583931e-01 -4.45651710e-02 -3.41883093e-01 1.59904450e-01
3.75631332e-01 1.23095429e+00 2.83015389e-02 -3.38694781e-01
-2.52654105e-01 8.96900713e-01 -1.35338809e-02 7.94535816e-01
-1.00162625e+00 -5.39407074e-01 1.04147160e+00 -1.65123239e-01
5.02599299e-01 8.06415305e-02 -3.96240562e-01 -1.31718826e+00
2.56515354e-01 -1.13131988e+00 -7.13688359e-02 -5.97427189e-01
-1.30170310e+00 1.14847469e+00 -5.71376443e-01 -1.12454081e+00
1.06036946e-01 -4.28542644e-01 -6.74880862e-01 1.21048808e+00
-1.77121639e+00 -9.77469444e-01 -1.46776050e-01 5.06830037e-01
7.30531394e-01 -2.98140675e-01 7.11700857e-01 6.70912206e-01
-9.14398611e-01 1.17127383e+00 2.53235817e-01 1.69292074e-02
9.09731209e-01 -9.74173963e-01 4.01296347e-01 1.37002575e+00
-7.30367005e-02 6.90312922e-01 9.36144054e-01 -5.09082198e-01
-9.29632187e-01 -1.02147222e+00 8.85724008e-01 4.13776971e-02
6.14501238e-02 -6.04442000e-01 -1.27251720e+00 4.74647045e-01
5.06787181e-01 -1.50474653e-01 6.34084761e-01 -7.11969659e-02
-1.95456773e-01 -3.80344987e-01 -1.41344392e+00 4.35566604e-01
9.81879354e-01 -7.56781042e-01 -9.17661488e-01 -1.29055142e-01
9.62974250e-01 -3.96194130e-01 -6.12676084e-01 4.37833846e-01
3.99711698e-01 -9.29173529e-01 1.00279713e+00 -1.92161158e-01
-9.85023305e-02 -3.37672323e-01 -3.15628767e-01 -1.61892951e+00
-1.58177376e-01 -8.07713866e-01 -1.95026964e-01 1.97625935e+00
6.09072804e-01 -8.88324738e-01 2.83665538e-01 6.13271534e-01
-5.16555905e-01 -1.56854153e-01 -1.09815216e+00 -8.70958447e-01
-3.62913877e-01 -6.40163839e-01 7.84664869e-01 6.65020585e-01
-1.35717914e-01 3.21141243e-01 -7.23694623e-01 5.72397351e-01
7.87093818e-01 -7.25844085e-01 6.32668197e-01 -6.92009211e-01
-4.13477093e-01 -2.50862479e-01 -1.34719282e-01 -1.02546906e+00
1.91305414e-01 -4.45305675e-01 3.99376214e-01 -1.21454191e+00
-1.57548979e-01 8.92144144e-02 -4.56692725e-01 -7.70427100e-03
-6.59988344e-01 -7.48576373e-02 7.95725808e-02 -1.69246435e-01
4.98953611e-02 1.10420859e+00 1.16350448e+00 -2.63756901e-01
-3.18338662e-01 2.79609710e-01 -7.09968209e-01 5.74857771e-01
6.39117122e-01 -1.89360052e-01 -3.45706880e-01 -4.00324792e-01
-6.76290870e-01 1.29754767e-01 1.43828139e-01 -1.09091020e+00
4.13163692e-01 4.23451841e-01 3.84042382e-01 -6.48182690e-01
7.79250324e-01 -8.85007679e-01 -3.06059606e-02 1.96575329e-01
-4.09936845e-01 -5.20577669e-01 4.39490974e-01 4.04405773e-01
-6.32026076e-01 -1.98295102e-01 8.80860686e-01 2.43472382e-01
-2.65299797e-01 -3.13803703e-02 -3.57359141e-01 -4.55714524e-01
4.68878955e-01 -1.22953884e-01 -2.75984734e-01 -6.21868789e-01
-7.91093230e-01 -1.44808218e-01 -8.53045359e-02 3.80496413e-01
8.40920985e-01 -1.17305171e+00 -8.89742672e-01 5.44133365e-01
-2.79701322e-01 -3.50281119e-01 8.01580787e-01 7.36867726e-01
1.44803241e-01 2.87003487e-01 1.28432542e-01 -4.03566837e-01
-1.74805963e+00 5.17401397e-01 5.76428115e-01 -3.28097157e-02
-3.70142579e-01 1.10350883e+00 3.96875352e-01 -6.82661057e-01
7.99418807e-01 -3.30962867e-01 -3.36369365e-01 -2.58669049e-01
8.56547117e-01 6.22939765e-01 4.85099047e-01 -9.01006758e-01
-3.48924935e-01 2.06643611e-01 -1.49248406e-01 -4.56488580e-01
1.31778204e+00 -5.42040527e-01 1.91158786e-01 1.75991114e-02
1.13214505e+00 2.90911853e-01 -1.13295913e+00 -4.34520394e-01
-4.43684071e-01 -3.80598873e-01 5.27230799e-01 -8.64787757e-01
-1.23666775e+00 1.05345178e+00 9.07663107e-01 -5.21654338e-02
1.39725852e+00 -5.55930257e-01 9.11065519e-01 -4.76611704e-02
-1.58495575e-01 -1.15233076e+00 -1.37458488e-01 1.93472877e-01
1.18978155e+00 -1.25977290e+00 -4.07771319e-01 -7.00957239e-01
-6.67377412e-01 8.69590104e-01 6.71777964e-01 4.81341094e-01
5.95261872e-01 3.50501895e-01 4.20229256e-01 1.88675344e-01
-5.08589327e-01 -3.40789929e-02 5.53190172e-01 7.81292379e-01
4.10614461e-01 1.40569089e-02 -8.69837925e-02 1.25744867e+00
-2.98605680e-01 -3.18842143e-01 2.06643358e-01 6.03246033e-01
-3.02589178e-01 -9.52809095e-01 -9.93368983e-01 5.35548069e-02
-6.38295531e-01 -2.47251377e-01 -2.17440575e-01 2.98984796e-01
1.15034536e-01 1.60440743e+00 -5.13191402e-01 -6.09618366e-01
5.93161225e-01 3.93278599e-01 2.35499486e-01 -1.56557828e-01
-9.71093535e-01 5.29018342e-01 2.04436883e-01 -9.70696956e-02
-2.10397124e-01 -6.04654253e-01 -5.69723070e-01 -6.98283464e-02
-9.72948074e-01 1.19386189e-01 9.69548225e-01 8.60258043e-01
2.80297786e-01 9.14015293e-01 7.49735534e-01 -6.44462287e-01
-1.01607168e+00 -1.71000338e+00 -7.32116163e-01 2.94156134e-01
8.53377342e-01 -4.50984269e-01 -9.42691743e-01 5.82901314e-02]
|
[14.794355392456055, 5.976571083068848]
|
cfdbe0e0-4ed1-4e4b-9a51-6543864e3ed4
|
2020-cataracts-semantic-segmentation
|
2110.10965
| null |
https://arxiv.org/abs/2110.10965v2
|
https://arxiv.org/pdf/2110.10965v2.pdf
|
2020 CATARACTS Semantic Segmentation Challenge
|
Surgical scene segmentation is essential for anatomy and instrument localization which can be further used to assess tissue-instrument interactions during a surgical procedure. In 2017, the Challenge on Automatic Tool Annotation for cataRACT Surgery (CATARACTS) released 50 cataract surgery videos accompanied by instrument usage annotations. These annotations included frame-level instrument presence information. In 2020, we released pixel-wise semantic annotations for anatomy and instruments for 4670 images sampled from 25 videos of the CATARACTS training set. The 2020 CATARACTS Semantic Segmentation Challenge, which was a sub-challenge of the 2020 MICCAI Endoscopic Vision (EndoVis) Challenge, presented three sub-tasks to assess participating solutions on anatomical structure and instrument segmentation. Their performance was assessed on a hidden test set of 531 images from 10 videos of the CATARACTS test set.
|
['Da-Han Wang', 'Feihong Huang', 'Joan M. Nunez Do Rio', 'Jeremy Birch', 'Martin Huber', 'Claudio Ravasio', 'Danail Stoyanov', 'Haili Ye', 'Jianyuan Hong', 'Uddhav Vaghela', 'Sophia Bano', 'Pal Halvorsen', 'Michael A. Riegler', 'Debesh Jha', 'Nikhil KumarTomar', 'Fucang Jia', 'Hongyu Chen', 'Christos Bergeles', 'Lyndon Da Cruz', 'Theodoros Pissas', 'Ren Hongliang', 'Bharat Giddwani', 'Mobarakol Islam', 'Guotai Wang', 'Lu Wang', 'Dong Guo', 'Haojie Wang', 'Jiacheng Wang', 'Heonjin Ha', 'Zeng-Guang Hou', 'Gui-Bin Bian', 'Chen-Chen Fan', 'Zhen-Liang Ni', 'Nicolas Padoy', 'Deepak Alapatt', 'Chinedu Innocent Nwoye', 'Chris Walsh', 'Rahim Mohammadi', 'Maria Grammatikopoulou', 'Imanol Luengo']
|
2021-10-21
| null | null | null | null |
['scene-segmentation']
|
['computer-vision']
|
[ 1.53028744e-03 3.02981377e-01 -6.58231080e-02 3.58868651e-02
-1.03992951e+00 -1.12254202e+00 1.08472966e-01 -1.59924671e-01
-5.33526897e-01 3.70600760e-01 6.45813942e-01 -5.05162597e-01
1.57292292e-01 1.23277649e-01 -6.76929832e-01 -3.27858299e-01
-3.37092608e-01 2.85443515e-01 1.08820565e-01 2.41321102e-01
1.21565543e-01 2.63334185e-01 -1.42420268e+00 3.33266288e-01
8.18257868e-01 1.04578757e+00 4.48776990e-01 1.08198357e+00
-3.64217092e-03 4.00487661e-01 -4.42011297e-01 -1.56800762e-01
5.65513432e-01 -3.37989211e-01 -1.21431172e+00 2.12555632e-01
1.16731167e+00 -1.25105813e-01 -1.02808185e-01 1.30584478e+00
6.80357397e-01 -5.93720078e-01 4.61382896e-01 -5.26944101e-01
1.04726560e-01 3.24098855e-01 -1.49905786e-01 5.33113182e-01
5.21576464e-01 3.17348093e-01 5.13950169e-01 -9.16144609e-01
1.30358088e+00 6.33142233e-01 8.97333264e-01 5.87132990e-01
-7.79364824e-01 -5.61845541e-01 -3.32204968e-01 6.67363778e-02
-8.75682533e-01 -2.10846275e-01 1.29432559e-01 -9.89113390e-01
6.75866127e-01 4.61934686e-01 1.68116796e+00 8.99378717e-01
3.44117850e-01 1.16608274e+00 9.91755784e-01 -3.06610823e-01
1.41304731e-01 -1.47021502e-01 4.35149074e-02 1.44746888e+00
5.97944140e-01 3.76060218e-01 -3.46115947e-01 1.55198388e-02
1.03496313e+00 -4.50046360e-01 -9.48922336e-01 -6.34736896e-01
-1.72955656e+00 5.48506439e-01 4.38115090e-01 1.43584594e-01
-1.23741828e-01 7.33092204e-02 8.34640205e-01 4.45658863e-02
1.52971037e-02 1.05868387e+00 -5.26148379e-01 -1.26486853e-01
-7.34071493e-01 -9.18514580e-02 9.58966911e-01 1.25521243e+00
-1.93395391e-02 -3.09938669e-01 -3.50622743e-01 5.01228929e-01
2.31463835e-01 2.03586984e-02 5.35143495e-01 -1.17724991e+00
1.13885231e-01 4.43676174e-01 2.87688613e-01 -5.38336160e-03
-8.32568645e-01 -5.09417772e-01 -5.06698668e-01 4.35393542e-01
6.21844530e-01 -3.02814335e-01 -1.67502201e+00 6.91115797e-01
6.22128844e-02 4.01149392e-01 -2.41203889e-01 1.10240519e+00
1.62389994e+00 -2.79260278e-01 1.16572626e-01 -2.43553579e-01
1.65899467e+00 -1.36273170e+00 -6.58216357e-01 -9.94210914e-02
9.99777734e-01 -1.30645823e+00 1.15401435e+00 6.27196193e-01
-1.31130981e+00 -4.07941222e-01 -7.65424609e-01 -1.49750531e-01
-1.87872887e-01 6.79242790e-01 9.92114067e-01 5.81341445e-01
-1.17118037e+00 -1.27296418e-01 -8.29012811e-01 -4.34496343e-01
9.00010526e-01 6.48619771e-01 -3.35026473e-01 1.22353453e-02
-4.73952770e-01 8.64203572e-01 2.26696879e-01 1.47689581e-01
-1.15530157e+00 -1.26322699e+00 -1.00803173e+00 -5.20023525e-01
3.97009552e-01 -1.02014303e+00 1.50944090e+00 -8.32398951e-01
-1.35347223e+00 1.82104158e+00 -4.87288944e-02 -5.66628098e-01
8.39734435e-01 -2.90663511e-01 -1.44289210e-01 4.12045926e-01
3.79451141e-02 8.16866815e-01 6.39800251e-01 -1.21194530e+00
-9.75286901e-01 -1.20781735e-01 1.44419089e-01 1.69596121e-01
3.72354358e-01 2.30526894e-01 -9.99287784e-01 -7.56133497e-01
-2.50738680e-01 -1.36295342e+00 -4.20854151e-01 5.21849394e-01
-7.87075520e-01 9.61217880e-02 4.35549021e-01 -1.13576138e+00
7.43276775e-01 -1.98443604e+00 3.96670967e-01 -5.75414523e-02
4.79528904e-01 2.76290089e-01 -3.44514847e-02 -6.79056168e-01
-2.77640373e-01 5.43590300e-02 1.00971371e-01 -5.89364707e-01
-7.68613994e-01 2.87501235e-02 2.34889060e-01 6.24492824e-01
-5.83619773e-01 9.22964513e-01 -1.03884602e+00 -7.79337227e-01
7.52943635e-01 1.10063903e-01 -7.28773177e-01 2.54266500e-01
-7.05562010e-02 9.81016576e-01 2.23790929e-02 1.38232863e+00
3.49698275e-01 -7.58468807e-02 -3.18945087e-02 -8.66115868e-01
-8.13333392e-02 -3.50440234e-01 -6.21744096e-01 2.58897781e+00
-6.24327958e-01 8.09436619e-01 4.05977935e-01 -3.96199286e-01
-1.53252468e-01 6.94663167e-01 9.82143104e-01 -4.28840160e-01
3.81672055e-01 4.85514283e-01 5.24124056e-02 -8.15813780e-01
-5.88041395e-02 3.36688101e-01 -9.12313312e-02 -7.92212486e-01
4.37692136e-01 -6.56150103e-01 2.28629559e-01 -1.29111975e-01
9.97454941e-01 4.81853485e-02 3.15553367e-01 -5.66542089e-01
3.75858575e-01 5.97462296e-01 1.65992498e-01 7.36985743e-01
-6.75660372e-01 8.83459210e-01 3.93958062e-01 -7.91886687e-01
-6.99935555e-01 -1.32330847e+00 -2.32716799e-01 1.24648802e-01
4.94911522e-01 -6.53916359e-01 -9.91803885e-01 -1.20122790e+00
-2.92523831e-01 2.94353459e-02 -9.40962732e-01 1.93074018e-01
-4.13514078e-01 -2.90866286e-01 8.13082382e-02 5.25530279e-01
1.46610200e-01 -8.09456825e-01 -8.71395111e-01 -1.53085887e-01
-3.97123694e-01 -1.32276320e+00 -7.49849141e-01 7.51499161e-02
-6.02698326e-01 -2.05578327e+00 -1.06472826e+00 -1.44343793e+00
8.51923048e-01 6.17235005e-02 1.22972846e+00 -2.22033635e-01
-1.34804940e+00 8.39080513e-01 -5.04982136e-02 -9.72571611e-01
-2.09010050e-01 -7.56025240e-02 -2.57418841e-01 -4.33852673e-01
-1.91926882e-01 2.64659882e-01 -1.09076548e+00 2.09423810e-01
-2.89036065e-01 3.27599674e-01 6.46021068e-01 8.59967947e-01
8.05682659e-01 -5.52430391e-01 -5.78924894e-01 -6.80268288e-01
3.87187153e-02 4.61850278e-02 -1.05306506e+00 2.93111086e-01
-1.13748536e-01 -5.14732361e-01 5.76564223e-02 -8.17298219e-02
-8.58958483e-01 5.35853624e-01 2.24273860e-01 -6.55408680e-01
-4.69110347e-02 2.23137230e-01 3.75092953e-01 -7.20711768e-01
5.34239709e-01 -2.25389764e-01 3.25214654e-01 -2.20202416e-01
2.05754250e-01 3.70489269e-01 1.13982403e+00 -2.15532631e-01
3.30680102e-01 4.89784747e-01 -3.84446699e-03 -6.46967471e-01
-1.17381346e+00 -1.11001563e+00 -5.46973646e-01 -5.48548698e-01
1.46681809e+00 -1.03918898e+00 -6.86880469e-01 4.13396895e-01
-1.18370879e+00 -5.15294075e-01 -4.72629637e-01 7.18498111e-01
-6.00751340e-01 1.11713059e-01 -4.54277694e-01 -5.44641875e-02
-4.00219530e-01 -1.88685083e+00 1.53165770e+00 2.44040698e-01
-2.60697722e-01 -1.14080787e+00 -1.85931355e-01 5.84742546e-01
1.56303957e-01 5.00516117e-01 6.15394771e-01 -7.81139284e-02
-6.32019639e-01 -2.55659908e-01 -2.02860117e-01 5.37377238e-01
2.26883233e-01 -9.13134292e-02 -5.36103249e-01 -4.10785913e-01
-4.35545325e-01 -8.87378231e-02 9.71162915e-01 1.17458379e+00
1.73301291e+00 8.34581107e-02 -5.03549457e-01 1.13032794e+00
1.55218530e+00 6.42043650e-02 6.47963285e-01 4.37435985e-01
7.55412340e-01 3.13771784e-01 6.90398455e-01 -1.96198951e-02
3.08594126e-02 6.45697773e-01 8.06924760e-01 -7.72062480e-01
-7.63494730e-01 4.35392559e-01 -1.57642379e-01 4.83814716e-01
-6.98517799e-01 3.16052109e-01 -1.28191102e+00 7.42782891e-01
-1.11118484e+00 -3.27379078e-01 -3.89214635e-01 2.18058586e+00
8.30824614e-01 -2.28093028e-01 -1.09980991e-02 -4.11841840e-01
1.57785967e-01 -3.72942269e-01 -7.01804832e-02 2.47089013e-01
6.28574267e-02 6.37701511e-01 1.12932456e+00 5.42847216e-01
-1.60453868e+00 9.58594561e-01 7.15617943e+00 6.43724144e-01
-9.57308233e-01 2.92836159e-01 2.08535910e-01 -2.67586797e-01
2.34677196e-01 -2.28496313e-01 -5.53075850e-01 6.12516522e-01
4.91022050e-01 2.66789883e-01 1.70378372e-01 6.76686466e-01
1.00561820e-01 -4.97825056e-01 -8.78624141e-01 1.10857987e+00
1.40816808e-01 -2.02111268e+00 -1.38432547e-01 2.35154852e-01
9.06089962e-01 3.40317994e-01 -3.75949554e-02 4.81149107e-02
-2.69169249e-02 -1.17577267e+00 4.71333295e-01 7.29007840e-01
1.40046310e+00 -1.70155555e-01 1.12447095e+00 -7.51583576e-01
-1.11166704e+00 -2.15017162e-02 2.86292851e-01 7.36519396e-01
4.89171781e-02 -7.04758940e-03 -1.12686193e+00 3.79079431e-01
9.31262612e-01 9.30376828e-01 -7.34450281e-01 2.06489539e+00
-1.26736939e-01 6.12390041e-01 -1.73611920e-02 6.43437445e-01
1.69271082e-01 -1.23900600e-01 9.07638669e-01 1.35803759e+00
1.02412499e-01 -2.68714249e-01 3.32084805e-01 9.06396806e-02
-1.08589984e-01 2.15025302e-02 -3.57579708e-01 2.32226193e-01
-4.68992665e-02 1.24589646e+00 -8.17621529e-01 -3.87559325e-01
-8.57641160e-01 1.16946590e+00 -5.60807586e-01 3.81624788e-01
-9.22756076e-01 -3.54479998e-02 6.77844167e-01 1.89624265e-01
-1.06617227e-01 -3.06023993e-02 -2.69812286e-01 -1.01556885e+00
-2.76579082e-01 -6.89600408e-01 4.89444733e-01 -1.05761850e+00
-5.84830582e-01 2.27617025e-01 -4.65076089e-01 -1.83167839e+00
2.78781027e-01 -1.19713366e+00 -3.75936568e-01 5.54939568e-01
-1.41395128e+00 -1.60591948e+00 -1.15632832e+00 6.43305421e-01
8.86996686e-01 -3.36886615e-01 1.08635318e+00 1.85854241e-01
-1.89961195e-01 3.07611614e-01 -1.85044721e-01 3.09852332e-01
9.98404920e-01 -1.81824815e+00 -1.18452378e-01 5.89841604e-01
-2.51601841e-02 3.45150381e-01 6.29926383e-01 -4.76386935e-01
-1.40869474e+00 -1.39752519e+00 6.72063157e-02 -8.20360124e-01
5.17425895e-01 3.68457772e-02 -6.84136078e-02 9.45008695e-01
1.74315080e-01 5.84406435e-01 7.41920769e-01 -2.24437460e-01
1.24979250e-01 5.43217480e-01 -1.02078485e+00 4.18845147e-01
1.28577685e+00 -3.28569323e-01 -6.28413320e-01 1.05994987e+00
6.00633919e-01 -1.63599122e+00 -1.09173095e+00 6.91047490e-01
6.46461666e-01 -7.63151109e-01 1.17141867e+00 -5.78858435e-01
4.62761998e-01 -2.27089331e-01 1.39988139e-01 -1.05088949e+00
3.93311977e-01 -8.33099067e-01 -1.00908421e-01 1.65978335e-02
8.95380303e-02 -1.55887127e-01 9.46688116e-01 -4.91196178e-02
-8.67048323e-01 -7.08253145e-01 -1.06246841e+00 -5.34838676e-01
-2.00400144e-01 -3.38613987e-01 -4.05304134e-01 6.54461205e-01
-5.05663455e-01 -4.59211111e-01 1.94464549e-01 3.11341494e-01
7.13459611e-01 -1.58195838e-01 6.63053751e-01 -1.28787661e+00
-7.63614029e-02 -7.80088961e-01 -8.49150658e-01 -2.22731248e-01
-1.39204681e-01 -1.02684462e+00 -1.60139903e-01 -1.94076502e+00
3.11296105e-01 7.34531283e-02 -1.20210767e-01 3.78982902e-01
-1.69927313e-03 7.43808031e-01 -1.32264137e-01 3.25957775e-01
-4.38279837e-01 -2.71484017e-01 1.62641716e+00 -3.91512007e-01
-2.12888598e-01 1.75497144e-01 -3.15841407e-01 1.05846429e+00
1.64498955e-01 1.53693810e-01 -1.20836712e-01 -2.48371094e-01
-1.88210830e-01 2.26650406e-02 8.95089865e-01 -1.29060829e+00
8.99164826e-02 3.70432138e-01 5.19718647e-01 -4.39850420e-01
2.79000819e-01 -8.11899483e-01 -1.81878999e-01 9.54209507e-01
2.48284116e-02 -7.05792665e-01 5.53786099e-01 5.09926677e-01
-4.63025868e-01 1.58299267e-01 8.13472450e-01 -5.51376402e-01
-1.02817345e+00 4.71515149e-01 -2.83389777e-01 2.45051682e-01
1.36670196e+00 -5.85927367e-01 -3.11768055e-03 3.01979333e-01
-1.61601484e+00 3.07765454e-01 4.67992276e-01 5.57048857e-01
6.08201027e-01 -5.98713875e-01 -7.10307300e-01 1.57855421e-01
5.19001484e-01 4.63197052e-01 3.57870460e-01 1.66379488e+00
-1.07054210e+00 6.48619533e-01 -9.25388001e-03 -9.70882297e-01
-1.82852900e+00 2.20285743e-01 1.22012544e+00 -1.12363044e-03
-1.02028191e+00 1.46791947e+00 3.59661698e-01 -9.07763764e-02
5.89905441e-01 -9.91599083e-01 3.40177445e-03 -2.47840565e-02
3.25089902e-01 -1.07161380e-01 2.97241747e-01 -6.96693733e-02
-3.79920870e-01 9.71124589e-01 -1.68794785e-02 2.72921860e-01
7.94306695e-01 1.69193611e-01 -1.21508367e-01 -6.53888807e-02
1.08276606e+00 1.78633288e-01 -9.10924613e-01 7.59041086e-02
-4.77131866e-02 -4.23360020e-01 5.22641242e-01 -1.51753032e+00
-1.14358115e+00 4.08514023e-01 1.31237400e+00 -4.09672618e-01
1.10286355e+00 2.65197009e-01 7.43134141e-01 -3.37520778e-01
6.11745656e-01 -9.34354067e-01 -2.33398199e-01 1.18480101e-01
1.09963357e+00 -1.43867362e+00 -3.73358607e-01 -1.00407422e+00
-4.66552913e-01 1.35683417e+00 5.02265096e-01 1.62864342e-01
5.69130123e-01 5.43353975e-01 1.67035073e-01 -5.09745300e-01
2.61084199e-01 -3.07750702e-01 1.00712228e+00 6.53453708e-01
4.12628919e-01 3.25640440e-01 -1.43736556e-01 2.79479921e-01
-2.76230127e-01 3.01759124e-01 6.37384474e-01 8.74080837e-01
-1.48589745e-01 -4.20617819e-01 2.42969431e-02 7.52185464e-01
-3.31672937e-01 -2.65260249e-01 -1.22539975e-01 8.05632710e-01
5.37577629e-01 4.94227141e-01 -1.68183312e-01 1.82144478e-01
4.72674817e-01 -3.73775214e-01 8.68623853e-01 -8.20228755e-01
-6.15563929e-01 5.99137396e-02 4.23630595e-01 -1.31031263e+00
-5.10296881e-01 -6.04300559e-01 -1.26446056e+00 6.05750680e-01
3.80280502e-02 6.82508051e-02 9.52040136e-01 7.49259174e-01
1.17818154e-01 1.16810334e+00 -3.68128121e-02 -1.05790055e+00
9.53612179e-02 -6.96372271e-01 -3.49279344e-01 1.14594899e-01
5.95562637e-01 -6.20871127e-01 -3.00394416e-01 4.45589006e-01]
|
[14.016230583190918, -3.263078212738037]
|
98375d3c-b944-4747-a346-0f7a7e861ff2
|
enhanced-direct-speech-to-speech-translation
|
2204.02967
| null |
https://arxiv.org/abs/2204.02967v3
|
https://arxiv.org/pdf/2204.02967v3.pdf
|
Enhanced Direct Speech-to-Speech Translation Using Self-supervised Pre-training and Data Augmentation
|
Direct speech-to-speech translation (S2ST) models suffer from data scarcity issues as there exists little parallel S2ST data, compared to the amount of data available for conventional cascaded systems that consist of automatic speech recognition (ASR), machine translation (MT), and text-to-speech (TTS) synthesis. In this work, we explore self-supervised pre-training with unlabeled speech data and data augmentation to tackle this issue. We take advantage of a recently proposed speech-to-unit translation (S2UT) framework that encodes target speech into discrete representations, and transfer pre-training and efficient partial finetuning techniques that work well for speech-to-text translation (S2T) to the S2UT domain by studying both speech encoder and discrete unit decoder pre-training. Our experiments on Spanish-English translation show that self-supervised pre-training consistently improves model performance compared with multitask learning with an average 6.6-12.1 BLEU gain, and it can be further combined with data augmentation techniques that apply MT to create weakly supervised training data. Audio samples are available at: https://facebookresearch.github.io/speech_translation/enhanced_direct_s2st_units/index.html .
|
['Ann Lee', 'Wei-Ning Hsu', 'Jiatao Gu', 'Yossi Adi', 'Juan Pino', 'Changhan Wang', 'Peng-Jen Chen', 'Sravya Popuri']
|
2022-04-06
| null | null | null | null |
['speech-to-text-translation', 'speech-to-speech-translation']
|
['natural-language-processing', 'speech']
|
[ 5.85997999e-01 3.43665630e-01 -5.54341972e-01 -4.42404389e-01
-1.63411415e+00 -4.62089658e-01 7.74598837e-01 -3.63130242e-01
-1.26404300e-01 8.47379446e-01 5.78747869e-01 -9.96288121e-01
7.32360721e-01 -2.41488934e-01 -9.00091171e-01 -4.14015085e-01
6.09915912e-01 7.75373936e-01 -2.00663898e-02 -5.00066459e-01
-4.41531062e-01 -5.75421145e-03 -8.94513369e-01 7.57699370e-01
9.16037261e-01 7.07019866e-01 4.30337489e-01 8.21799994e-01
-2.19880715e-01 6.35131657e-01 -6.20163620e-01 -4.15332079e-01
2.51986712e-01 -8.76928568e-01 -7.59393394e-01 2.75972448e-02
2.07925975e-01 -3.79318714e-01 -4.13199306e-01 7.13975430e-01
7.32106030e-01 -1.68782733e-02 5.06675184e-01 -9.81805146e-01
-9.21567142e-01 9.41881299e-01 -2.67370760e-01 1.99320495e-01
1.89251959e-01 -1.26648143e-01 8.15764964e-01 -1.30053568e+00
3.89735997e-01 1.49251854e+00 3.18609446e-01 9.64326501e-01
-1.19627643e+00 -7.59271920e-01 -2.31341034e-01 -5.10643283e-03
-1.10235393e+00 -1.24786007e+00 5.17363012e-01 -1.90648347e-01
1.24195647e+00 3.32606584e-01 2.11931825e-01 1.75164795e+00
-1.36109635e-01 1.12600195e+00 1.19427907e+00 -8.57812464e-01
8.43168050e-02 2.73966461e-01 -4.03212309e-01 4.00734037e-01
-4.61370856e-01 2.53273427e-01 -8.00954282e-01 1.07470624e-01
8.19454432e-01 -2.96140760e-01 -5.66954128e-02 3.61105740e-01
-1.55084109e+00 6.98921859e-01 1.04759961e-01 3.31503510e-01
-1.58864468e-01 -9.42364112e-02 5.92810631e-01 8.50027382e-01
9.39996958e-01 9.21744034e-02 -8.33569705e-01 -2.88152486e-01
-1.13162124e+00 -3.67329299e-01 5.72609186e-01 1.31738162e+00
6.35096669e-01 6.51874721e-01 -3.58328819e-01 1.31691074e+00
2.08524868e-01 9.26132202e-01 1.04849398e+00 -6.74001992e-01
1.08920538e+00 1.74305186e-01 -2.27059245e-01 5.47754802e-02
2.33113497e-01 -4.65270460e-01 -1.01090515e+00 -1.96623504e-01
6.31638244e-02 -4.30160344e-01 -1.16852033e+00 1.63719809e+00
1.18424244e-01 1.53552875e-01 4.08790231e-01 7.43223846e-01
8.38858485e-01 1.14923131e+00 -1.11051947e-01 -5.10613620e-01
1.03822398e+00 -1.55052161e+00 -9.41211939e-01 -4.85242009e-01
8.17852020e-01 -1.18832529e+00 1.38724291e+00 -1.01396963e-01
-1.34288418e+00 -7.94261038e-01 -7.91703403e-01 -1.89166456e-01
-1.90592602e-01 5.26253641e-01 2.44786986e-03 5.55172741e-01
-1.34843361e+00 1.46722913e-01 -9.64429855e-01 -4.42004234e-01
1.13332830e-01 3.89959365e-01 -2.44474381e-01 1.12289714e-03
-1.40145218e+00 1.05097365e+00 2.70128727e-01 -1.60153002e-01
-1.15882695e+00 -3.17564636e-01 -8.18849027e-01 -1.51741832e-01
3.81309688e-01 -6.59182787e-01 1.78032792e+00 -1.32292676e+00
-2.09223032e+00 5.62720656e-01 -6.45382285e-01 -4.89362001e-01
3.76176953e-01 -1.27044365e-01 -5.75748742e-01 3.49831879e-02
2.22714040e-02 6.94904566e-01 1.15958786e+00 -1.02420413e+00
-4.73109186e-01 -2.42693558e-01 -4.84531075e-01 4.93857324e-01
-4.82264966e-01 4.65805411e-01 -3.44616383e-01 -1.12045085e+00
-2.96253543e-02 -9.05139863e-01 -5.44256205e-03 -5.92227578e-01
-4.70634967e-01 -1.23689994e-01 8.27273846e-01 -1.09226990e+00
1.30943322e+00 -1.93705416e+00 3.15638512e-01 -3.39493334e-01
-3.22036445e-01 6.76912546e-01 -5.20476580e-01 7.02190399e-01
-8.10945407e-02 3.34059708e-02 -3.18860382e-01 -9.21365321e-01
-1.59299582e-01 3.54166329e-01 -4.16234016e-01 -9.34523568e-02
4.73105192e-01 1.26569462e+00 -6.68929636e-01 -3.40927362e-01
8.35036039e-02 3.47078919e-01 -2.21127063e-01 4.56453323e-01
-3.20431143e-01 7.82562375e-01 -2.59925097e-01 8.20024490e-01
2.17465326e-01 -7.39474371e-02 3.91278006e-02 3.00219417e-01
-4.39350046e-02 1.04841328e+00 -5.24000883e-01 1.70210361e+00
-7.55756319e-01 6.56858742e-01 1.66868031e-01 -1.03576159e+00
1.09808469e+00 9.33737755e-01 1.85970172e-01 -8.39461386e-01
8.94440487e-02 6.62827730e-01 -6.32083714e-02 -2.64577359e-01
3.37452650e-01 -2.59493589e-01 1.11737415e-01 5.21656215e-01
4.15150583e-01 -1.58254564e-01 -1.01053640e-01 1.91964954e-02
8.66970479e-01 1.67704657e-01 1.31000504e-01 1.66213930e-01
4.30961192e-01 -1.66394301e-02 3.48847598e-01 3.90178710e-01
1.09266445e-01 6.85136020e-01 -4.72808741e-02 1.04760669e-01
-1.50883579e+00 -9.50535297e-01 2.64032960e-01 1.51755333e+00
-5.78468025e-01 -3.26936781e-01 -9.97738004e-01 -6.64227724e-01
-3.85828018e-01 8.70931625e-01 -1.83082923e-01 -1.60305068e-01
-7.53745794e-01 -5.07529914e-01 8.85546207e-01 5.30615568e-01
2.99300104e-01 -1.04840851e+00 4.62248951e-01 4.48421776e-01
-5.86833894e-01 -1.34547865e+00 -8.72336507e-01 3.96201342e-01
-1.05153000e+00 -1.99949056e-01 -1.15176547e+00 -1.05521953e+00
4.87177730e-01 2.13757157e-01 9.08847570e-01 -4.03574795e-01
5.37669420e-01 -2.04522423e-02 -6.51151896e-01 -3.04921687e-01
-1.34498596e+00 4.52261031e-01 3.75013351e-01 8.34762957e-03
2.42148221e-01 -5.01897633e-01 -1.26119867e-01 5.32170713e-01
-6.34950042e-01 4.15679514e-01 1.12827313e+00 1.08575833e+00
5.33565164e-01 -6.65810227e-01 9.56985772e-01 -6.24277771e-01
6.95757389e-01 -4.21195537e-01 -2.92174011e-01 4.10294682e-01
-7.73184836e-01 -4.42914069e-02 7.30490565e-01 -6.89394176e-01
-1.10467136e+00 4.24173884e-02 -3.65844190e-01 -7.46774971e-01
-9.76175964e-02 5.04205406e-01 -1.52800754e-01 3.70654434e-01
7.38716900e-01 6.66470826e-01 1.42543957e-01 -6.50832117e-01
4.51199621e-01 1.47437072e+00 3.99357170e-01 -4.20165986e-01
8.31292450e-01 -1.63092524e-01 -6.58161223e-01 -9.04604495e-01
-7.02788711e-01 -4.18833792e-01 -7.55127013e-01 1.61431804e-01
6.68154001e-01 -1.28253222e+00 1.30436555e-01 2.79873103e-01
-1.22806871e+00 -8.67744267e-01 -2.77321130e-01 6.50030136e-01
-7.03131258e-01 1.51368827e-01 -8.79021585e-01 -7.83132553e-01
-6.81636989e-01 -1.27135527e+00 1.24617362e+00 -4.38511550e-01
-6.37093857e-02 -8.49402070e-01 -2.83794980e-02 8.64485025e-01
5.80018580e-01 -5.42142212e-01 6.11745298e-01 -9.18952525e-01
-3.01765025e-01 1.13376111e-01 5.15820198e-02 9.92586195e-01
3.06216627e-01 -3.87443960e-01 -9.69503760e-01 -4.90127265e-01
-4.87242751e-02 -4.80984896e-01 6.74297452e-01 2.94840366e-01
5.90062141e-01 -7.70157993e-01 3.08838915e-02 4.64412332e-01
6.67452335e-01 2.01683521e-01 4.36448812e-01 2.52397563e-02
6.90293729e-01 3.52688819e-01 4.71938014e-01 3.04079764e-02
4.58234429e-01 9.01246727e-01 -1.23117484e-01 -3.84759992e-01
-8.12371671e-01 -6.08605385e-01 1.09314382e+00 1.77419841e+00
1.17642537e-01 -3.45351785e-01 -8.85270059e-01 4.65835452e-01
-1.72118807e+00 -5.98663092e-01 -4.59996201e-02 2.12318206e+00
1.21090341e+00 7.59176314e-02 2.33577132e-01 1.26009658e-01
8.71364892e-01 -8.62582996e-02 -4.57444817e-01 -5.70993960e-01
-1.39048815e-01 2.91247398e-01 5.18908381e-01 8.41018736e-01
-7.23051727e-01 1.42323816e+00 5.78090668e+00 1.23751271e+00
-1.25482786e+00 7.78989136e-01 7.68012106e-01 5.38276462e-03
-3.08866769e-01 -3.55875120e-03 -8.99980307e-01 4.85727191e-01
1.67961526e+00 -1.17899381e-01 6.89647734e-01 5.75950801e-01
6.07928216e-01 6.02819383e-01 -1.01769805e+00 9.06056106e-01
-4.27080179e-03 -1.17554522e+00 3.38323414e-01 -6.22849353e-02
8.99325550e-01 5.41521311e-01 1.62493229e-01 5.92470467e-01
2.46087193e-01 -8.17111194e-01 7.84961104e-01 -1.00113377e-01
1.55783463e+00 -3.14667910e-01 4.99274701e-01 5.79247653e-01
-1.02289522e+00 3.93105904e-03 -2.46599585e-01 5.16609438e-02
2.48816311e-01 3.63507122e-01 -1.62094295e+00 6.16117179e-01
2.84279525e-01 6.26170576e-01 -2.41551891e-01 2.93904245e-01
-4.84972060e-01 1.29653227e+00 -2.54020095e-01 -3.34724449e-02
3.30583930e-01 -2.22917095e-01 5.64318299e-01 1.47453368e+00
6.59367561e-01 -1.14370584e-01 4.82521318e-02 4.17371452e-01
-4.32169855e-01 3.51232946e-01 -5.06177664e-01 -3.50190312e-01
6.55059099e-01 7.44441509e-01 -2.32931912e-01 -7.45360970e-01
-5.26984394e-01 1.27773356e+00 1.33994624e-01 6.22925818e-01
-4.87488359e-01 -1.25108054e-02 4.91867691e-01 2.31718108e-01
1.90318659e-01 -4.16712999e-01 -2.40313336e-01 -1.40337706e+00
5.67071065e-02 -1.24985290e+00 6.84123263e-02 -7.97519624e-01
-1.03834331e+00 9.48540807e-01 -2.80423284e-01 -1.41115916e+00
-7.48575330e-01 -2.53631651e-01 -3.30250829e-01 1.07315719e+00
-1.54011190e+00 -1.63849044e+00 3.31856102e-01 7.01696038e-01
1.35982621e+00 -6.55473948e-01 9.33048666e-01 4.55596119e-01
-5.64727783e-01 9.75225508e-01 5.24328172e-01 2.23463684e-01
9.93809700e-01 -1.03118777e+00 9.79514360e-01 8.96543682e-01
4.02581513e-01 3.30079705e-01 3.03019464e-01 -7.13625491e-01
-1.34517157e+00 -1.28283036e+00 1.27051556e+00 -4.31788445e-01
7.24726796e-01 -7.13084877e-01 -7.44793892e-01 9.71424282e-01
4.65304196e-01 -1.45700634e-01 5.88699043e-01 -1.43420950e-01
-2.04522282e-01 -1.23553894e-01 -6.86392128e-01 6.66004658e-01
1.02523422e+00 -8.83482456e-01 -5.74235260e-01 4.44583565e-01
1.22216475e+00 -4.56750602e-01 -7.20874190e-01 3.06283474e-01
1.77045673e-01 -2.41792649e-01 8.02367985e-01 -6.22097492e-01
4.48379457e-01 -3.29099521e-02 -4.91496503e-01 -1.65281677e+00
-5.59765212e-02 -1.10142910e+00 -2.45412007e-01 1.35815287e+00
9.70650375e-01 -5.61668038e-01 4.47580695e-01 -1.25259802e-01
-7.74826229e-01 -5.38251102e-01 -1.25107348e+00 -9.68333781e-01
2.70924270e-01 -4.00478363e-01 4.50650066e-01 1.03648722e+00
1.07417755e-01 9.32860911e-01 -6.74732208e-01 8.98432285e-02
2.91316152e-01 -2.72805512e-01 7.31331527e-01 -5.64563930e-01
-4.38833833e-01 -3.64055634e-01 1.87219888e-01 -1.52173936e+00
1.29845530e-01 -1.23170090e+00 5.88466674e-02 -1.54001427e+00
-1.78805783e-01 -2.35743672e-01 -1.52339280e-01 8.01478684e-01
-1.39696106e-01 2.97419459e-01 2.48480565e-03 4.83164668e-01
-2.22046673e-01 8.62899601e-01 1.52430868e+00 -1.44221485e-01
-3.46059948e-01 3.35692376e-01 -3.95934820e-01 1.16640531e-01
9.41609085e-01 -4.69243109e-01 -3.98438603e-01 -8.03863943e-01
-3.30225796e-01 5.44769168e-01 -1.99941874e-01 -5.96959591e-01
1.18866041e-01 -5.39300293e-02 7.93753266e-02 -3.72023940e-01
5.78568339e-01 -4.84919280e-01 -1.70013577e-01 2.92927295e-01
-6.41756475e-01 6.37610555e-02 1.53224126e-01 2.50702262e-01
-4.91908133e-01 9.38612819e-02 6.38694763e-01 -6.89946786e-02
-2.13302672e-01 2.60094613e-01 -6.45063639e-01 -1.40709177e-01
5.39880693e-01 5.12977652e-02 -1.95461050e-01 -6.97399497e-01
-9.08653259e-01 1.74486823e-02 4.28545028e-02 6.99917555e-01
6.72746122e-01 -1.57894397e+00 -1.33871520e+00 3.98449600e-01
9.34185982e-02 -1.36927843e-01 -2.63290256e-01 8.92857194e-01
6.19337149e-02 6.34365082e-01 1.05435550e-01 -6.43048763e-01
-1.17877710e+00 2.90796906e-01 6.07969463e-02 -4.21058945e-02
-3.52131426e-01 7.17666686e-01 -1.47324968e-02 -8.44404221e-01
1.28341958e-01 -3.85154545e-01 3.21654469e-01 -9.62485373e-02
2.70616889e-01 3.01206410e-01 4.04798537e-01 -7.74152815e-01
-7.57425874e-02 2.31639221e-02 -1.22238897e-01 -7.19793856e-01
1.22154498e+00 -4.53682363e-01 1.32881716e-01 5.72812736e-01
1.19350231e+00 -1.55217081e-01 -1.00033474e+00 -7.32371569e-01
-2.78897416e-02 -1.17351629e-01 -3.11490800e-02 -1.04817605e+00
-7.35078037e-01 1.11163867e+00 2.93972313e-01 -1.46300532e-02
1.05457819e+00 1.62828982e-01 1.15901613e+00 5.19065022e-01
1.15300007e-01 -1.14465594e+00 1.55985028e-01 8.08792174e-01
1.05690289e+00 -1.39101553e+00 -7.23251343e-01 -2.51372308e-01
-8.71083319e-01 9.75408375e-01 3.39423537e-01 3.28444034e-01
3.44801843e-01 4.58202004e-01 5.47711849e-01 4.93798673e-01
-1.04609442e+00 -2.40791321e-01 3.40560436e-01 4.68079239e-01
6.31135523e-01 3.01718116e-01 -2.53205895e-02 3.99585336e-01
-3.93523604e-01 -1.11865468e-01 3.10254753e-01 6.69405043e-01
-3.87428194e-01 -1.50568306e+00 -4.73160684e-01 3.10927182e-01
-4.12177533e-01 -5.66794395e-01 -5.50218582e-01 2.90344834e-01
-3.08289498e-01 1.28737438e+00 -2.18603238e-01 -5.59232652e-01
2.71741569e-01 5.00069320e-01 2.17984617e-01 -9.42309260e-01
-5.42362809e-01 7.48498559e-01 3.35355848e-01 -2.36848146e-01
-1.84745178e-01 -6.22385025e-01 -9.53724563e-01 -1.23310611e-01
-3.92764539e-01 2.18616679e-01 9.46204722e-01 9.56758976e-01
3.72439712e-01 5.72125971e-01 9.69690979e-01 -7.75903404e-01
-8.07244241e-01 -1.60706878e+00 -1.55472299e-02 7.14109838e-02
7.05159366e-01 -7.04083070e-02 -2.95673281e-01 3.99023354e-01]
|
[14.531233787536621, 7.127433776855469]
|
db84d439-64ee-418e-b75e-d75e84eefef3
|
prompting-for-multimodal-hateful-meme
|
2302.04156
| null |
https://arxiv.org/abs/2302.04156v1
|
https://arxiv.org/pdf/2302.04156v1.pdf
|
Prompting for Multimodal Hateful Meme Classification
|
Hateful meme classification is a challenging multimodal task that requires complex reasoning and contextual background knowledge. Ideally, we could leverage an explicit external knowledge base to supplement contextual and cultural information in hateful memes. However, there is no known explicit external knowledge base that could provide such hate speech contextual information. To address this gap, we propose PromptHate, a simple yet effective prompt-based model that prompts pre-trained language models (PLMs) for hateful meme classification. Specifically, we construct simple prompts and provide a few in-context examples to exploit the implicit knowledge in the pre-trained RoBERTa language model for hateful meme classification. We conduct extensive experiments on two publicly available hateful and offensive meme datasets. Our experimental results show that PromptHate is able to achieve a high AUC of 90.96, outperforming state-of-the-art baselines on the hateful meme classification task. We also perform fine-grained analyses and case studies on various prompt settings and demonstrate the effectiveness of the prompts on hateful meme classification.
|
['Jing Jiang', 'Wen-Haw Chong', 'Roy Ka-Wei Lee', 'Rui Cao']
|
2023-02-08
| null | null | null | null |
['meme-classification']
|
['natural-language-processing']
|
[-2.19672516e-01 -3.94290239e-01 -2.28140861e-01 -8.54982883e-02
-6.07970595e-01 -8.85985792e-01 9.23053324e-01 1.55633271e-01
-4.34506506e-01 6.34780705e-01 6.15822196e-01 -2.49688141e-03
4.04002339e-01 -4.61061835e-01 -2.20598489e-01 -4.54708219e-01
2.41540045e-01 -7.50001241e-03 6.76215887e-02 -5.13778806e-01
5.79385161e-01 9.45112854e-02 -9.49743092e-01 6.98288381e-01
9.82572079e-01 4.17306781e-01 -2.56082505e-01 7.76549280e-01
-2.24832714e-01 1.70277607e+00 -1.11011004e+00 -1.00762701e+00
-2.26807997e-01 -2.57575601e-01 -1.00910962e+00 -1.07032515e-01
9.39509928e-01 -6.99184299e-01 -4.75338638e-01 8.66320193e-01
6.08731508e-01 6.03082664e-02 8.31932425e-01 -9.88566995e-01
-1.33037484e+00 3.72300625e-01 -4.23055381e-01 4.74342108e-01
4.27524984e-01 3.18813890e-01 8.24851811e-01 -1.13458943e+00
7.13133156e-01 1.24801469e+00 8.69525552e-01 7.77476311e-01
-1.11169958e+00 -7.13624120e-01 -1.45311162e-01 2.72783220e-01
-1.22847867e+00 -2.84083456e-01 9.96583343e-01 -9.43727016e-01
9.55291808e-01 2.14641616e-01 1.76266477e-01 1.98149002e+00
1.09068900e-01 9.63707089e-01 1.42850745e+00 -2.24983543e-01
-1.00922838e-01 2.98708230e-01 5.18338680e-01 9.68475699e-01
-2.67422080e-01 -3.27219427e-01 -9.95382607e-01 -6.76787198e-01
3.23466063e-01 1.96525544e-01 -5.22494651e-02 4.21957642e-01
-1.06699121e+00 1.02941322e+00 3.19319040e-01 2.43702114e-01
-1.04743794e-01 -2.14755982e-02 8.03712666e-01 2.65853375e-01
9.44513798e-01 9.90992069e-01 1.80961505e-01 -1.07059866e-01
-9.38195586e-01 1.15473218e-01 7.61656642e-01 3.66203606e-01
6.79673731e-01 -1.22957438e-01 -5.66559911e-01 1.25819051e+00
-2.64286011e-01 9.19067442e-01 -4.66629565e-02 -6.66472197e-01
6.70094073e-01 5.78359008e-01 2.31406614e-01 -1.53977668e+00
-4.37001526e-01 -1.70885120e-03 -6.24062121e-01 -4.14329261e-01
1.94552436e-01 -3.03114802e-01 -8.78876448e-01 1.59751451e+00
8.01042691e-02 1.91508576e-01 -3.72756720e-01 7.96589613e-01
8.66558492e-01 9.44403827e-01 6.82371318e-01 1.05676115e-01
1.16072595e+00 -1.09605360e+00 -9.16250706e-01 -4.40065086e-01
9.14919496e-01 -8.80442679e-01 1.24881792e+00 4.53369506e-02
-5.15552223e-01 -9.02267992e-02 -4.99313563e-01 -2.38531604e-01
-6.43690109e-01 2.25755602e-01 6.01340294e-01 4.71214592e-01
-5.81710458e-01 1.52168423e-01 -3.69047970e-01 -6.39460921e-01
4.15303826e-01 -3.61704648e-01 -6.19153976e-01 -1.76682502e-01
-1.44298232e+00 1.22801387e+00 2.76253730e-01 -1.06595621e-01
-1.08538318e+00 -7.71652937e-01 -6.25813127e-01 -3.76477331e-01
5.40187001e-01 -3.31510276e-01 9.31702733e-01 -7.86180794e-01
-1.00286853e+00 1.25940967e+00 -3.51004824e-02 2.25871485e-02
9.17289853e-02 -6.51494741e-01 -4.57237929e-01 4.17749465e-01
2.09191829e-01 4.01389331e-01 1.07760262e+00 -1.25740159e+00
-6.93327636e-02 -1.26718700e-01 1.86265156e-01 -1.49349660e-01
-1.07441175e+00 6.91888690e-01 -9.22860391e-03 -9.81360614e-01
-8.38009000e-01 -1.04729617e+00 3.85884166e-01 -5.69092810e-01
-7.35543430e-01 -2.53671348e-01 1.01904583e+00 -1.29020977e+00
1.95783103e+00 -2.20777011e+00 9.04017687e-02 -3.54664922e-02
4.93358284e-01 7.21881688e-01 -3.72817606e-01 8.87013495e-01
3.07682097e-01 2.89676934e-01 -1.40641123e-01 -3.84242594e-01
1.55285120e-01 -8.24442357e-02 -8.46094966e-01 3.38443428e-01
4.64142591e-01 1.05995953e+00 -1.17251790e+00 -5.04605830e-01
4.21560789e-03 5.21985412e-01 -5.43857694e-01 7.90171921e-01
-1.71562254e-01 2.38063544e-01 -3.90555441e-01 8.59910965e-01
3.71585667e-01 -2.95042247e-01 3.66416611e-02 2.41456833e-02
2.51379795e-02 2.04748705e-01 -2.23262995e-01 1.22410882e+00
-2.27872506e-01 8.67315769e-01 8.40492453e-03 -1.65893033e-01
9.45186615e-01 2.19909370e-01 -7.61020929e-02 -6.31990016e-01
2.30587497e-01 1.53193459e-01 -6.18807018e-01 -9.00784731e-01
7.27359116e-01 -2.39977255e-01 -4.78519052e-01 5.75982809e-01
1.16526134e-01 3.04482073e-01 8.82317051e-02 5.86908519e-01
1.47178745e+00 -2.85349805e-02 2.77808309e-01 -1.35453567e-01
4.38810289e-01 1.31316856e-01 1.48771867e-01 8.41549337e-01
-6.12845421e-01 3.34696054e-01 6.50847852e-01 -5.65029740e-01
-9.54184115e-01 -8.72586191e-01 2.67128617e-01 1.85272539e+00
-2.51752406e-01 -8.36075842e-01 -7.94352949e-01 -1.04345953e+00
-6.04262315e-02 7.49113798e-01 -1.02080846e+00 -1.34620816e-01
-4.08228189e-01 -8.43046188e-01 1.02584052e+00 4.04884815e-01
6.41080797e-01 -1.09682226e+00 -4.21195388e-01 -6.34315908e-02
-4.42911178e-01 -1.06946993e+00 -5.72994411e-01 -4.60468680e-02
-1.97152734e-01 -1.00670159e+00 -6.41771972e-01 -4.73522961e-01
3.73351932e-01 3.39870244e-01 1.17149329e+00 4.61919218e-01
-5.30430749e-02 5.03385544e-01 -5.92036247e-01 -8.12418088e-02
-4.77482080e-01 3.27022016e-01 -9.89924371e-02 8.96061683e-05
8.86544347e-01 -1.65150866e-01 -2.36353129e-01 2.16659948e-01
-9.55713093e-01 1.49001300e-01 1.95736662e-01 1.00678170e+00
-2.63758123e-01 -5.61802626e-01 5.78945518e-01 -1.06625235e+00
9.62493837e-01 -1.12740576e+00 9.23719928e-02 4.08429027e-01
1.61098495e-01 -3.55817199e-01 7.47876942e-01 -6.25175714e-01
-1.15074050e+00 -4.51344341e-01 8.31001550e-02 -4.72797424e-01
-2.88844198e-01 5.11349618e-01 4.20708805e-01 1.27530560e-01
1.05367148e+00 1.83235630e-02 -4.56061721e-01 -6.11978233e-01
4.45204943e-01 9.80128050e-01 7.36466110e-01 -8.48679721e-01
9.77057874e-01 1.14435479e-01 -3.41436476e-01 -1.00221586e+00
-1.50272596e+00 -6.35965586e-01 -5.25245905e-01 -4.12721753e-01
1.07138669e+00 -7.78237641e-01 -5.31269789e-01 7.65977621e-01
-1.45475316e+00 -3.30042005e-01 6.44290566e-01 -1.88678712e-01
-9.96379852e-02 3.83785486e-01 -1.23806691e+00 -1.09328234e+00
-4.06446427e-01 -5.56813419e-01 9.85918105e-01 -2.88474839e-02
-5.46125531e-01 -1.24581420e+00 5.67403316e-01 8.10813725e-01
5.05616546e-01 4.86815989e-01 1.14265049e+00 -9.03312802e-01
-1.56076401e-01 -1.20413765e-01 -3.64365608e-01 3.05714488e-01
-5.18473573e-02 -4.30580378e-02 -1.26465714e+00 -1.05864957e-01
-2.98293352e-01 -1.17148876e+00 1.02905083e+00 -5.15171111e-01
6.69551313e-01 -7.19748139e-01 -1.92377269e-01 1.83297157e-01
1.13025486e+00 -3.01836610e-01 5.13216019e-01 4.88537222e-01
1.03464293e+00 8.16446006e-01 4.48347032e-01 6.72151089e-01
3.34657282e-01 4.60431874e-01 3.54088756e-04 3.29421670e-03
-4.83585075e-02 -6.07338011e-01 6.30439937e-01 1.00263679e+00
3.95943709e-02 -2.23838478e-01 -1.35251987e+00 7.49070525e-01
-1.89798379e+00 -1.45567667e+00 4.23687324e-02 1.51817977e+00
1.12562144e+00 -4.00358707e-01 3.09639454e-01 -3.67031485e-01
7.26723015e-01 6.67070985e-01 -1.32518038e-01 -6.79241717e-01
-1.73100919e-01 -2.65439779e-01 8.26096460e-02 4.98348117e-01
-1.49120486e+00 1.40408123e+00 6.57587814e+00 8.63454998e-01
-9.86947894e-01 5.54909706e-01 3.69962364e-01 -3.59436125e-01
-1.13667689e-01 -3.95006180e-01 -5.49509764e-01 7.85030901e-01
9.88136649e-01 2.97764301e-01 6.85671926e-01 6.71616375e-01
-2.58786559e-01 -3.36463302e-02 -9.29967821e-01 9.86990273e-01
5.49413741e-01 -1.44336259e+00 -6.75936788e-02 -1.42111003e-01
9.33273911e-01 3.09254806e-02 2.28888974e-01 6.80764914e-01
2.61996210e-01 -1.13663745e+00 5.07306218e-01 5.70236385e-01
5.14937162e-01 -7.04427958e-01 7.16538429e-01 3.84603649e-01
-5.99299431e-01 -2.72875667e-01 -2.27884784e-01 -4.49878313e-02
-2.13891435e-02 3.22840303e-01 -8.46787691e-01 -5.70694320e-02
6.06005847e-01 6.34263575e-01 -1.11361277e+00 3.75668675e-01
-4.67988878e-01 1.12397087e+00 1.10582605e-01 -1.88398108e-01
4.88525271e-01 2.43435621e-01 5.53890586e-01 2.09572697e+00
-2.09152520e-01 3.36961240e-01 3.01967353e-01 9.04831946e-01
-3.53877306e-01 2.45901331e-01 -9.74769115e-01 -8.60573888e-01
5.88477850e-01 1.28004098e+00 1.71885695e-02 -2.79656023e-01
-5.95703363e-01 1.23951900e+00 8.99177849e-01 4.61779296e-01
-9.43689406e-01 -4.14425462e-01 5.90187788e-01 -1.22656927e-01
-1.60970747e-01 -4.05464917e-01 -1.66389626e-02 -1.34680676e+00
-4.43268806e-01 -1.07790732e+00 5.87529421e-01 -8.60460341e-01
-2.08603859e+00 5.51346958e-01 -6.32484853e-02 -4.91801590e-01
-1.75756782e-01 -7.46084809e-01 -5.69175541e-01 8.02830696e-01
-1.24421251e+00 -1.60490859e+00 -1.65805176e-01 5.60013235e-01
3.27856779e-01 -3.93159598e-01 8.25204551e-01 1.57860532e-01
-8.66242945e-01 5.05173922e-01 -1.86840415e-01 7.76352108e-01
1.41421747e+00 -1.03303552e+00 1.42112687e-01 9.47147787e-01
-2.75307924e-01 1.06682241e+00 6.57606363e-01 -1.21534634e+00
-1.27344656e+00 -1.05325413e+00 1.03963697e+00 -1.58943152e+00
1.37281144e+00 -4.52419549e-01 -1.28722298e+00 7.92590261e-01
6.64596379e-01 -2.67183125e-01 1.26024294e+00 5.08025050e-01
-1.21349049e+00 6.53226674e-01 -8.66627216e-01 7.58673847e-01
9.66023088e-01 -1.33523977e+00 -1.09884465e+00 3.54600489e-01
5.39397657e-01 -1.01931445e-01 -8.16297412e-01 -4.52768318e-02
3.34337503e-01 -7.79784620e-01 6.82130754e-01 -1.24300313e+00
9.95673180e-01 -1.01214521e-01 -3.47719252e-01 -1.27698004e+00
-4.86197323e-01 -6.56783581e-01 -5.20407319e-01 1.56333232e+00
1.15099654e-01 -2.60349512e-01 2.36005574e-01 7.41234362e-01
1.04593627e-01 -5.11903584e-01 -4.23633248e-01 -6.48112535e-01
2.59927392e-01 -1.25273332e-01 1.09976992e-01 1.76124024e+00
3.28250974e-01 7.94851780e-01 -1.26835310e+00 1.01651959e-02
5.41764736e-01 -1.64476842e-01 8.83167684e-01 -8.09448302e-01
-4.79459623e-03 -1.46664709e-01 -7.68213943e-02 -4.73241866e-01
8.24585378e-01 -9.75898504e-01 -2.29904205e-01 -1.35069132e+00
1.03532982e+00 -8.43690708e-02 -3.07868868e-01 9.54677582e-01
-6.09768271e-01 6.13384664e-01 4.94106859e-01 3.79115820e-01
-1.03782547e+00 3.65320951e-01 8.56621027e-01 -4.26646680e-01
-4.67909612e-02 -9.44188952e-01 -5.92138648e-01 6.75768971e-01
7.26017058e-01 -5.13785303e-01 -1.35380909e-01 -4.37124521e-01
4.70955819e-01 -4.46172357e-01 7.05125511e-01 -6.67474449e-01
1.30055696e-01 -6.81527436e-01 3.07019413e-01 -3.42394114e-01
4.54669088e-01 -2.09179476e-01 -3.38550746e-01 4.67956141e-02
-5.52364647e-01 -4.74227071e-02 2.55409867e-01 3.81752700e-01
-1.41266058e-03 -1.83405474e-01 7.56892741e-01 -1.96718369e-02
-9.72267270e-01 4.22607884e-02 -5.97328901e-01 4.68876958e-01
7.22123802e-01 3.78846347e-01 -1.50913596e+00 -3.14663351e-01
-2.36327574e-01 -8.35474581e-03 6.94001019e-01 6.73063993e-01
4.76120889e-01 -1.44855428e+00 -8.38893533e-01 -3.22436839e-01
6.38794601e-01 -1.24742937e+00 3.14809620e-01 7.94807494e-01
-1.89664289e-01 4.71347153e-01 -4.60596055e-01 -1.73704579e-01
-1.19350207e+00 8.94880831e-01 1.17301814e-01 -1.97273027e-02
-3.71420920e-01 6.34186268e-01 2.29674101e-01 -5.43695807e-01
4.67169546e-02 7.18900740e-01 -1.54715791e-01 1.14528827e-01
9.98097897e-01 4.16834533e-01 -4.83089298e-01 -8.91936541e-01
-5.20093858e-01 2.99872309e-01 -1.01005122e-01 8.47112313e-02
1.04919493e+00 1.24166291e-02 -5.38053453e-01 4.85538036e-01
1.18745315e+00 4.75677520e-01 -7.74585605e-01 -2.98669338e-01
1.63091525e-01 -8.59323919e-01 -9.10941139e-02 -1.36410868e+00
-3.61307651e-01 9.98153150e-01 1.56048864e-01 2.16601878e-01
7.24460483e-01 6.07544594e-02 1.04740596e+00 8.85114253e-01
2.71622807e-01 -1.43983436e+00 6.98404789e-01 9.82063651e-01
1.06398892e+00 -1.34103012e+00 -2.88346201e-01 -1.47012323e-01
-1.01396286e+00 9.59693670e-01 1.08506882e+00 1.39513761e-01
1.10127181e-01 4.08157334e-02 4.59493458e-01 -4.01576847e-01
-9.23527718e-01 -2.14871168e-02 5.12784600e-01 4.79046911e-01
6.62563860e-01 6.80422336e-02 -1.67158455e-01 6.17931545e-01
1.48685485e-01 -2.92427689e-01 3.84294540e-01 9.68124986e-01
-6.75921023e-01 -5.82493246e-01 -6.19514287e-01 2.03055814e-01
-5.50083220e-01 -3.81488830e-01 -1.32158339e+00 5.61742723e-01
1.73280016e-02 1.10810840e+00 -3.76959592e-01 -8.78203392e-01
-8.03730488e-02 4.72176611e-01 3.64512324e-01 -7.51600802e-01
-1.02504122e+00 -3.42297673e-01 6.13978326e-01 -5.36988378e-01
-2.22675890e-01 -4.92379777e-02 -7.88788259e-01 -9.27118719e-01
1.04413785e-01 -6.27911761e-02 2.74709821e-01 9.63809788e-01
3.51538718e-01 -7.61647671e-02 5.65138102e-01 -6.86705768e-01
-4.08545017e-01 -1.04879677e+00 -3.25771064e-01 9.44649458e-01
4.41608369e-01 -7.49392211e-01 -2.61923462e-01 -4.39329110e-02]
|
[8.533036231994629, 10.648653984069824]
|
eb7d89e1-80f7-4899-8ae7-30c15c321e5b
|
enabling-data-diversity-efficient-automatic
|
2103.16493
| null |
https://arxiv.org/abs/2103.16493v1
|
https://arxiv.org/pdf/2103.16493v1.pdf
|
Enabling Data Diversity: Efficient Automatic Augmentation via Regularized Adversarial Training
|
Data augmentation has proved extremely useful by increasing training data variance to alleviate overfitting and improve deep neural networks' generalization performance. In medical image analysis, a well-designed augmentation policy usually requires much expert knowledge and is difficult to generalize to multiple tasks due to the vast discrepancies among pixel intensities, image appearances, and object shapes in different medical tasks. To automate medical data augmentation, we propose a regularized adversarial training framework via two min-max objectives and three differentiable augmentation models covering affine transformation, deformation, and appearance changes. Our method is more automatic and efficient than previous automatic augmentation methods, which still rely on pre-defined operations with human-specified ranges and costly bi-level optimization. Extensive experiments demonstrated that our approach, with less training overhead, achieves superior performance over state-of-the-art auto-augmentation methods on both tasks of 2D skin cancer classification and 3D organs-at-risk segmentation.
|
['Dimitris Metaxas', 'Mu Zhou', 'Zhiqiang Tang', 'Yunhe Gao']
|
2021-03-30
| null | null | null | null |
['skin-cancer-classification']
|
['medical']
|
[ 5.23641229e-01 2.94051558e-01 -2.32404664e-01 -4.21151996e-01
-8.52696121e-01 -4.30554628e-01 3.97797853e-01 2.40239903e-01
-5.98665416e-01 5.95661223e-01 -2.54636139e-01 -4.17801827e-01
3.70767325e-01 -5.89168429e-01 -6.52096272e-01 -8.26643646e-01
6.76342174e-02 5.12394369e-01 3.37800756e-02 -1.21899486e-01
-2.93631822e-01 6.33609533e-01 -8.03083718e-01 -9.09939110e-02
1.17470992e+00 1.03757203e+00 -2.57238865e-01 5.03318787e-01
-1.82250664e-01 3.85851979e-01 -5.19624710e-01 -4.73543108e-01
3.06782246e-01 -3.82294923e-01 -6.83881938e-01 3.71478647e-01
4.35432732e-01 -5.63123487e-02 -2.27805361e-01 1.10164499e+00
5.52134693e-01 3.25989798e-02 7.05753267e-01 -9.87725258e-01
-7.19409287e-01 1.50979266e-01 -7.47600853e-01 1.11097470e-01
-1.51208147e-01 2.82253712e-01 4.38469619e-01 -5.50738275e-01
2.62341619e-01 8.81009340e-01 1.07724059e+00 8.81771505e-01
-1.49529541e+00 -4.10397619e-01 2.24712193e-01 -5.25881410e-01
-1.08532524e+00 5.40662035e-02 8.23974431e-01 -5.09729505e-01
4.58644509e-01 4.09253776e-01 6.71666682e-01 1.06855571e+00
7.38427565e-02 8.40822995e-01 1.20953953e+00 -3.20481747e-01
1.13005064e-01 1.27479985e-01 -1.20214112e-01 9.50149536e-01
2.03652084e-01 -2.48439729e-01 3.56509089e-01 -1.45538658e-01
1.17263627e+00 1.40710101e-01 -2.16788575e-01 -5.89126706e-01
-1.06089509e+00 8.27010036e-01 6.00602269e-01 1.61610380e-01
-4.53881383e-01 1.71539918e-01 5.27746558e-01 3.71242240e-02
8.22119236e-01 6.10579491e-01 -5.92632830e-01 3.16848099e-01
-5.82323551e-01 6.12403117e-02 3.93621266e-01 5.54412484e-01
4.56883013e-01 2.12654516e-01 -2.14812800e-01 1.02593982e+00
-3.26169990e-02 3.03568751e-01 7.28062272e-01 -6.19128585e-01
2.46859029e-01 8.76877248e-01 -1.63262069e-01 -7.36253679e-01
-4.91792560e-01 -8.32876861e-01 -1.18884158e+00 3.37868601e-01
5.87593198e-01 -1.93570197e-01 -1.57125378e+00 1.82761252e+00
5.67162931e-01 1.18936203e-01 -7.22404048e-02 7.97186077e-01
7.40947723e-01 1.40601590e-01 4.22428191e-01 -1.05083108e-01
1.32679296e+00 -1.02884007e+00 -6.36921287e-01 -4.64975625e-01
6.93176746e-01 -4.59169030e-01 1.37103987e+00 5.28064929e-02
-1.16411161e+00 -3.97365481e-01 -9.34605420e-01 4.24340181e-02
-1.98513135e-01 2.03659013e-01 1.11827600e+00 8.64565015e-01
-6.98500693e-01 4.52949405e-01 -1.19058752e+00 -5.60883880e-02
9.46978807e-01 6.38299108e-01 -4.51328963e-01 1.65871695e-01
-8.36705983e-01 7.94224560e-01 2.00741440e-01 -9.70513895e-02
-6.87592208e-01 -1.12859833e+00 -1.04759109e+00 -1.92664534e-01
4.92362678e-01 -8.84033620e-01 1.14235389e+00 -1.16712880e+00
-1.53402460e+00 1.24727702e+00 4.25820410e-01 -4.29160833e-01
7.08272219e-01 -1.51654303e-01 -1.38884127e-01 2.57732570e-02
-1.89551041e-01 7.57363558e-01 7.90543675e-01 -1.15959895e+00
-6.83789849e-02 -6.46383941e-01 -1.92788780e-01 2.93516874e-01
-5.91673732e-01 -2.14053109e-01 -8.34002793e-01 -1.15290260e+00
3.58443335e-02 -1.06948268e+00 -9.30359364e-01 3.74129713e-01
-3.31454337e-01 2.68203527e-01 7.85472631e-01 -7.19915390e-01
7.73838818e-01 -2.07182741e+00 8.62022415e-02 1.67750999e-01
1.57731548e-01 5.61890125e-01 -1.89265043e-01 -4.72367346e-01
-1.12013303e-01 1.80549175e-01 -6.19080961e-01 -4.69392329e-01
-3.15492481e-01 3.72745872e-01 3.34383063e-02 5.33897102e-01
4.08903807e-01 1.02629995e+00 -8.54207873e-01 -6.30161166e-01
3.62832695e-01 6.75860584e-01 -5.33379078e-01 2.23757818e-01
-3.91596913e-01 7.83705115e-01 -5.35475850e-01 8.06418836e-01
5.38146377e-01 -4.93910968e-01 -2.22892296e-02 -1.87890545e-01
4.79076117e-01 -1.68606952e-01 -7.02535510e-01 1.79493248e+00
-5.16865373e-01 2.96975374e-01 1.08156279e-01 -1.22507024e+00
7.55062342e-01 2.34635666e-01 8.98807108e-01 -4.15540516e-01
4.93537128e-01 5.63626178e-03 -6.32466003e-02 -2.66064107e-01
9.92490677e-04 -1.17643632e-01 -6.44732267e-02 1.91066071e-01
2.57226154e-02 -3.64673883e-01 -9.55394581e-02 -2.14736983e-01
9.36980486e-01 5.39799966e-02 2.12553069e-01 -2.99709849e-02
5.31495929e-01 -4.61988561e-02 6.67314410e-01 4.44735944e-01
-1.37070015e-01 7.28382945e-01 5.14039695e-01 -5.82452476e-01
-1.01286364e+00 -9.50046659e-01 -1.32807240e-01 8.28441918e-01
-6.61587864e-02 2.79565245e-01 -1.00511956e+00 -1.17633128e+00
3.52739505e-02 3.82844537e-01 -1.05318201e+00 -1.51299134e-01
-5.96667588e-01 -1.26165223e+00 6.15143955e-01 8.87887895e-01
3.70900363e-01 -8.14852536e-01 -3.92330468e-01 -7.06671644e-03
1.68702066e-01 -1.35345280e+00 -7.64978111e-01 7.43267760e-02
-1.18760180e+00 -9.92218435e-01 -1.10068893e+00 -7.93910205e-01
1.40706205e+00 -1.97429270e-01 1.00385666e+00 1.33464903e-01
-6.99796975e-01 2.78273791e-01 1.75553598e-02 -7.03327358e-01
-6.15757287e-01 5.65650389e-02 -1.86195031e-01 -7.22324997e-02
-4.90305796e-02 -4.94732052e-01 -5.86558700e-01 3.93078923e-01
-1.06611490e+00 1.82867095e-01 8.48562717e-01 1.15858245e+00
9.89281058e-01 -3.23265374e-01 3.28403324e-01 -1.08647048e+00
6.05941117e-01 -1.72000993e-02 -5.01518726e-01 2.81992435e-01
-6.66154504e-01 7.30066672e-02 5.02224445e-01 -8.07191730e-01
-8.96756232e-01 4.04865444e-01 -1.69546515e-01 -7.34396696e-01
-1.62852928e-01 3.83362591e-01 1.27427680e-02 -4.84942853e-01
7.89703846e-01 5.00455238e-02 6.18981421e-01 -2.84312934e-01
3.34271312e-01 6.03965782e-02 7.73403406e-01 -4.39818084e-01
7.82268763e-01 5.45809925e-01 2.08530158e-01 -6.22924805e-01
-9.98209238e-01 -1.07700236e-01 -6.99145138e-01 -5.08775339e-02
9.36840713e-01 -7.57909954e-01 -4.24639374e-01 6.99312806e-01
-7.90859461e-01 -5.09653747e-01 -5.50567567e-01 4.08612996e-01
-3.94723415e-01 3.28211904e-01 -7.04695165e-01 -4.22472209e-01
-8.02419901e-01 -1.25640202e+00 8.98067236e-01 2.88986623e-01
3.38241793e-02 -1.31104529e+00 -6.45200834e-02 4.47712570e-01
4.74242926e-01 8.32454264e-01 9.81458962e-01 -7.16162205e-01
-2.00874165e-01 -5.80098152e-01 -1.69722810e-01 6.78577185e-01
4.90506262e-01 -2.49809757e-01 -6.95175767e-01 -3.76337707e-01
-1.63395047e-01 -4.28241760e-01 7.41161823e-01 5.89246392e-01
1.76663971e+00 -3.37031871e-01 -3.83152515e-01 9.51571882e-01
1.11530411e+00 3.85830738e-02 3.75796586e-01 8.33371505e-02
8.92085135e-01 3.57042372e-01 3.78655404e-01 1.88670605e-01
2.73311678e-02 5.51770687e-01 6.59239590e-01 -9.12748694e-01
-1.53390422e-01 6.99456185e-02 -1.95521817e-01 3.18397135e-01
-1.87244236e-01 1.07542619e-01 -8.41163993e-01 5.42390466e-01
-1.57197762e+00 -4.83428955e-01 1.27921686e-01 2.09001446e+00
1.23070431e+00 2.12869450e-01 1.66076481e-01 -5.67695610e-02
5.65675378e-01 -1.13360465e-01 -9.32095945e-01 -6.59186393e-02
1.94791436e-01 4.81814414e-01 6.23289704e-01 2.36804292e-01
-1.44112575e+00 8.37602317e-01 6.61313105e+00 6.39580786e-01
-1.24639142e+00 6.54520169e-02 1.11980855e+00 6.47155941e-02
-2.36262769e-01 -5.13482153e-01 -2.48205096e-01 2.63034791e-01
3.27750862e-01 3.40113163e-01 7.92687237e-02 9.59138691e-01
-1.69503823e-01 2.91161239e-01 -1.00554574e+00 7.34903812e-01
1.53072402e-02 -1.42912138e+00 -2.17144378e-02 4.83075045e-02
8.76947582e-01 -1.34374663e-01 4.12149847e-01 1.40453175e-01
3.61367941e-01 -1.08444583e+00 1.36181995e-01 2.72137642e-01
9.81773078e-01 -5.84316790e-01 6.91410899e-01 3.01170480e-02
-8.00866425e-01 1.57300726e-01 1.03892915e-01 4.28783894e-01
7.97629263e-03 4.65225905e-01 -1.00009215e+00 3.37854147e-01
3.72360200e-01 3.07676256e-01 -5.03270507e-01 7.83971667e-01
-1.60594836e-01 4.58171099e-01 -3.07873428e-01 7.65153468e-02
2.82485694e-01 -2.44993791e-01 3.38374943e-01 1.07621050e+00
-2.61280723e-02 1.06240086e-01 3.11039060e-01 6.90889597e-01
-1.76378056e-01 2.27327526e-01 -2.92525768e-01 -8.42121169e-02
-6.02093001e-04 1.40509915e+00 -9.11622047e-01 -1.63855195e-01
-3.40966403e-01 9.86264825e-01 -7.62927206e-03 1.50428340e-01
-1.06146765e+00 -1.09802991e-01 6.45537257e-01 2.54727185e-01
6.02846853e-02 -1.06145255e-01 -6.34330571e-01 -9.68616664e-01
-7.18518347e-02 -8.54044557e-01 4.88200277e-01 -4.06446666e-01
-1.27783561e+00 7.11175084e-01 -2.24728093e-01 -1.00323427e+00
-1.35141060e-01 -7.62054265e-01 -6.14802122e-01 7.57685125e-01
-1.37650478e+00 -1.58114672e+00 -6.01063848e-01 6.21539235e-01
3.56856495e-01 -3.04656655e-01 9.68088090e-01 4.94573802e-01
-6.92325354e-01 1.01762176e+00 -2.34495580e-01 4.76176351e-01
5.32763243e-01 -1.41963851e+00 4.36508745e-01 6.65016353e-01
-1.64334401e-01 1.94740787e-01 5.45121729e-01 -4.89012808e-01
-1.16793132e+00 -1.26199114e+00 2.65384233e-03 -2.48648122e-01
5.38684726e-01 -1.42950684e-01 -9.45720494e-01 6.44665778e-01
-3.22448872e-02 5.99136412e-01 8.62250090e-01 -1.12645130e-03
-1.12702109e-01 -1.44838005e-01 -1.41230059e+00 7.15252221e-01
7.49508321e-01 -1.22242160e-01 -1.46603793e-01 5.94819605e-01
6.53225064e-01 -9.66872573e-01 -1.04160881e+00 8.43025744e-01
2.49789536e-01 -3.47382039e-01 1.24872899e+00 -9.60312843e-01
2.94391900e-01 1.75749231e-03 2.58264810e-01 -1.23223674e+00
-2.13820599e-02 -5.82658350e-01 1.78206135e-02 8.36993515e-01
4.81730908e-01 -5.97876191e-01 1.12396669e+00 8.01871955e-01
-3.23282510e-01 -1.22184718e+00 -6.61787927e-01 -5.95013738e-01
1.73193961e-01 -2.23837018e-01 4.06764567e-01 1.14977217e+00
-5.03308773e-01 8.50808173e-02 -1.38849080e-01 2.86512822e-01
6.73945308e-01 -1.45362154e-01 8.53186786e-01 -1.09852219e+00
-3.12732339e-01 -6.38438284e-01 -4.20281351e-01 -6.81137323e-01
2.09125698e-01 -8.09451520e-01 -1.19802199e-01 -1.30247867e+00
4.71097305e-02 -7.83315480e-01 -4.40691769e-01 1.02230871e+00
-5.14882147e-01 5.43359995e-01 -2.34884664e-01 -9.89320315e-03
-1.52249306e-01 4.88526791e-01 1.71414137e+00 -4.28619832e-01
-4.07420605e-01 3.83687377e-01 -7.31192827e-01 8.24850559e-01
7.48668134e-01 -2.40733445e-01 -4.22247916e-01 -4.55019057e-01
-1.97853982e-01 -3.06673255e-02 4.74650413e-01 -8.51295650e-01
-1.42058119e-01 -1.33601815e-01 6.52014315e-01 -2.25045860e-01
3.56907159e-01 -8.19941401e-01 1.18557857e-02 6.66539550e-01
-4.26645368e-01 -4.43918817e-02 5.14373541e-01 3.86495173e-01
-1.42274976e-01 -4.02833186e-02 1.15343451e+00 -2.22134069e-01
-2.26546913e-01 7.86520898e-01 2.45716516e-02 1.09300591e-01
1.32794058e+00 -1.15086176e-01 4.94287275e-02 -1.76197678e-01
-9.75516975e-01 1.70447871e-01 4.37365323e-01 3.08473259e-01
4.86162722e-01 -1.27219033e+00 -6.35902882e-01 2.03053996e-01
-1.34570330e-01 4.53818709e-01 3.13682914e-01 1.02093112e+00
-5.31885624e-01 -8.92719440e-03 -3.20078433e-01 -7.17968702e-01
-1.39718449e+00 5.36119580e-01 6.01783395e-01 -7.08605409e-01
-6.23888314e-01 8.98816586e-01 3.13459277e-01 -5.53130150e-01
3.52905720e-01 -4.40045476e-01 -1.58613399e-02 -3.75056624e-01
2.17591375e-01 -5.22433128e-03 1.23836249e-01 -1.98370829e-01
-2.09821165e-01 4.94078040e-01 -3.66369605e-01 2.41878614e-01
1.27190399e+00 3.89361024e-01 1.66925803e-01 -1.49042085e-01
9.94991541e-01 -2.69506216e-01 -1.51240551e+00 -3.18581283e-01
-4.18701738e-01 -3.15593123e-01 2.95377284e-01 -9.67607319e-01
-1.57295597e+00 7.42299914e-01 7.04809368e-01 1.27128646e-01
1.31397986e+00 -1.02001518e-01 8.47691178e-01 9.04754624e-02
-7.08522052e-02 -8.88499022e-01 3.27033222e-01 1.07105389e-01
7.69690692e-01 -1.43767071e+00 1.85517505e-01 -6.60822034e-01
-8.65530074e-01 7.66280890e-01 9.55689132e-01 -4.00568098e-02
4.95729178e-01 4.48782325e-01 3.38481843e-01 3.88128161e-02
-2.08125666e-01 5.39306588e-02 7.06336856e-01 4.87473786e-01
3.60519826e-01 -3.87212932e-02 -1.45850927e-01 3.70881826e-01
2.19210431e-01 -2.14288250e-01 5.35667539e-02 8.82328928e-01
5.67948855e-02 -1.26646793e+00 -1.59993678e-01 4.34376448e-01
-8.35724056e-01 1.07051596e-01 -1.89177126e-01 1.18210340e+00
-7.02181756e-02 1.98403791e-01 1.54410109e-01 -7.44324476e-02
3.33580136e-01 -1.58197209e-01 7.68625379e-01 -5.62859178e-01
-5.71881354e-01 1.43041536e-01 -2.33534381e-01 -3.94764960e-01
-4.20367807e-01 -6.26593590e-01 -1.33401012e+00 1.55027181e-01
-3.60692322e-01 -1.52366862e-01 8.26978207e-01 9.30812716e-01
2.80224591e-01 9.34761584e-01 6.29643381e-01 -7.09464669e-01
-7.38754272e-01 -8.79725397e-01 -9.96288285e-02 6.75718188e-01
2.24542767e-01 -5.18991530e-01 4.83829156e-02 1.20453082e-01]
|
[14.494277000427246, -2.104868173599243]
|
8d48e482-675c-4e04-bac5-cd9d80589ea2
|
vibertgrid-a-jointly-trained-multi-modal-2d
|
2105.11672
| null |
https://arxiv.org/abs/2105.11672v1
|
https://arxiv.org/pdf/2105.11672v1.pdf
|
ViBERTgrid: A Jointly Trained Multi-Modal 2D Document Representation for Key Information Extraction from Documents
|
Recent grid-based document representations like BERTgrid allow the simultaneous encoding of the textual and layout information of a document in a 2D feature map so that state-of-the-art image segmentation and/or object detection models can be straightforwardly leveraged to extract key information from documents. However, such methods have not achieved comparable performance to state-of-the-art sequence- and graph-based methods such as LayoutLM and PICK yet. In this paper, we propose a new multi-modal backbone network by concatenating a BERTgrid to an intermediate layer of a CNN model, where the input of CNN is a document image and the BERTgrid is a grid of word embeddings, to generate a more powerful grid-based document representation, named ViBERTgrid. Unlike BERTgrid, the parameters of BERT and CNN in our multimodal backbone network are trained jointly. Our experimental results demonstrate that this joint training strategy improves significantly the representation ability of ViBERTgrid. Consequently, our ViBERTgrid-based key information extraction approach has achieved state-of-the-art performance on real-world datasets.
|
['Qiang Huo', 'Qin Ren', 'Kai Hu', 'Zhuoyao Zhong', 'Lei Sun', 'Qifang Gao', 'WeiHong Lin']
|
2021-05-25
| null | null | null | null |
['key-information-extraction']
|
['natural-language-processing']
|
[ 8.24154448e-03 1.08171873e-01 -1.79003075e-01 -1.68797538e-01
-8.71969342e-01 -8.36967945e-01 9.22205031e-01 4.13954526e-01
-3.18278730e-01 1.63294062e-01 1.35651350e-01 -2.69751191e-01
5.00412937e-03 -1.06976461e+00 -6.55561090e-01 -5.42970657e-01
6.80139363e-02 7.92786717e-01 3.18356961e-01 -5.49828000e-02
3.94632310e-01 5.44638574e-01 -1.24110591e+00 3.39647591e-01
7.25826323e-01 1.08781421e+00 4.28268939e-01 9.97079313e-01
-6.87911570e-01 4.62739229e-01 -6.61058664e-01 -2.84470767e-01
-1.20203625e-02 -2.16864079e-01 -8.16030800e-01 3.43143702e-01
7.40449607e-01 -4.05261755e-01 -4.73080844e-01 9.56163108e-01
4.76139605e-01 1.05449162e-01 6.48398578e-01 -1.17252469e+00
-8.32321167e-01 7.43584156e-01 -7.70256817e-01 -4.04545963e-01
1.31022349e-01 -1.17562301e-01 1.22135580e+00 -8.20969045e-01
8.77077818e-01 1.39793050e+00 3.65603209e-01 1.23916626e-01
-1.07194126e+00 -3.25949579e-01 5.68336546e-01 6.85165599e-02
-1.48758674e+00 7.11017624e-02 9.07430410e-01 -2.38991782e-01
9.81058478e-01 1.64761961e-01 8.95570576e-01 5.97334921e-01
3.45947370e-02 1.31680548e+00 6.50923610e-01 -5.30068517e-01
6.39984161e-02 -1.30245522e-01 1.17508285e-01 1.23823476e+00
1.47593215e-01 -5.99268675e-01 -4.95517135e-01 -1.66347828e-02
1.02163339e+00 -2.04044953e-02 -1.24533281e-01 -5.54410458e-01
-1.14410913e+00 8.74454856e-01 7.70785332e-01 3.61741364e-01
-2.03335896e-01 5.39192915e-01 3.77184421e-01 -2.48384431e-01
4.35171306e-01 4.25277084e-01 -9.38930660e-02 -1.07450135e-01
-1.36003375e+00 4.50717062e-01 7.31175900e-01 1.15379763e+00
8.25426280e-01 -2.48071596e-01 -4.31564242e-01 7.44111478e-01
3.99189293e-01 3.05963397e-01 2.12973341e-01 -6.36810005e-01
5.80014050e-01 1.05965257e+00 -1.49221957e-01 -1.05869424e+00
-3.81349713e-01 -4.38786209e-01 -7.06651986e-01 -1.04715593e-01
2.40234897e-01 2.00732708e-01 -1.48911619e+00 1.09051752e+00
1.22046493e-01 -1.46047369e-01 -2.42366895e-01 7.71217525e-01
8.89250755e-01 8.10331345e-01 -2.53883183e-01 5.05270422e-01
1.47206354e+00 -1.35425234e+00 -6.17019594e-01 -2.51294434e-01
5.53122163e-01 -8.21116686e-01 1.04084074e+00 5.39216399e-01
-1.12007582e+00 -5.57884693e-01 -1.09209955e+00 -3.84038508e-01
-7.93005168e-01 4.59894806e-01 7.53956616e-01 3.28788787e-01
-1.15642762e+00 1.87163249e-01 -6.97728336e-01 -6.13820672e-01
5.64452171e-01 2.33564600e-01 -3.69707525e-01 -4.89128798e-01
-5.55773616e-01 5.10365963e-01 8.10715675e-01 8.56535956e-02
-9.44996476e-01 -1.65233806e-01 -9.09167647e-01 2.77946502e-01
5.58223903e-01 -4.16527182e-01 1.18509960e+00 -6.67804629e-02
-1.12740755e+00 7.74823487e-01 -1.58085689e-01 -8.26821774e-02
2.90298313e-01 -2.79201806e-01 -3.51865217e-02 5.75928330e-01
6.81435019e-02 1.03390634e+00 6.49696350e-01 -1.54298556e+00
-6.29726946e-01 -2.37346172e-01 1.59779310e-01 4.16393876e-02
-5.19550443e-01 -8.56993496e-02 -1.51787698e+00 -5.08397162e-01
4.91284095e-02 -6.99226737e-01 -1.01377763e-01 1.58199623e-01
-1.01356697e+00 -2.82595724e-01 1.17921245e+00 -5.16276538e-01
1.23045278e+00 -1.94259906e+00 1.83151454e-01 3.67546231e-01
2.68296599e-01 2.73966700e-01 -5.82068384e-01 9.34378803e-01
4.34968591e-01 3.39038193e-01 -1.30729064e-01 -7.48701096e-01
2.76799858e-01 2.59074569e-01 -1.72732726e-01 3.46493065e-01
1.13408990e-01 1.33350146e+00 -8.32207501e-01 -7.44014561e-01
3.22492510e-01 6.07930839e-01 -4.56364304e-01 2.93417275e-01
-7.55107701e-01 -9.32446420e-02 -4.52243537e-01 8.04445446e-01
7.92850912e-01 -5.22333980e-01 3.32810104e-01 -2.98910797e-01
-2.88788211e-02 8.93341079e-02 -9.71817970e-01 2.10180545e+00
-2.62449861e-01 7.71768987e-01 1.75244078e-01 -7.97716856e-01
1.01134598e+00 3.69058400e-02 1.63906664e-01 -6.42913520e-01
2.02898353e-01 -8.86003822e-02 -4.41695541e-01 -1.78682625e-01
1.06703258e+00 6.57554984e-01 -2.83446640e-01 5.19194245e-01
4.15616274e-01 -5.44718206e-01 5.60971141e-01 7.35232055e-01
8.62994909e-01 1.73030376e-01 -5.86153679e-02 -3.10838073e-02
2.80319512e-01 9.90271419e-02 1.18532158e-01 1.08989203e+00
4.11594450e-01 8.86599660e-01 7.08003819e-01 -1.07120685e-01
-1.07925928e+00 -8.47747505e-01 3.52925748e-01 9.65150177e-01
1.49997950e-01 -9.47729826e-01 -1.04928732e+00 -7.57129848e-01
1.44629508e-01 3.68772209e-01 -6.16677940e-01 2.30223298e-01
-6.09401107e-01 -6.10268652e-01 7.28146434e-01 8.29644859e-01
7.78610766e-01 -8.10516357e-01 -2.42094398e-01 2.64929742e-01
-8.71979594e-02 -1.28631210e+00 -6.61443233e-01 2.77040064e-01
-5.18023252e-01 -1.11150074e+00 -6.71866179e-01 -9.52564001e-01
9.70461607e-01 3.34636688e-01 1.07962072e+00 3.89427245e-01
-4.62287515e-01 3.90069544e-01 -4.84933257e-01 -2.80036747e-01
7.50394911e-02 4.73204851e-01 -7.50157416e-01 -8.73229057e-02
3.05342413e-02 -1.67784274e-01 -7.03879356e-01 3.43860686e-02
-1.15443575e+00 2.36566797e-01 5.77718914e-01 7.56046891e-01
6.52833104e-01 1.73503056e-01 6.27567247e-02 -8.80988598e-01
8.87848973e-01 1.61621626e-02 -7.04467177e-01 3.70233744e-01
-5.17837107e-01 -6.27445802e-02 4.28785235e-01 -3.16230297e-01
-7.57499278e-01 -9.04078111e-02 -1.58560574e-01 -1.43471584e-01
-1.68827161e-01 9.05374229e-01 -3.28310877e-01 -1.25484675e-01
9.89076775e-03 2.96848387e-01 -3.73251021e-01 -7.20049739e-01
9.80192423e-01 7.97539353e-01 9.11593199e-01 -6.35188699e-01
8.80062759e-01 6.08241618e-01 -1.65655628e-01 -6.38143182e-01
-8.33910763e-01 -8.27451289e-01 -9.66540217e-01 -1.63335875e-02
1.05757332e+00 -7.58835614e-01 -6.80912435e-01 7.69926071e-01
-1.37522018e+00 -4.88775373e-01 -5.78386746e-02 -5.62044941e-02
-4.21136677e-01 2.43920058e-01 -6.83053851e-01 -6.91489160e-01
-2.86264300e-01 -1.10719633e+00 1.69155991e+00 3.68559718e-01
1.51517302e-01 -1.09513259e+00 -2.63797134e-01 3.41526747e-01
2.09542900e-01 2.44526625e-01 1.13849163e+00 -4.75506723e-01
-9.28730607e-01 -4.81882662e-01 -7.47759819e-01 8.09201151e-02
2.39602942e-03 1.23328678e-01 -8.51519346e-01 -3.41078907e-01
-7.28382409e-01 -3.15993339e-01 1.32071531e+00 2.40784377e-01
1.29820704e+00 -2.98194498e-01 -6.27805769e-01 7.91532278e-01
1.73333120e+00 1.25913441e-01 6.97142780e-01 5.86708665e-01
1.19777536e+00 1.52112782e-01 4.12321270e-01 3.24159712e-01
5.74868262e-01 5.45387387e-01 6.59966052e-01 -4.73838985e-01
-2.33819634e-01 -4.30375487e-01 -3.39030214e-02 7.51740277e-01
2.17433751e-01 -9.73018885e-01 -1.06187475e+00 7.40422070e-01
-2.19237566e+00 -5.40761590e-01 4.84645776e-02 1.65754223e+00
6.49174690e-01 2.17576236e-01 -1.00430474e-01 -1.11114256e-01
4.38261688e-01 5.05910099e-01 -2.61577994e-01 -4.48591620e-01
-1.24671303e-01 -3.96523364e-02 6.37693405e-01 2.41430163e-01
-1.33609128e+00 1.34414458e+00 6.43113232e+00 9.18180168e-01
-1.06506240e+00 -7.30940104e-02 4.36854124e-01 4.44966182e-02
-3.98252964e-01 -1.11399435e-01 -1.00033760e+00 1.02471478e-01
5.75872540e-01 2.26730928e-01 2.70994484e-01 6.44357681e-01
-2.64636334e-02 -2.09097028e-01 -9.34359193e-01 8.73470962e-01
4.81758267e-01 -1.69829679e+00 3.69689524e-01 3.00277293e-01
6.69433653e-01 1.26268389e-02 -8.32066759e-02 8.52165371e-02
6.46656632e-01 -1.18409991e+00 9.30332363e-01 3.82124335e-01
9.02588010e-01 -9.51248050e-01 5.93361676e-01 3.03260177e-01
-1.42859602e+00 1.27117977e-01 -3.02938789e-01 4.15449470e-01
2.40381598e-01 5.58155477e-01 -9.98070776e-01 8.07042003e-01
4.90391314e-01 6.21072590e-01 -8.76815796e-01 1.29990530e+00
-4.77185637e-01 5.76790333e-01 -1.46504208e-01 -2.04553962e-01
9.01632249e-01 -5.54799251e-02 1.87475801e-01 1.59636211e+00
1.38401717e-01 -3.63866597e-01 6.29926860e-01 9.75566268e-01
-4.00344372e-01 -9.29864347e-02 -4.07206982e-01 -5.65507174e-01
3.18357974e-01 1.64637244e+00 -1.15596700e+00 -5.10996401e-01
-5.12150168e-01 1.10437393e+00 4.40752089e-01 5.30115545e-01
-4.36838269e-01 -7.62371719e-01 3.76268119e-01 -2.24627797e-02
9.31756377e-01 -7.82703280e-01 -1.16691962e-01 -1.02528834e+00
-1.48742706e-01 -6.61476076e-01 2.14313835e-01 -9.69733119e-01
-9.44117725e-01 5.21601915e-01 1.07052261e-02 -6.10961497e-01
-1.70857742e-01 -7.77023196e-01 -5.70309877e-01 8.82361054e-01
-1.60667264e+00 -1.89423156e+00 -5.04514933e-01 5.17770112e-01
4.48060334e-01 -4.77623641e-02 9.96868372e-01 -1.45542338e-01
-5.11132479e-01 4.35590535e-01 3.09057325e-01 7.43587434e-01
4.03455913e-01 -1.65202951e+00 8.01031828e-01 7.79920816e-01
4.02103931e-01 6.04941249e-01 6.80061951e-02 -7.40931511e-01
-1.65696239e+00 -1.34407794e+00 5.03571987e-01 -1.00895144e-01
5.87485373e-01 -1.05166376e+00 -7.20960438e-01 6.61679089e-01
5.85130632e-01 -2.30480749e-02 3.89510751e-01 1.64940923e-01
-5.36485851e-01 1.74594432e-01 -8.33816648e-01 6.89713955e-01
8.94535244e-01 -7.60096967e-01 -2.35026643e-01 4.02839422e-01
7.59748399e-01 -5.26467562e-01 -6.83227658e-01 -4.56279889e-02
4.73089129e-01 -4.78238910e-01 9.42993343e-01 -4.00980026e-01
2.43857324e-01 -3.03588122e-01 -1.03830852e-01 -1.33431411e+00
-3.73768479e-01 -5.32129347e-01 -3.23444128e-01 1.43890703e+00
1.66632235e-01 -2.97537237e-01 8.53422403e-01 -1.37205040e-02
-2.23969251e-01 -8.41276288e-01 -6.09010577e-01 -6.72382534e-01
2.29659416e-02 -3.79865438e-01 7.70497799e-01 5.45651495e-01
-1.83032647e-01 1.70338333e-01 -6.52501732e-03 8.78486410e-02
5.37377417e-01 2.92915761e-01 8.87761295e-01 -1.24321830e+00
-3.02802194e-02 -4.99828696e-01 -4.75248456e-01 -1.41991436e+00
1.36573184e-02 -8.89567137e-01 2.31329858e-01 -2.52066565e+00
3.93322408e-01 -1.59365281e-01 -2.18518838e-01 7.00902581e-01
-6.01441450e-02 4.66656268e-01 4.67746705e-01 4.29572612e-02
-8.73573005e-01 4.30973202e-01 1.55690873e+00 -7.77553856e-01
-1.08930960e-01 -7.63937831e-01 -8.43367457e-01 6.29600465e-01
5.28321743e-01 -1.51953176e-01 -4.46220547e-01 -6.91470265e-01
1.98925585e-01 -2.19901904e-01 4.57922459e-01 -7.58547544e-01
5.18592954e-01 6.80812746e-02 6.24932587e-01 -1.16014385e+00
3.53756130e-01 -5.00373304e-01 -5.31362712e-01 -1.57406911e-01
-2.10194290e-01 -6.44253343e-02 4.51103181e-01 4.40009594e-01
-3.63222659e-01 -2.80487359e-01 1.65634945e-01 -3.60136807e-01
-8.43911111e-01 3.80839884e-01 -3.62153649e-01 -1.54150978e-01
5.93678057e-01 -2.20156014e-01 -7.60263801e-01 -3.25377524e-01
-3.93387496e-01 3.36588949e-01 5.76043010e-01 5.07061958e-01
8.07031691e-01 -1.14965653e+00 -5.44466197e-01 1.37613177e-01
1.98709622e-01 3.77314329e-01 -1.12643681e-01 5.04406810e-01
-5.32829821e-01 7.88513482e-01 9.93710756e-02 -6.65144801e-01
-1.24476469e+00 3.30424964e-01 8.66195932e-02 -5.85966945e-01
-7.38361239e-01 9.34671462e-01 2.84623742e-01 -5.42519987e-01
2.73459554e-01 -5.03981352e-01 -4.96694557e-02 2.62146920e-01
4.37387288e-01 -1.20710351e-01 2.02338621e-01 -4.40475732e-01
-3.53875548e-01 5.44412494e-01 -5.31811953e-01 -3.92581850e-01
1.39768779e+00 6.17706552e-02 -2.05477193e-01 9.78554040e-02
1.24530768e+00 -1.14750460e-01 -1.17114580e+00 -1.44460633e-01
-9.21670794e-02 -4.22507286e-01 3.80109310e-01 -1.00536585e+00
-1.08862293e+00 9.18182552e-01 1.93743020e-01 1.93048611e-01
1.06140471e+00 2.17379957e-01 8.59801829e-01 5.01112342e-01
3.32796156e-01 -1.31487787e+00 3.44277054e-01 6.50032103e-01
8.34330380e-01 -9.34664607e-01 2.78421938e-02 -4.85958010e-01
-2.64570594e-01 1.23931706e+00 4.31746274e-01 -8.06859285e-02
3.94591749e-01 4.25914973e-01 3.91191952e-02 -5.40355265e-01
-5.57074130e-01 -4.34782565e-01 5.78453481e-01 6.71822906e-01
3.32538784e-01 -1.91776697e-02 6.98896497e-02 3.28722507e-01
-7.71694705e-02 -1.98300645e-01 4.54090506e-01 1.32950342e+00
-5.07905364e-01 -1.15043330e+00 -3.49808246e-01 5.04628718e-01
-1.54562369e-01 -2.43170112e-01 -6.34416759e-01 9.66306031e-01
-4.62766886e-02 8.14617932e-01 3.69681418e-01 -2.05670908e-01
2.77791381e-01 6.67339424e-03 5.16856551e-01 -8.49032700e-01
-5.17726243e-01 4.30131108e-01 7.73676634e-02 -7.03860462e-01
-1.87689677e-01 -3.33855093e-01 -1.47493827e+00 -5.74806333e-02
-4.58186299e-01 9.75590199e-02 9.08903360e-01 9.13028181e-01
3.93414050e-01 8.24363589e-01 4.30188291e-02 -1.09169531e+00
7.23806620e-02 -9.26662743e-01 -7.53604352e-01 8.73563290e-02
2.48850986e-01 -4.67441797e-01 2.32760474e-01 6.43036291e-02]
|
[11.61681079864502, 2.4540212154388428]
|
6a5b752a-ef99-4970-9309-68c8bddec54f
|
blinkflow-a-dataset-to-push-the-limits-of
|
2303.07716
| null |
https://arxiv.org/abs/2303.07716v1
|
https://arxiv.org/pdf/2303.07716v1.pdf
|
BlinkFlow: A Dataset to Push the Limits of Event-based Optical Flow Estimation
|
Event cameras provide high temporal precision, low data rates, and high dynamic range visual perception, which are well-suited for optical flow estimation. While data-driven optical flow estimation has obtained great success in RGB cameras, its generalization performance is seriously hindered in event cameras mainly due to the limited and biased training data. In this paper, we present a novel simulator, BlinkSim, for the fast generation of large-scale data for event-based optical flow. BlinkSim consists of a configurable rendering engine and a flexible engine for event data simulation. By leveraging the wealth of current 3D assets, the rendering engine enables us to automatically build up thousands of scenes with different objects, textures, and motion patterns and render very high-frequency images for realistic event data simulation. Based on BlinkSim, we construct a large training dataset and evaluation benchmark BlinkFlow that contains sufficient, diversiform, and challenging event data with optical flow ground truth. Experiments show that BlinkFlow improves the generalization performance of state-of-the-art methods by more than 40% on average and up to 90%. Moreover, we further propose an Event optical Flow transFormer (E-FlowFormer) architecture. Powered by our BlinkFlow, E-FlowFormer outperforms the SOTA methods by up to 91% on MVSEC dataset and 14% on DSEC dataset and presents the best generalization performance.
|
['Guofeng Zhang', 'Zhaopeng Cui', 'Hujun Bao', 'Hongsheng Li', 'Xiaoyu Shi', 'Shuo Chen', 'Zhaoyang Huang', 'Yijin Li']
|
2023-03-14
| null | null | null | null |
['event-based-optical-flow']
|
['computer-vision']
|
[-2.95069814e-01 -7.75935650e-01 1.11523516e-01 -1.10724971e-01
-1.35975331e-01 -5.16308129e-01 5.57203233e-01 -1.15853332e-01
-3.94655406e-01 5.65705419e-01 2.62511194e-01 -4.54758108e-02
-3.67818307e-03 -7.83041120e-01 -4.60857421e-01 -4.53859627e-01
-1.90034062e-01 -4.82828403e-03 5.97028911e-01 -3.93238887e-02
1.93830118e-01 7.47023284e-01 -2.06484747e+00 2.35556901e-01
8.98871899e-01 1.25313556e+00 2.44830891e-01 1.07044876e+00
3.11190281e-02 1.16564441e+00 -7.32627690e-01 -2.96032906e-01
3.61272603e-01 -3.90831083e-01 -4.34813678e-01 -1.15915880e-01
1.00315583e+00 -9.51266229e-01 -5.49718142e-01 4.95330542e-01
7.62773931e-01 3.27844054e-01 9.33064241e-03 -1.54899383e+00
-1.88728362e-01 -9.45329070e-02 -1.72652274e-01 7.22609639e-01
6.94232702e-01 8.11850667e-01 5.89287817e-01 -7.73023903e-01
9.38510954e-01 1.14781487e+00 4.10401702e-01 5.69453478e-01
-1.06123173e+00 -7.80448854e-01 -3.30607295e-02 6.28112018e-01
-9.50373411e-01 -4.88507330e-01 4.80738044e-01 -5.22032917e-01
1.18548429e+00 2.92951882e-01 1.21750784e+00 1.44016838e+00
3.15720618e-01 6.20504558e-01 8.41608346e-01 2.38833711e-01
3.67011786e-01 -1.79117888e-01 -2.89495349e-01 5.97527921e-01
6.15711184e-03 5.61315060e-01 -1.06045580e+00 1.36645779e-01
1.04185259e+00 -1.67090371e-01 -7.62795985e-01 -2.32604519e-01
-1.51243556e+00 3.53920728e-01 4.62326258e-01 -1.77632302e-01
-2.87915140e-01 2.72218704e-01 6.23759747e-01 1.45423263e-01
1.88005835e-01 3.72649282e-01 -1.84322193e-01 -7.36267567e-01
-8.28802466e-01 4.36255693e-01 8.10489416e-01 9.45205569e-01
6.80577457e-01 3.26390922e-01 -3.11742276e-01 2.50134349e-01
-5.64254858e-02 7.29955852e-01 5.06788433e-01 -1.27659106e+00
3.94052178e-01 4.37142432e-01 1.91859514e-01 -1.10481620e+00
-4.50202823e-01 -3.95791531e-01 -7.43244350e-01 5.00284851e-01
4.24289852e-01 -3.76139916e-02 -5.11193097e-01 1.51334238e+00
4.44610655e-01 8.33996117e-01 -1.06708094e-01 1.31261301e+00
7.71073878e-01 7.21313715e-01 9.15317759e-02 -2.97343820e-01
1.00335956e+00 -7.07165956e-01 -5.36401749e-01 2.59142723e-02
2.84677535e-01 -7.01851606e-01 1.23865819e+00 7.23255873e-01
-1.07159829e+00 -8.20771754e-01 -9.64783132e-01 -1.54585496e-01
-1.13148130e-02 -2.01012909e-01 8.95865381e-01 5.72396338e-01
-9.31858063e-01 6.16556227e-01 -1.01535046e+00 -2.90735841e-01
5.85607529e-01 -8.90291408e-02 -4.67176348e-01 -1.35897681e-01
-8.74077976e-01 5.16153932e-01 2.50811696e-01 -7.40116686e-02
-1.10689688e+00 -1.30769897e+00 -8.33577096e-01 -4.32144590e-02
1.65790930e-01 -9.04458106e-01 9.94930387e-01 -5.53665936e-01
-1.69522929e+00 4.22484815e-01 -1.12774007e-01 -4.80419397e-01
8.23959768e-01 -3.63446563e-01 -5.31310976e-01 5.59368491e-01
-3.37333120e-02 7.19685793e-01 8.18742275e-01 -8.09625328e-01
-7.74693251e-01 7.72590637e-02 1.84875846e-01 5.45321405e-02
-3.20601285e-01 -9.49283764e-02 -3.89476806e-01 -4.97581422e-01
-4.13711995e-01 -6.95785522e-01 7.68908113e-02 4.24682796e-01
3.50487456e-02 1.43762439e-01 1.00537348e+00 -2.46636406e-01
1.23833501e+00 -2.28281999e+00 -7.77902454e-02 -2.60191798e-01
3.52489054e-01 5.60696840e-01 -6.40985882e-03 2.54543006e-01
2.30812337e-02 -4.01856989e-01 1.95845902e-01 -4.03990686e-01
-2.04815552e-01 2.89216399e-01 -5.78090012e-01 3.46921772e-01
2.85417080e-01 6.70779288e-01 -1.21959531e+00 -5.00052929e-01
9.14072752e-01 6.90608382e-01 -1.08123064e+00 4.62523490e-01
-3.59044299e-02 7.18337297e-01 -1.44722447e-01 4.49768245e-01
7.00910747e-01 -3.95822734e-01 -2.26763278e-01 -4.62737828e-01
-2.38126114e-01 1.80402502e-01 -1.56015301e+00 2.10331988e+00
-5.47230303e-01 1.09488308e+00 -4.78246152e-01 -2.49517396e-01
7.66063452e-01 1.16141193e-01 6.54084027e-01 -8.74168277e-01
9.16575566e-02 -2.80723581e-03 -2.43279800e-01 -6.73565328e-01
6.69558823e-01 1.91432819e-01 4.09643710e-01 3.18081349e-01
2.90792167e-01 -7.29897097e-02 6.21442974e-01 2.77412981e-01
1.26076400e+00 4.79716301e-01 4.25740369e-02 1.60252899e-02
5.28013706e-01 -2.01306745e-01 7.51451731e-01 5.94209969e-01
-4.81411755e-01 7.51651287e-01 3.14896941e-01 -7.85620093e-01
-7.91993380e-01 -1.46117055e+00 -2.92554647e-01 7.15776682e-01
4.32063431e-01 -7.77560771e-01 -5.31361997e-01 -3.33132684e-01
-1.03503503e-01 4.85684991e-01 -3.60192537e-01 -1.13106310e-01
-6.47596776e-01 -6.14929378e-01 3.51863116e-01 5.24115562e-01
9.58655596e-01 -1.00625563e+00 -1.25209272e+00 3.39337140e-01
-2.44985670e-01 -1.65196347e+00 -3.94149512e-01 -5.59716940e-01
-6.86779618e-01 -1.33067679e+00 -2.67921984e-01 -1.10992812e-01
1.80274263e-01 2.29188204e-01 1.31021357e+00 -1.69393852e-01
-7.29737639e-01 3.25789124e-01 -3.61957163e-01 -2.59829640e-01
-2.60617375e-01 -2.75997877e-01 -6.91829994e-02 2.95604825e-01
-1.04350276e-01 -7.83675194e-01 -1.22991133e+00 4.08823907e-01
-9.81988788e-01 1.48871377e-01 4.80295457e-02 4.49058264e-01
3.27556312e-01 -2.96624303e-01 2.54693031e-01 -2.66511768e-01
1.97855040e-01 -2.49109343e-01 -8.88815284e-01 -2.15708122e-01
-4.67094421e-01 -2.34179780e-01 8.81793857e-01 -5.98225236e-01
-1.19208896e+00 -2.18766883e-01 6.57198206e-02 -7.85159409e-01
-1.21061094e-01 -1.76035017e-01 9.30745900e-02 -1.52014554e-01
1.04890621e+00 1.09061353e-01 -1.72672510e-01 -1.05224945e-01
2.63978958e-01 3.59931648e-01 8.46726954e-01 -4.49742466e-01
5.12282014e-01 9.56459224e-01 2.43079334e-01 -7.23569930e-01
-5.38094282e-01 -3.39054853e-01 -2.60223329e-01 -7.72494137e-01
7.83802867e-01 -1.04946280e+00 -1.22657287e+00 8.26881051e-01
-1.06397521e+00 -4.65688169e-01 -5.09228647e-01 8.24058115e-01
-6.17439628e-01 2.29919136e-01 -5.67938864e-01 -4.80104655e-01
-2.33199805e-01 -1.28428853e+00 1.02534580e+00 3.61137688e-01
-1.33642331e-02 -8.60582948e-01 1.25528812e-01 6.72167540e-02
6.10978186e-01 5.72112083e-01 1.44683167e-01 1.47332966e-01
-1.13827920e+00 8.44679251e-02 -3.80094886e-01 3.74908775e-01
1.30011626e-02 3.33280563e-01 -1.17307794e+00 -3.55494082e-01
-2.05822945e-01 -3.49440426e-01 5.80980897e-01 2.46626511e-01
1.17340767e+00 1.23131856e-01 -1.25886053e-01 1.28158784e+00
1.49922252e+00 -1.23138465e-02 8.42058122e-01 1.12563007e-01
7.04340339e-01 2.43309975e-01 5.26924372e-01 9.74577665e-01
4.58469957e-01 6.31767213e-01 7.45873988e-01 2.15223238e-01
-5.27159691e-01 -1.82587892e-01 3.56688529e-01 5.33365428e-01
-3.45167488e-01 -4.10446167e-01 -6.94450438e-01 4.57009822e-01
-1.58806622e+00 -1.26673663e+00 -2.22952351e-01 2.25262618e+00
6.02686644e-01 2.33914524e-01 1.01763077e-01 1.76988333e-01
3.82404506e-01 2.98273832e-01 -5.69491506e-01 -9.56164747e-02
-1.37263685e-01 3.66761237e-01 2.15375498e-01 3.38379890e-01
-7.97665775e-01 7.61868596e-01 5.79256201e+00 4.98955846e-01
-1.55172706e+00 -1.05309211e-01 1.97836667e-01 -5.81990361e-01
7.49693289e-02 -4.45636250e-02 -8.22281837e-01 7.73596168e-01
9.70485628e-01 -2.05479071e-01 5.06091893e-01 5.58335245e-01
5.48406720e-01 -3.07895660e-01 -1.06573880e+00 1.35780847e+00
6.56529069e-02 -1.72535217e+00 1.63742498e-01 8.94308556e-03
5.95581174e-01 1.97660267e-01 -2.36871332e-01 -2.64526214e-02
1.05346486e-01 -7.99181223e-01 7.74188101e-01 5.38600564e-01
9.42963064e-01 -5.12131810e-01 3.18558276e-01 1.41761243e-01
-1.29507852e+00 -1.03923567e-01 -2.69217253e-01 -2.14306697e-01
5.88837028e-01 7.10067511e-01 -4.02974576e-01 5.71058333e-01
1.02736080e+00 1.27678788e+00 -5.00447512e-01 1.44752455e+00
-2.23050535e-01 5.40800512e-01 -4.16008353e-01 1.73708543e-01
-1.11271376e-02 2.83845663e-02 6.51820719e-01 1.23175967e+00
2.93603331e-01 -1.29639432e-01 7.30128065e-02 7.08490491e-01
1.71784848e-01 -1.39171198e-01 -2.97741443e-01 4.65645760e-01
4.45434690e-01 1.27427661e+00 -2.67091662e-01 -3.92513096e-01
-4.25616831e-01 1.07664931e+00 2.17157468e-01 3.72515947e-01
-1.09327531e+00 -3.11686605e-01 1.29402685e+00 1.17838539e-01
2.03262553e-01 -4.52893585e-01 2.32399747e-01 -1.49999416e+00
1.76550359e-01 -6.01832390e-01 3.26645136e-01 -1.14709187e+00
-1.02439809e+00 8.04627836e-01 6.40556589e-03 -1.65514493e+00
-3.98196161e-01 -7.05964863e-01 -4.32071328e-01 7.10542560e-01
-1.80683601e+00 -7.60144770e-01 -1.24284422e+00 9.92052138e-01
3.24053884e-01 -6.38476536e-02 6.28201723e-01 5.71316957e-01
-5.87217093e-01 2.07655326e-01 -2.50729054e-01 1.50606468e-01
8.57725501e-01 -1.07274950e+00 4.72027719e-01 9.94353056e-01
1.43455446e-01 2.12573543e-01 5.74846089e-01 -2.86509454e-01
-1.51140881e+00 -1.18934548e+00 3.48436028e-01 -4.74987715e-01
6.52004600e-01 -3.85664761e-01 -7.78681815e-01 4.62262154e-01
1.49701694e-02 8.96684527e-01 3.08054835e-01 -5.64180315e-01
-3.88598174e-01 -5.68894446e-01 -1.03815353e+00 6.06184542e-01
1.39536989e+00 -4.37145323e-01 -1.35458037e-01 1.77898914e-01
4.94804919e-01 -8.40478361e-01 -1.01753247e+00 2.73811787e-01
5.30452967e-01 -1.71264517e+00 1.10765111e+00 -3.50836754e-01
4.63524312e-01 -3.83659095e-01 8.20474327e-02 -1.22909653e+00
-9.43915471e-02 -9.97808158e-01 -4.22353685e-01 9.86573339e-01
-2.68468231e-01 -7.09486723e-01 7.15394318e-01 2.11808309e-01
-7.43935257e-02 -4.04728442e-01 -9.08772349e-01 -8.28100085e-01
-5.87907016e-01 -9.00129497e-01 6.57082736e-01 6.95179045e-01
-3.35948646e-01 9.03282687e-02 -1.63541421e-01 1.87777966e-01
7.77125597e-01 1.82934508e-01 1.06965482e+00 -9.20167744e-01
-4.05700356e-01 -3.04947793e-01 -9.50549185e-01 -1.27810824e+00
-3.11909500e-03 -5.92034042e-01 -3.30851167e-01 -1.22981036e+00
-3.29352111e-01 -3.07814121e-01 1.20796105e-02 1.74582824e-02
-2.11976588e-01 4.20113236e-01 4.36797440e-01 5.74547127e-02
-5.29391229e-01 5.29640913e-01 1.58458126e+00 3.94555062e-01
-1.80478871e-01 -3.53229433e-01 2.11563278e-02 4.71618205e-01
3.99511725e-01 3.35526504e-02 -6.64891481e-01 -6.12411380e-01
1.26605526e-01 3.52302551e-01 8.63731086e-01 -1.54168820e+00
2.00663447e-01 -1.97561935e-01 4.52409238e-01 -3.86456937e-01
3.97882223e-01 -6.65859938e-01 2.87434906e-01 3.43626082e-01
-2.28540264e-02 2.17236936e-01 3.66415083e-01 4.17083621e-01
-1.72281742e-01 4.06019002e-01 7.59893477e-01 3.78769450e-02
-1.11486995e+00 6.39829397e-01 -1.08035639e-01 5.14861166e-01
9.96912301e-01 -3.29954147e-01 -7.21559882e-01 -3.07909042e-01
-3.68265420e-01 -4.07708809e-02 4.91680503e-01 5.81522048e-01
6.93853617e-01 -1.23989272e+00 -6.89170122e-01 6.16858959e-01
3.01089406e-01 1.27632797e-01 4.23735768e-01 5.86218834e-01
-9.48414922e-01 5.52362353e-02 -5.82749486e-01 -1.04419255e+00
-7.23114312e-01 3.93403262e-01 4.80785459e-01 8.74062777e-02
-9.84022081e-01 6.99429214e-01 1.50162876e-01 2.40639821e-01
2.28940621e-02 -6.11544967e-01 8.37393031e-02 -6.15811907e-02
9.81838226e-01 7.74737597e-01 2.26523191e-01 -4.08139497e-01
-4.28599536e-01 4.97483045e-01 3.55839401e-01 1.39845997e-01
1.07536709e+00 -3.50850224e-02 3.33910882e-01 1.59947544e-01
1.11858130e+00 -1.47151081e-02 -1.95288730e+00 1.17653176e-01
-6.21537268e-01 -1.05373526e+00 1.43484026e-01 -6.42532647e-01
-1.27597821e+00 9.43139672e-01 7.27466106e-01 5.89204356e-02
1.41984856e+00 -4.35468554e-01 9.83423650e-01 -4.14582230e-02
6.03573203e-01 -5.68886161e-01 2.79080451e-01 2.61349380e-01
7.02541173e-01 -1.15745711e+00 -1.11837469e-01 -4.26477849e-01
-3.42817873e-01 1.10220349e+00 7.83601046e-01 -2.59061873e-01
4.18052882e-01 4.63050783e-01 1.07178479e-01 5.44828586e-02
-1.08026242e+00 -1.84342265e-01 6.93425909e-02 7.92067587e-01
4.22771461e-02 -3.77911180e-01 2.64057040e-01 -2.03859940e-01
-3.90401721e-01 4.66786265e-01 7.31431127e-01 6.57463014e-01
-5.81677221e-02 -6.87336087e-01 -2.49676853e-01 9.38988775e-02
-1.57270983e-01 6.00631535e-02 4.89769042e-01 8.31093013e-01
1.39971301e-01 8.84550273e-01 5.02698302e-01 -2.72551626e-01
6.04184270e-01 -4.90360111e-01 6.20832860e-01 -1.09618925e-01
-6.56580329e-01 -2.91933507e-01 -4.05032039e-02 -1.37673795e+00
-4.73091990e-01 -6.13098264e-01 -1.19171941e+00 -7.26876020e-01
1.51968807e-01 -2.94620216e-01 5.86635292e-01 6.36213124e-01
6.71065986e-01 6.23209953e-01 6.00556672e-01 -9.95397449e-01
-8.04219116e-03 -5.83693504e-01 -2.40386456e-01 7.41565466e-01
5.70307493e-01 -7.45355427e-01 -5.38601875e-01 2.48984039e-01]
|
[8.641474723815918, -1.4310989379882812]
|
a0db111c-acf3-4c9d-a277-c85da8b117cb
|
a-novel-end-to-end-framework-for-occluded
|
2304.07721
| null |
https://arxiv.org/abs/2304.07721v1
|
https://arxiv.org/pdf/2304.07721v1.pdf
|
A Novel end-to-end Framework for Occluded Pixel Reconstruction with Spatio-temporal Features for Improved Person Re-identification
|
Person re-identification is vital for monitoring and tracking crowd movement to enhance public security. However, re-identification in the presence of occlusion substantially reduces the performance of existing systems and is a challenging area. In this work, we propose a plausible solution to this problem by developing effective occlusion detection and reconstruction framework for RGB images/videos consisting of Deep Neural Networks. Specifically, a CNN-based occlusion detection model classifies individual input frames, followed by a Conv-LSTM and Autoencoder to reconstruct the occluded pixels corresponding to the occluded frames for sequential (video) and non-sequential (image) data, respectively. The quality of the reconstructed RGB frames is further refined and fine-tuned using a Conditional Generative Adversarial Network (cGAN). Our method is evaluated on four well-known public data sets of the domain, and the qualitative reconstruction results are indeed appealing. Quantitative evaluation in terms of re-identification accuracy of the Siamese network showed an exceptional Rank-1 accuracy after occluded pixel reconstruction on various datasets. A comparative analysis with state-of-the-art approaches also demonstrates the robustness of our work for use in real-life surveillance systems.
|
['Santosh Kumar', 'Satyanarayana Vollala', 'Praneeth Nemani', 'Ghanta Sai Krishna', 'Prathistith Raj Medi']
|
2023-04-16
| null | null | null | null |
['person-re-identification']
|
['computer-vision']
|
[ 2.75750339e-01 -1.35852158e-01 2.37739339e-01 -2.77707428e-01
-6.23229802e-01 -1.94506109e-01 4.66127783e-01 -2.61180252e-01
-7.25543320e-01 8.55021775e-01 2.18610808e-01 7.54381567e-02
2.35566616e-01 -7.37306237e-01 -8.94965291e-01 -8.93554509e-01
1.51712760e-01 3.25369716e-01 9.06927288e-02 -1.21578053e-01
-2.62063801e-01 6.79936469e-01 -1.82992852e+00 2.72008687e-01
5.79951763e-01 1.04126227e+00 -4.15369868e-01 8.24975312e-01
3.22918892e-01 6.93092823e-01 -9.43005264e-01 -8.59276056e-01
4.64244276e-01 -2.04213131e-02 -5.90149045e-01 4.30319041e-01
8.66854966e-01 -8.05297852e-01 -6.49523079e-01 8.80233526e-01
5.88984311e-01 2.48468399e-01 3.32624912e-01 -1.43785262e+00
-5.31251252e-01 -2.32836735e-02 -3.72379869e-01 2.60714501e-01
7.76809692e-01 4.56448495e-01 2.51731247e-01 -5.25799572e-01
5.44962525e-01 1.22617936e+00 1.29696620e+00 7.39350557e-01
-1.11012781e+00 -7.77891219e-01 3.21477242e-02 1.10447176e-01
-1.45051932e+00 -5.73623419e-01 7.84996033e-01 -4.48549896e-01
7.11971462e-01 3.60734165e-01 8.92877817e-01 1.62527609e+00
-7.90631026e-02 6.86377525e-01 9.21050906e-01 -3.15551162e-01
-4.99029458e-03 1.56247959e-01 -3.52249146e-02 6.56979144e-01
4.42731082e-01 4.33985770e-01 -4.22434151e-01 -1.89901087e-02
6.35526597e-01 4.94659185e-01 -3.19217712e-01 -1.17086530e-01
-9.70350623e-01 7.04026222e-01 6.48790658e-01 2.40674421e-01
-4.25281882e-01 1.46474764e-01 3.50895405e-01 -1.53809683e-02
5.05911589e-01 -5.16978279e-03 -1.20715052e-01 2.82777935e-01
-1.10072541e+00 3.34145844e-01 5.03130138e-01 6.10565901e-01
6.63461268e-01 1.35610357e-01 -4.34117079e-01 5.32387733e-01
2.95492649e-01 6.89660370e-01 3.77490759e-01 -1.03248751e+00
3.89375597e-01 5.77924728e-01 4.21587050e-01 -1.27272749e+00
-6.36194348e-01 -3.25071990e-01 -1.01691175e+00 3.50509435e-01
7.48602927e-01 -1.46353617e-01 -6.78921044e-01 1.69884443e+00
6.60667539e-01 2.52310693e-01 2.48682097e-01 8.70829284e-01
1.08430028e+00 3.11093867e-01 2.97044367e-01 -7.64826834e-02
1.39233446e+00 -1.04763258e+00 -7.19857335e-01 -1.82760224e-01
1.41538903e-01 -4.03841197e-01 4.97645587e-01 1.48255542e-01
-9.23870564e-01 -1.04936934e+00 -8.43089819e-01 -1.16825979e-02
-3.95643413e-01 3.71867746e-01 3.58945280e-01 1.05937207e+00
-1.13044572e+00 4.10206079e-01 -7.14142799e-01 -6.27331197e-01
7.60676444e-01 6.25908494e-01 -7.85425305e-01 -3.64748426e-02
-1.01814568e+00 6.18311584e-01 1.90371320e-01 4.82992977e-01
-9.50826049e-01 -4.12355483e-01 -8.22736979e-01 -3.34679335e-01
-7.62546659e-02 -7.97711968e-01 6.71572685e-01 -1.28155625e+00
-1.34860468e+00 1.18574536e+00 -2.26352856e-01 -7.56582260e-01
9.74481940e-01 -2.14811161e-01 -6.21968448e-01 2.38790989e-01
2.51737893e-01 8.62265170e-01 1.17724895e+00 -1.38785577e+00
-7.54652679e-01 -4.04665440e-01 1.40516758e-01 -2.14482378e-03
-4.91241366e-01 4.98824269e-02 -4.94921833e-01 -7.25164056e-01
-1.18935741e-01 -8.76685679e-01 1.05826417e-02 1.10652864e-01
-4.88485515e-01 -1.10763460e-01 9.30539846e-01 -1.27445829e+00
9.27946210e-01 -1.98283446e+00 -3.21354195e-02 -4.21236642e-02
3.95543158e-01 4.64899510e-01 -5.89142889e-02 -1.13875503e-02
-7.89272189e-02 -1.34788617e-01 -2.82076806e-01 -9.03624892e-01
-1.62912577e-01 -5.52425906e-02 -8.86005834e-02 9.32696462e-01
6.20729178e-02 1.04789972e+00 -6.77485108e-01 -3.92307252e-01
6.39650524e-01 9.94704664e-01 -2.78989732e-01 5.51190257e-01
2.50076562e-01 8.79736304e-01 -1.98397972e-03 9.21867967e-01
7.41153955e-01 1.34042846e-02 -2.82684684e-01 -4.83580559e-01
3.37073654e-02 -4.95627940e-01 -1.11544859e+00 1.44185650e+00
-1.02895051e-01 9.30871427e-01 -2.65050344e-02 -7.24690735e-01
6.43556654e-01 2.89165467e-01 6.63665652e-01 -6.34673476e-01
3.15683603e-01 -1.20667368e-01 -5.69578111e-01 -9.00045335e-01
6.12726212e-01 2.91845471e-01 9.39135998e-02 1.30488366e-01
-2.37898156e-01 5.12884080e-01 -1.92830153e-02 -2.06404567e-01
8.84037197e-01 1.73170850e-01 -1.04125701e-01 9.09642950e-02
8.32547247e-01 -1.19010031e-01 4.07029599e-01 7.30156362e-01
-7.88936675e-01 8.26863229e-01 -3.37369852e-02 -1.02035081e+00
-1.18047571e+00 -7.72902608e-01 -4.92084138e-02 8.47429514e-01
3.14619809e-01 6.70512468e-02 -1.32226300e+00 -6.42351866e-01
-2.75228620e-02 3.62503171e-01 -1.00222540e+00 3.65405232e-02
-8.39463353e-01 -9.48884606e-01 9.39712405e-01 5.47146499e-01
1.01132047e+00 -1.27188218e+00 -7.75448620e-01 7.34880418e-02
-6.31261051e-01 -1.38624144e+00 -2.77630001e-01 -5.09861231e-01
-4.37644243e-01 -1.44172716e+00 -9.73330259e-01 -7.11015165e-01
7.26824105e-01 3.82359892e-01 9.19754326e-01 3.43471795e-01
-3.69194120e-01 6.46848261e-01 -8.71268883e-02 -3.91940773e-02
-2.97330052e-01 -1.64327666e-01 2.07384691e-01 5.63414991e-01
4.06229138e-01 -1.14403971e-01 -9.16082144e-01 4.46020275e-01
-8.64586949e-01 -2.24721342e-01 2.93753296e-02 8.20193708e-01
4.55495000e-01 9.33166072e-02 1.87664792e-01 -5.29967844e-01
2.71000117e-01 -6.26856834e-02 -5.03582239e-01 3.85906488e-01
-2.07399074e-02 -4.50521082e-01 3.21230948e-01 -6.08213067e-01
-1.10775721e+00 3.44664544e-01 -2.55477965e-01 -4.39269394e-01
-5.29896796e-01 -5.09766579e-01 -2.79530317e-01 -3.79695654e-01
5.64274311e-01 1.37703359e-01 -1.07847251e-01 -1.29400849e-01
7.85545558e-02 5.54069400e-01 1.03363550e+00 -1.26889959e-01
7.68529475e-01 1.05438042e+00 -2.15418443e-01 -9.39784467e-01
-7.09683120e-01 -2.58590728e-01 -8.21802974e-01 -6.07693374e-01
1.33875239e+00 -1.13051319e+00 -1.16196728e+00 9.77258146e-01
-1.42535686e+00 -3.08763921e-01 -3.21643978e-01 7.35027418e-02
-2.04553202e-01 5.57922661e-01 -4.74642605e-01 -1.03498161e+00
-3.79826277e-01 -1.39547324e+00 1.35566568e+00 4.52926576e-01
-1.71910562e-02 -9.50383008e-01 -4.86986060e-03 8.81122887e-01
4.26197618e-01 8.38956773e-01 -1.33556738e-01 -3.30797911e-01
-5.74216425e-01 -6.77663326e-01 -2.54488647e-01 1.85745209e-01
-1.24102086e-01 -1.32249072e-01 -1.45981801e+00 -5.50493181e-01
-7.59365931e-02 -1.15893207e-01 9.49423611e-01 5.24691463e-01
1.16049826e+00 -3.90601397e-01 -4.70787257e-01 9.46666658e-01
1.15728998e+00 -2.10695580e-01 8.78778517e-01 5.53624928e-01
1.08552182e+00 7.36320019e-01 1.53165922e-01 3.72326523e-01
6.51852727e-01 8.51107538e-01 5.35479546e-01 -4.47540939e-01
-4.07932699e-01 -2.04258129e-01 3.62002492e-01 1.74083225e-02
-3.90709847e-01 -3.96387011e-01 -6.14585578e-01 2.88769364e-01
-1.78741908e+00 -1.21755612e+00 -9.44417641e-02 2.24732995e+00
2.47690946e-01 -2.74234027e-01 4.29125696e-01 3.48937660e-01
1.07450438e+00 1.30510986e-01 -4.22133952e-01 3.95822585e-01
-6.25748813e-01 -1.92637578e-01 6.53403580e-01 5.47824025e-01
-1.64967597e+00 9.10108328e-01 6.21829462e+00 6.29410923e-01
-8.12187910e-01 3.66556436e-01 9.18410957e-01 7.21253157e-02
3.07727754e-01 -5.30073285e-01 -8.84105802e-01 5.75684547e-01
6.76610589e-01 7.35410213e-01 4.88923460e-01 6.43491566e-01
2.39375588e-02 -8.17091540e-02 -8.38170648e-01 1.36909997e+00
5.39335728e-01 -1.09184289e+00 -1.97756037e-01 9.24805701e-02
8.48815203e-01 -3.09645355e-01 1.92722872e-01 1.07343338e-01
9.39226970e-02 -1.15746653e+00 8.54829788e-01 9.17483747e-01
7.29322195e-01 -7.11189091e-01 9.49607551e-01 -5.69333918e-02
-1.26667905e+00 -2.67302334e-01 -3.37583303e-01 1.29109070e-01
2.18108475e-01 1.76332369e-01 -3.65852565e-01 2.51685679e-01
1.16540408e+00 6.41024113e-01 -8.90010417e-01 9.58596528e-01
-2.12980673e-01 1.36407450e-01 -2.28662416e-01 3.58590841e-01
-1.68680921e-01 1.00595817e-01 5.14460862e-01 1.32466996e+00
2.29862481e-02 1.71826780e-02 1.38106614e-01 6.01812124e-01
-1.53861701e-01 -3.58682692e-01 -5.31382442e-01 6.35909140e-01
2.39546061e-01 1.10403383e+00 -5.73418260e-01 -4.76044804e-01
-2.41668284e-01 1.38424420e+00 4.00411129e-01 6.10220015e-01
-1.13319552e+00 1.70362428e-01 6.75715208e-01 1.66203797e-01
2.40361527e-01 4.56161834e-02 1.27509553e-02 -1.20169401e+00
4.31738794e-02 -8.85890484e-01 4.50508028e-01 -6.06436253e-01
-1.33215857e+00 1.01737440e+00 -5.12851104e-02 -1.15343571e+00
-1.94041267e-01 -4.82416034e-01 -4.30835456e-01 4.65607613e-01
-1.38754666e+00 -1.61007082e+00 -1.06278515e+00 8.76703143e-01
3.90743345e-01 -4.29531336e-01 7.59676397e-01 5.70599377e-01
-1.10243320e+00 9.14775431e-01 -2.95494357e-03 5.23617208e-01
5.24325550e-01 -7.88426995e-01 3.19071978e-01 1.20485663e+00
-1.04041025e-01 1.43829450e-01 6.94490910e-01 -6.14480019e-01
-1.10998690e+00 -1.42642367e+00 6.53305769e-01 -7.27904379e-01
5.75627275e-02 -1.65586486e-01 -7.20224500e-01 6.22461319e-01
1.41821191e-01 3.27416897e-01 5.60536206e-01 -5.19635201e-01
-1.86866626e-01 -2.89864719e-01 -1.54049230e+00 3.02615792e-01
1.05874050e+00 -4.50738966e-01 -2.67499000e-01 4.84962881e-01
5.83460271e-01 -4.56083059e-01 -7.76911557e-01 3.29176277e-01
7.40285993e-01 -1.46502638e+00 1.62592816e+00 -3.40048581e-01
3.43013182e-02 -4.07960474e-01 -2.60760456e-01 -6.58438325e-01
-7.81074241e-02 -4.09847915e-01 -2.64285475e-01 1.22956967e+00
-2.47963279e-01 -6.14780545e-01 1.03991139e+00 1.00371575e+00
3.24772090e-01 -2.09848106e-01 -1.14760804e+00 -3.87096941e-01
-4.53880847e-01 -4.31855887e-01 7.91153371e-01 5.90326786e-01
-8.98187339e-01 -3.12165886e-01 -7.80852318e-01 4.94309306e-01
1.05978501e+00 -2.70262688e-01 1.24469674e+00 -1.08695602e+00
-4.58803251e-02 -2.24149853e-01 -5.95346630e-01 -7.23065019e-01
3.70097041e-01 -3.59344631e-01 -1.99286476e-01 -1.25524509e+00
1.16724044e-01 -2.72678286e-01 -9.47863702e-03 3.04499567e-01
-2.88715214e-01 9.81380224e-01 2.36781031e-01 3.34781885e-01
-8.43536615e-01 6.32417798e-01 8.35067987e-01 -4.70545441e-01
-2.62196630e-01 2.18762875e-01 -2.64061898e-01 8.05044889e-01
6.70484960e-01 -2.59817719e-01 2.39464223e-01 -5.48341930e-01
-3.00791711e-01 -1.77292287e-01 1.09541452e+00 -1.60659397e+00
4.44931418e-01 4.54229563e-01 1.01631868e+00 -6.65132046e-01
7.10778534e-01 -1.00730491e+00 5.57839811e-01 7.90069997e-01
-1.42706618e-01 1.95302397e-01 2.18286812e-01 7.15960145e-01
-4.63092560e-03 -1.52649488e-02 7.91316867e-01 3.21282409e-02
-6.09624088e-01 4.40517426e-01 -6.71662092e-02 -1.50321856e-01
1.07497513e+00 -6.37784362e-01 -2.14573294e-01 -3.48724693e-01
-6.28456175e-01 -1.64356545e-01 6.00393713e-01 2.84984827e-01
7.32551455e-01 -1.30143094e+00 -7.65506327e-01 5.30547321e-01
2.22413745e-02 -2.10111544e-01 4.89965647e-01 6.54621720e-01
-8.02906930e-01 1.46017835e-01 -3.87470543e-01 -7.52671480e-01
-1.33114743e+00 6.76839471e-01 6.44606948e-01 -3.03581297e-01
-6.16904378e-01 7.84845233e-01 8.22784454e-02 -2.95188844e-01
6.85316026e-01 -3.69221531e-02 -4.53668028e-01 -2.98313070e-02
8.28840971e-01 7.16792583e-01 -1.94395915e-01 -1.31788230e+00
-3.90530407e-01 6.51892126e-01 3.69645745e-01 4.70004454e-02
1.21190536e+00 -3.59527558e-01 -1.92850064e-02 -1.43594667e-01
1.18465292e+00 -2.76181757e-01 -1.70897436e+00 -1.47275329e-01
-5.26476920e-01 -5.25078475e-01 -2.58123547e-01 -3.34766805e-01
-1.41998053e+00 6.90311372e-01 1.20017469e+00 1.44785672e-01
1.10949624e+00 -3.66641045e-01 8.93215477e-01 2.81378657e-01
1.12197250e-01 -8.39722216e-01 1.86623454e-01 -2.05155443e-02
7.77124524e-01 -1.77034140e+00 6.34815842e-02 -2.14470059e-01
-4.71862704e-01 9.68852580e-01 5.59143901e-01 -1.58505991e-01
5.01665413e-01 -8.41125399e-02 5.14794812e-02 -3.90217789e-02
2.84825981e-01 -2.07128376e-01 2.64577657e-01 1.00699925e+00
-6.95415512e-02 6.45466298e-02 3.27007860e-01 5.06701887e-01
-1.76350802e-01 -3.16517621e-01 8.93975142e-03 4.33905840e-01
-3.78765799e-02 -6.93854511e-01 -1.02244270e+00 -8.21379386e-03
-3.86745065e-01 2.85030305e-01 -1.57537162e-01 8.51663053e-01
6.28582954e-01 1.18031824e+00 1.35439605e-01 -4.61293787e-01
2.48522997e-01 -3.53609800e-01 3.31180483e-01 2.19118044e-01
-9.95344162e-01 -1.49307519e-01 -1.63929313e-01 -6.67808473e-01
-1.05675721e+00 -7.98149049e-01 -4.88943964e-01 -5.03152966e-01
-1.42004684e-01 -3.59586149e-01 5.30167580e-01 8.84339690e-01
9.80473384e-02 6.19461894e-01 4.40892071e-01 -1.37347782e+00
1.05636809e-02 -8.50055158e-01 -3.23356651e-02 9.64978218e-01
5.93922555e-01 -7.42971539e-01 -2.64307857e-01 3.47908348e-01]
|
[14.636724472045898, 0.9541665315628052]
|
86285f61-fcd6-484f-bbe5-344749637c32
|
learning-discriminative-data-fitting
| null | null |
http://openaccess.thecvf.com/content_iccv_2017/html/Pan_Learning_Discriminative_Data_ICCV_2017_paper.html
|
http://openaccess.thecvf.com/content_ICCV_2017/papers/Pan_Learning_Discriminative_Data_ICCV_2017_paper.pdf
|
Learning Discriminative Data Fitting Functions for Blind Image Deblurring
|
Solving blind image deblurring usually requires defining a data fitting function and image priors. While existing algorithms mainly focus on developing image priors for blur kernel estimation and non-blind deconvolution, only a few methods consider the effect of data fitting functions. In contrast to the state-of-the-art methods that use a single or a fixed data fitting term, we propose a data-driven approach to learn effective data fitting functions from a large set of motion blurred images with associated ground truth blur kernels. The learned data fitting function facilitates estimating accurate blur kernels for generic images and domain-specific problems with corresponding image priors. In addition, we extend the learning approach for data fitting function to latent image restoration and non-uniform deblurring. Extensive experiments on challenging motion blurred images demonstrate the proposed algorithm performs favorably against the state-of-the-art methods.
|
['Ming-Hsuan Yang', 'Yu-Wing Tai', 'Zhixun Su', 'Jinshan Pan', 'Jiangxin Dong']
|
2017-10-01
| null | null | null |
iccv-2017-10
|
['blind-image-deblurring']
|
['computer-vision']
|
[ 1.38693765e-01 -5.92303932e-01 1.02435008e-01 -3.54157925e-01
-6.36520863e-01 -5.23844242e-01 4.45358872e-01 -8.13500762e-01
-2.41823316e-01 6.86396897e-01 6.85433149e-01 -3.06485556e-02
-4.53210741e-01 -6.57563210e-02 -5.91138363e-01 -7.90634334e-01
1.53530568e-01 2.76767649e-02 -6.74340129e-02 3.76565546e-01
4.90872264e-01 1.75451308e-01 -1.10603571e+00 -5.90598322e-02
1.30773926e+00 6.46786451e-01 7.76024699e-01 7.97766566e-01
2.80280322e-01 9.70001221e-01 -3.69983464e-01 -1.33445114e-02
3.79307330e-01 -4.67806727e-01 -7.88513899e-01 6.14127636e-01
5.86452305e-01 -9.30463552e-01 -8.77492070e-01 1.51024699e+00
4.01391804e-01 2.30257064e-01 6.86542213e-01 -8.04746151e-01
-1.54373622e+00 2.02761501e-01 -7.11357534e-01 4.19199705e-01
1.53834492e-01 2.40210757e-01 3.45884144e-01 -1.03327119e+00
2.68874526e-01 1.01098561e+00 8.43160331e-01 4.73397046e-01
-1.37111378e+00 -2.73926973e-01 -1.92037329e-01 4.40608442e-01
-1.32628274e+00 -7.44458735e-01 7.32679963e-01 -7.37265348e-01
3.54782999e-01 2.72682935e-01 1.72574967e-01 8.59879315e-01
-1.11544058e-01 6.29125893e-01 1.46738791e+00 -2.98385799e-01
2.61368871e-01 -2.56601572e-01 3.91361475e-01 4.43540424e-01
4.40111369e-01 2.11912647e-01 -3.06027055e-01 -3.43446672e-01
1.36379206e+00 1.43442363e-01 -1.16611183e+00 -4.20755476e-01
-1.40071833e+00 4.92331862e-01 3.61061871e-01 9.89087820e-02
-5.99966228e-01 3.26482385e-01 -7.76234195e-02 -3.13388160e-03
6.27530396e-01 3.19986939e-01 -3.58711213e-01 7.12800492e-03
-1.43648970e+00 2.38118246e-01 5.52012801e-01 8.80494893e-01
9.32958901e-01 5.09952717e-02 -5.51336110e-01 1.10973907e+00
4.03756440e-01 5.57354391e-01 5.38234293e-01 -1.27850389e+00
3.89385745e-02 -2.10485786e-01 9.15916860e-01 -6.20209157e-01
8.54943469e-02 -2.02858701e-01 -8.97508621e-01 1.60530150e-01
6.39424205e-01 -1.62912175e-01 -1.23932672e+00 1.51510203e+00
-2.94603799e-02 9.27996814e-01 6.41340688e-02 1.57920706e+00
4.46594745e-01 5.70319414e-01 -4.06226367e-01 -3.42932016e-01
1.28984082e+00 -1.29007411e+00 -1.11136162e+00 -4.63666022e-01
-1.82606116e-01 -1.15993190e+00 9.97797191e-01 2.99990565e-01
-1.17426193e+00 -6.48024440e-01 -7.79353499e-01 -2.80782223e-01
2.37243354e-01 4.11064565e-01 5.03418803e-01 6.24350071e-01
-1.27429914e+00 3.77690196e-01 -7.47364104e-01 -3.01226377e-01
3.83772522e-01 9.41088498e-02 -1.33437872e-01 -5.65090895e-01
-9.02842879e-01 1.19951999e+00 3.20394747e-02 2.83050120e-01
-1.37770092e+00 -9.41163421e-01 -8.07405353e-01 -1.31376773e-01
3.24511975e-02 -1.15835452e+00 1.41069591e+00 -9.40140903e-01
-1.61640608e+00 5.88374972e-01 -4.01684076e-01 -4.27307129e-01
5.58442712e-01 -8.74605060e-01 -2.08586961e-01 3.28031957e-01
-1.04466110e-01 4.10544187e-01 1.68930507e+00 -1.60384631e+00
-1.57522410e-01 1.36041477e-01 -1.13096230e-01 2.61368364e-01
-1.54512301e-01 1.76526055e-01 -6.04263604e-01 -1.01325750e+00
4.04613614e-02 -6.44122481e-01 -1.75950170e-01 4.43037115e-02
-2.17428535e-01 4.28705841e-01 1.06381595e+00 -1.17723608e+00
1.18821514e+00 -2.07986546e+00 3.17792892e-01 -4.17725563e-01
2.32979223e-01 2.01021656e-01 -9.64327902e-02 -4.83751297e-02
-1.44956380e-01 -4.82708693e-01 -6.08564913e-01 -3.47474635e-01
-1.85783170e-02 4.89490852e-02 -4.91648346e-01 9.82644320e-01
-1.38574481e-01 7.33521283e-01 -1.02035952e+00 2.07110550e-02
5.97796977e-01 6.71396315e-01 -4.36595082e-01 6.96583807e-01
7.20625222e-02 5.41724741e-01 -1.16575144e-01 3.89664978e-01
1.18245792e+00 -5.37073314e-01 -1.28728792e-01 -6.67731047e-01
-1.50135472e-01 -1.42515093e-01 -1.08922946e+00 1.88253748e+00
-2.90495753e-01 8.21409047e-01 5.49539924e-01 -6.97765827e-01
4.84903723e-01 4.18953747e-01 3.19389999e-01 1.46611124e-01
-9.52442065e-02 2.10075170e-01 -2.41576031e-01 -8.86388302e-01
6.79621577e-01 -1.12498440e-01 5.23489356e-01 4.23647851e-01
6.35878788e-03 -2.06494957e-01 -2.83126026e-01 -3.76195759e-02
9.95371342e-01 3.19873124e-01 5.60357682e-02 -5.83946884e-01
5.57992458e-01 -1.30655214e-01 2.98245490e-01 8.89117599e-01
-3.86304557e-01 1.28391850e+00 -4.09834057e-01 -2.60736853e-01
-1.23666942e+00 -1.07768428e+00 -2.54179746e-01 6.80640161e-01
5.93761027e-01 6.54181391e-02 -1.31775022e+00 -2.37480789e-01
-1.41670167e-01 7.14240313e-01 -5.08265853e-01 1.16240434e-01
-2.85493702e-01 -1.11268246e+00 2.85761505e-01 1.95691735e-01
8.95080566e-01 -5.15733778e-01 -2.62966156e-01 9.56992209e-02
-5.17443061e-01 -1.20240188e+00 -1.35173118e+00 -3.73502254e-01
-8.15809011e-01 -1.05315208e+00 -1.47319388e+00 -7.59068608e-01
1.03269815e+00 9.90500927e-01 7.68055558e-01 -2.98243850e-01
-1.56906933e-01 7.07061231e-01 -5.97708859e-02 2.35312894e-01
-3.51437658e-01 -6.72133029e-01 1.85096800e-01 3.39361727e-01
1.13912143e-01 -3.55506420e-01 -1.02781594e+00 4.29422766e-01
-1.04236054e+00 3.18092048e-01 5.99157751e-01 9.38818872e-01
1.50803491e-01 1.46512017e-01 2.76999354e-01 -4.82538939e-01
9.27696466e-01 -4.75603700e-01 -6.33325160e-01 2.98048168e-01
-6.94637120e-01 1.52768001e-01 2.81814545e-01 -8.27562332e-01
-1.66142833e+00 -4.64465134e-02 6.36842787e-01 -9.26106751e-01
-2.52400249e-01 3.54097515e-01 5.82364462e-02 -1.58464447e-01
8.98761272e-01 3.97705078e-01 1.17068842e-01 -7.38280416e-01
6.67442739e-01 7.76986599e-01 1.09726870e+00 -4.59395856e-01
9.14837837e-01 6.52796805e-01 -4.57960606e-01 -8.61862361e-01
-7.19054341e-01 -7.15072036e-01 -4.92820859e-01 -1.66399017e-01
9.56037641e-01 -1.21922255e+00 -2.60383815e-01 1.14481151e+00
-1.37913275e+00 -4.47693467e-01 7.31311589e-02 6.86566472e-01
-7.86332488e-01 9.53047097e-01 -8.92295003e-01 -7.73114026e-01
-2.39577129e-01 -1.34308684e+00 8.24038744e-01 2.83266217e-01
5.21969721e-02 -1.19959354e+00 9.30142626e-02 6.13352299e-01
7.82095015e-01 -3.43055695e-01 4.30774271e-01 2.67267793e-01
-9.38481212e-01 1.88359112e-01 -7.66047955e-01 6.19040549e-01
7.31360078e-01 -6.52494431e-01 -1.35806596e+00 -4.28514808e-01
5.17832279e-01 1.51266113e-01 1.00430095e+00 1.12716210e+00
1.17429316e+00 -5.28092265e-01 -1.20523334e-01 9.63642716e-01
1.39755166e+00 -2.31223732e-01 8.73719096e-01 2.17954621e-01
8.45337331e-01 1.69974416e-01 3.37207437e-01 2.58747756e-01
3.23695123e-01 6.53216541e-01 2.04926729e-01 -1.59571484e-01
-6.04882002e-01 1.56845137e-01 4.59227622e-01 5.71891427e-01
-1.81705877e-01 4.77242917e-02 -6.19229078e-01 8.01312089e-01
-2.10168672e+00 -9.10806417e-01 -2.51149297e-01 2.14519715e+00
1.21548152e+00 -5.88225007e-01 -2.72381186e-01 -4.20659482e-01
1.08548236e+00 2.48334169e-01 -6.17291093e-01 5.29147923e-01
-1.87757805e-01 -2.24470884e-01 7.30246007e-01 9.45985436e-01
-1.12251770e+00 9.08784509e-01 7.18630552e+00 6.30679309e-01
-8.87874722e-01 3.52740705e-01 3.13593775e-01 2.73903757e-01
-3.05258721e-01 2.45300159e-01 -2.37419516e-01 6.52845562e-01
6.02339864e-01 -9.64123011e-02 1.28129387e+00 5.90777814e-01
7.25988865e-01 -3.06924194e-01 -8.86108935e-01 1.50157189e+00
6.61105588e-02 -1.34769917e+00 -2.09449008e-01 -3.40895914e-02
9.97774780e-01 1.14937067e-01 2.97705438e-02 -5.27635455e-01
4.15095538e-01 -9.74647582e-01 8.03973198e-01 1.07729626e+00
7.43369222e-01 6.26419932e-02 4.70379770e-01 1.47578314e-01
-5.41418195e-01 -2.34410502e-02 -4.95954961e-01 -1.65806055e-01
2.56080806e-01 9.77945030e-01 -4.83839571e-01 4.28367496e-01
7.79971898e-01 8.96410763e-01 -3.93491119e-01 1.60922909e+00
-2.04464108e-01 8.38339269e-01 2.27048233e-01 7.15347707e-01
-1.71301395e-01 -5.53889751e-01 6.90104485e-01 1.33159912e+00
5.64287066e-01 1.34577647e-01 -2.88376838e-01 1.17017233e+00
2.08123755e-02 -5.94586432e-01 -2.36341923e-01 1.72653496e-01
4.55239207e-01 1.12452567e+00 -2.62262642e-01 -3.19746852e-01
-5.40690601e-01 1.63503265e+00 3.55005488e-02 1.12979841e+00
-7.46719956e-01 -5.10964803e-02 1.03866971e+00 -2.35174466e-02
4.24139410e-01 -5.01077652e-01 -4.41611677e-01 -1.64789021e+00
-2.90968746e-01 -8.43652427e-01 -1.27853807e-02 -1.50235057e+00
-1.79230309e+00 3.09004486e-01 2.04681575e-01 -1.19379425e+00
8.78274664e-02 -5.85093975e-01 -6.14567220e-01 1.33118439e+00
-1.84296310e+00 -1.16133690e+00 -7.58403897e-01 6.35537446e-01
6.65135443e-01 -2.68359315e-02 4.75960761e-01 1.57270849e-01
-3.45212907e-01 -1.25684455e-01 4.73272562e-01 -5.21580540e-02
1.16754854e+00 -1.19395363e+00 3.59396756e-01 1.40714753e+00
-3.19795400e-01 1.10455084e+00 1.15405238e+00 -5.86660504e-01
-1.48674583e+00 -9.58038330e-01 1.42383516e-01 -4.16110784e-01
7.09426522e-01 9.09444392e-02 -1.12819350e+00 5.87989926e-01
7.20586717e-01 3.45223278e-01 2.77758509e-01 -4.02086169e-01
-2.39444047e-01 2.52504125e-02 -1.12192309e+00 3.58378202e-01
7.74316669e-01 -6.36563301e-01 -7.75204599e-01 2.42598712e-01
3.75323653e-01 -5.48420966e-01 -5.69441378e-01 2.31923074e-01
2.96307653e-01 -7.79094994e-01 1.36635530e+00 -2.55342692e-01
2.92119980e-01 -1.01799846e+00 -9.16073695e-02 -1.60564387e+00
-9.28797305e-01 -1.01029122e+00 -5.66005409e-01 1.03531158e+00
-3.22669894e-01 -4.26247448e-01 3.67900521e-01 6.95974588e-01
-2.64781684e-01 1.40327767e-01 -4.48103040e-01 -6.84899330e-01
-2.52818674e-01 -2.30803728e-01 3.35301965e-01 1.22866738e+00
-4.35532600e-01 -5.65761002e-03 -9.85547543e-01 6.99756384e-01
1.29992259e+00 1.43663540e-01 5.78337133e-01 -8.35555315e-01
-4.67901200e-01 -2.26883277e-01 7.77112916e-02 -1.63219082e+00
7.57368580e-02 -4.03416485e-01 5.15572786e-01 -1.74386668e+00
5.68766773e-01 -2.53873199e-01 -9.52434614e-02 1.09870337e-01
-6.52411520e-01 1.92046002e-01 -1.54419631e-01 6.69142008e-01
-2.95609087e-01 3.67729485e-01 1.49122977e+00 -1.08868703e-01
5.35070561e-02 -1.71115682e-01 -7.63399482e-01 6.55211031e-01
4.66571093e-01 -1.20545208e-01 -6.16934717e-01 -8.98458183e-01
-3.45378399e-01 -4.66455035e-02 8.94427955e-01 -6.61395550e-01
5.19223094e-01 -5.63607275e-01 4.17478025e-01 -2.40950376e-01
2.28452116e-01 -7.26552606e-01 5.10736048e-01 -4.03381139e-02
-4.48799394e-02 -5.06682456e-01 1.26557425e-01 7.39073157e-01
-3.37556638e-02 -4.53679413e-01 9.77799058e-01 -1.60194457e-01
-8.42616856e-01 1.43887758e-01 -4.12027478e-01 -2.22998951e-02
5.24709761e-01 -2.95800328e-01 -7.25684226e-01 -6.03609145e-01
-4.74756330e-01 -2.29371503e-01 8.03335905e-01 3.63912821e-01
8.10999811e-01 -1.14388573e+00 -7.62340486e-01 1.34204358e-01
-2.16661587e-01 -1.17679508e-02 4.04912829e-01 8.71850908e-01
-6.28335357e-01 1.17021300e-01 -1.30092800e-01 -4.90235925e-01
-8.91644418e-01 7.12402403e-01 6.69540584e-01 3.65804732e-01
-4.74883258e-01 8.65231037e-01 6.56122208e-01 7.71819800e-02
-4.14189883e-02 -4.77961898e-01 3.10385466e-01 -5.53995013e-01
7.36978829e-01 6.46571875e-01 -2.66351104e-01 -6.43045545e-01
-8.00127164e-02 5.19415557e-01 8.11268985e-02 -2.32783556e-01
1.26606643e+00 -6.95357442e-01 -3.95160824e-01 1.11000426e-02
8.04110885e-01 2.35514771e-02 -2.07469416e+00 -3.95803243e-01
-9.12334397e-02 -1.00822175e+00 7.02059567e-01 -8.90976191e-01
-9.39893365e-01 5.52915394e-01 7.65999913e-01 2.10788473e-02
1.43826616e+00 -1.29436597e-01 5.54953635e-01 -8.37795511e-02
2.37741157e-01 -7.21529484e-01 2.18154818e-01 3.97763729e-01
1.04817295e+00 -1.18118167e+00 2.22574696e-01 -4.98176545e-01
-4.05431151e-01 9.58548188e-01 2.78915882e-01 -1.35313094e-01
6.24788344e-01 2.77363658e-01 2.74741679e-01 8.56050327e-02
-1.94147244e-01 -1.43139854e-01 6.82461262e-01 8.80293250e-01
3.12189281e-01 -2.52858013e-01 4.22676187e-03 3.80061954e-01
3.30133080e-01 5.70759058e-01 7.27019489e-01 5.79083264e-01
-5.22059143e-01 -9.15559232e-01 -1.02814245e+00 9.56844613e-02
-3.69539350e-01 -5.10031402e-01 1.54502466e-01 8.42440426e-02
-1.48213640e-01 1.09980130e+00 -1.63484216e-01 1.01876013e-01
-1.35397062e-01 -2.12217942e-01 7.82205701e-01 -3.44256014e-01
-2.91674398e-02 3.88861537e-01 -2.73631781e-01 -1.13765523e-01
-7.46220410e-01 -6.93173766e-01 -8.53002012e-01 -2.49050617e-01
-5.32935798e-01 -1.99280027e-02 3.58837873e-01 8.31994951e-01
4.09806490e-01 2.78308600e-01 2.95075357e-01 -1.21332645e+00
-5.60971498e-01 -1.30535758e+00 -6.70531571e-01 5.55509388e-01
9.96482790e-01 -5.85486889e-01 -4.75733101e-01 9.26533401e-01]
|
[11.622532844543457, -2.7307403087615967]
|
571020d3-c52d-4cd5-ae62-701dad1dde40
|
attention-based-transformation-from-latent
|
2112.05324
| null |
https://arxiv.org/abs/2112.05324v1
|
https://arxiv.org/pdf/2112.05324v1.pdf
|
Attention-based Transformation from Latent Features to Point Clouds
|
In point cloud generation and completion, previous methods for transforming latent features to point clouds are generally based on fully connected layers (FC-based) or folding operations (Folding-based). However, point clouds generated by FC-based methods are usually troubled by outliers and rough surfaces. For folding-based methods, their data flow is large, convergence speed is slow, and they are also hard to handle the generation of non-smooth surfaces. In this work, we propose AXform, an attention-based method to transform latent features to point clouds. AXform first generates points in an interim space, using a fully connected layer. These interim points are then aggregated to generate the target point cloud. AXform takes both parameter sharing and data flow into account, which makes it has fewer outliers, fewer network parameters, and a faster convergence speed. The points generated by AXform do not have the strong 2-manifold constraint, which improves the generation of non-smooth surfaces. When AXform is expanded to multiple branches for local generations, the centripetal constraint makes it has properties of self-clustering and space consistency, which further enables unsupervised semantic segmentation. We also adopt this scheme and design AXformNet for point cloud completion. Considerable experiments on different datasets show that our methods achieve state-of-the-art results.
|
['Cheng Jin', 'Yuan Wu', 'Ximing Yang', 'Kaiyi Zhang']
|
2021-12-10
| null | null | null | null |
['point-cloud-completion', 'unsupervised-semantic-segmentation', 'point-cloud-generation']
|
['computer-vision', 'computer-vision', 'computer-vision']
|
[-2.40224123e-01 -1.64614208e-02 1.33447990e-01 -1.63947240e-01
-2.76893824e-01 -2.99592346e-01 4.28824425e-01 -3.58401276e-02
6.81068108e-04 3.20957989e-01 -3.24476570e-01 6.21451549e-02
-1.54172778e-01 -1.19101095e+00 -8.28201294e-01 -6.52589917e-01
2.17074960e-01 7.03910232e-01 6.10131860e-01 5.16959233e-03
3.13674420e-01 7.49939561e-01 -1.66245639e+00 -9.44711547e-03
1.57671213e+00 8.12489033e-01 3.68190408e-01 9.32517871e-02
-6.85290992e-01 -2.41732314e-01 -3.47275138e-01 -1.40447229e-01
3.73757005e-01 -5.38726114e-02 -6.39352441e-01 3.96076441e-01
3.35267723e-01 -5.38213626e-02 1.17695428e-01 1.08959937e+00
2.15498015e-01 5.86197153e-02 5.90111077e-01 -1.57549977e+00
-5.95362186e-01 2.00636759e-01 -6.38125777e-01 -5.77332556e-01
-3.30983400e-02 4.94784527e-02 6.32749140e-01 -1.22907734e+00
4.62586880e-01 1.50819790e+00 8.89142692e-01 3.17439735e-01
-1.13048625e+00 -8.91230762e-01 4.01627272e-01 -8.43250528e-02
-1.56987977e+00 -1.53119341e-01 8.89850438e-01 -5.00539541e-01
6.05195522e-01 3.56519789e-01 9.22177196e-01 4.10958856e-01
-2.29912922e-01 6.32944643e-01 6.67772591e-01 -1.19307801e-01
3.73738557e-01 -2.30593696e-01 -2.04952061e-02 5.35314500e-01
4.10138071e-01 -2.77280957e-01 -8.93071964e-02 -2.27740973e-01
1.23474073e+00 5.71177840e-01 -1.80000454e-01 -5.42807817e-01
-1.40865183e+00 8.15713584e-01 8.82138073e-01 2.87242979e-01
-4.00299668e-01 9.42435414e-02 -2.19004191e-02 4.42994684e-02
6.47311091e-01 3.61489385e-01 -2.82496542e-01 1.93793282e-01
-1.02635419e+00 2.39170626e-01 3.99507731e-01 1.34333026e+00
1.30366957e+00 -1.65813550e-01 1.00907996e-01 8.10163856e-01
5.05486667e-01 3.98736566e-01 3.64911377e-01 -9.74780142e-01
5.88519096e-01 1.26352143e+00 -1.01721361e-01 -1.34094203e+00
-3.89593989e-01 -2.07421198e-01 -1.09344435e+00 3.96016896e-01
2.56461967e-02 1.27625391e-01 -1.18474615e+00 1.19686937e+00
5.38473964e-01 4.78751600e-01 -1.74001664e-01 8.25764120e-01
7.00060904e-01 8.56545568e-01 -2.74951935e-01 -2.95677364e-01
9.81013238e-01 -1.02630854e+00 -5.02656341e-01 1.64314151e-01
4.55489427e-01 -7.51399934e-01 1.18232942e+00 4.27479386e-01
-1.00626695e+00 -6.35721505e-01 -8.85145426e-01 -6.18612580e-02
-1.78713217e-01 8.44195113e-02 7.36017466e-01 1.19271621e-01
-9.21001494e-01 9.28449631e-01 -1.04379487e+00 -2.51142234e-01
7.18099177e-01 4.25174206e-01 -2.17901558e-01 7.88528994e-02
-5.53589463e-01 2.02961311e-01 4.17034090e-01 1.32685259e-01
-2.32997134e-01 -6.86352313e-01 -6.17357254e-01 8.28018636e-02
3.03550243e-01 -8.50751519e-01 6.88538074e-01 -8.75496984e-01
-1.47054064e+00 3.86806369e-01 -3.12312663e-01 1.10691093e-01
7.33647048e-01 -6.86296299e-02 -1.80949554e-01 9.86637026e-02
2.71280020e-01 8.57503295e-01 9.37605143e-01 -1.40221167e+00
-5.67021191e-01 -3.75352800e-01 -3.45992178e-01 1.92701757e-01
-4.29311574e-01 -3.07612926e-01 -8.70660245e-01 -6.68345392e-01
1.06358016e+00 -9.16449189e-01 -4.11047727e-01 2.74325311e-01
-5.55535674e-01 -5.32088578e-01 1.17354167e+00 -2.73650736e-01
1.15400088e+00 -2.27199602e+00 1.41349956e-01 6.99951231e-01
4.04378146e-01 1.52626500e-01 -7.17701167e-02 3.03647339e-01
-1.07822679e-01 5.13238907e-01 -5.74756980e-01 -4.83873338e-01
-1.28399894e-01 2.68332928e-01 -1.92519918e-01 2.45779768e-01
2.65430063e-01 7.53549874e-01 -7.40907192e-01 -6.23174131e-01
3.10251206e-01 3.06568116e-01 -8.22218657e-01 -6.38794824e-02
-3.92185569e-01 5.03741860e-01 -7.73461521e-01 7.12482333e-01
1.17501593e+00 -2.36500144e-01 -3.76309514e-01 -2.16542721e-01
-3.78124774e-01 -1.12951495e-01 -1.40655577e+00 1.85318172e+00
-2.82619685e-01 1.26908466e-01 -3.72339189e-02 -5.58799386e-01
1.41481733e+00 7.07556382e-02 5.58679044e-01 -2.02927664e-01
-4.16175090e-02 4.43438023e-01 -4.11389410e-01 -1.82200879e-01
3.35283726e-01 9.55090448e-02 2.69068986e-01 -6.06626086e-03
-3.73357028e-01 -4.20639366e-01 5.08698970e-02 1.85638949e-01
5.98177314e-01 2.10138142e-01 -4.13614899e-01 -2.56720394e-01
4.25893396e-01 1.21681556e-01 9.95128274e-01 1.60637915e-01
4.96249259e-01 1.14143944e+00 3.74146372e-01 -3.38931888e-01
-1.13310146e+00 -9.87057805e-01 -3.19580168e-01 1.60938710e-01
5.56493282e-01 -5.88101268e-01 -9.38481927e-01 -5.35367250e-01
1.73900321e-01 5.14426112e-01 -1.77911550e-01 -8.34971666e-02
-5.65115929e-01 -5.56619942e-01 1.35807751e-03 4.82020676e-01
6.44904733e-01 -1.02949643e+00 -8.91348124e-02 3.73876721e-01
-1.79711223e-01 -7.01237500e-01 -4.19459075e-01 -5.13903201e-01
-1.49740922e+00 -1.01168144e+00 -7.86739707e-01 -8.38439465e-01
1.43293035e+00 3.97942901e-01 7.95503676e-01 6.86376333e-01
1.07131591e-02 -2.20261320e-01 -4.44962114e-01 -3.38443547e-01
-2.72062961e-02 1.17266685e-01 1.36454582e-01 2.08034277e-01
2.63194442e-01 -8.64028990e-01 -6.69259727e-01 6.29842699e-01
-1.03337955e+00 3.11558157e-01 4.77030724e-01 6.01352930e-01
1.03228986e+00 2.86204189e-01 3.49066615e-01 -6.81990027e-01
3.43854457e-01 -3.78878444e-01 -7.77068555e-01 -7.43100494e-02
-6.50811851e-01 -2.26290748e-01 6.61622941e-01 -3.61086965e-01
-7.95927167e-01 2.76966900e-01 -9.61532742e-02 -9.95253444e-01
-1.89098045e-01 1.79074213e-01 -4.23093647e-01 -2.19513942e-02
4.26331133e-01 3.65989469e-02 1.65092424e-01 -8.14875066e-01
3.24418306e-01 6.98121727e-01 3.67720813e-01 -4.18875754e-01
1.16983318e+00 7.04871356e-01 -7.16884881e-02 -7.81928599e-01
-2.61967182e-01 -4.00249571e-01 -9.36270356e-01 -2.02834606e-02
1.02273047e+00 -6.90367818e-01 -5.14116049e-01 6.64730549e-01
-1.24357092e+00 -5.30598462e-02 -3.47759604e-01 3.66371632e-01
-5.19332469e-01 4.81852323e-01 -3.36098760e-01 -5.26471615e-01
-3.20754260e-01 -1.09953427e+00 1.03321862e+00 2.37777114e-01
3.85798700e-02 -7.91119933e-01 -2.18126938e-01 1.08295314e-01
6.85826018e-02 4.77906108e-01 8.42960417e-01 -1.62597656e-01
-9.61857140e-01 -4.62165922e-02 -2.46156275e-01 3.03815246e-01
2.79632956e-01 4.99375463e-01 -6.15455210e-01 -3.33302945e-01
-1.27471030e-01 3.13344032e-01 6.78702116e-01 2.98582017e-01
1.52464890e+00 -3.60462993e-01 -6.64041936e-01 1.15369308e+00
1.32519901e+00 -8.59157555e-03 7.94196546e-01 3.78813148e-01
1.13569415e+00 6.09553456e-01 7.64184415e-01 1.55831724e-01
4.13566440e-01 5.40103555e-01 5.76918900e-01 -5.29661000e-01
1.06287539e-01 -3.97710323e-01 1.32721230e-01 1.27564931e+00
-2.38242984e-01 1.73431486e-01 -1.00742066e+00 5.35722613e-01
-2.09347391e+00 -6.38276637e-01 -8.20053041e-01 2.05871630e+00
6.02798522e-01 2.62381267e-02 -4.69048601e-03 1.12259515e-01
9.72267628e-01 -3.28505278e-01 -5.18579662e-01 1.88232549e-02
1.31186815e-02 -4.04556133e-02 3.96045536e-01 2.58200258e-01
-8.67304027e-01 1.14971352e+00 5.38064384e+00 1.05585206e+00
-1.05787253e+00 -2.15532724e-03 4.66113061e-01 4.38710377e-02
-6.12414956e-01 4.21305120e-01 -6.75620556e-01 7.37578869e-01
6.55847862e-02 7.50484020e-02 1.50856614e-01 9.22927558e-01
2.35346675e-01 1.36400402e-01 -7.97968268e-01 1.21991813e+00
-3.02888006e-01 -1.54327714e+00 2.93040365e-01 1.51281029e-01
9.92829144e-01 -1.77794993e-02 -2.92089105e-01 -1.12871394e-01
1.28314257e-01 -8.23790610e-01 7.26472259e-01 7.17817903e-01
9.44788098e-01 -1.00573838e+00 5.05024135e-01 6.28641546e-01
-1.37427151e+00 1.46961585e-01 -8.56626153e-01 2.37953544e-01
1.94919631e-01 1.01166737e+00 -4.19393480e-01 8.50028336e-01
9.03272390e-01 8.65375936e-01 -5.45821309e-01 1.29785001e+00
-1.40891299e-01 3.43012542e-01 -6.86407685e-01 2.16214925e-01
1.31127506e-01 -8.31949234e-01 6.13076389e-01 8.01174700e-01
6.32548153e-01 9.07909051e-02 4.44667041e-01 1.26770282e+00
1.56912699e-01 1.77063853e-01 -4.06360269e-01 2.92703956e-01
7.50386655e-01 1.40186417e+00 -1.07398629e+00 -3.84695381e-01
-1.96916431e-01 8.79169047e-01 1.42356068e-01 2.94164866e-01
-5.14869869e-01 -6.47305369e-01 5.02894700e-01 4.89973992e-01
2.32867941e-01 -4.08836186e-01 -5.70983410e-01 -1.18592834e+00
3.10647160e-01 -4.39649284e-01 1.54658239e-02 -8.59044135e-01
-1.38022542e+00 5.79748511e-01 1.32981325e-02 -1.66895807e+00
2.85593867e-01 -2.29654089e-01 -1.02961075e+00 8.93848479e-01
-1.41944802e+00 -1.12901306e+00 -7.48729587e-01 7.73035347e-01
4.78501976e-01 -6.21944666e-02 4.05165881e-01 2.91978180e-01
-5.67398250e-01 2.57414103e-01 1.24982201e-01 -1.01839058e-01
3.57744753e-01 -1.05400038e+00 6.73533559e-01 8.38897347e-01
-1.28358677e-01 7.31047988e-01 9.78571326e-02 -1.10148847e+00
-1.12519205e+00 -1.44445205e+00 4.42806393e-01 -2.94771791e-01
1.66054636e-01 -4.74750489e-01 -1.34587955e+00 4.60194498e-01
-4.72477138e-01 -1.20029757e-02 1.92252561e-01 -7.79741406e-02
8.82034302e-02 -1.36798844e-01 -1.16297829e+00 5.24082243e-01
1.27000153e+00 -2.15458851e-02 -5.49250543e-01 5.18006563e-01
9.20491874e-01 -2.42394865e-01 -7.94245362e-01 5.34322917e-01
-1.53691769e-01 -9.10327435e-01 8.28074932e-01 -1.13115735e-01
1.98514268e-01 -8.43085110e-01 4.39536184e-01 -1.26470923e+00
-6.80301964e-01 -7.78609812e-01 1.96085110e-01 1.49765599e+00
2.43221179e-01 -9.02966917e-01 9.36270356e-01 3.74113917e-01
-5.50791323e-01 -8.21029842e-01 -8.19542646e-01 -9.75247025e-01
8.93737674e-02 -1.99022204e-01 1.18430877e+00 1.20481205e+00
-4.16371882e-01 -3.62114385e-02 2.49441117e-01 2.18220115e-01
6.21187747e-01 3.30947280e-01 1.02797890e+00 -1.86777592e+00
2.85853803e-01 -5.06098688e-01 -3.56477916e-01 -1.08620656e+00
-2.28281721e-01 -1.03081298e+00 -5.20789474e-02 -1.88292873e+00
-2.29169503e-01 -1.22920370e+00 2.27218017e-01 6.38629496e-01
-1.90420091e-01 1.03389539e-01 2.14645758e-01 8.88143122e-01
-1.32379383e-01 8.69800985e-01 1.45106745e+00 1.81399614e-01
-6.26590967e-01 1.32819131e-01 -4.50009227e-01 1.08441877e+00
7.88698912e-01 -4.98995781e-01 -3.05144817e-01 -5.48756361e-01
1.00101151e-01 -4.38163906e-01 1.94217920e-01 -1.05853772e+00
2.86596030e-01 -2.79234529e-01 4.38199878e-01 -9.89791155e-01
2.51732200e-01 -9.92173553e-01 4.82681781e-01 3.55018020e-01
4.74603087e-01 1.41455740e-01 1.17920637e-02 4.33522075e-01
-3.28259200e-01 -1.37707621e-01 6.87998772e-01 -1.44374341e-01
-3.13949496e-01 8.20566356e-01 2.21695870e-01 -3.80487680e-01
1.19414043e+00 -8.16220045e-01 5.10148145e-02 5.76816984e-02
-6.19613588e-01 6.17225587e-01 1.00202620e+00 5.00086248e-01
9.03889894e-01 -1.67617214e+00 -7.47183800e-01 5.24592161e-01
-8.23819786e-02 1.17839396e+00 1.07639320e-01 7.90874839e-01
-9.69275773e-01 -6.69400468e-02 -4.63224389e-02 -1.10134828e+00
-1.00564957e+00 4.41304058e-01 7.24402517e-02 2.88700253e-01
-1.00947142e+00 7.49664724e-01 4.34181929e-01 -7.08152056e-01
-9.09135193e-02 -5.19590199e-01 -7.68464357e-02 9.49273109e-02
1.11775860e-01 4.93266702e-01 5.90998754e-02 -5.16813457e-01
-2.19297200e-01 1.15776420e+00 1.40428618e-01 1.02972880e-01
1.38355541e+00 8.80683959e-02 -4.73446101e-01 2.31156200e-01
1.02844238e+00 1.53139859e-01 -1.35752678e+00 -8.39787871e-02
-2.32780159e-01 -7.37869799e-01 4.26415093e-02 -7.29697421e-02
-1.34022522e+00 8.72869492e-01 3.28946680e-01 3.21255237e-01
1.02645445e+00 -1.14564210e-01 9.68031168e-01 2.94049317e-03
4.76245880e-01 -1.14560354e+00 3.72012518e-02 4.77294505e-01
1.06986380e+00 -7.34756291e-01 2.11422164e-02 -1.04150414e+00
-3.52267981e-01 1.16809261e+00 8.42304230e-01 -2.76436865e-01
5.54862738e-01 -1.31406173e-01 -9.30984989e-02 -2.85285860e-01
-2.83169478e-01 3.65479179e-02 2.30190784e-01 4.23115969e-01
-2.18190625e-01 6.28455309e-04 -1.70512185e-01 3.31382006e-01
-5.54069400e-01 -1.07381865e-01 1.82683855e-01 7.80527651e-01
-5.99700332e-01 -1.08913827e+00 -7.24485338e-01 5.53640723e-01
1.83908612e-01 1.00695260e-01 -3.01957935e-01 6.28444493e-01
4.65763986e-01 7.55474508e-01 5.72339475e-01 -3.78330857e-01
5.74118137e-01 -2.06749052e-01 -2.01523956e-02 -7.44781196e-01
-3.73837918e-01 3.90050977e-01 -5.01583576e-01 -7.01549470e-01
-1.96535259e-01 -6.55619204e-01 -1.79187822e+00 -3.95424932e-01
-8.23540628e-01 2.57517546e-01 8.24320793e-01 5.86417675e-01
7.98278570e-01 4.43568528e-01 8.64145219e-01 -1.00314271e+00
-8.12790543e-02 -8.11221778e-01 -4.19946223e-01 5.64509392e-01
-5.45686781e-02 -7.26830304e-01 -4.08968329e-01 -2.64351964e-02]
|
[8.293785095214844, -3.4513468742370605]
|
763efde5-f2a0-4b5a-b8cb-1be4d815df25
|
deep-relational-metric-learning
|
2108.10026
| null |
https://arxiv.org/abs/2108.10026v1
|
https://arxiv.org/pdf/2108.10026v1.pdf
|
Deep Relational Metric Learning
|
This paper presents a deep relational metric learning (DRML) framework for image clustering and retrieval. Most existing deep metric learning methods learn an embedding space with a general objective of increasing interclass distances and decreasing intraclass distances. However, the conventional losses of metric learning usually suppress intraclass variations which might be helpful to identify samples of unseen classes. To address this problem, we propose to adaptively learn an ensemble of features that characterizes an image from different aspects to model both interclass and intraclass distributions. We further employ a relational module to capture the correlations among each feature in the ensemble and construct a graph to represent an image. We then perform relational inference on the graph to integrate the ensemble and obtain a relation-aware embedding to measure the similarities. Extensive experiments on the widely-used CUB-200-2011, Cars196, and Stanford Online Products datasets demonstrate that our framework improves existing deep metric learning methods and achieves very competitive results.
|
['Jie zhou', 'Jiwen Lu', 'Borui Zhang', 'Wenzhao Zheng']
|
2021-08-23
| null |
http://openaccess.thecvf.com//content/ICCV2021/html/Zheng_Deep_Relational_Metric_Learning_ICCV_2021_paper.html
|
http://openaccess.thecvf.com//content/ICCV2021/papers/Zheng_Deep_Relational_Metric_Learning_ICCV_2021_paper.pdf
|
iccv-2021-1
|
['image-clustering']
|
['computer-vision']
|
[-2.12135881e-01 -3.87038946e-01 -4.22669500e-01 -1.05716765e+00
-6.90946043e-01 -4.70644534e-01 3.95288259e-01 2.17822969e-01
-1.33718729e-01 2.66442090e-01 7.83559680e-03 2.50511747e-02
-6.63043976e-01 -1.05167675e+00 -4.66412485e-01 -7.20257223e-01
2.28030272e-02 6.51896894e-01 -5.92123950e-03 1.10805854e-01
2.38286123e-01 5.80906630e-01 -1.60186279e+00 3.58967572e-01
7.66989946e-01 1.37181664e+00 1.25472574e-02 2.72802263e-01
5.30956825e-03 8.79202306e-01 -3.60495031e-01 -4.48379964e-01
3.90047938e-01 -4.20303106e-01 -6.60916567e-01 2.64579117e-01
4.49612558e-01 -5.21256447e-01 -8.26992631e-01 1.25697017e+00
3.22282314e-01 2.34159261e-01 8.68265569e-01 -1.49317491e+00
-1.21680772e+00 6.92116857e-01 -6.33448362e-01 1.09063290e-01
3.30248252e-02 -1.56292930e-01 1.48745906e+00 -7.58602202e-01
4.57164615e-01 1.30575120e+00 3.32946539e-01 2.19036832e-01
-1.24462116e+00 -6.55689657e-01 1.14597179e-01 7.24932671e-01
-1.81038749e+00 -2.32207007e-03 9.72393930e-01 -2.80108809e-01
5.70240200e-01 9.67937857e-02 4.01150733e-01 7.18227863e-01
-3.17010507e-02 8.44831645e-01 8.30623567e-01 2.64797918e-02
1.20001376e-01 1.03598997e-01 2.51856983e-01 9.22061682e-01
7.90340230e-02 1.12055592e-01 -4.62431550e-01 -4.36905175e-02
3.60325426e-01 4.02883083e-01 -8.85995179e-02 -9.37419593e-01
-1.07814884e+00 1.08970070e+00 9.50258613e-01 2.35166982e-01
-6.12902083e-03 2.07347259e-01 3.08509409e-01 5.78663170e-01
4.48940754e-01 2.89145112e-01 -1.22964166e-01 2.42783383e-01
-6.20269716e-01 1.75432786e-01 7.25709498e-01 9.98078048e-01
1.18882096e+00 -5.81764579e-01 -2.37224028e-01 1.10370553e+00
3.95140678e-01 2.72807389e-01 3.72037351e-01 -1.20166087e+00
4.01580513e-01 8.61617744e-01 -1.90607682e-01 -1.49047530e+00
-1.48355946e-01 -3.00119728e-01 -7.21101165e-01 -1.12238619e-02
1.11978389e-01 5.39817154e-01 -7.41962135e-01 1.59945822e+00
3.14513743e-01 1.87343299e-01 -3.54045063e-01 1.03224325e+00
6.34515703e-01 3.67080569e-01 -2.72040427e-01 2.72795528e-01
8.14231873e-01 -9.21266615e-01 -4.83055800e-01 1.90826491e-01
9.24838185e-01 -5.29357672e-01 1.04831994e+00 2.34623611e-01
-5.91327846e-01 -5.38826644e-01 -1.48052990e+00 -3.29188436e-01
-6.64468348e-01 1.09461807e-01 5.81441164e-01 5.54542124e-01
-9.23713088e-01 9.15388167e-01 -6.69157147e-01 -5.97592816e-02
7.31223762e-01 2.66923428e-01 -3.20026070e-01 -6.32326722e-01
-1.02901447e+00 5.56100130e-01 2.98963904e-01 -6.86370954e-03
-9.65231359e-01 -6.00402772e-01 -9.22707677e-01 1.05924100e-01
1.31530657e-01 -4.16663349e-01 7.22207189e-01 -6.30884945e-01
-1.01516712e+00 1.07701623e+00 1.37336090e-01 -3.75830859e-01
3.25806558e-01 1.38879031e-01 -4.80083942e-01 2.48167843e-01
3.52939256e-02 7.06992924e-01 6.13166213e-01 -1.18452561e+00
-6.31682217e-01 -6.13492727e-01 2.44326770e-01 1.90817654e-01
-6.57427013e-01 -2.34808236e-01 -3.51930976e-01 -5.36583126e-01
3.95444900e-01 -7.62281120e-01 3.29859070e-02 4.66691226e-01
-4.11059290e-01 -4.10071999e-01 1.00928676e+00 -3.08343530e-01
1.10070944e+00 -2.44826198e+00 3.04130107e-01 3.99680287e-01
5.67640841e-01 -9.36355069e-02 -5.43162048e-01 1.12317026e-01
-2.91237179e-02 1.42238848e-02 -1.50983348e-01 -3.70524466e-01
3.22488308e-01 5.17453074e-01 -4.52255607e-02 7.36556053e-01
1.89148173e-01 8.75812292e-01 -1.02195120e+00 -6.23603642e-01
3.57378602e-01 5.22527516e-01 -4.62080687e-01 2.38029420e-01
-4.09667147e-03 -3.27872597e-02 -4.54896927e-01 8.22194517e-01
7.94679105e-01 -2.74542779e-01 8.87695178e-02 -4.04727459e-01
5.65713167e-01 1.83922201e-01 -1.02584779e+00 1.98801959e+00
-1.60121813e-01 5.83051503e-01 -2.79675841e-01 -1.58376050e+00
1.17331755e+00 -2.59577334e-01 7.76159108e-01 -8.24569046e-01
1.10017471e-01 4.78759371e-02 2.05181409e-02 -2.45722473e-01
2.67464668e-01 2.10263491e-01 7.43565336e-02 5.89338362e-01
1.75111428e-01 -8.35493300e-03 8.29826966e-02 4.00175333e-01
1.05856693e+00 -1.72116295e-01 -1.72116131e-01 -2.71247596e-01
5.95263302e-01 -4.65383410e-01 5.27319729e-01 5.62050581e-01
-4.60995197e-01 8.16536427e-01 6.78986788e-01 -6.88981593e-01
-8.59460890e-01 -1.50235796e+00 -2.20180839e-01 9.53146935e-01
5.92422068e-01 -5.33053815e-01 -5.04219472e-01 -1.14356244e+00
3.27147394e-01 4.16529298e-01 -8.32936704e-01 -8.12335908e-01
-2.15742350e-01 -6.49346352e-01 1.96816102e-01 5.52044272e-01
4.66701418e-01 -5.61388314e-01 1.89188659e-01 -9.33279693e-02
-2.95778036e-01 -9.71236587e-01 -6.11002624e-01 1.50752231e-01
-6.62551463e-01 -1.35554111e+00 -1.17073245e-01 -9.46786821e-01
6.69961095e-01 4.96457040e-01 1.11858153e+00 1.02399237e-01
-5.09596586e-01 3.53978902e-01 -4.55419093e-01 2.26647146e-02
-1.30837649e-01 -1.12281907e-02 -9.36914682e-02 3.44924182e-01
9.09750402e-01 -4.38967854e-01 -7.56205320e-01 6.21730745e-01
-8.05196464e-01 -6.72025502e-01 2.95722604e-01 9.05117035e-01
8.33414614e-01 3.50317866e-01 4.67414290e-01 -6.50418878e-01
3.72772545e-01 -5.75182676e-01 -5.94525695e-01 5.69544256e-01
-8.28561366e-01 2.00951144e-01 3.09935719e-01 -4.12968844e-01
-5.37701011e-01 -5.27982861e-02 4.16649431e-01 -7.43286192e-01
2.79606402e-01 3.75214189e-01 -5.33592224e-01 -5.40193282e-02
4.60480958e-01 -6.62405044e-02 1.21218577e-01 -3.24119806e-01
6.65157735e-01 7.18526661e-01 3.63291264e-01 -7.10971177e-01
8.04371774e-01 6.94637895e-01 -4.33228258e-03 -2.39929825e-01
-1.08828104e+00 -6.05775297e-01 -8.21244895e-01 -8.15995932e-02
8.62845719e-01 -8.42123330e-01 -5.14142156e-01 1.86279982e-01
-7.16026306e-01 -1.43348694e-01 -4.59162652e-01 5.15088856e-01
-5.85164726e-01 3.00791293e-01 -5.67567945e-01 -2.48410448e-01
9.35822278e-02 -1.28151667e+00 1.23530006e+00 -2.85952240e-02
9.26204305e-03 -1.02137160e+00 1.99912228e-02 3.43963504e-01
8.89547095e-02 2.01963887e-01 1.13586652e+00 -6.05820417e-01
-7.29907095e-01 -3.76944691e-01 -6.17132425e-01 7.41779029e-01
3.62386018e-01 6.10827729e-02 -7.32007802e-01 -3.59856635e-01
-1.93565130e-01 -5.70401549e-01 1.13892734e+00 1.25265494e-01
1.80488944e+00 3.06571415e-03 -3.95313829e-01 9.13346708e-01
1.32022440e+00 -1.09760491e-02 6.33503258e-01 2.05238849e-01
8.73421311e-01 6.80338979e-01 7.49669492e-01 4.44317728e-01
5.21571398e-01 4.21226054e-01 6.70556724e-01 1.79357290e-01
-7.38119334e-03 -3.54150772e-01 5.53149506e-02 7.02192843e-01
4.48846519e-01 -1.49229333e-01 -6.34382367e-01 4.92261589e-01
-1.85237610e+00 -8.39259863e-01 3.31241071e-01 2.08112836e+00
6.77795708e-01 -3.74662131e-02 5.04065305e-03 1.64224103e-01
8.47490013e-01 3.91647488e-01 -7.30231106e-01 -2.06657648e-01
-2.45778695e-01 -3.12125683e-03 3.95871192e-01 3.33344549e-01
-1.26754642e+00 6.74570143e-01 5.85458326e+00 7.08817661e-01
-7.69444883e-01 -7.78082162e-02 8.42951536e-01 -2.71709025e-01
-4.94148552e-01 -1.41420946e-01 -4.42350268e-01 4.01588649e-01
6.21263206e-01 -3.22008491e-01 6.46597207e-01 8.43813360e-01
-3.17204833e-01 3.52450848e-01 -1.57647312e+00 1.28391576e+00
2.89905161e-01 -1.11700022e+00 3.59695286e-01 2.74016291e-01
6.93962097e-01 1.59773439e-01 4.38409686e-01 3.52596492e-01
5.05137265e-01 -1.09951019e+00 4.87409294e-01 6.54569566e-01
6.12650037e-01 -1.10828435e+00 5.74770629e-01 9.66038406e-02
-1.21090102e+00 -2.76332110e-01 -9.16705012e-01 3.00673634e-01
-3.62398297e-01 6.27732933e-01 -3.74879748e-01 5.14461577e-01
8.62033010e-01 1.40961957e+00 -8.53735983e-01 8.20381999e-01
-4.13466617e-02 1.04624018e-01 -5.54858372e-02 1.09088168e-01
2.83210754e-01 -5.56676030e-01 1.25341237e-01 6.66174233e-01
3.01743478e-01 -1.30514920e-01 2.32631102e-01 1.04437244e+00
-4.97931987e-01 -9.91669521e-02 -7.34580815e-01 -1.19607285e-01
6.21629357e-01 1.24456668e+00 -5.34669936e-01 -1.85855091e-01
-5.46981096e-01 1.11177886e+00 6.64642513e-01 3.60641569e-01
-9.69049156e-01 -6.47455931e-01 1.08990395e+00 -9.06511471e-02
4.46262330e-01 -2.00071037e-01 -1.25610724e-01 -1.11192536e+00
1.74949124e-01 -7.98254788e-01 7.76687205e-01 -5.09189188e-01
-1.65994120e+00 3.18522811e-01 -1.96651697e-01 -1.33095372e+00
-2.94739474e-02 -7.38107264e-01 -2.70581871e-01 5.34118533e-01
-1.46665370e+00 -1.05064964e+00 -3.96977842e-01 7.37033606e-01
5.02710342e-02 -3.03534359e-01 6.17598712e-01 4.79343623e-01
-4.91341412e-01 8.93627763e-01 4.65437710e-01 4.13962156e-01
8.09388995e-01 -1.39895570e+00 -1.05424151e-01 4.63566273e-01
5.42122424e-01 5.38022816e-01 1.00240476e-01 -1.60043344e-01
-1.26801252e+00 -1.36394417e+00 6.81479931e-01 -6.05960608e-01
8.77254009e-01 -4.85893011e-01 -8.40554476e-01 5.61577141e-01
-1.57872096e-01 6.23640001e-01 8.60478580e-01 2.09258795e-02
-1.20461047e+00 -7.20824897e-01 -1.25669336e+00 2.40240335e-01
1.33506584e+00 -1.02332354e+00 -2.75176466e-01 4.88116980e-01
8.76918137e-01 6.50896505e-02 -1.04392052e+00 3.98242474e-01
6.19300604e-01 -1.09729075e+00 1.01372766e+00 -6.05911195e-01
4.58463430e-01 -3.97073716e-01 -7.34782755e-01 -1.28763080e+00
-4.77839619e-01 -1.74818724e-01 -2.41797119e-01 1.27964544e+00
1.65630952e-01 -4.78783548e-01 7.85191000e-01 5.21590054e-01
8.75178799e-02 -7.60254562e-01 -8.77968967e-01 -9.46458876e-01
1.99779212e-01 -3.29088539e-01 9.90467310e-01 1.12408233e+00
-1.66887388e-01 1.27098396e-01 -4.10048142e-02 1.15930170e-01
9.17428672e-01 3.79754156e-01 4.33499694e-01 -1.24180377e+00
-1.78295985e-01 -5.73365688e-01 -1.08203089e+00 -9.20792460e-01
5.36063850e-01 -1.26635754e+00 -1.06611624e-01 -1.26736701e+00
5.18132806e-01 -4.85524714e-01 -1.01356769e+00 1.88061938e-01
1.81493033e-02 2.97415167e-01 -1.66830674e-01 1.27105430e-01
-1.07990420e+00 7.71048546e-01 1.02203429e+00 -7.19568968e-01
2.37361744e-01 -3.23776841e-01 -8.03995252e-01 2.98250496e-01
5.81346691e-01 -4.64478761e-01 -7.63596535e-01 -7.14726388e-01
1.43902436e-01 -5.27476072e-01 3.85619968e-01 -7.74153173e-01
2.34110817e-01 4.60126773e-02 4.88407016e-01 -7.05241740e-01
1.26627028e-01 -9.51358080e-01 -2.94495553e-01 2.20038265e-01
-6.15433335e-01 -9.21775103e-02 -4.99972135e-01 9.51890647e-01
-5.45186162e-01 1.04006395e-01 8.70697498e-01 2.28376180e-01
-5.35212874e-01 8.68094802e-01 3.56136084e-01 -1.12000033e-02
1.09814978e+00 -5.92813967e-03 -3.66921574e-01 -3.30765009e-01
-6.27463818e-01 5.32091260e-01 5.36024392e-01 7.23059297e-01
8.86442959e-01 -1.92105746e+00 -6.08170152e-01 1.36545211e-01
6.02405608e-01 -1.23089761e-01 6.28917292e-02 3.88505459e-01
-4.17319208e-01 1.15391895e-01 -1.30109712e-01 -8.07379544e-01
-9.38286245e-01 8.36985171e-01 5.74332356e-01 -1.50568202e-01
-2.99166322e-01 1.00689995e+00 3.22287410e-01 -6.38085067e-01
5.01067460e-01 -2.42182627e-01 -1.23724295e-02 2.47486636e-01
5.40012181e-01 4.87355292e-01 1.11318558e-01 -5.41181386e-01
-4.48675275e-01 4.50188726e-01 -3.44571203e-01 1.73547179e-01
1.23011816e+00 -2.31370464e-01 -2.98239142e-01 3.85028422e-01
2.12026906e+00 -7.07497776e-01 -1.20582736e+00 -6.19831681e-01
1.88441053e-01 -8.83206785e-01 1.19719341e-01 -3.26519102e-01
-1.43509042e+00 9.59587157e-01 7.86843061e-01 7.33772367e-02
1.18913007e+00 3.18845987e-01 7.28698432e-01 5.92870355e-01
2.66703397e-01 -1.23649454e+00 4.28086251e-01 1.94706410e-01
8.05449486e-01 -1.48289621e+00 7.50047863e-02 -1.19896866e-01
-2.71813124e-01 1.03496027e+00 6.27495110e-01 -2.76138276e-01
1.13426924e+00 -4.58023651e-03 -1.88188031e-01 -3.82116675e-01
-5.63135624e-01 -1.87044218e-01 5.11908591e-01 8.07751000e-01
2.81273901e-01 2.09095746e-01 -4.68964875e-02 2.27881372e-01
-6.16223812e-02 -4.51367408e-01 3.45424861e-02 6.13211274e-01
-1.99012175e-01 -1.18357110e+00 1.77677140e-01 6.30330741e-01
-1.83039811e-02 1.41507983e-01 -6.68882966e-01 6.65742695e-01
5.35806306e-02 9.74969983e-01 3.48691404e-01 -7.95642257e-01
7.48042464e-02 -2.36935183e-01 5.79639256e-01 -5.97900450e-01
3.33380117e-03 -3.52476597e-01 -3.32197934e-01 -9.96749699e-01
-4.27076191e-01 -7.50234246e-01 -9.92590666e-01 -3.08571219e-01
-2.00250074e-01 -3.79761122e-02 5.10074019e-01 6.78110063e-01
2.93692976e-01 2.17678472e-01 1.15844703e+00 -2.66307473e-01
-8.51467550e-01 -7.36620307e-01 -1.08094978e+00 8.22288036e-01
2.73887932e-01 -8.08700383e-01 -6.48167670e-01 -2.03641817e-01]
|
[9.42182731628418, 3.1103851795196533]
|
5ca0d601-9528-4e95-a4bd-f5b9c941e851
|
impact-of-the-reference-choice-on-scalp-eeg
|
1812.00794
| null |
http://arxiv.org/abs/1812.00794v2
|
http://arxiv.org/pdf/1812.00794v2.pdf
|
Impact of the reference choice on scalp EEG connectivity estimation
|
Several scalp EEG functional connectivity studies, mostly clinical, seem to
overlook the reference electrode impact. The subsequent interpretation of brain
connectivity is thus often biased by the choice a non-neutral reference. This
study aims at systematically investigating these effects. As EEG reference, we
examined: the vertex electrode (Cz), the digitally linked mastoids (DLM), the
average reference (AVE), and the Reference Electrode Standardization Technique
(REST). As a connectivity metric, we used the imaginary part of coherency. We
tested simulated and real data (eyes open resting state), by evaluating the
influence of electrode density, effect of head model accuracy in the REST
transformation, and impact on the characterization of the topology of
functional networks from graph analysis. Simulations demonstrated that REST
significantly reduced the distortion of connectivity patterns when compared to
AVE, Cz and DLM references. Moreover, the availability of high-density EEG
systems and an accurate knowledge of the head model are crucial elements to
improve REST performance. For real data, a systematic change of the spatial
pattern of functional connectivity depending on the chosen reference was also
observed. The distortion of connectivity patterns was larger for the Cz
reference, and progressively decreases when using the DLM, the AVE, the REST.
Strikingly, we also showed that network attributes derived from graph analysis,
i.e., node degree and local efficiency, are significantly influenced by the EEG
reference choice. Overall, this study highlights that significant differences
arise in scalp EEG functional connectivity and graph network properties, in
dependence of the chosen reference. We hope our study will convey the message
that caution should be taken when interpreting and comparing results obtained
from different laboratories when using different reference schemes.
|
[]
|
2019-01-24
| null | null | null | null |
['connectivity-estimation']
|
['graphs']
|
[ 2.38905451e-03 -1.24735892e-01 4.71472770e-01 1.58665646e-02
3.08454782e-02 -6.09542012e-01 5.65337300e-01 4.66894001e-01
-5.65004826e-01 7.98227727e-01 1.73533559e-01 -2.47928560e-01
-7.27714956e-01 -6.47841990e-01 -4.71672386e-01 -9.15556431e-01
-4.34198946e-01 1.49420723e-01 1.78195223e-01 -8.45176727e-02
3.54728669e-01 7.73571908e-01 -1.26489937e+00 -2.69625813e-01
8.43939006e-01 6.40301466e-01 2.05977216e-01 1.84434980e-01
2.20016390e-01 1.95789233e-01 -1.00409365e+00 -2.90556997e-01
4.84208986e-02 -5.86882174e-01 -4.27235782e-01 -2.41203338e-01
-2.51444250e-01 1.67565256e-01 -3.41111422e-02 9.41278398e-01
9.06920314e-01 -9.55124348e-02 6.86218560e-01 -1.15804088e+00
-6.22181557e-02 8.31006527e-01 -5.18266261e-01 6.46152377e-01
4.63423282e-01 1.67569950e-01 3.87961328e-01 -4.84125972e-01
7.87903547e-01 5.07630646e-01 6.87376678e-01 -7.96366259e-02
-1.51608932e+00 -7.61625648e-01 2.61818804e-02 3.53125185e-01
-2.00363684e+00 -3.66023868e-01 7.29467273e-01 -5.88595629e-01
7.65758634e-01 2.39184558e-01 1.27331400e+00 1.12065673e+00
7.50228047e-01 -5.96163332e-01 1.35665345e+00 -4.16229784e-01
2.74672121e-01 1.75784796e-01 1.28268495e-01 1.84307178e-03
9.11887527e-01 -2.11709544e-01 -3.53832722e-01 -3.98617625e-01
4.99440074e-01 -5.78663707e-01 -9.75361705e-01 -1.26846999e-01
-1.21347582e+00 4.24384356e-01 2.15072274e-01 1.01510298e+00
-5.42323172e-01 -1.44208953e-01 2.36512855e-01 2.20972210e-01
2.63287544e-01 6.97141171e-01 -1.33946702e-01 -2.95054495e-01
-1.02801609e+00 -2.55529523e-01 7.33951986e-01 3.54369551e-01
2.66175956e-01 -1.28550619e-01 -1.44071579e-01 5.62912464e-01
9.24546346e-02 2.96615571e-01 4.47451621e-01 -4.73962605e-01
2.66936123e-02 3.24997783e-01 -1.15876853e-01 -1.21960020e+00
-1.01248777e+00 -7.77317822e-01 -8.40084612e-01 4.51161675e-02
4.75730002e-01 -3.88710916e-01 -1.88774809e-01 1.72885334e+00
1.05098169e-02 -1.52744697e-02 -2.08603248e-01 8.40934753e-01
4.53325301e-01 7.31781311e-03 1.03561603e-01 -4.60874557e-01
1.19423521e+00 1.25485390e-01 -7.52613664e-01 1.16862640e-01
6.35665417e-01 -6.82959497e-01 7.05368280e-01 5.79313934e-01
-9.25822794e-01 -2.85466552e-01 -1.13928056e+00 6.50581360e-01
-2.91763812e-01 -1.76127180e-01 1.39999121e-01 8.38991344e-01
-1.35645998e+00 7.42003500e-01 -7.84908116e-01 -6.09464824e-01
-6.45382795e-04 4.46602672e-01 -6.65176988e-01 1.98964417e-01
-1.20363009e+00 1.32209849e+00 2.71491528e-01 3.73983026e-01
-2.36522555e-02 -6.93573117e-01 -3.29474241e-01 2.02670366e-01
-3.13670933e-01 -5.92026711e-01 3.05236667e-01 -8.55006635e-01
-1.32056701e+00 6.21801198e-01 1.66376844e-01 -3.62230122e-01
5.19909620e-01 4.02446657e-01 -5.72772264e-01 2.53517479e-01
-1.18563533e-01 2.86241591e-01 5.95816314e-01 -9.65786636e-01
2.23388642e-01 -4.77878749e-01 -3.11443239e-01 -2.02828106e-02
1.34741604e-01 -2.14104965e-01 1.46003991e-01 -3.02095711e-01
2.63324857e-01 -8.37763965e-01 1.82591558e-01 -5.76028287e-01
-2.23930880e-01 1.15834810e-01 1.21556543e-01 -5.22864282e-01
1.26315331e+00 -2.37172151e+00 1.08166747e-01 8.83778095e-01
4.64218855e-01 -2.08350837e-01 4.34734412e-02 6.29898548e-01
-4.76307034e-01 1.64969712e-01 -5.02403006e-02 4.37565297e-01
-2.04662919e-01 -3.59961361e-01 2.22566277e-01 1.19882631e+00
-7.29672471e-03 6.83852255e-01 -5.43391049e-01 -1.82393506e-01
1.78044364e-01 7.93049991e-01 -2.34230429e-01 -1.37803212e-01
6.67847037e-01 5.94101965e-01 1.97402090e-02 -1.05242655e-01
7.67922938e-01 -2.00083390e-01 4.36306506e-01 -3.79376739e-01
-4.78191525e-01 2.63487518e-01 -1.01783371e+00 1.46871686e+00
-1.69014320e-01 8.19237947e-01 -1.93614647e-01 -6.73052788e-01
1.12185216e+00 5.39354265e-01 3.42183679e-01 -9.41221893e-01
4.43230569e-01 2.75116295e-01 1.08668625e+00 -2.05324143e-01
-3.95425223e-02 8.76173377e-03 5.45796454e-01 3.47454071e-01
-4.61278260e-02 -1.20948218e-01 8.62686187e-02 1.26768723e-01
1.02778351e+00 -3.16446364e-01 3.71389627e-01 -1.02316225e+00
3.71211261e-01 -5.31133413e-01 -7.05284905e-03 4.52913284e-01
6.42274618e-02 7.73325324e-01 1.27493823e+00 1.47891849e-01
-3.59349251e-01 -1.00268805e+00 -7.46749103e-01 9.00054798e-02
2.72889763e-01 -3.68434995e-01 -9.19267356e-01 -1.00485489e-01
-9.82958302e-02 8.44240665e-01 -7.22986341e-01 -5.56508780e-01
-2.22108886e-01 -1.04214525e+00 4.72481191e-01 -1.28522441e-02
1.44701645e-01 -6.50647163e-01 -9.48883176e-01 1.31930918e-01
-1.02679901e-01 -8.66413593e-01 -7.54118562e-02 2.65179336e-01
-8.11369419e-01 -1.18328488e+00 -7.42571890e-01 -5.21157309e-02
6.92659557e-01 -1.02376072e-02 9.76371408e-01 3.06208223e-01
6.28879592e-02 6.03687167e-01 -3.40765387e-01 -3.04376572e-01
-7.23565891e-02 1.88404098e-01 -1.19670872e-02 7.34618232e-02
1.79201335e-01 -1.05703127e+00 -7.33142734e-01 3.84013474e-01
-6.83369458e-01 -2.51882046e-01 5.06157696e-01 1.90345734e-01
2.22795591e-01 -8.80083963e-02 7.59371221e-01 -5.23251712e-01
1.08418238e+00 -6.48984671e-01 -2.73390651e-01 1.33800626e-01
-7.72958457e-01 1.17686331e-01 2.79277325e-01 -3.75954449e-01
-5.42264223e-01 -6.66239142e-01 -7.84524307e-02 9.28789005e-02
-1.79092854e-01 5.40906847e-01 -1.93134934e-01 -3.36817652e-01
8.63562047e-01 3.07777748e-02 2.24251673e-02 -1.13197520e-01
-3.36332858e-01 5.59917152e-01 1.02981769e-01 -3.30288410e-01
1.28181905e-01 3.04484993e-01 1.24025688e-01 -1.08105719e+00
2.94469953e-01 -5.48585840e-02 -7.37249017e-01 -4.26433563e-01
8.43265355e-01 -5.09477854e-01 -7.27263570e-01 3.04871112e-01
-1.04018915e+00 -3.14493656e-01 -7.57561848e-02 8.71112227e-01
-1.68341137e-02 3.39774162e-01 -2.47225892e-02 -5.46951354e-01
-4.01915520e-01 -1.23645604e+00 5.92535019e-01 -1.58581629e-01
-6.94195747e-01 -1.06485081e+00 1.06959604e-01 -4.53620672e-01
5.29073358e-01 6.53047919e-01 1.11352515e+00 -5.66247344e-01
-1.37321100e-01 -1.74815223e-01 2.79489625e-02 -1.48318574e-01
2.69063652e-01 1.02734752e-01 -8.13108087e-01 -2.29672849e-01
1.60699844e-01 5.27319491e-01 3.52032214e-01 6.31132126e-01
6.01932943e-01 2.21462354e-01 -2.24326402e-01 3.74102592e-01
1.38376057e+00 4.14507419e-01 9.50512648e-01 2.12813199e-01
2.04594597e-01 5.60978711e-01 -1.90209910e-01 3.48522842e-01
6.33285269e-02 6.04141533e-01 2.33224481e-01 -2.71288585e-02
-1.44286230e-01 2.11685717e-01 1.71695337e-01 7.94478357e-01
-4.50016320e-01 -3.58762771e-01 -9.87410069e-01 3.30035418e-01
-9.91576433e-01 -5.55989146e-01 -5.67982316e-01 2.58539677e+00
3.31774622e-01 3.38181257e-01 1.36919260e-01 5.11008382e-01
7.54061043e-01 -1.64690971e-01 -1.64715260e-01 -3.54040772e-01
-2.86555767e-01 3.75004441e-01 4.07359630e-01 3.25550824e-01
-1.29668757e-01 1.57530718e-02 6.15313196e+00 3.02612841e-01
-1.68251872e+00 2.94906288e-01 2.57573545e-01 -2.96272308e-01
-3.38094980e-01 -1.09288089e-01 -1.91701323e-01 8.40138853e-01
1.27609253e+00 -4.29051787e-01 4.86776918e-01 5.90786664e-03
4.46759135e-01 -6.70450151e-01 -8.38275671e-01 1.00539088e+00
4.07430157e-02 -8.62775564e-01 -3.91170681e-01 3.37432474e-01
1.29559860e-01 1.92957103e-01 -2.90285289e-01 -3.56038928e-01
-7.10376263e-01 -8.97758842e-01 8.13592732e-01 8.90740395e-01
9.48916137e-01 -6.11820757e-01 9.95402455e-01 4.89561707e-02
-9.68164086e-01 2.94909865e-01 -1.11064062e-01 -1.45511732e-01
5.90153039e-02 8.21938992e-01 -4.84535784e-01 6.04551673e-01
4.52496976e-01 3.92716646e-01 -7.28382230e-01 1.13787460e+00
-1.61468655e-01 5.77490985e-01 -4.79878426e-01 3.68792042e-02
-1.65386185e-01 -5.20100772e-01 4.12301958e-01 1.03844452e+00
4.22152281e-01 7.99364373e-02 -8.09115112e-01 9.76354480e-01
3.52154911e-01 1.84075147e-01 -6.66497529e-01 7.72531405e-02
4.25783783e-01 1.28485882e+00 -1.33499813e+00 2.90693164e-01
-5.22027850e-01 6.49197876e-01 1.18172780e-01 5.90452075e-01
-7.07021356e-01 -4.99135047e-01 3.67604494e-01 7.84414411e-01
-4.07276563e-02 -1.39490709e-01 -4.60184336e-01 -8.62647176e-01
1.95834339e-01 -4.57705796e-01 -9.26396772e-02 -7.79457271e-01
-8.22669327e-01 9.78446424e-01 4.03578937e-01 -9.45121944e-01
-3.31164561e-02 -2.24867731e-01 -6.83138549e-01 1.22411728e+00
-9.72167492e-01 -2.24827871e-01 -2.84401238e-01 4.85957921e-01
-4.07965869e-01 3.89283508e-01 6.29708111e-01 2.59807974e-01
-4.95461166e-01 4.64602917e-01 -1.44485071e-01 -2.68919408e-01
5.21839738e-01 -8.39241922e-01 -9.83662829e-02 6.70917928e-01
7.75027499e-02 8.76541078e-01 6.99224710e-01 -4.89298940e-01
-1.01007867e+00 -4.56637949e-01 7.25029230e-01 -2.27640063e-01
5.41651309e-01 -5.14315307e-01 -8.55265915e-01 2.54967451e-01
4.76952761e-01 -3.05605114e-01 5.64435482e-01 3.05606909e-02
1.86353311e-01 -1.59619138e-01 -1.08790517e+00 5.12008190e-01
9.34594214e-01 -3.93358350e-01 -4.80478793e-01 1.16071943e-02
1.30086109e-01 1.05511606e-01 -1.24130118e+00 1.97122753e-01
8.18964660e-01 -1.32635748e+00 5.86696208e-01 2.86377102e-01
-1.58100322e-01 -6.06758744e-02 4.11779672e-01 -1.59428227e+00
-6.47795796e-01 -2.46605456e-01 4.97762531e-01 1.11924350e+00
5.22796154e-01 -1.28637326e+00 2.24856958e-01 5.96596539e-01
3.87006402e-02 -6.73416018e-01 -1.09515274e+00 -5.53582013e-01
4.32832303e-06 -2.20236689e-01 5.84294021e-01 6.86492145e-01
3.73637617e-01 3.14121544e-01 3.25728834e-01 1.11463713e-02
7.61184171e-02 -4.30542141e-01 1.95984617e-01 -1.52506793e+00
-6.46764860e-02 -6.65416241e-01 -7.88505018e-01 -5.88106550e-02
2.40797903e-02 -1.03065109e+00 -5.63985765e-01 -1.50284922e+00
-2.08343729e-01 -2.51785189e-01 -2.77518600e-01 -4.61840294e-02
4.67684083e-02 1.14176415e-01 1.26482978e-01 3.90799381e-02
2.38702878e-01 1.20012186e-01 1.02065313e+00 3.21150362e-01
-4.40760612e-01 -2.28367925e-01 -7.15293169e-01 5.23246229e-01
8.75019193e-01 -6.16144180e-01 -5.93870521e-01 -2.20257014e-01
4.42175478e-01 8.77516344e-02 2.06441864e-01 -1.24230361e+00
2.97151864e-01 4.73957330e-01 5.37777126e-01 -2.02780530e-01
1.33153081e-01 -9.75211918e-01 8.99072826e-01 6.25889838e-01
5.43851703e-02 4.55895633e-01 4.00844306e-01 8.09604824e-02
1.26978144e-01 -2.32325658e-01 6.31316245e-01 2.17132524e-01
1.57875940e-01 -2.39885092e-01 -6.41630054e-01 -6.04100265e-02
9.67215002e-01 -5.67183137e-01 -1.75251529e-01 -3.45427513e-01
-7.44657457e-01 -3.74603152e-01 4.72604275e-01 8.92281979e-02
3.19051951e-01 -9.35495794e-01 -7.16760635e-01 4.92534071e-01
-1.78263858e-01 -6.57040894e-01 4.53125983e-01 1.98163843e+00
-5.34133852e-01 3.69707137e-01 -4.04193729e-01 -4.89744633e-01
-9.55156028e-01 4.28143740e-01 5.59805036e-01 1.73339441e-01
-5.23578525e-01 4.26609933e-01 -1.77229252e-02 4.08660352e-01
4.52018268e-02 -5.68444550e-01 -3.56512249e-01 5.22822320e-01
2.34472975e-01 6.07488275e-01 6.37558639e-01 -5.80211759e-01
-6.50321782e-01 6.38031483e-01 4.40305322e-01 -1.54022917e-01
1.01995599e+00 -2.62091875e-01 -4.43316400e-01 6.21210754e-01
8.92651737e-01 2.41539612e-01 -6.36281431e-01 6.94747806e-01
-1.23406194e-01 -2.25453749e-01 -3.63220312e-02 -7.46234357e-01
-1.11940324e+00 8.17072749e-01 7.63577342e-01 5.10080338e-01
1.26059949e+00 -1.18223131e-01 -1.52183756e-01 -1.79533541e-01
7.54911363e-01 -5.79831719e-01 -7.79824913e-01 -8.90414789e-02
1.19258189e+00 -2.76167691e-01 1.08671613e-01 -2.30388150e-01
-4.37705278e-01 1.00589216e+00 1.65708318e-01 -2.17723995e-01
1.01677394e+00 3.33970487e-01 -2.33017236e-01 -3.74701113e-01
-4.30250764e-01 -1.15509499e-02 1.84952557e-01 5.65260231e-01
7.29648232e-01 7.71527039e-03 -1.06367779e+00 6.36444211e-01
-5.55793345e-01 -3.73424552e-02 6.94981873e-01 4.49914396e-01
9.11043212e-02 -9.09233212e-01 -4.74136949e-01 7.47618020e-01
-3.46570462e-01 -2.15622216e-01 -5.66212296e-01 1.15842366e+00
2.83816993e-01 9.81389880e-01 3.40696692e-01 -3.10738742e-01
5.55878699e-01 2.25313947e-01 7.55166411e-01 -2.81032443e-01
-8.84524822e-01 -1.03979714e-01 -1.14415325e-01 -3.38830590e-01
-3.19361061e-01 -7.69838035e-01 -9.97929215e-01 -2.84386337e-01
-6.23305738e-01 5.90571940e-01 7.55090594e-01 8.11264634e-01
6.48561180e-01 6.10162318e-01 7.64186606e-02 -8.06108177e-01
7.24077001e-02 -1.35408509e+00 -1.19332492e+00 4.97859437e-03
5.07453568e-02 -9.00127947e-01 -7.07468212e-01 -4.17627066e-01]
|
[12.982353210449219, 3.3796401023864746]
|
5cee1e63-dd38-4ec8-97b5-650a653f19c5
|
efficient-temporal-sequence-comparison-and
| null | null |
http://openaccess.thecvf.com/content_cvpr_2016/html/Zhang_Efficient_Temporal_Sequence_CVPR_2016_paper.html
|
http://openaccess.thecvf.com/content_cvpr_2016/papers/Zhang_Efficient_Temporal_Sequence_CVPR_2016_paper.pdf
|
Efficient Temporal Sequence Comparison and Classification Using Gram Matrix Embeddings on a Riemannian Manifold
|
In this paper we propose a new framework to compare and classify temporal sequences. The proposed approach captures the underlying dynamics of the data while avoiding expensive estimation procedures, making it suitable to process large numbers of sequences. The main idea is to first embed the sequences into a Riemannian manifold by using positive definite regularized Gram matrices of their Hankelets. The advantages of the this approach are: 1) it allows for using non-Euclidean similarity functions on the Positive Definite matrix manifold, which capture better the underlying geometry than directly comparing the sequences or their Hankel matrices; and 2) Gram matrices inherit desirable properties from the underlying Hankel matrices: their rank measure the complexity of the underlying dynamics, and the rank and the coefficients of the associated regressive models are invariant to affine transformations and varying initial conditions. The benefits of this approach are illustrated with extensive experiments in 3D action recognition using 3D joints sequences. In spite of its simplicity, the performance of this approach is competitive or better than using state-of-art approaches for this problem. Further, these results hold across a variety of metrics, supporting the idea that the improvement stems from the embedding itself, rather than from using one of these metrics.
|
['Yin Wang', 'Mario Sznaier', 'Xikang Zhang', 'Octavia Camps', 'Mengran Gou']
|
2016-06-01
| null | null | null |
cvpr-2016-6
|
['3d-human-action-recognition']
|
['computer-vision']
|
[-3.35283652e-02 -1.14941582e-01 -4.83720563e-02 -4.92014289e-02
-1.70271203e-01 -5.49299479e-01 9.12060261e-01 -4.25544567e-02
-5.65388799e-01 3.14106405e-01 2.50299305e-01 1.49779141e-01
-4.10641611e-01 -3.09695929e-01 -4.46167320e-01 -8.90492558e-01
-6.38168991e-01 3.70272309e-01 3.75972629e-01 -3.18599969e-01
3.49394202e-01 8.08721244e-01 -1.52028322e+00 -3.07433903e-01
4.76050943e-01 7.99888074e-01 6.30062073e-02 7.80411720e-01
2.30560109e-01 8.69707584e-01 -2.54007518e-01 -3.23478132e-01
3.97489488e-01 -5.39128959e-01 -7.77043343e-01 4.93525267e-01
2.68087059e-01 7.92643279e-02 -5.41722238e-01 1.00442243e+00
1.10481560e-01 3.46838921e-01 8.10928285e-01 -1.09924448e+00
-3.66063148e-01 9.35727730e-02 -3.16578060e-01 1.53568819e-01
5.32775760e-01 -1.46055102e-01 1.30773628e+00 -9.17051554e-01
7.19521403e-01 1.19585085e+00 6.71361029e-01 3.15297872e-01
-1.35763121e+00 2.11350732e-02 -8.23884085e-02 3.88118535e-01
-1.40846801e+00 -3.53962183e-01 6.84787035e-01 -7.90884793e-01
6.25919700e-01 4.35173512e-01 7.63525128e-01 7.80988514e-01
3.46259475e-01 8.27437162e-01 9.76812482e-01 -3.69277805e-01
8.17830265e-02 2.31393315e-02 2.11999312e-01 7.69122720e-01
1.66320816e-01 1.42423785e-03 -3.22096169e-01 -1.18288010e-01
7.98872292e-01 8.66707936e-02 -4.09725845e-01 -1.20608282e+00
-1.57982659e+00 7.81395197e-01 9.01865065e-02 8.00147474e-01
-3.28606009e-01 -6.55666506e-03 4.29634809e-01 2.70581454e-01
2.87896007e-01 3.12127769e-01 1.48193449e-01 -4.95338500e-01
-6.35710418e-01 1.59337148e-01 8.12556922e-01 6.51654840e-01
6.60612881e-01 8.11970904e-02 3.42031926e-01 5.48138618e-01
2.48963788e-01 4.09853905e-01 5.99720418e-01 -8.80075455e-01
5.50128579e-01 5.98559856e-01 -8.89573917e-02 -1.33171737e+00
-5.02506077e-01 -1.19017556e-01 -7.89761782e-01 3.57493043e-01
6.93128884e-01 3.80341977e-01 -3.10422480e-01 1.69048941e+00
2.49341339e-01 1.84844472e-02 -2.77914107e-02 8.55670094e-01
-9.93629396e-02 3.97381097e-01 -4.80638653e-01 -5.12627549e-02
1.07467628e+00 -6.07341707e-01 -7.22647190e-01 1.49630576e-01
6.57681584e-01 -7.59427428e-01 9.53580439e-01 2.26763308e-01
-1.00686681e+00 -5.41994512e-01 -1.23035550e+00 1.89889058e-01
-2.74698019e-01 2.31405482e-01 3.60372990e-01 5.78081429e-01
-1.08110416e+00 1.04662335e+00 -1.02567577e+00 -5.23860455e-01
-3.37180853e-01 3.22544694e-01 -6.82732821e-01 3.41027528e-01
-1.03086782e+00 1.10275149e+00 3.89247566e-01 1.77836880e-01
-4.59248215e-01 -2.17538953e-01 -9.28157270e-01 -1.45109877e-01
1.15336895e-01 -1.26946852e-01 7.65492141e-01 -8.13073874e-01
-1.68413115e+00 7.44151533e-01 2.55606882e-02 -2.95235515e-01
7.84529865e-01 -3.40489268e-01 -2.76366115e-01 4.44800168e-01
-2.05650121e-01 9.20690820e-02 9.62674081e-01 -7.76549935e-01
-2.46938020e-01 -5.03313661e-01 6.39535189e-02 2.36259252e-01
-3.91031355e-01 -3.52439582e-01 -2.68772542e-01 -7.44554698e-01
1.65142089e-01 -1.38291156e+00 -2.31485069e-01 -1.76962733e-01
-5.63548878e-02 -5.93940280e-02 8.15740287e-01 -7.69194782e-01
1.21897030e+00 -2.44549799e+00 9.73301947e-01 3.97117317e-01
4.80634607e-02 1.01818800e-01 -1.98456034e-01 7.70183623e-01
-3.67897868e-01 -1.70356467e-01 -2.93022215e-01 -8.90565589e-02
8.63985345e-03 2.86105454e-01 -1.59801483e-01 9.95801866e-01
1.88368440e-01 5.97887039e-01 -8.13433886e-01 -2.64643043e-01
4.98345137e-01 5.22844851e-01 -2.63747036e-01 7.47857336e-03
3.41005027e-01 4.45340067e-01 -4.17384237e-01 -3.50808352e-02
1.31176665e-01 3.94385569e-02 1.80772886e-01 -1.83336064e-01
-5.54073788e-02 2.47562051e-01 -1.68811715e+00 1.60745203e+00
-1.35509044e-01 6.58401847e-01 -1.05835937e-01 -1.28652763e+00
9.43695068e-01 3.64061803e-01 8.42120528e-01 -2.13398308e-01
3.58939990e-02 2.77097076e-01 1.84692740e-01 -3.51315111e-01
4.03986514e-01 1.20099250e-03 1.74980670e-01 5.53010941e-01
1.05077259e-01 -5.62940650e-02 5.23759484e-01 2.12261915e-01
9.71339703e-01 2.34543443e-01 3.59239370e-01 -5.84819674e-01
1.03215349e+00 -4.56630915e-01 3.25461268e-01 1.58183143e-01
4.02182676e-02 3.97973984e-01 6.23293698e-01 -2.95486212e-01
-1.12575090e+00 -1.07443392e+00 2.40234286e-02 3.32273543e-01
7.53302574e-02 -4.84135091e-01 -7.85935581e-01 -5.16742826e-01
-3.22235189e-02 2.40830287e-01 -8.18784595e-01 -2.39297882e-01
-5.82726002e-01 -4.21789646e-01 3.63113612e-01 3.82059515e-01
3.23270380e-01 -4.66385543e-01 -6.99560404e-01 1.92115307e-01
-3.09937865e-01 -1.16990519e+00 -7.42115259e-01 -1.16996251e-01
-1.38513041e+00 -1.31104553e+00 -8.46185088e-01 -4.82180119e-01
4.32263196e-01 3.47767323e-01 6.78626478e-01 -2.31601566e-01
-3.14742357e-01 8.53698075e-01 -3.93074930e-01 1.15783356e-01
-5.20410478e-01 -5.41245230e-02 4.30612653e-01 6.37133062e-01
7.27821663e-02 -6.78952634e-01 -2.83712983e-01 6.25170708e-01
-1.11671925e+00 -3.27078849e-01 5.24994016e-01 7.42947102e-01
2.86503255e-01 1.16536550e-01 2.80254334e-02 -2.48329863e-01
5.12739182e-01 -3.32125346e-03 -4.13423866e-01 1.38125151e-01
-5.21452427e-01 4.61777687e-01 4.26827610e-01 -6.00001812e-01
-5.42569101e-01 1.66975290e-01 2.99985617e-01 -4.25992697e-01
1.91988394e-01 2.60426134e-01 -8.74556676e-02 -8.58248919e-02
4.29449260e-01 2.90372431e-01 5.27337074e-01 -6.16637111e-01
3.62232625e-01 2.31705815e-01 2.66410440e-01 -2.85887361e-01
1.00993025e+00 7.42795825e-01 4.07354563e-01 -1.24722338e+00
-3.28700662e-01 -6.71972394e-01 -1.18956935e+00 -4.07181323e-01
1.05673015e+00 -5.21242499e-01 -7.13282108e-01 5.50999343e-01
-8.71990025e-01 -2.85156332e-02 -3.58771354e-01 9.67637599e-01
-9.37476218e-01 1.11340034e+00 -8.31886590e-01 -8.77650678e-01
1.81561515e-01 -9.88814235e-01 7.59892404e-01 -3.30075145e-01
-4.34689671e-01 -1.30736804e+00 3.74133468e-01 7.03998581e-02
9.92718339e-03 3.25580388e-01 8.57610762e-01 -4.21579659e-01
-3.97555649e-01 -5.29346585e-01 3.18246007e-01 7.97115624e-01
3.32079679e-01 1.27013355e-01 -4.31285113e-01 -4.10513312e-01
3.37244719e-01 1.74947247e-01 4.56052721e-01 5.47783114e-02
4.25456882e-01 -3.48219164e-02 6.22748211e-02 4.80036996e-02
1.21960700e+00 1.00969732e-01 6.23535275e-01 4.62950855e-01
6.92060888e-01 8.76534462e-01 6.66971684e-01 3.74696851e-01
2.51237378e-02 1.05411577e+00 3.73050898e-01 7.83850104e-02
2.07961455e-01 -1.13512687e-01 9.38747704e-01 1.33067274e+00
-4.69958365e-01 4.29842293e-01 -7.57191300e-01 4.90370184e-01
-2.03785634e+00 -1.10624623e+00 -3.49418581e-01 2.53376126e+00
4.10994917e-01 1.10877618e-01 5.32906294e-01 6.03497148e-01
6.72183573e-01 3.11343938e-01 -1.94490701e-01 -4.50988322e-01
-2.05851704e-01 1.74613334e-02 4.72259372e-01 5.61333358e-01
-1.03284860e+00 4.20656383e-01 6.17037773e+00 5.90490460e-01
-9.82775331e-01 -2.05717474e-01 -6.06759861e-02 1.64164260e-01
2.00283125e-01 -7.70089254e-02 -3.72340739e-01 2.35116318e-01
7.98255742e-01 -1.78880632e-01 4.56659079e-01 6.63174570e-01
2.53158391e-01 -1.00423515e-01 -1.41085243e+00 9.93537486e-01
3.05961221e-01 -9.83318985e-01 5.75169511e-02 5.51135063e-01
4.08537686e-01 -2.98794746e-01 6.43646568e-02 -1.33539468e-01
-7.32145384e-02 -6.07940674e-01 9.54705715e-01 7.15811610e-01
2.11286679e-01 -6.77078307e-01 5.89763045e-01 2.12836489e-01
-1.29373741e+00 1.50547773e-01 -3.52325946e-01 -1.45469546e-01
1.45408928e-01 4.57327932e-01 -5.36675572e-01 7.93051541e-01
4.22438532e-01 1.04610515e+00 -7.12058842e-01 8.67623985e-01
-1.91804357e-02 2.67090172e-01 -2.14859545e-01 -5.59194796e-02
4.03134078e-01 -9.34933186e-01 8.83404195e-01 1.09694469e+00
3.70240837e-01 -1.66585580e-01 -1.05896346e-01 4.36329782e-01
6.44596577e-01 4.59278554e-01 -7.53668427e-01 -4.34430629e-01
-2.89352387e-01 1.11515927e+00 -7.05675244e-01 -1.61987215e-01
-5.26909769e-01 1.10398996e+00 4.74071205e-02 2.04201847e-01
-6.94192231e-01 -3.63404393e-01 9.80166674e-01 1.23218261e-02
4.96787935e-01 -8.30952883e-01 4.59441282e-02 -1.37699819e+00
3.87803704e-01 -1.01154661e+00 3.17117095e-01 -4.16184485e-01
-8.58447313e-01 5.36217570e-01 7.58633614e-02 -1.69367468e+00
-5.32144547e-01 -8.83617759e-01 -2.75757700e-01 5.65860748e-01
-8.63493145e-01 -5.74206591e-01 9.69196111e-02 8.12356651e-01
2.43866220e-01 -2.15699766e-02 7.75739253e-01 3.01697314e-01
-3.52284610e-01 1.16964690e-01 3.83430928e-01 1.44568801e-01
4.21988189e-01 -1.35404110e+00 3.32419127e-01 9.28494692e-01
6.21647000e-01 6.24730051e-01 8.55266750e-01 -2.28272289e-01
-1.52183402e+00 -5.11556745e-01 8.57752740e-01 -5.53607106e-01
1.21604908e+00 -3.90618235e-01 -8.29566896e-01 5.40735066e-01
-1.17532454e-01 -4.70213890e-01 6.65919662e-01 4.50659096e-02
-3.46171647e-01 -6.60513937e-02 -5.82879305e-01 7.09952474e-01
9.93556261e-01 -5.72698295e-01 -6.60424411e-01 2.08806127e-01
2.91223396e-02 -3.43267322e-02 -1.04514933e+00 3.73145580e-01
6.53295159e-01 -1.10622537e+00 8.85594845e-01 -6.34436488e-01
6.64080605e-02 -5.39777756e-01 -2.93550342e-01 -1.25208735e+00
-4.50571537e-01 -6.77696407e-01 -1.88610449e-01 8.95639777e-01
2.19242752e-01 -5.64780354e-01 5.86455166e-01 3.06214511e-01
1.99629605e-01 -5.04072070e-01 -1.02093863e+00 -1.35860217e+00
-1.75930217e-01 -3.16562265e-01 1.07617691e-01 8.78156245e-01
2.05062076e-01 4.12794441e-01 -6.14603281e-01 -6.04024716e-02
6.30112410e-01 -4.13129516e-02 9.23201144e-01 -1.24687624e+00
-4.90679353e-01 -4.57591414e-01 -1.25829148e+00 -1.08195531e+00
6.88373446e-02 -8.51717353e-01 -2.06192464e-01 -9.68112171e-01
-1.09932862e-01 -1.13341786e-01 -1.77703708e-01 -1.18325930e-02
1.01905808e-01 1.31240264e-01 5.55630207e-01 4.12666291e-01
-3.94910634e-01 7.44170487e-01 1.09985709e+00 -1.56127661e-03
-1.05933912e-01 4.70325276e-02 6.92428201e-02 8.21666777e-01
5.03433406e-01 -4.43099290e-02 -3.74311715e-01 -1.47592336e-01
-3.02380091e-03 -4.46031280e-02 3.79248381e-01 -1.16224837e+00
-1.48991078e-01 1.11563995e-01 -1.20764293e-01 -3.18548292e-01
4.37382370e-01 -9.84108627e-01 4.86936599e-01 7.14397252e-01
-2.15585515e-01 3.85764599e-01 -8.06663036e-02 7.27617621e-01
-4.97383147e-01 -4.05579001e-01 9.14164186e-01 1.38864100e-01
-6.85536563e-01 2.45507415e-02 -6.70496881e-01 -4.89060581e-02
1.07416165e+00 -5.71726799e-01 4.13581908e-01 -6.29347146e-01
-9.39283073e-01 -2.57220775e-01 7.05540776e-01 7.00735748e-01
4.20714766e-01 -1.66335464e+00 -4.97627199e-01 5.48477210e-02
1.00460231e-01 -7.71799088e-01 -6.36405684e-03 1.36936474e+00
-5.32424510e-01 5.68385839e-01 -3.58652920e-01 -8.81486297e-01
-1.55989242e+00 5.86277664e-01 3.02685678e-01 -4.06431645e-01
-7.15662301e-01 2.57014006e-01 2.72519559e-01 -3.07283521e-01
2.03827918e-01 -3.19484025e-01 -1.84824795e-01 1.91174060e-01
4.99232888e-01 7.45021105e-01 5.07322550e-02 -1.09359062e+00
-3.91613424e-01 8.85426283e-01 2.17246026e-01 -3.95271003e-01
1.27200019e+00 -1.56897202e-01 -1.75021663e-01 9.64516461e-01
1.47763050e+00 -2.70364974e-02 -1.09258890e+00 -2.09576175e-01
4.64173406e-01 -5.04788816e-01 -3.63136202e-01 1.18748471e-01
-9.37828481e-01 9.52355027e-01 5.57013690e-01 3.77375007e-01
8.93911839e-01 -2.27821320e-01 3.77472848e-01 3.92891645e-01
3.77827108e-01 -1.10052466e+00 2.51608104e-01 5.86467087e-01
1.04233849e+00 -6.24505639e-01 -5.48213683e-02 -4.93090898e-01
-6.02950335e-01 1.36441052e+00 -1.06551431e-01 -4.36933011e-01
5.07033825e-01 -3.19136441e-01 4.56635803e-02 -9.28558111e-02
-3.45792174e-01 -3.31916004e-01 5.22477984e-01 4.09289867e-01
4.90664899e-01 -1.26395887e-02 -6.53173447e-01 -3.12358677e-01
-1.38536990e-01 -4.38287139e-01 6.18582904e-01 8.04338038e-01
-2.75719345e-01 -1.40136445e+00 -4.22831774e-01 -7.97249898e-02
-2.64525324e-01 3.92086357e-01 -6.52612567e-01 1.09393442e+00
-4.32320982e-01 8.33182335e-01 -2.62679309e-01 -5.16054571e-01
6.08033121e-01 1.94141567e-01 7.55358279e-01 -3.46927702e-01
-1.42204732e-01 1.91669464e-02 -1.75679289e-02 -9.08340633e-01
-7.03769803e-01 -1.31319404e+00 -8.52847040e-01 -4.11266722e-02
-2.51189142e-01 4.24288809e-01 6.11248851e-01 8.59681547e-01
1.90724328e-01 1.17628323e-02 8.39306831e-01 -8.71260405e-01
-8.84061992e-01 -8.23079646e-01 -1.02409291e+00 8.20979357e-01
1.43529713e-01 -1.02420962e+00 -5.33492982e-01 1.52920008e-01]
|
[7.747391223907471, 3.9204232692718506]
|
1d8d30e2-67e7-473d-b88a-0ec9f85a95a6
|
image-demoireing-with-learnable-bandpass
|
2004.00406
| null |
https://arxiv.org/abs/2004.00406v1
|
https://arxiv.org/pdf/2004.00406v1.pdf
|
Image Demoireing with Learnable Bandpass Filters
|
Image demoireing is a multi-faceted image restoration task involving both texture and color restoration. In this paper, we propose a novel multiscale bandpass convolutional neural network (MBCNN) to address this problem. As an end-to-end solution, MBCNN respectively solves the two sub-problems. For texture restoration, we propose a learnable bandpass filter (LBF) to learn the frequency prior for moire texture removal. For color restoration, we propose a two-step tone mapping strategy, which first applies a global tone mapping to correct for a global color shift, and then performs local fine tuning of the color per pixel. Through an ablation study, we demonstrate the effectiveness of the different components of MBCNN. Experimental results on two public datasets show that our method outperforms state-of-the-art methods by a large margin (more than 2dB in terms of PSNR).
|
['Ales Leonardis', 'Shanxin Yuan', 'Gregory Slabaugh', 'Bolun Zheng']
|
2020-04-01
|
image-demoireing-with-learnable-bandpass-1
|
http://openaccess.thecvf.com/content_CVPR_2020/html/Zheng_Image_Demoireing_with_Learnable_Bandpass_Filters_CVPR_2020_paper.html
|
http://openaccess.thecvf.com/content_CVPR_2020/papers/Zheng_Image_Demoireing_with_Learnable_Bandpass_Filters_CVPR_2020_paper.pdf
|
cvpr-2020-6
|
['tone-mapping']
|
['computer-vision']
|
[ 5.17260134e-01 -4.93037909e-01 2.03277215e-01 -6.53310418e-02
-1.00510228e+00 -2.57633269e-01 2.49250188e-01 -5.36834657e-01
-2.43365765e-01 6.26793265e-01 2.78174281e-01 -2.46220231e-01
3.49575393e-02 -8.68531168e-01 -9.87482488e-01 -9.44653988e-01
3.99353474e-01 -4.60734427e-01 3.30115169e-01 -3.28007579e-01
4.20663863e-01 4.00325716e-01 -1.37836587e+00 5.00454426e-01
1.13469601e+00 1.27349007e+00 2.43049443e-01 6.37285292e-01
1.59797490e-01 8.40562820e-01 -3.50852311e-01 -3.30916524e-01
5.07512629e-01 -3.41404259e-01 -6.19718254e-01 2.86954641e-01
8.67905319e-01 -6.58520281e-01 -5.88480890e-01 1.32952464e+00
4.75783110e-01 2.31217295e-01 3.55066657e-01 -6.81144774e-01
-1.15562773e+00 6.19353019e-02 -9.00735617e-01 1.57885514e-02
-1.10575808e-02 1.93123147e-01 7.36593187e-01 -1.14265251e+00
2.49553829e-01 1.32845771e+00 7.11282611e-01 3.14067215e-01
-1.41374838e+00 -6.63840652e-01 -7.06956163e-02 4.16826487e-01
-1.39181483e+00 -4.13243324e-01 7.51909435e-01 -6.34992793e-02
5.61205983e-01 2.86642015e-01 3.65097553e-01 7.72040188e-01
4.65699941e-01 3.82139713e-01 1.58231461e+00 -4.63429511e-01
4.90689725e-02 -4.84388202e-01 -3.81131053e-01 6.32658780e-01
9.19514522e-02 4.55998093e-01 -6.57069623e-01 9.20792073e-02
1.19801629e+00 -1.52301922e-01 -5.48587978e-01 1.72921509e-01
-1.04054105e+00 3.97882938e-01 5.50137520e-01 -1.77940160e-01
-8.58888850e-02 5.13015687e-01 8.22351277e-02 2.96446860e-01
8.71702552e-01 6.92387298e-02 -2.44166613e-01 2.31508106e-01
-7.18653560e-01 7.10347742e-02 3.79131496e-01 5.15829027e-01
9.06061113e-01 -5.26922904e-02 -4.34791535e-01 1.21148002e+00
2.23982304e-01 7.68163621e-01 -2.89640650e-02 -1.17644632e+00
1.38166934e-01 1.93602279e-01 2.90283293e-01 -1.04287302e+00
-8.01042095e-02 -2.18378425e-01 -1.16621149e+00 5.79583585e-01
1.88597307e-01 1.09799668e-01 -1.28507817e+00 1.40195870e+00
3.25150371e-01 5.71174622e-01 -2.12305680e-01 1.21169090e+00
7.76636243e-01 7.78242767e-01 -1.05484679e-01 -6.44774586e-02
1.43213785e+00 -1.05274510e+00 -7.26585031e-01 -1.99354202e-01
-2.59299368e-01 -1.26180887e+00 1.16717315e+00 6.64572001e-01
-1.02043211e+00 -6.07167840e-01 -1.18345737e+00 -5.24812877e-01
1.32480174e-01 3.64172369e-01 4.54312891e-01 6.93267822e-01
-1.22439265e+00 7.53232360e-01 -6.86345935e-01 -1.48964852e-01
3.57086539e-01 8.47226530e-02 -6.60520568e-02 -5.79947829e-01
-9.16950285e-01 5.94706535e-01 -2.23312508e-02 4.78724509e-01
-8.43133032e-01 -8.31655562e-01 -6.21711552e-01 2.42411718e-02
3.11036259e-01 -6.84621394e-01 8.96165848e-01 -9.19753134e-01
-1.94252110e+00 6.99174881e-01 -2.17684343e-01 -1.65155940e-02
4.88389492e-01 -3.92377973e-01 -5.60602903e-01 3.90361875e-01
8.89705643e-02 4.08621550e-01 1.24523592e+00 -1.46821415e+00
-6.31781816e-01 1.42903700e-01 -3.40959802e-02 1.48543835e-01
-3.11018318e-01 -3.32279243e-02 -1.01929772e+00 -1.13980138e+00
3.27158988e-01 -4.83119994e-01 -5.88055588e-02 3.51416171e-01
-5.81759334e-01 2.34861240e-01 7.43507087e-01 -1.02174282e+00
1.00840664e+00 -2.31120944e+00 -4.36589643e-02 1.15393996e-01
2.71610558e-01 -1.92509182e-02 -4.17118043e-01 1.40669495e-01
-2.98898946e-02 -6.19314350e-02 -5.46971023e-01 -4.35241401e-01
-1.80399299e-01 -1.22340865e-01 -4.77662295e-01 6.69133365e-01
2.31240869e-01 6.07938170e-01 -4.55024004e-01 -1.15403518e-01
3.12053412e-01 9.40832675e-01 -7.04024613e-01 8.02248940e-02
-5.46802916e-02 3.36615801e-01 5.36271073e-02 9.32831824e-01
1.34196615e+00 -2.40305021e-01 1.56543717e-01 -8.34178030e-01
-2.44502336e-01 -2.03455284e-01 -1.16973221e+00 1.60173142e+00
-4.83155549e-01 5.60641468e-01 1.75510049e-01 -5.81815660e-01
9.41738665e-01 3.37247271e-03 3.45539391e-01 -1.21489990e+00
-1.72868837e-02 3.79234105e-01 -5.08960485e-01 -3.38415802e-01
5.67045331e-01 -2.49609128e-01 2.00261891e-01 2.53148377e-01
-5.55477180e-02 -8.00724775e-02 1.08086355e-01 -3.09613142e-02
9.21552181e-01 1.91616774e-01 -2.66691595e-01 -4.55998361e-01
6.20546401e-01 -3.82720202e-01 7.48854339e-01 7.67854571e-01
-1.96613044e-01 1.16410363e+00 4.56350952e-01 -6.66015625e-01
-1.10858095e+00 -1.24131632e+00 -1.48644388e-01 1.01211107e+00
6.65071666e-01 -2.57989615e-01 -8.22873831e-01 -2.03741908e-01
-1.01275884e-01 1.43911451e-01 -6.49382830e-01 -1.67551428e-01
-6.67773008e-01 -1.06572425e+00 2.44322211e-01 1.41422555e-01
1.19898224e+00 -8.61251414e-01 -1.03333786e-01 1.74023032e-01
-5.56861460e-01 -1.20709050e+00 -8.88898849e-01 -2.43251324e-01
-5.64791143e-01 -1.10513425e+00 -7.71758318e-01 -7.44703054e-01
4.84209955e-01 6.95871055e-01 9.37716901e-01 3.42728645e-01
-4.90822375e-01 9.99416038e-02 -3.38243216e-01 2.50142187e-01
1.97935980e-02 -3.31030965e-01 -2.46909514e-01 6.33906424e-01
-1.69166028e-01 -7.10413516e-01 -1.24263239e+00 3.50070864e-01
-1.13982868e+00 2.49476105e-01 7.52282798e-01 9.56561446e-01
8.28745842e-01 1.48803592e-01 2.92204320e-01 -6.75959885e-01
6.05019450e-01 2.02053264e-02 -7.39829540e-01 3.95425826e-01
-5.97590744e-01 -2.40674287e-01 5.07595718e-01 -3.35145921e-01
-1.39040804e+00 -1.17494099e-01 -6.41350746e-02 -4.24896568e-01
8.17195922e-02 4.22879517e-01 -1.28692418e-01 -7.04447806e-01
5.89628696e-01 4.44903553e-01 -1.29279837e-01 -7.58151472e-01
4.08519804e-01 3.62648308e-01 9.59668636e-01 -6.67670190e-01
9.51587498e-01 8.04718018e-01 -3.45294476e-02 -6.79959595e-01
-9.00500000e-01 -4.05478440e-02 -1.77645013e-01 -3.74833226e-01
8.98637772e-01 -1.12066102e+00 -1.02105594e+00 9.07741547e-01
-9.44187403e-01 -7.38580406e-01 5.54096885e-02 1.35531291e-01
-4.44225669e-01 5.96549451e-01 -9.37871516e-01 -2.21329853e-01
-4.92545873e-01 -1.08460748e+00 1.04045498e+00 4.77086663e-01
6.55470669e-01 -6.52283728e-01 -9.27716047e-02 4.76959944e-01
8.61284435e-01 9.14475918e-02 8.60080838e-01 6.39970124e-01
-1.02481210e+00 2.56991178e-01 -9.12223697e-01 4.94671553e-01
2.58067459e-01 -1.26263991e-01 -1.06087351e+00 -5.66055298e-01
-7.56435990e-02 -2.07333311e-01 1.44311619e+00 5.67337632e-01
1.51590610e+00 -1.39580041e-01 1.28801465e-01 1.30749524e+00
1.83037543e+00 -8.12096670e-02 1.23266375e+00 5.94408512e-01
8.22624385e-01 7.06992596e-02 4.19132531e-01 3.68891865e-01
3.39689404e-01 5.96270680e-01 4.02982384e-01 -6.57203019e-01
-8.97107422e-01 2.85964347e-02 3.10383767e-01 4.67694163e-01
-1.36075586e-01 -1.42212242e-01 -6.63602412e-01 2.35699996e-01
-1.59456456e+00 -6.92292273e-01 4.95606521e-03 2.00993586e+00
9.62270617e-01 -1.87104419e-01 -3.94565552e-01 -1.09688550e-01
8.05846572e-01 3.56491953e-01 -5.79273760e-01 1.23773674e-02
-5.45759678e-01 5.08568823e-01 7.20946908e-01 7.86987305e-01
-1.29226899e+00 1.14564848e+00 6.52501202e+00 1.27001524e+00
-1.32520163e+00 6.83676973e-02 1.07238293e+00 1.31854787e-02
-3.30855638e-01 1.00288000e-02 -2.96158254e-01 4.07924354e-01
3.70263338e-01 2.57203043e-01 1.07430851e+00 9.19189900e-02
2.99603343e-01 -3.28985691e-01 -3.34266275e-01 1.01397014e+00
1.18963867e-01 -1.49344218e+00 9.34955701e-02 -1.51751667e-01
9.82961476e-01 -1.22793421e-01 4.04581606e-01 -9.98389348e-02
2.13981733e-01 -1.09678721e+00 7.55209446e-01 7.23832250e-01
1.32379913e+00 -9.37482536e-01 2.88653731e-01 -4.22607332e-01
-1.25644851e+00 -2.05267239e-02 -4.45117474e-01 3.10590148e-01
9.54388976e-02 9.70104992e-01 1.83003590e-01 6.59405529e-01
1.26530135e+00 7.98650861e-01 -4.82493550e-01 1.24511731e+00
-3.83611441e-01 6.11639321e-01 -6.75651878e-02 6.73681974e-01
-1.23691827e-01 -5.04876316e-01 3.49178582e-01 1.06970036e+00
4.00711924e-01 2.56231636e-01 1.22345343e-01 1.11571622e+00
-3.05561036e-01 -1.99878737e-01 2.51149237e-01 3.62761617e-01
3.11540067e-01 1.48227394e+00 -7.68514574e-01 -6.30678385e-02
-4.43537772e-01 1.48726869e+00 1.68134674e-01 9.03136551e-01
-6.56704485e-01 -6.67060316e-01 7.80643940e-01 -1.62287220e-01
4.47491735e-01 -3.83952893e-02 -4.13124174e-01 -1.14795935e+00
1.67195112e-01 -9.31305647e-01 -2.43687234e-03 -9.76546645e-01
-1.39900827e+00 4.40725625e-01 -6.12886906e-01 -1.24556971e+00
7.21068680e-01 -7.33657837e-01 -5.21212995e-01 1.17988825e+00
-2.07065892e+00 -1.40950167e+00 -6.43536866e-01 8.91454101e-01
2.16590196e-01 1.60041556e-01 4.86873537e-01 6.11104727e-01
-5.91592968e-01 3.91363770e-01 2.99241424e-01 1.63730800e-01
1.14620101e+00 -9.98696148e-01 3.64022732e-01 1.33040941e+00
-4.80203688e-01 3.69016856e-01 5.91769278e-01 -7.09658623e-01
-1.57900584e+00 -1.20707095e+00 3.75247866e-01 2.39125043e-01
4.34540868e-01 -2.75290817e-01 -9.11482275e-01 3.82347673e-01
3.11785340e-01 3.61459225e-01 3.15468937e-01 -1.82597518e-01
-7.85393119e-01 -5.65604210e-01 -1.00787282e+00 6.21416688e-01
7.85107195e-01 -6.73732996e-01 1.93431705e-01 2.33429611e-01
7.30381310e-01 -6.21021152e-01 -7.97574997e-01 5.32398701e-01
5.20100713e-01 -1.29590297e+00 1.13624620e+00 9.02841613e-02
8.74486744e-01 -7.34502196e-01 -4.32848364e-01 -1.23607206e+00
-5.58820546e-01 -7.47839987e-01 1.58056945e-01 9.55567896e-01
6.67288601e-02 -5.84843695e-01 4.94694471e-01 7.70469382e-02
-3.14208031e-01 -5.64389467e-01 -9.77469623e-01 -4.77753729e-01
4.91700310e-04 -1.53498530e-01 4.50153083e-01 8.23249221e-01
-5.95106125e-01 -9.39908624e-02 -8.35520983e-01 4.26845998e-01
1.07265985e+00 2.83527344e-01 4.59311336e-01 -7.79149413e-01
-2.46463135e-01 -4.04715449e-01 2.75341631e-03 -1.24234641e+00
-2.37155259e-01 -3.39836210e-01 3.57498527e-01 -1.47544527e+00
4.76582408e-01 -1.68639958e-01 -4.46009904e-01 3.65588576e-01
-4.18810636e-01 9.57096398e-01 7.44786859e-02 1.63233906e-01
-4.39213037e-01 7.37214446e-01 1.56080770e+00 -3.01289946e-01
8.97257775e-03 -4.12181139e-01 -9.71374691e-01 4.07368332e-01
5.49146533e-01 -1.69244394e-01 -1.17538825e-01 -7.83837736e-01
3.05569410e-01 -2.46012527e-02 7.39695370e-01 -1.08081019e+00
2.68105656e-01 -3.11501324e-01 6.76890790e-01 -4.09719199e-01
3.33411217e-01 -5.32735825e-01 -5.19673824e-02 1.59557462e-01
-3.44064459e-02 -3.95896912e-01 3.58676642e-01 7.17703342e-01
-3.39281350e-01 5.32319129e-01 1.22806871e+00 1.63195372e-01
-7.22560406e-01 4.97195721e-01 -2.92774439e-01 -2.21858010e-01
5.45181036e-01 -8.43855888e-02 -8.42148244e-01 -2.76979566e-01
-4.38809454e-01 -1.22679181e-01 4.71067935e-01 2.30818197e-01
9.00758982e-01 -1.48983908e+00 -8.36883485e-01 3.65055263e-01
-1.00444980e-01 -3.11295062e-01 7.70712376e-01 9.65716541e-01
-9.47528064e-01 -3.00092548e-01 -2.55388260e-01 -3.91801357e-01
-9.99514878e-01 1.87559634e-01 7.78070211e-01 9.89358649e-02
-9.58332837e-01 9.91006732e-01 3.20042908e-01 -3.12430084e-01
1.29162684e-01 -1.20706797e-01 1.69094503e-01 -4.01104271e-01
8.03213596e-01 2.00066566e-01 2.15314552e-01 -3.60218227e-01
-3.72813046e-02 9.22087193e-01 -1.92720979e-01 -1.29015952e-01
1.52372992e+00 -5.04347920e-01 -6.40723646e-01 -1.73702762e-01
1.10172117e+00 1.18604258e-01 -1.76563621e+00 -4.03384030e-01
-5.45343876e-01 -9.41133082e-01 5.17197013e-01 -1.16238403e+00
-1.55308247e+00 5.95480859e-01 9.47292686e-01 -9.26070362e-02
1.86390758e+00 -4.52505648e-01 8.88946056e-01 8.06393400e-02
-2.74113864e-02 -1.13887739e+00 4.38489020e-01 6.02161229e-01
1.00923228e+00 -1.03911185e+00 1.65947154e-01 -7.68285453e-01
-1.41534209e-01 1.29665959e+00 5.89456737e-01 -2.96328127e-01
7.24345744e-01 7.35077336e-02 4.38206226e-01 -3.98167633e-02
-4.06178236e-01 -6.44419566e-02 3.81161511e-01 4.14004296e-01
2.57387102e-01 -4.63113673e-02 -1.51341945e-01 1.99023724e-01
1.64356530e-01 -2.55543962e-02 5.51093698e-01 6.16550565e-01
-5.89143097e-01 -9.46164191e-01 -6.18419647e-01 1.04081459e-01
-4.99644309e-01 -5.27416468e-01 -2.35755816e-02 4.14179653e-01
5.39870486e-02 1.09751177e+00 2.03861739e-03 -5.41596949e-01
2.71696895e-01 -6.58415735e-01 4.74533498e-01 9.06461850e-02
-2.70154625e-01 6.02619290e-01 -2.27815330e-01 -1.03255916e+00
-3.62785310e-01 -1.86994970e-01 -7.99242973e-01 -6.61166906e-01
4.67437133e-02 -5.50819993e-01 3.43452752e-01 7.27421701e-01
3.71506393e-01 6.76399529e-01 7.71588326e-01 -1.01750147e+00
-1.42153233e-01 -8.13673019e-01 -8.93115759e-01 4.41281706e-01
6.52572691e-01 -4.75925386e-01 -4.10291195e-01 1.61959693e-01]
|
[11.07185173034668, -2.182785987854004]
|
c6b6654e-aacd-4036-b430-1d47a18d339d
|
enhance-nerf-multiple-performance-evaluation
|
2306.05303
| null |
https://arxiv.org/abs/2306.05303v1
|
https://arxiv.org/pdf/2306.05303v1.pdf
|
Enhance-NeRF: Multiple Performance Evaluation for Neural Radiance Fields
|
The quality of three-dimensional reconstruction is a key factor affecting the effectiveness of its application in areas such as virtual reality (VR) and augmented reality (AR) technologies. Neural Radiance Fields (NeRF) can generate realistic images from any viewpoint. It simultaneously reconstructs the shape, lighting, and materials of objects, and without surface defects, which breaks down the barrier between virtuality and reality. The potential spatial correspondences displayed by NeRF between reconstructed scenes and real-world scenes offer a wide range of practical applications possibilities. Despite significant progress in 3D reconstruction since NeRF were introduced, there remains considerable room for exploration and experimentation. NeRF-based models are susceptible to interference issues caused by colored "fog" noise. Additionally, they frequently encounter instabilities and failures while attempting to reconstruct unbounded scenes. Moreover, the model takes a significant amount of time to converge, making it even more challenging to use in such scenarios. Our approach, coined Enhance-NeRF, which adopts joint color to balance low and high reflectivity objects display, utilizes a decoding architecture with prior knowledge to improve recognition, and employs multi-layer performance evaluation mechanisms to enhance learning capacity. It achieves reconstruction of outdoor scenes within one hour under single-card condition. Based on experimental results, Enhance-NeRF partially enhances fitness capability and provides some support to outdoor scene reconstruction. The Enhance-NeRF method can be used as a plug-and-play component, making it easy to integrate with other NeRF-based models. The code is available at: https://github.com/TANQIanQ/Enhance-NeRF
|
['Baohua Zhang', 'Shuwan Yu', 'Yinling Xie', 'Tao Liu', 'Qianqiu Tan']
|
2023-06-08
| null | null | null | null |
['3d-reconstruction']
|
['computer-vision']
|
[ 1.58345670e-01 -3.36285442e-01 5.95298231e-01 -1.30078867e-01
-5.18776238e-01 -3.98476720e-01 3.00104827e-01 -5.22411108e-01
-1.55484691e-01 6.14691198e-01 -1.51277140e-01 -4.27615404e-01
3.10346838e-02 -8.92395973e-01 -8.33104908e-01 -6.37434542e-01
1.80283532e-01 -2.62998223e-01 3.12887788e-01 -4.89692122e-01
1.91403016e-01 7.51969397e-01 -2.00942755e+00 3.11975360e-01
8.29730630e-01 1.07749200e+00 9.66904461e-01 7.43997037e-01
-2.00021669e-01 8.01833153e-01 -6.41807497e-01 -3.03579718e-01
3.63509059e-01 -9.93708521e-02 -2.85927970e-02 -1.81308791e-01
2.61355221e-01 -5.17247736e-01 -5.07281363e-01 8.22734356e-01
7.41524935e-01 1.71175241e-01 2.62627691e-01 -9.19964492e-01
-6.98226452e-01 -2.65179694e-01 -5.37138581e-01 1.41462788e-01
8.09439182e-01 1.21247984e-01 3.06254923e-01 -1.07509232e+00
3.55047286e-01 8.81344497e-01 7.40478873e-01 5.78536451e-01
-9.79448020e-01 -6.66030228e-01 3.77731863e-03 -1.76980257e-01
-1.53053868e+00 -4.67519045e-01 7.40813255e-01 -1.25122651e-01
9.07201171e-01 7.75916100e-01 8.89511108e-01 1.01860785e+00
3.98419797e-01 4.87998039e-01 1.22248423e+00 -3.67592901e-01
1.91385016e-01 4.03555512e-01 -2.45771557e-01 7.15750396e-01
2.05891594e-01 3.37344199e-01 -5.12663186e-01 -8.59984290e-03
1.30211127e+00 2.62980256e-02 -7.73388863e-01 -2.31418312e-01
-8.11048567e-01 2.92578936e-01 5.18257320e-01 7.42587671e-02
-3.28284949e-01 1.33302391e-01 -2.93234855e-01 2.55292267e-01
2.84839720e-01 4.69015867e-01 3.91889885e-02 -1.82807565e-01
-6.40409887e-01 2.90934257e-02 3.70122820e-01 7.38076210e-01
4.80823815e-01 4.44999456e-01 2.54350483e-01 1.01202142e+00
4.06468540e-01 7.25725412e-01 2.28427753e-01 -1.12998223e+00
2.67030030e-01 2.61033356e-01 2.84836620e-01 -1.08603859e+00
-2.60056674e-01 -4.83720332e-01 -8.03034306e-01 5.57341039e-01
-8.04488063e-02 4.20020595e-02 -8.39537799e-01 1.43027163e+00
4.02275413e-01 4.78770673e-01 5.87195121e-02 1.43467772e+00
1.06826091e+00 9.39629018e-01 -4.41657126e-01 2.33096331e-02
1.01521349e+00 -5.28452098e-01 -5.24134219e-01 -2.29600400e-01
7.79070780e-02 -8.51578891e-01 1.37775469e+00 6.77936673e-01
-1.15628314e+00 -5.64102411e-01 -1.21839833e+00 1.21394597e-01
-1.95827290e-01 8.38234574e-02 9.28958833e-01 9.22211528e-01
-1.30308175e+00 3.45744163e-01 -5.33094049e-01 -1.15620352e-01
5.59821688e-02 2.94690758e-01 -3.69771272e-01 -3.98300022e-01
-9.38516438e-01 8.01546454e-01 -4.88523282e-02 6.01251721e-01
-5.73520005e-01 -5.65973938e-01 -6.48611963e-01 -1.42917961e-01
2.30266362e-01 -6.83622956e-01 8.58129323e-01 -7.90119469e-01
-1.92848921e+00 4.94647264e-01 6.08063154e-02 2.23717526e-01
5.67820191e-01 -2.54080474e-01 -7.17371285e-01 1.07338605e-02
-3.91512394e-01 4.60091352e-01 6.94545627e-01 -1.73500597e+00
-2.99631834e-01 -2.64242083e-01 2.07223788e-01 5.58452189e-01
-2.35642456e-02 -5.05067967e-02 -6.54304862e-01 -3.81738544e-01
4.61961418e-01 -6.06858611e-01 -3.03254098e-01 2.62688428e-01
-1.20587070e-02 5.76407969e-01 6.63149297e-01 -6.65740907e-01
8.67271006e-01 -2.08581543e+00 -3.04642498e-01 2.30533957e-01
1.67528346e-01 3.42454642e-01 -1.46339193e-01 2.49448955e-01
6.23934679e-02 -1.08267054e-01 3.10809836e-02 -2.50377595e-01
-2.87126392e-01 7.85726979e-02 -2.92012542e-01 4.58804220e-01
7.69709377e-03 6.02269053e-01 -5.62244236e-01 -1.03277573e-02
5.44154525e-01 1.07893217e+00 -4.63776886e-01 2.90296793e-01
8.75056833e-02 6.81269050e-01 -3.58108789e-01 6.32108271e-01
1.01969850e+00 -1.74663469e-01 1.34434015e-01 -4.54888009e-02
-3.46453011e-01 4.22913879e-02 -1.66513193e+00 1.54832304e+00
-8.22692394e-01 7.29975700e-01 2.25381061e-01 -5.40470064e-01
1.14191222e+00 1.48186535e-01 1.88953623e-01 -1.39514256e+00
2.27474377e-01 1.94827095e-01 -5.38215995e-01 -6.17528558e-01
1.05342150e+00 -5.23809008e-02 1.92643404e-01 1.39908075e-01
-4.07718956e-01 -1.66547745e-01 -4.39854264e-01 8.52532089e-02
9.49745953e-01 4.89794016e-01 -1.02363899e-01 1.41170636e-01
1.90231249e-01 -2.07984388e-01 4.67357159e-01 7.28222370e-01
2.80536324e-01 1.05757511e+00 -2.58114606e-01 -3.20687264e-01
-9.11432624e-01 -1.30135620e+00 -1.82074144e-01 6.27204120e-01
7.95298636e-01 -1.24576874e-01 -3.55146438e-01 7.85295218e-02
-2.60498941e-01 7.11461186e-01 -1.94108814e-01 -9.25230086e-02
-4.89609987e-01 -7.71423221e-01 2.89034665e-01 3.57848644e-01
5.65231025e-01 -7.63595283e-01 -9.67149675e-01 3.10157910e-02
-3.20955843e-01 -1.01236331e+00 3.05965189e-02 -4.70236503e-02
-7.86950529e-01 -8.69920969e-01 -8.31397891e-01 -3.59145463e-01
5.67429423e-01 9.00376379e-01 1.16878033e+00 2.10983694e-01
-4.48485106e-01 5.22187889e-01 -3.04435492e-01 -2.26039156e-01
-1.65150031e-01 -6.74822152e-01 -1.76642719e-03 -4.59596887e-02
-2.80722827e-01 -5.85877240e-01 -7.48789966e-01 6.22837782e-01
-8.14077258e-01 4.41595286e-01 2.88824290e-01 6.31877959e-01
5.47524333e-01 8.89212862e-02 2.80840486e-01 -4.88980621e-01
4.86041754e-01 -3.57985944e-01 -8.71843696e-01 2.19556987e-01
-3.38590026e-01 -3.21759313e-01 5.80146909e-01 -4.14926350e-01
-1.37880778e+00 -1.85887903e-01 -1.88089818e-01 -6.69907808e-01
-5.33175841e-02 2.64619231e-01 -2.76906312e-01 -3.42837334e-01
6.89309776e-01 1.36999056e-01 -1.98476970e-01 -4.92373317e-01
5.92012815e-02 8.13965619e-01 6.26708329e-01 -2.94746399e-01
6.87807143e-01 4.38350827e-01 -1.98830023e-01 -1.05935669e+00
-3.18267226e-01 -2.46982947e-01 7.64220208e-02 -8.08151901e-01
5.13278425e-01 -1.17842126e+00 -8.01920891e-01 4.58304942e-01
-1.04595280e+00 -5.26248217e-01 -1.13407560e-01 6.86700702e-01
-1.00428447e-01 1.74350411e-01 -3.57752353e-01 -1.34180772e+00
3.71235423e-03 -9.77854908e-01 9.84641075e-01 6.18209422e-01
3.14073771e-01 -6.97107732e-01 -1.62172467e-01 6.35232151e-01
5.53960860e-01 1.59139693e-01 4.45029646e-01 4.75180745e-01
-1.04789841e+00 -1.66683182e-01 -3.06230664e-01 1.68549925e-01
-1.62249386e-01 -1.68957427e-01 -1.34412515e+00 -3.37515235e-01
1.40750691e-01 -1.71129093e-01 6.86230659e-01 3.18574339e-01
1.06281519e+00 1.60259634e-01 -4.70491461e-02 9.15032625e-01
1.72558784e+00 4.08789456e-01 1.27796698e+00 3.28965575e-01
7.57738829e-01 4.31298226e-01 6.08474195e-01 4.88621116e-01
4.58423555e-01 9.12718236e-01 5.99358022e-01 -4.10277814e-01
-2.48337075e-01 -1.22067362e-01 3.58949989e-01 7.89644003e-01
-3.08390766e-01 -5.34543157e-01 -6.50695503e-01 -3.40867229e-02
-1.43113911e+00 -8.67246568e-01 -2.71879971e-01 2.51514506e+00
2.09557891e-01 -1.44224212e-01 -3.43887180e-01 6.05292767e-02
6.03751242e-01 2.67907064e-02 -5.47585964e-01 -4.21908885e-01
-3.37257534e-01 1.26280516e-01 4.42084670e-01 4.71184194e-01
-5.26709259e-01 5.88307798e-01 5.85105991e+00 5.42556703e-01
-1.53681695e+00 9.81626213e-02 5.10617733e-01 -1.50950402e-01
-6.57936990e-01 -1.97359145e-01 -4.06913847e-01 3.32029074e-01
6.85605526e-01 4.80639547e-01 8.24463427e-01 5.28367519e-01
3.70284557e-01 -5.15773773e-01 -4.85947132e-01 1.40294755e+00
2.35317841e-01 -1.28990245e+00 -3.01430494e-01 1.71959903e-02
5.94649971e-01 4.61243652e-02 2.78756320e-01 1.17681600e-01
1.44316852e-01 -1.05162132e+00 9.07544076e-01 7.63041317e-01
1.23466682e+00 -5.28014183e-01 6.38417244e-01 3.25616181e-01
-1.07940042e+00 -5.71146309e-02 -4.78549182e-01 -9.78820994e-02
2.43634135e-01 4.90635306e-01 -6.03092134e-01 6.88504517e-01
8.91928077e-01 -1.11641651e-02 -3.61799300e-01 1.20839119e+00
2.24925447e-02 2.70086497e-01 -3.76571864e-01 2.42929570e-02
-2.95708358e-01 -3.70938480e-01 5.55619061e-01 7.22432673e-01
6.81823552e-01 3.52005839e-01 -3.25163491e-02 7.71661282e-01
1.91646174e-01 -5.45335189e-02 -6.41689360e-01 3.62596244e-01
3.50210369e-01 1.05676210e+00 -7.37134576e-01 1.10528767e-01
-3.37593496e-01 1.14850569e+00 1.35215446e-01 5.58933914e-01
-1.09828377e+00 -1.87340662e-01 3.96176904e-01 3.58297050e-01
1.61447033e-01 -3.43490452e-01 -3.96499902e-01 -1.30044794e+00
3.34628016e-01 -7.38287270e-01 -2.34406248e-01 -1.52993810e+00
-8.63714099e-01 8.52262318e-01 -2.93898016e-01 -1.41961467e+00
9.84662548e-02 -5.47498763e-01 -2.94848025e-01 9.21233594e-01
-1.63453090e+00 -7.91905403e-01 -7.78236449e-01 5.73976696e-01
3.09482187e-01 1.74762290e-02 8.39356124e-01 5.95841229e-01
-5.23936391e-01 4.03483063e-01 2.89432377e-01 -5.41603982e-01
2.91841239e-01 -7.94426560e-01 3.05217445e-01 9.52842653e-01
2.29917601e-01 3.22796136e-01 6.59573257e-01 -5.85845172e-01
-1.85569906e+00 -8.33685637e-01 1.96578905e-01 -2.52637833e-01
-1.37376860e-01 -6.13741875e-01 -8.85486424e-01 5.51822744e-02
-1.78828806e-01 1.47860423e-01 5.81130505e-01 -1.63341478e-01
-2.12713093e-01 -1.68257803e-01 -1.30044353e+00 8.11525524e-01
1.22685575e+00 -4.42396671e-01 1.27292112e-01 2.17797793e-02
5.03649116e-01 -8.82606566e-01 -5.13106287e-01 2.48383224e-01
7.37715244e-01 -1.52682829e+00 1.06180167e+00 1.94031268e-01
2.01857492e-01 -5.05508900e-01 -5.60675681e-01 -1.13957465e+00
-1.93552941e-01 -6.00488782e-01 -4.73899171e-02 9.02576447e-01
3.17357421e-01 -8.31665576e-01 6.87278807e-01 7.55177677e-01
-2.78313100e-01 -6.40102088e-01 -7.71261692e-01 -6.02368891e-01
-5.67663729e-01 -7.64268577e-01 6.69476449e-01 7.99692333e-01
-6.25250578e-01 1.24489795e-02 -5.08587599e-01 6.45150661e-01
4.85667229e-01 6.29598647e-02 8.25526893e-01 -1.13970780e+00
-6.03158891e-01 -8.72917622e-02 -3.22653413e-01 -1.18401110e+00
-3.42615426e-01 -5.67776799e-01 1.61257744e-01 -1.75564373e+00
-1.39530674e-01 -8.87067199e-01 -1.98471814e-01 6.26010522e-02
3.18937339e-02 7.43105829e-01 4.15954411e-01 2.54285008e-01
-3.01183313e-01 6.21065080e-01 1.33130312e+00 2.40403399e-01
-4.27636832e-01 -3.66798788e-02 -6.78710580e-01 4.73114848e-01
8.24386597e-01 -1.36966020e-01 -5.38930595e-01 -8.14653635e-01
4.08913702e-01 3.99935544e-01 6.66273177e-01 -1.24316132e+00
1.15515634e-01 -5.68022281e-02 7.23191977e-01 -3.30244094e-01
9.38774884e-01 -9.64028180e-01 8.26206326e-01 8.53557605e-03
2.11373717e-01 1.53582901e-01 2.12826729e-01 4.19992507e-01
1.11480482e-01 -4.56460007e-02 6.68689132e-01 -1.64779082e-01
-8.33947599e-01 -8.87143016e-02 -2.99263388e-01 -3.51953447e-01
8.74065101e-01 -8.46280873e-01 -4.10517931e-01 -5.22589862e-01
-3.14351261e-01 -1.09172694e-01 6.96806014e-01 3.58759642e-01
1.05959058e+00 -1.09826398e+00 -3.91734153e-01 5.25347590e-01
-1.16693683e-01 4.10094581e-05 8.29132080e-01 4.98655528e-01
-8.44977438e-01 3.50449495e-02 -1.50146291e-01 -5.89310348e-01
-1.22799718e+00 2.71046072e-01 4.41766292e-01 1.03373803e-01
-9.14642513e-01 8.12625468e-01 2.08222285e-01 -4.92948234e-01
1.46905586e-01 2.15342771e-02 -7.60075524e-02 -4.88110870e-01
6.28919005e-01 3.48710418e-01 2.04030976e-01 -5.65389156e-01
-1.05743967e-01 5.22448838e-01 2.91538447e-01 -2.98247427e-01
1.38287294e+00 -3.51026207e-01 3.30513507e-01 3.75816107e-01
7.86493778e-01 1.26471341e-01 -1.42340505e+00 3.61272544e-02
-7.08969951e-01 -9.62193966e-01 5.82274914e-01 -9.98470724e-01
-1.24029934e+00 8.18016112e-01 1.00133598e+00 -1.47720221e-02
1.41993344e+00 -3.40119630e-01 6.60765946e-01 1.64543808e-01
8.02358985e-01 -8.26038361e-01 4.09840941e-02 2.29879513e-01
7.70530164e-01 -9.22850072e-01 -1.20281175e-01 -6.27935171e-01
-6.35920405e-01 8.44322503e-01 6.62709832e-01 1.10855252e-01
3.18170816e-01 5.52035272e-01 3.85648787e-01 -1.27519086e-01
-5.89392006e-01 2.90173367e-02 5.26816994e-02 7.65237212e-01
2.76976943e-01 3.50920595e-02 2.09099606e-01 4.03116196e-01
-2.37424552e-01 6.93777129e-02 6.86366796e-01 9.10136700e-01
-2.52352804e-01 -7.23778546e-01 -7.39745617e-01 2.92165548e-01
-1.94201872e-01 -7.92942196e-02 2.11969107e-01 5.36741078e-01
7.60008842e-02 9.59786654e-01 1.14935555e-01 -6.26147866e-01
5.31791210e-01 -4.77621943e-01 6.88862085e-01 -2.04046190e-01
-5.66048622e-01 1.36795744e-01 1.34173661e-01 -7.68456817e-01
-1.39402613e-01 -4.15651202e-01 -1.08340693e+00 -4.41088825e-01
-5.64035773e-01 -7.06761628e-02 1.06271696e+00 4.64790404e-01
6.01320982e-01 6.82477415e-01 8.11678290e-01 -1.04984260e+00
-5.91959096e-02 -3.30479503e-01 -7.25894392e-01 -1.06000140e-01
2.01734573e-01 -7.41173863e-01 -1.37406975e-01 -3.85628581e-01]
|
[9.738831520080566, -2.9175660610198975]
|
215dcfb8-b558-4c82-bd69-12acef4ec972
|
when-can-i-speak-predicting-initiation-points
|
2208.03812
| null |
https://arxiv.org/abs/2208.03812v1
|
https://arxiv.org/pdf/2208.03812v1.pdf
|
When can I Speak? Predicting initiation points for spoken dialogue agents
|
Current spoken dialogue systems initiate their turns after a long period of silence (700-1000ms), which leads to little real-time feedback, sluggish responses, and an overall stilted conversational flow. Humans typically respond within 200ms and successfully predicting initiation points in advance would allow spoken dialogue agents to do the same. In this work, we predict the lead-time to initiation using prosodic features from a pre-trained speech representation model (wav2vec 1.0) operating on user audio and word features from a pre-trained language model (GPT-2) operating on incremental transcriptions. To evaluate errors, we propose two metrics w.r.t. predicted and true lead times. We train and evaluate the models on the Switchboard Corpus and find that our method outperforms features from prior work on both metrics and vastly outperforms the common approach of waiting for 700ms of silence.
|
['Christopher D. Manning', 'Ashwin Paranjape', 'Siyan Li']
|
2022-08-07
| null |
https://aclanthology.org/2022.sigdial-1.22
|
https://aclanthology.org/2022.sigdial-1.22.pdf
|
sigdial-acl-2022-9
|
['spoken-dialogue-systems']
|
['speech']
|
[ 1.64871722e-01 5.44915557e-01 1.51209101e-01 -9.22038734e-01
-9.37680304e-01 -8.54403079e-01 8.44036818e-01 9.30278599e-02
-5.81962645e-01 8.52572262e-01 8.35332096e-01 -4.36149299e-01
4.47929889e-01 -2.16286212e-01 1.17341399e-01 -1.68206632e-01
-1.58647284e-01 6.89238369e-01 1.21855512e-01 -7.39287496e-01
4.87456053e-01 1.22307174e-01 -1.01265287e+00 5.38325667e-01
4.91544038e-01 7.20231414e-01 -1.47714764e-01 1.40154600e+00
-5.44081748e-01 1.23439062e+00 -1.16567159e+00 2.00427826e-02
-8.30959529e-02 -6.90906346e-01 -1.39024127e+00 6.22792616e-02
-7.28077516e-02 -5.53456604e-01 -4.29587245e-01 2.18184888e-01
5.99921465e-01 6.82452798e-01 4.14556921e-01 -1.11231911e+00
1.53871300e-02 8.77138138e-01 2.19048820e-02 3.26701313e-01
1.12700677e+00 3.80182505e-01 1.05837178e+00 -6.03353262e-01
4.58637416e-01 1.43394041e+00 5.91577113e-01 7.26022780e-01
-1.19560325e+00 -2.25246057e-01 1.05764717e-01 -2.30522856e-01
-1.00895894e+00 -1.02982664e+00 3.19841176e-01 -3.36179554e-01
1.74422097e+00 4.50018883e-01 3.12553763e-01 1.23754191e+00
1.16246767e-01 7.84960210e-01 7.78748870e-01 -5.06341577e-01
1.40895203e-01 2.45831534e-01 4.30533946e-01 3.39278877e-01
-1.07875657e+00 -8.52977112e-02 -8.13444316e-01 -2.72358507e-01
3.03448021e-01 -4.81833667e-01 -4.24611777e-01 6.82629347e-01
-1.06604993e+00 8.10695529e-01 3.66704427e-02 3.51701647e-01
-4.06342834e-01 -3.35240573e-01 7.29406238e-01 8.71036232e-01
3.53953540e-01 7.30856478e-01 -6.54975891e-01 -1.13315809e+00
-9.07500505e-01 2.41876423e-01 1.61503959e+00 7.95534372e-01
4.37148690e-01 1.10738426e-01 -3.47823799e-01 1.08499050e+00
7.40264431e-02 -7.52923489e-02 6.88765347e-01 -1.09581685e+00
5.59915602e-01 1.68453261e-01 3.63300771e-01 -5.72170973e-01
-7.55722880e-01 2.11772054e-01 -4.33670223e-01 -2.47096896e-01
6.96625769e-01 -7.65289545e-01 -4.23380136e-01 1.39507985e+00
5.49093932e-02 -1.12696841e-01 3.67245167e-01 7.13679314e-01
8.01050246e-01 1.28016126e+00 -2.69847363e-01 -5.08403301e-01
8.66103768e-01 -1.30707538e+00 -8.96333516e-01 -3.53493184e-01
7.35279441e-01 -9.04109120e-01 1.26859391e+00 6.48379028e-01
-1.12308860e+00 -5.37119865e-01 -7.37089574e-01 6.76314980e-02
8.77210572e-02 -4.11329448e-01 3.25721592e-01 5.43805122e-01
-1.27359354e+00 6.13663852e-01 -6.33629441e-01 -4.16820079e-01
-5.10728419e-01 4.56272691e-01 -2.76367158e-01 2.13044479e-01
-1.14754581e+00 1.01156366e+00 -3.48567180e-02 1.98234931e-01
-4.64790016e-01 -4.61923689e-01 -7.85111725e-01 -7.37329945e-04
2.42999092e-01 9.36682001e-02 2.17956948e+00 -8.19394410e-01
-2.44105387e+00 3.92282516e-01 -3.01635712e-01 -5.27803063e-01
4.56277251e-01 -4.90677774e-01 -2.82290071e-01 1.06209219e-02
-1.99280858e-01 5.61908662e-01 4.45118964e-01 -8.69339168e-01
-6.62992060e-01 1.32218510e-01 -6.38992786e-02 4.23492193e-01
7.03622326e-02 3.13817233e-01 -1.47928447e-01 -1.15373828e-01
-2.69288719e-01 -8.09999108e-01 -3.27432454e-01 -9.13684130e-01
-4.68374819e-01 -4.23517078e-01 3.99609476e-01 -7.40769327e-01
1.55860853e+00 -1.89747894e+00 3.15265171e-03 3.27053592e-02
-5.57595417e-02 1.94681048e-01 -3.20202380e-01 8.90612185e-01
2.36690804e-01 1.86186716e-01 1.36206582e-01 -4.33514655e-01
-1.99841280e-02 3.24665487e-01 -3.30691397e-01 1.42533839e-01
1.86183766e-01 6.70771062e-01 -9.10664380e-01 -1.20421648e-01
2.11044684e-01 2.47197405e-01 -5.23499668e-01 8.87572527e-01
-1.03337124e-01 6.54596031e-01 -3.58549058e-02 1.59373075e-01
2.23536976e-03 8.39619637e-02 2.50785142e-01 4.91585284e-01
-5.22578597e-01 1.17409968e+00 -6.94774687e-01 1.39215553e+00
-7.40503132e-01 7.75312364e-01 1.90805271e-01 -5.86790264e-01
1.11340189e+00 7.32790887e-01 2.16634154e-01 -6.44072354e-01
2.50296980e-01 -5.62277958e-02 2.74044782e-01 -4.24738079e-01
5.82155228e-01 4.49495576e-02 -3.64062816e-01 8.20817053e-01
2.15775087e-01 -3.78467113e-01 -1.02938183e-01 3.17784965e-01
1.40715563e+00 -4.87703025e-01 3.31196278e-01 1.50439754e-01
5.48199415e-01 -1.61203980e-01 2.89677292e-01 9.17786479e-01
-4.22282279e-01 5.35590470e-01 7.28690386e-01 -5.40493488e-01
-8.05404842e-01 -5.87671041e-01 3.95225108e-01 1.64728951e+00
-4.84444171e-01 -6.23645961e-01 -6.26896501e-01 -6.36344552e-01
-3.46602380e-01 1.08494258e+00 -3.28475446e-01 8.38011280e-02
-7.74686575e-01 -2.65093576e-02 6.31661296e-01 3.77077460e-01
-4.00513820e-02 -1.43877423e+00 -5.00344634e-01 7.47128010e-01
-3.26454043e-01 -1.03395450e+00 -8.48421812e-01 2.61806309e-01
-3.61237913e-01 -5.95373750e-01 -4.41981047e-01 -4.90085572e-01
4.08256762e-02 4.32179011e-02 1.23358953e+00 5.21804951e-03
4.49935580e-03 2.99783796e-01 -7.24800050e-01 -2.21093461e-01
-8.76651168e-01 1.24824628e-01 1.98806718e-01 -1.89503327e-01
3.64911079e-01 -2.59732842e-01 -1.39521077e-01 4.10096973e-01
-3.26394141e-01 -6.18201941e-02 7.15321228e-02 1.04364371e+00
-2.29720369e-01 -5.20032883e-01 9.90849018e-01 -7.44103551e-01
1.25609541e+00 -3.89384687e-01 -2.43701354e-01 -1.10372035e-02
-5.29537737e-01 -7.17746764e-02 6.52278066e-01 -4.02363837e-01
-1.08023274e+00 -1.46987528e-01 -5.63879788e-01 1.70983940e-01
-5.19559085e-01 4.74163830e-01 2.37110808e-01 3.79694551e-01
6.46919727e-01 1.31226137e-01 2.40910560e-01 -3.01731080e-01
4.50475603e-01 1.25367939e+00 4.83094037e-01 -1.43070370e-01
8.11975151e-02 -3.98695141e-01 -1.02260888e+00 -1.21053159e+00
-6.56032860e-01 -7.62541592e-01 -5.17340302e-01 -3.25688094e-01
4.33369786e-01 -6.12801433e-01 -9.23171282e-01 3.18001121e-01
-1.44054353e+00 -8.79481137e-01 3.72448415e-02 4.02877718e-01
-7.37231374e-01 2.61704057e-01 -1.02348721e+00 -1.33856106e+00
-4.66929317e-01 -9.37350094e-01 6.85590386e-01 2.83098161e-01
-1.10471940e+00 -8.05892766e-01 2.86279261e-01 4.25310224e-01
5.92857301e-01 -3.29332590e-01 3.69320750e-01 -1.32183003e+00
1.03304036e-01 -2.56077439e-01 2.39201427e-01 3.80586237e-01
3.16879034e-01 7.20075890e-02 -1.16203916e+00 -8.96996334e-02
-1.05424985e-01 -6.60600245e-01 4.23098862e-01 1.97674572e-01
5.21349609e-01 -8.09806824e-01 9.22331885e-02 -1.06394403e-02
5.96678376e-01 6.52749419e-01 2.92695135e-01 -1.71104039e-03
1.85042262e-01 9.43854034e-01 7.26566911e-01 7.41056144e-01
4.78737831e-01 5.33751786e-01 -7.00796809e-05 2.65749723e-01
4.71535474e-01 -1.50550067e-01 6.82511032e-01 1.27954066e+00
2.57631510e-01 -5.73861003e-01 -1.09924006e+00 5.76542735e-01
-1.72373819e+00 -1.00268567e+00 3.38997319e-02 2.04173660e+00
1.24203527e+00 4.99185860e-01 4.62102741e-01 1.26643598e-01
3.05494547e-01 2.83554077e-01 -1.34996861e-01 -1.40524650e+00
2.91019261e-01 2.82236814e-01 5.20846508e-02 1.25959635e+00
-8.63678157e-01 1.17190719e+00 7.09198713e+00 1.41751364e-01
-1.26146185e+00 -1.70596004e-01 6.45160198e-01 -1.85562521e-01
-4.70430404e-02 -1.66716650e-01 -7.00079978e-01 2.65882194e-01
1.78953385e+00 -4.09312189e-01 6.31064117e-01 5.53952873e-01
6.73726261e-01 -1.07458964e-01 -1.29709983e+00 6.54587090e-01
-2.18145642e-02 -1.04184210e+00 -5.28450489e-01 -2.50685632e-01
1.90826863e-01 1.63939774e-01 -3.58274221e-01 8.32136810e-01
5.22835851e-01 -1.22874594e+00 2.68297195e-01 4.77876186e-01
5.38635135e-01 -7.86015332e-01 8.61848652e-01 7.18462229e-01
-5.74294627e-01 -1.15314275e-02 -3.67878075e-03 -4.56663311e-01
4.86919910e-01 -8.71858373e-02 -1.86749279e+00 8.79725665e-02
2.64685839e-01 2.91168779e-01 6.81831837e-02 6.12806439e-01
-1.04016066e-01 1.13018572e+00 -3.62038165e-01 -3.22472572e-01
5.55009246e-01 8.48718137e-02 3.47972095e-01 1.59578586e+00
-3.71183641e-02 3.76278132e-01 4.73481417e-01 1.81161240e-01
1.00097820e-01 2.67871499e-01 -5.74125528e-01 -2.93867797e-01
7.46192336e-01 1.03002858e+00 -2.42217690e-01 -3.64056736e-01
-5.63818991e-01 1.06256974e+00 7.03124106e-02 3.07005644e-01
-4.26947892e-01 -5.28570056e-01 6.29399300e-01 -3.09086919e-01
-7.99029991e-02 -3.17805946e-01 -7.57708773e-02 -8.22633803e-01
-3.63035053e-01 -1.05549657e+00 -2.36325059e-02 -4.67034578e-01
-1.22851670e+00 1.06545687e+00 -4.43154871e-01 -7.12093234e-01
-1.13422990e+00 -1.38698667e-01 -1.01331782e+00 1.13430667e+00
-1.08734143e+00 -4.52152610e-01 3.65332030e-02 1.85236037e-01
1.28296626e+00 -1.33264959e-01 1.32742465e+00 -1.61117211e-01
-4.71079171e-01 6.92430913e-01 -2.76342779e-01 2.69643605e-01
1.07379842e+00 -1.32284141e+00 7.85277843e-01 3.63177694e-02
5.89903966e-02 4.80048686e-01 1.01746094e+00 -2.40031764e-01
-1.06448615e+00 -5.42605102e-01 1.28157794e+00 -5.44494450e-01
7.54223406e-01 -2.20084071e-01 -1.17483854e+00 6.87550962e-01
8.11999023e-01 -6.01370931e-01 1.14852309e+00 5.48821270e-01
-1.15993373e-01 2.28903785e-01 -6.63636804e-01 4.58138049e-01
3.57044727e-01 -5.79791486e-01 -9.32483494e-01 1.62885442e-01
8.73191893e-01 -3.92070413e-01 -8.11533034e-01 -1.87482983e-01
7.38775849e-01 -1.02617419e+00 3.59881282e-01 -7.82516956e-01
1.65010720e-01 4.10870284e-01 -3.83917652e-02 -1.52386773e+00
1.24294765e-01 -1.48825276e+00 1.10601261e-01 1.16497111e+00
7.67904460e-01 -4.46802497e-01 4.62450266e-01 1.01528132e+00
-2.75242716e-01 -7.06311882e-01 -7.48511374e-01 -4.66365606e-01
-8.53160322e-02 -3.76695931e-01 3.60542387e-01 6.87685907e-01
7.66787946e-01 9.90147352e-01 -4.42665398e-01 -1.52610034e-01
-2.18804479e-01 -2.28546470e-01 1.03427541e+00 -1.01364958e+00
-1.93457097e-01 -5.79210579e-01 -3.40360925e-02 -1.47882676e+00
1.66530997e-01 -2.59547114e-01 6.31113470e-01 -1.27740705e+00
-4.46534425e-01 -2.39758834e-01 -1.33710235e-01 4.58088130e-01
-8.25167298e-02 -4.54191357e-01 2.06792191e-01 -1.25755653e-01
-6.83989644e-01 5.51314235e-01 7.01195896e-01 -2.96834279e-02
-9.34251189e-01 3.88199061e-01 -4.08409446e-01 5.84199905e-01
9.66758549e-01 -9.60053056e-02 -3.57980549e-01 -1.41853318e-01
-2.07519323e-01 8.41422319e-01 -2.63233453e-01 -6.31902754e-01
4.52782422e-01 -1.22714333e-01 -2.46216148e-01 -4.65938598e-01
6.06439173e-01 -7.09268078e-02 -6.54161870e-01 -9.53277871e-02
-9.92669225e-01 1.03681423e-02 2.54338056e-01 3.39012176e-01
-4.60953653e-01 -1.97206616e-01 4.54672903e-01 -1.31478608e-01
-4.34096962e-01 -2.23523453e-01 -1.18080652e+00 1.52037367e-01
5.19108951e-01 -1.87855259e-01 -7.94950277e-02 -1.18558526e+00
-8.51425350e-01 5.31599224e-01 5.70839383e-02 6.87209368e-01
5.97456574e-01 -7.25754499e-01 -7.27906525e-01 2.47244045e-01
-2.41342589e-01 -1.88210279e-01 7.94006810e-02 6.80020630e-01
-4.74396586e-01 7.13249207e-01 -5.81159256e-03 -5.10435760e-01
-1.44918048e+00 -8.32410902e-02 2.46003628e-01 -9.64598730e-02
-4.73636985e-01 1.26026082e+00 -2.28090942e-01 -7.39560366e-01
7.49756873e-01 -3.68143052e-01 -4.19987887e-01 2.27228478e-01
7.18739033e-01 1.02993280e-01 8.83508325e-02 -5.56274593e-01
-2.08769262e-01 -2.93481052e-01 -3.57075423e-01 -7.29530096e-01
1.15217924e+00 -3.20848554e-01 1.90911219e-01 1.02500093e+00
1.08021212e+00 1.79987341e-01 -1.23282719e+00 -2.11415380e-01
3.55451852e-01 -2.08875567e-01 -1.79035783e-01 -9.29791510e-01
-1.81058004e-01 9.61865842e-01 9.09167249e-03 7.85017669e-01
6.62933290e-01 -1.56765670e-01 8.04992318e-01 6.89281166e-01
2.85550863e-01 -1.29759860e+00 2.83944398e-01 1.26614773e+00
1.08074617e+00 -1.25005221e+00 -5.69509625e-01 4.61009741e-02
-1.31297958e+00 1.24092376e+00 6.53154194e-01 1.28876969e-01
4.46647823e-01 4.20170277e-01 7.32147396e-01 2.25700915e-01
-1.52433038e+00 1.81239754e-01 -1.29515693e-01 4.06888664e-01
1.01683044e+00 1.76093474e-01 -3.25040109e-02 7.54050314e-01
-5.64602673e-01 -2.48882085e-01 8.64868104e-01 9.43733633e-01
-7.76805401e-01 -1.16646421e+00 -1.71674952e-01 3.59310806e-01
-3.56220990e-01 -5.76504432e-02 -8.82534742e-01 4.48049724e-01
-7.12928176e-01 1.77031553e+00 2.87962198e-01 -6.31147146e-01
5.07713377e-01 5.63140690e-01 -4.02514637e-03 -8.81997705e-01
-1.10668504e+00 3.39660615e-01 8.00297976e-01 -7.21474528e-01
-3.61419208e-02 -6.67626143e-01 -1.48480010e+00 -3.69528443e-01
-4.56985384e-01 5.05427182e-01 4.33855802e-01 8.94282043e-01
1.71174139e-01 4.59396183e-01 1.14207971e+00 -9.33742046e-01
-9.18337524e-01 -1.57246971e+00 -2.76954323e-01 4.05259021e-02
5.68566978e-01 -2.27284078e-02 -4.29513752e-01 -1.52616128e-01]
|
[12.865425109863281, 7.886548042297363]
|
194e3a0b-18b8-405f-a008-802637eecc0b
|
adaptive-convolutional-dictionary-network-for
|
2205.07471
| null |
https://arxiv.org/abs/2205.07471v2
|
https://arxiv.org/pdf/2205.07471v2.pdf
|
Adaptive Convolutional Dictionary Network for CT Metal Artifact Reduction
|
Inspired by the great success of deep neural networks, learning-based methods have gained promising performances for metal artifact reduction (MAR) in computed tomography (CT) images. However, most of the existing approaches put less emphasis on modelling and embedding the intrinsic prior knowledge underlying this specific MAR task into their network designs. Against this issue, we propose an adaptive convolutional dictionary network (ACDNet), which leverages both model-based and learning-based methods. Specifically, we explore the prior structures of metal artifacts, e.g., non-local repetitive streaking patterns, and encode them as an explicit weighted convolutional dictionary model. Then, a simple-yet-effective algorithm is carefully designed to solve the model. By unfolding every iterative substep of the proposed algorithm into a network module, we explicitly embed the prior structure into a deep network, \emph{i.e.,} a clear interpretability for the MAR task. Furthermore, our ACDNet can automatically learn the prior for artifact-free CT images via training data and adaptively adjust the representation kernels for each input CT image based on its content. Hence, our method inherits the clear interpretability of model-based methods and maintains the powerful representation ability of learning-based methods. Comprehensive experiments executed on synthetic and clinical datasets show the superiority of our ACDNet in terms of effectiveness and model generalization. {\color{blue}{{\textit{Code is available at {\url{https://github.com/hongwang01/ACDNet}}.}}}}
|
['Yefeng Zheng', 'Deyu Meng', 'Yuexiang Li', 'Hong Wang']
|
2022-05-16
| null | null | null | null |
['metal-artifact-reduction']
|
['medical']
|
[ 1.72639966e-01 8.65204334e-02 -2.60055307e-02 -3.15760404e-01
-5.82856476e-01 -5.75217232e-02 2.35176131e-01 1.27301842e-01
-1.61257029e-01 4.39300716e-01 8.60431045e-02 -3.56685936e-01
-4.18919951e-01 -7.62528718e-01 -6.35840535e-01 -7.52564788e-01
8.92800838e-02 1.05887271e-01 2.52161175e-01 -9.83090401e-02
7.21624345e-02 5.44467688e-01 -9.04935360e-01 1.98221311e-01
1.02297556e+00 1.04842985e+00 4.59319293e-01 2.34818414e-01
1.69949651e-01 1.07778180e+00 -9.41226557e-02 -1.13117911e-01
1.41272947e-01 -4.71932501e-01 -6.80138469e-01 1.57696202e-01
-1.44344717e-01 -4.83685583e-01 -9.42452252e-01 9.75724995e-01
6.53227508e-01 4.98962700e-02 5.99090934e-01 -7.64715075e-01
-7.46491432e-01 5.48393846e-01 -6.39950395e-01 5.93857825e-01
-1.12175442e-01 3.16840470e-01 6.86246693e-01 -1.04629731e+00
3.35785687e-01 4.51825142e-01 6.45501912e-01 4.22791392e-01
-1.08809733e+00 -7.61773407e-01 1.56923220e-01 1.48516029e-01
-1.39461040e+00 -1.75151423e-01 1.07860279e+00 -5.49377322e-01
3.64211977e-01 3.04314405e-01 7.58989036e-01 9.08107281e-01
1.76138684e-01 8.72068226e-01 1.03684258e+00 -2.22922176e-01
1.82202354e-01 -1.35396808e-01 5.11596315e-02 8.40287209e-01
3.57302964e-01 6.77979589e-02 -7.47707337e-02 -2.12610364e-02
1.38709295e+00 3.79958063e-01 -6.12136066e-01 -3.83767784e-01
-1.14099860e+00 6.80088282e-01 7.14731038e-01 3.98815095e-01
-5.37237883e-01 2.81812578e-01 3.10424387e-01 -2.37836167e-01
2.83085942e-01 2.74283290e-01 -1.67864874e-01 2.24391699e-01
-7.13977873e-01 -6.49303338e-03 2.40803286e-01 8.97563756e-01
6.73476279e-01 2.85313278e-01 -2.40086794e-01 1.03515899e+00
2.83317506e-01 2.11424947e-01 6.46086752e-01 -6.53930068e-01
2.23433688e-01 7.34976113e-01 -1.93057463e-01 -1.08100355e+00
-5.19393265e-01 -8.51427913e-01 -1.24491656e+00 5.67769781e-02
6.37776479e-02 -3.12995128e-02 -9.63275731e-01 1.54406106e+00
3.92550290e-01 4.60821331e-01 -4.33897316e-01 1.07595730e+00
7.69614816e-01 3.42306376e-01 5.67321889e-02 -2.00573802e-01
1.38422322e+00 -8.02741230e-01 -6.17550731e-01 -1.14382863e-01
5.87480724e-01 -6.07288301e-01 1.08206379e+00 4.91148621e-01
-1.02336526e+00 -4.27053154e-01 -1.29910374e+00 1.27506061e-02
1.86930984e-01 2.61422247e-01 7.02795506e-01 3.70495737e-01
-6.64362431e-01 5.04226387e-01 -1.24472666e+00 9.18326527e-02
7.15214610e-01 4.27279174e-01 2.23934658e-05 -2.82727063e-01
-1.11994839e+00 6.11189246e-01 2.50296384e-01 4.15616035e-01
-1.04553604e+00 -7.83753037e-01 -7.77518630e-01 -4.69375774e-02
4.67496812e-01 -6.73398972e-01 1.15175092e+00 -9.16643918e-01
-1.46235621e+00 5.24581611e-01 1.43124178e-01 -1.28249526e-01
3.84018004e-01 -1.67301580e-01 -3.65065038e-01 4.25148964e-01
-1.65217325e-01 1.96862012e-01 7.62575507e-01 -1.43897283e+00
-3.12857687e-01 -1.00401357e-01 2.42203042e-01 -2.04172343e-01
-6.93739533e-01 -1.04235053e-01 -5.41295230e-01 -1.15291238e+00
3.96466017e-01 -8.47349465e-01 -3.90212715e-01 2.06759527e-01
-4.01414573e-01 4.27631661e-02 4.61424440e-01 -6.47957444e-01
1.62072217e+00 -2.06957674e+00 3.53697427e-02 3.69005740e-01
6.39051616e-01 3.19814205e-01 6.68196082e-02 2.38954693e-01
-3.61842990e-01 -9.76991467e-03 -5.74699402e-01 1.13967694e-01
-1.39098421e-01 2.70609617e-01 -1.51998131e-02 5.49791932e-01
2.58762836e-01 8.62657845e-01 -9.25897419e-01 -4.64194149e-01
4.08164114e-01 5.67674458e-01 -6.44019723e-01 1.96826741e-01
-5.08414283e-02 7.51111686e-01 -9.33135688e-01 4.40476328e-01
8.47762883e-01 -6.39215469e-01 3.20945114e-01 -4.56849873e-01
-8.10547248e-02 9.54427123e-02 -1.15438128e+00 1.81530440e+00
-5.36033750e-01 2.40921736e-01 9.63450298e-02 -1.11901164e+00
6.97531343e-01 3.62310737e-01 8.69988441e-01 -6.34340286e-01
2.63318151e-01 4.30184901e-01 1.32289022e-01 -6.58869743e-01
1.32261142e-02 -2.89632261e-01 2.94220835e-01 5.11305332e-01
-1.99349433e-01 6.56208321e-02 -2.79326916e-01 1.03731975e-01
1.09001696e+00 1.29158854e-01 1.85512692e-01 -4.25065428e-01
6.42217636e-01 -2.54400462e-01 6.93377316e-01 5.48317611e-01
7.20306206e-03 8.69833589e-01 2.00557798e-01 -6.99895859e-01
-9.44286168e-01 -9.22571301e-01 -4.87289727e-01 5.75248063e-01
2.91729152e-01 -2.96385586e-01 -7.44870365e-01 -4.86772001e-01
-3.10539544e-01 4.05546755e-01 -7.75564373e-01 -3.19543749e-01
-9.28523064e-01 -1.00237501e+00 4.15093243e-01 7.77515709e-01
6.03382409e-01 -9.28935349e-01 -6.06490910e-01 2.72333652e-01
-1.90652534e-01 -9.54249203e-01 -5.08561254e-01 9.28428173e-02
-9.33132052e-01 -9.68002796e-01 -7.16606259e-01 -7.30048776e-01
9.29702580e-01 2.52497345e-01 8.77559841e-01 6.85804129e-01
-4.92943585e-01 3.03360045e-01 -3.53034139e-01 -2.81508684e-01
-1.43997684e-01 -1.31557912e-01 -8.74379948e-02 1.57816738e-01
3.98050845e-02 -8.82260323e-01 -1.23133552e+00 3.71883482e-01
-1.40777838e+00 3.76002133e-01 8.66512418e-01 9.62566018e-01
8.56342196e-01 9.56000537e-02 5.13264537e-01 -8.42162132e-01
4.15336758e-01 -5.57731748e-01 -3.81700307e-01 5.39768562e-02
-5.19558311e-01 -6.34276133e-04 6.28911018e-01 -3.24417263e-01
-9.05476213e-01 2.80395709e-02 -3.54447186e-01 -5.40118575e-01
-4.03639264e-02 6.86943829e-01 -1.53787643e-01 -8.43374133e-02
5.49597859e-01 4.33081299e-01 -1.09183356e-01 -4.20059830e-01
1.47996843e-03 4.18404400e-01 4.03554142e-01 -8.51080716e-01
7.29262948e-01 7.00743556e-01 -7.77434260e-02 -4.97791857e-01
-8.18108916e-01 -1.93700477e-01 -5.54213822e-01 -1.78080648e-01
8.61961722e-01 -8.12297165e-01 -5.22596538e-01 5.55026650e-01
-9.73499656e-01 -4.27074492e-01 -2.27531582e-01 6.02486789e-01
-4.59523201e-01 6.50869846e-01 -7.42033362e-01 -4.11354154e-01
-3.58192056e-01 -1.33062458e+00 6.73287988e-01 1.15477346e-01
9.52923391e-03 -1.07844555e+00 -8.26646835e-02 1.91869631e-01
3.69173437e-01 3.85487080e-01 1.11825609e+00 -5.41221201e-01
-7.71567702e-01 -1.53116152e-01 -3.74212861e-01 5.22046924e-01
4.43533242e-01 -4.32643235e-01 -9.54580903e-01 -3.41629356e-01
3.18726718e-01 -9.02871490e-02 6.96250081e-01 4.35175419e-01
1.88175046e+00 -1.47648484e-01 -2.37404943e-01 7.52996624e-01
1.56426001e+00 2.38684088e-01 8.06584954e-01 3.89664233e-01
7.92515337e-01 1.31637722e-01 1.52324542e-01 7.34917164e-01
2.56855488e-01 6.05981171e-01 5.45612037e-01 -4.54591453e-01
-1.78400040e-01 -7.04894066e-02 -1.06152847e-01 9.58163440e-01
-3.00678015e-01 6.59422651e-02 -9.83638406e-01 4.48339641e-01
-1.82466543e+00 -6.00074649e-01 -1.19423389e-01 2.06096148e+00
9.55870628e-01 1.69864923e-01 -2.35617802e-01 2.28066593e-01
6.07908249e-01 1.00935660e-01 -6.98295236e-01 1.01103522e-01
1.45836920e-01 5.28304458e-01 2.35485524e-01 1.55137509e-01
-9.83460724e-01 4.63836014e-01 4.87936401e+00 9.33787167e-01
-1.28784728e+00 2.47319743e-01 6.34725094e-01 3.84948589e-02
-5.36373317e-01 -8.59387666e-02 -2.67295271e-01 4.75100338e-01
4.37032998e-01 -2.86350846e-02 2.96672404e-01 5.11220157e-01
4.96326536e-01 1.14943862e-01 -1.02590847e+00 8.50449383e-01
-6.80790171e-02 -1.47500193e+00 1.93056703e-01 -2.21823566e-02
5.03632724e-01 -1.59755260e-01 1.59736723e-01 1.67117476e-01
7.84459263e-02 -9.98911679e-01 6.61290467e-01 5.96218228e-01
8.58875632e-01 -4.56846297e-01 7.17060864e-01 1.23714760e-01
-1.15808904e+00 -2.47177891e-02 -2.70877361e-01 9.90983695e-02
5.53939492e-04 7.92413712e-01 -3.57083291e-01 8.84784520e-01
6.79298580e-01 7.80916810e-01 -3.83335054e-01 1.13448560e+00
-3.67187887e-01 7.20782638e-01 -1.48717850e-01 3.67581725e-01
1.69770092e-01 -1.05959818e-01 3.69078308e-01 1.23420167e+00
2.07659930e-01 6.69619679e-01 2.48736903e-01 9.54424679e-01
-6.61861226e-02 7.51324892e-02 -2.27705404e-01 3.08830351e-01
2.86178887e-01 1.27131498e+00 -7.49869943e-01 -2.29913637e-01
-4.76018727e-01 7.32809722e-01 2.93469608e-01 3.62678319e-01
-1.02903390e+00 -4.31770623e-01 2.99031824e-01 3.98896158e-01
3.67588401e-01 -1.86516136e-01 -5.45611322e-01 -1.15475297e+00
2.47414738e-01 -8.94540131e-01 4.10899818e-01 -7.37780333e-01
-1.24497807e+00 8.56392860e-01 -2.97259130e-02 -1.62156296e+00
4.07161415e-01 -5.74312329e-01 -6.16005480e-01 7.36975372e-01
-1.64186788e+00 -1.17380238e+00 -6.28878891e-01 6.79926157e-01
5.24104655e-01 1.65820524e-01 6.03146136e-01 5.63838959e-01
-8.21578383e-01 5.59814990e-01 8.13845769e-02 4.44638163e-01
3.77011120e-01 -9.84743476e-01 -1.18207857e-01 7.90612817e-01
-2.56910294e-01 8.37678552e-01 3.19850147e-01 -5.22448719e-01
-1.36838496e+00 -1.20457554e+00 1.77815005e-01 -2.57029146e-01
8.91114831e-01 -1.20312355e-01 -1.05729294e+00 6.19338751e-01
1.21232487e-01 3.99899542e-01 6.30955577e-01 -3.88565719e-01
-1.56367272e-01 -1.71354055e-01 -8.76173198e-01 4.57169533e-01
1.01051080e+00 -3.40961635e-01 -4.46087301e-01 4.77842152e-01
5.90770245e-01 -6.47810757e-01 -1.12803674e+00 6.03325009e-01
4.10344064e-01 -8.69034290e-01 1.00408375e+00 -3.97060484e-01
7.90228248e-01 -4.08352286e-01 7.05135614e-02 -9.92676914e-01
-6.41478360e-01 -3.70675296e-01 -1.03212953e-01 8.46884549e-01
1.74500376e-01 -6.18479431e-01 5.52175343e-01 5.57748377e-01
-7.29231477e-01 -1.29750097e+00 -8.76921475e-01 -5.13585925e-01
2.63865888e-01 -5.16659141e-01 5.94224989e-01 1.02002001e+00
-1.53605178e-01 -1.30223423e-01 -3.53085577e-01 5.27675331e-01
6.23715520e-01 5.47387786e-02 2.63539374e-01 -9.16401267e-01
-4.78788853e-01 -3.73353958e-01 -2.57789910e-01 -1.18536019e+00
-2.97831774e-01 -1.02991891e+00 -1.35086089e-01 -1.58633447e+00
4.14535612e-01 -8.72219384e-01 -8.18893790e-01 6.30202472e-01
-3.21537375e-01 2.30633289e-01 -5.45449927e-02 4.35940564e-01
-4.83210266e-01 6.79820240e-01 1.64018309e+00 -1.22450352e-01
-7.15888431e-03 -1.74437612e-01 -7.45639503e-01 8.03786278e-01
6.88577354e-01 -5.07335842e-01 -4.58352655e-01 -7.73573041e-01
1.59031913e-01 7.64269382e-02 6.55649304e-01 -9.93123233e-01
1.69585779e-01 -2.51564115e-01 4.67391908e-01 -3.04498136e-01
-1.07646100e-02 -8.90873253e-01 2.56474108e-01 6.01027310e-01
-1.28877252e-01 -1.23204209e-01 2.91772664e-01 5.66099524e-01
-1.55886129e-01 -1.94081292e-01 9.79819417e-01 -2.68798888e-01
-4.89564031e-01 7.02863514e-01 -3.27133983e-01 -2.67781578e-02
7.73462653e-01 -2.53874838e-01 -3.43279876e-02 -6.47292212e-02
-7.71758199e-01 1.47588223e-01 2.80458927e-01 1.09188817e-01
9.01018739e-01 -1.41157663e+00 -6.31551325e-01 2.18916818e-01
1.39977500e-01 4.12985802e-01 5.79336822e-01 1.32042503e+00
-5.98410964e-01 1.13428555e-01 -1.36214465e-01 -6.46560311e-01
-6.89171433e-01 4.98743832e-01 6.22896910e-01 -3.29865068e-01
-9.70245004e-01 7.25429893e-01 5.97833395e-01 -1.22806214e-01
-2.90504340e-02 -4.63756949e-01 -5.83677478e-02 -5.04096150e-01
4.12186533e-01 1.03367724e-01 2.45558888e-01 -2.26404518e-01
-3.25893641e-01 5.05299866e-01 -2.67357588e-01 3.43578517e-01
1.66305697e+00 6.48073945e-03 -1.33336306e-01 4.43882532e-02
1.00615501e+00 -1.09180368e-01 -1.30869591e+00 -4.60831553e-01
-6.10274784e-02 -3.66765499e-01 2.52124399e-01 -6.52563751e-01
-1.52198637e+00 8.63873303e-01 6.45671070e-01 -1.79466084e-01
1.48082805e+00 -1.44518286e-01 9.45268035e-01 3.26341577e-02
2.99627721e-01 -8.38018894e-01 4.39796507e-01 1.81988209e-01
8.56830478e-01 -1.01086378e+00 3.00266773e-01 -4.41749096e-01
-4.48589236e-01 1.12865996e+00 5.39007127e-01 -2.77518809e-01
8.99250984e-01 1.88482434e-01 -2.75227912e-02 -5.71137428e-01
-2.76747465e-01 2.94862092e-01 1.27049834e-01 2.50718832e-01
3.22754860e-01 -1.37248725e-01 -4.52205569e-01 1.09775496e+00
2.14952707e-01 4.21186626e-01 4.24249411e-01 1.11140084e+00
-2.33705088e-01 -9.53093827e-01 -2.12178558e-01 3.87535334e-01
-3.38812888e-01 -2.91691482e-01 1.27547249e-01 7.54992604e-01
2.73754835e-01 6.74764156e-01 -3.51235390e-01 -4.29844946e-01
4.91478741e-01 -3.99939477e-01 4.19762522e-01 -6.04801536e-01
-6.13283575e-01 1.83013320e-01 -3.08874547e-01 -3.51846099e-01
-3.93571883e-01 -4.42297459e-01 -1.46728730e+00 -2.30732299e-02
-3.39225829e-01 1.21315452e-03 1.83950961e-01 8.96974564e-01
2.61810511e-01 9.35171008e-01 7.58407295e-01 -5.98268688e-01
-4.01024878e-01 -7.75889754e-01 -5.16934574e-01 4.48880762e-01
1.46911293e-01 -8.07175636e-01 -2.40844876e-01 1.87923014e-03]
|
[13.50600814819336, -2.546527862548828]
|
873689a9-fec0-4244-ac57-cb9bf17bbb12
|
effective-connectivity-from-single-trial-fmri
|
1803.05840
| null |
http://arxiv.org/abs/1803.05840v1
|
http://arxiv.org/pdf/1803.05840v1.pdf
|
Effective Connectivity from Single Trial fMRI Data by Sampling Biologically Plausible Models
|
The estimation of causal network architectures in the brain is fundamental
for understanding cognitive information processes. However, access to the
dynamic processes underlying cognition is limited to indirect measurements of
the hidden neuronal activity, for instance through fMRI data. Thus, estimating
the network structure of the underlying process is challenging. In this
article, we embed an adaptive importance sampler called Adaptive Path Integral
Smoother (APIS) into the Expectation-Maximization algorithm to obtain point
estimates of causal connectivity. We demonstrate on synthetic data that this
procedure finds not only the correct network structure but also the direction
of effective connections from random initializations of the connectivity
matrix. In addition--motivated by contradictory claims in the literature--we
examine the effect of the neuronal timescale on the sensitivity of the BOLD
signal to changes in the connectivity and on the maximum likelihood solutions
of the connectivity. We conclude with two warnings: First, the connectivity
estimates under the assumption of slow dynamics can be extremely biased if the
data was generated by fast neuronal processes. Second, the faster the time
scale, the less sensitive the BOLD signal is to changes in the incoming
connections to a node. Hence, connectivity estimation using realistic neural
dynamics timescale requires extremely high-quality data and seems infeasible in
many practical data sets.
|
[]
|
2018-03-15
| null | null | null | null |
['connectivity-estimation']
|
['graphs']
|
[ 2.87497818e-01 1.14406459e-01 2.61302620e-01 -1.74353212e-01
2.30015308e-01 -4.20201957e-01 6.08079314e-01 2.16824368e-01
-4.53675419e-01 8.58615279e-01 3.85199040e-01 -3.02452922e-01
-5.96417367e-01 -6.67675436e-01 -6.09230280e-01 -7.42550671e-01
-6.06461525e-01 3.16667497e-01 -2.07595546e-02 6.30372316e-02
4.14560497e-01 3.86544496e-01 -1.00362527e+00 -3.92221242e-01
8.91510367e-01 6.32156968e-01 5.83836250e-02 4.61168855e-01
2.27177981e-02 5.53759158e-01 -3.09148163e-01 -1.63628340e-01
1.60390660e-01 -6.82839513e-01 -5.23987770e-01 -1.03751481e-01
1.15472637e-02 -1.04952596e-01 -4.37392861e-01 1.31909394e+00
1.99245691e-01 2.19623148e-01 8.43039513e-01 -1.02974927e+00
-4.82359268e-02 9.08262014e-01 -6.04995906e-01 8.21170211e-01
-1.12519868e-01 3.54062647e-01 9.31175947e-01 -4.98879045e-01
6.34983778e-01 1.21850717e+00 5.22982597e-01 1.76260680e-01
-1.74912238e+00 -7.16769516e-01 2.59923100e-01 2.21914008e-01
-1.31820273e+00 -5.22442639e-01 7.53730237e-01 -7.66241252e-01
3.16087484e-01 -1.53708369e-01 9.32150185e-01 1.24604118e+00
6.68135405e-01 -1.27208680e-01 1.38215959e+00 5.38729280e-02
5.28057218e-01 -3.05941373e-01 4.48663980e-01 4.69031304e-01
4.78906512e-01 2.63253331e-01 -5.92768788e-01 -3.15641552e-01
1.08792174e+00 -3.66428673e-01 -3.66009891e-01 -9.88537371e-02
-1.40920126e+00 7.51328409e-01 5.62042594e-01 2.65636802e-01
-7.04049647e-01 3.70947897e-01 1.13975823e-01 3.22096527e-01
4.02433932e-01 5.26039660e-01 -2.80086756e-01 -1.48985684e-02
-9.91533875e-01 8.79222527e-02 7.75329053e-01 1.13125727e-01
5.06927311e-01 8.07347372e-02 -7.12209120e-02 2.16055989e-01
3.22381526e-01 3.60635638e-01 1.30790800e-01 -1.12021148e+00
1.63680926e-01 1.31595984e-01 7.34433979e-02 -1.10107458e+00
-6.69465661e-01 -6.71986878e-01 -1.04345942e+00 1.04193576e-01
9.40131724e-01 -6.65651143e-01 -3.57783407e-01 2.06127477e+00
9.12859887e-02 2.03064114e-01 -6.52865529e-01 8.30864370e-01
5.31769404e-03 4.22763884e-01 1.62312850e-01 -6.68587685e-01
1.12752497e+00 -7.28926510e-02 -7.93203056e-01 -2.93641329e-01
3.57979722e-02 -2.23487884e-01 6.95660412e-01 1.89013764e-01
-1.00304973e+00 -3.41253549e-01 -8.92068803e-01 4.71827000e-01
1.49239182e-01 -4.01319146e-01 7.45230615e-01 3.40320647e-01
-9.21820164e-01 8.24037969e-01 -9.70232666e-01 -2.67421395e-01
2.96480983e-01 2.55872786e-01 -1.21590175e-01 9.57741812e-02
-1.06047773e+00 8.55026424e-01 2.56683767e-01 2.75401652e-01
-8.98669124e-01 -7.21366048e-01 -1.81589842e-01 3.59053850e-01
2.30099469e-01 -7.17804193e-01 7.55957246e-01 -1.12912452e+00
-1.27135181e+00 7.24983364e-02 -1.78693503e-01 -5.07662714e-01
5.59521258e-01 2.11362422e-01 1.73779316e-02 2.92792678e-01
-3.55680622e-02 6.69755220e-01 1.05914927e+00 -7.72784591e-01
3.09348732e-01 -5.19502103e-01 -3.95142913e-01 1.58478558e-01
1.27647102e-01 -3.55306476e-01 1.86201453e-01 -4.21750426e-01
4.46622729e-01 -9.89090383e-01 -3.15429658e-01 2.04806209e-01
-4.89537597e-01 2.67373145e-01 3.05612717e-04 -6.66516542e-01
8.33570361e-01 -2.09635234e+00 2.06829444e-01 6.36289001e-01
5.69528401e-01 -4.90838856e-01 -5.00571355e-02 2.63531655e-01
-3.75375062e-01 3.37961018e-02 -4.28755134e-01 3.50637734e-01
-2.43788302e-01 -2.35169575e-01 -1.72016293e-01 9.41584766e-01
8.70257169e-02 7.80753255e-01 -1.03861094e+00 -2.08698943e-01
-9.66833755e-02 5.57450593e-01 -4.20094699e-01 -2.45656688e-02
-5.63201234e-02 9.38678682e-01 -2.08046108e-01 -4.18689661e-02
4.55230594e-01 -3.88147324e-01 5.27634501e-01 -1.97416902e-01
-3.41987878e-01 3.47093344e-01 -1.09794402e+00 1.18868124e+00
-6.58854097e-02 9.11867857e-01 6.79858625e-02 -9.85619366e-01
6.46553993e-01 2.80712456e-01 4.54131693e-01 -7.03236580e-01
1.42647043e-01 -4.26176004e-02 1.14430451e+00 -1.60931736e-01
-1.52472198e-01 -1.21759832e-01 4.32422251e-01 7.35314190e-01
-4.43721376e-02 6.29160404e-02 1.49418548e-01 3.77978444e-01
1.25238299e+00 -2.80652225e-01 9.58873853e-02 -7.85070062e-01
-2.31987499e-02 -4.31834847e-01 3.81991476e-01 7.71326244e-01
-2.48828903e-01 6.72747418e-02 1.26138270e+00 9.58478823e-02
-1.09647727e+00 -1.22306967e+00 -3.28695685e-01 1.12364657e-01
-2.95343786e-01 4.90588509e-02 -8.23365748e-01 -1.70969255e-02
-5.91209270e-02 5.81492245e-01 -8.78220677e-01 -4.12869871e-01
-2.02913135e-01 -1.13303661e+00 1.90160573e-01 4.22678981e-03
2.77586341e-01 -9.04259264e-01 -6.44937813e-01 4.12416548e-01
-2.64009476e-01 -9.59550500e-01 -3.70431334e-01 4.75710109e-02
-1.28067100e+00 -1.11071622e+00 -4.60618466e-01 4.19528633e-02
9.12639022e-01 3.09390072e-02 8.44464898e-01 6.83364570e-02
-3.66396725e-01 2.03551158e-01 2.49832839e-01 -4.70839441e-02
-9.01872143e-02 -1.30775258e-01 5.30097373e-02 3.15728903e-01
-1.31030530e-01 -1.00631630e+00 -9.30649996e-01 9.81053263e-02
-4.44861174e-01 1.11531308e-02 5.83103955e-01 4.92848486e-01
4.06538993e-01 2.31863320e-01 5.25027335e-01 -7.35367239e-01
9.33688521e-01 -8.30957770e-01 -6.58001482e-01 -1.96871758e-01
-7.10774958e-01 5.69187820e-01 2.29082271e-01 -6.73063278e-01
-1.01640010e+00 -2.98008144e-01 3.01165968e-01 5.29822595e-02
8.18829313e-02 7.28022873e-01 1.84738815e-01 9.14629847e-02
6.82347655e-01 -6.76044356e-03 1.10700898e-01 -1.47025779e-01
3.24577957e-01 -1.21608943e-01 1.84325263e-01 -6.51750326e-01
4.33320642e-01 6.11170053e-01 4.71088320e-01 -1.10157609e+00
-5.31614244e-01 1.26984045e-01 -7.99275637e-01 -4.95164961e-01
7.29496419e-01 -7.22708464e-01 -8.83646727e-01 4.50158864e-01
-1.02153742e+00 -6.01706743e-01 -2.39478409e-01 8.10729504e-01
-3.85869056e-01 2.42138058e-01 -5.99909067e-01 -9.16999698e-01
-1.31640121e-01 -9.85996008e-01 3.84309255e-02 5.72088212e-02
-4.86502767e-01 -1.17215836e+00 5.72178513e-02 -2.21437767e-01
5.46583891e-01 3.82411808e-01 1.10400999e+00 -1.20288566e-01
-4.79499847e-01 -2.31871475e-02 -3.67469579e-01 -1.17997333e-01
-6.87703043e-02 2.83004850e-01 -6.83602571e-01 2.47536283e-02
3.77552390e-01 2.65661925e-01 6.50905788e-01 1.11979830e+00
8.12752485e-01 -2.17801675e-01 -2.29021490e-01 2.58968025e-01
1.22893214e+00 -1.22672446e-01 4.03000087e-01 -2.24993005e-01
4.37073439e-01 1.07482481e+00 -3.60277928e-02 4.43491220e-01
1.02909513e-01 3.09554935e-01 3.25646639e-01 3.10103565e-01
-8.39446299e-03 -3.00958395e-01 3.55788231e-01 8.73345256e-01
-6.83985576e-02 8.60912055e-02 -9.20244873e-01 4.01920676e-01
-1.55486798e+00 -9.96743977e-01 -6.70470059e-01 2.52238679e+00
9.11203742e-01 3.88663530e-01 1.05865948e-01 -1.42374814e-01
7.35728145e-01 -8.05356726e-02 -8.96886289e-01 1.72781684e-02
5.97205274e-02 1.74983628e-02 5.15335381e-01 7.87689209e-01
-2.59767592e-01 2.89615422e-01 6.65256166e+00 1.02853052e-01
-9.48473573e-01 1.96677625e-01 7.25559890e-01 -2.98229724e-01
-2.59994000e-01 2.68964738e-01 -4.33797956e-01 6.75297856e-01
1.40625596e+00 -4.08582479e-01 8.16918433e-01 -6.74611330e-02
8.43317986e-01 -5.54072261e-01 -9.76610065e-01 6.16880774e-01
-5.52955031e-01 -9.39192533e-01 -4.30346131e-01 4.69428152e-01
4.98865008e-01 2.02069223e-01 -1.53480127e-01 -3.20583045e-01
3.96483451e-01 -8.19317818e-01 6.47504508e-01 1.02223647e+00
3.96055460e-01 -5.19994080e-01 4.37592477e-01 4.93701637e-01
-7.66764879e-01 1.44585237e-01 -4.56629157e-01 -4.04483467e-01
2.81269699e-01 1.44961905e+00 -4.96133804e-01 -3.95818025e-01
2.10854068e-01 5.01767039e-01 -5.41602433e-01 1.18436551e+00
-4.76178676e-01 9.44201946e-01 -5.19708693e-01 1.02542944e-01
-3.37759368e-02 -5.83458841e-01 5.99450946e-01 7.37242758e-01
1.53894201e-01 1.60293743e-01 -3.91015172e-01 1.40030503e+00
4.25835885e-02 -2.23351404e-01 -4.59467500e-01 -3.31605077e-01
6.53941154e-01 1.33502746e+00 -1.15683937e+00 1.61488913e-02
-1.24036811e-01 5.53776741e-01 5.30672014e-01 6.18688643e-01
-5.09167731e-01 2.76107937e-02 5.99005103e-01 3.98622632e-01
-2.11650040e-02 -7.80256033e-01 -3.50683481e-01 -9.37745869e-01
-2.57842004e-01 -5.28165340e-01 1.83091068e-03 -6.97264791e-01
-1.23938370e+00 2.23768622e-01 2.11646006e-01 -2.55888700e-01
-1.70287281e-01 -2.72201389e-01 -8.21453393e-01 1.21937180e+00
-1.15274775e+00 5.16320467e-02 -5.37154377e-02 5.07254958e-01
1.65744171e-01 5.90668857e-01 4.40650374e-01 -9.10776258e-02
-6.81126118e-01 -1.32339984e-01 4.59943935e-02 -7.65361339e-02
4.91207898e-01 -9.87400472e-01 4.84422147e-01 8.47433746e-01
8.87501538e-02 8.77744138e-01 9.28849757e-01 -8.55930328e-01
-1.13371599e+00 -5.69967628e-01 4.95239913e-01 -2.81112224e-01
1.18493664e+00 -4.22017276e-01 -9.80338395e-01 4.87411529e-01
2.35437259e-01 -1.78990904e-02 5.38164020e-01 1.65184587e-01
2.05985755e-02 8.95056874e-02 -9.39426959e-01 6.47708833e-01
8.83537233e-01 -4.50151145e-01 -1.88290894e-01 2.42936283e-01
3.20499092e-01 3.71704847e-01 -9.23758507e-01 -2.08283648e-01
5.89646578e-01 -8.23737681e-01 6.66249156e-01 -4.73654866e-01
2.32627317e-01 1.55582523e-03 2.56878972e-01 -1.56751943e+00
-6.52854443e-01 -4.15365607e-01 -1.40143886e-01 9.86974061e-01
5.01301467e-01 -7.26901710e-01 4.12747234e-01 9.03477371e-01
6.35158658e-01 -2.74348974e-01 -8.81935954e-01 -3.80762726e-01
3.68961468e-02 -3.16123277e-01 -1.12059005e-01 9.67186034e-01
5.71351387e-02 6.48006022e-01 7.71823106e-03 2.10372984e-01
1.22877145e+00 -3.98931086e-01 5.16488217e-02 -1.61303949e+00
-3.23878706e-01 -4.44571853e-01 -8.81731138e-02 -5.93980193e-01
1.50067121e-01 -5.36751211e-01 -6.49431050e-02 -1.17529571e+00
2.81319767e-01 -3.38051736e-01 -1.35017931e-01 -3.24078761e-02
-2.53854007e-01 -2.68851548e-01 -1.35085462e-02 3.63023043e-01
7.06031322e-02 4.49797779e-01 1.19387782e+00 1.94298968e-01
-1.80783868e-02 -2.69026812e-02 -5.11816442e-01 6.29622102e-01
7.96539664e-01 -7.34001815e-01 -6.97486341e-01 -6.58341497e-02
7.02140331e-01 4.21426177e-01 4.70682770e-01 -8.77504230e-01
2.44195968e-01 -8.38291720e-02 4.57163990e-01 -1.02181211e-01
6.60896823e-02 -5.53119242e-01 2.37665921e-01 7.54997313e-01
-6.31580949e-01 2.45623559e-01 4.14989293e-02 8.06546509e-01
2.99824268e-01 -2.78677225e-01 7.55727768e-01 -3.96117240e-01
-5.74883558e-02 3.22167397e-01 -7.75784850e-01 2.84252405e-01
3.74899328e-01 2.74771273e-01 -2.31493235e-01 -6.32051945e-01
-8.43417943e-01 -5.68836741e-02 1.12475313e-01 -7.32287839e-02
3.95329326e-01 -8.36977482e-01 -6.92185402e-01 1.46920338e-01
-6.28712952e-01 -5.73572338e-01 1.68340996e-01 1.40773475e+00
-1.66127488e-01 2.31081977e-01 -3.81362736e-01 -2.62511104e-01
-7.04252422e-01 3.27581912e-01 5.70237875e-01 9.42307711e-03
-4.47757900e-01 5.78022957e-01 1.46336943e-01 8.66336524e-02
-8.61468241e-02 -2.88352013e-01 -4.65227291e-02 4.18492407e-01
4.75743771e-01 4.87134188e-01 -2.12132141e-01 -2.71222353e-01
-1.40686944e-01 1.54814739e-02 2.14724913e-02 -5.57504892e-01
1.17068994e+00 -3.80800128e-01 -4.24845904e-01 9.32973981e-01
7.31123865e-01 -1.82187676e-01 -1.87877309e+00 -1.38950244e-01
1.44996913e-02 -2.06720069e-01 5.47407985e-01 -5.84532082e-01
-1.05038452e+00 1.08835518e+00 6.48087204e-01 4.08056349e-01
5.24215698e-01 -2.71116048e-01 -1.40124708e-02 1.89139694e-01
2.72144169e-01 -8.35458934e-01 -1.22559164e-02 2.29416430e-01
6.38149858e-01 -8.39307487e-01 2.79139727e-01 -2.28384554e-01
-3.13663334e-01 1.08363724e+00 3.20224494e-01 -3.44830006e-01
9.92915332e-01 9.97660235e-02 -3.72587144e-01 -5.29911220e-01
-9.56909359e-01 1.11661091e-01 9.31936428e-02 3.38245630e-01
5.23877680e-01 1.60314500e-01 -4.96593922e-01 2.94388179e-02
-1.27247795e-01 -2.72842526e-01 8.06787610e-01 2.96667308e-01
-3.85134518e-01 -5.97179294e-01 -2.14252487e-01 6.89733386e-01
-4.08789158e-01 -4.52001899e-01 -3.25795442e-01 5.10843635e-01
-2.39076912e-01 7.20382035e-01 3.45197141e-01 3.18331093e-01
5.23632765e-02 1.26293600e-01 7.13773370e-01 -4.29373264e-01
-1.54666170e-01 -1.30259898e-02 -9.51344371e-02 -3.94209146e-01
-2.75638402e-01 -1.22961676e+00 -1.10858202e+00 -4.95631844e-01
-3.67007375e-01 1.25660002e-01 8.20009172e-01 1.02959335e+00
3.01237136e-01 6.76243067e-01 2.43846416e-01 -8.72091949e-01
-4.82130438e-01 -1.04457140e+00 -7.20791757e-01 8.74476731e-02
2.23942623e-01 -6.50584638e-01 -6.80088401e-01 -1.31747842e-01]
|
[12.34194564819336, 3.4416794776916504]
|
f2dfed4b-f43e-4aaf-b779-023b5c99646a
|
towards-high-quality-and-efficient-video
|
2303.08331
| null |
https://arxiv.org/abs/2303.08331v2
|
https://arxiv.org/pdf/2303.08331v2.pdf
|
Towards High-Quality and Efficient Video Super-Resolution via Spatial-Temporal Data Overfitting
|
As deep convolutional neural networks (DNNs) are widely used in various fields of computer vision, leveraging the overfitting ability of the DNN to achieve video resolution upscaling has become a new trend in the modern video delivery system. By dividing videos into chunks and overfitting each chunk with a super-resolution model, the server encodes videos before transmitting them to the clients, thus achieving better video quality and transmission efficiency. However, a large number of chunks are expected to ensure good overfitting quality, which substantially increases the storage and consumes more bandwidth resources for data transmission. On the other hand, decreasing the number of chunks through training optimization techniques usually requires high model capacity, which significantly slows down execution speed. To reconcile such, we propose a novel method for high-quality and efficient video resolution upscaling tasks, which leverages the spatial-temporal information to accurately divide video into chunks, thus keeping the number of chunks as well as the model size to minimum. Additionally, we advance our method into a single overfitting model by a data-aware joint training technique, which further reduces the storage requirement with negligible quality drop. We deploy our models on an off-the-shelf mobile phone, and experimental results show that our method achieves real-time video super-resolution with high video quality. Compared with the state-of-the-art, our method achieves 28 fps streaming speed with 41.6 PSNR, which is 14$\times$ faster and 2.29 dB better in the live video resolution upscaling tasks. Code available in https://github.com/coulsonlee/STDO-CVPR2023.git
|
['Xiaolong Ma', 'Linke Guo', 'Fatemeh Afghah', 'Bin Ren', 'Wei Niu', 'Minghai Qin', 'Jie Ji', 'Gen Li']
|
2023-03-15
| null |
http://openaccess.thecvf.com//content/CVPR2023/html/Li_Towards_High-Quality_and_Efficient_Video_Super-Resolution_via_Spatial-Temporal_Data_Overfitting_CVPR_2023_paper.html
|
http://openaccess.thecvf.com//content/CVPR2023/papers/Li_Towards_High-Quality_and_Efficient_Video_Super-Resolution_via_Spatial-Temporal_Data_Overfitting_CVPR_2023_paper.pdf
|
cvpr-2023-1
|
['video-super-resolution']
|
['computer-vision']
|
[-6.21424317e-02 -1.58524171e-01 -3.43322545e-01 -2.60136753e-01
-7.35032082e-01 -2.17651546e-01 -1.20558769e-01 -4.74218220e-01
-4.28125888e-01 5.70401013e-01 2.10878581e-01 -2.38275439e-01
2.08510369e-01 -8.81606758e-01 -1.09680188e+00 -5.52962005e-01
-1.43743306e-01 -1.41638964e-01 5.66120505e-01 6.32768199e-02
-2.37513497e-01 9.67848077e-02 -1.49442685e+00 3.70405972e-01
9.05429542e-01 1.35033119e+00 6.14892006e-01 8.02068233e-01
9.84592810e-02 1.14236796e+00 -2.26701856e-01 -3.40721905e-01
4.92325544e-01 -2.89640967e-02 -4.55274224e-01 3.28735821e-02
7.25348651e-01 -1.35998511e+00 -1.08044600e+00 1.02570260e+00
3.69759113e-01 -6.67566666e-03 -3.70331556e-02 -9.51898694e-01
-5.83758891e-01 6.80077136e-01 -8.37617159e-01 4.87095743e-01
-8.81749243e-02 1.31321907e-01 8.38519335e-01 -6.35731459e-01
3.41186315e-01 1.05138159e+00 6.05202556e-01 6.58626258e-01
-9.91225719e-01 -1.22087216e+00 4.53055985e-02 3.77758324e-01
-1.25777292e+00 -8.80662203e-01 3.35071534e-01 -8.94455239e-02
8.55930567e-01 9.19865370e-02 5.80904186e-01 7.44076729e-01
-9.37468633e-02 7.29897141e-01 3.77756536e-01 1.17498480e-01
1.91725835e-01 -2.97955364e-01 -3.75114381e-01 5.60087204e-01
1.92461237e-01 -8.39346796e-02 -6.67415500e-01 3.47569346e-01
1.64359057e+00 2.29634240e-01 -6.07096016e-01 3.49670090e-02
-9.60789502e-01 5.81954479e-01 5.87629259e-01 -2.55955048e-02
-5.22503972e-01 4.21403646e-01 4.52620864e-01 2.05033153e-01
5.03704131e-01 -1.86935320e-01 -3.52149189e-01 -4.32432115e-01
-1.42125952e+00 1.41602606e-01 4.17829841e-01 1.20000482e+00
6.14364386e-01 1.99636593e-01 1.10352196e-01 9.42522347e-01
1.04114309e-01 5.38563132e-01 2.64475018e-01 -1.71376908e+00
9.41309750e-01 1.51291952e-01 9.01695639e-02 -8.38292956e-01
-2.26388592e-02 -2.87753701e-01 -1.04673707e+00 6.41867518e-02
2.63972223e-01 -1.83387682e-01 -6.04357183e-01 1.77379692e+00
8.37792158e-02 6.29514933e-01 -4.06315736e-02 1.36764014e+00
6.42783225e-01 9.97003436e-01 -1.81063205e-01 -5.50908566e-01
1.26157546e+00 -1.22759128e+00 -5.28974116e-01 -8.66693258e-02
2.88654327e-01 -6.06540084e-01 9.51444268e-01 4.33541715e-01
-1.58479512e+00 -6.06510878e-01 -1.11106789e+00 -2.77941287e-01
6.77227795e-01 -1.74683839e-01 4.87251759e-01 2.46351570e-01
-1.11541843e+00 6.70130432e-01 -1.10820413e+00 1.27837881e-01
8.42581332e-01 3.69709909e-01 -1.81791812e-01 -3.12664777e-01
-1.03380358e+00 1.22661196e-01 1.91525370e-01 -2.23116890e-01
-6.92759275e-01 -1.08456552e+00 -5.67943692e-01 3.19711208e-01
3.86359930e-01 -6.69527411e-01 1.43231297e+00 -1.05936074e+00
-1.48330247e+00 2.37492248e-01 -2.50681996e-01 -6.67798758e-01
6.47386372e-01 -4.28734779e-01 -3.39515179e-01 6.30778015e-01
-1.14682809e-01 8.44910026e-01 9.32150781e-01 -9.30746138e-01
-1.03918242e+00 -1.07072681e-01 1.86424136e-01 2.47256428e-01
-7.78715491e-01 -5.40752336e-02 -1.02864218e+00 -6.08715355e-01
7.47747496e-02 -5.73494852e-01 -4.46253307e-02 3.80273253e-01
1.57767594e-01 1.56258419e-01 8.18012238e-01 -8.27560365e-01
1.41217244e+00 -2.22579288e+00 -3.15466300e-02 -2.51752019e-01
7.81164467e-01 3.09441388e-01 -1.00054525e-01 -1.87688440e-01
1.90558419e-01 9.95365158e-02 7.87236616e-02 -3.82954508e-01
-3.28981161e-01 1.70073748e-01 -3.20905179e-01 1.42582700e-01
-1.20477416e-01 6.51303411e-01 -7.02013612e-01 -4.27702069e-01
4.29588296e-02 9.55194771e-01 -9.59035218e-01 2.98175275e-01
-6.14251792e-02 1.73123673e-01 -3.56111586e-01 4.53900725e-01
9.50276613e-01 -5.45366287e-01 2.35233992e-01 -5.14859259e-01
-1.68301370e-02 2.86988348e-01 -1.15774775e+00 1.53958142e+00
-6.47683859e-01 8.11344624e-01 3.72392774e-01 -7.48590708e-01
4.75701541e-01 3.86294216e-01 6.46682858e-01 -1.03940403e+00
-5.75228333e-02 1.91528499e-01 -4.16488796e-01 -5.39687455e-01
5.78142822e-01 1.04920045e-01 4.76306617e-01 1.15897089e-01
-1.60734683e-01 5.35666943e-01 2.07004383e-01 2.97763199e-01
1.22828233e+00 5.00025898e-02 -1.55314550e-01 1.75401732e-01
-5.42603098e-02 -3.88471246e-01 9.77154672e-01 3.26544851e-01
-2.36511901e-01 7.07492769e-01 4.25322950e-01 -4.12281781e-01
-1.56482577e+00 -1.14810514e+00 9.42889675e-02 9.83524024e-01
2.47723147e-01 -4.44046527e-01 -1.01933360e+00 -6.18044995e-02
-2.01819226e-01 9.77417007e-02 -8.08073953e-02 1.45589769e-01
-8.20891976e-01 -3.09865415e-01 4.51086670e-01 7.11288393e-01
1.00319254e+00 -6.10729694e-01 -7.51050532e-01 3.73458028e-01
-5.52459836e-01 -1.60247350e+00 -6.91205323e-01 -4.44187313e-01
-1.11005902e+00 -7.45404005e-01 -9.79261398e-01 -6.22551143e-01
3.72424811e-01 8.94832909e-01 1.10437608e+00 2.66408682e-01
-9.10676550e-03 -4.60109971e-02 -3.41464728e-01 3.01635176e-01
-1.03977956e-01 -1.42115699e-02 1.96923763e-01 -2.34789371e-01
1.43158948e-02 -1.02003610e+00 -1.07819355e+00 1.55452475e-01
-1.09002936e+00 4.99777913e-01 6.85283005e-01 5.68373859e-01
7.21925735e-01 3.43634903e-01 4.01607454e-01 -5.03811181e-01
7.20685199e-02 -5.59452653e-01 -7.81765819e-01 -1.75838575e-01
-4.75290745e-01 -1.44855857e-01 9.22317266e-01 -6.36698246e-01
-9.21803951e-01 -2.25596160e-01 -1.01424903e-01 -1.07637680e+00
2.93050498e-01 1.81016803e-01 -1.64188102e-01 2.39996724e-02
2.99755275e-01 3.10943693e-01 1.15672089e-01 -4.27029729e-01
1.71387747e-01 6.89091146e-01 6.88411534e-01 -1.31776094e-01
7.54411876e-01 4.40690249e-01 -2.07274944e-01 -7.25638211e-01
-5.77474773e-01 -2.01257288e-01 -5.88790001e-03 -2.63974577e-01
7.01585889e-01 -1.55353355e+00 -8.77513409e-01 4.48349506e-01
-9.01777625e-01 -6.84927821e-01 1.14320040e-01 5.39969504e-01
-5.60311079e-01 4.24960792e-01 -1.11655748e+00 -4.90370810e-01
-5.68200052e-01 -1.23904109e+00 8.17084491e-01 4.48603004e-01
3.65541101e-01 -4.76008415e-01 -4.55906957e-01 7.29687154e-01
6.32124484e-01 -2.69899398e-01 4.46786702e-01 2.13133276e-01
-1.11958921e+00 3.21460813e-01 -9.45279181e-01 3.64133000e-01
-9.24011245e-02 -2.15420976e-01 -7.89878368e-01 -5.60474575e-01
-1.53679237e-01 -2.20023811e-01 9.48778570e-01 4.70101684e-01
1.36533344e+00 -5.94885170e-01 -1.12632280e-02 1.11716735e+00
1.65337968e+00 -6.81766048e-02 8.79155993e-01 2.92633116e-01
8.14334273e-01 4.51459317e-03 5.74666917e-01 7.99381554e-01
6.25719726e-01 7.66738594e-01 6.49944007e-01 4.71773837e-03
-3.12456667e-01 -2.07849011e-01 4.84652281e-01 8.93377900e-01
-3.71238202e-01 -2.10902527e-01 -4.73347932e-01 3.23543459e-01
-1.92521966e+00 -1.14997816e+00 1.58120185e-01 2.32416034e+00
1.01997173e+00 2.25065514e-01 2.89440244e-01 1.15209008e-02
6.06148601e-01 2.72603840e-01 -7.54958153e-01 1.46352261e-01
1.45359322e-01 -1.04868338e-01 8.49806070e-01 4.29687500e-01
-9.03345525e-01 8.16458941e-01 5.17337418e+00 1.01465821e+00
-1.25451136e+00 1.46506414e-01 6.94841206e-01 -8.18458200e-01
-1.25245064e-01 -3.56728911e-01 -7.27547169e-01 8.14112663e-01
1.24898684e+00 -8.99755955e-02 1.00836432e+00 7.86484957e-01
7.37007022e-01 5.32016456e-02 -8.10946226e-01 1.29462898e+00
-1.81237772e-01 -1.71379054e+00 1.22722499e-01 2.53095686e-01
5.69182336e-01 2.28940979e-01 5.13486378e-02 1.10663176e-01
-8.79474506e-02 -8.14666629e-01 8.41015100e-01 3.59152108e-01
1.14710259e+00 -9.77470279e-01 5.04403651e-01 1.98741928e-01
-1.47484899e+00 -2.92215884e-01 -8.07328999e-01 8.24168511e-03
2.42687643e-01 8.23455095e-01 -1.93581522e-01 2.45164320e-01
1.28557229e+00 6.32234752e-01 -1.39535710e-01 9.03940856e-01
2.36082852e-01 7.70392060e-01 -3.34076375e-01 4.97056454e-01
-5.14546745e-02 1.33306421e-02 3.17609161e-01 9.98515368e-01
5.68951190e-01 4.98123527e-01 -1.53384283e-02 6.10731959e-01
-6.04347885e-01 -1.95329159e-01 -7.23435655e-02 1.72257304e-01
9.20190752e-01 1.18937242e+00 -3.39452505e-01 -3.43666583e-01
-7.02716231e-01 1.13870585e+00 3.05573016e-01 3.55463624e-01
-1.31034148e+00 -2.12711483e-01 1.04693508e+00 3.76296580e-01
6.40186369e-01 -3.71919930e-01 -1.41452119e-01 -1.36221755e+00
4.87094939e-01 -8.44807923e-01 -2.89798137e-02 -7.58367598e-01
-8.60160887e-01 6.07553244e-01 -3.81731868e-01 -1.46672189e+00
7.21214106e-03 -2.62852997e-01 -2.46044755e-01 5.45772076e-01
-1.78506935e+00 -8.71580660e-01 -7.00513840e-01 5.94373524e-01
8.11901093e-01 1.60171688e-02 2.45871559e-01 8.39887202e-01
-7.00526416e-01 8.86158288e-01 1.93728730e-01 1.67339444e-01
7.71268785e-01 -6.35794044e-01 3.36447597e-01 9.11116540e-01
-2.19648600e-01 2.90741831e-01 7.14251876e-01 -3.23632151e-01
-1.53025830e+00 -1.15623295e+00 4.60596681e-01 2.48235181e-01
6.65142953e-01 -2.64087498e-01 -1.15491509e+00 4.18390691e-01
1.85857713e-02 3.33233804e-01 4.22302157e-01 -3.55462044e-01
-6.24646723e-01 -5.18467069e-01 -9.21885610e-01 6.90903485e-01
1.23960006e+00 -2.95808285e-01 1.94909960e-01 1.69833928e-01
1.20504045e+00 -5.30885816e-01 -1.18092561e+00 2.47170642e-01
7.11739361e-01 -1.30098403e+00 1.01794851e+00 -4.65048291e-02
9.86209929e-01 -2.01521695e-01 -3.92760992e-01 -7.64572620e-01
-4.20339793e-01 -6.48326218e-01 -6.58518493e-01 1.11506033e+00
2.46194124e-01 -3.41639936e-01 1.00127316e+00 7.74865985e-01
2.81759221e-02 -1.00960994e+00 -9.78742301e-01 -7.66632974e-01
-1.45341039e-01 -4.39490020e-01 5.23310065e-01 5.99421740e-01
-1.15074925e-01 1.84863925e-01 -5.96571982e-01 3.46322060e-01
8.99648488e-01 -1.30389795e-01 6.34563863e-01 -7.87974596e-01
-5.69450498e-01 -3.33994806e-01 -3.41349185e-01 -1.75060749e+00
-8.72580484e-02 -2.64273256e-01 -1.57668531e-01 -1.44612086e+00
4.80775982e-01 -5.19030511e-01 -7.20487610e-02 4.11102921e-01
-1.21506400e-01 4.92287785e-01 5.17092884e-01 6.05243623e-01
-9.79771376e-01 4.54661578e-01 1.28038001e+00 2.88330793e-01
-1.60029665e-01 -1.18182644e-01 -6.93150342e-01 6.32182598e-01
6.82610214e-01 -3.52074690e-02 -5.70324719e-01 -9.85654831e-01
1.07091889e-01 5.80173254e-01 2.63432533e-01 -1.19510150e+00
3.40409487e-01 -1.45164534e-01 3.84688288e-01 -2.11163953e-01
4.87260759e-01 -9.14233029e-01 5.68037555e-02 1.94246233e-01
-8.68847296e-02 1.54340232e-03 1.19276963e-01 5.50669730e-01
-9.43769366e-02 2.97970831e-01 9.85987127e-01 1.19456202e-01
-7.77348042e-01 8.21106136e-01 -1.95038974e-01 -9.66574159e-03
8.11255634e-01 -2.50401378e-01 -5.42871356e-01 -7.39293993e-01
-2.91523010e-01 2.80210137e-01 7.02776790e-01 3.64000976e-01
7.75646448e-01 -1.28876638e+00 -7.66966045e-01 6.57481477e-02
-4.53010619e-01 2.98523754e-01 8.39436471e-01 8.54265034e-01
-7.50789881e-01 1.76616624e-01 -2.14027658e-01 -5.21908224e-01
-1.38521791e+00 4.86651659e-01 2.43789986e-01 -1.15655087e-01
-9.04351413e-01 7.88645327e-01 1.58655554e-01 4.38351423e-01
3.36131454e-01 -1.52694464e-01 -2.04074057e-03 -3.33468616e-01
9.88740385e-01 5.73154926e-01 -1.92027852e-01 -4.16354746e-01
-6.16486883e-03 3.94149274e-01 -2.88150489e-01 8.10816512e-02
1.49429417e+00 -5.32398760e-01 1.46802127e-01 -1.17500126e-01
1.23057270e+00 -1.34958804e-01 -2.12876177e+00 -4.65003401e-01
-6.29594445e-01 -7.37884820e-01 5.12968540e-01 -2.86139607e-01
-1.71687496e+00 6.28440738e-01 5.59533596e-01 9.60974693e-02
1.59209073e+00 -1.73064888e-01 1.60387433e+00 4.21954095e-02
5.03839850e-01 -9.61538136e-01 1.05160207e-01 2.46784896e-01
4.92409527e-01 -1.27411437e+00 2.84399744e-02 -4.62688893e-01
-5.84674597e-01 1.13893628e+00 6.92362249e-01 -2.05211625e-01
3.96180540e-01 6.71155989e-01 -1.15337156e-01 4.40588474e-01
-1.02301455e+00 1.98271871e-01 -1.70947030e-01 3.99965316e-01
3.37043196e-01 7.41598085e-02 2.76234839e-02 7.37462521e-01
-5.11115193e-02 3.58631134e-01 6.56976640e-01 5.41058600e-01
-7.16491342e-01 -7.94351161e-01 -9.07524824e-02 4.80166674e-01
-6.70895755e-01 -3.45991343e-01 6.25433862e-01 3.97662401e-01
-1.26889080e-01 9.99291718e-01 3.87539536e-01 -5.96212268e-01
1.52413681e-01 -6.54975772e-01 3.85304183e-01 -1.03117511e-01
-1.22234393e-02 2.38495231e-01 -2.48193126e-02 -1.03688121e+00
-3.32484961e-01 -3.28232020e-01 -1.45310938e+00 -1.11446309e+00
-1.68348216e-02 -2.49271005e-01 4.54648405e-01 7.09215820e-01
7.03146696e-01 6.44624352e-01 7.25476265e-01 -1.00601637e+00
-4.65513915e-01 -6.80346727e-01 -4.01224583e-01 -2.48412192e-02
4.61093813e-01 -3.49924326e-01 -2.34138221e-01 2.76985168e-01]
|
[11.04907512664795, -1.7668216228485107]
|
55335fb7-58db-4d33-8dcb-f34467cf3f5a
|
multi-resolution-networks-for-flexible
|
1905.00125
| null |
http://arxiv.org/abs/1905.00125v1
|
http://arxiv.org/pdf/1905.00125v1.pdf
|
Multi-resolution Networks For Flexible Irregular Time Series Modeling (Multi-FIT)
|
Missing values, irregularly collected samples, and multi-resolution signals
commonly occur in multivariate time series data, making predictive tasks
difficult. These challenges are especially prevalent in the healthcare domain,
where patients' vital signs and electronic records are collected at different
frequencies and have occasionally missing information due to the imperfections
in equipment or patient circumstances. Researchers have handled each of these
issues differently, often handling missing data through mean value imputation
and then using sequence models over the multivariate signals while ignoring the
different resolution of signals. We propose a unified model named
Multi-resolution Flexible Irregular Time series Network (Multi-FIT). The
building block for Multi-FIT is the FIT network. The FIT network creates an
informative dense representation at each time step using signal information
such as last observed value, time difference since the last observed time stamp
and overall mean for the signal. Vertical FIT (FIT-V) is a variant of FIT which
also models the relationship between different temporal signals while creating
the informative dense representations for the signal. The multi-FIT model uses
multiple FIT networks for sets of signals with different resolutions, further
facilitating the construction of flexible representations. Our model has three
main contributions: a.) it does not impute values but rather creates
informative representations to provide flexibility to the model for creating
task-specific representations b.) it models the relationship between different
signals in the form of support signals c.) it models different resolutions in
parallel before merging them for the final prediction task. The FIT, FIT-V and
Multi-FIT networks improve upon the state-of-the-art models for three
predictive tasks, including the forecasting of patient survival.
|
['Bryon Kucharski', 'Akhila Josyula', 'Madalina Fiterau', 'Rheeya Uppaal', 'Bhanu Pratap Singh', 'Iman Deznabi', 'Bharath Narasimhan']
|
2019-04-30
| null | null | null | null |
['irregular-time-series']
|
['time-series']
|
[ 3.33815396e-01 -1.83605120e-01 -2.33770683e-01 -3.99256736e-01
-5.51214278e-01 -2.05415800e-01 2.01578215e-01 1.33748502e-01
-3.83250713e-02 8.58401537e-01 7.79256523e-01 -1.03286862e-01
-9.38302815e-01 -8.87075782e-01 -5.15076697e-01 -6.79937661e-01
-4.98969615e-01 4.49270606e-01 -3.55619758e-01 -3.99402268e-02
-2.57942498e-01 5.46742976e-01 -1.23152101e+00 6.27404213e-01
7.04442680e-01 9.61254299e-01 -1.63975656e-01 3.28413993e-01
-1.61084175e-01 8.22681069e-01 -6.59902215e-01 1.99582338e-01
1.08620182e-01 -4.71132159e-01 -1.23925321e-01 -6.29785135e-02
-2.06806540e-01 -8.58790129e-02 -4.28783953e-01 5.53828955e-01
4.38993573e-01 7.86447152e-03 5.73669136e-01 -1.23582613e+00
-2.52190530e-01 9.07253504e-01 -6.84824586e-01 2.62415797e-01
2.57450879e-01 -4.83704470e-02 4.15251374e-01 -7.40461707e-01
4.96183246e-01 1.05606437e+00 1.37387049e+00 2.86046922e-01
-1.63429713e+00 -6.77059114e-01 -4.39503193e-02 4.14333828e-02
-1.39706135e+00 -3.57117742e-01 8.78048003e-01 -5.22389889e-01
7.69918859e-01 5.38666189e-01 7.84444988e-01 1.43594563e+00
5.37289798e-01 4.08490092e-01 7.63690889e-01 1.18612930e-01
1.27681836e-01 -5.34070730e-01 2.27486998e-01 9.69160944e-02
1.91579670e-01 3.21061224e-01 -5.21017849e-01 -6.81531847e-01
1.15212858e+00 1.08365273e+00 -4.99404103e-01 4.08447012e-02
-1.59292316e+00 7.65812397e-01 2.00704381e-01 5.99467576e-01
-9.70605850e-01 -5.55712692e-02 4.87935811e-01 3.75551075e-01
4.88094479e-01 1.28712550e-01 -5.90317905e-01 -6.90950379e-02
-1.23527384e+00 7.60407895e-02 7.70325541e-01 7.11214483e-01
4.11261141e-01 4.04681236e-01 -4.71471578e-01 7.27413177e-01
7.81797059e-03 7.59277195e-02 6.52078569e-01 -8.66439819e-01
5.16083539e-01 4.25129861e-01 3.29826698e-02 -1.16217232e+00
-1.05363214e+00 -7.67422795e-01 -1.84607220e+00 -4.45791930e-01
4.64795738e-01 -4.41452205e-01 -9.71931875e-01 1.75591779e+00
-4.40436862e-02 8.99015069e-01 -6.55034557e-02 5.62022150e-01
8.98075581e-01 6.23512387e-01 -1.16356246e-01 -7.90507317e-01
1.15990269e+00 -3.45809609e-01 -1.23848617e+00 1.07539639e-01
3.38213891e-01 -4.71747309e-01 2.76186705e-01 4.71621752e-01
-9.19244766e-01 -5.92542946e-01 -8.05587769e-01 3.18426639e-01
-1.21493533e-01 -1.63214490e-01 5.90275407e-01 7.29521737e-02
-7.94596970e-01 8.04406762e-01 -8.95646751e-01 1.30960003e-01
2.95100302e-01 3.23439002e-01 -4.38868850e-01 -3.84673357e-01
-1.32057571e+00 5.70202529e-01 -3.19545716e-02 3.84002596e-01
-4.91754740e-01 -1.06972516e+00 -9.69732344e-01 1.08231656e-01
1.28638402e-01 -1.08980215e+00 4.70861971e-01 -8.55629206e-01
-8.27414274e-01 2.25774527e-01 -3.56704384e-01 -6.27343833e-01
5.77086210e-01 1.39343664e-02 -9.23006952e-01 -1.84676215e-01
4.22191694e-02 7.16966018e-02 9.54309642e-01 -8.10463250e-01
-1.08181939e-01 -4.77353334e-01 -7.39842534e-01 -3.43198210e-01
4.20190364e-01 -3.65394324e-01 -1.47211343e-01 -1.18900490e+00
4.21574771e-01 -5.36401927e-01 -5.20984888e-01 -1.72203422e-01
-3.84362340e-01 1.05176367e-01 7.19521344e-01 -7.46786296e-01
1.59597754e+00 -2.39185047e+00 9.59442034e-02 4.43401486e-01
5.19887924e-01 -3.06661546e-01 -1.88951567e-01 6.78211391e-01
-5.66740811e-01 -1.46834012e-02 -4.04853493e-01 -4.89624590e-01
-5.21809697e-01 6.77566588e-01 -3.01448375e-01 4.86111730e-01
3.25934812e-02 8.08170438e-01 -6.39442623e-01 -1.69895098e-01
2.81317502e-01 9.96211886e-01 -3.41749698e-01 -1.11326408e-02
3.08761865e-01 8.55089784e-01 -2.44337767e-01 5.13461292e-01
7.35130548e-01 -3.71402442e-01 1.55899569e-01 -3.83622348e-01
4.58408729e-04 -4.88853119e-02 -1.44642794e+00 1.62867224e+00
-9.06383693e-02 1.79474652e-01 1.38006479e-01 -1.10264134e+00
1.21774316e+00 6.88201368e-01 1.19286776e+00 -5.72858453e-01
-6.83064461e-02 -3.13365124e-02 -6.13838248e-02 -5.79960346e-01
2.48267502e-01 -2.95709491e-01 -1.09851770e-01 1.21079825e-01
-1.76396310e-01 3.81163090e-01 -1.61686406e-01 -1.17373370e-01
1.31391621e+00 -7.17321038e-02 5.19123495e-01 9.37504985e-04
1.44849330e-01 -4.36393559e-01 1.22095454e+00 7.21243739e-01
3.88759151e-02 9.37908888e-01 5.10827065e-01 -9.93703485e-01
-7.76513219e-01 -1.03481579e+00 -4.75891113e-01 5.18936038e-01
-3.97394150e-01 -3.30834448e-01 -7.01725408e-02 -1.99555069e-01
3.67527813e-01 5.11871159e-01 -9.82473552e-01 -2.25965008e-02
-7.55960703e-01 -9.85056937e-01 3.93080592e-01 6.36084318e-01
8.27063620e-03 -9.92800772e-01 -3.73687536e-01 7.37973928e-01
-6.07510507e-01 -6.85012281e-01 -3.83314192e-01 5.93312919e-01
-1.22764337e+00 -9.35451865e-01 -6.17386878e-01 -1.98189840e-01
5.04352808e-01 -9.36230421e-02 1.25086498e+00 8.85603484e-03
-3.08023602e-01 2.07811132e-01 -2.23904759e-01 -3.45726967e-01
-4.11491022e-02 -3.15986961e-01 1.07295707e-01 2.76334763e-01
2.39160076e-01 -8.58269632e-01 -5.85094094e-01 2.77733505e-01
-1.11247599e+00 -9.81165543e-02 4.04127330e-01 1.16982079e+00
1.09269881e+00 1.14135839e-01 1.06279802e+00 -9.12172496e-01
6.92581832e-01 -1.17924953e+00 -8.64326060e-02 -9.73860174e-02
-3.18361938e-01 -1.07088849e-01 7.34575629e-01 -6.51997924e-01
-3.92992258e-01 1.14609487e-01 -1.72803819e-01 -8.48926544e-01
-1.94218084e-01 8.54263186e-01 2.13946834e-01 5.87973118e-01
5.36666095e-01 3.01085830e-01 4.08594936e-01 -7.02838898e-01
-1.01324409e-01 3.11049819e-01 6.46273017e-01 -2.50695676e-01
2.72609562e-01 3.93203348e-01 3.14965874e-01 -6.91550434e-01
-6.64554179e-01 -4.55417901e-01 -6.08560741e-01 -2.97368318e-02
6.07410848e-01 -9.80149150e-01 -5.38375497e-01 3.10202092e-01
-8.88922453e-01 9.89252776e-02 -7.48068154e-01 6.77344978e-01
-5.67636967e-01 1.12879768e-01 -5.71154356e-01 -6.94714010e-01
-3.50395828e-01 -8.87868106e-01 8.27199280e-01 -1.61422819e-01
-7.14127362e-01 -1.14188862e+00 5.01238592e-02 -8.62351656e-02
6.08267486e-01 1.04989433e+00 9.12564874e-01 -5.89572191e-01
-2.21443415e-01 -2.17330962e-01 1.20108180e-01 -7.12517723e-02
4.86176699e-01 -2.89953858e-01 -6.61368787e-01 -2.57086217e-01
3.99323553e-01 4.18298930e-01 8.36921155e-01 1.21065736e+00
1.54663503e+00 -5.10171354e-01 -4.41837460e-01 9.30076540e-01
1.25683355e+00 1.16074122e-01 8.52810323e-01 -1.92646012e-02
4.15427744e-01 4.47171539e-01 2.00782970e-01 9.14494216e-01
2.92797893e-01 4.34856862e-01 4.68905777e-01 -3.17844868e-01
1.84966296e-01 -3.80189507e-03 -6.91730380e-02 9.24331009e-01
-1.14968210e-01 1.10303752e-01 -6.94570184e-01 7.47882664e-01
-2.20338202e+00 -1.48975289e+00 -4.87848490e-01 2.29158068e+00
7.66723454e-01 -1.47336960e-01 2.94987649e-01 4.81748670e-01
5.24423420e-01 7.79897720e-02 -7.31291771e-01 -2.18137980e-01
-2.08190426e-01 2.31104523e-01 3.19686830e-01 2.44423658e-01
-8.55491936e-01 -2.62564868e-01 6.90513992e+00 4.57160562e-01
-9.88068938e-01 -1.25587210e-01 7.08425701e-01 -3.08938920e-01
-3.60014290e-01 -4.37663883e-01 -5.66485286e-01 7.90448129e-01
1.19381249e+00 -3.03256474e-02 3.65154624e-01 3.04112762e-01
6.28448963e-01 1.45183563e-01 -1.15187168e+00 1.18958187e+00
-9.78759900e-02 -1.55192065e+00 -5.87497093e-02 1.10784896e-01
4.83497143e-01 6.37739301e-02 -6.52613863e-02 1.58010140e-01
8.46854225e-02 -1.32177091e+00 1.90182149e-01 1.40356588e+00
7.36863136e-01 -8.42454195e-01 8.87309074e-01 5.66889942e-01
-1.31793678e+00 -2.31307402e-01 -7.93785155e-02 -8.37054476e-02
5.15258312e-01 1.19615972e+00 -4.84970093e-01 9.17590678e-01
7.61464775e-01 1.18624425e+00 -7.49954879e-02 1.22461593e+00
4.80866313e-01 6.56592786e-01 -4.04689670e-01 7.91446269e-01
-1.63927302e-01 -1.90203428e-01 5.38088500e-01 1.04421532e+00
7.60147214e-01 -8.42619240e-02 2.51834422e-01 8.52605641e-01
3.61586154e-01 -4.42920744e-01 -7.82598138e-01 3.23247790e-01
5.97688854e-01 9.10851240e-01 -2.71822959e-01 -2.31419638e-01
-4.83133882e-01 4.18597698e-01 -1.96552902e-01 4.65495348e-01
-6.55013025e-01 -7.73402527e-02 7.15292692e-01 4.25877571e-01
1.22883663e-01 -1.33094266e-01 -5.91621459e-01 -1.00670302e+00
-1.65528983e-01 -9.99768674e-01 9.03061390e-01 -5.57036817e-01
-1.81434917e+00 6.23287201e-01 -1.03932783e-01 -1.67382777e+00
-6.77808821e-01 7.57886991e-02 -3.82488132e-01 1.13550556e+00
-1.24443829e+00 -8.05284917e-01 -3.69830817e-01 1.21158779e+00
4.66078579e-01 4.77008242e-03 1.13514197e+00 5.42945445e-01
-5.29081762e-01 2.50012308e-01 1.94346026e-01 8.74886811e-02
6.07622087e-01 -7.01699078e-01 9.20478851e-02 4.31295395e-01
-1.08934745e-01 7.36707985e-01 5.21467328e-01 -8.34904969e-01
-1.16359973e+00 -1.27933526e+00 8.37777138e-01 -3.27336580e-01
4.74537104e-01 -6.98821992e-03 -1.28627980e+00 7.78211117e-01
-1.35612190e-01 3.66935104e-01 1.22280693e+00 2.40882412e-01
-8.18798095e-02 -1.87493101e-01 -1.39004457e+00 4.30730470e-02
7.36960471e-01 -2.54649758e-01 -6.30335450e-01 1.03604168e-01
5.36793232e-01 -4.73204195e-01 -1.38538122e+00 5.76189399e-01
5.11988342e-01 -8.18877399e-01 1.27180731e+00 -3.63678485e-01
1.62331000e-01 -3.98076355e-01 -2.21457034e-01 -1.27991557e+00
-8.47365856e-01 -6.52134717e-01 -5.42054117e-01 8.64850163e-01
2.83155233e-01 -8.28029692e-01 3.65722269e-01 6.54469192e-01
-9.28432271e-02 -8.53871465e-01 -1.28490365e+00 -5.32750249e-01
-3.83696824e-01 -5.76151729e-01 1.18711519e+00 1.41737974e+00
-1.40006483e-01 -7.44345933e-02 -9.18126941e-01 6.43298626e-02
6.33476377e-01 1.04403049e-01 3.62782925e-01 -1.78329444e+00
-4.05382395e-01 -1.52665302e-01 -2.69471198e-01 -6.21650279e-01
-3.32590491e-01 -7.00043201e-01 -4.08427089e-01 -1.60799778e+00
9.89173725e-02 -4.18043405e-01 -5.38091481e-01 7.36915171e-01
6.82499856e-02 -2.88786441e-02 -4.95575555e-02 5.54412425e-01
-4.01434489e-02 3.35071713e-01 1.22066081e+00 -2.29449361e-03
-5.43540537e-01 4.01431262e-01 -6.79106057e-01 6.18575275e-01
6.58504665e-01 -5.73588073e-01 -3.18209887e-01 -2.04252049e-01
1.46068424e-01 9.79469836e-01 4.52104896e-01 -1.00907004e+00
1.59587860e-01 -1.76931500e-01 1.01154912e+00 -1.01964176e+00
4.81759489e-01 -1.28293633e+00 1.08256698e+00 3.55502069e-01
-1.05309352e-01 4.79453146e-01 2.69122154e-01 7.09835052e-01
-4.21882868e-01 1.22355245e-01 5.21841347e-01 -1.54379755e-01
-1.58912778e-01 4.27538723e-01 -3.53796273e-01 -1.86567083e-01
8.10072839e-01 -4.37934369e-01 -2.88471490e-01 -5.44841528e-01
-1.44313526e+00 2.34100625e-01 -5.98844178e-02 4.23682511e-01
8.61959457e-01 -1.53420520e+00 -8.28459382e-01 7.64847994e-01
-1.82370633e-01 2.98205614e-01 6.59582853e-01 1.17727232e+00
1.38629332e-01 2.84224581e-02 -1.96107358e-01 -7.76992321e-01
-9.83344257e-01 5.88493407e-01 2.68239200e-01 -5.41757226e-01
-1.08353019e+00 1.68048337e-01 2.50240285e-02 -3.46445143e-01
1.62873209e-01 -6.26353979e-01 -4.25159574e-01 4.18910027e-01
7.64652252e-01 4.67141896e-01 -1.32798245e-02 -6.16829634e-01
-2.25564599e-01 5.13687074e-01 4.00065258e-02 3.49620610e-01
1.95715439e+00 -1.08586699e-02 -1.81334510e-01 7.79869258e-01
1.00628161e+00 -3.41697097e-01 -1.11372328e+00 -2.93311626e-01
-2.57236451e-01 -3.29704940e-01 -9.69334245e-02 -8.94749343e-01
-1.21419084e+00 5.15028954e-01 3.85352224e-01 3.12839776e-01
1.44073987e+00 -3.56904060e-01 5.44164062e-01 -2.44115978e-01
4.39037025e-01 -7.48803020e-01 -2.93299347e-01 5.28979778e-01
1.17010641e+00 -6.98680758e-01 6.85206894e-03 -1.23326458e-01
-5.85216522e-01 1.03204894e+00 4.87983003e-02 -1.93130840e-02
9.89643991e-01 6.29900157e-01 -6.08159751e-02 -1.36860684e-01
-8.38438272e-01 3.30139875e-01 4.11617070e-01 7.83930957e-01
5.03188908e-01 2.24018171e-01 -3.40834595e-02 1.09097481e+00
-1.17276326e-01 4.92954850e-01 3.30634147e-01 6.48222446e-01
8.64001662e-02 -9.19008195e-01 -6.91694498e-01 1.06018722e+00
-6.16772115e-01 6.59600794e-02 1.50984779e-01 5.02487242e-01
2.36006618e-01 8.63547087e-01 3.26048166e-01 -2.24945679e-01
5.52465796e-01 1.77930340e-01 -1.68843307e-02 -3.25807244e-01
-7.72988915e-01 4.55912620e-01 -1.71514973e-02 -8.32636178e-01
-3.17354590e-01 -7.94048667e-01 -1.18157470e+00 -3.22625160e-01
1.09451033e-01 -8.32742304e-02 3.67739916e-01 7.11099446e-01
7.61409760e-01 1.13938701e+00 5.59232652e-01 -8.27698290e-01
-3.99913758e-01 -1.04494476e+00 -8.75005722e-01 4.72460896e-01
9.62906063e-01 -5.74165285e-01 -3.89490485e-01 -9.79030803e-02]
|
[7.11093807220459, 3.2410898208618164]
|
261ae703-3b05-4f4e-97d9-e53743225c26
|
vmav-c-a-deep-attention-based-reinforcement
|
1812.09968
| null |
http://arxiv.org/abs/1812.09968v1
|
http://arxiv.org/pdf/1812.09968v1.pdf
|
VMAV-C: A Deep Attention-based Reinforcement Learning Algorithm for Model-based Control
|
Recent breakthroughs in Go play and strategic games have witnessed the great
potential of reinforcement learning in intelligently scheduling in uncertain
environment, but some bottlenecks are also encountered when we generalize this
paradigm to universal complex tasks. Among them, the low efficiency of data
utilization in model-free reinforcement algorithms is of great concern. In
contrast, the model-based reinforcement learning algorithms can reveal
underlying dynamics in learning environments and seldom suffer the data
utilization problem. To address the problem, a model-based reinforcement
learning algorithm with attention mechanism embedded is proposed as an
extension of World Models in this paper. We learn the environment model through
Mixture Density Network Recurrent Network(MDN-RNN) for agents to interact, with
combinations of variational auto-encoder(VAE) and attention incorporated in
state value estimates during the process of learning policy. In this way, agent
can learn optimal policies through less interactions with actual environment,
and final experiments demonstrate the effectiveness of our model in control
problem.
|
['Qi. Wang', 'Xingxing Liang', 'Zhong Liu', 'Yanghe Feng', 'Jincai Huang']
|
2018-12-24
| null | null | null | null |
['deep-attention', 'deep-attention']
|
['computer-vision', 'natural-language-processing']
|
[-3.44560653e-01 7.46723711e-02 -1.78100228e-01 1.37999654e-01
-3.78285237e-02 -8.63391235e-02 4.48148906e-01 -3.26106876e-01
-6.12044096e-01 1.15164971e+00 1.68518946e-01 -6.24923669e-02
-3.77255470e-01 -8.29414070e-01 -6.59694254e-01 -8.43744338e-01
-2.70268738e-01 7.00872540e-01 2.26608783e-01 -5.75361371e-01
2.47869670e-01 5.52493781e-02 -1.48204279e+00 -4.62276310e-01
8.73721659e-01 5.31701624e-01 9.57628369e-01 6.72399938e-01
-1.83353260e-01 1.47804081e+00 -5.95790982e-01 2.05340669e-01
2.60406405e-01 -3.41726899e-01 -5.91419876e-01 1.50022879e-01
-9.67600942e-01 -5.76964796e-01 -8.15330088e-01 9.63487446e-01
6.10880494e-01 6.95940971e-01 6.40141368e-01 -1.30805910e+00
-6.37355745e-01 6.36175334e-01 -6.06229901e-01 2.67130852e-01
-7.23327771e-02 5.19417524e-01 6.39081359e-01 -2.10291892e-01
2.68776059e-01 1.41519129e+00 8.72302577e-02 8.52604568e-01
-7.22801924e-01 -4.24662799e-01 5.74727058e-01 4.97255713e-01
-8.91370654e-01 -4.26018015e-02 6.29485309e-01 -4.85329665e-02
1.11329591e+00 -2.65493721e-01 9.51937973e-01 1.35069764e+00
4.93138820e-01 1.10230958e+00 1.12874579e+00 -1.56146064e-01
5.77085853e-01 -2.22925842e-02 -4.24285442e-01 6.48407519e-01
5.81791764e-03 6.58057511e-01 -3.05554420e-01 1.13809921e-01
1.22275305e+00 3.13872397e-01 -5.65581881e-02 -4.23389584e-01
-9.54156458e-01 8.49470317e-01 1.42819747e-01 -1.50578946e-01
-7.45023966e-01 4.12674487e-01 3.98021489e-01 5.84084928e-01
2.39032671e-01 5.21112442e-01 -6.32046402e-01 -6.13774419e-01
-3.01099509e-01 6.21406734e-01 6.96849525e-01 9.72205043e-01
5.19940078e-01 7.43396819e-01 -1.53409332e-01 8.09499383e-01
3.98598671e-01 6.54800951e-01 9.19457018e-01 -1.12600398e+00
3.10942769e-01 1.37288705e-01 5.13411701e-01 -6.20092809e-01
-5.43575406e-01 -5.37033498e-01 -7.09628224e-01 4.50744867e-01
-3.96893099e-02 -5.94423413e-01 -8.80616307e-01 1.86441481e+00
4.88233089e-01 6.36039317e-01 3.69734347e-01 8.78735006e-01
1.19680271e-01 7.94121802e-01 -6.52975738e-02 -5.66374302e-01
1.01554465e+00 -1.23862088e+00 -1.06560481e+00 -1.51546538e-01
1.94106668e-01 -2.67941386e-01 9.56624508e-01 3.85889322e-01
-1.08361959e+00 -6.39352798e-01 -8.67321968e-01 5.29776037e-01
-2.80722380e-01 -5.23464620e-01 6.33749962e-01 1.88526139e-01
-8.91928077e-01 6.41118169e-01 -1.18274701e+00 -2.49907449e-01
7.07126483e-02 6.35655880e-01 2.45692790e-01 3.14223349e-01
-1.43747330e+00 1.09514844e+00 5.33795476e-01 1.01806492e-01
-1.59058499e+00 -2.18740627e-01 -7.39513338e-01 1.39244944e-01
1.13344991e+00 -6.69038296e-01 1.72322869e+00 -9.66591597e-01
-2.28087473e+00 -2.87956089e-01 1.50412872e-01 -6.28960609e-01
3.82700890e-01 -2.30001584e-01 -3.30656767e-01 -1.82400569e-01
-1.06229097e-01 3.04177433e-01 9.13245738e-01 -1.18025708e+00
-9.07317638e-01 2.38028411e-02 2.73629874e-01 7.72395968e-01
-5.36941811e-02 -3.90671015e-01 -5.79193234e-02 -2.92585164e-01
-4.84561294e-01 -7.83899546e-01 -8.73850226e-01 -7.75292158e-01
1.54584810e-01 -4.45627511e-01 8.42957616e-01 -3.45453858e-01
1.20906472e+00 -1.75741339e+00 4.36495036e-01 -8.65142494e-02
7.72283152e-02 3.69943470e-01 -2.94424832e-01 6.35265172e-01
4.41087991e-01 -9.46214199e-02 7.55847469e-02 -1.36247903e-01
3.05085480e-01 7.92163491e-01 -3.54439169e-01 1.26890078e-01
-1.26711559e-02 8.84585261e-01 -1.13298166e+00 -3.79817516e-01
3.24808389e-01 1.18117496e-01 -6.88366294e-01 5.61956406e-01
-7.27838874e-01 5.69200277e-01 -1.02406931e+00 2.79014379e-01
3.33188951e-01 -3.56013216e-02 2.71968931e-01 4.82058525e-01
-1.62194267e-01 -1.16371494e-02 -1.28544700e+00 1.51093256e+00
-5.79018295e-01 -4.62062173e-02 2.30322108e-01 -1.33267558e+00
6.80068076e-01 4.81325209e-01 4.84902322e-01 -9.36428249e-01
2.30631545e-01 -8.22822601e-02 3.52259696e-01 -9.37635601e-01
6.54317856e-01 -1.42610013e-01 8.68466310e-03 5.11057615e-01
7.62613267e-02 -6.45296127e-02 7.62420893e-02 -8.74068290e-02
8.54471266e-01 6.77051365e-01 3.39669794e-01 -1.00122392e-01
2.14746699e-01 -9.32351127e-02 8.81865680e-01 1.03903854e+00
-4.03380841e-01 -1.01540595e-01 4.67268258e-01 -4.62078214e-01
-1.12300217e+00 -6.63050294e-01 4.71131682e-01 1.12719440e+00
2.97026068e-01 -1.45564407e-01 -7.04691768e-01 -4.20942873e-01
2.04987312e-03 5.93144059e-01 -7.03165591e-01 -2.05709636e-01
-5.20787418e-01 -7.76900172e-01 -9.39905457e-03 3.60780388e-01
6.62980556e-01 -1.58243597e+00 -8.97551954e-01 6.50334001e-01
1.60694838e-01 -7.34809339e-01 -2.54723459e-01 2.59664625e-01
-6.72317326e-01 -8.92172456e-01 -4.19952661e-01 -4.75529313e-01
2.96237558e-01 1.83196545e-01 9.38095629e-01 -1.38364688e-01
-3.63195911e-02 5.76111972e-01 -3.37561905e-01 -6.54482901e-01
-2.97926337e-01 1.74535334e-01 5.55249810e-01 -1.80576682e-01
9.11574438e-02 -5.33735931e-01 -5.93331456e-01 1.60698101e-01
-8.51041496e-01 9.75846052e-02 5.36206841e-01 1.27162158e+00
3.79636616e-01 3.05841118e-01 9.06246722e-01 -7.50164092e-01
1.08052337e+00 -8.43718469e-01 -8.61549139e-01 2.24728927e-01
-7.26056576e-01 4.23914343e-01 9.50648487e-01 -5.34896553e-01
-1.26289237e+00 -4.22535390e-01 3.80048566e-02 -5.38047910e-01
-4.72611636e-02 5.51865399e-01 -9.58741605e-02 4.92588490e-01
9.14769694e-02 5.72439313e-01 4.03186321e-01 -2.83907980e-01
1.06097475e-01 5.83436847e-01 -2.64468286e-02 -8.29972506e-01
4.26755220e-01 -2.22403482e-02 1.06833078e-01 -6.18983686e-01
-5.37587166e-01 -1.86220743e-02 -2.07480378e-02 -2.47975066e-01
7.70959258e-01 -8.11646938e-01 -1.28050482e+00 5.11870265e-01
-7.26292789e-01 -7.66975164e-01 -3.42751652e-01 7.21871316e-01
-1.21634650e+00 7.24314302e-02 -6.26975536e-01 -1.39788163e+00
1.44242018e-01 -1.30964792e+00 4.47739512e-01 6.09508932e-01
4.44323391e-01 -1.13570189e+00 3.45751196e-01 -4.70961444e-02
5.52370667e-01 -8.66860226e-02 7.22580373e-01 -5.05670547e-01
-5.69754481e-01 2.65372038e-01 3.15256536e-01 8.21101516e-02
2.08041057e-01 -2.84793228e-01 -6.31348312e-01 -4.42595094e-01
2.28699461e-01 -4.31477815e-01 4.43738639e-01 6.14458382e-01
1.21038306e+00 -3.70744526e-01 -1.45472199e-01 3.37029397e-01
1.44760132e+00 8.13600659e-01 2.73163766e-01 4.14559036e-01
4.27622765e-01 3.66482019e-01 6.88289523e-01 8.99327517e-01
5.23100853e-01 4.87753004e-01 8.17366302e-01 1.45149991e-01
4.84165788e-01 -5.53227782e-01 5.28689265e-01 9.45758939e-01
-3.71384859e-01 -3.50632578e-01 -6.23400509e-01 4.14107084e-01
-2.45570350e+00 -1.19808567e+00 3.96406978e-01 2.04081535e+00
5.05197167e-01 7.02283978e-02 2.68678814e-01 -4.17865962e-01
5.42693555e-01 2.39820436e-01 -1.12771821e+00 -3.88766617e-01
1.26056284e-01 -1.17809594e-01 4.61862147e-01 6.51413023e-01
-7.74680078e-01 1.19627631e+00 6.53271866e+00 9.81209457e-01
-8.72950852e-01 1.17146485e-01 4.65390712e-01 -2.75557548e-01
-6.83424324e-02 -1.89473063e-01 -5.79216242e-01 6.93326354e-01
1.09896195e+00 -2.08728299e-01 1.06266129e+00 9.36744928e-01
7.47893810e-01 -2.03691438e-01 -5.98548710e-01 8.40974867e-01
-3.35630119e-01 -1.04469192e+00 -2.46450365e-01 2.37704232e-01
7.05677330e-01 1.45491809e-01 2.85346776e-01 9.49208081e-01
1.01308036e+00 -1.03266990e+00 5.55831492e-01 7.09279001e-01
1.43071637e-01 -1.14877570e+00 8.55468750e-01 7.73811102e-01
-9.82668161e-01 -5.62698066e-01 -7.91540086e-01 -6.20092273e-01
1.25856787e-01 -2.71439493e-01 -7.32648015e-01 6.65463448e-01
3.04424614e-01 7.30041325e-01 4.92555872e-02 8.27920258e-01
-3.67233008e-02 5.67710936e-01 -1.97741855e-02 -5.39298952e-01
6.13671780e-01 -4.36187267e-01 5.44577181e-01 5.99105597e-01
2.10066468e-01 2.19557405e-01 7.03491688e-01 7.13155210e-01
3.64939332e-01 -1.25480518e-01 -7.41299450e-01 -1.20914862e-01
4.06473011e-01 9.87653136e-01 -5.90540409e-01 -2.19938293e-01
-3.55295628e-01 8.10368657e-01 5.61539114e-01 5.96794069e-01
-9.53050017e-01 -4.70742993e-02 7.37139344e-01 -2.62420684e-01
2.88847923e-01 -4.53857094e-01 4.35514539e-01 -1.14904547e+00
-4.12228078e-01 -1.04037750e+00 -7.06580514e-03 -5.55764675e-01
-1.16011739e+00 3.66260082e-01 -2.65574977e-02 -1.11969543e+00
-6.64988101e-01 -4.56965864e-01 -4.76182103e-01 7.09978878e-01
-1.66846275e+00 -6.19970202e-01 2.80363530e-01 7.58321524e-01
9.92535770e-01 -5.56952357e-01 6.69241190e-01 -1.20980725e-01
-8.08920324e-01 3.30601707e-02 6.68208957e-01 -6.55971467e-02
3.25294852e-01 -1.46193635e+00 1.09153248e-01 4.14452374e-01
-1.89014688e-01 3.64087671e-01 9.15025234e-01 -5.44939399e-01
-1.65917826e+00 -1.01447821e+00 -8.03622231e-02 -5.01830839e-02
6.85506940e-01 -1.42379373e-01 -6.83619916e-01 5.72084963e-01
5.07862031e-01 -2.77396947e-01 2.85259724e-01 7.79195055e-02
2.08968237e-01 7.36010373e-02 -8.68542254e-01 8.21471512e-01
8.86969030e-01 -1.32205129e-01 -4.99064773e-01 1.25039056e-01
9.78443980e-01 -5.22591710e-01 -5.32275736e-01 8.87963325e-02
4.84641403e-01 -7.31166542e-01 5.66956460e-01 -1.19550550e+00
2.52202541e-01 -3.10388416e-01 -7.62597620e-02 -1.80227077e+00
-5.07394075e-01 -1.02684009e+00 -4.40378755e-01 7.99511731e-01
1.59719080e-01 -5.76810360e-01 7.70334125e-01 4.27516103e-01
-2.23287642e-01 -8.10914040e-01 -8.26296151e-01 -7.61158347e-01
1.48011938e-01 -3.54501382e-02 7.96010613e-01 6.74866378e-01
1.23546764e-01 2.14319348e-01 -1.06432152e+00 1.00163147e-01
4.38519716e-01 -9.91605222e-02 6.97509944e-01 -7.95660257e-01
-7.84368455e-01 -3.58968288e-01 -9.54039919e-04 -1.26322210e+00
4.24894273e-01 -2.13590860e-01 2.46466056e-01 -1.50772583e+00
1.82055891e-01 -3.60847473e-01 -6.96971774e-01 1.34542063e-01
-3.67908508e-01 -7.63580561e-01 3.85942906e-01 8.58607739e-02
-1.04795539e+00 1.17906868e+00 1.74876547e+00 8.99453610e-02
-2.13422045e-01 1.24771237e-01 -4.50677544e-01 5.53097367e-01
9.78300750e-01 -3.61755013e-01 -1.05593324e+00 -4.93367940e-01
1.39835894e-01 7.04391301e-01 -1.04237750e-01 -8.59184325e-01
2.87726849e-01 -9.22120571e-01 6.50186613e-02 -1.06989577e-01
3.59009683e-01 -8.40557992e-01 -3.22400331e-02 8.14016044e-01
-3.81635219e-01 4.65866625e-01 2.23828461e-02 1.06457555e+00
-2.10251287e-02 -2.96840131e-01 4.41741616e-01 -6.06644094e-01
-9.02102232e-01 7.46244192e-01 -9.10706103e-01 7.92225525e-02
9.77433681e-01 3.12349498e-02 -4.51952918e-03 -6.95300102e-01
-7.78355420e-01 7.54789233e-01 7.76667520e-03 4.67307508e-01
4.60564911e-01 -1.11042798e+00 -4.75857645e-01 9.79165211e-02
-5.23790240e-01 9.05088261e-02 5.69332778e-01 4.97585028e-01
-2.19937742e-01 3.69766474e-01 -3.99509937e-01 -1.64996162e-01
-5.64314783e-01 8.96579802e-01 6.59829974e-01 -6.47466302e-01
-3.70631456e-01 4.59558100e-01 2.05907136e-01 -4.97101605e-01
2.88834214e-01 -1.86344951e-01 -4.13242459e-01 -1.95595056e-01
5.30884027e-01 3.75265926e-01 -5.48274577e-01 -7.71941990e-02
1.88091934e-01 1.25121409e-02 -1.89423978e-01 -4.74338502e-01
1.46253860e+00 -3.77114445e-01 2.99578846e-01 7.25613475e-01
3.29487383e-01 -5.93607724e-01 -1.86530221e+00 -3.38298470e-01
-1.33086249e-01 -1.26169950e-01 1.49187684e-01 -7.32195675e-01
-8.46926033e-01 7.97077000e-01 5.69468498e-01 4.48103666e-01
8.05661619e-01 -5.35003603e-01 6.33109510e-01 5.49265623e-01
6.52049363e-01 -1.61963534e+00 2.77503371e-01 9.73580182e-01
5.37567556e-01 -1.27502227e+00 -3.80311489e-01 4.25635338e-01
-9.96451080e-01 8.47511470e-01 9.88098145e-01 -4.90918726e-01
7.97588706e-01 4.10362482e-01 -6.07486330e-02 2.29246914e-02
-1.49795175e+00 -4.66713428e-01 -5.72094381e-01 9.52244878e-01
-3.67623195e-02 1.72395781e-01 -1.23544976e-01 6.68194354e-01
9.60798562e-02 8.52352977e-02 8.04994881e-01 1.14222419e+00
-8.39512467e-01 -1.14286876e+00 -1.22240439e-01 2.71405131e-01
-3.63170952e-01 1.07888184e-01 3.70340079e-01 7.49372661e-01
-5.82064688e-02 9.60633337e-01 1.56678557e-01 -2.36345276e-01
1.86705321e-01 -9.03267190e-02 4.15065318e-01 -5.11535466e-01
-5.68185747e-01 2.80483395e-01 -2.36934036e-01 -5.91538250e-01
-3.28726977e-01 -5.59184015e-01 -1.21663797e+00 -2.29675591e-01
-1.52434677e-01 5.68838239e-01 4.10738826e-01 9.96412456e-01
3.84901136e-01 9.94500518e-01 8.33712637e-01 -7.06171572e-01
-1.32347596e+00 -1.07457602e+00 -1.05883873e+00 4.42471355e-03
4.86472458e-01 -1.15536118e+00 -1.02183782e-01 -3.01995248e-01]
|
[3.8850739002227783, 2.0258281230926514]
|
fcdc847f-cd52-4592-93bd-ca3391ed35d6
|
an-efficient-approach-to-the-online-multi
|
2301.04446
| null |
https://arxiv.org/abs/2301.04446v1
|
https://arxiv.org/pdf/2301.04446v1.pdf
|
An Efficient Approach to the Online Multi-Agent Path Finding Problem by Using Sustainable Information
|
Multi-agent path finding (MAPF) is the problem of moving agents to the goal vertex without collision. In the online MAPF problem, new agents may be added to the environment at any time, and the current agents have no information about future agents. The inability of existing online methods to reuse previous planning contexts results in redundant computation and reduces algorithm efficiency. Hence, we propose a three-level approach to solve online MAPF utilizing sustainable information, which can decrease its redundant calculations. The high-level solver, the Sustainable Replan algorithm (SR), manages the planning context and simulates the environment. The middle-level solver, the Sustainable Conflict-Based Search algorithm (SCBS), builds a conflict tree and maintains the planning context. The low-level solver, the Sustainable Reverse Safe Interval Path Planning algorithm (SRSIPP), is an efficient single-agent solver that uses previous planning context to reduce duplicate calculations. Experiments show that our proposed method has significant improvement in terms of computational efficiency. In one of the test scenarios, our algorithm can be 1.48 times faster than SOTA on average under different agent number settings.
|
['Lujia Wang', 'Ming Liu', 'Hongji Liu', 'Yuanhang Li', 'Boyi Liu', 'Mingkai Tang']
|
2023-01-11
| null | null | null | null |
['multi-agent-path-finding']
|
['playing-games']
|
[ 6.79054409e-02 1.30242124e-01 -2.61652261e-01 1.45082861e-01
-3.78981620e-01 -7.28900850e-01 3.08925569e-01 4.82199878e-01
-4.61496204e-01 1.40353143e+00 -2.70743757e-01 -3.60938400e-01
-6.56266928e-01 -1.22697186e+00 -4.14113253e-01 -6.57339692e-01
-7.33115673e-01 1.07051039e+00 1.02520454e+00 -2.61551142e-01
4.27585930e-01 4.45710599e-01 -1.04944670e+00 -1.82726264e-01
9.18666363e-01 6.30367637e-01 4.61427212e-01 5.68019927e-01
-1.82361558e-01 7.14605451e-01 -6.10285580e-01 2.88697898e-01
5.65773129e-01 -3.71553034e-01 -1.15275049e+00 -7.95725435e-02
-7.48237550e-01 -6.50378764e-01 -2.10783020e-01 8.44226182e-01
1.71998799e-01 3.00702393e-01 2.62393169e-02 -2.11211491e+00
2.82173604e-01 8.43851030e-01 -9.09070790e-01 2.49475732e-01
6.83564842e-01 6.16823770e-02 4.12288547e-01 -2.78826356e-01
1.02063560e+00 1.11390638e+00 3.86209548e-01 1.03603713e-01
-6.94880307e-01 -4.28084791e-01 7.13767767e-01 6.28347754e-01
-1.43304169e+00 1.03298105e-01 3.71336997e-01 1.28496334e-01
1.37096131e+00 3.47565204e-01 7.81351149e-01 3.59535925e-02
6.92104697e-01 1.93830997e-01 1.15454018e+00 -4.70431983e-01
5.58218956e-01 -2.92301416e-01 1.09855860e-01 5.14491200e-01
4.64581400e-01 3.93992454e-01 -5.03940582e-01 -3.82967412e-01
5.57159245e-01 -2.10318059e-01 -2.07909252e-02 -2.92965353e-01
-1.30628347e+00 7.41726696e-01 1.85054585e-01 4.56078462e-02
-6.11320138e-01 2.19559044e-01 3.16510677e-01 3.83617580e-01
-1.13502912e-01 1.65238395e-01 -3.12710434e-01 -1.90798134e-01
-4.56111252e-01 7.45734394e-01 1.06677353e+00 1.16913033e+00
7.47585833e-01 -2.98560053e-01 1.73846453e-01 1.44195825e-01
2.73350149e-01 3.57376873e-01 -2.27080598e-01 -1.45373034e+00
5.40146947e-01 6.33859873e-01 6.73257649e-01 -1.05729938e+00
-8.30930293e-01 -1.66784167e-01 -2.26217791e-01 6.12219572e-01
2.53369480e-01 -1.58905700e-01 -6.64078116e-01 1.68553889e+00
9.49155331e-01 3.26561540e-01 2.74986744e-01 8.24989021e-01
2.25876570e-01 1.06671011e+00 -3.54484916e-01 -9.58068073e-01
1.10305250e+00 -1.40823483e+00 -7.95651615e-01 -3.58827144e-01
6.26704574e-01 -6.76325619e-01 1.10373266e-01 3.96171957e-01
-1.38484812e+00 3.40120673e-01 -1.32313049e+00 5.47563732e-01
-5.70550919e-01 -7.70259559e-01 7.34036326e-01 2.34757200e-01
-1.14693236e+00 3.09422731e-01 -8.83697748e-01 -3.95455450e-01
-2.08516479e-01 6.09602690e-01 -2.15722978e-01 -3.40160578e-01
-1.06464183e+00 1.00812912e+00 7.46385396e-01 -2.96128588e-03
-1.01968431e+00 -2.41149366e-01 -6.39633119e-01 -5.79678714e-02
1.33121037e+00 -6.84984028e-01 1.34434032e+00 -3.05814385e-01
-1.52218878e+00 -1.99293508e-03 -2.11918384e-01 -3.51178259e-01
3.55763972e-01 5.74606776e-01 -2.11919859e-01 -7.14641735e-02
4.23219234e-01 4.06022042e-01 5.46612479e-02 -1.46372890e+00
-1.17729843e+00 -2.21071228e-01 5.56528926e-01 5.90709329e-01
4.33402568e-01 1.21758819e-01 -5.14336407e-01 2.00459175e-02
2.66402096e-01 -1.15667868e+00 -9.10892963e-01 -4.73310500e-01
-4.91984934e-02 -2.81580865e-01 5.74457109e-01 -3.02187890e-01
1.48897207e+00 -1.64126492e+00 2.38555938e-01 6.22878909e-01
-9.26155522e-02 -2.99749911e-01 -1.87463552e-01 1.23334956e+00
3.77509385e-01 -1.78851962e-01 -2.94214282e-02 2.27185845e-01
-8.12759250e-02 5.68322062e-01 -4.86182719e-02 3.30850780e-01
-5.71887314e-01 3.41195196e-01 -1.13572633e+00 -6.41947210e-01
-1.28881419e-02 -2.49457285e-01 -4.69750077e-01 -2.06076711e-01
-3.99673939e-01 -6.92806393e-02 -5.82310438e-01 4.57584828e-01
1.12408447e+00 -6.28893822e-02 6.43080771e-01 5.65512598e-01
-9.89317358e-01 3.04667830e-01 -1.76618826e+00 1.73158073e+00
-9.63966027e-02 8.98057893e-02 3.28770220e-01 -4.66230899e-01
4.02393579e-01 4.88920044e-03 6.12171173e-01 -9.27272201e-01
-5.37537672e-02 3.35984528e-01 1.52576312e-01 -1.03671513e-01
5.84418595e-01 4.37258244e-01 -3.10790896e-01 9.75241542e-01
-8.71152103e-01 2.19293550e-01 8.41253817e-01 5.28195500e-01
1.55327809e+00 4.83387373e-02 5.70251703e-01 -6.74901307e-02
6.19724333e-01 8.76629055e-01 1.00401461e+00 8.91084611e-01
-4.36116308e-01 -5.33436596e-01 4.48555529e-01 -7.90952146e-01
-5.47128379e-01 -9.14280236e-01 5.84691763e-01 7.95339644e-01
1.13383043e+00 -5.83015144e-01 -5.34732878e-01 -2.15766579e-01
-1.72446594e-01 7.53999829e-01 -4.12820399e-01 2.94992894e-01
-9.48704243e-01 -5.30941665e-01 -2.07780898e-01 1.87243134e-01
6.44877255e-01 -9.09909904e-01 -1.21386766e+00 8.43809247e-01
-3.03150356e-01 -1.03014612e+00 -5.92069983e-01 -2.24474430e-01
-3.83872211e-01 -1.29539084e+00 1.23476021e-01 -7.49213278e-01
7.68903852e-01 8.69988441e-01 5.89424491e-01 3.47378999e-01
2.86156498e-02 2.03805521e-01 -5.02019763e-01 -4.08699006e-01
-2.15611935e-01 -1.33876652e-01 3.55835594e-02 -7.15260983e-01
5.62221603e-03 -4.35717583e-01 -4.56362396e-01 5.41018009e-01
-4.01801109e-01 2.89731592e-01 3.82297933e-01 5.38378537e-01
8.37199390e-01 9.36498404e-01 6.53317988e-01 -6.03482425e-01
7.37487555e-01 -5.67469597e-01 -1.01661813e+00 3.74479562e-01
-7.20724165e-01 -2.96345562e-01 6.40324593e-01 -2.90204175e-02
-9.11652267e-01 -3.82930301e-02 6.50881171e-01 1.80039853e-01
2.75293231e-01 7.65117526e-01 1.45568445e-01 -2.41185054e-01
1.03751257e-01 2.33676776e-01 -4.82233539e-02 1.92962810e-01
-3.49604525e-02 6.56129569e-02 4.23756629e-01 -4.14136708e-01
8.10068250e-01 4.09183472e-01 5.69607317e-01 -6.17481582e-02
-1.86417922e-02 -2.78121918e-01 -1.21646479e-01 -4.84106064e-01
3.13570708e-01 -5.02242148e-01 -1.20494449e+00 3.10832292e-01
-1.32831979e+00 -6.04541302e-01 2.21320868e-01 2.10982606e-01
-6.37077570e-01 1.39321566e-01 -4.13738430e-01 -1.22833896e+00
-1.51391968e-01 -1.09059513e+00 3.05817723e-01 4.85185087e-01
1.55651160e-02 -5.78148246e-01 3.42933744e-01 -1.62744371e-04
3.12698960e-01 8.12895775e-01 6.09660864e-01 -3.49139482e-01
-1.16310692e+00 1.78041220e-01 -1.21771380e-01 -1.02364445e+00
-4.60352600e-02 -1.71517029e-01 2.01104850e-01 -5.93945563e-01
-3.32093209e-01 3.24808180e-01 -4.98101711e-02 3.51470590e-01
4.43214238e-01 -4.62338597e-01 -1.03339386e+00 -6.13390170e-02
1.68153429e+00 1.18101096e+00 5.00315964e-01 1.08205879e+00
-8.74395669e-02 6.19194329e-01 1.48556960e+00 7.33890772e-01
9.85275149e-01 8.73820186e-01 6.65503204e-01 2.96791643e-01
1.07979700e-01 1.08679496e-01 4.04860646e-01 4.48328614e-01
-1.62974715e-01 -4.91448939e-01 -1.06820369e+00 5.61586320e-01
-2.47488713e+00 -9.94287491e-01 -1.79278210e-01 1.99869359e+00
4.84130055e-01 3.18742454e-01 3.99938494e-01 1.42696902e-01
7.49447286e-01 -2.80392677e-01 -8.19577813e-01 -8.30148101e-01
3.31505150e-01 -3.63816231e-01 8.45224977e-01 1.02514613e+00
-5.58944523e-01 9.00084913e-01 5.82289028e+00 3.38502586e-01
-5.96901298e-01 1.96592823e-01 -2.12187424e-01 -2.13108286e-01
-1.59911998e-02 3.84977460e-01 -5.33574045e-01 5.63487709e-01
9.60001707e-01 -8.49178612e-01 1.00349414e+00 7.57385135e-01
6.63826346e-01 -8.10513973e-01 -8.35324943e-01 5.81906915e-01
-1.68668941e-01 -1.47821152e+00 -5.43043911e-01 2.85600126e-01
7.10210383e-01 -2.70990938e-01 -4.81766194e-01 3.55026573e-02
5.44243038e-01 -5.71541607e-01 7.91917205e-01 1.88829333e-01
8.85355771e-02 -1.41212487e+00 8.59886885e-01 6.75952554e-01
-1.67406511e+00 -3.29844654e-01 -1.58504874e-01 -5.41414738e-01
6.93217993e-01 -1.00374028e-01 -1.13273871e+00 1.11950076e+00
7.67925799e-01 6.65359721e-02 3.00119221e-01 1.36458516e+00
1.05088472e-01 -3.22476983e-01 -6.23209238e-01 -1.33283153e-01
5.83886564e-01 -4.90358979e-01 8.86854529e-01 4.98081535e-01
1.91757366e-01 6.33875668e-01 8.20078194e-01 5.58263063e-01
5.70141256e-01 -1.34374812e-01 -3.48685563e-01 2.60501385e-01
1.19686890e+00 9.48738277e-01 -1.15531445e+00 -1.56908020e-01
-2.18276441e-01 5.09736478e-01 1.34095266e-01 3.47276658e-01
-1.07977569e+00 -5.47852039e-01 4.59438622e-01 1.41497422e-02
-6.59202561e-02 -5.79410732e-01 -2.03087017e-01 -1.46010816e-01
-5.75595684e-02 -8.10286343e-01 6.13222599e-01 -6.55607283e-01
-4.16372210e-01 5.56860089e-01 3.22284758e-01 -1.06733465e+00
-3.77387255e-01 -3.84622514e-02 -7.67954588e-01 6.19216263e-01
-1.58852744e+00 -8.32654536e-01 -5.08751392e-01 6.09152436e-01
8.28706563e-01 4.63966541e-02 5.67485809e-01 1.13226034e-01
-7.63262749e-01 1.59708396e-01 -1.21718228e-01 -6.86201632e-01
1.21912189e-01 -7.79163897e-01 2.71364907e-03 1.18445063e+00
-7.99504817e-01 4.95289713e-01 9.15886104e-01 -1.03659999e+00
-1.83666158e+00 -7.64456809e-01 6.20518208e-01 2.77337581e-01
4.92844403e-01 1.85827211e-01 -4.11963046e-01 8.08950126e-01
3.63493621e-01 -6.12573087e-01 5.23357868e-01 -3.95793825e-01
2.89002419e-01 -1.41293570e-01 -1.36703312e+00 8.24632347e-01
1.09344399e+00 5.09610653e-01 -1.82194665e-01 4.11649346e-01
8.32701981e-01 -6.63618386e-01 -4.59877461e-01 1.81669280e-01
3.64339441e-01 -8.09012711e-01 6.69916868e-01 -2.24781856e-01
-1.65875897e-01 -9.82757926e-01 4.37637381e-02 -1.30972171e+00
-5.38609743e-01 -8.73821557e-01 4.94649867e-03 7.57615924e-01
4.34903085e-01 -1.21528411e+00 6.82727039e-01 8.18414032e-01
-4.29716080e-01 -8.45385134e-01 -1.19020021e+00 -1.17975521e+00
-3.44943792e-01 -8.14100951e-02 1.10257351e+00 8.01653326e-01
4.81617600e-01 -2.34459192e-02 -2.65342623e-01 7.81622887e-01
8.80408287e-01 5.22284150e-01 9.25919831e-01 -8.09829235e-01
-2.76899427e-01 -3.12191486e-01 1.39984846e-01 -5.92340589e-01
3.24579887e-02 -5.03418505e-01 1.21002980e-01 -2.18863130e+00
2.12700114e-01 -7.37989962e-01 9.32360534e-03 6.22387886e-01
4.35509503e-01 -5.41125655e-01 4.55991268e-01 2.59960294e-01
-9.20199215e-01 4.63103913e-02 1.38017440e+00 4.35214154e-02
-6.55870199e-01 -2.68398315e-01 -3.17241699e-01 5.97630620e-01
9.47283804e-01 -6.23563051e-01 -7.24959254e-01 -4.95191514e-01
2.97987014e-01 7.97281563e-01 -2.46141359e-01 -7.95766532e-01
9.43915546e-01 -1.11373866e+00 -4.02586490e-01 -1.04357672e+00
3.67566943e-01 -1.03555667e+00 9.28122461e-01 1.18169451e+00
1.71043545e-01 4.89270806e-01 2.93135077e-01 5.51269770e-01
-7.04661198e-03 -3.64081711e-01 2.67867655e-01 -2.16982067e-01
-1.00471091e+00 2.29663730e-01 -6.45369828e-01 -4.71134365e-01
1.88128257e+00 -5.21375418e-01 -8.59547317e-01 -2.55849928e-01
-3.91604573e-01 1.17226255e+00 5.11610866e-01 -5.73653840e-02
7.11848497e-01 -1.14052105e+00 -4.23584789e-01 -3.93357426e-01
-3.64540309e-01 4.92628738e-02 3.00651163e-01 9.94565189e-01
-9.67552900e-01 4.86178577e-01 -4.06993419e-01 -5.00972085e-02
-1.11871588e+00 8.26772392e-01 2.05686077e-01 -6.67091131e-01
-4.67588723e-01 4.71944392e-01 -2.59389937e-01 4.66066785e-02
-9.61786602e-03 1.18947424e-01 -1.47799432e-01 -1.94712564e-01
8.18063557e-01 1.12496805e+00 -2.77519673e-01 -2.48935178e-01
-1.03114021e+00 5.34174562e-01 -1.85021609e-01 -4.80992734e-01
1.31333899e+00 -5.57188869e-01 -6.14679396e-01 -2.16699928e-01
3.10674220e-01 -2.98964754e-02 -8.55274796e-01 1.15111656e-01
7.48815760e-02 -7.77297616e-01 1.26052812e-01 -1.03674436e+00
-6.68106854e-01 -2.49210328e-01 4.29003648e-02 5.02889156e-01
1.22318184e+00 -3.06565940e-01 8.74851525e-01 3.61587554e-01
1.49233246e+00 -1.27813041e+00 -3.38160813e-01 7.35049009e-01
7.47698426e-01 -6.91461980e-01 2.30246603e-01 -1.12214494e+00
-5.40183902e-01 9.76322353e-01 9.30100441e-01 -8.38400647e-02
9.85917151e-02 5.79428732e-01 -3.93867046e-01 -2.83975098e-02
-1.14985979e+00 -1.42612576e-01 -1.00138879e+00 7.36338258e-01
-8.12934577e-01 1.57926306e-01 -8.82690191e-01 5.11444807e-01
-1.93423748e-01 1.36384055e-01 1.33551133e+00 1.49526227e+00
-7.90889442e-01 -1.22486496e+00 -6.96593344e-01 -5.09318709e-02
3.23766887e-01 5.20453215e-01 -3.03919256e-01 1.00519502e+00
7.82445371e-02 1.38801599e+00 1.94566384e-01 -1.28142789e-01
3.72051567e-01 -4.31672484e-01 5.65017164e-01 -3.22046012e-01
-3.21125627e-01 8.15161914e-02 7.48644769e-01 -9.00698483e-01
-3.34265709e-01 -6.93889260e-01 -2.10471153e+00 -8.10380578e-01
-1.03867009e-01 4.62211698e-01 5.49887419e-01 6.98525250e-01
3.33757222e-01 5.70460558e-01 7.97702014e-01 -7.31173575e-01
-1.35961846e-01 -1.69710845e-01 -1.76304311e-01 -5.69683075e-01
-7.20485859e-03 -1.07861745e+00 -1.60633132e-01 -5.77536881e-01]
|
[4.959190845489502, 1.8060814142227173]
|
9bbf943e-3eac-4d79-8bd8-246d5ec99f61
|
lone-pine-at-semeval-2021-task-5-fine-grained
|
2104.03506
| null |
https://arxiv.org/abs/2104.03506v1
|
https://arxiv.org/pdf/2104.03506v1.pdf
|
Lone Pine at SemEval-2021 Task 5: Fine-Grained Detection of Hate Speech Using BERToxic
|
This paper describes our approach to the Toxic Spans Detection problem (SemEval-2021 Task 5). We propose BERToxic, a system that fine-tunes a pre-trained BERT model to locate toxic text spans in a given text and utilizes additional post-processing steps to refine the boundaries. The post-processing steps involve (1) labeling character offsets between consecutive toxic tokens as toxic and (2) assigning a toxic label to words that have at least one token labeled as toxic. Through experiments, we show that these two post-processing steps improve the performance of our model by 4.16% on the test set. We also studied the effects of data augmentation and ensemble modeling strategies on our system. Our system significantly outperformed the provided baseline and achieved an F1-score of 0.683, placing Lone Pine in the 17th place out of 91 teams in the competition. Our code is made available at https://github.com/Yakoob-Khan/Toxic-Spans-Detection
|
['Soroush Vosoughi', 'Weicheng Ma', 'Yakoob Khan']
|
2021-04-08
| null |
https://aclanthology.org/2021.semeval-1.132
|
https://aclanthology.org/2021.semeval-1.132.pdf
|
semeval-2021
|
['toxic-spans-detection']
|
['natural-language-processing']
|
[ 2.43883625e-01 -1.23695277e-01 1.56324446e-01 -1.36970878e-01
-1.25814569e+00 -7.20809937e-01 4.99542117e-01 5.57212710e-01
-6.93109870e-01 6.69052541e-01 1.80371732e-01 -3.78203958e-01
1.48460492e-01 -5.10755062e-01 -8.07663977e-01 -4.96172071e-01
-4.66496386e-02 4.21054900e-01 4.84049797e-01 -7.62289315e-02
7.45665193e-01 2.04610139e-01 -1.27675116e+00 6.23519540e-01
8.60221744e-01 8.21866691e-01 2.07035959e-01 1.00254571e+00
9.30115655e-02 6.64764285e-01 -8.09827030e-01 -5.29785275e-01
1.68267682e-01 -1.76820487e-01 -8.53122175e-01 -1.44869894e-01
4.52612698e-01 1.71600673e-02 -1.79102376e-01 7.43743479e-01
5.55793166e-01 -4.26625758e-02 7.62327671e-01 -9.93993342e-01
-5.76572895e-01 1.02049279e+00 -9.55846071e-01 5.55547774e-01
2.75734328e-02 1.72454149e-01 1.08533525e+00 -1.07907724e+00
3.47304970e-01 9.72243488e-01 8.76959801e-01 3.58820885e-01
-1.07263076e+00 -6.38432264e-01 1.23737283e-01 3.63948494e-01
-1.47921777e+00 -4.18369114e-01 2.44463235e-01 -5.46277285e-01
1.06026268e+00 4.01214153e-01 3.08198512e-01 1.03745508e+00
2.98665315e-01 7.79463708e-01 9.84349370e-01 -6.52937591e-01
1.76845372e-01 -3.09853047e-01 4.06700969e-01 5.12343347e-01
2.10897133e-01 -2.72770941e-01 -7.59609520e-01 -7.21048117e-02
-1.20522335e-01 -3.42509329e-01 -5.72608188e-02 4.49735016e-01
-9.86150622e-01 6.99452043e-01 2.73563236e-01 2.06086397e-01
-3.47819626e-01 2.27309570e-01 7.25805640e-01 -6.06497824e-02
7.82505274e-01 5.72281480e-01 -5.44199288e-01 -2.52062649e-01
-1.06453943e+00 2.35900789e-01 5.76431632e-01 8.72353554e-01
1.76017329e-01 -4.38204676e-01 -5.45627534e-01 9.73178387e-01
-3.71233486e-02 2.69799620e-01 3.61458272e-01 -5.48550546e-01
6.27095461e-01 4.64607537e-01 4.34875116e-02 -3.44389647e-01
-3.19998831e-01 -3.71292233e-01 -3.31031919e-01 -1.48438558e-01
4.42648470e-01 -2.68274188e-01 -1.19922721e+00 1.43132162e+00
8.79920572e-02 -6.12184405e-02 -2.72692382e-01 4.35283989e-01
4.72848862e-01 7.08445251e-01 3.88143569e-01 -3.41174416e-02
1.52598834e+00 -1.09185898e+00 -3.91449779e-01 -1.68494791e-01
8.62224340e-01 -1.08090317e+00 1.33877456e+00 6.65215254e-01
-9.48167861e-01 -1.42474532e-01 -1.10830414e+00 -7.16206431e-02
-5.08597791e-01 2.25928351e-01 1.41069414e-02 7.08935678e-01
-7.92864382e-01 5.92066765e-01 -8.11014354e-01 -4.00386274e-01
3.44930172e-01 6.15435913e-02 -1.04494259e-01 -7.48267919e-02
-1.02509630e+00 8.42387915e-01 4.26088810e-01 -1.19121060e-01
-1.05041766e+00 -7.44325638e-01 -5.77357709e-01 -3.46066267e-03
5.37704706e-01 -3.55127484e-01 1.59772980e+00 -2.69071192e-01
-7.40103722e-01 8.28822553e-01 -2.12240905e-01 -4.28922594e-01
6.74974144e-01 -7.79065251e-01 -2.90757686e-01 2.72868364e-03
3.63423556e-01 4.33194160e-01 2.45835438e-01 -9.81830716e-01
-7.51549721e-01 -3.68182957e-01 -1.87497213e-01 1.53956890e-01
-3.89279962e-01 3.87001991e-01 -7.57465780e-01 -6.03876948e-01
-2.17003599e-01 -9.18878317e-01 -1.47741899e-01 -5.22182286e-01
-1.02559876e+00 -4.30716604e-01 5.04034162e-01 -8.87869537e-01
1.53687525e+00 -1.93544924e+00 -3.99493784e-01 -6.10216707e-03
1.18349239e-01 1.58557057e-01 -2.57940382e-01 7.83200681e-01
-2.96185613e-02 6.22185647e-01 -2.72880524e-01 -6.39279664e-01
4.98669073e-02 -3.63494188e-01 -3.30223680e-01 4.71049637e-01
2.52585500e-01 4.50827330e-01 -7.72941053e-01 -4.60439771e-01
-1.53926998e-01 3.05231154e-01 -2.64023453e-01 -5.23305088e-02
-3.96399647e-01 1.99367069e-02 -2.20462948e-01 6.31929934e-01
6.01783454e-01 1.08497985e-01 -1.29417747e-01 1.56682149e-01
-4.18034494e-01 7.21214354e-01 -8.29597354e-01 1.35478616e+00
-3.08055758e-01 6.95427001e-01 -3.15351337e-01 -4.22017604e-01
7.22175002e-01 1.76279724e-01 2.34034002e-01 -5.07891774e-01
1.40796095e-01 2.05905482e-01 -1.10994339e-01 -4.93820161e-01
8.04103732e-01 8.82626846e-02 -3.32098573e-01 4.26281452e-01
-3.04182500e-01 1.91325054e-01 8.09056878e-01 3.68020266e-01
1.67922997e+00 -6.73588514e-02 -6.86288066e-03 -1.81975797e-01
3.49451751e-01 5.81734292e-02 5.71269393e-01 7.21995056e-01
-1.39839977e-01 8.56275499e-01 7.46557593e-01 -7.04034865e-02
-1.10309386e+00 -8.34141731e-01 -8.83734226e-02 1.36506784e+00
-1.73607528e-01 -9.49249983e-01 -7.98765481e-01 -8.81782591e-01
-6.59724399e-02 1.18717289e+00 -7.99331427e-01 -1.81220278e-01
-3.44787329e-01 -7.02513456e-01 1.10060668e+00 6.19296789e-01
3.04742068e-01 -1.17017639e+00 -5.57795942e-01 7.08851293e-02
-4.01055008e-01 -8.04127336e-01 -7.59722590e-01 7.06635773e-01
-3.86019886e-01 -1.04485738e+00 -4.75214094e-01 -5.54100156e-01
3.81299555e-01 3.11715361e-02 1.00967550e+00 1.57606691e-01
-4.41934526e-01 -3.55670094e-01 -7.30781615e-01 -6.76580250e-01
-3.27132821e-01 2.88415104e-01 -1.61543533e-01 -4.44565535e-01
5.61460197e-01 -1.72236264e-01 -4.96728808e-01 3.47171456e-01
-8.10511112e-01 -4.44568098e-02 4.82639313e-01 6.43077910e-01
5.93253136e-01 -6.07365482e-02 5.59188545e-01 -1.04260969e+00
6.37416720e-01 -4.14758891e-01 -4.19667631e-01 2.80170113e-01
-6.74380779e-01 -6.54825196e-02 5.47133386e-01 -2.86795110e-01
-8.19733381e-01 1.72236010e-01 -4.28226948e-01 -1.13185808e-01
-7.13787228e-02 6.72141135e-01 -1.63671881e-01 4.92456973e-01
8.06138039e-01 2.35928483e-02 -6.39890254e-01 -7.76015222e-01
3.75251055e-01 9.54195142e-01 5.98292828e-01 -5.02564669e-01
6.09924376e-01 -1.06234811e-02 -3.69184881e-01 -5.36555052e-01
-1.12199152e+00 -7.13205099e-01 -7.15130329e-01 -7.84995183e-02
7.76842415e-01 -8.18635643e-01 -4.60139036e-01 6.23748362e-01
-1.12955809e+00 -4.65486884e-01 -2.73949206e-02 1.27486944e-01
-2.21832633e-01 3.51028562e-01 -8.17205787e-01 -8.62517238e-01
-7.11617053e-01 -7.26460636e-01 1.03108537e+00 6.69019669e-02
-4.92704749e-01 -4.73250628e-01 2.56837964e-01 5.73279679e-01
4.64936830e-02 1.71230018e-01 8.67919147e-01 -9.57866430e-01
5.00449091e-02 -3.64620030e-01 -5.87876365e-02 1.95370406e-01
-2.54746675e-01 2.77333796e-01 -1.04687297e+00 -2.16166154e-02
-5.81587493e-01 -5.02147734e-01 1.25294340e+00 2.41828039e-01
1.35003531e+00 -7.39024878e-02 -3.59378308e-01 2.26876870e-01
1.21439135e+00 1.78257465e-01 7.30690300e-01 6.87544227e-01
5.60524642e-01 5.15759051e-01 7.57783294e-01 6.72104001e-01
2.18981683e-01 4.34008390e-01 5.35282016e-01 2.51681536e-01
-7.59828016e-02 -4.04065311e-01 5.10228157e-01 4.86377269e-01
2.15094432e-01 -8.16914558e-01 -1.17729235e+00 9.79972303e-01
-1.73271537e+00 -9.13006425e-01 -6.90722644e-01 2.13066149e+00
9.90441740e-01 4.43115860e-01 2.68028319e-01 3.70406270e-01
9.38174665e-01 -4.73925062e-02 -5.15908003e-01 -7.24282682e-01
9.53659788e-02 1.15581468e-01 6.68880880e-01 2.89492697e-01
-1.31037819e+00 1.14043617e+00 5.97534084e+00 1.06967831e+00
-6.89909935e-01 1.42816722e-01 8.18606257e-01 -3.87578517e-01
-1.43306300e-01 -1.75885379e-01 -1.01497567e+00 6.12626433e-01
1.27759433e+00 -1.66052997e-01 5.77086136e-02 5.88261187e-01
3.21811646e-01 -4.11210716e-01 -1.02466631e+00 4.18080658e-01
2.34409109e-01 -9.71093893e-01 -2.54776269e-01 1.74657330e-01
5.96526742e-01 2.94415742e-01 4.46339510e-02 4.50867414e-01
5.33979535e-01 -1.12518072e+00 1.18806899e+00 1.59330934e-01
8.02753687e-01 -9.69802558e-01 7.36580789e-01 4.39305037e-01
-9.67459917e-01 -1.53897917e-02 -2.82802224e-01 -1.03612151e-02
1.83211733e-02 9.85124290e-01 -1.16816270e+00 3.19150478e-01
9.45558190e-01 3.81189734e-01 -8.88430476e-01 1.45342314e+00
-5.14770925e-01 1.16773391e+00 -3.65905792e-01 -5.43236323e-02
1.70412317e-01 2.29370341e-01 4.72745150e-01 1.51342952e+00
3.85748804e-01 -8.33920613e-02 -3.23622152e-02 6.38178647e-01
-4.63472486e-01 2.21582025e-01 -2.72943527e-01 -1.42894387e-01
6.41556859e-01 1.33775973e+00 -9.50977147e-01 -2.78339326e-01
-5.36978953e-02 9.83742118e-01 4.09313232e-01 -1.11664981e-01
-1.05650914e+00 -8.46204758e-01 3.41515034e-01 8.26996490e-02
4.69549567e-01 -5.64671382e-02 -8.00023198e-01 -5.95341861e-01
9.19833258e-02 -6.25125825e-01 5.86855292e-01 -7.63606608e-01
-1.28170395e+00 6.81751311e-01 -2.75169909e-01 -8.77639890e-01
8.81021917e-02 -3.89694840e-01 -8.62598658e-01 9.95139360e-01
-1.05219054e+00 -1.08028293e+00 -2.47273430e-01 5.97782061e-02
8.19099128e-01 2.35581532e-01 6.89787507e-01 2.11535886e-01
-9.30242896e-01 7.72194266e-01 2.58233845e-01 4.40842398e-02
1.10516572e+00 -1.41287494e+00 8.48146439e-01 1.19560242e+00
2.01719701e-02 3.49347323e-01 8.22406530e-01 -9.09889579e-01
-6.83529437e-01 -1.32251954e+00 1.45065653e+00 -7.09123909e-01
7.56278336e-01 -5.02584040e-01 -7.46657550e-01 6.51533365e-01
3.99080902e-01 -4.34734523e-01 6.80134714e-01 3.54685187e-01
-5.53497016e-01 2.85980701e-01 -9.09410894e-01 5.38010895e-01
9.77799773e-01 -2.65380859e-01 -4.54926997e-01 5.56081891e-01
5.73803246e-01 -2.70278066e-01 -5.95016420e-01 1.74523741e-01
3.32272619e-01 -7.45292127e-01 5.09283602e-01 -4.64655548e-01
8.55489373e-01 -2.43520766e-01 -1.15900405e-01 -1.34046137e+00
-3.65123630e-01 -4.26788777e-01 1.70101196e-01 1.50633287e+00
8.75602186e-01 -3.84329148e-02 5.78776419e-01 3.90172541e-01
-5.59948206e-01 -9.55627024e-01 -6.65276766e-01 -9.32105660e-01
3.26330274e-01 -5.91409504e-01 2.59832084e-01 4.10699755e-01
1.90808609e-01 4.25017715e-01 -2.88166076e-01 1.26038179e-01
2.96248108e-01 -1.99730769e-01 3.86391521e-01 -9.39588249e-01
-8.28225762e-02 -3.66193801e-01 1.07169487e-01 -7.25937486e-01
3.55230644e-02 -8.73118937e-01 5.78648150e-01 -1.58376408e+00
6.05413914e-01 -2.60193825e-01 -4.92906630e-01 9.15596545e-01
-4.23378497e-01 7.49153078e-01 1.97798267e-01 1.23005494e-01
-7.77037382e-01 3.69482011e-01 6.63336694e-01 -1.62975967e-01
-1.45355567e-01 -8.71769935e-02 -1.03278768e+00 5.58335125e-01
1.29783523e+00 -7.09550738e-01 -7.87524804e-02 -3.41822028e-01
2.62141526e-01 -5.53442717e-01 -4.34996784e-02 -8.70954990e-01
1.05717510e-01 -4.60846275e-02 4.39465612e-01 -1.00769556e+00
5.87774701e-02 -2.93433011e-01 -1.96256191e-01 5.40806115e-01
-5.67352176e-01 3.68674606e-01 4.83336806e-01 4.23460692e-01
2.91213691e-01 -6.14103794e-01 6.08788073e-01 1.73367798e-01
-5.95853150e-01 -1.00479454e-01 -7.55279124e-01 2.63942778e-01
1.15795076e+00 9.06786397e-02 -8.03101897e-01 -5.58020361e-02
-5.68073571e-01 4.87354398e-01 4.25503105e-01 3.65981907e-01
3.22387874e-01 -7.43322313e-01 -8.91745746e-01 -3.40260655e-01
4.46078062e-01 -1.63280576e-01 8.02972168e-02 8.90618026e-01
-7.74018764e-01 1.90718621e-01 5.74626476e-02 -2.25587457e-01
-1.61306083e+00 4.07503009e-01 1.28231989e-02 -4.99306172e-01
-5.87682784e-01 1.23559439e+00 -1.53997511e-01 -2.44742796e-01
3.45292807e-01 -1.00157380e-01 -1.22898631e-01 1.14315666e-01
6.07250035e-01 6.50194585e-01 4.96407986e-01 -5.66357493e-01
-5.78597128e-01 1.67527139e-01 -5.66248775e-01 -3.32854211e-01
1.29502678e+00 1.01952441e-01 8.27791542e-02 5.28406858e-01
9.52087402e-01 2.35050097e-01 -1.07997537e+00 9.94691718e-03
4.38843250e-01 -2.16965988e-01 5.39274290e-02 -1.46458673e+00
-6.15352452e-01 6.77510023e-01 3.10787588e-01 2.37292834e-02
1.01616859e+00 4.20512594e-02 8.61779034e-01 2.64802963e-01
-1.14867456e-01 -1.18196750e+00 8.56722668e-02 8.83169711e-01
9.03680801e-01 -7.84310997e-01 8.07050802e-03 -5.88490665e-01
-6.94135368e-01 8.16137791e-01 7.09385216e-01 3.50912623e-02
2.22209543e-01 4.19988781e-01 1.28981575e-01 -7.98511356e-02
-1.20124400e+00 -2.73931295e-01 2.02177614e-01 2.87448406e-01
6.99535966e-01 2.14183647e-02 -6.87941790e-01 6.63892329e-01
-4.70956206e-01 -2.38791853e-01 6.76171303e-01 9.97434437e-01
-6.31055415e-01 -9.86018598e-01 -4.08809394e-01 5.68778932e-01
-7.93339729e-01 -4.58359420e-01 -9.24382865e-01 5.64185500e-01
1.74010545e-01 1.24270976e+00 7.67067298e-02 -7.17181981e-01
4.84784633e-01 3.93997490e-01 2.57419497e-01 -8.59407604e-01
-9.07631993e-01 3.61523718e-01 4.76112574e-01 -2.48875245e-01
3.30269158e-01 -8.52864385e-01 -1.38132179e+00 -4.74165767e-01
-4.77452457e-01 1.93232507e-01 6.91707909e-01 6.59694850e-01
3.65330637e-01 7.29803085e-01 4.16996837e-01 -6.13429904e-01
-7.07073629e-01 -1.45106518e+00 -5.68022847e-01 2.63681620e-01
-2.45364290e-03 -3.91784221e-01 -3.40860069e-01 3.35482389e-01]
|
[8.978611946105957, 10.637084007263184]
|
9de4d96e-b7e9-4f95-b987-b7e18243298b
|
teaching-a-new-dog-old-tricks-resurrecting
|
1912.13080
| null |
https://arxiv.org/abs/1912.13080v1
|
https://arxiv.org/pdf/1912.13080v1.pdf
|
Teaching a New Dog Old Tricks: Resurrecting Multilingual Retrieval Using Zero-shot Learning
|
While billions of non-English speaking users rely on search engines every day, the problem of ad-hoc information retrieval is rarely studied for non-English languages. This is primarily due to a lack of data set that are suitable to train ranking algorithms. In this paper, we tackle the lack of data by leveraging pre-trained multilingual language models to transfer a retrieval system trained on English collections to non-English queries and documents. Our model is evaluated in a zero-shot setting, meaning that we use them to predict relevance scores for query-document pairs in languages never seen during training. Our results show that the proposed approach can significantly outperform unsupervised retrieval techniques for Arabic, Chinese Mandarin, and Spanish. We also show that augmenting the English training collection with some examples from the target language can sometimes improve performance.
|
['Nazli Goharian', 'Sean MacAvaney', 'Luca Soldaini']
|
2019-12-30
| null | null | null | null |
['ad-hoc-information-retrieval']
|
['natural-language-processing']
|
[-1.54880390e-01 -3.98047537e-01 -2.63286650e-01 -1.95146367e-01
-1.59045541e+00 -7.72073627e-01 1.00617135e+00 4.01523113e-01
-8.73831213e-01 6.96322858e-01 1.83352649e-01 -4.33232278e-01
-1.83632255e-01 -5.45282602e-01 -4.18929189e-01 -2.27597058e-01
-2.79227793e-02 8.80657256e-01 3.94785494e-01 -8.28654051e-01
1.74879566e-01 3.27703685e-01 -1.46013784e+00 4.49573576e-01
1.04551959e+00 4.86396164e-01 3.66699010e-01 6.91454291e-01
-3.84647518e-01 5.52826643e-01 -5.31742990e-01 -3.03738028e-01
1.90726787e-01 -3.37608486e-01 -9.25934494e-01 -3.28895509e-01
3.89970511e-01 -4.01588261e-01 -1.34697959e-01 7.92694807e-01
6.25250340e-01 2.79751688e-01 8.34522009e-01 -6.83079898e-01
-8.08738291e-01 5.86553276e-01 -1.98032886e-01 1.08125106e-01
5.39930344e-01 -5.56516349e-01 1.25784361e+00 -1.24220634e+00
1.00172937e+00 1.07989466e+00 1.64831147e-01 4.07031775e-01
-7.53045559e-01 -3.09529155e-01 -1.33291304e-01 3.94392192e-01
-1.63311565e+00 -5.82008123e-01 4.76143211e-01 -5.90035953e-02
1.11794567e+00 3.29258859e-01 2.26184502e-01 9.12589550e-01
-1.97021946e-01 1.05352306e+00 8.78164947e-01 -1.06842804e+00
-1.35342330e-01 5.38606405e-01 2.56455123e-01 3.78609836e-01
-3.77964601e-02 -2.40118936e-01 -3.55474561e-01 -2.38496810e-01
2.88136844e-02 -1.30401224e-01 -1.48922220e-01 -2.49980718e-01
-1.06940103e+00 8.72580349e-01 2.95217782e-01 6.71845973e-01
-1.93763167e-01 -3.22190255e-01 4.19996798e-01 7.99849272e-01
5.91759324e-01 7.40063906e-01 -5.40996850e-01 1.01559564e-01
-1.11498988e+00 1.37489542e-01 1.00000298e+00 1.13133264e+00
9.76281703e-01 -5.79649389e-01 -1.17120102e-01 1.31856239e+00
2.83024490e-01 8.37981820e-01 7.53127158e-01 -3.68717641e-01
4.54817504e-01 4.39553529e-01 3.24845910e-01 -6.26818657e-01
-9.07192826e-02 -1.36962086e-01 -2.20982850e-01 -5.44478118e-01
3.57564688e-01 1.25487760e-01 -7.37336218e-01 1.25885773e+00
-1.16147781e-02 -5.14503002e-01 5.92552483e-01 8.21393430e-01
7.11222708e-01 9.37075853e-01 -1.06333889e-01 -3.58523190e-01
1.18856430e+00 -1.24554825e+00 -6.11185491e-01 -3.34315568e-01
9.08456504e-01 -1.26697302e+00 1.41123283e+00 1.31380245e-01
-7.85071254e-01 -2.48521611e-01 -7.02488601e-01 -4.54413265e-01
-8.96043599e-01 2.52948314e-01 3.49760950e-01 3.12837332e-01
-1.01764679e+00 1.15191840e-01 -5.28358698e-01 -9.52897310e-01
-3.07509243e-01 1.69064760e-01 -2.37227112e-01 -6.70987129e-01
-1.47078192e+00 1.00780261e+00 3.84235263e-01 -2.53493041e-01
-7.09250748e-01 -9.22003016e-02 -7.05678880e-01 -7.37264827e-02
3.22744757e-01 -1.88024685e-01 1.38720679e+00 -8.67238402e-01
-1.09830201e+00 9.99937713e-01 -2.56674021e-01 -1.53788656e-01
2.75088459e-01 -4.12252009e-01 -6.54576004e-01 2.71680355e-01
2.91331828e-01 5.28344870e-01 5.77554345e-01 -1.12019598e+00
-5.99843383e-01 -2.89218754e-01 1.22889332e-01 3.99702251e-01
-7.21125722e-01 5.81029177e-01 -1.21803820e+00 -5.16132176e-01
-1.70451358e-01 -9.27932203e-01 -1.53168589e-02 -4.67583746e-01
-1.03387468e-01 -5.12309253e-01 6.67615950e-01 -7.75117338e-01
1.44125307e+00 -1.90714943e+00 -1.98382229e-01 3.80424619e-01
-5.90681076e-01 4.07464206e-01 -6.11176133e-01 1.04863679e+00
5.10658979e-01 2.16072604e-01 9.82345045e-02 -7.77400807e-02
-5.62195703e-02 2.56733000e-01 -4.05608684e-01 -9.14669931e-02
1.03047214e-01 8.92178655e-01 -1.31804717e+00 -7.77084351e-01
-1.93521947e-01 5.32045901e-01 -3.49351525e-01 2.24822059e-01
-1.71918973e-01 6.51844293e-02 -7.36714900e-01 8.96165192e-01
1.95238397e-01 -5.84974810e-02 2.69426316e-01 3.06523442e-01
-4.36530896e-02 5.07544279e-01 -7.03050911e-01 1.87050879e+00
-8.64795446e-01 6.23982787e-01 -2.50782698e-01 -5.53505063e-01
7.30104804e-01 4.30906653e-01 3.00007612e-01 -1.19374144e+00
-2.19370589e-01 8.67759287e-01 -9.84722748e-02 -5.37070334e-01
9.27440464e-01 2.02731356e-01 -3.52636427e-01 7.03562021e-01
4.59914915e-02 -1.67216226e-01 7.64656365e-01 3.17588210e-01
1.01764023e+00 -2.10544970e-02 1.78573459e-01 -2.76673526e-01
6.22159839e-01 3.37601632e-01 -1.20389655e-01 1.13644195e+00
1.96586490e-01 3.54927093e-01 -2.35988453e-01 2.77512148e-03
-1.04851210e+00 -8.89570653e-01 -3.24577391e-01 1.83286452e+00
-1.07846791e-02 -6.41234875e-01 -5.19351780e-01 -7.63852358e-01
-2.92949438e-01 6.82047129e-01 -1.31271482e-01 -9.50310007e-02
-5.66160679e-01 -4.13088232e-01 3.94717336e-01 -2.19925158e-02
1.15466066e-01 -1.15937686e+00 -3.55861217e-01 2.92086035e-01
-2.86417365e-01 -1.00383568e+00 -6.57296717e-01 1.86797827e-01
-6.24036252e-01 -8.87482941e-01 -1.22149444e+00 -1.08489656e+00
5.99674344e-01 5.98090947e-01 1.46313679e+00 3.35942000e-01
-9.97716039e-02 7.79616237e-01 -8.72829437e-01 -4.94396657e-01
-5.93042612e-01 6.82749927e-01 3.13651301e-02 -4.16707844e-01
8.85590851e-01 4.02089506e-02 -5.08922517e-01 2.52359450e-01
-1.23872554e+00 -4.69051898e-01 5.78260303e-01 8.97006452e-01
3.20510238e-01 -2.63461918e-01 5.79007268e-01 -8.11940372e-01
1.03247559e+00 -5.87355196e-01 -5.23376763e-01 1.01807225e+00
-7.58508801e-01 3.02803725e-01 4.77088928e-01 -5.03819942e-01
-8.06777120e-01 -2.50842422e-01 1.33839920e-01 -3.07338908e-02
1.83308706e-01 1.04134309e+00 3.21342140e-01 5.94755672e-02
7.71297991e-01 1.94828168e-01 -4.06688035e-01 -6.80847704e-01
3.58116686e-01 1.15714872e+00 1.34055614e-01 -7.05947578e-01
6.19890153e-01 4.71679308e-03 -5.70483208e-01 -1.07046127e+00
-8.21403027e-01 -1.11536396e+00 -6.70325994e-01 -9.23820660e-02
3.96030903e-01 -9.47915077e-01 2.03243464e-01 1.23020530e-01
-1.13253617e+00 -1.43980190e-01 -7.35198334e-02 5.44406891e-01
-1.11097790e-01 2.46570110e-01 -6.63855910e-01 -8.93561423e-01
-5.79034805e-01 -1.03892410e+00 1.32096732e+00 2.40347143e-02
5.29485010e-02 -1.08718657e+00 4.86793458e-01 5.21046408e-02
6.16383970e-01 -7.86592245e-01 1.00321651e+00 -1.12199175e+00
-4.68823910e-01 -6.73984706e-01 -9.43884701e-02 2.81811059e-01
1.20480768e-01 -1.99944690e-01 -8.71203423e-01 -4.56866890e-01
-3.25943530e-01 -9.60383356e-01 8.26484084e-01 -1.91684753e-01
6.81686103e-01 -2.06027299e-01 -2.73087084e-01 -9.08182859e-02
1.30117655e+00 8.15413147e-02 3.95240456e-01 6.38818860e-01
2.38810495e-01 8.70970309e-01 9.04582024e-01 8.29688981e-02
4.37498897e-01 8.65557671e-01 -2.50010401e-01 2.24503092e-02
7.85122365e-02 -2.41483539e-01 5.16264677e-01 1.28403938e+00
1.34423330e-01 -3.89090538e-01 -1.10391366e+00 8.55730176e-01
-1.70862186e+00 -6.53697491e-01 3.69111121e-01 2.43945360e+00
1.27118778e+00 -3.43816608e-01 -5.99241704e-02 -4.00201410e-01
3.29578429e-01 -1.65897697e-01 -1.43272266e-01 -2.70971090e-01
-2.38260463e-01 4.80795115e-01 2.24056631e-01 6.98361158e-01
-1.01283240e+00 1.34054291e+00 6.32632971e+00 9.33018684e-01
-1.09147549e+00 7.19860122e-02 3.06277990e-01 1.79093197e-01
-4.25068378e-01 1.29815266e-01 -8.60548735e-01 -3.56802717e-02
1.01050615e+00 -4.81196970e-01 5.60708880e-01 7.77007878e-01
-1.92979455e-01 1.27600192e-03 -1.10848641e+00 7.44732797e-01
4.05612707e-01 -5.24975419e-01 2.55878508e-01 -9.38193724e-02
8.13842058e-01 3.84965867e-01 -8.64759237e-02 7.67825723e-01
3.56898695e-01 -7.80654550e-01 3.65872711e-01 4.60329145e-01
5.87541163e-01 -5.99649549e-01 6.55583978e-01 5.96828520e-01
-7.78712749e-01 2.60781556e-01 -6.55284047e-01 4.88701642e-01
-1.79109126e-02 9.72917303e-02 -1.05155349e+00 4.77997452e-01
6.29987061e-01 4.58294719e-01 -9.90592539e-01 1.13130546e+00
-2.14291960e-01 3.07873011e-01 -5.02995074e-01 -4.03066069e-01
5.28724790e-01 -1.97155863e-01 4.75222468e-01 1.39796877e+00
7.62908340e-01 -3.05689842e-01 2.96811879e-01 3.45650733e-01
-1.27475858e-01 8.55397463e-01 -9.56837893e-01 -2.83694059e-01
3.41176301e-01 1.24956942e+00 -4.69672918e-01 -4.36532289e-01
-6.81598425e-01 1.18350458e+00 3.94140154e-01 6.92262590e-01
-1.17091119e-01 -6.68980598e-01 -1.79868005e-02 -1.81843996e-01
6.85122162e-02 -2.25345060e-01 6.90509140e-01 -1.46863508e+00
1.71681136e-01 -1.17566872e+00 6.34804368e-01 -7.64367878e-01
-1.47898579e+00 8.85908842e-01 1.82307646e-01 -1.28633368e+00
-9.79705811e-01 -5.75546741e-01 -1.30198851e-01 9.13773239e-01
-1.94785118e+00 -1.18786991e+00 1.48371935e-01 7.16716528e-01
6.15377367e-01 -4.95132536e-01 1.20241785e+00 5.53374112e-01
-1.36739165e-01 5.49394071e-01 6.97755396e-01 2.41156936e-01
1.25256455e+00 -1.11134839e+00 7.06670284e-02 8.25578868e-01
8.27828050e-01 1.06969333e+00 4.90128875e-01 -3.58730108e-01
-1.46396744e+00 -6.84485435e-01 1.53357828e+00 -5.26382089e-01
9.36736166e-01 -3.12316328e-01 -8.92551601e-01 3.84026319e-01
5.02345026e-01 -2.04155654e-01 7.50809789e-01 3.78554076e-01
-5.25588036e-01 -1.04145259e-01 -5.97006679e-01 7.64283359e-01
5.13042331e-01 -1.07670546e+00 -7.31836081e-01 6.89237416e-01
5.15977621e-01 3.45073417e-02 -6.06954873e-01 2.10646808e-01
4.48393047e-01 -2.71130115e-01 1.03859520e+00 -9.19716537e-01
2.53649443e-01 -1.02572367e-01 -4.02468503e-01 -1.30707157e+00
3.33568037e-01 -3.05921584e-01 -1.72869265e-02 1.07767653e+00
7.95883477e-01 -3.30967665e-01 1.76647589e-01 5.00684977e-01
1.16417557e-01 -2.44865730e-01 -8.28140616e-01 -8.82843912e-01
3.67423266e-01 -3.72872025e-01 2.46816278e-01 8.98536980e-01
1.59109399e-01 7.50592470e-01 -2.89430022e-01 -2.38135189e-01
1.98403820e-01 2.10208043e-01 8.29489887e-01 -1.03760612e+00
-2.25371972e-01 -3.22344124e-01 1.51520371e-01 -1.11592257e+00
3.50172400e-01 -1.18839777e+00 3.51557523e-01 -1.43757999e+00
4.18673933e-01 -5.85234404e-01 -5.25780797e-01 4.67660695e-01
-3.69629413e-01 4.16865647e-01 5.14437407e-02 5.46773314e-01
-1.08318496e+00 3.25000614e-01 8.94883871e-01 -3.35320562e-01
-2.26936862e-01 -1.09069094e-01 -4.60547894e-01 3.50561827e-01
4.63183820e-01 -5.64832568e-01 -4.74828660e-01 -8.68649423e-01
5.73778510e-01 -1.11371033e-01 -1.29509047e-01 -6.55183136e-01
3.11312288e-01 6.01015612e-02 -6.13163374e-02 -4.22662973e-01
1.22318670e-01 -9.40111697e-01 -5.76892674e-01 1.02017954e-01
-5.29314399e-01 3.71407002e-01 7.13277161e-02 2.42419511e-01
-6.58407331e-01 -7.81279802e-01 1.32346541e-01 -1.94126353e-01
-8.49216759e-01 1.10571466e-01 -4.41226512e-01 2.31160581e-01
4.30850357e-01 4.14574087e-01 -2.49170080e-01 -6.67351067e-01
-3.37953210e-01 3.93816918e-01 5.25485635e-01 7.82174945e-01
4.06283528e-01 -1.37463474e+00 -7.97125340e-01 7.57200294e-04
9.05447841e-01 -4.54800189e-01 -2.49135137e-01 5.98213732e-01
-5.47246873e-01 1.02331245e+00 2.11814508e-01 -4.73660678e-01
-1.17616642e+00 4.84028339e-01 -3.60943675e-02 -7.32535899e-01
-1.78011298e-01 4.75240201e-01 -2.83861637e-01 -7.71456420e-01
3.96598220e-01 1.93813354e-01 -4.02256757e-01 2.52400249e-01
7.01250434e-01 -1.30748704e-01 3.57664138e-01 -6.92724466e-01
-2.06779554e-01 4.55889583e-01 -5.84493399e-01 -6.54783368e-01
1.14916766e+00 -2.23806053e-01 -2.65015870e-01 5.79422534e-01
1.51561928e+00 3.22790742e-01 -1.85595721e-01 -5.55560827e-01
6.44348443e-01 -2.45033309e-01 6.91784248e-02 -8.13478589e-01
-4.69062030e-01 8.81971717e-01 6.45062029e-01 3.41152065e-02
1.18055892e+00 2.54303724e-01 7.36813247e-01 1.42477393e+00
7.07676709e-01 -1.42634189e+00 1.58926640e-02 1.03272879e+00
7.46949673e-01 -1.56846404e+00 -2.41917372e-01 1.67760283e-01
-4.28317308e-01 9.95517135e-01 1.42343059e-01 2.24962309e-01
5.94761729e-01 -3.28917831e-01 5.07679760e-01 -1.62188113e-01
-7.63157547e-01 -7.26970792e-01 7.98309147e-01 2.92613387e-01
9.34794366e-01 -3.51107895e-01 -6.07192993e-01 2.34854281e-01
-8.56534913e-02 -4.32996340e-02 1.35742068e-01 1.33859015e+00
-3.55965257e-01 -1.73055100e+00 -1.57440022e-01 2.66915411e-01
-7.16828823e-01 -7.47726500e-01 -6.37081087e-01 9.15316999e-01
-5.55014789e-01 9.86374557e-01 -7.05624670e-02 4.71863821e-02
1.93162113e-01 5.21466494e-01 3.19762588e-01 -9.39963341e-01
-6.12871170e-01 5.04800558e-01 1.58375964e-01 -1.68641329e-01
-5.50124407e-01 -5.15828371e-01 -7.02869892e-01 2.69332796e-01
-4.95211929e-01 6.98893249e-01 7.10293412e-01 8.36771548e-01
1.50330931e-01 -1.19088650e-01 5.67432165e-01 -3.55393261e-01
-6.69531524e-01 -1.24390996e+00 -4.17132258e-01 5.75388908e-01
1.61940128e-01 -2.05584958e-01 -2.54212976e-01 -1.39045745e-01]
|
[11.380242347717285, 9.821402549743652]
|
ba4631e1-51b8-400c-8de5-3864f15e8745
|
3d-densely-convolutional-networks-for-1
| null | null |
https://arxiv.org/abs/1709.03199
|
https://arxiv.org/pdf/1709.03199.pdf
|
3D Densely Convolutional Networks for VolumetricSegmentation
|
In the isointense stage, the accurate volumetric image segmentation is a challenging task due to the low contrast between tissues. In this paper, we propose a novel very deep network architecture based on densely convolutional network for volumetric brain segmentation. The proposed network architecture provides a dense connection between layers that aims to improve the information flow in the network. By concatenating features map of fine and coarse dense blocks, it allows capturing multi-scale contextual information. Experimental results demonstrate significant advantages of the proposed method over existing methods, in terms of both segmentation accuracy and parameter efficiency in MICCAI grand challenge on 6-month infant brain MRI segmentation
|
['Jitae Shin', 'Toan Duc Bui', 'Taesup Moon']
|
2017-09-13
| null | null | null |
arxiv-preprint-2017-9
|
['3d-medical-imaging-segmentation', 'infant-brain-mri-segmentation', 'volumetric-medical-image-segmentation']
|
['medical', 'medical', 'medical']
|
[ 2.85091419e-02 9.95623842e-02 8.07347819e-02 -6.17621303e-01
-1.81519210e-01 -9.49170440e-03 -2.52680518e-02 2.29077876e-01
-5.29515207e-01 6.39239669e-01 2.55859494e-01 -4.47224230e-02
-1.35285422e-01 -7.87942708e-01 -6.65537179e-01 -5.75024247e-01
-3.78806889e-01 4.76192325e-01 5.29999614e-01 7.78810540e-03
8.76589119e-02 6.17636442e-01 -9.72489774e-01 3.12702805e-01
9.95261967e-01 1.15759206e+00 3.42177957e-01 3.49343568e-01
-4.47476834e-01 8.22408080e-01 -3.07882667e-01 -1.01545863e-01
2.31745496e-01 -2.03655183e-01 -1.14662540e+00 -2.26915479e-01
3.27086031e-01 -6.15848064e-01 -2.02480301e-01 1.15161180e+00
6.21807337e-01 -5.99598326e-03 5.39533019e-01 -6.37929857e-01
-2.28746921e-01 8.88327897e-01 -8.18288028e-01 9.36044872e-01
-3.15747112e-01 -6.71691447e-02 5.00206172e-01 -6.20508850e-01
2.59025753e-01 1.03986669e+00 7.65169978e-01 4.97794300e-01
-7.83154130e-01 -7.95603931e-01 2.79781930e-02 2.48768851e-01
-1.25280643e+00 -1.19880080e-01 6.70365632e-01 -6.86649621e-01
8.15686464e-01 -4.12649801e-03 1.00258577e+00 4.16165262e-01
3.87366891e-01 6.48932219e-01 1.04527628e+00 9.52318236e-02
1.35159552e-01 -3.64442974e-01 4.19269979e-01 8.21390808e-01
3.80685717e-01 -2.39775702e-01 -9.67426002e-02 1.71510264e-01
1.14082730e+00 1.20863333e-01 -3.18783909e-01 -4.11266118e-01
-1.01628506e+00 8.10402215e-01 9.71611738e-01 7.38743424e-01
-6.34928524e-01 1.61188975e-01 4.39096332e-01 -2.52032638e-01
5.30787408e-01 1.04298733e-01 -2.34945983e-01 1.53779224e-01
-1.27490091e+00 -1.29701927e-01 3.51426899e-01 5.72221041e-01
4.36641484e-01 2.14352787e-01 -3.48245561e-01 8.85872662e-01
4.84043837e-01 1.14967180e-02 7.54332423e-01 -5.72952211e-01
4.40542102e-01 5.10571599e-01 -4.31240290e-01 -8.80051911e-01
-9.98346984e-01 -6.64864242e-01 -1.25531316e+00 4.39769998e-02
2.94130743e-01 -2.16891393e-01 -1.26063812e+00 1.46874058e+00
4.78455484e-01 2.27903143e-01 -2.29839683e-01 1.09332490e+00
1.43229067e+00 3.88934880e-01 2.06556529e-01 -2.21091226e-01
1.28194880e+00 -1.03170884e+00 -7.12508082e-01 -1.59244150e-01
2.84686774e-01 -4.34631407e-01 6.35303020e-01 1.16207547e-01
-1.36418521e+00 -5.46644032e-01 -1.12155604e+00 2.84418222e-02
-5.11095822e-02 -2.56606638e-01 9.95273292e-01 6.17567182e-01
-1.45824981e+00 5.24901271e-01 -1.01426649e+00 1.33400992e-01
1.21347165e+00 8.44403982e-01 -1.28321394e-01 1.01908837e-02
-1.09971619e+00 6.65655553e-01 7.23444045e-01 2.59314805e-01
-7.92388678e-01 -1.01446486e+00 -7.65066743e-01 3.14068228e-01
-1.23691075e-01 -3.56295556e-01 9.22453701e-01 -6.70438468e-01
-1.16384029e+00 8.27988982e-01 1.34295508e-01 -6.65355146e-01
4.41949248e-01 1.21419288e-01 7.51793459e-02 7.32673347e-01
-6.75086379e-02 9.89188492e-01 6.54200017e-01 -7.67031014e-01
-4.35529143e-01 -5.65701485e-01 -7.36811757e-02 1.23575814e-01
-2.23703966e-01 2.87639555e-02 -2.68642098e-01 -6.60084903e-01
3.93010587e-01 -3.89563620e-01 -4.41603839e-01 -2.21937820e-01
-1.82017386e-01 -1.27644660e-02 8.42492163e-01 -9.16466773e-01
7.96578825e-01 -1.74879014e+00 -1.25365108e-01 1.79173842e-01
7.72029400e-01 3.36513788e-01 2.93683261e-02 -5.60297787e-01
-2.41303399e-01 5.09415157e-02 -3.63796085e-01 3.80828679e-02
-5.50846517e-01 -1.37505323e-01 3.09090018e-01 5.75525701e-01
7.86474645e-02 1.19849062e+00 -6.11532867e-01 -7.45015442e-01
3.43850672e-01 7.64342129e-01 -5.80218136e-01 2.64499277e-01
1.87673286e-01 9.11582768e-01 -4.77286547e-01 4.49598700e-01
1.10331023e+00 -1.81031734e-01 -2.11799785e-01 -3.55659008e-01
-8.48435462e-02 -2.97327429e-01 -6.48868620e-01 1.92866766e+00
-3.27537119e-01 4.43158746e-01 2.96427637e-01 -1.43620396e+00
7.53088236e-01 3.40719819e-01 9.51062322e-01 -9.38051283e-01
6.66700125e-01 1.04674399e-01 2.67620116e-01 -4.62924987e-01
-8.42395574e-02 -1.99784175e-01 4.09858733e-01 3.63488972e-01
3.01473260e-01 -2.38028392e-01 1.72046632e-01 4.49158140e-02
7.66036332e-01 -3.02537739e-01 1.53932571e-01 -7.06623495e-01
6.38495862e-01 -3.36360961e-01 6.26148462e-01 4.58258033e-01
-5.24697483e-01 6.83218539e-01 3.37425053e-01 -6.47354245e-01
-1.02902138e+00 -9.39161599e-01 -4.50024575e-01 6.01680517e-01
1.42266154e-01 2.78447360e-01 -1.36185682e+00 -5.36202729e-01
-4.54782993e-01 1.69467553e-03 -7.72659302e-01 4.21424992e-02
-9.66399312e-01 -1.08509743e+00 3.84825140e-01 6.57778800e-01
9.92763758e-01 -1.17488980e+00 -8.83587718e-01 3.68899435e-01
-2.21964404e-01 -1.18680358e+00 -5.10189891e-01 1.45442903e-01
-1.32947743e+00 -9.39625323e-01 -1.16774845e+00 -1.20226550e+00
7.51203716e-01 1.16123166e-02 1.07580519e+00 4.47715014e-01
-6.77386165e-01 -2.01956168e-01 -1.69237211e-01 -3.08678150e-01
-7.26381466e-02 2.40439966e-01 -4.91800040e-01 -1.57176137e-01
1.33529425e-01 -7.64275372e-01 -1.14752245e+00 1.91724915e-02
-8.18469703e-01 2.52643168e-01 6.31103873e-01 8.39109063e-01
6.13634884e-01 -9.72900260e-03 6.08971775e-01 -7.67945528e-01
5.02872050e-01 -5.32415986e-01 -6.11410618e-01 8.74546692e-02
-3.99301320e-01 -4.79305442e-03 4.36175674e-01 -1.47989511e-01
-9.96360362e-01 5.56854792e-02 -6.84749484e-01 -6.69345930e-02
-1.10921353e-01 2.32011780e-01 1.62265033e-01 -5.60534835e-01
5.94854876e-02 1.39040217e-01 1.95723101e-02 -4.87365246e-01
5.30403666e-02 3.99818659e-01 6.00782156e-01 -4.55838025e-01
1.78927317e-01 4.85290974e-01 3.46875601e-02 -3.84221435e-01
-7.25051403e-01 -3.29867154e-01 -1.09222114e+00 -2.50710279e-01
1.38500476e+00 -7.45845675e-01 -7.37236321e-01 6.23044014e-01
-1.16084421e+00 -3.44617814e-01 -2.16704533e-01 7.06662536e-01
-3.82911116e-01 3.06427062e-01 -9.94375527e-01 -5.99469207e-02
-9.37511086e-01 -1.74360383e+00 6.12484753e-01 5.97549558e-01
1.72801942e-01 -1.17777050e+00 -1.55146554e-01 5.21762371e-01
8.89821470e-01 4.87360686e-01 1.09780395e+00 -5.52796483e-01
-4.44995642e-01 1.56107530e-01 -7.77281404e-01 3.91208738e-01
2.14310866e-02 -3.73508602e-01 -8.36135387e-01 -3.98448795e-01
3.62083942e-01 -2.59216845e-01 1.04293847e+00 1.08633161e+00
1.69564986e+00 -1.75214112e-02 -8.58184397e-02 1.14145088e+00
1.48728681e+00 4.02341574e-01 5.61697125e-01 9.92321670e-02
1.05352330e+00 6.53414965e-01 1.61032483e-01 3.35352570e-01
4.09328580e-01 9.29891318e-02 6.70665562e-01 -7.00905025e-01
-3.88494968e-01 4.06822801e-01 -5.64719379e-01 1.18688893e+00
-1.27058789e-01 3.63954306e-01 -1.11886334e+00 7.82344997e-01
-1.45155561e+00 -5.77980399e-01 -6.56916425e-02 1.67556286e+00
8.80615175e-01 9.46554914e-02 -7.94192497e-03 -8.93090293e-02
7.88362920e-01 1.69366285e-01 -5.92973113e-01 -5.58183610e-01
1.81222513e-01 7.57346690e-01 5.52875578e-01 4.20695513e-01
-1.16374958e+00 8.15574944e-01 7.07407951e+00 9.02774990e-01
-1.22326612e+00 6.39686704e-01 1.09201145e+00 -8.80310386e-02
-7.03388974e-02 -6.81571186e-01 -5.51556349e-01 5.52891552e-01
7.11203814e-01 2.51466990e-01 2.67734259e-01 5.59412181e-01
6.08107671e-02 -1.84224665e-01 -7.95671225e-01 9.27784026e-01
-7.92559087e-02 -1.50801694e+00 -8.96027908e-02 -1.64777696e-01
8.04630101e-01 2.14215398e-01 7.63295665e-02 1.28632650e-01
-4.07841913e-02 -1.49602532e+00 6.22410238e-01 2.58434743e-01
7.50343323e-01 -1.10580635e+00 7.74599195e-01 2.18034461e-01
-1.25508177e+00 5.41890264e-02 -4.41674173e-01 1.36842176e-01
-6.07696027e-02 6.09244823e-01 -8.52403104e-01 9.29902494e-02
8.91094387e-01 3.59944463e-01 -4.49044645e-01 1.35249400e+00
1.99028417e-01 6.04208887e-01 -7.04435781e-02 1.08117446e-01
4.33295041e-01 -3.01179349e-01 9.91648063e-02 1.52372777e+00
-1.48334410e-02 3.65579247e-01 6.22306541e-02 9.85418081e-01
-2.43429646e-01 3.77344489e-01 -2.38504276e-01 3.98088753e-01
3.44163971e-03 1.37846494e+00 -1.30192065e+00 -3.99934620e-01
-4.09779459e-01 5.40625036e-01 5.03020644e-01 1.13002084e-01
-8.05062354e-01 -3.36775124e-01 2.31496647e-01 3.03580910e-02
3.62736255e-01 -7.96615034e-02 -6.81134284e-01 -8.72641504e-01
-2.84778416e-01 -4.44243550e-01 2.12069198e-01 -3.11132222e-01
-6.15214348e-01 8.59023690e-01 -6.76738843e-02 -6.49423659e-01
1.05789252e-01 -2.59157777e-01 -7.43202507e-01 9.32870090e-01
-1.69769812e+00 -8.92104506e-01 -5.91328323e-01 7.06121206e-01
5.27211666e-01 -6.21300638e-02 3.60603869e-01 5.68227649e-01
-5.15278041e-01 5.68960130e-01 -1.04916789e-01 3.81181508e-01
5.11243455e-02 -1.07250881e+00 3.32822531e-01 8.69062066e-01
-2.98833400e-01 5.29792070e-01 3.21749002e-01 -5.45911729e-01
-7.24667013e-01 -1.18732846e+00 5.90922236e-02 3.68594259e-01
3.34926695e-01 -2.09040046e-01 -1.05990791e+00 4.18582380e-01
3.10996056e-01 4.08552468e-01 4.76232499e-01 -4.27871108e-01
6.71249256e-02 -9.35636610e-02 -1.73290706e+00 7.66475424e-02
7.18878806e-01 3.20331031e-03 -5.45436442e-01 3.30453753e-01
9.73987460e-01 -6.69433355e-01 -1.16282582e+00 7.95888722e-01
3.62903684e-01 -9.49823916e-01 1.07203770e+00 -1.33262902e-01
4.32440370e-01 9.51015726e-02 3.11826557e-01 -1.18713212e+00
-3.52644861e-01 -6.82061389e-02 -1.00527652e-01 8.75486135e-01
1.11404717e-01 -5.32038152e-01 9.39184368e-01 4.58099753e-01
-2.37560511e-01 -1.18516862e+00 -9.99738872e-01 -2.81545758e-01
3.89320135e-01 -1.29957646e-01 9.65352595e-01 8.23805213e-01
-4.50960845e-01 -2.07889620e-02 3.57821421e-03 -4.81226481e-02
8.31610382e-01 -7.81320855e-02 -1.89990252e-01 -1.15434527e+00
9.71943140e-02 -5.71296751e-01 -5.29317677e-01 -8.58267844e-01
3.14426757e-02 -9.15313244e-01 1.81186777e-02 -1.87104905e+00
6.77839279e-01 -5.13725460e-01 -6.66848242e-01 3.02049011e-01
-1.25252143e-01 6.55638456e-01 -1.84412941e-01 -8.12600255e-02
-4.69756484e-01 2.13217244e-01 1.76564336e+00 -3.13230842e-01
-8.35399032e-02 -1.40214965e-01 -5.69830775e-01 8.40017140e-01
9.25076842e-01 -2.90434718e-01 -4.74458754e-01 -7.78857350e-01
-4.64976698e-01 2.62763143e-01 7.94814676e-02 -1.20985329e+00
1.38472795e-01 8.21867958e-02 8.15398157e-01 -8.63578856e-01
5.37125431e-02 -7.11880803e-01 -3.00692528e-01 7.31763422e-01
-2.41781801e-01 -1.54191300e-01 2.78676838e-01 -5.91338053e-02
-3.82782489e-01 -2.13924825e-01 1.36881101e+00 -3.48864377e-01
-5.99725544e-01 8.88301790e-01 -1.72626257e-01 2.46988177e-01
1.09744763e+00 -2.10750982e-01 -4.53844201e-03 1.38571784e-01
-7.90212452e-01 2.20897093e-01 1.77104084e-03 1.67296067e-01
9.40425098e-01 -1.14156139e+00 -6.69084251e-01 2.35603228e-01
-5.27744353e-01 4.63815600e-01 5.30929804e-01 1.19474292e+00
-1.15782869e+00 5.15934885e-01 -8.32620502e-01 -8.07404041e-01
-1.23446226e+00 2.13673040e-01 6.53112531e-01 -3.78065169e-01
-8.25778723e-01 1.35909235e+00 6.02390945e-01 -1.36080697e-01
3.66599143e-01 -7.00127423e-01 -7.73949444e-01 -8.43446031e-02
7.46862411e-01 1.92236736e-01 2.27613613e-01 -8.34430754e-01
-4.16151613e-01 7.04522908e-01 -3.49793315e-01 2.73632437e-01
1.52872014e+00 -1.17776781e-01 -5.15412807e-01 -1.09874539e-01
1.47370589e+00 -4.28441852e-01 -1.26010084e+00 -1.94281310e-01
-1.72081992e-01 -3.82207721e-01 4.88479435e-01 -6.30539000e-01
-1.94158161e+00 1.27159488e+00 1.17126656e+00 -1.43796369e-01
1.20745945e+00 -6.37734458e-02 1.23291934e+00 -2.09373534e-01
3.31031978e-01 -8.26722443e-01 -5.28924689e-02 3.40736777e-01
7.37410307e-01 -1.40082490e+00 -2.85589565e-02 -3.32446277e-01
-2.99984694e-01 1.20285153e+00 8.70250642e-01 -3.72554153e-01
7.04926968e-01 5.42435169e-01 -7.47831771e-03 -5.01898944e-01
-1.11541249e-01 3.44244502e-02 4.40529346e-01 5.67706704e-01
5.89555621e-01 4.27338555e-02 -4.82607961e-01 6.63963377e-01
-1.12807363e-01 -7.75699317e-02 3.17139566e-01 6.01150990e-01
-7.31644511e-01 -5.34173191e-01 -3.19597900e-01 8.22476983e-01
-9.95635748e-01 -2.72346318e-01 1.85945719e-01 4.32963997e-01
3.97214442e-01 6.85005724e-01 2.29102954e-01 7.55793303e-02
1.92921367e-02 -5.17291665e-01 6.87253058e-01 -5.66462338e-01
-7.85774648e-01 6.54852539e-02 -4.23941523e-01 -6.28314197e-01
-3.61934066e-01 -3.96987528e-01 -1.81812119e+00 -9.96432453e-02
1.85349006e-02 9.47001427e-02 8.22452486e-01 1.03574026e+00
-1.05867134e-02 9.83595908e-01 6.35697007e-01 -8.80032897e-01
-8.19183290e-02 -9.93541300e-01 -5.92497826e-01 2.54079729e-01
4.85477030e-01 -6.33053601e-01 4.81949970e-02 -2.36393631e-01]
|
[14.258218765258789, -2.3778746128082275]
|
abd2ca0b-b314-4ac3-aed0-0449781632f3
|
surgical-fine-tuning-for-grape-bunch
|
2307.00837
| null |
https://arxiv.org/abs/2307.00837v1
|
https://arxiv.org/pdf/2307.00837v1.pdf
|
Surgical fine-tuning for Grape Bunch Segmentation under Visual Domain Shifts
|
Mobile robots will play a crucial role in the transition towards sustainable agriculture. To autonomously and effectively monitor the state of plants, robots ought to be equipped with visual perception capabilities that are robust to the rapid changes that characterise agricultural settings. In this paper, we focus on the challenging task of segmenting grape bunches from images collected by mobile robots in vineyards. In this context, we present the first study that applies surgical fine-tuning to instance segmentation tasks. We show how selectively tuning only specific model layers can support the adaptation of pre-trained Deep Learning models to newly-collected grape images that introduce visual domain shifts, while also substantially reducing the number of tuned parameters.
|
['Matteo Matteucci', 'Matteo Gatti', 'Nico Catalano', 'Riccardo Bertoglio', 'Agnese Chiatti']
|
2023-07-03
| null | null | null | null |
['instance-segmentation']
|
['computer-vision']
|
[ 5.05146623e-01 1.61814064e-01 -5.60193846e-04 -4.78617907e-01
2.00978950e-01 -1.11331773e+00 1.78550154e-01 4.60552037e-01
-3.99780333e-01 4.55781311e-01 -7.86939561e-01 -4.16523755e-01
-2.08216123e-02 -8.31392705e-01 -8.76326382e-01 -5.34658551e-01
-1.22902304e-01 5.56198359e-01 3.58014345e-01 -6.15849435e-01
1.49887130e-02 8.27118158e-01 -1.51556003e+00 1.12270936e-01
9.11573648e-01 7.01093197e-01 1.03912961e+00 1.04407048e+00
3.30752671e-01 4.92875800e-02 -4.50304329e-01 1.71889931e-01
2.16731116e-01 -1.57569602e-01 -6.07330501e-01 3.77927840e-01
7.77509287e-02 -3.18613261e-01 2.39341289e-01 1.06049323e+00
2.26326957e-02 2.99824420e-02 7.31897116e-01 -9.63885963e-01
-6.49741113e-01 5.46706796e-01 -5.22533715e-01 -1.62613783e-02
-2.70623922e-01 3.97629082e-01 4.66133893e-01 -2.52479434e-01
8.04839015e-01 1.16732001e+00 7.62012661e-01 4.04531777e-01
-1.45067227e+00 -3.27286939e-03 5.32947719e-01 3.09804231e-01
-8.44199002e-01 -4.96158272e-01 4.01782990e-01 -4.21498954e-01
8.78506601e-01 -1.77991748e-01 6.70682609e-01 8.82151186e-01
1.46982223e-01 4.42800581e-01 6.02085710e-01 -3.89123827e-01
3.25166225e-01 -1.13751464e-01 -1.60115480e-01 5.00117123e-01
6.30498946e-01 -5.39070070e-02 2.32996732e-01 4.47711349e-01
8.59651029e-01 -2.95623928e-01 -6.74549639e-02 -9.98160958e-01
-1.08241391e+00 7.85792053e-01 9.23588932e-01 3.22638631e-01
-7.14183748e-01 1.96439683e-01 5.10209143e-01 8.15169960e-02
-1.37614608e-01 9.56926882e-01 -1.11947858e+00 2.05345213e-01
-4.12850142e-01 -9.01769996e-02 7.91727901e-01 1.07747066e+00
6.06872678e-01 2.42714770e-02 1.78206548e-01 7.76107907e-01
3.90611261e-01 7.00357556e-01 1.91255465e-01 -1.25388396e+00
-1.39725700e-01 6.09904110e-01 3.51030767e-01 -8.06959212e-01
-7.73660719e-01 -9.76524055e-02 -6.57525241e-01 3.38008434e-01
6.10113502e-01 -4.00306970e-01 -1.33239925e+00 1.50413895e+00
3.78271520e-01 -5.05248725e-01 2.74267286e-01 5.92003226e-01
5.99260271e-01 4.22892928e-01 3.71356636e-01 -2.22924165e-02
1.44747126e+00 -6.66417420e-01 -2.49017045e-01 -8.32136214e-01
4.99431163e-01 -5.94103396e-01 1.01454008e+00 2.34580383e-01
-7.02990294e-01 -6.87027156e-01 -1.35286701e+00 1.69895366e-01
-8.62900198e-01 4.82516021e-01 8.65039468e-01 4.46348965e-01
-8.97520244e-01 5.90910971e-01 -1.06388438e+00 -1.08421481e+00
6.73646569e-01 3.54791999e-01 -4.68040287e-01 -7.88759813e-03
-7.75481045e-01 1.18141770e+00 8.75021696e-01 6.76168203e-01
-1.12545204e+00 -2.51012117e-01 -9.50646639e-01 1.29933372e-01
5.33923388e-01 -3.03643614e-01 1.51597214e+00 -1.00872314e+00
-1.71335149e+00 1.11896813e+00 7.06050768e-02 -5.76292276e-01
2.62335241e-01 -4.88398783e-02 2.39802867e-01 1.98201880e-01
4.52385917e-02 1.18791640e+00 7.26772964e-01 -1.55697250e+00
-9.52155054e-01 -5.76035559e-01 3.04513693e-01 1.20501660e-01
-1.04456998e-01 -4.14942682e-01 -1.19086444e-01 -1.93560924e-02
2.24602520e-01 -1.22469449e+00 -6.15128100e-01 2.60119498e-01
-1.23052321e-01 2.05027297e-01 9.62245762e-01 -3.93998861e-01
7.44524971e-02 -2.07756805e+00 2.22443298e-01 -2.17689872e-01
-4.67581414e-02 7.15611756e-01 -3.94272417e-01 1.44996613e-01
3.89042169e-01 -1.14791133e-01 -4.82037425e-01 3.23942930e-01
-1.34836212e-01 6.19128287e-01 8.21559578e-02 3.23607117e-01
4.06029552e-01 9.92543519e-01 -9.43571687e-01 -1.08058214e-01
3.77972871e-01 1.24377841e-02 -1.14838921e-01 2.66964845e-02
-6.98010266e-01 5.65097094e-01 -3.04470837e-01 7.93121696e-01
9.83125687e-01 -1.02371357e-01 6.29117250e-01 -6.51658624e-02
-1.75170928e-01 -3.51148725e-01 -7.43101954e-01 1.50656962e+00
-5.61981022e-01 7.24630237e-01 8.60227406e-01 -1.36773753e+00
1.01016390e+00 -1.20565578e-01 2.23137572e-01 -5.32981515e-01
2.37711594e-01 1.67008758e-01 2.21495450e-01 -5.78086853e-01
4.52383667e-01 1.54681340e-01 -1.09219775e-01 -2.86628872e-01
1.81823164e-01 -6.82995915e-01 4.19488221e-01 -2.51324743e-01
6.40571415e-01 4.80417103e-01 5.18817544e-01 -2.68010765e-01
3.35611254e-01 7.10516810e-01 5.64110160e-01 7.96959817e-01
-6.94598556e-01 1.35842338e-01 5.42501569e-01 -3.83251399e-01
-1.09878051e+00 -8.50537241e-01 -2.79633760e-01 1.26319063e+00
3.92445117e-01 3.29601854e-01 -8.44335020e-01 -5.72419763e-01
2.95078337e-01 5.61638772e-01 -6.41649961e-01 -3.41919810e-01
-4.34908032e-01 -1.01263070e+00 5.13045609e-01 6.43923461e-01
6.55105233e-01 -1.49395978e+00 -1.46573019e+00 5.07552743e-01
1.11514553e-01 -1.48465812e+00 4.14432645e-01 1.00095487e+00
-8.80624652e-01 -1.14911318e+00 -6.28424168e-01 -1.22165751e+00
5.91960609e-01 6.16118014e-01 9.43985403e-01 -1.78279594e-01
-5.35935462e-01 1.36979759e-01 -6.34426057e-01 -8.14164877e-01
-7.34632611e-01 7.53509879e-01 -2.91959226e-01 -5.59776425e-01
3.85529071e-01 -3.76817882e-01 -3.48990262e-01 2.66229033e-01
-7.87989497e-01 -1.55129418e-01 7.00841308e-01 7.36695945e-01
5.43973863e-01 -5.39826639e-02 6.06180966e-01 -8.59443605e-01
3.05688739e-01 -3.83604288e-01 -8.67360473e-01 4.30283010e-01
-3.85223567e-01 -1.34063467e-01 8.09578359e-01 -4.33887482e-01
-9.29456413e-01 8.37968707e-01 1.95739567e-01 2.69530386e-01
-8.15788627e-01 4.80394661e-01 -3.63042653e-01 -1.33005202e-01
9.02556360e-01 -2.62396455e-01 1.87530853e-02 -2.52892315e-01
7.95636952e-01 6.35592341e-01 8.12025428e-01 -2.48620659e-01
5.75066805e-01 5.98079860e-01 2.14852512e-01 -1.14658689e+00
-4.46938783e-01 -2.92098403e-01 -1.32340360e+00 -4.24895436e-02
9.58677888e-01 -7.00356305e-01 -6.71742857e-01 7.25688875e-01
-1.00594652e+00 -1.00850117e+00 -3.54645222e-01 1.57512993e-01
-7.44216561e-01 2.20715210e-01 -3.62710714e-01 -3.52566838e-01
-2.04179585e-01 -1.12005818e+00 9.67553318e-01 6.66205466e-01
3.29831727e-02 -7.86169052e-01 -1.51271656e-01 -1.75551310e-01
3.41091156e-01 3.78299892e-01 8.07704806e-01 -2.40756154e-01
-2.29263261e-01 -2.51308501e-01 -4.50814188e-01 3.20179880e-01
3.82452458e-01 3.46448660e-01 -1.11113083e+00 -1.14092924e-01
-2.73080349e-01 -2.60094404e-01 9.00456905e-01 9.31150675e-01
6.53893650e-01 2.12317914e-01 -5.32915711e-01 8.22084904e-01
1.18617117e+00 6.21105194e-01 4.23498034e-01 7.26140738e-01
6.67795777e-01 9.11861658e-01 9.63146687e-01 2.14204311e-01
3.68155003e-01 4.27452564e-01 9.87392485e-01 -1.66463360e-01
3.39101732e-01 1.79634877e-02 -9.31870099e-03 -3.65380570e-02
-2.10677879e-03 -2.62822807e-01 -7.99230099e-01 8.10889006e-01
-2.08056521e+00 -5.80700874e-01 -2.38341182e-01 1.68858969e+00
5.08252025e-01 8.49681124e-02 -1.69532359e-01 -1.31575853e-01
6.99678719e-01 -1.11223221e-01 -1.10116744e+00 -7.20316291e-01
-2.73803800e-01 -2.99327206e-02 1.06921387e+00 3.42164367e-01
-1.59910321e+00 1.65456831e+00 6.40987539e+00 5.64362220e-02
-1.09447479e+00 -3.11576277e-01 3.54410857e-01 6.11866593e-01
3.47747594e-01 1.08766720e-01 -5.37134588e-01 -5.96423000e-02
8.80950332e-01 4.85408723e-01 5.29241621e-01 1.02090871e+00
2.88877249e-01 -3.86026978e-01 -7.89007425e-01 3.99150699e-01
-3.13187838e-01 -9.34338689e-01 -4.48567450e-01 -8.59635994e-02
6.45506620e-01 -1.85780376e-02 -2.12367140e-02 1.69168726e-01
8.24639142e-01 -9.28528965e-01 5.08904219e-01 3.30537170e-01
5.62309384e-01 -3.19122642e-01 6.62721276e-01 3.40661645e-01
-1.06198955e+00 -3.94217372e-01 -8.03800583e-01 -3.28132920e-02
3.91008295e-02 3.53692651e-01 -1.36632609e+00 1.65276363e-01
9.91291881e-01 7.92366922e-01 -7.69672751e-01 9.37186837e-01
-3.37372780e-01 2.37134457e-01 -3.82965833e-01 2.78549995e-02
3.50812942e-01 -2.61201173e-01 3.67051125e-01 1.02992535e+00
2.50037223e-01 -2.46465519e-01 1.27052302e-02 6.67922616e-01
1.12727098e-01 -2.91618466e-01 -8.77715766e-01 -1.62057549e-01
4.68945146e-01 1.48621738e+00 -1.11640096e+00 -1.17018148e-01
1.34863809e-01 1.22804415e+00 1.43301338e-01 2.75652021e-01
-4.88891423e-01 -4.55565214e-01 6.91171527e-01 -2.43350297e-01
8.81407738e-01 -5.63134432e-01 -5.52214265e-01 -6.49143338e-01
-2.25757197e-01 -6.74010038e-01 6.48582503e-02 -8.23107660e-01
-8.70923102e-01 3.02223057e-01 -1.72879621e-01 -6.42121077e-01
-2.28336200e-01 -1.15305269e+00 -3.15549046e-01 6.06707454e-01
-1.56709409e+00 -1.44869137e+00 -8.68042767e-01 -2.56047882e-02
7.12082624e-01 -7.85083137e-03 1.06047678e+00 -3.32307786e-01
-4.81185704e-01 8.72160271e-02 5.21387100e-01 -2.52161592e-01
7.37454951e-01 -1.42978632e+00 6.49960697e-01 8.63393605e-01
-2.08918095e-01 2.35705942e-01 7.98621655e-01 -5.03534317e-01
-1.22078514e+00 -1.17322719e+00 3.21949959e-01 -2.74404347e-01
3.69885862e-01 -3.01836312e-01 -8.48010421e-01 8.07702661e-01
-6.06473908e-03 3.71821150e-02 9.96251479e-02 -6.63265735e-02
1.08110346e-01 -2.89124876e-01 -1.42006791e+00 4.85316664e-01
9.37753856e-01 -1.23215646e-01 -9.68762264e-02 1.60792425e-01
5.31188190e-01 -5.50469637e-01 -7.89640069e-01 6.48170412e-01
5.65393209e-01 -3.41193825e-01 1.07042503e+00 -6.97261751e-01
1.25966996e-01 -4.04724330e-01 -1.52631849e-01 -1.57060206e+00
-5.91173887e-01 -3.99244845e-01 2.48428434e-01 1.12545812e+00
3.76703560e-01 -4.52811450e-01 7.17432678e-01 1.30752251e-01
-1.36252806e-01 -6.11402281e-02 -4.64198291e-01 -5.59028625e-01
3.51005197e-01 1.39359742e-01 6.03067517e-01 6.81528091e-01
-2.26492956e-01 1.71388045e-01 1.58877611e-01 5.71596444e-01
2.70545632e-01 -7.85854757e-02 7.31479585e-01 -1.67640221e+00
6.55069426e-02 -3.14805359e-01 -5.63620329e-01 -8.55645537e-01
-1.85420532e-02 -5.05226254e-01 8.49825442e-01 -1.75436401e+00
-2.48804510e-01 -4.62977231e-01 1.74983039e-01 5.73667347e-01
-1.90322250e-01 -2.93795746e-02 2.36114115e-01 -5.69586270e-02
-2.55856395e-01 2.47410148e-01 1.19841492e+00 -2.98945874e-01
-6.20838284e-01 2.74990678e-01 -9.66687620e-01 8.32747161e-01
1.37258887e+00 -1.67362168e-01 -3.68250012e-01 -5.85885584e-01
-2.58307569e-02 -4.38472033e-01 4.21846926e-01 -9.36407983e-01
-3.54840308e-02 -4.87204552e-01 5.64384460e-01 -2.92102754e-01
4.59286012e-02 -9.06913519e-01 -3.70886207e-01 6.83652103e-01
-2.06394166e-01 -1.57386303e-01 7.58909404e-01 5.12149751e-01
2.78650850e-01 -7.00613916e-01 1.03509307e+00 -5.20283222e-01
-1.35230505e+00 -8.69795457e-02 -1.00825965e+00 -3.21734130e-01
1.35287941e+00 -1.77959204e-01 -3.85695070e-01 7.85330087e-02
-9.54154849e-01 4.63404894e-01 6.90805078e-01 6.17793858e-01
1.75736025e-01 -5.77977717e-01 -5.77082455e-01 2.41846919e-01
2.86916345e-01 5.09089112e-01 1.50590062e-01 2.91584402e-01
-1.17949426e+00 4.30090696e-01 -7.62495458e-01 -8.07316422e-01
-1.03139484e+00 7.23978639e-01 6.12673044e-01 -3.97490263e-02
-5.32507122e-01 7.66606867e-01 2.16271490e-01 -8.80505860e-01
1.71561670e-02 -7.36628532e-01 -5.58826208e-01 1.71438575e-01
2.85699405e-02 8.90129283e-02 1.43832356e-01 -5.10268211e-01
-3.91336799e-01 3.86449188e-01 4.54267487e-02 1.73733711e-01
1.44458818e+00 -2.81332344e-01 8.53920504e-02 3.55296105e-01
5.29628158e-01 -8.13073635e-01 -1.99525118e+00 2.20573500e-01
2.60924399e-01 -6.94260225e-02 -6.03758618e-02 -8.55627835e-01
-9.49895859e-01 9.94254529e-01 1.09174800e+00 2.89082289e-01
1.02778792e+00 -1.52289957e-01 4.03363228e-01 8.29475284e-01
3.45566392e-01 -1.29527259e+00 -2.89754331e-01 6.15880370e-01
6.98547959e-01 -1.52590775e+00 -2.72002876e-01 -4.77631956e-01
-8.34249854e-01 1.22415447e+00 7.72766769e-01 -2.01773927e-01
5.03792703e-01 2.75317192e-01 3.54438215e-01 -1.17795505e-01
-3.78437489e-01 -7.02243090e-01 -3.38472664e-01 1.70501328e+00
6.25753701e-02 4.04490650e-01 2.02639043e-01 1.98034227e-01
6.31688833e-02 3.93672800e-03 7.49947965e-01 1.14632440e+00
-9.45595324e-01 -9.79402602e-01 -4.42254066e-01 1.91118822e-01
1.53959110e-01 3.28061998e-01 -6.42135561e-01 6.92649961e-01
2.90475756e-01 9.95174468e-01 1.02354307e-03 -1.90656595e-02
6.55526340e-01 -1.41067415e-01 6.27081513e-01 -6.91212237e-01
-3.13491404e-01 -2.01409772e-01 -1.06600307e-01 -9.69515219e-02
-2.78457075e-01 -8.96388412e-01 -1.22780299e+00 7.74283335e-02
-1.93335101e-01 -3.52501482e-01 9.57696795e-01 7.73244262e-01
3.42133820e-01 8.22068155e-01 3.36601198e-01 -1.26159775e+00
-5.40656567e-01 -1.05499375e+00 -6.32513285e-01 -3.56002450e-02
3.18483651e-01 -6.58780217e-01 1.00650951e-01 5.12152433e-01]
|
[9.113179206848145, -1.555993914604187]
|
369d8d21-0592-419e-b7e8-d1198cb200c0
|
utrnet-high-resolution-urdu-text-recognition
|
2306.15782
| null |
https://arxiv.org/abs/2306.15782v2
|
https://arxiv.org/pdf/2306.15782v2.pdf
|
UTRNet: High-Resolution Urdu Text Recognition In Printed Documents
|
In this paper, we propose a novel approach to address the challenges of printed Urdu text recognition using high-resolution, multi-scale semantic feature extraction. Our proposed UTRNet architecture, a hybrid CNN-RNN model, demonstrates state-of-the-art performance on benchmark datasets. To address the limitations of previous works, which struggle to generalize to the intricacies of the Urdu script and the lack of sufficient annotated real-world data, we have introduced the UTRSet-Real, a large-scale annotated real-world dataset comprising over 11,000 lines and UTRSet-Synth, a synthetic dataset with 20,000 lines closely resembling real-world and made corrections to the ground truth of the existing IIITH dataset, making it a more reliable resource for future research. We also provide UrduDoc, a benchmark dataset for Urdu text line detection in scanned documents. Additionally, we have developed an online tool for end-to-end Urdu OCR from printed documents by integrating UTRNet with a text detection model. Our work not only addresses the current limitations of Urdu OCR but also paves the way for future research in this area and facilitates the continued advancement of Urdu OCR technology. The project page with source code, datasets, annotations, trained models, and online tool is available at abdur75648.github.io/UTRNet.
|
['Chetan Arora', 'Arjun Ghosh', 'Abdur Rahman']
|
2023-06-27
| null | null | null | null |
['optical-character-recognition', 'line-detection']
|
['computer-vision', 'computer-vision']
|
[ 2.71892071e-01 -2.32877523e-01 -3.54305357e-02 -1.94170788e-01
-9.73425210e-01 -7.01463640e-01 4.98921782e-01 -1.92312956e-01
-2.02545539e-01 5.41422904e-01 1.45997286e-01 -3.80270034e-01
4.47411776e-01 -6.06572568e-01 -7.86061764e-01 -6.23833761e-02
6.69664383e-01 3.96235466e-01 3.80115509e-01 -1.32535234e-01
5.19846201e-01 7.54430234e-01 -1.38076913e+00 4.82948691e-01
8.15281808e-01 1.07942665e+00 2.82665074e-01 1.01940036e+00
-1.68236643e-01 8.73976707e-01 -1.08798647e+00 -3.75352859e-01
2.69880652e-01 -5.60578942e-01 -5.77111721e-01 1.75845381e-02
9.56549287e-01 -7.94992566e-01 -7.94615209e-01 8.79562140e-01
9.39072192e-01 -1.67300165e-01 4.57856178e-01 -7.17135549e-01
-1.33866370e+00 4.07650650e-01 -5.82277954e-01 1.48985952e-01
4.89564389e-01 -2.25840092e-01 9.22002017e-01 -7.02997208e-01
1.08181560e+00 8.97213042e-01 6.65175617e-01 7.59382665e-01
-4.92619485e-01 -4.67928439e-01 -4.91887271e-01 -2.72712521e-02
-1.08994007e+00 -2.61034787e-01 4.08540934e-01 -2.17672199e-01
1.04645109e+00 2.19822958e-01 2.28984967e-01 1.36436403e+00
-2.40852218e-02 1.44031489e+00 8.09787869e-01 -8.39959204e-01
-1.03268988e-01 1.23150207e-01 4.74736124e-01 7.40930974e-01
2.48435333e-01 -2.19599277e-01 -3.40849221e-01 1.70956627e-01
9.20518816e-01 -3.22540015e-01 -4.78902638e-01 -5.18684648e-03
-1.05452204e+00 5.80905020e-01 -7.55232386e-03 5.92889249e-01
1.72257781e-01 6.35286272e-02 5.80709696e-01 2.27543920e-01
5.19146502e-01 3.57371002e-01 -5.14384627e-01 -6.40173256e-01
-1.04469359e+00 2.49970201e-02 8.30649912e-01 1.38043356e+00
3.86053503e-01 1.24468639e-01 -2.20808238e-01 1.21967661e+00
1.42489180e-01 4.72279608e-01 6.99615955e-01 -7.49739707e-01
6.56305790e-01 4.96451944e-01 3.39296833e-02 -5.52187324e-01
-2.02064812e-01 2.32198015e-02 -1.74758777e-01 -3.23495120e-02
6.39086187e-01 -3.36205870e-01 -1.16034865e+00 8.25417161e-01
-1.39653459e-01 -2.74786711e-01 1.53505504e-02 9.34802711e-01
9.12050784e-01 6.23192906e-01 -6.81084692e-01 4.39747930e-01
1.26759958e+00 -1.16855049e+00 -9.50207055e-01 1.51559532e-01
7.92546749e-01 -1.37330067e+00 1.39827704e+00 4.82594609e-01
-6.88419819e-01 -3.23113233e-01 -1.45114100e+00 -5.21636724e-01
-7.77778089e-01 7.62352943e-01 4.54375476e-01 1.02608979e+00
-9.63962615e-01 5.44937611e-01 -3.05474669e-01 -7.74567842e-01
3.25725943e-01 -6.76259995e-02 -1.36647642e-01 -2.73530543e-01
-1.03108263e+00 7.57865250e-01 3.31780165e-01 1.22237056e-01
-3.58232677e-01 -4.70065117e-01 -7.15752304e-01 -3.18344861e-01
4.46208388e-01 2.85332143e-01 1.55316961e+00 -5.03007472e-01
-1.70529282e+00 7.49168158e-01 1.25392303e-01 -2.26336598e-01
1.07317364e+00 -5.73411405e-01 -8.57144117e-01 2.31827289e-01
-9.21434630e-03 4.15612131e-01 6.82884812e-01 -1.10661495e+00
-6.41204596e-01 -1.77402139e-01 -3.94477278e-01 -2.67797559e-01
-3.51398289e-01 2.09621221e-01 -7.92530656e-01 -9.13988769e-01
-2.33753949e-01 -6.52398646e-01 6.37626886e-01 -5.23445085e-02
-5.86831331e-01 -2.31162593e-01 1.45464718e+00 -1.08558631e+00
1.06431067e+00 -2.02616119e+00 -5.35740376e-01 -7.29191396e-03
-1.27821397e-02 5.40220797e-01 -2.85236508e-01 4.13058013e-01
4.65412289e-02 2.27665678e-01 3.06241717e-02 -3.64924639e-01
1.66502401e-01 -2.17165440e-01 -2.77481705e-01 4.54550087e-01
-2.60475092e-02 1.02009177e+00 -6.13015711e-01 -3.02706629e-01
3.51770550e-01 4.09349859e-01 1.68712988e-01 2.45524701e-02
-3.68340492e-01 -2.41902635e-01 -1.86884657e-01 1.14611554e+00
6.63649678e-01 -9.64943096e-02 -1.94147021e-01 -2.65430182e-01
-2.77855337e-01 -1.01060703e-01 -1.19452763e+00 1.64089990e+00
-2.54588693e-01 1.43447948e+00 -4.21324819e-01 -1.96179807e-01
1.15046203e+00 2.75872409e-01 1.48962453e-01 -1.00636888e+00
5.38722515e-01 6.38110936e-01 -5.04895508e-01 -4.95302349e-01
1.07581794e+00 7.18054533e-01 1.03607200e-01 6.57225728e-01
4.07244787e-02 -2.60918677e-01 6.79169118e-01 1.61002263e-01
1.07586956e+00 4.45066869e-01 -4.39794511e-02 7.01160207e-02
3.18735510e-01 -6.36827722e-02 1.05339959e-01 8.36924076e-01
-2.58648098e-01 1.29234028e+00 5.85585535e-01 -3.57205778e-01
-1.58132243e+00 -7.05054343e-01 -3.95897686e-01 5.98638833e-01
-2.01770052e-01 -2.43050128e-01 -1.00543463e+00 -9.10381913e-01
-1.06815919e-01 8.97359729e-01 -6.81072831e-01 4.82368141e-01
-5.69261312e-01 -6.61968827e-01 1.20283186e+00 6.21783733e-01
8.24791014e-01 -1.05575144e+00 -2.72798836e-01 1.29926205e-01
-5.32984920e-02 -1.65611196e+00 -8.58804822e-01 -1.55958226e-02
-5.78937232e-01 -1.30085838e+00 -1.28800058e+00 -7.33340979e-01
2.99952865e-01 1.62856311e-01 7.46216595e-01 -1.05085000e-01
-7.49976814e-01 5.04970610e-01 -6.52455270e-01 -5.37798405e-01
-6.02390409e-01 1.24484979e-01 -4.20216471e-01 -2.50684917e-01
6.43580377e-01 2.24827647e-01 -2.64141411e-01 2.74664342e-01
-9.17473674e-01 -2.33295336e-01 4.27078247e-01 5.24742901e-01
3.33943844e-01 -3.36609036e-01 4.20293480e-01 -7.08027899e-01
9.42492604e-01 1.64590374e-01 -8.41530561e-01 4.44356352e-01
-4.67091411e-01 -3.00989389e-01 5.19606829e-01 -2.41280898e-01
-1.15414357e+00 -2.68033326e-01 -2.74115860e-01 -2.55821377e-01
-2.61040658e-01 2.65861657e-02 -7.82821421e-03 -9.45857074e-03
7.48977542e-01 3.67362440e-01 -2.43838593e-01 -7.31813073e-01
3.64910036e-01 1.43184245e+00 5.94714880e-01 -4.83777046e-01
3.97459924e-01 3.95742267e-01 -5.90900242e-01 -1.26806009e+00
-7.67843783e-01 -5.32420099e-01 -7.47579455e-01 -4.22552899e-02
9.27863777e-01 -7.35276520e-01 -2.77032048e-01 1.11380434e+00
-1.36914527e+00 -5.56321681e-01 -2.10723311e-01 2.07149163e-01
-4.84353393e-01 9.07198608e-01 -1.20311749e+00 -6.86295867e-01
-5.44881821e-01 -9.87647772e-01 1.30426407e+00 4.43264395e-01
-1.72058031e-01 -8.68202507e-01 1.06942035e-01 5.62340438e-01
1.50347561e-01 2.91334838e-01 7.73330510e-01 -7.23292291e-01
-5.19913256e-01 -7.27312565e-01 -8.25320184e-01 6.30283058e-01
1.56877175e-01 4.85414863e-01 -1.02652431e+00 -1.90517807e-03
-4.83438402e-01 -6.48657262e-01 6.81342244e-01 -2.06117760e-02
1.26698160e+00 2.41072431e-01 4.13616635e-02 2.90628016e-01
1.41605425e+00 4.95262355e-01 9.86977577e-01 8.55406523e-01
8.50325167e-01 2.18266100e-01 5.58735490e-01 4.35273945e-01
8.15680847e-02 5.58301151e-01 2.94832643e-02 -1.86593860e-01
-6.50951266e-01 -2.62403190e-01 1.17159091e-01 6.26036048e-01
8.35023299e-02 -8.16212296e-01 -7.96234787e-01 2.75356978e-01
-1.57237613e+00 -6.10656857e-01 -4.87530440e-01 1.85312855e+00
6.14803731e-01 7.76416734e-02 6.23765178e-02 1.72950074e-01
9.92300034e-01 2.23076437e-02 -4.37432021e-01 -5.95122337e-01
-4.97670561e-01 6.60051927e-02 7.79748857e-01 6.50710911e-02
-1.02331221e+00 1.26944637e+00 6.59457970e+00 9.20416594e-01
-1.13496077e+00 8.32490996e-02 4.28229481e-01 2.51180589e-01
4.19803746e-02 -6.24603033e-01 -1.06927025e+00 4.73826081e-01
8.33357871e-01 3.86697978e-01 3.76624912e-01 7.58642077e-01
2.02848941e-01 -6.06914721e-02 -9.71936703e-01 9.67282593e-01
6.89164400e-01 -1.53576601e+00 3.50770429e-02 -6.94676023e-03
7.35992134e-01 3.96221757e-01 -2.26505116e-01 -2.53836382e-02
2.09015861e-01 -7.33477652e-01 9.65168893e-01 1.96264699e-01
1.23577499e+00 -5.58514357e-01 8.91543329e-01 -1.81139231e-01
-7.83637166e-01 2.67652392e-01 -4.05597836e-01 5.62573433e-01
1.74930301e-02 4.93555754e-01 -8.81273806e-01 4.62421894e-01
7.08490908e-01 8.32287967e-01 -8.37509036e-01 9.41350162e-01
-2.55506635e-01 4.17503327e-01 -1.53041229e-01 -4.15615499e-01
3.73104304e-01 -2.02887833e-01 1.76854283e-01 1.62846899e+00
5.72284400e-01 -4.56776798e-01 -5.14553249e-01 7.64822662e-01
-5.18108964e-01 3.09688240e-01 -4.50550407e-01 -6.72629058e-01
1.87250733e-01 1.24635577e+00 -9.73423064e-01 -3.41836542e-01
-7.30642736e-01 1.48853076e+00 1.26022726e-01 3.63092929e-01
-7.62631238e-01 -1.23718500e+00 2.18770325e-01 -2.71243185e-01
3.77570570e-01 -3.30763608e-01 -3.17695409e-01 -1.38503063e+00
2.30299324e-01 -1.05823231e+00 2.14486897e-01 -1.17256486e+00
-1.01731610e+00 5.89865685e-01 -5.40387571e-01 -1.16031909e+00
5.84551767e-02 -1.21882689e+00 -4.12123710e-01 6.38689041e-01
-1.52175462e+00 -1.26459861e+00 -4.85647708e-01 3.81254882e-01
1.05710566e+00 -4.49633360e-01 5.82669079e-01 3.87382239e-01
-7.78671443e-01 8.90886247e-01 7.12457120e-01 7.32754052e-01
1.03599608e+00 -1.34945619e+00 7.10092127e-01 8.86797607e-01
1.66382208e-01 2.61209577e-01 5.99134028e-01 -8.80934954e-01
-1.21682703e+00 -9.33316529e-01 7.39776433e-01 -6.09821916e-01
9.10975933e-01 -7.17857420e-01 -7.55848408e-01 7.84438372e-01
2.45974958e-01 -2.48127058e-01 4.97188091e-01 -3.80909890e-01
-4.23724473e-01 2.38201201e-01 -1.17986465e+00 7.24062383e-01
7.78518438e-01 -6.29974484e-01 -6.09926701e-01 5.54375350e-01
6.79045081e-01 -7.77903914e-01 -9.81642187e-01 -6.24971539e-02
7.96575010e-01 -8.48460376e-01 4.23973203e-01 -4.98296991e-02
6.94909573e-01 -1.03392899e-01 -1.85902879e-01 -9.66129899e-01
2.54940957e-01 -4.52010572e-01 -3.49325538e-01 1.44608784e+00
5.94614148e-01 -6.28635347e-01 8.53739142e-01 2.32646137e-01
-3.78159404e-01 -5.24176359e-01 -6.70987308e-01 -7.85403728e-01
7.71733671e-02 -6.06128514e-01 5.34583449e-01 7.20736861e-01
-2.04304114e-01 1.63375959e-01 -5.12314498e-01 -2.55812883e-01
3.32093805e-01 -1.02098770e-01 6.06188059e-01 -8.56755435e-01
-1.88331500e-01 -4.09791380e-01 -2.70808190e-01 -1.18886924e+00
-5.72171807e-02 -5.65005243e-01 -9.87484679e-02 -1.73482323e+00
4.91593070e-02 1.94401145e-02 3.88080448e-01 3.61871094e-01
2.17187271e-01 6.54503644e-01 3.93008769e-01 2.75450736e-01
-4.86759245e-01 2.81662524e-01 1.37463641e+00 -2.33634874e-01
-1.12267829e-01 -2.88845360e-01 -3.42817277e-01 3.77234221e-01
8.30410302e-01 -6.90663699e-03 -3.09927892e-02 -4.60256010e-01
-3.57560115e-03 -3.55667144e-01 -1.15895003e-01 -9.84715402e-01
-6.06018379e-02 4.99836445e-01 8.62629890e-01 -1.07524455e+00
1.79211423e-01 -6.08353794e-01 -7.72441149e-01 1.05460048e-01
-2.42655128e-01 1.51815349e-02 3.05682987e-01 2.85007805e-01
1.89163424e-02 -8.08278263e-01 5.95590234e-01 -1.63102761e-01
-9.46965814e-01 2.15192381e-02 -7.31725156e-01 -3.19807082e-02
9.23227429e-01 -6.06341898e-01 -9.61188316e-01 -3.08436066e-01
-1.81379423e-01 -2.29295297e-03 6.15509152e-01 9.19773400e-01
5.10277867e-01 -9.43217874e-01 -5.56996644e-01 5.73762320e-02
2.43443668e-01 -3.15889806e-01 -8.19795653e-02 2.75060833e-01
-1.22716570e+00 8.88464272e-01 -2.21395761e-01 -2.09990755e-01
-1.16435373e+00 1.52759120e-01 4.57042158e-01 2.27161366e-02
-7.35667348e-01 3.67769808e-01 -5.49578726e-01 -5.94811857e-01
2.93970019e-01 -2.95770943e-01 -1.57761857e-01 1.44717649e-01
6.66627944e-01 8.14939737e-01 5.52875638e-01 -3.86629552e-01
1.95118383e-01 6.61645353e-01 -3.76957744e-01 2.29558144e-02
1.13246000e+00 -1.22467123e-01 5.02758585e-02 5.66206932e-01
1.36307740e+00 2.15175390e-01 -1.16845667e+00 7.94869512e-02
1.34351745e-01 -4.31086570e-01 3.03364731e-02 -1.12083054e+00
-9.08850789e-01 7.77486324e-01 9.03007388e-01 3.90320420e-02
6.77385330e-01 -1.54090583e-01 1.13868439e+00 6.74427509e-01
2.03789890e-01 -1.74021971e+00 3.40734720e-01 8.26469541e-01
8.58930767e-01 -1.26025188e+00 -1.29054829e-01 -3.51650387e-01
-3.88028592e-01 1.76224136e+00 8.26675951e-01 5.92050292e-02
1.98871698e-02 4.52191681e-01 6.12897336e-01 -3.41640078e-02
4.43213433e-02 -4.95610982e-02 1.29477575e-01 5.75454950e-01
7.98184872e-01 -3.65908891e-01 -2.59047538e-01 1.21724598e-01
-1.82869196e-01 2.89052904e-01 9.81732965e-01 1.16207087e+00
-2.96919674e-01 -1.14870322e+00 -4.93798912e-01 7.43015349e-01
-6.08459353e-01 -1.10201731e-01 -7.71928370e-01 1.18102515e+00
-2.78200001e-01 6.20510638e-01 1.12980910e-01 -1.27065390e-01
6.19158685e-01 9.87062752e-02 4.94127125e-01 -3.61608505e-01
-2.89216161e-01 1.57109872e-01 2.47771889e-01 -3.56587321e-01
1.49806976e-01 -7.66529799e-01 -1.16800189e+00 -2.44694829e-01
-4.36027944e-01 -2.71751732e-01 9.99934673e-01 7.70263910e-01
3.05566430e-01 6.22872770e-01 2.09071949e-01 -6.64187491e-01
-4.59738821e-01 -1.22032654e+00 -8.16331208e-01 2.28676908e-02
1.17956400e-01 -1.10284448e-01 -2.75662482e-01 -4.19033021e-02]
|
[11.836586952209473, 2.5635507106781006]
|
cf414641-c257-438a-b62b-37856a6767af
|
redditbias-a-real-world-resource-for-bias
|
2106.03521
| null |
https://arxiv.org/abs/2106.03521v1
|
https://arxiv.org/pdf/2106.03521v1.pdf
|
RedditBias: A Real-World Resource for Bias Evaluation and Debiasing of Conversational Language Models
|
Text representation models are prone to exhibit a range of societal biases, reflecting the non-controlled and biased nature of the underlying pretraining data, which consequently leads to severe ethical issues and even bias amplification. Recent work has predominantly focused on measuring and mitigating bias in pretrained language models. Surprisingly, the landscape of bias measurements and mitigation resources and methods for conversational language models is still very scarce: it is limited to only a few types of bias, artificially constructed resources, and completely ignores the impact that debiasing methods may have on the final performance in dialog tasks, e.g., conversational response generation. In this work, we present RedditBias, the first conversational data set grounded in the actual human conversations from Reddit, allowing for bias measurement and mitigation across four important bias dimensions: gender, race, religion, and queerness. Further, we develop an evaluation framework which simultaneously 1) measures bias on the developed RedditBias resource, and 2) evaluates model capability in dialog tasks after model debiasing. We use the evaluation framework to benchmark the widely used conversational DialoGPT model along with the adaptations of four debiasing methods. Our results indicate that DialoGPT is biased with respect to religious groups and that some debiasing techniques can remove this bias while preserving downstream task performance.
|
['Goran Glavaš', 'Ivan Vulić', 'Anne Lauscher', 'Soumya Barikeri']
|
2021-06-07
| null |
https://aclanthology.org/2021.acl-long.151
|
https://aclanthology.org/2021.acl-long.151.pdf
|
acl-2021-5
|
['conversational-response-generation']
|
['natural-language-processing']
|
[-5.62804341e-02 5.94152331e-01 -4.96613234e-01 -7.13302433e-01
-1.03872500e-01 -4.82178181e-01 1.10904849e+00 4.75225560e-02
-4.07761812e-01 1.03401101e+00 1.07881367e+00 -2.52244323e-01
2.18677282e-01 -7.32300043e-01 -1.42061099e-01 -4.50557828e-01
5.19380212e-01 8.43763351e-01 -2.31854022e-01 -8.59147847e-01
3.63641232e-01 2.05255255e-01 -1.04084313e+00 4.00357187e-01
1.13500488e+00 4.14057285e-01 -4.45827425e-01 5.24401665e-01
-1.10953853e-01 1.12779248e+00 -1.06744397e+00 -1.07185972e+00
-1.46774545e-01 -5.42099416e-01 -1.11523461e+00 -4.32893068e-01
6.54416382e-01 -8.49552155e-01 -2.12016374e-01 8.65029275e-01
9.38308299e-01 3.05393655e-02 7.88951814e-01 -1.32762349e+00
-9.84678805e-01 1.18780363e+00 -3.51474583e-02 -2.85585728e-02
3.50255609e-01 1.13277689e-01 8.77040863e-01 -6.61916018e-01
7.03019023e-01 2.02568436e+00 8.13178360e-01 1.42414248e+00
-1.45523798e+00 -9.89466131e-01 9.06051248e-02 -2.64341116e-01
-5.88643074e-01 -6.80227935e-01 8.39556098e-01 -8.00953090e-01
6.56439185e-01 4.65176672e-01 4.42749381e-01 2.04043460e+00
6.11297339e-02 5.45090139e-01 1.55294323e+00 -1.88840583e-01
1.28201425e-01 7.95335054e-01 8.38247836e-01 3.77394438e-01
2.18124643e-01 3.23056102e-01 -8.10002267e-01 -6.35176599e-01
2.82391440e-02 -2.80107439e-01 -1.06078297e-01 -3.74552868e-02
-1.10770810e+00 1.25197899e+00 4.36746746e-01 2.17253938e-01
-7.06652105e-02 -3.06067199e-01 5.35209417e-01 4.36276495e-01
9.01619315e-01 6.20279253e-01 -2.34273195e-01 -1.42472357e-01
-6.40572548e-01 5.27398944e-01 1.24705887e+00 7.06114948e-01
6.43793881e-01 4.51681428e-02 -7.25558996e-01 1.45877576e+00
6.80457801e-02 6.91659451e-01 5.69798112e-01 -9.02559936e-01
4.18339759e-01 8.23879242e-01 2.41618305e-01 -1.12562895e+00
-6.20370328e-01 -8.52874108e-03 -9.74936187e-01 -1.41733244e-01
6.62720382e-01 -7.78539538e-01 -3.93755764e-01 2.21096659e+00
2.85028487e-01 -9.20180500e-01 7.52969012e-02 8.63564014e-01
1.22930241e+00 2.29496181e-01 3.33009422e-01 9.41407457e-02
1.21080530e+00 -8.08543205e-01 -8.39972377e-01 -6.59848332e-01
8.03778291e-01 -6.29103422e-01 1.38940179e+00 2.88083479e-02
-1.07329941e+00 -3.01962495e-01 -7.62885094e-01 -3.41954172e-01
-5.87244213e-01 -1.23274431e-01 6.65603578e-01 1.29364061e+00
-9.59354877e-01 4.13112968e-01 1.48778737e-01 -5.33790350e-01
2.11010560e-01 2.93585330e-01 -4.82022204e-02 6.42588139e-02
-1.87490296e+00 1.52513731e+00 -3.47414240e-02 -1.30122021e-01
-7.80709803e-01 -9.88560081e-01 -5.59849560e-01 -6.56853467e-02
2.18396991e-01 -6.75279200e-01 1.17526972e+00 -1.13314199e+00
-1.65944695e+00 1.17352104e+00 -6.08159378e-02 -4.10681993e-01
8.80669057e-01 -3.13849419e-01 -2.21341833e-01 -5.14465988e-01
-6.43638968e-02 8.58861089e-01 6.64039314e-01 -1.26963615e+00
-1.02478392e-01 -4.22382861e-01 1.65546238e-01 2.28795886e-01
-6.89282417e-01 2.60206789e-01 5.84931612e-01 -6.30960107e-01
-5.57204545e-01 -1.04025888e+00 4.95754406e-02 -4.25845176e-01
-5.27390480e-01 -2.42025390e-01 3.29415023e-01 -8.04003775e-01
1.15920866e+00 -1.74489570e+00 2.15383470e-01 -1.54827207e-01
2.40509480e-01 2.21022576e-01 3.91854234e-02 6.33353651e-01
1.76059440e-01 4.81317431e-01 -2.93404162e-02 -4.45950568e-01
3.05234611e-01 1.17757551e-01 -6.30722880e-01 1.99392170e-01
-7.85972774e-02 6.51525974e-01 -8.66330862e-01 -4.22642767e-01
1.11403972e-01 2.55647182e-01 -1.04429257e+00 3.68784279e-01
-1.04459062e-01 3.85973155e-01 8.10731351e-02 4.01677549e-01
6.56877100e-01 3.92494082e-01 2.60464877e-01 1.65049165e-01
-9.83225256e-02 8.77247393e-01 -4.35083598e-01 1.15988731e+00
-4.19274360e-01 6.28836095e-01 3.52683097e-01 -3.50501001e-01
1.18384600e+00 1.39617827e-02 3.26032415e-02 -5.02138972e-01
3.09707999e-01 6.96277842e-02 3.33535641e-01 -1.77529514e-01
9.69226122e-01 -4.00660664e-01 -2.81144708e-01 6.31063461e-01
4.09432082e-03 -3.29882175e-01 -1.68617174e-01 4.09183502e-01
5.88981032e-01 -6.77481592e-01 1.54478192e-01 -8.10196936e-01
5.20498276e-01 1.02378100e-01 3.60647887e-01 8.29913139e-01
-5.11577189e-01 2.01821417e-01 8.53067458e-01 -1.67886853e-01
-8.06629360e-01 -7.06389189e-01 -1.70598730e-01 1.84969413e+00
-1.89132333e-01 -5.61826639e-02 -9.14447725e-01 -9.61035788e-01
3.27926815e-01 1.26363325e+00 -9.76291895e-01 -6.63132846e-01
-4.17161435e-01 -1.06502938e+00 9.36501205e-01 2.30520248e-01
6.16641819e-01 -9.82301950e-01 -1.20855831e-01 -9.08074975e-02
-4.24480706e-01 -5.98903894e-01 -4.75230366e-01 -1.52395710e-01
-7.62163281e-01 -7.61981905e-01 -6.81946933e-01 -3.43608558e-01
2.84483045e-01 1.32422760e-01 1.52886045e+00 3.31728607e-01
3.95742536e-01 2.00826481e-01 -6.70225173e-02 -7.70298898e-01
-1.12828577e+00 3.69818658e-01 1.34769762e-02 -2.73189574e-01
6.57916427e-01 -1.09159537e-01 -5.82481563e-01 6.20645463e-01
-4.64591175e-01 4.07047421e-02 2.41340250e-01 1.16347694e+00
-8.40215921e-01 -8.60601425e-01 1.15969467e+00 -1.41852522e+00
1.36771095e+00 -5.24410248e-01 -3.14163719e-03 3.93795036e-02
-9.36640084e-01 -2.58474886e-01 2.24428356e-01 -6.17229819e-01
-1.70744777e+00 -8.18056345e-01 -1.95943683e-01 4.94246811e-01
-7.14369491e-02 1.17942937e-01 -9.51945782e-02 2.01579586e-01
1.10894549e+00 -2.55840510e-01 3.28802556e-01 -3.77429754e-01
4.43158686e-01 1.19294727e+00 5.41005656e-02 -8.05315197e-01
3.48055184e-01 3.08143169e-01 -6.23102486e-01 -8.96013439e-01
-7.37909853e-01 -5.61581701e-02 -5.39563954e-01 -4.52730536e-01
6.55343711e-01 -7.86089659e-01 -7.66351700e-01 6.42786443e-01
-1.09488940e+00 -8.04820418e-01 8.27114433e-02 2.02995181e-01
-7.27100484e-03 1.19412690e-01 -8.56008112e-01 -1.02427673e+00
-5.92677653e-01 -7.89810479e-01 6.34814262e-01 -7.39651918e-02
-1.15414488e+00 -1.17455900e+00 4.07094449e-01 8.24222445e-01
7.68296421e-01 -2.55295429e-02 1.50322664e+00 -1.05492043e+00
2.75794268e-01 7.47708157e-02 -2.25601867e-01 4.72467750e-01
-5.28087327e-03 -1.23429708e-01 -1.31085575e+00 -1.89929530e-01
1.40914828e-01 -8.09578657e-01 7.17705786e-01 1.16065212e-01
4.84434128e-01 -6.59980476e-01 -8.47204402e-02 -9.41341892e-02
5.09211719e-01 -2.58476310e-03 3.76343727e-01 4.45055142e-02
6.24946535e-01 1.41462374e+00 4.55479592e-01 3.69997650e-01
6.24972284e-01 5.31204879e-01 2.41990283e-01 -1.63512394e-01
-1.17527083e-01 -4.34644401e-01 5.45285821e-01 7.33067453e-01
1.67583421e-01 -1.07132792e-01 -9.04273331e-01 6.16845489e-01
-1.59619701e+00 -8.12608540e-01 -1.79022133e-01 1.89720535e+00
1.35190880e+00 -1.22002266e-01 3.35054368e-01 -2.63819158e-01
6.92996144e-01 5.38990438e-01 -5.14429867e-01 -9.94260907e-01
-1.48408636e-01 -3.01111490e-01 2.75155216e-01 8.63931358e-01
-4.69184607e-01 1.18539667e+00 6.88248301e+00 2.65814513e-01
-9.73808050e-01 1.22177348e-01 8.67944360e-01 -2.02897891e-01
-5.58339298e-01 -2.21971259e-01 -7.41006672e-01 4.27224487e-01
7.92846441e-01 -2.99728215e-01 5.01232743e-01 1.01614344e+00
1.44617364e-01 -2.53461357e-02 -1.40074325e+00 5.08122206e-01
2.15141222e-01 -8.85028899e-01 2.91948378e-01 7.35609159e-02
7.31113195e-01 -2.06542835e-01 9.00588632e-02 8.95106912e-01
8.96188974e-01 -1.19196141e+00 7.29123592e-01 1.49407849e-01
3.48579139e-01 -5.52514315e-01 8.55467200e-01 5.14216781e-01
3.01849633e-01 -1.79189324e-01 -5.32675862e-01 -4.17128205e-01
-2.45129064e-01 4.13351685e-01 -1.16212881e+00 -9.55954716e-02
5.92266142e-01 4.13526535e-01 -6.80837691e-01 -1.02871232e-01
-1.50451483e-02 8.09987187e-01 2.38725185e-01 -5.43401599e-01
-4.90118973e-02 -1.74578726e-01 4.95266289e-01 1.59540296e+00
-1.33732244e-01 3.50276493e-02 -3.63859206e-01 9.51079071e-01
-4.70235914e-01 2.47134477e-01 -8.40843141e-01 1.05282083e-01
7.65905499e-01 1.12941027e+00 2.28133332e-02 -5.22284687e-01
-1.87579662e-01 5.32915711e-01 5.62652528e-01 2.42569789e-01
-5.23384988e-01 1.83297783e-01 1.10826242e+00 -9.46035385e-02
-6.64807975e-01 2.31005475e-01 -8.25419247e-01 -1.02776694e+00
-5.97370088e-01 -1.62736893e+00 4.59177226e-01 -5.71335435e-01
-1.63490105e+00 2.73505002e-01 1.74230903e-01 -1.34134322e-01
-4.12020624e-01 -5.56805670e-01 -5.70990324e-01 1.07603252e+00
-1.06877542e+00 -1.03521979e+00 -4.69667107e-01 5.03853440e-01
5.55529952e-01 -2.73289740e-01 8.04894030e-01 1.41268015e-01
-7.00100005e-01 7.18105197e-01 -2.90961683e-01 2.63865292e-01
1.50664270e+00 -1.36500943e+00 1.75750688e-01 2.48040706e-01
-7.55319953e-01 9.56602633e-01 8.89618158e-01 -8.32249522e-01
-8.58295441e-01 -7.53509402e-01 1.13485825e+00 -9.76982534e-01
6.38748705e-01 -8.28406036e-01 -9.96528327e-01 8.03907812e-01
6.57707632e-01 -1.19715190e+00 7.75421262e-01 7.00438559e-01
-4.57394719e-01 -5.25555164e-02 -1.37367976e+00 1.12592661e+00
1.20784426e+00 -5.14681995e-01 -8.14001799e-01 3.40957105e-01
5.85646987e-01 -3.26869100e-01 -7.82253087e-01 2.21316561e-01
6.19052231e-01 -1.33589530e+00 8.53610337e-01 -1.02702129e+00
7.24870801e-01 8.21375310e-01 -7.92791769e-02 -1.87262166e+00
-2.64884353e-01 -5.93019128e-01 3.28777462e-01 1.86582029e+00
5.71632445e-01 -9.88889575e-01 6.77788258e-01 1.37008536e+00
8.63481909e-02 -2.99872249e-01 -6.70519412e-01 -2.07411334e-01
9.89665449e-01 4.19342034e-02 8.85257065e-01 1.50519383e+00
1.32540405e-01 8.33611608e-01 -8.75730872e-01 -3.61243278e-01
4.09702331e-01 -4.25575003e-02 1.29501927e+00 -1.30386209e+00
2.54654467e-01 -6.21792078e-01 5.65682530e-01 -8.77678335e-01
6.10497594e-01 -8.15490961e-01 2.18422571e-03 -1.08621752e+00
3.53899181e-01 -4.78126943e-01 4.24096823e-01 1.98851183e-01
-4.49088216e-01 -2.68363088e-01 1.19245075e-01 1.31471723e-01
2.25607425e-01 7.25007176e-01 1.07775044e+00 -1.90997675e-01
-2.68436283e-01 -1.03448868e-01 -1.24389648e+00 8.07146251e-01
9.72581506e-01 -5.16361356e-01 -3.52348328e-01 -3.74300569e-01
2.86058724e-01 -2.04399019e-01 4.61693376e-01 -2.91513592e-01
-1.64795518e-01 -3.55926543e-01 1.77028239e-01 -1.56522006e-01
3.92924696e-01 -4.89228636e-01 -3.42007309e-01 5.57387054e-01
-1.08608902e+00 -2.48616561e-02 1.00892074e-01 1.84696361e-01
1.12566538e-01 -2.47855768e-01 8.16925287e-01 -4.00309116e-02
3.40479277e-02 -3.75399888e-01 -6.21863484e-01 5.32351196e-01
3.61394852e-01 3.40969473e-01 -1.18709576e+00 -7.30254173e-01
-1.69518143e-01 2.55666524e-01 5.00280380e-01 7.20992506e-01
4.40939106e-02 -1.23038793e+00 -1.02345121e+00 -9.47673544e-02
3.95824052e-02 -5.89682400e-01 2.40246207e-01 6.61954880e-01
-3.62421274e-02 3.78154397e-01 -3.28278422e-01 -2.80793458e-01
-1.55699265e+00 4.19004053e-01 4.81686532e-01 -2.42142290e-01
-9.97392181e-03 6.74512446e-01 3.49397928e-01 -1.09158695e+00
2.49092236e-01 -3.11903715e-01 -3.76813143e-01 5.15265644e-01
3.00586671e-01 8.93985212e-01 -1.27185807e-01 -5.92930973e-01
-2.72682726e-01 -1.97791174e-01 -1.51488096e-01 -3.01785737e-01
8.01406443e-01 -2.23082036e-01 -3.49354357e-01 5.82383156e-01
7.53645778e-01 3.32006067e-01 -7.50444353e-01 -1.40255556e-01
5.61928451e-02 -3.61542434e-01 -1.89999297e-01 -1.39005232e+00
-5.75538218e-01 9.57337916e-01 2.41642043e-01 3.91251922e-01
3.58723909e-01 -2.90471345e-01 3.83672923e-01 4.16959256e-01
1.37042310e-02 -1.38838744e+00 9.74561274e-02 7.70968139e-01
1.41309834e+00 -1.44355416e+00 -4.55975868e-02 -3.38610560e-01
-1.15918875e+00 5.01353323e-01 9.60435271e-01 1.86636195e-01
4.34089154e-01 -5.88918664e-02 6.21504009e-01 -9.18516293e-02
-1.02632523e+00 2.44199470e-01 1.57660589e-01 6.92386329e-01
8.76201212e-01 7.87673667e-02 -5.24279892e-01 7.63286591e-01
-8.42077434e-01 -4.48988199e-01 5.38774312e-01 3.25145811e-01
-1.47194222e-01 -8.88347864e-01 -5.43032885e-01 3.63066763e-01
-4.35002506e-01 -3.12556833e-01 -1.73745787e+00 1.00621390e+00
-1.59012705e-01 1.33707392e+00 -1.57678351e-01 -4.88458782e-01
5.11279106e-01 4.98466581e-01 4.16988041e-03 -4.75294054e-01
-1.48982227e+00 -5.20988345e-01 9.93364275e-01 -2.62442619e-01
-3.32031012e-01 -5.03235936e-01 -5.41217625e-01 -1.04090297e+00
-2.33489871e-01 1.58655927e-01 3.56978565e-01 6.09354138e-01
3.50072712e-01 2.16465548e-01 5.14233053e-01 -6.62030518e-01
-8.89984667e-01 -1.71294665e+00 -2.23291472e-01 8.27949882e-01
2.47647807e-01 -6.51009083e-01 -5.89335918e-01 -4.61338162e-01]
|
[9.283846855163574, 10.14884090423584]
|
2373f572-6e0a-4a9c-91f2-03bd120aea19
|
segmental-recurrent-neural-networks-for-end
|
1603.00223
| null |
http://arxiv.org/abs/1603.00223v2
|
http://arxiv.org/pdf/1603.00223v2.pdf
|
Segmental Recurrent Neural Networks for End-to-end Speech Recognition
|
We study the segmental recurrent neural network for end-to-end acoustic
modelling. This model connects the segmental conditional random field (CRF)
with a recurrent neural network (RNN) used for feature extraction. Compared to
most previous CRF-based acoustic models, it does not rely on an external system
to provide features or segmentation boundaries. Instead, this model
marginalises out all the possible segmentations, and features are extracted
from the RNN trained together with the segmental CRF. In essence, this model is
self-contained and can be trained end-to-end. In this paper, we discuss
practical training and decoding issues as well as the method to speed up the
training in the context of speech recognition. We performed experiments on the
TIMIT dataset. We achieved 17.3 phone error rate (PER) from the first-pass
decoding --- the best reported result using CRFs, despite the fact that we only
used a zeroth-order CRF and without using any language model.
|
['Steve Renals', 'Noah A. Smith', 'Lingpeng Kong', 'Liang Lu', 'Chris Dyer']
|
2016-03-01
| null | null | null | null |
['acoustic-modelling']
|
['speech']
|
[ 3.07979822e-01 4.25127983e-01 2.07591400e-01 -6.81930721e-01
-1.17945051e+00 -4.75820154e-01 3.55558366e-01 -1.18999422e-01
-6.99529409e-01 4.74486738e-01 2.77913630e-01 -7.45991588e-01
4.60324019e-01 -4.06764746e-01 -7.29953468e-01 -5.89876950e-01
6.26057833e-02 5.45817137e-01 2.16643900e-01 3.31150517e-02
-3.66346389e-02 1.12858392e-01 -1.27900302e+00 2.31910329e-02
5.95826864e-01 8.29100013e-01 6.83043659e-01 1.11715424e+00
-1.41171530e-01 6.90524161e-01 -6.85386658e-01 -8.79075229e-02
-3.42049599e-01 -5.36846876e-01 -8.50525796e-01 -1.15702666e-01
1.04878113e-01 -9.22340974e-02 -4.91797067e-02 6.61451817e-01
3.85890871e-01 2.95759767e-01 5.38030863e-01 -3.96670073e-01
-2.66005397e-01 9.07707214e-01 1.15434915e-01 -1.30541921e-01
3.04018378e-01 -3.46499085e-01 8.41654897e-01 -9.79194343e-01
4.23366189e-01 1.12266278e+00 7.17205822e-01 7.01192617e-01
-1.13974488e+00 -2.73266613e-01 -1.42757334e-02 -2.23198757e-01
-1.37842131e+00 -8.60312581e-01 4.41261113e-01 -3.68541032e-01
1.47381449e+00 1.61109164e-01 3.04256827e-01 9.86754000e-01
6.54941052e-02 7.41026759e-01 9.30025280e-01 -9.82928514e-01
3.23531866e-01 -7.98567235e-02 2.50745058e-01 3.45919788e-01
-6.77921772e-01 3.16376954e-01 -4.06515270e-01 7.97488913e-02
6.51097536e-01 -5.29893100e-01 1.22806206e-01 3.36976558e-01
-9.88872051e-01 7.77997434e-01 2.84679793e-02 5.01645565e-01
-4.35714036e-01 2.72758126e-01 4.28968340e-01 8.32246691e-02
5.06483793e-01 -7.82429576e-02 -7.97358453e-01 -4.91244584e-01
-1.44580448e+00 -3.61527026e-01 8.80959392e-01 7.54539490e-01
6.82907581e-01 3.54664236e-01 9.18862820e-02 1.30702412e+00
7.82267213e-01 5.37399530e-01 5.12767434e-01 -8.85172904e-01
3.87529284e-01 -2.17706025e-01 -3.55981439e-01 -2.53441066e-01
-3.25419545e-01 -5.01064777e-01 -6.64822161e-01 -8.42863098e-02
2.62522995e-01 -4.71834898e-01 -1.39041293e+00 1.69122756e+00
-6.64785579e-02 3.51806343e-01 1.69319689e-01 5.67170799e-01
8.25058281e-01 8.05084229e-01 3.21552724e-01 -3.31166536e-01
1.09689236e+00 -1.06167638e+00 -8.93858314e-01 -2.97627717e-01
8.51934433e-01 -1.13233602e+00 7.32063293e-01 4.43536371e-01
-1.16734648e+00 -8.80444586e-01 -8.71383131e-01 -2.87089460e-02
-3.20668578e-01 3.23532820e-01 1.93259165e-01 8.86658430e-01
-1.50760555e+00 4.95336980e-01 -1.13272178e+00 -2.48136431e-01
-3.38095427e-01 6.23234034e-01 -1.67437598e-01 2.81274289e-01
-1.21073520e+00 8.49558234e-01 3.71239781e-01 4.92793500e-01
-8.12160313e-01 -3.50794420e-02 -9.08506215e-01 -1.04769371e-01
1.60122991e-01 -3.22609812e-01 1.72860754e+00 -7.31763899e-01
-2.18793678e+00 4.98438954e-01 -7.57397294e-01 -5.36900997e-01
-3.16539258e-02 -4.73266691e-01 -4.86910105e-01 -1.87919065e-01
-3.69471133e-01 8.14839184e-01 5.45372367e-01 -9.82564330e-01
-4.32970732e-01 8.63959417e-02 -5.26959777e-01 1.13520600e-01
2.86374420e-01 4.36616123e-01 -6.34198368e-01 -5.58751881e-01
2.26164281e-01 -1.01858306e+00 -6.33014441e-01 -1.07011247e+00
-6.89699233e-01 -2.44866282e-01 5.16376674e-01 -1.05068052e+00
1.24112844e+00 -2.16828823e+00 -2.34490633e-02 3.02426606e-01
-4.90183294e-01 4.41716075e-01 -5.09998500e-02 3.97577167e-01
-8.11583325e-02 2.71643460e-01 -4.97093052e-01 -9.93602812e-01
-7.65459612e-02 5.10130167e-01 -4.48322177e-01 1.15505412e-01
1.98527992e-01 8.20580900e-01 -4.83165979e-01 -4.86575484e-01
5.23034096e-01 8.87356877e-01 -3.24598700e-01 4.92209196e-01
-2.34209076e-01 6.69488549e-01 -5.13788238e-02 1.74744830e-01
5.43372929e-01 4.64107782e-01 2.77802467e-01 3.48044515e-01
-3.11315417e-01 1.12967110e+00 -1.00995684e+00 1.87523496e+00
-8.84996653e-01 4.69623029e-01 1.49042994e-01 -1.02958679e+00
1.14602375e+00 7.47429669e-01 2.99612861e-02 -5.33454001e-01
4.96513173e-02 4.36645150e-01 -2.19825953e-01 -5.37820421e-02
5.60412884e-01 -1.95717856e-01 -2.76663333e-01 2.04348475e-01
4.79877025e-01 -1.79734588e-01 -1.37632117e-01 -5.27263843e-02
7.94722915e-01 4.24117684e-01 9.86537486e-02 -3.66973057e-02
4.73224074e-01 -4.25047159e-01 5.39885998e-01 6.28975928e-01
1.48095652e-01 8.67810607e-01 1.11893967e-01 -2.82251369e-03
-8.23174000e-01 -1.05102229e+00 -9.76823419e-02 1.31888509e+00
-5.22761524e-01 -5.03960669e-01 -1.00695455e+00 -4.46367800e-01
-7.66195059e-01 1.02275026e+00 -1.42730519e-01 3.99814993e-01
-8.01259279e-01 -3.75064015e-01 7.90229976e-01 5.07088304e-01
-7.04120472e-02 -1.47366703e+00 -1.64869219e-01 7.17336118e-01
-3.71648967e-01 -1.29395783e+00 -3.30997556e-01 8.14803481e-01
-9.19145405e-01 -2.21092150e-01 -5.51223576e-01 -1.13782227e+00
3.27152014e-01 -5.48991203e-01 1.26095939e+00 -1.00851022e-01
3.15142721e-01 1.08099736e-01 -4.49746639e-01 -2.01166809e-01
-1.01746213e+00 3.37771147e-01 -1.70039102e-01 -2.78305948e-01
3.44513297e-01 -6.16873205e-01 -4.50064912e-02 3.02068114e-01
-5.91155767e-01 -8.30938593e-02 5.86748183e-01 6.71080947e-01
6.74402058e-01 -2.32231155e-01 6.99988604e-01 -8.49031389e-01
2.96813667e-01 -6.30886406e-02 -4.13852096e-01 1.28546432e-01
-2.48698011e-01 -1.16820857e-02 6.34965420e-01 -1.26019549e-02
-1.15621018e+00 6.15776539e-01 -1.19687688e+00 -1.03954412e-01
-8.98128331e-01 6.10748768e-01 -1.74264997e-01 5.46360970e-01
3.20300102e-01 3.27841729e-01 -2.82287985e-01 -9.08943176e-01
6.37328744e-01 1.14731228e+00 7.07847059e-01 -3.41027409e-01
2.61243463e-01 -2.02027932e-01 -4.15398479e-01 -1.14403117e+00
-6.07039571e-01 -6.39036953e-01 -1.05131662e+00 -4.09313589e-02
1.06431234e+00 -9.06561434e-01 -4.79906201e-01 5.68441629e-01
-1.51210177e+00 -5.57862937e-01 -2.67195523e-01 8.63424838e-01
-7.73341417e-01 4.33169961e-01 -9.73568559e-01 -1.29030490e+00
-3.34338039e-01 -1.00573719e+00 9.58354652e-01 1.98136494e-02
-2.09474564e-01 -1.04821622e+00 3.09733510e-01 1.81872636e-01
5.44005573e-01 -2.19630659e-01 5.94266891e-01 -7.61119664e-01
-1.40372097e-01 1.33725554e-01 2.05055460e-01 8.88288438e-01
-1.40299365e-01 1.06768511e-01 -1.47490191e+00 4.44763713e-02
1.74926966e-01 -3.35739374e-01 1.00476038e+00 6.11797094e-01
8.70796740e-01 -1.16799600e-01 -1.52879670e-01 3.00248682e-01
1.23944640e+00 3.45439732e-01 8.39642167e-01 -2.45842472e-01
7.13847578e-01 5.89183211e-01 4.26634133e-01 -7.25928135e-03
3.00430268e-01 6.47705376e-01 2.22906679e-01 -2.86263406e-01
-2.90320724e-01 -3.72080773e-01 7.80524194e-01 1.74396694e+00
1.46428108e-01 -3.02177757e-01 -9.53525782e-01 6.94946468e-01
-1.56586492e+00 -4.55116510e-01 -3.45443189e-01 2.24485230e+00
8.84875655e-01 2.80454606e-01 8.64519849e-02 6.75442517e-02
7.98373401e-01 6.21895911e-03 1.45814151e-01 -9.98373866e-01
6.88927174e-02 6.25522614e-01 3.29456776e-01 1.08243036e+00
-1.07011032e+00 1.32888854e+00 7.09751463e+00 1.07385457e+00
-1.11019421e+00 2.57735550e-01 5.73860407e-01 3.35362822e-01
-9.00069922e-02 3.36411357e-01 -1.10252762e+00 3.04360509e-01
1.97974491e+00 9.32761610e-01 3.59240264e-01 6.64891183e-01
2.43250579e-01 -2.28545755e-01 -8.66167307e-01 4.11918491e-01
-2.81706244e-01 -8.34976196e-01 -4.14172202e-01 9.23733488e-02
2.82612562e-01 6.26057029e-01 -1.79253757e-01 4.84832615e-01
4.36312586e-01 -1.27250671e+00 8.13683689e-01 5.67432046e-01
1.01301146e+00 -8.89731467e-01 9.27001178e-01 5.40604115e-01
-1.24470389e+00 4.95116889e-01 -2.94969440e-01 -4.79586013e-02
7.13922441e-01 8.02589059e-01 -1.21716201e+00 5.78940094e-01
2.77202725e-01 3.33182454e-01 -1.94285721e-01 8.40406179e-01
-4.38734740e-01 1.49251652e+00 -7.17072666e-01 9.99979973e-02
3.15758586e-01 -1.15119152e-01 2.86679834e-01 1.91401064e+00
2.44102493e-01 -3.78030092e-01 2.26796135e-01 3.85330290e-01
1.60307929e-01 2.52435714e-01 -2.80375332e-01 -2.16834005e-02
2.91555941e-01 9.51342702e-01 -6.68206632e-01 -2.41496995e-01
-3.64277661e-01 8.84270430e-01 4.14850414e-01 3.68002504e-01
-6.86500728e-01 -4.72119778e-01 1.00674719e-01 -2.60746151e-01
8.01204205e-01 -7.70842791e-01 -2.96554655e-01 -7.75388956e-01
-1.83043629e-01 -4.94163752e-01 -6.35002404e-02 -4.94302183e-01
-1.08014822e+00 1.04665375e+00 -2.37529978e-01 -6.05253398e-01
-9.04295266e-01 -4.09398884e-01 -5.81349969e-01 1.24827909e+00
-1.44266176e+00 -1.12288606e+00 6.45784080e-01 3.01666915e-01
6.10247433e-01 9.49888155e-02 1.43781996e+00 2.62301415e-01
-3.00297171e-01 5.07889152e-01 2.68915981e-01 3.20423216e-01
4.49034870e-01 -1.23531103e+00 8.55061769e-01 8.77015054e-01
4.82834697e-01 7.39285707e-01 5.79824090e-01 -5.87619305e-01
-6.23669624e-01 -9.25645769e-01 1.63816941e+00 -3.19755554e-01
4.73039925e-01 -7.64691710e-01 -9.50633824e-01 7.27883995e-01
4.59062576e-01 -1.72418743e-01 8.98947001e-01 5.65073788e-01
-7.76218623e-02 3.26769471e-01 -5.48644304e-01 1.34968191e-01
7.02763498e-01 -8.15290689e-01 -6.86869681e-01 -2.89556477e-02
1.02815044e+00 -3.70833486e-01 -8.12004030e-01 3.81496042e-01
4.51954335e-01 -8.59755278e-01 5.84342003e-01 -6.87908232e-02
2.41529290e-03 -2.70307243e-01 -5.58773518e-01 -1.24035990e+00
1.83260813e-02 -7.63951898e-01 3.66204977e-02 1.57064545e+00
9.32849526e-01 -3.89078170e-01 6.58992171e-01 2.83485860e-01
-5.97300768e-01 -5.16746223e-01 -1.26261103e+00 -7.13942707e-01
1.74970031e-01 -1.18645263e+00 2.44401276e-01 4.61160064e-01
-1.73016548e-01 6.13266587e-01 -4.67281848e-01 1.55039772e-01
2.01637551e-01 -3.10427964e-01 4.23529863e-01 -9.01262522e-01
-6.26390100e-01 -6.45248741e-02 -9.76777375e-02 -1.64101863e+00
3.19468766e-01 -7.83677697e-01 8.37629378e-01 -1.54255879e+00
-3.77899945e-01 -7.35901952e-01 -3.23028564e-01 6.24063134e-01
1.17449872e-01 1.30767196e-01 1.12298511e-01 7.37667009e-02
-5.35748601e-01 4.82996017e-01 6.83644414e-01 4.08916056e-01
-4.52465892e-01 3.25726330e-01 -2.34042019e-01 7.45245814e-01
1.00165379e+00 -7.82853544e-01 -1.16292067e-01 -3.63603055e-01
-9.63070765e-02 3.37454975e-01 6.84305727e-02 -9.93306577e-01
3.90271306e-01 2.99140960e-01 8.03778246e-02 -1.06686330e+00
5.36678076e-01 -5.10376573e-01 -7.25415815e-03 8.46116096e-02
-2.90578455e-01 -2.63353914e-01 2.18988836e-01 3.74942392e-01
-5.52316666e-01 -5.44403255e-01 6.04372442e-01 6.08074665e-02
-3.73736113e-01 -3.63814861e-01 -1.01320016e+00 -2.63582766e-01
3.26405019e-01 6.12213984e-02 1.72704428e-01 -4.27740246e-01
-1.24691272e+00 -2.42582813e-01 9.62941721e-02 1.23184316e-01
4.52114195e-01 -8.11918795e-01 -7.29577661e-01 4.58708346e-01
-4.53056067e-01 2.25794137e-01 1.10270835e-01 6.04204237e-01
-2.70336032e-01 7.05877602e-01 3.41005027e-01 -7.85943925e-01
-1.05862939e+00 2.71215558e-01 3.51263434e-01 -5.16879022e-01
-2.54646599e-01 9.22419608e-01 -5.58258907e-04 -1.02732241e+00
4.33792263e-01 -4.06505316e-01 -2.94885933e-01 -2.26474002e-01
1.04128525e-01 2.77209133e-02 3.91271234e-01 -9.92556810e-01
-4.84407037e-01 5.43353677e-01 3.03843934e-02 -7.39483178e-01
1.21729124e+00 -3.57498974e-01 -1.11934528e-01 8.47022235e-01
1.25269270e+00 2.50892252e-01 -1.01963556e+00 -1.88259423e-01
2.28652120e-01 2.42172360e-01 2.21955642e-01 -1.00691962e+00
-6.44440830e-01 1.18733120e+00 4.07435507e-01 2.75223494e-01
9.48049188e-01 3.15805003e-02 9.78879094e-01 2.28986964e-01
1.12428263e-01 -1.28947079e+00 -5.98413825e-01 1.25015640e+00
5.70796609e-01 -8.24239790e-01 -5.25832415e-01 -4.47218716e-01
-6.15475297e-01 1.09300876e+00 7.33906105e-02 -2.41224587e-01
1.08153439e+00 6.07413292e-01 4.51349854e-01 2.84220695e-01
-7.24272251e-01 -2.95248598e-01 3.00514728e-01 6.14392698e-01
8.86132181e-01 3.35791945e-01 -9.10368413e-02 5.61304867e-01
-6.37229443e-01 -1.49550244e-01 3.06723237e-01 7.53576338e-01
-5.30244231e-01 -1.62503803e+00 -1.91817075e-01 1.25758901e-01
-8.38199794e-01 -3.97013783e-01 -2.35154495e-01 1.80513769e-01
9.91511419e-02 1.55226064e+00 -2.87680216e-02 -5.56025147e-01
1.60512999e-01 5.25713146e-01 9.38306823e-02 -8.82538199e-01
-8.49336982e-01 7.92358100e-01 4.13803101e-01 -3.51837546e-01
-5.09843469e-01 -6.96891308e-01 -1.48063779e+00 1.68633759e-01
-8.01767409e-01 4.21266317e-01 1.29081583e+00 1.20078671e+00
1.81131929e-01 6.09400570e-01 6.23642862e-01 -8.49509180e-01
-6.68995455e-02 -1.31376529e+00 -5.58349788e-01 -3.24659735e-01
5.01824021e-01 -4.22171839e-02 -3.67911696e-01 3.09799135e-01]
|
[14.471124649047852, 6.7585344314575195]
|
1ef2f75d-801a-4fe1-8fbd-127553dfa0ea
|
child-care-provider-survival-analysis
|
2208.02154
| null |
https://arxiv.org/abs/2208.02154v1
|
https://arxiv.org/pdf/2208.02154v1.pdf
|
Child Care Provider Survival Analysis
|
The aggregate ability of child care providers to meet local demand for child care is linked to employment rates in many sectors of the economy. Amid growing concern regarding child care provider sustainability due to the COVID-19 pandemic, state and local governments have received large amounts of new funding to better support provider stability. In response to this new funding aimed at bolstering the child care market in Florida, this study was devised as an exploratory investigation into features of child care providers that lead to business longevity. In this study we used optimal survival trees, a machine learning technique designed to better understand which providers are expected to remain operational for longer periods of time, supporting stabilization of the child care market. This tree-based survival analysis detects and describes complex interactions between provider characteristics that lead to differences in expected business survival rates. Results show that small providers who are religiously affiliated, and all providers who are serving children in Florida's universal Prekindergarten program and/or children using child care subsidy, are likely to have the longest expected survival rates.
|
['Courtney K. Blackwell', 'Maya Schreiber', 'Robert Chapman', 'Herman T. Knopf', 'Phillip Sherlock']
|
2022-08-03
| null | null | null | null |
['survival-analysis']
|
['miscellaneous']
|
[-2.66666621e-01 3.14804882e-01 -1.04157066e+00 -2.84111053e-01
-4.02884543e-01 -2.24755958e-01 -5.59202991e-02 8.70811343e-01
-1.84766188e-01 4.72755224e-01 6.71251535e-01 -8.84005904e-01
-2.21630290e-01 -6.94681406e-01 -1.85275495e-01 -6.17580056e-01
-3.44613552e-01 4.49198455e-01 -6.38095856e-01 -2.01896176e-01
1.14135452e-01 7.04710126e-01 -1.05532897e+00 -1.19653232e-01
9.86466765e-01 4.74014491e-01 -1.86000496e-01 2.97559023e-01
2.11466074e-01 8.43312263e-01 4.45395596e-02 -3.64780128e-01
-2.57123616e-02 -4.94225442e-01 -7.65810430e-01 -2.73345560e-01
-1.65968746e-01 -6.82456613e-01 1.04631901e-01 3.22292328e-01
4.26095217e-01 -1.74683556e-01 8.73050988e-01 -1.60307360e+00
-4.51714426e-01 1.10069907e+00 -4.00483251e-01 3.81491929e-01
5.02691150e-01 -5.96606061e-02 8.38238239e-01 -4.61360276e-01
5.91100991e-01 1.08428144e+00 7.65154719e-01 5.19061923e-01
-1.76652336e+00 -8.68141055e-01 -2.38647580e-01 -2.03069896e-01
-1.08905351e+00 -9.38556790e-01 4.35908169e-01 -7.06893563e-01
9.57225442e-01 -4.05133031e-02 9.60811496e-01 7.74213135e-01
5.20237386e-01 -1.60088345e-01 6.17269576e-01 -4.03839290e-01
1.06821761e-01 1.72725141e-01 -3.52465510e-01 5.03618538e-01
6.20091915e-01 -3.17159109e-02 -2.03085020e-01 -7.47230947e-01
6.16035342e-01 2.10435405e-01 -2.59276271e-01 -2.58424640e-01
-7.37807453e-01 1.23417056e+00 3.72168630e-01 5.49447715e-01
-5.73564708e-01 9.44563597e-02 1.04155272e-01 6.11586124e-02
2.48711914e-01 3.44648242e-01 -5.80937862e-01 -3.93108577e-01
-7.46643364e-01 1.32747486e-01 7.24305630e-01 3.54268909e-01
3.93567353e-01 5.44938780e-02 2.09225237e-01 5.73936582e-01
2.59003013e-01 4.15975600e-01 5.15446365e-02 -1.09763718e+00
4.02306408e-01 5.00858963e-01 9.83312055e-02 -1.05902493e+00
-7.52571762e-01 -5.79467356e-01 -5.23875237e-01 -1.06153719e-01
4.38657373e-01 -2.77286142e-01 -4.18798327e-01 1.83695734e+00
2.60578722e-01 -2.41576254e-01 1.99719191e-01 9.54591408e-02
2.49092460e-01 4.40841198e-01 3.39170337e-01 -8.39863777e-01
9.62541401e-01 -4.69229043e-01 -4.30669159e-01 6.11804388e-02
1.05909514e+00 -4.37343925e-01 4.60119396e-01 -1.37397066e-01
-1.29564619e+00 9.46468338e-02 -2.96558261e-01 5.70784926e-01
1.29215375e-01 -7.13141382e-01 8.48831296e-01 9.20182765e-01
-1.05885434e+00 7.26615608e-01 -1.19060230e+00 -7.86824286e-01
7.29349732e-01 1.38168961e-01 -5.62431872e-01 -2.56942529e-02
-5.08494616e-01 1.03383827e+00 -8.73548239e-02 -2.50430048e-01
-6.89611018e-01 -7.01078594e-01 -9.24673676e-01 6.20601177e-01
-5.18808551e-02 -7.18757927e-01 1.04488146e+00 -9.49151218e-01
-3.03106487e-01 5.34756422e-01 -2.26919681e-01 -3.42603356e-01
-1.81996986e-01 4.82977420e-01 -3.83690298e-01 3.04184169e-01
6.00909054e-01 5.34336746e-01 2.65193045e-01 -9.33386147e-01
-7.40385592e-01 -8.52521062e-01 -3.50667834e-01 1.56866610e-02
-5.29381037e-01 4.37524974e-01 6.83461964e-01 -5.14870465e-01
5.06700099e-01 -8.03647518e-01 -4.26258504e-01 -8.17096233e-01
2.46232271e-01 -2.27140576e-01 1.12127207e-01 -5.98314345e-01
1.25604677e+00 -1.89732659e+00 -2.52319664e-01 1.64884135e-01
8.67573395e-02 -6.50497854e-01 1.72353849e-01 9.46489573e-01
-2.09946260e-01 5.70119739e-01 1.68090314e-01 -3.02363634e-02
-3.22741240e-01 2.84263819e-01 2.17393667e-01 5.75190365e-01
1.47892967e-01 3.72876614e-01 -8.14336717e-01 -6.08277380e-01
-2.95584470e-01 1.08977981e-01 -7.05115318e-01 6.64700270e-02
5.23151815e-01 5.90746701e-01 -3.76750231e-01 1.33074403e+00
6.09625936e-01 -2.84735441e-01 3.33161682e-01 7.71282434e-01
-2.58382678e-01 2.67079145e-01 -3.43083233e-01 1.18533826e+00
4.06873450e-02 9.29774269e-02 4.66926455e-01 -9.25554395e-01
5.87118030e-01 5.16852856e-01 9.63351846e-01 -4.27817225e-01
2.37700999e-01 4.61675972e-01 2.62466431e-01 -6.77789927e-01
2.32937500e-01 -6.81980848e-01 1.93761230e-01 2.65742928e-01
-3.99731845e-01 -4.99147773e-02 -1.61822051e-01 3.73645216e-01
1.10296571e+00 -2.92373270e-01 3.21845710e-01 -6.55165017e-01
-1.28280878e-01 2.91442454e-01 1.15192020e+00 1.80734098e-01
-5.85244596e-01 -3.17235431e-03 6.44095600e-01 -2.88120300e-01
-8.40105176e-01 -9.19523120e-01 -5.42517841e-01 9.66292500e-01
-2.55769342e-01 1.52070403e-01 -3.91582131e-01 -1.86605528e-01
2.28734627e-01 9.67161357e-01 -4.14103299e-01 -9.09157619e-02
-2.73016274e-01 -4.94803458e-01 4.00011182e-01 4.72506493e-01
1.40669137e-01 -7.07719207e-01 -1.16830528e+00 4.35454130e-01
-2.77152687e-01 -5.49491107e-01 -4.41842824e-01 1.05679914e-01
-1.32835937e+00 -9.90691841e-01 -7.29547560e-01 -7.11886585e-01
6.02756083e-01 1.09086670e-01 9.15071070e-01 3.40151221e-01
-1.92995310e-01 3.83744359e-01 -4.19919848e-01 -3.82884741e-01
-6.44819438e-01 1.65510967e-01 4.77656454e-01 -5.90539515e-01
2.11022899e-01 -8.24869275e-01 -6.38143778e-01 -5.39944395e-02
-5.75637817e-01 -8.26588869e-02 1.53930292e-01 7.22192645e-01
-1.37023762e-01 2.44817480e-01 1.30372596e+00 -5.48886955e-01
2.21612960e-01 -1.09029162e+00 -2.07401171e-01 1.25061020e-01
-1.61807513e+00 -4.55821693e-01 -1.06211580e-01 -2.71967769e-01
-9.78579819e-01 -2.56764323e-01 -1.45613551e-01 1.78846166e-01
-4.47252914e-02 8.41972947e-01 3.98486793e-01 4.50256139e-01
5.94236672e-01 -6.09304607e-01 2.50402302e-01 -2.91939855e-01
-4.56068605e-01 6.77896321e-01 3.13440621e-01 -4.88940299e-01
5.60972750e-01 2.94128835e-01 1.58186495e-01 -3.76818180e-01
-1.10502206e-01 -2.52674282e-01 -1.72713213e-02 -3.02200019e-01
1.01154351e+00 -9.21264350e-01 -9.94989336e-01 8.99358690e-02
-5.66399872e-01 -3.10963959e-01 1.73473090e-01 8.92536938e-01
-3.05785120e-01 -1.62995860e-01 -5.71885467e-01 -1.45458698e+00
-3.36601824e-01 -8.81280363e-01 2.42247701e-01 5.23018181e-01
-2.46381760e-01 -8.84655476e-01 2.24027693e-01 5.85603833e-01
5.56272984e-01 9.23590362e-01 1.48119974e+00 -3.89972389e-01
-8.11031461e-02 1.06389619e-01 2.79056519e-01 -1.19474098e-01
3.61138135e-01 2.71746576e-01 -4.09099758e-01 -7.68693984e-01
-2.45015379e-02 -2.77279168e-02 2.77233928e-01 9.15129781e-01
3.29701871e-01 -6.96975768e-01 -6.52168751e-01 4.39196169e-01
1.53393221e+00 7.50771403e-01 2.35895410e-01 3.71883959e-01
-9.68846753e-02 1.27155352e+00 4.90157396e-01 7.49436140e-01
7.97211826e-01 1.84382096e-01 6.58697963e-01 -9.66788381e-02
7.11160362e-01 -5.25638640e-01 2.30486333e-01 3.81176651e-01
-1.67537421e-01 4.15547490e-02 -1.41066825e+00 1.03205788e+00
-1.50466502e+00 -1.11072838e+00 -1.55678838e-01 2.15880704e+00
4.41152602e-01 2.98256993e-01 5.66968262e-01 -1.45381883e-01
4.86325324e-01 -5.04914284e-01 -4.67985123e-01 -7.85662770e-01
1.43089548e-01 1.44862104e-02 6.84766233e-01 1.47540897e-01
-1.16241440e-01 3.29913944e-01 7.24520016e+00 2.61593491e-01
-8.07095051e-01 -2.73547471e-02 1.36322308e+00 -1.72311261e-01
-7.14582205e-01 6.22508347e-01 -2.53355563e-01 1.88788682e-01
1.13205457e+00 -4.13907021e-01 3.68395686e-01 9.31841075e-01
7.25047827e-01 -2.80842245e-01 -9.54604805e-01 3.45876515e-01
-1.59495413e-01 -1.36984444e+00 -7.42189109e-01 5.14755726e-01
8.52240205e-01 -1.81852892e-01 1.72191560e-01 5.56509309e-02
6.61792085e-02 -1.29401064e+00 6.97056174e-01 2.63555646e-01
8.36202323e-01 -9.87573028e-01 6.86873555e-01 6.47778511e-01
-8.70729566e-01 -8.86755586e-01 2.07700491e-01 -5.66073000e-01
3.81315917e-01 3.37663978e-01 -1.15272021e+00 -9.53185186e-03
1.32365417e+00 6.41507953e-02 -5.16927913e-02 4.45128530e-01
3.91690522e-01 8.40479970e-01 7.87496939e-02 4.86994594e-01
1.24426834e-01 -1.42995849e-01 8.63323063e-02 4.81502861e-01
6.18953764e-01 6.31993175e-01 -2.31098101e-01 6.27065122e-01
1.52094424e-01 2.95402974e-01 -8.52392375e-01 -6.56870782e-01
7.18566775e-01 9.63248372e-01 -8.20943713e-01 -7.62099475e-02
-4.20272112e-01 1.05944216e-01 7.36753643e-02 1.80244729e-01
-5.26555181e-01 -7.98986703e-02 8.91044319e-01 7.27697968e-01
-2.08083168e-01 -4.92618941e-02 -3.77194613e-01 -8.20531547e-01
-6.41486287e-01 -8.77480865e-01 4.96466249e-01 -4.29019362e-01
-5.98649383e-01 1.01879761e-01 1.41808331e-01 -3.83454651e-01
-3.61827791e-01 3.02723646e-01 -6.83175147e-01 9.11698401e-01
-1.23769224e+00 -1.04236329e+00 3.72136384e-01 2.36999184e-01
2.03886837e-01 -1.57489534e-02 7.85993397e-01 -1.11046322e-01
-8.42822731e-01 5.82780540e-01 1.55748986e-02 -4.02319789e-01
4.06087697e-01 -6.81200564e-01 -4.00066167e-01 7.16447890e-01
-5.61062694e-01 9.72108066e-01 5.97125351e-01 -9.96279597e-01
-9.82277155e-01 -4.45162505e-01 1.50261617e+00 -2.38337860e-01
3.55903685e-01 -1.43887633e-02 -6.06194794e-01 9.06573474e-01
2.59991914e-01 -9.87460971e-01 9.27598655e-01 3.25845361e-01
3.30767259e-02 -2.31951550e-01 -1.71573782e+00 3.10954988e-01
9.08172607e-01 -6.94849566e-02 -3.37604463e-01 2.26116374e-01
8.77793968e-01 1.39186397e-01 -1.32704091e+00 6.37383580e-01
6.97786748e-01 -1.10717356e+00 5.21200955e-01 -6.61465943e-01
6.82638645e-01 5.76927304e-01 -2.01269791e-01 -8.72677088e-01
-6.64871156e-01 -7.97186792e-01 5.01819551e-01 1.42236698e+00
5.91899216e-01 -1.05566812e+00 8.44197094e-01 1.25224292e+00
2.90760063e-02 -1.07214606e+00 -1.16376281e+00 -4.16454166e-01
3.65096480e-01 1.84828639e-01 1.08047426e+00 1.30869341e+00
4.09611285e-01 6.15198463e-02 1.28404442e-02 2.96562135e-01
4.22041655e-01 1.85659125e-01 4.10482019e-01 -1.04414856e+00
-1.16773315e-01 -6.90985203e-01 -1.82777002e-01 4.15852033e-02
-2.04884812e-01 -5.50911427e-01 -4.90611672e-01 -1.86306787e+00
5.87244868e-01 -9.20268357e-01 -1.25774115e-01 9.11935389e-01
6.70214668e-02 -6.14460468e-01 -7.51951486e-02 2.09622100e-01
5.53175628e-01 1.20157972e-01 7.18860865e-01 1.77411094e-01
-5.74711680e-01 2.45999172e-01 -1.27312803e+00 4.11537588e-01
7.81936944e-01 -5.66270769e-01 -3.11329752e-01 -3.43779653e-01
3.52179319e-01 1.09110057e+00 -9.31013823e-02 -5.13438463e-01
-1.51986733e-01 -9.99237776e-01 1.46591902e-01 -4.87315416e-01
-6.91639632e-02 -1.01425874e+00 8.82319987e-01 1.21952152e+00
-1.99588567e-01 5.78875184e-01 -6.36859238e-02 -8.18176344e-02
2.84398317e-01 -2.37823635e-01 8.90716791e-01 2.78996348e-01
5.31808317e-01 2.33691365e-01 -7.20379233e-01 -2.80548096e-01
1.17562962e+00 -2.75699973e-01 -3.20192665e-01 -6.11102104e-01
-5.66721201e-01 4.82959002e-01 1.05979145e+00 -8.22894275e-02
3.77098083e-01 -1.23973286e+00 -8.58056068e-01 -1.13689348e-01
-1.52891040e-01 -3.74908626e-01 3.12636614e-01 1.26125133e+00
-7.55917907e-01 4.80124533e-01 -3.19451004e-01 -1.32667348e-01
-1.13396466e+00 8.02260518e-01 1.64255843e-01 -3.20956297e-02
-4.99184519e-01 9.10033345e-01 -2.56059170e-01 2.75717556e-01
9.86272246e-02 -2.49069825e-01 -2.56542057e-01 2.07158789e-01
8.09904784e-02 8.73827875e-01 -4.28450465e-01 -8.63877952e-01
-5.84174991e-01 2.00212430e-02 2.83471495e-01 -6.21109530e-02
1.51455140e+00 -2.66025752e-01 -4.50586557e-01 2.88705289e-01
1.15645730e+00 2.18928069e-01 -7.71894038e-01 5.11804998e-01
-9.56361890e-02 -7.38894165e-01 -9.92096961e-02 -4.55695271e-01
-1.09974766e+00 4.17062581e-01 5.21367967e-01 6.75211191e-01
1.51066041e+00 1.33019999e-01 6.57703161e-01 -4.97098684e-01
3.46751362e-01 -8.54403675e-01 -2.19940975e-01 -3.72212350e-01
5.65449536e-01 -1.23074663e+00 -1.69911474e-01 2.19816715e-02
-4.25294071e-01 7.27368176e-01 3.26649874e-01 6.11216843e-01
5.04591465e-01 -9.57434475e-02 -1.25163272e-01 -1.33854672e-01
-9.65688884e-01 2.67420322e-01 -6.65449619e-01 5.74408293e-01
4.27254409e-01 5.76033175e-01 -8.25433254e-01 4.35498446e-01
-1.32129654e-01 -1.82978466e-01 7.38309681e-01 8.38912845e-01
-7.47325420e-01 -1.06051683e+00 -4.71835792e-01 9.34250891e-01
-8.51546407e-01 -1.37831151e-01 2.54762582e-02 5.32135069e-01
5.14511883e-01 1.59971738e+00 9.22535807e-02 1.24067359e-01
-1.81357644e-03 1.84491917e-01 3.77812833e-02 -5.53257883e-01
-6.06823027e-01 3.91144186e-01 2.86163896e-01 -2.04052344e-01
-4.00162071e-01 -1.17368650e+00 -1.58304322e+00 -9.58565891e-01
-5.35213888e-01 5.43599486e-01 9.20529366e-01 5.02135873e-01
2.48126388e-01 -1.00220973e-02 1.11940432e+00 -2.25607753e-01
-5.03557682e-01 -7.99327552e-01 -9.85524714e-01 -9.52640772e-02
3.25748116e-01 -5.00679672e-01 -4.60732192e-01 -2.36618832e-01]
|
[8.022582054138184, 5.518388748168945]
|
873a8905-4af2-4a26-88b8-9d0480feefef
|
reading-like-her-human-reading-inspired
| null | null |
https://aclanthology.org/D19-1300
|
https://aclanthology.org/D19-1300.pdf
|
Reading Like HER: Human Reading Inspired Extractive Summarization
|
In this work, we re-examine the problem of extractive text summarization for long documents. We observe that the process of extracting summarization of human can be divided into two stages: 1) a rough reading stage to look for sketched information, and 2) a subsequent careful reading stage to select key sentences to form the summary. By simulating such a two-stage process, we propose a novel approach for extractive summarization. We formulate the problem as a contextual-bandit problem and solve it with policy gradient. We adopt a convolutional neural network to encode gist of paragraphs for rough reading, and a decision making policy with an adapted termination mechanism for careful reading. Experiments on the CNN and DailyMail datasets show that our proposed method can provide high-quality summaries with varied length, and significantly outperform the state-of-the-art extractive methods in terms of ROUGE metrics.
|
['Feiyang Pan', 'Yan Song', 'Xiang Ao', 'Min Yang', 'Qing He', 'Ling Luo']
|
2019-11-01
| null | null | null |
ijcnlp-2019-11
|
['extractive-document-summarization']
|
['natural-language-processing']
|
[ 5.42447567e-01 3.01095158e-01 -4.62358326e-01 -3.28531116e-01
-1.17185473e+00 -4.81667876e-01 6.37761354e-01 3.62778544e-01
-5.04481792e-01 7.81383991e-01 1.03216636e+00 -4.51931924e-01
-2.27167942e-02 -6.37419879e-01 -6.94432080e-01 -4.03111458e-01
3.03249806e-01 4.48265165e-01 -1.10983581e-03 -1.09788075e-01
9.28991616e-01 2.38792166e-01 -9.69360828e-01 6.13464952e-01
1.28917646e+00 8.45112205e-01 2.73016810e-01 1.13994133e+00
-5.25667548e-01 9.15030658e-01 -1.04274154e+00 -4.16227311e-01
8.12215440e-04 -6.16435230e-01 -1.27049160e+00 3.39984447e-01
5.14220595e-01 -8.25531483e-01 -2.28977978e-01 9.27484274e-01
6.97318256e-01 4.87715453e-01 7.30681002e-01 -5.34206033e-01
-8.59932363e-01 1.00410044e+00 -7.00013936e-01 4.86062348e-01
4.76303786e-01 8.68117064e-02 1.33034408e+00 -6.02644682e-01
2.83049583e-01 1.31474316e+00 3.32145095e-01 6.44351900e-01
-9.60630953e-01 -1.79497391e-01 5.53260565e-01 -5.34735247e-02
-5.48110366e-01 -6.64180219e-01 7.98806012e-01 -1.44868508e-01
1.03067374e+00 6.04451299e-01 8.25770259e-01 9.74698424e-01
3.10059696e-01 1.55278599e+00 5.04727900e-01 -3.92147034e-01
5.66777647e-01 -4.39150602e-01 7.10968614e-01 6.47496581e-01
4.05278057e-01 -6.43172383e-01 -6.01793826e-01 -1.21601626e-01
3.60558867e-01 1.18708745e-01 -3.98448825e-01 4.17647570e-01
-1.16485214e+00 8.99418414e-01 4.62189674e-01 -5.55148954e-03
-7.81097054e-01 2.15111956e-01 6.30094409e-01 -7.02419365e-03
9.53964412e-01 7.51795173e-01 -1.32644743e-01 -3.77433784e-02
-1.51206958e+00 4.61701900e-01 1.03194308e+00 1.03808844e+00
3.90510738e-01 -3.68987285e-02 -8.96280646e-01 9.08873141e-01
-1.12398034e-02 3.36206764e-01 6.89009547e-01 -7.89744377e-01
9.78885889e-01 4.21418548e-01 1.33299738e-01 -7.95982957e-01
-1.61918625e-01 -3.55493993e-01 -1.12920475e+00 -4.17297244e-01
-8.66645649e-02 -4.40145224e-01 -9.49947834e-01 1.00469768e+00
-9.23424810e-02 -5.66811673e-02 9.38471258e-02 8.20497274e-01
1.31956077e+00 1.07443416e+00 -1.73983917e-01 -3.95113319e-01
1.35493553e+00 -1.58369803e+00 -9.48167562e-01 -5.37443995e-01
1.82069838e-01 -4.07975137e-01 1.01351011e+00 3.80028874e-01
-1.58266509e+00 -2.83611506e-01 -1.27374649e+00 -5.08323014e-01
2.19877586e-02 4.23661411e-01 4.44737107e-01 1.03006154e-01
-1.04828823e+00 8.47677886e-01 -7.64405787e-01 -4.16039735e-01
5.76014936e-01 8.48864391e-02 1.93048000e-01 2.12426066e-01
-8.23026896e-01 5.19666314e-01 5.80108047e-01 2.55372733e-01
-5.91234386e-01 -4.36354965e-01 -7.34706998e-01 6.24906838e-01
4.97552931e-01 -1.21126509e+00 1.73239601e+00 -6.81603372e-01
-1.97483003e+00 5.51632583e-01 -4.85523254e-01 -6.51548684e-01
5.75060427e-01 -9.20160174e-01 1.86291769e-01 4.29787099e-01
1.81244984e-01 5.04732966e-01 9.15912390e-01 -1.02814507e+00
-8.71624231e-01 -1.50099143e-01 1.90645173e-01 5.19683242e-01
-2.63036698e-01 3.23060378e-02 -6.11733973e-01 -7.50502288e-01
-1.22549102e-01 -5.08596480e-01 -2.91348130e-01 -6.81304514e-01
-1.03655457e+00 -4.36751634e-01 3.88039380e-01 -1.13971651e+00
1.59033704e+00 -1.44838309e+00 3.20959449e-01 -2.36500084e-01
4.58825469e-01 3.95383388e-01 -1.34363964e-01 6.49731636e-01
2.65262395e-01 3.98638546e-01 -3.04454267e-01 -6.80748045e-01
7.02504069e-02 -2.55671740e-01 -7.77291417e-01 -6.77052885e-02
2.25236475e-01 1.15551925e+00 -1.18336570e+00 -6.77720487e-01
-2.72373587e-01 -1.76297724e-01 -4.08317894e-01 4.57355499e-01
-4.44860697e-01 -3.87061127e-02 -8.91715765e-01 3.60909432e-01
3.85592282e-01 -4.15930599e-01 -1.39073431e-01 1.25730053e-01
-1.48037463e-01 7.02217817e-01 -6.74059212e-01 1.72134364e+00
-2.76879430e-01 8.22838068e-01 -3.82412151e-02 -8.44638705e-01
7.53048062e-01 5.61588956e-03 3.36834192e-02 -4.02219504e-01
1.61257774e-01 -8.64172652e-02 -4.74433154e-01 -5.71958244e-01
1.37951219e+00 2.68081367e-01 -1.91975102e-01 8.99741888e-01
-9.02592987e-02 -2.99689710e-01 3.72659624e-01 5.77108800e-01
1.20218301e+00 -8.06186050e-02 3.53208572e-01 -2.34015808e-01
3.92538011e-01 -9.25166458e-02 4.29159760e-01 1.20198476e+00
1.42559305e-01 7.14663923e-01 8.68499279e-01 -3.75675529e-01
-9.98375773e-01 -7.33404040e-01 5.57627857e-01 1.12980890e+00
7.81179813e-04 -5.20148516e-01 -1.15581155e+00 -8.75814915e-01
-1.47745892e-01 9.81001198e-01 -4.05287683e-01 -2.78428365e-02
-8.52460265e-01 -5.22567749e-01 3.18719596e-01 6.40048504e-01
9.46416140e-01 -1.41853952e+00 -7.84011245e-01 1.82583153e-01
-3.86753976e-01 -7.66393661e-01 -9.57162797e-01 1.62093919e-02
-1.05668938e+00 -6.43569291e-01 -1.00198936e+00 -7.54989386e-01
6.21824980e-01 3.43683004e-01 1.15670228e+00 2.51532972e-01
3.00788432e-01 8.89926627e-02 -3.99297357e-01 -6.43192530e-01
-3.60366076e-01 7.77061522e-01 -4.24513429e-01 -6.09887652e-02
1.15445442e-01 -2.82598436e-01 -1.09040463e+00 -3.52438301e-01
-1.03193998e+00 5.43528438e-01 9.10125673e-01 8.33895504e-01
4.08076614e-01 -3.38741601e-01 7.81921208e-01 -1.04679799e+00
1.67622530e+00 -2.83061266e-01 -1.29718140e-01 5.07341385e-01
-4.96000767e-01 2.89716095e-01 8.44174087e-01 -2.92458028e-01
-1.31404603e+00 -3.27514559e-01 -4.90107276e-02 8.33093971e-02
9.64452326e-02 7.46794641e-01 1.00880466e-01 7.06829369e-01
5.39219201e-01 5.30385554e-01 -2.33559996e-01 -5.75424850e-01
4.72401112e-01 1.05022705e+00 6.81226254e-01 -5.95422864e-01
5.03840923e-01 3.33606005e-01 -5.47140241e-01 -9.48163688e-01
-1.36050606e+00 -5.26420414e-01 -6.24649167e-01 -6.31776825e-02
8.72345626e-01 -6.19008303e-01 -5.87880492e-01 4.62184936e-01
-1.49136472e+00 -4.38158423e-01 -3.57775927e-01 5.75456098e-02
-6.35786831e-01 6.48334265e-01 -1.00909126e+00 -6.39260888e-01
-1.41924548e+00 -8.18680644e-01 1.32148147e+00 7.04639792e-01
-3.74916255e-01 -7.35870540e-01 3.99284922e-02 4.90406454e-01
3.38103205e-01 -3.23104896e-02 8.24196637e-01 -9.04853880e-01
-6.82333291e-01 -2.39705727e-01 -3.43828827e-01 3.23907405e-01
1.37829166e-02 -3.58795747e-02 -6.92732692e-01 -3.33152145e-01
-6.59129173e-02 -2.79668331e-01 1.50003099e+00 6.03846610e-01
1.21866083e+00 -9.96515155e-01 -2.22951829e-01 6.07318103e-01
9.36448574e-01 -1.12722218e-02 5.55199564e-01 4.26051378e-01
6.50836825e-01 4.01638567e-01 6.13402367e-01 4.60960716e-01
3.53961051e-01 -2.62543559e-02 1.28529489e-01 -9.42945927e-02
8.86268914e-02 -4.44961786e-01 1.98649243e-01 1.15933752e+00
-6.13519363e-03 -6.94375575e-01 -6.58339858e-01 6.15930676e-01
-2.26693130e+00 -1.04592240e+00 2.97230959e-01 1.66564333e+00
9.10821617e-01 3.48974615e-01 1.16706155e-01 -1.11590981e-01
7.29978442e-01 8.44890714e-01 -7.09838867e-01 -8.20524633e-01
2.62167543e-01 -8.06872994e-02 3.34231734e-01 5.06117165e-01
-1.09079719e+00 1.18589091e+00 6.26574039e+00 8.31506610e-01
-1.07168257e+00 -5.03844917e-01 8.60017657e-01 -3.06852341e-01
-3.46249431e-01 -1.36532143e-01 -8.98665965e-01 5.49186051e-01
7.32671380e-01 -4.04871166e-01 3.94409597e-01 5.59512794e-01
3.69961411e-01 -2.82404959e-01 -1.00414634e+00 6.51675463e-01
3.55017364e-01 -1.71176815e+00 4.75912571e-01 -2.98640460e-01
8.52746665e-01 -1.64390117e-01 -2.70717233e-01 2.78384924e-01
4.57014352e-01 -9.31177318e-01 7.24753916e-01 8.00727785e-01
5.11833847e-01 -7.48067081e-01 5.93939245e-01 7.21854746e-01
-7.65563786e-01 -2.98868138e-02 -5.12499690e-01 8.89191218e-03
2.34860748e-01 4.98568565e-01 -7.19056845e-01 6.24251425e-01
4.71081406e-01 8.13712835e-01 -3.92942816e-01 1.07781148e+00
-5.84698498e-01 7.27003515e-01 5.92854992e-02 -6.45620823e-01
5.24059176e-01 -3.17656636e-01 7.17566669e-01 1.54968190e+00
2.26945922e-01 3.20149571e-01 2.06056200e-02 7.21822858e-01
-4.90805417e-01 1.18734770e-01 -9.10950154e-02 -3.36292610e-02
3.53849858e-01 1.22618926e+00 -8.14194739e-01 -7.81735778e-01
1.25250947e-02 1.26387239e+00 5.96486747e-01 6.42456770e-01
-4.61167395e-01 -8.64998460e-01 2.08956376e-02 -1.39277309e-01
4.83341306e-01 -3.74186113e-02 -5.89215219e-01 -1.38896668e+00
1.78860053e-01 -9.57734406e-01 3.25572759e-01 -8.60448897e-01
-8.60933065e-01 5.91125131e-01 -3.46973841e-03 -7.72678077e-01
-2.47784168e-01 -6.50592223e-02 -1.31742847e+00 7.42823660e-01
-1.56922460e+00 -9.22007859e-01 -2.74681330e-01 -8.09036419e-02
1.34161091e+00 -5.07848300e-02 4.49439913e-01 -4.52277958e-01
-8.60700309e-01 2.29493439e-01 4.06837970e-01 2.05767766e-01
4.91823256e-01 -1.63720500e+00 9.78384018e-01 1.00422919e+00
-2.16813266e-01 6.32164657e-01 7.45682240e-01 -7.77195930e-01
-1.14625371e+00 -9.96214926e-01 9.21821654e-01 -1.87342111e-02
3.56663555e-01 -2.17899069e-01 -8.14003050e-01 7.72307754e-01
8.60130250e-01 -8.62668037e-01 4.23580199e-01 1.54315844e-01
1.52509987e-01 -7.18595684e-02 -7.23506033e-01 9.15239453e-01
9.50162470e-01 -1.29565716e-01 -1.16310132e+00 4.08167154e-01
1.01829660e+00 -5.04510880e-01 -3.89459997e-01 -1.11159869e-02
5.09834707e-01 -8.07989180e-01 7.40102410e-01 -7.37454832e-01
1.14006901e+00 2.86691636e-01 4.90482867e-01 -1.77200115e+00
-4.21875507e-01 -1.14805722e+00 -4.35811400e-01 1.25311601e+00
3.37374389e-01 -4.14945155e-01 7.85531580e-01 5.19033432e-01
-4.04594511e-01 -1.16378200e+00 -3.72886419e-01 -2.76957422e-01
1.39699310e-01 1.94606125e-01 7.60794520e-01 2.84527063e-01
2.54968792e-01 8.96317899e-01 -3.49918246e-01 -2.91419446e-01
3.84969354e-01 5.49001276e-01 8.74153554e-01 -9.18760955e-01
-1.56472966e-01 -9.50093210e-01 5.04647255e-01 -2.03725100e+00
2.66755939e-01 -6.50534153e-01 4.41056848e-01 -2.31870747e+00
6.18847370e-01 2.73088247e-01 -1.27005145e-01 1.39487401e-01
-7.16270864e-01 -5.48585176e-01 2.00493902e-01 3.24732929e-01
-1.07855260e+00 7.12596655e-01 1.36015034e+00 -5.14625371e-01
-4.19285536e-01 2.92235166e-01 -1.27475250e+00 5.67409098e-01
8.10261071e-01 -2.13228136e-01 -3.06174636e-01 -6.65872157e-01
3.08357000e-01 4.52014953e-01 -1.30798399e-01 -6.96581066e-01
5.52102506e-01 -1.22412086e-01 3.55533808e-01 -1.33271205e+00
5.98318002e-04 4.24723588e-02 -7.94138551e-01 3.14816624e-01
-1.03535163e+00 1.68301612e-01 6.15354255e-02 6.65087223e-01
-1.34663552e-01 -5.23765981e-01 2.75898159e-01 -2.49579772e-01
-2.34165698e-01 2.28749707e-01 -5.78140616e-01 2.22995967e-01
4.63720381e-01 -4.78469208e-02 -6.48193955e-01 -7.52668440e-01
-1.93869367e-01 7.14954853e-01 2.27654934e-01 2.06713960e-01
7.42885888e-01 -9.14480209e-01 -1.12506962e+00 -2.21661195e-01
-3.60879123e-01 4.98981684e-01 1.11505866e-01 4.44772989e-01
-8.50850344e-01 6.42070949e-01 1.01847671e-01 -3.87467682e-01
-1.18589199e+00 4.57700133e-01 -1.79382004e-02 -8.00777912e-01
-9.05487061e-01 7.83853412e-01 1.49628237e-01 -1.20577672e-02
4.79724497e-01 -7.21237659e-01 -4.30021614e-01 3.44899744e-01
8.07369292e-01 6.22630775e-01 4.07507131e-03 1.57803185e-02
1.60505652e-01 2.89035678e-01 -5.98107457e-01 -9.99148786e-02
1.44858360e+00 -2.49144882e-01 -2.03240484e-01 2.91490406e-01
1.00330389e+00 8.43810588e-02 -1.24749100e+00 -2.81154871e-01
1.90387219e-01 -1.74069688e-01 2.13459097e-02 -7.60435820e-01
-6.43779457e-01 8.39586675e-01 -4.10001904e-01 6.60814285e-01
1.12356305e+00 -2.05224410e-01 1.39701200e+00 8.06058288e-01
-3.84654284e-01 -1.37441194e+00 2.44352072e-01 7.66021073e-01
1.27182412e+00 -1.10393715e+00 4.01878685e-01 2.92675421e-02
-7.63842821e-01 1.29561579e+00 3.93323153e-01 -4.21899229e-01
6.15326390e-02 -1.63135424e-01 -1.31135255e-01 -2.18231753e-01
-9.79455948e-01 1.28849998e-01 5.45235455e-01 -4.00691899e-03
4.01057363e-01 -5.85268997e-02 -5.29931426e-01 7.19121099e-01
-5.01616061e-01 -4.98307496e-02 6.31584823e-01 9.03325021e-01
-8.89070272e-01 -4.66861486e-01 -2.09710211e-01 9.06518459e-01
-7.37495601e-01 -2.74136335e-01 -7.40445793e-01 3.49288106e-01
-8.89847815e-01 1.00762784e+00 1.63507164e-01 -1.13857880e-01
3.36116821e-01 -1.17136696e-02 1.83704585e-01 -8.01966429e-01
-8.75768661e-01 1.20988324e-01 3.49586487e-01 -3.23769152e-01
-1.87097639e-02 -4.59212154e-01 -1.07737505e+00 -4.76487249e-01
-3.19641978e-01 3.48084092e-01 5.70546508e-01 9.42519784e-01
4.30628926e-01 8.96552265e-01 5.91340423e-01 -1.19739473e+00
-9.86230314e-01 -1.27962828e+00 -2.82904744e-01 1.89577118e-01
7.56169677e-01 3.10441077e-01 -2.45093048e-01 1.23924486e-01]
|
[12.519285202026367, 9.477912902832031]
|
3ca71281-b019-473d-8f48-1141a17e2de7
|
exploring-grammatical-error-correction-with
| null | null |
https://aclanthology.org/W12-2005
|
https://aclanthology.org/W12-2005.pdf
|
Exploring Grammatical Error Correction with Not-So-Crummy Machine Translation
| null |
['Nitin Madnani', 'Martin Chodorow', 'Joel Tetreault']
|
2012-06-01
| null | null | null |
ws-2012-6
|
['grammatical-error-detection']
|
['natural-language-processing']
|
[-8.63703638e-02 1.71006292e-01 -6.22772932e-01 -4.08054382e-01
-8.41685571e-03 -9.08429027e-01 6.55310392e-01 -6.53472245e-01
-2.85945535e-01 1.06888819e+00 -4.63127941e-02 -1.01159286e+00
-3.91567826e-01 -9.63214397e-01 -4.95059669e-01 -6.31337762e-01
-9.79754329e-01 7.25764990e-01 3.30370307e-01 -6.93831444e-01
7.03166842e-01 7.88774848e-01 -1.68942046e+00 7.18545914e-01
7.04417467e-01 8.52217197e-01 2.49141872e-01 1.14950800e+00
-1.95044339e-01 1.55633950e+00 -7.48382092e-01 -5.46825826e-01
3.13719302e-01 -1.23176083e-01 -7.22945035e-01 -1.01074085e-01
9.28529128e-02 -8.59008506e-02 -2.09758401e-01 9.22211111e-01
5.37373662e-01 4.49454933e-02 1.08379531e+00 -1.42548037e+00
-5.91619551e-01 6.10313773e-01 -4.01565880e-02 1.21627934e-01
1.03678203e+00 -5.39447069e-01 1.19919395e+00 -1.13026452e+00
7.20913768e-01 1.26888943e+00 8.66221786e-01 5.44149756e-01
-1.22286928e+00 -1.94712028e-01 -3.26822817e-01 -9.51717794e-02
-1.46558487e+00 -3.25250506e-01 4.25783843e-02 -2.08119690e-01
1.66093647e+00 1.26596653e+00 1.20609856e+00 1.01401424e+00
1.26658809e+00 8.34431887e-01 1.04267764e+00 -5.13792276e-01
3.35295945e-01 3.66983831e-01 1.54683650e-01 6.33519173e-01
8.40953708e-01 5.26628852e-01 -7.06372619e-01 -9.13127720e-01
9.33553874e-01 -2.94925272e-01 1.71355158e-01 -5.05680561e-01
-9.05919552e-01 6.91228509e-01 1.78732842e-01 3.83959889e-01
-1.39880210e-01 9.89067405e-02 1.26390755e-01 5.30987144e-01
-2.58292928e-02 6.47037446e-01 -9.11868811e-01 -1.33165747e-01
-8.71728659e-01 5.10332465e-01 1.25398111e+00 1.52653182e+00
1.24482810e-01 2.94908643e-01 -9.34252143e-02 3.17179203e-01
8.92314315e-01 1.01808000e+00 4.28362608e-01 -1.36146402e+00
-6.87414408e-02 1.72361732e-01 5.01781464e-01 -8.52631688e-01
-6.33224547e-01 -9.64177120e-03 -8.93263519e-01 4.49267089e-01
3.49161088e-01 4.57367361e-01 -8.02827001e-01 5.07305264e-01
4.33481112e-02 -2.34125629e-01 4.53833073e-01 5.55570945e-02
4.99930978e-01 3.76208365e-01 -1.34477139e-01 -5.73289394e-01
1.06082785e+00 -1.36716676e+00 -1.35299087e+00 2.33215362e-01
9.05734658e-01 -1.07320261e+00 4.35900748e-01 5.33875942e-01
-1.55548143e+00 -1.37560293e-01 -1.08699942e+00 1.78573877e-01
-7.27255583e-01 -3.14239264e-01 8.57801437e-01 1.43120694e+00
-1.60129595e+00 9.73287821e-01 -4.91727620e-01 6.59165755e-02
1.20568443e-02 8.24621081e-01 -2.64718989e-03 4.62812334e-01
-1.33193445e+00 1.08501506e+00 2.22979754e-01 -1.21242590e-01
-1.65216476e-01 -2.13068098e-01 -8.23704481e-01 -5.63443303e-01
-4.78693932e-01 -5.29636025e-01 1.44139910e+00 -2.59346128e-01
-1.65295815e+00 9.71794367e-01 -1.42069459e-01 -1.97814897e-01
6.14786744e-01 -1.28011424e-02 -8.31891418e-01 2.42498964e-01
-1.89849049e-01 5.76383233e-01 9.28263724e-01 -1.35132408e+00
-7.59897232e-01 -1.67359829e-01 -1.23336017e-01 2.66287565e-01
-1.25510961e-01 1.89734384e-01 2.11616129e-01 -1.12999000e-01
3.27147305e-01 -7.20919967e-01 -2.53068686e-01 -5.32041907e-01
-1.46512717e-01 -7.10518599e-01 7.70373225e-01 -4.51523662e-01
1.83705616e+00 -1.67618537e+00 -2.06720144e-01 3.98590982e-01
3.57815564e-01 -1.24705513e-03 2.40583986e-01 1.08380008e+00
-2.76906848e-01 7.32199550e-01 4.11965609e-01 -1.10722095e-01
2.24991128e-01 4.85861301e-01 -4.16602850e-01 3.05609167e-01
-1.29282743e-01 1.15307164e+00 -1.16605783e+00 -5.23096442e-01
4.11106765e-01 9.42391157e-02 -4.58732933e-01 4.79237735e-01
2.99364805e-01 1.47170946e-01 -3.56553018e-01 1.39399457e+00
1.15709066e+00 -1.31984919e-01 1.45911396e-01 5.30878425e-01
-3.79135728e-01 3.55090618e-01 -6.96863770e-01 1.05554795e+00
7.08333924e-02 5.00173986e-01 1.02364108e-01 -7.94621468e-01
3.33247900e-01 8.61540735e-01 4.37155962e-01 -9.67555881e-01
-2.26398129e-02 6.92409754e-01 1.12803578e-01 -6.07703328e-01
7.58228302e-01 3.94563079e-02 -3.64872098e-01 6.40070081e-01
-2.37588286e-01 -6.59476995e-01 9.64643434e-02 3.08100313e-01
6.22585893e-01 -5.18246442e-02 5.62923312e-01 -1.05613089e+00
7.86340594e-01 -1.82965681e-01 -1.81299388e-01 1.03415680e+00
-3.09923887e-01 3.19085121e-01 2.99841821e-01 -6.65102363e-01
-6.45341039e-01 -1.12307119e+00 -4.89381433e-01 1.30636716e+00
3.24267983e-01 -4.39044595e-01 -9.54439282e-01 -2.49762803e-01
1.77620783e-01 6.89606130e-01 -5.90509653e-01 3.84124845e-01
-5.03739953e-01 -8.32535863e-01 7.39044368e-01 3.45434904e-01
-5.07752821e-02 -1.33414865e+00 -6.58416986e-01 1.25490099e-01
-2.20292807e-01 -6.63697243e-01 -6.23428151e-02 4.48765576e-01
-1.35989368e+00 -5.18594682e-01 -6.66252747e-02 -8.20914626e-01
5.87345481e-01 2.46782884e-01 1.27047324e+00 5.39230824e-01
-2.31483161e-01 4.26904231e-01 -1.21292919e-01 -4.95818377e-01
-4.59671497e-01 -8.00336525e-02 5.28869390e-01 -5.87835789e-01
5.19427478e-01 -2.50617653e-01 -7.29350567e-01 5.37953973e-01
-6.88540697e-01 1.62748516e-01 1.79803044e-01 1.04410267e+00
1.35816500e-01 -9.34035778e-02 1.22507080e-01 -6.38007045e-01
8.72274399e-01 -1.69219792e-01 -3.78732830e-01 5.77745810e-02
-6.77108407e-01 -3.74140263e-01 3.21430594e-01 -3.25342178e-01
-1.01981449e+00 -4.87835288e-01 -9.82677937e-02 2.45538145e-01
1.11353043e-02 -1.46784872e-01 6.47139177e-02 -5.24923325e-01
8.02199244e-01 9.25758183e-02 1.99174434e-02 -6.80815242e-03
3.01039815e-01 7.09525108e-01 -6.82967342e-03 -6.68678164e-01
8.44880998e-01 4.91470337e-01 7.98524171e-02 -9.57177758e-01
-1.52186140e-01 -2.60129690e-01 -9.51962709e-01 -6.54426932e-01
6.56643391e-01 -6.78531289e-01 -9.10833478e-01 3.91110867e-01
-9.38691139e-01 -3.38627815e-01 -3.91645581e-01 4.25431967e-01
-1.01278400e+00 2.75717527e-02 -3.90154392e-01 -1.27895141e+00
-5.10977268e-01 -1.02017939e+00 9.43384409e-01 5.30070923e-02
-5.10597289e-01 -1.26927447e+00 5.87685481e-02 2.71537274e-01
1.81734428e-01 -1.73075795e-01 6.90226793e-01 -2.38256708e-01
-4.24233019e-01 -1.53791070e-01 2.34436691e-02 -1.39755070e-01
1.70832314e-02 4.95917559e-01 -9.81751978e-01 -5.31145096e-01
6.65065646e-02 -1.92070693e-01 -1.08835101e-01 6.52520418e-01
5.91872573e-01 -2.29931593e-01 -8.56000841e-01 5.40386558e-01
1.38545322e+00 3.85070026e-01 5.32770038e-01 7.28214979e-01
1.41836226e-01 5.53460240e-01 9.17806149e-01 4.63203549e-01
1.30579369e-02 3.28798652e-01 2.40537539e-01 1.49327129e-01
1.11720070e-01 -1.54819340e-01 3.77893507e-01 1.16112018e+00
-8.18235934e-01 -2.69281328e-01 -5.07867396e-01 4.42987174e-01
-1.72482407e+00 -1.40330648e+00 -4.32368398e-01 6.90478683e-01
6.25676990e-01 1.56016424e-01 -1.48347050e-01 3.35214496e-01
4.99015123e-01 -2.03574806e-01 -1.19133167e-01 -1.06291151e+00
-1.43546045e-01 3.15233678e-01 7.37729073e-01 1.00061214e+00
-7.20721722e-01 1.03317809e+00 1.29781246e+01 1.02230716e+00
2.21112028e-01 1.03134915e-01 5.16071796e-01 3.48020852e-01
-4.36954498e-01 -4.56139445e-02 -1.04416132e+00 2.72933897e-02
1.38140702e+00 -4.30666685e-01 6.85999811e-01 5.44219851e-01
3.44648361e-01 -4.23268199e-01 -1.26188684e+00 5.26221812e-01
9.73738134e-02 -1.40886843e+00 -2.83300440e-04 6.85225725e-01
7.73699820e-01 -5.08050561e-01 6.22419357e-01 3.24184299e-01
6.09259963e-01 -1.14389277e+00 8.60300779e-01 2.53660440e-01
1.03040910e+00 -6.05088234e-01 5.67372203e-01 1.68872893e-01
-1.14389896e+00 -2.20873043e-01 -8.77727985e-01 -1.00755692e+00
3.93533185e-02 -1.81779593e-01 -4.29956943e-01 3.48861217e-01
9.58353162e-01 2.99398601e-01 -3.93658698e-01 9.95779395e-01
-4.78694476e-02 1.04875881e-02 -2.71853864e-01 -4.48467314e-01
4.83122796e-01 -3.54241252e-01 4.66730654e-01 1.00164843e+00
2.48499006e-01 3.51035744e-01 -9.84472036e-02 4.01770771e-01
5.45058846e-01 3.29446048e-02 -1.19659424e+00 -1.78908288e-01
2.83276141e-01 9.16795909e-01 -4.83487815e-01 -4.22520459e-01
-2.00212970e-01 8.62069130e-01 -3.55488248e-02 5.01107454e-01
-6.11489356e-01 -4.35615242e-01 9.72222984e-01 -1.27327025e-01
-1.14700586e-01 -3.48497719e-01 -6.23769283e-01 -7.30352640e-01
-5.89872956e-01 -4.54965204e-01 5.93606755e-02 -5.53365827e-01
-1.39813089e+00 5.79277515e-01 -2.27688253e-02 -1.40553558e+00
-6.99901402e-01 -1.27676582e+00 -4.76714373e-01 4.92853165e-01
-1.11898029e+00 -1.10984349e+00 2.50124663e-01 4.52870727e-01
1.64141744e-01 -5.34416080e-01 1.39563632e+00 3.57715860e-02
1.00637585e-01 9.24474537e-01 6.69434488e-01 -7.37814724e-01
5.56605101e-01 -1.27867436e+00 5.68737745e-01 -1.37897313e-01
-4.31265175e-01 9.05828118e-01 6.28349900e-01 -5.39804697e-01
-1.41196322e+00 -3.66917729e-01 1.08350635e+00 -9.83769417e-01
6.55218959e-01 -3.86345625e-01 4.23767231e-02 7.88592756e-01
7.15902448e-01 -6.18741751e-01 8.21781039e-01 -1.83753878e-01
1.80774391e-01 5.75296998e-01 -1.39248300e+00 6.12354755e-01
1.66275799e+00 -4.63594139e-01 -6.25784039e-01 7.60327101e-01
8.13696027e-01 -6.94087505e-01 -1.30082703e+00 3.34633321e-01
8.65424156e-01 -8.75409484e-01 1.61978090e+00 -1.32660246e+00
-4.43697497e-02 2.81152606e-01 -2.61993498e-01 -9.32519078e-01
-5.96193194e-01 -1.23518765e+00 -5.33532679e-01 -5.83747849e-02
5.96577883e-01 -1.13057327e+00 3.42365682e-01 8.74560475e-01
-2.82833427e-01 -6.42737269e-01 -1.06996536e+00 -1.32016802e+00
-3.35779637e-02 -1.45572275e-01 4.90409225e-01 7.63798356e-01
6.83744550e-01 1.09839931e-01 -6.36873543e-02 -1.00294888e-01
5.33176839e-01 9.55312885e-03 4.41501856e-01 -1.34294486e+00
3.86843324e-01 -5.75816095e-01 -3.07655483e-01 -9.45992947e-01
-8.85957032e-02 -8.12076271e-01 -6.53862000e-01 -1.28511906e+00
-8.32044985e-03 -1.91056758e-01 -1.11109078e-01 -1.65725678e-01
3.67937148e-01 2.13746816e-01 1.20859891e-02 1.03788137e-01
-3.71160030e-01 6.18435517e-02 1.29639816e+00 6.91750320e-05
-1.62315920e-01 4.85058486e-01 -4.81304944e-01 7.84440815e-01
8.58408585e-02 -2.99253196e-01 -6.78878546e-01 6.11881316e-02
6.69384480e-01 4.61409837e-02 3.24159935e-02 -7.42885649e-01
5.37211418e-01 -3.75702560e-01 4.78586555e-01 -1.32223868e+00
1.30741090e-01 -9.61415648e-01 6.95283338e-02 9.40189242e-01
2.74610907e-01 1.20424610e-02 8.66204947e-02 5.39500564e-02
-1.44871444e-01 -5.70943117e-01 9.21121240e-01 -4.07591403e-01
-4.92852688e-01 -5.20386267e-03 -1.03226590e+00 8.97834301e-02
9.92593169e-01 -7.84614205e-01 -3.59281451e-01 -4.20183957e-01
-8.29068601e-01 -1.95836127e-02 6.50830388e-01 3.03609259e-02
7.30431557e-01 -1.51530886e+00 -2.30721906e-01 7.18729138e-01
-3.22939813e-01 -3.74400020e-01 -1.70157343e-01 6.58265352e-01
-1.32361674e+00 1.02442718e+00 -5.41665435e-01 -4.55340147e-01
-1.14228773e+00 4.78126436e-01 4.28307921e-01 -2.41845414e-01
-2.15481281e-01 1.11954463e+00 2.71224789e-02 -8.23025763e-01
1.85185194e-01 -9.21545625e-02 -7.54407048e-01 3.55081353e-03
6.88606799e-01 1.05194807e+00 -2.91290224e-01 -6.04341030e-01
-4.56784427e-01 6.65885091e-01 2.32151806e-01 -2.87484169e-01
9.17833567e-01 -2.19243199e-01 -9.89108324e-01 4.28274393e-01
8.38715494e-01 -1.36269778e-01 -3.69319022e-02 4.16855574e-01
1.25943512e-01 -8.02164078e-01 -4.32406247e-01 -3.55811834e-01
-1.85641110e-01 5.54822803e-01 5.34874737e-01 8.99602413e-01
8.57008278e-01 -3.02566767e-01 8.18335712e-01 9.66778398e-01
5.72402716e-01 -1.68019545e+00 -2.13140488e-01 6.89524531e-01
9.12339568e-01 -9.28350806e-01 5.44190466e-01 -7.27165341e-01
-4.14997995e-01 1.32979155e+00 4.68304873e-01 -1.55325383e-01
1.27306652e+00 5.74917436e-01 1.14069022e-02 -3.70670199e-01
-9.44949508e-01 1.12705544e-01 3.75366658e-01 1.11147714e+00
5.06513238e-01 5.07374525e-01 -9.66500878e-01 3.21953118e-01
-7.28706717e-01 -2.34555230e-01 4.90474731e-01 1.41972518e+00
-6.43810987e-01 -1.20391917e+00 -7.23931909e-01 4.61561680e-01
-5.49773455e-01 -1.16372630e-01 -5.28106689e-01 8.46754074e-01
-5.76629937e-02 1.49448860e+00 -1.97535474e-03 -5.03491640e-01
4.11356747e-01 1.41089618e-01 7.21762300e-01 -1.23501487e-01
-9.04846430e-01 4.15413082e-01 3.84890139e-01 -1.23056793e+00
-8.58632207e-01 -1.05834293e+00 -1.40667629e+00 -1.19437599e+00
-5.12782812e-01 1.89310342e-01 3.83317530e-01 3.90289724e-01
-2.06836104e-01 2.85260603e-02 9.80917513e-01 -1.07949340e+00
-3.90341938e-01 -9.39418614e-01 -1.00026262e+00 -7.84516707e-02
2.89751232e-01 -8.00943017e-01 -7.83523321e-01 2.83909619e-01]
|
[-7.261100769042969, 3.6947214603424072]
|
033eca93-2765-48c4-ad65-cb69e1dd645d
|
self-supervised-learning-based-cervical
|
2302.05195
| null |
https://arxiv.org/abs/2302.05195v2
|
https://arxiv.org/pdf/2302.05195v2.pdf
|
Self-supervised learning-based cervical cytology for the triage of HPV-positive women in resource-limited settings and low-data regime
|
Screening Papanicolaou test samples has proven to be highly effective in reducing cervical cancer-related mortality. However, the lack of trained cytopathologists hinders its widespread implementation in low-resource settings. Deep learning-based telecytology diagnosis emerges as an appealing alternative, but it requires the collection of large annotated training datasets, which is costly and time-consuming. In this paper, we demonstrate that the abundance of unlabeled images that can be extracted from Pap smear test whole slide images presents a fertile ground for self-supervised learning methods, yielding performance improvements relative to readily available pre-trained models for various downstream tasks. In particular, we propose \textbf{C}ervical \textbf{C}ell \textbf{C}opy-\textbf{P}asting ($\texttt{C}^{3}\texttt{P}$) as an effective augmentation method, which enables knowledge transfer from open-source and labeled single-cell datasets to unlabeled tiles. Not only does $\texttt{C}^{3}\texttt{P}$ outperforms naive transfer from single-cell images, but we also demonstrate its advantageous integration into multiple instance learning methods. Importantly, all our experiments are conducted on our introduced \textit{in-house} dataset comprising liquid-based cytology Pap smear images obtained using low-cost technologies. This aligns with our objective of leveraging deep learning-based telecytology for diagnosis in low-resource settings.
|
['Jean-Philippe Thiran', 'Pierre Vassilakos', 'Patrick Petignat', 'Holly Clarke', 'Behzad Bozorgtabar', 'Christian Abbet', 'Thomas Stegmüller']
|
2023-02-10
| null | null | null | null |
['whole-slide-images', 'multiple-instance-learning']
|
['computer-vision', 'methodology']
|
[ 6.05869949e-01 3.32092911e-01 -4.71879423e-01 4.39148471e-02
-1.56728327e+00 -7.63280272e-01 2.39094257e-01 4.90664005e-01
-5.80348611e-01 1.12639046e+00 -2.52994955e-01 -9.87286210e-01
-1.09586297e-02 -9.63637054e-01 -9.66187119e-01 -1.26915956e+00
3.89420033e-01 6.97691858e-01 4.21171449e-03 2.26783082e-01
-1.59766540e-01 5.10780513e-01 -1.21541464e+00 5.22296488e-01
8.82452607e-01 9.38845754e-01 2.11515903e-01 9.75835800e-01
-3.06039274e-01 6.19636595e-01 -2.85387903e-01 -5.79511404e-01
1.11421151e-02 -5.15207350e-01 -6.41796947e-01 -1.13798194e-01
5.38226664e-01 -3.51268828e-01 1.35608271e-01 7.15337932e-01
5.64621985e-01 -4.69169199e-01 9.11632478e-01 -7.26229608e-01
-3.89820576e-01 -1.53948441e-01 -6.71901286e-01 2.53058702e-01
-1.73670501e-01 2.52347231e-01 6.74975455e-01 -7.14006782e-01
9.78247464e-01 4.48647052e-01 8.73691022e-01 6.98012888e-01
-1.05840182e+00 -8.21925044e-01 -4.38826531e-01 -2.32365236e-01
-1.03221083e+00 -1.87991738e-01 -9.36415570e-04 -3.27444911e-01
9.71158445e-01 4.53758955e-01 7.83061922e-01 1.03799069e+00
5.62099934e-01 1.01245713e+00 1.20977592e+00 -6.39632761e-01
7.77279288e-02 2.64797419e-01 -3.47279042e-01 1.04121745e+00
6.26054049e-01 -2.22092807e-01 -3.13037902e-01 3.50891203e-02
1.02374208e+00 3.49739432e-01 9.77874994e-02 -6.04623593e-02
-9.03670728e-01 5.38096726e-01 2.01657891e-01 4.75401908e-01
1.73664197e-01 2.10507199e-01 4.76512492e-01 1.25994300e-02
5.68840146e-01 1.97292075e-01 -3.50376725e-01 -1.57081544e-01
-7.79662967e-01 -3.61950964e-01 4.97539580e-01 7.33421326e-01
8.84311140e-01 -3.71533483e-01 -2.45425329e-01 5.90939343e-01
-1.08178914e-01 8.47119570e-01 2.42707729e-01 -7.49912977e-01
2.34502748e-01 7.82917023e-01 1.85668379e-01 -5.66002429e-01
-4.34118241e-01 -4.31699425e-01 -1.08026028e+00 -1.17554627e-01
6.39782965e-01 -3.57342064e-01 -1.14635456e+00 1.13800287e+00
3.24149221e-01 1.75312907e-01 -1.43879885e-02 3.95046324e-01
6.96753085e-01 4.23568964e-01 1.77879453e-01 -1.64186254e-01
1.44145763e+00 -7.50540912e-01 -4.78576958e-01 9.54799205e-02
1.29992378e+00 -6.76231086e-01 1.01495087e+00 2.85386324e-01
-8.78774881e-01 -1.62733033e-01 -6.88675404e-01 -2.41467550e-01
-6.73038602e-01 3.87234420e-01 7.43974984e-01 1.09332073e+00
-1.31094623e+00 1.95276476e-02 -1.10148108e+00 -7.41192102e-01
8.98083329e-01 6.56399608e-01 -5.33882141e-01 -3.61482441e-01
-3.75071466e-01 5.31809986e-01 -2.09828857e-02 1.80663675e-01
-8.24383914e-01 -7.74892092e-01 -7.99094141e-01 -7.85961077e-02
3.85992974e-01 -4.86459911e-01 1.07969272e+00 -5.50939083e-01
-1.44784045e+00 1.37516201e+00 -4.07278270e-01 -3.12330909e-02
5.89958429e-01 1.75532833e-01 -1.44376084e-01 4.65011299e-01
1.78917736e-01 8.40704739e-01 4.64530438e-01 -9.36361194e-01
-9.32453215e-01 -4.04363036e-01 -5.06935239e-01 -3.91306765e-02
-7.61033118e-01 -2.76890486e-01 -7.76740968e-01 -3.16529244e-01
-1.05731592e-01 -9.35899138e-01 -1.47081807e-01 2.35706761e-01
-3.40483963e-01 -1.36012524e-01 8.97504807e-01 -3.35788220e-01
7.43569672e-01 -1.98237622e+00 -4.09609854e-01 3.42040181e-01
2.27233738e-01 6.14852071e-01 4.41034846e-02 3.55341822e-01
4.78518844e-01 4.38759059e-01 6.83043674e-02 -1.25375345e-01
-3.03623617e-01 2.30530828e-01 1.59252644e-01 4.83384818e-01
5.19786000e-01 1.37963510e+00 -9.84185159e-01 -8.97385001e-01
2.89829224e-01 4.15408969e-01 -3.99044693e-01 -1.42406449e-01
-4.34050232e-01 6.34155989e-01 -6.74669206e-01 1.09192610e+00
4.79749501e-01 -7.34255791e-01 3.85575086e-01 2.43383706e-01
8.07126760e-02 -2.86557198e-01 -4.97164041e-01 1.39311993e+00
-1.33865297e-01 5.88113606e-01 9.29070413e-02 -8.28065813e-01
7.43606925e-01 3.17962885e-01 5.36301017e-01 -7.95890450e-01
3.00404906e-01 5.73328435e-01 -2.78371871e-01 -6.67589903e-01
2.03669235e-01 -2.71279663e-01 1.78836972e-01 4.64046150e-01
1.00242123e-02 -4.06336844e-01 3.40991139e-01 -3.08258068e-02
1.38567603e+00 3.77305485e-02 3.46657895e-02 -4.47551727e-01
4.70688224e-01 3.65902066e-01 3.12399298e-01 8.33799481e-01
-1.62117213e-01 5.28250098e-01 6.51241243e-01 -3.44115794e-01
-1.04863274e+00 -9.15809214e-01 -4.91850078e-01 9.12201822e-01
-1.51120782e-01 1.71645477e-01 -5.25794566e-01 -8.26106310e-01
2.90136654e-02 -1.18471399e-01 -1.04811502e+00 3.65610301e-01
-5.55832267e-01 -1.03906727e+00 1.08809471e+00 6.55917048e-01
5.26413500e-01 -7.66040683e-01 -3.44658911e-01 1.57865211e-01
-1.04704171e-01 -7.62255490e-01 -1.60656080e-01 4.61756051e-01
-8.53653312e-01 -1.26328027e+00 -1.16149163e+00 -9.96625483e-01
1.07572019e+00 1.37021057e-02 8.54919732e-01 3.54339570e-01
-7.45528698e-01 2.51768261e-01 -1.99796706e-01 -9.65079963e-01
-3.05244118e-01 3.06766152e-01 -4.67131078e-01 -1.02040820e-01
7.11517453e-01 9.32633504e-02 -9.48236585e-01 1.71460032e-01
-9.69138801e-01 2.50734072e-02 9.70932364e-01 1.13735378e+00
1.26273441e+00 -3.52725744e-01 6.01057291e-01 -1.37324560e+00
4.42686267e-02 -3.31346184e-01 -5.28800964e-01 3.39487612e-01
-2.50013858e-01 -4.88794267e-01 5.72045386e-01 -1.15630932e-01
-1.07287753e+00 -2.00965762e-01 -2.41013736e-01 -5.63049614e-02
-2.15724707e-01 5.44528484e-01 3.34863901e-01 -3.86501625e-02
6.69526398e-01 2.30079770e-01 1.69146240e-01 -4.73689772e-02
-2.27668390e-01 5.75066328e-01 5.19721985e-01 -4.17068183e-01
4.64707762e-01 1.09224629e+00 2.26364255e-01 -8.24200153e-01
-7.69756198e-01 -6.48555040e-01 -3.67281348e-01 -1.03718899e-01
1.05057883e+00 -9.37875986e-01 -1.04133856e+00 4.00492102e-01
-4.29271698e-01 -8.58692944e-01 -2.26000935e-01 3.21689039e-01
-2.93005198e-01 5.75872175e-02 -1.04833663e+00 -7.13437200e-01
-4.78355974e-01 -1.03178430e+00 1.63206875e+00 4.64068711e-01
5.01810275e-02 -1.24985874e+00 8.11108127e-02 6.69751585e-01
3.45946610e-01 5.28694987e-01 1.10165417e+00 -4.63523507e-01
-7.19362140e-01 -6.93076611e-01 -6.26949310e-01 3.62521447e-02
4.05257970e-01 1.27788901e-01 -1.12269604e+00 -5.76129436e-01
-5.05267501e-01 -6.71469390e-01 7.66396224e-01 4.93836731e-01
1.43806672e+00 8.18366110e-02 -1.02412760e+00 6.43051088e-01
1.54584169e+00 8.40746164e-02 6.29400373e-01 1.93925902e-01
4.54024076e-01 2.15924904e-01 4.00049895e-01 9.88380313e-02
3.07268620e-01 -4.89365868e-02 1.34795085e-01 -8.00862730e-01
-6.03411645e-02 -1.10936582e-01 -1.98881134e-01 2.60325938e-01
-9.59265903e-02 -5.05011559e-01 -1.28908360e+00 7.06890523e-01
-1.45401931e+00 -5.51555037e-01 -1.08535223e-01 1.79317498e+00
9.31846023e-01 -5.65415658e-02 -2.82267034e-01 -3.02570146e-02
4.28045422e-01 -5.42969167e-01 -6.31496191e-01 -2.44150311e-01
-1.52343974e-01 7.82993257e-01 3.23851943e-01 1.26897573e-01
-9.80839908e-01 7.21944451e-01 5.42631531e+00 1.12263238e+00
-1.43261969e+00 1.07925735e-01 1.42054236e+00 -1.51227161e-01
-2.26710156e-01 -5.65593362e-01 -8.85741174e-01 2.57544696e-01
7.78444529e-01 1.65352002e-01 -3.64785612e-01 3.41954857e-01
2.95767579e-02 -6.29397273e-01 -1.25750673e+00 5.84716558e-01
1.00485189e-02 -1.89958787e+00 -3.83338798e-03 4.86097813e-01
1.08224106e+00 1.76755175e-01 3.14179063e-01 2.59255737e-01
4.07198071e-01 -1.31737149e+00 -1.21235982e-01 1.47011071e-01
1.59998512e+00 -4.24641728e-01 1.14399219e+00 2.48830408e-01
-1.01254189e+00 1.31147459e-01 -9.79890451e-02 2.59642959e-01
-4.48177904e-01 7.20754802e-01 -1.30333757e+00 6.01267278e-01
9.56629097e-01 4.11946625e-01 -5.12606442e-01 8.01014900e-01
1.10029645e-01 7.07197785e-01 -3.25348139e-01 -3.24806184e-01
3.56852323e-01 7.31786862e-02 -3.51851821e-01 1.59386361e+00
5.54614782e-01 1.68230563e-01 -2.61812210e-02 3.58332932e-01
-3.73643339e-01 1.58930615e-01 -5.95922649e-01 -1.21592477e-01
1.68812767e-01 1.30604219e+00 -1.33240855e+00 -3.10030609e-01
-3.58731657e-01 6.14879549e-01 3.44246328e-01 3.57497513e-01
-5.29201031e-01 -2.38445044e-01 -4.95736375e-02 1.83971301e-01
3.91256571e-01 2.76037693e-01 -3.71809632e-01 -8.80746067e-01
-1.20132647e-01 -6.87782645e-01 6.08393312e-01 -4.31459606e-01
-1.21667242e+00 1.55494019e-01 -4.48697776e-01 -1.15144813e+00
1.06615998e-01 -7.90593982e-01 -3.56960595e-01 7.50372648e-01
-1.95426428e+00 -1.27761829e+00 -4.98759359e-01 5.64124227e-01
1.65077627e-01 3.31924967e-02 1.18832803e+00 9.39951539e-02
-6.44156694e-01 8.51920247e-01 7.76759088e-01 3.05864692e-01
7.35575259e-01 -1.18717897e+00 -2.54067183e-01 3.80408973e-01
-4.66719747e-01 4.28257704e-01 7.62899220e-02 -6.22574985e-01
-1.65141141e+00 -1.33490193e+00 7.42642283e-01 -7.74684846e-01
5.11008918e-01 -3.57008904e-01 -6.88166559e-01 7.91101158e-01
1.26309395e-01 2.73684740e-01 1.30485189e+00 -1.74030185e-01
-6.53373748e-02 -2.16233537e-01 -1.41215527e+00 8.01883757e-01
6.26467526e-01 -4.63886499e-01 2.19254285e-01 6.72849715e-01
1.56599343e-01 -6.40645325e-01 -9.74905968e-01 4.90468591e-01
5.74042678e-01 -8.55032444e-01 5.94788551e-01 -6.35995120e-02
6.41935945e-01 6.84643313e-02 3.30592245e-01 -5.21451175e-01
1.88626379e-01 -5.15379190e-01 3.63265783e-01 9.51115787e-01
7.81188488e-01 -6.78409994e-01 1.62875497e+00 5.56582868e-01
-9.41108689e-02 -8.86545599e-01 -1.15670013e+00 -4.05250162e-01
6.49349868e-01 -7.50732794e-03 1.54386058e-01 9.14661467e-01
1.22265913e-01 -3.81050885e-01 2.66935766e-01 -2.11999342e-01
4.10971969e-01 -2.36698873e-02 7.96793401e-01 -9.42861199e-01
-2.24593520e-01 -2.59770870e-01 -9.46947262e-02 -6.02145433e-01
-2.93305784e-01 -9.35866773e-01 -8.93520936e-03 -1.50940371e+00
5.98247647e-01 -8.94936442e-01 -4.00556803e-01 6.78177357e-01
-2.80715257e-01 8.91202569e-01 -3.01959604e-01 1.70096338e-01
-7.08452761e-01 -7.63980374e-02 1.65372777e+00 -2.44463056e-01
-7.38175064e-02 -2.25408703e-01 -6.68334544e-01 4.33312029e-01
7.17649937e-01 -2.74542093e-01 -1.68632433e-01 -6.13152266e-01
2.47606784e-01 1.89374015e-01 1.07789852e-01 -8.44331145e-01
3.21460634e-01 -1.62146792e-01 7.32919574e-01 -5.56258917e-01
3.17515075e-01 -6.16539001e-01 -6.41109720e-02 7.36706495e-01
4.43508476e-02 -3.79114509e-01 6.57066643e-01 5.65755129e-01
3.86165618e-03 5.43150790e-02 5.63703418e-01 -4.00290877e-01
-1.17230646e-01 2.16858968e-01 -7.05988407e-01 -2.73126215e-01
1.16290379e+00 -7.34923601e-01 -6.90814674e-01 6.59891292e-02
-4.16162282e-01 4.17072594e-01 5.10602534e-01 -1.80603474e-01
2.73053914e-01 -7.49716759e-01 -7.75305688e-01 3.62716675e-01
3.01312327e-01 5.18093288e-01 4.75233644e-01 1.14913046e+00
-9.58888650e-01 7.40768850e-01 2.20197253e-02 -1.05165625e+00
-1.03635752e+00 1.37408316e-01 3.94220203e-01 -7.04497099e-01
-3.94457608e-01 1.22985137e+00 9.07099545e-02 -6.70458674e-01
2.28894532e-01 -4.31112260e-01 1.65228322e-01 -1.33522019e-01
2.73902953e-01 2.25473866e-01 5.08286536e-01 1.03550047e-01
-3.90306748e-02 3.39723438e-01 -4.18131411e-01 7.99520686e-02
1.25691831e+00 1.85666978e-01 -2.01673791e-01 1.60764128e-01
1.12008858e+00 -2.34647289e-01 -1.23367190e+00 -5.98685369e-02
-4.72658165e-02 -2.93019652e-01 -2.95546681e-01 -9.95496452e-01
-8.33260953e-01 8.16888392e-01 5.97949624e-01 -1.55378371e-01
1.02624810e+00 -2.75581721e-02 6.30497098e-01 7.38333225e-01
5.20087063e-01 -1.07821679e+00 3.18383306e-01 1.41330510e-01
1.62448868e-01 -1.47304070e+00 -1.14203453e-01 -5.23650706e-01
-1.05788492e-01 9.20109808e-01 5.73380828e-01 2.00747803e-01
4.57921892e-01 6.29734337e-01 4.14754719e-01 -2.19502762e-01
-7.55925834e-01 -7.65648410e-02 -4.06388760e-01 6.77184165e-01
7.27159798e-01 1.69709455e-02 -2.36448392e-01 1.66766405e-01
1.01834469e-01 5.36150217e-01 5.21801472e-01 1.40432692e+00
-2.70362943e-01 -1.20448363e+00 -3.22133273e-01 1.07312739e+00
-7.59475052e-01 -6.82588965e-02 -3.64670187e-01 9.54339504e-01
2.46760994e-01 7.94827223e-01 1.93606243e-01 3.07240218e-01
-1.31467491e-01 1.52975814e-02 5.19002795e-01 -7.48589694e-01
-7.75920212e-01 5.20425975e-01 -1.83003590e-01 8.80226120e-02
-3.69614691e-01 -7.21561790e-01 -1.29383767e+00 -2.11202830e-01
-2.59536207e-01 8.34969729e-02 3.18330884e-01 9.53878760e-01
2.97853351e-01 4.28029090e-01 1.89292580e-01 -5.84006667e-01
1.76966861e-01 -7.05528498e-01 -6.17436409e-01 4.36574034e-02
4.23040986e-01 -1.75743431e-01 7.95546323e-02 2.53666759e-01]
|
[15.065332412719727, -2.979355573654175]
|
e289da9a-0308-4ff0-9cb2-ce4862de1032
|
hierarchical-graph-neural-networks-for-causal
|
2302.01987
| null |
https://arxiv.org/abs/2302.01987v1
|
https://arxiv.org/pdf/2302.01987v1.pdf
|
Hierarchical Graph Neural Networks for Causal Discovery and Root Cause Localization
|
In this paper, we propose REASON, a novel framework that enables the automatic discovery of both intra-level (i.e., within-network) and inter-level (i.e., across-network) causal relationships for root cause localization. REASON consists of Topological Causal Discovery and Individual Causal Discovery. The Topological Causal Discovery component aims to model the fault propagation in order to trace back to the root causes. To achieve this, we propose novel hierarchical graph neural networks to construct interdependent causal networks by modeling both intra-level and inter-level non-linear causal relations. Based on the learned interdependent causal networks, we then leverage random walks with restarts to model the network propagation of a system fault. The Individual Causal Discovery component focuses on capturing abrupt change patterns of a single system entity. This component examines the temporal patterns of each entity's metric data (i.e., time series), and estimates its likelihood of being a root cause based on the Extreme Value theory. Combining the topological and individual causal scores, the top K system entities are identified as root causes. Extensive experiments on three real-world datasets with case studies demonstrate the effectiveness and superiority of the proposed framework.
|
['Haifeng Chen', 'Yanjie Fu', 'Zheng Wang', 'Liang Tong', 'Jingchao Ni', 'Zhengzhang Chen', 'Dongjie Wang']
|
2023-02-03
| null | null | null | null |
['causal-discovery']
|
['knowledge-base']
|
[-1.77108452e-01 1.43407151e-01 2.73398142e-02 1.14290588e-01
-6.36033192e-02 -5.78988969e-01 5.05284786e-01 4.44912732e-01
7.61890590e-01 6.07815802e-01 3.71745706e-01 -5.12315571e-01
-1.09151316e+00 -1.13151312e+00 -7.82273591e-01 -4.83112723e-01
-1.02429056e+00 2.16274321e-01 4.52015221e-01 1.54984653e-01
4.12103534e-01 5.08016944e-01 -1.00546098e+00 1.16006039e-01
8.53705645e-01 6.66822851e-01 -4.12106887e-02 6.18663132e-01
4.82268631e-02 1.32534790e+00 -6.10837281e-01 2.18048245e-01
-2.66059637e-01 -2.24391848e-01 -8.23907256e-01 -3.61031711e-01
2.51230765e-02 -1.01008005e-01 -6.31470799e-01 7.42520213e-01
9.67991203e-02 -2.99698859e-01 6.27992690e-01 -1.69634676e+00
-7.97001898e-01 9.19399798e-01 -7.89744377e-01 4.23285544e-01
1.79788098e-01 3.87898870e-02 1.16756618e+00 -7.41688013e-01
4.01294321e-01 1.46713126e+00 7.42871881e-01 -4.12671596e-01
-1.36504090e+00 -6.79070115e-01 3.93857539e-01 3.58540237e-01
-1.43606734e+00 -2.38000974e-02 1.05506015e+00 -7.95852304e-01
7.76462495e-01 -9.87406150e-02 3.65234286e-01 6.86258435e-01
7.20562518e-01 2.84790367e-01 6.82115734e-01 6.99332282e-02
3.20139229e-01 -6.28066778e-01 2.68977642e-01 7.50603139e-01
3.74471962e-01 -3.90135534e-02 -5.69653749e-01 -5.71246803e-01
1.07906246e+00 1.66590124e-01 -2.62151361e-01 -1.08809620e-01
-1.11444819e+00 4.88668084e-01 1.13556612e+00 6.04645252e-01
-6.84051752e-01 7.74098933e-01 2.13895231e-01 2.72123158e-01
3.83432716e-01 4.32438940e-01 -5.79212785e-01 3.28800559e-01
-6.30494952e-01 -3.33474934e-01 8.25141251e-01 6.85205400e-01
7.28841782e-01 -7.81405494e-02 -1.22260273e-01 5.34352541e-01
4.70175236e-01 3.04186434e-01 -2.70123005e-01 -5.15082300e-01
3.79742295e-01 1.22332716e+00 -4.34163697e-02 -1.81836390e+00
-5.51748753e-01 -6.15509093e-01 -1.24596870e+00 -1.41516060e-01
1.03351876e-01 -5.51158115e-02 -6.74228370e-01 1.93783426e+00
4.11594301e-01 1.00645876e+00 -3.23430777e-01 5.83150923e-01
4.86615628e-01 8.33346903e-01 -9.12006758e-03 -3.95298123e-01
8.53234053e-01 -3.94931942e-01 -5.87097168e-01 3.13658744e-01
2.56802350e-01 -4.27486122e-01 5.58761835e-01 -4.01254222e-02
-7.58727551e-01 -2.09384561e-01 -7.12263942e-01 7.85309672e-01
-2.41626740e-01 -4.82042553e-03 3.76826942e-01 -1.48886904e-01
-1.07069051e+00 7.50822186e-01 -8.49999249e-01 -3.44295889e-01
-6.10305369e-02 1.38383955e-01 -1.68135241e-01 1.54008627e-01
-1.61194575e+00 2.41067648e-01 4.30425137e-01 1.57450691e-01
-1.06033456e+00 -9.70895290e-01 -3.57108384e-01 3.64108503e-01
6.61913931e-01 -6.03538096e-01 5.07345319e-01 -8.02025050e-02
-6.26721740e-01 -2.87102848e-01 -2.58561391e-02 -9.90419611e-02
6.23309761e-02 -1.60804659e-01 -6.73118293e-01 3.35874632e-02
4.61942077e-01 -2.65783757e-01 5.75563669e-01 -1.39250016e+00
-6.01300895e-01 -1.64083987e-01 4.37191576e-02 -1.87532246e-01
-4.68927771e-01 -4.42405462e-01 -4.48681325e-01 -4.50724274e-01
3.55374366e-01 -6.05558097e-01 1.01105785e-02 -5.85036218e-01
-1.09798992e+00 -5.44850051e-01 9.18733418e-01 -7.26224661e-01
1.85280442e+00 -2.06935549e+00 2.31107622e-01 7.18504965e-01
8.57667029e-01 -5.15252292e-01 8.25351477e-02 1.07007945e+00
-4.17857409e-01 4.58056420e-01 5.12203909e-02 3.82994086e-01
-3.79778087e-01 -7.24075288e-02 -4.62788910e-01 3.31976414e-01
3.70207757e-01 6.14130080e-01 -1.02934730e+00 -2.86806911e-01
9.13879275e-02 2.57413745e-01 -1.25685349e-01 2.69621551e-01
4.81481180e-02 3.68899047e-01 -5.74846327e-01 3.87538671e-01
3.58032912e-01 -8.51358056e-01 2.71209598e-01 -3.23740900e-01
-1.68807790e-01 1.35260627e-01 -1.39449549e+00 8.43661785e-01
-3.60071421e-01 4.01440889e-01 -2.21694574e-01 -8.06393802e-01
7.84048021e-01 3.62859488e-01 7.53183544e-01 -4.00347263e-01
-2.45786518e-01 9.65775028e-02 -2.52075512e-02 -4.64790732e-01
-3.48974913e-02 4.64629292e-01 -2.06816778e-01 5.00071108e-01
-1.29587889e-01 7.16466367e-01 1.85762957e-01 5.96773028e-01
1.92864215e+00 -3.65050375e-01 1.82841182e-01 -3.17139357e-01
3.85253519e-01 -4.38795201e-02 7.31031120e-01 5.81126928e-01
1.18326284e-01 1.09941894e-02 1.13485944e+00 -5.18876910e-01
-8.94699991e-01 -1.42226315e+00 5.01278713e-02 6.34957790e-01
5.12301803e-01 -4.07098383e-01 -4.24842834e-01 -6.53045356e-01
4.36793923e-01 7.27141321e-01 -1.00052094e+00 -3.88985813e-01
-4.89654481e-01 -4.89207238e-01 4.21750546e-01 4.98668760e-01
3.59591514e-01 -9.00474846e-01 -2.98884865e-02 4.56412911e-01
-2.44381785e-01 -5.89868844e-01 -1.31986201e-01 7.02137128e-02
-8.32008660e-01 -1.40805149e+00 -5.19986041e-02 -4.02548045e-01
6.62365258e-01 9.73400995e-02 1.09249842e+00 3.25420350e-01
-3.17834556e-01 2.30658069e-01 -5.03974268e-03 3.07731569e-01
-3.02720368e-01 -7.92193189e-02 3.11399043e-01 1.14999369e-01
-1.93181381e-01 -9.82667744e-01 -7.99519122e-01 6.50881350e-01
-5.21679878e-01 -2.07648754e-01 7.89495528e-01 7.01042831e-01
5.77797115e-01 1.05612922e+00 8.97082269e-01 -4.93047476e-01
8.70654702e-01 -1.21502054e+00 -4.88836229e-01 4.29281592e-01
-1.07521451e+00 1.26995757e-01 6.93214715e-01 -4.75815237e-01
-7.96944618e-01 -3.87888700e-01 7.50824451e-01 -6.53243363e-01
-2.13838533e-01 1.05838919e+00 -4.18423563e-02 4.56938505e-01
5.93240976e-01 -1.41269062e-02 -6.85140967e-01 -2.73132712e-01
4.90786612e-01 1.60852700e-01 5.94701707e-01 -4.74858254e-01
1.15348375e+00 2.60872871e-01 2.74173319e-01 -4.93039817e-01
-3.73493850e-01 -5.01743495e-01 -6.64586246e-01 -7.09957421e-01
7.42853940e-01 -7.01265931e-01 -8.23230922e-01 2.30077907e-01
-1.16752839e+00 -5.08103408e-02 1.35078490e-01 1.88832641e-01
-1.17184274e-01 -1.94595784e-01 -8.65455687e-01 -8.04649949e-01
-1.58297688e-01 -5.83160043e-01 8.12393427e-01 1.27822548e-01
-3.32524776e-01 -1.37135303e+00 3.71046096e-01 -3.27852726e-01
1.48988247e-01 5.53797901e-01 1.27853954e+00 -4.57941741e-01
-8.57215583e-01 -2.58686751e-01 -5.93660712e-01 -4.78333950e-01
6.09603167e-01 4.53917265e-01 -2.73524940e-01 -1.23248674e-01
-3.97849530e-01 4.17832583e-01 4.79775161e-01 5.06949425e-01
6.31032765e-01 -5.18130362e-01 -9.00379777e-01 8.67314935e-02
1.45930815e+00 2.28417903e-01 4.68679845e-01 5.23597822e-02
1.18046403e+00 6.77358747e-01 4.07883257e-01 3.48704576e-01
5.08233666e-01 2.33788341e-01 7.33563900e-01 -1.76093787e-01
-3.73100489e-02 -5.67089617e-01 1.12572692e-01 1.17274940e+00
6.17927760e-02 -2.06014946e-01 -1.54002523e+00 9.23073053e-01
-2.08284926e+00 -1.07562244e+00 -9.04981554e-01 1.91936696e+00
4.63295907e-01 3.12451988e-01 -2.79436465e-02 -7.90768191e-02
1.28787577e+00 -1.23695135e-01 -8.62836182e-01 1.96875155e-01
2.36882791e-01 -6.28241241e-01 3.24693084e-01 3.45687389e-01
-8.83672059e-01 5.83107173e-01 5.18141937e+00 5.47729790e-01
-8.16980302e-01 -5.98394759e-02 5.31653166e-01 1.82978287e-01
-2.78521985e-01 2.19249308e-01 -2.68553555e-01 5.61800599e-01
9.77539301e-01 -3.81500691e-01 3.57566804e-01 5.79741836e-01
5.39619327e-01 1.93369955e-01 -1.00551057e+00 4.46288168e-01
-6.18527591e-01 -1.36834335e+00 -1.10605873e-01 1.81679085e-01
8.38443220e-01 -9.04148258e-03 -2.11754009e-01 -6.00584820e-02
8.85378778e-01 -7.55620956e-01 4.66427773e-01 9.55383062e-01
6.08577967e-01 -8.45952213e-01 6.32757664e-01 2.60130554e-01
-1.94718933e+00 -2.57544607e-01 8.78689736e-02 -2.24792417e-02
3.02693754e-01 1.37950373e+00 -9.30220068e-01 9.03874576e-01
9.14841890e-01 9.70011950e-01 -4.14392203e-01 1.06469893e+00
-6.11198604e-01 9.92282808e-01 -3.48415405e-01 8.83483887e-02
-1.15075640e-01 2.31999140e-02 6.92397296e-01 8.48391533e-01
3.15428257e-01 -4.91163880e-02 2.53429502e-01 1.24721014e+00
-2.22973861e-02 -1.76550969e-01 -6.12220347e-01 -8.80929008e-02
9.74054277e-01 1.22689533e+00 -1.04160905e+00 -1.21578224e-01
2.92254500e-02 5.50372839e-01 4.47659671e-01 6.41070902e-01
-9.29103613e-01 -4.61923867e-01 4.28529352e-01 1.04271121e-01
-1.43226944e-02 -2.83279508e-01 -5.90821207e-01 -8.80005181e-01
7.33822631e-03 -1.14015035e-01 4.94909525e-01 -6.60995781e-01
-1.63184834e+00 3.30565751e-01 2.43251938e-02 -1.14260113e+00
-7.77293742e-02 5.81910573e-02 -1.31390131e+00 9.65783656e-01
-9.35731649e-01 -8.86457860e-01 -4.28415090e-01 6.17443919e-01
-3.01577784e-02 1.66302234e-01 3.43319476e-01 1.54444799e-01
-9.76008534e-01 1.61665142e-01 1.69461280e-01 1.10168613e-01
5.26087582e-01 -1.51818943e+00 6.39690459e-01 9.71933186e-01
-2.51421422e-01 1.01374507e+00 5.12436926e-01 -1.47449982e+00
-1.32266510e+00 -1.38355529e+00 7.11797893e-01 -4.37821448e-01
1.58174825e+00 -2.55297154e-01 -1.02809393e+00 4.53413039e-01
-9.48684514e-02 -2.05595940e-01 2.98028737e-01 4.98652577e-01
-4.22283381e-01 -1.60816565e-01 -8.49230826e-01 5.29905260e-01
1.11952746e+00 -5.33655465e-01 -4.10596699e-01 2.69618422e-01
1.15594959e+00 3.72882515e-01 -1.37685752e+00 5.31377614e-01
2.57236838e-01 -6.01806521e-01 9.06253040e-01 -1.94542751e-01
8.34300399e-01 -7.97844231e-01 2.18743756e-01 -1.38210106e+00
-1.09565604e+00 -3.09367478e-01 -5.63639879e-01 1.58555210e+00
4.47100103e-01 -6.80397391e-01 2.96357125e-01 1.65667281e-01
6.01288490e-02 -7.61612713e-01 -8.40034425e-01 -9.48365986e-01
-1.99844137e-01 -2.75726318e-01 6.84997439e-01 1.45536530e+00
1.02543131e-01 6.73679590e-01 -3.42544049e-01 1.13875461e+00
1.01303458e+00 3.18135530e-01 4.51370478e-01 -1.50258672e+00
-2.80796587e-01 -5.24378061e-01 -4.36495245e-01 -3.95888329e-01
-1.46827355e-01 -6.23644829e-01 -9.86543763e-03 -1.88894558e+00
1.31206602e-01 -6.16033435e-01 -7.99886167e-01 6.07210875e-01
-4.21268374e-01 -3.83963466e-01 -3.60174328e-01 7.80195296e-01
-4.08809155e-01 3.88256371e-01 7.30480552e-01 -8.55396092e-02
-4.20825660e-01 -2.78084934e-01 -3.75381082e-01 6.22364879e-01
7.31626749e-01 -5.40059447e-01 -6.06558919e-01 -1.08126082e-01
5.22328794e-01 5.21962762e-01 6.36016488e-01 -1.01021838e+00
6.18278205e-01 -3.77746731e-01 2.69998908e-01 -7.74684548e-01
-4.50440705e-01 -8.53958905e-01 5.59545994e-01 6.70639515e-01
-2.56909847e-01 6.44588917e-02 6.48713019e-03 1.11218786e+00
-2.71639556e-01 5.58012903e-01 1.92290887e-01 3.76814455e-01
-5.87064385e-01 3.91664147e-01 -1.08349144e-01 -2.22424418e-01
1.07930028e+00 3.31447363e-01 -7.86578000e-01 -3.65632057e-01
-8.46128464e-01 7.56075501e-01 8.93753916e-02 4.71028924e-01
5.92876315e-01 -1.56255805e+00 -6.64415836e-01 -3.32798004e-01
2.53105730e-01 -2.88302779e-01 3.93018275e-01 1.14193046e+00
-1.26054108e-01 2.55594105e-01 2.71001071e-01 -7.38264918e-01
-8.98605347e-01 7.57005751e-01 2.83231109e-01 -5.63087404e-01
-7.37285197e-01 7.29166210e-01 5.53228438e-01 -3.45048815e-01
-1.41675919e-02 -2.42386349e-02 -2.68749326e-01 -6.02071173e-03
2.65464276e-01 7.33889818e-01 -2.51620233e-01 -2.58389235e-01
-5.77673852e-01 4.60377693e-01 1.67454496e-01 2.53781080e-01
1.37260270e+00 -2.69756168e-01 -7.48767912e-01 9.36814785e-01
9.88862038e-01 -2.80003518e-01 -1.06557798e+00 -2.04957724e-01
3.10357422e-01 -1.45628244e-01 -3.44928354e-03 -9.90880549e-01
-1.01334536e+00 5.33458352e-01 4.11159456e-01 9.69074905e-01
1.31250405e+00 3.64571154e-01 4.49108601e-01 8.82773548e-02
5.81747651e-01 -6.24972284e-01 2.11824983e-01 4.86166090e-01
9.39024985e-01 -5.20222902e-01 -2.30477512e-01 -6.08370841e-01
2.68939316e-01 1.09820783e+00 5.91475964e-01 -2.69680887e-01
8.48673284e-01 2.95744061e-01 -4.50691104e-01 -8.36671412e-01
-1.01171613e+00 1.59178406e-01 3.50153565e-01 6.39192909e-02
1.20234028e-01 4.49957609e-01 9.58475173e-02 2.18883604e-01
8.17730352e-02 -2.16385871e-01 4.52892095e-01 5.58656216e-01
-4.30665314e-01 -4.98402894e-01 -5.26787937e-01 6.50258958e-01
7.05625713e-02 -1.64486155e-01 -7.11884022e-01 5.29805720e-01
-6.34626523e-02 1.23285818e+00 5.89966699e-02 -1.05852211e+00
4.58610028e-01 -1.79864556e-01 -3.89198303e-01 -2.95369625e-01
-1.82471246e-01 -1.80856455e-02 -4.91087250e-02 -8.10361445e-01
1.53186545e-01 -5.61519742e-01 -1.44213128e+00 -6.06314778e-01
-4.96187180e-01 6.15407936e-02 3.92845362e-01 8.21654618e-01
6.88183010e-01 1.26939178e+00 1.00615299e+00 -3.68102849e-01
1.00250863e-01 -9.23899651e-01 -6.08924925e-01 4.05426532e-01
3.89681339e-01 -9.11289334e-01 -6.36991262e-01 -3.83577980e-02]
|
[7.659582138061523, 5.060122013092041]
|
3a5f648d-c58a-4c64-b402-5676c3be0d34
|
high-fidelity-audio-compression-with-improved
|
2306.06546
| null |
https://arxiv.org/abs/2306.06546v1
|
https://arxiv.org/pdf/2306.06546v1.pdf
|
High-Fidelity Audio Compression with Improved RVQGAN
|
Language models have been successfully used to model natural signals, such as images, speech, and music. A key component of these models is a high quality neural compression model that can compress high-dimensional natural signals into lower dimensional discrete tokens. To that end, we introduce a high-fidelity universal neural audio compression algorithm that achieves ~90x compression of 44.1 KHz audio into tokens at just 8kbps bandwidth. We achieve this by combining advances in high-fidelity audio generation with better vector quantization techniques from the image domain, along with improved adversarial and reconstruction losses. We compress all domains (speech, environment, music, etc.) with a single universal model, making it widely applicable to generative modeling of all audio. We compare with competing audio compression algorithms, and find our method outperforms them significantly. We provide thorough ablations for every design choice, as well as open-source code and trained model weights. We hope our work can lay the foundation for the next generation of high-fidelity audio modeling.
|
['Kundan Kumar', 'Ishaan Kumar', 'Alejandro Luebs', 'Prem Seetharaman', 'Rithesh Kumar']
|
2023-06-11
| null | null | null | null |
['audio-generation', 'quantization']
|
['audio', 'methodology']
|
[ 4.13808912e-01 7.57549107e-02 -2.05182567e-01 -1.64543077e-01
-1.37704587e+00 -3.08128297e-01 4.83873338e-01 -2.09258214e-01
-1.18778192e-01 5.79933763e-01 7.60354042e-01 -4.70619760e-02
2.01831758e-01 -6.36127830e-01 -1.01261175e+00 -4.06763703e-01
-3.67664635e-01 3.50576103e-01 -7.65888765e-02 -1.55467346e-01
-1.78308144e-01 2.04328880e-01 -1.71645463e+00 6.44998968e-01
2.04728067e-01 9.43143785e-01 1.32710159e-01 1.22818279e+00
2.00888872e-01 9.49457109e-01 -6.61453009e-01 -3.77263546e-01
5.08920252e-02 -4.17392194e-01 -6.75290525e-01 -3.76634330e-01
4.33018446e-01 -7.90652335e-01 -1.16475129e+00 7.52476394e-01
8.20312023e-01 2.85070087e-03 8.05866003e-01 -1.11547065e+00
-7.48855412e-01 1.09595776e+00 -1.65412232e-01 -3.08895577e-03
2.56675661e-01 2.60522515e-01 1.10814571e+00 -7.08463788e-01
4.82887715e-01 1.39354300e+00 8.02146614e-01 8.09956849e-01
-1.18364918e+00 -9.21870112e-01 -4.76719916e-01 2.43521228e-01
-1.60508239e+00 -1.02260828e+00 6.55582428e-01 -1.82580128e-01
1.06363750e+00 2.09718123e-01 6.50914788e-01 1.42898333e+00
1.60621732e-01 8.86764586e-01 3.34920853e-01 -3.84839684e-01
2.54158646e-01 -1.63000420e-01 -5.56804657e-01 2.78082192e-01
-2.89618254e-01 2.38024592e-01 -8.40157092e-01 -3.30422312e-01
9.07592952e-01 -1.63805649e-01 -3.24483693e-01 2.55964428e-01
-1.02992582e+00 7.79181302e-01 2.04178154e-01 7.78655559e-02
-1.77807599e-01 9.80474353e-01 5.44292092e-01 3.61555099e-01
9.29986835e-02 1.62339345e-01 -1.86155140e-02 -7.57136822e-01
-1.36343896e+00 5.65696120e-01 5.76299906e-01 1.12816691e+00
3.10934961e-01 8.00186753e-01 -3.63463126e-02 1.00218165e+00
2.78349459e-01 5.81322849e-01 7.77582705e-01 -1.49052906e+00
2.75279939e-01 -6.31075025e-01 -2.47511357e-01 -8.30012798e-01
1.70790985e-01 -4.60200191e-01 -1.02463078e+00 -1.45503819e-01
-2.64698952e-01 -1.08351685e-01 -8.70749533e-01 2.00782490e+00
-1.35998115e-01 7.36223578e-01 1.26131445e-01 6.56257451e-01
5.19279838e-01 1.04549062e+00 -2.72665750e-02 6.88905641e-02
1.11296690e+00 -6.91485286e-01 -7.55371392e-01 -3.25980298e-02
1.90626666e-01 -8.42776000e-01 1.15136111e+00 6.55223012e-01
-1.49000800e+00 -7.87015140e-01 -1.30888271e+00 -3.51614177e-01
1.06560051e-01 -3.85341197e-01 5.66742122e-01 5.60217381e-01
-1.16983449e+00 8.19029987e-01 -9.41612840e-01 2.51220584e-01
3.01099509e-01 1.76732942e-01 -9.82674584e-03 -1.65610731e-01
-1.57842422e+00 2.52035260e-01 2.61375695e-01 -6.35236204e-01
-1.56745946e+00 -9.29945111e-01 -8.71048152e-01 3.24938089e-01
-3.79109561e-01 -8.50286603e-01 1.69341695e+00 -5.87993920e-01
-1.50895941e+00 3.88328880e-01 -1.81123265e-03 -1.16786051e+00
1.14120916e-01 -2.70367414e-01 -6.25651538e-01 5.13584435e-01
-3.11104923e-01 1.15391874e+00 1.32978714e+00 -1.12949431e+00
-4.30085510e-01 1.10768259e-01 -2.80869097e-01 -6.47772430e-03
-7.62414515e-01 -1.21782891e-01 -4.08839196e-01 -1.19857407e+00
-3.06434065e-01 -7.44635046e-01 -5.57848252e-02 1.28280781e-02
-1.62400782e-01 2.00490564e-01 8.88373256e-01 -7.97245264e-01
1.29073846e+00 -2.56700945e+00 8.76359344e-02 -1.08283550e-01
2.49316528e-01 2.59736981e-02 -3.55583042e-01 5.13599873e-01
1.41217615e-02 3.29166561e-01 -1.98174387e-01 -8.78720462e-01
3.93582582e-01 2.12102816e-01 -1.02073276e+00 3.93272154e-02
1.28705382e-01 8.32674682e-01 -5.55659711e-01 -4.68669444e-01
5.29062450e-02 1.11924398e+00 -1.11744523e+00 1.18541121e-01
-3.05811077e-01 -8.80402396e-04 4.34699245e-02 5.26251972e-01
4.43650275e-01 3.85693721e-02 -2.71395415e-01 -1.30281851e-01
3.89795154e-01 6.20399237e-01 -9.54956949e-01 2.17204213e+00
-5.09479105e-01 8.32932770e-01 3.52918804e-01 -6.15117311e-01
6.45806730e-01 7.78517306e-01 5.15592515e-01 -4.03340518e-01
-4.59963307e-02 1.17955349e-01 -2.22287998e-01 -2.08741188e-01
7.59331942e-01 -3.66003901e-01 -1.03738092e-01 5.15823901e-01
5.42103648e-01 -7.54759073e-01 -2.14260221e-01 4.92095113e-01
9.84949589e-01 -2.51363635e-01 -1.78759739e-01 2.84443170e-01
-2.90914804e-01 -4.93181556e-01 2.27194548e-01 8.25237930e-01
-1.37379080e-01 9.37691629e-01 3.98046616e-03 -1.03724352e-03
-1.60259652e+00 -1.20712149e+00 -1.79388866e-01 1.00506639e+00
-4.32553381e-01 -9.12527919e-01 -8.86445284e-01 2.51825064e-01
6.66574985e-02 6.02909923e-01 -1.85287923e-01 -6.31175518e-01
-4.19875324e-01 -3.42202365e-01 1.41949165e+00 5.52943170e-01
1.74123242e-01 -8.41644108e-01 -1.71909675e-01 4.00538802e-01
-3.66538227e-01 -1.05286694e+00 -6.13374889e-01 2.95621008e-01
-8.34778309e-01 -1.44420788e-01 -8.72708738e-01 -8.10525715e-01
-4.41502005e-01 3.59795988e-02 1.15820003e+00 -1.77332744e-01
-2.80564666e-01 4.92462307e-01 -2.53609687e-01 -4.82501805e-01
-9.13304567e-01 2.90785674e-02 6.10200047e-01 -3.53261530e-01
-9.23097413e-03 -1.30909967e+00 -4.79001105e-01 -1.95939526e-01
-1.30241060e+00 -1.22836135e-01 4.11897779e-01 6.75232470e-01
8.82916033e-01 2.92562008e-01 7.89393783e-01 -2.54202157e-01
8.04436803e-01 -6.13278866e-01 -2.74285376e-02 -3.18030387e-01
-3.24411273e-01 -4.92592528e-02 7.78340578e-01 -6.72854841e-01
-4.81696516e-01 -1.17344521e-01 -7.16007769e-01 -9.72501576e-01
-1.17188968e-01 4.34807509e-01 4.72872779e-02 1.07511930e-01
9.34370279e-01 5.31927168e-01 -1.39472056e-02 -5.40642858e-01
6.99051499e-01 1.09897768e+00 1.13477290e+00 -8.30625117e-01
6.33671165e-01 2.44953454e-01 -1.96984217e-01 -9.45700824e-01
-2.22121909e-01 4.07679789e-02 3.44360732e-02 1.20527565e-01
5.01141787e-01 -1.45183563e+00 -3.54923874e-01 2.85771281e-01
-1.00922775e+00 -5.08720756e-01 -7.43073106e-01 5.26710749e-01
-1.08864486e+00 4.56559569e-01 -1.28595912e+00 -7.87485301e-01
-4.39023256e-01 -9.76886213e-01 1.20840061e+00 -1.37061328e-01
-2.79911011e-01 -6.09347820e-01 1.59112707e-01 1.24535881e-01
6.76379859e-01 -2.07541175e-02 8.71982038e-01 -2.36131147e-01
-6.42064989e-01 -1.36769965e-01 2.57771730e-01 8.20566237e-01
-2.48386472e-01 -1.40657853e-02 -1.23304665e+00 -4.53738481e-01
1.23000227e-01 -8.85008812e-01 1.07834804e+00 5.46469510e-01
1.59221697e+00 -3.89236540e-01 8.19348246e-02 9.72773075e-01
1.14344501e+00 9.90213528e-02 8.28628719e-01 -1.81309655e-01
3.89793694e-01 -7.98526928e-02 8.64619762e-02 7.38737643e-01
6.14345670e-02 6.37041748e-01 4.15215999e-01 -8.35075788e-03
-5.08092463e-01 -6.76410735e-01 6.02491736e-01 1.40075147e+00
1.51869908e-01 -2.66130924e-01 -6.04188144e-01 6.64949954e-01
-1.25320065e+00 -1.23233926e+00 6.08318686e-01 2.00036621e+00
1.29589307e+00 2.42863968e-01 1.58758149e-01 6.36239529e-01
3.51554841e-01 1.68331832e-01 -4.82105076e-01 -4.44304287e-01
-1.95836425e-01 6.86613917e-01 1.30849570e-01 5.43081462e-01
-1.04659116e+00 7.31011331e-01 7.85211563e+00 1.33466864e+00
-1.04364419e+00 9.81503353e-02 5.67100227e-01 -6.17315948e-01
-5.37578344e-01 -4.16247427e-01 -6.95687532e-01 5.05568624e-01
1.84090757e+00 -4.34018672e-01 9.93604481e-01 8.25583220e-01
1.52585685e-01 7.36988902e-01 -1.16560817e+00 1.33978081e+00
-8.20857659e-03 -1.53401089e+00 5.86759388e-01 1.41168505e-01
6.00702286e-01 2.13712052e-01 5.50216138e-01 4.76173818e-01
1.92237020e-01 -1.38458955e+00 8.32649231e-01 3.54339182e-01
1.46430492e+00 -1.01837027e+00 1.95062190e-01 1.91903278e-01
-1.23051965e+00 -5.87159507e-02 -5.91547966e-01 1.18381254e-01
4.18056875e-01 4.52796787e-01 -7.34927893e-01 2.67456383e-01
6.70866251e-01 7.80039787e-01 -1.74640007e-02 7.27010608e-01
2.36681879e-01 1.06149244e+00 -5.03040016e-01 4.04837430e-01
8.82233381e-02 5.13509214e-01 5.01683235e-01 1.47705209e+00
5.97396553e-01 3.57349366e-02 -5.26150391e-02 8.29411149e-01
-5.13135791e-01 -2.46356100e-01 -7.13333905e-01 -4.13419187e-01
8.39146316e-01 5.20461261e-01 1.58050388e-01 -3.88653487e-01
-1.03637248e-01 1.00565338e+00 -1.05476342e-01 3.12272936e-01
-1.10242462e+00 -6.54182494e-01 1.09168208e+00 1.37408465e-01
2.63110876e-01 -3.77522528e-01 -6.56356812e-02 -1.07280433e+00
-2.46339947e-01 -1.24674189e+00 5.02571315e-02 -1.00178075e+00
-1.04705226e+00 6.41159713e-01 -1.12988189e-01 -1.21368599e+00
-8.85931671e-01 -1.44496709e-01 -1.10046804e-01 7.46939480e-01
-1.34765923e+00 -1.00619698e+00 1.03889026e-01 8.87520313e-01
6.07452393e-01 -4.50329572e-01 1.21021187e+00 7.91498244e-01
-9.12719034e-03 1.09788620e+00 3.14934015e-01 4.17226106e-02
6.92148685e-01 -8.63636255e-01 6.94122076e-01 5.89181185e-01
4.91971493e-01 5.37788451e-01 7.77895212e-01 -2.68729955e-01
-1.46824288e+00 -1.21018612e+00 4.66224968e-01 -1.94755673e-01
5.54308951e-01 -4.55857068e-01 -8.15976143e-01 8.48508894e-01
2.91057467e-01 -2.48515427e-01 1.01127219e+00 -2.47163698e-01
-5.62287271e-01 -4.17201668e-02 -1.02321935e+00 6.35432541e-01
9.75646496e-01 -1.15464032e+00 -2.86727637e-01 2.78995484e-01
1.60911655e+00 -4.35637683e-01 -1.30264187e+00 2.44371295e-01
6.57556772e-01 -5.92548013e-01 1.41776049e+00 -6.31636024e-01
9.31461632e-01 -1.76461097e-02 -7.59263813e-01 -1.12008095e+00
-3.77653450e-01 -1.08989072e+00 -7.81315327e-01 1.18888581e+00
2.02457920e-01 -1.99148413e-02 6.99930966e-01 1.81576595e-01
-3.82239997e-01 -5.68095446e-01 -1.15062022e+00 -7.69105673e-01
3.15240294e-01 -9.63427961e-01 8.07117820e-01 6.60269618e-01
6.34258017e-02 1.99611783e-01 -9.51169789e-01 -5.21050245e-02
6.27493024e-01 -4.07941431e-01 7.63966560e-01 -7.73267567e-01
-8.62584233e-01 -3.58759671e-01 -5.74683905e-01 -1.50878859e+00
4.99386601e-02 -8.57536018e-01 -7.43738636e-02 -1.02165616e+00
3.80582102e-02 -3.48186076e-01 -3.03587139e-01 2.94009119e-01
4.41408038e-01 7.10656106e-01 2.88861811e-01 2.37395018e-01
-2.57815242e-01 9.53389645e-01 1.06734848e+00 -3.61842245e-01
-1.79623812e-02 -3.23069811e-01 -8.19161057e-01 3.84675711e-01
7.47904003e-01 -4.96795535e-01 -6.29697323e-01 -7.88722336e-01
-6.22933432e-02 2.85582185e-01 2.88033992e-01 -1.60482454e+00
1.11635074e-01 6.79607093e-02 3.01668078e-01 -1.95860654e-01
9.74811018e-01 -5.77763677e-01 3.49387735e-01 3.07674974e-01
-7.08304226e-01 -1.91672012e-01 3.78544152e-01 6.27741039e-01
-5.56928217e-01 3.21958102e-02 9.37048435e-01 -7.99392257e-03
-1.73284978e-01 4.04144436e-01 -3.76385659e-01 1.99826077e-01
3.82940471e-01 6.87897354e-02 -9.55664665e-02 -1.27787590e+00
-9.39691663e-01 -3.04897398e-01 4.57020193e-01 4.23766851e-01
9.08529103e-01 -1.67580366e+00 -1.05062795e+00 3.78658831e-01
-3.27741295e-01 -7.04336911e-02 4.15911645e-01 4.82659712e-02
-4.43138957e-01 4.65569824e-01 -9.57334191e-02 -5.22720993e-01
-1.04207218e+00 4.93407130e-01 2.30563149e-01 6.82446063e-02
-6.24722183e-01 8.08104098e-01 -1.60681501e-01 2.02951267e-01
6.52047276e-01 -3.44102442e-01 3.31896752e-01 -5.48930883e-01
1.05511439e+00 1.14146605e-01 -9.51217562e-02 -4.90619063e-01
1.38317421e-01 1.92248076e-01 9.93677601e-02 -7.08709180e-01
1.29415321e+00 4.67976183e-02 2.97717243e-01 4.74607944e-01
1.43728268e+00 4.82724793e-02 -1.22188854e+00 -1.16786338e-01
-8.54942024e-01 -3.02794427e-01 1.56489611e-01 -3.49638373e-01
-9.54270065e-01 1.17179465e+00 5.97275198e-01 2.05129027e-01
1.29217887e+00 -5.55363810e-03 1.41200197e+00 2.35226020e-01
4.19029415e-01 -9.02518630e-01 2.22756237e-01 5.23866653e-01
9.83418345e-01 -4.58749533e-01 -1.61166325e-01 1.56168193e-02
-4.81362909e-01 8.37630332e-01 2.50782445e-03 -2.44701609e-01
7.17574298e-01 8.33983600e-01 -2.46066242e-01 3.37003857e-01
-1.10193765e+00 3.45907182e-01 -5.82564399e-02 9.17594016e-01
4.34524536e-01 8.21176618e-02 2.82529056e-01 8.08977544e-01
-8.08501363e-01 1.46128044e-01 4.52074051e-01 8.24781001e-01
-5.05982876e-01 -1.03687799e+00 -3.76314193e-01 4.13464218e-01
-6.61118150e-01 -4.39185828e-01 -6.97021186e-02 2.86690772e-01
-6.03105910e-02 8.91908824e-01 2.25974828e-01 -8.48472416e-01
-5.45664318e-02 2.65509933e-01 3.50241035e-01 -2.74142087e-01
-3.26130748e-01 2.51777053e-01 -9.26880911e-02 -5.08778036e-01
-5.77927381e-02 -5.05462289e-01 -1.17346442e+00 -7.13747203e-01
1.18250921e-01 9.39898044e-02 7.31783092e-01 2.04292521e-01
7.41812170e-01 6.87971950e-01 5.43779433e-01 -1.18272555e+00
-7.53253400e-01 -9.72056508e-01 -6.97090089e-01 3.62440884e-01
6.81290030e-01 -1.31074116e-01 -4.48392272e-01 5.40742755e-01]
|
[15.525093078613281, 5.8619384765625]
|
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.