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
|
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
817c73b0-ec47-47a9-a5df-1d477a1eb79e
|
starss23-an-audio-visual-dataset-of-spatial
|
2306.09126
| null |
https://arxiv.org/abs/2306.09126v1
|
https://arxiv.org/pdf/2306.09126v1.pdf
|
STARSS23: An Audio-Visual Dataset of Spatial Recordings of Real Scenes with Spatiotemporal Annotations of Sound Events
|
While direction of arrival (DOA) of sound events is generally estimated from multichannel audio data recorded in a microphone array, sound events usually derive from visually perceptible source objects, e.g., sounds of footsteps come from the feet of a walker. This paper proposes an audio-visual sound event localization and detection (SELD) task, which uses multichannel audio and video information to estimate the temporal activation and DOA of target sound events. Audio-visual SELD systems can detect and localize sound events using signals from a microphone array and audio-visual correspondence. We also introduce an audio-visual dataset, Sony-TAu Realistic Spatial Soundscapes 2023 (STARSS23), which consists of multichannel audio data recorded with a microphone array, video data, and spatiotemporal annotation of sound events. Sound scenes in STARSS23 are recorded with instructions, which guide recording participants to ensure adequate activity and occurrences of sound events. STARSS23 also serves human-annotated temporal activation labels and human-confirmed DOA labels, which are based on tracking results of a motion capture system. Our benchmark results show that the audio-visual SELD system achieves lower localization error than the audio-only system. The data is available at https://zenodo.org/record/7880637.
|
['Yuki Mitsufuji', 'Tuomas Virtanen', 'Shusuke Takahashi', 'Naoya Takahashi', 'Yuichiro Koyama', 'Aapo Hakala', 'Sharath Adavanne', 'Kengo Uchida', 'Daniel Krause', 'Parthasaarathy Sudarsanam', 'Archontis Politis', 'Kazuki Shimada']
|
2023-06-15
| null | null | null | null |
['sound-event-localization-and-detection']
|
['audio']
|
[-2.33354941e-01 -8.04310858e-01 3.22037637e-01 -7.40321307e-03
-1.27824593e+00 -6.34834886e-01 1.13524236e-01 2.65986085e-01
-3.20472956e-01 2.23271027e-01 5.11067033e-01 6.62364289e-02
2.82430649e-01 -2.52029270e-01 -5.36409259e-01 -4.02667135e-01
-4.13568497e-01 -3.29003304e-01 6.83879972e-01 3.62465769e-01
2.02435136e-01 2.78048843e-01 -1.95232356e+00 6.23690844e-01
-1.90280050e-01 1.41847932e+00 3.72282743e-01 1.46930718e+00
1.86360374e-01 7.44653761e-01 -8.10263157e-01 7.29382455e-01
6.76530078e-02 -4.90137160e-01 -3.56862955e-02 -4.04622525e-01
7.17055082e-01 -1.91328168e-01 -5.53811967e-01 8.08247626e-01
9.38764811e-01 3.48875135e-01 2.61317194e-01 -1.76414454e+00
-1.18988760e-01 1.45055920e-01 -2.62405068e-01 8.08012366e-01
1.08352947e+00 1.29793033e-01 8.34382594e-01 -1.32790399e+00
2.97543079e-01 1.05977213e+00 7.08702028e-01 2.69521505e-01
-7.06502557e-01 -9.91145611e-01 -1.14710540e-01 6.66793764e-01
-1.90136111e+00 -8.33985806e-01 7.89853156e-01 -6.17443502e-01
7.51179993e-01 7.59282887e-01 7.34886169e-01 1.18239713e+00
1.26962019e-02 5.30471444e-01 6.43303394e-01 -4.01744843e-01
5.83195269e-01 -1.03682391e-01 -9.69042405e-02 3.83546591e-01
-3.82417828e-01 3.23112398e-01 -1.49311256e+00 -2.65718490e-01
7.21657932e-01 -1.27938047e-01 -6.07382953e-01 2.23609865e-01
-1.51131153e+00 8.03472102e-02 1.61870673e-01 3.47684234e-01
-2.84951091e-01 4.26761329e-01 3.93383026e-01 2.60986388e-01
1.69169039e-01 -1.65481016e-01 -1.26883075e-01 -7.74262190e-01
-1.00076604e+00 -1.33022115e-01 5.45423925e-01 8.98470163e-01
3.21840107e-01 3.57756555e-01 -3.48559022e-02 7.08531916e-01
5.65680504e-01 9.22978759e-01 3.40849072e-01 -1.23286259e+00
4.31617379e-01 -7.75496336e-03 2.96806097e-01 -1.27777505e+00
-4.71930295e-01 7.70343319e-02 -4.45955336e-01 3.97055089e-01
4.53488886e-01 -9.69234407e-02 -4.19385374e-01 1.48061335e+00
4.93313253e-01 9.33765948e-01 -6.09174848e-01 1.36602187e+00
9.28768516e-01 8.81234348e-01 5.49070798e-02 -4.02787268e-01
1.58090103e+00 -5.50558865e-01 -1.08066845e+00 1.04380339e-01
3.27115394e-02 -7.41333127e-01 1.30988407e+00 6.32854939e-01
-8.94997001e-01 -9.62512970e-01 -7.78539240e-01 2.88145274e-01
-2.91620195e-01 1.35702357e-01 -4.08356115e-02 4.69099969e-01
-8.09033036e-01 -1.66992560e-01 -1.10810328e+00 -1.90597251e-01
3.14517245e-02 -1.93774432e-01 -3.93893659e-01 4.52536255e-01
-1.25229299e+00 -2.39440538e-02 -2.54593611e-01 1.73624814e-01
-1.47662222e+00 -8.85648906e-01 -8.85031223e-01 -7.69135132e-02
3.22933882e-01 -3.85916457e-02 1.52898121e+00 -5.58147192e-01
-1.24137950e+00 2.57433295e-01 -4.46370929e-01 -1.99944451e-01
1.61483079e-01 -2.88141608e-01 -1.05848908e+00 6.03582442e-01
3.74746472e-01 4.92877960e-01 9.42533493e-01 -9.75127041e-01
-9.28443670e-01 -4.58362699e-02 -3.10044110e-01 5.82095087e-02
-5.21611869e-01 4.25618529e-01 -4.70252097e-01 -7.15525568e-01
1.44667430e-02 -6.27151072e-01 3.34884733e-01 3.67127627e-01
-5.01961887e-01 -3.17797400e-02 8.00376892e-01 -6.98838949e-01
1.65474725e+00 -2.59204936e+00 -4.74057406e-01 7.03629181e-02
-2.17433982e-02 -1.84067503e-01 -1.72567770e-01 3.49958837e-01
-6.92360625e-02 -2.73544371e-01 2.94067383e-01 -3.57482791e-01
-9.07522719e-03 -4.30989712e-01 -6.87893152e-01 4.59827185e-01
-4.48275208e-01 2.87867427e-01 -1.07967079e+00 -6.36904776e-01
2.96509594e-01 5.60725391e-01 -4.55929399e-01 2.18402773e-01
1.98790178e-01 5.22234499e-01 -1.37839004e-01 9.67135429e-01
1.72665000e-01 1.29773140e-01 -6.17403030e-01 -2.21605971e-01
-4.15816128e-01 2.62138158e-01 -1.77977109e+00 1.74115849e+00
-3.84183556e-01 1.23645449e+00 3.27163875e-01 -2.88535237e-01
7.39079416e-01 9.14235890e-01 5.20507455e-01 -7.10386693e-01
-1.80117071e-01 1.02802210e-01 -5.02762377e-01 -8.55983138e-01
4.39240128e-01 3.37483197e-01 -1.63392156e-01 2.58485109e-01
-2.37970114e-01 8.42173994e-02 6.40226156e-02 1.66941807e-01
1.54985547e+00 -2.46568725e-01 1.50179625e-01 2.89713353e-01
4.79449868e-01 -2.60764897e-01 5.27479112e-01 9.31445062e-01
-3.93794924e-01 8.86641383e-01 -9.61201563e-02 -6.80584982e-02
-4.11405414e-01 -1.65546775e+00 -2.62376126e-02 1.43479085e+00
5.46643361e-02 -9.32245612e-01 -5.14390349e-01 -3.45531374e-01
-2.19671309e-01 5.22403419e-01 -3.94469827e-01 2.41646573e-01
-4.91849095e-01 1.19309694e-01 9.44804728e-01 7.23621666e-01
2.27696046e-01 -1.17566466e+00 -7.63272643e-01 2.63018698e-01
-5.29634058e-01 -1.20605350e+00 -1.03123951e+00 -1.37180850e-01
-1.84901521e-01 -1.12812805e+00 -4.94079202e-01 -6.62122488e-01
1.39298141e-01 5.40400863e-01 7.14392543e-01 -3.64804417e-01
-6.55666888e-01 1.26970053e+00 -2.71571457e-01 -5.82474470e-01
-4.01956998e-02 -6.58280611e-01 4.59729135e-01 2.91081131e-01
3.28132778e-01 -8.83350551e-01 -7.62997985e-01 6.46812499e-01
-4.68314290e-01 -2.30780482e-01 -1.97517544e-01 3.07041444e-02
7.25902557e-01 8.06461498e-02 6.57383442e-01 2.69172519e-01
3.90476555e-01 -6.32239103e-01 -4.07704175e-01 -2.31355995e-01
1.10732503e-01 -1.05364645e+00 6.52134836e-01 -8.02692890e-01
-7.46125996e-01 2.64532894e-01 5.69384210e-02 -8.50436687e-01
-7.41774023e-01 1.49570435e-01 -7.14098960e-02 3.73423666e-01
9.47786987e-01 3.03738236e-01 -5.79062045e-01 -5.34900486e-01
-6.28085509e-02 1.24291289e+00 1.08173239e+00 -6.55722320e-02
2.23809138e-01 6.63181603e-01 -1.71399310e-01 -1.19433510e+00
-3.32519352e-01 -9.12978292e-01 -2.17627153e-01 -9.86644626e-01
7.60430753e-01 -1.32273602e+00 -9.91709530e-01 4.41167712e-01
-1.16660535e+00 -2.61918575e-01 -3.02882325e-02 1.04055583e+00
-4.97357517e-01 6.75245002e-02 -5.77285171e-01 -1.19173682e+00
6.37175515e-02 -5.42276144e-01 1.27777421e+00 5.99943399e-02
-5.72871327e-01 -5.51956773e-01 3.00250322e-01 2.22832635e-01
1.14029735e-01 1.29498988e-01 6.05761521e-02 -2.66078770e-01
-4.17407066e-01 -2.66110033e-01 3.91415000e-01 1.09541766e-01
2.81730741e-01 -1.09546036e-01 -1.36233962e+00 8.17387477e-02
-1.50970683e-01 -1.68509595e-02 4.33779806e-01 6.87148511e-01
1.25097406e+00 -1.77368268e-01 -2.51242131e-01 2.13882983e-01
7.82038152e-01 7.45949268e-01 3.01758945e-01 -6.21749721e-02
7.47442186e-01 4.45841074e-01 7.78230429e-01 7.53006518e-01
3.33857596e-01 1.09020174e+00 5.01174271e-01 1.82753056e-01
-4.69150573e-01 -4.54191744e-01 8.77583981e-01 1.05503118e+00
1.07435390e-01 -3.87510359e-01 -9.62190449e-01 7.47156620e-01
-1.48682582e+00 -1.10790539e+00 -5.65931082e-01 2.25124884e+00
7.58461535e-01 -7.22699463e-02 4.31085855e-01 6.52887881e-01
9.99455690e-01 1.64609149e-01 -2.74548262e-01 6.63318411e-02
1.71024382e-01 1.47858122e-02 -5.36064021e-02 6.85823023e-01
-1.03661931e+00 2.79935509e-01 5.79936504e+00 8.58616590e-01
-1.17756295e+00 3.37733656e-01 -1.74003035e-01 -8.56927037e-01
4.46185581e-02 -5.54141223e-01 -6.03221655e-01 9.00244594e-01
1.20297611e+00 1.71123166e-02 4.65111941e-01 5.07829189e-01
1.06080163e+00 -3.66619110e-01 -1.34543133e+00 1.41247857e+00
1.19316466e-01 -1.07661796e+00 -4.93003368e-01 -1.97955638e-01
3.25523950e-02 -7.28441998e-02 4.55563180e-02 -6.04309961e-02
-2.93739349e-01 -7.27015972e-01 1.21765244e+00 6.51552081e-01
9.15745854e-01 -3.97992998e-01 -1.37543261e-01 1.95632547e-01
-1.96802986e+00 -4.90669981e-02 1.50160134e-01 -2.56919533e-01
6.00607395e-01 5.13519347e-01 -7.03810871e-01 -1.70938205e-02
1.42668176e+00 8.10923040e-01 -3.57188582e-01 1.35591590e+00
-3.24943542e-01 1.26686430e+00 -5.87522089e-01 2.18239147e-03
-2.89094478e-01 4.04494852e-01 1.15412283e+00 1.55090761e+00
7.74875760e-01 1.66538253e-01 9.37229395e-02 5.65909863e-01
3.05124611e-01 1.39129153e-02 -7.89188206e-01 1.81676224e-01
1.21680939e+00 1.12326872e+00 -5.57623148e-01 -1.04967028e-01
-3.73216242e-01 5.35057008e-01 -6.38604403e-01 5.77539027e-01
-1.07205880e+00 -7.01480806e-01 7.03522742e-01 2.06632167e-01
6.73486814e-02 -3.89043301e-01 1.23127989e-01 -6.11071169e-01
1.86218202e-01 -5.89280844e-01 5.14429390e-01 -1.45074368e+00
-9.03521180e-01 4.52222735e-01 -6.33162260e-02 -2.03075576e+00
-2.09838554e-01 -2.60713726e-01 -8.27213645e-01 3.30648869e-01
-7.36517251e-01 -4.90116060e-01 -6.06530726e-01 9.60897028e-01
6.54435992e-01 -4.92942035e-02 6.59591019e-01 8.55391681e-01
-4.73273247e-01 4.26434308e-01 -3.64869803e-01 2.01990828e-01
9.87999439e-01 -9.80895638e-01 2.35872686e-01 6.44411564e-01
7.78588951e-01 8.03635046e-02 6.95844233e-01 -5.73224962e-01
-1.33480775e+00 -1.17508554e+00 9.55030560e-01 -4.98732895e-01
7.73840904e-01 -6.06872320e-01 -7.82942891e-01 6.39903724e-01
-1.07009165e-01 5.10776460e-01 8.51375103e-01 -4.16675568e-01
-1.17322594e-01 -2.50938565e-01 -5.50898790e-01 4.30482060e-01
1.07556760e+00 -1.13435435e+00 -5.33903897e-01 1.24164365e-01
4.94084358e-01 -5.98828495e-01 -5.44542372e-01 -1.01673856e-01
6.92921102e-01 -7.36194313e-01 1.19165230e+00 -9.10400227e-02
-7.07464367e-02 -9.37765300e-01 -4.52285081e-01 -1.16365981e+00
1.06708817e-02 -6.49333358e-01 -3.25601459e-01 1.37815058e+00
-5.98822795e-02 -1.90405890e-01 2.05630168e-01 -6.28487095e-02
-1.70258284e-01 6.90606162e-02 -1.46607888e+00 -9.50154722e-01
-9.94202137e-01 -1.42900956e+00 2.19477579e-01 8.79988074e-01
2.92377472e-01 1.28526732e-01 -5.78096569e-01 6.34811461e-01
4.90267903e-01 -2.34233111e-01 5.61593473e-01 -9.73101676e-01
-1.36223122e-01 -5.60150528e-03 -5.64310849e-01 -1.15120864e+00
-9.26940739e-02 -5.28219461e-01 4.61025417e-01 -1.09864914e+00
-4.94907886e-01 -2.14739870e-02 -5.56741714e-01 4.47827041e-01
3.69220108e-01 6.20941162e-01 8.09635893e-02 3.28730117e-03
-9.85244334e-01 3.94903809e-01 7.19120979e-01 -4.04303148e-02
-4.70825493e-01 4.63624895e-02 7.70630240e-02 1.09040058e+00
5.01837373e-01 -7.05046237e-01 -2.92301625e-01 -3.21754932e-01
2.90837467e-01 6.74663007e-01 1.06787944e+00 -1.36319935e+00
9.60880637e-01 -1.57551557e-01 2.84904569e-01 -7.99057782e-01
7.06524611e-01 -9.55853879e-01 3.42122108e-01 2.61413515e-01
-4.86277461e-01 2.41341777e-02 4.57887143e-01 9.44418371e-01
-4.11237538e-01 2.45722890e-01 2.68716544e-01 2.09584013e-01
-8.87780368e-01 1.38808554e-02 -1.08551383e+00 1.55293718e-01
8.34154427e-01 -2.40788400e-01 -3.02871227e-01 -7.16498256e-01
-9.30745721e-01 4.87250686e-02 -3.07010114e-01 7.25357473e-01
8.61284375e-01 -2.02195048e+00 -5.07382631e-01 2.49943525e-01
1.20147109e-01 -3.19394618e-01 5.17566979e-01 8.67448628e-01
-2.21934333e-01 3.05514127e-01 -7.03573525e-02 -1.01255417e+00
-1.59250939e+00 8.82968158e-02 2.43119195e-01 8.78967226e-01
-7.73515046e-01 8.90355647e-01 2.56796867e-01 2.48970881e-01
9.14000511e-01 -4.94553238e-01 -2.34248534e-01 2.02064976e-01
1.07984936e+00 1.11005449e+00 -9.07727703e-02 -5.65528035e-01
-9.21778798e-01 5.61864197e-01 7.17651129e-01 -7.56307721e-01
7.91433156e-01 -6.12189531e-01 3.83342236e-01 1.06635952e+00
1.07606995e+00 5.82757235e-01 -1.19191039e+00 -1.08715072e-01
-2.89608479e-01 -5.85910857e-01 2.16409370e-01 -7.05694795e-01
-8.74897122e-01 1.19672954e+00 9.34016585e-01 3.96756083e-01
1.25606656e+00 3.96638960e-02 8.00324738e-01 2.18966246e-01
4.90118086e-01 -1.04950535e+00 4.00847942e-01 2.76958644e-01
1.04971349e+00 -7.86475897e-01 -6.70320213e-01 -1.51307866e-01
-4.21915889e-01 1.06372726e+00 6.20231628e-01 2.98847705e-01
8.63171458e-01 7.16194689e-01 3.32365632e-01 -5.87194711e-02
-8.25818837e-01 -6.97517842e-02 3.02068263e-01 7.21405506e-01
8.28929394e-02 -6.61517456e-02 5.26317954e-01 1.00339329e+00
-2.72795975e-01 -8.60559717e-02 3.48033935e-01 8.29419315e-01
-4.94707704e-01 -2.04000115e-01 -1.02518880e+00 -1.50813311e-01
-3.83687377e-01 -4.61374186e-02 -2.65361190e-01 3.40599567e-01
1.87399849e-01 1.57820439e+00 4.46247816e-01 -5.96411109e-01
7.32558250e-01 1.14528544e-01 6.28423989e-02 -4.07025576e-01
-5.30463934e-01 4.14670080e-01 -6.58701807e-02 -9.19949234e-01
-2.71201432e-01 -7.72541821e-01 -1.57162726e+00 9.01421905e-02
9.74585712e-02 3.32861185e-01 4.80910391e-01 4.34966356e-01
5.33448935e-01 9.47244287e-01 7.57919014e-01 -1.09454620e+00
1.15741491e-01 -7.84464657e-01 -9.13314998e-01 1.50307715e-01
7.16239750e-01 -4.85706329e-01 -7.43700206e-01 6.03920043e-01]
|
[15.069684028625488, 5.221979141235352]
|
2c52b100-95b3-4d8a-8a1e-db90615f1653
|
responsive-listening-head-generation-a
|
2112.13548
| null |
https://arxiv.org/abs/2112.13548v3
|
https://arxiv.org/pdf/2112.13548v3.pdf
|
Responsive Listening Head Generation: A Benchmark Dataset and Baseline
|
We present a new listening head generation benchmark, for synthesizing responsive feedbacks of a listener (e.g., nod, smile) during a face-to-face conversation. As the indispensable complement to talking heads generation, listening head generation has seldomly been studied in literature. Automatically synthesizing listening behavior that actively responds to a talking head, is critical to applications such as digital human, virtual agents and social robots. In this work, we propose a novel dataset "ViCo", highlighting the listening head generation during a face-to-face conversation. A total number of 92 identities (67 speakers and 76 listeners) are involved in ViCo, featuring 483 clips in a paired "speaking-listening" pattern, where listeners show three listening styles based on their attitudes: positive, neutral, negative. Different from traditional speech-to-gesture or talking-head generation, listening head generation takes as input both the audio and visual signals from the speaker, and gives non-verbal feedbacks (e.g., head motions, facial expressions) in a real-time manner. Our dataset supports a wide range of applications such as human-to-human interaction, video-to-video translation, cross-modal understanding and generation. To encourage further research, we also release a listening head generation baseline, conditioning on different listening attitudes. Code & ViCo dataset: https://project.mhzhou.com/vico.
|
['Tao Mei', 'Tiejun Zhao', 'Ting Yao', 'Wei zhang', 'Yalong Bai', 'Mohan Zhou']
|
2021-12-27
| null | null | null | null |
['talking-head-generation']
|
['computer-vision']
|
[ 2.64597684e-01 4.54010010e-01 1.54333711e-01 -6.05047166e-01
-6.43569767e-01 -6.54840767e-01 8.65605593e-01 -5.30133486e-01
2.41855383e-01 3.56382102e-01 8.05233479e-01 1.89211965e-01
4.72071558e-01 -4.39990610e-01 -4.15938526e-01 -7.11946011e-01
1.27980232e-01 3.39061528e-01 -2.77542800e-01 -5.29372096e-01
-3.38894799e-02 2.06961408e-01 -2.21143818e+00 6.42022789e-01
7.23413453e-02 1.23224223e+00 -2.53772121e-02 1.04693604e+00
1.34263765e-02 9.16851699e-01 -8.03488910e-01 -3.70438933e-01
-5.44460416e-02 -6.77777290e-01 -6.51000440e-01 1.41128361e-01
3.10651004e-01 -1.54068619e-01 1.88885033e-01 9.08806324e-01
1.21597588e+00 2.26321355e-01 5.52079618e-01 -1.81231129e+00
-3.81514847e-01 8.34661067e-01 -2.94365942e-01 -3.42646480e-01
1.70540166e+00 5.27644455e-01 7.43335009e-01 -9.43898022e-01
8.96155000e-01 1.82656705e+00 3.67333502e-01 1.13895476e+00
-9.05332506e-01 -1.06788898e+00 -1.09932020e-01 -5.53785302e-02
-1.10595703e+00 -1.06329000e+00 1.05588090e+00 -5.08052111e-01
5.80991507e-01 7.86257088e-01 6.47355080e-01 1.93722153e+00
-2.02485010e-01 7.45120525e-01 1.05620515e+00 -1.68122962e-01
7.19981939e-02 2.57766217e-01 -3.67046744e-01 -1.52352834e-02
-8.26288998e-01 1.88512415e-01 -1.05716050e+00 -9.03921500e-02
2.35493630e-01 -5.53097963e-01 -5.48249245e-01 2.79444426e-01
-1.39002025e+00 7.87756801e-01 7.30563700e-02 1.04333259e-01
-3.03855032e-01 -1.73945159e-01 5.83981574e-01 5.69799304e-01
4.10902709e-01 3.12579125e-02 2.87104456e-04 -4.73163038e-01
-4.82408375e-01 5.64079463e-01 1.12745607e+00 1.02491701e+00
5.12032509e-01 1.52015060e-01 -5.21246254e-01 8.88601840e-01
2.87253380e-01 7.17585802e-01 5.88451028e-01 -1.20515156e+00
2.09283933e-01 3.40005577e-01 7.16276765e-02 -1.15973866e+00
-5.70610821e-01 1.75572768e-01 -7.56066203e-01 1.36251986e-01
-1.62290305e-01 -6.44561172e-01 -2.61046231e-01 2.14909124e+00
7.12912440e-01 -1.55319762e-03 2.27470011e-01 1.21028972e+00
1.62977934e+00 7.51786768e-01 -3.07228535e-01 -7.99727440e-01
1.50553000e+00 -7.38222003e-01 -1.15379345e+00 -1.69248432e-02
2.84308583e-01 -1.08074033e+00 1.33366442e+00 5.56741416e-01
-1.31287038e+00 -6.23439550e-01 -3.96807373e-01 -2.69762054e-03
-1.61325380e-01 -2.65117168e-01 2.34848410e-01 5.93329251e-01
-1.33694434e+00 -9.62176472e-02 -6.42957240e-02 -5.15445828e-01
-1.25448540e-01 3.02413732e-01 -6.12981021e-01 4.12442982e-01
-1.31300163e+00 3.59334558e-01 -1.67161554e-01 3.82263251e-02
-7.48638093e-01 -4.68080223e-01 -9.41663861e-01 -3.95502239e-01
2.16120154e-01 -5.61914921e-01 1.78074765e+00 -1.39373982e+00
-2.24982953e+00 1.14695823e+00 -4.23518181e-01 -1.54980883e-01
8.63249481e-01 -2.68032491e-01 -7.14111924e-01 9.77290943e-02
2.72877961e-02 1.11211514e+00 1.13103950e+00 -1.31377947e+00
-1.57568648e-01 -2.94621676e-01 -4.34595533e-02 4.29100037e-01
-1.99697405e-01 7.73544550e-01 -2.02643856e-01 -6.91347361e-01
-2.79432982e-01 -1.15887737e+00 3.49922866e-01 -2.74814665e-01
-7.58300602e-01 -2.55947709e-01 1.19644308e+00 -4.43065524e-01
9.96691465e-01 -2.38400483e+00 4.27679628e-01 -7.03945814e-04
1.29429832e-01 -1.05743691e-01 9.79221333e-03 5.76590538e-01
-3.36169988e-01 -1.06324635e-01 4.51142117e-02 -7.16517448e-01
2.75876254e-01 -2.45127693e-01 -2.75976419e-01 3.51026803e-01
-2.12380588e-01 7.64743865e-01 -8.36129665e-01 -6.28305674e-01
2.56218091e-02 6.38134539e-01 -7.75310338e-01 8.08698654e-01
-6.73759133e-02 9.06813383e-01 -1.50307596e-01 6.14494562e-01
5.55430055e-01 3.14540267e-01 -8.74464288e-02 -1.71100199e-01
-1.57451794e-01 6.94660544e-02 -1.11054993e+00 1.40850854e+00
-4.90897268e-01 8.25689256e-01 5.23903430e-01 -3.03429872e-01
1.28717029e+00 8.53033841e-01 2.59696782e-01 -5.16669273e-01
5.61306119e-01 -7.11034387e-02 4.01440971e-02 -8.90743971e-01
4.39662308e-01 -2.78376609e-01 -4.10134315e-01 5.67194819e-01
-8.86422843e-02 -5.78368723e-01 3.75230387e-02 -3.61051261e-02
6.97235525e-01 -9.71478671e-02 2.45345935e-01 1.56958893e-01
6.26497209e-01 -6.66987956e-01 1.44893229e-01 1.67463928e-01
-2.43088543e-01 9.22996998e-01 5.82250059e-01 1.52470604e-01
-2.59170473e-01 -7.00051904e-01 3.97836000e-01 1.85590529e+00
-2.05926344e-01 -2.80826330e-01 -1.23158622e+00 1.10531978e-01
-3.68704945e-01 6.28448784e-01 -7.29494452e-01 -3.04970473e-01
-3.53678256e-01 -1.25198349e-01 8.28710675e-01 1.81599379e-01
2.89048523e-01 -1.95998406e+00 -6.17375910e-01 7.21616820e-02
-6.54620945e-01 -1.23272622e+00 -8.92214179e-01 -3.59109700e-01
-2.49159727e-02 -5.45342803e-01 -7.66259849e-01 -7.94777155e-01
2.56523401e-01 3.04082427e-02 1.11260283e+00 -3.87220383e-01
-1.31627068e-01 7.09839284e-01 -6.56537056e-01 -4.85916913e-01
-7.10542083e-01 -2.77026623e-01 5.18903852e-01 6.76942110e-01
2.53652092e-02 -7.17963219e-01 -5.49782455e-01 7.15337813e-01
-6.57786608e-01 1.70318246e-01 1.27521858e-01 4.25615638e-01
-1.22938551e-01 -7.87458599e-01 7.48241961e-01 -6.60829246e-01
1.23620462e+00 -6.28255367e-01 1.40143812e-01 -2.38799587e-01
3.04426402e-01 -5.37138224e-01 5.11268198e-01 -1.04011190e+00
-1.21840298e+00 9.29598808e-02 -3.46454233e-01 -6.82967663e-01
-4.67164367e-01 2.97090948e-01 -5.73948920e-01 2.54009247e-01
7.13457525e-01 5.84470527e-03 2.34993622e-01 -1.32439792e-01
5.57560623e-01 1.19039571e+00 9.91570652e-01 -4.38877225e-01
5.28688669e-01 1.79973289e-01 -5.98938584e-01 -1.01072311e+00
-2.61410177e-01 -3.98930043e-01 -1.41585216e-01 -9.84637737e-01
9.29494560e-01 -6.94783568e-01 -1.52758944e+00 7.64960349e-01
-1.38878179e+00 -5.24055004e-01 3.75174172e-02 3.09952438e-01
-8.96381557e-01 -1.06489278e-01 -5.04030883e-01 -9.56406713e-01
-6.13028646e-01 -1.12771726e+00 1.37507641e+00 2.55447000e-01
-9.19808507e-01 -5.22526503e-01 1.61205634e-01 6.29547000e-01
4.75960732e-01 3.53999943e-01 2.49742955e-01 -4.61713374e-01
3.05455238e-01 -8.56936201e-02 2.93175995e-01 6.99599460e-02
2.60548800e-01 3.25169593e-01 -1.44211555e+00 -5.35228737e-02
4.79320027e-02 -7.23070264e-01 8.24232325e-02 1.51852384e-01
7.99864948e-01 -7.63320863e-01 1.63677335e-01 4.13695186e-01
2.43889391e-01 3.60014319e-01 7.19269395e-01 -4.31298405e-01
4.89084005e-01 1.28256774e+00 6.72218084e-01 8.60704839e-01
4.12146062e-01 8.58566880e-01 5.77251852e-01 1.35739762e-02
-7.59184211e-02 -3.68890136e-01 1.08270121e+00 7.91865647e-01
-1.76016808e-01 -5.18434167e-01 -6.22316122e-01 4.40702476e-02
-1.48878217e+00 -1.27139914e+00 -8.14649016e-02 1.93096495e+00
8.63545597e-01 -2.32352644e-01 5.23322284e-01 2.84799606e-01
9.46266115e-01 3.67819220e-01 -5.51616192e-01 -7.86378324e-01
-2.41091147e-01 -2.34671347e-02 -3.53128046e-01 7.57457614e-01
-7.60568440e-01 9.98258114e-01 5.34194136e+00 5.90426087e-01
-1.78670442e+00 3.96238118e-02 5.40908992e-01 -3.24946612e-01
-3.38559091e-01 -5.56519985e-01 -6.54391944e-01 3.28022629e-01
7.87371099e-01 -3.06464851e-01 6.18971825e-01 6.74065888e-01
8.20740938e-01 2.36311540e-01 -1.34023666e+00 1.41984379e+00
5.94408870e-01 -8.60795498e-01 -3.70841995e-02 -1.84012294e-01
2.98292011e-01 -3.18940133e-01 4.05963331e-01 3.13099146e-01
2.73492709e-02 -1.15561044e+00 1.20476913e+00 3.50653112e-01
1.11183894e+00 -3.58907968e-01 4.53379273e-01 2.29927167e-01
-1.23505592e+00 8.29575583e-02 5.81710875e-01 -1.70034945e-01
5.36528111e-01 -1.81064475e-02 -8.95471215e-01 1.13250658e-01
8.02475035e-01 4.88625735e-01 -2.22550347e-01 2.51246661e-01
-3.27065662e-02 5.55755615e-01 -2.03291461e-01 -2.67958611e-01
-2.76074946e-01 1.07257245e-02 8.96743834e-01 1.30613005e+00
3.45943928e-01 2.66148120e-01 -1.82718690e-02 6.55644953e-01
-1.55324951e-01 3.80833328e-01 -8.74385297e-01 2.61163898e-02
4.52569127e-01 1.32173944e+00 -3.94726783e-01 -1.62515253e-01
-5.52997515e-02 1.00161684e+00 -3.43681604e-01 4.22674626e-01
-9.38554883e-01 -1.83126643e-01 5.62679052e-01 2.78732449e-01
-3.29281420e-01 2.58237213e-01 2.87453473e-01 -7.21033514e-01
6.85667172e-02 -1.44426811e+00 -1.29970293e-02 -1.40842474e+00
-9.18070436e-01 9.48797107e-01 7.24092172e-03 -1.55032074e+00
-8.69029939e-01 -1.79332227e-01 -9.40174997e-01 5.55204511e-01
-6.78215206e-01 -1.07669151e+00 -7.33251452e-01 9.42815125e-01
7.37514377e-01 -3.94911356e-02 7.52908945e-01 3.55207711e-01
-3.97676378e-01 7.70166636e-01 -9.65997100e-01 -3.37151140e-02
1.10659945e+00 -6.35174572e-01 1.04811840e-01 -3.45154516e-02
-2.66764492e-01 5.30851841e-01 1.21059835e+00 -1.23225637e-01
-1.52442455e+00 -8.89701605e-01 7.00073361e-01 -1.61822274e-01
5.07437110e-01 -6.65163696e-01 -5.87856174e-01 6.10439718e-01
5.36365032e-01 -1.35292172e-01 7.77602613e-01 -2.62799561e-01
-1.55872688e-01 -8.20762515e-02 -1.24792767e+00 7.96959221e-01
1.19702566e+00 -5.63888431e-01 -4.20542210e-01 2.62605399e-01
9.98470426e-01 -5.71570456e-01 -5.67558050e-01 2.62141228e-01
7.92893529e-01 -1.38756073e+00 7.16062367e-01 -3.07575971e-01
3.30350190e-01 -1.42199576e-01 -2.61219233e-01 -1.29121864e+00
3.71600062e-01 -1.60746765e+00 2.60255754e-01 1.61767077e+00
2.98779130e-01 -6.81696594e-01 4.38981920e-01 2.91083485e-01
-2.69695073e-01 -1.85757220e-01 -8.25489938e-01 -2.54487753e-01
-4.48761463e-01 -7.23605931e-01 7.71459937e-01 9.92367923e-01
3.24250609e-01 6.67018831e-01 -7.38560855e-01 -1.91932723e-01
-3.14862616e-02 -1.39312744e-01 1.45875657e+00 -1.23468661e+00
-2.10799918e-01 -6.49498403e-01 -2.30852127e-01 -9.09187317e-01
4.90619302e-01 -6.59715116e-01 2.38367423e-01 -8.89900029e-01
-5.19222505e-02 2.52128422e-01 5.37937105e-01 3.41315538e-01
2.61211842e-01 3.89536202e-01 3.43897730e-01 -7.28511736e-02
-3.75080109e-01 7.18568683e-01 1.38123345e+00 -4.78171147e-02
-4.37940598e-01 2.01558441e-01 -5.69890082e-01 9.07729924e-01
7.16345608e-01 -9.42718685e-02 -3.81610304e-01 5.78805320e-02
3.25541705e-01 7.90668547e-01 3.31818223e-01 -7.60847390e-01
2.17984572e-01 -1.55951053e-01 -3.84371281e-01 -3.96417290e-01
9.41692293e-01 -3.75237107e-01 4.45482373e-01 1.19832665e-01
-4.17175651e-01 2.29664922e-01 -9.61263478e-02 1.67281497e-02
-4.69836324e-01 2.27234036e-01 6.53436005e-01 7.79922446e-03
-2.85375714e-01 1.67950243e-01 -4.86024708e-01 2.21138805e-01
9.33873296e-01 -2.90627986e-01 -3.46375018e-01 -1.40882671e+00
-9.20812368e-01 1.01382263e-01 2.54786193e-01 8.29057157e-01
6.48686886e-01 -1.44711041e+00 -8.60504627e-01 2.91619509e-01
3.07918608e-01 -1.06165133e-01 4.68752235e-01 8.79319429e-01
-2.03675926e-01 1.97682623e-03 -2.00606823e-01 -8.23177934e-01
-1.67937946e+00 2.60835648e-01 2.33907774e-01 5.26046932e-01
-2.48995319e-01 9.71385300e-01 4.63328063e-01 -5.61557472e-01
4.72032487e-01 -2.32327446e-01 -5.11879504e-01 6.49356544e-01
5.80181360e-01 4.98713702e-01 -2.22748756e-01 -1.42939723e+00
-2.53051847e-01 6.75568402e-01 5.09560168e-01 -6.27056360e-01
7.75489748e-01 -1.78452954e-01 -2.19566803e-02 8.55710089e-01
1.32731211e+00 2.86122203e-01 -6.69184387e-01 1.93275079e-01
-5.69095433e-01 -2.58653909e-01 -3.93876910e-01 -4.35068429e-01
-1.02363074e+00 1.00679874e+00 5.14424860e-01 3.62147480e-01
1.10016847e+00 2.82081306e-01 7.40944982e-01 1.48786068e-01
2.66447425e-01 -9.67013001e-01 4.98901367e-01 3.89293104e-01
1.71187222e+00 -1.33971894e+00 -7.03320444e-01 -4.61543560e-01
-1.14740467e+00 7.21568286e-01 6.48526669e-01 5.05314171e-01
4.72589225e-01 4.43356484e-01 6.48646235e-01 6.21159561e-02
-1.11044502e+00 -9.86164436e-02 1.28591031e-01 8.35670352e-01
6.45056486e-01 4.54973370e-01 -2.39256751e-02 7.80957043e-01
-1.24492288e+00 -3.64519134e-02 5.06668448e-01 4.73243475e-01
-1.74009323e-01 -5.62763929e-01 -8.73583853e-01 -1.28217310e-01
-2.76589662e-01 1.03072263e-01 -9.45090711e-01 4.18883085e-01
4.01561335e-02 1.58395183e+00 3.45063619e-02 -7.53492832e-01
5.45196712e-01 4.74743731e-02 1.15118980e-01 -5.35961926e-01
-9.10515547e-01 3.03147554e-01 2.39546850e-01 -6.40837193e-01
-6.87980294e-01 -7.21834540e-01 -1.25291348e+00 -5.00014007e-01
1.16014346e-01 -2.09266506e-02 4.19872671e-01 6.50178909e-01
2.95790702e-01 2.94821739e-01 1.02703810e+00 -1.76571047e+00
-5.87729514e-02 -1.45494187e+00 -2.65454352e-01 7.06045866e-01
4.53802913e-01 -7.08792865e-01 -8.32296193e-01 2.53344893e-01]
|
[13.227076530456543, -0.40146520733833313]
|
e46a3c43-28d7-4755-bd94-dbb2033f0034
|
san-francisco-crime-classification
|
1607.03626
| null |
http://arxiv.org/abs/1607.03626v1
|
http://arxiv.org/pdf/1607.03626v1.pdf
|
San Francisco Crime Classification
|
San Francisco Crime Classification is an online competition administered by
Kaggle Inc. The competition aims at predicting the future crimes based on a
given set of geographical and time-based features. In this paper, I achieved a
an accuracy that ranks at top %18, as of May 19th, 2016. I will explore the
data, and explain in details the tools I used to achieve that result.
|
['Yehya Abouelnaga']
|
2016-07-13
| null | null | null | null |
['crime-prediction']
|
['miscellaneous']
|
[-5.44228137e-01 -3.98592263e-01 -1.99303895e-01 -6.41683936e-01
-6.90784037e-01 -3.94886672e-01 7.66093671e-01 6.72843635e-01
-8.07566881e-01 7.72213221e-01 3.76717567e-01 -4.16177362e-01
-4.43248689e-01 -8.77434134e-01 -1.76865578e-01 2.31303647e-02
-5.50341547e-01 1.51153818e-01 2.77884249e-02 -5.14272928e-01
9.27964389e-01 5.37979722e-01 -1.37846959e+00 6.85153186e-01
6.72439098e-01 6.17695868e-01 -2.70507872e-01 6.45981789e-01
1.78525537e-01 1.10804904e+00 -9.37108517e-01 -1.02704442e+00
9.91647765e-02 8.43844116e-02 -1.15817320e+00 -6.74697518e-01
2.93334454e-01 -4.10829335e-01 -7.24774957e-01 6.87187731e-01
-5.97344749e-02 2.40985677e-01 8.53394508e-01 -1.24364364e+00
-3.77947390e-01 4.03628081e-01 -3.38527828e-01 1.24779379e+00
7.24620283e-01 -1.00423858e-01 8.73949051e-01 -9.46412444e-01
8.30990553e-01 6.12901390e-01 9.44534659e-01 3.68718266e-01
-6.75494075e-01 -6.95333481e-01 1.02247141e-01 8.64865541e-01
-1.59452081e+00 -1.85184643e-01 6.54446423e-01 -8.00225914e-01
1.55358434e+00 3.60198289e-01 6.78771913e-01 9.94037747e-01
4.28104252e-01 9.19575453e-01 1.05346811e+00 -2.49945283e-01
1.84869647e-01 -1.88348636e-01 6.62827313e-01 4.28471893e-01
4.56945971e-03 1.82123706e-01 -7.49373138e-01 -5.22811770e-01
2.88826823e-01 -3.78288925e-02 4.38280076e-01 4.52452302e-01
-4.65322703e-01 1.35215724e+00 3.52820963e-01 4.04888004e-01
-1.47595331e-01 -2.40707114e-01 7.02111304e-01 1.36986256e-01
1.00097120e+00 5.24154842e-01 -3.14326257e-01 -8.06019902e-01
-1.08879173e+00 7.94109225e-01 3.56337875e-01 2.62389123e-01
3.87078047e-01 -3.18325102e-01 1.59748137e-01 9.28080082e-01
-1.26865223e-01 1.25322670e-01 1.88350812e-01 -8.76140654e-01
1.01623487e+00 4.88067448e-01 6.91452771e-02 -1.44174290e+00
-8.44351411e-01 2.59560999e-03 -6.16945028e-01 4.63793352e-02
7.37689912e-01 -3.56617033e-01 -7.32521534e-01 1.13405883e+00
-9.62061435e-02 1.38436511e-01 -3.87445390e-01 6.33181095e-01
6.28458560e-01 5.05301952e-01 2.16691241e-01 3.05823535e-01
6.61999106e-01 -3.63217413e-01 -4.14343566e-01 -1.92906946e-01
8.43192518e-01 -3.89059663e-01 3.02903593e-01 6.71819687e-01
-9.15261567e-01 -3.34247649e-01 -8.93285275e-01 1.55300334e-01
-9.34537172e-01 -1.46134108e-01 1.07905912e+00 9.64907706e-01
-8.02850962e-01 9.05904889e-01 -1.04940617e+00 -3.90439361e-01
6.64445937e-01 7.26585835e-02 -5.38179874e-01 1.18824162e-01
-1.38224351e+00 1.18976200e+00 1.53105676e-01 -1.72862679e-01
-3.69175106e-01 -5.58429182e-01 -8.99712741e-01 -4.03459430e-01
-4.35502082e-01 2.22956568e-01 6.51563525e-01 -1.10171966e-01
-5.98595917e-01 1.26663232e+00 -1.60099655e-01 -7.90231228e-01
6.55044913e-01 -2.52600640e-01 -6.98189557e-01 1.57973859e-02
8.11629295e-01 1.19171977e-01 -5.66068739e-02 -4.85024542e-01
-1.27227855e+00 -8.27862561e-01 2.05999166e-01 -2.37081908e-02
-5.34148812e-01 7.78641105e-01 -7.66485110e-02 -5.99561512e-01
-1.42586734e-02 -4.23626691e-01 -3.54716748e-01 -1.02593613e+00
-2.27550030e-01 -5.83538592e-01 2.03518718e-01 -1.07063913e+00
1.61221468e+00 -1.82755041e+00 -3.88517290e-01 4.74452734e-01
6.20216131e-02 2.18676060e-01 2.57083476e-01 6.79452300e-01
-2.75614977e-01 4.18803453e-01 -1.03332423e-01 -4.47972476e-01
-2.80927688e-01 -3.52212459e-01 -4.08010781e-01 7.18425214e-01
2.12197050e-01 4.64620531e-01 -9.32765603e-01 -2.94311285e-01
4.14857924e-01 1.71123460e-01 -5.96429348e-01 -2.62661159e-01
8.41745436e-01 1.65179342e-01 -3.07507128e-01 5.42004645e-01
6.64440274e-01 3.55898082e-01 -1.86200738e-01 8.57767999e-01
-4.12080646e-01 5.13042450e-01 -6.98584139e-01 1.52540040e+00
-8.47874880e-02 9.67936575e-01 -3.50003809e-01 -1.11701226e+00
1.03538442e+00 1.90154999e-04 6.88142478e-01 -1.02526307e+00
-1.95695177e-01 1.45540312e-01 -3.61089498e-01 -2.15360045e-01
7.18527138e-01 3.36951673e-01 -5.35801649e-01 3.66841465e-01
-2.22689152e-01 7.74220675e-02 4.07532781e-01 2.25464553e-01
1.62616551e+00 -1.14442058e-01 8.42918977e-02 -3.33010197e-01
2.88979799e-01 3.51192504e-01 6.27479911e-01 1.03715694e+00
-4.76380140e-01 8.33483100e-01 5.64931214e-01 -1.18864632e+00
-8.84238839e-01 -9.03291762e-01 -5.12961268e-01 7.46083140e-01
-3.39858234e-01 -5.87250769e-01 -8.01551521e-01 -5.01056612e-01
-2.14368746e-01 6.83044255e-01 -1.03500009e+00 1.54539466e-01
-6.84415877e-01 -8.64246607e-01 8.86700273e-01 4.79274124e-01
4.12958324e-01 -1.07242632e+00 -7.82979429e-01 1.36263505e-01
-3.02009732e-01 -7.39251137e-01 -4.03208397e-02 4.20840597e-03
-6.19408071e-01 -1.13409758e+00 -3.18890691e-01 -3.99968714e-01
1.66809723e-01 -3.59537564e-02 9.81223226e-01 3.04595441e-01
-6.21921003e-01 -1.39616668e-01 -6.87367916e-01 -7.09796131e-01
3.16166401e-01 5.11008441e-01 -1.76531188e-02 -1.06752232e-01
9.61937785e-01 -4.31742787e-01 -3.20445746e-01 -2.60897964e-01
-3.81651759e-01 -7.61991963e-02 -4.92456615e-01 7.12684095e-01
-2.42332786e-01 4.06123161e-01 3.89866143e-01 -7.43226171e-01
6.59635186e-01 -9.02673483e-01 -3.94694448e-01 -6.00336790e-02
-3.71138006e-01 -7.31493831e-01 3.51228774e-01 3.86754811e-01
-6.94202304e-01 5.73547371e-02 -2.10442007e-01 1.00728951e-01
-3.95556897e-01 6.22224569e-01 4.50850815e-01 4.49696481e-01
8.51936817e-01 -6.08590692e-02 -6.51644886e-01 -3.19262087e-01
-2.05854088e-01 8.78711283e-01 5.76630116e-01 -5.70636392e-01
4.86306936e-01 6.42801285e-01 -1.28046274e-01 -6.68299139e-01
-7.87673652e-01 -5.04958451e-01 -9.94129598e-01 -3.72562706e-01
9.48240697e-01 -6.19615614e-01 -7.93883741e-01 1.03909326e+00
-1.39195764e+00 -3.17534119e-01 3.28365594e-01 5.26474833e-01
-2.67116308e-01 -1.61600590e-01 -8.39255035e-01 -1.23627627e+00
-3.77682690e-03 -5.56000531e-01 4.42279100e-01 2.85137594e-01
-5.80260217e-01 -1.18489099e+00 4.04962748e-01 8.56670618e-01
2.27728426e-01 6.02550924e-01 4.01868761e-01 -6.95195258e-01
2.73855031e-01 -5.92895865e-01 -1.72308326e-01 -2.12330222e-02
2.23989692e-03 2.64838427e-01 -1.03915584e+00 6.19811676e-02
-3.67419571e-01 -7.52857327e-02 9.86462653e-01 5.90091825e-01
1.34800696e+00 -6.48571923e-03 -2.95286119e-01 4.16711628e-01
1.26331890e+00 4.47182536e-01 7.44034946e-01 1.04430544e+00
4.58836228e-01 9.66870546e-01 9.16694999e-01 7.69945383e-01
7.83774555e-01 7.10948229e-01 1.97760776e-01 1.52758285e-01
4.69440460e-01 -3.78386617e-01 -2.12974893e-03 2.34483294e-02
-5.65721095e-01 2.63080835e-01 -1.81412125e+00 1.10877275e+00
-2.03728127e+00 -1.34526443e+00 -4.45943713e-01 1.85346818e+00
4.44523841e-01 5.80371797e-01 4.60743368e-01 5.09408832e-01
7.57253766e-01 2.81762704e-03 5.84730059e-02 -8.15328717e-01
-2.26610944e-01 6.72046125e-01 6.69957399e-01 5.94111800e-01
-1.45146382e+00 1.20423627e+00 7.69530058e+00 1.16755581e+00
-7.38482416e-01 1.34546384e-01 1.11116481e+00 -4.18551356e-01
1.85043871e-01 2.20980290e-02 -7.48597443e-01 5.74877143e-01
1.06389916e+00 -2.21411794e-01 4.05052096e-01 6.01802111e-01
3.40231001e-01 -3.86062980e-01 -6.21937394e-01 9.82780933e-01
1.08350351e-01 -1.80380809e+00 -3.24299514e-01 3.30449104e-01
5.29854000e-01 2.62268305e-01 1.43745959e-01 3.50457907e-01
3.69603723e-01 -1.60409582e+00 8.38643670e-01 8.38273525e-01
2.29381353e-01 -1.53271604e+00 7.51972973e-01 5.48457563e-01
-8.82181823e-01 -5.58398902e-01 -1.70171142e-01 -9.76904333e-01
3.60235746e-04 3.23840141e-01 -4.40222234e-01 6.98441207e-01
1.17276251e+00 1.04125905e+00 -6.88205838e-01 9.74851251e-01
-9.69875827e-02 7.41400540e-01 -1.54860392e-01 1.20899202e-02
5.36101937e-01 5.67891449e-03 1.40487850e-01 1.32607329e+00
9.64481160e-02 6.38792455e-01 -2.05879007e-03 4.06867594e-01
4.88694370e-01 2.49570161e-02 -9.26301837e-01 4.97629613e-01
4.80743408e-01 8.52415740e-01 -4.57401693e-01 -3.92213054e-02
-4.28561091e-01 5.98666668e-01 6.78943515e-01 -1.17267504e-01
-7.69426703e-01 -7.14927554e-01 7.20675528e-01 2.96781898e-01
-2.52517968e-01 -3.30036253e-01 -9.00544524e-01 -9.75358009e-01
2.79410444e-02 -3.90960217e-01 7.05529273e-01 -6.30790472e-01
-1.33726203e+00 6.20320618e-01 3.93789530e-01 -1.02130699e+00
-2.61538267e-01 -5.54229677e-01 -1.08366466e+00 8.65405798e-01
-9.72894728e-01 -9.45924819e-01 2.46966422e-01 6.47498131e-01
4.39155012e-01 -6.22153580e-01 8.83968472e-01 5.64570665e-01
-7.29341269e-01 6.57397687e-01 3.11709605e-02 8.06918025e-01
2.65499860e-01 -1.05934024e+00 8.44711304e-01 7.49489486e-01
1.50952324e-01 5.58671474e-01 6.29134357e-01 -6.75067663e-01
-3.68577093e-01 -7.63747334e-01 1.73097384e+00 -1.00666320e+00
8.73853803e-01 -1.53829515e-01 -4.04983610e-01 5.30118287e-01
-9.94527861e-02 -4.08201247e-01 7.53567159e-01 5.12007475e-01
-2.06623331e-01 1.15758143e-01 -1.31012130e+00 2.83939630e-01
9.66400981e-01 -6.94362879e-01 -6.29327118e-01 4.19327259e-01
-2.27667168e-01 -5.58159530e-01 -9.59522486e-01 3.30769233e-02
3.60758513e-01 -1.26982057e+00 8.02320600e-01 -1.07653165e+00
8.56376529e-01 3.53004634e-01 -1.16985202e-01 -1.15524697e+00
-4.82415646e-01 -2.16233566e-01 4.75215644e-01 1.07395315e+00
6.69825435e-01 -5.10765254e-01 1.18095303e+00 1.04774630e+00
7.44248042e-03 -6.64018512e-01 -1.49150789e+00 -8.55145693e-01
8.40389490e-01 -1.30907679e+00 8.81628513e-01 1.21102250e+00
8.29640329e-01 -2.82189369e-01 -3.35162759e-01 -7.41515756e-02
6.72442138e-01 -4.90854234e-01 7.12332308e-01 -1.05680537e+00
4.46160108e-01 -6.61694467e-01 -1.00408101e+00 -2.35925511e-01
2.20655859e-01 -7.39603341e-01 -7.58527339e-01 -1.23178577e+00
3.46622854e-01 -6.46230042e-01 -3.36767226e-01 7.32274473e-01
-5.76217212e-02 5.62124372e-01 1.50314197e-01 5.48853576e-02
-4.57324475e-01 -4.48199250e-02 2.25056231e-01 -2.88445413e-01
-8.28294009e-02 5.22950478e-02 -5.94692290e-01 8.62501919e-01
1.18027449e+00 -4.87576991e-01 1.16577866e-02 -6.32835865e-01
3.05113912e-01 -3.83621752e-02 4.09402460e-01 -1.19262707e+00
1.10897101e-01 -6.48784399e-01 5.74824095e-01 -7.59834647e-01
5.57882249e-01 -2.78830349e-01 -4.47275974e-02 4.22837615e-01
-2.42319703e-01 1.56125262e-01 3.43662322e-01 2.50242669e-02
-4.95358884e-01 -1.79962575e-01 2.33525708e-01 2.35710740e-02
-5.47774076e-01 3.32428455e-01 -6.30456030e-01 -9.79386568e-02
1.11898327e+00 -4.42995369e-01 -3.89772147e-01 -4.21105415e-01
-9.72026050e-01 3.34633142e-01 9.12831426e-02 6.09272599e-01
6.57356918e-01 -1.32949543e+00 -1.16334331e+00 -9.28021297e-02
1.98639512e-01 -7.62325227e-01 4.18268472e-01 5.40678501e-01
-6.87042892e-01 6.24786377e-01 -5.78509212e-01 -9.68711227e-02
-1.25217223e+00 -2.17021387e-02 1.51131898e-01 -7.04506695e-01
-5.54466248e-01 7.29655147e-01 -5.73250473e-01 -4.90257353e-01
-2.74825990e-01 3.16240698e-01 -9.61451590e-01 4.55286838e-02
9.40154493e-01 9.12532568e-01 3.36132795e-01 -9.00746584e-01
-8.53084445e-01 2.24453673e-01 1.16678579e-02 -5.78052461e-01
1.93722570e+00 1.77678138e-01 2.34181285e-02 3.11639905e-01
1.21465576e+00 -4.05574560e-01 -8.12971115e-01 9.77330878e-02
4.04505789e-01 -9.94788289e-01 -5.00864163e-02 -5.13744116e-01
-9.80361640e-01 7.56043911e-01 4.63144273e-01 5.50999701e-01
7.64026403e-01 -3.49194676e-01 6.94409966e-01 7.35057071e-02
8.36719871e-01 -1.72018957e+00 -4.60710347e-01 1.02624631e+00
6.63515747e-01 -1.16639197e+00 1.26940131e-01 -7.93139637e-02
-6.02256954e-01 1.19674110e+00 4.27420855e-01 -6.40410542e-01
1.12855101e+00 -7.92835876e-02 -8.06189999e-02 -5.73444009e-01
-5.98983169e-01 1.19489484e-01 8.77402425e-02 1.05633998e+00
4.36950296e-01 5.13232708e-01 -5.60300112e-01 7.33867049e-01
-7.64441192e-01 -1.49497017e-01 5.06356716e-01 9.47202504e-01
-2.79546082e-01 -9.91287529e-01 -5.91788292e-01 8.78388107e-01
-8.43983471e-01 1.06135448e-02 -4.47530568e-01 6.65695071e-01
3.42670351e-01 1.75430107e+00 3.37209105e-01 -8.11198175e-01
3.89822274e-01 -3.45056176e-01 1.70565799e-01 -6.33318424e-01
-9.61944222e-01 -8.12457979e-01 3.15864325e-01 -7.10600376e-01
-4.42651920e-02 -1.30802441e+00 -1.05046105e+00 -1.03421700e+00
3.65064256e-02 2.91752517e-01 6.24400973e-01 1.06626630e+00
-1.37682661e-01 -1.29603893e-01 1.02058697e+00 -7.32266366e-01
1.01816602e-01 -9.78172421e-01 -8.15514863e-01 3.83859247e-01
-1.08257839e-02 -5.34960032e-01 -4.06722456e-01 -3.86940181e-01]
|
[6.721374034881592, 1.9105924367904663]
|
d2d8fd3b-12af-462d-a684-210409639262
|
expectile-quadrangle-and-applications
|
2306.16351
| null |
https://arxiv.org/abs/2306.16351v1
|
https://arxiv.org/pdf/2306.16351v1.pdf
|
Expectile Quadrangle and Applications
|
The paper explores the concept of the \emph{expectile risk measure} within the framework of the Fundamental Risk Quadrangle (FRQ) theory. According to the FRQ theory, a quadrangle comprises four stochastic functions associated with a random variable: ``error'', ``regret'', ``risk'', and ``deviation''. These functions are interconnected through a stochastic function known as the ``statistic''. Expectile is a risk measure that, similar to VaR (quantile) and CVaR (superquantile), can be employed in risk management. While quadrangles based on VaR and CVaR statistics are well-established and widely used, the paper focuses on the recently proposed quadrangles based on expectile. The aim of this paper is to rigorously examine the properties of these Expectile Quadrangles, with particular emphasis on a quadrangle that encompasses expectile as both a statistic and a measure of risk.
|
['Stan Uryasev', 'Anton Malandii', 'Viktor Kuzmenko']
|
2023-06-28
| null | null | null | null |
['management']
|
['miscellaneous']
|
[-1.75318643e-01 3.73004407e-01 1.27547964e-01 -1.40416443e-01
-1.00662768e+00 -6.41349971e-01 5.04625738e-01 6.88186467e-01
-3.76244396e-01 8.43602598e-01 -9.44885090e-02 -4.33408171e-01
-7.21268475e-01 -1.01611280e+00 -3.60310286e-01 -6.84617400e-01
-4.65634733e-01 7.63752982e-02 -9.74248052e-02 -1.62043408e-01
5.17705441e-01 4.77296382e-01 -1.34391582e+00 -4.39639091e-01
1.13772202e+00 1.58176363e+00 -4.16632295e-01 3.21064174e-01
-1.33252963e-02 4.11478460e-01 -9.39319968e-01 -9.06987607e-01
4.32638913e-01 -5.15991569e-01 -5.46202898e-01 -7.75759757e-01
-6.38362393e-02 1.77309871e-01 5.39033353e-01 1.28842521e+00
1.22037619e-01 7.16585964e-02 1.18275976e+00 -1.22776437e+00
-3.03591996e-01 3.66282880e-01 -1.77212864e-01 2.48177424e-01
2.28346467e-01 -1.69209510e-01 1.34734702e+00 -7.58146703e-01
1.41175762e-01 1.00894392e+00 7.30636418e-01 1.04852736e-01
-1.18061388e+00 -3.85546893e-01 -1.66849881e-01 -2.46853948e-01
-1.31574154e+00 2.91918099e-01 3.86971474e-01 -8.85506213e-01
3.61048758e-01 7.78152466e-01 5.34446418e-01 5.37942111e-01
1.34208882e+00 5.56796193e-01 1.20554471e+00 -2.55439937e-01
6.09638095e-01 1.90092087e-01 -9.65036228e-02 2.38858849e-01
4.91095811e-01 6.49780273e-01 -1.36179879e-01 -4.05560732e-01
6.32007301e-01 -6.55292720e-02 -6.84556589e-02 -4.37867671e-01
-6.00639045e-01 1.11769807e+00 1.02853484e-01 1.09302402e-01
-5.88430703e-01 1.83001742e-01 4.54349995e-01 3.03458929e-01
8.00579011e-01 5.92289746e-01 -1.41717985e-01 -2.37986267e-01
-6.69811904e-01 5.22046745e-01 9.79435384e-01 5.14532208e-01
3.46638352e-01 1.16311118e-01 -4.51675713e-01 4.71975088e-01
4.88265514e-01 8.58866990e-01 -1.97577789e-01 -5.93892753e-01
5.26430368e-01 3.56115699e-01 2.49738216e-01 -1.04365063e+00
-4.29622680e-01 -9.36110675e-01 -7.93121874e-01 7.27085471e-01
4.53962654e-01 -2.07226783e-01 -2.84340113e-01 1.65386391e+00
1.87962968e-02 -3.11255723e-01 2.57123611e-03 2.51433402e-01
1.14390701e-02 4.16533917e-01 3.73464823e-02 -6.29151821e-01
1.13020897e+00 -3.14474218e-02 -6.18627787e-01 4.45189476e-01
2.94603914e-01 -6.22159004e-01 7.13148355e-01 8.98885727e-01
-1.36437201e+00 -9.67052281e-02 -8.98441970e-01 6.71594560e-01
-1.81274965e-01 -6.54083788e-01 3.02920997e-01 1.31292140e+00
-1.07916570e+00 9.77534413e-01 -3.62182647e-01 3.69167149e-01
1.88263267e-01 -1.86446726e-01 -1.02481380e-01 4.21564341e-01
-1.37114120e+00 1.14652085e+00 2.26878613e-01 6.69463500e-02
-8.07365477e-01 -8.96295249e-01 -7.32339859e-01 2.94201910e-01
3.49297076e-01 -3.69351625e-01 1.10252273e+00 -3.03751498e-01
-1.46074426e+00 4.87998039e-01 5.61122596e-01 -3.56617749e-01
7.75039077e-01 -6.39263093e-01 -2.52711028e-01 2.76638400e-02
2.05347966e-02 -6.29830122e-01 7.70607531e-01 -9.73572731e-01
-5.44202387e-01 -3.22367549e-01 -7.28832483e-02 -1.88781619e-01
4.22810346e-01 4.24869508e-01 6.68085814e-01 -1.03382039e+00
-2.96638813e-02 -4.34016436e-01 -3.52746099e-01 -1.89735681e-01
-5.89039266e-01 -4.11747098e-01 -2.39773378e-01 -5.13107479e-01
1.75470805e+00 -1.89184165e+00 1.39525137e-03 8.25057685e-01
-2.53001433e-02 3.87946852e-02 4.53827202e-01 9.64036703e-01
-3.91923070e-01 4.33577150e-01 -5.77000797e-01 3.22188020e-01
3.20544511e-01 -2.75295883e-01 -3.24996829e-01 8.12833846e-01
2.44712770e-01 3.37650865e-01 -8.76290679e-01 -6.93332180e-02
3.61850858e-02 1.98902339e-02 -4.87823188e-01 2.58554667e-01
-1.37249827e-01 -1.81483719e-02 -5.96713543e-01 4.70027894e-01
8.11008930e-01 4.33185607e-01 -2.09869966e-01 1.96842536e-01
-5.73575497e-01 -9.65353847e-02 -1.21077728e+00 7.10084856e-01
-3.27307701e-01 1.19223699e-01 -5.47399297e-02 -9.53317761e-01
1.40083694e+00 2.63525695e-01 4.94077355e-01 -1.60205662e-01
3.75904828e-01 3.73086512e-01 -3.81118774e-01 -3.03463638e-02
5.17053604e-01 -8.17068994e-01 -5.90625167e-01 4.85065103e-01
-2.91179746e-01 -2.83717066e-01 -1.14401691e-01 -9.17073637e-02
1.05436873e+00 -2.39527822e-02 8.59661162e-01 -9.01254833e-01
6.75902069e-01 -6.35763049e-01 5.23262560e-01 4.56205666e-01
-3.87136817e-01 2.97091484e-01 1.23864436e+00 -1.25212327e-01
-7.64237225e-01 -1.93071294e+00 -4.28651452e-01 4.48517680e-01
-8.95972922e-03 -3.40583295e-01 -5.52273691e-01 -6.16689444e-01
3.57100278e-01 1.05209017e+00 -8.20551634e-01 -1.20150983e-01
-1.13743268e-01 -7.59828329e-01 3.21342289e-01 4.40365255e-01
2.30248466e-01 -5.79684377e-01 -7.31226087e-01 2.06984282e-01
2.95564711e-01 -1.13955311e-01 -3.06824356e-01 1.76322594e-01
-6.76003039e-01 -1.12790620e+00 -5.70217788e-01 3.30917597e-01
1.42291740e-01 -4.15063798e-01 1.49154913e+00 -4.35662478e-01
-2.05824375e-01 4.32215065e-01 -5.19321561e-01 -8.50725830e-01
-4.36570764e-01 -6.39301956e-01 1.33321825e-02 2.47345388e-01
-2.85734922e-01 -5.93849003e-01 -8.34347785e-01 3.40000957e-01
-9.82253075e-01 -8.50150347e-01 4.68675166e-01 6.98922038e-01
7.00777292e-01 2.92129993e-01 8.96895409e-01 -8.81998718e-01
8.70825946e-01 -7.62652755e-01 -8.38016570e-01 1.71296880e-01
-9.28641915e-01 1.60412714e-01 5.54245055e-01 2.36359850e-01
-8.14903438e-01 -7.55926371e-01 -2.94721216e-01 -3.29810791e-02
2.87712336e-01 5.74619591e-01 -2.80089617e-01 2.31313989e-01
5.25250852e-01 -1.12414628e-01 -2.15147406e-01 -4.90072310e-01
2.52610445e-01 3.18260312e-01 4.00102168e-01 -6.56736135e-01
9.41252589e-01 3.07959855e-01 5.68059266e-01 -2.09925488e-01
-9.85363603e-01 -2.38405779e-01 -2.65231431e-01 -5.15015841e-01
1.07602370e+00 -2.70612538e-01 -1.06342244e+00 2.17067778e-01
-7.29178250e-01 2.76128262e-01 -7.87854671e-01 5.81553638e-01
-9.59137976e-01 6.86032102e-02 -1.74363881e-01 -1.61004114e+00
-4.20134485e-01 -7.97458768e-01 9.06799614e-01 2.87875026e-01
-8.88687670e-02 -1.28720307e+00 2.36897632e-01 -2.62354821e-01
2.77155101e-01 1.24027812e+00 8.56574953e-01 -6.72538459e-01
-3.38624895e-01 -6.08412981e-01 -1.01913676e-01 8.67873907e-01
-8.30652192e-02 -1.30392546e-02 -4.86895591e-01 -2.15410113e-01
4.66765881e-01 1.27313271e-01 3.77504319e-01 3.47115338e-01
8.27738345e-01 -3.57659668e-01 1.58883378e-01 3.81025642e-01
1.86350703e+00 5.05774319e-01 6.86184108e-01 4.55537587e-01
-8.29368383e-02 7.29650259e-01 1.04830945e+00 8.19085181e-01
9.41765010e-02 4.69772726e-01 8.50340545e-01 5.86861551e-01
8.85336518e-01 -1.05122000e-01 5.17523885e-01 7.15153992e-01
-2.08031282e-01 -1.45045310e-01 -9.26035047e-01 2.72264659e-01
-1.44079065e+00 -1.03577292e+00 -1.01176500e-01 2.62140012e+00
4.13367748e-01 4.00207430e-01 3.51236939e-01 3.04425597e-01
7.61919260e-01 3.17979127e-01 -1.66684300e-01 -1.18336296e+00
1.39856547e-01 4.41213965e-01 6.39821231e-01 5.17926455e-01
-9.96339798e-01 1.72005892e-01 6.46847153e+00 1.06237233e+00
-6.01103187e-01 -1.34373397e-01 5.59069991e-01 3.32150847e-01
-6.16352797e-01 9.45482701e-02 -7.02330589e-01 7.49570668e-01
1.12173116e+00 -5.97017527e-01 -2.47958556e-01 9.59439695e-01
2.11765438e-01 -3.23341846e-01 -8.67135704e-01 5.38134038e-01
-1.42984912e-01 -9.31277812e-01 -1.65457949e-01 2.81419665e-01
4.96277153e-01 -6.73631310e-01 4.70295846e-01 2.36194983e-01
4.48219448e-01 -1.26382196e+00 9.06433403e-01 1.19965947e+00
8.45447242e-01 -1.39657652e+00 9.43284333e-01 1.45072386e-01
-1.11492765e+00 -1.44802099e-02 -1.00956924e-01 1.36225224e-01
4.09375817e-01 1.17386067e+00 -2.46742368e-01 1.20261514e+00
5.33816814e-01 1.99990854e-01 -1.07181899e-01 1.17919064e+00
-2.34957278e-01 4.29773927e-01 -1.30227566e-01 -4.22618948e-02
2.28286505e-01 -5.86701214e-01 8.50856364e-01 7.07038939e-01
7.08096564e-01 -1.26165599e-01 -8.30426067e-02 8.35078001e-01
5.12475371e-01 4.30357486e-01 -4.01139349e-01 2.55017012e-01
3.56257498e-01 7.60955811e-01 -4.49866861e-01 1.48894846e-01
-7.28304982e-02 3.75726551e-01 -1.53455868e-01 -8.91661048e-02
-6.85161233e-01 -7.13813722e-01 8.15629601e-01 9.08019543e-02
1.25991525e-02 -1.27517417e-01 -3.30721736e-01 -6.89504206e-01
6.75311983e-02 -3.96834403e-01 5.24218619e-01 -2.74530232e-01
-1.64824331e+00 7.16209054e-01 3.53690624e-01 -1.52794886e+00
-3.15399557e-01 -7.64400780e-01 -7.29272068e-01 1.11360633e+00
-1.19450450e+00 -3.60630214e-01 2.42152005e-01 3.51168782e-01
-1.57407776e-01 -1.46838546e-01 6.70915842e-01 -2.81956494e-01
-3.17458212e-01 7.65046716e-01 7.48099804e-01 -4.07478750e-01
1.34335458e-01 -1.78837061e+00 1.60848513e-01 6.00119829e-01
-5.59364557e-01 4.27619934e-01 1.17049634e+00 -6.19455755e-01
-9.12842572e-01 -8.43859434e-01 8.74259472e-01 -6.81739748e-01
8.72494638e-01 -3.33043970e-02 -6.19713008e-01 2.36330211e-01
-3.63522142e-01 -2.16228306e-01 9.01528656e-01 6.30462123e-03
-2.02060893e-01 -4.01213527e-01 -1.47703063e+00 4.02940899e-01
5.62138796e-01 -3.03848833e-01 -5.87127388e-01 2.61326194e-01
5.28513610e-01 4.59117955e-03 -1.51444256e+00 5.93124211e-01
7.03818321e-01 -1.67486787e+00 8.08238328e-01 -2.03866884e-01
2.52805382e-01 2.34330982e-01 -2.08796248e-01 -1.43380833e+00
-7.52639025e-02 -9.58389759e-01 2.19812736e-01 1.07864869e+00
1.38073921e-01 -1.12361312e+00 4.99031365e-01 2.91452736e-01
-9.10516009e-02 -1.28489673e+00 -1.50761557e+00 -1.58328187e+00
7.27533996e-01 -6.87681854e-01 6.86920464e-01 3.14040989e-01
1.62745491e-01 -2.80814081e-01 -2.97177345e-01 -7.83225000e-02
8.47100854e-01 -1.51355237e-01 1.94579169e-01 -1.63587165e+00
-4.17406976e-01 -6.78095818e-01 -7.03498006e-01 -3.69416207e-01
-3.23620468e-01 -8.83003592e-01 5.36856800e-02 -1.02777779e+00
-1.30918562e-01 -4.06502128e-01 -6.84078336e-01 -3.55781555e-01
-2.33343989e-01 -2.33587936e-01 2.87988305e-01 -1.08685263e-01
-1.46907508e-01 7.76288211e-01 9.48536098e-01 5.22253633e-01
1.18356086e-02 8.05354893e-01 -7.14836836e-01 8.01896632e-01
9.03006196e-01 -5.11170506e-01 -3.85505199e-01 4.73059535e-01
6.27956867e-01 6.34203672e-01 1.31066859e-01 -7.93987572e-01
-4.77117121e-01 -3.87894630e-01 -1.50947616e-01 -5.17524719e-01
2.48551946e-02 -5.73920965e-01 2.47709870e-01 6.30994558e-01
-1.75797045e-01 7.25214541e-01 -2.67218977e-01 6.94113553e-01
-3.15884769e-01 -3.51295769e-01 1.05068231e+00 5.88150620e-02
6.42417297e-02 2.78135866e-01 -3.48036557e-01 3.33056629e-01
1.47774208e+00 -8.45760927e-02 -2.70205379e-01 -5.91956377e-01
-7.91837573e-01 4.10797559e-02 1.08237147e-01 -1.21077836e-01
6.64222181e-01 -1.33636999e+00 -7.70773768e-01 -2.80666053e-01
-7.93136284e-02 -2.33638555e-01 2.40874201e-01 8.19045722e-01
-7.28076518e-01 5.32610476e-01 -5.19077741e-02 -2.26359203e-01
-7.23170877e-01 5.05729198e-01 3.60047400e-01 -8.14950585e-01
-4.51266468e-01 9.06294107e-01 3.92565936e-01 1.05439246e-01
3.89950387e-02 -2.41248548e-01 -1.66141897e-01 1.37893602e-01
4.77588475e-01 8.39426219e-01 -2.92410582e-01 -5.69985569e-01
-4.57373619e-01 5.22603512e-01 3.40842575e-01 -4.17343497e-01
1.11785591e+00 -2.35607296e-01 -3.46437305e-01 6.63774133e-01
1.01974046e+00 3.14765126e-01 -1.00931406e+00 2.42252901e-01
7.57377088e-01 -5.73918581e-01 -4.03916746e-01 -8.22762311e-01
-5.15459180e-01 7.95215964e-01 4.48531181e-01 8.26493621e-01
1.16762114e+00 -1.55997649e-01 5.55384099e-01 -4.12337422e-01
5.10711193e-01 -1.02523434e+00 -4.50895578e-02 4.81135160e-01
1.25758743e+00 -5.87777495e-01 -9.29956511e-02 -5.27119398e-01
-5.05752087e-01 9.44154680e-01 -2.09722016e-02 -3.86388600e-01
1.31431389e+00 1.08384617e-01 -1.14721112e-01 -4.91466781e-04
-4.33460653e-01 -7.72698922e-03 4.73544568e-01 5.93765080e-01
2.01745257e-01 3.33894640e-01 -9.90362942e-01 9.75994945e-01
-6.00390434e-01 -3.69072497e-01 5.82966805e-01 7.53096938e-01
-6.08789980e-01 -1.15056777e+00 -4.93595541e-01 4.83937204e-01
-9.07305300e-01 -3.12683694e-02 1.76403582e-01 7.84704864e-01
3.21663082e-01 8.95232379e-01 -5.44446744e-02 -3.76938134e-01
6.81646645e-01 -1.77839100e-01 6.95974529e-02 -6.30315840e-01
-7.97610283e-01 -2.89841801e-01 -3.72544937e-02 -7.48758435e-01
9.95887220e-02 -8.76682758e-01 -6.78647101e-01 -3.57613295e-01
-2.41433069e-01 7.10777044e-01 5.74874341e-01 9.28648353e-01
-3.58865589e-01 4.39185798e-01 1.10590482e+00 -3.99212360e-01
-1.29641128e+00 -6.16659403e-01 -1.41797459e+00 -1.87608693e-02
2.11468041e-02 -7.83942938e-01 -5.80978990e-01 -7.66343176e-01]
|
[5.062130451202393, 3.955767869949341]
|
5fc5db50-d324-41e0-aba3-ba1049254d1b
|
search-for-the-ugle-truth-an-investigation
|
2305.06026
| null |
https://arxiv.org/abs/2305.06026v2
|
https://arxiv.org/pdf/2305.06026v2.pdf
|
Search for the UGLE Truth: An Investigation into Unsupervised GNN Learning Environments
|
Graph Neural Networks (GNNs) are a pertinent tool for any machine learning task due to their ability to learn functions over graph structures, a powerful and expressive data representation. The detection of communities, an unsupervised task has increasingly been performed with GNNs. Clustering nodes in a graph using the multi-dimensionality of node features with the connectivity of the graph has many applications to real world tasks from social networks to genomics. Unfortunately, there is currently a gap in the literature with no established sufficient benchmarking environment for fairly and rigorously evaluating GNN based community detection, thereby potentially impeding progress in this nascent field. We observe the particular difficulties in this setting is the ambiguous hyperparameter tuning environments combined with conflicting metrics of performance and evaluation datasets. In this work, we propose and evaluate frameworks for the consistent comparisons of community detection algorithms using GNNs. With this, we show the strong dependence of the performance to the experimental settings, exacerbated by factors such as the use of GNNs and the unsupervised nature of the task, providing clear motivation for the use of a framework to facilitate congruent research in the field.
|
['Ryan McConville', 'Will Leeney']
|
2023-05-10
| null | null | null | null |
['community-detection']
|
['graphs']
|
[ 3.81738633e-01 -1.06370293e-01 -6.00213222e-02 -2.11787686e-01
-4.37484980e-02 -6.60551488e-01 6.73248768e-01 5.34420609e-01
-5.25331199e-01 6.01533711e-01 1.09695747e-01 -4.25846130e-01
-6.61246121e-01 -7.56741703e-01 -1.85465187e-01 -7.96965301e-01
-6.58883750e-01 5.59840262e-01 -4.55334829e-03 -2.37247601e-01
2.59985834e-01 6.22355044e-01 -1.40107965e+00 -7.92093575e-02
5.38295805e-01 4.82363164e-01 -4.36655544e-02 6.90123439e-01
-2.57203355e-02 6.42540216e-01 -4.12040561e-01 -2.52564818e-01
2.43036270e-01 -4.36061621e-01 -6.58572078e-01 1.12471886e-01
1.93345696e-01 2.02277347e-01 -3.29866827e-01 9.88326192e-01
5.40954053e-01 4.17362079e-02 6.86800063e-01 -1.43322110e+00
-3.92566830e-01 6.48256958e-01 -4.84546810e-01 4.15931940e-01
1.42189056e-01 1.39540628e-01 1.36253035e+00 -3.63830566e-01
8.55600893e-01 1.06020808e+00 9.52672482e-01 2.83473521e-01
-1.42500448e+00 -4.07008320e-01 1.87082682e-03 -3.19412723e-02
-1.46382093e+00 -3.46060574e-01 6.73133194e-01 -6.98164761e-01
8.82302999e-01 1.37943000e-01 8.16391826e-01 9.97110367e-01
-2.19140992e-01 5.82458675e-02 9.93358493e-01 -4.89604682e-01
3.24387729e-01 -4.13304381e-02 1.94180742e-01 6.04355752e-01
6.48942888e-01 -2.03110296e-02 -3.31876576e-01 -3.34811091e-01
9.00951862e-01 -1.57933030e-02 -1.85587466e-01 -8.92910600e-01
-1.06535101e+00 1.04352248e+00 5.83590925e-01 6.22443199e-01
-1.34706154e-01 3.06203425e-01 6.49747491e-01 4.03478205e-01
4.03052270e-01 8.16040337e-01 -1.14263684e-01 -9.05678347e-02
-8.54209661e-01 6.57311007e-02 1.17915618e+00 5.07149160e-01
5.40533304e-01 -6.32824078e-02 2.67907321e-01 6.56238914e-01
3.30659211e-01 -2.31321737e-01 1.19862393e-01 -6.73730135e-01
2.35368356e-01 8.51112068e-01 -3.98883790e-01 -1.35802436e+00
-6.38120472e-01 -6.20436072e-01 -1.03136551e+00 1.46390229e-01
6.73562407e-01 -3.21474940e-01 -4.98673886e-01 1.86695731e+00
3.05535287e-01 -9.03279856e-02 -1.34537652e-01 7.66743243e-01
6.13049567e-01 1.72481552e-01 8.08865279e-02 -3.23906392e-02
1.15531385e+00 -4.70854729e-01 -2.33741298e-01 -1.06058925e-01
7.18125105e-01 -4.55739141e-01 7.07063615e-01 2.28609547e-01
-6.35864973e-01 6.77654073e-02 -1.08427727e+00 1.44345775e-01
-5.65193415e-01 -3.92510146e-01 1.11917543e+00 6.74919903e-01
-1.45494092e+00 8.58112514e-01 -7.69863486e-01 -1.01196635e+00
4.22887266e-01 6.37010574e-01 -5.94413280e-01 -1.10819913e-01
-8.91856313e-01 7.86541939e-01 6.69337332e-01 1.66624188e-01
-4.19551969e-01 -2.72952288e-01 -7.61067569e-01 1.92996040e-01
4.09781843e-01 -6.74611926e-01 7.20182657e-01 -8.51208091e-01
-7.89593875e-01 9.21495557e-01 4.74128723e-01 -4.91641581e-01
4.29874778e-01 5.42342663e-01 -2.83283651e-01 8.37442204e-02
2.96899006e-02 4.73776311e-01 4.62106138e-01 -9.95191097e-01
-2.21773714e-01 -4.01191145e-01 3.89678665e-02 1.96382806e-01
-4.61321384e-01 8.68641064e-02 -1.37140378e-01 -3.69040310e-01
1.94646195e-01 -7.80735612e-01 -4.78587270e-01 -5.65415248e-03
-3.70317131e-01 -3.60862345e-01 7.00283110e-01 -1.17746159e-01
1.22450197e+00 -2.22373438e+00 7.03149214e-02 6.11525953e-01
9.75748539e-01 6.94482401e-02 -1.65767774e-01 8.19376469e-01
-8.20930451e-02 3.44232500e-01 -2.50920475e-01 6.51259571e-02
4.08304017e-03 2.01493606e-01 1.98589727e-01 7.51925349e-01
2.56019920e-01 8.30951393e-01 -1.04601717e+00 -4.32481498e-01
3.81574333e-02 6.57713234e-01 -4.03420419e-01 -1.00224458e-01
-1.20370902e-01 1.94709882e-01 -3.56756240e-01 3.09079647e-01
1.50654301e-01 -8.59530032e-01 7.56360352e-01 1.54382482e-01
-9.89407450e-02 1.78088412e-01 -1.32228100e+00 1.28386736e+00
2.48479441e-01 9.46227610e-01 3.64093214e-01 -1.28217220e+00
8.66241038e-01 3.62725705e-01 6.06142402e-01 -2.10362524e-01
1.02812015e-01 1.87948316e-01 6.49448633e-01 -4.07068163e-01
2.69245416e-01 -7.36999661e-02 3.20613474e-01 7.37674654e-01
2.35929474e-01 9.50137675e-02 6.27409935e-01 2.94377774e-01
1.51993859e+00 -2.68195122e-01 3.02271038e-01 -5.56656718e-01
1.03259809e-01 -1.44656450e-02 1.94185972e-01 8.63511384e-01
-3.30438972e-01 4.35822785e-01 8.85719955e-01 -4.44143921e-01
-1.25701284e+00 -6.84819937e-01 -1.48088232e-01 1.05156386e+00
-1.54554382e-01 -5.60384214e-01 -6.19300842e-01 -3.35888445e-01
7.97051340e-02 -9.01743844e-02 -6.53790414e-01 5.26511706e-02
-5.16067624e-01 -1.01996803e+00 7.75186121e-01 2.70392507e-01
3.60793360e-02 -1.12131321e+00 -4.62943345e-01 3.80430311e-01
1.55151963e-01 -1.07544839e+00 -4.92576994e-02 4.78002042e-01
-1.05771339e+00 -1.42594337e+00 -4.02568847e-01 -1.00531828e+00
6.30956233e-01 3.58316690e-01 1.34749973e+00 5.98577261e-01
-3.59341472e-01 4.61000741e-01 -1.91279173e-01 -2.31760710e-01
-4.94046211e-01 3.29952627e-01 1.06190264e-01 -2.20728099e-01
4.55957174e-01 -1.12299085e+00 -5.44289052e-01 6.18389621e-02
-1.10277224e+00 -1.14579208e-01 5.22687256e-01 6.77473724e-01
1.11307830e-01 1.45791262e-01 6.72601521e-01 -9.93992150e-01
1.14167845e+00 -7.30162263e-01 -5.19363999e-01 3.55940275e-02
-8.67987871e-01 7.43233562e-02 4.44079340e-01 -2.39837855e-01
-1.97252691e-01 -2.81409323e-01 2.24413663e-01 7.50416815e-02
-2.74374872e-01 1.00379455e+00 1.36722341e-01 -2.94227034e-01
9.08724010e-01 -1.63069040e-01 2.75156885e-01 -2.74780750e-01
2.82410592e-01 4.38539535e-01 4.29142535e-01 -5.25792658e-01
6.87556624e-01 5.00605762e-01 3.39043647e-01 -9.87004638e-01
-2.23322585e-01 -6.89399064e-01 -7.45294571e-01 -1.21581562e-01
6.51471317e-01 -6.78478837e-01 -5.74529111e-01 1.41447544e-01
-8.79005909e-01 -2.32343495e-01 -5.64056747e-02 3.58204514e-01
-3.33545625e-01 6.47290230e-01 -7.57316470e-01 -8.41570914e-01
-3.49912018e-01 -9.60688710e-01 5.21684349e-01 7.72333294e-02
-3.11761200e-01 -1.36069179e+00 2.56143987e-01 2.29448736e-01
5.65079212e-01 6.88449621e-01 1.15746915e+00 -1.00713122e+00
-5.59599102e-01 -1.69878244e-01 -4.75941956e-01 3.89543064e-02
1.60152406e-01 1.89230785e-01 -8.65480006e-01 -5.23155332e-01
-2.82436490e-01 -3.12735677e-01 6.11098468e-01 2.84874976e-01
7.80421734e-01 -1.06794469e-01 -3.09177428e-01 4.72572267e-01
1.60224867e+00 -2.05180496e-01 3.38812202e-01 4.19083208e-01
7.83088624e-01 7.58648515e-01 -2.94069707e-01 2.45460391e-01
1.56141371e-01 3.51304501e-01 4.65422750e-01 -2.67630666e-01
1.14743061e-01 7.54163042e-02 -1.16120398e-01 7.94169068e-01
-2.28143632e-01 -3.98614049e-01 -1.28618264e+00 4.35231328e-01
-1.84791076e+00 -1.01283288e+00 -2.62112975e-01 2.11574125e+00
6.71861291e-01 2.30603293e-01 3.42398137e-01 2.17510507e-01
1.00462747e+00 1.97502360e-01 -2.61645228e-01 -1.60449877e-01
-1.28710791e-01 5.12910560e-02 1.72323599e-01 2.20294580e-01
-9.58080351e-01 5.08567691e-01 6.36819839e+00 3.89902949e-01
-1.12196648e+00 -4.18453276e-01 5.82539976e-01 1.82116866e-01
-8.25556293e-02 2.48402841e-02 -1.78315774e-01 2.36765325e-01
9.01092410e-01 -1.72193900e-01 7.39462078e-01 5.63412011e-01
2.47895613e-01 1.39627740e-01 -1.35683811e+00 8.50029111e-01
-4.32996191e-02 -1.20321357e+00 -3.11600715e-01 4.04083371e-01
6.10511422e-01 5.13984621e-01 -1.46868140e-01 7.62523487e-02
5.58292449e-01 -1.37270153e+00 9.49511230e-02 -3.43597010e-02
6.75218642e-01 -4.80130166e-01 4.85792667e-01 3.01035762e-01
-1.08155572e+00 -8.61632638e-03 -8.96848962e-02 -3.68989229e-01
-1.53002754e-01 5.66718936e-01 -1.15693295e+00 3.01568329e-01
5.35019994e-01 6.24610126e-01 -6.47582293e-01 1.37153566e+00
1.75800160e-01 7.03305840e-01 -5.17206371e-01 -2.62062997e-01
4.03486907e-01 -4.33251768e-01 6.63870990e-01 1.34840822e+00
-1.45672187e-02 -3.58455896e-01 2.13452160e-01 8.44976723e-01
-1.56973556e-01 1.46733373e-01 -9.97513652e-01 -6.05038345e-01
4.05849904e-01 1.46167397e+00 -1.32647169e+00 7.23671615e-02
-3.85110587e-01 3.12399626e-01 6.97892904e-01 4.30978000e-01
-3.02486390e-01 -3.13758701e-01 4.99717027e-01 3.05242628e-01
1.85161576e-01 -5.00679851e-01 -2.13641852e-01 -8.87287378e-01
9.85527411e-02 -1.01422369e+00 6.51040196e-01 -4.93701816e-01
-1.67430353e+00 4.94771183e-01 4.88608144e-03 -7.55380452e-01
-3.23585153e-01 -6.81966245e-01 -6.47161067e-01 5.94114840e-01
-1.17380071e+00 -9.23601329e-01 -3.14304113e-01 2.81348437e-01
-1.31466046e-01 -1.86454877e-02 8.11412096e-01 3.02294880e-01
-5.19392908e-01 1.98637098e-01 3.16190481e-01 4.68035430e-01
4.58873391e-01 -1.42489028e+00 4.57031041e-01 7.42389619e-01
4.14935917e-01 9.20675159e-01 7.55806267e-01 -5.44904053e-01
-1.35225189e+00 -7.62270629e-01 7.87625372e-01 -5.69636464e-01
1.01685476e+00 -6.86872721e-01 -9.32825923e-01 5.86788476e-01
1.07330434e-01 5.68271019e-02 9.21123266e-01 6.59650385e-01
-4.95131880e-01 2.22893566e-01 -9.91357088e-01 6.27249479e-01
1.18200231e+00 -6.39158785e-01 -1.23699330e-01 3.92868489e-01
3.38539392e-01 1.53000103e-02 -9.95168269e-01 2.89114416e-01
4.22390521e-01 -1.15376568e+00 7.51694262e-01 -7.23273754e-01
2.71668881e-01 -2.17267171e-01 -1.01726167e-02 -1.10262752e+00
-5.68025112e-01 -6.41157389e-01 2.36083698e-02 1.26746309e+00
3.58466268e-01 -7.11164534e-01 1.08174384e+00 5.29599965e-01
2.01966241e-01 -8.50201905e-01 -8.39094102e-01 -3.66028041e-01
1.44213149e-02 -1.15363032e-01 4.14852470e-01 1.27451539e+00
1.51055291e-01 6.56924188e-01 1.37337625e-01 -1.06776483e-01
5.86240411e-01 -9.34734866e-02 6.92257285e-01 -1.87382925e+00
-3.08874220e-01 -8.17279100e-01 -9.00224030e-01 -3.76052290e-01
-1.24794193e-01 -1.05336690e+00 -1.51139706e-01 -1.50123048e+00
4.96246070e-01 -5.71442604e-01 -3.79762918e-01 1.36113182e-01
-4.61349823e-02 2.41971642e-01 4.73633036e-02 3.35667908e-01
-8.51589501e-01 -6.49718046e-02 8.93553853e-01 -2.10656542e-02
-2.58396506e-01 -2.32908860e-01 -9.16519642e-01 6.03132963e-01
8.18891048e-01 -4.31337386e-01 -4.65287060e-01 -3.16972911e-01
6.22200012e-01 -3.44043493e-01 4.59400833e-01 -1.05787420e+00
3.33598167e-01 1.03582419e-01 2.84799486e-01 -5.84100001e-02
7.51845597e-04 -7.45161414e-01 2.93338597e-01 3.79284441e-01
-2.39919528e-01 4.10680234e-01 -1.50614725e-02 7.84959197e-01
3.50253806e-02 -1.56597868e-01 4.70723212e-01 -2.92939395e-01
-5.77392995e-01 2.83740461e-01 -2.93190569e-01 2.72313118e-01
7.02539265e-01 -5.18560946e-01 -3.98396015e-01 -4.28332686e-01
-5.88731408e-01 2.46994033e-01 7.45056450e-01 2.55231291e-01
1.82059646e-01 -9.35572624e-01 -7.62113750e-01 9.11075696e-02
1.36849046e-01 9.77810323e-02 -2.05269769e-01 8.31787944e-01
-6.12004399e-01 2.76065290e-01 -2.82863051e-01 -7.06588864e-01
-1.26416981e+00 5.55871725e-01 3.35770607e-01 -4.74039823e-01
-5.72221100e-01 5.22801399e-01 -8.90593082e-02 -4.66153413e-01
3.52765888e-01 6.66303560e-02 -3.17008764e-01 4.52409945e-02
2.88581718e-02 3.73374939e-01 1.34892538e-01 -4.91215944e-01
-1.87849760e-01 -5.26368171e-02 3.78978029e-02 8.80690962e-02
1.59283447e+00 8.45684558e-02 -5.36265016e-01 3.31065476e-01
9.10586953e-01 -2.81108171e-01 -9.63983655e-01 -6.06320947e-02
4.59517390e-01 6.56961799e-02 -2.55536139e-01 -4.39492732e-01
-7.68801928e-01 5.91398656e-01 3.96294206e-01 9.18014586e-01
7.07553983e-01 2.39274297e-02 1.47610754e-01 4.29355413e-01
2.31993005e-01 -1.00764072e+00 -5.81779843e-03 4.40568417e-01
4.56286103e-01 -1.16553533e+00 3.20466012e-01 -2.92306900e-01
-9.41318721e-02 1.07755709e+00 3.39673281e-01 -1.36717454e-01
4.39317137e-01 1.74223587e-01 4.50902805e-02 -7.47227252e-01
-5.45791626e-01 -2.04599857e-01 7.78904781e-02 8.61599922e-01
6.64833486e-01 -1.06103525e-01 -2.33159348e-01 5.90170324e-02
-1.36962190e-01 -1.92328677e-01 4.68929619e-01 8.19497764e-01
-3.69497269e-01 -1.13052440e+00 -2.11946756e-01 8.35482836e-01
-4.94012326e-01 -1.87368110e-01 -8.04403961e-01 1.07915282e+00
-9.55389813e-02 9.44170535e-01 -5.56339212e-02 -1.71055570e-01
-1.40258878e-01 -4.03149538e-02 4.00211066e-01 -7.87379146e-01
-8.04285586e-01 -1.95233718e-01 2.40510046e-01 -1.54135570e-01
-4.71174896e-01 -6.56957924e-01 -8.47849786e-01 -4.49949831e-01
-5.33953547e-01 3.47007602e-01 7.44949758e-01 8.48323882e-01
4.52569872e-01 2.83088923e-01 2.68341303e-01 -7.43185341e-01
-6.13320053e-01 -8.94713223e-01 -7.28300631e-01 5.87480664e-01
2.44505495e-01 -5.48879206e-01 -5.89054942e-01 -3.05958390e-01]
|
[6.895699977874756, 5.7994794845581055]
|
79380c80-589a-4f43-b289-ecdcfcdd9dc8
|
additive-tree-structured-conditional
|
2010.03171
| null |
https://arxiv.org/abs/2010.03171v1
|
https://arxiv.org/pdf/2010.03171v1.pdf
|
Additive Tree-Structured Conditional Parameter Spaces in Bayesian Optimization: A Novel Covariance Function and a Fast Implementation
|
Bayesian optimization (BO) is a sample-efficient global optimization algorithm for black-box functions which are expensive to evaluate. Existing literature on model based optimization in conditional parameter spaces are usually built on trees. In this work, we generalize the additive assumption to tree-structured functions and propose an additive tree-structured covariance function, showing improved sample-efficiency, wider applicability and greater flexibility. Furthermore, by incorporating the structure information of parameter spaces and the additive assumption in the BO loop, we develop a parallel algorithm to optimize the acquisition function and this optimization can be performed in a low dimensional space. We demonstrate our method on an optimization benchmark function, on a neural network compression problem, on pruning pre-trained VGG16 and ResNet50 models as well as on searching activation functions of ResNet20. Experimental results show our approach significantly outperforms the current state of the art for conditional parameter optimization including SMAC, TPE and Jenatton et al. (2017).
|
['Matthew B. Blaschko', 'Xingchen Ma']
|
2020-10-06
| null | null | null | null |
['smac-1', 'smac']
|
['playing-games', 'playing-games']
|
[ 1.72131002e-01 2.60994257e-03 -1.28699848e-02 -5.19557595e-01
-7.24356711e-01 -3.20640355e-01 3.32932860e-01 -1.88384071e-01
-8.33245635e-01 8.42176080e-01 -1.89753160e-01 -4.24933404e-01
-5.33001304e-01 -5.07968366e-01 -8.30489397e-01 -8.36288273e-01
-7.47698396e-02 8.15645039e-01 3.09457272e-01 1.90572023e-01
3.09037268e-01 4.88845915e-01 -1.25103962e+00 -2.43532509e-01
8.70606422e-01 1.28973436e+00 5.00373006e-01 7.42219269e-01
2.30117589e-01 4.80967432e-01 -3.65482479e-01 -4.53643680e-01
2.77979046e-01 2.28197150e-05 -8.57754827e-01 -1.77563906e-01
6.40416205e-01 -2.48256505e-01 1.18375815e-01 9.28072274e-01
6.40383422e-01 4.04510021e-01 7.08284914e-01 -8.76922309e-01
-7.68920332e-02 9.60761905e-01 -1.55976593e-01 1.56709120e-01
-5.21981001e-01 1.05515033e-01 1.16052437e+00 -8.66522789e-01
4.69990581e-01 1.27420497e+00 8.42321336e-01 4.14948225e-01
-1.61382353e+00 -5.66775799e-01 3.45505625e-01 3.81077498e-01
-1.45511746e+00 -3.47005665e-01 4.46457863e-01 -4.58868414e-01
1.37810993e+00 9.97712091e-02 6.71105981e-01 6.77279413e-01
2.81492233e-01 9.12299454e-01 9.31488812e-01 -3.27818841e-01
4.94680315e-01 -1.65552288e-01 4.50866848e-01 9.07575428e-01
6.56222925e-02 3.56312692e-01 -5.59555233e-01 -1.53752804e-01
4.69932675e-01 -3.76229942e-01 -2.61827290e-01 -5.68253577e-01
-6.86118543e-01 1.03077006e+00 4.43346947e-01 -3.15764062e-02
-2.51272500e-01 6.74644291e-01 1.19467020e-01 1.96851278e-03
5.85750222e-01 5.24473190e-01 -9.55437005e-01 -3.36760700e-01
-1.19277012e+00 4.24463719e-01 1.05896962e+00 9.24284697e-01
7.58019388e-01 2.10547537e-01 -3.20768356e-01 1.05394745e+00
7.48198628e-01 2.89240628e-01 4.46172386e-01 -9.66755807e-01
4.85920042e-01 3.19243073e-01 -2.13042349e-01 -4.45337385e-01
-5.75599134e-01 -8.20099294e-01 -7.41368055e-01 1.20699480e-01
3.41811031e-01 -1.28222659e-01 -1.21230090e+00 1.73737609e+00
3.78291041e-01 1.69942901e-01 -2.79723674e-01 5.10191798e-01
6.99123621e-01 3.61487120e-01 3.71337272e-02 -9.57173482e-02
1.04187143e+00 -1.20625412e+00 -4.93033141e-01 -4.73494530e-01
8.53291214e-01 -4.82496530e-01 9.34418023e-01 6.59495652e-01
-1.13659322e+00 -2.37508789e-01 -1.19721973e+00 -4.84767221e-02
-2.41871253e-01 3.29060644e-01 8.14267039e-01 1.04396033e+00
-1.17458415e+00 1.30454314e+00 -1.18123639e+00 1.94759816e-01
8.35492015e-01 7.45697677e-01 5.87148918e-03 1.30824432e-01
-8.67789805e-01 1.03098845e+00 8.03868473e-01 1.80176288e-01
-1.07239509e+00 -1.27574122e+00 -6.37086928e-01 3.23652267e-01
5.59061468e-01 -7.90389359e-01 1.43505108e+00 -3.44851524e-01
-2.06140351e+00 4.08364594e-01 -5.33624031e-02 -1.01863873e+00
5.00389874e-01 -5.14766216e-01 3.83989066e-01 -2.40809187e-01
-5.47843337e-01 7.79386401e-01 9.88929927e-01 -7.10387170e-01
-2.12133631e-01 -4.21982586e-01 -3.09634507e-01 1.14352547e-01
-1.97842211e-01 1.65174250e-02 -6.16021693e-01 -5.66268027e-01
1.51751518e-01 -1.01633775e+00 -5.21458566e-01 -3.27621490e-01
-5.94196439e-01 -1.44315585e-01 3.20414513e-01 -8.06223392e-01
1.52298188e+00 -1.69860291e+00 4.24305260e-01 5.55545568e-01
1.47198364e-01 1.11927561e-01 -5.59335994e-03 -1.29875019e-01
-7.56283998e-02 3.92580032e-01 -7.37490773e-01 -7.31908858e-01
1.99225441e-01 2.98021913e-01 -6.97514489e-02 5.36548376e-01
9.12580863e-02 1.06256068e+00 -3.20297986e-01 -5.31136215e-01
2.57138133e-01 4.63714510e-01 -1.04910111e+00 -1.39876902e-02
-4.41578478e-01 1.60667002e-01 -1.97284371e-01 5.20341933e-01
8.39139760e-01 -2.87555337e-01 1.38073534e-01 -1.24338903e-01
1.26661599e-01 4.10181224e-01 -1.29228747e+00 1.61076939e+00
-5.37233233e-01 5.40489495e-01 1.09781392e-01 -1.09345520e+00
8.38120759e-01 -7.44375437e-02 4.42616224e-01 -9.11237970e-02
4.06522065e-01 8.38995129e-02 1.23437576e-01 1.44933581e-01
2.48343214e-01 1.22606099e-01 4.21068549e-01 1.87774390e-01
5.49448788e-01 -5.57501853e-01 4.69914049e-01 -1.86369076e-01
9.82229531e-01 4.15311843e-01 2.35122442e-01 -6.10354424e-01
4.59020674e-01 -4.59515393e-01 3.67761463e-01 8.89145970e-01
1.66511655e-01 4.55054283e-01 4.07669306e-01 -5.92664629e-02
-7.03189969e-01 -7.69360960e-01 -5.51149130e-01 1.32414031e+00
-5.76244831e-01 -8.35806847e-01 -1.03856027e+00 -5.40239871e-01
-1.20144308e-01 1.02516079e+00 -6.39149845e-01 -5.03016040e-02
-6.45446897e-01 -1.22762275e+00 3.60197991e-01 5.77464163e-01
5.85136116e-01 -7.66704321e-01 -6.49558783e-01 4.04009998e-01
1.79021075e-01 -9.27496493e-01 -3.52429032e-01 6.41383767e-01
-1.54965353e+00 -8.49227190e-01 -4.77892011e-01 -4.01798785e-01
4.84544873e-01 -5.45934856e-01 1.21958256e+00 -1.09818123e-01
-3.70838881e-01 1.18825920e-01 8.89830589e-02 -6.94563627e-01
-1.47828162e-01 5.75107515e-01 -3.37293386e-01 -2.40512550e-01
3.43264230e-02 -7.17270732e-01 -3.18756312e-01 3.88143152e-01
-6.54774070e-01 2.62810320e-01 3.86741668e-01 1.05562770e+00
9.28219438e-01 -1.21989898e-01 1.13073245e-01 -9.11883950e-01
5.52984834e-01 -2.52124637e-01 -1.28986764e+00 3.67118090e-01
-1.22824383e+00 7.77299285e-01 3.22350502e-01 -3.95748049e-01
-1.04642117e+00 2.62711376e-01 -2.48623848e-01 -3.98297042e-01
3.11971694e-01 6.59336448e-01 -9.63568389e-02 -3.71804446e-01
6.93117559e-01 2.54222732e-02 -1.13698281e-01 -7.95319080e-01
3.76900494e-01 -4.47645783e-02 2.50485390e-01 -8.64262283e-01
5.55541992e-01 1.57513797e-01 5.00163555e-01 -6.52468503e-01
-9.02879775e-01 -2.65509427e-01 -6.22114360e-01 -5.59711307e-02
7.72910476e-01 -5.95256686e-01 -7.13864028e-01 4.52485412e-01
-1.06295824e+00 -7.34762311e-01 -2.36346945e-01 6.67673588e-01
-8.97999525e-01 1.48452267e-01 -3.45598429e-01 -7.97205925e-01
-5.39019465e-01 -1.38796985e+00 1.02164388e+00 6.07816838e-02
3.63512076e-02 -1.15077329e+00 1.02276877e-01 2.77935803e-01
5.56436181e-01 -5.99123016e-02 9.04248118e-01 -7.15710163e-01
-8.54849458e-01 2.83481944e-02 1.28927696e-02 6.41489685e-01
-5.32102942e-01 9.36931744e-02 -8.51863921e-01 -1.67118907e-01
2.17062965e-01 -2.86352038e-01 1.39171147e+00 9.45743263e-01
1.70519280e+00 -3.38277876e-01 -2.82967657e-01 1.36758876e+00
1.39440906e+00 -7.83012211e-02 5.89385033e-01 3.18176925e-01
6.43026471e-01 3.50560308e-01 4.86987501e-01 4.56225783e-01
1.33098528e-01 6.72658205e-01 6.03696942e-01 4.31373447e-01
-3.29400077e-02 -1.18478192e-02 2.43175000e-01 8.03860784e-01
-4.01643604e-01 -1.38115406e-01 -9.67563748e-01 2.36881778e-01
-1.95382440e+00 -5.10880589e-01 -7.56209567e-02 2.18153834e+00
1.14665353e+00 8.88330489e-02 -3.49918038e-01 -8.84305499e-03
4.22429115e-01 -5.07788435e-02 -7.54006743e-01 -2.69232482e-01
-2.83792708e-02 8.29268694e-01 1.01636159e+00 6.42242134e-01
-1.21349514e+00 1.22501695e+00 6.74641418e+00 1.32833922e+00
-6.66807652e-01 2.23913088e-01 6.59998477e-01 -3.01437020e-01
-7.60252327e-02 1.65155455e-01 -1.41084731e+00 1.01535082e-01
1.26160026e+00 2.09152773e-01 8.30591202e-01 1.08155775e+00
-1.49466887e-01 -1.72921360e-01 -1.24426317e+00 1.02225041e+00
-2.29160726e-01 -1.45965171e+00 -4.31490004e-01 1.63533121e-01
6.69130921e-01 5.21135926e-01 9.36560407e-02 4.38806981e-01
6.14984095e-01 -1.40308428e+00 7.61581957e-01 4.46558744e-01
2.58261323e-01 -8.25657129e-01 5.74220181e-01 2.99361825e-01
-9.61776972e-01 -3.48672196e-02 -4.68467295e-01 4.17465627e-01
1.63292721e-01 6.62346542e-01 -1.02665448e+00 3.64006311e-01
8.90019834e-01 6.45425081e-01 -8.19584310e-01 1.18565059e+00
-4.69846725e-01 1.09421337e+00 -8.59549820e-01 -3.76150370e-01
1.46501124e-01 -5.40387690e-01 6.28171861e-01 1.18106914e+00
1.93508402e-01 -4.05041575e-01 -2.29301155e-01 1.00986922e+00
-6.46431148e-02 1.58018038e-01 -4.63304184e-02 -5.30674756e-02
2.16444671e-01 1.15448141e+00 -5.72414398e-01 -3.52873169e-02
1.82135805e-01 3.18675250e-01 6.67668104e-01 3.42759401e-01
-1.09123254e+00 -2.11049989e-01 3.75950336e-01 -3.41026515e-01
8.56327951e-01 -3.04248095e-01 -5.97673833e-01 -9.78691697e-01
-6.66075051e-02 -8.32087457e-01 4.02102351e-01 -5.02523839e-01
-8.57584655e-01 3.30396056e-01 4.92819875e-01 -6.06368542e-01
-3.40112388e-01 -1.00415254e+00 -3.05488825e-01 8.18072498e-01
-1.32531035e+00 -8.20631921e-01 5.85556045e-05 5.18456936e-01
3.82418424e-01 -2.38243371e-01 7.38149464e-01 1.47635221e-01
-6.66064620e-01 7.90167689e-01 4.09918129e-01 -3.27758014e-01
1.96899533e-01 -1.29511356e+00 3.73117894e-01 6.94231153e-01
1.55401513e-01 7.59276092e-01 5.46306610e-01 -5.19140840e-01
-1.34287107e+00 -9.42124724e-01 4.36610222e-01 -2.76570737e-01
6.71433687e-01 -6.07622921e-01 -7.99009919e-01 7.07750857e-01
-2.41071209e-02 -2.20732287e-01 6.39143944e-01 3.55239958e-01
-1.98264569e-01 -1.55613869e-01 -1.09467018e+00 5.28560877e-01
1.12032247e+00 -4.05417234e-02 -4.13467407e-01 5.92461467e-01
6.91173196e-01 -6.93817616e-01 -1.05481791e+00 7.24822223e-01
5.73488116e-01 -7.26649225e-01 1.16758811e+00 -7.39675641e-01
5.02734371e-02 2.83626951e-02 -3.33055437e-01 -1.42618835e+00
-1.30633321e-02 -9.88196135e-01 -6.13937378e-01 1.05013144e+00
7.04452455e-01 -8.64732504e-01 9.39879179e-01 7.40929604e-01
-3.11443895e-01 -1.29251564e+00 -1.23879206e+00 -9.39279556e-01
2.12932080e-01 -8.66899252e-01 6.68887556e-01 1.41325191e-01
-6.55254602e-01 3.10925633e-01 -1.70425296e-01 -1.77145764e-01
7.38047421e-01 -3.05756658e-01 7.33787656e-01 -1.39672649e+00
-8.28357816e-01 -7.37834930e-01 -8.02350789e-02 -1.25683451e+00
2.78424174e-01 -1.11074388e+00 2.14308754e-01 -1.20970011e+00
2.27339827e-02 -4.67815727e-01 -8.40315297e-02 5.12413681e-01
-1.55434966e-01 -3.39605749e-01 1.38010919e-01 -1.93342399e-02
-3.16401541e-01 8.71829331e-01 1.08727884e+00 4.65334617e-02
-1.51718304e-01 2.01007545e-01 -2.98455685e-01 8.26538146e-01
9.73219991e-01 -8.57185423e-01 -4.37069118e-01 -4.80939806e-01
2.80625522e-01 -4.44150239e-01 2.29199573e-01 -9.72030938e-01
2.88935363e-01 -2.65647210e-02 2.42565885e-01 -8.29452693e-01
6.15270138e-01 -6.77278936e-01 4.99306656e-02 4.91502881e-01
-3.38229090e-01 -1.40213713e-01 3.65427196e-01 5.84404647e-01
-4.16623130e-02 -9.60823119e-01 9.55912292e-01 8.92510638e-02
-2.78706759e-01 6.44068897e-01 -9.81912166e-02 2.17963099e-01
5.86954296e-01 -6.90092295e-02 -1.54904217e-01 1.17364945e-02
-7.93820083e-01 3.19436342e-01 -8.81306976e-02 -5.97472712e-02
5.84835887e-01 -1.02761149e+00 -6.19447351e-01 8.48807320e-02
-3.44648808e-01 3.62376064e-01 7.34316930e-02 1.04506481e+00
-6.91263914e-01 5.37119806e-01 2.00968146e-01 -8.64032865e-01
-1.23996866e+00 6.66019619e-02 4.85682964e-01 -9.09163117e-01
-2.07248747e-01 1.42333174e+00 -6.21188954e-02 -6.42869473e-01
4.99591529e-01 -6.73020244e-01 -4.64191437e-02 -1.59423053e-02
9.38046873e-02 5.38711548e-01 3.28777969e-01 -1.60758331e-01
-3.93196732e-01 5.35654724e-01 -5.21109104e-02 -4.49942887e-01
1.69731915e+00 4.19839174e-01 -1.54506311e-01 2.04158410e-01
1.36630416e+00 -4.89704400e-01 -1.40476370e+00 -2.54911155e-01
2.01139867e-01 -3.24363291e-01 5.49951196e-01 -6.73626482e-01
-1.28167760e+00 9.33380604e-01 7.10170805e-01 -1.50851220e-01
1.10038769e+00 -1.29035279e-01 2.91209966e-01 9.18454051e-01
2.05052659e-01 -1.47752714e+00 -1.49887159e-01 9.38452125e-01
9.65951443e-01 -8.84277046e-01 3.58478338e-01 -2.77301937e-01
-3.75825256e-01 1.22228265e+00 4.44640577e-01 -1.19130380e-01
1.13919103e+00 2.64197260e-01 -7.05282032e-01 -1.56376570e-01
-8.75019431e-01 -1.04113251e-01 6.23193204e-01 4.73740369e-01
2.50912756e-01 -1.35028109e-01 -2.12038636e-01 6.81323528e-01
-6.72306657e-01 -1.37932628e-01 -5.28208800e-02 7.58554518e-01
-3.65027636e-01 -1.20758319e+00 -8.62595364e-02 6.65807784e-01
-5.83097816e-01 -6.60417676e-01 -1.77021667e-01 7.45053470e-01
-2.10331723e-01 6.81289613e-01 -2.61324853e-01 -1.80583209e-01
4.93822508e-02 3.67940187e-01 8.29813004e-01 -5.96852243e-01
-7.33049273e-01 5.94701152e-03 3.58064115e-01 -7.44835973e-01
-2.64353245e-01 -7.80762434e-01 -1.00640428e+00 1.34019941e-01
-8.62004280e-01 -1.34866849e-01 1.34435439e+00 1.05938971e+00
3.46990615e-01 7.12896109e-01 9.35163721e-02 -9.24483180e-01
-1.10543633e+00 -1.24634397e+00 -3.15319657e-01 -2.33274803e-01
5.95580507e-03 -8.47760081e-01 -4.70287234e-01 -2.00362176e-01]
|
[7.840514659881592, 3.5923988819122314]
|
31ab486d-cf6d-4cfd-bea6-8cdb8b655c57
|
accurate-tree-roots-positioning-and-sizing
|
2205.13731
| null |
https://arxiv.org/abs/2205.13731v1
|
https://arxiv.org/pdf/2205.13731v1.pdf
|
Accurate Tree Roots Positioning and Sizing over Undulated Ground Surfaces by Common Offset GPR Measurements
|
Tree roots detection is a popular application of the Ground-penetrating radar (GPR). Normally, the ground surface above the tree roots is assumed to be flat, and standard processing methods based on hyperbolic fitting are applied to the hyperbolae reflection patterns of tree roots for detection purposes. When the surface of the land is undulating (not flat), these typical hyperbolic fitting methods becomes inaccurate. This is because, the reflection patterns change with the uneven ground surfaces. When the soil surface is not flat, it is inaccurate to use the peak point of an asymmetric reflection pattern to identify the depth and horizontal position of the underground target. The reflection patterns of the complex shapes due to extreme surface variations results in analysis difficulties. Furthermore, when multiple objects are buried under an undulating ground, it is hard to judge their relative positions based on a B-scan that assumes a flat ground. In this paper, a roots fitting method based on electromagnetic waves (EM) travel time analysis is proposed to take into consideration the realistic undulating ground surface. A wheel-based (WB) GPR and an antenna-height-fixed (AHF) GPR System are presented, and their corresponding fitting models are proposed. The effectiveness of the proposed method is demonstrated and validated through numerical examples and field experiments.
|
['Abdulkadir C. Yucel', 'Mohamed Lokman Mohd Yusof', 'Lai Fern Ow', 'Yee Hui Lee', 'Wenhao Luo']
|
2022-05-27
| null | null | null | null |
['gpr', 'gpr']
|
['computer-vision', 'miscellaneous']
|
[ 4.50634599e-01 -5.24190180e-02 6.94536746e-01 -5.13216592e-02
-4.22002316e-01 -2.23714486e-02 -1.01934455e-01 6.73131049e-02
1.49731413e-01 4.56025243e-01 -4.49247420e-01 -4.75822240e-01
-3.52508485e-01 -1.31141961e+00 -3.20381910e-01 -8.11321080e-01
-4.19937313e-01 2.53535032e-01 5.46441913e-01 -4.47630018e-01
3.21554512e-01 6.43146574e-01 -1.62397981e+00 4.51155100e-03
9.91747320e-01 9.53988135e-01 7.14627564e-01 3.01591456e-01
6.30974993e-02 1.45296603e-01 -5.61174631e-01 -1.12440474e-01
5.85066080e-02 -8.21998417e-02 -2.88345277e-01 -6.24837168e-02
-2.87371576e-01 -4.35026258e-01 5.83750829e-02 1.09556496e+00
3.23215306e-01 -4.63275723e-02 9.44498122e-01 -7.34016061e-01
-7.27266073e-02 9.88352001e-02 -1.17255747e+00 1.38739228e-01
4.21492696e-01 -8.45225513e-01 3.09372604e-01 -1.07773650e+00
-5.19123189e-02 1.00761199e+00 1.23840952e+00 -1.72659680e-01
-5.81395745e-01 -5.19323468e-01 -3.05927485e-01 4.71731454e-01
-1.53902423e+00 -1.29610419e-01 6.80912852e-01 -3.25334579e-01
4.48013842e-01 3.93068373e-01 6.84642017e-01 3.56603891e-01
6.72976375e-01 1.80571675e-01 8.84058475e-01 -7.99156964e-01
2.86067516e-01 -3.03681105e-01 3.80631924e-01 4.12599921e-01
1.13599956e+00 -6.05422165e-03 3.50130320e-01 -4.45537686e-01
5.60816228e-01 -1.16346292e-01 -6.98601902e-01 -1.13008447e-01
-2.73520589e-01 6.13006473e-01 3.35535973e-01 4.97681260e-01
-9.02250409e-01 -3.41072947e-01 5.17816246e-02 -1.16803981e-01
3.67853582e-01 3.94230634e-02 4.64537041e-03 1.73104525e-01
-9.89886582e-01 2.71267295e-01 1.02516925e+00 6.68522596e-01
4.68263268e-01 3.20191145e-01 3.65715533e-01 7.77193189e-01
6.17620111e-01 1.23110282e+00 8.40332732e-02 -1.50858119e-01
1.61075458e-01 4.58247773e-02 4.01717603e-01 -1.61177528e+00
-6.07803524e-01 -4.84332055e-01 -6.24035180e-01 1.44424453e-01
3.02761108e-01 -2.43877456e-01 -7.68271744e-01 9.34513569e-01
4.58330303e-01 2.54838258e-01 4.13384944e-01 7.52969086e-01
9.60541666e-01 1.02288246e+00 -2.90532887e-01 -5.24174333e-01
1.58484662e+00 -3.21598090e-02 -8.98557782e-01 -4.10033703e-01
6.99602664e-01 -8.64786088e-01 4.21063483e-01 5.41630447e-01
-7.29935586e-01 -1.34481028e-01 -1.20638597e+00 9.88816321e-01
1.27401620e-01 1.93516001e-01 2.98427284e-01 7.94855773e-01
-5.06506205e-01 5.23557961e-01 -9.49075758e-01 -4.53091562e-01
-2.25699127e-01 -2.82285929e-01 3.66015099e-02 -4.59366322e-01
-1.27739334e+00 9.56380844e-01 -2.56447494e-01 8.08214366e-01
-2.29595341e-02 -4.59921986e-01 -6.78096771e-01 -6.00553956e-03
-8.19274485e-02 -5.07282652e-02 1.21446896e+00 -3.30814362e-01
-1.27191770e+00 4.05398756e-01 1.91813096e-01 7.12421164e-02
2.79333264e-01 -2.02098325e-01 -8.49998057e-01 2.07580745e-01
3.35517555e-01 -1.06580853e+00 5.59517324e-01 -1.59222674e+00
-5.42101562e-01 -7.35530317e-01 -6.21976614e-01 1.37849346e-01
3.36343259e-01 -1.44961411e-02 3.08418423e-01 -5.38581014e-02
1.04854393e+00 -5.29051423e-01 -2.47744069e-01 -5.61362684e-01
-5.64340055e-01 3.61127675e-01 1.00866520e+00 -1.01716805e+00
1.21255946e+00 -2.07635236e+00 -8.66001129e-01 7.36657321e-01
-3.01409960e-01 1.67943984e-01 2.98003197e-01 8.44382882e-01
-9.69609022e-02 -4.64309156e-01 -4.43998039e-01 5.69352150e-01
-4.66137737e-01 -1.77758217e-01 -2.80543894e-01 8.73490870e-01
-5.38877308e-01 -1.69813586e-03 -6.79594755e-01 -1.91391259e-01
-3.22111666e-01 5.14797747e-01 1.91086754e-02 -8.01586173e-03
7.01661587e-01 -1.98428735e-01 -9.95252907e-01 8.41250300e-01
1.62750220e+00 4.31134135e-01 1.49300126e-02 -3.65603566e-01
-6.66926324e-01 -2.07473338e-01 -1.49750566e+00 3.57528061e-01
-4.83490646e-01 2.42088109e-01 4.60138142e-01 -1.10760272e+00
1.63685393e+00 3.60330999e-01 2.65176743e-01 -8.67227495e-01
-9.69762430e-02 7.76341915e-01 8.33637919e-03 -8.07280779e-01
3.41937512e-01 -3.68318856e-01 2.34781995e-01 -5.88391982e-02
-8.44059467e-01 -4.06925797e-01 -3.50539863e-01 -3.31946284e-01
1.09921050e+00 2.70777941e-02 4.12588209e-01 -6.03650331e-01
2.36803189e-01 2.77328826e-02 7.16964066e-01 3.02403361e-01
5.69330335e-01 5.45149028e-01 1.27167359e-01 -3.53836060e-01
-5.85649729e-01 -9.60576415e-01 -8.35285008e-01 2.34170198e-01
5.69695532e-01 1.33647546e-02 -5.56162834e-01 4.39092666e-01
1.45886494e-02 9.59089994e-01 -3.11540365e-01 -1.22290790e-01
-4.96094942e-01 -1.28099823e+00 4.64541137e-01 2.32015416e-01
7.53923416e-01 -9.59878027e-01 -1.13223124e+00 6.30851746e-01
-4.32679862e-01 -9.64582860e-01 7.61123717e-01 4.23409045e-02
-1.03477728e+00 -1.25777411e+00 -8.90276492e-01 -6.97999716e-01
4.91708189e-01 7.55798876e-01 8.18379700e-01 1.28939390e-01
-5.02818525e-01 4.16112840e-01 -8.88949871e-01 -7.27440834e-01
-1.80341586e-01 -6.50402844e-01 -1.51375413e-01 8.86614323e-02
1.13932014e-01 -6.05179429e-01 -4.81934458e-01 7.45294452e-01
-5.16403496e-01 -2.79118717e-01 3.61592442e-01 5.30236125e-01
4.14059639e-01 7.72894621e-01 6.01888776e-01 -5.24933398e-01
5.39535463e-01 -6.96533680e-01 -5.94899952e-01 1.30236551e-01
-1.63021132e-01 -7.36774743e-01 2.92833209e-01 -5.05598336e-02
-1.16457951e+00 -6.74430013e-01 -7.78580904e-02 2.48643622e-01
-5.25813997e-02 9.19741511e-01 -2.52890944e-01 -2.27679849e-01
8.81032646e-01 2.90911764e-01 -1.29378349e-01 -4.06761199e-01
-6.31103396e-01 6.70005918e-01 4.99137014e-01 -3.97112310e-01
7.54225433e-01 2.95387954e-01 3.67981732e-01 -1.90146983e+00
-5.17128646e-01 -3.86834562e-01 2.28809845e-02 -4.50562656e-01
5.55413008e-01 -4.44025248e-01 -4.76664811e-01 7.92304754e-01
-1.31323862e+00 -1.47108763e-01 1.55389458e-01 1.08175004e+00
-4.40129489e-01 8.07834983e-01 -3.56866091e-01 -1.44390631e+00
-5.02393186e-01 -5.37597060e-01 5.41134477e-01 3.28414440e-01
7.57652447e-02 -7.22281754e-01 1.13531224e-01 -1.47257954e-01
4.67182189e-01 5.18417120e-01 9.62333918e-01 -1.40553087e-01
5.96902184e-02 -6.68245316e-01 -1.19164832e-01 -1.25921547e-01
2.30235934e-01 3.95187497e-01 -6.63218081e-01 -1.86919719e-01
5.49673855e-01 2.32507467e-01 3.47601891e-01 9.87589061e-01
5.39622843e-01 2.83782114e-03 -7.87086666e-01 7.08657920e-01
1.67853332e+00 6.30412281e-01 1.27790713e+00 7.32040226e-01
1.70240715e-01 7.79823303e-01 1.26434422e+00 8.01740587e-01
1.73174068e-01 3.94224495e-01 7.97187686e-01 -1.27102956e-01
6.66516364e-01 2.87881225e-01 1.47524282e-01 6.11107767e-01
-7.59063005e-01 -4.72949624e-01 -1.04188013e+00 3.06051344e-01
-1.30248129e+00 -1.15681911e+00 -1.06729066e+00 2.53994942e+00
-4.80464175e-02 -2.20804274e-01 -5.49184859e-01 6.16851330e-01
1.03139412e+00 -3.62108737e-01 -1.41368983e-02 -4.78526086e-01
-8.11457783e-02 2.38332614e-01 6.61750317e-01 3.96939486e-01
-4.78003860e-01 1.68584734e-01 5.61592913e+00 4.87735808e-01
-1.19316840e+00 -3.73058736e-01 -2.21412435e-01 9.79690373e-01
-3.36878240e-01 1.49573060e-02 -6.71746254e-01 2.30974764e-01
3.67738545e-01 -1.00179285e-01 -2.03619674e-01 6.27701342e-01
3.44903499e-01 -5.85992217e-01 -1.91629931e-01 1.02475441e+00
-4.74251986e-01 -4.57845628e-01 -3.30286175e-01 3.69322151e-02
4.69358638e-02 -3.09087008e-01 -2.33031541e-01 1.62121356e-01
-2.33023152e-01 -6.50097907e-01 5.36753833e-01 6.99330389e-01
6.16253078e-01 -7.82660246e-01 1.03323245e+00 5.75552106e-01
-1.45633018e+00 1.18459769e-01 -8.31119239e-01 -1.61920041e-01
5.18655598e-01 1.15929317e+00 -8.46360028e-01 1.16374946e+00
8.85814130e-01 1.17669441e-01 8.37992951e-02 1.55451000e+00
-1.34033978e-01 7.33475626e-01 -7.62080669e-01 2.04490479e-02
2.08212763e-01 -6.44000828e-01 7.14636147e-01 9.40487266e-01
1.47686768e+00 7.59387553e-01 -1.00101247e-01 6.51745975e-01
9.21999395e-01 7.30367005e-01 -7.74619818e-01 3.25730681e-01
5.39060831e-01 1.21857524e+00 -7.43598104e-01 2.57600486e-01
-3.29548627e-01 4.05821323e-01 -7.32281148e-01 4.74451691e-01
-7.62586892e-01 -7.24818349e-01 8.06843266e-02 9.70237792e-01
2.37879187e-01 -4.61605303e-02 -6.42667115e-02 -5.16135216e-01
2.09251285e-01 -2.32357129e-01 1.55468270e-01 -1.08106053e+00
-9.37455177e-01 6.52508438e-01 2.16818005e-01 -1.43485701e+00
1.76012725e-01 -4.16867733e-01 -1.30111587e+00 1.00529206e+00
-1.24504447e+00 -7.95010090e-01 -8.38768721e-01 3.13914388e-01
-1.49883345e-01 1.36790648e-01 1.03077459e+00 2.44347125e-01
-3.05117756e-01 8.85944664e-02 4.50243056e-01 -4.25535262e-01
2.79571153e-02 -5.72031438e-01 -4.24894542e-02 6.59203053e-01
-7.62961924e-01 4.10394296e-02 1.31758511e+00 -1.00282729e+00
-1.23390639e+00 -6.71748877e-01 5.06367385e-01 7.13671744e-01
2.75179565e-01 1.45318031e-01 -1.29087126e+00 2.91308999e-01
-4.10693467e-01 -5.03611006e-02 5.56461692e-01 -9.42038149e-02
2.81768411e-01 -1.16706397e-02 -1.42486668e+00 2.81786710e-01
4.34686095e-01 4.31842238e-01 -5.68833053e-01 2.13762522e-01
-1.95087940e-01 -1.88499495e-01 -6.27655685e-01 1.07446384e+00
7.74197459e-01 -1.05896521e+00 7.99261630e-01 3.00176293e-01
-1.32088186e-02 -1.94203809e-01 -5.82765400e-01 -1.35743284e+00
-6.51169538e-01 -8.27701539e-02 5.40061891e-01 8.39635551e-01
1.35108873e-01 -1.13452625e+00 6.06371105e-01 -1.82287902e-01
-4.14670736e-01 -8.01746428e-01 -8.43128681e-01 -8.19085598e-01
-3.15520048e-01 -1.53160349e-01 2.38549799e-01 8.12565148e-01
6.66053817e-02 6.12509921e-02 1.93335831e-01 1.01886320e+00
9.77794766e-01 2.50845253e-01 3.99380445e-01 -1.89196587e+00
-1.01683512e-02 2.43761733e-01 -6.08016551e-01 -7.48232424e-01
-3.63275558e-01 -3.45136523e-01 4.03648317e-01 -2.04627037e+00
-2.54009753e-01 -9.05790508e-01 2.58639544e-01 -1.05274193e-01
2.03564584e-01 -1.82791576e-01 -4.83116090e-01 1.93849072e-01
6.47998631e-01 6.06809616e-01 1.21943796e+00 2.58671343e-01
-3.21180105e-01 7.88510799e-01 -1.47151172e-01 1.17022347e+00
6.85123324e-01 -3.80874842e-01 -3.51361066e-01 -1.34110898e-01
3.23518872e-01 6.65071249e-01 2.41372496e-01 -1.19808006e+00
-5.54920267e-03 -6.73275515e-02 1.68841287e-01 -1.20889270e+00
3.86758417e-01 -9.38997269e-01 5.70000589e-01 7.32393444e-01
9.94817257e-01 -1.96665436e-01 2.04631671e-01 4.12283361e-01
-1.82076901e-01 -7.21225619e-01 1.06132078e+00 4.92298640e-02
-4.76473838e-01 6.99467957e-02 -7.00172067e-01 -3.78673494e-01
1.05245864e+00 -8.65207791e-01 -1.71189472e-01 -4.73535061e-01
-5.23478329e-01 -1.80373877e-01 1.85503349e-01 -4.82219815e-01
8.81716847e-01 -1.14905488e+00 -9.15015578e-01 2.58602709e-01
4.91698310e-02 1.02396272e-01 6.98820651e-01 1.01608896e+00
-9.89973009e-01 1.38789248e-02 -2.53497362e-01 -5.55606365e-01
-1.15418649e+00 -1.04233734e-02 5.95301628e-01 -1.45539930e-02
-9.63552773e-01 5.15166044e-01 4.66201425e-01 -1.14562102e-01
-4.10593182e-01 -2.52455294e-01 -5.95529854e-01 1.33386344e-01
4.99329686e-01 8.01874518e-01 4.37630147e-01 -5.48720777e-01
-3.14129233e-01 1.13068175e+00 4.02763754e-01 9.44433361e-02
1.50720668e+00 1.60630614e-01 -1.02186039e-01 2.86060393e-01
5.32279968e-01 4.54543173e-01 -4.43042636e-01 -7.78062269e-02
-3.32828045e-01 -4.79258448e-01 9.42584351e-02 -1.32618338e-01
-7.19920516e-01 6.69732511e-01 2.39516377e-01 6.79420054e-01
1.06636298e+00 -5.24921536e-01 6.01096690e-01 5.51436782e-01
8.18500817e-01 -9.05392945e-01 -4.04078424e-01 5.14937103e-01
1.06709576e+00 -2.49525249e-01 1.41834617e-01 -1.14742720e+00
-1.54484674e-01 1.48789203e+00 3.17370474e-01 -2.42402539e-01
1.05372047e+00 7.00672090e-01 -1.01752609e-01 -2.96810955e-01
4.76275869e-02 4.29815706e-03 -4.07606900e-01 7.87747383e-01
2.07590297e-01 1.61140367e-01 -6.62155092e-01 8.89324009e-01
-3.00611466e-01 -1.62977129e-01 9.31214869e-01 1.28618014e+00
-1.09262049e+00 -4.08178717e-01 -1.17168069e+00 4.45844084e-01
-5.87176323e-01 9.36742350e-02 4.68714029e-01 9.93792892e-01
-2.93981284e-01 8.40869546e-01 2.21542511e-02 -1.03327595e-01
8.71848285e-01 -2.70541072e-01 6.01612270e-01 -4.53441590e-01
4.53106523e-01 4.19139713e-02 2.35437497e-01 -5.38935848e-02
-5.28480262e-02 -7.53462136e-01 -1.70440984e+00 -1.60818234e-01
-8.74063134e-01 4.89976197e-01 8.55491936e-01 7.07984686e-01
-3.22575212e-01 3.09855580e-01 8.81284595e-01 -7.97200322e-01
-7.41554379e-01 -1.02420843e+00 -1.72554719e+00 -2.74184287e-01
-2.15100512e-01 -1.32374942e+00 -7.41459966e-01 -5.27511537e-01]
|
[6.810297966003418, 1.404147744178772]
|
ee242d53-db4b-43b2-8b4f-af54eff13e27
|
final-adaptation-reinforcement-learning-for-n
|
2111.14375
| null |
https://arxiv.org/abs/2111.14375v1
|
https://arxiv.org/pdf/2111.14375v1.pdf
|
Final Adaptation Reinforcement Learning for N-Player Games
|
This paper covers n-tuple-based reinforcement learning (RL) algorithms for games. We present new algorithms for TD-, SARSA- and Q-learning which work seamlessly on various games with arbitrary number of players. This is achieved by taking a player-centered view where each player propagates his/her rewards back to previous rounds. We add a new element called Final Adaptation RL (FARL) to all these algorithms. Our main contribution is that FARL is a vitally important ingredient to achieve success with the player-centered view in various games. We report results on seven board games with 1, 2 and 3 players, including Othello, ConnectFour and Hex. In most cases it is found that FARL is important to learn a near-perfect playing strategy. All algorithms are available in the GBG framework on GitHub.
|
['Samineh Bagheri', 'Wolfgang Konen']
|
2021-11-29
| null | null | null | null |
['board-games']
|
['playing-games']
|
[-4.56839323e-01 1.00351609e-01 -3.00405264e-01 2.95149237e-01
-9.63078082e-01 -8.26271355e-01 2.17433736e-01 -4.61367704e-02
-6.19730115e-01 1.27897191e+00 9.52662975e-02 -2.64192641e-01
-5.84464490e-01 -9.69714403e-01 -5.13951540e-01 -6.76122069e-01
-5.29202878e-01 7.57476091e-01 4.37563866e-01 -9.31382000e-01
3.30959260e-01 2.35386528e-02 -1.41948247e+00 2.03656331e-01
6.09656870e-01 7.17205405e-01 -1.27644762e-02 1.24131775e+00
4.38064113e-02 1.35836172e+00 -5.96398830e-01 -6.64157748e-01
7.56002188e-01 -7.58486867e-01 -1.05455470e+00 -1.35872811e-01
-5.52796662e-01 -5.09806335e-01 -1.31317928e-01 6.61458254e-01
8.92578185e-01 2.80966610e-01 2.18041524e-01 -1.35821486e+00
1.01187184e-01 1.15848291e+00 -6.18111432e-01 1.39734581e-01
5.96745253e-01 3.59218538e-01 1.34048676e+00 -2.96726257e-01
4.64851677e-01 9.89864886e-01 4.55765069e-01 5.88804364e-01
-6.82671547e-01 -4.88331050e-01 1.26714706e-01 3.23401272e-01
-8.29120517e-01 4.03823666e-02 5.45749843e-01 1.01516701e-01
9.48945761e-01 3.73701155e-01 1.30384576e+00 6.53084040e-01
2.06546694e-01 1.16874886e+00 1.32578444e+00 -5.40466547e-01
6.53811634e-01 -2.89531469e-01 -4.27327335e-01 4.44128305e-01
9.25827697e-02 5.58957219e-01 -8.04324627e-01 -2.01159656e-01
1.05262148e+00 -3.99545997e-01 3.78442317e-01 -6.55116558e-01
-6.84994042e-01 9.76551414e-01 -4.45675701e-02 -1.19365871e-01
-6.22989297e-01 5.75721085e-01 2.60603666e-01 9.20320749e-01
1.41615927e-01 7.60231078e-01 -4.87596273e-01 -1.14204931e+00
-5.49407721e-01 5.47965229e-01 1.01800954e+00 8.01884651e-01
6.74483836e-01 1.41925544e-01 -1.99260265e-01 5.92403054e-01
-1.19337469e-01 9.23918039e-02 2.90039539e-01 -1.67373574e+00
2.55592525e-01 3.90526295e-01 6.28615081e-01 -4.19100016e-01
-4.45629328e-01 -6.20619297e-01 -4.21736538e-01 6.70341074e-01
3.43463451e-01 -6.89173281e-01 -4.02830631e-01 1.66565526e+00
3.54320765e-01 2.71667182e-01 2.74676412e-01 5.57803988e-01
6.91415846e-01 3.54705542e-01 -2.29079753e-01 -3.55748117e-01
8.52549732e-01 -1.04594898e+00 -4.19127613e-01 -9.83159840e-02
6.44942284e-01 -4.61299658e-01 9.10422206e-01 8.75146389e-01
-1.82389700e+00 -1.34743780e-01 -8.52340639e-01 6.52661443e-01
-1.10577531e-01 -4.56086189e-01 9.02977407e-01 1.00086069e+00
-1.34443533e+00 6.88586295e-01 -5.92207253e-01 -2.74651289e-01
2.73896277e-01 8.28208804e-01 1.05247553e-02 6.75262287e-02
-1.29151285e+00 7.49915659e-01 5.56216836e-01 -5.66043973e-01
-9.93576109e-01 -3.47072631e-01 -3.63306493e-01 -1.62257180e-01
1.12126303e+00 -7.08827257e-01 1.76766002e+00 -8.43634307e-01
-2.23090124e+00 6.15927577e-01 4.49253142e-01 -5.99372208e-01
7.47499526e-01 -1.09030314e-01 2.16048539e-01 -4.83381450e-02
1.40855297e-01 5.53721368e-01 4.70948458e-01 -1.21650231e+00
-1.14243519e+00 -1.13014385e-01 7.34679699e-01 1.09732866e+00
-5.25509287e-03 6.15208596e-03 -4.32278663e-01 -2.73089647e-01
-1.44465208e-01 -6.96837127e-01 -7.27125645e-01 -7.49578476e-01
-6.29262999e-02 -4.37727809e-01 7.14448616e-02 -1.12421341e-01
1.28861833e+00 -1.53013146e+00 2.85129786e-01 2.11423486e-01
2.93320805e-01 1.83262736e-01 -2.45453805e-01 1.14545810e+00
2.95355171e-01 2.83796452e-02 4.24473494e-01 -1.19296543e-01
1.01141207e-01 5.41688502e-01 3.00829589e-01 3.13755497e-02
-3.39985669e-01 1.10081398e+00 -1.07609379e+00 -1.91583693e-01
1.02659278e-01 -5.62561572e-01 -9.82724428e-01 4.05058324e-01
-3.32949638e-01 4.29318875e-01 -5.09498656e-01 4.23295200e-01
4.05679792e-01 2.49718763e-02 3.25670689e-01 8.46100390e-01
-2.70573914e-01 2.98105329e-01 -1.76946867e+00 1.63942742e+00
-7.93380961e-02 -2.46326640e-01 1.44851446e-01 -9.48925912e-01
6.56146288e-01 3.60000074e-01 8.35800529e-01 -9.33536232e-01
2.09996730e-01 1.88686356e-01 1.39274821e-01 -2.76743680e-01
7.22977638e-01 -1.35278136e-01 -4.46479589e-01 1.01474285e+00
5.15263490e-02 -2.73017019e-01 6.08930707e-01 3.98175895e-01
1.23095965e+00 1.29579738e-01 6.05098307e-01 -5.83081171e-02
1.87481746e-01 6.70681000e-02 7.31289387e-01 1.45703828e+00
-3.09040457e-01 1.26835614e-01 1.03160036e+00 -4.71481234e-01
-8.66594970e-01 -9.53537226e-01 7.63759732e-01 1.74208891e+00
7.63361901e-02 -6.54798150e-01 -5.04655421e-01 -3.65502715e-01
-1.93138212e-01 5.05790412e-01 -7.96636045e-01 -1.68502569e-01
-4.11959797e-01 -4.11346316e-01 5.15143275e-01 4.87608105e-01
6.34442449e-01 -1.26768804e+00 -8.02340448e-01 5.25277793e-01
-2.50133798e-02 -5.54798663e-01 -3.02587390e-01 5.03567636e-01
-7.63909161e-01 -1.03024435e+00 -4.85424846e-01 -4.19754952e-01
-1.90532550e-01 2.20859855e-01 1.33057785e+00 -7.49673462e-03
1.08914874e-01 6.92257047e-01 -7.96411574e-01 -5.56993723e-01
-2.85728931e-01 2.20912367e-01 -5.40603995e-02 -5.41662097e-01
5.46314940e-02 -7.29915619e-01 -6.02088094e-01 3.26959610e-01
-4.92580712e-01 -8.79176185e-02 3.20552766e-01 7.92306721e-01
4.09843415e-01 2.77258784e-01 5.01281381e-01 -9.53755796e-01
1.11876988e+00 -4.10823673e-01 -5.71328044e-01 6.05203547e-02
-3.33548605e-01 -7.02941865e-02 5.86623967e-01 -1.36608869e-01
-6.30794525e-01 -1.63793132e-01 -2.03742847e-01 4.36213873e-02
5.83126210e-02 4.11557853e-01 4.71435711e-02 5.32350168e-02
7.00191617e-01 2.84924269e-01 1.17981575e-01 7.54877739e-03
5.88036954e-01 4.15819079e-01 5.24895191e-02 -8.93497884e-01
5.02856255e-01 7.83633217e-02 -2.05476359e-02 -2.32753903e-01
-5.40382624e-01 -5.34909725e-01 -3.55444700e-01 -4.88544613e-01
3.23085457e-01 -9.36562538e-01 -1.58136904e+00 7.27122724e-01
-4.59222436e-01 -8.69126320e-01 -9.07102823e-01 3.43961924e-01
-1.21943247e+00 7.85094202e-02 -6.51148558e-01 -1.08664131e+00
-2.49050379e-01 -9.59511459e-01 3.16352814e-01 5.93600869e-01
1.16053782e-01 -1.01741552e+00 5.06271422e-01 4.08493578e-01
3.79161149e-01 -5.66164702e-02 5.30246079e-01 -6.62337899e-01
-4.88701254e-01 2.19590738e-01 4.41354066e-01 -1.65212322e-02
-1.22806914e-02 -2.81979173e-01 -3.98912728e-01 -5.82561135e-01
-2.29603142e-01 -9.22286391e-01 4.98406947e-01 5.79110324e-01
6.65303886e-01 -8.71939436e-02 2.12512016e-01 3.57205302e-01
1.47366846e+00 6.20548487e-01 7.47059047e-01 7.81249285e-01
1.43750787e-01 2.54099280e-01 9.44919407e-01 1.17935050e+00
6.65196598e-01 6.26033604e-01 7.77256429e-01 1.42896682e-01
3.25668156e-01 -2.99907655e-01 4.39440608e-01 7.72749782e-01
-5.95730662e-01 -5.03579378e-01 -6.96262658e-01 3.30690295e-01
-2.28791523e+00 -1.19009054e+00 1.09679207e-01 2.23600769e+00
8.91287565e-01 2.93934256e-01 1.01209116e+00 2.91591197e-01
3.27862412e-01 -5.49820028e-02 -7.54657626e-01 -8.37316096e-01
-1.42564446e-01 6.23660505e-01 6.17827237e-01 6.81612253e-01
-7.25225925e-01 1.41118979e+00 7.09780931e+00 1.13433003e+00
-4.75605249e-01 1.76938206e-01 6.04790390e-01 -4.49748278e-01
-1.95229232e-01 6.91758990e-02 -5.33028126e-01 1.38042301e-01
8.07298958e-01 -4.04215544e-01 9.26370144e-01 7.02864110e-01
4.57289249e-01 -6.11741066e-01 -6.51412010e-01 8.64384055e-01
-4.49468225e-01 -1.40747833e+00 -3.42712730e-01 2.07613528e-01
9.99327362e-01 -1.24942638e-01 1.66486353e-01 6.84642673e-01
1.46691000e+00 -1.14016163e+00 4.94530708e-01 2.19061434e-01
6.67214632e-01 -1.50480556e+00 7.25404263e-01 5.29863656e-01
-1.11544049e+00 -4.03106958e-01 -3.92110735e-01 -7.38395631e-01
-2.32147828e-01 -9.27974954e-02 -7.05378890e-01 8.81162167e-01
6.97145343e-01 5.81612885e-01 -1.77305117e-01 1.22285771e+00
-3.08943987e-01 6.86595082e-01 -2.40687728e-01 -3.22373241e-01
6.57409072e-01 -3.72962028e-01 3.21155787e-01 4.31581408e-01
2.02043235e-01 5.24097264e-01 4.34899777e-01 2.93255091e-01
8.09989795e-02 1.17841542e-01 -3.94254357e-01 1.39508024e-01
5.02617836e-01 1.14625967e+00 -5.99089324e-01 1.22607097e-01
-8.86678025e-02 8.78715277e-01 4.35497403e-01 2.37503760e-02
-6.97627485e-01 -1.27442867e-01 8.32068443e-01 -7.08767772e-02
4.07251567e-01 -2.76164021e-02 -6.45813420e-02 -5.94165146e-01
-3.76926750e-01 -1.28252816e+00 7.30483830e-01 -6.02400661e-01
-1.09564877e+00 2.27723718e-01 -2.89255142e-01 -1.25298750e+00
-7.34366715e-01 -3.17018837e-01 -8.21732700e-01 4.71842498e-01
-1.21821404e+00 -8.36588442e-01 9.66826007e-02 1.00122702e+00
4.05368090e-01 -5.65627217e-01 8.28254104e-01 -1.24607407e-01
-2.83772498e-01 7.60863364e-01 3.95952344e-01 -1.42351195e-01
3.88406605e-01 -1.67275524e+00 2.20710218e-01 4.07797128e-01
5.86563163e-02 -6.62050620e-02 6.05391860e-01 -4.47015584e-01
-1.53109455e+00 -4.38197315e-01 8.36287737e-02 -3.25820833e-01
6.09305263e-01 1.98431523e-03 -1.00537352e-01 4.99950260e-01
4.99681681e-01 -3.35873157e-01 8.59680891e-01 3.32763106e-01
3.30134064e-01 -8.33679512e-02 -9.25022781e-01 6.97601080e-01
1.17590451e+00 -2.17666515e-04 -1.75162584e-01 2.07226172e-01
4.59812224e-01 -9.20097470e-01 -7.04383016e-01 -8.75040591e-02
2.79731214e-01 -1.48948479e+00 8.36216927e-01 -9.93055284e-01
3.21863353e-01 -4.38528918e-02 6.24088123e-02 -1.71594346e+00
-4.88348812e-01 -1.37047029e+00 -1.56015912e-02 6.76455081e-01
1.49034575e-01 -5.74867666e-01 1.41425979e+00 1.98664457e-01
1.20610222e-01 -9.58361566e-01 -1.09162915e+00 -8.13822746e-01
5.57175100e-01 -6.51963413e-01 6.85258269e-01 4.97752190e-01
4.42954302e-01 2.18263224e-01 -9.66951668e-01 -2.89020270e-01
5.13816237e-01 -6.42595738e-02 1.03347993e+00 -9.91066992e-01
-9.63545620e-01 -3.88105094e-01 -2.09022880e-01 -1.18749535e+00
-4.97420549e-01 -4.89293188e-01 -2.76560009e-01 -1.60688186e+00
2.55098730e-01 -6.38524055e-01 -4.80787694e-01 4.18706506e-01
-5.35572134e-02 1.25878975e-01 3.83942127e-01 -8.84208009e-02
-1.43731248e+00 5.64328372e-01 1.66381252e+00 4.68035489e-01
-5.34160793e-01 7.22685635e-01 -9.72502768e-01 4.79407519e-01
1.20901752e+00 -4.88287330e-01 -6.59092486e-01 -6.94280043e-02
7.28867412e-01 6.59756184e-01 -1.45670474e-01 -8.82982373e-01
4.43593204e-01 -5.65996468e-01 -8.81545916e-02 -4.12616342e-01
4.62987900e-01 -2.40676686e-01 -6.03346759e-03 7.40627050e-01
-2.71556735e-01 3.42116863e-01 7.26222321e-02 4.67160910e-01
8.11635517e-03 -2.82325566e-01 5.04754841e-01 -6.94689512e-01
-6.23685837e-01 2.01890677e-01 -6.82980418e-01 6.56813383e-01
1.33294666e+00 -4.72234815e-01 -1.01935118e-01 -1.27780426e+00
-8.90121818e-01 6.17456794e-01 2.34881192e-01 -6.23807944e-02
4.36312109e-01 -1.21083963e+00 -8.89685810e-01 -9.20229778e-02
-2.19931692e-01 -3.49043846e-01 2.62444526e-01 6.17443442e-01
-4.47012752e-01 2.18357816e-01 -6.59825981e-01 -5.75875305e-02
-1.31729078e+00 1.35918260e-01 6.90935910e-01 -1.01243830e+00
-1.54556602e-01 1.22067750e+00 -4.93829906e-01 -5.54957688e-01
3.80699962e-01 2.55987227e-01 -2.23505452e-01 -1.77097935e-02
2.66174823e-01 4.94456351e-01 -2.08561450e-01 1.09802922e-02
-5.14614508e-02 2.95327604e-01 -1.01122275e-01 -5.48539162e-01
1.56386495e+00 4.18730900e-02 9.08860862e-02 3.00370961e-01
1.99777350e-01 -1.04716010e-01 -1.48800063e+00 -2.44757950e-01
-2.52528816e-01 -3.18119645e-01 -2.13824898e-01 -8.62840414e-01
-1.07775879e+00 4.83010560e-01 4.24541324e-01 4.13329184e-01
1.14086485e+00 -6.79464266e-02 5.08994818e-01 3.32406789e-01
1.06727731e+00 -1.46643591e+00 6.57959998e-01 9.06107306e-01
3.84100497e-01 -1.05466378e+00 -7.27138594e-02 1.07303336e-01
-1.13514614e+00 8.89691830e-01 7.71797657e-01 -2.36603007e-01
2.90126354e-01 3.53963077e-01 -1.41968831e-01 -7.30151460e-02
-1.35526979e+00 -7.85619497e-01 -5.13859689e-01 8.76240134e-01
1.04074620e-01 2.76979446e-01 -4.16549504e-01 7.54672170e-01
-4.28710401e-01 4.96542938e-02 1.01924562e+00 1.24424529e+00
-7.24043787e-01 -1.96535957e+00 -2.35099927e-01 4.53506440e-01
-5.94953477e-01 -2.04648264e-02 -3.76151413e-01 8.44511509e-01
6.28767461e-02 1.21930170e+00 -1.08507365e-01 -6.39232755e-01
2.81686425e-01 -4.09186393e-01 7.15688229e-01 -6.56203866e-01
-1.15102601e+00 -3.40588018e-02 1.68130174e-01 -8.18832040e-01
-2.86718518e-01 -6.46095932e-01 -1.24118912e+00 -8.34912181e-01
5.05815409e-02 5.17698526e-01 -7.97434896e-03 7.72353888e-01
1.07385077e-01 4.27850306e-01 8.04369092e-01 -4.29657340e-01
-7.38631368e-01 -5.82839608e-01 -1.05047607e+00 -4.94798534e-02
-2.53423870e-01 -6.04534328e-01 5.53331636e-02 -7.40967870e-01]
|
[3.5775461196899414, 1.5411176681518555]
|
c15dc6b6-6493-48c1-b66d-940dd8d67548
|
learning-audio-driven-viseme-dynamics-for-3d
|
2301.06059
| null |
https://arxiv.org/abs/2301.06059v1
|
https://arxiv.org/pdf/2301.06059v1.pdf
|
Learning Audio-Driven Viseme Dynamics for 3D Face Animation
|
We present a novel audio-driven facial animation approach that can generate realistic lip-synchronized 3D facial animations from the input audio. Our approach learns viseme dynamics from speech videos, produces animator-friendly viseme curves, and supports multilingual speech inputs. The core of our approach is a novel parametric viseme fitting algorithm that utilizes phoneme priors to extract viseme parameters from speech videos. With the guidance of phonemes, the extracted viseme curves can better correlate with phonemes, thus more controllable and friendly to animators. To support multilingual speech inputs and generalizability to unseen voices, we take advantage of deep audio feature models pretrained on multiple languages to learn the mapping from audio to viseme curves. Our audio-to-curves mapping achieves state-of-the-art performance even when the input audio suffers from distortions of volume, pitch, speed, or noise. Lastly, a viseme scanning approach for acquiring high-fidelity viseme assets is presented for efficient speech animation production. We show that the predicted viseme curves can be applied to different viseme-rigged characters to yield various personalized animations with realistic and natural facial motions. Our approach is artist-friendly and can be easily integrated into typical animation production workflows including blendshape or bone based animation.
|
['Di Kang', 'Xuefei Zhe', 'Changhai Chen', 'Tangli Xue', 'Yue Qian', 'Haoxian Zhang', 'Linchao Bao']
|
2023-01-15
| null | null | null | null |
['3d-face-animation']
|
['computer-vision']
|
[ 1.21271203e-03 1.30872250e-01 5.55735491e-02 -9.08410251e-02
-9.77072239e-01 -5.26741505e-01 4.51375544e-01 -5.04343271e-01
1.75787106e-01 1.72311410e-01 3.00801277e-01 -1.94933526e-02
2.94940948e-01 -3.87467653e-01 -7.68075824e-01 -5.14912069e-01
-1.48728773e-01 5.04414737e-01 1.16745241e-01 -5.18749595e-01
-1.55681282e-01 8.18993747e-01 -1.98571968e+00 3.41104090e-01
5.70672214e-01 7.03471482e-01 5.17736226e-02 1.29456162e+00
-1.44623563e-01 2.71246850e-01 -7.48501360e-01 -4.40202594e-01
3.12162787e-01 -4.18805033e-01 -1.89449236e-01 8.87235999e-02
6.24746740e-01 -4.90304232e-01 -2.78774470e-01 6.67917550e-01
6.53459430e-01 -8.75560269e-02 8.46779406e-01 -1.49082410e+00
-4.05442387e-01 5.33189118e-01 -5.47479093e-01 -2.68956333e-01
5.31408131e-01 4.43517268e-01 7.79256225e-01 -8.58580172e-01
7.38427341e-01 1.70321476e+00 4.52074170e-01 8.00572634e-01
-1.18038177e+00 -9.27246094e-01 -1.19672351e-01 1.20159343e-01
-1.24951005e+00 -6.90947354e-01 9.81055319e-01 -4.60391015e-01
6.93148434e-01 4.08192754e-01 1.28386271e+00 1.18961143e+00
1.39208898e-01 6.75932586e-01 7.15721011e-01 -4.00546044e-01
-1.86363623e-01 7.22111762e-02 -9.02488768e-01 8.98193240e-01
-7.04592049e-01 3.47961485e-01 -7.69183934e-01 -1.22640893e-01
1.06853724e+00 -7.21811831e-01 -2.05596283e-01 -1.64335325e-01
-1.09884346e+00 5.80718994e-01 -9.42010656e-02 -5.88162243e-02
-6.11904636e-02 3.82095009e-01 4.09089774e-01 1.59197509e-01
2.62260348e-01 1.75894529e-01 -3.26356262e-01 -4.95649725e-01
-1.23231041e+00 4.50389326e-01 6.11997068e-01 1.07218397e+00
6.00140691e-01 7.79393733e-01 1.40989234e-03 8.62593412e-01
4.18835133e-01 1.05673432e+00 2.64718711e-01 -1.33644104e+00
5.61190918e-02 1.40611967e-02 -2.06718072e-01 -8.21235597e-01
-4.01345968e-01 9.33639780e-02 -3.76116961e-01 4.61374551e-01
2.75166720e-01 -1.59897119e-01 -8.89667273e-01 1.80528998e+00
4.79655117e-01 5.41387081e-01 -2.63319999e-01 8.93374562e-01
9.12963033e-01 9.24264550e-01 5.37960865e-02 -2.52025485e-01
1.32380795e+00 -6.80568159e-01 -8.75742376e-01 3.06856781e-01
3.09673250e-01 -1.09168875e+00 1.52606452e+00 4.84771639e-01
-1.41875041e+00 -5.88852525e-01 -7.68620968e-01 -1.92518130e-01
3.84420082e-02 1.55259401e-01 3.54836732e-01 7.83012390e-01
-1.34528220e+00 3.75979632e-01 -7.72312999e-01 1.18298650e-01
4.28650491e-02 4.26164299e-01 -2.53725052e-01 6.70585811e-01
-1.05678165e+00 5.57002544e-01 -9.31322109e-03 -2.17039153e-01
-1.09991825e+00 -1.02582037e+00 -1.07388651e+00 -1.67687777e-02
3.35383974e-02 -5.88412583e-01 1.42588520e+00 -1.33121407e+00
-2.48195672e+00 7.10403204e-01 -1.24859042e-01 -3.17127369e-02
6.10885084e-01 6.19921014e-02 -7.30094671e-01 5.61543524e-01
-4.81174469e-01 1.39633262e+00 1.45536780e+00 -1.21285975e+00
-3.68273020e-01 2.33960956e-01 -3.42911571e-01 1.45473719e-01
-3.50099564e-01 2.87029535e-01 -7.61059344e-01 -9.26489711e-01
-4.91030782e-01 -1.02097666e+00 3.42096716e-01 4.86461103e-01
-2.73429334e-01 2.20937002e-02 1.14404690e+00 -8.05802047e-01
1.07252324e+00 -2.06125927e+00 2.72044927e-01 2.89633989e-01
3.01996828e-03 1.80936471e-01 -5.38425863e-01 2.08041966e-01
-1.67508930e-01 -6.18989617e-02 1.20298095e-01 -3.81271005e-01
-1.14541380e-02 -1.67201489e-01 -2.77779132e-01 3.14700067e-01
1.52878910e-01 8.60436678e-01 -6.85133100e-01 -7.50604093e-01
3.44377339e-01 1.01824296e+00 -9.71795857e-01 2.74526089e-01
-4.74156499e-01 7.11600304e-01 1.22680187e-01 7.55611539e-01
5.17298579e-01 3.87758940e-01 -2.64327526e-02 -3.15644830e-01
-2.58141588e-02 6.70113042e-02 -1.02111351e+00 1.70255053e+00
-8.03997934e-01 1.06270635e+00 5.12619376e-01 -2.51092166e-01
1.07537079e+00 5.17517447e-01 6.44237101e-01 -3.30862075e-01
3.26385945e-01 1.70038193e-01 -6.01274744e-02 -5.42024493e-01
4.44771528e-01 -1.26786754e-01 -4.79075918e-03 4.32954311e-01
2.73256570e-01 -8.82212937e-01 -1.56820402e-01 -2.17289142e-02
2.62586534e-01 4.35602129e-01 -5.11711240e-02 -1.00166380e-01
5.56030571e-01 -3.73528123e-01 2.83084422e-01 -1.88147187e-01
3.57058318e-03 5.13714671e-01 4.65124547e-01 -1.14841282e-01
-1.29061151e+00 -1.29574084e+00 8.47815424e-02 1.35021830e+00
-3.12025011e-01 -5.33984482e-01 -1.14601302e+00 -8.42867643e-02
-1.56262368e-01 5.30821443e-01 -4.48165953e-01 4.36091572e-02
-8.90148520e-01 2.63520986e-01 8.30433309e-01 3.23794484e-01
-9.16643441e-02 -1.22983205e+00 -3.35828424e-01 2.32691709e-02
-6.41234964e-02 -1.10425186e+00 -1.03331006e+00 -6.43353045e-01
-4.30798560e-01 -7.12891459e-01 -8.88895690e-01 -9.77481186e-01
2.58382291e-01 -1.82169899e-01 7.59593725e-01 -1.00609146e-01
-3.96717459e-01 6.56991839e-01 1.05307229e-01 -6.40521824e-01
-1.09096348e+00 -2.04572320e-01 4.80624437e-01 -2.39456110e-02
-3.08642656e-01 -7.62667179e-01 -4.97133911e-01 3.50234926e-01
-7.19483495e-01 3.61110449e-01 -7.10681155e-02 4.96948302e-01
6.49273694e-01 -3.60923886e-01 4.48072881e-01 -3.24819446e-01
5.36598265e-01 -2.00741757e-02 -7.92057931e-01 6.09058701e-03
-5.00299372e-02 -3.38130929e-02 1.02663183e+00 -1.00148416e+00
-9.94233668e-01 1.04496807e-01 -4.88381088e-01 -1.07735455e+00
-1.41816005e-01 -9.78766456e-02 -4.16944385e-01 -9.61487144e-02
6.11578763e-01 1.25400618e-01 5.24224460e-01 -2.01424271e-01
7.80497193e-01 8.31892908e-01 9.69050467e-01 -5.99744499e-01
9.14146543e-01 4.31695193e-01 2.15467587e-02 -1.33863163e+00
2.05648050e-01 1.65518403e-01 -4.05441493e-01 -7.87527323e-01
7.15401649e-01 -9.78528857e-01 -1.08867002e+00 7.06107974e-01
-1.06910706e+00 -7.31129766e-01 -2.54826367e-01 3.66780162e-01
-1.15148962e+00 3.13679159e-01 -6.67099774e-01 -6.73955083e-01
-4.42794234e-01 -1.63073385e+00 1.29558587e+00 2.60686845e-01
-4.25946146e-01 -8.91264617e-01 -5.60829928e-03 1.13720171e-01
3.17018479e-01 2.29007974e-01 9.19506013e-01 -8.37399736e-02
-4.85611737e-01 5.27834594e-02 3.09510201e-01 1.70745570e-02
2.76620656e-01 1.07455325e+00 -9.34597075e-01 -1.59373760e-01
-6.27953112e-01 -2.92905807e-01 2.21051648e-01 6.47440374e-01
1.19564378e+00 -5.46330929e-01 4.41126823e-02 9.24427986e-01
5.44071078e-01 -5.52764768e-03 4.69625443e-01 -2.26066321e-01
7.47675002e-01 8.19068730e-01 5.92832029e-01 4.86266285e-01
3.20640773e-01 1.12729001e+00 3.81326854e-01 -6.70558140e-02
-6.12321436e-01 -6.26440942e-01 8.18229854e-01 1.12946546e+00
-1.03962094e-01 -1.12103999e-01 -5.09292543e-01 3.93653542e-01
-1.00048256e+00 -8.84342551e-01 1.64760306e-01 1.94105136e+00
9.68635082e-01 -2.24792853e-01 7.77113259e-01 2.67820759e-03
8.37320745e-01 5.23426086e-02 -4.76632714e-01 -9.43243384e-01
1.14066355e-01 5.14382601e-01 1.13444932e-01 6.13038063e-01
-6.17859185e-01 1.37937462e+00 6.70621252e+00 1.14190924e+00
-1.75798190e+00 -5.01308180e-02 3.38423193e-01 -4.27745461e-01
-7.99327195e-01 -3.29921544e-01 -5.31259298e-01 3.13432187e-01
1.29779077e+00 -2.25310192e-01 8.20988417e-01 6.74108922e-01
7.50997186e-01 4.72799540e-01 -9.73973513e-01 1.20468545e+00
-1.25418693e-01 -1.49753308e+00 3.70251149e-01 -1.08577393e-01
4.92400080e-01 -4.54405457e-01 4.82559502e-01 9.48981047e-02
1.13929860e-01 -1.12644482e+00 1.14418912e+00 3.63825977e-01
1.84636521e+00 -9.40244377e-01 -2.84822017e-01 1.57334611e-01
-1.53546512e+00 1.59997314e-01 -6.29980769e-03 5.68452537e-01
5.15545487e-01 -3.29720646e-01 -1.03282690e+00 8.35242048e-02
4.29369777e-01 4.20257181e-01 -5.63542768e-02 8.03478420e-01
-2.21782550e-01 6.04938209e-01 -4.83275831e-01 -7.18288198e-02
-1.43619031e-01 -6.27002493e-02 8.34869444e-01 1.41355610e+00
6.60362601e-01 -4.21539880e-02 -2.36643806e-01 7.20253229e-01
2.46825572e-02 5.97289145e-01 -5.31050086e-01 -2.49472901e-01
6.40280247e-01 1.18862152e+00 -2.60332704e-01 -1.79358590e-02
-9.99221206e-02 7.69305468e-01 -1.03867248e-01 3.32926393e-01
-9.81812477e-01 -1.68751270e-01 9.72970784e-01 3.74832898e-01
1.47493690e-01 -3.09935182e-01 4.72085830e-03 -7.48655736e-01
-5.10493577e-01 -1.28930306e+00 -5.34994006e-02 -1.00040698e+00
-7.05159843e-01 8.36720169e-01 1.12738863e-01 -1.44074833e+00
-7.12334991e-01 -5.47560275e-01 -7.68784225e-01 6.35192037e-01
-1.25416720e+00 -1.35500574e+00 -2.40041748e-01 9.00063753e-01
7.89020240e-01 -4.26803380e-01 8.81540239e-01 2.00913176e-01
-2.53408760e-01 1.09759426e+00 -5.15435576e-01 3.12457941e-02
8.31390798e-01 -7.84406722e-01 6.00356400e-01 3.39129359e-01
2.69866019e-01 1.49643555e-01 9.53654289e-01 -4.74169403e-01
-1.38474667e+00 -9.42210615e-01 4.25466776e-01 -1.51119858e-01
7.50342429e-01 -5.08751333e-01 -7.25275576e-01 4.00457263e-01
3.04946482e-01 -1.41700223e-01 7.30510294e-01 -5.04140019e-01
-3.83316964e-01 -2.82560289e-01 -1.04221916e+00 9.23984170e-01
9.18507576e-01 -6.32502854e-01 -9.47382450e-02 1.89900711e-01
7.61477470e-01 -8.49316716e-01 -9.53524292e-01 7.81523362e-02
9.36467171e-01 -7.89578974e-01 9.10750270e-01 -4.82483447e-01
2.79986888e-01 -2.73259044e-01 2.28292078e-01 -1.49548900e+00
7.83603191e-02 -1.45114517e+00 -1.74196303e-01 1.21601808e+00
3.72283936e-01 -2.27444932e-01 6.13501966e-01 1.16284363e-01
-2.81813711e-01 -4.50568318e-01 -1.02069628e+00 -5.51212013e-01
2.27329537e-01 -6.70154035e-01 9.02306855e-01 6.55478835e-01
9.29897744e-03 -1.18747875e-01 -6.74640119e-01 1.93278879e-01
4.10223722e-01 -5.70412911e-02 1.16867936e+00 -1.00039685e+00
-3.41660500e-01 -7.13544965e-01 -2.90932655e-01 -9.93420243e-01
5.27010798e-01 -8.23825479e-01 5.55725060e-02 -9.15972590e-01
-3.17674398e-01 -3.37782085e-01 4.35557306e-01 4.09807414e-01
6.46939874e-02 2.49141946e-01 5.21848738e-01 -9.30496827e-02
1.78906322e-01 7.17237175e-01 1.69018590e+00 -9.28482637e-02
-5.99961102e-01 5.36253490e-02 -9.24446434e-02 7.98785746e-01
6.15899026e-01 -2.45022595e-01 -6.26247942e-01 -1.57788575e-01
-6.58154339e-02 5.29491723e-01 2.03789786e-01 -8.83893132e-01
-7.41794184e-02 -1.62336230e-01 6.37107342e-02 -4.69579279e-01
8.47173870e-01 -6.21833503e-01 3.60581309e-01 1.05451316e-01
-3.60125005e-01 2.06771776e-01 6.03543758e-01 2.05526233e-01
-1.77383311e-02 2.99051017e-01 1.02539802e+00 1.95523128e-01
-2.08676785e-01 6.35341942e-01 -5.84152460e-01 -3.54946479e-02
8.89490962e-01 -4.28144157e-01 -1.22715808e-01 -1.03362274e+00
-8.10892820e-01 -3.64499912e-02 5.96012115e-01 5.86186469e-01
7.90957570e-01 -1.42136657e+00 -9.26562309e-01 4.63974983e-01
-1.17950097e-01 -2.95827478e-01 3.71863842e-01 6.40908062e-01
-8.97930324e-01 1.40547469e-01 -3.91729087e-01 -8.94329488e-01
-1.55845833e+00 3.99992257e-01 4.89742786e-01 4.46437329e-01
-7.13909924e-01 8.94762337e-01 1.98226437e-01 -3.19849044e-01
3.71959686e-01 -3.59788060e-01 -6.98181987e-02 5.49424626e-02
5.20736992e-01 3.01834822e-01 -9.91120711e-02 -1.10434449e+00
-7.26633668e-02 8.65080416e-01 3.94563198e-01 -5.31648874e-01
1.09940815e+00 -2.40194444e-02 4.23157185e-01 2.82090962e-01
1.21854115e+00 6.67322695e-01 -1.83400333e+00 2.17696339e-01
-7.99769223e-01 -3.32355499e-01 -8.99709687e-02 -4.29912329e-01
-1.51679456e+00 1.06652284e+00 4.39238548e-01 -4.02112097e-01
1.15710056e+00 -1.84851676e-01 1.06117964e+00 -2.75094301e-01
2.43552849e-01 -9.95070755e-01 3.18269163e-01 1.92312479e-01
1.08005822e+00 -5.17147601e-01 -4.74054188e-01 -5.59075296e-01
-9.54448342e-01 1.49882996e+00 4.65493381e-01 1.72611460e-01
7.30954230e-01 7.90537894e-01 4.59562749e-01 1.32687390e-01
-7.13237107e-01 1.85175762e-01 6.37422979e-01 9.88099754e-01
2.23968327e-01 6.21215627e-02 2.26867273e-01 2.01973721e-01
-8.98519754e-01 -2.40005463e-01 5.89536011e-01 2.47903571e-01
-2.91949958e-01 -1.03048837e+00 -4.10858303e-01 -3.32274698e-02
-2.49608636e-01 -1.39853507e-01 -2.14947045e-01 7.59197772e-01
-7.27986619e-02 7.23998785e-01 2.37970427e-01 -3.96049112e-01
2.54497111e-01 5.76489419e-02 7.66017199e-01 -3.49579901e-01
-6.02638364e-01 6.45444512e-01 -5.16882986e-02 -6.21267200e-01
-1.03570715e-01 -6.84523344e-01 -1.59881198e+00 -5.80311358e-01
-5.89233711e-02 -1.61717489e-01 7.93883801e-01 5.32455862e-01
4.35218602e-01 3.52956891e-01 7.52472043e-01 -1.32906950e+00
1.06052153e-01 -7.40636528e-01 -4.81137246e-01 1.89697355e-01
5.00322700e-01 -6.53906405e-01 -2.85788000e-01 2.12807745e-01]
|
[13.18025016784668, -0.43315181136131287]
|
07dd8e90-34b3-4801-8d0f-46f170250eb0
|
backpropagation-free-4d-continuous-ant-based
|
2305.06715
| null |
https://arxiv.org/abs/2305.06715v1
|
https://arxiv.org/pdf/2305.06715v1.pdf
|
Backpropagation-Free 4D Continuous Ant-Based Neural Topology Search
|
Continuous Ant-based Topology Search (CANTS) is a previously introduced novel nature-inspired neural architecture search (NAS) algorithm that is based on ant colony optimization (ACO). CANTS utilizes a continuous search space to indirectly-encode a neural architecture search space. Synthetic ant agents explore CANTS' continuous search space based on the density and distribution of pheromones, strongly inspired by how ants move in the real world. This continuous search space allows CANTS to automate the design of artificial neural networks (ANNs) of any size, removing a key limitation inherent to many current NAS algorithms that must operate within structures of a size that is predetermined by the user. This work expands CANTS by adding a fourth dimension to its search space representing potential neural synaptic weights. Adding this extra dimension allows CANTS agents to optimize both the architecture as well as the weights of an ANN without applying backpropagation (BP), which leads to a significant reduction in the time consumed in the optimization process. The experiments of this study - using real-world data - demonstrate that the BP-Free~CANTS algorithm exhibits highly competitive performance compared to both CANTS and ANTS while requiring significantly less operation time.
|
['Travis Desell', 'Alexander Ororbia', 'Zeming Lyu', 'Karl Ricanek', 'AbdElRahman ElSaid']
|
2023-05-11
| null | null | null | null |
['architecture-search']
|
['methodology']
|
[ 3.21790487e-01 -4.90856282e-02 -1.50556825e-02 -1.64979659e-02
6.16025269e-01 -3.41318876e-01 3.84961247e-01 1.56589568e-01
-7.89300740e-01 8.92599344e-01 -4.30019468e-01 -5.36092579e-01
-2.94280350e-01 -9.83607233e-01 -4.50821400e-01 -7.89004147e-01
4.53551896e-02 6.04247630e-01 6.01802289e-01 -4.21132416e-01
5.30701458e-01 8.82830441e-01 -1.71394420e+00 -3.67314994e-01
7.01639891e-01 8.54755223e-01 5.26498318e-01 3.84524882e-01
-3.62399250e-01 1.15064070e-01 -7.94817269e-01 3.66481654e-02
4.11876857e-01 -4.29794103e-01 -3.90317053e-01 -3.69115829e-01
-3.56212139e-01 6.89660847e-01 5.51384799e-02 9.22219098e-01
4.54454929e-01 1.26407772e-01 4.43158329e-01 -1.14113677e+00
-5.42721331e-01 5.01646757e-01 -4.04574752e-01 4.69074160e-01
-1.20011851e-01 9.01238844e-02 7.79013932e-01 -4.75296646e-01
5.90417087e-01 1.02677119e+00 4.72568601e-01 5.74327767e-01
-1.23410201e+00 -8.88935387e-01 2.25769877e-01 3.82407337e-01
-1.46514237e+00 -2.12831736e-01 9.63887453e-01 2.09967703e-01
1.22484231e+00 6.65423930e-01 1.34258115e+00 5.43122411e-01
4.73974645e-01 4.73695338e-01 8.17152441e-01 -8.34764183e-01
9.28841650e-01 1.16001345e-01 -2.45181262e-01 7.96144605e-01
8.04274619e-01 4.06228423e-01 -5.68648934e-01 -3.27339247e-02
8.24770927e-01 -4.97189283e-01 -1.03342801e-01 -6.36835933e-01
-1.02254045e+00 8.70581865e-01 7.86776066e-01 6.85706854e-01
-5.07117569e-01 1.63336232e-01 3.71026769e-02 2.40356803e-01
-2.74381638e-01 1.07552016e+00 -4.96261775e-01 -2.05093607e-01
-5.53242922e-01 2.41114125e-02 8.68782282e-01 3.86888206e-01
4.99458432e-01 6.52255595e-01 4.94094104e-01 7.33461559e-01
5.05335391e-01 4.32775080e-01 8.65187764e-01 -7.33557701e-01
2.55421512e-02 1.15354836e+00 -2.81950653e-01 -1.02165234e+00
-5.69170773e-01 -8.19049299e-01 -6.99243486e-01 7.06077278e-01
-5.08412085e-02 -1.29697666e-01 -1.07860386e+00 1.68585610e+00
4.66431171e-01 -1.14565544e-01 -3.45259625e-03 6.32840991e-01
3.61383855e-01 4.49045897e-01 -3.08198690e-01 -4.25787628e-01
1.09017622e+00 -1.04857945e+00 -4.29837346e-01 -4.88476813e-01
1.41699314e-01 -2.66896605e-01 8.70931447e-01 3.29151630e-01
-9.51809943e-01 -4.58567254e-02 -1.53593361e+00 7.01018929e-01
-8.66411805e-01 -3.49002481e-01 9.09039259e-01 9.70947385e-01
-1.37503493e+00 3.99318576e-01 -9.49774265e-01 -3.64897877e-01
3.37371826e-01 1.05519831e+00 -4.65092249e-02 4.86923635e-01
-8.15488517e-01 1.04978764e+00 7.28360832e-01 5.41593656e-02
-9.99424532e-02 -2.73063302e-01 -5.44437408e-01 1.40907675e-01
3.51388782e-01 -7.19695270e-01 7.78631091e-01 -7.98774898e-01
-1.72159946e+00 1.61650941e-01 -2.15115815e-01 -7.42488086e-01
-1.95583880e-01 7.41837382e-01 -4.88293648e-01 -2.64545232e-02
-1.99576065e-01 9.23808873e-01 8.41682971e-01 -1.17633307e+00
-6.24282658e-01 -2.49496549e-01 -2.44661942e-01 5.05170345e-01
-7.34363437e-01 -2.70616561e-01 -4.09585476e-01 -7.11876392e-01
4.00483906e-01 -1.15805411e+00 -4.99845475e-01 2.97569018e-02
-5.35365343e-02 -1.52094930e-01 7.45055377e-01 2.27143884e-01
1.39493859e+00 -1.87508070e+00 2.44631991e-01 7.29898453e-01
8.56777951e-02 5.68912923e-01 -5.08848250e-01 2.69600272e-01
5.07512167e-02 1.26524359e-01 -4.07791078e-01 1.63742527e-01
-1.65431559e-01 6.45351768e-01 1.96391374e-01 5.81110083e-02
-4.96792197e-02 8.47989500e-01 -6.92812264e-01 -3.02756339e-01
-2.68440153e-02 2.78851599e-01 -3.07806641e-01 -2.68662363e-01
-9.71700773e-02 -3.85935232e-02 -3.96170229e-01 6.89291358e-01
1.26338974e-01 -7.63316154e-02 9.28155258e-02 2.26054624e-01
-4.73253965e-01 8.49426165e-02 -1.12482786e+00 1.28456330e+00
-2.73305148e-01 6.85591757e-01 1.32998213e-01 -9.13072288e-01
1.30624807e+00 3.38130221e-02 4.44309652e-01 -9.99444067e-01
3.75410646e-01 4.10632819e-01 6.11606002e-01 1.00316010e-01
1.31011918e-01 3.12680036e-01 5.33727825e-01 6.94906116e-01
-3.49872231e-01 -1.43453315e-01 7.17496336e-01 -2.50822753e-01
1.41644573e+00 -2.88937926e-01 1.14884101e-01 -4.30380821e-01
6.89136207e-01 3.07520837e-01 6.51784360e-01 5.81923604e-01
-6.57863319e-02 -1.19389653e-01 -6.28695413e-02 -8.38414550e-01
-5.68473697e-01 -9.23841834e-01 -1.62420906e-02 8.36675167e-01
2.83483207e-01 -1.43532470e-01 -6.15822017e-01 -1.65602788e-01
-1.49863228e-01 8.45070302e-01 -8.52400661e-01 -2.27053225e-01
-6.07241213e-01 -1.22092509e+00 3.47406447e-01 1.49739027e-01
4.65138108e-01 -1.59650481e+00 -1.46612191e+00 6.34891391e-01
3.69491845e-01 -5.71866989e-01 -1.20618381e-01 7.67670095e-01
-1.12162828e+00 -9.79946852e-01 -1.81536078e-01 -1.01078355e+00
9.72735286e-01 -1.01179801e-01 7.81600356e-01 3.34505022e-01
-8.73424888e-01 7.07380548e-02 -3.03876787e-01 -9.56019819e-01
-4.09254998e-01 3.96182030e-01 4.28701714e-02 -3.03911567e-01
4.60348099e-01 -1.12578654e+00 -5.45708954e-01 2.00957760e-01
-6.73374355e-01 4.68834117e-02 9.62450802e-01 8.70160997e-01
6.51873171e-01 6.15608096e-01 5.05530000e-01 -2.42689416e-01
1.06058383e+00 -5.66864051e-02 -8.80126536e-01 2.67895758e-01
-1.39431286e+00 4.20538396e-01 5.06092846e-01 -5.77123165e-01
-4.36435789e-01 2.40937456e-01 2.69918203e-01 -1.34173274e-01
-7.49944849e-03 4.28070664e-01 3.61745119e-01 -7.14928806e-01
6.76611245e-01 6.15462065e-01 3.46327752e-01 -3.50924820e-01
-1.68417320e-01 3.27211022e-01 5.36664307e-01 -1.35727763e-01
8.88040245e-01 2.61062294e-01 3.64264131e-01 -8.69398534e-01
1.32207349e-01 -2.72484124e-01 -3.23707014e-01 6.05289126e-03
4.87193286e-01 -1.56147629e-01 -8.85729611e-01 1.32136896e-01
-8.51603568e-01 -1.30990654e-01 -4.42319959e-01 3.02438051e-01
-2.27328882e-01 -2.08788618e-01 -2.07102597e-01 -9.35789883e-01
-7.43054271e-01 -1.08778691e+00 1.47374287e-01 4.33835536e-01
-3.05555820e-01 -9.33680058e-01 1.11398838e-01 -2.41827399e-01
7.56392539e-01 3.50242466e-01 9.20621634e-01 -5.13698936e-01
-3.33495975e-01 -9.75971743e-02 -2.99409200e-02 -2.83956915e-01
1.97057351e-01 -1.22223571e-01 -2.00107902e-01 -3.24989706e-01
2.00605795e-01 2.78847992e-01 4.28724587e-01 2.96559691e-01
5.50546169e-01 -2.93189526e-01 -4.97279018e-01 5.14648497e-01
1.59130251e+00 1.09730566e+00 4.31101471e-01 1.30055678e+00
1.39167398e-01 2.66370207e-01 3.19262892e-01 2.38810971e-01
8.95660892e-02 4.63950366e-01 9.01805162e-01 -1.73747778e-01
-1.22067071e-01 2.68386245e-01 3.99044566e-02 9.89465058e-01
-7.32671469e-02 -1.51605368e-01 -9.35167611e-01 6.12666070e-01
-1.64027083e+00 -7.54050672e-01 1.12151384e-01 1.79469740e+00
9.08984244e-01 3.27143103e-01 -1.14283405e-01 5.94615102e-01
5.61736703e-01 -1.79016113e-01 -9.99783933e-01 -1.11254323e+00
-1.12707019e-01 5.71606100e-01 6.51537240e-01 2.27930859e-01
-5.72076380e-01 7.20130563e-01 6.23803616e+00 5.14331222e-01
-1.35366809e+00 -4.14918661e-01 1.37823899e-05 -3.08377415e-01
-1.20056309e-01 -1.20020330e-01 -5.69033086e-01 5.06439447e-01
9.31165934e-01 -2.75690615e-01 1.07332933e+00 5.95680475e-01
1.21757120e-01 -2.32844755e-01 -5.89839280e-01 1.04873645e+00
8.24068952e-03 -1.46263981e+00 1.77117258e-01 2.12096408e-01
6.26367152e-01 2.99014926e-01 5.71319368e-03 -2.62678683e-01
1.93005547e-01 -9.42989945e-01 6.91161752e-01 8.50832015e-02
3.06363225e-01 -8.43424141e-01 5.75326741e-01 2.35200465e-01
-1.30008876e+00 -6.47656024e-01 -2.78628111e-01 -4.10707220e-02
-1.48391038e-01 1.39851868e-01 -9.86938417e-01 1.51999936e-01
7.85974622e-01 6.40351102e-02 -7.74846196e-01 1.37735200e+00
-2.21683495e-02 3.27734649e-01 -7.36936331e-01 -9.01016533e-01
5.48390210e-01 -3.36522371e-01 7.56832659e-01 6.88702822e-01
2.24105582e-01 2.44899616e-02 -3.28667462e-01 8.30515206e-01
2.35177591e-01 -2.34232731e-02 -2.42628470e-01 -1.87041402e-01
1.05731487e+00 8.61608922e-01 -1.31073022e+00 1.55801803e-01
2.46993110e-01 6.35132492e-01 1.91799216e-02 2.43706673e-01
-5.42561889e-01 -8.56231153e-01 2.49025285e-01 -2.71479655e-02
5.96538723e-01 -2.16314524e-01 -6.72968626e-01 -2.09036887e-01
-1.22487217e-01 -8.67321551e-01 5.37178246e-03 -4.42638457e-01
-6.14719272e-01 1.15640390e+00 -3.63878578e-01 -8.36911738e-01
-1.76917806e-01 -4.48624194e-01 -6.23339474e-01 6.51991189e-01
-1.45419753e+00 -6.81331515e-01 1.66739747e-02 5.96188009e-01
5.57504594e-01 -7.63211727e-01 9.00253594e-01 -2.30402857e-01
-6.65683746e-01 5.08764625e-01 9.60600302e-02 -3.07199568e-01
-3.51829939e-02 -1.01405478e+00 5.26814222e-01 6.74160540e-01
5.07831037e-01 6.78181887e-01 7.19836414e-01 -4.37632352e-01
-1.50618041e+00 -5.39589584e-01 7.23406792e-01 1.24810368e-01
5.98426104e-01 -4.16167043e-02 -4.82864439e-01 -1.39228374e-01
1.65810689e-01 -3.25160056e-01 5.25547922e-01 -1.34027734e-01
9.83051211e-03 -3.38429987e-01 -1.13203144e+00 8.24035406e-01
9.84549403e-01 2.14450136e-01 -5.52441180e-01 -2.82603707e-02
3.88666213e-01 -1.65191203e-01 -3.59734386e-01 4.30737823e-01
6.71141028e-01 -8.71818244e-01 9.43807185e-01 1.93309058e-02
-4.11620259e-01 -5.06193161e-01 1.72479764e-01 -1.18322635e+00
-4.74819750e-01 -5.08044899e-01 -1.96835220e-01 5.20689964e-01
7.40205765e-01 -1.22155476e+00 1.00110006e+00 4.81359303e-01
-4.81246263e-02 -1.16425347e+00 -1.26698363e+00 -1.00159788e+00
-3.71503234e-01 2.86831916e-03 9.76543427e-01 7.03677773e-01
-7.64302835e-02 2.93229550e-01 6.75413191e-01 -3.00907996e-03
6.14392161e-01 3.03272642e-02 1.45732507e-01 -1.67558408e+00
-2.17452779e-01 -9.03983951e-01 -4.82163668e-01 -4.41412896e-01
-2.35038534e-01 -6.00616276e-01 7.67012537e-02 -1.54680908e+00
-4.59715605e-01 -9.17530775e-01 -5.35266101e-01 8.01901460e-01
4.79964942e-01 4.29100484e-01 1.87942237e-01 2.75877893e-01
-2.11044788e-01 5.68971395e-01 1.09930825e+00 -7.67019242e-02
-7.39226162e-01 -1.63787872e-01 -5.51369131e-01 7.36796856e-01
1.11211133e+00 -7.15043962e-01 -8.76839757e-01 -5.47272265e-01
2.01556116e-01 -4.17441130e-01 -2.21353889e-01 -1.28965604e+00
7.30070770e-01 -2.45006099e-01 3.34728479e-01 -3.00799519e-01
3.69021863e-01 -1.09483743e+00 2.30649740e-01 1.16208863e+00
-5.89333028e-02 6.39300764e-01 4.54273105e-01 6.49973750e-01
2.72899177e-02 -4.46003288e-01 6.29904091e-01 -5.11965081e-02
-5.71129084e-01 -1.42303079e-01 -6.20163560e-01 -4.48623002e-01
1.19904387e+00 -1.03398681e+00 1.09763723e-03 1.50327131e-01
-5.99774718e-01 8.05037618e-02 3.99162173e-01 4.34889346e-01
7.74809301e-01 -9.42675710e-01 -2.13034019e-01 5.19714177e-01
-3.18428874e-01 -1.73408389e-02 -7.29378819e-01 4.12676245e-01
-9.24693167e-01 6.47888720e-01 -5.63993216e-01 -3.16912770e-01
-1.21965814e+00 4.34163362e-01 1.38243869e-01 -1.00462191e-01
-4.35296535e-01 1.04408026e+00 -4.95494932e-01 -2.81173885e-01
3.48992258e-01 1.18322000e-01 -4.42231417e-01 -4.04156089e-01
2.99752027e-01 2.98747927e-01 1.30718872e-01 -2.66788781e-01
-6.20968997e-01 2.80673444e-01 3.30471955e-02 -2.87595600e-01
1.54808116e+00 5.16419299e-02 -4.30935919e-01 1.46856502e-01
5.73980331e-01 -2.97056496e-01 -7.53401279e-01 2.90095240e-01
3.46266687e-01 -1.23902790e-01 3.75329822e-01 -1.25195444e+00
-1.16735351e+00 3.14534843e-01 8.14432502e-01 2.55438805e-01
1.45992637e+00 -5.64704716e-01 6.38511002e-01 7.99635530e-01
5.25467098e-01 -1.23021078e+00 -4.96056639e-02 4.13457572e-01
7.69131720e-01 -5.32686055e-01 8.62148777e-02 -1.26670316e-01
-1.83028653e-01 1.21579504e+00 7.56989181e-01 -3.09527926e-02
5.83531618e-01 4.61534828e-01 -4.94636483e-02 -2.34922543e-01
-9.31841493e-01 1.96539182e-02 6.05173968e-03 7.56153703e-01
-2.20446318e-01 -2.35544354e-01 -7.39747882e-01 5.28580248e-02
-3.93320352e-01 -2.22318187e-01 3.33598703e-01 1.35307610e+00
-6.96674228e-01 -1.50332832e+00 -4.87102419e-01 4.79854107e-01
1.49402678e-01 -1.83763444e-01 -5.84027112e-01 7.18208432e-01
2.47240230e-01 1.02143204e+00 1.34528786e-01 -3.39014530e-01
7.43480697e-02 -2.63217062e-01 3.09384912e-01 -3.42614502e-01
-6.95939422e-01 -2.63865560e-01 -2.05192924e-01 -4.88997608e-01
-2.22585857e-01 -4.35075402e-01 -1.59138370e+00 -2.38370731e-01
-5.78551590e-01 6.55777395e-01 1.41609085e+00 7.00684547e-01
5.87743938e-01 6.60772204e-01 5.27767003e-01 -6.04039013e-01
-2.49750331e-01 -3.22187722e-01 -3.16313863e-01 -4.74009663e-01
-8.35550874e-02 -7.33547807e-01 -1.89059123e-01 -2.88258433e-01]
|
[8.174603462219238, 3.250377893447876]
|
0c534966-c4b2-4d82-9728-5df10bdcef5f
|
one-ruler-for-all-languages-multi-lingual
|
1805.02914
| null |
http://arxiv.org/abs/1805.02914v1
|
http://arxiv.org/pdf/1805.02914v1.pdf
|
One "Ruler" for All Languages: Multi-Lingual Dialogue Evaluation with Adversarial Multi-Task Learning
|
Automatic evaluating the performance of Open-domain dialogue system is a
challenging problem. Recent work in neural network-based metrics has shown
promising opportunities for automatic dialogue evaluation. However, existing
methods mainly focus on monolingual evaluation, in which the trained metric is
not flexible enough to transfer across different languages. To address this
issue, we propose an adversarial multi-task neural metric (ADVMT) for
multi-lingual dialogue evaluation, with shared feature extraction across
languages. We evaluate the proposed model in two different languages.
Experiments show that the adversarial multi-task neural metric achieves a high
correlation with human annotation, which yields better performance than
monolingual ones and various existing metrics.
|
['Rui Yan', 'Mingyue Shang', 'Xiaowei Tong', 'Zhenxin Fu', 'Dongyan Zhao']
|
2018-05-08
| null | null | null | null |
['dialogue-evaluation']
|
['natural-language-processing']
|
[-3.87831837e-01 -1.23196192e-01 1.25875771e-01 -5.27214468e-01
-1.04119313e+00 -7.28185892e-01 8.78610134e-01 -3.42958830e-02
-6.51737869e-01 1.20311582e+00 4.12531495e-01 -1.05799794e-01
3.13391954e-01 -6.34799659e-01 -9.43539590e-02 -4.42933410e-01
2.51877904e-01 6.33903027e-01 1.41305238e-01 -8.89863610e-01
4.67109913e-03 -3.00247222e-02 -5.92784762e-01 3.03193241e-01
9.99120831e-01 6.81283355e-01 -2.43000031e-01 8.09670925e-01
-2.47587621e-01 7.23052502e-01 -1.23866940e+00 -8.77062082e-01
-5.20240888e-03 -6.37708783e-01 -1.27595699e+00 -5.44954777e-01
2.79427290e-01 -4.38360751e-01 -1.80876374e-01 1.06379473e+00
8.85018766e-01 3.02537438e-02 9.01247799e-01 -1.09212029e+00
-8.70967329e-01 6.97146356e-01 -1.19666882e-01 -3.01810540e-02
5.71953595e-01 1.00465812e-01 1.23172557e+00 -6.98412180e-01
4.23588037e-01 1.51522899e+00 6.74244165e-01 7.65216351e-01
-1.04153717e+00 -3.48243356e-01 -2.46708512e-01 -5.09429239e-02
-1.04282212e+00 -1.43292204e-01 7.58338928e-01 -3.68554860e-01
7.79459536e-01 9.50110257e-02 -4.91512157e-02 1.55066431e+00
1.68189406e-01 8.48242819e-01 1.53803372e+00 -3.78958821e-01
-3.44659567e-01 5.32445908e-01 8.80480260e-02 6.42552972e-01
-4.23375309e-01 -2.90915012e-01 -3.92638773e-01 -1.17016360e-01
4.12189990e-01 -6.84760273e-01 -2.58567721e-01 5.11219874e-02
-1.41844714e+00 1.17784238e+00 7.66506791e-02 6.83130860e-01
1.46795198e-01 -3.73093843e-01 1.20787251e+00 8.22181463e-01
7.57344663e-01 7.79128850e-01 -7.29911804e-01 -3.75085115e-01
-4.27516222e-01 2.63241768e-01 1.17565024e+00 5.65268040e-01
5.30537307e-01 1.78526640e-01 -7.35987246e-01 1.52246773e+00
5.37889898e-02 5.44842958e-01 8.38516116e-01 -6.76472485e-01
5.81902802e-01 5.65818906e-01 -4.72606122e-02 -8.63494515e-01
-5.49788296e-01 7.87777733e-03 -1.03492427e+00 1.16780683e-01
6.98750198e-01 -7.10103214e-01 6.44952431e-02 1.82112718e+00
1.09619908e-01 -6.07708931e-01 5.61872363e-01 8.12563062e-01
1.21582592e+00 7.29977548e-01 -6.24519947e-04 -4.25391011e-02
1.14078951e+00 -1.38498020e+00 -1.03244972e+00 2.02464908e-01
7.45455921e-01 -1.01880491e+00 1.52626038e+00 3.25061411e-01
-8.56418133e-01 -6.65513515e-01 -1.09823442e+00 -1.36815935e-01
-6.37999952e-01 1.93495840e-01 3.71081978e-01 8.12742770e-01
-1.00524497e+00 3.22091460e-01 -5.27939759e-02 -2.82616913e-01
-3.17083359e-01 1.73627466e-01 -3.16842169e-01 3.50706935e-01
-1.92702127e+00 1.47948766e+00 5.12346268e-01 -2.09989715e-02
-9.85689938e-01 -5.89480884e-02 -8.36583555e-01 -3.26859504e-01
1.76425368e-01 -1.55570790e-01 1.57234728e+00 -8.73730183e-01
-2.13382173e+00 9.37755764e-01 4.35942352e-01 -4.32681829e-01
7.11929798e-01 -2.63407558e-01 -5.07617593e-01 -3.57427269e-01
-4.22160514e-03 4.31240112e-01 2.93112069e-01 -8.93076003e-01
-3.68640125e-01 -1.56716272e-01 5.03264606e-01 4.82058704e-01
-9.41036642e-01 2.53491819e-01 -1.46233946e-01 -5.44621885e-01
-8.14964831e-01 -6.93096042e-01 -7.18491599e-02 -3.97046626e-01
-5.15845716e-01 -7.41459012e-01 7.15557396e-01 -7.30033338e-01
1.21694231e+00 -1.60325074e+00 2.57328719e-01 -3.98004085e-01
6.13274872e-02 6.29231751e-01 -1.19267717e-01 5.22636890e-01
6.39488339e-01 1.13025382e-02 -1.57342970e-01 -3.39594066e-01
1.86666787e-01 5.94009720e-02 1.58287808e-01 1.46170527e-01
1.14998713e-01 7.71767557e-01 -8.48365843e-01 -7.44027495e-01
5.43494523e-02 2.13804111e-01 -2.07313448e-01 7.02175438e-01
-2.76126593e-01 8.44699621e-01 -5.94879806e-01 3.74638319e-01
3.86499912e-01 5.31404018e-02 1.58754379e-01 -1.70305997e-01
-8.51140264e-03 2.40350157e-01 -7.05130100e-01 2.03443456e+00
-7.62019217e-01 7.78375626e-01 2.69993506e-02 -7.75588632e-01
1.44427955e+00 5.27388096e-01 2.67968297e-01 -6.88836455e-01
2.94350535e-01 2.67884135e-01 1.54625580e-01 -4.70262498e-01
8.11351597e-01 -7.08548306e-03 -6.89229071e-01 6.32614255e-01
4.57227975e-01 -2.35817015e-01 2.78640151e-01 3.95597219e-02
9.37843800e-01 2.82621622e-01 3.09656829e-01 -3.26010376e-01
1.19355190e+00 -1.15456045e-01 4.05838579e-01 4.34100419e-01
-6.58267558e-01 3.69282603e-01 6.04177654e-01 -3.17580044e-01
-1.11396313e+00 -8.10475230e-01 -2.05730900e-01 1.60351205e+00
-1.81043237e-01 -1.12888396e-01 -1.09227574e+00 -1.08445156e+00
-1.35413662e-01 4.00137126e-01 -5.72673440e-01 -3.25793922e-02
-6.16608322e-01 -8.99209082e-01 1.31777501e+00 1.49420932e-01
9.27284777e-01 -1.16720152e+00 4.81746998e-03 2.36392081e-01
-4.95404869e-01 -1.16048908e+00 -6.22575939e-01 6.99176779e-03
-2.44715601e-01 -1.00999343e+00 -1.13146794e+00 -9.45736587e-01
-1.16701238e-01 -3.29825729e-01 1.33440161e+00 -3.97389889e-01
2.67402500e-01 2.51349539e-01 -4.00175601e-01 -2.06934050e-01
-1.12983906e+00 4.66310948e-01 2.55239438e-02 3.71757336e-02
3.85471284e-01 -8.26412365e-02 -1.54511139e-01 4.55774546e-01
-5.33526599e-01 -6.69568107e-02 3.40346843e-01 1.23350859e+00
-5.55199757e-03 -6.10192001e-01 1.13399601e+00 -1.01992416e+00
1.79864883e+00 -4.10637349e-01 -1.63180336e-01 5.90288877e-01
-6.33897662e-01 2.35817745e-01 1.01284659e+00 -4.57424104e-01
-1.25836563e+00 -6.14979565e-01 -3.44090998e-01 1.28408462e-01
-9.74862501e-02 5.97856283e-01 -2.65993297e-01 -1.23239629e-01
8.46453190e-01 2.65863109e-02 5.81453405e-02 -5.25696695e-01
4.46532011e-01 1.01607573e+00 5.04787803e-01 -9.15876269e-01
3.09590399e-01 -2.64777541e-01 -3.10929775e-01 -5.95814526e-01
-8.37509274e-01 -1.75850451e-01 -8.90470386e-01 -4.13809657e-01
1.13063359e+00 -6.85924292e-01 -7.92536914e-01 6.35067523e-01
-1.52104306e+00 -3.90395880e-01 1.46689549e-01 4.51091975e-01
-5.51882684e-01 4.09344852e-01 -9.82172012e-01 -5.15872180e-01
-6.91112757e-01 -1.40462065e+00 8.28088522e-01 9.36649293e-02
-2.32825980e-01 -1.55481923e+00 7.85985172e-01 4.03396785e-01
7.02825546e-01 4.80731845e-01 6.68949425e-01 -1.29092050e+00
2.07616523e-01 -9.24836621e-02 -1.90711081e-01 9.31447983e-01
9.26344246e-02 -1.04326770e-01 -1.06108952e+00 -1.73944160e-01
8.79522040e-02 -1.23769832e+00 3.66734773e-01 -1.64180890e-01
7.36022472e-01 -2.31016040e-01 3.32267374e-01 1.44088179e-01
1.02985203e+00 1.94703471e-02 3.02325994e-01 4.47959781e-01
6.86797976e-01 5.68656683e-01 6.65486276e-01 3.98835897e-01
6.46423578e-01 8.59495819e-01 4.05671308e-03 -3.08899939e-01
9.74162444e-02 1.16421044e-01 6.40129507e-01 1.38940120e+00
7.08152801e-02 -2.73294955e-01 -1.01896942e+00 3.29673141e-01
-1.80105770e+00 -5.95679998e-01 -2.04206690e-01 1.96387124e+00
1.37284744e+00 1.36129946e-01 2.39554569e-01 -2.02734485e-01
7.48307228e-01 4.89371806e-01 -2.78371930e-01 -1.07875538e+00
-5.05879939e-01 1.32259289e-02 9.02697369e-02 5.33785939e-01
-1.30046749e+00 1.07190943e+00 6.19123363e+00 9.42340612e-01
-8.75151217e-01 6.65416956e-01 5.97921610e-01 4.36893523e-01
-1.16415858e-01 -3.95792335e-01 -8.29805851e-01 1.73506454e-01
1.10879707e+00 -4.20988888e-01 8.26013088e-02 7.39505410e-01
-8.43357295e-03 2.29327083e-01 -1.05337954e+00 6.38634503e-01
3.94770354e-01 -6.86393201e-01 5.89741394e-02 -1.27095550e-01
8.35544527e-01 -6.78688474e-03 -3.63288261e-02 8.20940077e-01
6.52543366e-01 -1.02843940e+00 1.35213017e-01 4.08578724e-01
8.87354076e-01 -7.53599703e-01 1.04879987e+00 2.45236561e-01
-8.19390178e-01 4.36254591e-01 -4.30212557e-01 4.29100730e-02
1.16425253e-01 2.20727734e-02 -1.05894113e+00 5.29098153e-01
2.01015159e-01 4.57410753e-01 -5.68350613e-01 5.24009228e-01
-1.54914603e-01 4.03408557e-01 1.12528831e-01 -5.89471281e-01
6.41451299e-01 -3.25325906e-01 4.53342706e-01 1.67413688e+00
-1.59222513e-01 -6.38621271e-01 4.38100219e-01 5.79171002e-01
-3.13705206e-01 8.25380802e-01 -9.84350324e-01 1.13778509e-01
1.59718916e-01 1.39999866e+00 3.17396484e-02 -1.87423736e-01
-6.77814007e-01 1.13142216e+00 6.35577679e-01 8.56607407e-03
-8.68008673e-01 -5.74466228e-01 4.47866708e-01 -7.67292738e-01
-4.72259760e-01 -3.39946508e-01 -1.67602107e-01 -1.15238822e+00
-1.09619334e-01 -1.25649953e+00 4.57118422e-01 -3.33003789e-01
-1.52339244e+00 1.14429486e+00 -2.06208989e-01 -1.26024961e+00
-5.47948360e-01 -7.63665795e-01 -6.32183194e-01 1.03712583e+00
-1.35099196e+00 -1.38625669e+00 -1.36490345e-01 8.02514017e-01
8.29216599e-01 -9.52224076e-01 1.41516209e+00 4.40482378e-01
-5.85687101e-01 1.17868149e+00 4.49827373e-01 4.66356188e-01
1.46256006e+00 -1.53449249e+00 1.60629004e-01 3.54035050e-01
-9.12477151e-02 2.38558650e-01 4.31860864e-01 -1.89228386e-01
-9.60149407e-01 -9.53745425e-01 8.48352432e-01 -5.50461829e-01
9.30847287e-01 -1.70917481e-01 -9.27625775e-01 3.07906717e-01
9.60371375e-01 -5.10851920e-01 9.36646163e-01 4.35505688e-01
-3.04693222e-01 -8.69403407e-02 -1.24075186e+00 5.51521599e-01
3.44148904e-01 -6.71382964e-01 -4.93731439e-01 6.60133004e-01
8.18340421e-01 -3.98424685e-01 -1.50857615e+00 4.91407126e-01
4.58401680e-01 -9.50165272e-01 6.77713275e-01 -8.36257577e-01
6.25240207e-01 2.29743645e-01 -1.83467090e-01 -1.62108326e+00
1.94561452e-01 -6.00265980e-01 2.61311889e-01 1.65117824e+00
7.02180028e-01 -5.87188065e-01 2.37680554e-01 3.50654870e-01
-1.66544318e-01 -7.92171776e-01 -8.89057696e-01 -6.42816424e-01
7.62585461e-01 6.11260720e-03 4.35029358e-01 1.30148053e+00
1.00171529e-01 9.97785687e-01 -8.14802110e-01 -4.17264223e-01
2.65838951e-01 -2.03321338e-01 1.07735920e+00 -1.19813168e+00
-2.34673694e-01 -6.60776496e-01 -1.56296194e-01 -9.02146518e-01
5.78198612e-01 -8.66945744e-01 -1.51002882e-02 -1.27784693e+00
1.09379657e-01 -3.76145840e-01 -2.93725252e-01 3.22849005e-01
-2.16009498e-01 1.55391693e-01 2.87756771e-02 1.18421637e-01
-8.77241671e-01 9.95942175e-01 1.45907438e+00 -4.27110732e-01
-1.25812916e-02 -9.44931060e-02 -3.32319796e-01 6.89948559e-01
1.05539465e+00 -1.31044969e-01 -1.32603928e-01 -5.37348688e-01
-2.05715567e-01 2.38483682e-01 -2.02042997e-01 -1.00659895e+00
-6.58150017e-02 -5.18997908e-02 -7.40559846e-02 -1.83748633e-01
3.30187768e-01 -3.28399569e-01 -6.93198264e-01 2.31828868e-01
-6.89695954e-01 3.57365429e-01 4.70774285e-02 1.12209372e-01
-6.08913481e-01 -5.50573885e-01 7.99353778e-01 -2.09112063e-01
-4.73849118e-01 1.35897785e-01 -1.90962687e-01 5.92937231e-01
6.92609310e-01 3.99018854e-01 -5.40726960e-01 -4.30402488e-01
-5.64237833e-01 4.35653389e-01 5.45420311e-02 6.61343038e-01
2.86525726e-01 -1.50727105e+00 -1.30850446e+00 -3.05688679e-01
2.30437964e-01 -4.15145129e-01 -1.29994735e-01 6.06932759e-01
-5.54808319e-01 6.98781490e-01 -5.43212533e-01 -4.56611842e-01
-1.16063571e+00 9.72818732e-02 7.22232640e-01 -9.11414206e-01
5.08669913e-02 5.81829965e-01 -5.17063849e-02 -1.22934878e+00
3.80589098e-01 2.44047701e-01 -6.94889069e-01 1.74169183e-01
1.94220886e-01 3.55178595e-01 -5.13851419e-02 -8.39806855e-01
1.40008867e-01 3.04665118e-01 -1.52214780e-01 -4.24214244e-01
9.18837368e-01 -9.10532624e-02 -1.56495392e-01 9.65823650e-01
1.43318021e+00 6.43005595e-02 -7.69319892e-01 -5.98789752e-01
-4.85104248e-02 1.52354101e-02 -3.33567291e-01 -9.77942646e-01
-6.83276057e-01 1.03465044e+00 6.93399668e-01 4.42842871e-01
6.24625504e-01 -4.44915652e-01 8.92539620e-01 8.82479072e-01
4.91217852e-01 -1.42530107e+00 4.26122904e-01 1.27241158e+00
1.21937335e+00 -1.66985130e+00 -3.87875289e-01 1.54447615e-01
-1.09297514e+00 1.25378203e+00 1.25397110e+00 6.03832006e-02
4.12423700e-01 9.73254535e-03 5.79025626e-01 -3.73322330e-02
-5.22936344e-01 -1.38907898e-02 3.09642553e-01 4.93284106e-01
1.07170630e+00 1.24172881e-01 -7.56704867e-01 5.53645849e-01
-2.55473584e-01 -4.43723917e-01 4.29500759e-01 3.58481109e-01
-2.79291779e-01 -1.52595317e+00 -1.62068114e-01 1.58722743e-01
-7.66807079e-01 -5.85849509e-02 -7.01451004e-01 7.02455401e-01
-2.33875811e-01 1.04697931e+00 -4.36261415e-01 -6.43988311e-01
4.02463198e-01 2.86122590e-01 1.99445441e-01 -4.96348768e-01
-1.08882391e+00 -2.58144945e-01 7.19804168e-01 -1.05755538e-01
-5.71190000e-01 -2.47181728e-01 -7.45099425e-01 -1.73283696e-01
-3.62785012e-01 5.47124386e-01 6.03401244e-01 8.48406374e-01
-1.78148225e-01 5.20882130e-01 1.06845450e+00 -4.85695571e-01
-1.00711060e+00 -1.81018567e+00 -3.48936826e-01 5.41925192e-01
1.05444016e-02 -6.48181677e-01 -1.71992391e-01 -2.61546701e-01]
|
[12.636343955993652, 8.235668182373047]
|
fd0c095c-114c-4558-a628-092d8c772126
|
graph-self-attention-for-learning-graph
|
2201.12787
| null |
https://arxiv.org/abs/2201.12787v3
|
https://arxiv.org/pdf/2201.12787v3.pdf
|
GRPE: Relative Positional Encoding for Graph Transformer
|
We propose a novel positional encoding for learning graph on Transformer architecture. Existing approaches either linearize a graph to encode absolute position in the sequence of nodes, or encode relative position with another node using bias terms. The former loses preciseness of relative position from linearization, while the latter loses a tight integration of node-edge and node-topology interaction. To overcome the weakness of the previous approaches, our method encodes a graph without linearization and considers both node-topology and node-edge interaction. We name our method Graph Relative Positional Encoding dedicated to graph representation learning. Experiments conducted on various graph datasets show that the proposed method outperforms previous approaches significantly. Our code is publicly available at https://github.com/lenscloth/GRPE.
|
['Seung-won Hwang', 'Juntae Kim', 'Donggeon Lee', 'WoongGi Chang', 'Wonpyo Park']
|
2022-01-30
| null | null | null | null |
['graph-regression']
|
['graphs']
|
[-2.12444305e-01 4.48762804e-01 -5.51334679e-01 -2.69017607e-01
-4.48315769e-01 -7.76947141e-01 4.81248558e-01 3.82086813e-01
2.47928813e-01 5.95022082e-01 2.88814545e-01 -5.03436148e-01
-2.19017327e-01 -1.08785093e+00 -8.59639108e-01 -5.26445210e-01
-2.55780458e-01 3.21868658e-01 2.54176229e-01 -2.30655000e-01
1.34499416e-01 3.85850042e-01 -7.87544727e-01 -4.64210436e-02
6.03531003e-01 8.84592235e-01 1.75778374e-01 4.54362422e-01
2.80301068e-02 1.11931324e+00 -2.75188684e-01 -4.02938396e-01
3.61472130e-01 -3.25518280e-01 -7.71636546e-01 -1.32803544e-01
5.66377461e-01 -5.06811738e-02 -9.37972009e-01 1.27380860e+00
3.42426866e-01 -1.69410229e-01 2.45176598e-01 -1.37403297e+00
-8.63231480e-01 1.10117054e+00 -6.19054139e-01 2.03803495e-01
4.53189045e-01 -3.33476394e-01 1.41447926e+00 -5.93043447e-01
5.66077232e-01 9.64715958e-01 7.84975767e-01 1.31250471e-01
-1.21220672e+00 -5.38407266e-01 5.27908146e-01 2.07429931e-01
-1.62564766e+00 -2.35851571e-01 1.38357449e+00 -4.41834539e-01
7.78256774e-01 2.81218886e-01 8.80454004e-01 7.84324467e-01
2.25480840e-01 5.82868636e-01 8.85050297e-01 -2.82618672e-01
-2.83859149e-02 -1.72980145e-01 4.51994449e-01 1.32340336e+00
2.89949566e-01 -1.80878453e-02 -3.25734705e-01 3.68480794e-02
9.63235378e-01 -1.15958564e-01 -4.95910257e-01 -9.81846452e-01
-9.26647365e-01 6.52001500e-01 1.18668294e+00 3.76818269e-01
2.28201598e-02 6.36541545e-01 2.49365255e-01 4.54102010e-01
4.22106504e-01 1.89086929e-01 -9.36726108e-02 1.24097325e-01
-7.15672553e-01 7.10766902e-03 7.51485765e-01 1.38455439e+00
9.39560115e-01 6.80385754e-02 7.94834793e-02 5.51992238e-01
3.71611297e-01 1.77717313e-01 2.02623948e-01 -4.79996234e-01
6.34626329e-01 8.11558366e-01 -4.63499069e-01 -1.55933249e+00
-4.25891668e-01 -6.78733587e-01 -9.93487298e-01 -2.84076989e-01
1.92236617e-01 7.64423460e-02 -8.57664227e-01 1.81041360e+00
-6.52487278e-02 4.84781563e-01 -2.86783427e-01 5.39035618e-01
1.15297163e+00 4.55591798e-01 -3.12283218e-01 -5.22200838e-02
1.13474095e+00 -1.15789354e+00 -5.72063029e-01 -2.52985388e-01
6.60313427e-01 -4.45661843e-01 7.87051141e-01 7.84027204e-02
-1.08738744e+00 -2.62631088e-01 -1.09663582e+00 -2.20700458e-01
-3.73286486e-01 1.64911494e-01 7.77203918e-01 4.67171907e-01
-1.52006066e+00 5.60665429e-01 -7.90564001e-01 -1.69797853e-01
1.51883483e-01 4.48114604e-01 -4.78363395e-01 -4.35616896e-02
-1.17433345e+00 5.23658395e-01 3.68706942e-01 1.41695023e-01
-5.29518545e-01 -5.75127125e-01 -1.22073555e+00 1.62931502e-01
4.20325488e-01 -6.61399603e-01 1.08067679e+00 -5.05823195e-01
-1.31194210e+00 5.95661342e-01 -8.24406520e-02 -5.04546046e-01
2.83934355e-01 6.33010566e-02 -1.96013972e-01 -7.12942751e-03
-1.87159061e-01 3.77016783e-01 5.64574599e-01 -1.30352080e+00
-1.33449957e-01 -4.19958323e-01 5.44847727e-01 2.23685145e-01
-2.34188035e-01 -5.34931660e-01 -8.42168927e-01 -5.87200165e-01
6.55185759e-01 -9.13802683e-01 -1.21520668e-01 -2.38791645e-01
-6.33518934e-01 -3.20460558e-01 8.81552041e-01 -4.84023333e-01
1.60286391e+00 -1.94752586e+00 4.92432237e-01 3.13285112e-01
8.17779601e-01 2.14613024e-02 -1.09317794e-01 8.14990878e-01
-4.42554682e-01 1.73733085e-01 8.28868225e-02 -2.33616680e-01
-1.83271114e-02 2.80464262e-01 -2.17525333e-01 5.46611071e-01
-3.27958614e-01 1.15447485e+00 -1.22346532e+00 -4.24690962e-01
2.21233621e-01 6.31422043e-01 -6.07886672e-01 6.71730377e-03
-3.50474683e-03 1.51466042e-01 -5.53745985e-01 5.74402332e-01
6.72201574e-01 -5.40179789e-01 5.79681277e-01 -6.75158560e-01
1.12074830e-01 6.58849776e-01 -1.09485459e+00 1.64128268e+00
-3.45552146e-01 5.98120153e-01 5.80766387e-02 -1.07921433e+00
9.52935159e-01 9.95498076e-02 4.93975341e-01 -5.26951373e-01
1.52541213e-02 -7.21698776e-02 -4.06136289e-02 -7.24614412e-02
4.06456232e-01 1.50387481e-01 -8.55968222e-02 1.37291715e-01
1.76580682e-01 -2.00990573e-01 2.06489369e-01 7.71571815e-01
1.30815709e+00 1.54767543e-01 4.92820024e-01 -2.97552228e-01
4.36251789e-01 -4.39518690e-01 4.81048524e-01 3.84435833e-01
-7.20999017e-02 5.19360721e-01 1.03774488e+00 -3.63892466e-01
-7.28115618e-01 -1.24109066e+00 9.36342552e-02 7.90408969e-01
2.72982419e-01 -1.27306068e+00 -5.78933239e-01 -9.75726962e-01
1.94166526e-02 4.22332466e-01 -6.57469571e-01 -1.89894006e-01
-7.85066485e-01 -3.07905257e-01 3.64056468e-01 6.78467929e-01
3.72591436e-01 -6.51265621e-01 7.96854272e-02 -5.20943515e-02
-1.70213416e-01 -1.02247012e+00 -6.73912704e-01 6.05272353e-02
-8.17365348e-01 -1.06823683e+00 -3.92238855e-01 -9.43603039e-01
1.02134728e+00 2.64824122e-01 1.21640539e+00 3.87065083e-01
2.53330171e-02 5.13706088e-01 -3.67362618e-01 2.17624411e-01
-1.90736204e-02 3.11835289e-01 -1.56745315e-01 -2.09657937e-01
-1.89527959e-01 -8.94960701e-01 -5.49485505e-01 7.72533789e-02
-5.23729503e-01 3.09445620e-01 5.09175360e-01 6.69235766e-01
6.48463428e-01 1.82403162e-01 1.74103364e-01 -1.15988564e+00
5.77977717e-01 -4.95890230e-01 -7.54494131e-01 2.81525224e-01
-6.44825399e-01 2.24365234e-01 4.98508841e-01 4.42191027e-03
-4.72530276e-01 6.78758249e-02 -5.14166057e-02 -5.04310131e-01
3.60758871e-01 8.19686294e-01 -3.62051725e-01 -4.56807703e-01
2.08134070e-01 1.32127360e-01 -1.52899623e-01 -4.90360171e-01
5.32055080e-01 4.57491018e-02 5.32381833e-01 -6.72942996e-01
8.99755359e-01 1.15209728e-01 2.17059061e-01 -5.40977597e-01
-4.98472780e-01 -2.07138687e-01 -8.49230111e-01 -2.09393904e-01
2.45250657e-01 -7.88880885e-01 -6.23168945e-01 3.17894161e-01
-9.69577610e-01 -3.20288867e-01 -8.86812881e-02 1.72106415e-01
-3.22011977e-01 5.58668911e-01 -6.97403669e-01 -2.52969652e-01
-2.06064984e-01 -1.15288317e+00 9.47155774e-01 -7.61384889e-02
-3.21246497e-02 -1.37838113e+00 1.00855090e-01 -6.18659817e-02
2.35282704e-01 3.41260672e-01 1.00073063e+00 -3.85847807e-01
-9.35203552e-01 -3.55447054e-01 -3.05329025e-01 7.62789026e-02
3.00152987e-01 -9.22113881e-02 -2.54525244e-01 -6.44898415e-01
-4.31180894e-01 -8.86089355e-02 9.01602805e-01 2.73308039e-01
1.13702810e+00 -5.50156593e-01 -6.07311428e-01 8.59409750e-01
1.71485722e+00 -1.82762906e-01 6.83487415e-01 8.66812542e-02
1.17723370e+00 2.36279741e-01 1.66015804e-01 1.58202603e-01
7.56527185e-01 7.95353532e-01 6.29694343e-01 -5.29200658e-02
-4.91810054e-01 -6.21446788e-01 3.12886029e-01 1.15774500e+00
-1.65115356e-01 -5.93997180e-01 -1.06285226e+00 5.39240360e-01
-1.89883304e+00 -7.50825524e-01 -8.12216923e-02 2.03663564e+00
4.50458676e-01 2.43143275e-01 -4.66912426e-02 5.72321676e-02
6.21127307e-01 5.83045542e-01 -2.05322519e-01 -7.70214349e-02
1.02424704e-01 -5.40075032e-03 7.35794246e-01 9.28849280e-01
-8.08344066e-01 1.08975124e+00 6.38910913e+00 4.90127951e-01
-1.15403020e+00 -2.16655284e-02 2.73451149e-01 9.67710987e-02
-6.28376484e-01 2.13856548e-01 -5.53393304e-01 3.67904365e-01
5.43907940e-01 -4.89724308e-01 5.88383138e-01 6.36407852e-01
-2.95662493e-01 2.80372858e-01 -1.25974751e+00 9.00180638e-01
8.71498957e-02 -1.40956104e+00 1.94800064e-01 1.22622393e-01
4.18256849e-01 4.12314646e-02 7.54332021e-02 1.78433031e-01
5.13955772e-01 -9.80466366e-01 6.43540263e-01 4.85662758e-01
8.25901508e-01 -5.87026358e-01 5.06290495e-01 3.24134119e-02
-1.86523902e+00 1.64097205e-01 -1.53426886e-01 -1.68080062e-01
5.17347567e-02 6.08297706e-01 -8.64816189e-01 1.03770590e+00
4.82324511e-01 1.06317878e+00 -8.19192052e-01 7.50244915e-01
-6.55169964e-01 6.58234656e-01 -1.90379843e-01 2.05381200e-01
1.89060718e-01 -5.07022798e-01 7.54002094e-01 9.85035956e-01
2.84100175e-01 -2.07484071e-03 3.51183176e-01 7.38266885e-01
-3.29421967e-01 -2.90128943e-02 -9.41998541e-01 -8.78315493e-02
7.67657638e-01 1.28207541e+00 -7.91539848e-01 -2.10661933e-01
-3.58762145e-01 7.41836667e-01 8.87797296e-01 4.43867236e-01
-8.63105893e-01 -4.40622091e-01 3.77998263e-01 3.87453824e-01
2.59961486e-01 -5.16853452e-01 2.27505323e-02 -1.09142005e+00
1.62063703e-01 -6.18630886e-01 5.35463035e-01 -7.63732195e-01
-9.24691856e-01 6.68877661e-01 2.22447425e-01 -1.04308724e+00
-1.46539330e-01 -4.55659360e-01 -6.42399549e-01 5.02090752e-01
-1.21196163e+00 -1.45858824e+00 -3.74767512e-01 5.56070924e-01
2.11205706e-02 -1.75900497e-02 5.82076192e-01 2.42252126e-01
-6.05624378e-01 8.57340038e-01 -3.35900784e-01 2.73996651e-01
3.99115890e-01 -1.49890757e+00 5.62378824e-01 8.01600277e-01
4.40594107e-01 9.04883444e-01 5.88745594e-01 -6.67630017e-01
-1.56310940e+00 -1.11216879e+00 9.09489989e-01 -3.17163706e-01
7.92243659e-01 -6.07515335e-01 -8.45504344e-01 1.47672617e+00
3.61978173e-01 2.78865218e-01 3.54633689e-01 2.53211766e-01
-6.28514349e-01 -2.58603960e-01 -8.78129661e-01 6.31759763e-01
1.30021381e+00 -5.90694845e-01 -4.31227237e-01 3.72778088e-01
7.32772946e-01 -7.48664260e-01 -9.29428875e-01 5.94781041e-01
2.91644454e-01 -8.29177737e-01 1.02811372e+00 -1.81137443e-01
1.99263930e-01 -4.66860712e-01 -7.53935277e-02 -1.62111199e+00
-7.15302050e-01 -7.10324824e-01 -3.39033931e-01 1.26255751e+00
3.07985455e-01 -8.72203052e-01 8.48002374e-01 4.63952310e-02
-3.34017664e-01 -9.95982528e-01 -8.01926970e-01 -6.61752522e-01
5.78599162e-02 -9.49178189e-02 6.56245053e-01 9.57992673e-01
3.21534812e-01 5.18077791e-01 -3.84292305e-01 4.48621184e-01
6.38202012e-01 2.14354560e-01 6.00425720e-01 -1.01982033e+00
-3.11099946e-01 -5.83666146e-01 -9.35193837e-01 -1.36821008e+00
2.61111110e-01 -1.46882713e+00 -3.31432104e-01 -2.00226188e+00
1.16863102e-01 -3.51694614e-01 -4.57883447e-01 6.95554554e-01
1.41423151e-01 2.96820998e-01 1.10957086e-01 -8.74498263e-02
-6.69691205e-01 5.44411838e-01 1.18047869e+00 -2.02562094e-01
3.26206163e-02 -2.29909644e-01 -7.39166975e-01 7.23896921e-01
1.03275764e+00 -3.08119088e-01 -7.97194541e-01 -5.74630260e-01
4.21881288e-01 1.00560486e-01 4.32515353e-01 -1.00085354e+00
5.07701159e-01 2.15163231e-01 8.35778862e-02 -5.09640872e-01
2.07831904e-01 -7.34510899e-01 2.22003579e-01 4.97672468e-01
-2.54768163e-01 3.18596035e-01 9.00466219e-02 6.20291352e-01
-2.51536101e-01 6.81657046e-02 4.99997795e-01 -2.17414778e-02
-6.49398744e-01 4.94759351e-01 1.81010395e-01 -1.26523497e-02
9.57895577e-01 4.20998037e-02 -5.34620464e-01 -6.95175171e-01
-8.15091372e-01 3.01262110e-01 5.37392139e-01 5.06107986e-01
6.18569553e-01 -1.59118032e+00 -4.47314709e-01 2.55080879e-01
7.51616061e-02 -1.77246973e-01 -1.90994442e-02 9.37069356e-01
-5.69564700e-01 5.11503398e-01 -2.52191555e-02 -3.83015543e-01
-1.38307714e+00 5.89936972e-01 5.22731543e-01 -4.16229308e-01
-8.04700077e-01 1.06799543e+00 2.72725403e-01 -5.08148015e-01
3.21112961e-01 -4.72702533e-01 -5.81486486e-02 -1.61348701e-01
6.21879213e-02 3.23344558e-01 4.85217944e-02 -7.69595265e-01
-4.26582426e-01 7.62675464e-01 -2.33626410e-01 2.73476928e-01
1.15371883e+00 -7.15364069e-02 -5.02164841e-01 4.92319494e-01
1.27609479e+00 3.50982428e-01 -8.47429693e-01 -3.66478741e-01
-1.13341115e-01 -4.56855863e-01 2.12481141e-01 -4.96604621e-01
-1.46193874e+00 7.21282303e-01 2.22013369e-01 3.56043726e-01
1.08900714e+00 1.60700023e-01 5.97727537e-01 1.24643028e-01
7.10041761e-01 -3.73583436e-01 9.09044594e-02 5.11168003e-01
1.20997262e+00 -7.62062252e-01 2.77483493e-01 -1.01745951e+00
-2.83263266e-01 9.67075646e-01 7.03592181e-01 -3.87173444e-01
9.08555090e-01 4.45140839e-01 -2.43694752e-01 -4.44524407e-01
-7.28075087e-01 -6.51353449e-02 5.42254627e-01 4.43424433e-01
7.18306720e-01 2.29966611e-01 -2.43260577e-01 3.32006067e-01
-5.56299329e-01 -3.73243123e-01 4.22217935e-01 7.68910766e-01
-4.70495038e-02 -1.40894878e+00 1.29002437e-01 2.33399034e-01
-1.09976716e-01 -2.61930972e-01 -6.62157774e-01 9.52820003e-01
-1.62870362e-01 6.11198664e-01 2.34881476e-01 -6.78419709e-01
2.85066307e-01 -2.43796542e-01 7.66946018e-01 -6.03443563e-01
-1.34282961e-01 1.08879758e-02 9.15050134e-02 -6.37523949e-01
-1.72894970e-01 -4.50575113e-01 -1.37262225e+00 -3.90121758e-01
-2.09326044e-01 2.14632884e-01 2.95325428e-01 3.33992273e-01
4.44792002e-01 8.01817417e-01 4.98033017e-01 -6.94029331e-01
-2.49751940e-01 -7.59956539e-01 -6.75608814e-01 1.68401122e-01
5.24098754e-01 -7.69758880e-01 -3.36222976e-01 -2.29402587e-01]
|
[7.009444713592529, 6.286719799041748]
|
fe358efc-7f14-4230-88e2-7f9100b83545
|
a-context-based-approach-for-dialogue-act
|
1805.06280
| null |
http://arxiv.org/abs/1805.06280v1
|
http://arxiv.org/pdf/1805.06280v1.pdf
|
A Context-based Approach for Dialogue Act Recognition using Simple Recurrent Neural Networks
|
Dialogue act recognition is an important part of natural language
understanding. We investigate the way dialogue act corpora are annotated and
the learning approaches used so far. We find that the dialogue act is
context-sensitive within the conversation for most of the classes.
Nevertheless, previous models of dialogue act classification work on the
utterance-level and only very few consider context. We propose a novel
context-based learning method to classify dialogue acts using a character-level
language model utterance representation, and we notice significant improvement.
We evaluate this method on the Switchboard Dialogue Act corpus, and our results
show that the consideration of the preceding utterances as a context of the
current utterance improves dialogue act detection.
|
['Cornelius Weber', 'Sven Magg', 'Stefan Wermter', 'Chandrakant Bothe']
|
2018-05-16
|
a-context-based-approach-for-dialogue-act-1
|
https://aclanthology.org/L18-1307
|
https://aclanthology.org/L18-1307.pdf
|
lrec-2018-5
|
['dialogue-act-classification']
|
['natural-language-processing']
|
[ 5.38460255e-01 5.11814237e-01 -2.03697905e-01 -8.83582354e-01
-3.31715107e-01 -7.55470991e-01 1.22028196e+00 3.95225078e-01
-3.52009177e-01 1.05984318e+00 1.00366974e+00 -4.82569665e-01
1.58017397e-01 -5.82440376e-01 4.11911339e-01 -2.87944794e-01
1.14417247e-01 7.10058033e-01 4.58420366e-01 -9.40838814e-01
6.90862715e-01 1.54522091e-01 -9.97913837e-01 8.57831538e-01
4.23039526e-01 4.52859342e-01 -2.85293490e-01 1.28877032e+00
-6.31025195e-01 1.53330910e+00 -1.20932341e+00 -8.75381380e-02
-1.65220439e-01 -1.06891322e+00 -1.90495193e+00 4.27067041e-01
1.16303181e-02 -3.84553343e-01 -1.69513181e-01 4.05864745e-01
1.89167783e-01 2.96831131e-01 7.46015429e-01 -8.67159009e-01
3.25424254e-01 7.12975979e-01 3.40070367e-01 4.21216846e-01
1.21563590e+00 -3.52728628e-02 1.09232426e+00 -2.20134020e-01
7.38753021e-01 1.38642848e+00 3.74047577e-01 1.06735551e+00
-1.06906688e+00 1.22435041e-01 1.97954714e-01 7.48130679e-02
-5.18769920e-01 -5.03811419e-01 7.57198334e-01 -4.32178646e-01
1.53163445e+00 4.47236210e-01 5.94378531e-01 9.38833356e-01
1.72010079e-01 1.04546535e+00 1.21167088e+00 -1.06983399e+00
3.22155416e-01 5.55048622e-02 8.00364017e-01 5.11280656e-01
-8.84200990e-01 -3.67364496e-01 -1.35591313e-01 -4.29151684e-01
5.92098057e-01 -3.27232808e-01 8.72387923e-03 3.72174233e-01
-6.55930817e-01 1.17261744e+00 -2.93115139e-01 8.64611030e-01
-7.93548822e-02 -2.74265289e-01 9.94281709e-01 7.48620391e-01
6.25997365e-01 7.05662787e-01 -7.02534616e-01 -9.42593813e-01
-2.81820983e-01 3.61795008e-01 1.68573916e+00 5.70059538e-01
5.62708378e-01 -1.90884456e-01 -2.34280854e-01 1.11972082e+00
1.87348858e-01 -3.81227046e-01 4.62688863e-01 -1.09842801e+00
3.12348366e-01 8.80751014e-01 1.37652084e-01 -5.15099466e-01
-6.41082525e-01 6.74376786e-01 -9.07528773e-02 5.20465411e-02
7.60936379e-01 -4.62591708e-01 -3.30296278e-01 1.24015629e+00
3.57387662e-01 -3.25672060e-01 3.12083244e-01 4.08562422e-01
8.57387900e-01 8.17440629e-01 3.37725222e-01 -6.45591795e-01
1.50573432e+00 -8.97260249e-01 -1.17211509e+00 -1.66000769e-01
1.16434801e+00 -6.92866981e-01 7.84797072e-01 3.94682676e-01
-9.56759334e-01 -3.66768271e-01 -8.24393034e-01 3.25437896e-02
-3.54768425e-01 -1.60030067e-01 7.78455555e-01 8.76877964e-01
-8.01067352e-01 2.99163699e-01 -4.85194862e-01 -5.28842688e-01
-3.46906394e-01 2.94620603e-01 -1.75205246e-01 3.18484306e-01
-1.26950061e+00 1.25051606e+00 6.45442545e-01 -3.93285215e-01
-3.04654092e-01 1.34102628e-01 -1.01275325e+00 -1.63704917e-01
3.57875407e-01 9.55355465e-02 1.96498096e+00 -1.00945890e+00
-2.01206207e+00 8.77476811e-01 -2.88493395e-01 -4.09124315e-01
2.02802032e-01 4.58422713e-02 -3.11755478e-01 2.21374527e-01
-4.89557564e-01 9.48139727e-02 5.66651583e-01 -9.81895864e-01
-1.10145855e+00 -8.30479339e-02 7.27454841e-01 5.59464216e-01
5.88353211e-03 6.24931872e-01 9.86902267e-02 -1.31100997e-01
-1.80960864e-01 -7.36866772e-01 -1.96016252e-01 -6.80840313e-01
1.13353007e-01 -1.04450810e+00 8.29881787e-01 -6.23308063e-01
1.52084959e+00 -1.79046786e+00 -5.81004396e-02 -1.83345288e-01
1.87444910e-01 3.37402940e-01 3.07436109e-01 8.98882091e-01
-1.14490181e-01 5.49264923e-02 -1.97110519e-01 -2.71630913e-01
2.00666294e-01 6.80921257e-01 -1.32895544e-01 1.66742310e-01
-1.49173085e-02 4.75829273e-01 -9.78609443e-01 -4.90981936e-01
6.76723480e-01 -1.03577755e-01 -3.56154561e-01 7.61292398e-01
-4.92436320e-01 5.23139238e-01 -5.24536490e-01 1.96178198e-01
7.64894998e-03 2.50015587e-01 6.76112652e-01 3.72900397e-01
-2.09865451e-01 1.19579554e+00 -7.29327619e-01 1.35187793e+00
-5.88469565e-01 8.71136725e-01 1.16558641e-01 -9.52789187e-01
1.03781426e+00 9.29421782e-01 3.27687949e-01 -3.30458850e-01
9.12214667e-02 -7.62207434e-03 5.59783816e-01 -6.28577352e-01
4.38340247e-01 -4.13664311e-01 -4.78857756e-01 8.32510710e-01
2.01638490e-01 -4.40325499e-01 4.87128317e-01 2.49627113e-01
1.24212790e+00 -3.95116955e-02 1.08137989e+00 2.54383255e-02
1.00753367e+00 1.58158898e-01 1.76340848e-01 6.05136991e-01
-7.20522046e-01 1.32307634e-01 8.43024790e-01 -8.71723592e-01
-7.55535185e-01 -2.56558150e-01 -1.71730191e-01 1.75811613e+00
-4.62233365e-01 -6.99551105e-01 -9.22762513e-01 -1.11473596e+00
-5.65381467e-01 9.47757125e-01 -3.10468405e-01 1.96801901e-01
-1.04887080e+00 -5.47125995e-01 6.16466522e-01 3.29252630e-01
4.95373636e-01 -1.41682768e+00 -5.09681046e-01 5.86366773e-01
-2.73287922e-01 -1.02474749e+00 -8.92492533e-02 2.71528959e-01
-6.86105728e-01 -1.20590186e+00 4.71603125e-02 -8.79105866e-01
2.74879783e-01 -3.26676935e-01 1.16313255e+00 3.52426827e-01
7.45034516e-02 6.93084359e-01 -9.06112552e-01 -3.42070043e-01
-1.36243951e+00 3.71388858e-03 -2.56738335e-01 -2.73891389e-01
8.85306060e-01 -1.81853399e-01 9.52298492e-02 4.08132434e-01
-5.18328786e-01 -1.22890301e-01 -2.49371022e-01 1.01945233e+00
-5.82714498e-01 2.03111153e-02 5.70869148e-01 -1.22338009e+00
1.37279439e+00 -1.61106452e-01 1.13738887e-02 1.46373838e-01
-2.47210860e-01 -6.03038631e-02 5.36876976e-01 -1.06798997e-02
-1.55584514e+00 4.68499660e-02 -7.59638667e-01 7.26173818e-01
-1.03946662e+00 2.36444354e-01 -5.94815016e-02 1.02945656e-01
7.37048209e-01 8.43935683e-02 6.48469552e-02 -3.12264740e-01
6.32320419e-02 9.94952738e-01 3.74248959e-02 -7.22410679e-01
-3.91203612e-02 -8.46954584e-02 -4.24421012e-01 -1.30347717e+00
-5.24467647e-01 -9.84478533e-01 -9.02585626e-01 -6.15648866e-01
9.32622313e-01 -3.19105417e-01 -9.92446899e-01 2.46673420e-01
-1.45338476e+00 -4.69534218e-01 -1.60912126e-01 3.53344828e-01
-8.69867682e-01 6.70866013e-01 -1.09134972e+00 -1.44766533e+00
-7.63145313e-02 -8.40981364e-01 7.09779203e-01 1.34340301e-01
-1.01271319e+00 -1.51069248e+00 4.26897466e-01 4.72417295e-01
4.10486050e-02 6.38125390e-02 1.05408347e+00 -1.53947854e+00
3.68579447e-01 -2.42681652e-01 2.71483898e-01 2.59134203e-01
5.47097325e-01 -9.99193937e-02 -9.85438049e-01 1.19329892e-01
6.81660771e-01 -7.15696037e-01 3.41748953e-01 -9.97730419e-02
4.01039630e-01 -6.91541493e-01 -9.06272531e-02 -5.20068824e-01
9.05600011e-01 9.84245360e-01 7.17042804e-01 5.43754511e-02
1.82288080e-01 1.13759720e+00 8.20830464e-01 5.10264099e-01
3.24436963e-01 6.60676658e-01 -1.95012311e-03 1.24412365e-01
3.07845265e-01 1.36135351e-02 4.48532730e-01 9.17111516e-01
-2.85563558e-01 -2.84516454e-01 -1.13458693e+00 3.99541855e-01
-2.00365710e+00 -1.21508956e+00 -1.96613982e-01 1.56892395e+00
1.34617698e+00 4.45426077e-01 5.52148819e-01 4.62495655e-01
5.23819208e-01 4.78431076e-01 2.23553538e-01 -1.20226574e+00
3.91673177e-01 2.27920383e-01 -2.45919123e-01 1.12425995e+00
-1.37954819e+00 1.05027425e+00 7.59224558e+00 4.61458325e-01
-6.10637486e-01 5.09847514e-02 4.90938038e-01 4.43431824e-01
2.95191079e-01 -2.31263749e-02 -7.49912500e-01 1.49205953e-01
1.14184892e+00 2.16847472e-02 1.68131873e-01 8.23745251e-01
2.78725713e-01 -4.07362431e-01 -1.45837355e+00 5.73920250e-01
1.86364174e-01 -1.17880404e+00 -2.84801543e-01 -1.71249852e-01
2.15157241e-01 -5.89140832e-01 -8.91324401e-01 6.95461333e-01
4.33194369e-01 -8.54011238e-01 -6.25935271e-02 1.47622183e-01
1.01188786e-01 -5.97816229e-01 9.35046196e-01 6.49127066e-01
-9.90649641e-01 -4.76132631e-02 -1.86091259e-01 -7.83353925e-01
2.28070214e-01 -2.12704718e-01 -1.60987628e+00 1.70244440e-01
-7.51426667e-02 5.35791337e-01 -3.55828315e-01 4.62564439e-01
-4.18238372e-01 9.81252611e-01 -1.11634828e-01 -5.50447583e-01
3.58408719e-01 -3.07020247e-01 6.53741300e-01 1.70012963e+00
-7.63130784e-01 7.10594714e-01 7.35293806e-01 2.29439642e-02
2.90178567e-01 2.61668980e-01 -6.46620929e-01 -7.98655525e-02
1.04149655e-01 9.40898299e-01 -6.53725982e-01 -5.25740206e-01
-5.26203990e-01 7.30092824e-01 -3.02933101e-02 -1.33219436e-01
-1.44124597e-01 -2.27451354e-01 5.52458704e-01 -3.90677869e-01
-1.63625583e-01 -2.56942004e-01 -1.63895696e-01 -8.49081337e-01
-2.46791080e-01 -1.13882458e+00 5.44685066e-01 -1.45769447e-01
-1.16811609e+00 5.29630184e-01 2.28319630e-01 -8.46740425e-01
-1.03395259e+00 -8.64776254e-01 -1.00734818e+00 7.08871722e-01
-8.69293988e-01 -7.45071530e-01 1.49172857e-01 3.72549713e-01
1.42763078e+00 -3.37192833e-01 1.41773415e+00 -1.91168681e-01
-7.62700066e-02 1.65589228e-01 -4.01077211e-01 7.20141828e-01
6.52015686e-01 -1.67780793e+00 3.06944042e-01 1.07038975e-01
5.00278138e-02 6.13176405e-01 9.49259877e-01 -5.49441755e-01
-7.58322120e-01 -2.11450920e-01 1.35483515e+00 -6.65025413e-01
5.99538147e-01 -3.98803979e-01 -1.03861201e+00 5.91841877e-01
8.56682956e-01 -8.26229036e-01 1.07460225e+00 6.02637231e-01
-7.07037449e-02 4.86503422e-01 -1.13885820e+00 4.52054739e-01
6.01647973e-01 -9.54334497e-01 -1.66770518e+00 5.58021069e-01
5.59753418e-01 -3.79452914e-01 -8.81894708e-01 3.46194580e-02
3.87683004e-01 -9.25603390e-01 4.28257287e-01 -9.95061696e-01
3.77611637e-01 1.53631866e-01 9.10466760e-02 -1.24883509e+00
1.19428225e-01 -8.61906171e-01 1.31031826e-01 1.19465411e+00
4.46386278e-01 -4.31473255e-01 7.42107928e-01 1.11962962e+00
-2.73656219e-01 -3.46152067e-01 -9.20560539e-01 -2.23068237e-01
7.89778829e-02 -2.67159909e-01 1.73208192e-01 1.09176362e+00
1.32690167e+00 1.08993399e+00 -3.43908399e-01 -8.33677888e-01
-2.21301436e-01 -1.86562508e-01 7.60739028e-01 -1.26486731e+00
-3.13283622e-01 -4.84042317e-01 -4.27124768e-01 -1.43823528e+00
2.45867014e-01 -3.08732212e-01 3.59921873e-01 -1.43157578e+00
-3.22924137e-01 -4.16437956e-03 4.38579202e-01 5.59403479e-01
-2.22332582e-01 -4.76598501e-01 6.20755106e-02 1.03404574e-01
-5.80046058e-01 2.66986370e-01 8.60478282e-01 -1.80290759e-01
-3.94103199e-01 4.50604320e-01 -9.70089138e-02 9.48809683e-01
1.07603538e+00 -1.22149289e-01 -3.31756145e-01 3.39099407e-01
-2.95732796e-01 3.72592062e-01 -4.09147650e-01 -6.71110809e-01
1.79551929e-01 -4.48885649e-01 -4.71831374e-02 -3.60070646e-01
5.31132877e-01 -7.00755775e-01 -6.27843797e-01 4.67859566e-01
-9.86133039e-01 -2.27219209e-01 2.83860173e-02 3.96371961e-01
-3.84469450e-01 -8.94541442e-01 6.66554987e-01 -5.19836903e-01
-8.30435872e-01 -5.19650221e-01 -1.38209367e+00 7.12311491e-02
9.01939571e-01 -2.25379258e-01 -2.94695109e-01 -7.95415938e-01
-8.76655638e-01 7.55363330e-02 1.39180109e-01 2.95502245e-01
3.47092956e-01 -7.33270288e-01 -7.39857674e-01 -3.82458158e-02
4.83847708e-02 -4.33011889e-01 -2.45741040e-01 2.45328158e-01
-5.42872906e-01 7.12215602e-01 -2.12606579e-01 -4.75822955e-01
-1.83914483e+00 6.49851039e-02 5.24260342e-01 -5.08695602e-01
-3.64279896e-01 5.80278873e-01 -1.64930522e-01 -7.20328271e-01
4.71201479e-01 -2.21539497e-01 -1.00060284e+00 3.01322699e-01
8.00114512e-01 1.82004601e-01 -1.28591940e-01 -6.77134633e-01
-8.20435286e-02 -1.41024336e-01 -3.46435785e-01 -4.57387060e-01
9.81811166e-01 -2.00145617e-01 -3.06984037e-01 1.03871369e+00
8.92590284e-01 -1.89051837e-01 -8.28316033e-01 -3.57572645e-01
5.49527407e-01 -4.28044021e-01 -5.55324495e-01 -8.58317971e-01
2.49444265e-02 6.60368443e-01 1.40846923e-01 1.23694646e+00
8.06038737e-01 -4.80214916e-02 3.83958787e-01 7.74333119e-01
5.15325069e-01 -1.68975830e+00 2.22618088e-01 1.41490281e+00
8.97857428e-01 -1.32557559e+00 1.64190084e-01 -5.99030077e-01
-9.48028564e-01 1.46024823e+00 8.57810855e-01 4.55055088e-02
5.24674237e-01 1.44135937e-01 3.83912504e-01 -2.35130787e-01
-1.10981500e+00 -1.90457493e-01 -2.13316113e-01 5.61209917e-01
1.04341626e+00 -5.31032402e-03 -9.64663804e-01 4.14028317e-02
-1.33896887e-01 -3.52169216e-01 7.33883977e-01 1.46900833e+00
-7.76725590e-01 -1.66949821e+00 -3.06337327e-01 2.59786636e-01
-4.70144272e-01 1.14431620e-01 -1.38207150e+00 8.13726485e-01
-3.56934637e-01 1.56294155e+00 9.85842124e-02 -3.82971466e-01
2.60016710e-01 1.01032317e+00 3.83616865e-01 -1.29504991e+00
-1.47356665e+00 1.04388237e-01 1.16260874e+00 -2.43515134e-01
-9.46700513e-01 -6.50130987e-01 -1.33639705e+00 -2.69457906e-01
-3.46752375e-01 6.93671644e-01 3.40627193e-01 1.61239743e+00
-5.26624203e-01 2.65050262e-01 9.94640887e-01 -6.12949133e-01
-5.09528160e-01 -1.40979588e+00 -3.03967953e-01 6.40330195e-01
3.24578524e-01 -1.53729603e-01 -3.71015638e-01 1.52899384e-01]
|
[12.918439865112305, 7.931952476501465]
|
bafa5ac6-dc7b-4672-bd09-3c14723aa2da
|
rfr-wwanet-weighted-window-attention-based
|
2305.04236
| null |
https://arxiv.org/abs/2305.04236v2
|
https://arxiv.org/pdf/2305.04236v2.pdf
|
RFR-WWANet: Weighted Window Attention-Based Recovery Feature Resolution Network for Unsupervised Image Registration
|
The Swin transformer has recently attracted attention in medical image analysis due to its computational efficiency and long-range modeling capability. Owing to these properties, the Swin Transformer is suitable for establishing more distant relationships between corresponding voxels in different positions in complex abdominal image registration tasks. However, the registration models based on transformers combine multiple voxels into a single semantic token. This merging process limits the transformers to model and generate coarse-grained spatial information. To address this issue, we propose Recovery Feature Resolution Network (RFRNet), which allows the transformer to contribute fine-grained spatial information and rich semantic correspondences to higher resolution levels. Furthermore, shifted window partitioning operations are inflexible, indicating that they cannot perceive the semantic information over uncertain distances and automatically bridge the global connections between windows. Therefore, we present a Weighted Window Attention (WWA) to build global interactions between windows automatically. It is implemented after the regular and cyclic shift window partitioning operations within the Swin transformer block. The proposed unsupervised deformable image registration model, named RFR-WWANet, detects the long-range correlations, and facilitates meaningful semantic relevance of anatomical structures. Qualitative and quantitative results show that RFR-WWANet achieves significant improvements over the current state-of-the-art methods. Ablation experiments demonstrate the effectiveness of the RFRNet and WWA designs. Our code is available at \url{https://github.com/MingR-Ma/RFR-WWANet}.
|
['Guixia Liu', 'Weijie Wang', 'Lei Song', 'Tao Wang', 'Mingrui Ma']
|
2023-05-07
| null | null | null | null |
['image-registration', 'long-range-modeling']
|
['computer-vision', 'natural-language-processing']
|
[ 2.87709355e-01 1.28330097e-01 -2.92216986e-01 -3.63967359e-01
-7.62930393e-01 -1.51274651e-01 2.94077545e-01 2.51488209e-01
-3.88292283e-01 3.26618522e-01 5.38305879e-01 4.09713686e-02
-7.93493807e-01 -8.18148196e-01 -4.16169524e-01 -7.85523713e-01
-3.13114285e-01 2.51169294e-01 5.79750359e-01 -3.46279472e-01
7.14597106e-02 3.58251005e-01 -1.05966580e+00 3.96679163e-01
1.17427516e+00 8.61675620e-01 6.75711215e-01 -4.96837236e-02
-2.03721598e-01 1.22087210e-01 -2.41255149e-01 -4.79951315e-02
8.21501836e-02 -2.03554228e-01 -8.57982874e-01 -4.86445755e-01
-3.77082196e-03 3.39485034e-02 -3.18690300e-01 1.08403683e+00
6.71464086e-01 4.00984675e-01 3.50190252e-01 -7.48672724e-01
-5.65509379e-01 9.19196725e-01 -8.39295328e-01 5.96762002e-01
6.23275638e-02 -2.01637149e-01 8.80779207e-01 -1.02704966e+00
5.84271550e-01 8.96735847e-01 7.72073090e-01 3.51385981e-01
-1.18406677e+00 -8.30888271e-01 3.28981519e-01 -4.46419604e-02
-1.66095197e+00 -1.23638794e-01 8.18601787e-01 -3.17335486e-01
7.56665170e-01 6.89584970e-01 6.73683405e-01 8.43658447e-01
5.25409877e-01 3.70807350e-01 9.85894620e-01 -5.67744784e-02
-1.68379247e-01 -2.07677230e-01 2.11477146e-01 6.47308230e-01
2.32086033e-01 8.70750248e-02 -4.86703992e-01 -4.90587465e-02
1.26528573e+00 3.71623576e-01 -5.50098658e-01 -2.26320833e-01
-1.60584891e+00 7.17157543e-01 1.14089286e+00 8.87902975e-01
-2.36338377e-01 -2.86706351e-02 2.45060742e-01 -1.57915384e-01
6.23188317e-01 6.81003034e-01 -1.75742626e-01 3.38448614e-01
-9.64553654e-01 -1.23610996e-01 -6.59107417e-02 7.84263909e-01
3.62134308e-01 -4.44483995e-01 -4.69092578e-01 9.71017182e-01
3.27443272e-01 -7.17019439e-02 8.58306885e-01 -5.58882177e-01
3.65209639e-01 7.25383997e-01 -3.55519354e-01 -1.26082253e+00
-8.95234644e-01 -7.81574607e-01 -1.16595805e+00 -9.93652642e-02
1.56352460e-01 3.59551519e-01 -8.91784668e-01 1.59057879e+00
4.38346475e-01 3.11330676e-01 -3.34747672e-01 1.16969752e+00
1.27143645e+00 3.48705590e-01 1.16964296e-01 -1.52214050e-01
1.56009459e+00 -8.95076036e-01 -7.31212974e-01 -1.60835087e-01
5.78699946e-01 -7.22971559e-01 1.07883024e+00 -1.34257197e-01
-1.06363988e+00 -6.48869693e-01 -9.46631193e-01 -1.23240255e-01
-2.94410735e-01 -5.90963028e-02 6.15382195e-01 9.04096738e-02
-9.00950193e-01 7.36517966e-01 -1.11462784e+00 -2.22741738e-01
6.53832257e-01 5.67535102e-01 -4.35027868e-01 -7.74871334e-02
-1.39443398e+00 8.40665936e-01 2.26848647e-01 6.36389732e-01
-3.48502785e-01 -9.92201984e-01 -1.00558424e+00 -6.77385181e-02
2.91901886e-01 -5.27274132e-01 6.64700866e-01 -3.99040908e-01
-1.03881419e+00 9.06079590e-01 -2.00672269e-01 -8.27146694e-02
5.38597107e-01 -8.38202685e-02 -3.58938843e-01 1.04441240e-01
2.69878000e-01 5.16111255e-01 4.40461606e-01 -1.23597288e+00
-1.99976832e-01 -6.33437157e-01 -1.59045205e-01 4.27217603e-01
-2.59614736e-01 3.38910371e-02 -4.87621546e-01 -1.34022367e+00
6.44514501e-01 -6.59808040e-01 -5.30604124e-01 1.11540839e-01
-5.41087866e-01 -1.99510515e-01 3.98151726e-01 -6.22787595e-01
1.48267126e+00 -2.17211413e+00 2.42293656e-01 6.01984322e-01
5.83185315e-01 -8.17686766e-02 -1.42814502e-01 3.70038338e-02
-3.99036169e-01 2.28356764e-01 -4.49187726e-01 -5.14239408e-02
-3.63834500e-01 1.31095439e-01 8.04453939e-02 5.54761946e-01
-1.21971019e-01 1.10823786e+00 -9.47463036e-01 -7.22653747e-01
2.73433030e-01 5.81182957e-01 -3.49383831e-01 1.99297536e-02
4.54911649e-01 8.03605795e-01 -8.02319348e-01 5.92170715e-01
8.14508200e-01 -2.73835391e-01 -1.40658133e-02 -7.78953075e-01
-1.36371493e-01 3.86901014e-02 -9.51608777e-01 2.21630049e+00
-3.20245534e-01 1.15005039e-01 -6.16988055e-02 -1.01865351e+00
1.00767124e+00 1.83456346e-01 9.54667985e-01 -9.12768900e-01
2.64598787e-01 2.49955505e-01 8.34144726e-02 -4.92540717e-01
3.44272226e-01 -1.01511247e-01 -7.91732371e-02 9.03925207e-03
-2.24736303e-01 6.30486310e-02 -1.80497497e-01 1.15574617e-02
8.14110100e-01 6.73379898e-02 3.08321297e-01 -7.14848578e-01
4.75021869e-01 -1.89993963e-01 7.63263762e-01 4.80187207e-01
-7.70718381e-02 8.56542230e-01 -6.39497936e-02 -4.88238752e-01
-4.72856164e-01 -1.31685853e+00 -6.11768842e-01 9.09556985e-01
9.28026676e-01 -3.08100671e-01 -5.57690620e-01 -5.51518261e-01
-1.57000825e-01 2.09343001e-01 -1.03377891e+00 -3.55796337e-01
-8.19672644e-01 -8.95705640e-01 2.91674554e-01 7.80240417e-01
3.92778158e-01 -1.04083300e+00 -5.97617865e-01 2.83131093e-01
-5.38593054e-01 -8.40487778e-01 -9.58267391e-01 3.66842896e-02
-1.11126399e+00 -9.60285664e-01 -9.68462825e-01 -8.40027452e-01
1.02949250e+00 3.43122035e-01 8.64579916e-01 3.04942489e-01
-6.69379592e-01 -2.06322074e-01 -4.46797639e-01 2.48459578e-02
1.69616476e-01 2.25137815e-01 -1.39161333e-01 -1.47897854e-01
-9.90708242e-04 -6.99854970e-01 -9.99935031e-01 7.02349246e-01
-8.59438241e-01 3.19689006e-01 5.85772097e-01 9.99926686e-01
9.26743329e-01 -1.19603910e-02 3.72716129e-01 -8.92914534e-01
4.87719536e-01 -3.46244544e-01 -3.24889213e-01 3.60450208e-01
-4.79073137e-01 1.23309910e-01 1.58984095e-01 -5.50686955e-01
-1.00084019e+00 -2.35989958e-01 -2.53150225e-01 -3.29857796e-01
1.50541469e-01 5.77580094e-01 4.39177901e-02 -1.68031931e-01
4.81728792e-01 1.47376567e-01 -2.35064533e-02 -3.33786517e-01
1.15260363e-01 2.04463840e-01 4.76470351e-01 -5.84381580e-01
7.48921990e-01 6.21973693e-01 -1.45914584e-01 -4.46306080e-01
-7.76337981e-01 -4.99697894e-01 -8.30665886e-01 -5.69572188e-02
1.08925974e+00 -7.40429282e-01 -4.68296051e-01 2.03170657e-01
-9.41801488e-01 -1.75443813e-01 -5.60002267e-01 7.97093570e-01
-3.15731347e-01 1.75849512e-01 -6.38686597e-01 -1.46570325e-01
-6.28122866e-01 -1.43266881e+00 1.00913775e+00 3.55408639e-01
-4.34863061e-01 -1.03504241e+00 -8.05287436e-02 3.62873256e-01
5.76538920e-01 5.12813210e-01 8.24440241e-01 -6.12111986e-01
-4.86285269e-01 7.28169531e-02 -3.96299601e-01 -3.24510068e-01
3.86353940e-01 -4.94897306e-01 -6.91264033e-01 -3.26869905e-01
-1.20524883e-01 3.31074417e-01 9.41555321e-01 8.64277124e-01
1.64180827e+00 -9.16369185e-02 -6.67665899e-01 9.60695982e-01
1.20387399e+00 2.33945668e-01 6.82639062e-01 3.52605194e-01
8.08145344e-01 5.99967003e-01 7.35283196e-01 2.40727186e-01
2.91482985e-01 8.27970922e-01 4.44929659e-01 -7.15172708e-01
-1.03300355e-01 -6.60601184e-02 -2.71904886e-01 9.16214824e-01
-6.73431039e-01 2.21058860e-01 -7.67792046e-01 4.70645010e-01
-1.73254836e+00 -7.42237747e-01 -2.49535628e-02 2.13960361e+00
9.15706396e-01 8.30878615e-02 -3.81664217e-01 -1.65045813e-01
8.03608358e-01 4.06090319e-01 -3.95521700e-01 -2.47223321e-02
1.27000406e-01 4.78739232e-01 4.97812301e-01 4.76990759e-01
-1.11604774e+00 6.41590595e-01 4.80803299e+00 1.13114476e+00
-8.99871945e-01 4.38345522e-01 6.22372329e-01 -6.03825971e-03
-5.36418855e-01 -3.31596136e-01 -4.71359044e-01 3.32130402e-01
-1.34454388e-02 -6.78441375e-02 2.38654986e-01 3.51138085e-01
3.13766718e-01 3.34384590e-02 -8.32599282e-01 8.48351121e-01
-7.40846545e-02 -1.43819582e+00 -3.03882249e-02 6.79336488e-02
4.07476097e-01 -6.27595261e-02 1.36361048e-01 -1.03006244e-01
-9.21309218e-02 -1.34887147e+00 4.13389295e-01 5.50867558e-01
1.00815105e+00 -5.56663513e-01 9.15217042e-01 -4.31031659e-02
-1.77927351e+00 5.66488653e-02 -3.29311192e-01 5.15055239e-01
5.35374247e-02 6.36583686e-01 -3.91306013e-01 9.52559650e-01
1.00734377e+00 7.82901764e-01 -5.38593888e-01 1.12363434e+00
5.87579720e-02 3.95211391e-02 -2.14607686e-01 3.38332593e-01
2.39125732e-02 -1.39548153e-01 5.73946714e-01 1.09312499e+00
5.45877516e-01 2.85563856e-01 1.19345687e-01 9.34539914e-01
1.66880146e-01 2.02544302e-01 -2.25615144e-01 4.84621465e-01
4.99791652e-01 1.45279849e+00 -1.17790496e+00 -1.84493922e-02
-8.87513459e-02 8.02155852e-01 1.75677389e-01 2.52828360e-01
-1.05077529e+00 -3.66912901e-01 5.19821048e-01 5.06223500e-01
-8.50942656e-02 2.56284345e-02 -4.23123747e-01 -1.06298971e+00
1.04262397e-01 -4.41841662e-01 5.98178089e-01 -5.21334529e-01
-1.25905633e+00 9.96059477e-01 1.83622599e-01 -1.39608645e+00
5.32386839e-01 -1.55482262e-01 -7.20153511e-01 8.59778285e-01
-1.53530610e+00 -1.18919528e+00 -7.01879144e-01 7.52336860e-01
6.82506979e-01 3.49538386e-01 8.94619763e-01 3.09177756e-01
-4.19140548e-01 7.37060487e-01 -1.62243903e-01 2.39390403e-01
7.16632426e-01 -9.19844627e-01 5.28545305e-02 5.63878894e-01
-2.07204908e-01 9.63987231e-01 3.10846865e-01 -7.19554245e-01
-8.24677050e-01 -1.09624267e+00 4.94405806e-01 -1.17932878e-01
4.28503513e-01 -2.74091423e-01 -1.14893544e+00 4.48612273e-01
-2.35757977e-02 5.42854249e-01 6.38409078e-01 -9.16300621e-03
-1.84431314e-01 -2.74947941e-01 -1.19177198e+00 4.14739639e-01
1.32736337e+00 -1.85402855e-01 -7.95010507e-01 1.90293387e-01
8.92454445e-01 -7.39005327e-01 -1.45403969e+00 8.69355857e-01
5.42236209e-01 -7.79167831e-01 1.45699525e+00 -2.85265706e-02
2.55921513e-01 -3.00082862e-01 1.31753772e-01 -1.24870968e+00
-4.74503785e-01 -3.64589125e-01 3.57711226e-01 1.02541173e+00
4.03970599e-01 -8.02986085e-01 4.05551523e-01 5.77322960e-01
-3.65718901e-01 -1.13622391e+00 -1.34363496e+00 -4.88529861e-01
9.61227641e-02 -1.47688836e-01 7.21705079e-01 1.31896257e+00
1.78014040e-01 -1.58509552e-01 3.83813605e-02 2.71954417e-01
6.30132616e-01 3.41741383e-01 -9.49106291e-02 -1.30796361e+00
-4.25548106e-02 -6.71000659e-01 -3.99067789e-01 -7.84001946e-01
-2.92650700e-01 -1.20594776e+00 -1.19971903e-02 -1.85974908e+00
3.41215670e-01 -8.98269117e-01 -6.54735506e-01 5.71876049e-01
-2.94374734e-01 4.85399812e-01 2.74704657e-02 3.97230268e-01
-3.69332939e-01 5.90201020e-01 1.80287600e+00 -1.01007983e-01
-3.93394172e-01 5.92016913e-02 -6.79062426e-01 7.79631495e-01
7.74765491e-01 -4.66923743e-01 -3.86715382e-01 -4.55812067e-01
-1.43869162e-01 -4.79492964e-03 4.08425927e-01 -8.15497279e-01
3.61074477e-01 -7.96138868e-02 7.65201092e-01 -5.32595694e-01
2.72920996e-01 -8.75587881e-01 1.88618124e-01 5.63878298e-01
-4.08351123e-01 1.32452965e-01 5.64194992e-02 2.33343452e-01
-2.87021041e-01 1.75254300e-01 7.67239630e-01 -2.56751746e-01
-4.89239633e-01 6.28253520e-01 -1.72891952e-02 7.07525909e-02
1.15898395e+00 -4.42980081e-01 -1.79813102e-01 2.42950812e-01
-1.12454939e+00 3.69598091e-01 6.97070211e-02 5.70004165e-01
1.07590342e+00 -1.32518494e+00 -4.80232716e-01 3.40137988e-01
1.80107176e-01 6.32354856e-01 8.16599011e-01 1.41809809e+00
-3.37868154e-01 1.37955397e-01 -7.28802979e-02 -7.04045475e-01
-1.33390296e+00 4.03715819e-01 4.68554199e-01 -6.01652861e-01
-1.09225702e+00 9.46337581e-01 6.89763010e-01 -4.46933061e-01
1.78519581e-02 -6.66168928e-01 -5.64332545e-01 3.90858091e-02
3.38214397e-01 1.31426245e-01 5.73318265e-02 -7.19909966e-01
-6.28012419e-01 1.18773222e+00 -3.50698829e-02 3.28775436e-01
1.35648465e+00 -1.90761447e-01 -2.82792419e-01 8.40888247e-02
1.07042170e+00 3.68866278e-03 -8.82269621e-01 -4.06665921e-01
-2.50091702e-01 -5.65045834e-01 1.58931047e-01 -5.46562195e-01
-1.29682231e+00 7.45927989e-01 8.76643717e-01 -1.46603942e-01
1.30373871e+00 3.60765815e-01 8.70873392e-01 -3.17871958e-01
2.99153775e-01 -7.38624215e-01 -1.47063807e-01 3.57614346e-02
1.08334947e+00 -1.13239551e+00 1.61128432e-01 -6.99702621e-01
-5.74844539e-01 9.96173322e-01 7.53246248e-01 -7.85477310e-02
8.87978435e-01 4.19358432e-01 -9.99630690e-02 -3.80416334e-01
-1.77338421e-02 -7.23578632e-02 7.21055210e-01 4.72584188e-01
4.93264675e-01 1.40535980e-01 -5.43052018e-01 9.12936747e-01
-1.88694820e-01 -1.48993760e-01 -1.20238766e-01 6.89369440e-01
-2.34454662e-01 -8.70485902e-01 -3.23224813e-01 4.43838090e-01
-4.25093234e-01 -2.36527994e-01 2.95967877e-01 7.97690988e-01
2.41636828e-01 6.19561791e-01 1.92651898e-01 -4.05877888e-01
4.60568160e-01 -5.32357037e-01 3.03637117e-01 -4.42274123e-01
-7.38421321e-01 4.80645597e-01 -3.27366114e-01 -8.83426905e-01
-5.15292406e-01 -3.88576806e-01 -1.73160279e+00 1.72879517e-01
-4.52074826e-01 2.37630457e-01 2.75295913e-01 7.52417207e-01
2.05006033e-01 1.09576416e+00 5.63552082e-01 -8.99875939e-01
-9.83330533e-02 -8.31695318e-01 -4.73800242e-01 3.53720874e-01
1.85301840e-01 -8.36631477e-01 -1.00080231e-02 -3.82898033e-01]
|
[14.097367286682129, -2.6507930755615234]
|
ff5ab668-a30a-4f59-8be8-0e49737ab3c5
|
generating-programmatic-referring-expressions
| null | null |
https://proceedings.icml.cc/static/paper_files/icml/2020/1158-Paper.pdf
|
https://proceedings.icml.cc/static/paper_files/icml/2020/1158-Paper.pdf
|
Generating Programmatic Referring Expressions via Program Synthesis
|
Incorporating symbolic reasoning into machine learning algorithms is a promising approach to improve performance on learning tasks that require logical reasoning. We study the problem of generating a programmatic variant of referring expressions that we call referring relational programs. In particular, given a symbolic representation of an image and a target object in that image, the goal is to generate a relational program that uniquely identifies the target object in terms of its attributes and its relations to other objects in the image. We propose a neurosymbolic program synthesis algorithm that combines a policy neural network with enumerative search to generate such relational programs. The policy neural network employs a program interpreter that provides immediate feedback on the consequences of the decisions made by the policy, and also takes into account the uncertainty in the symbolic representation of the image. We evaluate our algorithm on challenging benchmarks based on the CLEVR dataset, and demonstrate that our approach significantly outperforms several baselines.
|
['Aws Albarghouthi', 'Calvin Smith', 'Mayur Naik', 'Jiani Huang', 'Rishabh Singh', 'Osbert Bastani']
| null | null |
https://proceedings.icml.cc/static/paper_files/icml/2020/1158-Paper.pdf
|
https://proceedings.icml.cc/static/paper_files/icml/2020/1158-Paper.pdf
|
icml-2020-1
|
['enumerative-search']
|
['computer-code']
|
[ 5.34332633e-01 7.26406336e-01 -4.54609990e-01 -5.50637841e-01
-6.57504618e-01 -4.92309868e-01 8.00293922e-01 2.83934355e-01
-1.86617672e-01 4.02086169e-01 -1.00626983e-02 -5.24652183e-01
4.77995314e-02 -1.20381975e+00 -1.29471862e+00 -3.34696770e-01
1.40500680e-01 7.19755888e-01 1.17495872e-01 -6.58416152e-02
1.74399406e-01 5.17278969e-01 -1.60458362e+00 6.77057087e-01
5.96082807e-01 8.53187919e-01 -1.81543268e-02 5.96687615e-01
-1.81072250e-01 1.43371785e+00 -5.94610989e-01 -5.02939165e-01
4.25736839e-03 -3.16026777e-01 -1.06856084e+00 -5.80004267e-02
4.61775124e-01 -2.50599474e-01 -1.91043660e-01 1.44068241e+00
-2.17396855e-01 2.64747441e-01 4.62968916e-01 -1.43749249e+00
-5.98215938e-01 1.18329048e+00 -1.13337658e-01 -2.95961171e-01
5.07741988e-01 4.93029326e-01 1.26512372e+00 -3.68077964e-01
7.15634465e-01 1.77262509e+00 6.22501373e-02 7.43025601e-01
-1.71139169e+00 -2.97617555e-01 2.96162546e-01 2.51830816e-02
-1.05688751e+00 -2.98066884e-01 5.82349539e-01 -5.08716285e-01
1.08742726e+00 2.35241249e-01 3.94298732e-01 8.34173977e-01
1.80089369e-01 8.81642938e-01 9.29654062e-01 -6.71477318e-01
5.85918427e-01 2.01599866e-01 2.74227709e-01 1.23667896e+00
-1.66694984e-01 4.02367920e-01 -4.83501494e-01 -5.79982936e-01
5.00921130e-01 -3.67778838e-01 8.34136307e-02 -6.15847647e-01
-1.20823121e+00 8.71107221e-01 6.62960589e-01 -2.20301032e-01
-2.64478922e-01 7.87198365e-01 5.09488344e-01 1.62268102e-01
-1.61028221e-01 8.75162125e-01 -1.70110762e-01 1.27129346e-01
-4.20291632e-01 6.79296076e-01 8.28998685e-01 8.28507066e-01
4.46181267e-01 -6.84897974e-03 -4.16664034e-01 4.23968345e-01
2.80499667e-01 4.62930650e-01 1.85310155e-01 -1.50918460e+00
3.61317515e-01 8.53030145e-01 -1.18271872e-01 -8.68752956e-01
1.86894372e-01 3.08367521e-01 -1.62670523e-01 5.81863165e-01
1.20408401e-01 8.06399062e-03 -6.18574619e-01 1.81594145e+00
2.04976305e-01 2.13687956e-01 2.72643715e-01 6.78892791e-01
7.88732171e-01 7.62205005e-01 2.24197254e-01 1.06286757e-01
1.11672986e+00 -9.35661972e-01 -2.23024338e-01 -3.02509725e-01
7.72222281e-01 2.54596174e-02 1.24054492e+00 3.95596385e-01
-1.10125566e+00 -3.26085508e-01 -1.03191555e+00 2.30157673e-02
-2.04050437e-01 1.77074522e-01 8.86347532e-01 2.23924920e-01
-1.07075119e+00 6.75327599e-01 -8.92678201e-01 -6.21148869e-02
5.14490962e-01 4.89482701e-01 -5.10702841e-02 1.77104160e-01
-7.27625489e-01 7.63242602e-01 8.63526583e-01 -2.92141467e-01
-1.17266560e+00 -4.96944010e-01 -9.80449319e-01 2.68909723e-01
5.88181496e-01 -6.28922164e-01 1.80957580e+00 -1.27415419e+00
-1.44012582e+00 1.19802952e+00 -2.82343656e-01 -8.54865730e-01
3.78365070e-01 5.45637170e-03 -5.13910204e-02 1.60483152e-01
-8.53910074e-02 1.03372002e+00 7.65896022e-01 -1.10655534e+00
-7.30117202e-01 -1.81561664e-01 6.85179651e-01 1.18033193e-01
2.94852912e-01 3.01930368e-01 -3.59448642e-01 -3.03707153e-01
-9.83949155e-02 -1.11632264e+00 -2.87306339e-01 2.56225079e-01
-5.63040793e-01 -4.93619084e-01 7.18021750e-01 -4.83870432e-02
6.85859144e-01 -2.19241929e+00 4.59013015e-01 3.04810047e-01
3.23473960e-02 -5.84277771e-02 -1.40579604e-02 -2.78697699e-01
-1.53320029e-01 2.26851523e-01 -1.92563340e-01 2.34272614e-01
1.62814736e-01 4.66463953e-01 -1.09669685e+00 7.47987255e-02
4.41239774e-01 8.38846385e-01 -1.02271318e+00 -6.07605159e-01
3.79922874e-02 -7.25103766e-02 -7.97482610e-01 2.55062371e-01
-1.44268048e+00 2.03021429e-02 -7.13784814e-01 4.32205766e-01
4.29417156e-02 -2.58783191e-01 2.71737456e-01 5.07845543e-02
2.93264747e-01 4.23319161e-01 -8.66034329e-01 1.33999157e+00
-3.39782000e-01 5.33205211e-01 -3.19418550e-01 -7.21301258e-01
8.33180845e-01 2.70831808e-02 -4.21241559e-02 -2.42654562e-01
6.83943927e-02 -2.73168851e-02 -1.58936650e-01 -5.26586473e-01
2.59223431e-01 -3.72692989e-03 -2.85916895e-01 6.68990135e-01
-1.40499264e-01 -6.05976999e-01 9.85503048e-02 4.16025698e-01
1.19921839e+00 7.07354188e-01 3.35043550e-01 -1.03683043e-02
7.07394540e-01 4.83678281e-01 3.01483512e-01 1.04336143e+00
1.21457532e-01 8.56665596e-02 1.29741430e+00 -6.05706096e-01
-7.39787340e-01 -1.03269398e+00 2.25605220e-01 1.23985159e+00
-2.34057363e-02 -3.65958691e-01 -7.37208843e-01 -8.54410529e-01
1.73328638e-01 1.23279321e+00 -6.00545347e-01 -3.90623629e-01
-6.53199971e-01 -1.74388677e-01 6.28628373e-01 5.18451214e-01
2.32587755e-01 -1.48357570e+00 -1.04775333e+00 -2.10550621e-01
7.65885562e-02 -1.08434141e+00 -5.07618822e-02 3.13921481e-01
-9.26288366e-01 -1.20001316e+00 3.46890718e-01 -7.62821972e-01
1.08474541e+00 -4.51388896e-01 1.22767067e+00 1.42526716e-01
-4.80888933e-02 3.01956207e-01 1.01966813e-01 -2.26119712e-01
-1.01169324e+00 -7.00379461e-02 -3.75047743e-01 -3.67601067e-01
1.55452818e-01 6.11760770e-04 2.96483457e-01 -1.24072239e-01
-7.49302506e-01 4.20199603e-01 1.92504153e-01 7.96043158e-01
7.84058213e-01 1.63577229e-01 5.46924956e-02 -1.31754851e+00
5.90796888e-01 -2.56911844e-01 -1.34909594e+00 4.77375567e-01
-2.45993376e-01 7.60016918e-01 6.92257702e-01 -3.05098563e-01
-1.22468793e+00 5.26432812e-01 3.91019583e-01 -5.26822805e-01
-1.17676526e-01 5.08353353e-01 -2.03043059e-01 8.10901523e-02
1.08256412e+00 2.50743721e-02 7.39267766e-02 1.33301139e-01
5.95184267e-01 2.39078164e-01 1.05970979e+00 -1.63507783e+00
5.71674168e-01 2.64450788e-01 3.15622360e-01 -1.51580513e-01
-8.28436911e-01 3.63635540e-01 -3.43013287e-01 1.85406040e-02
6.61869526e-01 -7.51860559e-01 -1.10392439e+00 -2.99389325e-02
-1.43702328e+00 -7.21877277e-01 -4.21366632e-01 -3.94802727e-02
-1.11502326e+00 -4.14566517e-01 -3.71505469e-01 -6.41995311e-01
-7.28858784e-02 -1.69975030e+00 9.83598232e-01 1.04177542e-01
-6.07117891e-01 -6.59180284e-01 -3.10147464e-01 9.01066661e-02
4.40177582e-02 4.92976695e-01 1.72674799e+00 -7.00839102e-01
-1.13258791e+00 -1.30415842e-01 -1.68730661e-01 1.30097315e-01
-2.85284430e-01 3.97949547e-01 -9.03126955e-01 3.91295880e-01
-2.40225017e-01 -7.10707068e-01 4.36454386e-01 4.86238673e-02
1.67445862e+00 -6.63283765e-01 -4.55219239e-01 5.79471707e-01
1.25211561e+00 3.70394021e-01 3.66494536e-01 3.92707050e-01
5.15586853e-01 5.84519863e-01 6.88746810e-01 1.99501842e-01
1.97879255e-01 6.17250621e-01 6.11366987e-01 4.79887396e-01
3.33151482e-02 -4.75762308e-01 5.05419075e-01 -4.34277654e-01
3.68446052e-01 3.08721542e-01 -1.06863570e+00 3.62654269e-01
-2.07077622e+00 -9.44939911e-01 3.24282289e-01 1.95373607e+00
1.28974843e+00 3.00148070e-01 -1.38330489e-01 -3.95639181e-01
5.83499134e-01 6.18369691e-02 -9.08874094e-01 -8.34158123e-01
1.29546136e-01 4.54027474e-01 2.93227255e-01 6.59066498e-01
-9.70057309e-01 1.28752029e+00 6.66114855e+00 3.90048712e-01
-1.14998388e+00 -2.81765133e-01 7.71457076e-01 -1.59978062e-01
-4.17474717e-01 2.30269238e-01 -8.86870205e-01 -4.05036211e-02
8.62645507e-01 -6.49686873e-01 1.00155520e+00 1.41824150e+00
-2.57641226e-01 -2.50271082e-01 -1.95220363e+00 4.57602590e-01
2.63504405e-02 -1.59373772e+00 2.17311352e-01 -4.13098037e-01
6.21651649e-01 -5.70536032e-02 2.60253698e-01 4.42128837e-01
1.03720820e+00 -1.20318687e+00 1.06534731e+00 4.99055237e-01
6.12897635e-01 -8.77570570e-01 1.36251375e-01 4.26355153e-01
-4.47093934e-01 -2.89021462e-01 -2.69026086e-02 4.76013012e-02
-5.24626315e-01 1.17453367e-01 -1.18086231e+00 -1.41216308e-01
3.73603195e-01 4.63138849e-01 -6.41974926e-01 4.52533007e-01
-9.46774125e-01 2.91568279e-01 -1.68400943e-01 -9.29472074e-02
1.41107306e-01 7.83457607e-02 3.73299569e-01 9.53266084e-01
-2.42495928e-02 1.84341818e-01 1.50358438e-01 1.78696942e+00
-3.15820247e-01 -3.96775484e-01 -8.74276459e-01 -2.49991387e-01
4.99589026e-01 1.01410079e+00 -4.25654233e-01 -6.34187341e-01
-1.81320921e-01 4.02366757e-01 5.43038964e-01 3.09762329e-01
-9.16436911e-01 -1.27797335e-01 4.78256941e-01 -1.62980765e-01
1.08407155e-01 1.45612776e-01 -4.19535369e-01 -9.42317188e-01
-1.20464936e-01 -1.43239963e+00 5.12955129e-01 -1.33201396e+00
-5.63891888e-01 4.57429260e-01 3.70741516e-01 -4.93349910e-01
-6.83379591e-01 -7.64010310e-01 -5.39727926e-01 9.13952172e-01
-1.14031863e+00 -8.98710549e-01 1.37824053e-02 4.10130590e-01
2.79528230e-01 -3.81498426e-01 1.12203991e+00 -6.03918552e-01
-4.76401061e-01 3.15370739e-01 -6.31140947e-01 1.25117600e-01
2.75938749e-01 -1.18756461e+00 3.32555860e-01 7.74599731e-01
4.88751493e-02 9.12107050e-01 7.37659216e-01 -5.39540708e-01
-1.67569840e+00 -1.37177372e+00 6.30262136e-01 -5.33477366e-01
8.21611822e-01 1.01925489e-02 -6.95905805e-01 1.41081047e+00
5.56225739e-02 2.18376398e-01 1.37509346e-01 -4.42416221e-02
-7.06618249e-01 1.38794910e-02 -1.22602475e+00 9.42823291e-01
8.24227095e-01 -7.18754351e-01 -8.37381005e-01 4.75553751e-01
9.66944516e-01 -1.03786397e+00 -4.62695003e-01 1.03700265e-01
1.64053127e-01 -6.43385887e-01 9.13267314e-01 -1.05068302e+00
1.04378247e+00 -4.21336800e-01 -4.07462269e-01 -1.07561553e+00
2.74803005e-02 -4.66113985e-01 -2.77784884e-01 8.42001379e-01
3.49236608e-01 -4.73913610e-01 8.95220339e-01 1.26846719e+00
4.51190248e-02 -6.64569855e-01 -5.72172642e-01 -3.94463152e-01
2.29193643e-02 -4.05898780e-01 7.92319179e-01 5.91775179e-01
2.80383259e-01 3.11326176e-01 2.68875301e-01 3.78906429e-01
5.60326576e-01 6.18929565e-01 7.23958611e-01 -8.97111237e-01
-6.19083941e-01 -6.43409252e-01 -1.80504739e-01 -6.05509937e-01
1.17889655e+00 -1.35224104e+00 3.11309457e-01 -9.63409781e-01
1.91209421e-01 -5.70326805e-01 1.75155848e-01 1.26423848e+00
1.39824348e-02 -5.53806841e-01 2.41248995e-01 4.92592938e-02
-5.87346137e-01 1.26116171e-01 9.00561810e-01 -4.36617970e-01
-1.47534728e-01 -1.15007818e-01 -8.84179294e-01 1.06037021e+00
5.82864285e-01 -5.84795713e-01 -4.56562489e-01 -5.82905710e-01
5.15352726e-01 5.13277650e-01 7.62033999e-01 -7.30056822e-01
4.85238343e-01 -7.05362439e-01 3.43128443e-02 -1.82951853e-01
1.51301011e-01 -8.38020563e-01 1.02675654e-01 6.50153816e-01
-1.23508811e+00 -4.71423706e-03 2.55481452e-01 1.86467648e-01
-1.97965533e-01 -4.77912277e-01 8.34619939e-01 -3.18842232e-01
-9.15229499e-01 -6.40360266e-02 -1.69611290e-01 4.05237917e-03
1.18074775e+00 4.58752543e-01 -5.82990706e-01 -8.29951689e-02
-6.97517395e-01 4.02502507e-01 5.57540238e-01 3.13212961e-01
6.27060294e-01 -1.28455114e+00 -1.67492449e-01 2.34576941e-01
2.70798594e-01 2.63721019e-01 -6.25410140e-01 2.76619732e-01
-6.32243395e-01 3.61612171e-01 -1.69067726e-01 -5.19874573e-01
-1.41335809e+00 6.74214482e-01 6.90489888e-01 -1.46042839e-01
-4.80430216e-01 8.00180972e-01 1.24313630e-01 -6.90817058e-01
5.69144905e-01 -6.77166164e-01 7.27140084e-02 -5.66305876e-01
3.97859931e-01 -2.01852277e-01 -2.00169742e-01 -3.45352471e-01
-2.35593125e-01 1.49185036e-03 -3.68083827e-02 -2.78272539e-01
1.17242742e+00 8.76685262e-01 -6.35592163e-01 3.82660389e-01
7.80111372e-01 -3.19802016e-01 -1.16267323e+00 -6.50787950e-01
1.58580497e-01 -4.56720501e-01 -1.99628286e-02 -8.22039366e-01
-9.36980367e-01 5.80312431e-01 -6.54749274e-02 -8.22909996e-02
9.00639892e-01 2.47669831e-01 9.34788305e-03 1.05486929e+00
4.35558468e-01 -6.65284991e-01 2.91230619e-01 7.33244836e-01
9.88885999e-01 -1.00722384e+00 8.88587981e-02 -4.22950774e-01
-8.27368200e-01 1.21323395e+00 8.45393598e-01 -3.21000785e-01
4.99167480e-02 5.56001186e-01 -3.54969203e-01 -2.61247396e-01
-1.26474845e+00 2.29510561e-01 6.03858195e-03 4.87204313e-01
2.30728999e-01 2.29043588e-01 3.47408354e-01 3.69230807e-01
-3.38350534e-01 4.48446393e-01 4.38752741e-01 1.03796220e+00
-2.98638701e-01 -1.08923209e+00 -5.81788182e-01 3.60601366e-01
-2.24887103e-01 -1.28162131e-02 -3.45816404e-01 5.20267546e-01
1.76459234e-02 4.04424101e-01 8.86789560e-02 -1.64857328e-01
3.12656820e-01 2.08519146e-01 7.61260808e-01 -1.02582705e+00
-5.79393208e-01 -4.58224624e-01 2.34910458e-01 -7.95555949e-01
-1.87771976e-01 -7.65670538e-01 -1.96654248e+00 -1.77438080e-01
2.27517113e-01 -1.26330584e-01 5.56976855e-01 8.03463042e-01
3.25366437e-01 5.77379107e-01 2.21784726e-01 -3.99367869e-01
-9.90848720e-01 -3.27401944e-02 -4.93788049e-02 3.36501956e-01
3.13207120e-01 -2.88873076e-01 -2.40191072e-02 4.95538622e-01]
|
[8.666534423828125, 7.148620128631592]
|
f319a97e-5896-4614-9987-16c9a5c505a4
|
nmbr9-as-a-constraint-programming-challenge
|
2001.04238
| null |
https://arxiv.org/abs/2001.04238v1
|
https://arxiv.org/pdf/2001.04238v1.pdf
|
Nmbr9 as a Constraint Programming Challenge
|
Modern board games are a rich source of interesting and new challenges for combinatorial problems. The game Nmbr9 is a solitaire style puzzle game using polyominoes. The rules of the game are simple to explain, but modelling the game effectively using constraint programming is hard. This abstract presents the game, contributes new generalized variants of the game suitable for benchmarking and testing, and describes a model for the presented variants. The question of the top possible score in the standard game is an open challenge.
|
['Mikael Zayenz Lagerkvist']
|
2020-01-13
| null | null | null | null |
['board-games', 'solitaire']
|
['playing-games', 'playing-games']
|
[-2.80595690e-01 4.59687918e-01 3.16117145e-02 1.11722179e-01
-8.61595646e-02 -9.38587248e-01 2.11366966e-01 -1.33496478e-01
-3.14285576e-01 1.19472671e+00 -4.66315180e-01 -3.77988786e-01
-6.88226819e-01 -1.05338323e+00 -3.55145961e-01 -5.69820881e-01
-2.90086448e-01 1.33055353e+00 9.65617597e-01 -1.22589016e+00
2.28974506e-01 1.81449592e-01 -1.56879473e+00 4.80443776e-01
3.74807686e-01 7.34713078e-01 2.89417386e-01 1.06608272e+00
-2.09967315e-01 4.46883738e-01 -9.74658728e-01 -6.36528492e-01
7.38404810e-01 -5.00135481e-01 -9.21613634e-01 -2.73601532e-01
-3.57563555e-01 4.06211764e-01 5.88727668e-02 8.14609766e-01
4.14252698e-01 1.61889210e-01 2.96026617e-01 -1.61956239e+00
4.91995782e-01 5.68389714e-01 -6.01528287e-02 1.81291059e-01
8.40347111e-01 -1.17067561e-01 1.28434360e+00 -1.57870233e-01
1.03507388e+00 6.17054164e-01 4.30607229e-01 6.33810282e-01
-9.58341539e-01 -3.92782688e-01 -1.69538587e-01 3.30861628e-01
-1.51765680e+00 3.07348251e-01 2.30829075e-01 1.55928776e-01
1.30809474e+00 1.20226526e+00 1.26639450e+00 5.78665435e-01
2.80642211e-01 3.08597803e-01 1.36080980e+00 -4.98332888e-01
5.29080391e-01 -7.83564597e-02 -1.33528292e-01 3.43135446e-01
5.90001285e-01 1.81114838e-01 -4.13715452e-01 -2.41743222e-01
9.07533824e-01 -6.27741814e-01 1.42620429e-01 -5.99898636e-01
-6.75727010e-01 9.30339336e-01 -1.14806555e-01 1.75272480e-01
5.40400483e-02 1.73781931e-01 3.38849455e-01 4.75629240e-01
-2.02211156e-01 1.12637699e+00 -8.90356123e-01 -8.65092099e-01
-6.80897236e-01 1.15263617e+00 1.56259644e+00 1.26481748e+00
2.19712093e-01 -2.52782345e-01 1.95080295e-01 5.37279487e-01
1.90980613e-01 -1.08164556e-01 -1.82196856e-01 -9.27613258e-01
5.89395523e-01 5.97756863e-01 1.08224101e-01 -6.03912175e-01
-1.06062925e+00 -2.43877977e-01 -3.30763549e-01 9.34939265e-01
8.11936677e-01 -2.33087376e-01 -3.35384071e-01 9.97432947e-01
1.59285307e-01 -8.08511078e-02 -1.05442494e-01 8.29598010e-01
9.41462398e-01 6.77302480e-01 -4.07354176e-01 -2.72565424e-01
1.27053714e+00 -1.08701432e+00 -4.29972619e-01 -3.60662341e-02
3.87984335e-01 -6.16813421e-01 5.57380319e-01 1.10973763e+00
-1.83576512e+00 1.24970585e-01 -1.20348132e+00 3.36792469e-01
-6.14417493e-01 -6.52865410e-01 1.03644228e+00 1.31178796e+00
-1.23970902e+00 4.33094621e-01 -3.02402347e-01 -1.38249189e-01
-2.46968880e-01 1.09339690e+00 -3.50381613e-01 3.55360322e-02
-1.07064271e+00 1.20277607e+00 1.04949665e+00 -1.12456247e-01
-9.91030410e-02 -3.27038199e-01 -6.84039116e-01 2.96703517e-01
1.03714168e+00 -4.76708949e-01 1.21835864e+00 -2.66247720e-01
-2.00566173e+00 1.00522149e+00 5.98639369e-01 -8.02667364e-02
7.63774514e-01 7.74956703e-01 -1.80783659e-01 -2.78425902e-01
-3.52191031e-01 3.00779670e-01 -1.27620157e-02 -8.78579021e-01
-9.03471529e-01 2.93241322e-01 1.07045257e+00 4.41146761e-01
6.61391318e-01 3.81075740e-01 -6.96767628e-01 -3.52195174e-01
1.29491806e-01 -8.95472646e-01 -8.10471356e-01 -6.41671419e-01
-2.82630734e-02 -6.91040903e-02 -3.66948605e-01 6.55904179e-03
1.41473866e+00 -1.42439902e+00 4.05716360e-01 7.17971504e-01
2.31069356e-01 1.26404902e-02 -1.37758777e-01 9.38947618e-01
-1.92335710e-01 3.03564250e-01 1.97927237e-01 4.20685947e-01
6.70477152e-01 7.41108596e-01 3.77714634e-01 -2.20739786e-02
-2.17923328e-01 8.46862197e-01 -6.44553006e-01 -1.61217675e-01
2.06257682e-02 -5.35700083e-01 -9.83015716e-01 -1.71605751e-01
-7.02588797e-01 -4.15300652e-02 -1.26224905e-01 4.02023077e-01
9.00700152e-01 4.18235779e-01 4.92161334e-01 7.07675755e-01
-4.01608855e-01 3.68712246e-01 -1.80540252e+00 1.53466678e+00
-1.73768461e-01 2.16731057e-01 2.93429583e-01 -7.79326797e-01
8.93584430e-01 1.44354939e-01 3.90812218e-01 -6.95397317e-01
2.96021730e-01 2.86729485e-01 4.85275120e-01 -3.00479800e-01
7.31392682e-01 -2.42804751e-01 -5.38016319e-01 4.79335040e-01
-8.07506964e-02 -1.03211224e+00 1.13316715e+00 -1.52050659e-01
1.51809478e+00 2.41182279e-02 6.07921243e-01 -5.57811141e-01
4.08725411e-01 4.77423608e-01 6.45438254e-01 8.47138941e-01
1.59189671e-01 6.78477883e-01 1.23968339e+00 -8.29934061e-01
-1.03422320e+00 -8.86607885e-01 1.56817168e-01 9.49893892e-01
3.53222877e-01 -1.18611264e+00 -4.61733699e-01 -3.25840592e-01
-3.72488290e-01 3.09134513e-01 -5.34203410e-01 4.05676961e-01
-6.13381088e-01 -9.95137513e-01 3.71726960e-01 1.23754404e-01
-1.78145897e-03 -1.13690841e+00 -6.84527993e-01 3.81071866e-01
-2.28964955e-01 -1.06192350e+00 1.74185380e-01 4.85255808e-01
-5.98808944e-01 -1.48994255e+00 -8.64539966e-02 -7.60973215e-01
1.73450693e-01 2.13166829e-02 1.64463353e+00 5.86517692e-01
-4.85900998e-01 -4.51802742e-03 -7.42171586e-01 -7.29326129e-01
-1.94098532e-01 2.47483566e-01 -2.48977616e-01 -1.07931149e+00
5.60330264e-02 -4.76017833e-01 9.04804841e-03 7.05942571e-01
-7.63143480e-01 1.38364909e-02 1.08887225e-01 8.80502462e-01
4.21495467e-01 6.61006391e-01 -2.23830894e-01 -9.45319057e-01
9.16645586e-01 -2.75181174e-01 -1.13640261e+00 6.12403974e-02
-1.34110376e-01 -5.08140214e-02 5.10940731e-01 -7.00930208e-02
-3.65668118e-01 -1.15100935e-01 -4.06675905e-01 7.01307714e-01
2.08063759e-02 6.36539698e-01 -1.88004985e-01 -5.75099647e-01
5.98581493e-01 -2.31460944e-01 -3.52858216e-01 -1.03376769e-01
3.86109203e-02 -1.00148089e-01 3.74129862e-02 -1.15049136e+00
7.85367489e-01 4.56690043e-02 5.82370579e-01 -6.09906971e-01
8.82548690e-02 -5.77311695e-01 -3.84449840e-01 -3.16678911e-01
4.59604502e-01 -2.46849954e-01 -1.46368527e+00 1.22635908e-01
-1.18409121e+00 -6.62354708e-01 -6.91707850e-01 1.30737096e-01
-1.09129918e+00 -3.86741459e-02 -3.59891057e-01 -8.30976725e-01
3.73941809e-01 -1.10402179e+00 4.70820397e-01 4.14940178e-01
-2.13414460e-01 -7.74878502e-01 5.92802405e-01 4.15555507e-01
3.26405317e-01 4.58341330e-01 6.15611613e-01 -6.27139688e-01
-5.64946115e-01 -4.45359737e-01 5.01488373e-02 -3.26862305e-01
-4.56379771e-01 9.63005573e-02 -2.19985068e-01 -2.00274326e-02
-6.60784021e-02 -1.12680607e-01 -3.20879295e-02 3.14602047e-01
7.78981805e-01 8.09779465e-02 -7.64788361e-03 4.84818518e-01
1.73528397e+00 3.65225166e-01 1.36321628e+00 1.00664747e+00
-3.35045338e-01 4.53539521e-01 7.27111280e-01 6.15514874e-01
3.40766221e-01 1.26497102e+00 7.55039692e-01 2.47294858e-01
4.51048911e-01 4.36222076e-01 1.09057287e-02 7.88189530e-01
-1.05646753e+00 -2.74976581e-01 -1.03069246e+00 2.59683505e-02
-1.92934263e+00 -1.00805593e+00 -6.53275907e-01 1.77406156e+00
4.41413641e-01 7.46955097e-01 5.72005749e-01 5.25913596e-01
3.52254301e-01 -8.00623447e-02 4.36044961e-01 -1.54600704e+00
-3.43478262e-01 1.32607508e+00 6.32455587e-01 6.56161129e-01
-6.66672647e-01 9.90842342e-01 7.81885242e+00 1.21766770e+00
-2.82314867e-01 1.27633587e-01 -2.06232861e-01 -4.21354532e-01
1.37666479e-01 2.10570768e-01 -6.14529312e-01 2.07025692e-01
6.71749115e-01 -4.34540480e-01 9.02717054e-01 6.51268601e-01
8.19557533e-02 -7.14148402e-01 -6.93062127e-01 7.90652871e-01
-2.95070428e-02 -1.67902553e+00 -3.60750377e-01 3.19322169e-01
8.15423429e-01 -2.75442809e-01 -2.03165472e-01 5.81662953e-01
5.56494296e-01 -1.33176410e+00 7.99806714e-01 1.54396653e-01
4.52438653e-01 -9.52665687e-01 1.05605638e+00 3.70866239e-01
-1.21086001e+00 -1.11462452e-01 -7.87409306e-01 -1.09548879e+00
5.07119261e-02 -3.15053970e-01 -3.24036926e-01 9.86762702e-01
6.18891060e-01 -2.49917045e-01 -3.83116752e-01 2.00629973e+00
-3.85136813e-01 1.44139603e-01 -7.79061556e-01 -5.22850692e-01
5.43800950e-01 -5.55784523e-01 6.73599720e-01 9.14651632e-01
1.66758910e-01 6.77808285e-01 5.28936945e-02 4.51575547e-01
5.37655056e-01 2.40953371e-01 -2.74791807e-01 3.39710921e-01
-8.34570173e-03 1.06320071e+00 -1.26127553e+00 4.06796485e-01
-1.24037892e-01 5.27151644e-01 8.40845555e-02 -1.09434672e-01
-9.58538711e-01 -4.01297778e-01 6.81855321e-01 1.34343222e-01
3.26561511e-01 -3.07259500e-01 -6.40359044e-01 -8.94857526e-01
1.02496654e-01 -1.06180608e+00 5.78450859e-01 -8.74878228e-01
-7.75200188e-01 5.92632115e-01 2.60554403e-01 -1.15149724e+00
-2.41032988e-01 -1.22016478e+00 -9.72049475e-01 6.96114063e-01
-9.42032158e-01 -7.03609526e-01 -1.79963201e-01 5.00210762e-01
3.81973982e-01 -2.45745718e-01 1.24048543e+00 -6.98569641e-02
-2.07811922e-01 5.62744319e-01 9.56042260e-02 -5.41622698e-01
-4.62764949e-02 -1.65493655e+00 4.81560767e-01 3.54629517e-01
-1.23459343e-02 -3.95775437e-02 1.06982577e+00 -2.74900764e-01
-1.34959352e+00 -2.45868683e-01 7.13667095e-01 -5.39390564e-01
7.90765285e-01 -7.28043079e-01 3.53145450e-02 1.59294114e-01
9.36377645e-02 -4.84858841e-01 7.07560062e-01 3.41841817e-01
2.41705365e-02 2.48519108e-01 -1.15581989e+00 7.74392545e-01
1.12889707e+00 2.75474459e-01 -4.08152640e-01 4.44307148e-01
1.87672690e-01 -1.17344773e+00 -6.20676935e-01 -7.03756735e-02
7.30949342e-01 -1.39722192e+00 8.28046620e-01 -7.31933951e-01
3.26424092e-01 -1.88185021e-01 -3.16086896e-02 -1.34796536e+00
-4.72671658e-01 -1.16739702e+00 3.47228527e-01 4.38607395e-01
7.80805171e-01 -4.33927953e-01 1.49466276e+00 8.68753076e-01
-7.66100064e-02 -7.12017715e-01 -1.32699740e+00 -1.29036367e+00
3.25805068e-01 -7.76790619e-01 8.04361105e-01 6.34096563e-01
1.02726090e+00 -8.94398317e-02 -1.73794240e-01 -8.38514492e-02
1.68526575e-01 -2.70712525e-01 1.02670336e+00 -1.46724474e+00
-8.65973234e-01 -9.43001986e-01 -1.10382366e+00 -5.86451530e-01
-4.12945062e-01 -8.40176463e-01 -2.66952515e-01 -1.73066866e+00
-1.62785619e-01 -5.53251028e-01 2.58529067e-01 8.41866881e-02
3.81276041e-01 6.14260912e-01 7.16512799e-01 -5.39673150e-01
-9.86049294e-01 -2.22606778e-01 1.20220661e+00 -6.11835271e-02
-2.27156475e-01 4.29524809e-01 -4.80780810e-01 5.17097533e-01
1.03140938e+00 -5.85116804e-01 -1.47851795e-01 1.37268472e-03
1.49198103e+00 2.02700347e-01 -2.83487350e-01 -9.71317828e-01
4.55137789e-01 -7.70793676e-01 -3.89331520e-01 -2.79789001e-01
5.91548383e-01 -9.81216371e-01 6.63687289e-01 6.22735143e-01
3.10982227e-01 4.63420123e-01 5.88778079e-01 -4.43325518e-03
-1.20976768e-01 -1.00347650e+00 2.30845138e-01 -4.19865638e-01
-5.64745963e-01 -3.51790451e-02 -8.22411478e-01 3.56304318e-01
1.56359994e+00 -8.24466050e-01 -3.43463510e-01 -4.58967358e-01
-1.27363646e+00 4.31010604e-01 5.01852214e-01 -3.78162339e-02
3.02414626e-01 -1.05833149e+00 -8.11282814e-01 -1.70091182e-01
-7.18426108e-02 4.27039713e-02 1.29303575e-01 4.08984154e-01
-1.67830431e+00 5.21120012e-01 -9.49133694e-01 1.35023579e-01
-1.39813757e+00 4.50987756e-01 4.57202286e-01 -1.02732182e+00
-3.41563880e-01 8.93642783e-01 -3.12083364e-01 -5.16105056e-01
7.20335320e-02 -2.87817478e-01 -3.95116836e-01 -3.15926611e-01
3.48812938e-01 5.50361037e-01 3.31394583e-01 -1.17978014e-01
-4.33913052e-01 3.67556483e-01 4.22055602e-01 -2.92564631e-01
1.67078817e+00 2.76399076e-01 -5.71530044e-01 -1.37220714e-02
5.52832112e-02 4.39907908e-02 -3.17228705e-01 5.73377252e-01
2.11255229e-03 -4.19388354e-01 -5.84762812e-01 -1.08690882e+00
-8.12597454e-01 3.11622769e-01 -1.90489501e-01 8.84561300e-01
9.91074681e-01 -3.26304048e-01 2.42467806e-01 4.40053225e-01
1.18697023e+00 -1.38176739e+00 -3.71171057e-01 1.26115882e+00
1.06124902e+00 -4.74258304e-01 2.24862844e-01 -1.06449425e+00
-5.34451306e-01 1.33589292e+00 6.78050041e-01 -6.13814712e-01
7.43549526e-01 8.40928555e-01 -2.07886383e-01 -3.86904776e-01
-1.02901852e+00 -5.17972946e-01 -1.76642150e-01 1.09751725e+00
-1.37773097e-01 2.57491291e-01 -1.32476687e+00 1.26944625e+00
-1.03879821e+00 -2.15109438e-01 1.36648619e+00 1.08527470e+00
-1.47813335e-01 -1.71284127e+00 -7.65311241e-01 3.11860353e-01
-3.12694550e-01 -4.22409885e-02 -8.78808022e-01 1.20451283e+00
4.71864015e-01 1.03675389e+00 -1.61227033e-01 -5.10060787e-01
6.87893510e-01 -2.54197836e-01 1.08444595e+00 -8.38172019e-01
-1.16665280e+00 -3.52308601e-01 6.50480986e-01 -5.11201799e-01
-6.46866532e-03 -3.98999602e-01 -7.89753318e-01 -8.23362470e-01
-6.10157013e-01 7.47195959e-01 7.79035509e-01 8.36513340e-01
-4.05469030e-01 5.27874589e-01 3.48076254e-01 -9.47554767e-01
-8.30199495e-02 -3.44066501e-01 -1.10447621e+00 -3.95852417e-01
-8.22007596e-01 -8.03954184e-01 -1.40421884e-02 -6.76250517e-01]
|
[3.4511685371398926, 1.4808011054992676]
|
7c481799-e564-465c-ac3d-e9c0a856b134
|
hmm-based-writer-identification-in-music
|
1707.06828
| null |
http://arxiv.org/abs/1707.06828v2
|
http://arxiv.org/pdf/1707.06828v2.pdf
|
HMM-based Writer Identification in Music Score Documents without Staff-Line Removal
|
Writer identification from musical score documents is a challenging task due
to its inherent problem of overlapping of musical symbols with staff lines.
Most of the existing works in the literature of writer identification in
musical score documents were performed after a preprocessing stage of staff
lines removal. In this paper we propose a novel writer identification framework
in musical documents without removing staff lines from documents. In our
approach, Hidden Markov Model has been used to model the writing style of the
writers without removing staff lines. The sliding window features are extracted
from musical score lines and they are used to build writer specific HMM models.
Given a query musical sheet, writer specific confidence for each musical line
is returned by each writer specific model using a loglikelihood score. Next, a
loglikelihood score in page level is computed by weighted combination of these
scores from the corresponding line images of the page. A novel Factor Analysis
based feature selection technique is applied in sliding window features to
reduce the noise appearing from staff lines which proves efficiency in writer
identification performance.In our framework we have also proposed a novel score
line detection approach in musical sheet using HMM. The experiment has been
performed in CVC-MUSCIMA dataset and the results obtained that the proposed
approach is efficient for score line detection and writer identification
without removing staff lines. To get the idea of computation time of our
method, detail analysis of execution time is also provided.
|
['Ayan Kumar Bhunia', 'Umapada Pal', 'Partha Pratim Roy']
|
2017-07-21
| null | null | null | null |
['line-detection']
|
['computer-vision']
|
[ 4.11022753e-01 -6.23526573e-01 2.46412098e-01 -3.28211263e-02
-6.48966789e-01 -7.76588380e-01 3.93758744e-01 3.97976860e-02
-4.08129722e-01 2.88708448e-01 1.05414316e-01 2.72340834e-01
-5.11917174e-01 -4.09870744e-01 -9.21837091e-02 -5.01087844e-01
3.23918939e-01 5.05530894e-01 5.92854977e-01 -1.28552377e-01
8.17328990e-01 6.59826994e-01 -1.47167706e+00 2.76997358e-01
1.86732858e-01 4.02886033e-01 4.70180988e-01 1.36097538e+00
-2.00510368e-01 4.11643624e-01 -9.39807594e-01 -4.95367907e-02
3.70588332e-01 -7.76384056e-01 -4.24166352e-01 4.51395452e-01
1.42025292e-01 -2.28574336e-01 1.80007800e-01 8.89748216e-01
4.97771382e-01 3.68746787e-01 1.11281204e+00 -5.39134324e-01
9.66861621e-02 6.06739283e-01 -9.45143938e-01 1.67528659e-01
5.57701170e-01 -5.03644228e-01 8.49963665e-01 -9.32407200e-01
4.83333796e-01 9.27553475e-01 7.50369549e-01 1.07137419e-01
-1.21219456e+00 -6.80653334e-01 -6.05467975e-01 1.90859094e-01
-1.54904890e+00 -3.53974581e-01 8.20335746e-01 -6.22713149e-01
8.79341602e-01 5.82890153e-01 4.55829084e-01 4.01029915e-01
2.12278783e-01 5.10114133e-01 1.19975567e+00 -1.13824153e+00
-1.16120197e-01 2.12044328e-01 5.85190237e-01 7.50067890e-01
-5.03744148e-02 -2.51626343e-01 -9.17402446e-01 -1.90545708e-01
8.45101297e-01 -2.80921519e-01 1.91600844e-01 7.46082738e-02
-8.71888757e-01 6.24632537e-01 -7.17751563e-01 5.84133625e-01
-6.06086217e-02 -2.57689059e-01 5.37301958e-01 2.36094102e-01
-3.56342673e-01 1.30383968e-01 -1.22987553e-01 -4.47856188e-01
-1.57566655e+00 3.11955065e-01 9.55496371e-01 8.64272058e-01
2.03372419e-01 -1.20594010e-01 -6.00592680e-02 9.69014168e-01
3.88212472e-01 4.21846241e-01 7.72375703e-01 -4.76249605e-01
4.03980166e-01 4.04958040e-01 2.78048456e-01 -1.00009239e+00
-2.59408355e-01 -4.75758910e-01 -5.31826496e-01 4.81047809e-01
4.11745548e-01 2.90263649e-02 -6.46374404e-01 1.01381528e+00
1.65573880e-02 4.29298282e-02 8.07559341e-02 7.08281875e-01
3.49637479e-01 7.22329557e-01 -4.67359573e-01 -3.96302730e-01
1.77491963e+00 -7.76471257e-01 -9.66378391e-01 3.64566863e-01
2.16309756e-01 -1.54835045e+00 8.53116155e-01 1.05311894e+00
-9.19012725e-01 -8.37561369e-01 -1.43363845e+00 2.13497087e-01
3.62442434e-03 9.00179386e-01 9.35720727e-02 9.93239164e-01
-3.83508265e-01 5.54213047e-01 -7.42282212e-01 -5.15846729e-01
-3.47999603e-01 5.04311621e-01 -1.61654711e-01 8.82477820e-01
-5.02990365e-01 6.47555470e-01 1.81454852e-01 1.73015356e-01
-4.29676622e-01 1.74004957e-01 -1.53482571e-01 1.58334196e-01
1.80103153e-01 -1.57068118e-01 1.31529009e+00 -1.03009415e+00
-1.73220944e+00 6.45975113e-01 -3.61140668e-01 -2.26626962e-01
7.36529350e-01 -1.56757846e-01 -6.78297520e-01 6.67106186e-04
1.07367106e-01 -4.07049298e-01 1.24514699e+00 -8.35564494e-01
-8.46314847e-01 -4.18913513e-01 -9.53679144e-01 2.41214037e-01
-1.91435859e-01 4.95953321e-01 -6.51658297e-01 -8.37337375e-01
4.44194466e-01 -8.97238553e-01 2.40361452e-01 -9.83022749e-01
-4.36046392e-01 -6.58396930e-02 6.91040635e-01 -1.10278547e+00
1.66611028e+00 -2.19347048e+00 -1.46645263e-01 8.54141533e-01
-2.26109654e-01 7.01027503e-03 4.57797796e-01 6.13475323e-01
1.60401061e-01 -4.01492059e-01 -6.01719357e-02 -2.70837784e-01
-8.15910846e-02 -2.23963574e-01 -3.13372165e-01 3.30979794e-01
-2.65359014e-01 2.62118392e-02 -1.17328890e-01 -6.85259104e-01
1.05697155e-01 3.03860813e-01 -9.93013605e-02 5.07405773e-02
3.52785498e-01 2.01189026e-01 -2.95575321e-01 3.29554021e-01
6.72373652e-01 5.36789894e-01 1.66889235e-01 -1.26987770e-01
-5.17561853e-01 -1.79533780e-01 -1.85955608e+00 1.36099648e+00
-2.19990164e-01 6.23286724e-01 -1.78867787e-01 -4.46641684e-01
1.58581960e+00 4.70819920e-01 9.56787691e-02 -2.86179274e-01
6.81484193e-02 2.53689140e-01 4.78769056e-02 -6.07182264e-01
8.65677834e-01 -9.14596170e-02 -9.34574232e-02 5.20088673e-01
8.26279297e-02 6.33033514e-02 4.44672376e-01 -2.03142792e-01
9.43225384e-01 1.94317117e-01 4.00133580e-01 -8.02845135e-03
9.84234691e-01 -1.89974979e-01 3.75957966e-01 9.56117749e-01
-8.04218948e-02 4.57911342e-01 2.62806118e-01 -9.00100693e-02
-1.21517873e+00 -9.84986246e-01 -2.22587902e-02 1.10351539e+00
-4.49698865e-01 -5.31343043e-01 -8.78164649e-01 -2.06102729e-01
-2.58437127e-01 2.37248212e-01 -2.47670114e-01 5.32168090e-01
-7.68698633e-01 -5.85392356e-01 8.42104733e-01 2.37432048e-01
3.16418201e-01 -1.13543296e+00 -7.98005521e-01 4.77728754e-01
3.38900864e-01 -7.27231979e-01 -4.62394446e-01 3.32863510e-01
-9.40948129e-01 -8.38010013e-01 -7.30280936e-01 -1.02515483e+00
4.87542421e-01 -1.30768701e-01 3.33116889e-01 -1.67181596e-01
-5.57699084e-01 5.90158589e-02 -3.88390571e-01 -6.99500203e-01
-3.97290289e-01 2.00892583e-01 6.79152831e-02 1.80155873e-01
5.50735831e-01 -2.27765858e-01 -2.63925850e-01 1.91534594e-01
-5.11521220e-01 -1.85373425e-01 5.06263196e-01 6.91561520e-01
3.32001507e-01 5.42406440e-01 2.93124944e-01 -7.64909208e-01
9.52650428e-01 1.88297540e-01 -8.01844656e-01 1.09714493e-01
-3.83543462e-01 1.93278238e-01 6.70393050e-01 -1.94375768e-01
-1.22786033e+00 4.24645633e-01 -5.53478748e-02 2.75897741e-01
-2.26328075e-01 2.32875019e-01 7.22316056e-02 2.22779304e-01
4.59200472e-01 5.44191480e-01 -1.43461272e-01 -1.01352358e+00
-2.65862286e-01 1.00030994e+00 7.12208986e-01 -3.36583704e-01
7.41874099e-01 2.71938205e-01 1.43184230e-01 -1.06876636e+00
-3.84263486e-01 -9.88526881e-01 -9.55957532e-01 -4.12828773e-01
9.03407931e-01 -2.87971228e-01 -8.35727274e-01 5.63352585e-01
-9.53824997e-01 4.62421894e-01 2.43126333e-01 7.09173024e-01
-6.70565426e-01 7.03255653e-01 -6.16584241e-01 -1.46370685e+00
-4.11042422e-01 -9.32254016e-01 6.84446812e-01 3.85436237e-01
-7.69312918e-01 -5.56739509e-01 4.18124467e-01 3.56266648e-01
-2.33808741e-01 -8.03083405e-02 8.66832852e-01 -6.53850138e-01
-4.47335355e-02 -6.72884941e-01 2.12691039e-01 3.32844287e-01
1.63329139e-01 2.11524814e-01 -9.38388586e-01 -3.88270454e-03
2.52643019e-01 4.01281446e-01 7.03489125e-01 9.25937817e-02
6.49262369e-01 1.00021437e-01 -3.94122750e-02 3.85383904e-01
1.55281711e+00 5.65306246e-01 2.52728552e-01 6.42343640e-01
5.34750462e-01 3.03640008e-01 7.60776043e-01 8.59785616e-01
-4.39532727e-01 8.35006118e-01 -5.09249866e-01 2.60793984e-01
8.77913907e-02 -4.25845720e-02 6.20526433e-01 1.15003705e+00
-3.70252371e-01 -8.51873588e-03 -7.86254168e-01 3.03176433e-01
-1.79577744e+00 -1.19447708e+00 -5.95076203e-01 2.33712554e+00
7.06809878e-01 4.99622405e-01 4.85932738e-01 1.05616200e+00
7.98684180e-01 -2.95126885e-01 2.35370249e-01 -9.98454988e-01
-1.23637855e-01 5.26558578e-01 6.60541177e-01 7.07590818e-01
-1.13742518e+00 8.44331145e-01 5.49082661e+00 8.66574347e-01
-8.46132994e-01 -1.07317135e-01 -2.03938529e-01 -1.31302670e-01
5.47178149e-01 3.01983114e-02 -1.18558252e+00 5.00581563e-01
7.04379678e-01 1.88465007e-02 2.52353758e-01 5.19602537e-01
6.32147908e-01 -4.99777585e-01 -6.67267621e-01 1.18431878e+00
3.15431565e-01 -7.25451410e-01 1.10171093e-02 -4.25953232e-02
5.50940812e-01 -6.75641596e-01 -2.40057427e-02 -1.71063557e-01
-2.30393857e-01 -8.96791816e-01 8.82000029e-01 1.03544295e+00
4.34225947e-01 -1.24524081e+00 8.17587912e-01 3.69771689e-01
-1.17013156e+00 -3.80384475e-02 -3.36885124e-01 -2.85941303e-01
1.54810408e-02 1.87379763e-01 -1.35389853e+00 4.05120611e-01
3.19184840e-01 9.46655348e-02 -7.71346748e-01 1.20285141e+00
1.80290639e-01 9.08978462e-01 -3.75948906e-01 -2.36505315e-01
2.84170881e-02 -4.21846628e-01 5.88177204e-01 1.71944070e+00
7.75551975e-01 -3.15606207e-01 -1.70088917e-01 4.21908647e-01
6.42342329e-01 7.48284280e-01 -1.96578696e-01 4.43264842e-03
3.92624438e-01 1.24697495e+00 -1.26520789e+00 -3.84742469e-01
-2.77118146e-01 1.25024414e+00 -2.61692554e-01 -8.70397091e-02
-3.65771979e-01 -7.57674515e-01 -1.85164332e-01 1.95702985e-01
5.05277991e-01 -5.08658350e-01 -6.91186488e-01 -5.36430657e-01
-1.58888437e-02 -7.50415325e-01 4.35989827e-01 -4.38680053e-01
-7.46895492e-01 3.76660496e-01 -3.56539220e-01 -1.40366590e+00
-3.01975042e-01 -5.90687990e-01 -5.57465494e-01 1.12703609e+00
-4.12293613e-01 -8.68681490e-01 -1.77399523e-03 4.20642346e-01
9.01404142e-01 -8.39348018e-01 1.09509826e+00 3.91413830e-02
-4.47332531e-01 5.56885302e-01 6.60712540e-01 3.03529322e-01
9.83358800e-01 -1.31149507e+00 7.18825981e-02 8.94015014e-01
5.89080572e-01 8.96407783e-01 1.00777483e+00 -9.58248913e-01
-1.03018582e+00 -1.93310246e-01 1.10108554e+00 -3.67646992e-01
5.62010229e-01 -4.91492808e-01 -5.72593927e-01 4.41158473e-01
1.12514652e-01 -9.60066795e-01 1.02592647e+00 8.79859924e-02
1.20668478e-01 -8.73343945e-02 -6.56893432e-01 3.96557361e-01
2.28869081e-01 -4.37597752e-01 -8.46302807e-01 -1.38751343e-01
-4.17613536e-01 2.14517452e-02 -7.35870242e-01 -1.69954509e-01
1.07651246e+00 -1.15013731e+00 5.19287586e-01 -6.58214539e-02
-4.31647636e-02 -8.04865062e-01 1.81917146e-01 -7.02652454e-01
-4.86216307e-01 -7.14749813e-01 2.71003455e-01 1.54934013e+00
2.87717909e-01 1.43544912e-01 8.78970206e-01 2.48261401e-03
1.62601694e-01 3.17793861e-02 -5.34395695e-01 -6.90007031e-01
-6.66045308e-01 -4.21795398e-01 2.68937171e-01 5.99438965e-01
-3.94593701e-02 3.89836103e-01 -6.04770601e-01 -1.78798549e-02
7.72067904e-01 -1.87617783e-02 8.31346035e-01 -1.42655718e+00
-8.62440467e-01 -3.65324795e-01 -5.50188005e-01 -3.55355293e-01
-3.05121094e-01 -6.27404392e-01 -1.18689306e-01 -1.12611556e+00
3.83197516e-01 -4.59692590e-02 -3.68065774e-01 -8.72875080e-02
2.08805403e-04 3.75679731e-01 4.06318128e-01 4.52900350e-01
-1.88411847e-01 -2.59246320e-01 8.17570865e-01 2.17374265e-01
-6.09484673e-01 4.44264680e-01 5.63130714e-02 8.95801723e-01
7.65844285e-01 -5.62505305e-01 -2.34247714e-01 8.77467319e-02
2.24471629e-01 1.27069250e-01 -6.52093999e-03 -1.33412385e+00
3.85742426e-01 2.59228587e-01 9.11299109e-01 -1.12914228e+00
2.28136748e-01 -6.71834111e-01 2.02681392e-01 4.49095100e-01
-1.78068995e-01 2.53890753e-01 -4.39734794e-02 3.45178127e-01
-1.41117290e-01 -9.77815628e-01 5.47250330e-01 -7.75464997e-02
-6.43527925e-01 -5.82521081e-01 -9.12062049e-01 -7.87910819e-01
6.98332727e-01 -7.51212955e-01 4.92950678e-01 -6.54303506e-02
-1.19435060e+00 -5.41934848e-01 2.65734166e-01 2.44463399e-01
4.05985415e-01 -9.99788284e-01 -6.87138379e-01 4.62147295e-01
2.89927516e-02 -7.30062246e-01 2.95591295e-01 7.94018328e-01
-1.04227960e+00 4.91600335e-01 -4.40931708e-01 -3.24820817e-01
-2.27628469e+00 1.30808577e-01 -2.11945221e-01 -2.50588894e-01
-5.16272306e-01 7.60332286e-01 -3.30468863e-01 6.88200817e-02
3.13400120e-01 -8.53024572e-02 -6.76950395e-01 3.10862988e-01
5.53066075e-01 7.39601254e-01 2.35536575e-01 -7.20258713e-01
-2.78099000e-01 1.01383972e+00 -1.66481182e-01 -9.60754514e-01
1.07732964e+00 -1.27042815e-01 -3.51228379e-02 8.95283520e-01
7.73748159e-01 9.03613091e-01 -7.92668879e-01 1.19278997e-01
7.62546778e-01 -4.62974995e-01 -1.44620702e-01 -7.76250422e-01
-1.15243658e-01 4.52696741e-01 9.93371487e-01 6.02300139e-03
9.90558147e-01 -6.16283238e-01 5.41567504e-01 3.72292548e-01
7.00251982e-02 -1.76973510e+00 -2.66366243e-01 5.41236341e-01
7.08935618e-01 -7.25839853e-01 -3.61388177e-02 -2.02753037e-01
-5.79118133e-01 1.60699475e+00 1.44254953e-01 -3.36619943e-01
3.88213217e-01 6.70504212e-01 1.50136799e-01 8.89979824e-02
-2.39734784e-01 -6.51927711e-03 1.50273517e-01 4.05969203e-01
8.62906575e-01 1.69730052e-01 -9.41965699e-01 1.12281871e+00
-8.23957145e-01 2.73181940e-04 7.67971694e-01 1.06415677e+00
-7.98305631e-01 -1.42119241e+00 -1.16433597e+00 4.17865306e-01
-8.67217779e-01 3.73287648e-02 -6.08986259e-01 4.27117586e-01
2.33343959e-01 8.99937570e-01 -7.29678497e-02 -4.40100759e-01
2.41681263e-01 4.56058353e-01 7.97747076e-01 -5.01190960e-01
-8.99036646e-01 8.45597684e-01 6.96419552e-02 1.01393327e-01
-4.69391234e-02 -1.21167088e+00 -1.40118480e+00 9.22224596e-02
-3.68416131e-01 3.72996271e-01 1.04270124e+00 7.41421521e-01
-3.27765107e-01 4.22477275e-01 4.37891960e-01 -3.80908966e-01
-4.65147108e-01 -1.23891616e+00 -1.19460082e+00 2.00502604e-01
4.74685989e-02 -3.28030974e-01 -1.03537813e-01 4.22422469e-01]
|
[15.860721588134766, 5.294817924499512]
|
13c573d6-fcb5-4103-9fad-417e5c6949db
|
a-deep-cnn-architecture-with-novel-pooling
|
2201.12664
| null |
https://arxiv.org/abs/2201.12664v1
|
https://arxiv.org/pdf/2201.12664v1.pdf
|
A Deep CNN Architecture with Novel Pooling Layer Applied to Two Sudanese Arabic Sentiment Datasets
|
Arabic sentiment analysis has become an important research field in recent years. Initially, work focused on Modern Standard Arabic (MSA), which is the most widely-used form. Since then, work has been carried out on several different dialects, including Egyptian, Levantine and Moroccan. Moreover, a number of datasets have been created to support such work. However, up until now, less work has been carried out on Sudanese Arabic, a dialect which has 32 million speakers. In this paper, two new publicly available datasets are introduced, the 2-Class Sudanese Sentiment Dataset (SudSenti2) and the 3-Class Sudanese Sentiment Dataset (SudSenti3). Furthermore, a CNN architecture, SCM, is proposed, comprising five CNN layers together with a novel pooling layer, MMA, to extract the best features. This SCM+MMA model is applied to SudSenti2 and SudSenti3 with accuracies of 92.75% and 84.39%. Next, the model is compared to other deep learning classifiers and shown to be superior on these new datasets. Finally, the proposed model is applied to the existing Saudi Sentiment Dataset and to the MSA Hotel Arabic Review Dataset with accuracies 85.55% and 90.01%.
|
['Ephrem A. Retta', 'Eiad Almekhlafi', 'Jun Feng', 'Xia Sun', 'Richard Sutcliffe', 'Mustafa Mhamed']
|
2022-01-29
| null | null | null | null |
['arabic-sentiment-analysis']
|
['natural-language-processing']
|
[-4.82239276e-01 -2.35019594e-01 2.33105734e-01 -6.14515841e-01
-3.99597794e-01 -4.20764089e-01 6.47340119e-01 1.68265954e-01
-5.49631715e-01 8.28006446e-01 -2.28882387e-01 -2.98289079e-02
2.21463785e-01 -9.22025263e-01 -1.44353211e-01 -8.06966126e-01
-1.11763753e-01 1.93845302e-01 -7.25106746e-02 -1.07056177e+00
5.40293217e-01 5.43479145e-01 -1.37995172e+00 4.11521554e-01
1.01929367e+00 9.04576182e-01 -2.89346009e-01 3.14408153e-01
2.48178706e-01 5.86085081e-01 -9.75079477e-01 -8.69914114e-01
2.26488277e-01 -4.21586454e-01 -7.30071664e-01 -2.02783793e-02
2.92891651e-01 -1.75038457e-01 9.40196365e-02 7.95388818e-01
6.57050192e-01 1.71402656e-02 6.19608521e-01 -1.03125346e+00
-1.03439844e+00 7.28551745e-01 -8.82113874e-01 -1.33307591e-01
6.84831366e-02 -5.89527071e-01 8.78324866e-01 -1.19017291e+00
4.39417422e-01 1.33595490e+00 5.93568921e-01 3.03651571e-01
-2.61929810e-01 -9.64688420e-01 3.83653707e-04 3.76898825e-01
-1.20009816e+00 -7.58447424e-02 7.15020001e-01 2.08471436e-03
8.16770792e-01 -1.51045725e-01 6.80429697e-01 6.04052365e-01
5.35705030e-01 8.38992417e-01 1.61763668e+00 -6.43040061e-01
1.15583800e-01 4.11397338e-01 3.71748507e-01 7.49642849e-01
3.21319193e-01 -5.60336769e-01 -5.71824491e-01 2.81068176e-01
9.42429677e-02 -2.04953641e-01 -5.96727431e-02 3.26276183e-01
-6.85382426e-01 1.15010571e+00 5.01872301e-01 4.30236757e-01
-2.63890445e-01 -4.19645131e-01 3.68751466e-01 5.64541996e-01
6.39874995e-01 2.97333747e-01 -6.43241584e-01 -2.33550500e-02
-7.53625989e-01 4.40796077e-01 8.65306139e-01 7.43426144e-01
5.88049173e-01 3.42685610e-01 7.40285933e-01 1.18354988e+00
7.42920160e-01 6.79274321e-01 6.64880276e-01 -1.77469403e-01
4.49540794e-01 9.45071876e-01 -4.79933433e-02 -1.26170409e+00
-6.51931167e-01 -1.40085086e-01 -7.82791734e-01 3.35451663e-01
3.83561254e-01 -5.62175035e-01 -1.01374078e+00 1.27452493e+00
1.06619917e-01 -6.46025896e-01 4.88144666e-01 9.68706310e-01
9.81027842e-01 9.22822297e-01 -1.65530697e-01 2.66184688e-01
1.33217776e+00 -9.52469110e-01 -7.00563014e-01 -1.36621311e-01
3.95039141e-01 -1.23355448e+00 8.29183221e-01 8.87057185e-01
-1.03864563e+00 -3.06271374e-01 -1.51467240e+00 -2.87779048e-03
-1.01190758e+00 2.30806723e-01 5.96474648e-01 1.23160708e+00
-1.07245731e+00 2.49349266e-01 -6.79829419e-01 -5.94796002e-01
5.45149684e-01 6.51444674e-01 -5.61573744e-01 -2.61916190e-01
-1.43131316e+00 1.34057283e+00 2.31668383e-01 5.93430817e-01
-5.85957885e-01 -8.29260126e-02 -8.91379178e-01 -1.88859686e-01
-1.94935828e-01 3.84072274e-01 9.12927568e-01 -1.39389992e+00
-1.77396142e+00 8.30712855e-01 1.05728716e-01 -1.49420261e-01
9.54798758e-02 -2.93988019e-01 -9.21013594e-01 -1.01951979e-01
-2.10194439e-01 4.93622869e-01 5.32792747e-01 -1.29814720e+00
-7.96354949e-01 -7.60400414e-01 4.39379990e-01 2.07747072e-01
-5.36941767e-01 4.89101201e-01 -2.64939815e-01 -8.38960648e-01
2.20543608e-01 -1.20756102e+00 1.21134296e-01 -7.42362857e-01
-2.03761131e-01 -4.47332524e-02 8.79165530e-01 -9.84849751e-01
1.18387473e+00 -1.84387589e+00 -8.41755886e-03 4.75583255e-01
-3.54466110e-01 4.94767815e-01 -2.06250921e-01 6.52561903e-01
-3.76899205e-02 -1.88682020e-01 -5.27711034e-01 -5.61854728e-02
-6.17368072e-02 -2.14480627e-02 2.99004484e-02 5.46554804e-01
7.08041012e-01 5.34271896e-01 -5.13976872e-01 -1.59222782e-01
-1.26496822e-01 5.49218357e-01 -3.17467123e-01 -2.96592474e-01
2.53740311e-01 8.08683485e-02 -1.69665709e-01 1.39868534e+00
1.16606927e+00 2.01190278e-01 4.78361070e-01 -5.33862077e-02
-3.06212485e-01 -2.00886637e-01 -9.33826387e-01 1.20502245e+00
-4.52304751e-01 7.23700941e-01 -7.04817986e-03 -8.48200679e-01
1.29896617e+00 1.76561192e-01 1.69756219e-01 -8.48908067e-01
7.51182497e-01 8.33713353e-01 3.61233771e-01 -2.93099493e-01
9.53382611e-01 -1.37766108e-01 -2.30156586e-01 4.41943526e-01
2.00154390e-02 -1.76724389e-01 5.57568610e-01 -1.23971403e-01
3.47865909e-01 8.94144922e-02 2.23084599e-01 -4.97232556e-01
9.55293417e-01 8.91323984e-02 3.22335213e-01 -2.76560541e-02
-3.60640824e-01 7.49219537e-01 4.71914083e-01 -6.58851445e-01
-6.47158682e-01 -5.95137358e-01 -3.58501315e-01 1.06816840e+00
6.51471838e-02 -2.69199431e-01 -8.65264535e-01 -8.75330567e-01
-1.67353943e-01 2.68332541e-01 -8.12304914e-01 1.52654096e-01
-8.68149638e-01 -1.51516664e+00 5.90521455e-01 3.33081216e-01
1.14718533e+00 -1.32828510e+00 -6.64973021e-01 1.90519795e-01
5.71311861e-02 -6.20064795e-01 8.79287198e-02 1.59225196e-01
-6.47067487e-01 -1.19373775e+00 -8.25226068e-01 -1.08302832e+00
4.35421646e-01 -4.11490165e-02 9.48006809e-01 1.34415373e-01
1.37186497e-01 -2.69039750e-01 -8.04144561e-01 -1.07228148e+00
-2.75370181e-01 3.90733093e-01 -5.96709289e-02 2.38885716e-01
8.92208219e-01 7.08307251e-02 -2.34823123e-01 3.02517951e-01
-9.23670650e-01 -3.84368718e-01 6.21350706e-01 8.26134264e-01
-8.33990201e-02 -8.05451795e-02 1.22146487e+00 -1.04675078e+00
5.92578888e-01 -6.38509393e-01 -4.37860101e-01 1.31570309e-01
-5.23515880e-01 -5.56111217e-01 6.54210448e-01 6.13709576e-02
-1.25121093e+00 -2.02574313e-01 -4.64791834e-01 5.75383127e-01
-5.43709286e-03 1.02425551e+00 -2.60960191e-01 2.92828977e-02
4.88148123e-01 -6.87443167e-02 2.36656785e-01 -9.60903466e-02
-3.03892861e-03 1.32706177e+00 -7.60312751e-02 -2.20900148e-01
3.06146324e-01 5.07994890e-01 -2.07458422e-01 -1.03796577e+00
-7.81463504e-01 -2.26011071e-02 -8.02900493e-01 -1.47246882e-01
6.65306866e-01 -9.55416143e-01 -3.59229624e-01 1.40256274e+00
-7.55264401e-01 -1.33901149e-01 5.11689544e-01 2.98661560e-01
-3.91715989e-02 4.08982448e-02 -6.83645427e-01 -7.11488187e-01
-4.87732559e-01 -1.43353462e+00 6.44992828e-01 5.41363657e-01
-1.33255795e-01 -1.01754510e+00 -1.08772531e-01 2.90322870e-01
5.20179927e-01 3.64342690e-01 7.82740951e-01 -7.90662766e-01
3.61004286e-02 -4.23014104e-01 -2.51998842e-01 7.78515816e-01
3.40312332e-01 2.93756455e-01 -1.14800811e+00 -4.98419613e-01
-1.29810423e-01 -6.04872823e-01 6.17031693e-01 1.87980920e-01
3.98802310e-01 2.39704534e-01 3.48580629e-01 1.29872322e-01
1.29836369e+00 8.97942901e-01 7.35792816e-01 1.00494874e+00
2.83592820e-01 7.44789422e-01 9.29982662e-01 3.43571573e-01
7.82544434e-01 1.57565981e-01 3.72598946e-01 2.41125636e-02
1.65987611e-01 6.23917103e-01 6.99044585e-01 1.34353197e+00
-2.33943790e-01 -1.92058474e-01 -1.09082294e+00 5.94818056e-01
-1.54304123e+00 -4.29918110e-01 -2.59312421e-01 1.70497108e+00
7.06419468e-01 -1.28062114e-01 2.70441249e-02 7.13934600e-01
3.63889903e-01 9.93844196e-02 -1.93238422e-01 -9.17781889e-01
-6.48073375e-01 7.92290986e-01 2.13666540e-03 3.75785470e-01
-1.50723028e+00 1.05430067e+00 5.09205770e+00 3.50565016e-01
-1.52498114e+00 -4.34030481e-02 7.05608308e-01 2.23614618e-01
2.97638983e-01 -3.89846176e-01 -7.61106551e-01 4.27287430e-01
1.08569872e+00 2.57803887e-01 1.32026792e-01 5.52807629e-01
-1.63405821e-01 -2.11947024e-01 -4.01520640e-01 6.53357208e-01
8.29151392e-01 -8.66190195e-01 6.73239008e-02 -1.93274811e-01
9.68464553e-01 -6.15283586e-02 3.62073720e-01 4.97391105e-01
8.61814097e-02 -1.18345749e+00 6.40321553e-01 1.36899292e-01
4.44859535e-01 -1.52456617e+00 1.54585755e+00 -1.49539933e-01
-6.91438615e-01 -3.32769543e-01 -3.30202192e-01 -1.16606280e-01
-2.61623830e-01 2.12294206e-01 -5.69472790e-01 7.22416759e-01
1.19397080e+00 9.35957730e-01 -8.43754172e-01 6.12900853e-01
-2.99288064e-01 6.55900955e-01 -1.35149434e-01 -4.02803004e-01
7.82377660e-01 -5.48120558e-01 -9.26496461e-02 1.25306451e+00
4.76062924e-01 -6.02255724e-02 -2.95397997e-01 -5.87557591e-02
-1.48129359e-01 6.19289279e-01 -2.66261995e-01 7.64784664e-02
8.87327865e-02 1.48008859e+00 -8.41956079e-01 -9.13763121e-02
-5.41170657e-01 7.05039442e-01 8.08393881e-02 1.44313872e-01
-7.11512864e-01 -1.14020967e+00 4.50806201e-01 -4.12702918e-01
8.04399103e-02 -6.99488446e-02 -2.87233502e-01 -1.08829105e+00
-2.28068903e-01 -1.50923121e+00 4.62071240e-01 -5.36922812e-01
-1.19303524e+00 9.62490499e-01 -3.12402844e-01 -1.06984234e+00
-3.48370685e-03 -1.28270948e+00 -3.03765863e-01 1.18929100e+00
-1.83901298e+00 -1.47568381e+00 -3.02274019e-01 4.40317869e-01
5.97252071e-01 -9.07198489e-01 9.52349782e-01 6.65014029e-01
-7.45196640e-01 7.21278787e-01 3.05507511e-01 3.82390738e-01
1.01581120e+00 -1.25209820e+00 7.07313493e-02 7.28831530e-01
-3.54256094e-01 8.13327491e-01 2.10015655e-01 -3.99283201e-01
-1.31940472e+00 -9.69201088e-01 1.09455621e+00 -4.17477369e-01
5.89538813e-01 -1.98751032e-01 -8.72427583e-01 5.96844852e-01
7.65169203e-01 -5.48380375e-01 8.73853564e-01 -1.47364726e-02
-2.04848155e-01 -3.17936659e-01 -1.36775982e+00 6.84531689e-01
-8.75273868e-02 -1.62180141e-02 -4.25655097e-01 6.62311260e-03
-1.05721042e-01 -4.64551240e-01 -8.83905411e-01 2.56912649e-01
8.85716140e-01 -9.33492899e-01 3.96505833e-01 -4.29361463e-01
8.55679572e-01 -2.64810681e-01 -4.81171012e-01 -1.64363432e+00
1.84172735e-01 -1.40207335e-01 1.59807369e-01 1.07209432e+00
6.30163848e-01 -8.22969735e-01 6.86433554e-01 2.18862873e-02
-2.11816132e-01 -9.85733032e-01 -5.32700777e-01 -2.46441767e-01
6.84436560e-01 -2.30466902e-01 8.91877532e-01 1.22429156e+00
-1.97800025e-01 2.79165596e-01 -2.01200008e-01 8.96004960e-02
1.29984133e-02 1.16485260e-01 6.27969682e-01 -1.24952388e+00
5.89305758e-01 -3.94929707e-01 -2.71342874e-01 -2.30420589e-01
8.43045190e-02 -7.30577648e-01 -1.18566930e-01 -1.41747212e+00
-3.05628955e-01 -4.48736042e-01 -4.14894223e-01 5.48569441e-01
-3.95969003e-02 8.54875565e-01 2.46443689e-01 -1.85794860e-01
1.67841613e-02 3.50525081e-01 1.15568006e+00 -1.51245907e-01
-2.01432616e-01 -1.65885717e-01 -9.47220922e-01 9.71986115e-01
1.16066468e+00 -4.71498482e-02 -1.23159207e-01 -3.81485790e-01
4.43850815e-01 -5.10570824e-01 -3.77746940e-01 -8.85433376e-01
-2.10492909e-01 6.68719187e-02 6.37352049e-01 -9.30088222e-01
3.83610964e-01 -6.64868474e-01 -4.73257959e-01 3.96782160e-01
8.48691240e-02 5.88639319e-01 3.83150280e-01 -1.80870891e-01
-6.73282444e-01 -2.58814514e-01 1.12179494e+00 7.26689994e-02
-8.15244496e-01 9.76933762e-02 -4.82148737e-01 -3.44824493e-01
1.11555862e+00 -1.10372864e-01 -3.77708077e-01 -1.10635832e-01
-4.40406770e-01 2.30520174e-01 6.25002161e-02 6.99289203e-01
7.70594001e-01 -1.13076258e+00 -1.06333876e+00 3.64943355e-01
1.73064709e-01 -3.21791917e-01 1.24920182e-01 7.29377210e-01
-1.27424073e+00 3.36293727e-01 -7.95325577e-01 -1.33110628e-01
-1.27355826e+00 -1.82046399e-01 1.03997678e-01 1.89064983e-02
-4.74732295e-02 7.52184808e-01 -5.37428319e-01 -9.86330450e-01
-1.88696414e-01 -4.57439087e-02 -9.94682968e-01 7.91566133e-01
4.35571402e-01 4.68728960e-01 5.21830320e-01 -1.19569659e+00
-3.60166997e-01 5.92414856e-01 -3.22725296e-01 -1.68170750e-01
1.50020790e+00 1.95987616e-02 -7.21812725e-01 3.94030660e-01
9.51110482e-01 3.19746047e-01 -4.93368298e-01 2.18566895e-01
1.65220723e-01 -2.40568966e-01 -1.80938318e-01 -1.15344405e+00
-1.31618071e+00 1.05342281e+00 7.79845774e-01 2.25582525e-01
1.29097736e+00 -7.45031893e-01 7.03275859e-01 6.03562176e-01
3.27188879e-01 -1.36477315e+00 -2.35967070e-01 1.19315195e+00
9.69142258e-01 -1.52154219e+00 -6.31047189e-02 -2.53137350e-02
-7.43104756e-01 1.36777318e+00 7.42569923e-01 -3.31962824e-01
1.01845551e+00 -5.83987162e-02 8.76417875e-01 -6.92628175e-02
-2.29466334e-01 8.74630064e-02 1.88410610e-01 4.71690804e-01
1.08215988e+00 5.56205362e-02 -6.14326894e-01 8.18858385e-01
-6.57903552e-01 -2.12746896e-02 1.03788745e+00 1.32799017e+00
-2.12304682e-01 -1.14921951e+00 -4.74574834e-01 4.65491980e-01
-8.95064950e-01 -1.38524026e-01 -4.30969983e-01 1.49414170e+00
1.94408730e-01 1.23945296e+00 -1.20376796e-02 -3.06368858e-01
3.53012681e-01 1.28519282e-01 2.38040715e-01 -4.68956143e-01
-1.19501245e+00 -5.94497621e-02 6.52584955e-02 7.06772953e-02
-7.00667381e-01 -4.89101022e-01 -1.39027119e+00 -5.46207368e-01
-3.55100662e-01 2.24155009e-01 1.04788935e+00 7.48057902e-01
-2.99529620e-02 2.75808573e-01 7.33508408e-01 -6.98024213e-01
-2.45366365e-01 -1.30549753e+00 -9.11899090e-01 2.69111484e-01
1.88029021e-01 -5.00167668e-01 5.68083338e-02 -9.02280807e-02]
|
[11.144991874694824, 7.060960292816162]
|
8801fe14-f8de-462b-8195-1b3832a1c61a
|
demonstration-of-machine-learning-enhanced
|
2212.08032
| null |
https://arxiv.org/abs/2212.08032v1
|
https://arxiv.org/pdf/2212.08032v1.pdf
|
Demonstration of machine-learning-enhanced Bayesian quantum state estimation
|
Machine learning (ML) has found broad applicability in quantum information science in topics as diverse as experimental design, state classification, and even studies on quantum foundations. Here, we experimentally realize an approach for defining custom prior distributions that are automatically tuned using ML for use with Bayesian quantum state estimation methods. Previously, researchers have looked to Bayesian quantum state tomography due to its unique advantages like natural uncertainty quantification, the return of reliable estimates under any measurement condition, and minimal mean-squared error. However, practical challenges related to long computation times and conceptual issues concerning how to incorporate prior knowledge most suitably can overshadow these benefits. Using both simulated and experimental measurement results, we demonstrate that ML-defined prior distributions reduce net convergence times and provide a natural way to incorporate both implicit and explicit information directly into the prior distribution. These results constitute a promising path toward practical implementations of Bayesian quantum state tomography.
|
['Brian T. Kirby', 'Thomas A. Searles', 'Ryan T. Glasser', 'Daniel E. Jones', 'Sangita Regmi', 'Amirali Khannejad', 'Atiyya A. Davis', 'Joseph M. Lukens', 'Sanjaya Lohani']
|
2022-12-15
| null | null | null | null |
['quantum-state-tomography', 'experimental-design']
|
['medical', 'methodology']
|
[ 3.48587066e-01 -2.66926110e-01 -3.02171111e-01 -6.43536091e-01
-1.13214338e+00 -3.84103864e-01 7.49422133e-01 2.55836435e-02
-6.86204493e-01 1.10164201e+00 -4.33493108e-02 -6.97606683e-01
-1.99463755e-01 -7.10872114e-01 -3.33245784e-01 -8.77471566e-01
1.85427070e-01 4.84893799e-01 1.17963320e-02 2.16195490e-02
7.14252234e-01 5.60622394e-01 -1.04824781e+00 -4.27716136e-01
6.38582230e-01 7.42611468e-01 -2.39839539e-01 5.04334450e-01
1.60688818e-01 3.77001315e-01 -5.04306436e-01 -2.17485756e-01
3.29008549e-02 -6.04729891e-01 -6.92260921e-01 -4.16480154e-01
1.02905653e-01 -5.29960752e-01 -6.11213803e-01 1.43128002e+00
6.26759291e-01 1.51207745e-01 8.41592371e-01 -8.31696391e-01
-1.02527179e-01 7.00672805e-01 -2.28515342e-01 2.86934555e-01
3.02814245e-01 4.25639689e-01 9.15196419e-01 -4.23834085e-01
3.86170566e-01 1.00962591e+00 3.75320166e-01 5.31893551e-01
-1.84707725e+00 -7.17038274e-01 -7.48741567e-01 4.26729918e-01
-1.54429054e+00 -9.15162504e-01 6.90767109e-01 -2.91327059e-01
9.34699774e-01 -2.28677299e-02 4.41868454e-01 7.96591699e-01
5.74574769e-01 3.78536195e-01 1.51613641e+00 -8.89978766e-01
6.11593723e-01 3.48098338e-01 4.18998450e-01 5.17335415e-01
3.45921010e-01 6.11258209e-01 -4.75988775e-01 -3.45462263e-01
6.87619567e-01 -5.14755905e-01 -1.39122963e-01 -5.05021632e-01
-1.11286914e+00 8.73551786e-01 1.46963775e-01 5.83138317e-02
-1.24728657e-01 5.41985750e-01 2.50896424e-01 3.16690147e-01
-1.57674074e-01 6.01519763e-01 -9.18639004e-02 -3.51742148e-01
-8.71257603e-01 1.56096697e-01 8.65529060e-01 8.55339766e-01
1.18438160e+00 -4.66512749e-03 8.28544572e-02 2.17156485e-01
4.33631331e-01 8.08954597e-01 -1.36281058e-01 -1.17486620e+00
4.76129390e-02 -2.72644937e-01 4.87629652e-01 -4.58345145e-01
-3.67348969e-01 -1.70971319e-01 -5.62499762e-01 1.49851248e-01
4.12391096e-01 -2.01232731e-01 -8.18647683e-01 1.77734530e+00
2.54129797e-01 -1.21847898e-01 2.35004529e-01 7.26096809e-01
4.13537532e-01 5.89286387e-01 -8.62688273e-02 -4.44320261e-01
1.03151977e+00 -1.04691051e-01 -8.56868446e-01 -9.23331082e-02
5.99853456e-01 -8.01969528e-01 6.44288480e-01 3.31009865e-01
-9.46173310e-01 -2.30757892e-01 -1.26400292e+00 5.59546379e-03
9.32044908e-02 -2.12415621e-01 1.05900776e+00 1.34112442e+00
-8.32246423e-01 1.01959276e+00 -1.01087999e+00 -1.78082034e-01
1.21306643e-01 6.26570821e-01 -6.77996203e-02 -7.51323476e-02
-1.04263151e+00 1.22527373e+00 6.25760615e-01 -2.54592649e-03
-6.31789982e-01 -3.34239542e-01 -6.03507161e-01 -4.27505821e-02
3.73758942e-01 -6.26337945e-01 1.42797887e+00 -6.12431653e-02
-2.06457853e+00 4.22291815e-01 -3.31936181e-01 -3.16686213e-01
-2.33533010e-02 2.72278041e-01 -5.14313281e-01 1.89231187e-01
-5.64452633e-02 3.55498761e-01 7.14004099e-01 -6.27361059e-01
1.46938674e-02 -1.89124331e-01 -5.21506518e-02 -1.28449991e-01
1.34969071e-01 -9.53349099e-02 -8.71375799e-02 3.92835706e-01
7.22265065e-01 -1.01324725e+00 -3.66101533e-01 -6.08659685e-01
-4.06573534e-01 3.29437144e-02 3.69311064e-01 -7.18999878e-02
9.13810849e-01 -1.86989272e+00 3.60218808e-02 4.81799036e-01
-3.80466580e-02 4.10822267e-03 1.95318729e-01 5.54952681e-01
1.63804412e-01 -3.56483244e-04 1.41315153e-02 1.45178542e-01
3.00868094e-01 1.09075986e-01 -1.65226549e-01 7.21737862e-01
6.97909892e-02 7.72303581e-01 -9.45531666e-01 -2.68907428e-01
6.59415543e-01 1.54988021e-01 -6.68724477e-01 -4.44654301e-02
-6.45906031e-02 9.28474724e-01 -4.72914666e-01 4.32258695e-01
5.62372029e-01 -2.61649281e-01 5.26021302e-01 -5.19217610e-01
-2.62984365e-01 1.04667306e+00 -1.41488504e+00 1.49240613e+00
-2.61122972e-01 3.89192730e-01 2.58362859e-01 -9.34262276e-01
5.94642401e-01 1.78445756e-01 2.53515542e-01 -8.02580774e-01
4.88766253e-01 4.06205773e-01 4.41513956e-01 -2.42613062e-01
7.85707831e-01 -8.74769151e-01 -2.96912670e-01 6.91637635e-01
3.73337924e-01 -9.82549548e-01 7.27200285e-02 2.59477764e-01
7.96919584e-01 2.33403325e-01 4.07926649e-01 -5.96625268e-01
2.03239501e-01 -1.18671224e-01 3.80811960e-01 1.15789485e+00
-4.76695001e-01 -1.91808399e-02 2.34276488e-01 -1.42064840e-01
-1.22235310e+00 -1.25312793e+00 -8.02318156e-01 5.70363045e-01
4.32866961e-01 -5.46755791e-01 -4.39444214e-01 4.39791791e-02
-2.45013759e-01 1.04943907e+00 -1.43781817e-02 -3.47831070e-01
-2.38483027e-01 -1.07119393e+00 4.26838845e-01 1.25954583e-01
3.16708118e-01 -6.63976371e-01 -3.22721839e-01 4.37034309e-01
-1.88231096e-02 -1.05906749e+00 2.53436584e-02 5.23012757e-01
-8.65486860e-01 -7.30789125e-01 3.24381068e-02 -1.03864148e-02
3.48704964e-01 2.48981137e-02 7.83047199e-01 -5.81420779e-01
-4.44261014e-01 5.38237751e-01 1.16793595e-01 -2.53574789e-01
-8.60599041e-01 -1.72043994e-01 4.91255641e-01 -5.18915117e-01
7.08566368e-01 -6.71081603e-01 -3.51240247e-01 4.18941788e-02
-5.20345032e-01 -2.35215098e-01 6.74403131e-01 8.93932164e-01
2.79162198e-01 -2.62073129e-02 4.28737611e-01 -6.63001001e-01
5.31061113e-01 -1.02088757e-01 -9.95029509e-01 1.07423775e-01
-8.29926252e-01 6.71719372e-01 1.01453297e-01 1.12926094e-02
-1.07421446e+00 -3.77203435e-01 -2.89546758e-01 2.65973657e-01
-1.69760957e-01 5.76908112e-01 4.74475697e-02 -4.37948197e-01
8.32196772e-01 1.97977096e-01 3.05358265e-02 -2.64840350e-02
4.17058736e-01 6.66369855e-01 3.30880880e-01 -1.14844251e+00
6.19197488e-01 2.47619778e-01 5.57515204e-01 -9.81706500e-01
-9.88417625e-01 -4.51192170e-01 -7.15348840e-01 2.13066101e-01
6.32606268e-01 -7.47347891e-01 -9.38259900e-01 8.45642090e-02
-8.07314694e-01 6.35242509e-03 -9.79337320e-02 1.11416948e+00
-7.04306364e-01 7.77555466e-01 -6.81587756e-01 -1.22291720e+00
-1.80964783e-01 -1.44096017e+00 8.01842034e-01 3.87593150e-01
-1.81958452e-01 -1.03448379e+00 2.27878783e-02 3.29129547e-01
6.28823221e-01 -3.66514593e-01 1.01739037e+00 -1.60730764e-01
-1.03647232e+00 -2.30233073e-01 -2.66644061e-01 1.82832941e-01
5.43887131e-02 -2.11016938e-01 -1.12009180e+00 -5.34839392e-01
3.89242172e-02 -6.99157834e-01 8.31666529e-01 3.91962677e-01
7.93902874e-01 2.05756500e-01 -4.00838733e-01 5.32222927e-01
1.40796959e+00 1.01254791e-01 7.81221211e-01 -1.40206948e-01
2.85688698e-01 7.39295334e-02 2.08550185e-01 3.76223326e-01
6.69932179e-03 6.62167847e-01 -8.95860419e-02 5.48840880e-01
2.65654176e-01 -4.70818132e-02 2.88701862e-01 1.08838367e+00
-4.85364571e-02 8.05628896e-02 -6.63471699e-01 3.30665633e-02
-1.37547839e+00 -1.26584411e+00 -1.21715046e-01 2.62628269e+00
9.76614594e-01 3.62348437e-01 -2.80581445e-01 -1.02394493e-02
5.21782994e-01 -1.31210759e-01 -5.08264899e-01 -3.37851107e-01
1.63059920e-01 7.00813174e-01 6.11145437e-01 7.15962291e-01
-1.00637972e+00 8.94464970e-01 7.84654713e+00 9.28639412e-01
-1.06619525e+00 1.59537435e-01 5.29006198e-02 2.00855926e-01
-4.18541461e-01 6.01669550e-01 -1.02611852e+00 2.59821117e-01
1.36673450e+00 -9.38507989e-02 6.19090915e-01 2.93977022e-01
1.52439445e-01 -5.35022497e-01 -1.18413734e+00 1.01364470e+00
-4.08434927e-01 -1.52034593e+00 -3.24035734e-01 2.17290372e-01
7.19380677e-01 3.00177455e-01 -1.20521717e-01 4.70071137e-01
2.48557016e-01 -9.54593062e-01 4.91996348e-01 4.83730584e-01
1.04017746e+00 -6.30407989e-01 7.08761811e-01 4.35534716e-01
-5.70278466e-01 1.94136411e-01 -6.87214434e-01 -3.85990024e-01
3.43556941e-01 8.19383860e-01 -1.14285469e+00 4.45332229e-01
5.39450832e-02 2.16905147e-01 1.26384094e-01 9.24515545e-01
-2.31878966e-01 8.74807954e-01 -7.08153069e-01 -3.64952385e-01
2.08926320e-01 -6.31346524e-01 4.82794821e-01 9.55318451e-01
4.04901266e-01 2.18309477e-01 1.65463865e-01 1.24693143e+00
2.19321817e-01 -3.12528968e-01 -5.06220043e-01 -4.20098394e-01
7.54223645e-01 1.07068694e+00 -7.50709295e-01 -2.10983455e-01
-2.24091396e-01 4.17463213e-01 6.88253045e-02 2.29558617e-01
-5.87490797e-01 -3.32369447e-01 3.70283097e-01 -1.34203568e-01
1.64619192e-01 -6.48259342e-01 -3.27043384e-01 -1.39678562e+00
-4.18119431e-01 -5.05462110e-01 -1.58874169e-01 -3.78379524e-01
-1.20712602e+00 -8.73517618e-02 2.53558427e-01 -7.38515198e-01
-4.29473877e-01 -8.44313622e-01 -4.72692937e-01 1.05050957e+00
-1.25328684e+00 -6.57653749e-01 3.19164515e-01 2.03414142e-01
-4.45429713e-01 1.80611927e-02 1.06286490e+00 2.15824798e-01
-4.27056015e-01 3.40891480e-01 5.53114653e-01 -1.07891373e-01
6.45056069e-01 -1.20347750e+00 1.58590883e-01 7.40071714e-01
1.92132577e-01 1.07608044e+00 1.07843697e+00 -4.76616889e-01
-1.79453409e+00 -3.78457278e-01 3.68190348e-01 -5.81280231e-01
9.34611320e-01 -1.81842625e-01 -4.55342561e-01 5.06873608e-01
-9.74420905e-02 -1.69546977e-01 7.44133294e-01 7.09190071e-01
-3.25823277e-01 6.49464950e-02 -1.18879807e+00 4.34493363e-01
5.83404899e-01 -1.11686134e+00 -4.25498873e-01 2.46377885e-01
1.67638838e-01 -3.41147333e-01 -1.08151126e+00 4.70212787e-01
6.39633536e-01 -1.03875625e+00 7.90809870e-01 -2.76844680e-01
-3.19404095e-01 -3.56496304e-01 -4.73576218e-01 -1.11149704e+00
-3.38216573e-01 -1.02644587e+00 6.63346201e-02 7.76793242e-01
3.73471290e-01 -7.33122587e-01 7.81672597e-01 7.91565955e-01
-2.31042594e-01 -8.28839839e-02 -1.15738714e+00 -6.59986019e-01
3.16327572e-01 -7.65778720e-01 1.58239260e-01 5.92286944e-01
3.17589492e-01 6.28137290e-01 -2.75008112e-01 2.44296163e-01
1.01602030e+00 3.60480100e-01 6.89883053e-01 -1.10866392e+00
-5.94260514e-01 -3.82448107e-01 -6.38534129e-01 -1.39092696e+00
-1.76117029e-02 -9.25954878e-01 2.84704596e-01 -9.76485252e-01
5.39718270e-01 -7.65731215e-01 -3.91298980e-01 -3.68863754e-02
-3.32306810e-02 1.50021493e-01 -7.55572021e-02 3.74224596e-02
-6.03695512e-01 5.86438477e-01 1.05370045e+00 1.06014855e-01
1.85977239e-02 7.10725188e-02 -4.44991410e-01 4.91814315e-01
6.70856953e-01 -4.86069322e-01 -1.56172678e-01 -1.39380374e-03
3.48400861e-01 3.51540267e-01 4.30999666e-01 -1.08356762e+00
2.91600138e-01 -3.14075708e-01 1.96785152e-01 -5.34445226e-01
6.97523236e-01 -2.32701644e-01 7.77452290e-02 3.60053927e-01
-4.22975831e-02 -6.71533763e-01 3.87131497e-02 5.76794028e-01
1.52651832e-01 -7.82193661e-01 1.11320889e+00 -1.97439834e-01
-6.72559083e-01 7.08120614e-02 -4.56813693e-01 -8.68165419e-02
6.16663158e-01 1.50807112e-01 -2.57819414e-01 -5.52155137e-01
-7.84298480e-01 -2.73333490e-01 5.08552969e-01 -4.21713114e-01
2.25825936e-01 -9.95592117e-01 -3.10521811e-01 3.67596298e-01
-1.13556860e-02 -4.82208669e-01 1.15659960e-01 1.09267139e+00
-3.97929013e-01 8.06820631e-01 -5.93526661e-02 -8.52522850e-01
-5.51931024e-01 2.72415876e-01 3.02932918e-01 -1.74876284e-02
-3.65301788e-01 6.59409523e-01 -4.43356276e-01 -6.67781830e-01
-1.59586355e-01 -2.80515581e-01 6.26979232e-01 -5.16774237e-01
4.01400894e-01 2.25401103e-01 8.68028104e-02 -3.44642341e-01
-2.10801810e-01 3.18098426e-01 -3.81457359e-02 -4.78694320e-01
9.89495218e-01 -1.59410417e-01 -2.28677884e-01 5.57300627e-01
8.92764866e-01 2.01803088e-01 -1.06278884e+00 -4.63262171e-01
2.13717073e-01 -2.04860240e-01 5.51121652e-01 -6.38519585e-01
-1.60167843e-01 1.11130905e+00 5.45065820e-01 2.73740709e-01
3.76852900e-01 9.94700193e-02 5.25057852e-01 1.10259962e+00
1.13815749e+00 -1.11203825e+00 -1.78013727e-01 6.67371333e-01
-9.67931151e-02 -1.35639393e+00 4.44006354e-01 -1.17815107e-01
-8.81199390e-02 1.14064121e+00 1.58255294e-01 1.01038396e-01
6.97625816e-01 2.12114155e-01 -2.93323100e-01 -1.81194633e-01
-4.77564931e-01 -1.51943684e-01 1.56298250e-01 4.84253198e-01
5.16542792e-01 3.69967967e-01 -1.29792750e-01 -6.89096153e-02
-2.52656162e-01 -1.16520815e-01 8.98571551e-01 9.58156407e-01
-7.79347420e-01 -1.52141261e+00 -3.82118493e-01 5.32960951e-01
-3.29588234e-01 -2.39796087e-01 3.23251814e-01 5.06083786e-01
-1.96661785e-01 9.89929020e-01 -1.65656731e-01 -1.89803362e-01
-2.74885625e-01 3.22823077e-01 1.43832707e+00 -8.10186207e-01
3.66859734e-01 1.59941241e-01 3.08772445e-01 -4.71461266e-01
-5.53597271e-01 -7.70426631e-01 -1.30773711e+00 -5.24582684e-01
-9.40516591e-01 5.70073545e-01 9.53711927e-01 1.42843354e+00
2.55868644e-01 1.51863515e-01 3.52528006e-01 -9.97024298e-01
-1.34664834e+00 -1.09180200e+00 -6.58419669e-01 -8.45199376e-02
3.05084050e-01 -7.69397438e-01 -3.86924088e-01 -4.18628067e-01]
|
[5.601589202880859, 4.905053615570068]
|
81cd8975-1b31-4c7a-a719-acdac5853836
|
counter-hypothetical-particle-filters-for
|
2305.17828
| null |
https://arxiv.org/abs/2305.17828v1
|
https://arxiv.org/pdf/2305.17828v1.pdf
|
Counter-Hypothetical Particle Filters for Single Object Pose Tracking
|
Particle filtering is a common technique for six degree of freedom (6D) pose estimation due to its ability to tractably represent belief over object pose. However, the particle filter is prone to particle deprivation due to the high-dimensional nature of 6D pose. When particle deprivation occurs, it can cause mode collapse of the underlying belief distribution during importance sampling. If the region surrounding the true state suffers from mode collapse, recovering its belief is challenging since the area is no longer represented in the probability mass formed by the particles. Previous methods mitigate this problem by randomizing and resetting particles in the belief distribution, but determining the frequency of reinvigoration has relied on hand-tuning abstract heuristics. In this paper, we estimate the necessary reinvigoration rate at each time step by introducing a Counter-Hypothetical likelihood function, which is used alongside the standard likelihood. Inspired by the notions of plausibility and implausibility from Evidential Reasoning, the addition of our Counter-Hypothetical likelihood function assigns a level of doubt to each particle. The competing cumulative values of confidence and doubt across the particle set are used to estimate the level of failure within the filter, in order to determine the portion of particles to be reinvigorated. We demonstrate the effectiveness of our method on the rigid body object 6D pose tracking task.
|
['Odest Chadwicke Jenkins', 'Jasmine A. Berry', 'Jana Pavlasek', 'Elizabeth A. Olson']
|
2023-05-28
| null | null | null | null |
['pose-tracking', 'pose-estimation', '6d-pose-estimation-1']
|
['computer-vision', 'computer-vision', 'computer-vision']
|
[ 1.39470264e-01 2.75351733e-01 -1.81923844e-02 3.47768664e-02
-4.62328792e-01 -6.43445909e-01 7.87256360e-01 3.92911702e-01
-4.38449144e-01 8.60722721e-01 2.63202846e-01 -3.58049184e-01
-4.10388112e-01 -8.30174387e-01 -8.89813721e-01 -7.14506686e-01
4.21202034e-02 7.94289410e-01 4.55630422e-01 1.54227093e-01
4.34869587e-01 6.43629313e-01 -1.66336608e+00 -1.21510223e-01
8.35814416e-01 8.25038850e-01 3.81216526e-01 5.35397530e-01
1.85047552e-01 5.77429593e-01 -9.15578842e-01 -3.66044164e-01
1.92789093e-01 -2.16334224e-01 -4.17341352e-01 7.18535334e-02
3.86934042e-01 -4.41317409e-01 1.48409128e-01 1.24883485e+00
1.40454724e-01 4.89559442e-01 8.73593569e-01 -1.11647344e+00
-2.43395120e-01 4.44718540e-01 -1.67106107e-01 1.86904386e-01
6.41366661e-01 1.47230715e-01 8.18659067e-01 -9.03086424e-01
4.18709040e-01 1.30804002e+00 6.36573493e-01 3.11506569e-01
-1.29813719e+00 -4.43758011e-01 1.91100121e-01 -6.46408796e-02
-1.26790750e+00 -2.15344161e-01 7.94047058e-01 -5.17433822e-01
7.16773629e-01 5.01702607e-01 1.12339163e+00 7.59037673e-01
6.01802289e-01 4.61761713e-01 1.18242288e+00 -3.45295101e-01
6.45437062e-01 1.08279347e-01 -1.50690764e-01 6.14057958e-01
7.05916345e-01 3.74622107e-01 -5.86793005e-01 -6.71901405e-01
8.77296805e-01 -1.19975589e-01 -1.98732480e-01 -3.80282074e-01
-9.78345156e-01 7.45145679e-01 3.19575787e-01 -1.61986932e-01
-4.56795305e-01 7.92184025e-02 -2.18572482e-01 -3.00291836e-01
3.51014227e-01 5.88119745e-01 -2.57136166e-01 -4.02916931e-02
-9.32623863e-01 6.44187391e-01 9.39082682e-01 4.46548373e-01
4.46984738e-01 -3.80724192e-01 -2.72279203e-01 3.19347143e-01
7.63442516e-01 5.20461917e-01 7.70502239e-02 -9.21533942e-01
8.63229483e-03 5.38750291e-01 6.15470946e-01 -1.03496563e+00
-2.06272244e-01 -5.26625216e-01 -4.49053705e-01 8.08685780e-01
5.90665579e-01 4.55514975e-02 -1.11510432e+00 1.77888775e+00
7.38519430e-01 3.20101976e-01 -2.72467107e-01 9.66665387e-01
1.02227531e-01 3.84822726e-01 2.33832337e-02 -3.23267758e-01
1.39312184e+00 -3.61168608e-02 -5.32485247e-01 -2.25864351e-01
-1.64743915e-01 -7.70699322e-01 6.84751272e-01 5.12447536e-01
-1.16280842e+00 -4.05323237e-01 -1.16961324e+00 3.09891105e-01
1.54813714e-02 -5.69938064e-01 6.33276105e-01 5.96219897e-01
-6.19667113e-01 7.17916250e-01 -1.02651751e+00 1.13576017e-01
2.33400643e-01 3.96280974e-01 -1.45158526e-02 1.82366326e-01
-1.16956019e+00 1.48786330e+00 4.28178191e-01 1.99003354e-01
-8.01412165e-01 -7.19961405e-01 -5.93436003e-01 -2.01633126e-01
5.29879749e-01 -8.21343005e-01 9.20547545e-01 3.00932936e-02
-1.27322316e+00 4.40647542e-01 -1.37533739e-01 -5.25317371e-01
7.06784904e-01 -1.65917009e-01 -1.02398396e-01 -1.23420440e-01
6.60322458e-02 4.78131562e-01 1.21061695e+00 -1.49478924e+00
-6.40921593e-01 -3.83154035e-01 3.11187804e-01 4.07187045e-01
3.80906314e-01 -5.84362447e-01 -2.82851607e-01 -4.45742518e-01
6.15825891e-01 -1.08110535e+00 -4.11894649e-01 -5.28008975e-02
-3.91006619e-01 -3.55109781e-01 4.11115319e-01 -3.10363710e-01
1.08602428e+00 -2.03438640e+00 5.16981967e-02 4.59774971e-01
4.43222463e-01 -1.62618339e-01 4.72920895e-01 5.89072704e-02
3.78692240e-01 -2.81179305e-02 -2.04600915e-01 5.32780513e-02
1.58091486e-01 2.11822823e-01 -3.07354569e-01 7.05963373e-01
1.39716730e-01 3.27607721e-01 -1.09762621e+00 -5.11428356e-01
4.79628205e-01 5.03034532e-01 -7.12170720e-01 -5.87426648e-02
-2.93408841e-01 1.53697878e-01 -5.71741104e-01 4.05583978e-01
8.48596692e-01 -1.88147068e-01 -4.39379625e-02 -4.69356626e-01
-9.23682675e-02 4.18490648e-01 -1.47880375e+00 1.15205467e+00
-7.09392950e-02 -8.99386555e-02 -4.84216772e-02 -2.96920717e-01
7.93644130e-01 3.60919945e-02 4.21855927e-01 -4.89893854e-02
1.83159947e-01 1.04124904e-01 4.73644845e-02 -4.63231280e-02
6.99599862e-01 -6.45290554e-01 -1.55397341e-01 3.74805540e-01
-2.13363186e-01 -7.14090586e-01 1.25694200e-02 1.08746812e-01
1.05289114e+00 1.26809850e-01 3.14843416e-01 -3.84419024e-01
3.14798117e-01 1.19484633e-01 5.51536143e-01 1.02623844e+00
-2.44675025e-01 5.49967170e-01 2.79553711e-01 -2.10123822e-01
-8.42760444e-01 -1.59703958e+00 -5.96986055e-01 4.32055354e-01
4.81851399e-01 -1.19183667e-01 -4.32242155e-01 -6.34110928e-01
4.13165390e-01 1.18713355e+00 -6.99254096e-01 -4.09618556e-01
-3.72873336e-01 -8.17394137e-01 9.94570330e-02 2.26150692e-01
7.56612048e-03 -8.13890278e-01 -1.10633588e+00 3.48415554e-01
-1.17860578e-01 -5.38583815e-01 -1.20490342e-01 3.32368433e-01
-8.49775016e-01 -1.18221128e+00 -3.56538415e-01 -1.22612201e-01
9.42320585e-01 -7.45004937e-02 1.09891653e+00 2.25383759e-01
-1.60470475e-02 2.16410860e-01 -1.04370207e-01 -4.10173923e-01
-5.03838360e-01 -5.82720518e-01 2.83224016e-01 -5.08653104e-01
1.29028007e-01 -2.16972470e-01 -6.87618971e-01 3.37310642e-01
-6.20897412e-01 -2.77048349e-02 2.41215959e-01 7.90711761e-01
7.83465385e-01 7.01859176e-01 8.32058489e-02 -5.68606257e-01
7.64086902e-01 -4.18909013e-01 -7.60103941e-01 -1.90551039e-02
-4.50344235e-01 2.61774808e-01 -4.34306711e-02 -7.83119261e-01
-1.13755715e+00 -5.89564219e-02 -6.20317683e-02 -2.91640162e-01
8.74787495e-02 5.45801282e-01 1.63612589e-02 2.09660530e-01
6.29850447e-01 -2.21200958e-01 7.58659989e-02 -2.67489791e-01
3.56454551e-01 1.66001886e-01 6.24532938e-01 -7.18992531e-01
7.71935701e-01 6.13315105e-01 1.32067323e-01 -3.30869675e-01
-1.01440310e+00 -5.73291667e-02 -4.42412108e-01 -5.16571462e-01
7.16825783e-01 -6.08890533e-01 -1.02024996e+00 3.49451415e-02
-9.76524889e-01 1.23505414e-01 -7.04950511e-01 7.06310451e-01
-5.75156689e-01 3.16153258e-01 -9.88075584e-02 -1.14487982e+00
1.11421458e-01 -1.16567385e+00 9.29257810e-01 2.15653539e-01
-5.90224087e-01 -7.67698944e-01 5.84119139e-03 1.99688241e-01
1.45760924e-02 3.67980987e-01 8.48303318e-01 -3.16105843e-01
-4.98519540e-01 -6.05155110e-01 2.21635416e-01 7.36342147e-02
1.99019108e-02 -1.65979087e-01 -5.56189299e-01 -3.11733991e-01
6.31137848e-01 1.58624258e-02 6.50273144e-01 6.64881349e-01
6.84533119e-01 -2.25035623e-01 -4.21783000e-01 -4.15498279e-02
1.22478604e+00 1.61879003e-01 5.69387019e-01 3.38254839e-01
2.33666807e-01 5.40444195e-01 6.94928944e-01 5.74540913e-01
1.35422677e-01 5.40303826e-01 7.87381053e-01 4.24635708e-01
-1.17992446e-01 -4.12885398e-01 1.73742220e-01 3.27598006e-01
-7.30860084e-02 -3.47074240e-01 -7.20978260e-01 3.77964497e-01
-1.64667690e+00 -9.45237935e-01 -1.62009001e-02 2.53604984e+00
8.30654740e-01 9.34076607e-01 -8.12196508e-02 1.81139186e-01
7.79390514e-01 -6.21988811e-02 -6.91555798e-01 2.68677194e-02
1.07868046e-01 -1.67309642e-01 4.71500784e-01 8.54574740e-01
-7.41607070e-01 5.39154649e-01 6.67485046e+00 8.41199577e-01
-4.81926262e-01 -4.80994545e-02 1.22698501e-01 -1.51825145e-01
-4.08076704e-01 3.04171920e-01 -1.07835841e+00 7.12624609e-01
5.08855343e-01 -1.64831430e-01 2.60072052e-01 6.22866035e-01
-5.97943813e-02 -7.40922034e-01 -1.04301071e+00 4.65896040e-01
-1.02972621e-02 -1.13175249e+00 -2.41325181e-02 1.12024322e-01
4.21060145e-01 -3.28551531e-01 2.23520231e-02 -6.92379624e-02
5.58034539e-01 -7.84582376e-01 1.23204851e+00 6.87564373e-01
4.31427687e-01 -6.74633503e-01 6.54706001e-01 5.48658550e-01
-8.52028668e-01 7.53592998e-02 -4.75296080e-01 -2.86038846e-01
6.10213339e-01 1.09445345e+00 -1.24096823e+00 5.23456335e-02
3.47202271e-01 -7.26775080e-02 -2.38562569e-01 1.05938244e+00
-3.86482120e-01 3.57647508e-01 -9.35891986e-01 -1.57772645e-01
-1.97566271e-01 -8.67146552e-02 1.09114754e+00 4.97833937e-01
2.91032881e-01 1.23650499e-01 1.75687239e-01 1.16571355e+00
4.09021318e-01 -7.22008467e-01 -1.24268152e-01 2.05796957e-01
8.75250995e-01 7.28534579e-01 -1.05622733e+00 -3.34763646e-01
3.79826203e-02 5.22313654e-01 3.67846936e-02 -1.64008886e-02
-9.24449503e-01 6.88319281e-02 5.66154003e-01 2.88256228e-01
2.45622680e-01 -2.11911455e-01 -4.72269028e-01 -8.06739569e-01
-7.62386620e-02 -4.85697985e-01 2.74138898e-01 -8.10096383e-01
-1.45376790e+00 2.73026675e-01 4.79543418e-01 -1.14523268e+00
-2.69170970e-01 -5.22025764e-01 -3.36691916e-01 1.01401496e+00
-7.98541784e-01 -6.93088114e-01 9.98249799e-02 3.40961844e-01
4.14354146e-01 3.63106340e-01 5.84076285e-01 -1.94703847e-01
-4.87846397e-02 1.42970383e-01 -4.19662863e-01 -3.96554828e-01
3.03736985e-01 -1.28269529e+00 2.23616496e-01 7.33989835e-01
-2.02499330e-01 8.61077428e-01 1.41486728e+00 -1.42186320e+00
-1.42317474e+00 -6.33165658e-01 7.23644257e-01 -9.49915111e-01
7.45476127e-01 -3.22792351e-01 -8.64202917e-01 4.01826143e-01
-5.68084776e-01 -1.64739802e-01 2.30857536e-01 2.71054178e-01
-4.71745210e-04 4.18045670e-01 -1.43799734e+00 7.29722142e-01
8.13740373e-01 -3.33465219e-01 -1.18746066e+00 2.57427692e-01
4.99136835e-01 -6.74534976e-01 -9.75074232e-01 6.92451298e-01
7.35989094e-01 -8.12495232e-01 1.33424556e+00 -6.16519451e-02
1.33849466e-02 -7.32274652e-01 -1.90302685e-01 -1.04233861e+00
-7.13319838e-01 -3.69610548e-01 -4.66225058e-01 9.29384708e-01
2.72161335e-01 -4.50045586e-01 9.75078404e-01 7.58492887e-01
-3.77156958e-02 -6.84454322e-01 -1.26192522e+00 -7.41267741e-01
-1.19094603e-01 -6.37946486e-01 6.35053039e-01 4.55743730e-01
-1.87755600e-01 2.11700588e-01 -2.01417863e-01 5.22532642e-01
1.04867077e+00 1.78248752e-02 4.01596755e-01 -1.54601038e+00
-3.66000265e-01 -4.18563187e-01 -2.60886341e-01 -1.02289939e+00
-2.84729451e-01 -5.30413151e-01 4.61308181e-01 -1.46858084e+00
2.08224624e-01 -6.00507677e-01 -2.57905900e-01 1.96907878e-01
-3.08667868e-01 1.20123960e-01 3.53707403e-01 3.18242133e-01
-3.68402123e-01 3.90936971e-01 1.40960598e+00 9.04128551e-02
-2.57161677e-01 1.31517217e-01 -5.66124439e-01 1.12696970e+00
5.24760127e-01 -6.77257657e-01 -2.63238490e-01 3.25741954e-02
5.86932361e-01 2.68798023e-01 7.82397509e-01 -9.97073770e-01
2.23502234e-01 -2.21603706e-01 8.07269454e-01 -1.17832732e+00
5.51846147e-01 -8.92919660e-01 6.35265648e-01 7.03718007e-01
-7.63167441e-02 -1.14787944e-01 2.49358304e-02 9.45110738e-01
2.98478216e-01 -4.65913326e-01 6.83556437e-01 -2.39177570e-01
-3.30498189e-01 -1.07168771e-01 -4.72539306e-01 -2.02702582e-01
7.78705120e-01 -5.32185972e-01 -8.87305811e-02 4.70548868e-02
-8.74219656e-01 7.03230277e-02 6.07489526e-01 2.49113217e-01
7.46875286e-01 -1.23377860e+00 -4.72978681e-01 2.65249610e-01
-2.47083649e-01 1.54094636e-01 1.17726259e-01 7.28351772e-01
-1.73614472e-01 -2.63664667e-02 5.58708087e-02 -7.26245403e-01
-8.40144396e-01 4.47108805e-01 6.85306713e-02 -1.14835627e-01
-7.41551280e-01 9.71016705e-01 4.45731776e-03 -7.01275766e-02
1.45577282e-01 -5.17401457e-01 -4.78119440e-02 -2.44731400e-02
3.44513685e-01 6.16645634e-01 -6.80745095e-02 -5.38284779e-01
-5.04244804e-01 2.99337357e-01 3.65335457e-02 -4.21535462e-01
9.50977266e-01 -2.05617130e-01 -5.45067415e-02 5.32562733e-01
5.34161329e-01 2.67986178e-01 -1.70403516e+00 3.66987765e-01
-1.93359181e-01 -6.69025958e-01 1.48644730e-01 -1.00246799e+00
-2.01493949e-01 2.15235755e-01 4.43250954e-01 3.47794712e-01
6.69471443e-01 2.19049618e-01 4.91965950e-01 5.11727296e-02
6.76819563e-01 -9.42584395e-01 -1.51143581e-01 4.48053837e-01
8.99561286e-01 -1.01493895e+00 5.40452898e-01 -4.39888746e-01
-3.00896198e-01 6.55540407e-01 4.16487902e-01 -3.45064312e-01
8.44310999e-01 5.19075394e-01 -4.15964097e-01 -3.83982718e-01
-5.01819074e-01 3.66236307e-02 4.79265571e-01 4.88075256e-01
-7.68977553e-02 2.52335459e-01 -3.63784343e-01 4.63590920e-01
-6.45056725e-01 -1.42687500e-01 3.14168990e-01 1.11650670e+00
-7.21074402e-01 -7.18963981e-01 -9.19393480e-01 6.69677198e-01
-2.65913874e-01 1.94849432e-01 -5.48582524e-02 6.71807349e-01
4.68518525e-01 9.37244117e-01 2.55062491e-01 -2.70696014e-01
3.19077313e-01 -2.03651309e-01 9.08224702e-01 -5.52644432e-01
-1.75004661e-01 3.15671593e-01 1.78219490e-02 -2.44657472e-01
-2.12109312e-01 -8.98685277e-01 -1.41791785e+00 -1.40008733e-01
-8.96867096e-01 3.39917570e-01 6.79925978e-01 1.06853616e+00
-7.60100111e-02 3.69829327e-01 1.65773798e-02 -1.13033843e+00
-8.04435790e-01 -7.29710162e-01 -4.21257317e-01 5.22599638e-01
4.62401658e-01 -1.34459651e+00 -7.50602663e-01 -2.68233567e-01]
|
[7.35066556930542, -1.0441350936889648]
|
b4483be9-419e-4167-b1fc-74decbcbcbd3
|
survaival-survival-analysis-with-the-eyes-of
|
2305.18222
| null |
https://arxiv.org/abs/2305.18222v1
|
https://arxiv.org/pdf/2305.18222v1.pdf
|
survAIval: Survival Analysis with the Eyes of AI
|
In this study, we propose a novel approach to enrich the training data for automated driving by using a self-designed driving simulator and two human drivers to generate safety-critical corner cases in a short period of time, as already presented in~\cite{kowol22simulator}. Our results show that incorporating these corner cases during training improves the recognition of corner cases during testing, even though, they were recorded due to visual impairment. Using the corner case triggering pipeline developed in the previous work, we investigate the effectiveness of using expert models to overcome the domain gap due to different weather conditions and times of day, compared to a universal model from a development perspective. Our study reveals that expert models can provide significant benefits in terms of performance and efficiency, and can reduce the time and effort required for model training. Our results contribute to the progress of automated driving, providing a pathway for safer and more reliable autonomous vehicles on the road in the future.
|
['Hanno Gottschalk', 'Stefan Bracke', 'Kamil Kowol']
|
2023-05-23
| null | null | null | null |
['autonomous-vehicles', 'survival-analysis']
|
['computer-vision', 'miscellaneous']
|
[-1.05056964e-01 1.69418350e-01 2.22394578e-02 -5.92015028e-01
-2.54309773e-01 -5.71805656e-01 4.29186434e-01 9.51782539e-02
-4.43344921e-01 5.63778877e-01 -2.98371404e-01 -8.41753721e-01
-2.56843239e-01 -7.86400914e-01 -7.68693686e-01 -9.66996774e-02
2.16047361e-01 3.41862708e-01 5.53760469e-01 -7.20744669e-01
2.72048742e-01 7.25570202e-01 -2.52915025e+00 6.46893680e-02
1.26936400e+00 5.85312188e-01 3.71505588e-01 5.01170635e-01
2.28896752e-01 3.92162532e-01 -6.35518551e-01 -2.60880768e-01
3.55428278e-01 6.47699879e-03 -2.17347383e-01 -6.97170198e-02
3.63641441e-01 -4.49720055e-01 -4.33670044e-01 5.56797981e-01
5.90369880e-01 9.86786261e-02 3.85346889e-01 -1.73701394e+00
9.25980434e-02 5.05196825e-02 1.00717016e-01 8.28722790e-02
2.00189963e-01 6.51411116e-01 2.05135360e-01 -5.67203879e-01
6.65798068e-01 7.25031912e-01 7.60288715e-01 3.97449106e-01
-9.75805342e-01 -1.09206593e+00 -2.78864279e-02 7.65320301e-01
-1.50121188e+00 -6.08314097e-01 6.83815181e-01 -6.79325581e-01
1.03843319e+00 -7.43310824e-02 6.42399728e-01 1.08210671e+00
1.81182340e-01 5.74056089e-01 9.61434066e-01 -3.38686079e-01
3.17774653e-01 6.52829885e-01 1.58863023e-01 4.19353664e-01
3.65331203e-01 7.54545033e-01 -3.42938095e-01 2.56402820e-01
-4.32876758e-02 -3.71559024e-01 1.52394831e-01 -2.65359223e-01
-7.23218441e-01 6.93506062e-01 -1.17118396e-01 2.21315667e-01
-3.56736839e-01 -1.36858851e-01 3.12042743e-01 3.37023258e-01
1.13528408e-01 5.54231167e-01 -4.02667582e-01 -5.96446991e-01
-9.11635816e-01 7.00981200e-01 8.35794508e-01 1.15068734e+00
1.19516277e+00 1.48198739e-01 -2.59746999e-01 3.88910413e-01
9.85493883e-02 8.03342462e-01 2.54841857e-02 -1.00483727e+00
3.72533590e-01 8.50230217e-01 4.84399498e-01 -8.05087030e-01
-6.55138135e-01 -4.14502323e-01 1.13819323e-01 4.95602667e-01
9.97688025e-02 -5.04328787e-01 -9.81019735e-01 1.50095296e+00
8.28547776e-02 5.20615518e-01 1.75018951e-01 7.28951752e-01
6.23536885e-01 2.89184600e-01 6.64767027e-02 2.89549291e-01
9.99466181e-01 -6.57565236e-01 -8.07552397e-01 -5.55668712e-01
1.14313710e+00 -6.05806828e-01 9.70831513e-01 5.96084774e-01
-7.11841702e-01 -9.66371298e-01 -1.38642001e+00 3.90763134e-01
-7.64408469e-01 4.86170232e-01 4.10346746e-01 1.07951474e+00
-9.63088751e-01 4.66011018e-01 -4.78678763e-01 -4.15215343e-01
1.58200324e-01 3.21299076e-01 -2.56410599e-01 -2.11676240e-01
-1.51139295e+00 1.41645598e+00 2.73726672e-01 6.04559407e-02
-1.00185490e+00 -7.54578054e-01 -1.01646960e+00 -3.45777929e-01
3.50632966e-01 -1.29342765e-01 1.32053506e+00 -4.09170419e-01
-1.29738081e+00 4.67254192e-01 -3.27289045e-01 -6.79153025e-01
6.12746716e-01 -3.33139062e-01 -9.44995165e-01 -1.69270858e-01
1.64984643e-01 7.12133825e-01 3.80894244e-01 -1.21875668e+00
-8.92493129e-01 7.07841814e-02 9.26962271e-02 -1.10559396e-01
-3.97761017e-02 -8.01943541e-02 -4.85329717e-01 5.19890115e-02
-7.73503900e-01 -1.20999956e+00 -3.56793433e-01 -4.69487697e-01
1.06616907e-01 -9.59615037e-02 1.10378408e+00 -8.08576345e-01
1.49645996e+00 -2.18572474e+00 -4.02661234e-01 4.60771173e-01
-1.88515633e-01 9.70083535e-01 -7.64510483e-02 5.08932412e-01
1.32167637e-01 2.61900593e-02 -1.71475708e-01 -1.25722572e-01
8.18223506e-02 3.87697756e-01 -1.86349675e-01 1.16468213e-01
5.38603842e-01 6.75497651e-01 -8.48399401e-01 -2.51994818e-01
7.33286440e-01 4.42195147e-01 -3.60656410e-01 2.11663112e-01
-5.96756414e-02 3.93948168e-01 -1.78714171e-01 2.01189488e-01
7.71229386e-01 7.05553293e-01 -2.42290571e-01 2.00123608e-01
-5.60857534e-01 2.83296585e-01 -1.33877611e+00 1.40784824e+00
-6.84960783e-01 9.05163407e-01 -1.88110098e-01 -7.04860806e-01
1.19748271e+00 3.05090934e-01 7.28700683e-02 -1.15155876e+00
3.59683096e-01 2.61327088e-01 2.11320847e-01 -8.74851167e-01
6.42971516e-01 3.56067598e-01 -1.05056360e-01 1.22242451e-01
-2.80095696e-01 -3.66645664e-01 6.95862591e-01 -9.29490998e-02
9.49149728e-01 4.30851638e-01 -2.37240195e-01 -3.21587711e-03
6.09669209e-01 5.33065557e-01 4.99296784e-01 6.23425841e-01
-4.94762450e-01 1.23426564e-01 2.82974362e-01 -3.26999098e-01
-9.59893525e-01 -6.58005536e-01 -9.42023247e-02 6.34763181e-01
2.67361283e-01 -4.44740146e-01 -7.90091336e-01 -3.73782575e-01
2.18695462e-01 1.48292279e+00 -3.73964071e-01 -5.81433535e-01
-4.58756208e-01 -1.98868841e-01 8.85428607e-01 6.45211875e-01
4.54242200e-01 -9.22115326e-01 -1.07831931e+00 2.34790906e-01
5.17622046e-02 -1.46941531e+00 1.40614361e-01 7.39659593e-02
-3.90908003e-01 -1.00971639e+00 -5.45192435e-02 -5.92077672e-01
4.65844363e-01 3.44723672e-01 8.51507843e-01 1.72559351e-01
-3.78720075e-01 1.65296867e-01 -3.54562998e-01 -1.14386344e+00
-7.64600515e-01 6.29865304e-02 3.97756025e-02 -3.18934590e-01
8.41686487e-01 -1.69366732e-01 -5.56476295e-01 9.28151131e-01
-6.37409210e-01 5.69869541e-02 6.06325626e-01 2.31944904e-01
1.15641132e-01 2.68350989e-01 9.00323808e-01 -6.09093368e-01
7.14827061e-01 -6.03298247e-01 -9.01360393e-01 -1.70773489e-03
-9.85744476e-01 2.44962126e-02 5.16395509e-01 -1.25342131e-01
-1.05168962e+00 1.11286163e-01 -2.58426279e-01 -1.91302687e-01
-7.64650106e-01 4.37667370e-01 -8.48331526e-02 -2.16010913e-01
8.64203036e-01 -6.04411215e-02 2.33896941e-01 -1.20055415e-01
1.30648553e-01 8.81030738e-01 3.54078889e-01 -2.77705014e-01
9.37015891e-01 1.29992276e-01 6.42719492e-02 -6.18409872e-01
-5.67007475e-02 -5.89257538e-01 -4.99583900e-01 -6.29616380e-01
6.37355030e-01 -1.03343415e+00 -5.84404051e-01 4.08359379e-01
-8.31805110e-01 -5.11114240e-01 4.54438776e-02 4.92586941e-01
-2.64009535e-01 1.06479429e-01 3.39838833e-01 -8.77455652e-01
1.82092518e-01 -1.43035376e+00 7.51037896e-01 3.08125675e-01
-3.44552249e-01 -8.19127798e-01 -2.31836196e-02 4.87071902e-01
6.14253163e-01 1.68504164e-01 4.17868167e-01 -4.85846311e-01
-3.03915441e-01 -6.35378301e-01 -8.58002901e-02 5.48938751e-01
-1.90454692e-01 1.67589918e-01 -1.01823151e+00 1.86211092e-03
-5.40872395e-01 9.46943089e-02 4.24161971e-01 -1.52312443e-01
7.58534670e-01 4.66836452e-01 -4.49978501e-01 1.26981467e-01
1.11469162e+00 6.61258638e-01 8.45629811e-01 5.02373040e-01
3.03578556e-01 9.42642450e-01 1.21289432e+00 2.15382367e-01
7.63387620e-01 7.20620513e-01 4.06363726e-01 -2.44126767e-01
-1.43095180e-01 -1.37422130e-01 3.87627691e-01 -2.27737110e-02
-9.22529120e-03 -1.29352594e-02 -1.29899442e+00 9.79156077e-01
-1.81007743e+00 -8.32136333e-01 -4.96035963e-01 2.35416865e+00
3.18047434e-01 3.41821045e-01 1.68584675e-01 3.85845870e-01
5.59522450e-01 -4.87648278e-01 -4.28002328e-01 -7.78832018e-01
8.04495886e-02 3.28455627e-01 9.07070279e-01 4.48394686e-01
-8.04470658e-01 9.09550190e-01 6.90515280e+00 4.34963405e-01
-1.31286192e+00 -6.09727297e-03 5.02088405e-02 5.88866584e-02
-1.62485257e-01 1.54795006e-01 -1.05529177e+00 4.71470296e-01
1.58113682e+00 -1.67328358e-01 2.52030343e-01 9.75201428e-01
7.60891259e-01 -3.07771802e-01 -1.00248528e+00 5.16264737e-01
1.18812084e-01 -1.07890558e+00 -5.73887706e-01 6.57992363e-02
4.97542292e-01 9.52322409e-02 -5.90901114e-02 8.32841158e-01
1.22969471e-01 -8.98608923e-01 7.00530767e-01 5.11491060e-01
4.65775132e-01 -9.95519698e-01 1.05382490e+00 4.52684730e-01
-9.39482033e-01 -2.25626677e-01 -2.22975994e-03 -3.86617452e-01
1.25321373e-01 2.51785725e-01 -1.54955387e+00 7.20833659e-01
5.56108654e-01 3.74086380e-01 -9.61079299e-01 1.27318025e+00
-3.43158305e-01 7.66231060e-01 -2.32781529e-01 -4.07239348e-02
1.24144569e-01 -1.44390212e-02 4.28286165e-01 1.17518556e+00
4.25366879e-01 -4.26560313e-01 -1.09761067e-01 7.38100410e-01
6.73431814e-01 -1.25172764e-01 -9.77568388e-01 2.30571210e-01
5.94659388e-01 1.10887754e+00 -2.72948653e-01 -1.42151341e-01
-4.47847605e-01 4.33826298e-01 -2.28283808e-01 4.26895469e-01
-1.18826628e+00 -8.59813035e-01 6.64147556e-01 5.31817555e-01
3.02790999e-01 -2.98943073e-01 -4.38813627e-01 -3.71839434e-01
8.91970471e-02 -7.67082274e-01 -2.14370921e-01 -9.23196733e-01
-4.21242803e-01 5.53639770e-01 4.90038753e-01 -1.48452425e+00
-5.46879590e-01 -7.57561147e-01 -6.64013565e-01 9.93008614e-01
-1.78332269e+00 -1.01847005e+00 -6.77808642e-01 2.55126745e-01
2.68303424e-01 -3.08090895e-01 5.55483460e-01 6.98346376e-01
-6.16902769e-01 6.75134063e-01 -3.67748648e-01 -4.41806704e-01
8.51259410e-01 -7.84362495e-01 6.87828362e-01 1.05848515e+00
-5.49058974e-01 4.07034636e-01 1.15480030e+00 -7.08650947e-01
-1.26198792e+00 -1.15127182e+00 9.45402324e-01 -6.14444375e-01
4.25808042e-01 -3.86780053e-01 -7.68744648e-01 3.81911725e-01
1.62868947e-01 -4.71363395e-01 6.57077968e-01 -3.85614932e-02
5.20629436e-02 -3.46641749e-01 -1.20682836e+00 6.21977746e-01
9.24782813e-01 -4.30295885e-01 -6.13264322e-01 -1.80826724e-01
5.59670389e-01 -3.82332295e-01 -6.01244569e-01 6.17801785e-01
5.24726570e-01 -9.58538890e-01 4.51376110e-01 -3.19058210e-01
8.40870589e-02 -6.61222041e-01 1.76794276e-01 -1.50420475e+00
4.54215519e-02 -5.95787883e-01 1.81643277e-01 9.88784373e-01
8.04038703e-01 -7.75415063e-01 6.76934123e-01 9.17164743e-01
-7.91864216e-01 -5.31452835e-01 -8.78013432e-01 -8.81520867e-01
-1.20366998e-01 -1.09869182e+00 9.05241311e-01 3.51230204e-01
-1.73624590e-01 -7.69962668e-02 -2.11772755e-01 3.66014451e-01
1.43072173e-01 -6.14298463e-01 1.29637396e+00 -1.41546309e+00
1.79292083e-01 -2.31882349e-01 -8.16891670e-01 -4.02752936e-01
3.84899557e-01 -4.63213652e-01 2.56112695e-01 -1.62745273e+00
-4.51319456e-01 -4.67688948e-01 -1.29927650e-01 4.57166970e-01
-8.11912939e-02 5.49834184e-02 9.63935256e-02 -4.19296026e-01
-2.92195678e-01 6.08726516e-02 8.99673641e-01 1.21761113e-01
-3.30419809e-01 1.16537891e-01 -7.45620191e-01 4.61273432e-01
8.30572486e-01 -2.16889724e-01 -7.19761550e-01 -2.94783831e-01
3.28029782e-01 -2.93387741e-01 4.44550812e-01 -1.45325637e+00
2.70763129e-01 -1.43018454e-01 -6.63636625e-02 -8.29472244e-01
2.63909072e-01 -1.16148961e+00 2.27436647e-01 5.13724506e-01
1.84104800e-01 1.48997143e-01 1.01720381e+00 2.23181039e-01
-1.93862498e-01 -3.02515835e-01 3.37561995e-01 4.40305173e-01
-1.25041306e+00 -2.06550658e-01 -7.99397409e-01 -3.77580881e-01
1.59948874e+00 -5.62531710e-01 -4.07283753e-01 -3.45014602e-01
-3.80585760e-01 7.09078074e-01 5.09049416e-01 8.74471962e-01
4.21116680e-01 -9.18850541e-01 -5.34355223e-01 6.17138028e-01
5.78298211e-01 -3.57037693e-01 3.78678292e-01 7.90192187e-01
-5.29224575e-01 6.35933578e-01 -5.34688175e-01 -5.24654865e-01
-1.13601482e+00 3.07648897e-01 2.95927465e-01 1.29126132e-01
-3.12063754e-01 1.79172769e-01 -3.18153828e-01 -5.68369567e-01
2.68191099e-01 -3.04416597e-01 -3.77427757e-01 -4.01834622e-02
4.09545511e-01 6.93533838e-01 5.98146796e-01 -7.80609310e-01
-4.44608510e-01 4.19833928e-01 3.99104767e-02 -1.92725137e-01
8.73843193e-01 6.66583376e-03 6.49299502e-01 3.19132417e-01
6.21627033e-01 -1.35019213e-01 -1.25700486e+00 3.35517526e-01
-2.83846576e-02 -3.25336426e-01 2.14657590e-01 -1.03384769e+00
-6.43087268e-01 7.87116945e-01 1.15683985e+00 -1.04717039e-01
9.18336868e-01 -4.61656153e-01 7.74886191e-01 4.69599217e-01
4.39633310e-01 -1.54925549e+00 -5.26261806e-01 5.59614062e-01
5.92455387e-01 -1.30039537e+00 -2.95535028e-01 -3.95471305e-01
-1.00409865e+00 8.20478737e-01 8.94690871e-01 1.22431397e-01
5.49797833e-01 1.94526687e-01 3.45450848e-01 -4.54993825e-03
-7.04201877e-01 -5.88756084e-01 1.60603896e-01 1.12422431e+00
4.72490601e-02 9.46966857e-02 -3.53537828e-01 6.58176780e-01
-2.55122870e-01 4.51793343e-01 6.15802646e-01 1.21221256e+00
-4.69719768e-01 -1.27775776e+00 -2.88508862e-01 2.86651641e-01
1.04707547e-01 2.52076447e-01 -3.11675876e-01 1.03448164e+00
7.48088241e-01 1.42983258e+00 8.84310901e-02 -8.81804705e-01
1.15151489e+00 2.70394802e-01 9.73881409e-02 -4.11048055e-01
-5.88026285e-01 -6.56500757e-01 6.67258441e-01 -4.50991899e-01
-1.24045208e-01 -6.47015393e-01 -1.25099289e+00 -2.17815623e-01
-2.08114401e-01 6.35012165e-02 1.18991828e+00 1.00957549e+00
9.96470809e-01 8.01334560e-01 6.74383342e-01 -7.31064796e-01
-2.17375591e-01 -8.12807560e-01 2.66395397e-02 3.29600841e-01
1.65055960e-01 -1.13340437e+00 -3.73819798e-01 -1.33867458e-01]
|
[5.688711643218994, 1.1444401741027832]
|
2799a3c4-dcd3-4e61-840c-cfb135641964
|
variational-distillation-for-multi-view
|
2206.09548
| null |
https://arxiv.org/abs/2206.09548v1
|
https://arxiv.org/pdf/2206.09548v1.pdf
|
Variational Distillation for Multi-View Learning
|
Information Bottleneck (IB) based multi-view learning provides an information theoretic principle for seeking shared information contained in heterogeneous data descriptions. However, its great success is generally attributed to estimate the multivariate mutual information which is intractable when the network becomes complicated. Moreover, the representation learning tradeoff, {\it i.e.}, prediction-compression and sufficiency-consistency tradeoff, makes the IB hard to satisfy both requirements simultaneously. In this paper, we design several variational information bottlenecks to exploit two key characteristics ({\it i.e.}, sufficiency and consistency) for multi-view representation learning. Specifically, we propose a Multi-View Variational Distillation (MV$^2$D) strategy to provide a scalable, flexible and analytical solution to fitting MI by giving arbitrary input of viewpoints but without explicitly estimating it. Under rigorously theoretical guarantee, our approach enables IB to grasp the intrinsic correlation between observations and semantic labels, producing predictive and compact representations naturally. Also, our information-theoretic constraint can effectively neutralize the sensitivity to heterogeneous data by eliminating both task-irrelevant and view-specific information, preventing both tradeoffs in multiple view cases. To verify our theoretically grounded strategies, we apply our approaches to various benchmarks under three different applications. Extensive experiments to quantitatively and qualitatively demonstrate the effectiveness of our approach against state-of-the-art methods.
|
['DaCheng Tao', 'Yuan Xie', 'Zongze Wu', 'Lizhuang Ma', 'Yanyun Qu', 'Wensheng Zhang', 'Cong Wang', 'Zhizhong Zhang', 'Xudong Tian']
|
2022-06-20
| null | null | null | null |
['multi-view-learning']
|
['computer-vision']
|
[ 1.03278197e-01 8.70318711e-02 -5.49992800e-01 -4.24419731e-01
-1.01079500e+00 -5.67569435e-01 2.80668318e-01 -1.53943926e-01
1.26881659e-01 4.07796651e-01 2.75914729e-01 -2.77494062e-02
-5.84059536e-01 -4.49923784e-01 -4.05121088e-01 -8.32746267e-01
7.53673539e-02 4.81356651e-01 -4.63794321e-02 -6.16522878e-03
2.66170323e-01 7.35969841e-02 -1.59696531e+00 3.88150483e-01
8.88244867e-01 1.08708000e+00 4.81231034e-01 2.52546489e-01
-9.99599770e-02 9.41191971e-01 -1.44628063e-01 -4.59286153e-01
3.97434473e-01 -3.98116916e-01 -8.38554800e-01 2.70254284e-01
2.86863744e-01 -2.36824840e-01 -2.63572752e-01 1.26614559e+00
3.78900290e-01 6.54148608e-02 8.52013230e-01 -1.48689771e+00
-9.01513159e-01 6.33476198e-01 -9.75134969e-01 2.10200220e-01
2.43267149e-01 -2.05128521e-01 1.46704245e+00 -9.91748452e-01
7.11284637e-01 1.27407587e+00 3.47853541e-01 4.13711339e-01
-1.23983288e+00 -4.57558215e-01 5.21275878e-01 3.46941650e-01
-1.52884507e+00 -5.39546132e-01 1.04913712e+00 -4.10414100e-01
5.18534541e-01 2.77878284e-01 3.01000386e-01 1.04978073e+00
4.24145535e-02 1.12205231e+00 1.03537500e+00 -1.52268931e-01
2.12847382e-01 5.09764433e-01 1.67811960e-02 6.88615263e-01
2.84642756e-01 -1.75932035e-01 -7.34181881e-01 -2.04079881e-01
7.01693237e-01 2.77713299e-01 -3.76926780e-01 -8.72714460e-01
-1.20579767e+00 8.68677318e-01 2.41222993e-01 1.50076626e-02
-1.50323376e-01 -7.74896294e-02 5.00429809e-01 4.81642783e-01
4.99151289e-01 1.60850585e-01 -5.72553337e-01 6.83428720e-02
-6.40154302e-01 -8.63449648e-02 6.84888601e-01 1.43165445e+00
8.49803507e-01 -1.63076505e-01 -1.46584675e-01 9.25220430e-01
5.73749423e-01 5.03402591e-01 1.54757738e-01 -1.28335238e+00
7.11370528e-01 6.87103391e-01 -1.42716058e-03 -1.29274619e+00
-3.35030884e-01 -4.52051938e-01 -1.22090387e+00 -3.88193488e-01
2.06270516e-01 2.51493067e-01 -3.70033145e-01 1.97920012e+00
2.46393055e-01 -2.82006174e-01 9.84232798e-02 8.92440677e-01
8.16416740e-01 4.33091223e-01 -2.98324853e-01 -7.03239918e-01
1.27579379e+00 -9.75092113e-01 -8.04049313e-01 -1.18069723e-01
6.78945482e-01 -5.30655682e-01 1.08375728e+00 2.91534007e-01
-1.09651017e+00 -3.56580585e-01 -9.65809047e-01 -1.10502511e-01
4.72247824e-02 -7.68757313e-02 5.69932282e-01 3.44362080e-01
-8.54418635e-01 3.81580651e-01 -5.90781748e-01 -7.66970217e-02
3.96484584e-01 2.61232525e-01 -4.08094466e-01 -4.88676637e-01
-9.95030761e-01 5.51795363e-01 2.01317042e-01 -1.43844828e-01
-8.64592969e-01 -7.19923854e-01 -8.05127442e-01 1.11188710e-01
7.74419367e-01 -9.32661593e-01 8.70001495e-01 -7.30232537e-01
-1.22546756e+00 8.03193748e-01 -4.28141564e-01 1.51201949e-01
5.31001866e-01 -7.01211691e-02 -3.04530919e-01 2.32947871e-01
3.27355295e-01 5.25039136e-01 7.83414006e-01 -1.64226449e+00
-3.13698024e-01 -5.62072337e-01 3.10591877e-01 4.52506036e-01
-3.82873029e-01 -2.37190992e-01 -6.42184675e-01 -4.12603647e-01
7.22536802e-01 -8.65894318e-01 -1.09934248e-01 2.22724020e-01
-5.19205034e-01 -2.00871512e-01 5.48932791e-01 -3.94236743e-01
1.16823578e+00 -2.21125603e+00 5.48514307e-01 1.54514208e-01
6.51612461e-01 -2.70586789e-01 -3.84554081e-02 5.65970659e-01
2.21783027e-01 2.80696481e-01 -1.03474863e-01 -5.32070220e-01
1.47800362e-02 2.19870776e-01 -7.59635717e-02 6.66430056e-01
-2.33750448e-01 7.72182822e-01 -1.07145226e+00 -7.55264699e-01
1.22789316e-01 3.10786366e-01 -8.06367934e-01 2.35328838e-01
3.11524840e-03 5.78085482e-01 -5.89807570e-01 7.71921456e-01
7.72489190e-01 -8.80831838e-01 5.00618279e-01 -5.51349282e-01
1.44100919e-01 8.64141509e-02 -1.31891537e+00 1.91249561e+00
-5.38863003e-01 3.78463864e-01 1.45901889e-01 -1.11849046e+00
7.48034000e-01 2.01945916e-01 7.23111868e-01 -4.75358993e-01
3.98063026e-02 2.07733717e-02 -2.62873173e-01 -4.97259200e-01
2.56618708e-01 -2.44208738e-01 -2.05653012e-02 5.38633049e-01
1.24993652e-01 8.53503048e-02 -1.44434571e-01 3.86976808e-01
6.95734978e-01 6.34470209e-03 4.98815149e-01 -4.66327786e-01
4.79020506e-01 -4.74402815e-01 7.82886803e-01 8.39652658e-01
-3.63226354e-01 7.13957727e-01 5.88266969e-01 -3.33963871e-01
-9.79569256e-01 -1.08388233e+00 2.63506733e-03 1.08954406e+00
5.77146113e-01 -6.06811345e-01 -3.17152470e-01 -7.34868526e-01
-1.82207316e-01 5.59354544e-01 -5.98797500e-01 -1.98474050e-01
-8.74235183e-02 -5.52539468e-01 -7.20352307e-02 4.28236306e-01
5.11833549e-01 -3.04837912e-01 -3.04037243e-01 -1.43140301e-01
-6.66044712e-01 -1.07435310e+00 -4.99802858e-01 1.16787218e-01
-8.66005182e-01 -9.37055290e-01 -4.19473678e-01 -3.68696272e-01
6.91068113e-01 9.55343008e-01 1.24020469e+00 -8.49873945e-02
8.81987363e-02 5.89795113e-01 -2.76183456e-01 -1.34819254e-01
-2.05904290e-01 7.49376640e-02 1.57205135e-01 -2.54531279e-02
2.50126958e-01 -8.45124602e-01 -7.58272827e-01 6.37068868e-01
-9.26355422e-01 3.20349485e-01 5.88738739e-01 8.87663484e-01
9.28760707e-01 -1.22241363e-01 4.80861366e-01 -9.63458896e-01
2.69284666e-01 -7.07981110e-01 -4.64144915e-01 6.37419224e-01
-7.92339921e-01 3.23794514e-01 5.40602624e-01 -1.93887755e-01
-1.03751791e+00 -1.87282786e-01 4.29470927e-01 -8.13610673e-01
1.94330782e-01 6.73496902e-01 -5.11095822e-01 1.75437823e-01
3.29992294e-01 3.18611562e-01 4.15775627e-02 -3.86729717e-01
6.14922822e-01 4.46089029e-01 1.29448116e-01 -6.30492449e-01
4.81875241e-01 8.24995875e-01 1.30038321e-01 -6.28804684e-01
-1.42029119e+00 -7.00027049e-01 -7.28384197e-01 -1.67319834e-01
6.23036861e-01 -1.15737665e+00 -1.00157547e+00 -7.07430989e-02
-9.48499918e-01 5.03661215e-01 -1.01379409e-01 4.25130069e-01
-8.45534027e-01 5.77891648e-01 -3.61864507e-01 -9.03809130e-01
-1.05651401e-01 -1.20098042e+00 1.02320063e+00 -1.62874460e-01
1.09621935e-01 -1.11366189e+00 8.94241780e-03 6.86570108e-01
2.50778407e-01 -8.87265801e-02 9.79712486e-01 -6.01153076e-01
-7.20316589e-01 1.28084540e-01 -5.57804167e-01 2.38306537e-01
2.44147480e-01 -2.58720994e-01 -1.05089402e+00 -4.30831879e-01
2.70320773e-01 -4.05243546e-01 9.20239329e-01 3.75311226e-01
1.27318561e+00 -4.08732742e-01 -2.65530288e-01 7.76271105e-01
1.48679185e+00 -1.16064750e-01 2.67584801e-01 -1.93846989e-02
9.12835777e-01 7.03446984e-01 4.81403440e-01 8.64319265e-01
8.20921957e-01 6.48324966e-01 5.66779613e-01 2.83827990e-01
1.36417434e-01 -2.66632527e-01 2.45267659e-01 1.46719623e+00
-1.96482390e-01 -4.22976285e-01 -6.41994834e-01 3.82256389e-01
-2.03664827e+00 -1.02901793e+00 1.68921649e-01 2.19947886e+00
5.67105174e-01 -1.88380137e-01 -8.70976504e-03 -1.94217786e-01
6.31866097e-01 4.48015541e-01 -8.20190847e-01 9.55080613e-02
-1.85935974e-01 -7.61661291e-01 3.91098320e-01 3.01160663e-01
-1.00201344e+00 5.85914791e-01 6.26849413e+00 7.41116464e-01
-6.40591919e-01 2.69140005e-01 6.77345097e-01 -3.09165120e-01
-8.91662121e-01 -8.73797834e-02 -7.08797395e-01 1.67251781e-01
6.78940117e-01 -1.67078644e-01 5.24190724e-01 8.41145337e-01
-1.10916533e-02 8.14305097e-02 -1.37343287e+00 1.32884634e+00
1.82576030e-01 -1.42569232e+00 3.01958591e-01 1.33468077e-01
8.23484302e-01 1.15995063e-02 2.40611672e-01 1.71932518e-01
3.60409081e-01 -6.71488523e-01 6.88321173e-01 4.94448125e-01
7.61040926e-01 -7.09002733e-01 5.73948205e-01 5.25878549e-01
-1.38544488e+00 -2.30695039e-01 -5.63341796e-01 1.43267468e-01
1.20584585e-01 6.17292285e-01 -2.71142513e-01 1.11050677e+00
5.35572708e-01 1.08345973e+00 -3.29027623e-01 4.42627519e-01
1.97012484e-01 7.85643533e-02 -1.39368713e-01 2.69531041e-01
1.40012920e-01 -3.91469359e-01 6.06498837e-01 7.60337770e-01
1.10218272e-01 1.40163034e-01 4.07364547e-01 9.70834970e-01
1.92318037e-02 1.22020416e-01 -8.93118441e-01 1.34761259e-01
7.74806261e-01 1.00920355e+00 -6.36313200e-01 -2.16187924e-01
-6.72997653e-01 8.47471356e-01 6.82075977e-01 4.67303038e-01
-6.70923054e-01 2.84443080e-01 7.72316515e-01 -9.77493823e-02
3.49501729e-01 -7.30404407e-02 -3.35968077e-01 -1.65323544e+00
1.89909548e-01 -8.29320133e-01 5.80629945e-01 -4.39984560e-01
-1.63156199e+00 4.02041793e-01 3.53989333e-01 -1.54460466e+00
-1.98250979e-01 -3.39311391e-01 -5.18418401e-02 6.12777650e-01
-1.51547956e+00 -1.21241117e+00 -2.25411534e-01 7.29833961e-01
5.98329127e-01 -8.67704898e-02 6.96356118e-01 3.08424652e-01
-6.42682552e-01 6.86016858e-01 3.60119075e-01 -2.30300799e-01
6.31583452e-01 -1.03405428e+00 -1.90471500e-01 4.23884779e-01
1.19936168e-01 6.68910205e-01 5.62223971e-01 -3.22669387e-01
-1.76568604e+00 -8.73437405e-01 4.65471268e-01 -4.80028361e-01
6.05062068e-01 -4.00548816e-01 -8.40522945e-01 7.61961877e-01
-2.64653116e-02 1.65013671e-01 1.04950821e+00 3.44407231e-01
-7.54863739e-01 -1.29975736e-01 -1.00461769e+00 4.54491824e-01
1.42725933e+00 -7.60851026e-01 -2.98845053e-01 4.41499352e-01
9.41443563e-01 -5.69264144e-02 -1.05245459e+00 3.73537809e-01
5.64942420e-01 -1.31817138e+00 1.15872264e+00 -6.98475659e-01
5.44551253e-01 -7.80978501e-02 -9.05418038e-01 -1.08891380e+00
-5.24405599e-01 -5.58473766e-01 -4.04247284e-01 1.24664593e+00
3.83924216e-01 -5.69335163e-01 5.69202721e-01 7.18684494e-01
-1.70710534e-02 -7.99151301e-01 -9.78825271e-01 -9.34662104e-01
-5.33124059e-02 -3.76069337e-01 4.41394508e-01 1.08825529e+00
5.97716719e-02 3.61961961e-01 -8.02799642e-01 2.45756775e-01
9.52928066e-01 5.10958850e-01 4.80174422e-01 -1.31661904e+00
-3.71939003e-01 -3.22758287e-01 -2.64256775e-01 -1.36008644e+00
1.38674468e-01 -8.98010731e-01 -2.32898355e-01 -1.34567964e+00
8.33112836e-01 -5.41744947e-01 -6.21937931e-01 4.55657244e-02
-2.53999799e-01 -3.74999702e-01 3.34822059e-01 7.83599794e-01
-1.00326252e+00 7.34259963e-01 1.48336565e+00 -9.08898935e-02
-6.21569082e-02 -1.31399989e-01 -1.00139892e+00 8.10178995e-01
4.54715729e-01 -5.10253251e-01 -9.31430757e-01 -5.55765390e-01
6.22595787e-01 3.51079941e-01 1.89756960e-01 -4.26740140e-01
2.24692941e-01 -3.98337305e-01 -6.93138689e-03 -6.77914977e-01
4.99207556e-01 -8.25360894e-01 1.64661944e-01 6.38529286e-02
-4.95875448e-01 8.03084765e-03 -3.22936386e-01 1.06193697e+00
-2.34176606e-01 -7.53164589e-02 6.15744293e-01 -4.47177052e-01
-6.38552010e-01 5.34551561e-01 6.09862618e-02 4.14292276e-01
7.54021525e-01 7.40110455e-03 -3.82552415e-01 -6.56347573e-01
-5.63265502e-01 3.91383588e-01 5.42134464e-01 2.94701636e-01
6.10031426e-01 -1.50523186e+00 -6.05808675e-01 1.46578237e-01
5.17310202e-01 3.73117849e-02 5.50976157e-01 1.07370043e+00
-3.76506485e-02 3.37258369e-01 6.86543137e-02 -8.76585662e-01
-8.95323455e-01 9.77771521e-01 3.48797627e-02 -4.04914856e-01
-6.66298509e-01 7.44235694e-01 6.44081235e-01 -4.02747244e-01
2.64854342e-01 -6.84590340e-02 -1.48664907e-01 2.71418333e-01
4.81045693e-01 5.06280422e-01 -3.42277706e-01 -5.10496855e-01
-3.83758903e-01 5.79869211e-01 -2.03868225e-01 -1.14214689e-01
1.17691886e+00 -7.83985257e-01 -1.10714599e-01 8.93678427e-01
1.49816847e+00 -3.05861562e-01 -1.36229622e+00 -5.86798787e-01
-1.24446787e-01 -6.91365063e-01 1.05114818e-01 -3.73865128e-01
-1.19964206e+00 9.18704867e-01 2.77381510e-01 3.33535314e-01
1.00368524e+00 4.21800196e-01 4.22841817e-01 5.78723907e-01
6.00478232e-01 -9.91838753e-01 2.17135906e-01 1.90098092e-01
8.48807216e-01 -1.66003060e+00 2.65116811e-01 -6.35881543e-01
-9.29945886e-01 8.26540589e-01 5.52535236e-01 3.85720015e-01
9.00059938e-01 -1.48146585e-01 -1.05780333e-01 -5.16384184e-01
-1.03145742e+00 5.78004569e-02 4.01482195e-01 5.12277246e-01
2.76383400e-01 9.60900486e-02 1.33055210e-01 5.10254204e-01
2.41270140e-01 -4.05071467e-01 2.71845281e-01 7.56675243e-01
-3.53573918e-01 -6.78434908e-01 6.33477867e-02 4.39788938e-01
-2.86841065e-01 -6.14640862e-02 -2.09145248e-01 7.83938050e-01
-1.28346056e-01 9.69871342e-01 -1.12516306e-01 -3.83526891e-01
2.31515858e-02 -1.98814958e-01 4.82501060e-01 -3.90550166e-01
-1.54905282e-02 2.37838939e-01 -1.50159299e-01 -7.30227292e-01
-8.09262693e-01 -7.89251387e-01 -7.91002154e-01 -3.50826740e-01
-5.03935874e-01 -8.42030570e-02 3.26769114e-01 9.81345415e-01
6.55429006e-01 3.46839190e-01 9.72461343e-01 -5.94286621e-01
-9.13254023e-01 -7.32342482e-01 -7.94142783e-01 4.41751122e-01
2.86220402e-01 -1.02478468e+00 -6.18081212e-01 -9.15087536e-02]
|
[8.511631965637207, 4.477954387664795]
|
4e4a7ace-8a84-46db-8a6c-d5f681109a54
|
pgtask-introducing-the-task-of-profile
|
2304.06634
| null |
https://arxiv.org/abs/2304.06634v1
|
https://arxiv.org/pdf/2304.06634v1.pdf
|
PGTask: Introducing the Task of Profile Generation from Dialogues
|
Recent approaches have attempted to personalize dialogue systems by leveraging profile information into models. However, this knowledge is scarce and difficult to obtain, which makes the extraction/generation of profile information from dialogues a fundamental asset. To surpass this limitation, we introduce the Profile Generation Task (PGTask). We contribute with a new dataset for this problem, comprising profile sentences aligned with related utterances, extracted from a corpus of dialogues. Furthermore, using state-of-the-art methods, we provide a benchmark for profile generation on this novel dataset. Our experiments disclose the challenges of profile generation, and we hope that this introduces a new research direction.
|
['Luísa Coheur', 'Joao P. Carvalho', 'Rui Ribeiro']
|
2023-04-13
| null | null | null | null |
['pgtask']
|
['natural-language-processing']
|
[ 4.19320285e-01 6.61722183e-01 -3.60451490e-01 -6.39672220e-01
-8.42622459e-01 -6.77522719e-01 9.98697817e-01 -6.69721812e-02
-1.86911970e-01 1.22617245e+00 6.93652391e-01 6.41474500e-02
1.38102189e-01 -5.93177617e-01 -1.46640375e-01 -3.47899884e-01
8.09418410e-02 7.35671103e-01 3.44225951e-02 -7.35810578e-01
3.43497962e-01 7.35246763e-02 -1.38489330e+00 7.48103440e-01
1.05480886e+00 8.16815734e-01 2.96298414e-01 6.01263285e-01
-2.86412686e-01 7.68309236e-01 -1.06443989e+00 -7.29352653e-01
-1.19270548e-01 -7.58993030e-01 -1.45290077e+00 2.12801591e-01
-1.19836964e-01 -3.16268593e-01 -1.95417911e-01 8.30649912e-01
6.72896624e-01 3.65278333e-01 6.50435627e-01 -1.27361369e+00
-6.09014392e-01 9.91980076e-01 1.38681576e-01 -1.98185429e-01
8.65135193e-01 6.15981966e-02 1.25219941e+00 -5.45284510e-01
9.06804621e-01 1.21592152e+00 4.45838571e-01 1.11401188e+00
-9.71497834e-01 -9.63981226e-02 6.41130656e-02 1.20617516e-01
-8.59838009e-01 -7.91074276e-01 1.04692364e+00 -2.73778647e-01
9.47227120e-01 3.72135252e-01 5.79851866e-01 1.88620198e+00
-3.67887408e-01 1.24361598e+00 1.26503909e+00 -4.63204980e-01
-1.24652982e-02 4.66548264e-01 4.57987823e-02 3.34503651e-01
-4.07637179e-01 -1.87937602e-01 -7.70502448e-01 -4.29362774e-01
2.81271636e-01 -7.37482846e-01 -6.14094794e-01 -1.10458963e-01
-1.06955326e+00 9.07854140e-01 -2.85543859e-01 1.70631990e-01
-1.69605598e-01 -8.47623110e-01 4.15297896e-01 4.29752946e-01
7.44229019e-01 9.04301763e-01 -5.70561409e-01 -8.87678206e-01
-6.83149397e-01 5.77840030e-01 1.56629467e+00 1.15643442e+00
6.53928399e-01 -3.39326292e-01 -5.66920877e-01 1.29925191e+00
1.71394527e-01 2.16813669e-01 7.76295960e-01 -8.83343220e-01
7.28870451e-01 6.97395504e-01 4.17570651e-01 -6.37075841e-01
-5.56445956e-01 9.45177600e-02 -6.66794300e-01 -5.36360323e-01
4.77461249e-01 -5.62486410e-01 -1.35863982e-02 1.70756018e+00
3.06964129e-01 -2.58373111e-01 4.48232949e-01 5.47282279e-01
1.11836898e+00 5.41715384e-01 -9.34567675e-02 -4.36992526e-01
1.46396840e+00 -1.25459135e+00 -9.05912459e-01 -1.91085190e-01
6.00121319e-01 -8.18878949e-01 1.34851933e+00 1.47866443e-01
-1.02905917e+00 -2.67593980e-01 -8.96348059e-01 1.88436061e-01
-2.16259286e-01 4.90376763e-02 7.91568518e-01 9.18921471e-01
-9.14381564e-01 7.90786028e-01 -3.71056795e-01 -3.98842245e-01
1.01227395e-01 7.32476041e-02 -2.42931738e-01 4.98198539e-01
-1.57654774e+00 1.04923332e+00 4.71324354e-01 -1.43817291e-02
-5.93867600e-01 -3.81428421e-01 -9.71456170e-01 -2.33475447e-01
5.11182845e-01 -7.02499807e-01 1.84713471e+00 -5.44313252e-01
-2.46787572e+00 5.75498104e-01 -1.65738195e-01 -4.52253759e-01
5.06940246e-01 -7.29406551e-02 -4.09641564e-01 -5.78359962e-02
-6.96921647e-02 2.70390451e-01 4.14874136e-01 -1.09446371e+00
-8.05844188e-01 -1.50995016e-01 4.45378542e-01 5.34581900e-01
-5.71189940e-01 2.36883029e-01 -1.50978282e-01 -2.00166225e-01
-6.01455331e-01 -8.87560487e-01 -2.71767944e-01 -8.95683169e-01
-7.75632739e-01 -6.40587628e-01 6.70677960e-01 -4.94160235e-01
1.24935544e+00 -1.58997381e+00 2.45416105e-01 -3.77038479e-01
1.07043378e-01 5.38388610e-01 -7.83185139e-02 8.32799911e-01
4.71190572e-01 1.87872499e-01 -2.01949850e-01 -7.22692966e-01
3.51262808e-01 1.35419682e-01 -3.50955814e-01 -1.35753512e-01
1.86212361e-01 8.35904896e-01 -1.35085678e+00 -4.38642085e-01
-1.61372483e-01 1.14609733e-01 -4.22251850e-01 8.93544853e-01
-7.47574568e-01 7.91388452e-01 -7.02280521e-01 3.12130332e-01
2.79593110e-01 -2.93136865e-01 4.47070271e-01 3.82093340e-02
-2.45377794e-03 8.86135340e-01 -7.41191864e-01 1.79295182e+00
-6.13042295e-01 3.30177605e-01 -6.06861971e-02 -6.93558514e-01
1.09679055e+00 6.03221118e-01 3.83840382e-01 -2.43274778e-01
9.66754463e-03 1.54359952e-01 -8.59272927e-02 -8.17875087e-01
1.22914708e+00 1.21594109e-01 -5.11833608e-01 8.75606418e-01
3.03513885e-01 -4.01666731e-01 6.06968641e-01 1.10474780e-01
9.39921558e-01 1.33287460e-01 5.66154420e-01 1.18973851e-01
7.23282993e-01 8.77106935e-02 4.98364449e-01 7.48264611e-01
-3.60796422e-01 4.80960310e-01 7.42880702e-01 -8.47702101e-02
-7.91378021e-01 -6.31913245e-01 2.27519982e-02 1.14876425e+00
-2.28695273e-01 -7.84781933e-01 -9.57876682e-01 -1.19430459e+00
-3.64580989e-01 5.16919017e-01 -5.40867746e-01 1.58782214e-01
-5.71782351e-01 -9.32565093e-01 7.68400013e-01 1.41353056e-01
6.22928381e-01 -1.35357976e+00 -1.42352298e-01 4.36380208e-01
-7.63787210e-01 -1.31401145e+00 -5.15186965e-01 -3.62297177e-01
-5.85450709e-01 -1.20622385e+00 -6.24343514e-01 -4.62780803e-01
2.39535823e-01 -1.50051462e-02 1.44880307e+00 -9.00110230e-02
4.61145401e-01 2.51465559e-01 -7.75186181e-01 -3.02821457e-01
-9.77929831e-01 8.17533135e-01 -7.61942044e-02 3.45823690e-02
2.90403903e-01 -5.08031547e-01 -1.85704902e-01 2.87802279e-01
-4.47984487e-01 1.88583985e-01 4.14678633e-01 1.21463788e+00
5.24194129e-02 -3.85263145e-01 1.02953601e+00 -1.55420589e+00
1.62556553e+00 -4.54310775e-01 -3.68542135e-01 3.17596912e-01
-5.02328038e-01 1.79243013e-02 6.69552565e-01 -1.76199079e-01
-1.71360052e+00 -2.42873117e-01 -5.31164944e-01 3.90525490e-01
-2.72903621e-01 6.96221232e-01 -5.67842066e-01 1.79533660e-01
6.51703119e-01 2.72987992e-01 1.36635676e-01 -6.54748678e-01
6.80014729e-01 1.20685613e+00 4.20715690e-01 -8.07534039e-01
3.42721522e-01 5.24130948e-02 -4.28731918e-01 -9.47608173e-01
-1.17840266e+00 -4.62929010e-01 -6.94121480e-01 -1.30415738e-01
4.57840949e-01 -4.97436166e-01 -6.81596577e-01 5.56283355e-01
-1.46456969e+00 -4.98411417e-01 -1.09403193e-01 8.20453241e-02
-8.45476210e-01 6.48261011e-01 -7.43738055e-01 -1.00783718e+00
-6.39815986e-01 -1.15909266e+00 7.40534961e-01 4.07268167e-01
-7.65986860e-01 -1.20961511e+00 3.26072633e-01 7.20443964e-01
5.44888377e-01 -2.39292514e-02 6.33476794e-01 -1.32391977e+00
-2.77008951e-01 5.15236966e-02 2.01146409e-01 3.69247496e-01
4.11906272e-01 -1.49356648e-01 -1.12894452e+00 7.37339854e-02
1.86333165e-01 -5.82933128e-01 3.57453078e-01 -3.91310573e-01
1.06853855e+00 -6.40793860e-01 -2.65064031e-01 2.78439939e-01
4.01857466e-01 1.10483684e-01 5.24692357e-01 4.53941733e-01
3.50875705e-01 1.07135010e+00 9.62995350e-01 8.15721452e-01
1.00132215e+00 1.04326797e+00 -1.01142451e-01 2.89582431e-01
-3.63495685e-02 -3.46870065e-01 1.67242840e-01 1.11492801e+00
-1.17490567e-01 -4.12691116e-01 -6.52733564e-01 6.24006212e-01
-1.99453294e+00 -1.00879347e+00 6.00625649e-02 1.83438230e+00
1.66649377e+00 -1.30901396e-01 3.42146516e-01 -1.72312528e-01
6.77648902e-01 4.90474969e-01 -1.09437153e-01 -3.29813242e-01
-2.13238746e-01 4.62647192e-02 -2.51555771e-01 5.20972431e-01
-1.21887791e+00 1.08277857e+00 6.52157974e+00 7.08432913e-01
-6.97890460e-01 8.14975053e-02 4.34693813e-01 2.11195841e-01
-3.96771371e-01 -6.80493563e-02 -1.16841578e+00 7.82656431e-01
1.22241569e+00 -6.88184679e-01 2.88023859e-01 9.62722003e-01
8.65309313e-02 1.72293454e-01 -1.30564940e+00 6.62132442e-01
2.17648700e-01 -1.24185503e+00 1.42531842e-02 -2.29501855e-02
6.64246857e-01 -1.35199800e-01 -1.23041138e-01 5.57980537e-01
6.45319343e-01 -8.69635642e-01 2.17802420e-01 3.31693649e-01
4.09666032e-01 -5.27373552e-01 8.02953184e-01 5.92264473e-01
-6.40984774e-01 2.99784839e-01 -3.16950113e-01 -1.19221441e-01
5.61153293e-01 4.83976334e-01 -1.58673513e+00 8.06063294e-01
1.87904552e-01 5.21425128e-01 -3.91427338e-01 8.42093170e-01
-5.73855042e-01 4.11939353e-01 1.11281417e-01 -3.62745941e-01
8.60906221e-06 -2.20279023e-01 6.20007277e-01 1.26007152e+00
5.64760044e-02 -1.11460708e-01 1.82337165e-01 8.78440857e-01
-3.06869000e-01 2.09839568e-01 -6.17114067e-01 -3.42400789e-01
8.51390898e-01 1.54663384e+00 -1.05412506e-01 -5.00062525e-01
-3.39497000e-01 9.22939718e-01 4.68672216e-01 3.20625789e-02
-2.43766785e-01 -4.03286308e-01 8.64865184e-01 -4.86265153e-01
-3.08458030e-01 -1.94249433e-02 -1.30896747e-01 -1.41426396e+00
-7.99546763e-02 -1.36659718e+00 3.05011064e-01 -2.04936638e-01
-1.55707705e+00 9.89853024e-01 2.13685617e-01 -1.09374702e+00
-1.21181107e+00 -3.50560069e-01 -4.33431089e-01 1.00897467e+00
-1.60002995e+00 -1.13340104e+00 -2.82989144e-01 3.76884222e-01
6.71565592e-01 -4.74048197e-01 1.28092849e+00 5.80332614e-02
-4.48379785e-01 7.39319742e-01 -1.65015221e-01 2.00728297e-01
1.06676877e+00 -1.32900548e+00 7.75823832e-01 2.35174626e-01
-3.88490297e-02 6.36227250e-01 7.50171542e-01 -5.51120162e-01
-1.15150118e+00 -9.45765793e-01 1.38138163e+00 -8.42688560e-01
6.83462858e-01 -4.89622653e-01 -9.53635693e-01 5.51529884e-01
5.26204050e-01 -6.94887161e-01 1.15141428e+00 4.12369341e-01
-2.61827633e-02 4.19801712e-01 -1.15233386e+00 8.19036484e-01
1.16755950e+00 -6.08159661e-01 -8.85767579e-01 4.35293585e-01
6.99590325e-01 -8.81369889e-01 -1.16022050e+00 4.21188176e-01
3.94939303e-01 -9.09892023e-01 6.84405863e-01 -5.47980905e-01
4.33730960e-01 2.04615414e-01 1.35645643e-01 -2.02525091e+00
3.64835918e-01 -1.17015016e+00 -2.52086967e-01 1.82016730e+00
7.05941200e-01 -6.71050072e-01 9.31182981e-01 6.56207740e-01
-3.85363013e-01 -7.53500342e-01 -7.06015587e-01 -7.10671604e-01
9.31333676e-02 -9.18310583e-02 1.21192837e+00 9.35038924e-01
6.71383739e-01 8.55130970e-01 -7.08700299e-01 -3.53824586e-01
2.26518244e-01 2.14189485e-01 1.15311790e+00 -1.24383342e+00
-4.39519882e-01 -5.32412767e-01 1.90693244e-01 -1.40553975e+00
6.98673666e-01 -6.56684220e-01 9.61515531e-02 -1.47790051e+00
1.46168381e-01 -4.38514322e-01 3.82995695e-01 3.00333798e-01
-5.07982254e-01 -1.15502693e-01 1.25776112e-01 2.05881670e-01
-7.25539148e-01 1.10345471e+00 1.41020811e+00 -1.60588726e-01
-2.78408080e-01 4.98221368e-01 -9.10011351e-01 5.95098495e-01
1.07455003e+00 -1.54719621e-01 -5.36865950e-01 -7.27201030e-02
-1.17789708e-01 4.62845325e-01 -3.04271817e-01 -3.99615467e-01
2.43406370e-01 -3.59914005e-01 -1.91263124e-01 -4.14213210e-01
7.96890140e-01 -2.56619230e-02 -3.91556799e-01 -2.63965279e-01
-4.77898568e-01 -3.35788801e-02 -9.11032483e-02 2.68449545e-01
-4.30942029e-01 -6.82622433e-01 2.47361377e-01 -3.77756178e-01
-7.28088975e-01 1.48418605e-01 -4.32025015e-01 5.43912768e-01
5.99212408e-01 3.09476167e-01 -8.29345882e-01 -8.28108788e-01
-4.39679652e-01 3.43973488e-01 5.47057211e-01 5.46896756e-01
4.23550040e-01 -1.17368209e+00 -8.12215686e-01 4.71590534e-02
2.51011550e-01 -2.32710257e-01 2.09782362e-01 4.73744392e-01
1.46981969e-01 5.24687588e-01 -2.20462456e-01 -2.42933348e-01
-1.27310514e+00 1.78264096e-01 2.03695938e-01 -6.32911623e-01
-2.33325556e-01 7.74501801e-01 -2.33271658e-01 -9.65754449e-01
-4.81802747e-02 4.72191542e-01 -8.36793184e-01 3.17602277e-01
7.17203379e-01 2.19629779e-01 1.23392679e-01 -5.80967128e-01
2.09025033e-02 -2.60510266e-01 -3.70380729e-01 -3.93685728e-01
1.07219863e+00 -3.47105324e-01 -3.89505416e-01 4.26339090e-01
7.67848432e-01 2.26853833e-01 -1.16717863e+00 -4.76512372e-01
4.15904194e-01 -5.67400217e-01 -6.76122189e-01 -8.61021340e-01
-4.29936796e-01 6.40101969e-01 -2.31017977e-01 6.17420554e-01
6.09570503e-01 -1.15714055e-02 9.32352781e-01 5.52512944e-01
5.73633969e-01 -1.33088124e+00 2.13589862e-01 1.14350712e+00
1.05643594e+00 -1.28578234e+00 -3.51430923e-01 -5.56724072e-01
-1.11392117e+00 9.08809245e-01 7.72060156e-01 5.59657097e-01
2.62118638e-01 2.63078883e-02 3.81232619e-01 4.34948690e-02
-1.11072040e+00 -2.69432724e-01 1.33332461e-01 9.50677037e-01
8.58241141e-01 1.92323074e-01 -5.20763576e-01 1.13989031e+00
-8.30298305e-01 -1.55437618e-01 8.69107187e-01 7.80790448e-01
-1.24917395e-01 -2.09269428e+00 1.02257058e-01 3.58920962e-01
-5.21083593e-01 -1.94112301e-01 -8.64431620e-01 5.85295498e-01
-6.08245492e-01 1.50881565e+00 -4.24006671e-01 -5.71314454e-01
4.48588490e-01 4.27732617e-01 4.33813453e-01 -1.05432081e+00
-6.88108921e-01 -4.56304193e-01 1.09821868e+00 -2.21636698e-01
-4.24807042e-01 -6.63942039e-01 -5.63940167e-01 -3.02446246e-01
-1.44615293e-01 7.16453969e-01 4.78459179e-01 9.22594786e-01
4.08981949e-01 3.20744574e-01 8.54030013e-01 -8.23641658e-01
-1.06704342e+00 -1.53073442e+00 -3.26260924e-01 5.38458228e-01
-1.51511759e-01 -7.06514359e-01 -1.66074336e-01 -6.11178316e-02]
|
[12.738418579101562, 8.159120559692383]
|
69d8aa6d-649a-470f-b9d3-44288f9b4118
|
usim-dal-uncertainty-aware-statistical-image
|
2305.17520
| null |
https://arxiv.org/abs/2305.17520v1
|
https://arxiv.org/pdf/2305.17520v1.pdf
|
USIM-DAL: Uncertainty-aware Statistical Image Modeling-based Dense Active Learning for Super-resolution
|
Dense regression is a widely used approach in computer vision for tasks such as image super-resolution, enhancement, depth estimation, etc. However, the high cost of annotation and labeling makes it challenging to achieve accurate results. We propose incorporating active learning into dense regression models to address this problem. Active learning allows models to select the most informative samples for labeling, reducing the overall annotation cost while improving performance. Despite its potential, active learning has not been widely explored in high-dimensional computer vision regression tasks like super-resolution. We address this research gap and propose a new framework called USIM-DAL that leverages the statistical properties of colour images to learn informative priors using probabilistic deep neural networks that model the heteroscedastic predictive distribution allowing uncertainty quantification. Moreover, the aleatoric uncertainty from the network serves as a proxy for error that is used for active learning. Our experiments on a wide variety of datasets spanning applications in natural images (visual genome, BSD100), medical imaging (histopathology slides), and remote sensing (satellite images) demonstrate the efficacy of the newly proposed USIM-DAL and superiority over several dense regression active learning methods.
|
['Biplab Banerjee', 'Zeynep Akata', 'Uddeshya Upadhyay', 'Vikrant Rangnekar']
|
2023-05-27
| null | null | null | null |
['image-super-resolution', 'active-learning', 'active-learning']
|
['computer-vision', 'methodology', 'natural-language-processing']
|
[ 5.46101034e-01 3.49178940e-01 -4.61644888e-01 -4.55050498e-01
-1.54178584e+00 8.23077094e-03 4.66547966e-01 1.92583472e-01
-8.03266704e-01 9.30718720e-01 2.72605687e-01 2.88379967e-01
-2.23044172e-01 -6.35932505e-01 -5.54275036e-01 -1.27796686e+00
2.48740092e-01 6.32525146e-01 2.58832365e-01 4.23153102e-01
1.98561192e-01 5.47222495e-01 -1.37465155e+00 9.40054208e-02
1.13278973e+00 9.47833657e-01 3.35478067e-01 2.98640251e-01
-2.75449485e-01 1.08847511e+00 -3.60898286e-01 -1.97671026e-01
-7.84774590e-03 3.16906013e-02 -6.22039020e-01 1.45407647e-01
1.87888771e-01 -2.37388432e-01 2.36315385e-01 9.25619364e-01
5.89151740e-01 1.41808540e-01 7.56776333e-01 -8.09640884e-01
-5.31382799e-01 5.78787923e-01 -9.71507072e-01 2.79940933e-01
-1.79579079e-01 7.24784611e-03 8.05588305e-01 -1.00707233e+00
6.48615539e-01 1.14155960e+00 5.97974241e-01 5.02719223e-01
-1.41418064e+00 -6.93908811e-01 1.04388431e-01 2.19482988e-01
-1.51640046e+00 -5.98712683e-01 6.36173427e-01 -4.70091701e-01
5.16058564e-01 1.28550813e-01 4.92352664e-01 1.00394595e+00
1.36368319e-01 1.19924319e+00 1.18009269e+00 -4.39100921e-01
4.57396060e-01 2.70636857e-01 5.95749393e-02 7.18665361e-01
2.25935549e-01 5.43630272e-02 -6.62637472e-01 -2.13247448e-01
9.75199699e-01 6.51045367e-02 -2.09941104e-01 -5.52228630e-01
-8.17460299e-01 9.08126056e-01 5.75383306e-01 -1.97183788e-02
-3.29165518e-01 1.69555396e-02 -9.26038623e-02 -4.01721269e-01
1.03696930e+00 5.09611368e-01 -1.69528425e-01 2.69292235e-01
-1.17408657e+00 -2.32533380e-01 2.19257548e-01 3.63444775e-01
8.39764655e-01 -1.65263191e-01 -3.32194179e-01 1.08179677e+00
6.09643161e-01 9.64721963e-02 2.29337513e-01 -1.00581193e+00
-1.11400671e-01 8.33319783e-01 2.48318389e-02 -7.59629309e-01
-4.02664185e-01 -5.41915059e-01 -1.04527712e+00 5.68324208e-01
9.31920856e-02 1.67636171e-01 -1.06850946e+00 1.23878121e+00
5.24405718e-01 5.12939095e-01 -3.10875535e-01 8.83140802e-01
1.06489456e+00 5.08052170e-01 2.71126747e-01 -6.80823147e-01
7.76867151e-01 -8.26458156e-01 -6.99200511e-01 -3.04904640e-01
3.53143245e-01 -4.28207666e-01 7.12371171e-01 5.26587129e-01
-9.03587103e-01 -2.35900521e-01 -8.57718349e-01 -1.88623980e-01
4.91381949e-03 1.31111234e-01 7.40229666e-01 3.12565386e-01
-8.36696088e-01 1.69221938e-01 -1.26738060e+00 1.16862074e-01
1.12282658e+00 4.71082240e-01 -3.81436557e-01 -2.17077523e-01
-7.57473886e-01 7.94391572e-01 3.07984978e-01 1.64728716e-01
-8.93230557e-01 -9.84442294e-01 -8.75925660e-01 -2.42503926e-01
5.35332799e-01 -4.69792008e-01 8.10863614e-01 -9.00864184e-01
-1.45561337e+00 9.21782255e-01 -2.45943270e-03 -5.24915040e-01
4.75493431e-01 -2.88198978e-01 1.28931716e-01 1.93079133e-02
-8.40804130e-02 9.52874899e-01 9.18085039e-01 -1.17001712e+00
-4.55139339e-01 -6.72164261e-01 -1.82689786e-01 3.68128330e-01
-2.03636631e-01 -1.03661813e-01 -6.87259436e-01 -4.98208344e-01
2.69729197e-01 -8.52069378e-01 -6.54288471e-01 3.70453745e-01
-3.29213291e-01 -1.74020082e-01 5.93320489e-01 -5.65195918e-01
8.64499688e-01 -2.18339562e+00 3.74855399e-01 3.44450250e-02
4.66282248e-01 1.09048486e-01 1.30183712e-01 -3.92157316e-01
2.42876381e-01 9.48763639e-03 -6.83858633e-01 -5.81940353e-01
-5.25126040e-01 3.46745640e-01 1.17560148e-01 5.85424423e-01
5.89675426e-01 8.91265035e-01 -6.99857473e-01 -1.04681516e+00
3.28068584e-01 9.39019084e-01 -3.03155243e-01 3.43365878e-01
-2.06971914e-01 8.78339350e-01 -3.74454588e-01 6.56238973e-01
6.75934076e-01 -5.72336793e-01 -1.20539956e-01 -1.33400932e-01
-8.16969275e-02 -9.04983655e-02 -1.14032114e+00 1.78753781e+00
-4.92641628e-01 6.25488698e-01 6.44165426e-02 -1.09588814e+00
9.35168982e-01 6.82562813e-02 7.36692727e-01 -7.31485724e-01
-6.33865446e-02 -9.03085321e-02 -3.23658556e-01 -5.47682464e-01
3.66623640e-01 3.48419882e-02 2.28016004e-01 1.12743117e-01
3.03361360e-02 -1.55309722e-01 -1.57472286e-02 9.10853818e-02
8.90135288e-01 3.28773618e-01 5.46189606e-01 -4.21750769e-02
3.62921536e-01 -4.09150030e-03 9.07811165e-01 6.61777020e-01
-8.31416845e-02 7.97877610e-01 3.47028404e-01 1.55258188e-02
-7.51069427e-01 -9.26051915e-01 -4.57589120e-01 1.04552841e+00
2.34071468e-03 1.63834020e-02 -4.71718252e-01 -7.33371258e-01
-2.48290718e-01 5.30902445e-01 -9.16675627e-01 8.76000747e-02
-3.78579557e-01 -1.38614357e+00 2.35127553e-01 5.06119668e-01
4.26698446e-01 -8.40539694e-01 -3.96677822e-01 -7.93551952e-02
-1.26953036e-01 -9.06876028e-01 2.02796265e-01 3.29795837e-01
-1.01886380e+00 -9.24665093e-01 -7.92994916e-01 -2.69903868e-01
8.51445735e-01 1.97712332e-03 1.23098207e+00 -3.09027910e-01
-7.85693228e-01 2.95523793e-01 -2.51893491e-01 -5.49277723e-01
-2.04134807e-01 3.25485831e-03 -4.80406225e-01 -5.63998744e-02
1.66694582e-01 -1.24166191e-01 -5.13748050e-01 1.65622130e-01
-9.41685438e-01 1.89529374e-01 9.89049256e-01 1.00080609e+00
1.34083760e+00 -9.55790952e-02 2.57850051e-01 -1.45922053e+00
1.91936847e-02 -5.30737877e-01 -9.10963893e-01 3.82251352e-01
-6.87648952e-01 2.22319379e-01 -1.17968105e-01 -3.40469420e-01
-1.46363175e+00 4.33610201e-01 -1.92889377e-01 -9.17279422e-02
5.86624406e-02 4.88496602e-01 -5.41639589e-02 -1.12040110e-01
7.60954380e-01 -8.04993883e-02 1.06106654e-01 -2.76941717e-01
7.93552920e-02 5.97432017e-01 3.25003713e-01 1.02560911e-02
5.29894054e-01 7.08993852e-01 1.10013448e-01 -8.90231192e-01
-1.41361272e+00 -6.31325603e-01 -7.04529285e-01 -9.78174284e-02
8.21070611e-01 -1.09803069e+00 -3.73431295e-01 2.08504081e-01
-7.60759532e-01 -2.27813318e-01 -3.39333057e-01 7.45261431e-01
-5.68101287e-01 2.62875885e-01 -3.14954042e-01 -1.00297379e+00
-5.15017450e-01 -1.37781239e+00 1.24119294e+00 4.08292830e-01
-8.45304206e-02 -1.10352004e+00 2.71350771e-01 8.51816714e-01
2.78149009e-01 3.16242784e-01 7.28564441e-01 -2.58273602e-01
-9.01201487e-01 -1.03667982e-01 -2.77955979e-01 3.31388056e-01
-6.70840740e-02 1.04315147e-01 -1.19366586e+00 -1.81971919e-02
-1.71584234e-01 -4.97607678e-01 1.35408282e+00 9.83586967e-01
1.49136615e+00 3.88485119e-02 -3.36216599e-01 7.80588984e-01
1.55217457e+00 -1.23408921e-01 9.70140517e-01 1.33769870e-01
5.39848506e-01 5.20733416e-01 9.10460353e-01 3.26358587e-01
5.39546795e-02 5.83627045e-01 6.23180926e-01 -4.20085728e-01
-1.35415927e-01 3.30084950e-01 1.27957612e-01 4.74248409e-01
-2.76654065e-01 3.67482379e-03 -8.99853230e-01 2.93051124e-01
-1.81494617e+00 -7.47694612e-01 5.88644296e-03 2.34172702e+00
1.08687437e+00 1.59104262e-02 -2.83815324e-01 1.40051335e-01
6.07279658e-01 -1.05471879e-01 -7.68060148e-01 3.01209211e-01
-1.99744508e-01 2.07083866e-01 4.48147088e-01 4.75113392e-01
-1.28112876e+00 6.87397301e-01 5.97990179e+00 9.64184582e-01
-1.09327805e+00 3.00757945e-01 1.09687626e+00 -1.27603710e-01
-1.24702506e-01 -1.72279209e-01 -1.02625322e+00 1.31552756e-01
5.35030127e-01 4.55904812e-01 -1.82307199e-01 7.99962521e-01
1.73333943e-01 -5.81227064e-01 -9.60398793e-01 1.07902610e+00
2.29821146e-01 -1.61742961e+00 -1.74279451e-01 9.03450791e-03
8.93045843e-01 1.39012367e-01 2.12581083e-01 5.49978353e-02
4.01843190e-01 -1.32163751e+00 2.65287906e-01 8.82829607e-01
8.14868569e-01 -6.82332158e-01 8.32064688e-01 5.29130578e-01
-5.01019895e-01 -1.29587844e-01 -4.47810084e-01 4.72723573e-01
-4.67165224e-02 1.00818264e+00 -1.26871192e+00 2.43704617e-01
6.61544621e-01 8.17033947e-01 -7.16003180e-01 1.32134771e+00
-1.01379529e-01 6.33734107e-01 -1.31093249e-01 2.77161986e-01
-1.94337636e-01 -4.83200997e-02 4.28745031e-01 1.05752647e+00
1.66275740e-01 -5.43098226e-02 -3.59825492e-02 9.90157127e-01
-6.51641786e-02 7.25759268e-02 -7.71099031e-02 -5.50314113e-02
3.30059171e-01 1.39976919e+00 -9.95516777e-01 1.44819662e-01
-2.32301280e-01 6.69631302e-01 4.91804242e-01 2.30855823e-01
-6.34998500e-01 3.18880439e-01 1.68331549e-01 1.61638245e-01
-1.20084742e-02 4.42088693e-02 -4.37962770e-01 -9.13389802e-01
-5.11021614e-01 -5.85441113e-01 4.49228376e-01 -8.02419960e-01
-1.27881110e+00 3.44646513e-01 3.75430323e-02 -1.05936468e+00
-2.33859956e-01 -4.36518043e-01 -9.17755812e-02 7.12248206e-01
-1.56035447e+00 -1.21484399e+00 -6.46989882e-01 3.79959762e-01
6.29762590e-01 -2.08776951e-01 8.52749527e-01 1.46571055e-01
-8.12608957e-01 2.70236254e-01 3.20539594e-01 1.08256727e-01
7.81168520e-01 -1.23549557e+00 -2.22808450e-01 6.73888326e-01
2.08403826e-01 3.42298031e-01 5.22071719e-01 -5.57790101e-01
-1.05632925e+00 -1.07036150e+00 1.40828490e-01 -4.03661996e-01
1.64514333e-01 -2.17695221e-01 -9.32529688e-01 4.85137254e-01
-3.47230524e-01 5.20958364e-01 9.11382735e-01 1.33012325e-01
-1.60028398e-01 -2.09270060e-01 -1.25471568e+00 1.07995443e-01
5.00002503e-01 -2.23552093e-01 -5.78862280e-02 1.87118798e-01
5.14975250e-01 -3.50217193e-01 -7.45579481e-01 6.16656005e-01
4.32103544e-01 -9.77606535e-01 1.11603630e+00 -1.27739623e-01
5.14663279e-01 -6.67772591e-02 4.02764864e-02 -1.02851880e+00
-4.18402135e-01 7.20324740e-02 -2.94322044e-01 1.22900856e+00
5.51037371e-01 -1.99258700e-01 9.93580937e-01 5.22837043e-01
-6.07604869e-02 -9.11739290e-01 -9.07709479e-01 -8.68476927e-02
-1.19314015e-01 -3.79563600e-01 7.99662545e-02 8.43341529e-01
-6.61362290e-01 2.31267840e-01 -2.99519360e-01 1.31433129e-01
1.09422171e+00 -2.92831212e-01 5.43840587e-01 -1.71981525e+00
-3.04326594e-01 -2.34860126e-02 -4.71073061e-01 -6.52480960e-01
1.90201834e-01 -5.94572365e-01 3.29130083e-01 -1.78321934e+00
6.47854507e-01 -8.35817933e-01 -2.37246901e-01 3.05265158e-01
-3.52845818e-01 6.65165782e-01 -3.71304244e-01 4.70435560e-01
-8.45764339e-01 4.34183270e-01 1.14687312e+00 -3.77044648e-01
-1.56858727e-01 1.79122761e-01 -5.94613135e-01 8.75776768e-01
4.02161509e-01 -5.42088449e-01 -4.78969544e-01 -3.53831738e-01
6.01493776e-01 1.14188664e-01 3.83958459e-01 -7.56432176e-01
2.58391947e-01 -3.37151259e-01 7.85960853e-01 -6.98837221e-01
6.37766182e-01 -7.54971564e-01 1.35512307e-01 1.66150495e-01
-7.12801218e-01 -8.37914526e-01 -9.26511511e-02 8.91330838e-01
-2.54330277e-01 -3.65526825e-01 1.31813192e+00 -2.33567134e-01
-7.66825259e-01 4.70252782e-01 -1.77437752e-01 -4.01994847e-02
9.56853867e-01 -2.57191777e-01 -2.22473621e-01 -1.27363373e-02
-8.39578450e-01 7.27891893e-05 2.63569117e-01 9.29267034e-02
8.08582246e-01 -8.62811148e-01 -9.87634480e-01 -4.83441502e-02
2.84612298e-01 7.06275344e-01 3.51941854e-01 1.12583828e+00
-4.60086197e-01 1.51524752e-01 -2.98568234e-02 -1.10873425e+00
-1.45827854e+00 1.57530442e-01 3.17823261e-01 -4.67602193e-01
-5.84031582e-01 1.32278955e+00 3.60092193e-01 -2.34024346e-01
3.82575333e-01 1.15645699e-01 -6.95878029e-01 2.46546924e-01
5.15684843e-01 5.06061316e-01 2.63417691e-01 -3.96258742e-01
-1.69361413e-01 3.67431045e-01 -2.72263259e-01 1.56358331e-01
1.74442685e+00 -3.16502422e-01 -3.32388073e-01 7.21084476e-01
8.85506809e-01 -1.80658966e-01 -1.54219210e+00 -6.45557702e-01
1.67366847e-01 -4.26437110e-01 7.49973059e-01 -7.70913363e-01
-1.09452188e+00 8.56692374e-01 1.01886654e+00 -2.94477969e-01
1.12701440e+00 9.76688564e-02 1.10626355e-01 1.94957837e-01
2.13015810e-01 -1.20372224e+00 2.37692013e-01 1.05320871e-01
6.54737294e-01 -1.83054972e+00 4.62873638e-01 -5.77956498e-01
-7.75841594e-01 8.19763184e-01 5.31977057e-01 -3.50295193e-02
5.33763766e-01 6.57131493e-01 -1.21352654e-02 3.29221860e-02
-7.64040768e-01 -1.92220688e-01 4.26164895e-01 6.76828861e-01
6.38370275e-01 -1.22510575e-01 3.89146321e-02 1.63185477e-01
3.63905877e-01 1.89277247e-01 4.55763191e-01 6.65920913e-01
-4.80508119e-01 -7.26209044e-01 -4.67998713e-01 5.83956063e-01
-5.69395006e-01 4.64894762e-03 -2.70773500e-01 4.58841532e-01
2.54159659e-01 6.12097323e-01 4.41917360e-01 2.64380842e-01
-4.62506711e-01 -1.95587546e-01 6.65726781e-01 -8.93123507e-01
-1.63036883e-01 2.64708281e-01 -1.17083557e-01 -4.06494856e-01
-9.39946949e-01 -5.90177536e-01 -1.26583195e+00 3.59199703e-01
-6.58002913e-01 -4.22071218e-02 7.83853352e-01 1.00547481e+00
7.80823603e-02 4.97701824e-01 4.40757513e-01 -5.04884899e-01
-2.14332327e-01 -1.03121793e+00 -6.27237499e-01 1.28666591e-02
3.46442461e-01 -9.20718014e-01 -2.72275031e-01 1.66891500e-01]
|
[14.712980270385742, -2.1740567684173584]
|
b0a9048d-e11a-40f9-a6bb-5019da4922e5
|
is-more-data-better-re-thinking-the
|
2209.10193
| null |
https://arxiv.org/abs/2209.10193v1
|
https://arxiv.org/pdf/2209.10193v1.pdf
|
Is More Data Better? Re-thinking the Importance of Efficiency in Abusive Language Detection with Transformers-Based Active Learning
|
Annotating abusive language is expensive, logistically complex and creates a risk of psychological harm. However, most machine learning research has prioritized maximizing effectiveness (i.e., F1 or accuracy score) rather than data efficiency (i.e., minimizing the amount of data that is annotated). In this paper, we use simulated experiments over two datasets at varying percentages of abuse to demonstrate that transformers-based active learning is a promising approach to substantially raise efficiency whilst still maintaining high effectiveness, especially when abusive content is a smaller percentage of the dataset. This approach requires a fraction of labeled data to reach performance equivalent to training over the full dataset.
|
['Scott A. Hale', 'Bertie Vidgen', 'Hannah Rose Kirk']
|
2022-09-21
| null |
https://aclanthology.org/2022.trac-1.7
|
https://aclanthology.org/2022.trac-1.7.pdf
|
trac-coling-2022-10
|
['abusive-language']
|
['natural-language-processing']
|
[ 4.50164862e-02 4.07588065e-01 -9.07087386e-01 -5.20634115e-01
-9.25602496e-01 -6.52100861e-01 2.59382188e-01 7.48995304e-01
-1.13476932e+00 7.62186825e-01 9.39049795e-02 -3.83656740e-01
-5.81739433e-02 -5.32450318e-01 -3.02742958e-01 -2.69523531e-01
-4.07425202e-02 2.89794922e-01 -1.21067263e-01 1.66172445e-01
5.59285998e-01 6.49036527e-01 -7.74628043e-01 -1.07655548e-01
9.57589686e-01 4.01296794e-01 -6.13861978e-01 1.90866306e-01
-8.98575932e-02 1.25636458e+00 -1.04902172e+00 -1.09843171e+00
2.00053081e-01 -1.08754978e-01 -1.02906990e+00 -2.77845822e-02
4.89667565e-01 -8.03548813e-01 -5.88236153e-02 9.28112686e-01
4.21908319e-01 2.41427600e-01 4.12823826e-01 -1.40437424e+00
-5.42878091e-01 6.10845089e-01 -4.89609778e-01 5.91598213e-01
2.43213087e-01 4.18014199e-01 1.01815116e+00 -5.51911473e-01
6.65308833e-01 1.06550145e+00 5.33543706e-01 6.58233702e-01
-1.46892869e+00 -9.64114964e-01 -3.57768461e-02 -9.83474031e-02
-1.11534286e+00 -9.02975380e-01 7.92713881e-01 -4.22339380e-01
1.01917636e+00 3.07804078e-01 5.74046433e-01 1.30679297e+00
2.45637428e-02 6.40199482e-01 9.98165548e-01 -3.25723469e-01
4.70216751e-01 4.66756970e-01 6.63191319e-01 5.36334276e-01
8.84188175e-01 -5.35863414e-02 -6.34127378e-01 -8.18162262e-01
1.61580950e-01 -9.49463397e-02 1.53224960e-01 -1.58408642e-01
-9.01674479e-02 1.48073268e+00 1.41223758e-01 1.67616829e-01
-2.41956830e-01 -1.21806502e-01 5.25911272e-01 5.20571396e-02
7.58399069e-01 1.05611098e+00 -2.04640120e-01 -6.46856308e-01
-1.00428832e+00 3.42131644e-01 8.55687916e-01 4.70539659e-01
4.00181562e-01 8.35074633e-02 2.01398298e-01 9.67056632e-01
2.33289152e-01 9.84821767e-02 7.97469616e-02 -1.10673034e+00
6.00014210e-01 8.51774752e-01 2.22194508e-01 -9.27583396e-01
-2.63683766e-01 -8.04130659e-02 1.92030724e-02 2.12714523e-01
4.69868869e-01 -4.39607501e-01 -6.65755212e-01 1.89559245e+00
2.26583645e-01 -4.09267128e-01 -4.64479268e-01 4.96120006e-01
1.78585336e-01 3.00999254e-01 8.78250718e-01 -7.02917278e-01
6.37546539e-01 -3.31615835e-01 -8.89433682e-01 -7.16652930e-01
9.62135911e-01 -4.69255805e-01 1.17916298e+00 5.15717208e-01
-1.35059988e+00 3.00050884e-01 -8.76906574e-01 -1.91964537e-01
-1.96317360e-01 -8.85316491e-01 8.70535433e-01 1.26852787e+00
-5.11288106e-01 4.11263525e-01 -8.43528986e-01 -1.43968031e-01
1.06805325e+00 3.32291752e-01 -3.38544369e-01 1.12007022e-01
-9.78263259e-01 1.28135121e+00 3.56789380e-01 -4.28192198e-01
-7.25305319e-01 -9.12484765e-01 -8.86510134e-01 4.88658361e-02
4.46253687e-01 8.01070631e-02 1.18594456e+00 -9.32913005e-01
-8.31038535e-01 8.49854529e-01 2.87258863e-01 -4.50365275e-01
4.87092763e-01 -6.04186058e-01 -1.78009182e-01 -5.15138768e-02
1.56872775e-02 5.11365414e-01 4.31752384e-01 -1.10828626e+00
-2.81925291e-01 -6.00208819e-01 3.67473960e-01 2.08595589e-01
-1.09126890e+00 7.40187109e-01 2.55164713e-01 -4.67576653e-01
-5.82345545e-01 -8.86600137e-01 -1.84113115e-01 9.83967632e-03
-1.18008554e-01 -2.84580290e-01 8.79729867e-01 -9.23221886e-01
1.54188621e+00 -2.25765586e+00 -5.12040406e-02 2.23311245e-01
4.19126421e-01 5.58461070e-01 3.21442783e-01 2.22649097e-01
-4.22904305e-02 7.64008164e-01 -4.35443997e-01 -3.38598996e-01
-4.65696424e-01 3.20386887e-01 3.09246522e-03 7.40991414e-01
3.48971963e-01 7.21136093e-01 -9.11327481e-01 -7.32862055e-01
-4.10480127e-02 7.12339059e-02 -6.97782516e-01 3.15234005e-01
2.50759214e-01 6.73870370e-02 -3.17261107e-02 7.58588612e-01
3.90695214e-01 2.69344151e-01 2.14531153e-01 3.30033213e-01
-6.23289347e-02 4.52847719e-01 -6.67471945e-01 1.11356390e+00
-4.73675251e-01 5.38524210e-01 1.49583653e-01 -8.54704440e-01
8.22476625e-01 1.90637365e-01 6.19502723e-01 -9.75945115e-01
3.37313652e-01 1.06124960e-01 2.45188534e-01 -6.59539521e-01
4.38428611e-01 -3.00927699e-01 -2.76963413e-01 7.88093328e-01
-3.09742466e-02 9.56544057e-02 7.27770105e-02 3.14602077e-01
1.40706098e+00 -4.27899987e-01 5.82353413e-01 -2.20218182e-01
-3.70496839e-01 2.81861514e-01 6.29050612e-01 6.23554349e-01
-6.88549161e-01 -2.71144122e-01 5.87138593e-01 -6.51694313e-02
-1.45169663e+00 -9.48325038e-01 -3.84334624e-01 1.11764526e+00
-9.72727761e-02 -4.11231756e-01 -8.67095947e-01 -9.74880397e-01
-3.07184402e-02 1.34420133e+00 -5.02763808e-01 -6.25060380e-01
-6.69095755e-01 -9.98651743e-01 1.00648797e+00 4.36863542e-01
2.37838969e-01 -8.76866579e-01 -8.88352931e-01 2.34799273e-02
-1.66074261e-01 -7.63190806e-01 -1.95617811e-03 2.72813380e-01
-9.71410155e-01 -1.01235664e+00 1.33660715e-02 -1.94527693e-02
5.96232772e-01 -2.06135567e-02 1.07551479e+00 2.99644858e-01
-3.34047705e-01 1.11812375e-01 -4.69036102e-01 -6.74756050e-01
-1.91845253e-01 -5.05174398e-02 -1.20480053e-01 -4.86408979e-01
7.40776956e-01 -2.63848245e-01 -3.08247805e-01 -2.59429842e-01
-9.74726677e-01 -4.94737893e-01 2.15444103e-01 6.88987732e-01
-6.06501698e-02 -3.90438624e-02 6.78504527e-01 -1.48865926e+00
1.17077100e+00 -1.00412154e+00 -1.53269351e-01 9.53435451e-02
-9.62612808e-01 -5.25164306e-01 4.06917602e-01 -8.35008740e-01
-1.04440928e+00 -3.28349359e-02 -2.29511987e-02 1.77371856e-02
4.17277776e-02 3.88121545e-01 -1.60894558e-01 1.09925680e-01
1.11967754e+00 -5.50797880e-01 -5.19884005e-03 -3.23881805e-01
7.75387734e-02 8.66748273e-01 -9.59185287e-02 -3.58704001e-01
5.77543497e-01 5.60726598e-02 -4.29554582e-01 -6.15243852e-01
-1.01102674e+00 -3.63461792e-01 -6.26685500e-01 -2.54679382e-01
3.84901702e-01 -4.04418558e-01 -5.16489685e-01 2.34060645e-01
-6.77222550e-01 -4.86360878e-01 -2.95987189e-01 3.39520872e-01
-2.52201200e-01 1.21260442e-01 -4.82989818e-01 -1.34218907e+00
-5.38288713e-01 -7.26728618e-01 4.03594047e-01 8.90008211e-02
-1.05008090e+00 -9.34079230e-01 2.60093510e-01 6.81666553e-01
2.74949759e-01 4.20243055e-01 1.04589546e+00 -1.19224739e+00
1.55987546e-01 -5.91217279e-01 1.22616710e-02 3.01865578e-01
2.40519315e-01 8.86756256e-02 -9.03281689e-01 -4.19327170e-01
1.75761089e-01 -8.89034092e-01 2.31243685e-01 -1.09631911e-01
1.10063839e+00 -8.97222579e-01 -1.85996350e-02 -2.39054356e-02
1.29429913e+00 7.01231539e-01 6.30822122e-01 2.24420354e-01
5.55442691e-01 7.59254038e-01 8.20630312e-01 4.49282467e-01
-5.05810082e-02 5.91626525e-01 1.01753689e-01 -6.30621687e-02
2.29880214e-01 -1.62820786e-01 2.75292993e-01 7.34024076e-03
3.59216154e-01 -1.85670599e-01 -1.29062915e+00 5.86703598e-01
-1.48358357e+00 -1.13954675e+00 1.37497723e-01 2.24370480e+00
1.22719979e+00 6.82321548e-01 6.38163269e-01 4.50103223e-01
5.12161016e-01 -1.09125767e-02 -6.80159450e-01 -1.06714237e+00
3.70683014e-01 4.32200581e-01 5.37563384e-01 7.46943474e-01
-8.02661777e-01 5.84176004e-01 7.33522224e+00 6.07012033e-01
-1.02879333e+00 5.72279513e-01 9.52642500e-01 -7.60743976e-01
-2.08872542e-01 5.68591766e-02 -4.65107679e-01 8.12153518e-01
1.28773439e+00 -4.19547290e-01 3.40106398e-01 1.12504697e+00
2.09299967e-01 -2.70752877e-01 -8.72204185e-01 5.06972253e-01
2.00313881e-01 -8.43827724e-01 -1.59713447e-01 4.36404079e-01
4.02674317e-01 -4.14648682e-01 -1.65061444e-01 3.71783763e-01
5.10784566e-01 -1.10530925e+00 7.94598818e-01 1.99635327e-01
7.36724079e-01 -1.07213318e+00 7.08573520e-01 7.50523865e-01
-7.18416646e-02 -5.12479842e-01 -7.33775273e-02 -2.70410866e-01
3.12517583e-02 4.85165387e-01 -9.20950055e-01 -2.65708059e-01
7.03790069e-01 7.78870657e-02 -7.38306224e-01 9.61559951e-01
-1.21396266e-01 1.11045587e+00 -3.38718414e-01 -1.82291016e-01
1.15946561e-01 2.65258968e-01 2.90282249e-01 1.23753393e+00
-4.39339429e-01 3.16261232e-01 1.74753323e-01 5.70795178e-01
-2.78858691e-01 2.44402215e-01 -9.70376313e-01 -5.05678892e-01
9.42027628e-01 1.27044022e+00 -4.51358855e-01 -1.84391558e-01
-1.82229295e-01 3.58128607e-01 9.27218676e-01 -1.94998339e-01
-6.99890852e-01 -2.68758059e-01 3.51070911e-01 4.28300947e-01
-7.11381733e-01 -2.37265334e-01 -7.48587489e-01 -6.98582947e-01
-3.29409361e-01 -9.12554860e-01 5.77858329e-01 -3.23474914e-01
-1.23280478e+00 9.14285406e-02 3.81337076e-01 -6.14740551e-01
-3.01186740e-01 -3.27099204e-01 -4.81640160e-01 8.08998108e-01
-6.08563542e-01 -9.58602905e-01 -9.17345583e-02 1.80000812e-01
4.22537625e-01 2.04333782e-01 9.25403535e-01 5.21925628e-01
-9.35783625e-01 1.08592105e+00 -4.03301984e-01 3.51421446e-01
6.24013901e-01 -1.06015813e+00 3.77483368e-02 8.82222712e-01
-1.91452533e-01 1.16861022e+00 6.22727096e-01 -1.05365491e+00
-8.95704269e-01 -7.89992273e-01 7.60144711e-01 -8.05694461e-01
6.96886957e-01 -5.98669529e-01 -1.04358959e+00 9.03137326e-01
-6.79943338e-02 -4.21715677e-01 1.31339371e+00 4.32859629e-01
-4.06051248e-01 1.08974725e-01 -1.82183385e+00 5.97787261e-01
1.02690828e+00 -5.19704163e-01 -5.81434667e-01 1.79720089e-01
4.00111586e-01 -1.58247113e-01 -8.99633169e-01 4.39704090e-01
4.69151586e-01 -4.54450727e-01 6.48971915e-01 -1.06949401e+00
4.79458094e-01 6.67017758e-01 5.03959358e-02 -1.07624960e+00
-3.30068678e-01 -3.51425141e-01 -4.50638324e-01 1.52012682e+00
6.48358703e-01 -2.88315207e-01 7.97295272e-01 1.42902231e+00
1.84567556e-01 -8.98953438e-01 -9.29952860e-01 -6.23621464e-01
3.42414916e-01 -3.33321154e-01 1.86887622e-01 1.54588902e+00
3.04657221e-01 2.71787494e-01 -4.33241785e-01 -2.83857912e-01
7.05353439e-01 -6.30258560e-01 5.38399220e-01 -1.06482065e+00
1.51093319e-01 -1.88862637e-01 -3.14592063e-01 1.53341174e-01
2.97760606e-01 -5.75403273e-01 -2.52654731e-01 -1.11837840e+00
5.83280623e-01 -6.49584830e-01 -3.47394869e-02 9.19368923e-01
-3.93573821e-01 4.38428015e-01 4.12691385e-02 3.05021197e-01
-2.71714896e-01 -8.50393400e-02 6.26094878e-01 -4.34578657e-02
-2.88318694e-01 -4.26503867e-01 -1.13218868e+00 7.27818429e-01
1.02375162e+00 -9.16920006e-01 -5.95721543e-01 -2.33402103e-01
2.19428986e-01 -8.25761706e-02 2.35020474e-01 -3.99536520e-01
-4.97069731e-02 -8.01155269e-01 4.73600954e-01 5.84950745e-02
5.82537353e-01 -8.44669998e-01 -3.32345106e-02 5.91661930e-01
-8.56353104e-01 1.40493875e-02 2.20195428e-01 1.33474186e-01
2.36759871e-01 -8.70031536e-01 1.05180705e+00 -1.47186816e-01
-2.63144821e-01 6.66800961e-02 -2.92160541e-01 2.65333503e-01
1.41394269e+00 -1.90529466e-01 -6.89053535e-01 -1.35181695e-01
-5.11536002e-01 -2.61467665e-01 5.90854347e-01 2.77883768e-01
3.43476206e-01 -9.69484806e-01 -4.61695999e-01 -2.70478517e-01
2.90296882e-01 -4.33431298e-01 -1.01809204e-01 5.20221889e-01
-5.12752473e-01 1.56726494e-01 -3.40517014e-01 1.67284552e-02
-1.51225913e+00 5.60925961e-01 2.85744458e-01 -1.47367939e-01
-2.92325556e-01 6.93410933e-01 -5.26680052e-01 -2.29993165e-01
2.97772825e-01 7.79170036e-01 -1.32199958e-01 3.47213626e-01
4.19255942e-01 6.87530935e-01 2.36409456e-02 -4.31448936e-01
-4.34599876e-01 -4.47217256e-01 -4.93474931e-01 -3.54120374e-01
1.38245702e+00 2.80037314e-01 -7.72187050e-05 4.81980115e-01
1.18381333e+00 4.54037517e-01 -9.33100164e-01 1.01960758e-02
4.23022836e-01 -1.11717832e+00 3.57719749e-01 -1.02016211e+00
-7.18372047e-01 5.87776303e-01 5.26581407e-01 3.75490099e-01
8.44495714e-01 -1.53986350e-01 4.61849540e-01 2.10274205e-01
4.83556032e-01 -1.34679079e+00 4.71809685e-01 -2.02386558e-01
5.82161188e-01 -1.08590090e+00 4.99696136e-01 -3.72763306e-01
-8.18290114e-01 5.11028469e-01 1.04758859e+00 -6.38193786e-02
3.80905300e-01 5.52889943e-01 -8.47243965e-02 -3.34712863e-01
-5.25476158e-01 3.12840879e-01 -2.48282254e-01 3.78574431e-01
6.24041617e-01 1.27547875e-01 -9.06264484e-01 4.43026185e-01
-3.01076263e-01 -2.55836725e-01 5.25635779e-01 1.40873921e+00
-4.43840772e-01 -1.09192717e+00 -2.38891795e-01 9.63893592e-01
-9.59334552e-01 -4.09857333e-02 -1.20138776e+00 9.92354870e-01
1.54984742e-01 1.30752647e+00 2.82757670e-01 -2.06996351e-01
2.25464180e-01 2.66201079e-01 4.51985270e-01 -9.79232192e-01
-9.77445662e-01 -2.86896437e-01 5.92950284e-01 -5.33770859e-01
-2.54948109e-01 -7.18190670e-01 -9.24699962e-01 -9.61214185e-01
-5.99652410e-01 9.12245587e-02 5.60649335e-01 9.59090531e-01
3.27357836e-02 -2.35798880e-01 7.33468711e-01 1.84522122e-01
-6.94931984e-01 -1.15140235e+00 -3.40468854e-01 7.89632022e-01
-6.89837411e-02 -8.25849712e-01 -4.59092349e-01 -3.27244431e-01]
|
[8.689231872558594, 10.449369430541992]
|
3d0567ad-0bbf-42ef-9caf-42367452ef40
|
computational-choreography-using-human-motion
|
2210.04366
| null |
https://arxiv.org/abs/2210.04366v2
|
https://arxiv.org/pdf/2210.04366v2.pdf
|
Computational Choreography using Human Motion Synthesis
|
Should deep learning models be trained to analyze human performance art? To help answer this question, we explore an application of deep neural networks to synthesize artistic human motion. Problem tasks in human motion synthesis can include predicting the motions of humans in-the-wild, as well as generating new sequences of motions based on said predictions. We will discuss the potential of a less traditional application, where learning models are applied to predicting dance movements. There have been notable, recent efforts to analyze dance movements in a computational light, such as the Everybody Dance Now (EDN) learning model and a Cal Poly master's thesis, Take The Lead (TTL). We have effectively combined these two works along with our own deep neural network to produce a new system for dance motion prediction, image-to-image translation, and video generation.
|
['Trevor Kirkby', 'Patrick Perrine']
|
2022-10-09
| null | null | null | null |
['video-generation']
|
['computer-vision']
|
[ 2.29880452e-01 7.63253644e-02 -4.54487056e-02 1.00526989e-01
-4.17105913e-01 -4.09908503e-01 6.52693510e-01 -1.06940615e+00
-1.75279051e-01 7.02011645e-01 7.08995402e-01 -1.49296165e-01
4.24650669e-01 -8.03002715e-01 -7.34352887e-01 -4.91594791e-01
4.28497046e-01 5.42792141e-01 1.12138897e-01 -6.93026364e-01
2.46592194e-01 4.48793441e-01 -1.28941786e+00 7.69246399e-01
5.46276337e-03 3.73122483e-01 1.68716222e-01 1.41171408e+00
4.86522466e-02 1.62274444e+00 -8.31383348e-01 -4.45567846e-01
2.42519125e-01 -1.47478294e+00 -1.22357213e+00 3.83127704e-02
3.93270105e-01 -4.93715763e-01 -4.71267074e-01 5.64420998e-01
7.20228314e-01 5.86158693e-01 5.76896429e-01 -1.12020910e+00
-8.88616145e-01 6.20110571e-01 -3.48981112e-01 1.92265138e-01
6.33788049e-01 6.73220396e-01 7.56736875e-01 -6.13957107e-01
1.14524853e+00 1.26225162e+00 8.74385476e-01 1.30447876e+00
-9.50438201e-01 -4.38490093e-01 -6.51773155e-01 4.46185857e-01
-6.82889760e-01 -3.42347413e-01 6.49774432e-01 -3.77815485e-01
7.79960692e-01 4.19612229e-01 1.30059075e+00 1.64402783e+00
3.01346153e-01 1.35863757e+00 5.38247049e-01 -6.02198124e-01
-5.11278696e-02 -8.46504688e-01 -7.27481663e-01 8.08792651e-01
-2.85329908e-01 1.01662539e-01 -7.41948843e-01 3.49586755e-01
1.22904360e+00 -4.51265424e-01 -6.87649101e-02 3.06586146e-01
-1.76264191e+00 8.60440016e-01 4.09848273e-01 5.64825296e-01
-3.31340641e-01 8.53343904e-01 3.22610319e-01 1.45023346e-01
6.92384616e-02 9.46372330e-01 -6.21323325e-02 -6.16421878e-01
-1.26753807e+00 9.93541658e-01 7.18043208e-01 6.38386726e-01
4.29963589e-01 7.44825304e-01 -4.89491403e-01 2.56406426e-01
3.43506709e-02 2.85307407e-01 8.44946682e-01 -1.79425442e+00
1.41218171e-01 1.95596457e-01 3.42373885e-02 -1.16398799e+00
-5.67100406e-01 1.28978819e-01 -8.17374706e-01 3.96202266e-01
6.88177884e-01 -5.24986625e-01 -6.94782972e-01 1.38341963e+00
1.46553600e-02 4.41406101e-01 1.08363584e-01 1.19121218e+00
1.02798975e+00 9.29442823e-01 -4.89951074e-02 8.07852298e-02
9.78763402e-01 -1.43038201e+00 -8.13790381e-01 -6.92976341e-02
8.08194339e-01 -1.02580094e+00 1.23272073e+00 2.55656689e-01
-1.56416202e+00 -9.08626199e-01 -9.49740231e-01 -4.86076266e-01
2.01151922e-01 3.22813749e-01 7.05717206e-01 3.21344465e-01
-1.23488116e+00 1.11851263e+00 -1.06949985e+00 -5.10800421e-01
3.75518233e-01 2.67421365e-01 -2.07331143e-02 4.59696919e-01
-7.61149049e-01 8.96405995e-01 2.25762740e-01 2.78470159e-01
-7.53904283e-01 -3.49020243e-01 -6.01538539e-01 -2.88852423e-01
-2.40738150e-02 -1.26408863e+00 1.58686709e+00 -1.48690164e+00
-1.95158505e+00 9.55139816e-01 -8.49556774e-02 -4.56627578e-01
7.75869668e-01 -2.95078486e-01 -2.66109288e-01 -8.97796303e-02
1.25698864e-01 1.15426874e+00 3.33561420e-01 -9.50377584e-01
-5.89904726e-01 5.29832114e-03 -8.15618187e-02 2.08676443e-01
1.61490411e-01 -3.16611305e-02 -2.93751657e-01 -1.05069792e+00
-2.71721244e-01 -1.39983535e+00 -2.36782685e-01 -2.33143762e-01
-4.52338398e-01 -4.83667329e-02 8.54422867e-01 -8.06169689e-01
1.08432317e+00 -1.56936121e+00 6.45404816e-01 -3.42207104e-01
-3.16388644e-02 4.79414642e-01 -2.77575105e-01 3.39068741e-01
2.58919448e-01 -7.39244744e-02 6.02991320e-03 -1.51239187e-01
-7.59355957e-03 2.54274726e-01 -4.05843467e-01 -6.61436170e-02
2.38233849e-01 1.55215597e+00 -8.60682011e-01 -3.71079087e-01
1.86177827e-02 3.25597584e-01 -6.75383151e-01 2.26605475e-01
-6.00254059e-01 1.15048778e+00 -4.62604135e-01 4.44365114e-01
-2.50472754e-01 -2.75217414e-01 9.36470851e-02 -1.70957118e-01
5.81542123e-03 -1.27629846e-01 -8.47286761e-01 2.13527203e+00
-2.45607793e-01 1.46601224e+00 -5.06057858e-01 -5.93286932e-01
7.21438646e-01 4.87515688e-01 6.28897011e-01 -5.34869015e-01
2.63914347e-01 2.34188661e-01 1.84220225e-01 -9.09736574e-01
1.04062343e+00 -2.91497260e-01 -3.47804837e-02 7.77536869e-01
-9.37621947e-03 -1.89976349e-01 2.01522335e-01 -2.98591375e-01
1.18202353e+00 1.12645948e+00 7.66428411e-02 7.57743940e-02
3.60505313e-01 8.34461391e-01 2.19761550e-01 4.86490637e-01
-1.68728963e-01 9.83518898e-01 2.15125963e-01 -1.15858686e+00
-1.53841352e+00 -6.36978805e-01 1.02322555e+00 1.25343907e+00
-1.35682762e-01 -1.87425882e-01 -1.11411357e+00 -2.10010007e-01
-6.10650301e-01 6.91695571e-01 -6.32217050e-01 -6.21303276e-04
-1.54177225e+00 -7.24061012e-01 1.19612670e+00 8.39158833e-01
5.25703490e-01 -1.93183815e+00 -1.14038503e+00 3.77572745e-01
-3.14719170e-01 -8.91655266e-01 -5.86068273e-01 -4.67198968e-01
-7.06423759e-01 -9.69255805e-01 -1.13630354e+00 -1.09285057e+00
-1.52005062e-01 -3.09566855e-01 1.33867002e+00 2.15221018e-01
-4.55852121e-01 2.93772578e-01 -3.29009414e-01 -3.50456774e-01
-9.12330389e-01 1.91083804e-01 -4.17735614e-02 -3.22774678e-01
3.21913421e-01 -5.31082511e-01 -8.36003423e-01 9.67519507e-02
-8.60668659e-01 4.95900363e-01 3.95889372e-01 7.22215235e-01
3.46303105e-01 -6.54750168e-01 2.95097023e-01 -5.67452967e-01
7.55790472e-01 -2.99735162e-02 1.00464653e-02 3.88116762e-02
1.98258072e-01 1.35552168e-01 4.34379250e-01 -7.24460304e-01
-1.03255320e+00 3.96318972e-01 -4.53251004e-01 -4.09458280e-01
-1.01455795e-02 7.63379559e-02 2.60073781e-01 2.21806496e-01
9.78805661e-01 1.51352361e-01 -9.49869081e-02 -8.69601741e-02
7.05603361e-01 1.70268461e-01 1.11484599e+00 -4.87206101e-01
6.65461063e-01 5.69275618e-01 2.93952346e-01 -8.20719182e-01
-6.77034080e-01 1.96548283e-01 -9.46503162e-01 -5.18757403e-01
1.73109376e+00 -7.95133233e-01 -1.03957665e+00 3.17002177e-01
-1.44778204e+00 -1.05926681e+00 -6.19260132e-01 4.60129589e-01
-1.21137321e+00 -1.78674068e-02 -9.66808081e-01 -4.08488393e-01
-2.94681042e-01 -1.15499032e+00 1.13308442e+00 4.09330606e-01
-1.05664873e+00 -1.04865742e+00 5.27641356e-01 4.81394202e-01
1.90322340e-01 8.67271543e-01 6.35708869e-01 -1.45230323e-01
-7.00253725e-01 2.69875795e-01 3.35918844e-01 -5.63892052e-02
-1.42422980e-02 1.33800134e-01 -6.88767433e-01 1.77888408e-01
-3.75440180e-01 -6.29086018e-01 5.89859784e-01 5.73750675e-01
8.60858738e-01 -2.31440306e-01 3.50366980e-02 8.29908729e-01
9.23915148e-01 2.45891139e-01 9.37316775e-01 6.05771244e-01
1.21384132e+00 4.73321140e-01 3.94913167e-01 3.22046489e-01
3.18956971e-01 8.04261446e-01 5.09251542e-02 4.56271879e-02
-8.20315599e-01 -4.06953752e-01 4.99217302e-01 9.53783989e-01
-1.02084839e+00 -3.19251418e-01 -9.45607722e-01 2.83003628e-01
-1.96841741e+00 -1.59652841e+00 -6.33284628e-01 1.26696432e+00
6.97321594e-01 -2.57659346e-01 4.88149703e-01 -7.04861954e-02
3.42946321e-01 2.54798204e-01 -2.91487336e-01 -5.63104331e-01
-2.04346105e-01 5.08423567e-01 6.57042712e-02 3.57889354e-01
-9.62924421e-01 1.42217875e+00 7.00167847e+00 6.80403054e-01
-1.12571001e+00 -4.23793085e-02 6.78433418e-01 -2.54805565e-01
-1.59747705e-01 -1.47114515e-01 -5.88654160e-01 1.44459188e-01
9.80012000e-01 1.08317688e-01 5.76905727e-01 5.80578685e-01
5.17686725e-01 2.65445054e-01 -1.17845309e+00 9.72057998e-01
3.21808636e-01 -2.17179132e+00 3.17175597e-01 1.75649524e-01
1.24012399e+00 -4.35962915e-01 -1.38452798e-02 2.56268561e-01
5.56844234e-01 -1.49342704e+00 9.28705633e-01 1.00964880e+00
6.03456318e-01 -6.84070230e-01 5.07899880e-01 2.90156513e-01
-1.09210026e+00 1.56097457e-01 -1.58715561e-01 -5.29243946e-01
3.89323592e-01 -5.00494421e-01 -7.08221793e-01 2.50243366e-01
5.46512067e-01 7.87822843e-01 -4.54871804e-01 4.81623799e-01
2.70661004e-02 4.95306492e-01 3.62452626e-01 -3.35372776e-01
4.96012121e-01 -1.85394868e-01 3.81918192e-01 1.08963084e+00
8.09385955e-01 2.70200104e-01 -6.70435280e-02 1.15119600e+00
-1.31951734e-01 -1.43503442e-01 -4.79190350e-01 -1.67372748e-01
-1.26141950e-01 8.47071886e-01 -9.13200855e-01 -5.01121581e-01
-2.89599206e-02 1.47998822e+00 -9.85235423e-02 2.69155025e-01
-9.24392045e-01 6.44991249e-02 4.09928858e-01 2.17125207e-01
1.30848959e-01 -3.87913674e-01 -4.22636122e-01 -9.74075854e-01
-4.48977441e-01 -1.10938787e+00 -8.61948915e-03 -1.33579588e+00
-9.56364393e-01 6.69848859e-01 -3.59729558e-01 -1.37370181e+00
-9.81678367e-01 -6.88504934e-01 -9.28855121e-01 5.30058801e-01
-4.14433450e-01 -1.45687068e+00 -3.26817930e-01 6.16837084e-01
1.03619516e+00 -4.89104599e-01 7.90435255e-01 9.37824845e-02
-2.75655985e-01 2.02140763e-01 -1.37400568e-01 4.96930361e-01
6.20208085e-01 -1.14347506e+00 6.66262746e-01 5.82407236e-01
7.03005254e-01 2.73126308e-02 9.52059209e-01 -6.06367588e-01
-1.39503026e+00 -1.00258493e+00 7.58463681e-01 -1.07581103e+00
5.42926252e-01 3.15141678e-01 -5.03660500e-01 9.78193820e-01
4.99034554e-01 -3.93584073e-01 5.65937400e-01 -7.22210288e-01
4.57060784e-01 4.48021978e-01 -4.59394455e-01 1.17749381e+00
1.22505760e+00 -1.80584252e-01 -6.95316732e-01 2.58956462e-01
5.37351906e-01 -7.21099436e-01 -7.63213813e-01 1.28022805e-01
8.55842650e-01 -9.54059422e-01 1.06335306e+00 -1.03983676e+00
1.40146708e+00 -3.11689734e-01 -4.72527929e-02 -1.16155660e+00
-2.28021741e-01 -1.05692732e+00 -1.76951885e-01 8.42985630e-01
1.49180740e-01 6.74760282e-01 1.37506318e+00 3.42555106e-01
-3.66046757e-01 -3.77787590e-01 -5.16945302e-01 -4.32079375e-01
2.77936906e-01 -4.57257748e-01 5.42757571e-01 9.62016284e-01
-3.13538909e-01 4.83966261e-01 -1.06283116e+00 -5.34030795e-01
4.82668839e-02 1.35483012e-01 1.44493151e+00 -7.84806311e-01
-5.53020000e-01 -6.05420113e-01 -4.42077816e-01 -9.63152349e-01
1.77325562e-01 -7.67707229e-01 2.06785798e-01 -1.81512523e+00
-3.88006978e-02 4.84657705e-01 6.22180998e-01 3.12935412e-01
-7.29014724e-02 8.70660007e-01 6.38729274e-01 5.51906884e-01
-4.58958179e-01 4.39990610e-01 1.99483085e+00 -4.21706550e-02
-4.87128168e-01 1.81964580e-02 -3.43067676e-01 1.16457343e+00
6.49367690e-01 -1.46409482e-01 -5.59742153e-02 -6.89990997e-01
3.61435443e-01 3.16273630e-01 4.61272836e-01 -1.46120477e+00
1.46754816e-01 -3.15681726e-01 9.01434004e-01 -5.15556097e-01
5.10984242e-01 -2.25595132e-01 5.71653128e-01 8.48374605e-01
-5.97094595e-01 5.50322294e-01 5.00259735e-02 2.65258193e-01
-2.45311260e-01 9.70426798e-02 5.74742675e-01 -7.42603302e-01
-1.08644938e+00 -5.57087995e-02 -6.23457253e-01 6.25692308e-02
7.80936062e-01 -4.64459926e-01 -2.36683547e-01 -7.65183330e-01
-1.22426498e+00 -2.99263984e-01 3.44643146e-01 4.72623736e-01
5.10444224e-01 -1.84757864e+00 -9.53729570e-01 -3.22705954e-01
-3.25228304e-01 -2.16549933e-01 1.20356202e-01 4.77426022e-01
-1.59867537e+00 2.21348256e-01 -8.26452672e-01 -4.85203177e-01
-1.13361228e+00 2.46012762e-01 3.30666363e-01 -3.42251718e-01
-7.59272039e-01 7.27724850e-01 -2.22754776e-01 -2.57073611e-01
-1.92866892e-01 -2.20206439e-01 -2.40287423e-01 -2.38879845e-01
6.31334305e-01 6.90785766e-01 -5.71498573e-01 -1.05237997e+00
1.16462268e-01 6.32428586e-01 8.37589681e-01 -4.97048378e-01
1.39664721e+00 1.39955267e-01 8.07535648e-02 7.27021873e-01
7.23223150e-01 -1.42991124e-03 -1.43235648e+00 4.16721195e-01
1.63270518e-01 -1.09509677e-01 -6.45729542e-01 -5.17602146e-01
-1.04553092e+00 9.92249787e-01 5.47057092e-01 -5.28633408e-02
9.91075933e-01 -1.27901375e-01 1.39346802e+00 5.07580459e-01
9.02221054e-02 -1.07502806e+00 8.79780173e-01 6.04599476e-01
9.48818386e-01 -9.47040379e-01 -1.06855676e-01 3.38004261e-01
-1.13710976e+00 1.64945197e+00 7.41167486e-01 -6.02253020e-01
1.25797912e-02 3.99935432e-02 1.94460258e-01 -9.44283828e-02
-6.37411892e-01 -1.96233064e-01 5.25093615e-01 6.82951987e-01
8.55328977e-01 -5.73606677e-02 -1.50122449e-01 5.21154583e-01
-9.24730301e-01 5.28000712e-01 9.60753381e-01 6.66710138e-01
-3.10588568e-01 -1.16733873e+00 -6.18895888e-01 -3.56248803e-02
-4.27677572e-01 1.75675675e-01 -7.35617995e-01 1.07339644e+00
4.36526746e-01 6.36762559e-01 1.86117291e-01 -7.40958691e-01
2.95053065e-01 1.46135747e-01 7.77167141e-01 -4.44937766e-01
-7.96467423e-01 7.40867555e-02 1.14218697e-01 -6.91072166e-01
-9.13882077e-01 -5.04820764e-01 -1.31175470e+00 -7.97419429e-01
5.00788569e-01 -3.59132409e-01 1.28107578e-01 1.03943706e+00
-1.73673369e-02 9.40896451e-01 -6.01675510e-02 -1.50131023e+00
-2.81201825e-02 -1.04694772e+00 -2.78206319e-01 6.22105598e-01
1.19248442e-01 7.94236436e-02 2.88058341e-01 8.78929555e-01]
|
[7.271836757659912, -0.12449961155653]
|
fad679ee-6e00-4284-95fb-0a6c92b84e33
|
context-sensitive-neocortical-neurons
|
2207.07338
| null |
https://arxiv.org/abs/2207.07338v6
|
https://arxiv.org/pdf/2207.07338v6.pdf
|
Context-sensitive neocortical neurons transform the effectiveness and efficiency of neural information processing
|
Deep learning (DL) has big-data processing capabilities that are as good, or even better, than those of humans in many real-world domains, but at the cost of high energy requirements that may be unsustainable in some applications and of errors, that, though infrequent, can be large. We hypothesise that a fundamental weakness of DL lies in its intrinsic dependence on integrate-and-fire point neurons that maximise information transmission irrespective of whether it is relevant in the current context or not. This leads to unnecessary neural firing and to the feedforward transmission of conflicting messages, which makes learning difficult and processing energy inefficient. Here we show how to circumvent these limitations by mimicking the capabilities of context-sensitive neocortical neurons that receive input from diverse sources as a context to amplify and attenuate the transmission of relevant and irrelevant information, respectively. We demonstrate that a deep network composed of such local processors seeks to maximise agreement between the active neurons, thus restricting the transmission of conflicting information to higher levels and reducing the neural activity required to process large amounts of heterogeneous real-world data. As shown to be far more effective and efficient than current forms of DL, this two-point neuron study offers a possible step-change in transforming the cellular foundations of deep network architectures.
|
['Khubaib Ahmed', 'Mohsin Raza', 'Mario Franco', 'Ahsan Adeel']
|
2022-07-15
| null | null | null | null |
['audio-signal-processing']
|
['audio']
|
[ 3.28258306e-01 1.70103893e-01 2.99214751e-01 -6.54757470e-02
-1.30106419e-01 -2.90487409e-01 6.25498593e-01 5.91035485e-01
-9.13517356e-01 9.34154451e-01 2.71648075e-02 -3.88097800e-02
-1.54506043e-01 -1.13223326e+00 -6.91172838e-01 -1.08077180e+00
-2.85806149e-01 3.97935629e-01 4.13608760e-01 -3.33824992e-01
2.46849716e-01 5.91466606e-01 -1.79286265e+00 3.80784035e-01
2.90284812e-01 9.52059984e-01 3.95299166e-01 3.91461849e-01
-4.79530096e-02 1.04252875e+00 -8.56720686e-01 5.57000227e-02
1.50824308e-01 -5.24601758e-01 -5.37209868e-01 -3.73004943e-01
-4.22416478e-02 5.84181631e-03 -1.13430984e-01 8.32968295e-01
7.10327506e-01 -7.24296644e-03 4.03208524e-01 -1.02049077e+00
-2.83099115e-01 4.69237506e-01 -1.55320987e-01 6.50478601e-01
-2.81872183e-01 3.35939586e-01 6.39886498e-01 -5.86751640e-01
5.35868526e-01 9.91275787e-01 3.85022223e-01 6.28438652e-01
-1.39783514e+00 -5.48845112e-01 1.17916696e-01 -5.24438955e-02
-1.10914302e+00 -9.81767833e-01 5.07330179e-01 -7.17186704e-02
1.46614742e+00 1.86652228e-01 1.10007906e+00 8.33181202e-01
9.28320169e-01 3.71112883e-01 8.85296285e-01 -4.05174047e-01
8.72044027e-01 -2.12846130e-01 -1.17111504e-01 4.10294771e-01
5.45445621e-01 -1.65136028e-02 -8.46355975e-01 5.61369443e-03
7.55264699e-01 -6.77063018e-02 -2.76100785e-01 1.48295850e-01
-1.34569216e+00 4.82538581e-01 5.43198347e-01 6.75691962e-01
-6.64693058e-01 5.33890009e-01 2.36043260e-01 7.26968765e-01
5.96172139e-02 6.73051298e-01 -3.38986933e-01 2.28356756e-02
-7.26399243e-01 1.62276775e-01 7.56429613e-01 2.90919900e-01
8.48647416e-01 2.07051188e-01 1.46360278e-01 6.13826394e-01
3.44637990e-01 2.97759175e-01 6.62239015e-01 -1.23149967e+00
1.51844665e-01 7.10424244e-01 -3.44031602e-01 -9.55166698e-01
-6.25581086e-01 -7.54581451e-01 -1.20141220e+00 7.80941963e-01
5.49999893e-01 -2.23573908e-01 -4.18484867e-01 2.06778598e+00
-1.65635377e-01 -3.18611056e-01 1.00177847e-01 1.02896559e+00
2.24019453e-01 7.54196465e-01 1.76626563e-01 -4.68981683e-01
1.23361087e+00 -2.89679676e-01 -5.70144475e-01 -8.24094772e-01
5.29066861e-01 -4.66690689e-01 7.33691514e-01 3.71925712e-01
-1.50276196e+00 -3.03221226e-01 -1.14752829e+00 -3.60414051e-02
-3.48597586e-01 -8.56731236e-01 6.96047127e-01 3.70861948e-01
-1.41373610e+00 4.95515168e-01 -6.55288041e-01 -2.18257755e-01
6.60511553e-01 5.92394054e-01 -2.19737530e-01 2.14652836e-01
-1.17106056e+00 9.17412758e-01 2.34568208e-01 6.40363768e-02
-6.87788308e-01 -6.71317518e-01 -3.72741073e-01 5.83328426e-01
3.65031175e-02 -9.14452851e-01 1.03718066e+00 -1.44399703e+00
-9.90113616e-01 6.42900884e-01 2.13631410e-02 -4.58795071e-01
8.02996159e-02 4.42400813e-01 2.75742617e-02 1.69273257e-01
-4.08696920e-01 9.39601302e-01 5.81352949e-01 -9.58274007e-01
-3.79136980e-01 -7.19300032e-01 -1.16888501e-01 5.26749671e-01
-3.29794556e-01 -6.29264861e-02 -2.02220660e-02 -4.53007907e-01
1.33688420e-01 -7.67930210e-01 -4.62083399e-01 3.07755381e-01
3.49159420e-01 -5.74252494e-02 4.87792134e-01 2.23534763e-01
8.38037848e-01 -2.07029557e+00 1.04393639e-01 2.86934823e-01
5.78715980e-01 6.15025237e-02 -2.40170494e-01 6.10501409e-01
2.48521179e-01 1.52653724e-01 -2.02993453e-01 -3.95275280e-02
-3.72981042e-01 4.81075078e-01 -1.37288734e-01 4.05798763e-01
2.72793174e-01 8.88688385e-01 -7.89217234e-01 -3.30881357e-01
-3.00940365e-01 6.66798413e-01 -3.68257016e-01 -1.68513358e-01
-1.61179453e-01 4.05712947e-02 -2.21557260e-01 2.29834214e-01
1.86220095e-01 -2.81162113e-01 3.40902656e-01 1.26983836e-01
-2.95418173e-01 3.19475889e-01 -1.08919215e+00 1.27351058e+00
-2.92937100e-01 9.17272449e-01 4.43651795e-01 -1.09676826e+00
7.34772086e-01 3.24434161e-01 5.01935303e-01 -1.31511068e+00
4.23445880e-01 3.39873165e-01 6.03956997e-01 2.60753054e-02
-1.82969477e-02 -4.00355875e-01 9.65665579e-02 9.20556664e-01
6.02146946e-02 -1.31947473e-01 3.22384238e-01 2.64122516e-01
1.40474117e+00 -5.61408043e-01 1.62185252e-01 -5.18208206e-01
1.05362542e-01 -2.48436943e-01 5.74996591e-01 7.38130212e-01
-3.72728825e-01 2.95073926e-01 7.60051668e-01 -3.51882875e-01
-1.03280497e+00 -8.52825940e-01 1.24916486e-01 1.20010102e+00
1.22362904e-01 -1.25147551e-01 -6.28497422e-01 1.21591762e-01
-2.69993991e-01 4.92555261e-01 -3.33424121e-01 -4.25138354e-01
-5.98263681e-01 -9.32819784e-01 5.88666379e-01 3.59897882e-01
3.88852715e-01 -1.45563257e+00 -1.57227564e+00 4.49515611e-01
8.06930587e-02 -6.85420871e-01 3.74183292e-03 1.06369293e+00
-1.02593899e+00 -7.36500442e-01 -3.71689886e-01 -8.20611835e-01
8.56219828e-01 2.13467568e-01 1.27762330e+00 4.53974932e-01
-2.81023115e-01 -8.99337009e-02 2.17333153e-01 -6.23197079e-01
-1.76185042e-01 -1.19385444e-01 -1.10499352e-01 -3.52816552e-01
3.81079733e-01 -9.16923344e-01 -7.30560064e-01 6.43855631e-02
-1.15618443e+00 1.63838223e-01 8.98989022e-01 7.42139518e-01
6.00103855e-01 3.30499977e-01 9.02564704e-01 -6.77347362e-01
8.89445841e-01 -5.70848942e-01 -3.27274412e-01 -1.60185575e-01
-7.40106821e-01 3.53001684e-01 7.47617543e-01 -2.64231324e-01
-8.73494565e-01 -1.47660255e-01 -1.54803559e-01 2.14524269e-01
2.53818221e-02 2.97148794e-01 -4.58541000e-03 -2.31213704e-01
9.47649419e-01 2.43678525e-01 1.87737629e-01 5.26648276e-02
-9.70076546e-02 1.44847915e-01 1.47695959e-01 -3.15944016e-01
1.86023757e-01 4.26116645e-01 2.46921882e-01 -7.49010980e-01
-1.00468673e-01 -2.23222375e-02 -3.85465026e-01 -4.80913937e-01
6.14104688e-01 -7.07152963e-01 -8.85840535e-01 5.54040313e-01
-1.09999454e+00 -5.59397280e-01 -5.81554949e-01 1.47007301e-01
-5.56772292e-01 -1.90240115e-01 -8.14835489e-01 -6.22946918e-01
-3.69479239e-01 -9.92630243e-01 4.45969522e-01 3.58071953e-01
-2.25452498e-01 -8.84421408e-01 9.34365205e-03 -1.11172117e-01
8.50850165e-01 -8.10364261e-02 1.25068927e+00 -5.21813750e-01
-7.54876018e-01 1.82103589e-02 -1.36541307e-01 -1.51576415e-01
-2.79622942e-01 -1.37404621e-01 -1.01370430e+00 -1.85497418e-01
2.61724085e-01 -5.53668380e-01 9.52606142e-01 2.99765527e-01
8.67444158e-01 -4.53972846e-01 -3.24636966e-01 1.52485386e-01
1.65442920e+00 1.51272103e-01 5.86725473e-01 3.53950709e-02
-6.30243355e-03 7.11858451e-01 -3.25133979e-01 3.15833539e-01
1.39381304e-01 1.87113494e-01 6.62811637e-01 -1.19502760e-01
-1.40841275e-01 2.42244095e-01 1.32532746e-01 9.66980100e-01
-1.52747005e-01 -4.73329633e-01 -9.08758581e-01 5.27133465e-01
-1.83276701e+00 -1.21496034e+00 -1.97349284e-02 2.11088657e+00
1.14222503e+00 6.16124988e-01 -2.67300367e-01 4.18813497e-01
2.41172478e-01 8.65347013e-02 -5.22162378e-01 -7.64010549e-01
-3.98428559e-01 2.79475033e-01 5.07471442e-01 2.08806470e-01
-2.40766943e-01 3.43715459e-01 6.73908663e+00 5.01922011e-01
-1.16101623e+00 2.29782194e-01 7.83234119e-01 -6.87764227e-01
-5.45576990e-01 -3.06833088e-01 -3.61918956e-01 5.30881286e-01
1.49424231e+00 -4.30085883e-03 7.54914403e-01 2.82786518e-01
2.99203336e-01 -5.81404626e-01 -1.24271500e+00 7.74279654e-01
-2.95673817e-01 -1.60885584e+00 -1.11078873e-01 3.10608149e-01
4.49106097e-01 3.22835892e-01 -8.11677501e-02 1.22948110e-01
2.89651781e-01 -1.11914945e+00 8.53719413e-01 4.56336617e-01
1.09624319e-01 -8.25480402e-01 6.90209985e-01 9.14552271e-01
-7.20530510e-01 -3.67500335e-01 -5.72197795e-01 -5.06200194e-01
-1.02906473e-01 8.68372440e-01 -3.58917713e-01 -3.56865495e-01
4.91386026e-01 -8.06767568e-02 -3.42265993e-01 9.15451765e-01
1.34510741e-01 3.36619675e-01 -6.65136278e-01 -4.09649491e-01
1.58193707e-01 2.44783789e-01 2.27989227e-01 1.06843626e+00
1.02323256e-01 3.30707848e-01 -3.89849007e-01 9.51143801e-01
-4.06915575e-01 -5.92181794e-02 -7.50165880e-01 4.90113627e-03
6.70353174e-01 1.20251417e+00 -1.00863516e+00 -1.95770875e-01
-2.81377286e-01 5.54875374e-01 4.88092840e-01 3.48956734e-01
-9.55290869e-02 -1.97228551e-01 5.20712674e-01 2.18891129e-01
-2.52833396e-01 -3.33741635e-01 -6.96914434e-01 -6.32203937e-01
-1.87211141e-01 -7.53558338e-01 6.73534051e-02 -6.62738502e-01
-9.75174725e-01 4.42547292e-01 -5.79367340e-01 -4.84534085e-01
-2.37411827e-01 -3.37041110e-01 -6.03785753e-01 8.33421886e-01
-1.41700721e+00 -5.39238214e-01 -1.44345105e-01 7.31190503e-01
5.45863628e-01 8.23661238e-02 7.81826317e-01 2.49803022e-01
-3.29429984e-01 1.98276877e-01 -1.45381771e-03 -2.49367237e-01
3.41134578e-01 -9.39776361e-01 1.08807079e-01 6.42315865e-01
8.27037469e-02 5.55134535e-01 5.22902012e-01 -3.46527636e-01
-1.71819854e+00 -7.68189132e-01 1.05511737e+00 -1.64248258e-01
2.10703358e-01 -5.98398745e-01 -9.49521244e-01 1.94843829e-01
3.32557648e-01 -1.55467521e-02 7.56342113e-01 -1.42008662e-01
-1.51921824e-01 -2.92308897e-01 -1.04544508e+00 8.38580549e-01
9.10592079e-01 -2.80032039e-01 -4.45125431e-01 -2.54383720e-02
3.63770485e-01 3.91431332e-01 -5.75537026e-01 -3.86479460e-02
5.04040837e-01 -1.29051077e+00 8.19000006e-01 -2.16688961e-01
2.10328192e-01 -1.12782471e-01 -1.44539028e-01 -1.28704596e+00
-4.60256875e-01 -2.99048215e-01 4.87346761e-02 7.37988293e-01
4.80576277e-01 -6.98393464e-01 5.43910444e-01 6.69715703e-01
-9.08769220e-02 -7.41559803e-01 -1.25195289e+00 -4.49227661e-01
1.52544469e-01 -3.48900825e-01 1.54613018e-01 6.09618545e-01
3.37777436e-01 4.82323796e-01 3.27067047e-01 -3.65073770e-01
7.50048578e-01 -2.83578962e-01 7.61645958e-02 -1.30840564e+00
-6.59923926e-02 -7.80705333e-01 -3.18019599e-01 -6.25495672e-01
-1.57338902e-01 -6.80920303e-01 2.98935592e-01 -1.58769238e+00
-3.63774002e-02 -6.32068098e-01 -6.39221728e-01 6.29853725e-01
3.55595320e-01 5.15330672e-01 2.67791510e-01 4.90136832e-01
-4.52853620e-01 2.58145016e-02 1.11073017e+00 2.12012250e-02
-9.58392471e-02 -3.58049273e-01 -7.96950161e-01 6.43581867e-01
7.48955727e-01 -4.92542416e-01 -6.37489855e-01 -6.23316705e-01
9.50921297e-01 -1.05970792e-01 2.72256166e-01 -1.30980062e+00
7.89949358e-01 -2.57152587e-01 7.22546995e-01 -1.69280380e-01
3.11125785e-01 -9.39308941e-01 1.25218943e-01 8.05950224e-01
-6.91504359e-01 -3.01378220e-02 1.45673990e-01 5.07629633e-01
-3.58552597e-02 -9.98562127e-02 1.02212155e+00 -5.92090607e-01
-6.28007889e-01 1.40056189e-03 -1.21277690e+00 7.36566409e-02
1.07982826e+00 -2.96401143e-01 -7.39607573e-01 -2.25806355e-01
-1.74430043e-01 -3.63158211e-02 5.16038120e-01 5.90952933e-02
6.31495476e-01 -9.94381964e-01 -5.25401056e-01 5.57574809e-01
-2.65221506e-01 -8.45063031e-02 8.33970308e-02 7.73968458e-01
-5.08260071e-01 4.87051755e-01 -7.06609905e-01 -2.55656540e-01
-7.88701057e-01 5.38865983e-01 5.50870061e-01 -3.02896257e-02
-4.03074175e-01 9.02023554e-01 1.90007210e-01 2.77457774e-01
1.34840354e-01 -1.80899009e-01 -5.22095039e-02 2.47580573e-01
6.02853894e-01 7.66288340e-02 4.14131612e-01 -1.38635516e-01
-4.55438882e-01 2.40114436e-01 1.73787370e-01 -2.07248732e-01
1.24212003e+00 -1.64008006e-01 -2.37843528e-01 4.98495132e-01
7.55007863e-01 -1.68409646e-01 -1.16359997e+00 1.36813372e-02
-1.40834853e-01 -3.64872813e-02 4.13716346e-01 -8.25360298e-01
-1.11561096e+00 1.16228437e+00 3.38687867e-01 4.70174074e-01
1.40041625e+00 -5.74338138e-02 7.18697011e-01 6.07969284e-01
5.22998035e-01 -1.33448255e+00 1.90783814e-01 4.02085155e-01
5.60127795e-01 -8.22533906e-01 -3.47294480e-01 2.63696551e-01
-2.18183354e-01 1.19290006e+00 6.73767507e-01 -7.68469721e-02
6.13531411e-01 9.85302389e-01 -6.42909482e-02 -3.22533667e-01
-1.69927406e+00 -2.42391732e-02 -4.30713683e-01 6.78482473e-01
5.20281076e-01 -2.16713458e-01 -3.95025164e-01 1.90062627e-01
2.43274540e-01 1.91237703e-01 5.72681487e-01 1.11756563e+00
-1.12880826e+00 -7.19623446e-01 -2.16396049e-01 5.64430058e-01
-5.27285516e-01 -2.01363504e-01 -2.44102806e-01 3.01120102e-01
2.67689407e-01 8.83064389e-01 6.64667606e-01 8.94149765e-02
-9.24290493e-02 1.32534966e-01 5.11315584e-01 -5.30751109e-01
-9.11456525e-01 1.54115662e-01 -6.86824247e-02 -4.09164608e-01
-2.61198312e-01 -2.52472430e-01 -1.72237813e+00 -5.33684194e-01
1.15513526e-01 -1.57269835e-01 7.50729501e-01 8.79505396e-01
4.15230215e-01 6.98367715e-01 2.99398452e-01 -7.58235395e-01
-3.70405197e-01 -4.18070525e-01 -4.41766351e-01 1.46389157e-01
4.15852457e-01 -2.39376366e-01 -2.76016891e-01 6.07202873e-02]
|
[8.131750106811523, 2.8970937728881836]
|
70152e28-3d5a-4cf2-bc99-850f7b74994d
|
domain-knowledge-empowered-structured-neural
|
2009.07373
| null |
https://arxiv.org/abs/2009.07373v2
|
https://arxiv.org/pdf/2009.07373v2.pdf
|
Domain Knowledge Empowered Structured Neural Net for End-to-End Event Temporal Relation Extraction
|
Extracting event temporal relations is a critical task for information extraction and plays an important role in natural language understanding. Prior systems leverage deep learning and pre-trained language models to improve the performance of the task. However, these systems often suffer from two short-comings: 1) when performing maximum a posteriori (MAP) inference based on neural models, previous systems only used structured knowledge that are assumed to be absolutely correct, i.e., hard constraints; 2) biased predictions on dominant temporal relations when training with a limited amount of data. To address these issues, we propose a framework that enhances deep neural network with distributional constraints constructed by probabilistic domain knowledge. We solve the constrained inference problem via Lagrangian Relaxation and apply it on end-to-end event temporal relation extraction tasks. Experimental results show our framework is able to improve the baseline neural network models with strong statistical significance on two widely used datasets in news and clinical domains.
|
['Nanyun Peng', 'Yichao Zhou', 'Rujun Han']
|
2020-09-15
| null |
https://aclanthology.org/2020.emnlp-main.461
|
https://aclanthology.org/2020.emnlp-main.461.pdf
|
emnlp-2020-11
|
['temporal-relation-extraction']
|
['natural-language-processing']
|
[ 4.10527796e-01 3.24197620e-01 -8.47198725e-01 -6.66482091e-01
-7.68686175e-01 -2.13065237e-01 6.22132778e-01 4.53686118e-01
-8.50010633e-01 1.00992179e+00 4.79569376e-01 -2.99217939e-01
-3.42121035e-01 -6.63461745e-01 -7.65588403e-01 -3.30429345e-01
-3.93253624e-01 7.01906621e-01 1.66863412e-01 6.03894144e-02
-1.63450047e-01 4.74456362e-02 -8.00070584e-01 5.49115002e-01
7.76202202e-01 9.90759432e-01 -2.27225218e-02 1.88647076e-01
-1.37223247e-02 1.34362364e+00 -2.52694339e-01 -2.31224403e-01
-3.37773636e-02 -2.44726121e-01 -1.01626825e+00 -4.13930595e-01
-4.73628044e-01 -4.11801338e-01 -3.53433460e-01 8.50887001e-01
2.83976555e-01 2.33332515e-01 6.01383388e-01 -9.66903031e-01
-2.96116173e-01 9.11690176e-01 -4.74285334e-01 5.36508441e-01
4.70678471e-02 -1.29661217e-01 1.23480785e+00 -6.71063066e-01
7.15299606e-01 1.08311713e+00 6.93728924e-01 3.75152618e-01
-1.36139977e+00 -5.73252320e-01 5.50156355e-01 4.20649856e-01
-1.22707033e+00 -3.39252710e-01 6.77465975e-01 -3.29323560e-01
1.27267325e+00 5.46418130e-03 3.25253636e-01 1.29739916e+00
3.90812606e-01 9.92458463e-01 9.06530380e-01 -2.15891138e-01
2.31128871e-01 -9.37463120e-02 3.30739796e-01 5.27490735e-01
-1.79539308e-01 1.67774349e-01 -7.02521563e-01 -2.32615232e-01
4.21041518e-01 -1.59124229e-02 -3.22315514e-01 3.88148949e-02
-1.34627259e+00 8.91772509e-01 4.14747000e-01 3.07242006e-01
-5.75394750e-01 -5.47285751e-02 6.63023710e-01 4.73432578e-02
7.49024987e-01 3.12643856e-01 -9.22402322e-01 -1.28141254e-01
-9.45336759e-01 7.16395453e-02 7.69102871e-01 5.92565417e-01
2.20908314e-01 -5.11049569e-01 -4.15906876e-01 7.67354190e-01
2.07217753e-01 -1.30303241e-02 3.94740283e-01 -4.72317934e-01
5.54118097e-01 4.78770345e-01 -8.73301178e-02 -8.72248173e-01
-7.94960201e-01 -3.87276471e-01 -8.95635188e-01 -3.61483753e-01
5.18889725e-01 -2.68553019e-01 -8.94177318e-01 2.12150526e+00
2.19475284e-01 5.01511216e-01 1.98760964e-02 8.05888355e-01
7.90931344e-01 5.48413157e-01 3.61965954e-01 -5.59874177e-01
1.53642273e+00 -8.27699780e-01 -1.26027095e+00 -3.68476361e-01
7.75145650e-01 -4.44422781e-01 6.40528977e-01 4.51239467e-01
-8.01796556e-01 -8.02214369e-02 -9.24560189e-01 -6.56701699e-02
-2.43245110e-01 6.64868653e-02 8.03607106e-01 1.51976928e-01
-6.16089940e-01 6.09448016e-01 -1.43669665e+00 -1.69533417e-01
6.23340845e-01 4.14137274e-01 -3.20067823e-01 1.28923720e-02
-1.66603470e+00 9.52512026e-01 7.42696106e-01 2.51563787e-01
-8.21363568e-01 -8.28249454e-01 -9.78077710e-01 4.62164767e-02
7.81635880e-01 -7.52333105e-01 1.33547306e+00 -6.86118960e-01
-1.35868573e+00 6.65532291e-01 -2.09991738e-01 -8.74136329e-01
5.87862015e-01 -5.40597677e-01 -2.66319364e-01 1.47250101e-01
7.08283633e-02 4.28530127e-01 3.12771857e-01 -5.56260526e-01
-6.02512181e-01 -1.78854585e-01 -2.00242791e-02 2.50446552e-04
-3.67143929e-01 2.23194242e-01 -6.89923584e-01 -7.10997701e-01
-4.23485786e-02 -8.80384028e-01 -5.67006171e-01 -4.64388840e-02
-6.93947971e-01 -4.73565459e-01 5.09076774e-01 -8.86826694e-01
1.33861899e+00 -1.92494464e+00 2.40686480e-02 1.94478661e-01
1.99201539e-01 2.74708897e-01 1.41048163e-01 1.27910241e-01
-3.13925803e-01 1.12086765e-01 -3.15607041e-01 -4.04817075e-01
-1.06967181e-01 4.20809418e-01 -4.06025499e-01 3.50277960e-01
3.94390851e-01 9.18180168e-01 -1.10813427e+00 -7.85030663e-01
-4.09875326e-02 5.15530467e-01 -6.35332644e-01 1.67271331e-01
-4.88985896e-01 6.37650907e-01 -6.45640135e-01 2.90433258e-01
3.88779521e-01 -5.26644826e-01 5.59472144e-01 -2.66542375e-01
2.38766462e-01 8.73538375e-01 -7.50439048e-01 1.93945146e+00
-4.90484059e-01 3.66153806e-01 -4.18615639e-01 -1.34626007e+00
4.01373267e-01 6.08799577e-01 8.31644058e-01 -6.29853666e-01
2.19435990e-01 -1.18317030e-01 -3.27722318e-02 -6.46385133e-01
-9.89380181e-02 -3.59546483e-01 1.84916034e-02 3.20767075e-01
1.92677137e-02 4.08212572e-01 1.05241358e-01 2.72327006e-01
1.17838788e+00 3.05496991e-01 4.96007085e-01 -1.42274469e-01
2.89967448e-01 -6.35534599e-02 1.31140149e+00 6.30966008e-01
-1.47319540e-01 5.35613298e-01 1.03528130e+00 -5.12414455e-01
-5.36011040e-01 -8.33573401e-01 -3.60709518e-01 7.84794331e-01
-2.62171477e-01 -6.24382138e-01 -2.97759265e-01 -1.06656873e+00
-3.62044394e-01 8.92393887e-01 -8.64472747e-01 -2.73378253e-01
-6.55245900e-01 -1.19775164e+00 4.57774907e-01 8.39850664e-01
2.07816213e-01 -9.44216728e-01 -5.59258342e-01 4.96316642e-01
-5.88726640e-01 -1.62529635e+00 -3.41224164e-01 3.56052846e-01
-7.97014296e-01 -1.01138783e+00 -5.44782996e-01 -4.18115348e-01
5.35089552e-01 -4.68196660e-01 1.28537369e+00 -3.65735292e-01
-1.91900268e-01 -1.42259151e-01 -2.40233764e-01 -5.28817236e-01
-1.05751224e-01 6.16048276e-02 2.90905274e-02 -5.18141910e-02
5.41971207e-01 -6.55257225e-01 -5.77708483e-01 1.71338841e-01
-8.26559067e-01 2.82203019e-01 7.37766564e-01 1.02877939e+00
7.58664191e-01 1.51423171e-01 6.30309403e-01 -1.09864414e+00
3.76330972e-01 -6.24339879e-01 -4.49615061e-01 2.92028666e-01
-5.69453359e-01 4.12972480e-01 4.45814639e-01 -5.44511914e-01
-1.22727740e+00 1.04225449e-01 -9.59352478e-02 -4.04189974e-01
-3.17918770e-02 1.27204537e+00 2.77493726e-02 7.35279381e-01
4.78210330e-01 5.91388531e-02 -2.26814106e-01 -3.54986697e-01
2.42851302e-01 2.48564318e-01 5.57087421e-01 -7.40472615e-01
2.85122216e-01 6.63575292e-01 2.54387427e-02 -4.11409557e-01
-1.43828356e+00 -4.75858390e-01 -5.78730285e-01 2.88249105e-01
1.09524620e+00 -9.20213342e-01 -7.12216437e-01 8.13062191e-02
-1.15488303e+00 -3.94894511e-01 -1.76966637e-02 8.99175346e-01
-3.12817156e-01 3.23547304e-01 -7.76861727e-01 -7.38788664e-01
-4.22081321e-01 -8.98932338e-01 9.41309631e-01 -1.06616929e-01
-4.48932678e-01 -1.14669895e+00 1.35111988e-01 1.44664600e-01
8.14243108e-02 5.10058105e-01 1.09637845e+00 -8.22606504e-01
-3.70146751e-01 -3.86747383e-02 -3.21022600e-01 1.37577176e-01
1.74945280e-01 -2.48785675e-01 -8.74344349e-01 7.31025860e-02
-1.17173502e-02 -4.60881114e-01 1.22616148e+00 6.46552026e-01
1.32654262e+00 -4.87152070e-01 -7.81944096e-01 4.70387489e-01
1.02002990e+00 -7.46969273e-03 3.54688227e-01 -9.87756997e-03
4.57349956e-01 7.49539435e-01 6.66787803e-01 6.02140129e-01
6.34553552e-01 7.36909091e-01 1.88095048e-01 2.16319244e-02
1.99674383e-01 -2.24189222e-01 2.06264496e-01 4.44130570e-01
-4.30105105e-02 -2.46779487e-01 -1.16304219e+00 6.95859432e-01
-2.32616973e+00 -7.01965272e-01 1.04951121e-01 2.01295590e+00
1.42092860e+00 5.82166672e-01 -5.46824634e-02 8.55026674e-03
3.54897290e-01 1.12465313e-02 -5.45440614e-01 2.47607846e-02
3.07049192e-02 2.13863090e-01 2.96191990e-01 2.92724699e-01
-1.30632818e+00 8.41467559e-01 5.38205433e+00 6.85864806e-01
-1.19662118e+00 2.17254132e-01 7.28337944e-01 -2.78601915e-01
3.68296243e-02 -3.95753831e-02 -7.36400008e-01 4.52924818e-01
1.14381003e+00 -6.25144399e-04 -4.94403616e-02 3.87258470e-01
4.10418034e-01 -1.02300480e-01 -1.52182937e+00 8.57654512e-01
-1.74822107e-01 -1.34514821e+00 -2.66247302e-01 -6.42652288e-02
5.72315693e-01 2.10023448e-01 4.50922251e-02 4.75021183e-01
6.02576673e-01 -1.18322754e+00 4.83512461e-01 6.14654779e-01
6.12120748e-01 -6.31433904e-01 9.42064404e-01 5.68148196e-01
-8.91424179e-01 1.67656213e-01 3.27187218e-02 -4.87420894e-02
5.16544521e-01 1.12287617e+00 -1.14162171e+00 7.38570154e-01
5.44809103e-01 1.02647066e+00 -3.90410870e-02 7.84166217e-01
-5.96679628e-01 9.12420452e-01 -6.69468045e-01 2.24943593e-01
2.49918208e-01 1.71917021e-01 4.21972513e-01 1.23867142e+00
-1.61906987e-01 3.14726949e-01 4.40557659e-01 9.98837948e-01
-8.76376107e-02 -9.47505683e-02 -3.28110486e-01 -2.11457923e-01
1.55609578e-01 1.06390166e+00 -8.76232862e-01 -2.32679114e-01
-5.21554828e-01 7.18150139e-01 4.48649913e-01 3.11138958e-01
-1.08259380e+00 -5.84792458e-02 3.36535126e-01 -2.31468305e-01
8.91352445e-02 -1.70043930e-01 -4.16655958e-01 -1.26611531e+00
9.68966912e-03 -6.49490952e-01 9.53093767e-01 -3.16168457e-01
-1.32714558e+00 5.42147160e-01 2.30687782e-01 -9.76997137e-01
-5.22570610e-01 -4.83936757e-01 -3.68742377e-01 8.13028634e-01
-1.72372651e+00 -1.24681628e+00 2.30355948e-01 4.40645605e-01
3.53639901e-01 2.36124620e-01 8.66575658e-01 4.70807076e-01
-7.85881639e-01 5.39452851e-01 -3.87622178e-01 4.82968152e-01
8.67259502e-01 -1.18811309e+00 1.29226178e-01 7.64651239e-01
9.86770838e-02 6.65051043e-01 6.13407969e-01 -7.07615495e-01
-9.77405012e-01 -1.17229056e+00 1.32182002e+00 -5.33656359e-01
6.22399509e-01 -3.49825799e-01 -9.52876687e-01 9.87567067e-01
-4.36517037e-02 2.77858764e-01 9.09011364e-01 6.26506329e-01
-3.13425750e-01 1.89082175e-01 -8.07642400e-01 3.53263438e-01
8.79592240e-01 -4.92885888e-01 -7.46605277e-01 6.36331141e-01
7.64574766e-01 -6.38321757e-01 -9.14585412e-01 6.73126101e-01
3.67165685e-01 -4.16444629e-01 9.94855344e-01 -9.32014406e-01
7.33451605e-01 -1.70675784e-01 8.44847858e-02 -9.65808749e-01
4.70286272e-02 -5.87028444e-01 -3.36160749e-01 9.75278139e-01
7.80110657e-01 -4.46714103e-01 5.41641593e-01 9.14756536e-01
6.80325627e-02 -1.05449820e+00 -8.26119959e-01 -6.08194411e-01
3.15937679e-03 -7.12003708e-01 1.90192953e-01 1.31429970e+00
3.23460460e-01 5.62108994e-01 -4.11933541e-01 3.64842772e-01
3.76506269e-01 4.48754169e-02 1.66816011e-01 -1.17491269e+00
-4.51381981e-01 -2.83451885e-01 -1.57910809e-02 -8.84631753e-01
4.96065527e-01 -9.16178524e-01 1.25989392e-01 -1.57044470e+00
4.69474822e-01 -3.00420463e-01 -7.43537486e-01 1.05500805e+00
-4.61872011e-01 -7.90789127e-02 -2.34236106e-01 -3.11211981e-02
-7.76694238e-01 5.27267694e-01 8.32270324e-01 -3.41317467e-02
-3.02613646e-01 4.54282723e-02 -5.05223691e-01 8.56743693e-01
6.07770860e-01 -7.33361006e-01 -4.60188985e-01 -4.43392724e-01
6.34231746e-01 3.78347248e-01 2.62264252e-01 -4.82641369e-01
3.36107999e-01 -2.81884074e-01 1.90849215e-01 -5.78471184e-01
2.04048142e-01 -7.43189991e-01 -2.26032689e-01 4.73148167e-01
-7.18064010e-01 -3.76899421e-01 3.52707207e-01 8.45498145e-01
-3.79312336e-01 2.16013998e-01 6.29254162e-01 3.86722349e-02
-5.30498147e-01 4.26270515e-01 -1.48870558e-01 2.21187770e-01
7.47087955e-01 5.59143245e-01 -7.71855004e-03 -4.38971579e-01
-1.04906321e+00 5.35287023e-01 -3.98727983e-01 3.91495526e-01
4.21779752e-01 -1.10016119e+00 -8.80361259e-01 -1.47829145e-01
4.41419799e-03 3.33721340e-01 2.04185039e-01 1.26542759e+00
-6.76732808e-02 6.19524240e-01 2.77310759e-01 -7.20673978e-01
-9.46590304e-01 7.13106573e-01 2.72114635e-01 -1.05450976e+00
-7.13108540e-01 8.94575894e-01 4.73515511e-01 -2.67139196e-01
3.40552062e-01 -7.54168212e-01 -3.48200113e-01 2.88521200e-02
4.97289091e-01 -2.29547724e-01 1.95384860e-01 -2.00055435e-01
-6.78884327e-01 2.56549157e-02 -5.16437650e-01 -2.32550636e-01
1.58543873e+00 2.10611001e-01 3.31398994e-02 3.77338737e-01
1.06050277e+00 -3.75329465e-01 -1.13004184e+00 -8.59437585e-01
2.91047692e-01 4.34797304e-03 3.97615343e-01 -1.04845500e+00
-9.98664260e-01 8.68303180e-01 1.98540047e-01 -3.26259464e-01
1.09438872e+00 2.73077153e-02 8.86226714e-01 5.31369209e-01
7.44843930e-02 -9.72403884e-01 -2.24484056e-02 5.82929611e-01
6.31896973e-01 -1.38270164e+00 1.59418508e-01 -4.83245522e-01
-6.02959991e-01 9.08804476e-01 4.10646707e-01 2.07996815e-01
1.00484681e+00 4.66388136e-01 -7.34705478e-02 -2.14561880e-01
-1.19385386e+00 -1.27629951e-01 6.07648849e-01 1.18825115e-01
6.19308293e-01 -1.77929193e-01 -4.43862498e-01 1.13473856e+00
1.33454293e-01 4.18647766e-01 -2.49436703e-02 8.62259388e-01
2.52295107e-01 -1.10962486e+00 1.98996022e-01 3.21862757e-01
-1.03350115e+00 -3.49291593e-01 -9.52308699e-02 5.85439444e-01
1.32851340e-02 8.86425078e-01 -1.04429245e-01 -4.94911037e-02
2.03307167e-01 1.33560836e-01 2.18388841e-01 -5.12206078e-01
-2.83308774e-01 1.81539103e-01 3.18649650e-01 -7.12337136e-01
-5.56273937e-01 -6.19754791e-01 -1.58076656e+00 2.37997323e-01
-3.36217612e-01 -4.00317088e-03 2.53790647e-01 1.41277087e+00
3.76251876e-01 9.61929798e-01 1.16392657e-01 -2.11523369e-01
-4.37604755e-01 -9.37753141e-01 -1.99405953e-01 5.02698839e-01
4.11661625e-01 -5.43453515e-01 8.32987055e-02 2.46966079e-01]
|
[8.936067581176758, 8.968766212463379]
|
547f2f5b-4f50-45e8-b69e-d451b18f8187
|
walk-and-learn-facial-attribute
|
1604.06433
| null |
http://arxiv.org/abs/1604.06433v3
|
http://arxiv.org/pdf/1604.06433v3.pdf
|
Walk and Learn: Facial Attribute Representation Learning from Egocentric Video and Contextual Data
|
The way people look in terms of facial attributes (ethnicity, hair color,
facial hair, etc.) and the clothes or accessories they wear (sunglasses, hat,
hoodies, etc.) is highly dependent on geo-location and weather condition,
respectively. This work explores, for the first time, the use of this
contextual information, as people with wearable cameras walk across different
neighborhoods of a city, in order to learn a rich feature representation for
facial attribute classification, without the costly manual annotation required
by previous methods. By tracking the faces of casual walkers on more than 40
hours of egocentric video, we are able to cover tens of thousands of different
identities and automatically extract nearly 5 million pairs of images connected
by or from different face tracks, along with their weather and location
context, under pose and lighting variations. These image pairs are then fed
into a deep network that preserves similarity of images connected by the same
track, in order to capture identity-related attribute features, and optimizes
for location and weather prediction to capture additional facial attribute
features. Finally, the network is fine-tuned with manually annotated samples.
We perform an extensive experimental analysis on wearable data and two standard
benchmark datasets based on web images (LFWA and CelebA). Our method
outperforms by a large margin a network trained from scratch. Moreover, even
without using manually annotated identity labels for pre-training as in
previous methods, our approach achieves results that are better than the state
of the art.
|
['Rogerio Schmidt Feris', 'Jing Wang', 'Yu Cheng']
|
2016-04-21
|
walk-and-learn-facial-attribute-1
|
http://openaccess.thecvf.com/content_cvpr_2016/html/Wang_Walk_and_Learn_CVPR_2016_paper.html
|
http://openaccess.thecvf.com/content_cvpr_2016/papers/Wang_Walk_and_Learn_CVPR_2016_paper.pdf
|
cvpr-2016-6
|
['facial-attribute-classification']
|
['computer-vision']
|
[ 2.53154431e-03 -4.25277621e-01 -6.68054447e-02 -8.60037446e-01
-3.75316918e-01 -7.64409184e-01 7.97217607e-01 4.03723009e-02
-5.11686921e-01 6.47810161e-01 3.47259879e-01 5.16429603e-01
4.41268571e-02 -8.56539905e-01 -7.28340149e-01 -5.88298559e-01
-4.12095189e-02 4.44518954e-01 -2.89252877e-01 -1.51121303e-01
-1.32045493e-01 5.61836541e-01 -2.03105092e+00 8.21709931e-02
3.20225686e-01 1.33697629e+00 -3.97245973e-01 4.25350726e-01
2.80869752e-01 4.50012922e-01 -1.96688160e-01 -9.90945458e-01
5.15481353e-01 -1.94203064e-01 -4.90419924e-01 3.08058560e-01
1.30593979e+00 -3.99444520e-01 -3.79616022e-01 8.80210459e-01
6.69457376e-01 1.21734388e-01 6.10177636e-01 -1.27591968e+00
-7.28265166e-01 1.06514283e-01 -5.12476206e-01 -1.42041370e-01
6.59067690e-01 2.95152694e-01 9.15025532e-01 -8.73657823e-01
6.97369456e-01 1.16518712e+00 1.21855247e+00 4.97852772e-01
-1.24372804e+00 -8.60609114e-01 2.01926619e-01 4.05861706e-01
-1.82257628e+00 -8.20306778e-01 6.59419239e-01 -3.92025292e-01
4.58634645e-01 2.03697100e-01 1.03825343e+00 1.63424528e+00
-3.59071702e-01 3.79797608e-01 1.14648962e+00 -9.64156315e-02
6.72972649e-02 3.75791818e-01 -1.54239595e-01 9.70657945e-01
1.31073326e-01 -2.73242474e-01 -7.01045871e-01 -2.44859725e-01
4.64196682e-01 1.41197070e-01 -2.92255312e-01 -4.77282196e-01
-1.19760418e+00 6.01518035e-01 5.08679986e-01 4.29212525e-02
-4.14225221e-01 1.20716743e-01 3.98137182e-01 2.50508249e-01
5.13844788e-01 -1.62701122e-02 -3.82622957e-01 -1.08702593e-02
-7.73703516e-01 2.03222319e-01 7.77779281e-01 9.49599504e-01
1.15183735e+00 -1.56850308e-01 -5.41820526e-02 7.57112026e-01
6.10490851e-02 9.15509641e-01 2.90711045e-01 -8.92380476e-01
3.34300280e-01 6.67352259e-01 1.70854613e-01 -1.30237126e+00
-5.47448337e-01 -1.73479944e-01 -8.29530597e-01 5.86297736e-02
7.15292692e-01 -3.37538034e-01 -6.13919377e-01 2.02871299e+00
6.33378506e-01 4.18029010e-01 -3.82575691e-01 9.59372938e-01
6.11330688e-01 4.86677587e-02 7.46497065e-02 2.25088954e-01
1.79385948e+00 -6.17907524e-01 -4.81761068e-01 -1.98973015e-01
4.46479499e-01 -5.67206800e-01 1.24536264e+00 1.00165360e-01
-8.43366563e-01 -5.57302356e-01 -8.11352968e-01 8.86136368e-02
-7.25664675e-01 3.68699133e-01 5.16192198e-01 1.02020597e+00
-1.07689047e+00 5.84427714e-01 -5.21517873e-01 -9.38485742e-01
7.01525092e-01 4.67146546e-01 -9.14361358e-01 -1.55247822e-01
-1.12248123e+00 6.68059826e-01 -2.38555297e-01 3.65735069e-02
-6.41842782e-01 -6.96680546e-01 -1.04708099e+00 -2.16806188e-01
2.26114556e-01 -8.99972320e-01 7.19027817e-01 -1.36684656e+00
-1.41282129e+00 1.23334193e+00 -1.98376030e-01 -2.77314454e-01
6.15163326e-01 -2.81093538e-01 -6.49681091e-01 1.70008212e-01
5.43996170e-02 5.30530155e-01 9.39882517e-01 -9.86589134e-01
-5.97995520e-01 -8.88855696e-01 2.10684523e-01 1.44570798e-01
-8.05436373e-01 1.42855808e-01 -4.63037401e-01 -5.38021922e-01
-3.52954715e-01 -1.19197404e+00 2.03338265e-01 4.44147140e-01
-2.88263947e-01 -7.88197815e-02 6.14129961e-01 -6.99028313e-01
8.28358829e-01 -2.06023312e+00 4.43683416e-02 3.79173726e-01
3.88688222e-02 2.64099687e-02 -2.11946368e-01 1.27707347e-01
9.84346401e-03 -7.19340816e-02 -3.14076096e-02 -7.15539873e-01
2.13436633e-01 1.55047640e-01 2.40341946e-01 9.60713506e-01
9.06717777e-03 7.47259498e-01 -8.47606480e-01 -3.00070494e-01
9.20123309e-02 8.02226186e-01 -3.37811977e-01 6.69554388e-03
1.76482633e-01 5.10351479e-01 -2.11510792e-01 9.02400732e-01
5.90253413e-01 2.51538865e-02 1.42406136e-01 -3.79552066e-01
1.85555339e-01 -7.27412999e-02 -1.35391903e+00 1.81932652e+00
-7.09096909e-01 6.70763850e-01 1.72217309e-01 -7.56075859e-01
8.63307059e-01 3.43538165e-01 5.78837574e-01 -6.21224344e-01
2.24091902e-01 -3.27360001e-03 -6.27680898e-01 -7.09219992e-01
1.82274491e-01 1.79622680e-01 -1.40004545e-01 5.74861228e-01
4.24447171e-02 6.13402605e-01 -6.62148520e-02 -2.31510118e-01
7.71967292e-01 1.98939428e-01 1.20321717e-02 -2.16775030e-01
6.70913577e-01 -3.98733258e-01 4.59669948e-01 4.17720526e-01
-2.36456424e-01 6.90750301e-01 3.44831385e-02 -9.71676767e-01
-1.22261775e+00 -9.79180276e-01 -6.40910566e-02 1.41099322e+00
7.65445158e-02 -3.15434277e-01 -8.83172631e-01 -5.72409689e-01
1.32664919e-01 1.97560355e-01 -1.00597858e+00 -1.66241452e-01
-4.57667589e-01 -7.43452132e-01 7.15907872e-01 5.18839538e-01
7.58624494e-01 -7.57503271e-01 -2.77072996e-01 -1.83561072e-01
-3.07609469e-01 -1.38601422e+00 -5.73755980e-01 -5.56240141e-01
-1.54022440e-01 -1.08597386e+00 -6.38817430e-01 -4.46243852e-01
6.41705334e-01 1.39136270e-01 1.29999936e+00 1.04086213e-01
-4.36543584e-01 6.34196699e-01 -2.15960994e-01 -1.52484506e-01
3.90356600e-01 1.57279536e-01 4.84301955e-01 1.08637440e+00
6.69397593e-01 -9.28560197e-01 -8.27673435e-01 4.05820072e-01
-3.22239071e-01 -3.76423657e-01 2.64893770e-01 4.98882502e-01
3.86037201e-01 -2.16767758e-01 2.19473034e-01 -7.04114258e-01
4.01268750e-02 -6.51858211e-01 -3.69824350e-01 2.51882643e-01
-1.12313285e-01 -3.53360295e-01 5.49549878e-01 -5.35410941e-01
-8.92729104e-01 5.13993621e-01 2.80696098e-02 -3.08788449e-01
-5.53318560e-01 -1.00225642e-01 -3.92130375e-01 -2.85662383e-01
6.98466599e-01 -1.24080339e-03 -7.00107440e-02 -4.52677816e-01
3.80343974e-01 6.64632678e-01 6.30799890e-01 -4.93632108e-01
1.00135398e+00 1.01126611e+00 3.72267030e-02 -7.37131357e-01
-9.79335606e-01 -4.17468101e-01 -1.07287276e+00 -3.27859104e-01
1.04385948e+00 -1.18867183e+00 -9.74098384e-01 5.93079209e-01
-7.45427728e-01 -2.88611948e-01 1.12591358e-02 5.12633741e-01
-4.87233937e-01 1.49141669e-01 -3.44692677e-01 -8.19712520e-01
-1.47186905e-01 -6.64186239e-01 1.12439430e+00 2.73294121e-01
-3.90146345e-01 -9.40161347e-01 -6.96891248e-02 5.29442251e-01
2.89174855e-01 7.06396043e-01 3.37832510e-01 -4.67224568e-01
-3.94899040e-01 -3.83062541e-01 -2.48951256e-01 -3.79925929e-02
3.88005853e-01 -1.14186399e-01 -1.27688062e+00 -2.36808628e-01
-3.78336906e-01 -4.19562340e-01 6.17703855e-01 8.69560465e-02
1.19759476e+00 -6.11717224e-01 -2.07408771e-01 1.05052948e+00
1.41460145e+00 -5.47220349e-01 4.57989097e-01 3.33577514e-01
1.01050782e+00 7.92415559e-01 2.62201995e-01 6.96154356e-01
8.15423310e-01 1.06447029e+00 5.78188300e-01 -6.76978603e-02
-6.35535046e-02 -9.25428495e-02 3.92237186e-01 -5.44749051e-02
-6.42739236e-01 9.31131542e-02 -6.88636899e-01 7.34030962e-01
-1.69213390e+00 -1.32498252e+00 -2.10805088e-02 2.47882819e+00
5.31652510e-01 -3.73008817e-01 6.40720725e-01 -1.46130145e-01
8.90159309e-01 1.04053289e-01 -6.16469324e-01 -4.70004827e-02
-2.73250580e-01 1.97427750e-01 5.81923425e-01 2.00576663e-01
-1.43655801e+00 6.74537480e-01 5.42842388e+00 4.85011697e-01
-1.06640589e+00 9.49294940e-02 6.76317453e-01 -5.53980708e-01
-3.06327175e-02 -4.75704134e-01 -7.42419958e-01 5.10810554e-01
8.29208076e-01 3.34422290e-01 9.44209695e-01 7.16196716e-01
4.19939756e-02 2.60860533e-01 -1.16991591e+00 1.18117285e+00
3.97835433e-01 -1.06943691e+00 -2.68893540e-01 1.03831679e-01
7.76770771e-01 3.98787744e-02 2.68548280e-01 -2.07997281e-02
1.53947905e-01 -1.15461385e+00 7.90052712e-01 7.31078267e-01
7.97968626e-01 -8.20972741e-01 6.69704080e-01 -1.36599839e-01
-1.44191039e+00 -3.22730005e-01 -3.43560427e-01 -1.06586307e-01
2.80332472e-02 3.66156220e-01 -3.07787180e-01 4.59832400e-01
1.20171881e+00 8.39947939e-01 -8.69503975e-01 7.18794107e-01
-1.15478463e-01 2.91167587e-01 -5.17746568e-01 1.30680084e-01
9.51872095e-02 -2.55375952e-01 2.71597624e-01 1.11094654e+00
5.25298536e-01 -2.52068751e-02 1.09695144e-01 5.75635970e-01
-3.56505930e-01 9.79685783e-02 -7.86594093e-01 4.48649287e-01
5.41422844e-01 1.67318380e+00 -3.55042905e-01 -2.54802018e-01
-6.54702425e-01 1.15698361e+00 3.48361105e-01 3.83852035e-01
-9.77587104e-01 -5.74596412e-02 1.27905583e+00 5.02119780e-01
3.12955678e-01 -6.03070334e-02 8.96395221e-02 -1.14822412e+00
3.50771725e-01 -7.67892957e-01 4.56249684e-01 -7.43199468e-01
-1.66551185e+00 6.77716851e-01 -3.65669012e-01 -1.14214253e+00
-2.09850460e-01 -4.38093960e-01 -5.37337422e-01 8.08711231e-01
-1.42481446e+00 -1.69174433e+00 -7.98871279e-01 1.09137917e+00
6.85992986e-02 -3.13325286e-01 9.10317183e-01 7.60741472e-01
-6.60340548e-01 9.26304638e-01 9.93753597e-02 5.35939872e-01
9.93846714e-01 -1.09032321e+00 5.09095848e-01 3.88562202e-01
2.65390992e-01 4.56159353e-01 4.14701134e-01 -2.73641109e-01
-1.37125003e+00 -1.31638777e+00 9.25142407e-01 -8.02078605e-01
6.97262704e-01 -6.90088093e-01 -6.09248459e-01 7.56658554e-01
5.14702089e-02 4.63705897e-01 9.51195478e-01 3.55274856e-01
-6.75311983e-01 -4.60249424e-01 -1.30759120e+00 5.60085237e-01
1.56389081e+00 -7.10749865e-01 -1.78191476e-02 5.68435848e-01
1.16227739e-01 -1.18373185e-01 -1.02820730e+00 5.99380508e-02
9.90280330e-01 -1.19339502e+00 1.37278640e+00 -6.13878608e-01
1.54587507e-01 -2.30837330e-01 -2.66889036e-01 -1.17660725e+00
-3.00572157e-01 -4.90946949e-01 2.44878769e-01 1.60393763e+00
1.58063889e-01 -5.98971784e-01 8.21752787e-01 7.77431250e-01
2.82596558e-01 -6.38743162e-01 -8.23644280e-01 -6.47254765e-01
-4.17121708e-01 -1.18124463e-01 1.14698410e+00 1.18078578e+00
-5.22252679e-01 2.10941017e-01 -6.51320577e-01 2.39445448e-01
7.91524768e-01 6.58885464e-02 1.09883606e+00 -1.54388452e+00
9.21637192e-02 -3.91085267e-01 -7.27528870e-01 -3.27835739e-01
4.88096327e-01 -7.30245948e-01 -4.29325700e-01 -1.08572769e+00
1.58255458e-01 -4.53282565e-01 -9.40290093e-02 7.13081717e-01
-9.74922851e-02 1.18703353e+00 1.00765340e-01 -8.25525448e-02
-5.21567523e-01 5.42234004e-01 7.85352886e-01 -2.38145724e-01
1.04686111e-01 1.02453455e-01 -5.84543049e-01 1.05526054e+00
8.05116653e-01 -3.15894067e-01 -1.35714650e-01 -5.13626397e-01
4.12421763e-01 -4.97499347e-01 8.90315533e-01 -1.17152727e+00
7.33074918e-02 -1.72704663e-02 8.32637310e-01 -8.97870734e-02
7.35215962e-01 -1.16579127e+00 2.65292466e-01 6.54426068e-02
-1.45206124e-01 3.30494106e-01 -8.00938159e-02 6.09861553e-01
1.30533367e-01 2.98777577e-02 6.21745110e-01 -2.07831308e-01
-7.18535960e-01 7.21803784e-01 -1.04054600e-01 4.44858614e-03
9.64836538e-01 -4.46372271e-01 -1.25732139e-01 -6.11543655e-01
-8.69400442e-01 2.27126367e-02 9.45475578e-01 4.22880352e-01
1.46328583e-01 -1.91310358e+00 -8.50323617e-01 2.65714556e-01
3.91166270e-01 -4.92500305e-01 2.81238914e-01 8.45470905e-01
-2.42534935e-01 -2.18021244e-01 -5.64260542e-01 -5.67775607e-01
-1.38825655e+00 4.91665393e-01 3.94142926e-01 1.38976529e-01
-4.99893963e-01 5.87694407e-01 -3.59056853e-02 -6.56819284e-01
5.21297641e-02 2.68717855e-01 -2.94872552e-01 4.06567186e-01
6.68672800e-01 3.79514098e-01 4.11245711e-02 -1.26468599e+00
-4.74341065e-01 1.09489954e+00 4.34414685e-01 1.17122956e-01
1.40057850e+00 -4.50497836e-01 -4.75672707e-02 1.86650604e-01
1.32876921e+00 1.25959739e-01 -1.41534424e+00 -4.06272322e-01
-3.07392538e-01 -6.71424270e-01 -4.11954373e-01 -5.34951150e-01
-1.47821248e+00 7.57709682e-01 8.13360572e-01 6.32552877e-02
1.12013972e+00 -1.09394021e-01 8.61790597e-01 3.91709805e-01
3.78661513e-01 -1.24848449e+00 6.85141608e-02 2.45005444e-01
5.20458162e-01 -1.33250117e+00 3.67397186e-03 -1.60836339e-01
-5.36883235e-01 8.90677512e-01 5.03959656e-01 -4.34629768e-02
5.80962420e-01 -1.58112571e-01 1.13635406e-01 -1.21514790e-01
-2.01218411e-01 -3.75057667e-01 3.48702103e-01 8.98554265e-01
1.73383057e-01 7.73284212e-02 3.05038035e-01 4.24688786e-01
-4.99316752e-01 -2.33576044e-01 1.30464301e-01 4.18186992e-01
-9.76429135e-02 -8.49513233e-01 -5.53345740e-01 3.16140145e-01
-3.95778000e-01 1.43450171e-01 -5.71922362e-01 7.68263221e-01
5.61833382e-01 8.07241082e-01 4.17179763e-01 -4.03649896e-01
3.92626762e-01 1.25064030e-01 5.80843627e-01 -2.44184017e-01
-7.96420515e-01 -6.25794768e-01 3.16186249e-01 -7.60176420e-01
-7.24506736e-01 -9.23233867e-01 -6.12273276e-01 -7.23310888e-01
1.97798535e-01 -4.46689546e-01 6.24647737e-01 9.44690883e-01
4.03663069e-01 -7.28290826e-02 7.78761387e-01 -1.19382489e+00
-1.71107426e-01 -7.55977452e-01 -4.88782257e-01 1.02508855e+00
2.68550098e-01 -8.89961541e-01 -1.97797343e-01 3.15204382e-01]
|
[14.448375701904297, 0.9567000865936279]
|
e34f8e34-f386-4a27-8310-b5a7b2da6319
|
efficient-lightweight-3d-cnn-using-frame
|
2105.06340
| null |
https://arxiv.org/abs/2105.06340v4
|
https://arxiv.org/pdf/2105.06340v4.pdf
|
3D-CNN for Facial Micro- and Macro-expression Spotting on Long Video Sequences using Temporal Oriented Reference Frame
|
Facial expression spotting is the preliminary step for micro- and macro-expression analysis. The task of reliably spotting such expressions in video sequences is currently unsolved. The current best systems depend upon optical flow methods to extract regional motion features, before categorisation of that motion into a specific class of facial movement. Optical flow is susceptible to drift error, which introduces a serious problem for motions with long-term dependencies, such as high frame-rate macro-expression. We propose a purely deep learning solution which, rather than tracking frame differential motion, compares via a convolutional model, each frame with two temporally local reference frames. Reference frames are sampled according to calculated micro- and macro-expression duration. As baseline for MEGC2021 using leave-one-subject-out evaluation method, we show that our solution achieves F1-score of 0.105 in a high frame-rate (200 fps) SAMM long videos dataset (SAMM-LV) and is competitive in a low frame-rate (30 fps) (CAS(ME)2) dataset. On unseen MEGC2022 challenge dataset, the baseline results are 0.1176 on SAMM Challenge dataset, 0.1739 on CAS(ME)3 and overall performance of 0.1531 on both dataset.
|
['SuJing Wang', 'Jingting Li', 'Connah Kendrick', 'Ryan Cunningham', 'Adrian K. Davison', 'Moi Hoon Yap', 'Chuin Hong Yap']
|
2021-05-13
| null | null | null | null |
['micro-expression-spotting']
|
['computer-vision']
|
[ 1.04397915e-01 -3.39549631e-01 -1.05617523e-01 -7.05687702e-01
-8.96528244e-01 -3.13631058e-01 5.32814622e-01 -4.91416126e-01
-8.30764532e-01 6.51557148e-01 4.88800518e-02 3.03681821e-01
3.73537064e-01 -1.54720217e-01 -5.64414203e-01 -8.62392724e-01
-4.31277663e-01 -2.68333793e-01 1.79847747e-01 -2.80620098e-01
2.51029078e-02 5.79304814e-01 -1.51967657e+00 5.74827015e-01
-9.53586027e-02 1.28131342e+00 -1.47589475e-01 1.13512170e+00
2.36135229e-01 1.29528415e+00 -6.53187752e-01 -4.44693923e-01
-2.72200815e-02 -6.34260833e-01 -9.96549189e-01 1.76231340e-01
8.12876582e-01 -5.38356960e-01 -2.28344381e-01 9.07107174e-01
5.93537867e-01 1.39261156e-01 3.86140794e-01 -1.62098801e+00
4.13416289e-02 -2.76600659e-01 -7.90873587e-01 3.81974131e-01
5.61184108e-01 2.80002981e-01 5.42413354e-01 -9.37491059e-01
1.19950688e+00 1.39488673e+00 6.60897732e-01 9.59245741e-01
-1.02390003e+00 -6.50068581e-01 1.12913040e-04 2.84765810e-01
-1.31551540e+00 -9.29928362e-01 5.14551938e-01 -5.27933836e-01
1.00409698e+00 2.17982933e-01 5.10656774e-01 1.38771307e+00
3.48954111e-01 8.82108748e-01 8.30557048e-01 -9.66452360e-02
1.06343821e-01 -4.91715282e-01 -2.20924452e-01 7.39509404e-01
-5.73124647e-01 -4.49827425e-02 -8.42533588e-01 1.82302177e-01
6.58112168e-01 -5.62021494e-01 -3.46955806e-01 3.20593715e-01
-1.01408160e+00 6.55135036e-01 1.15015335e-01 3.60700458e-01
-3.22447747e-01 6.39632165e-01 8.15057755e-01 4.89481807e-01
5.88659525e-01 -1.63031399e-01 -5.42854905e-01 -7.92003512e-01
-1.32937801e+00 3.42159241e-01 3.06145310e-01 8.61560524e-01
6.30668223e-01 2.14965791e-01 -2.03470275e-01 6.16927803e-01
1.28338039e-01 4.77402955e-01 5.44473886e-01 -1.49453914e+00
1.49911687e-01 -9.31956097e-02 2.42204860e-01 -1.11581588e+00
-5.80967069e-01 3.10369700e-01 -7.86776721e-01 4.94148135e-01
6.80632174e-01 -4.50172484e-01 -6.58497214e-01 2.02863050e+00
3.47554743e-01 3.37198377e-01 -5.46409227e-02 1.22296774e+00
8.87991488e-01 8.20953846e-01 2.39729121e-01 -7.09138632e-01
1.26737428e+00 -9.06694055e-01 -9.23171878e-01 1.81838170e-01
1.00583434e+00 -7.32465386e-01 7.37558126e-01 5.50540268e-01
-1.20241499e+00 -6.47457182e-01 -7.81231880e-01 -1.80387378e-01
1.07558332e-01 7.04521835e-02 4.29628164e-01 5.29734612e-01
-1.46676290e+00 5.61795354e-01 -9.34908986e-01 -2.93274552e-01
5.79759777e-01 4.35689747e-01 -7.91495502e-01 2.98005909e-01
-1.16659915e+00 7.45112658e-01 -1.23588242e-01 2.68740028e-01
-9.16654348e-01 -6.30525053e-01 -9.93471920e-01 -4.71901566e-01
-2.13477239e-01 -2.41073564e-01 1.33585763e+00 -1.75425339e+00
-1.82163918e+00 1.24797606e+00 -6.39594555e-01 -4.71969366e-01
9.01404321e-01 -4.02008414e-01 -5.72174788e-01 6.11314416e-01
-1.07869901e-01 1.21824026e+00 1.00458372e+00 -6.92010164e-01
-5.95675409e-01 -1.76764458e-01 -4.44800742e-02 4.64382470e-02
5.48942350e-02 7.86487103e-01 -5.25481880e-01 -3.75189930e-01
-3.14740717e-01 -1.14599276e+00 6.95387870e-02 4.67231393e-01
2.43639186e-01 -1.91737786e-01 1.31203628e+00 -5.81754148e-01
1.12216628e+00 -2.17674017e+00 -1.39696717e-01 -4.14486587e-01
2.10834164e-02 4.90388364e-01 -3.62922519e-01 -2.10689425e-01
-3.73658866e-01 -1.67271703e-01 3.67310978e-02 -5.65245211e-01
-3.84731412e-01 7.39457384e-02 -6.99570179e-02 9.33929861e-01
4.67516541e-01 9.52169061e-01 -1.05101788e+00 -5.13137817e-01
2.07070246e-01 7.15929568e-01 -4.84025687e-01 1.88102081e-01
4.91653532e-02 6.79887116e-01 -7.57913888e-02 6.83324575e-01
7.66615629e-01 1.02406949e-01 -1.59050167e-01 -2.98805535e-01
-5.42490557e-02 -3.25001717e-01 -8.52608383e-01 1.91645706e+00
-4.27085131e-01 1.61138916e+00 2.07807124e-01 -8.18621576e-01
8.35285425e-01 5.47468603e-01 9.23192382e-01 -6.51279151e-01
1.57947019e-01 9.87702608e-02 -1.80784047e-01 -9.00437653e-01
4.25754339e-01 -1.28398195e-01 4.99388501e-02 -3.46535854e-02
3.50474417e-01 2.51213789e-01 2.95286447e-01 4.88040783e-02
1.21358085e+00 5.06111026e-01 1.67752966e-01 -2.32257083e-01
6.43272996e-01 -5.35618424e-01 5.62619865e-01 1.14351429e-01
-8.74319434e-01 8.91290665e-01 8.74519229e-01 -7.80449390e-01
-8.39186013e-01 -7.48659134e-01 -1.27867714e-01 1.01133370e+00
-3.82879041e-02 -5.74276984e-01 -8.78322780e-01 -5.82825661e-01
-4.23614323e-01 1.46652326e-01 -7.46868312e-01 1.11047234e-02
-8.17419171e-01 -7.36341476e-01 8.50485861e-01 4.18492496e-01
5.78557074e-01 -1.35701609e+00 -1.11427617e+00 2.95665413e-01
-5.09140491e-01 -1.56962574e+00 -4.15436655e-01 -2.73918688e-01
-5.33139050e-01 -9.93300557e-01 -9.32297707e-01 -5.25665522e-01
3.77408445e-01 -3.57268602e-02 1.19926023e+00 -2.96894112e-03
-2.67355829e-01 1.66928932e-01 -3.89963210e-01 1.72235314e-02
-2.26632997e-01 -5.76306403e-01 1.84905380e-02 4.44571704e-01
4.50598568e-01 -2.24940374e-01 -7.75006950e-01 5.11115730e-01
-6.75966144e-01 1.73162427e-02 1.20360935e-02 7.75906265e-01
3.27849120e-01 -5.90948343e-01 3.12212586e-01 -3.62739801e-01
1.01925381e-01 -2.31203496e-01 -5.79591095e-01 -1.26983374e-01
4.12151627e-02 -3.99681419e-01 2.46116757e-01 -5.14872730e-01
-8.75812709e-01 4.54720706e-01 -3.17535520e-01 -7.55324125e-01
-3.27306122e-01 1.15878908e-02 -6.92903772e-02 -4.99244221e-02
4.79270905e-01 -4.15283367e-02 2.14396641e-01 1.53063953e-01
1.71942204e-01 5.21172166e-01 8.30033422e-01 -2.60889709e-01
9.28060859e-02 7.38829732e-01 2.28595808e-01 -1.37792432e+00
-6.29254162e-01 -5.78138232e-01 -7.94191301e-01 -7.95330584e-01
1.28619683e+00 -1.16757226e+00 -1.09426033e+00 9.08600032e-01
-1.41283524e+00 -6.25026703e-01 2.01369330e-01 3.31078708e-01
-8.95708263e-01 3.23857427e-01 -6.53957784e-01 -9.29374754e-01
-2.73265421e-01 -1.23486292e+00 1.56396401e+00 9.22164023e-02
-8.40944946e-01 -9.05697167e-01 2.83821020e-02 1.82416916e-01
2.03035742e-01 8.67223859e-01 -8.15297216e-02 8.68865252e-02
-1.78890228e-01 -1.00444786e-01 -7.49733374e-02 3.49654108e-01
-3.25932242e-02 4.74486232e-01 -1.26972926e+00 -2.28785187e-01
-4.31569703e-02 -6.22798681e-01 7.89228380e-01 7.65255034e-01
1.14838696e+00 -1.82049006e-01 5.57063147e-02 8.61356437e-01
1.22991002e+00 2.73219287e-01 1.18569410e+00 2.07926512e-01
6.27891600e-01 6.85580313e-01 8.96954238e-01 5.12056351e-01
-4.92299907e-02 1.08037400e+00 2.73938835e-01 -1.58639863e-01
-1.33656502e-01 3.18355769e-01 9.17786360e-01 1.54386863e-01
-3.81758124e-01 -2.30470404e-01 -4.96018499e-01 5.04570961e-01
-1.89598000e+00 -1.49455416e+00 -4.16531861e-01 1.89041340e+00
7.15782285e-01 -1.73432902e-01 3.65288526e-01 6.18229918e-02
5.53949535e-01 4.30285931e-01 -1.82302058e-01 -5.39125502e-01
-4.83559757e-01 2.10683838e-01 1.32397339e-01 6.59591913e-01
-1.39332080e+00 1.10125315e+00 5.85489988e+00 7.70870626e-01
-1.59706414e+00 1.47521332e-01 1.08575654e+00 -4.53922480e-01
5.02093613e-01 -3.30597609e-01 -8.03949118e-01 8.04156542e-01
1.39408553e+00 2.48990729e-01 -1.76277339e-01 5.84612191e-01
7.51613498e-01 -3.94370168e-01 -1.02807510e+00 1.55686784e+00
1.36727288e-01 -1.31712592e+00 -3.49230766e-01 -8.31331611e-02
6.78019702e-01 -1.09048605e-01 -9.08220038e-02 7.44964704e-02
-3.27458382e-01 -1.17468965e+00 9.06809628e-01 4.62123781e-01
1.12126362e+00 -7.85006702e-01 7.81136334e-01 -1.73480898e-01
-1.14816999e+00 3.58206123e-01 -3.08533877e-01 -1.47726476e-01
4.67240363e-01 3.02205980e-01 -3.63371164e-01 1.08974665e-01
9.45168972e-01 1.15768564e+00 -2.91616142e-01 4.99857783e-01
-5.31303361e-02 5.27206421e-01 -2.61594713e-01 1.06357627e-01
3.87559235e-01 2.56186947e-02 3.00852180e-01 1.70447707e+00
3.51242602e-01 2.30623782e-01 -3.63909394e-01 2.97562003e-01
1.00896224e-01 -1.65505901e-01 -4.87708092e-01 1.88446388e-01
-1.61589146e-01 1.47644413e+00 -5.48102915e-01 -3.61034662e-01
-3.93227905e-01 1.36826742e+00 9.98174306e-03 4.53130066e-01
-1.17905450e+00 -3.92292924e-02 1.20597506e+00 3.29160020e-02
1.86986238e-01 -2.57783383e-01 3.61381173e-01 -1.08770561e+00
-1.74995456e-02 -6.41774654e-01 2.85541564e-01 -1.07291353e+00
-6.50829613e-01 7.20504999e-01 -1.39496669e-01 -1.36382151e+00
-7.00168192e-01 -7.51798928e-01 -4.81993020e-01 4.91465330e-01
-1.22824168e+00 -9.28559721e-01 -6.16805375e-01 8.26683700e-01
7.18161762e-01 5.90132922e-03 8.24823618e-01 4.55427110e-01
-3.40584606e-01 7.56574214e-01 -2.63572901e-01 4.60540086e-01
9.67613041e-01 -7.81955421e-01 1.83729187e-01 8.19169462e-01
2.20307276e-01 1.10821584e-02 7.77178526e-01 -8.34907815e-02
-1.05308163e+00 -1.21583831e+00 1.10981393e+00 -5.05023777e-01
5.73905766e-01 -3.55764985e-01 -7.37910748e-01 4.65730488e-01
1.90243125e-01 6.83661699e-01 4.57495451e-01 -4.34930086e-01
-5.67388497e-02 -8.68497863e-02 -1.04215181e+00 4.15832579e-01
9.24425066e-01 -5.04726350e-01 7.19363838e-02 2.11655125e-01
3.41956258e-01 -6.01048350e-01 -8.29892099e-01 3.61168474e-01
9.03920531e-01 -1.32560349e+00 7.22458541e-01 -7.30586529e-01
7.16962278e-01 -2.86103755e-01 -1.57132149e-01 -8.27901363e-01
1.71310246e-01 -1.24993849e+00 -8.02096725e-02 1.02531588e+00
-1.88560388e-03 7.45836645e-02 8.92470598e-01 4.78378028e-01
1.81716084e-01 -7.77610719e-01 -1.23467577e+00 -7.24523187e-01
-1.53977469e-01 -6.79485738e-01 6.65240176e-03 9.80526745e-01
-1.63281024e-01 1.57839686e-01 -7.48969376e-01 -1.84889913e-01
2.36024871e-01 -2.12297842e-01 1.04015648e+00 -6.56343639e-01
1.44450650e-01 -4.54412043e-01 -1.15839040e+00 -9.91782367e-01
5.41094661e-01 -3.65015805e-01 2.42283363e-02 -8.12246501e-01
-6.73385561e-02 3.05852294e-01 -1.28515556e-01 3.72116864e-01
4.70404103e-02 6.67732179e-01 2.13115454e-01 -6.96863532e-02
-8.46436918e-01 4.37748313e-01 1.06324899e+00 5.69393225e-02
5.49998917e-02 -2.56175280e-01 1.91921726e-01 8.35422218e-01
4.89432812e-01 -2.81624466e-01 -2.08637089e-01 -2.53163159e-01
-9.74297002e-02 3.91340941e-01 5.34388781e-01 -1.07882524e+00
-8.63592401e-02 -1.38799325e-01 5.76673090e-01 -5.21591663e-01
6.55328095e-01 -4.72580642e-01 1.30418792e-01 4.56393361e-01
-3.41692686e-01 3.24174732e-01 3.89827639e-01 2.23476112e-01
-4.38590854e-01 6.96811676e-02 1.22809303e+00 1.95765719e-01
-1.30553520e+00 4.56383735e-01 -5.11903226e-01 1.95743993e-01
1.09788132e+00 -4.35604602e-01 -3.53772603e-02 -7.86562264e-01
-8.52696478e-01 -9.87208933e-02 3.11439455e-01 5.94559431e-01
6.28680825e-01 -1.47416341e+00 -7.98038304e-01 7.16072544e-02
-3.85244824e-02 -2.46015355e-01 3.24856460e-01 1.14506245e+00
-8.83216500e-01 1.19737782e-01 -4.75056708e-01 -1.07180941e+00
-1.75239897e+00 2.41920780e-02 4.86299247e-01 2.21254334e-01
-4.16634053e-01 1.23334920e+00 2.59846389e-01 3.91699046e-01
-3.19546126e-02 -2.36850530e-01 -1.34226903e-01 2.08395898e-01
9.08003330e-01 2.37346947e-01 -2.62305457e-02 -1.43189633e+00
-6.27104342e-01 7.22975075e-01 2.22129226e-01 -2.91997373e-01
1.10953665e+00 -2.52386868e-01 3.12941559e-02 3.68463099e-01
2.06681204e+00 -3.71258914e-01 -1.73592377e+00 4.15089160e-01
4.99021374e-02 -5.74099243e-01 -9.26461667e-02 -2.87998885e-01
-1.30563235e+00 9.70851958e-01 9.48826671e-01 -5.01142919e-01
1.35283446e+00 -2.15179995e-01 7.02651978e-01 6.77896142e-02
1.33630097e-01 -1.02413809e+00 3.29523206e-01 6.10211611e-01
9.05820549e-01 -1.36022973e+00 -2.37871245e-01 -4.38263379e-02
-7.69734085e-01 1.41074610e+00 5.97658575e-01 -7.59792775e-02
5.40663958e-01 4.38299179e-01 5.26003599e-01 -7.12791607e-02
-9.80726182e-01 -2.98193675e-02 1.59995690e-01 5.28065145e-01
8.90197754e-01 -2.45300815e-01 -1.09979443e-01 -2.76882909e-02
-1.47961140e-01 2.89269716e-01 5.63672364e-01 7.30521977e-01
-1.30364478e-01 -6.04796112e-01 -8.45841840e-02 -2.50327468e-01
-8.31006706e-01 2.71968395e-01 -1.62954867e-01 9.90532577e-01
9.70577635e-03 8.29353392e-01 4.59391862e-01 -1.91704780e-01
5.47503680e-02 -3.09737325e-02 6.39938235e-01 1.42607421e-01
-3.03168774e-01 4.33269680e-01 2.63711512e-01 -1.23355472e+00
-1.09420168e+00 -7.91342139e-01 -1.37955749e+00 -3.44889253e-01
2.33203828e-01 -2.17706338e-01 4.02829975e-01 7.58534133e-01
2.53295362e-01 1.35096461e-01 5.38486600e-01 -1.09658694e+00
1.25087962e-01 -8.73435676e-01 -4.48746026e-01 9.72109258e-01
7.25612342e-01 -5.83089173e-01 -2.84261525e-01 7.04862833e-01]
|
[13.622672080993652, 1.758375644683838]
|
dff3060f-9e9b-41f0-8e6d-60e4b220381e
|
imbalance-agnostic-source-free-domain
|
2305.12649
| null |
https://arxiv.org/abs/2305.12649v1
|
https://arxiv.org/pdf/2305.12649v1.pdf
|
Imbalance-Agnostic Source-Free Domain Adaptation via Avatar Prototype Alignment
|
Source-free Unsupervised Domain Adaptation (SF-UDA) aims to adapt a well-trained source model to an unlabeled target domain without access to the source data. One key challenge is the lack of source data during domain adaptation. To handle this, we propose to mine the hidden knowledge of the source model and exploit it to generate source avatar prototypes. To this end, we propose a Contrastive Prototype Generation and Adaptation (CPGA) method. CPGA consists of two stages: Prototype generation and Prototype adaptation. Extensive experiments on three UDA benchmark datasets demonstrate the superiority of CPGA. However, existing SF.UDA studies implicitly assume balanced class distributions for both the source and target domains, which hinders their real applications. To address this issue, we study a more practical SF-UDA task, termed imbalance-agnostic SF-UDA, where the class distributions of both the unseen source domain and unlabeled target domain are unknown and could be arbitrarily skewed. This task is much more challenging than vanilla SF-UDA due to the co-occurrence of covariate shifts and unidentified class distribution shifts between the source and target domains. To address this task, we extend CPGA and propose a new Target-aware Contrastive Prototype Generation and Adaptation (T-CPGA) method. Specifically, for better prototype adaptation in the imbalance-agnostic scenario, T-CPGA applies a new pseudo label generation strategy to identify unknown target class distribution and generate accurate pseudo labels, by utilizing the collective intelligence of the source model and an additional contrastive language-image pre-trained model. Meanwhile, we further devise a target label-distribution-aware classifier to adapt the model to the unknown target class distribution. We empirically show that T-CPGA significantly outperforms CPGA and other SF-UDA methods in imbalance-agnostic SF-UDA.
|
['Yanxia Liu', 'Qing Du', 'Dong Liu', 'Shuaicheng Niu', 'Zhen Qiu', 'Yifan Zhang', 'Mingkui Tan', 'Hongbin Lin']
|
2023-05-22
| null | null | null | null |
['source-free-domain-adaptation', 'unsupervised-domain-adaptation', 'pseudo-label']
|
['computer-vision', 'methodology', 'miscellaneous']
|
[ 3.80374104e-01 -1.83960512e-01 -3.98488730e-01 -4.86955732e-01
-8.46079767e-01 -6.58289433e-01 5.66067934e-01 6.09910078e-02
-7.70306289e-02 9.25620258e-01 -5.71546815e-02 -1.98200226e-01
7.63485059e-02 -7.66710758e-01 -7.58466780e-01 -7.32236385e-01
4.95681345e-01 9.26661909e-01 1.12173051e-01 -2.20007256e-01
-7.55578801e-02 2.53394604e-01 -1.30266190e+00 2.39550054e-01
1.21953809e+00 7.63316512e-01 9.57055092e-02 2.37459436e-01
-3.76895607e-01 2.16678768e-01 -8.72286379e-01 -2.91690141e-01
3.91359121e-01 -8.33165646e-01 -6.35261178e-01 3.32922250e-01
2.10704699e-01 -1.35676309e-01 2.31453642e-01 9.84907031e-01
6.45498693e-01 3.85719202e-02 1.16260886e+00 -1.53888571e+00
-7.88638055e-01 3.84947598e-01 -8.04831266e-01 1.47215694e-01
3.55014987e-02 2.65993118e-01 5.85493743e-01 -9.30756092e-01
6.17228627e-01 1.15945184e+00 6.28983319e-01 9.11979377e-01
-1.36484063e+00 -1.03732336e+00 5.24173677e-01 1.15785383e-01
-1.52618766e+00 -3.35973084e-01 1.06503868e+00 -5.49533606e-01
2.68490255e-01 3.67286876e-02 2.94163376e-01 1.47973859e+00
-2.71576345e-01 8.44772935e-01 1.24899828e+00 -6.13841355e-01
6.77633584e-01 5.60085416e-01 3.19177192e-03 8.78129527e-02
2.33278364e-01 6.46064356e-02 -3.39841336e-01 -3.11740845e-01
4.93960053e-01 -1.64694086e-01 -1.90323025e-01 -8.29694450e-01
-1.17214358e+00 7.27199197e-01 2.30337903e-01 1.01909243e-01
-2.68775284e-01 -8.03882480e-01 4.50970650e-01 3.64301264e-01
6.29215717e-01 3.56328070e-01 -7.10095465e-01 3.38852376e-01
-6.80112898e-01 3.49146485e-01 6.35912657e-01 1.10992825e+00
8.35009396e-01 1.01414546e-01 -2.93482810e-01 1.15232420e+00
2.58924693e-01 7.06230462e-01 1.03373086e+00 -3.84042740e-01
4.48061615e-01 8.55283201e-01 1.58513188e-01 -5.83571792e-01
-2.93892324e-01 -7.11423814e-01 -9.52889442e-01 1.55104861e-01
6.26724541e-01 -1.66318819e-01 -1.16342998e+00 2.03000998e+00
1.03790593e+00 1.31964520e-01 4.38790500e-01 8.20149004e-01
5.34723580e-01 4.17990059e-01 1.52778640e-01 -3.10095280e-01
1.05322456e+00 -8.80183935e-01 -4.20663357e-01 -4.34356719e-01
6.30002916e-01 -6.44034743e-01 1.38738787e+00 7.49074817e-02
-4.36645597e-01 -6.84769750e-01 -1.18916416e+00 3.18577528e-01
-3.69108796e-01 2.20293939e-01 2.20536545e-01 7.39496350e-01
-4.58820045e-01 -3.89068760e-02 -3.66987348e-01 -4.97879475e-01
5.76433957e-01 2.03554288e-01 -2.96006501e-01 -3.46923351e-01
-1.26011419e+00 5.78028738e-01 6.77988887e-01 -2.39655867e-01
-8.46162200e-01 -8.27504218e-01 -8.47293973e-01 -3.38603318e-01
4.68249589e-01 -8.79224420e-01 1.32408917e+00 -1.60843825e+00
-1.55086672e+00 9.06822264e-01 -1.42970696e-01 -1.87618524e-01
6.09933197e-01 3.64567310e-01 -6.28020287e-01 -3.62788141e-01
3.95597368e-01 4.39644784e-01 1.08222222e+00 -1.49548328e+00
-6.26966536e-01 -4.55637246e-01 -3.74438763e-01 2.41431132e-01
-4.32716727e-01 -5.51619828e-01 -9.36831012e-02 -8.20478559e-01
6.35213926e-02 -8.02854002e-01 -1.71518512e-03 -1.01281345e-01
-3.43865067e-01 -1.41507253e-01 8.81763995e-01 -2.83252090e-01
1.13357735e+00 -2.16287160e+00 -5.92230223e-02 2.76045859e-01
4.74547483e-02 5.58695138e-01 -4.38299745e-01 2.89699398e-02
-2.67479599e-01 -1.53630480e-01 -5.91719151e-01 -9.95123982e-02
-4.36684396e-03 3.66186291e-01 -4.06859100e-01 1.57600105e-01
2.81912178e-01 6.39429986e-01 -1.03037572e+00 -4.33165401e-01
-2.44500309e-01 1.24986723e-01 -5.34793615e-01 3.85985196e-01
-4.16538775e-01 6.80386305e-01 -5.70636511e-01 8.28456283e-01
9.36003983e-01 -2.05578834e-01 3.31594288e-01 -9.19032171e-02
3.44485372e-01 -9.58346203e-02 -1.31373656e+00 1.43850005e+00
-3.88248712e-01 1.34414628e-01 -3.20089847e-01 -1.24603581e+00
1.34097958e+00 1.31019920e-01 2.07257628e-01 -8.59637439e-01
5.47093758e-03 4.63847607e-01 1.01109050e-01 -3.51324797e-01
2.11826146e-01 -6.09179258e-01 -2.12822348e-01 4.77681130e-01
1.78052813e-01 4.40854728e-02 -1.06019685e-02 -1.68964826e-03
8.23287070e-01 2.96002567e-01 5.45042515e-01 -1.42897502e-01
6.29352868e-01 1.87183663e-01 1.08847845e+00 7.10861206e-01
-5.30683815e-01 8.35060954e-01 2.95376778e-01 -3.10285866e-01
-1.07721603e+00 -1.26329112e+00 -1.30618155e-01 1.15068662e+00
1.91231728e-01 1.40209511e-01 -6.28868282e-01 -1.21075964e+00
2.44825426e-02 7.57205725e-01 -8.29986095e-01 -5.00051916e-01
-2.98887938e-01 -1.03580475e+00 3.81102175e-01 4.41568226e-01
4.74995524e-01 -8.57450545e-01 -6.68713078e-02 3.17080706e-01
-4.25877124e-01 -8.22910190e-01 -6.21613920e-01 1.59968942e-01
-7.87848830e-01 -1.03058171e+00 -9.48771775e-01 -7.92142510e-01
8.96882594e-01 1.14545360e-01 1.08607626e+00 -5.25042176e-01
4.80417013e-02 1.36517063e-01 -4.55031216e-01 -5.98883331e-01
-7.81039655e-01 2.66157597e-01 2.08660439e-01 3.92406881e-01
7.23621070e-01 -6.59003496e-01 -4.35532421e-01 6.10678554e-01
-1.03784132e+00 2.64602732e-02 6.55781865e-01 9.80082810e-01
8.12762260e-01 -1.68201830e-02 1.28429699e+00 -1.28486598e+00
6.19768500e-01 -9.58863616e-01 -4.27230418e-01 3.73156786e-01
-7.71399498e-01 -3.00623639e-03 7.82491088e-01 -1.28310215e+00
-1.20882964e+00 1.25847816e-01 1.67156830e-02 -6.04951203e-01
-4.02417421e-01 3.85232359e-01 -8.30741107e-01 4.17313308e-01
1.14755595e+00 3.42349648e-01 -1.44694880e-01 -5.01699984e-01
2.26483092e-01 9.27487135e-01 6.96999788e-01 -7.55689740e-01
1.06232727e+00 3.03124249e-01 -3.95104229e-01 -3.87801498e-01
-8.32130492e-01 -5.27337849e-01 -7.41007209e-01 1.01725757e-01
2.40510359e-01 -1.03774703e+00 9.81442258e-02 7.31720030e-01
-9.03032124e-01 -5.31010926e-01 -6.60478711e-01 3.26962799e-01
-4.70661074e-01 1.62991226e-01 6.59729019e-02 -5.90863705e-01
-3.14291388e-01 -8.99034142e-01 8.47820282e-01 3.39035809e-01
-2.09501237e-01 -1.00706995e+00 3.32404673e-01 3.21675152e-01
4.04802263e-01 2.41106674e-01 1.06202006e+00 -1.27117550e+00
-7.85466656e-03 -1.32887527e-01 -2.31900051e-01 4.91358310e-01
5.80232859e-01 -3.80037934e-01 -9.68003571e-01 -3.42088640e-01
6.45593088e-03 -3.38883221e-01 4.54091191e-01 9.43933278e-02
8.61501098e-01 -3.43827516e-01 -2.31618956e-01 4.72664893e-01
1.23166704e+00 2.82590985e-01 2.26446792e-01 4.96954769e-01
6.09185040e-01 5.33169448e-01 9.34193194e-01 5.20563960e-01
5.85019469e-01 7.40410686e-01 6.93456456e-02 -9.28354710e-02
-3.77088010e-01 -5.50129831e-01 3.88311952e-01 8.10342312e-01
5.31593263e-01 -3.50200176e-01 -1.02613270e+00 8.80724430e-01
-1.71902323e+00 -4.72811848e-01 1.11205399e-01 2.29296947e+00
1.25294709e+00 1.24090314e-01 3.61877739e-01 1.11828879e-01
9.72657204e-01 -3.02998275e-01 -1.23079467e+00 -2.26149738e-01
-3.06760252e-01 4.79659922e-02 9.90005732e-02 1.67165875e-01
-1.08100283e+00 7.25519180e-01 5.13860989e+00 1.10910738e+00
-1.27065051e+00 3.20186585e-01 6.75406516e-01 1.16929285e-01
-4.60855752e-01 -1.66666925e-01 -7.75391579e-01 6.06738091e-01
7.95849144e-01 -4.10633326e-01 8.07034224e-02 9.67267513e-01
-6.08122684e-02 7.81236663e-02 -1.09352016e+00 8.32779646e-01
2.00780526e-01 -7.70119250e-01 2.80542046e-01 -7.32695460e-02
1.01609504e+00 -2.22162381e-01 4.73240428e-02 7.47943461e-01
3.30059171e-01 -4.94344741e-01 5.98861694e-01 1.90293461e-01
1.02249086e+00 -6.56172574e-01 7.31683254e-01 5.00543296e-01
-8.97609770e-01 -1.66038275e-01 -4.33621168e-01 1.23106673e-01
-1.95790738e-01 7.47478783e-01 -9.58806098e-01 7.00748444e-01
4.74517137e-01 4.25067216e-01 -6.53449833e-01 9.95761871e-01
-1.94722384e-01 7.76826024e-01 -2.71842360e-01 3.97738129e-01
-1.12539791e-01 5.38812429e-02 5.76922834e-01 9.48429525e-01
3.27619255e-01 -6.43801987e-02 2.58270055e-01 8.82302046e-01
-1.99446753e-01 3.55197191e-01 -4.93018806e-01 2.07926616e-01
7.36665130e-01 9.50403452e-01 -5.19779861e-01 -4.95047331e-01
-1.49277821e-01 9.65160787e-01 2.43112221e-01 5.72696269e-01
-7.32480228e-01 -2.77433604e-01 5.96585572e-01 1.98346823e-01
1.73755467e-01 3.61990124e-01 -2.93602735e-01 -1.39741576e+00
3.88522819e-02 -1.21712160e+00 7.02244997e-01 -5.58519065e-01
-1.83475232e+00 5.52047133e-01 7.25701749e-02 -1.66627395e+00
-3.82186741e-01 -1.67781562e-01 -4.75880712e-01 9.37991083e-01
-1.75255561e+00 -1.42692757e+00 -3.65019947e-01 8.44905615e-01
6.94082558e-01 -3.34448248e-01 7.64118552e-01 3.53629619e-01
-6.69623852e-01 1.03724599e+00 3.90713125e-01 4.35114019e-02
1.25561559e+00 -1.15153491e+00 2.77843386e-01 8.01055491e-01
-1.36493668e-01 3.38301212e-01 5.39574265e-01 -6.71420991e-01
-8.19279313e-01 -1.64297831e+00 7.29890227e-01 -3.35010588e-01
4.02700573e-01 -4.54706192e-01 -1.38375187e+00 6.46231651e-01
-3.61722767e-01 2.64730245e-01 7.75686979e-01 -2.61553656e-02
-6.31875575e-01 -4.04456884e-01 -1.38686502e+00 4.45701152e-01
9.04097140e-01 -3.02804977e-01 -7.30749130e-01 2.92517811e-01
5.99709213e-01 -3.61108124e-01 -5.94662249e-01 5.93507230e-01
3.19069773e-01 -7.05795884e-01 9.16876078e-01 -4.90319550e-01
1.93857104e-01 -5.29517829e-01 -1.27434567e-01 -1.56049347e+00
-1.77943572e-01 -1.45552024e-01 -2.28022989e-02 1.71525109e+00
3.13618183e-01 -8.73885870e-01 6.62856638e-01 3.26735020e-01
5.30368276e-02 -4.56857890e-01 -9.97091532e-01 -1.09093499e+00
4.96650726e-01 -2.21058398e-01 1.07931995e+00 1.35232973e+00
-2.82037258e-01 3.71508956e-01 -2.45597020e-01 2.16923103e-01
6.32497191e-01 3.66598159e-01 1.14303553e+00 -1.18794990e+00
-2.48605609e-01 -2.44370356e-01 -8.55268314e-02 -8.01460326e-01
2.18486547e-01 -9.64314759e-01 2.22026870e-01 -1.04675615e+00
2.43150890e-01 -9.21427608e-01 -4.61080372e-01 6.08848214e-01
-4.73831326e-01 3.55588138e-01 -4.42620814e-02 3.98600370e-01
-4.42928553e-01 7.54136384e-01 1.13854373e+00 -3.18399370e-01
-5.06137431e-01 2.18167633e-01 -9.49809194e-01 4.65173334e-01
9.06520605e-01 -6.69011176e-01 -7.32428670e-01 -1.44195259e-01
-1.88555956e-01 -3.00180018e-01 2.14342907e-01 -8.73122752e-01
-4.17470932e-02 -4.19435948e-01 5.07016063e-01 -3.38074654e-01
-1.69499904e-01 -7.29278386e-01 7.30768219e-02 2.09577218e-01
-1.03489943e-01 -3.43714625e-01 3.05422097e-01 6.09804869e-01
-2.86052614e-01 -1.23714551e-01 1.00883055e+00 1.00236945e-01
-8.16510022e-01 1.72548920e-01 -1.10998675e-01 4.25697207e-01
1.08385432e+00 -3.57033849e-01 -3.72807711e-01 -1.04885571e-01
-5.47001481e-01 1.89630687e-01 5.85835516e-01 5.28582573e-01
4.03236240e-01 -1.46536112e+00 -8.07025254e-01 6.57888412e-01
6.65057003e-01 3.49271685e-01 4.79397178e-01 6.08085752e-01
1.81187466e-01 -7.15724379e-02 -2.71290511e-01 -7.45418787e-01
-8.53645027e-01 9.07268167e-01 3.72666717e-01 -3.17925304e-01
-2.75756806e-01 6.54820085e-01 6.89513922e-01 -1.09842491e+00
-4.17538844e-02 2.93235071e-02 -1.45541370e-01 1.20653369e-01
4.82162148e-01 5.63114099e-02 1.64450973e-01 -6.80441201e-01
-2.56074667e-01 4.64448333e-01 -1.64026916e-01 2.43192255e-01
1.03061616e+00 -4.55812335e-01 2.72432864e-01 5.44097364e-01
9.93107498e-01 -9.16165635e-02 -1.33640409e+00 -6.51934266e-01
4.33296971e-02 -3.21837068e-01 -4.50134367e-01 -1.23708344e+00
-9.77170229e-01 7.58048952e-01 7.77620673e-01 -3.15576166e-01
1.48369622e+00 -4.80974689e-02 6.31785989e-01 -8.58032703e-02
1.13322884e-01 -1.01535881e+00 1.97989851e-01 3.25326800e-01
7.66687095e-01 -1.46668482e+00 -1.66387141e-01 -1.40473917e-01
-7.39242971e-01 7.72639811e-01 1.08928156e+00 3.88065815e-01
4.12013084e-01 -9.95762274e-02 5.69832385e-01 3.06179553e-01
-5.54621458e-01 -1.12844966e-01 3.13612521e-01 1.09007823e+00
-1.63729638e-01 7.26942867e-02 -1.49387777e-01 1.06242657e+00
-9.19903256e-03 1.19182564e-01 3.59658927e-01 8.56380343e-01
-7.00326934e-02 -1.51423895e+00 -6.90164745e-01 3.14114064e-01
2.97881500e-03 1.88414305e-01 -2.92861074e-01 7.52924323e-01
4.12443995e-01 7.65731335e-01 -8.81738290e-02 -3.88793319e-01
5.96332312e-01 3.19947302e-01 1.94340959e-01 -7.22896457e-01
-3.51809859e-01 2.79808659e-02 -2.97077000e-01 -2.42413906e-03
-4.19077992e-01 -5.32782078e-01 -9.90605593e-01 3.99901420e-02
-3.99382919e-01 1.91200331e-01 5.33743858e-01 9.60418403e-01
4.86047417e-01 4.30005401e-01 7.97018945e-01 -5.86739063e-01
-7.24320173e-01 -9.69410300e-01 -5.99083483e-01 6.96424007e-01
3.71804208e-01 -8.91460240e-01 -3.63594919e-01 1.70018017e-01]
|
[10.338887214660645, 3.111502170562744]
|
05dff2c1-0836-41eb-a304-6ca7a81e5a1c
|
machine-learning-cryptanalysis-of-a-quantum
|
1905.02342
| null |
https://arxiv.org/abs/1905.02342v2
|
https://arxiv.org/pdf/1905.02342v2.pdf
|
Machine Learning Cryptanalysis of a Quantum Random Number Generator
|
Random number generators (RNGs) that are crucial for cryptographic applications have been the subject of adversarial attacks. These attacks exploit environmental information to predict generated random numbers that are supposed to be truly random and unpredictable. Though quantum random number generators (QRNGs) are based on the intrinsic indeterministic nature of quantum properties, the presence of classical noise in the measurement process compromises the integrity of a QRNG. In this paper, we develop a predictive machine learning (ML) analysis to investigate the impact of deterministic classical noise in different stages of an optical continuous variable QRNG. Our ML model successfully detects inherent correlations when the deterministic noise sources are prominent. After appropriate filtering and randomness extraction processes are introduced, our QRNG system, in turn, demonstrates its robustness against ML. We further demonstrate the robustness of our ML approach by applying it to uniformly distributed random numbers from the QRNG and a congruential RNG. Hence, our result shows that ML has potentials in benchmarking the quality of RNG devices.
|
['Syed Muhamad Assad', 'Nhan Duy Truong', 'Ping Koy Lam', 'Omid Kavehei', 'Jing Yan Haw']
|
2019-05-07
| null | null | null | null |
['cryptanalysis']
|
['miscellaneous']
|
[ 7.34948039e-01 -1.99963059e-02 1.52709529e-01 4.83737975e-01
-8.43947589e-01 -7.45023847e-01 8.21130335e-01 -2.13266805e-01
-2.79043674e-01 9.41860557e-01 -3.16782683e-01 -6.62257373e-01
-1.79114118e-01 -1.16552162e+00 -6.15115523e-01 -1.38751042e+00
-8.16582143e-02 1.56626523e-01 -1.30398348e-01 -4.85428542e-01
6.18262887e-01 7.34343350e-01 -1.07569885e+00 9.41352686e-04
4.63875562e-01 6.22017086e-01 -3.24886143e-01 1.20571101e+00
4.48323309e-01 8.72528672e-01 -1.18729496e+00 -4.04939741e-01
5.29489338e-01 -6.87789440e-01 -5.46814799e-01 -8.25017095e-01
-1.29205182e-01 -1.17125861e-01 -7.05437660e-01 1.28544283e+00
6.40858412e-01 -5.41337132e-01 6.95943773e-01 -1.20238793e+00
-3.05575520e-01 8.87458146e-01 2.75598288e-01 -2.60672212e-01
1.91939965e-01 7.04988360e-01 9.38660502e-01 -2.85809010e-01
7.14173973e-01 6.48939967e-01 4.30433959e-01 6.40435100e-01
-1.59912229e+00 -9.13097322e-01 -1.18185270e+00 6.52258620e-02
-1.39077640e+00 -5.68136454e-01 7.95375228e-01 -1.22042537e-01
7.94868767e-01 1.67889774e-01 4.30374295e-01 1.14978039e+00
1.01236939e+00 -8.17079768e-02 1.58528411e+00 -9.44191933e-01
5.56673944e-01 7.39568546e-02 -4.48860765e-01 2.43284062e-01
6.21884048e-01 9.69054520e-01 -4.67369974e-01 -4.30075347e-01
4.49637473e-01 -6.70728624e-01 -8.83725882e-02 -9.75990891e-02
-1.28451145e+00 6.44596338e-01 1.29100844e-01 3.56825471e-01
-2.47683719e-01 6.48390830e-01 1.17720075e-01 6.29922569e-01
-4.28267539e-01 1.02295732e+00 -1.48785263e-01 -2.15538815e-01
-6.82421565e-01 1.04426377e-01 9.86755192e-01 8.03527057e-01
7.47736275e-01 -3.78664955e-02 -8.01937357e-02 -2.65882045e-01
4.57207598e-02 1.22761190e+00 1.54119864e-01 -7.72009671e-01
1.80625021e-01 1.27324358e-01 2.73008764e-01 -7.47777581e-01
-4.73179758e-01 -2.92438775e-01 -1.00225544e+00 4.46330011e-01
4.52795953e-01 -3.01611900e-01 -6.72686160e-01 1.45682061e+00
-1.94726929e-01 -7.61370957e-02 5.30880451e-01 2.32737735e-01
4.82212782e-01 4.31231886e-01 -3.55447292e-01 -2.32805535e-01
7.01867282e-01 1.10130601e-01 -6.47336125e-01 2.88080841e-01
6.55564010e-01 -7.80441821e-01 4.26349759e-01 5.62740743e-01
-6.99082613e-01 -3.22432727e-01 -1.42296028e+00 3.56872112e-01
-3.53542455e-02 -2.22334683e-01 5.52350104e-01 1.44086778e+00
-8.36144865e-01 9.04562294e-01 -6.19296551e-01 3.84885728e-01
1.65130541e-01 9.05143440e-01 -7.45582283e-02 6.25676140e-02
-1.30338705e+00 7.68629193e-01 2.58849949e-01 2.64220089e-01
-6.56643569e-01 -2.60329217e-01 -2.21373454e-01 -2.23253533e-01
-5.45829162e-02 -5.72025478e-01 8.82916451e-01 -2.59861320e-01
-1.95177436e+00 6.32570028e-01 -3.48287709e-02 -6.31506562e-01
4.00019556e-01 2.95148939e-01 -6.87334776e-01 2.37991780e-01
-3.20171237e-01 -2.44284570e-01 9.88952994e-01 -9.38796639e-01
1.16845414e-01 -8.81475117e-03 -1.53824016e-01 -7.59914458e-01
3.36994082e-01 -4.38287370e-02 5.98957717e-01 -1.54972598e-01
2.18656823e-01 -1.25639129e+00 -5.41486144e-01 -1.10020125e+00
-9.16830540e-01 2.34462857e-01 1.90114200e-01 1.82973593e-01
1.13570654e+00 -1.81486404e+00 -2.17211127e-01 8.59183729e-01
2.58003145e-01 4.30779636e-01 -3.99406552e-02 1.11025655e+00
-7.07350150e-02 4.26213861e-01 7.80453300e-03 3.21829230e-01
-4.93352823e-02 -2.18974650e-01 -4.81420010e-01 7.26749837e-01
3.07009846e-01 1.14106655e+00 -8.40680718e-01 1.53819919e-01
2.26593316e-01 1.21304512e-01 -2.69261301e-01 1.25685468e-01
-2.63214380e-01 6.46135867e-01 -4.20534343e-01 6.36250257e-01
8.02814901e-01 -7.05346167e-02 3.02665740e-01 -2.27463543e-01
-6.60000518e-02 5.90863109e-01 -1.01017165e+00 9.75083768e-01
-2.55701095e-01 6.64999723e-01 -3.53591174e-01 -4.12279069e-01
9.85360801e-01 2.60141909e-01 3.64335775e-01 -8.33699763e-01
4.60853815e-01 7.01149642e-01 6.11185372e-01 -4.11621928e-01
7.41672873e-01 -4.82338488e-01 -4.93941158e-01 8.17274570e-01
7.73944557e-02 -7.61469185e-01 -1.90142795e-01 3.62946302e-01
1.57552457e+00 -7.60632679e-02 4.56933528e-01 -3.02131530e-02
4.48139668e-01 -2.41840884e-01 2.38458559e-01 1.36553895e+00
-1.74447581e-01 4.22220618e-01 9.22156930e-01 -1.73600003e-01
-1.30290854e+00 -1.05050111e+00 -1.46877915e-01 2.35053431e-02
3.69382530e-01 -6.15617812e-01 -3.98589045e-01 -9.46473703e-02
-2.44193017e-01 7.89712667e-01 -9.84870046e-02 -5.48903048e-01
-6.11781836e-01 -1.29058421e+00 1.00463331e+00 -2.06442043e-01
1.37940735e-01 -9.77585733e-01 -2.21222162e-01 4.21262920e-01
2.30500743e-01 -1.41534781e+00 5.80851555e-01 3.63676608e-01
-2.62486190e-01 -1.21943152e+00 1.86266795e-01 5.03782332e-02
4.00249064e-01 -3.87437791e-01 1.10328722e+00 -5.92090599e-02
-4.06204224e-01 1.84099138e-01 -2.85574615e-01 -4.68874484e-01
-1.44633257e+00 -1.51375458e-01 3.54047805e-01 -1.08971141e-01
3.45267206e-01 -7.79281139e-01 -6.00209594e-01 4.84049395e-02
-7.94646978e-01 -3.54347110e-01 7.77506769e-01 6.92783117e-01
1.90225154e-01 4.37780768e-01 4.93724853e-01 -9.34458792e-01
4.56608087e-01 -1.68091014e-01 -1.06667209e+00 5.90145364e-02
-5.45908093e-01 4.86190557e-01 9.88685608e-01 -2.59849370e-01
-3.63469481e-01 -1.55150503e-01 -8.82222429e-02 3.57281536e-01
7.00664381e-03 2.55619828e-02 -2.85014778e-01 -7.28221118e-01
1.00572085e+00 1.54258072e-01 -2.88565397e-01 3.70470643e-01
5.39588392e-01 7.43256569e-01 3.69897932e-01 -5.02038836e-01
1.64804578e+00 4.71391141e-01 1.05388939e+00 -9.75975573e-01
-8.54094848e-02 -2.77072378e-02 -4.49698508e-01 1.58551391e-02
2.86382079e-01 -6.03171647e-01 -1.23578453e+00 9.80554581e-01
-9.92615402e-01 -1.16113111e-01 -3.39967370e-01 5.01867950e-01
-6.01245403e-01 1.34496108e-01 -6.19901061e-01 -1.40246344e+00
-1.76984057e-01 -1.12128949e+00 7.82971501e-01 3.02183330e-01
8.68265107e-02 -6.30411446e-01 1.11412801e-01 -1.11947367e-02
6.14943564e-01 5.81408560e-01 1.00609553e+00 -6.04573905e-01
-1.27158666e+00 -7.51173139e-01 2.16728628e-01 2.27773964e-01
7.60913128e-03 2.27569282e-01 -1.19518065e+00 -2.72741169e-01
2.77650952e-01 -3.48244697e-01 6.63143516e-01 9.03967619e-02
5.84778726e-01 -1.93473235e-01 -1.66988328e-01 6.50384486e-01
1.82278895e+00 1.67391151e-01 1.15358865e+00 3.20210367e-01
4.60732251e-01 1.81014568e-01 1.52943149e-01 2.67044544e-01
-2.29209766e-01 4.63762492e-01 4.62292582e-01 2.35003456e-01
4.07234102e-01 -5.57411537e-02 6.59382224e-01 6.37082934e-01
-1.88547492e-01 -4.29512322e-01 -6.66534126e-01 -9.79215056e-02
-1.03569412e+00 -1.31357980e+00 -5.36947548e-01 2.44634128e+00
7.91556358e-01 3.03441346e-01 -1.60503164e-01 4.40059572e-01
5.97154081e-01 1.14624478e-01 -1.09831855e-01 -5.72042942e-01
-5.74195862e-01 1.12190545e+00 1.10002649e+00 3.85478824e-01
-9.59325492e-01 7.16277480e-01 6.44261026e+00 8.73689532e-01
-1.28651106e+00 -2.97003388e-01 3.41344088e-01 1.39321476e-01
-6.47007227e-01 3.17727536e-01 -8.11100245e-01 5.52152932e-01
1.36652517e+00 -1.15006894e-01 5.71482003e-01 1.83657616e-01
1.35050297e-01 -2.35083312e-01 -9.93100226e-01 8.04097235e-01
-3.73482972e-01 -1.43381941e+00 -3.22474986e-01 2.64570445e-01
7.42706656e-01 3.51731889e-02 3.30474943e-01 1.38504162e-01
3.79982471e-01 -1.35636747e+00 3.41041118e-01 6.56163037e-01
9.69305038e-01 -6.97089195e-01 1.03995991e+00 1.91026837e-01
-4.24982697e-01 -9.28926021e-02 -3.05947214e-01 -1.55921713e-01
9.57878679e-02 9.22287941e-01 -1.19795680e+00 7.32863903e-01
-1.56374931e-01 -1.19161032e-01 -5.10577381e-01 6.49572670e-01
-5.91084123e-01 9.68482137e-01 -3.68149191e-01 -3.52462620e-01
-1.27359092e-01 -4.55375075e-01 7.60862589e-01 7.39451468e-01
3.15358669e-01 -5.77014424e-02 -6.26581788e-01 1.07678974e+00
-8.78615305e-02 -2.97645867e-01 -6.71977341e-01 -4.40834194e-01
5.54951668e-01 1.14680386e+00 -5.49161613e-01 3.64452243e-01
-4.64426279e-02 7.29847193e-01 -3.81931126e-01 3.55175674e-01
-5.20516932e-01 -7.21101344e-01 4.16931272e-01 -3.06078270e-02
1.82496399e-01 -5.56647360e-01 -4.93636519e-01 -1.15450442e+00
-5.17101809e-02 -9.21844184e-01 -4.01710838e-01 -3.36366981e-01
-1.15080583e+00 4.06287730e-01 -5.50036967e-01 -1.42334199e+00
-5.29042184e-01 -8.04382324e-01 -5.39703846e-01 1.02044177e+00
-1.42537904e+00 -8.89002442e-01 2.85376668e-01 2.20189199e-01
-9.68246520e-01 -2.28293687e-01 1.12645793e+00 -2.23403841e-01
-2.36661285e-01 5.45772493e-01 5.81652999e-01 3.39504063e-01
4.86711323e-01 -1.23600185e+00 7.35051692e-01 1.14405668e+00
2.88647026e-01 9.63728487e-01 8.73477042e-01 -5.75992227e-01
-1.94498229e+00 -6.88901722e-01 7.45058596e-01 -5.93130887e-01
8.67981851e-01 -6.67686224e-01 -1.65099084e-01 1.71768710e-01
-3.97730879e-02 2.34035775e-01 7.76736856e-01 -2.80187249e-01
-3.39878887e-01 4.54887524e-02 -1.27557647e+00 5.70417941e-01
5.62810838e-01 -1.09721935e+00 -1.30559281e-01 4.51614469e-01
4.99243110e-01 -3.06732565e-01 -6.79396272e-01 3.85416001e-01
7.98366904e-01 -1.29336262e+00 9.93836522e-01 -4.52214122e-01
2.84080029e-01 -4.83121514e-01 -2.44301856e-01 -8.72441530e-01
5.57431057e-02 -1.56109500e+00 7.28709996e-02 9.16008353e-01
3.67715895e-01 -8.58934104e-01 8.76446843e-01 4.23643649e-01
4.91080791e-01 -1.67792633e-01 -1.10963929e+00 -8.97558689e-01
2.96790808e-01 -7.02029705e-01 5.05379975e-01 5.28870761e-01
-1.87989682e-01 1.30221248e-01 -5.17659247e-01 4.33582008e-01
7.01395750e-01 -1.87366698e-02 9.71204877e-01 -1.08526683e+00
-4.85459983e-01 -3.11263621e-01 -9.74417686e-01 -4.08310473e-01
-2.10861132e-01 -8.05629015e-01 3.58954296e-02 -5.96170008e-01
-2.78842360e-01 -7.43837416e-01 -3.65143478e-01 -3.09886932e-01
-9.01624933e-02 4.21368331e-01 6.05001673e-02 6.55246004e-02
-1.68385610e-01 3.67943555e-01 7.45769680e-01 -2.80176010e-03
-3.56810987e-01 5.81242859e-01 -4.61893976e-01 2.46723950e-01
9.39974129e-01 -7.44617283e-01 1.27868161e-01 4.54417795e-01
8.72072101e-01 -6.28667623e-02 5.55119574e-01 -1.27589631e+00
3.13420713e-01 -6.91111460e-02 1.59240946e-01 -1.99827358e-01
2.29667798e-01 -4.27785337e-01 3.74916673e-01 6.59637749e-01
3.31281759e-02 -3.59125584e-01 -1.78485021e-01 4.52043205e-01
1.49163932e-01 -5.02996266e-01 7.93835580e-01 -9.49809030e-02
-9.13071185e-02 6.96708411e-02 -5.14065027e-01 -2.48562936e-02
8.48064899e-01 2.66179466e-03 -6.32909119e-01 -3.74869853e-01
-2.29090914e-01 -5.55124581e-01 6.73152745e-01 -3.26943994e-01
4.50251222e-01 -9.80737746e-01 -6.19381309e-01 5.55585861e-01
-2.06714664e-02 -4.45739001e-01 -7.46223032e-02 7.46944845e-01
-8.40584397e-01 5.92913687e-01 3.09757330e-02 -4.77741331e-01
-8.90354037e-01 4.40789849e-01 5.49159586e-01 -3.53956252e-01
-1.49339765e-01 8.29606533e-01 -7.55548894e-01 -2.31079713e-01
-4.41725314e-01 -1.18155129e-01 2.35577285e-01 -4.37983245e-01
3.98978025e-01 1.75078943e-01 3.38211507e-01 -4.72979963e-01
-2.22445339e-01 4.55855876e-01 2.27251753e-01 -2.94928908e-01
1.00198209e+00 1.09522574e-01 -3.15056384e-01 2.40920678e-01
1.12347937e+00 9.28326249e-01 -5.44504166e-01 -5.74722998e-02
2.16492832e-01 -2.40115136e-01 -3.11599910e-01 -4.89677727e-01
-4.37136650e-01 6.90662205e-01 4.91743714e-01 5.27172804e-01
1.00952566e+00 -2.03472555e-01 6.20285451e-01 6.87243164e-01
9.23889756e-01 -8.11411619e-01 -2.47791439e-01 3.59196424e-01
-1.27137313e-02 -8.62969041e-01 1.54910952e-01 -2.52189606e-01
-1.10831177e-02 1.41887641e+00 -3.38845521e-01 -2.47065306e-01
3.32654893e-01 5.87686539e-01 9.51933786e-02 -7.15829358e-02
-4.61750984e-01 -1.36312202e-01 -1.26311392e-01 8.62159193e-01
2.18908712e-02 3.45236152e-01 -1.11881882e-01 3.11107963e-01
-7.77102411e-01 -1.69904768e-01 1.32880127e+00 7.40422606e-01
-7.69771561e-02 -1.79614079e+00 -4.76011932e-01 2.70499825e-01
-6.58997357e-01 -5.53714752e-01 -3.91686082e-01 6.41474903e-01
3.04506481e-01 1.15561831e+00 -5.72314262e-01 -7.78086603e-01
-7.51302615e-02 8.37035384e-03 7.81146884e-01 -3.60736370e-01
-5.37457764e-01 -3.49939853e-01 -6.20900979e-03 -5.85644364e-01
-4.79556859e-01 -6.65774643e-01 -1.08732331e+00 -4.13629293e-01
-7.12771297e-01 1.00871108e-01 7.15837836e-01 1.10321569e+00
2.04902843e-01 2.45351434e-01 1.14562738e+00 -3.60681653e-01
-8.94336641e-01 -5.49477220e-01 -8.04636538e-01 -6.51437938e-02
6.25455797e-01 4.06050906e-02 -8.33702028e-01 -4.16606456e-01]
|
[5.524283409118652, 5.077869415283203]
|
640a0402-3c5b-4108-8b0b-69d76f8b0ec8
|
geometric-perception-based-efficient-text
|
2302.03873
| null |
https://arxiv.org/abs/2302.03873v1
|
https://arxiv.org/pdf/2302.03873v1.pdf
|
Geometric Perception based Efficient Text Recognition
|
Every Scene Text Recognition (STR) task consists of text localization \& text recognition as the prominent sub-tasks. However, in real-world applications with fixed camera positions such as equipment monitor reading, image-based data entry, and printed document data extraction, the underlying data tends to be regular scene text. Hence, in these tasks, the use of generic, bulky models comes up with significant disadvantages compared to customized, efficient models in terms of model deployability, data privacy \& model reliability. Therefore, this paper introduces the underlying concepts, theory, implementation, and experiment results to develop models, which are highly specialized for the task itself, to achieve not only the SOTA performance but also to have minimal model weights, shorter inference time, and high model reliability. We introduce a novel deep learning architecture (GeoTRNet), trained to identify digits in a regular scene image, only using the geometrical features present, mimicking human perception over text recognition. The code is publicly available at https://github.com/ACRA-FL/GeoTRNet
|
['D. Y. Silva', 'D. R. Jayakodi', 'P. N. Deelaka']
|
2023-02-08
| null | null | null | null |
['scene-text-recognition']
|
['computer-vision']
|
[ 4.28796321e-01 -3.07944685e-01 9.40808728e-02 -4.33296353e-01
-3.03989887e-01 -4.27578598e-01 5.76016963e-01 6.18370622e-02
-3.67309242e-01 2.58483499e-01 -4.23231423e-01 -2.41823107e-01
8.14734846e-02 -7.43251443e-01 -6.89205348e-01 -6.59245789e-01
5.69657087e-01 3.66304457e-01 -6.29322231e-02 2.59999901e-01
5.42473912e-01 6.06705248e-01 -1.43948686e+00 1.59686163e-01
6.89981222e-01 1.40360510e+00 2.51295894e-01 6.69837892e-01
-1.69948295e-01 9.19611514e-01 -5.24691284e-01 -6.79786384e-01
2.51902580e-01 -9.23850238e-02 -4.81224537e-01 3.80197287e-01
4.47229922e-01 -6.35947347e-01 -6.45787776e-01 1.10316682e+00
5.48523664e-01 6.48930296e-02 5.67535996e-01 -9.74862397e-01
-8.25359046e-01 5.17121136e-01 -5.41283906e-01 -3.66177440e-01
3.35887104e-01 -6.09747060e-02 7.69639313e-01 -9.18255448e-01
2.95664370e-01 8.92730474e-01 6.11941218e-01 5.15420318e-01
-1.00156355e+00 -4.17345941e-01 1.28385291e-01 1.05211429e-01
-1.67133307e+00 -6.23846829e-01 7.70558059e-01 -4.36522901e-01
6.13980055e-01 4.64497149e-01 2.74696767e-01 1.43588793e+00
2.57859409e-01 1.14330804e+00 8.19978178e-01 -4.73688990e-01
2.29727060e-01 5.15571177e-01 1.89103991e-01 8.25148165e-01
3.00546736e-01 -3.56330991e-01 -6.06417179e-01 1.80618063e-01
8.47875774e-01 2.31259331e-01 -2.55174816e-01 -4.69673097e-01
-1.03400147e+00 5.16250789e-01 8.77077729e-02 2.95526564e-01
-6.98524015e-03 4.24249843e-02 3.06084365e-01 2.04690993e-01
3.32939506e-01 1.18125759e-01 -3.47803563e-01 1.62383430e-02
-9.42976654e-01 -5.41107021e-02 7.33032525e-01 1.24020016e+00
4.90744203e-01 2.02808872e-01 5.57049662e-02 9.03570950e-01
3.79239142e-01 7.11054265e-01 6.59926236e-01 -2.77256221e-01
5.90510666e-01 6.07098937e-01 -1.21956490e-01 -1.28492463e+00
-2.52082378e-01 -3.19078803e-01 -1.25096679e+00 -1.14045434e-01
3.45026195e-01 -1.32917523e-01 -9.36531544e-01 1.22024262e+00
-2.75094411e-03 -2.63567090e-01 -2.56492555e-01 7.58915901e-01
6.55741990e-01 4.89302665e-01 -3.61383259e-01 1.17887162e-01
1.19224954e+00 -7.43577719e-01 -7.46663392e-01 -3.50894064e-01
7.18881726e-01 -9.01985586e-01 1.10577679e+00 6.42953455e-01
-7.57487714e-01 -6.16565108e-01 -1.05890560e+00 -5.03466725e-01
-8.53698730e-01 9.18623507e-01 4.01261061e-01 8.22303236e-01
-9.03988481e-01 3.59551102e-01 -7.60497928e-01 -6.39489949e-01
5.34009576e-01 4.18417633e-01 -2.56456196e-01 -9.09817740e-02
-6.66522503e-01 6.76231742e-01 3.07398289e-01 5.29462695e-01
-5.43573797e-01 -7.47333467e-02 -8.57496083e-01 2.20614783e-02
3.98692280e-01 -3.21194857e-01 1.01224470e+00 -7.56507754e-01
-1.71781266e+00 9.07029569e-01 -2.78804660e-01 -3.10360342e-01
8.68871868e-01 -3.39202255e-01 -3.28312665e-01 -1.04896739e-01
-1.93729430e-01 2.58546442e-01 1.19605803e+00 -1.07212210e+00
-3.09008956e-01 -4.14038956e-01 -2.44483039e-01 1.65778771e-01
-3.36143345e-01 -6.27307780e-03 -6.97105885e-01 -7.20617354e-01
1.30805746e-01 -6.40155613e-01 1.19494302e-02 3.12164783e-01
-7.40762711e-01 -1.32618435e-02 1.12005079e+00 -7.47629583e-01
1.05771589e+00 -2.48479104e+00 -1.11084215e-01 2.08804742e-01
2.51261473e-01 3.16303849e-01 2.01643288e-01 3.97754788e-01
2.09056109e-01 3.31711844e-02 -2.03298405e-01 -4.92843479e-01
1.26851246e-01 -1.27365574e-01 -4.77801651e-01 6.37808144e-01
2.01894660e-02 7.77567744e-01 -3.20864707e-01 -4.42651451e-01
7.89266884e-01 2.85621673e-01 -1.11017577e-01 -3.03350780e-02
-2.39378363e-01 2.68930405e-01 -7.76881695e-01 8.82554352e-01
8.02804291e-01 -3.87151867e-01 6.97931573e-02 -1.74560651e-01
-5.43211550e-02 -5.52947558e-02 -1.36649764e+00 1.41415989e+00
-2.85828978e-01 8.50669205e-01 1.72778994e-01 -1.17360437e+00
1.09026921e+00 5.26352003e-02 2.85442412e-01 -7.09664583e-01
5.87618709e-01 1.95859849e-01 -5.43231726e-01 -6.28798604e-01
6.93730056e-01 5.55034876e-01 6.19225986e-02 3.74078423e-01
-2.53308505e-01 -6.00435846e-02 -1.14495926e-01 -3.78986932e-02
7.62883663e-01 -1.26172267e-02 1.11784838e-01 -4.86304238e-02
5.68608820e-01 -2.38739178e-01 6.18488938e-02 8.77098739e-01
5.61132506e-02 7.84047484e-01 2.55298674e-01 -5.40621400e-01
-8.10482979e-01 -7.66714931e-01 -3.75238538e-01 8.13083053e-01
3.23586971e-01 -4.00355965e-01 -8.24645281e-01 -5.17141759e-01
-9.51025486e-02 5.89452386e-01 -3.73130143e-01 8.73750746e-02
-5.03521085e-01 -5.99856973e-01 6.79597199e-01 3.95498633e-01
9.64393497e-01 -8.49538684e-01 -5.74001253e-01 -1.56414330e-01
1.01578003e-02 -1.29577911e+00 -5.19801736e-01 1.77928820e-01
-6.65742159e-01 -9.65555608e-01 -7.36892700e-01 -8.38144004e-01
9.85742986e-01 2.40325466e-01 5.26138842e-01 1.06511109e-01
-5.03814042e-01 3.39356631e-01 -3.75940263e-01 -4.06207293e-01
-2.04199608e-02 4.68787178e-02 -3.20619941e-02 3.25960129e-01
5.08802831e-01 -2.20457632e-02 -3.04248422e-01 3.42105836e-01
-1.06035483e+00 2.23353878e-01 6.97251737e-01 8.07064891e-01
5.73628306e-01 4.53900933e-01 2.66396143e-02 -6.42057776e-01
4.81226206e-01 6.48049116e-02 -7.95788109e-01 4.55801368e-01
-4.63790685e-01 -3.64788324e-01 8.99277031e-01 -3.18769038e-01
-9.71617758e-01 3.10114741e-01 -1.80782139e-01 -3.35161775e-01
-4.77218270e-01 3.55928689e-01 -3.34585011e-01 -1.16618268e-01
5.09227991e-01 8.17552030e-01 2.37419619e-03 -6.37689948e-01
-6.02181554e-02 1.15261245e+00 6.67135343e-02 -4.20383871e-01
6.21871829e-01 5.08433938e-01 -3.33127826e-02 -1.22558177e+00
-6.37650490e-01 -2.13166058e-01 -8.10187340e-01 -4.85591367e-02
7.43966758e-01 -7.84241557e-01 -8.60203147e-01 1.11906004e+00
-1.07742560e+00 -3.82168382e-01 -1.47087976e-01 5.18456221e-01
-5.41533113e-01 7.16635525e-01 -5.05289435e-01 -9.05996621e-01
-3.12817812e-01 -1.20227504e+00 1.39131093e+00 1.09798312e-01
2.13377506e-01 -8.95168483e-01 -5.42136192e-01 2.85690725e-01
2.71453023e-01 1.90136041e-02 8.19836676e-01 -7.25573897e-01
-7.98763514e-01 -6.45841956e-01 -4.11402076e-01 5.65703988e-01
1.30520508e-01 3.64759564e-02 -1.18863380e+00 -3.57180595e-01
1.16226025e-01 -3.01062196e-01 5.81896365e-01 3.52499336e-01
1.85810077e+00 -3.53132159e-01 -3.42082798e-01 8.84126306e-01
1.33451498e+00 1.69070065e-01 6.21437550e-01 2.08685827e-02
8.19701493e-01 5.41181087e-01 4.28368032e-01 5.61225772e-01
2.12776288e-01 5.77281415e-01 2.67082483e-01 -4.31246758e-02
1.86735988e-01 -1.61734968e-01 3.10152680e-01 8.22110772e-01
2.16949940e-01 -6.76175237e-01 -9.17790174e-01 1.74793556e-01
-1.89500964e+00 -6.23831630e-01 -1.29499406e-01 2.27389550e+00
4.07088906e-01 3.89612047e-03 -2.93136209e-01 1.63486019e-01
9.20299649e-01 1.40259728e-01 -6.73157454e-01 -2.83003122e-01
-2.18817860e-01 9.20671690e-03 7.18292534e-01 1.24592647e-01
-1.17295170e+00 1.00143003e+00 6.13607931e+00 9.92795885e-01
-1.47196102e+00 -2.92488843e-01 6.97745860e-01 4.18414287e-02
2.65963048e-01 -3.72881770e-01 -8.49904060e-01 4.72853422e-01
5.52465737e-01 5.94043732e-02 5.98228514e-01 1.01666045e+00
3.05416062e-02 -4.77334037e-02 -1.23128343e+00 1.35739183e+00
2.83829719e-01 -1.12985718e+00 2.85901606e-01 4.02441472e-02
1.72819167e-01 -4.20112871e-02 2.62047410e-01 -1.05311023e-03
-1.34281935e-02 -1.14586401e+00 7.76890099e-01 3.65911156e-01
1.10417151e+00 -4.20651972e-01 7.38627732e-01 3.83286715e-01
-9.80394006e-01 -1.42284418e-02 -6.35481060e-01 1.86570853e-01
-1.86468899e-01 6.36542082e-01 -6.13678396e-01 7.18507707e-01
7.78950036e-01 9.33204293e-01 -7.24054158e-01 8.34231794e-01
6.18961826e-02 2.89596140e-01 -4.97421622e-01 -3.31418455e-01
-3.21419686e-02 -2.92366773e-01 2.48644352e-01 1.27096307e+00
5.02724051e-01 -2.17896104e-01 1.26068536e-02 7.67506897e-01
-1.72574297e-01 2.43147612e-01 -8.52724254e-01 -1.68681860e-01
3.17261994e-01 1.10522199e+00 -8.04605305e-01 -8.54754820e-02
-3.88872772e-01 1.12607265e+00 6.44128472e-02 3.45811486e-01
-8.08499932e-01 -7.27886796e-01 2.24608287e-01 1.10006087e-01
3.11906040e-01 -4.07199740e-01 -6.86544716e-01 -1.34485281e+00
5.59567750e-01 -9.09580112e-01 8.86717960e-02 -8.09024632e-01
-1.13004339e+00 4.61114168e-01 -3.55046690e-01 -1.26914597e+00
8.48779380e-02 -1.21719754e+00 -4.20454174e-01 7.23479867e-01
-1.26002717e+00 -1.30998468e+00 -5.52529335e-01 8.85206044e-01
6.52880788e-01 -3.58770549e-01 7.46050358e-01 3.04136097e-01
-1.05934572e+00 1.04589260e+00 5.52674353e-01 6.04969382e-01
5.03006697e-01 -9.15591598e-01 3.17200094e-01 8.61162782e-01
6.21790215e-02 5.86055636e-01 3.48463565e-01 -4.47515726e-01
-1.76531363e+00 -1.14557779e+00 5.65923214e-01 -6.09869957e-01
5.82323730e-01 -9.47042644e-01 -7.14169145e-01 8.41555834e-01
-1.47810310e-01 -1.72000498e-01 4.70289856e-01 -2.38609776e-01
-2.08003357e-01 -2.19739020e-01 -1.31922817e+00 7.66898692e-01
8.78593445e-01 -5.99503815e-01 -9.24254134e-02 5.03529370e-01
2.72501439e-01 -5.84338844e-01 -7.29278445e-01 1.12663262e-01
6.15303695e-01 -7.94568002e-01 7.40720987e-01 -4.90657659e-03
3.05891573e-01 -2.72608817e-01 -3.60785216e-01 -5.91262698e-01
-6.72722757e-02 -4.98182982e-01 -1.30648360e-01 1.23678696e+00
2.07748279e-01 -9.23814535e-01 8.40704799e-01 6.46636546e-01
8.84824246e-02 -4.94421303e-01 -1.00221181e+00 -7.03262627e-01
-1.50581241e-01 -5.08203328e-01 5.55409193e-01 9.50517237e-01
-1.95896462e-01 9.28728580e-02 -5.31484187e-01 3.55171740e-01
6.26757085e-01 -1.33465827e-02 8.02002251e-01 -1.16246021e+00
-1.89447030e-01 -4.23760384e-01 -5.01436472e-01 -1.47072697e+00
-9.24492851e-02 -6.13267839e-01 1.17642254e-01 -1.30265474e+00
-1.50720626e-01 -4.29259866e-01 6.57307431e-02 3.51974219e-01
3.27810913e-01 3.39356288e-02 5.47117181e-02 2.67795563e-01
-6.20705605e-01 5.59051931e-01 1.12384212e+00 -3.01947743e-01
8.62511620e-02 1.37599215e-01 -6.04757905e-01 7.39479721e-01
1.00342178e+00 -1.71139359e-01 -2.86032081e-01 -9.54721510e-01
1.06606781e-01 -2.67391086e-01 5.60938179e-01 -1.05740893e+00
4.88249391e-01 -4.18722555e-02 7.13601768e-01 -7.88811624e-01
4.89804327e-01 -1.17776859e+00 -2.60192510e-02 2.46011317e-01
-3.34653527e-01 -2.98435688e-01 1.08234547e-01 4.68994260e-01
-1.53297603e-01 -4.94303703e-01 7.10600972e-01 1.18170530e-02
-4.71237361e-01 3.53255749e-01 -4.78382945e-01 -5.85683107e-01
1.00361907e+00 -5.41540742e-01 -4.48724419e-01 -2.48187184e-01
-3.92599136e-01 -9.86364111e-03 6.06838405e-01 3.97884011e-01
7.01170921e-01 -8.35258126e-01 -3.33106607e-01 5.07376671e-01
1.65828288e-01 4.51685637e-01 1.99378863e-01 7.40748107e-01
-6.82066023e-01 7.58534431e-01 1.77407011e-01 -7.21544445e-01
-1.00970924e+00 6.40210748e-01 2.48978823e-01 -3.78819704e-02
-7.63145208e-01 6.10415161e-01 2.86203325e-01 -5.17592549e-01
6.32232308e-01 -3.92844111e-01 1.54756576e-01 -2.79577762e-01
4.84602422e-01 3.42124820e-01 3.22566032e-01 -3.79898220e-01
-2.03284308e-01 8.61011863e-01 -3.41120541e-01 2.45516300e-01
9.09470737e-01 -3.49169821e-01 -1.26617206e-02 3.55590761e-01
9.08025503e-01 -1.09386491e-02 -1.06200266e+00 -4.08381373e-01
-9.49490592e-02 -5.91336191e-01 1.73327208e-01 -7.21848309e-01
-9.78293955e-01 1.01335073e+00 5.88427901e-01 3.44988942e-01
1.15996766e+00 -3.25985700e-01 5.51076055e-01 9.76018369e-01
5.24111807e-01 -1.40028965e+00 -5.73894866e-02 5.59775233e-01
8.29782903e-01 -1.22468078e+00 -5.20636402e-02 -3.08074355e-01
-5.72924554e-01 1.26115763e+00 6.42680049e-01 1.11489981e-01
8.15508962e-01 3.20982069e-01 -9.95439105e-03 -1.58276647e-01
-2.95553088e-01 4.42183614e-02 -7.52717955e-03 6.24071658e-01
2.95239091e-01 -1.44776404e-01 3.13968331e-01 3.56832594e-01
-1.87392682e-01 1.83840189e-02 3.86263967e-01 1.06340575e+00
-1.64428785e-01 -8.48562002e-01 -4.29387391e-01 8.23098898e-01
-4.74557579e-01 -7.22515136e-02 -6.65182710e-01 7.36750066e-01
-1.21059872e-01 8.78243983e-01 9.42587554e-02 -4.56182152e-01
2.73389548e-01 -1.45889157e-02 3.01246792e-01 -3.56884927e-01
-2.31020480e-01 -1.55096054e-01 -2.45665133e-01 -2.78662294e-01
-2.95663308e-02 -5.95321357e-01 -8.35455775e-01 -7.83361614e-01
-5.37896752e-01 -3.40673596e-01 9.87274587e-01 7.95131862e-01
4.70911175e-01 3.37837487e-01 8.18685114e-01 -6.31276131e-01
-4.38549072e-01 -8.90527129e-01 -8.67076337e-01 6.82228059e-02
2.05303133e-01 -2.70268500e-01 -2.58064538e-01 2.91811854e-01]
|
[11.97212028503418, 2.1942291259765625]
|
f017630d-c7ad-4214-82a6-fb51daf86bff
|
ms-unique-multi-model-and-sharpness-weighted
|
1811.08947
| null |
http://arxiv.org/abs/1811.08947v1
|
http://arxiv.org/pdf/1811.08947v1.pdf
|
MS-UNIQUE: Multi-model and Sharpness-weighted Unsupervised Image Quality Estimation
|
In this paper, we train independent linear decoder models to estimate the
perceived quality of images. More specifically, we calculate the responses of
individual non-overlapping image patches to each of the decoders and scale
these responses based on the sharpness characteristics of filter set. We use
multiple linear decoders to capture different abstraction levels of the image
patches. Training each model is carried out on 100,000 image patches from the
ImageNet database in an unsupervised fashion. Color space selection and ZCA
Whitening are performed over these patches to enhance the descriptiveness of
the data. The proposed quality estimator is tested on the LIVE and the TID 2013
image quality assessment databases. Performance of the proposed method is
compared against eleven other state of the art methods in terms of accuracy,
consistency, linearity, and monotonic behavior. Based on experimental results,
the proposed method is generally among the top performing quality estimators in
all categories.
|
['Ghassan AlRegib', 'Dogancan Temel', 'Mohit Prabhushankar']
|
2018-11-21
| null | null | null | null |
['image-quality-estimation']
|
['computer-vision']
|
[ 9.60911959e-02 -4.19729918e-01 6.45699948e-02 -4.99951094e-01
-9.12666500e-01 -3.54177654e-01 2.36698508e-01 1.76465809e-01
-3.65002960e-01 3.43291789e-01 2.19479203e-01 3.62949789e-01
-2.53483236e-01 -6.31511748e-01 -6.39914632e-01 -7.28295803e-01
-6.56319112e-02 -1.33813903e-01 3.51154149e-01 3.16672847e-02
5.81109345e-01 2.95639008e-01 -1.65164959e+00 3.75561774e-01
1.00753593e+00 1.41485310e+00 9.59291384e-02 7.72937536e-01
3.02083552e-01 5.65089822e-01 -5.95347643e-01 -4.44731772e-01
4.92554039e-01 -3.02221984e-01 -3.80045235e-01 4.87501830e-01
6.80127621e-01 -4.11708921e-01 -1.29587978e-01 1.23684943e+00
5.38685083e-01 1.85994849e-01 7.98228145e-01 -9.10129011e-01
-7.52780139e-01 1.12928599e-01 -8.80014122e-01 2.30177537e-01
3.30997586e-01 1.72704592e-01 9.65377808e-01 -9.37218368e-01
3.00525010e-01 1.16050041e+00 4.86380011e-01 1.18277356e-01
-1.36532605e+00 -8.11358333e-01 -2.26331815e-01 4.79108423e-01
-1.41322458e+00 -6.20823562e-01 7.14379609e-01 -3.92164201e-01
6.05179071e-01 -1.71296835e-01 4.33337331e-01 4.69708502e-01
6.39823139e-01 4.04816657e-01 1.39504445e+00 -3.68962437e-01
3.24936509e-01 2.91388571e-01 1.59477040e-01 7.04446495e-01
1.10289656e-01 2.37166002e-01 -7.18094110e-01 1.97828710e-01
7.19217062e-01 -4.61690098e-01 -2.52215445e-01 -2.01974452e-01
-7.99973190e-01 5.89960694e-01 5.06184578e-01 9.70633850e-02
-5.55198610e-01 -3.16427425e-02 1.32311031e-01 4.60218608e-01
4.14719224e-01 7.24662542e-02 -1.28284425e-01 9.22837183e-02
-1.18011045e+00 -2.32511103e-01 4.46814895e-01 8.46446037e-01
6.96029484e-01 5.97651005e-02 -1.78599715e-01 8.61419141e-01
4.72051114e-01 5.20852387e-01 1.92886397e-01 -9.07526731e-01
3.19797069e-01 5.45079648e-01 -5.22386655e-02 -1.19042695e+00
-1.32940382e-01 -4.50351834e-01 -9.19615865e-01 6.78592563e-01
3.17454666e-01 8.23197812e-02 -1.13413095e+00 1.26313376e+00
-1.08149886e-01 4.57058176e-02 4.37045097e-02 8.90554309e-01
8.15593779e-01 9.55695450e-01 1.86658993e-01 -9.69030783e-02
1.06895125e+00 -6.03762686e-01 -6.05731428e-01 2.79964022e-02
-3.25680882e-01 -9.80663657e-01 1.02422702e+00 7.55438387e-01
-1.18164027e+00 -1.27085757e+00 -1.23227477e+00 1.76565915e-01
-4.64845710e-02 5.83283901e-01 -6.12958893e-02 8.12363267e-01
-1.15030432e+00 4.14866716e-01 -5.71106553e-01 -1.04027390e-01
4.02823806e-01 2.54824072e-01 -3.04686666e-01 -8.01912919e-02
-8.22077453e-01 8.14984739e-01 1.51235372e-01 -5.36579080e-02
-1.16429830e+00 -3.28258425e-01 -6.37787282e-01 1.13275684e-01
-5.41756526e-02 -3.05029333e-01 6.81762636e-01 -1.41708982e+00
-1.47910905e+00 7.35241771e-01 9.39058289e-02 -3.72711957e-01
2.52175003e-01 -1.05707400e-01 -5.63993692e-01 3.10188115e-01
-1.23603828e-01 6.65262461e-01 1.16242051e+00 -1.41442919e+00
-6.34426177e-01 -2.32267439e-01 -2.04224318e-01 3.38489413e-01
-2.89557725e-01 4.03465666e-02 -7.78270125e-01 -5.61706424e-01
7.27889389e-02 -4.84061003e-01 9.95602384e-02 1.78793371e-01
-2.42219165e-01 1.52380154e-01 2.05553398e-01 -8.54564250e-01
1.15983486e+00 -2.45894670e+00 6.12312332e-02 5.53938270e-01
1.12774819e-01 -1.87734589e-01 -3.38992804e-01 1.09547213e-01
1.74090251e-01 -2.11694852e-01 -1.95411384e-01 -4.15292643e-02
-1.45016730e-01 -2.39560172e-01 3.69692653e-01 7.93556809e-01
2.76091844e-01 2.80868500e-01 -3.85292828e-01 -5.67938030e-01
3.69266093e-01 4.53040242e-01 -5.28575540e-01 4.07543361e-01
1.75292671e-01 3.19363445e-01 -2.42239493e-03 5.94668508e-01
8.59458268e-01 -6.34773970e-02 -1.89614281e-01 -6.61445022e-01
2.20020581e-02 -9.60631762e-03 -1.47887874e+00 1.50722611e+00
-4.78308439e-01 7.92677999e-01 -9.08684433e-02 -7.57254899e-01
1.07696688e+00 3.87676835e-01 4.46873128e-01 -9.38661873e-01
1.57801419e-01 3.67154218e-02 3.38394582e-01 -3.55112582e-01
3.49860936e-01 -1.83217358e-02 1.43489301e-01 1.12758085e-01
4.87926275e-01 -8.93372968e-02 1.74968958e-01 -1.04743354e-01
6.72537386e-01 -1.33317098e-01 2.81060129e-01 -3.16711605e-01
6.14564240e-01 -2.74913132e-01 3.72437716e-01 6.92550600e-01
-4.62468177e-01 7.00118124e-01 6.69081435e-02 2.23590638e-02
-1.19430614e+00 -1.49009299e+00 -2.72946864e-01 8.55657101e-01
5.77506363e-01 -5.55278733e-02 -6.76597416e-01 -2.61945188e-01
-2.84430325e-01 4.80563790e-01 -5.19574285e-01 -2.64137596e-01
-1.60643309e-01 -5.04236877e-01 3.62822056e-01 2.11931035e-01
9.58206773e-01 -1.07703745e+00 -5.48289180e-01 1.54961973e-01
3.22095081e-02 -9.50991631e-01 -3.67819846e-01 5.91534972e-02
-6.66928291e-01 -1.02550876e+00 -6.99030876e-01 -8.27870607e-01
6.22140646e-01 1.15875088e-01 1.12520492e+00 -1.87613800e-01
-1.43741131e-01 3.48569870e-01 -3.18408042e-01 -2.39122391e-01
-4.23400044e-01 -1.80210903e-01 -1.83580264e-01 4.08177525e-01
1.06893070e-01 -2.79475331e-01 -7.94940174e-01 3.20526958e-01
-1.00335360e+00 -1.01072177e-01 7.94714272e-01 5.33403099e-01
7.34794438e-01 8.48735213e-01 2.63930261e-01 -4.51895356e-01
9.32465851e-01 -1.89803824e-01 -7.52164006e-01 1.88861892e-01
-7.59109318e-01 6.25336841e-02 4.66397166e-01 -4.22849357e-01
-1.32828891e+00 3.06146336e-03 5.93799949e-02 6.05998561e-02
-2.75286674e-01 3.33238602e-01 -2.79655755e-01 -3.00610811e-01
6.72455907e-01 2.49297380e-01 -1.35539442e-01 -4.04444724e-01
2.96691626e-01 5.43574274e-01 6.52018070e-01 -1.72469929e-01
8.58920693e-01 1.95026532e-01 -2.32403740e-01 -7.18489766e-01
-2.14778438e-01 -4.31730479e-01 -3.19270819e-01 -7.09231377e-01
9.98816490e-01 -8.60564351e-01 -3.79714847e-01 7.14826703e-01
-8.49073172e-01 3.55621614e-02 1.48387924e-01 5.26094437e-01
-3.36657643e-01 2.58480042e-01 -7.24148631e-01 -7.10141063e-01
-5.10731280e-01 -1.46234024e+00 7.30301261e-01 5.30897319e-01
1.34470642e-01 -6.91172183e-01 -3.12230922e-02 1.44725785e-01
4.62783337e-01 -1.84339061e-02 9.07425284e-01 -1.83588326e-01
-3.32496434e-01 -2.84799993e-01 -3.98672879e-01 8.65005314e-01
2.27896705e-01 2.21083134e-01 -8.59923065e-01 -1.66660428e-01
7.45209530e-02 -1.81490883e-01 8.58288169e-01 7.38328636e-01
9.24837232e-01 -2.58375853e-02 2.89377302e-01 4.71085876e-01
1.91946578e+00 4.02427584e-01 8.13004553e-01 4.53118294e-01
2.32887357e-01 4.13930595e-01 4.19708580e-01 3.67004067e-01
1.32146686e-01 4.93081838e-01 5.06065786e-01 -5.08825839e-01
-3.09045523e-01 1.20042332e-01 5.19306421e-01 6.51837289e-01
4.96900678e-02 -3.18107724e-01 -7.29387403e-01 3.25728893e-01
-1.11262441e+00 -8.23526859e-01 2.58977013e-03 2.17944217e+00
6.61030054e-01 5.08330464e-01 5.17218076e-02 3.44625622e-01
7.03515291e-01 -3.02626602e-02 -3.30344796e-01 -4.18868184e-01
-1.56127706e-01 4.67349231e-01 6.91956818e-01 3.64288449e-01
-1.08783913e+00 6.29734516e-01 6.80863047e+00 7.41104782e-01
-1.09356642e+00 -9.34851319e-02 8.43538344e-01 1.34590045e-02
1.91651329e-01 -1.68324232e-01 -3.39923531e-01 4.53721911e-01
8.47446442e-01 -3.60770077e-02 5.12665033e-01 5.60816646e-01
3.66646022e-01 -6.19241416e-01 -9.01255369e-01 1.08333135e+00
1.35183424e-01 -8.83083940e-01 -3.10837701e-02 -2.22486839e-01
9.88638222e-01 -1.07696190e-01 5.53930640e-01 -1.82790667e-01
-9.40527022e-03 -1.06528723e+00 8.57844293e-01 9.42049980e-01
6.26949310e-01 -8.20249915e-01 8.13543022e-01 1.02106899e-01
-1.05550730e+00 -5.01457006e-02 -3.64132315e-01 2.80400544e-01
-1.31109282e-01 5.34599304e-01 -2.01829359e-01 3.46462309e-01
9.21316981e-01 5.21773756e-01 -8.99569035e-01 1.53458416e+00
2.20185090e-02 7.75486350e-01 -2.04394013e-02 2.89862156e-01
-1.21752368e-02 -2.80628324e-01 3.04804474e-01 1.23794365e+00
2.74385631e-01 2.49115918e-02 -7.56662786e-02 7.65154779e-01
-1.34049699e-01 1.89453840e-01 -1.80413201e-01 3.33083034e-01
3.10733289e-01 9.95187700e-01 -6.30560875e-01 -4.59361851e-01
-6.22245371e-01 9.01510715e-01 -1.77811265e-01 3.47575426e-01
-9.20070410e-01 -3.74813467e-01 3.01635265e-01 -2.99895331e-02
3.17349851e-01 -4.61411197e-03 -4.85313743e-01 -8.38049471e-01
4.08631861e-02 -1.28839314e+00 1.56737953e-01 -9.09694076e-01
-1.16921306e+00 7.57395744e-01 -2.06679292e-02 -1.60013056e+00
2.02046558e-01 -4.27965254e-01 -4.16650087e-01 8.87125313e-01
-1.34357452e+00 -6.88712955e-01 -4.53893334e-01 6.34360731e-01
6.77523315e-01 -3.37856919e-01 5.23111880e-01 3.91715676e-01
-3.81333470e-01 5.96317351e-01 2.31061518e-01 1.16781078e-01
8.32640588e-01 -1.01188827e+00 5.20468093e-02 1.14381361e+00
1.51023343e-01 2.48803839e-01 8.05548072e-01 -3.62077534e-01
-1.20742905e+00 -7.25579858e-01 2.79771507e-01 5.80139160e-02
1.92730099e-01 -1.49145216e-01 -9.30863261e-01 6.40867949e-02
5.51901698e-01 -1.77648425e-01 5.17815351e-01 -4.00541127e-01
-3.46997023e-01 -5.08744478e-01 -1.20366275e+00 8.79092887e-02
2.68060774e-01 -4.36530590e-01 -3.61453235e-01 -2.58417398e-01
-7.03008473e-02 -2.09515870e-01 -9.98114526e-01 3.48912895e-01
6.05936885e-01 -1.42294943e+00 8.27475071e-01 1.75614744e-01
6.66243911e-01 -4.87490594e-01 -2.74274856e-01 -1.29605520e+00
-5.56524038e-01 1.24705061e-01 3.70230138e-01 1.27247381e+00
5.29791415e-01 -2.24130243e-01 5.08756101e-01 2.93032706e-01
2.36926571e-01 -3.04323852e-01 -6.72366500e-01 -3.05975884e-01
-3.15066040e-01 -4.14620131e-01 3.41498107e-01 2.46080562e-01
-4.58612472e-01 2.12348685e-01 -4.24268514e-01 3.77815932e-01
9.41369295e-01 -2.15115282e-03 6.07331634e-01 -9.41692293e-01
-4.64023858e-01 -5.29574692e-01 -7.34184444e-01 -7.72698462e-01
-3.32353771e-01 -4.21385735e-01 2.64774203e-01 -1.50596988e+00
2.99987078e-01 2.53793187e-02 -6.21789038e-01 4.41655740e-02
-2.34203249e-01 5.75629234e-01 1.72236204e-01 1.31141499e-01
-6.34734750e-01 3.59322190e-01 1.16943991e+00 -4.15540218e-01
-2.14644626e-01 6.93789497e-02 -3.88267934e-01 4.88584816e-01
8.15864623e-01 -3.63116413e-01 -4.44799542e-01 -4.97328937e-01
-1.07912675e-01 1.54847309e-01 2.88903415e-01 -1.56600070e+00
1.42108008e-01 -4.19097207e-03 8.99370074e-01 -4.27513242e-01
1.64389223e-01 -9.08929944e-01 2.42480040e-01 3.31760734e-01
-6.21569991e-01 -7.99300428e-03 3.41452658e-01 4.12990898e-01
-4.71308559e-01 1.00062378e-02 1.37774146e+00 6.33480251e-02
-9.70607340e-01 1.37156114e-01 -4.07322735e-01 -2.33068004e-01
6.25131965e-01 -4.04859841e-01 5.84791079e-02 -5.75817347e-01
-4.74026412e-01 -2.91569948e-01 4.30476815e-01 2.22793326e-01
9.78140533e-01 -1.21289217e+00 -8.31978679e-01 3.76836747e-01
2.99189985e-01 -6.54603660e-01 1.88702017e-01 6.11358404e-01
-4.97492135e-01 -1.55186757e-01 -7.34927118e-01 -4.93871391e-01
-1.27849853e+00 2.80312032e-01 4.46537197e-01 4.56088744e-02
-3.58178079e-01 5.75259924e-01 9.43752676e-02 3.91276032e-01
4.29952323e-01 -3.61158162e-01 -5.25070012e-01 -2.13879675e-01
3.39139760e-01 4.46708500e-01 1.79804668e-01 -8.45205426e-01
-2.65013725e-01 8.42776000e-01 1.32400244e-01 -3.26643020e-01
1.14128768e+00 -2.72230864e-01 -6.16925359e-02 2.20492661e-01
1.24350703e+00 1.18427332e-02 -1.38127089e+00 -1.37141570e-01
-6.24437295e-02 -4.96315300e-01 4.90016520e-01 -8.97461534e-01
-1.35093415e+00 6.46215498e-01 1.42609334e+00 7.36529902e-02
1.93344748e+00 -4.00742948e-01 3.27891618e-01 -2.03647360e-01
-2.21333243e-02 -1.18913901e+00 1.61991343e-01 -3.37043107e-02
7.05656171e-01 -1.27658427e+00 9.23734065e-03 -3.14242952e-02
-6.21205330e-01 1.24039781e+00 6.32590652e-01 -5.30952871e-01
7.12983847e-01 1.07955813e-01 2.97971934e-01 -8.86885747e-02
-6.63204610e-01 -1.77233353e-01 8.04074168e-01 5.94181299e-01
6.06045425e-01 -1.39474766e-02 -3.59888881e-01 1.76746428e-01
-2.80447006e-02 -8.12576488e-02 4.27174628e-01 3.90114039e-01
-6.23434067e-01 -7.20527470e-01 -5.46591938e-01 4.20218229e-01
-4.12280709e-01 -1.10110074e-01 -1.76893979e-01 5.68447709e-01
2.22396463e-01 1.44458210e+00 -6.75808787e-02 -6.01245642e-01
4.67237264e-01 -4.60628599e-01 4.96974617e-01 -2.42537841e-01
-6.45633161e-01 3.64465624e-01 -1.93987533e-01 -4.84261245e-01
-3.25815916e-01 -2.99078077e-01 -1.00227427e+00 -1.49114370e-01
-2.97601938e-01 -1.19399831e-01 7.93896317e-01 5.80750942e-01
2.64635086e-02 4.44254816e-01 8.59737456e-01 -6.06545508e-01
-4.91947502e-01 -1.01995981e+00 -6.02298141e-01 6.93682432e-01
3.44978213e-01 -6.11301243e-01 -3.78597856e-01 3.43080759e-01]
|
[11.766498565673828, -1.9175348281860352]
|
889ba110-3a41-4aff-92d1-fce850c134d5
|
respiratory-diseases-recognition-through
| null | null |
https://ieeexplore.ieee.org/document/9080747
|
https://ieeexplore.ieee.org/document/9080747
|
Respiratory diseases recognition through respiratory sound with the help of deep neural network
|
Prediction of respiratory diseases such as COPD(Chronic obstructive pulmonary disease), URTI(upper respiratory tract infection), Bronchiectasis, Pneumonia, Bronchiolitis with the help of deep neural networks or deep learning. We have constructed a deep neural network model that takes in respiratory sound as input and classifies the condition of its respiratory system. It not only classifies among the above-mentioned disease but also classifies if a person’s respiratory system is healthy or not with higher accuracy and precision.
|
['Srinibas Rana', 'Victor Basu']
|
2020-04-30
| null | null | null | null |
['lung-disease-classification']
|
['medical']
|
[-1.58143207e-01 1.15462027e-01 -2.82943785e-01 1.81278065e-01
3.18103135e-01 -3.12449157e-01 2.37655081e-02 -1.80937216e-01
-4.07267697e-02 6.51587725e-01 3.07786375e-01 -7.65715718e-01
-5.76644778e-01 -1.19538975e+00 -6.96793795e-02 -5.30802369e-01
1.71034276e-01 1.15829909e+00 1.72458246e-01 1.80416510e-01
-2.67517895e-01 1.00769842e+00 -1.27737916e+00 3.65293890e-01
3.12979460e-01 7.68737435e-01 3.17004435e-02 1.30292416e+00
2.68046334e-02 8.99011552e-01 -7.79129744e-01 1.62836581e-01
-8.59957114e-02 -2.49150634e-01 -7.99682200e-01 -3.48446459e-01
8.08581561e-02 -7.18275130e-01 -6.05724990e-01 1.55597836e-01
7.60293901e-01 -8.13746899e-02 9.91787016e-01 -9.58204865e-01
-4.84314620e-01 -2.23708346e-01 2.19689652e-01 6.58863187e-01
-1.05894335e-01 7.27316439e-02 7.29634047e-01 -6.40558898e-01
-1.54729588e-02 1.07627559e+00 1.08762944e+00 8.96776259e-01
-4.89338845e-01 -5.05580127e-01 -5.57435870e-01 8.07262212e-03
-9.84617114e-01 1.98361382e-01 -3.41818458e-03 -7.10811138e-01
1.14886558e+00 3.93454999e-01 6.61239862e-01 9.08191323e-01
4.23701733e-01 1.34103775e-01 4.57876593e-01 1.21298634e-01
-4.51409034e-02 1.20188572e-01 7.25712627e-02 4.96389002e-01
4.42339569e-01 3.53230029e-01 5.79508603e-01 -2.34382659e-01
6.78873360e-01 9.07217085e-01 -2.37331942e-01 5.52123547e-01
-1.07483232e+00 7.59609282e-01 4.29579973e-01 9.07388926e-01
-7.09818304e-01 1.72210351e-01 2.47482121e-01 3.33613932e-01
-1.28265366e-01 2.91631639e-01 -8.19173515e-01 4.21755686e-02
-5.67786872e-01 -9.02030841e-02 7.97633290e-01 1.18623160e-01
3.60313803e-01 8.10793415e-02 -3.72473806e-01 1.14043128e+00
4.43143964e-01 6.78493142e-01 9.24880981e-01 -9.60830212e-01
9.56780761e-02 6.99249864e-01 5.86131215e-02 -6.36544585e-01
-6.20637655e-01 -1.96352005e-01 -1.37268054e+00 5.05238064e-02
1.03852795e-02 -5.30455828e-01 -1.05781746e+00 1.23931670e+00
2.20383987e-01 9.87837613e-02 1.01422541e-01 5.68083465e-01
1.30631793e+00 7.28965342e-01 3.18391807e-02 -8.28979611e-02
1.22524810e+00 -8.60627830e-01 -8.69124711e-01 2.20119759e-01
3.60964864e-01 -3.81305039e-01 5.40112793e-01 3.24940234e-01
-8.09845448e-01 -8.10792387e-01 -5.49796581e-01 3.01207006e-02
-5.56520164e-01 2.87655294e-01 1.43075526e-01 4.80992585e-01
-9.37083125e-01 8.13481629e-01 -8.08152378e-01 -5.03417909e-01
2.78042167e-01 7.22442269e-01 -2.14159414e-01 2.40474358e-01
-1.25099611e+00 9.40528452e-01 2.37903386e-01 -2.81984240e-01
-7.64031827e-01 -6.23128653e-01 -9.54615772e-02 3.51683050e-01
-6.55536652e-02 -1.27708805e+00 1.08945632e+00 -2.16281742e-01
-1.04310191e+00 7.59816885e-01 1.72633044e-02 -3.67970735e-01
2.50235558e-01 -1.60611928e-01 -7.53986120e-01 1.06023572e-01
-4.96963322e-01 3.41305345e-01 6.14325881e-01 -6.83409870e-01
-8.67360413e-01 -4.63674337e-01 -1.23568065e-01 -1.44194886e-01
-2.40216494e-01 8.54495391e-02 4.01679128e-01 -4.76885736e-01
-2.37277374e-01 -1.14259350e+00 -1.15363427e-01 3.25912535e-02
-5.72799027e-01 -9.57286179e-01 1.31707823e+00 -7.15503216e-01
1.19941556e+00 -1.94075453e+00 -1.66437566e-01 -1.60504222e-01
5.72412252e-01 1.14569640e+00 3.35679263e-01 5.85372329e-01
-1.54615462e-01 6.84383452e-01 -1.08893141e-02 3.42482239e-01
-3.85394916e-02 9.17193055e-01 1.76089704e-01 1.95051655e-01
3.36086482e-01 8.20132375e-01 -4.94975537e-01 -4.09694314e-01
5.34151018e-01 9.21590149e-01 -3.37113798e-01 6.40079141e-01
-6.56560576e-03 3.48156989e-01 -6.24088883e-01 2.71835327e-01
1.69059500e-01 -5.62281966e-01 -1.23302080e-01 5.96307814e-02
3.91313583e-02 6.19067311e-01 -7.91048169e-01 1.75315708e-01
-7.41048038e-01 6.18832886e-01 -2.71107107e-02 -9.77700949e-01
6.94722176e-01 1.17697847e+00 4.93046194e-01 -1.75170198e-01
2.81615138e-01 2.54766285e-01 3.24166745e-01 -1.37933075e+00
-4.82256472e-01 -4.34544891e-01 5.95889568e-01 7.38093376e-01
-2.82997608e-01 2.14843541e-01 -2.03058794e-01 -5.48480153e-01
1.31709576e+00 -5.79873145e-01 3.89411062e-01 1.40380323e-01
8.69768381e-01 -3.88558060e-01 1.81226373e-01 5.44872522e-01
-3.96804154e-01 8.64001989e-01 7.64588118e-02 -5.55781364e-01
-9.25637364e-01 -1.02641308e+00 -3.68318796e-01 8.62132788e-01
-4.98221338e-01 2.80596226e-01 -3.16505075e-01 -7.34160662e-01
2.66788065e-01 2.96662569e-01 -5.39771438e-01 -2.64592767e-01
-8.92070055e-01 -6.08193457e-01 6.01382792e-01 6.78702533e-01
4.18593287e-01 -1.60892045e+00 -5.34959614e-01 9.75392535e-02
-7.51060098e-02 -4.28629667e-01 1.32101104e-01 4.29446161e-01
-7.88892806e-01 -1.38644826e+00 -8.07326257e-01 -1.06310976e+00
2.48390600e-01 1.29931346e-01 1.13052785e+00 8.23898733e-01
-6.10898554e-01 2.74006248e-01 1.08860984e-01 -4.09494609e-01
-8.95954192e-01 4.92837746e-03 3.43647003e-01 -4.29390967e-01
2.80343711e-01 -6.15013957e-01 -9.64362085e-01 3.23080897e-01
-7.91874528e-01 -6.79505706e-01 5.90381026e-01 4.28891480e-01
5.02423167e-01 5.85029483e-01 1.10704792e+00 -7.03181803e-01
8.61822069e-01 -8.55352879e-01 2.68205971e-01 4.02203947e-02
-8.48262072e-01 -3.77404600e-01 9.19186890e-01 -5.08738339e-01
-3.80216897e-01 -4.26166594e-01 -8.49871337e-01 -6.08665049e-01
-8.70208502e-01 6.74369261e-02 2.70656586e-01 4.37858373e-01
5.60206652e-01 1.37200266e-01 -1.99220985e-01 -8.77506495e-01
-1.94189340e-01 1.28597820e+00 7.24860907e-01 2.79791594e-01
5.72792470e-01 1.85003713e-01 1.27570689e-01 -8.40204358e-01
-7.57381082e-01 -7.20390737e-01 -6.28906965e-01 8.97835046e-02
1.34110129e+00 -3.69359851e-01 -9.15495098e-01 3.29856217e-01
-1.01214433e+00 -6.19488768e-02 -5.41500628e-01 8.20926726e-01
-1.05487853e-01 -9.04709473e-02 -8.73142004e-01 -8.13176155e-01
-9.60311413e-01 -5.32319725e-01 6.62927032e-01 2.39310890e-01
-6.48817241e-01 -1.43222952e+00 6.72222257e-01 4.74471241e-01
6.89026475e-01 2.29108632e-01 1.30378234e+00 -1.23395729e+00
-6.81897402e-02 -4.63833511e-01 -3.03577721e-01 1.06915879e+00
5.56766510e-01 3.55318546e-01 -7.01388240e-01 -2.31164023e-02
2.28717223e-01 -2.42133304e-01 7.63702512e-01 5.20220876e-01
1.43190336e+00 -6.96411729e-01 -3.39762777e-01 3.45061243e-01
1.09751832e+00 7.77787387e-01 5.28477550e-01 -6.48770854e-02
7.71686912e-01 5.55544972e-01 3.00645437e-02 1.38906196e-01
9.89680458e-03 1.67045951e-01 6.20227575e-01 -3.02336067e-01
-4.68091607e-01 3.36366028e-01 -9.06795040e-02 9.34403539e-01
-1.64493203e-01 -8.34669888e-01 -9.16897655e-01 6.98339701e-01
-1.09028053e+00 -1.21819603e+00 -7.32296646e-01 1.91691077e+00
6.09137177e-01 -2.80142933e-01 1.16242968e-01 5.12593985e-01
8.21259379e-01 -1.53960422e-01 -5.55512309e-01 -8.11638176e-01
5.17470896e-01 8.43973815e-01 -2.74765259e-03 1.83701888e-01
-1.00451672e+00 2.79052798e-02 7.23038101e+00 1.36791661e-01
-1.36730456e+00 -3.93453427e-02 4.70428318e-01 2.35833436e-01
1.86221730e-02 -7.34989524e-01 -6.62914753e-01 4.40527380e-01
1.20460689e+00 2.76566774e-01 3.04535240e-01 5.74599206e-01
5.67522049e-02 4.49327111e-01 -8.98503482e-01 5.35592258e-01
-3.07731062e-01 -9.18159783e-01 -1.82771742e-01 3.60203683e-01
4.63100582e-01 4.13981169e-01 8.13575238e-02 5.69123864e-01
-1.37382120e-01 -1.44754696e+00 -4.24909413e-01 4.77452695e-01
7.27352977e-01 -5.49624622e-01 1.10410500e+00 5.41209936e-01
-6.91769719e-01 -2.76495367e-01 -1.39037222e-01 6.68466464e-02
-2.97727227e-01 3.09672415e-01 -1.48557961e+00 -2.26399422e-01
8.61022949e-01 3.77713829e-01 -2.02051625e-01 1.07437015e+00
-1.55251384e-01 1.05995333e+00 -4.87034470e-01 -9.95293930e-02
1.37202904e-01 1.05732426e-01 4.71258849e-01 1.01131845e+00
6.01368010e-01 2.74221301e-01 -1.59700111e-01 8.88714790e-01
-2.37522662e-01 -1.40878215e-01 -7.25580156e-01 -3.32378417e-01
3.54852021e-01 1.28108203e+00 -1.69073895e-01 -6.41056538e-01
-3.35006304e-02 4.33199823e-01 -4.51001972e-01 1.92869082e-01
-6.69948339e-01 -2.41068289e-01 5.01250327e-01 6.02117360e-01
5.82775414e-01 6.51544273e-01 1.36062084e-02 -4.46291804e-01
-5.82703054e-01 -5.34614325e-01 5.10142982e-01 -7.96137810e-01
-1.08675539e+00 6.93252981e-01 -4.76955980e-01 -7.70221233e-01
-5.91959119e-01 -8.30173254e-01 -1.12609267e+00 8.01352203e-01
-1.40298784e+00 -6.46380544e-01 -3.56700033e-01 5.80162466e-01
2.90255845e-01 -2.00116351e-01 1.08310091e+00 7.04954445e-01
-7.72798061e-01 3.26796174e-02 9.52758342e-02 4.51886170e-02
4.55743968e-01 -1.50725257e+00 6.10886291e-02 -4.47441861e-02
-7.18123734e-01 4.84309644e-01 3.47702831e-01 -5.03786325e-01
-5.76249063e-01 -1.68033814e+00 1.49064136e+00 -3.90036851e-01
4.02489275e-01 5.55799067e-01 -1.13593853e+00 3.79820406e-01
5.43072708e-02 2.07573876e-01 1.00625634e+00 -3.70378107e-01
8.59286189e-02 -2.60245390e-02 -1.38048112e+00 -1.79320239e-02
4.87943649e-01 -3.85037482e-01 -7.75861621e-01 9.54678893e-01
7.92920828e-01 7.91931227e-02 -1.35854244e+00 7.36234128e-01
5.40617287e-01 -9.92688000e-01 1.51572490e+00 -1.03355241e+00
5.96592009e-01 3.43413837e-03 1.70701481e-02 -7.27006495e-01
-7.25881517e-01 -1.27153724e-01 -5.86506546e-01 6.92174256e-01
-2.38700770e-02 -7.87770629e-01 6.59011245e-01 1.87567905e-01
-1.77506283e-01 -1.01543558e+00 -6.74394786e-01 -5.94512701e-01
2.68372744e-01 1.86563745e-01 5.85825622e-01 9.80626404e-01
-1.01340616e+00 5.29359102e-01 -2.45085984e-01 -3.21072377e-02
-1.59178093e-01 -2.59993654e-02 4.15310621e-01 -1.89146507e+00
-3.59312236e-01 -4.41914827e-01 -9.86522287e-02 -4.81134742e-01
-2.43171632e-01 -5.89866579e-01 -1.20660819e-01 -2.55472326e+00
-1.72808096e-01 -5.18149257e-01 -8.55711937e-01 5.82360089e-01
9.22258496e-02 3.76286119e-01 -3.72505009e-01 1.66072354e-01
-7.82053247e-02 -8.54574442e-02 1.41284895e+00 5.87040596e-02
-1.17457069e-01 9.98474538e-01 -4.34477776e-01 7.82029629e-01
1.12174964e+00 -7.31730282e-01 -2.93818146e-01 1.35570288e-01
2.46302724e-01 4.69586641e-01 4.35002297e-01 -1.04794931e+00
-3.32239181e-01 -3.30016106e-01 4.38071132e-01 -8.85135055e-01
2.10229322e-01 -1.15904975e+00 4.17737126e-01 1.35217118e+00
-3.95031780e-01 -8.33334178e-02 -8.95897150e-02 3.40309709e-01
-1.94990188e-02 -5.23427427e-01 1.00518060e+00 -2.76885778e-01
1.48393288e-01 5.20688415e-01 -1.13494682e+00 2.37211548e-02
6.68921292e-01 -5.10404930e-02 -3.77516419e-01 -2.25257680e-01
-1.07347965e+00 -6.61081308e-03 -1.80008858e-01 4.63529289e-01
4.87257361e-01 -1.28048873e+00 -5.41573465e-01 2.21682891e-01
-1.66240945e-01 4.33272421e-02 -9.13028792e-02 9.17644560e-01
-7.63087273e-01 8.95121694e-01 -6.58860877e-02 -3.07465255e-01
-1.46338594e+00 4.91117358e-01 9.40317333e-01 -4.18402880e-01
-5.84742188e-01 6.99163198e-01 7.61209428e-02 -6.91579282e-01
5.24100423e-01 -4.70076889e-01 -6.93163574e-01 -2.85302043e-01
3.59883040e-01 8.13434839e-01 3.24994214e-02 -6.02840781e-01
-3.23389798e-01 5.34105718e-01 3.20130318e-01 8.05308998e-01
1.12536597e+00 1.13169454e-01 -3.56884390e-01 4.18236762e-01
1.67976892e+00 -2.76122391e-01 -2.07010627e-01 1.21108769e-03
-6.33634388e-01 3.02840322e-01 4.36500400e-01 -1.10477579e+00
-1.12206185e+00 1.32746029e+00 1.03892004e+00 8.64193618e-01
9.80414510e-01 -4.76694293e-02 1.56476736e+00 8.26479912e-01
-5.60887992e-01 -7.94509768e-01 2.83080190e-01 4.02520329e-01
8.16564739e-01 -1.19721365e+00 -8.72107074e-02 2.15963826e-01
-1.45083100e-01 1.40242410e+00 4.60404187e-01 -3.18600982e-01
1.20462048e+00 -3.46170291e-02 7.84903392e-02 -4.98171002e-01
-1.09362090e+00 -1.84053019e-01 5.90602636e-01 7.09004700e-01
5.79568386e-01 1.54291064e-01 -1.43031240e-01 4.17352647e-01
-9.30707529e-02 3.06329548e-01 -4.19242680e-02 4.55138803e-01
-9.87659931e-01 -6.96667492e-01 -3.44671398e-01 9.79525208e-01
-1.03791571e+00 1.18593439e-01 -8.09597492e-01 7.48028815e-01
8.38467598e-01 8.84465516e-01 3.67701471e-01 -4.14289802e-01
1.65256187e-01 2.32678160e-01 8.61558914e-02 -8.12480271e-01
-1.14691520e+00 -1.66179225e-01 5.48816361e-02 5.01700938e-02
-2.80572653e-01 -1.47765756e-01 -1.52870131e+00 -2.18463957e-01
-2.52066225e-01 -1.18392207e-01 5.43990612e-01 1.01035404e+00
-7.59160845e-03 1.02201021e+00 7.29908228e-01 -4.12546098e-02
-5.30988634e-01 -1.17703235e+00 -4.88803655e-01 2.55600810e-01
8.87292325e-01 -4.25494164e-01 -8.02072942e-01 -1.31317049e-01]
|
[14.535948753356934, 3.8154642581939697]
|
193c9a3d-40a6-41b1-9aea-ef589f167e0c
|
a-deep-learning-framework-for-nuclear
|
2203.03420
| null |
https://arxiv.org/abs/2203.03420v2
|
https://arxiv.org/pdf/2203.03420v2.pdf
|
A Deep Learning Framework for Nuclear Segmentation and Classification in Histopathological Images
|
Nucleus segmentation and classification are the prerequisites in the workflow of digital pathology processing. However, it is very challenging due to its high-level heterogeneity and wide variations. This work proposes a deep neural network to simultaneously achieve nuclear classification and segmentation, which is designed using a unified framework with three different branches, including segmentation, HoVer mapping, and classification. The segmentation branch aims to generate the boundaries of each nucleus. The HoVer branch calculates the horizontal and vertical distances of nuclear pixels to their centres of mass. The nuclear classification branch is used to distinguish the class of pixels inside the nucleus obtained from segmentation.
|
['Jinxi Xiang', 'Xiyue Wang', 'Sen yang']
|
2022-03-04
| null | null | null | null |
['nuclear-segmentation']
|
['medical']
|
[ 1.71725988e-01 9.28340182e-02 -2.51982287e-02 -2.87818342e-01
-5.34579754e-01 -3.69451553e-01 3.10412228e-01 3.65583450e-01
-3.77696633e-01 6.04526281e-01 -2.34853327e-02 -2.34438851e-01
-3.64627838e-02 -1.07872927e+00 -1.37982070e-01 -1.21234453e+00
3.64295751e-01 5.57506263e-01 3.00234765e-01 9.22127515e-02
2.14996845e-01 1.07595527e+00 -1.24721992e+00 3.28915000e-01
8.00515831e-01 1.05426836e+00 4.02542986e-02 8.22302222e-01
-4.47890162e-01 3.71137112e-01 -6.40558481e-01 3.78123820e-02
5.74912429e-02 -6.07120037e-01 -1.03944993e+00 1.76228970e-01
2.50599924e-02 -2.50767112e-01 -2.16099560e-01 1.31664002e+00
5.94200611e-01 -1.17107064e-01 8.85077298e-01 -8.85819077e-01
-2.94181526e-01 5.97092032e-01 -3.24559689e-01 3.09507847e-01
-3.73697102e-01 3.41263190e-02 5.66132784e-01 -5.99657655e-01
8.24364424e-01 6.26242697e-01 5.61405241e-01 5.40238380e-01
-7.99352944e-01 -3.15998673e-01 -5.53854465e-01 2.17260215e-02
-1.45471597e+00 -6.25401884e-02 3.45634013e-01 -7.94061601e-01
3.96720141e-01 2.50790656e-01 9.87252295e-01 2.98024774e-01
4.45581287e-01 5.82008123e-01 8.99286091e-01 -2.16288850e-01
2.44338676e-01 -1.01873666e-01 4.55402195e-01 4.78175879e-01
2.30698198e-01 -4.36225981e-01 3.05629045e-01 2.25635916e-01
9.84435916e-01 1.12024926e-01 -2.66784728e-01 -2.09982231e-01
-1.12397015e+00 6.89948678e-01 5.25547445e-01 8.07475269e-01
-2.65899599e-01 -1.25614598e-01 5.26452422e-01 -4.51274186e-01
1.84166525e-02 2.78663523e-02 -6.41410202e-02 2.80120254e-01
-1.17007017e+00 -5.89072034e-02 3.46797496e-01 4.97901738e-01
3.96855891e-01 -1.18294351e-01 -4.61099982e-01 4.35067505e-01
3.77986968e-01 1.09574966e-01 6.86126590e-01 -7.94236958e-01
-1.30039394e-01 1.08604360e+00 -1.48283228e-01 -9.53322053e-01
-1.03325570e+00 -7.49692142e-01 -1.11341786e+00 2.54227608e-01
7.75472224e-01 -2.12628722e-01 -1.15398419e+00 9.79568362e-01
6.83678150e-01 -1.52608186e-01 -1.37725314e-02 8.95442009e-01
1.36900687e+00 4.87984568e-01 7.81797841e-02 -2.47468561e-01
1.47939491e+00 -8.23713720e-01 -9.23368633e-01 3.16795588e-01
9.73106861e-01 -5.23669124e-01 4.72422481e-01 4.69161011e-02
-9.92922068e-01 -4.22474653e-01 -1.03917456e+00 -4.63979900e-01
-6.45168245e-01 6.36312723e-01 4.56195951e-01 3.70559454e-01
-1.10755360e+00 3.57305259e-01 -8.75311077e-01 -3.14960420e-01
5.41531205e-01 4.99810457e-01 -4.49420482e-01 5.17023206e-01
-1.02509665e+00 7.12909997e-01 9.14979875e-01 5.96271932e-01
-4.60799992e-01 -4.72821712e-01 -7.60227501e-01 1.35744169e-01
-3.07305828e-02 -5.96262097e-01 1.10567546e+00 -5.90061605e-01
-1.31513941e+00 1.28486228e+00 7.31574073e-02 -3.18245053e-01
6.11898839e-01 7.62216508e-01 -1.02682725e-01 3.40297252e-01
5.19358590e-02 7.90457070e-01 2.56871581e-01 -1.03727186e+00
-9.59813356e-01 -6.55908704e-01 -4.46270049e-01 2.94354618e-01
1.46932816e-02 -2.06932813e-01 -5.97178340e-01 -3.25272739e-01
4.92389441e-01 -5.43009043e-01 -2.88119137e-01 -5.02908938e-02
-5.00681937e-01 -2.63982117e-01 8.76906574e-01 -9.65370119e-01
1.07731676e+00 -2.28548193e+00 2.08005272e-02 4.71489131e-01
5.74625373e-01 3.08837324e-01 5.41376054e-01 -2.22817659e-01
-1.19223014e-01 1.37324750e-01 -1.02863526e-02 8.57548118e-02
-1.27743065e-01 3.01474165e-02 3.13343495e-01 6.81544304e-01
-2.35513359e-01 8.24002504e-01 -7.17639029e-01 -8.80996346e-01
1.70443609e-01 4.90094364e-01 -2.05663722e-02 -8.12688395e-02
1.10725746e-01 2.28001177e-01 -2.06429690e-01 9.41426039e-01
8.02011847e-01 -2.30847448e-01 3.01403463e-01 -5.40386021e-01
-2.02847973e-01 -2.42578879e-01 -1.21251822e+00 1.00867057e+00
3.11548203e-01 4.64513987e-01 5.00367165e-01 -1.03506637e+00
1.01267600e+00 2.59775251e-01 5.40147483e-01 -3.56935143e-01
7.91305184e-01 3.82991791e-01 1.92880213e-01 -4.14330870e-01
6.24013543e-01 -2.33603850e-01 4.21432965e-02 9.60400179e-02
6.33254508e-03 -1.45307943e-01 5.34477234e-01 -1.40621318e-02
7.63527095e-01 -6.62534609e-02 5.24351656e-01 -2.14668155e-01
8.19489121e-01 5.07155061e-02 8.24361622e-01 3.19323689e-01
-6.22706950e-01 7.74431407e-01 9.27094519e-01 -5.53613782e-01
-8.84969234e-01 -9.56827044e-01 -4.08703417e-01 4.57225919e-01
2.98210084e-01 6.69991225e-02 -1.19680893e+00 -7.90973127e-01
-2.30425313e-01 2.22780973e-01 -9.65541542e-01 3.81133221e-02
-3.82315576e-01 -1.01707137e+00 4.63652313e-01 5.58868706e-01
5.67994118e-01 -1.12811792e+00 -6.25873029e-01 3.36992438e-03
-1.65269449e-01 -8.26942086e-01 -1.37652814e-01 3.25138897e-01
-7.40331352e-01 -1.04191244e+00 -7.62442291e-01 -9.25024569e-01
1.09122097e+00 -5.82458824e-03 6.52604461e-01 3.14344436e-01
-6.20037615e-01 -3.41775090e-01 -1.34115433e-02 -3.33887070e-01
-4.79201287e-01 1.50557309e-01 -3.87873262e-01 -1.20731279e-01
2.52087265e-01 -2.66292132e-02 -7.16336906e-01 2.47013703e-01
-1.00262654e+00 4.37410995e-02 7.30966449e-01 6.85504675e-01
1.13326550e+00 3.98044646e-01 -1.56947393e-02 -7.97688484e-01
3.24667573e-01 -3.88427854e-01 -6.70043886e-01 1.16542444e-01
-1.14029914e-01 -4.39556479e-01 5.15779078e-01 1.31612897e-01
-5.51421523e-01 2.28838891e-01 -2.86969662e-01 4.13994305e-02
-6.25533104e-01 3.60363424e-01 -2.19910756e-01 -9.94615853e-02
3.88410777e-01 3.53825539e-01 2.67333865e-01 -1.34788230e-01
-3.71645018e-02 7.04742730e-01 8.52177501e-01 6.01802878e-02
3.08825582e-01 7.60152876e-01 2.94146806e-01 -6.46974623e-01
-5.55102110e-01 -4.79383498e-01 -1.07778192e+00 -5.52075207e-01
1.29893434e+00 -2.29792669e-01 -7.34265804e-01 6.70094848e-01
-1.01083207e+00 -2.62210727e-01 -4.19837683e-01 4.78739679e-01
-2.71007359e-01 4.08395052e-01 -8.12295020e-01 -1.69803366e-01
-7.67327964e-01 -1.38514566e+00 8.44854593e-01 9.26914632e-01
-4.18742821e-02 -9.18072820e-01 -5.46202660e-02 4.32518005e-01
-4.49355319e-02 3.71777654e-01 1.11573207e+00 -7.48197496e-01
-3.61361206e-01 -5.19022644e-01 -4.05920893e-01 2.04080626e-01
-3.73023599e-02 6.86588109e-01 -5.51821768e-01 7.16959611e-02
-2.81138510e-01 1.27855077e-01 5.95159292e-01 8.37123334e-01
1.38058150e+00 3.60502690e-01 -6.97254002e-01 9.03148055e-01
1.29152668e+00 5.70186317e-01 8.77569735e-01 5.06478250e-01
4.52218741e-01 5.19787252e-01 5.19220591e-01 3.62919867e-02
8.64472240e-02 2.94878662e-01 4.19447809e-01 -6.26606762e-01
-1.45709217e-01 3.92082155e-01 -2.67031878e-01 5.43829441e-01
-9.13086068e-03 -6.33418709e-02 -1.25583577e+00 5.03609240e-01
-1.49645543e+00 -9.25347686e-01 -5.86695671e-01 1.72225273e+00
7.05739141e-01 1.33111537e-01 1.03629716e-02 3.45769733e-01
9.52813804e-01 -3.72342974e-01 -4.24136251e-01 -4.18447465e-01
7.70885646e-02 -6.57620803e-02 3.55881989e-01 3.07306647e-01
-1.21286869e+00 6.47084475e-01 7.09282398e+00 1.02081490e+00
-1.26574218e+00 -1.74521834e-01 1.00368166e+00 2.59529144e-01
2.01792821e-01 -4.02049333e-01 -1.19294024e+00 5.18789768e-01
4.17164296e-01 2.32805476e-01 -1.58447325e-01 4.79519457e-01
2.06493407e-01 -4.41463530e-01 -7.15371907e-01 6.51889622e-01
-5.03102839e-02 -1.47984672e+00 -8.61579087e-03 3.22898746e-01
5.13897121e-01 -2.47600064e-01 -2.20568195e-01 1.77662581e-01
-8.39324221e-02 -1.01716983e+00 4.24295008e-01 7.41442680e-01
6.10509932e-01 -7.80541241e-01 1.31535792e+00 3.38697672e-01
-9.76873636e-01 1.87839881e-01 -2.83504874e-01 3.67160320e-01
-6.60814494e-02 8.58601570e-01 -1.03313553e+00 5.55560112e-01
4.48540658e-01 1.50634170e-01 -5.02365410e-01 1.34405708e+00
-1.06946066e-01 2.62154639e-01 -1.43955931e-01 4.71390672e-02
2.64285892e-01 -6.47605300e-01 2.25852340e-01 1.19837391e+00
4.16542828e-01 -7.56719932e-02 7.06379935e-02 8.69807661e-01
-9.27115530e-02 4.29575980e-01 -1.16068169e-01 -8.78851041e-02
3.04502517e-01 1.89103723e+00 -1.68464184e+00 -4.94699717e-01
-8.64411965e-02 4.99314904e-01 1.54036433e-01 8.60089809e-02
-7.89612055e-01 -8.09844196e-01 1.87478960e-01 1.44965529e-01
-5.90118021e-03 -1.93436705e-02 -6.77410424e-01 -5.21182954e-01
-3.53837252e-01 -4.13993865e-01 5.56603253e-01 -6.50942147e-01
-7.77351558e-01 3.68572116e-01 -2.83218622e-01 -9.65147913e-01
1.63566083e-01 -6.89742088e-01 -7.30483711e-01 8.34181786e-01
-9.88985538e-01 -1.12925887e+00 -5.90129316e-01 1.94525525e-01
1.22359373e-01 1.81094743e-03 6.18051291e-01 2.82419980e-01
-9.79328215e-01 3.67854625e-01 3.43985260e-01 5.53424239e-01
2.59466052e-01 -1.48316729e+00 -3.94555897e-01 7.10113645e-01
-6.11441553e-01 3.45727831e-01 2.67380923e-01 -5.64778447e-01
-7.15504587e-01 -1.04801524e+00 9.57974613e-01 2.37091631e-01
4.46710140e-01 -1.59124751e-02 -6.89848661e-01 3.82158697e-01
-1.07254259e-01 2.51552224e-01 8.63652289e-01 -3.35968226e-01
5.11340976e-01 -1.25756338e-01 -1.36998034e+00 7.19047308e-01
2.57218391e-01 -9.65755954e-02 -2.48653218e-01 3.90138477e-01
1.82738736e-01 -7.18619466e-01 -1.02711761e+00 4.21978474e-01
4.41452950e-01 -9.37893629e-01 7.74453700e-01 4.30367552e-02
2.56257832e-01 -7.90354133e-01 3.59302521e-01 -9.74007070e-01
-5.12229681e-01 8.00143555e-02 3.18129718e-01 1.18939281e+00
3.70110273e-01 -3.68764967e-01 1.03479171e+00 3.19866806e-01
-4.37028021e-01 -9.35385168e-01 -8.93664002e-01 -1.99244991e-01
1.02472536e-01 1.61046222e-01 5.86664438e-01 7.63252616e-01
1.72413811e-02 1.15857981e-01 3.38325977e-01 1.08392261e-01
4.20146197e-01 1.46965146e-01 4.12323266e-01 -1.18438566e+00
2.71614224e-01 -9.30532873e-01 -8.08037937e-01 -4.89937246e-01
1.69062316e-02 -1.08849072e+00 -1.20496694e-02 -2.05771160e+00
2.56344527e-01 -5.97565435e-03 -2.17932895e-01 4.19911683e-01
4.36453940e-03 5.22310674e-01 -2.15262715e-02 1.74427271e-01
-3.93330812e-01 1.34780854e-01 1.60383201e+00 -1.85883507e-01
-2.57140934e-01 4.60721403e-02 -5.29677749e-01 1.00850129e+00
1.10236442e+00 -1.13892123e-01 -1.51335057e-02 8.84942263e-02
-5.80569170e-02 -3.39940749e-02 1.04856364e-01 -1.23528039e+00
3.47451895e-01 -1.25734121e-01 8.66954565e-01 -1.31998169e+00
-6.98813871e-02 -6.00557148e-01 1.61178902e-01 8.10032845e-01
-2.83221334e-01 -4.60228980e-01 -1.34922758e-01 6.52878508e-02
-1.86737463e-01 -4.90546167e-01 1.23774385e+00 -2.95911223e-01
-3.97221625e-01 1.76380306e-01 -7.78159499e-01 -3.40206772e-01
1.49040663e+00 -6.59671962e-01 -3.37403089e-01 9.83729288e-02
-1.19428217e+00 3.47148508e-01 2.44989023e-01 -4.67410982e-01
5.20985425e-01 -1.21838808e+00 -3.86281192e-01 2.12653428e-01
-1.54816344e-01 4.51915056e-01 4.69609588e-01 1.36508679e+00
-1.13667071e+00 4.81437325e-01 -3.47348183e-01 -6.72096133e-01
-1.50445831e+00 2.58601665e-01 9.51963365e-01 -4.19126272e-01
-6.16867900e-01 8.42761695e-01 2.01296002e-01 -5.17836094e-01
3.18508655e-01 -4.73841131e-01 -9.33313727e-01 2.67104924e-01
6.28228247e-01 6.41849101e-01 2.92750567e-01 -9.12595928e-01
-1.87179506e-01 5.21105766e-01 -8.84584337e-02 3.28636348e-01
7.33842671e-01 1.23285040e-01 -6.57476723e-01 2.31389463e-01
9.46907818e-01 -2.43959025e-01 -7.65638769e-01 1.74668983e-01
-2.30101407e-01 -4.28033434e-02 3.79763722e-01 -7.89954007e-01
-1.20064008e+00 7.92184949e-01 5.95761657e-01 2.51861334e-01
1.06426549e+00 -1.35825351e-01 8.42131317e-01 -5.43787442e-02
-1.14646509e-01 -1.44308305e+00 -4.61928636e-01 5.83198488e-01
3.78762335e-01 -7.06782579e-01 -7.40015060e-02 -5.46206713e-01
-2.57762998e-01 1.45405209e+00 7.53901243e-01 2.63777345e-01
5.10930538e-01 4.48426366e-01 1.49464056e-01 -3.86127412e-01
-1.54504776e-01 -2.13183105e-01 1.37569994e-01 6.30185306e-01
4.70885783e-01 8.56291652e-02 -8.72297943e-01 6.67194605e-01
-8.19590688e-02 2.33644605e-01 5.48938334e-01 7.20962346e-01
-6.73877299e-01 -6.27215564e-01 -6.76969409e-01 6.11860573e-01
-4.82794404e-01 3.16769391e-01 -5.67697227e-01 7.86072135e-01
6.90755486e-01 3.67770135e-01 3.19643378e-01 -1.31963670e-01
-5.82213374e-03 -8.37754086e-02 1.65176511e-01 -4.67766255e-01
-6.36680007e-01 3.81548464e-01 -2.82881677e-01 -1.45527482e-01
-1.52348578e-01 -5.08916199e-01 -2.22905302e+00 -1.26482129e-01
-1.77445084e-01 2.91027069e-01 7.24391639e-01 9.41999793e-01
-3.27034667e-02 7.13755190e-01 3.80282760e-01 -6.94400966e-01
-2.44801760e-01 -8.34686279e-01 -1.12086689e+00 1.83587447e-01
1.57239977e-02 -4.18888897e-01 -1.76434621e-01 1.20352447e-01]
|
[14.905811309814453, -3.072237968444824]
|
3781e629-1692-42e1-b684-6475cd795fbc
|
torchxrayvision-a-library-of-chest-x-ray
|
2111.00595
| null |
https://arxiv.org/abs/2111.00595v1
|
https://arxiv.org/pdf/2111.00595v1.pdf
|
TorchXRayVision: A library of chest X-ray datasets and models
|
TorchXRayVision is an open source software library for working with chest X-ray datasets and deep learning models. It provides a common interface and common pre-processing chain for a wide set of publicly available chest X-ray datasets. In addition, a number of classification and representation learning models with different architectures, trained on different data combinations, are available through the library to serve as baselines or feature extractors.
|
['Hadrien Bertrand', 'Mohammad Hashir', 'Rupert Brooks', 'Akshay Chaudhari', 'Matthew P Lungren', 'Matteo Guarrera', 'Parsa Torabian', 'Paul Morrison', 'Paul Bertin', 'Joseph D. Viviano', 'Joseph Paul Cohen']
|
2021-10-31
| null | null | null | null |
['medical-image-retrieval', 'small-data', 'medical-x-ray-image-segmentation', 'medical-image-retrieval']
|
['computer-vision', 'computer-vision', 'medical', 'medical']
|
[-1.60372928e-02 -3.09246659e-01 -5.86045563e-01 -9.58343685e-01
-1.37165570e+00 -7.31393322e-02 3.04925531e-01 5.86677939e-02
-2.34045565e-01 2.17060104e-01 2.91105092e-01 -3.51981938e-01
-2.13762879e-01 -6.16057992e-01 -2.14286089e-01 -5.42479813e-01
1.42360941e-01 6.36421800e-01 7.74315596e-02 -1.36799619e-01
-2.53025800e-01 6.59955561e-01 -1.03060234e+00 1.00170004e+00
-2.75182992e-01 1.06583762e+00 2.13155657e-01 1.19260180e+00
2.32271329e-01 1.10574043e+00 -3.27882648e-01 -3.59576242e-03
1.38826221e-01 -8.23514760e-02 -9.49346125e-01 -2.75867254e-01
5.73503375e-01 -7.17547596e-01 -1.03304827e+00 3.45425218e-01
1.02786243e+00 -1.53044406e-02 5.81822217e-01 -4.47848737e-01
-5.12467921e-01 3.53768855e-01 -6.43972397e-01 1.07030916e+00
5.21229208e-01 3.82821381e-01 8.25040221e-01 -7.94323921e-01
5.77053845e-01 6.90939903e-01 8.58371794e-01 6.07767045e-01
-8.09125960e-01 -4.62160677e-01 -7.13859618e-01 2.25632757e-01
-1.07958066e+00 -1.46696895e-01 5.65779746e-01 -4.00135666e-01
1.28553140e+00 7.24150836e-01 3.73369098e-01 1.60540688e+00
7.12145269e-01 7.76575863e-01 1.13408542e+00 -3.40632588e-01
-3.20383251e-01 -1.20120756e-01 8.59147727e-01 8.87929976e-01
-2.75305420e-01 3.32860559e-01 -3.39154094e-01 -5.21905839e-01
8.00417066e-01 5.57054818e-01 -2.33784080e-01 3.32132392e-02
-1.11927652e+00 9.21402216e-01 8.23737025e-01 3.81961375e-01
-7.51979828e-01 1.80050835e-01 6.50762796e-01 2.72879034e-01
2.74673671e-01 3.32297891e-01 -5.85787594e-01 3.92903723e-02
-6.56560779e-01 5.24044394e-01 2.90569037e-01 6.26454890e-01
1.07144170e-01 -3.33939761e-01 -3.74000907e-01 1.10492349e+00
2.69924045e-01 1.52060658e-01 8.96951020e-01 -4.24084336e-01
4.51786399e-01 4.34996456e-01 -8.06764841e-01 -4.71459150e-01
-1.11028433e+00 -3.39926004e-01 -7.06801891e-01 2.37786472e-01
1.25412330e-01 3.23752239e-02 -1.32024646e+00 9.21533108e-01
1.56417906e-01 -7.52837732e-02 -2.44003803e-01 1.09909046e+00
1.63561821e+00 2.39732161e-01 2.00373203e-01 3.08349967e-01
1.82968175e+00 -8.91695440e-01 -6.53424084e-01 -2.70968795e-01
6.14177346e-01 -8.89327645e-01 1.32518113e+00 5.38508773e-01
-1.19409585e+00 -6.77718401e-01 -1.03769207e+00 -5.11784375e-01
-4.08590019e-01 3.08069468e-01 8.47369373e-01 4.69750881e-01
-7.62740910e-01 6.62422895e-01 -1.41018450e+00 -2.59627223e-01
6.79202139e-01 3.62439275e-01 -4.12724882e-01 -2.21156433e-01
-7.58447051e-01 1.12737787e+00 3.42775844e-02 -4.20320213e-01
-6.84652925e-01 -1.00095022e+00 -6.06160045e-01 -1.43821627e-01
2.09175155e-01 -8.67601812e-01 1.76305437e+00 -2.61198163e-01
-1.27728128e+00 1.20368969e+00 5.10656834e-01 -1.93959773e-01
3.00406575e-01 -5.31030059e-01 -4.75189656e-01 6.19925141e-01
-1.30191863e-01 3.55446517e-01 7.04391718e-01 -5.54220557e-01
-2.04613283e-01 -5.63547075e-01 -2.14706749e-01 9.74929258e-02
1.56953663e-01 6.00287259e-01 -5.72394431e-01 -9.69354153e-01
1.03792958e-01 -8.34134400e-01 -3.27712595e-01 1.31657362e-01
-4.84456539e-01 -1.97507232e-01 7.68776000e-01 -7.09368408e-01
9.38617349e-01 -2.17781734e+00 -1.50834605e-01 1.07790284e-01
4.89657015e-01 1.33558214e-01 1.96477219e-01 1.87472329e-01
-7.55803764e-01 -2.06846282e-01 -7.02622086e-02 -2.23456532e-01
-3.78351837e-01 3.19524944e-01 -8.40296969e-02 6.35073006e-01
-2.13702582e-02 9.03038383e-01 -5.49498677e-01 -5.95786989e-01
7.66568005e-01 6.83681786e-01 -2.77387947e-01 4.30684388e-01
4.42894455e-03 4.63023603e-01 -4.11886781e-01 7.74507046e-01
3.32346737e-01 -5.45589030e-01 -1.51347056e-01 -2.74556309e-01
3.36281210e-01 4.30149436e-01 -6.84623539e-01 2.14630604e+00
-6.08913600e-01 4.72544372e-01 -1.50559396e-02 -5.55783093e-01
3.91920298e-01 4.83050555e-01 7.19284892e-01 -3.31518382e-01
6.91587746e-01 -1.82974949e-01 -1.99506313e-01 -9.57844734e-01
1.98392197e-01 -2.80791581e-01 1.76188484e-01 1.06424463e+00
4.20655638e-01 -1.59000352e-01 -1.46321312e-01 4.09726165e-02
1.58034635e+00 -2.42285669e-01 5.28540313e-01 1.54370703e-02
3.34099919e-01 -8.55421424e-02 -1.85979843e-01 8.09739590e-01
4.59979698e-02 1.32802534e+00 -1.48718625e-01 -1.03674066e+00
-6.83928013e-01 -1.37115848e+00 -9.22058463e-01 1.19364893e+00
-4.57921684e-01 -6.36174738e-01 -2.40527526e-01 -7.75454342e-01
-1.35706797e-01 3.90700579e-01 -8.61693442e-01 -5.08420616e-02
-5.52427530e-01 -8.94591928e-01 6.74108028e-01 1.02813876e+00
9.59232450e-02 -1.30352390e+00 -8.58846903e-01 -1.60219967e-01
6.69302121e-02 -5.97585976e-01 -4.29868214e-02 5.94497025e-01
-7.56449878e-01 -1.66953719e+00 -4.43648398e-01 -3.41042131e-01
2.94050604e-01 1.76912934e-01 1.40803027e+00 6.11472726e-01
-1.15428269e+00 7.48733878e-01 -2.42555216e-01 -4.49219465e-01
-2.84040064e-01 2.71893322e-01 -3.25891614e-01 -6.98559940e-01
3.41043860e-01 5.72272614e-02 -8.00520599e-01 -6.78424537e-02
-9.89449978e-01 -2.54652888e-01 3.98944587e-01 8.44252586e-01
1.02187574e+00 -4.09693509e-01 1.72642216e-01 -1.12925243e+00
8.19599926e-01 -9.51494157e-01 -1.44427121e-01 9.44551677e-02
-4.37496006e-01 -3.57237458e-01 3.90854627e-01 6.53209239e-02
-9.79979932e-01 -7.50563219e-02 -1.03763020e+00 -4.05274451e-01
-5.93155503e-01 3.86395752e-01 5.30011535e-01 -1.63622070e-02
1.06298733e+00 -9.50288102e-02 -2.56871879e-01 -9.62130725e-01
4.57173198e-01 7.34791875e-01 8.10933232e-01 -1.09359749e-01
1.19637355e-01 5.22898316e-01 -8.89972523e-02 -6.68090463e-01
-1.04324841e+00 -5.89360416e-01 -7.09615409e-01 2.18314063e-02
1.26769364e+00 -6.19029582e-01 -1.25447124e-01 3.08096111e-01
-5.68649709e-01 -2.02364530e-02 -3.55896235e-01 9.05952632e-01
-5.93541503e-01 1.46137521e-01 -9.88456011e-01 1.33579344e-01
-7.89711535e-01 -1.34091973e+00 1.18034887e+00 2.92802881e-02
-5.59842944e-01 -9.44752693e-01 6.47089601e-01 4.06784862e-01
2.98985362e-01 4.50936347e-01 9.90356863e-01 -9.26177800e-01
-1.50920972e-01 -7.12643743e-01 -1.52767777e-01 4.68591124e-01
2.82939434e-01 1.17847152e-01 -1.01687837e+00 -4.71042514e-01
4.45749491e-01 -7.40629375e-01 1.08127689e+00 7.32089996e-01
1.95247602e+00 2.01808646e-01 -5.78875363e-01 1.18447936e+00
1.27839661e+00 3.85118164e-02 6.29411280e-01 2.37617314e-01
5.32745421e-01 -2.50734016e-02 2.08407462e-01 4.12340611e-01
2.35049710e-01 3.80434752e-01 2.56060600e-01 -5.38697422e-01
-2.32425645e-01 3.27929705e-01 -5.58556497e-01 5.25001466e-01
-7.02424720e-02 2.48047784e-01 -1.39341474e+00 -1.31729126e-01
-1.13002026e+00 -8.74874115e-01 -4.39077914e-01 1.47324586e+00
7.52147734e-01 -1.48818661e-02 5.89034446e-02 2.49758944e-01
1.60072315e-02 2.99892217e-01 -2.49081194e-01 -2.06965685e-01
2.84195244e-01 1.24525666e+00 1.39590263e-01 9.74669978e-02
-1.40583849e+00 1.34698629e-01 8.72868633e+00 3.23607415e-01
-1.59547317e+00 4.68021572e-01 4.21455175e-01 -4.88137454e-01
3.64344984e-01 -4.84643489e-01 -3.33827943e-01 2.84389108e-01
7.78566301e-01 2.10765660e-01 -1.00249268e-01 1.36528265e+00
-3.38461250e-01 -1.08964056e-01 -1.19862390e+00 1.12830091e+00
1.43956050e-01 -1.91491187e+00 -4.93200094e-01 -2.26280674e-01
1.83401376e-01 1.19435692e+00 5.37180342e-02 2.66914666e-01
1.27768368e-01 -9.47731376e-01 5.96809387e-02 6.39558017e-01
1.07638419e+00 -5.26530445e-01 9.20646250e-01 -1.44270763e-01
-7.42154181e-01 4.96618077e-02 -1.69337884e-01 1.95072308e-01
-3.27103734e-02 5.40894307e-02 -8.76720190e-01 5.64345717e-01
1.13877940e+00 5.51378489e-01 -1.06270373e+00 8.98779035e-01
-1.40426680e-01 6.23413861e-01 -3.01903099e-01 7.27079749e-01
-5.19634522e-02 4.36868310e-01 1.28267586e-01 1.40793967e+00
-1.81612477e-01 5.70612252e-01 2.40500808e-01 3.98929387e-01
-8.64881948e-02 2.60543879e-02 -5.28396547e-01 4.12383586e-01
-5.62256314e-02 1.58312547e+00 -6.29937112e-01 -5.38127959e-01
-9.65727806e-01 4.20762360e-01 3.45589663e-03 -7.18238875e-02
-1.02011335e+00 -3.71272504e-01 3.65512490e-01 5.50331056e-01
-8.00185949e-02 -1.29895033e-02 -4.90463138e-01 -9.60200131e-01
-4.20981646e-01 -9.82588410e-01 1.11898732e+00 -1.09665632e+00
-1.50935912e+00 9.03930306e-01 2.67463654e-01 -1.06483757e+00
-3.20816249e-01 -8.98648798e-01 -9.80405509e-01 6.45463824e-01
-1.40004277e+00 -1.15561771e+00 -7.65043557e-01 1.06014478e+00
5.62374830e-01 -4.29757237e-01 1.17711020e+00 4.44972336e-01
-6.51801527e-01 4.39467043e-01 -2.19543904e-01 1.73963785e-01
5.78829765e-01 -1.40449846e+00 4.56385583e-01 2.08076105e-01
1.53879225e-01 6.22625411e-01 2.12819859e-01 -3.05280209e-01
-1.46442676e+00 -1.09143674e+00 1.68076053e-01 -9.54420745e-01
3.86601388e-01 3.28548104e-02 -1.15501511e+00 9.79133964e-01
3.49753290e-01 3.12466055e-01 1.31984460e+00 1.48682073e-01
-1.56287625e-01 1.37250006e-01 -1.12453532e+00 -1.42141566e-01
4.75749701e-01 -6.57059848e-01 -1.04266691e+00 9.31184709e-01
1.74984977e-01 -1.09217215e+00 -1.25978518e+00 6.45931482e-01
4.03865188e-01 -1.02920246e+00 1.28847134e+00 -5.81021667e-01
6.15832865e-01 2.20395267e-01 -9.15810242e-02 -9.23236430e-01
-3.50026339e-01 -1.04906015e-01 -1.96428165e-01 2.72890896e-01
2.06620216e-01 -3.61394227e-01 7.10906386e-01 3.50953668e-01
-4.14217263e-01 -1.30862212e+00 -8.97393525e-01 -2.81398855e-02
-1.48305958e-02 -9.03742254e-01 5.86113036e-01 7.78386056e-01
-1.98793218e-01 3.17131698e-01 1.27483353e-01 -9.45657715e-02
2.22827911e-01 1.19013697e-01 8.13247144e-01 -8.92083645e-01
-6.38680041e-01 -3.05937916e-01 -2.53063291e-01 -5.62516510e-01
-3.77287835e-01 -1.31985104e+00 -5.69931805e-01 -1.60575545e+00
1.79327905e-01 -4.91199583e-01 -6.37779951e-01 7.59127617e-01
-9.30734202e-02 5.57617962e-01 -7.33184814e-03 3.90329629e-01
-1.23534888e-01 -3.46604362e-02 1.03667521e+00 3.82841453e-02
1.39978640e-02 4.29410815e-01 -4.88241017e-01 1.11520576e+00
8.65018070e-01 -7.68686533e-01 -3.36548060e-01 -7.42568254e-01
-8.89337584e-02 1.96212724e-01 6.13117993e-01 -1.24213970e+00
-2.90652275e-01 1.07259266e-01 9.62643445e-01 -1.34110165e+00
5.85835814e-01 -6.75446510e-01 2.57815141e-02 4.20472085e-01
-4.95540619e-01 4.07900125e-01 1.53936177e-01 3.09943943e-03
-8.46762676e-03 -2.25715041e-01 1.09267581e+00 -5.87228000e-01
-3.55191767e-01 5.18439472e-01 -2.44129181e-01 -6.94496930e-02
1.00907886e+00 -1.49921216e-02 -3.78132582e-01 2.09496543e-01
-1.00773633e+00 -2.33813282e-02 1.55068889e-01 6.76222444e-01
9.13462460e-01 -1.33483458e+00 -6.54851079e-01 7.81581819e-01
2.32038721e-01 -5.11170104e-02 3.35779101e-01 7.70365536e-01
-8.56947184e-01 2.95559376e-01 -4.54841912e-01 -8.90379250e-01
-1.34733391e+00 4.91350710e-01 5.36673009e-01 -4.47388142e-01
-1.36509037e+00 9.24528599e-01 -1.12981267e-01 -3.28019410e-01
8.37928206e-02 -7.14504421e-01 -1.14492372e-01 -2.64398575e-01
9.06280875e-01 2.43423611e-01 6.40961587e-01 -3.83631945e-01
-5.87404132e-01 2.96761632e-01 -4.63550389e-01 3.24809045e-01
1.71137965e+00 2.96963483e-01 1.77714154e-01 3.32747221e-01
1.35979331e+00 -4.19380397e-01 -4.60594922e-01 -1.86687112e-01
-7.02249855e-02 -4.27039891e-01 3.97413999e-01 -8.44485819e-01
-1.33741140e+00 9.65845525e-01 1.09041262e+00 -3.63680013e-02
1.26516569e+00 7.73555696e-01 8.12020123e-01 4.99259293e-01
-1.72097057e-01 -8.06996882e-01 2.50084817e-01 1.12585589e-01
1.14054513e+00 -1.21615160e+00 6.33796811e-01 -7.89376721e-02
-6.22072637e-01 1.26336944e+00 5.98556161e-01 -5.06227970e-01
1.13791120e+00 5.13996303e-01 5.59574246e-01 -1.05900848e+00
-1.02707803e+00 1.01426393e-01 3.25897455e-01 7.38808632e-01
8.76415491e-01 1.38007969e-01 -6.13556914e-02 9.02881444e-01
-3.40936214e-01 1.16726346e-01 4.13028508e-01 1.06775987e+00
-3.15258279e-02 -1.07205153e+00 -4.57387298e-01 9.86214995e-01
-1.13581848e+00 8.39361325e-02 -9.03386772e-02 1.08315265e+00
5.40830493e-02 3.37944299e-01 9.88192484e-03 -3.38508427e-01
5.16100347e-01 -1.07259020e-01 6.65783525e-01 -1.07578170e+00
-1.13616073e+00 1.01040408e-01 -6.78607672e-02 -8.64798903e-01
-3.66185792e-02 -3.97891998e-01 -1.37465680e+00 -3.52988467e-02
-4.07441199e-01 -5.49198568e-01 6.78672016e-01 6.11428201e-01
-4.23117448e-03 9.85366464e-01 5.09568274e-01 -7.21364915e-01
-6.44028783e-01 -1.14871442e+00 -3.57211679e-01 5.22158444e-01
4.71252739e-01 -6.81364298e-01 6.71437755e-02 -5.77652492e-02]
|
[15.225720405578613, -1.949194073677063]
|
c6fa8d47-44a9-4e14-b5b3-a7e542cd8c68
|
segmented-convolutional-gated-recurrent
| null | null |
https://doi.org/10.1016/j.neucom.2018.11.109
|
https://sci-hub.se/10.1016/j.neucom.2018.11.109
|
Segmented convolutional gated recurrent neural networks for human activity recognition in ultra-wideband radar
|
The automatic detection and recognition of human activities are valuable for physical security, gaming, and intelligent interface. Compared to an optical recognition system, radar is more robust to variations in lighting conditions and occlusions. The centimeter-wave ultra-wideband radar can even track human motion when the target is fully occluded from it. In this work, we propose a neural network architecture, namely segmented convolutional gated recurrent neural network (SCGRNN), to recognize human activities based on micro-Doppler spectrograms measured by the ultra-wideband radar. Unlike most existing approaches which treat the micro-Doppler spectrograms the same way as natural images, we extract segmented features of spectrograms via convolution operation and encode the feature maps along the time axis with gated recurrent units. Taking advantage of regularities in both the time and Doppler frequency domains in this way, our model can detect activities with arbitrary lengths. The experiments show that our method outperforms existing models in fine temporal resolution, noise robustness, and generalization performance. The radar system can thus recognize human behavior when visible light is blocked by opaque objects.
Keywords:
Micro-Doppler spectrograms, Human activity recognition, Deep learning, Convolutional neural network, Recurrent neural network
|
['Yongpeng Dai', 'Yongping Song', 'Hao Du', 'Tian Jin', 'Yuan He']
|
2019-04-27
| null | null | null |
neurocomputing-2019-4
|
['rf-based-pose-estimation']
|
['computer-vision']
|
[ 4.55166996e-01 -6.35814965e-01 9.08799544e-02 -1.28787398e-01
-2.03800395e-01 -3.87943774e-01 6.97491229e-01 -6.96254373e-01
-4.92731422e-01 6.98223114e-01 1.10155165e-01 -2.59963423e-01
-2.23304421e-01 -8.92355740e-01 -1.48168296e-01 -8.99915338e-01
-5.15177369e-01 -2.28662133e-01 5.62267080e-02 -5.57226464e-02
7.63900355e-02 9.36232865e-01 -1.60328817e+00 -9.44120586e-02
4.60442454e-01 1.19187427e+00 -2.38357022e-01 1.15509665e+00
3.29528302e-01 1.09795463e+00 -9.39789474e-01 4.74427938e-01
3.06143403e-01 -5.00825346e-01 -7.38980770e-02 -1.54671252e-01
3.36157024e-01 -5.42910457e-01 -1.16208827e+00 7.43595302e-01
4.58382636e-01 5.26166320e-01 6.40392840e-01 -8.55685472e-01
-7.61204422e-01 -5.77156954e-02 -4.79960442e-01 7.62950659e-01
7.04306960e-01 1.41194999e-01 3.39888871e-01 -3.97575080e-01
1.45956531e-01 9.36801434e-01 6.17618024e-01 3.44558805e-01
-5.42697191e-01 -6.14457667e-01 -8.24084878e-02 4.16153044e-01
-1.22115719e+00 -1.76364228e-01 6.70102656e-01 -3.65303963e-01
1.18166721e+00 2.62044489e-01 7.22984850e-01 1.35918903e+00
5.97655416e-01 4.46882397e-01 6.87367380e-01 -3.31203789e-01
-7.26550072e-02 -5.42280614e-01 3.27166468e-01 7.82754004e-01
2.85628557e-01 8.42530608e-01 -5.37370205e-01 1.31241456e-01
9.75650728e-01 3.45435470e-01 -4.89267021e-01 6.67258538e-03
-1.37517202e+00 4.53722090e-01 3.29584777e-01 6.46755815e-01
-5.26267350e-01 4.96221423e-01 2.20378995e-01 3.67623001e-01
3.44992399e-01 4.32318777e-01 1.16409898e-01 -4.95886236e-01
-8.46703768e-01 -2.24189740e-02 6.75999224e-01 3.25702786e-01
4.83408570e-01 8.40793908e-01 -1.07248202e-01 4.60193604e-01
1.08466316e-02 1.06315100e+00 7.17783511e-01 -6.02249086e-01
5.46343550e-02 7.25357905e-02 2.29432791e-01 -1.10346484e+00
-9.11359072e-01 -7.41243064e-01 -9.61362958e-01 1.83870286e-01
2.87369937e-01 -6.02576613e-01 -1.12102079e+00 1.34279644e+00
7.48325586e-02 5.90066075e-01 1.77028164e-01 1.18564367e+00
6.43732011e-01 7.53485799e-01 -8.42323601e-02 -4.09381777e-01
1.29424894e+00 -5.65607488e-01 -9.91529644e-01 -3.41535032e-01
3.23672563e-01 -4.34770733e-01 3.13148856e-01 3.16888362e-01
-3.17116052e-01 -7.81512022e-01 -1.42277610e+00 3.64837319e-01
-4.99690920e-01 1.75348192e-01 8.59831691e-01 1.10477936e+00
-4.92062390e-01 3.42295021e-01 -1.11069441e+00 -1.65258899e-01
-1.49960322e-02 4.38509881e-02 -1.03082955e-01 2.05848306e-01
-1.45601428e+00 7.54906654e-01 6.41030222e-02 5.80366790e-01
-6.62118673e-01 -2.78020233e-01 -1.11392725e+00 -1.80861186e-02
1.48606867e-01 -1.59670666e-01 1.12278187e+00 -7.38524497e-01
-1.46101665e+00 3.15742552e-01 1.28691713e-03 -9.29054976e-01
1.93369940e-01 -4.19536471e-01 -1.22510636e+00 1.78786054e-01
-1.98393643e-01 -1.29844502e-01 9.30812180e-01 -2.99728364e-01
-5.76569498e-01 -3.88257712e-01 -2.26111244e-02 -1.00676730e-01
4.95561622e-02 -4.72599342e-02 1.67021692e-01 -7.60170221e-01
1.97173461e-01 -8.05930138e-01 -1.35051936e-01 -4.43356097e-01
-8.93947929e-02 1.00426957e-01 1.21350169e+00 -4.47100550e-01
1.11418569e+00 -2.14494824e+00 -4.30158675e-01 2.29572237e-01
2.97853611e-02 5.08815646e-01 1.53243631e-01 1.70882404e-01
9.79024544e-02 -7.36921966e-01 -6.26740530e-02 4.33581650e-01
-9.35463011e-02 7.08433464e-02 -4.17391390e-01 8.42649281e-01
-7.85831735e-02 8.90956700e-01 -7.85924554e-01 1.73950478e-01
4.97451872e-01 7.01879561e-01 2.50363767e-01 5.91561273e-02
1.99828580e-01 3.54509532e-01 -5.54677606e-01 7.51032412e-01
4.74553883e-01 -1.08997911e-01 -1.19918354e-01 -1.35019645e-01
-3.81121457e-01 6.04463555e-02 -1.02627826e+00 1.28931761e+00
-5.17411947e-01 1.32721591e+00 -3.24815631e-01 -1.20915663e+00
1.27603936e+00 4.44289654e-01 6.89339936e-01 -1.36464107e+00
2.03805387e-01 -1.44669652e-01 -6.10410087e-02 -6.52236104e-01
5.61843395e-01 -1.29982643e-02 -3.64940494e-01 8.14341486e-01
-4.45481390e-02 7.01310039e-01 1.36989251e-01 -1.81023866e-01
1.29694784e+00 6.97418004e-02 3.10362816e-01 3.74390423e-01
5.18505037e-01 -3.79705429e-01 5.65371633e-01 1.00880814e+00
-3.92807007e-01 2.04004452e-01 -2.82927006e-02 -8.52789998e-01
-4.79788601e-01 -1.19977188e+00 1.31316744e-02 1.19096017e+00
1.81725040e-01 7.96728879e-02 -2.39934936e-01 -2.70254284e-01
-1.42069787e-01 2.85836369e-01 -5.75720549e-01 -3.23828459e-01
-8.88403654e-01 -7.82431602e-01 1.05392635e+00 7.62892067e-01
9.95645881e-01 -1.04085040e+00 -1.39161777e+00 4.09067392e-01
-3.36925052e-02 -1.52227652e+00 -1.65379360e-01 1.58017203e-01
-4.73773301e-01 -1.06676066e+00 -8.21628690e-01 -4.68585402e-01
-3.53271291e-02 4.94473636e-01 6.20488286e-01 -4.47023481e-01
-9.99548912e-01 7.14115202e-01 -2.32825845e-01 -4.54532206e-01
3.81746441e-01 -3.61143023e-01 2.19690606e-01 4.21065122e-01
7.67407238e-01 -6.74956501e-01 -5.93910158e-01 1.42620176e-01
-4.67260003e-01 -3.67205888e-01 6.89477861e-01 4.51805919e-01
2.02878881e-02 2.08983365e-02 4.42969143e-01 -6.87875986e-01
5.77465594e-01 -2.37974674e-01 -5.37361205e-01 1.21974513e-01
-6.83010817e-02 -2.43797097e-02 6.26175642e-01 -5.04533589e-01
-1.18203986e+00 -2.31087375e-02 1.72570288e-01 -3.45140994e-01
-3.67722780e-01 4.36529696e-01 4.03847128e-01 -2.37475589e-01
9.69856739e-01 6.32911742e-01 -2.99788266e-01 -1.97236583e-01
1.69023275e-01 5.95968425e-01 1.10264087e+00 5.13724647e-02
7.04829752e-01 7.91003108e-01 4.55815345e-02 -1.37077856e+00
-7.00370550e-01 -6.02483034e-01 -5.36442578e-01 -5.77026069e-01
9.50192332e-01 -9.05682504e-01 -8.82564485e-01 5.26347578e-01
-9.55510676e-01 -4.79784459e-02 1.44724190e-01 1.24203765e+00
-4.62707579e-01 3.64053845e-01 -5.51297069e-01 -1.37032735e+00
-3.03534210e-01 -3.24190348e-01 9.66467023e-01 5.18552482e-01
-6.27434999e-02 -8.74887884e-01 2.63719052e-01 -2.58807885e-03
5.20303190e-01 6.30238712e-01 2.29865253e-01 -1.38615638e-01
-7.77245998e-01 -4.79533464e-01 -3.30223143e-02 1.18680112e-01
2.61363715e-01 -2.05909804e-01 -1.00112867e+00 -4.02127244e-02
1.42409086e-01 -1.25583857e-01 1.10694551e+00 8.55099916e-01
7.49618351e-01 1.49910916e-02 -3.52436900e-01 1.06383646e+00
1.02668345e+00 7.52415895e-01 9.25207615e-01 1.10451423e-01
5.72108746e-01 1.35501593e-01 5.52275717e-01 6.37362599e-01
-2.48188213e-01 3.70550483e-01 9.31962803e-02 -7.37392753e-02
8.17378610e-02 3.03969234e-01 5.53089380e-01 1.89208865e-01
-6.44926846e-01 -3.97034995e-02 -6.91330910e-01 2.31875628e-01
-1.63979566e+00 -1.67571998e+00 1.31092161e-01 1.99138403e+00
-2.02261791e-01 1.62600100e-01 1.45643979e-01 2.45972261e-01
6.40173554e-01 7.00824976e-01 -5.09824097e-01 -3.86040598e-01
-5.35398722e-02 3.40635777e-01 7.30319679e-01 2.94481486e-01
-1.34415710e+00 5.43113768e-01 6.74702597e+00 3.70535225e-01
-1.45404899e+00 -2.03606546e-01 -6.00323081e-02 -4.30988878e-01
3.21248323e-01 -7.21083879e-01 -7.00647116e-01 3.74847084e-01
1.15017259e+00 -2.52310019e-02 4.44929481e-01 5.12880743e-01
2.44348705e-01 -1.89961269e-01 -5.91697812e-01 1.22038412e+00
2.83883721e-01 -1.21323895e+00 -3.43222052e-01 -1.21991138e-03
3.18138063e-01 1.56557336e-01 1.52550101e-01 4.80454654e-01
1.79394618e-01 -1.17595303e+00 1.94183767e-01 8.51443291e-01
7.72277296e-01 -8.82863760e-01 4.23238933e-01 3.97797018e-01
-1.63429165e+00 -2.49190792e-01 -3.10409427e-01 -4.73301202e-01
2.40476370e-01 6.09468758e-01 -6.72392190e-01 4.92636174e-01
5.43808043e-01 6.66188419e-01 -1.96327224e-01 8.07563901e-01
-8.50577503e-02 4.73424047e-01 -2.71095514e-01 -3.29724371e-01
5.97608566e-01 -3.34911555e-01 4.19223130e-01 1.45818472e+00
5.86237013e-01 2.69228995e-01 2.14130476e-01 5.02046049e-01
4.75745916e-01 -6.84603751e-01 -1.03073812e+00 -4.08498168e-01
1.65166408e-01 1.05419397e+00 -5.45094669e-01 -8.97932276e-02
-5.12438357e-01 7.25354135e-01 -2.25237265e-01 7.99595952e-01
-1.14035690e+00 -1.00109828e+00 7.29899466e-01 -1.90734908e-01
4.00771379e-01 -8.16643178e-01 -3.87432650e-02 -1.00193751e+00
-6.67141005e-02 -4.13611799e-01 5.90417385e-01 -7.90809929e-01
-6.50704145e-01 8.97338152e-01 -3.05110753e-01 -1.53413606e+00
-7.82457829e-01 -7.04296350e-01 -5.10618865e-01 8.07058573e-01
-1.37652266e+00 -1.02725601e+00 -5.96158981e-01 8.34711909e-01
4.34278220e-01 -3.70629102e-01 8.03325474e-01 2.49780774e-01
-4.57609117e-01 1.98278442e-01 5.76366251e-03 7.43977249e-01
3.39566231e-01 -7.95228899e-01 5.75839639e-01 1.15791690e+00
4.26140726e-01 4.87466395e-01 6.86948180e-01 -6.36657119e-01
-1.34384251e+00 -1.04727912e+00 5.34120083e-01 -1.09566832e-02
5.60348868e-01 -2.89582461e-01 -7.01408923e-01 7.11740434e-01
-1.35011047e-01 3.50997090e-01 8.39941204e-01 1.04543179e-01
-3.49206597e-01 -1.11883700e-01 -7.41994619e-01 2.31228143e-01
1.06166470e+00 -8.39430809e-01 -7.15171099e-01 3.07362020e-01
3.53687227e-01 -3.55229765e-01 -3.68180066e-01 4.38631833e-01
1.10774505e+00 -9.51136768e-01 1.13148546e+00 -5.20177424e-01
-2.77517319e-01 -2.72721112e-01 -1.36529833e-01 -1.06692493e+00
-6.21431768e-01 -7.27272391e-01 -5.41614234e-01 3.47915813e-02
-9.89565067e-03 -8.07246447e-01 8.47892225e-01 -2.01821953e-01
-5.13969585e-02 -2.74430215e-01 -1.07450891e+00 -1.17708874e+00
-8.51635873e-01 -6.33619606e-01 3.59595209e-01 6.17585897e-01
-5.19703105e-02 2.62771219e-01 -8.65503252e-01 4.79504287e-01
7.16744304e-01 3.90165359e-01 5.87407410e-01 -1.29447603e+00
-3.53849441e-01 -1.22280508e-01 -7.22831309e-01 -1.29194188e+00
-8.33799392e-02 -2.05311477e-01 4.96749803e-02 -1.28122175e+00
-5.97796321e-01 1.14397049e-01 -4.53294933e-01 1.22443348e-01
3.75994116e-01 5.60214221e-01 -8.91770273e-02 -3.85784470e-02
-5.23785532e-01 3.54519188e-01 1.01165795e+00 -9.77910832e-02
-2.32460544e-01 3.20253104e-01 -1.94719166e-01 8.36814284e-01
5.97574472e-01 -1.92095339e-02 -3.22536170e-01 -2.04181001e-01
-1.68833658e-02 3.01093727e-01 5.81510723e-01 -1.72990072e+00
6.49748623e-01 -1.69887871e-01 7.96479583e-01 -7.22932279e-01
7.53071845e-01 -5.90771794e-01 6.06344268e-02 8.73415172e-01
-3.20611522e-02 -1.43055558e-01 -6.31717294e-02 9.72705185e-01
-2.63040215e-01 1.17548399e-01 5.38035095e-01 -1.78638667e-01
-1.16647410e+00 3.16034824e-01 -1.13694322e+00 -4.17911768e-01
8.66325974e-01 -6.76659882e-01 -6.59312427e-01 -6.57289326e-01
-5.99828541e-01 -1.31921545e-01 -5.32128334e-01 3.65208924e-01
7.02787578e-01 -1.38458085e+00 -5.04747331e-01 6.98223114e-01
-7.72783831e-02 -8.23146760e-01 4.90871757e-01 7.24997580e-01
-5.49538493e-01 9.95386720e-01 -5.59651613e-01 -5.82025647e-01
-1.09319937e+00 2.83131778e-01 6.47850871e-01 -1.42058656e-01
-7.86632061e-01 5.87110758e-01 6.87733442e-02 1.53232634e-01
2.91725457e-01 -3.59899819e-01 -5.53719461e-01 -4.72460613e-02
9.98841107e-01 5.04367411e-01 5.19606881e-02 -5.47116399e-01
-4.40830410e-01 6.45564854e-01 9.39046144e-02 -2.17465043e-01
1.09331214e+00 7.62577280e-02 4.56247777e-01 5.51775575e-01
8.82384598e-01 -1.85422033e-01 -1.11550128e+00 -3.52978289e-01
-2.40340054e-01 -4.65778679e-01 1.25184327e-01 -5.76248884e-01
-8.68479669e-01 9.12123561e-01 8.69961500e-01 6.41188800e-01
1.29864120e+00 -4.46785599e-01 8.89807940e-01 9.69596148e-01
2.92058945e-01 -1.04227388e+00 2.14549527e-01 9.57945645e-01
5.14264107e-01 -6.79819226e-01 -9.03351307e-02 -1.16362214e-01
-3.71583998e-01 1.35085034e+00 4.65025425e-01 -4.97815222e-01
6.02077425e-01 5.25460005e-01 3.46958786e-01 -2.73546875e-01
-5.59223711e-01 -6.07381523e-01 2.51823455e-01 9.99201655e-01
4.16225314e-01 1.22087725e-01 1.32336378e-01 2.22783566e-01
-2.76513189e-01 -8.93399268e-02 4.43874806e-01 9.97895956e-01
-9.84947383e-01 -1.70741916e-01 -6.78671241e-01 4.63963717e-01
-3.21710318e-01 3.01424772e-01 -1.31308481e-01 6.89081728e-01
-5.27749509e-02 8.47769558e-01 5.51752031e-01 -5.40548921e-01
3.84575039e-01 7.18182325e-02 6.03090644e-01 -2.99547672e-01
-1.71808288e-01 -4.00556475e-02 2.09705114e-01 -7.14690268e-01
-6.89730883e-01 -3.12542677e-01 -1.17292345e+00 -1.35929942e-01
-1.42858550e-01 2.17756070e-02 5.84636271e-01 1.18155003e+00
9.15120617e-02 9.36640680e-01 7.34274864e-01 -8.01085293e-01
-2.56224304e-01 -9.54226851e-01 -8.80627751e-01 -9.11523178e-02
7.32807577e-01 -8.33164930e-01 -2.26040810e-01 -6.03127629e-02]
|
[6.800650596618652, 0.41226014494895935]
|
0b2f3a68-a80c-4d1d-a54f-aef86f6716e2
|
skygpt-probabilistic-short-term-solar
|
2306.11682
| null |
https://arxiv.org/abs/2306.11682v1
|
https://arxiv.org/pdf/2306.11682v1.pdf
|
SkyGPT: Probabilistic Short-term Solar Forecasting Using Synthetic Sky Videos from Physics-constrained VideoGPT
|
In recent years, deep learning-based solar forecasting using all-sky images has emerged as a promising approach for alleviating uncertainty in PV power generation. However, the stochastic nature of cloud movement remains a major challenge for accurate and reliable solar forecasting. With the recent advances in generative artificial intelligence, the synthesis of visually plausible yet diversified sky videos has potential for aiding in forecasts. In this study, we introduce \emph{SkyGPT}, a physics-informed stochastic video prediction model that is able to generate multiple possible future images of the sky with diverse cloud motion patterns, by using past sky image sequences as input. Extensive experiments and comparison with benchmark video prediction models demonstrate the effectiveness of the proposed model in capturing cloud dynamics and generating future sky images with high realism and diversity. Furthermore, we feed the generated future sky images from the video prediction models for 15-minute-ahead probabilistic solar forecasting for a 30-kW roof-top PV system, and compare it with an end-to-end deep learning baseline model SUNSET and a smart persistence model. Better PV output prediction reliability and sharpness is observed by using the predicted sky images generated with SkyGPT compared with other benchmark models, achieving a continuous ranked probability score (CRPS) of 2.81 (13\% better than SUNSET and 23\% better than smart persistence) and a Winkler score of 26.70 for the test set. Although an arbitrary number of futures can be generated from a historical sky image sequence, the results suggest that 10 future scenarios is a good choice that balances probabilistic solar forecasting performance and computational cost.
|
['Adam Brandt', 'Quentin Paletta', 'Andea Scott', 'Eric Zelikman', 'Yuhao Nie']
|
2023-06-20
| null | null | null | null |
['video-prediction']
|
['computer-vision']
|
[-6.81800693e-02 -2.60355622e-01 2.13556826e-01 -2.95098186e-01
-8.06207240e-01 -8.66878569e-01 9.25319135e-01 -6.26399636e-01
6.79916501e-01 1.00705409e+00 2.80207038e-01 -2.54928976e-01
5.49352616e-02 -9.29799557e-01 -1.09985399e+00 -1.32718801e+00
-7.72267506e-02 3.03149611e-01 -5.94279496e-03 7.51774684e-02
-2.34067604e-01 2.46275917e-01 -1.93839896e+00 3.41482222e-01
1.21770930e+00 1.08207011e+00 8.46541405e-01 1.27791786e+00
1.68448240e-02 9.63280559e-01 -7.96328127e-01 -3.04186493e-01
2.28662014e-01 -4.38929945e-01 1.68973014e-01 -4.69904877e-02
6.85945690e-01 -2.47463360e-01 -4.26773489e-01 7.47877777e-01
6.95579708e-01 -1.32544041e-01 7.11558759e-01 -1.36701310e+00
-4.71031934e-01 4.42424208e-01 -2.89043248e-01 1.61639884e-01
3.25184852e-01 7.64989853e-01 6.19608819e-01 -6.91832781e-01
3.31614554e-01 8.11065018e-01 5.18219531e-01 1.48312584e-01
-9.35742319e-01 -6.70900881e-01 3.70917544e-02 3.13774198e-01
-1.30159390e+00 7.07254186e-02 6.24202847e-01 -6.77429557e-01
7.84413278e-01 3.83855224e-01 1.13150573e+00 1.49736512e+00
6.53539896e-01 7.87513971e-01 1.22607982e+00 2.50865854e-02
6.11283839e-01 -7.80625094e-04 -5.04899204e-01 1.33111820e-01
1.60665572e-01 5.50871015e-01 -5.99005640e-01 -1.65676754e-02
2.60214299e-01 -1.85512409e-01 -4.72084522e-01 -1.66691631e-01
-1.08298206e+00 7.34354496e-01 5.31306744e-01 -8.39108601e-02
-7.25002646e-01 3.85103196e-01 -2.20810175e-01 -3.12285036e-01
5.31686485e-01 2.33726874e-02 -2.45309517e-01 -2.00179771e-01
-1.39462602e+00 6.31368577e-01 8.65164101e-01 9.85309362e-01
2.84637392e-01 1.18857813e+00 -6.59711480e-01 2.19165191e-01
3.20337772e-01 1.60621583e+00 3.08600843e-01 -9.78698075e-01
2.12672576e-01 -2.35309452e-02 6.16750360e-01 -6.04766011e-01
-1.71074435e-01 -8.25081527e-01 -1.02793193e+00 3.78974974e-01
-4.63457741e-02 -5.62577307e-01 -1.30088615e+00 1.36556804e+00
1.80702701e-01 8.89513910e-01 3.62521857e-01 9.29961383e-01
8.12145591e-01 1.60392463e+00 2.88400743e-02 -2.62304932e-01
7.96832263e-01 -9.74285185e-01 -5.25774717e-01 -9.98563394e-02
4.01965678e-02 -5.65340400e-01 7.74401903e-01 5.52493274e-01
-6.75109148e-01 -6.94560885e-01 -9.95931447e-01 7.21368134e-01
-2.30908364e-01 1.66802049e-01 5.49671292e-01 7.94878066e-01
-1.36024904e+00 3.21075857e-01 -8.34581137e-01 -1.74397063e-02
2.24169701e-01 -2.10434154e-01 4.13362116e-01 -1.10634148e-01
-1.05411422e+00 6.08859539e-01 4.46999669e-01 2.14861125e-01
-1.51335669e+00 -9.04296517e-01 -6.19185030e-01 4.36078668e-01
5.21107912e-02 -9.83657539e-01 1.10423350e+00 -1.12031662e+00
-1.64761710e+00 -1.87103316e-01 -3.15788001e-01 -8.87574553e-01
5.29643476e-01 -2.00685948e-01 -6.74412906e-01 -1.27458379e-01
-2.04656273e-01 8.30935061e-01 1.21899509e+00 -1.53387284e+00
-8.47396016e-01 6.53738007e-02 -2.18259409e-01 5.89487553e-01
2.28468910e-01 -7.36071348e-01 -2.09708512e-01 -7.22527504e-01
-5.69041491e-01 -1.21060073e+00 -1.10295683e-01 -2.72953302e-01
-1.66570425e-01 -4.23678048e-02 1.02490759e+00 -9.73408639e-01
7.94005990e-01 -1.78928268e+00 -9.13415803e-04 9.12713259e-03
-2.69687265e-01 3.76100183e-01 6.17600791e-02 3.67890686e-01
2.97528170e-02 -8.76152664e-02 -3.18175554e-01 -3.06418478e-01
-3.60594131e-02 3.78873110e-01 -9.77456391e-01 -2.64132340e-02
9.09804832e-03 1.05279958e+00 -8.02520931e-01 1.18046455e-01
6.37542963e-01 8.22320044e-01 1.26967326e-01 2.64775693e-01
-1.08345509e+00 5.29750884e-01 -1.82914644e-01 5.68844676e-01
9.64187264e-01 -3.24574888e-01 5.48897572e-02 2.46457562e-01
5.01807891e-02 -2.12113559e-01 -1.02759004e+00 1.24948990e+00
-3.37391198e-01 9.03007030e-01 -4.98764127e-01 -2.22450390e-01
9.37941194e-01 2.53269821e-01 1.36060685e-01 -8.54940891e-01
-4.07499254e-01 -1.43792242e-01 -4.76562828e-01 -2.73899853e-01
7.21140742e-01 7.00418130e-02 3.68800849e-01 7.90097415e-02
-3.54665935e-01 -7.67023921e-01 6.56099916e-02 2.30560437e-01
5.64162254e-01 5.98549187e-01 -2.99275279e-01 -2.44580619e-02
2.30190799e-01 1.46405578e-01 5.88147223e-01 6.94073975e-01
3.46115828e-01 9.30719435e-01 4.15637977e-02 -5.21024168e-01
-1.16842270e+00 -1.41725194e+00 1.57874785e-02 3.18482161e-01
-4.96013388e-02 -2.05407441e-01 -5.88725030e-01 -3.79701555e-01
1.73782542e-01 1.56772351e+00 -1.67309135e-01 3.12883765e-01
-1.48153067e-01 -1.22614467e+00 4.85795327e-02 5.74953735e-01
3.87069285e-01 -8.91613543e-01 -6.37075841e-01 7.90667012e-02
-2.89240509e-01 -1.04837728e+00 1.95811182e-01 -2.85840452e-01
-6.16349101e-01 -7.75322497e-01 -1.14116228e+00 -1.22882985e-01
1.25027299e-01 5.17981946e-01 1.42278445e+00 -4.42201346e-01
-9.37790126e-02 6.78331912e-01 -3.20981830e-01 -6.48816168e-01
-5.58255315e-01 -4.56451416e-01 8.47349763e-02 -1.43041044e-01
-2.05527157e-01 -4.65164572e-01 -8.63639534e-01 6.10270314e-02
-9.22024310e-01 4.52432334e-01 2.61954308e-01 6.11451387e-01
6.35023594e-01 5.01433194e-01 4.21378046e-01 -2.07670182e-01
1.52164012e-01 -8.64185929e-01 -1.26615775e+00 3.97352159e-01
-7.91366041e-01 -1.87844411e-01 7.61655807e-01 -1.13404118e-01
-1.59741020e+00 3.64064336e-01 1.41057581e-01 -7.90245712e-01
-6.49624094e-02 5.19397080e-01 1.17723607e-01 6.15104400e-02
4.96313691e-01 8.93552601e-01 -5.38444459e-01 1.54008949e-03
5.92818916e-01 2.93910563e-01 5.88649809e-01 -1.64024770e-01
1.10159183e+00 5.32328904e-01 2.75001619e-02 -9.97749865e-01
-5.01968741e-01 7.03627989e-02 -4.77806181e-02 -6.28057241e-01
7.72282064e-01 -1.76396871e+00 -5.18114507e-01 8.68390858e-01
-8.69517505e-01 -5.57469666e-01 -2.61810035e-01 2.31847510e-01
-5.70759177e-01 3.42032969e-01 1.03768088e-01 -1.34083366e+00
-5.46559393e-01 -8.37042451e-01 1.19889772e+00 6.66371286e-01
3.01093280e-01 -8.43580723e-01 3.04770738e-01 3.24763417e-01
5.76983809e-01 5.50353825e-01 5.74896574e-01 9.50768217e-02
-1.42178464e+00 1.55956270e-02 3.85245718e-02 3.89380932e-01
-1.02491699e-01 4.94331747e-01 -1.22360539e+00 -4.72996265e-01
-5.71883097e-02 -2.87048779e-02 8.65067482e-01 8.71662021e-01
1.06634140e+00 -4.09285754e-01 -2.29132950e-01 7.03214049e-01
1.67689157e+00 4.77761269e-01 7.72977829e-01 1.70224030e-02
6.90319419e-01 8.74683782e-02 5.27835429e-01 8.22187066e-01
6.65475607e-01 3.02860975e-01 9.30812716e-01 2.13050812e-01
-1.86373234e-01 -4.22725260e-01 6.93535089e-01 4.12628442e-01
2.32437372e-01 -1.12002242e+00 -8.55022311e-01 7.22257853e-01
-1.72174788e+00 -1.18519306e+00 -1.64802417e-01 2.40908599e+00
6.47187382e-02 -7.47770723e-03 -1.03590339e-01 -2.10303500e-01
3.93224448e-01 6.38181627e-01 -8.41602862e-01 -4.08747117e-04
-7.91126788e-01 -7.22524300e-02 5.56718290e-01 3.53472471e-01
-8.57407928e-01 6.78628147e-01 6.24067879e+00 8.57725739e-01
-1.39809406e+00 -1.72837690e-01 9.87606466e-01 -2.82457083e-01
-8.45436215e-01 -3.39908823e-02 -9.95926678e-01 1.20745850e+00
1.38068807e+00 -2.77077824e-01 4.99813497e-01 8.43719125e-01
5.98022103e-01 -5.62775433e-01 -5.46655238e-01 9.59304452e-01
3.00861329e-01 -2.01870608e+00 2.75532812e-01 -1.74128428e-01
1.41500366e+00 5.33179939e-01 3.81449580e-01 1.44200310e-01
7.46245980e-01 -1.13299370e+00 6.85525775e-01 1.10250890e+00
4.31849152e-01 -5.79985559e-01 5.69524169e-01 5.37388682e-01
-1.25138962e+00 -2.24993408e-01 -2.95065850e-01 1.43672302e-01
4.06076372e-01 8.67769301e-01 -1.17654443e+00 9.62916076e-01
1.09589136e+00 6.80950046e-01 -4.47529823e-01 1.26506960e+00
-4.02523369e-01 9.90674794e-01 -7.61582375e-01 -2.79910892e-01
2.07444444e-01 -2.79016554e-01 7.16552913e-01 1.03177619e+00
1.11916435e+00 3.31822596e-03 -9.49570760e-02 8.41707468e-01
2.99558401e-01 -5.18760145e-01 -8.52434874e-01 -1.34594217e-01
2.61203021e-01 1.10323942e+00 -1.71331301e-01 -5.14639139e-01
-9.05929282e-02 9.45489764e-01 -4.62033927e-01 6.85037374e-01
-1.21410799e+00 4.55444753e-01 5.87504268e-01 -1.13315806e-01
9.07081485e-01 -2.96167254e-01 -2.67776132e-01 -1.14783502e+00
7.73474127e-02 -6.64000213e-01 2.84809262e-01 -1.90279281e+00
-1.03449678e+00 6.79207385e-01 7.29454458e-02 -1.36215413e+00
-7.58952677e-01 -3.32410634e-01 -1.10415578e+00 1.16596007e+00
-1.73463607e+00 -1.43575537e+00 -7.21195161e-01 3.40870291e-01
8.49388003e-01 -1.92482144e-01 6.13207221e-01 -5.16323447e-01
-1.46388426e-01 -1.05510645e-01 8.58170986e-01 -8.50245714e-01
1.63064554e-01 -1.39032996e+00 8.35504711e-01 1.18956387e+00
4.07841623e-01 -4.33021277e-01 1.01152360e+00 -8.63302231e-01
-1.55102146e+00 -1.42269313e+00 5.51628888e-01 -5.70147038e-01
3.54778111e-01 -8.61272663e-02 -7.43628681e-01 3.14904124e-01
6.28425419e-01 -3.45291287e-01 2.47145236e-01 -6.64530694e-01
-1.17731355e-02 -3.87722492e-01 -9.60743070e-01 4.99861628e-01
4.31261837e-01 -1.76456466e-01 -8.35769176e-02 7.62045205e-01
7.94033647e-01 -5.35291195e-01 -7.46745169e-01 5.74434757e-01
4.12865341e-01 -1.16502213e+00 9.08746779e-01 2.64968025e-03
3.36759299e-01 -7.32648671e-01 -3.88076276e-01 -1.62789690e+00
-3.53194952e-01 -7.55061984e-01 -5.14295995e-01 1.08394969e+00
4.18709695e-01 -3.24761033e-01 1.08457410e+00 6.09204412e-01
-1.05429806e-01 -3.76260102e-01 -9.21066105e-01 -7.71323323e-01
-2.12978199e-02 -5.08491218e-01 8.14189792e-01 5.04638195e-01
-8.57075810e-01 -1.95775732e-01 -6.85957849e-01 8.94595504e-01
7.23053634e-01 5.74638665e-01 1.00030363e+00 -7.91520834e-01
-3.64628226e-01 -3.05291750e-02 -6.67516515e-02 -7.67661929e-01
-7.50102028e-02 -4.22462136e-01 9.34107974e-02 -1.80646944e+00
9.44142193e-02 -1.73972487e-01 -1.95717022e-01 -1.39001384e-02
-2.29406327e-01 -6.18396103e-02 4.52746928e-01 1.16532199e-01
-8.75033513e-02 1.14581823e+00 8.87054265e-01 -3.08230281e-01
-8.89419839e-02 3.03249836e-01 -6.71179220e-02 3.50370258e-01
7.98554242e-01 -1.52781337e-01 -6.28924310e-01 -5.32712877e-01
3.99442554e-01 4.07811850e-01 6.02339208e-01 -1.50767672e+00
1.10620260e-01 -1.55292809e-01 7.84150898e-01 -1.20488107e+00
8.11307192e-01 -6.64239228e-01 1.01060259e+00 3.57566029e-01
4.72716719e-01 2.81422585e-02 4.26877618e-01 9.63774621e-01
-2.44935751e-02 1.89493701e-01 4.10509825e-01 -5.13279922e-02
-1.10677063e+00 3.38147998e-01 -5.57253420e-01 -4.64949042e-01
1.09401226e+00 -2.35625699e-01 -5.59856534e-01 -1.02187014e+00
-5.79473615e-01 4.86871272e-01 4.87451255e-01 6.77094519e-01
5.73415518e-01 -1.00960970e+00 -8.61332715e-01 1.34480953e-01
-3.01189069e-02 6.05356544e-02 7.81706154e-01 2.73768809e-02
-5.85788786e-01 4.97609794e-01 1.53789327e-01 -1.19361579e+00
-9.66799676e-01 3.18015397e-01 3.40120286e-01 -1.35807350e-04
-6.78606927e-01 6.43801868e-01 4.31676656e-01 -1.41027644e-01
6.87282607e-02 -4.71998394e-01 -3.92384306e-02 -2.39688843e-01
3.76880556e-01 3.74230713e-01 -5.28754182e-02 -5.34306586e-01
1.10737436e-01 3.61827374e-01 6.55458629e-01 2.30059680e-02
1.35014009e+00 -1.55675530e-01 5.00658214e-01 4.56261724e-01
2.34676600e-01 -3.11126232e-01 -2.11954570e+00 6.15286194e-02
-3.68494958e-01 -5.01217008e-01 1.95763275e-01 -1.49552786e+00
-1.16726649e+00 7.76872873e-01 1.15001523e+00 9.64294970e-02
1.09866679e+00 -3.49766880e-01 7.59373426e-01 2.63382047e-01
4.03139621e-01 -6.41650438e-01 -2.78550178e-01 2.89489716e-01
9.84889269e-01 -1.20411193e+00 -8.63012597e-02 -4.73913923e-02
-1.27125919e+00 8.04971695e-01 5.81540525e-01 2.67803501e-02
4.30494815e-01 2.48086289e-01 2.25776695e-02 1.65285230e-01
-1.02888846e+00 9.77050960e-02 3.49174887e-01 7.57271588e-01
-3.17720294e-01 5.45217335e-01 5.24641693e-01 2.52168357e-01
-2.58229852e-01 2.44212016e-01 5.87555885e-01 5.01664758e-01
-4.29615438e-01 -4.17842448e-01 -6.05310380e-01 4.23092932e-01
1.35844022e-01 -2.72956222e-01 1.23993553e-01 3.01216215e-01
-5.33505417e-02 9.88033414e-01 1.49661884e-01 -2.12284565e-01
-1.59436658e-01 7.04487860e-02 1.12793803e-01 3.75264212e-02
-1.73616976e-01 6.06800355e-02 -8.63020420e-02 -3.91223729e-01
-2.70768523e-01 -6.21346176e-01 -6.53506219e-01 -3.84691685e-01
-1.94154978e-01 -1.58670601e-02 1.02378237e+00 7.89982438e-01
7.86449611e-01 4.78872657e-01 9.38209414e-01 -1.17721581e+00
-2.01871917e-01 -6.59270585e-01 -4.65998769e-01 -2.01266438e-01
2.74343073e-01 -3.18470150e-01 -4.57523495e-01 1.03573151e-01]
|
[6.457332134246826, 2.702085256576538]
|
5d9f283d-2e6a-453b-8813-1bd82efdf5f4
|
ctds-centralized-teacher-with-decentralized
|
2203.08412
| null |
https://arxiv.org/abs/2203.08412v1
|
https://arxiv.org/pdf/2203.08412v1.pdf
|
CTDS: Centralized Teacher with Decentralized Student for Multi-Agent Reinforcement Learning
|
Due to the partial observability and communication constraints in many multi-agent reinforcement learning (MARL) tasks, centralized training with decentralized execution (CTDE) has become one of the most widely used MARL paradigms. In CTDE, centralized information is dedicated to learning the allocation of the team reward with a mixing network, while the learning of individual Q-values is usually based on local observations. The insufficient utility of global observation will degrade performance in challenging environments. To this end, this work proposes a novel Centralized Teacher with Decentralized Student (CTDS) framework, which consists of a teacher model and a student model. Specifically, the teacher model allocates the team reward by learning individual Q-values conditioned on global observation, while the student model utilizes the partial observations to approximate the Q-values estimated by the teacher model. In this way, CTDS balances the full utilization of global observation during training and the feasibility of decentralized execution for online inference. Our CTDS framework is generic which is ready to be applied upon existing CTDE methods to boost their performance. We conduct experiments on a challenging set of StarCraft II micromanagement tasks to test the effectiveness of our method and the results show that CTDS outperforms the existing value-based MARL methods.
|
['Houqiang Li', 'Jiangcheng Zhu', 'Wengang Zhou', 'Mingyu Yang', 'Xunhan Hu', 'Jian Zhao']
|
2022-03-16
| null | null | null | null |
['starcraft-ii']
|
['playing-games']
|
[-4.10758078e-01 1.51124761e-01 -2.78904855e-01 -4.86506559e-02
-5.78829408e-01 -2.32656822e-01 5.34221113e-01 2.17600256e-01
-5.42536438e-01 1.13289857e+00 -2.88124174e-01 -1.78323731e-01
-2.96972096e-01 -8.03038657e-01 -6.75195992e-01 -1.09469473e+00
-1.32611901e-01 7.38901615e-01 3.27564031e-01 -2.23215967e-01
-1.29648849e-01 2.03522995e-01 -1.43108332e+00 -1.13481127e-01
1.04073429e+00 1.02523887e+00 3.58192265e-01 4.86832500e-01
1.94440126e-01 1.16012788e+00 -7.96432436e-01 2.18009353e-02
3.26778740e-01 -4.91726190e-01 -4.49492335e-01 7.25595802e-02
-7.76080266e-02 -7.66139984e-01 -1.38431266e-01 9.55228984e-01
5.82279027e-01 5.91061234e-01 1.66414797e-01 -1.64716971e+00
6.02762364e-02 9.88440216e-01 -5.77020645e-01 -6.25698417e-02
-1.82623994e-02 3.47692192e-01 9.57129180e-01 -3.78904164e-01
3.81672472e-01 1.29742146e+00 2.84159660e-01 4.30197626e-01
-1.14477015e+00 -5.76056540e-01 5.95785975e-01 2.96505719e-01
-1.01432979e+00 -1.27447039e-01 6.66776538e-01 -2.42392272e-01
7.42007852e-01 -1.77416459e-01 7.65860617e-01 8.17245364e-01
2.77123839e-01 9.57645118e-01 1.22601068e+00 -1.27683550e-01
7.29473650e-01 1.36800736e-01 -1.53102830e-01 9.29271281e-01
3.97826731e-02 3.13963801e-01 -6.83212101e-01 -3.35155308e-01
8.05906057e-01 8.52157474e-02 3.03041637e-02 -5.69591522e-01
-1.10034001e+00 7.95033991e-01 4.33990180e-01 -1.28334701e-01
-6.07549489e-01 3.76154691e-01 5.40785074e-01 7.13967681e-01
3.89004052e-01 1.57819226e-01 -4.42257077e-01 -1.86713666e-01
-7.22127795e-01 4.65281963e-01 8.58925939e-01 7.09124744e-01
9.61732209e-01 4.46500033e-01 -1.63099468e-01 7.02972889e-01
3.97372693e-01 5.94366491e-01 5.93390107e-01 -1.16553271e+00
4.99857962e-01 6.44618928e-01 1.95325881e-01 -6.56904042e-01
-4.14371878e-01 -5.80903649e-01 -6.62257671e-01 4.68325704e-01
3.49002153e-01 -7.14712262e-01 -4.01534855e-01 1.78251922e+00
9.60180700e-01 2.51491815e-01 3.62020016e-01 1.15232658e+00
3.71272236e-01 6.01593733e-01 -1.90825805e-01 -3.29253614e-01
1.01607263e+00 -1.21635127e+00 -6.38678610e-01 -5.77127980e-03
6.64708912e-01 -3.32781613e-01 5.44357955e-01 6.29571319e-01
-9.29886758e-01 -5.64134002e-01 -9.28147972e-01 4.15932924e-01
1.35289937e-01 3.20000499e-01 5.63729048e-01 3.46329421e-01
-8.25239539e-01 7.06354618e-01 -1.30119145e+00 6.80155978e-02
1.30660459e-01 5.29662490e-01 -4.39510122e-02 7.51062408e-02
-9.71542776e-01 7.40605712e-01 5.12549579e-01 2.20722243e-01
-1.78400159e+00 -2.94182152e-01 -6.05814338e-01 1.80935532e-01
9.56368923e-01 -5.88863194e-01 1.53315556e+00 -9.49859023e-01
-2.16058159e+00 -4.37025204e-02 4.14307654e-01 -6.83945715e-01
6.76195681e-01 -1.49231657e-01 1.61061227e-01 3.08410347e-01
4.82020788e-02 3.48097652e-01 9.67647254e-01 -1.14378715e+00
-9.31740046e-01 -1.67260900e-01 4.37833279e-01 5.08750379e-01
-2.97855675e-01 -4.16950822e-01 1.23554319e-01 -2.67598152e-01
-3.10482055e-01 -9.77958918e-01 -4.38030273e-01 -2.37114906e-01
-1.30147949e-01 -4.85891461e-01 7.50839233e-01 -3.04951221e-01
9.55408037e-01 -2.04927754e+00 4.44243014e-01 2.38345116e-01
3.51364642e-01 1.39825223e-02 -2.56011784e-01 8.58722448e-01
4.57687497e-01 -5.83579421e-01 -1.20178116e-02 -4.26473379e-01
1.86066478e-01 5.93399763e-01 -2.98123181e-01 4.37025756e-01
-3.48172225e-02 3.90397131e-01 -1.14296675e+00 -5.43225169e-01
1.31252944e-01 -2.13815342e-03 -7.34600306e-01 6.73453867e-01
-6.27982020e-01 7.00518668e-01 -8.96309376e-01 5.02704680e-01
3.42654377e-01 -1.98384568e-01 7.15970755e-01 3.27295095e-01
-1.96714327e-01 1.89581394e-01 -1.49982500e+00 1.45606923e+00
-5.05381107e-01 6.06379136e-02 4.47138935e-01 -1.27862537e+00
9.00596499e-01 5.46123266e-01 6.71272814e-01 -5.57551682e-01
8.53127986e-02 2.55867928e-01 2.57879376e-01 -4.65860188e-01
2.77463198e-01 5.47535196e-02 1.02890514e-01 7.50432134e-01
2.45445013e-01 -1.30153731e-01 2.94800341e-01 3.22927088e-01
1.12468731e+00 4.03761685e-01 2.87530839e-01 -1.30624741e-01
5.28057277e-01 -2.08705906e-02 1.03283036e+00 7.89823294e-01
-1.75862834e-01 -1.89707875e-01 7.38262355e-01 -3.69823992e-01
-7.90060461e-01 -9.38638210e-01 5.36479950e-01 1.29011047e+00
1.97937593e-01 -3.26497316e-01 -5.69119930e-01 -9.72042441e-01
4.04817797e-02 3.87732327e-01 -3.80773455e-01 -2.10852828e-02
-4.57247317e-01 -4.79880214e-01 1.17340922e-01 3.54600996e-01
5.23012280e-01 -1.05725682e+00 -9.32358384e-01 5.16492903e-01
3.66433524e-03 -9.68483448e-01 -3.22762638e-01 3.37557703e-01
-9.32909906e-01 -1.03468096e+00 -3.97660315e-01 -4.29542512e-01
7.50120044e-01 2.16628850e-01 7.18007207e-01 2.23900333e-01
1.38970256e-01 5.01954734e-01 -4.56500679e-01 -3.77943903e-01
-3.25733423e-01 2.36024067e-01 3.20387900e-01 1.51526943e-01
-2.89306521e-01 -7.34849274e-01 -5.29299974e-01 3.70701134e-01
-8.55953336e-01 1.20110013e-01 6.16739988e-01 1.22624624e+00
3.97606164e-01 9.80165973e-02 7.28506088e-01 -7.70048141e-01
6.64362192e-01 -6.09546721e-01 -1.23340976e+00 6.16465472e-02
-6.52220964e-01 2.56213695e-01 1.14289224e+00 -7.82355666e-01
-1.16709471e+00 -3.15031670e-02 1.73666418e-01 -5.10025859e-01
1.71623468e-01 7.00483739e-01 1.20402977e-01 1.30232841e-01
2.86738276e-01 1.70658916e-01 3.47968191e-01 -3.79510939e-01
1.00561723e-01 4.99384373e-01 1.29189372e-01 -1.19445050e+00
7.26618588e-01 2.18295112e-01 1.30599350e-01 -4.52599376e-01
-8.01715612e-01 -1.18126519e-01 -1.81530714e-01 -4.34584200e-01
3.19523454e-01 -1.13580596e+00 -1.33738697e+00 3.65734309e-01
-7.91684091e-01 -8.15191567e-01 -4.49330151e-01 8.73735547e-01
-6.57478452e-01 2.03257471e-01 -6.40260935e-01 -1.10566700e+00
-2.22996876e-01 -1.41497970e+00 7.23894358e-01 3.10032636e-01
3.80804658e-01 -1.01782084e+00 3.79836977e-01 3.23823482e-01
3.19595456e-01 7.89695382e-02 5.40242672e-01 -6.56816900e-01
-7.25833833e-01 1.63807169e-01 2.33064309e-01 3.19109946e-01
-9.40886065e-02 -1.51210412e-01 -6.61852837e-01 -7.87669420e-01
-4.67042811e-02 -9.55461860e-01 4.30162132e-01 1.66757017e-01
8.35149467e-01 -4.75330830e-01 -5.91455214e-02 2.16088936e-01
1.43182278e+00 1.20576262e-01 -7.76466131e-02 4.86171097e-01
4.22222018e-01 4.59012091e-01 9.08399642e-01 1.11043119e+00
6.78308845e-01 5.97425222e-01 8.86991262e-01 1.51010305e-01
3.36410612e-01 -4.17444080e-01 9.19848502e-01 1.08711040e+00
-1.41161710e-01 -3.27646546e-02 -5.95729113e-01 4.14685309e-01
-2.53465557e+00 -7.94696927e-01 2.81655043e-01 2.15895510e+00
9.45877373e-01 -1.08163908e-01 3.39787722e-01 -1.89475670e-01
3.76561731e-01 5.92894405e-02 -8.01484644e-01 -3.72624397e-01
4.19690043e-01 -9.22577232e-02 2.77295440e-01 5.08354485e-01
-6.58088803e-01 7.45224178e-01 4.98719072e+00 8.52842927e-01
-9.27315354e-01 3.01155061e-01 1.59690142e-01 -1.22366399e-01
2.92094015e-02 2.54990697e-01 -8.83792579e-01 4.86878037e-01
9.04389799e-01 -1.33563340e-01 7.97578633e-01 1.12205291e+00
4.84847009e-01 -4.25185293e-01 -1.03886521e+00 5.54053485e-01
-3.84445727e-01 -8.99839163e-01 -3.18827569e-01 9.46392119e-02
8.87044549e-01 1.51793987e-01 -2.30659828e-01 6.44002795e-01
1.00338018e+00 -4.83991921e-01 7.23505080e-01 3.13488305e-01
1.88665852e-01 -8.12387228e-01 8.54632974e-01 9.62777555e-01
-1.11629081e+00 -4.92432356e-01 -4.84195024e-01 -2.87421316e-01
-6.81937709e-02 4.99559730e-01 -9.47951019e-01 8.83746564e-01
4.87935185e-01 5.46317101e-01 -1.10848963e-01 8.29514802e-01
-6.34913683e-01 7.18770385e-01 -4.40082669e-01 -1.96459264e-01
4.59498465e-01 -5.57810128e-01 5.15328646e-01 5.17795861e-01
1.24375656e-01 -3.99617804e-03 1.00309753e+00 5.24963498e-01
5.71986996e-02 3.98833342e-02 -3.07345718e-01 1.40613332e-01
5.99987984e-01 1.45081067e+00 -4.08286065e-01 -5.26412070e-01
-2.28369117e-01 4.08082396e-01 7.92286217e-01 2.53978372e-01
-8.10211420e-01 -1.32758915e-01 5.46254575e-01 -4.00508225e-01
3.84435773e-01 -3.45098406e-01 3.44329536e-01 -1.27754688e+00
-1.57964692e-01 -1.17740536e+00 5.08869648e-01 -4.46406901e-01
-1.22091854e+00 4.53405201e-01 2.21291948e-02 -1.35760212e+00
-6.01903737e-01 -1.64338127e-01 -6.98204517e-01 5.02032578e-01
-1.63311064e+00 -8.46700072e-01 -1.99865475e-01 6.60363853e-01
4.27264214e-01 -3.88909757e-01 5.78034937e-01 8.58057961e-02
-1.01356840e+00 3.58039021e-01 3.03307384e-01 6.52725066e-05
5.89124918e-01 -1.34236073e+00 -3.77380490e-01 6.16973162e-01
-8.54958370e-02 3.05627525e-01 7.34918475e-01 -4.57641244e-01
-1.56316042e+00 -8.35072875e-01 1.26795352e-01 1.39554307e-01
8.95640790e-01 -2.40523905e-01 -7.24044502e-01 6.57631218e-01
3.25663209e-01 9.64872167e-02 3.42835873e-01 -7.74797723e-02
-4.55251075e-02 -6.47104681e-01 -9.40836787e-01 4.53153700e-01
3.31615090e-01 -6.82140663e-02 -3.97431999e-01 3.21517617e-01
6.27696753e-01 -4.85879660e-01 -1.14886796e+00 2.69137435e-02
3.51849645e-01 -8.69024456e-01 3.42543840e-01 -4.84441459e-01
3.42269421e-01 -4.63040441e-01 2.01318294e-01 -1.80651760e+00
7.79750943e-02 -6.56326234e-01 -3.90230238e-01 1.16712928e+00
5.04157087e-03 -9.90904570e-01 7.32550740e-01 2.52836645e-01
-1.53037414e-01 -8.14157486e-01 -1.14524055e+00 -7.98403621e-01
-2.34410763e-01 1.03213280e-01 5.54891944e-01 7.44053006e-01
8.09101388e-02 3.76836538e-01 -5.96716344e-01 3.28709364e-01
8.70595455e-01 4.08735603e-01 1.18886089e+00 -8.64938378e-01
-9.76881564e-01 -6.09784201e-02 5.60507104e-02 -1.07377791e+00
2.58446485e-01 -5.86513579e-01 2.05500662e-01 -1.23620749e+00
1.29219830e-01 -4.81865495e-01 -3.37671638e-01 7.52112329e-01
-1.64523110e-01 -4.06213939e-01 5.03760636e-01 1.34485811e-01
-1.07547629e+00 1.10674453e+00 1.45704091e+00 -4.99060340e-02
-3.67841333e-01 1.51078478e-01 -1.31133422e-01 4.89836782e-01
9.44935679e-01 -7.33961761e-01 -6.51736200e-01 -4.18600619e-01
1.43548355e-01 6.07415676e-01 1.76124334e-01 -8.63580942e-01
5.13990581e-01 -5.93590915e-01 -4.82469099e-03 -3.59469086e-01
3.05683613e-01 -8.82871032e-01 -1.74467966e-01 8.87376845e-01
-2.86857098e-01 1.42019570e-01 -1.30762815e-01 6.52995229e-01
-3.11913282e-01 -2.18485132e-01 6.04987919e-01 -1.83956444e-01
-3.00301611e-01 3.70985210e-01 -5.49701035e-01 1.83382891e-02
1.06198847e+00 2.28272304e-01 -3.18419218e-01 -3.80164593e-01
-5.06905138e-01 1.07344198e+00 6.76545128e-02 1.97287202e-02
5.41342854e-01 -1.06459260e+00 -6.00225627e-01 7.35828131e-02
-2.52215326e-01 3.10161620e-01 2.77783811e-01 1.17312002e+00
-1.80611730e-01 4.10646759e-02 -3.77855241e-01 -5.38606346e-01
-1.12379277e+00 4.18519020e-01 4.10779208e-01 -7.41128802e-01
-5.25748253e-01 2.01535523e-01 1.16393320e-01 -7.17037499e-01
2.89449483e-01 -1.22772358e-01 -1.44836232e-01 -2.43166126e-02
2.49634057e-01 6.67084634e-01 -1.73937321e-01 9.55829918e-02
6.55183941e-02 3.54843847e-02 -1.35380253e-01 -3.21106434e-01
1.55960536e+00 -1.49642795e-01 -1.58741340e-01 5.43206990e-01
5.41354179e-01 -1.89949453e-01 -1.72648072e+00 -4.81848449e-01
-1.28165171e-01 -2.09832773e-01 4.34195176e-02 -6.34323418e-01
-1.26132846e+00 7.02487886e-01 1.78832531e-01 1.26688227e-01
9.78756309e-01 -3.70377481e-01 5.84929824e-01 7.42165446e-01
7.60248721e-01 -1.44034231e+00 3.68788451e-01 5.26020050e-01
5.09258270e-01 -1.16087520e+00 1.54771820e-01 1.19339615e-01
-8.41876149e-01 1.08477104e+00 1.04850101e+00 -3.48247975e-01
3.71718705e-01 2.33041644e-01 5.89778088e-02 -1.78315006e-02
-1.43956280e+00 -1.23035870e-01 -3.51409733e-01 2.47635886e-01
-1.34455904e-01 1.60235420e-01 -3.16173375e-01 4.47307616e-01
4.25458811e-02 -9.67474282e-02 6.12394691e-01 1.09610283e+00
-4.88845110e-01 -1.20983791e+00 -4.24406976e-01 3.42009991e-01
-1.96616501e-01 4.29416716e-01 8.57606083e-02 6.88937724e-01
-1.00418255e-02 9.54585135e-01 -1.75453320e-01 -3.20433117e-02
1.13829352e-01 -3.11938494e-01 2.81537443e-01 -5.71970284e-01
-1.00295770e+00 2.81936169e-01 -2.94550676e-02 -6.91195011e-01
-4.26722705e-01 -3.85885298e-01 -1.40379536e+00 -2.02644393e-01
-3.83115739e-01 7.25175679e-01 6.31545484e-01 1.03966141e+00
2.17752203e-01 7.20493913e-01 1.04859340e+00 -8.44040692e-01
-1.37344611e+00 -9.52938855e-01 -7.64750779e-01 -3.22867674e-03
4.47547585e-01 -8.71097147e-01 -2.73554355e-01 -3.15382659e-01]
|
[3.7441394329071045, 2.1084611415863037]
|
d1813e22-0de8-436e-be43-e2d89f6d99f2
|
focus-attention-promoting-faithfulness-and
|
2105.11921
| null |
https://arxiv.org/abs/2105.11921v1
|
https://arxiv.org/pdf/2105.11921v1.pdf
|
Focus Attention: Promoting Faithfulness and Diversity in Summarization
|
Professional summaries are written with document-level information, such as the theme of the document, in mind. This is in contrast with most seq2seq decoders which simultaneously learn to focus on salient content, while deciding what to generate, at each decoding step. With the motivation to narrow this gap, we introduce Focus Attention Mechanism, a simple yet effective method to encourage decoders to proactively generate tokens that are similar or topical to the input document. Further, we propose a Focus Sampling method to enable generation of diverse summaries, an area currently understudied in summarization. When evaluated on the BBC extreme summarization task, two state-of-the-art models augmented with Focus Attention generate summaries that are closer to the target and more faithful to their input documents, outperforming their vanilla counterparts on \rouge and multiple faithfulness measures. We also empirically demonstrate that Focus Sampling is more effective in generating diverse and faithful summaries than top-$k$ or nucleus sampling-based decoding methods.
|
['Ryan Mcdonald', 'Sascha Rothe', 'Joshua Maynez', 'Shashi Narayan', 'Rahul Aralikatte']
|
2021-05-25
| null |
https://aclanthology.org/2021.acl-long.474
|
https://aclanthology.org/2021.acl-long.474.pdf
|
acl-2021-5
|
['extreme-summarization']
|
['natural-language-processing']
|
[ 4.98221278e-01 5.67933679e-01 -2.97716141e-01 -3.15890282e-01
-1.58245325e+00 -6.24356627e-01 8.03038538e-01 3.50006402e-01
-2.57668942e-01 1.29898512e+00 1.46157670e+00 -1.35004520e-01
1.55067533e-01 -5.15770733e-01 -6.34844661e-01 -3.73621374e-01
2.63420314e-01 6.11913681e-01 -2.32460842e-01 -3.81972760e-01
1.04255128e+00 -2.25101411e-01 -1.14037788e+00 7.19156384e-01
1.39846277e+00 2.26926968e-01 4.63715047e-01 1.19864261e+00
-2.23706260e-01 8.12765300e-01 -1.10764110e+00 -5.08622169e-01
-3.96073431e-01 -1.11094093e+00 -9.20936108e-01 -1.02762900e-01
8.87464702e-01 -3.72171611e-01 -3.22514743e-01 9.15706158e-01
8.77401412e-01 2.79931635e-01 9.50267315e-01 -4.78714347e-01
-1.00809908e+00 1.50506425e+00 -5.20804584e-01 6.15646601e-01
7.72890985e-01 1.29046440e-01 1.38899279e+00 -6.55942738e-01
6.33123040e-01 1.30797029e+00 3.36401701e-01 8.93723071e-01
-1.06818545e+00 -3.79329175e-01 3.19795638e-01 -2.62754168e-02
-7.22432435e-01 -9.73549843e-01 5.38865328e-01 -5.36119863e-02
1.13483286e+00 5.79767644e-01 4.75480348e-01 1.46913135e+00
4.21526283e-01 1.07886815e+00 4.90008593e-01 -4.75675941e-01
2.00629294e-01 -1.70384660e-01 2.37504914e-01 2.75505155e-01
5.05342901e-01 -4.81062084e-01 -1.03513014e+00 -1.19907252e-01
8.44218060e-02 -5.26563287e-01 -6.50177717e-01 6.43285096e-01
-1.49288619e+00 8.20129395e-01 3.19598094e-02 1.17138147e-01
-7.52912700e-01 5.18906951e-01 4.77890909e-01 -1.03397809e-01
7.08885372e-01 9.47364569e-01 1.89965009e-03 -7.57484972e-01
-1.59027267e+00 5.36889493e-01 9.89143193e-01 1.21344292e+00
2.96652943e-01 3.72307271e-01 -1.21595812e+00 7.05057681e-01
1.63925946e-01 6.54575527e-01 6.17288411e-01 -9.23834383e-01
8.17442536e-01 2.91164458e-01 1.29874498e-01 -7.12656617e-01
1.25712320e-01 -6.99788153e-01 -7.99088955e-01 -5.06980121e-01
-1.28646567e-01 -4.13516223e-01 -8.39930177e-01 1.66541803e+00
-2.85224259e-01 -2.03107953e-01 4.02701885e-01 6.76236689e-01
1.16066766e+00 1.14476860e+00 -8.71491805e-02 -4.00702059e-01
1.25835240e+00 -1.10465300e+00 -9.95657444e-01 -5.73047757e-01
5.05010605e-01 -8.20997894e-01 1.09490228e+00 2.04173386e-01
-1.60143602e+00 -3.67529452e-01 -1.05981743e+00 -3.69765341e-01
2.49816000e-01 1.67055741e-01 4.02463585e-01 4.55817968e-01
-1.28340721e+00 5.53991795e-01 -5.01603186e-01 -2.74474949e-01
4.37750399e-01 -2.13862538e-01 2.11649761e-02 -2.02318933e-02
-1.05381858e+00 9.59288895e-01 5.51157475e-01 -1.79759845e-01
-1.06524742e+00 -7.89017737e-01 -8.88261735e-01 3.15668315e-01
1.05874129e-01 -1.02710903e+00 1.56372237e+00 -4.95324284e-01
-1.47463679e+00 4.53833163e-01 -6.17579818e-01 -5.88467062e-01
2.55280733e-01 -4.33278531e-01 -1.12917684e-01 2.51810580e-01
4.81291950e-01 9.34914112e-01 5.24618745e-01 -1.13978899e+00
-7.50323474e-01 1.54168367e-01 -6.16976731e-02 5.64933538e-01
-1.79279178e-01 -1.73480392e-01 -4.11420494e-01 -7.83570647e-01
-3.13889205e-01 -4.62846994e-01 -1.12142034e-01 -8.30661893e-01
-1.21355009e+00 -3.57303947e-01 2.62093246e-01 -8.79571617e-01
1.65212166e+00 -1.63839829e+00 3.55590165e-01 -2.28822842e-01
8.61222073e-02 1.63663536e-01 -2.28148684e-01 9.37892139e-01
4.23008919e-01 4.67950374e-01 -3.00819993e-01 -4.31648433e-01
2.42441952e-01 -1.16340786e-01 -6.43307626e-01 -6.76884595e-03
2.99911380e-01 1.05222321e+00 -1.20856988e+00 -6.04813337e-01
-4.17514622e-01 1.38918385e-01 -7.08836257e-01 1.23496935e-01
-4.83825088e-01 2.60028183e-01 -4.40033376e-01 2.63666242e-01
1.74825177e-01 -3.02950233e-01 -1.41049683e-01 1.44982055e-01
-1.36219501e-01 9.52927828e-01 -5.06076336e-01 1.86414373e+00
-2.93525040e-01 9.07911360e-01 -2.30083168e-01 -5.18509746e-01
8.37928891e-01 4.37212437e-01 -2.42978871e-01 -5.78795135e-01
-1.71620607e-01 2.64277697e-01 -7.27539742e-03 -2.67449975e-01
1.48562932e+00 -1.13201104e-01 -3.17146122e-01 7.33204603e-01
6.86207712e-02 -3.74997556e-01 7.71183789e-01 9.58928227e-01
1.00855935e+00 -2.61984728e-02 2.10066989e-01 -4.49359208e-01
1.64786637e-01 2.02698350e-01 2.97954082e-01 1.29926479e+00
1.99066982e-01 9.84515250e-01 7.89406478e-01 4.04791474e-01
-1.10487437e+00 -8.11217844e-01 2.52747506e-01 9.91675079e-01
-1.18886314e-01 -7.14317858e-01 -9.02351618e-01 -5.64383447e-01
-2.07484767e-01 1.69589484e+00 -6.25247240e-01 -3.08038294e-01
-6.11785650e-01 -4.27116513e-01 8.11836064e-01 4.47908193e-01
2.45653182e-01 -1.17088306e+00 -5.02215683e-01 3.81722361e-01
-7.13193417e-01 -6.28905118e-01 -1.12116301e+00 -5.96974082e-02
-8.52736175e-01 -5.49536407e-01 -1.14169228e+00 -5.67283869e-01
5.47416985e-01 1.22176379e-01 1.45603836e+00 -1.11393206e-01
2.37877220e-01 3.15651715e-01 -4.90277201e-01 -5.88937104e-01
-7.53155887e-01 6.77206933e-01 -4.14236665e-01 -4.91745949e-01
5.76431379e-02 -1.35663912e-01 -5.49918175e-01 -4.76131499e-01
-9.18178976e-01 4.02733266e-01 6.68213010e-01 7.51507223e-01
2.36020267e-01 -5.28787434e-01 1.16312563e+00 -1.07940876e+00
1.50681686e+00 -6.01017356e-01 2.63274103e-01 2.44640410e-01
-4.58151400e-01 2.42888287e-01 7.01055586e-01 -6.54780492e-02
-1.18010402e+00 -7.79018342e-01 -2.92506367e-01 2.62980193e-01
1.98861435e-01 8.02094698e-01 1.76456600e-01 8.64695489e-01
8.42361689e-01 8.08461845e-01 -3.59678976e-02 -1.98176160e-01
4.10460770e-01 7.02687562e-01 6.25067711e-01 -5.87836206e-01
2.25198567e-01 7.60028418e-03 -5.39022982e-01 -7.82738805e-01
-1.14548922e+00 -3.50628674e-01 -1.13128290e-01 -1.95390388e-01
6.50713980e-01 -9.05758023e-01 -1.76609844e-01 1.35381535e-01
-1.51294088e+00 -2.24067092e-01 -4.81622994e-01 3.50873917e-01
-6.05443418e-01 3.67298245e-01 -5.87279081e-01 -7.66757548e-01
-9.70395267e-01 -9.37142968e-01 1.22889984e+00 5.34192741e-01
-8.01138878e-01 -1.00357187e+00 1.64486557e-01 3.24759394e-01
5.62566102e-01 1.10298209e-01 9.74722683e-01 -8.29593956e-01
-2.82748967e-01 -1.07196182e-01 -5.59597611e-02 1.81516096e-01
1.28595486e-01 1.38151363e-01 -6.68837070e-01 -3.11881304e-01
-2.99974173e-01 -2.67432004e-01 1.18433738e+00 5.58897018e-01
7.65470088e-01 -7.88550556e-01 -2.75545925e-01 1.05590284e-01
1.03148258e+00 -1.46663291e-02 7.45572865e-01 -3.02921850e-02
4.64002222e-01 4.71212506e-01 4.69059348e-01 6.89285517e-01
6.34915829e-01 2.53940225e-01 6.96334168e-02 2.42704377e-01
-5.28699875e-01 -5.66474676e-01 5.59416413e-01 1.30791891e+00
1.32187769e-01 -9.87435043e-01 -6.69960856e-01 1.00395703e+00
-1.67936444e+00 -1.28598511e+00 -1.92085296e-01 1.85569382e+00
1.33901727e+00 2.75085777e-01 -7.67702013e-02 -2.07467198e-01
8.06919158e-01 6.49177790e-01 -3.84687543e-01 -7.65356004e-01
-3.24192524e-01 5.83262965e-02 1.39610097e-01 7.61813700e-01
-4.44503933e-01 1.02379060e+00 7.05932188e+00 8.38439524e-01
-9.53408897e-01 -2.32997373e-01 7.51987219e-01 -5.07110536e-01
-1.10717773e+00 -1.10974297e-01 -1.13249421e+00 7.32150197e-01
1.25259769e+00 -6.90038860e-01 1.41163334e-01 4.36909735e-01
6.12851620e-01 -3.42987269e-01 -1.28045118e+00 4.43327934e-01
5.20228863e-01 -1.79966366e+00 5.60847938e-01 -1.73378095e-01
1.00484848e+00 -2.37896949e-01 2.91512907e-02 4.82160866e-01
5.46673715e-01 -1.15115047e+00 9.71113324e-01 7.39641070e-01
7.46186614e-01 -7.83934355e-01 5.62196195e-01 4.57058430e-01
-4.44325864e-01 4.00957257e-01 -5.29751956e-01 -5.16818240e-02
6.20405495e-01 7.07535923e-01 -1.04941261e+00 4.92495507e-01
2.21779823e-01 7.15192378e-01 -4.07439590e-01 9.26009417e-01
-4.28316742e-01 1.02269077e+00 2.01059833e-01 -6.49508059e-01
6.35779381e-01 6.22950755e-02 8.64723325e-01 1.79201925e+00
5.66079378e-01 1.17129542e-01 -1.96827963e-01 9.83940542e-01
-3.16270083e-01 4.26148027e-02 -4.68265563e-01 -3.69562894e-01
5.98302543e-01 7.87551820e-01 -3.17318648e-01 -7.51654744e-01
1.79747537e-01 9.80479479e-01 2.41138995e-01 4.39152569e-01
-5.95694184e-01 -7.78599858e-01 2.58159220e-01 -2.15771452e-01
3.98569733e-01 -7.02377632e-02 -5.39669514e-01 -1.13886356e+00
-5.76744638e-02 -1.12834704e+00 7.83961713e-02 -9.67495978e-01
-9.57689762e-01 6.31049633e-01 2.87255552e-02 -8.83673370e-01
-6.17966235e-01 1.81185409e-01 -1.03709638e+00 1.17533565e+00
-1.68001533e+00 -6.48945451e-01 4.61013243e-02 -4.83588912e-02
1.12580311e+00 -1.07766353e-01 6.77830994e-01 -1.72335595e-01
-4.39915717e-01 5.36085129e-01 2.49372885e-01 -1.38405249e-01
7.96923399e-01 -1.48607719e+00 6.43793464e-01 1.11310494e+00
-1.27163837e-02 8.70826364e-01 1.18079531e+00 -9.67062175e-01
-1.40780640e+00 -1.07085204e+00 1.47294867e+00 -3.71433139e-01
2.18742549e-01 -1.06039800e-01 -6.99678659e-01 5.43778718e-01
1.04194367e+00 -9.69405472e-01 8.15021098e-01 1.49464309e-01
2.15938166e-02 2.06080690e-01 -8.16642046e-01 9.61553931e-01
8.99728298e-01 -2.08221599e-01 -1.04324472e+00 3.75404328e-01
9.14220989e-01 -6.38849020e-01 -4.94559765e-01 -9.37761143e-02
2.59336621e-01 -8.62119377e-01 5.35494089e-01 -7.17890739e-01
1.28876293e+00 4.08442318e-02 -4.40320969e-02 -1.96484911e+00
-5.23109555e-01 -9.82739270e-01 -2.57279783e-01 1.45375443e+00
7.85280228e-01 -1.36302620e-01 7.31843531e-01 2.39036888e-01
-9.35769737e-01 -5.95697165e-01 -6.24217689e-01 -2.98607945e-01
3.37114692e-01 -3.19945328e-02 4.54408586e-01 4.13076848e-01
4.00216758e-01 7.88888752e-01 -4.98538643e-01 -2.95491815e-01
3.89544815e-01 6.89539015e-02 6.21092319e-01 -6.93013906e-01
-1.20801076e-01 -7.76809275e-01 3.49571288e-01 -1.56565571e+00
1.02988236e-01 -1.03130007e+00 5.88045418e-01 -2.41458511e+00
3.91078830e-01 9.07542259e-02 2.10808352e-01 1.38238430e-01
-6.90876305e-01 3.53037827e-02 1.78345785e-01 4.26804423e-02
-8.52741301e-01 8.39475930e-01 1.63886714e+00 -2.50055343e-01
-1.64719652e-02 -1.26890153e-01 -1.69153953e+00 9.42284167e-02
5.73028803e-01 -3.97882015e-01 -5.56982815e-01 -7.16036141e-01
2.88490593e-01 3.80456686e-01 -1.60716772e-01 -7.88272142e-01
4.18345243e-01 -1.22403391e-01 2.13036105e-01 -1.01677525e+00
1.27195984e-01 2.75703937e-01 -2.28643775e-01 3.83818984e-01
-1.16222501e+00 1.82583466e-01 2.76814073e-01 6.48469627e-01
-2.23719627e-01 -6.47985816e-01 4.02228981e-01 -4.04518127e-01
-2.42428958e-01 -8.45828354e-02 -8.11495304e-01 7.72154868e-01
4.49370861e-01 -3.07277083e-01 -7.42609441e-01 -8.98151815e-01
-1.34045035e-01 3.78471583e-01 2.21739650e-01 2.64302760e-01
7.70362496e-01 -1.13760245e+00 -1.54646087e+00 -3.32547069e-01
3.56327631e-02 1.16636381e-01 2.80026108e-01 5.55343509e-01
-5.57928324e-01 8.75143349e-01 1.44203126e-01 -3.73184055e-01
-8.74712884e-01 4.61068228e-02 -1.61994636e-01 -3.60259116e-01
-6.46673441e-01 9.80503976e-01 -6.28843978e-02 -6.85185194e-02
2.33153850e-01 -4.58372921e-01 -2.63309211e-01 2.39185706e-01
8.75045061e-01 4.87101674e-01 3.42662334e-02 -3.60128641e-01
-9.73381773e-02 -1.04135750e-02 -5.09504735e-01 -3.88477385e-01
1.09150910e+00 -1.91761747e-01 -1.07749477e-01 4.47292060e-01
8.10272634e-01 4.71641719e-01 -1.21031356e+00 -2.41908375e-02
-1.09098880e-02 -2.85099953e-01 1.97713710e-02 -1.13905168e+00
-5.28793275e-01 7.78369784e-01 -5.41392684e-01 4.25366551e-01
6.86781883e-01 -6.30991831e-02 9.02538180e-01 2.85213947e-01
-2.01276690e-01 -1.19230521e+00 3.25532049e-01 9.06074405e-01
1.24478257e+00 -9.74277616e-01 1.03503473e-01 3.33579443e-02
-1.05259776e+00 1.03254747e+00 4.93851960e-01 -1.19312063e-01
-1.92734987e-01 -8.54553469e-03 -1.59261972e-01 -8.09479803e-02
-1.27977991e+00 1.87570512e-01 3.75351101e-01 5.03047347e-01
9.56894577e-01 -1.04580455e-01 -5.57744801e-01 4.80786502e-01
-7.09778845e-01 -3.34109306e-01 1.10746133e+00 6.80937231e-01
-8.86453927e-01 -6.40179813e-01 -2.36089051e-01 8.21374714e-01
-7.13825881e-01 -7.23171651e-01 -4.80067968e-01 3.20028454e-01
-6.90781593e-01 1.28966546e+00 2.49344870e-01 9.77004617e-02
8.78183097e-02 2.43508015e-02 2.77561843e-01 -1.14767349e+00
-8.65159571e-01 1.07038662e-01 6.44027948e-01 5.81301935e-02
-1.89359523e-02 -8.27456713e-01 -1.07630098e+00 -5.21156728e-01
-2.42858335e-01 6.17994726e-01 4.40080434e-01 8.32189500e-01
6.47395313e-01 9.63378787e-01 3.27435613e-01 -7.60945559e-01
-7.63845921e-01 -1.38720644e+00 -3.68747324e-01 1.18616655e-01
4.51910287e-01 1.65719897e-01 -2.71564156e-01 7.06290007e-02]
|
[12.313438415527344, 9.334712028503418]
|
f1344e87-41ab-4a08-af54-114b891110cb
|
aspect-specific-context-modeling-for-aspect
|
2207.08099
| null |
https://arxiv.org/abs/2207.08099v1
|
https://arxiv.org/pdf/2207.08099v1.pdf
|
Aspect-specific Context Modeling for Aspect-based Sentiment Analysis
|
Aspect-based sentiment analysis (ABSA) aims at predicting sentiment polarity (SC) or extracting opinion span (OE) expressed towards a given aspect. Previous work in ABSA mostly relies on rather complicated aspect-specific feature induction. Recently, pretrained language models (PLMs), e.g., BERT, have been used as context modeling layers to simplify the feature induction structures and achieve state-of-the-art performance. However, such PLM-based context modeling can be not that aspect-specific. Therefore, a key question is left under-explored: how the aspect-specific context can be better modeled through PLMs? To answer the question, we attempt to enhance aspect-specific context modeling with PLM in a non-intrusive manner. We propose three aspect-specific input transformations, namely aspect companion, aspect prompt, and aspect marker. Informed by these transformations, non-intrusive aspect-specific PLMs can be achieved to promote the PLM to pay more attention to the aspect-specific context in a sentence. Additionally, we craft an adversarial benchmark for ABSA (advABSA) to see how aspect-specific modeling can impact model robustness. Extensive experimental results on standard and adversarial benchmarks for SC and OE demonstrate the effectiveness and robustness of the proposed method, yielding new state-of-the-art performance on OE and competitive performance on SC.
|
['Dawei Song', 'Bo Zhang', 'Chen Zhang', 'Fang Ma']
|
2022-07-17
| null | null | null | null |
['aspect-based-sentiment-analysis']
|
['natural-language-processing']
|
[ 1.92558557e-01 4.45438437e-02 -1.71135306e-01 -6.56626761e-01
-7.73909986e-01 -6.24185026e-01 8.43489587e-01 6.71697855e-02
-1.29049048e-01 3.08284163e-01 1.02493927e-01 -4.47757959e-01
2.25469738e-01 -1.00703180e+00 -5.68742156e-01 -5.62104046e-01
2.18858674e-01 8.79216790e-02 5.32749966e-02 -7.57516623e-01
2.97706783e-01 1.52075797e-01 -1.15746677e+00 5.79425633e-01
9.44421113e-01 1.11609876e+00 -3.31593215e-01 4.39964861e-01
-3.22993726e-01 7.37422645e-01 -7.89696813e-01 -8.98184299e-01
4.31101210e-03 -4.74072337e-01 -7.45691717e-01 1.42849460e-01
-6.56502740e-03 3.11587900e-01 3.28017592e-01 1.02081966e+00
2.82254875e-01 -5.84916957e-02 6.11749351e-01 -1.21289062e+00
-5.05818665e-01 5.92019677e-01 -7.93277621e-01 3.93027589e-02
3.98691624e-01 3.31443071e-01 1.46114004e+00 -9.60315406e-01
3.47191215e-01 1.27111077e+00 5.23722053e-01 4.57257748e-01
-9.46131647e-01 -5.48894405e-01 8.45178485e-01 2.11846888e-01
-9.44996119e-01 -2.22541504e-02 1.27614951e+00 -2.10795440e-02
1.01309550e+00 6.34271741e-01 8.42440844e-01 1.31138992e+00
5.35267830e-01 1.22172403e+00 1.49915552e+00 -2.09298715e-01
2.69309700e-01 5.51214755e-01 3.85770977e-01 4.92890328e-01
1.53498754e-01 -9.73855183e-02 -6.07056737e-01 -2.31198907e-01
-1.66980159e-02 -3.87122750e-01 -7.92195499e-02 -2.69420072e-02
-8.27135205e-01 9.85483229e-01 2.48295575e-01 2.01770380e-01
-3.54351670e-01 -3.87380749e-01 4.41316307e-01 4.28943902e-01
6.15416288e-01 7.45523870e-01 -8.82317841e-01 5.17012039e-03
-5.65490782e-01 2.34734252e-01 9.60788667e-01 8.68831038e-01
8.99373233e-01 3.16469461e-01 -3.97545695e-01 7.25390673e-01
3.60834390e-01 6.09686911e-01 4.23835933e-01 -2.15761840e-01
4.86394286e-01 1.17678320e+00 -3.75698179e-01 -9.93427157e-01
-3.71105045e-01 -6.60375834e-01 -7.43492424e-01 1.95388630e-01
-1.24767967e-01 -3.16267610e-01 -9.09014463e-01 1.66251743e+00
4.56245124e-01 5.79510666e-02 3.62029582e-01 7.47506499e-01
8.15193295e-01 6.67480886e-01 6.65049478e-02 -1.49900243e-01
1.74377227e+00 -1.09199119e+00 -5.88300645e-01 -8.59184146e-01
6.05726302e-01 -9.21964824e-01 1.46830642e+00 4.35519516e-01
-7.19393790e-01 -2.43016511e-01 -1.19846213e+00 1.17068328e-01
-5.79690158e-01 -9.24791023e-02 8.95822048e-01 6.21557117e-01
-5.21498382e-01 1.81097150e-01 -6.69852555e-01 6.47481605e-02
3.91313434e-01 2.73208261e-01 -3.69052947e-01 2.03485042e-03
-1.40327954e+00 6.68731034e-01 3.31052691e-02 8.55140537e-02
-7.95258939e-01 -7.00515151e-01 -1.18755805e+00 -1.06373519e-01
6.80266261e-01 -9.18635726e-01 1.02968574e+00 -1.34566772e+00
-1.56627357e+00 6.96644366e-01 -4.95180279e-01 -1.64633870e-01
1.03410222e-01 -3.51727426e-01 -6.67907178e-01 -9.55557376e-02
1.84654444e-02 2.57839739e-01 1.09237051e+00 -1.51681638e+00
-5.58807790e-01 -4.19059992e-01 7.44021356e-01 2.70977855e-01
-3.69957685e-01 2.29586065e-01 -4.97004211e-01 -8.97487819e-01
-2.61602044e-01 -1.08063138e+00 -3.66323203e-01 -5.13863087e-01
-7.07969010e-01 -8.93211290e-02 8.86724472e-01 -5.07730424e-01
1.30981338e+00 -2.11974716e+00 -9.68023837e-02 2.27417454e-01
-1.98862538e-01 4.48920429e-01 -3.81940901e-01 1.40259370e-01
-7.50217810e-02 2.19932809e-01 -5.16066968e-01 -4.63505447e-01
-9.74542201e-02 6.40197545e-02 -4.35774863e-01 -6.65103942e-02
7.47070670e-01 1.06918752e+00 -7.39891112e-01 -2.64860243e-01
-6.75051585e-02 6.36239231e-01 -8.49141181e-01 2.54062712e-01
-4.10711974e-01 3.41169804e-01 -6.02535903e-01 1.05895209e+00
5.47975123e-01 -1.80863834e-03 -2.90427264e-02 -3.44707429e-01
3.29593658e-01 5.01707196e-01 -9.65259373e-01 1.09653854e+00
-9.52914953e-01 3.66562784e-01 -1.09778009e-01 -7.29689717e-01
1.09725380e+00 1.22027569e-01 4.06239592e-02 -5.30869901e-01
3.25049937e-01 -1.42929509e-01 2.15516686e-01 -2.93229222e-01
6.42969489e-01 -3.94515604e-01 -4.61030453e-01 3.45253080e-01
-1.39261276e-01 -3.49169254e-01 1.35928720e-01 1.46581888e-01
7.47135758e-01 -3.47875990e-02 5.31523705e-01 -3.41752380e-01
1.13609970e+00 -6.70972094e-02 8.45696747e-01 3.01839948e-01
-1.51954457e-01 7.06421733e-01 8.13501596e-01 -3.94072503e-01
-4.19136405e-01 -6.33865058e-01 1.46851093e-01 1.06282187e+00
2.43696555e-01 -7.89046884e-01 -6.51737154e-01 -1.12726939e+00
-4.14155334e-01 1.06807256e+00 -9.52142417e-01 -4.37106013e-01
-6.20781481e-01 -1.15842271e+00 1.89778775e-01 4.43112582e-01
6.16015732e-01 -1.35968757e+00 -1.35269389e-01 5.23597002e-03
-1.09755293e-01 -1.21462321e+00 -5.96091032e-01 1.76611364e-01
-6.88062012e-01 -1.03348601e+00 -1.79314792e-01 -5.68677545e-01
7.02035666e-01 1.60397932e-01 1.32379293e+00 -5.80099784e-02
3.24787796e-01 2.26613209e-01 -5.07531703e-01 -7.38036573e-01
-3.55177760e-01 2.24030003e-01 -1.67298183e-01 5.08505225e-01
5.31651556e-01 -7.06628799e-01 -5.60142279e-01 3.63858789e-01
-1.13808227e+00 -1.47326410e-01 8.95792305e-01 7.81953156e-01
7.86195755e-01 3.55363414e-02 5.41260183e-01 -1.56518006e+00
8.65730286e-01 -4.36240762e-01 -3.38002712e-01 2.35992864e-01
-7.85910666e-01 -2.80407183e-02 9.88004208e-01 -5.29504001e-01
-1.09418690e+00 -2.74730682e-01 -4.00334597e-01 -7.68097565e-02
-1.71742663e-01 8.12518239e-01 -8.15081179e-01 8.40584561e-02
3.74595135e-01 3.29720914e-01 -2.90749520e-01 -1.31442957e-02
2.81736553e-01 4.42239821e-01 5.99052981e-02 -5.42141974e-01
1.00003648e+00 3.95002693e-01 -7.58057162e-02 -6.24628425e-01
-1.29620874e+00 -3.56989026e-01 -1.65162697e-01 1.30682632e-01
5.90795994e-01 -9.98537004e-01 -3.10250282e-01 4.13480967e-01
-9.55458581e-01 -9.69425142e-02 -1.91919804e-01 1.03898940e-03
-3.03368628e-01 1.87803015e-01 -4.13642526e-01 -6.68285131e-01
-8.82903278e-01 -1.44917929e+00 1.11857510e+00 3.74279797e-01
-4.24869865e-01 -1.06795776e+00 7.02631772e-02 6.58855855e-01
4.18858767e-01 1.81584597e-01 1.03180063e+00 -8.95059109e-01
-3.91007453e-01 -4.21739191e-01 1.43572211e-01 5.87946594e-01
2.67047465e-01 7.70220580e-03 -1.23699999e+00 -7.13451877e-02
3.23042274e-01 -7.34607652e-02 6.06981754e-01 -1.29563138e-02
9.03043807e-01 -5.56261718e-01 5.62640093e-02 7.09169090e-01
1.20620346e+00 6.19112104e-02 6.47699773e-01 6.97181702e-01
7.27296472e-01 3.48149925e-01 1.17278600e+00 2.13890076e-01
5.60820818e-01 4.78352517e-01 4.11371678e-01 1.77211929e-02
1.06387608e-01 -1.29178867e-01 7.89488554e-01 1.04855406e+00
7.25665092e-02 -8.32444131e-02 -5.80298543e-01 6.62161708e-01
-1.44037163e+00 -5.98122537e-01 1.18000666e-02 1.69593859e+00
1.05847204e+00 6.02430582e-01 -2.93945838e-02 3.25848162e-01
2.03148425e-01 7.03130007e-01 -5.03628850e-01 -8.69140446e-01
-3.73172641e-01 8.96599442e-02 -1.54273316e-01 6.23306632e-01
-1.24279165e+00 1.13970649e+00 5.02474165e+00 9.06886220e-01
-1.33991432e+00 -4.32431102e-02 8.45505476e-01 5.76974964e-03
-8.49260390e-01 2.42441863e-01 -9.03169692e-01 3.30258191e-01
6.42622352e-01 -2.22812340e-01 2.44201228e-01 1.25250697e+00
7.60920625e-03 7.09216669e-02 -8.52157652e-01 5.60451508e-01
2.55163670e-01 -9.43602443e-01 5.01147568e-01 -9.63110998e-02
8.95604670e-01 -3.43035370e-01 2.71601051e-01 8.79936814e-01
-4.00075465e-02 -7.86735892e-01 4.67792749e-01 2.06699982e-01
3.93461466e-01 -1.04459834e+00 1.09868431e+00 1.62767708e-01
-1.21697736e+00 3.94395292e-02 -2.49695718e-01 6.51458977e-03
4.19949517e-02 7.97010779e-01 -6.27074420e-01 8.12463582e-01
5.24454176e-01 6.25103652e-01 -8.14296782e-01 5.83478272e-01
-8.36351573e-01 8.95385981e-01 8.19718763e-02 -1.20183967e-01
3.19896460e-01 -2.61263728e-01 9.19556975e-01 1.21990108e+00
-1.18125349e-01 -4.53146454e-03 -1.53776035e-01 6.53910935e-01
-1.03715025e-01 4.25356686e-01 -5.57973027e-01 5.57945780e-02
1.14298694e-01 1.69376850e+00 -4.35737640e-01 -1.66455001e-01
-6.25917554e-01 8.64402950e-01 1.75075769e-01 3.07289183e-01
-7.69299507e-01 -3.82420957e-01 1.01781476e+00 1.09346965e-02
4.33009923e-01 2.29308948e-01 -6.32301152e-01 -1.34075177e+00
3.46541435e-01 -1.54714334e+00 4.61589247e-01 -5.55515707e-01
-1.34414768e+00 1.06767714e+00 -3.09983075e-01 -1.33760607e+00
-2.12737069e-01 -5.00477374e-01 -1.21388352e+00 7.81298280e-01
-1.94536984e+00 -1.54884017e+00 -1.14357412e-01 5.86520910e-01
6.97216570e-01 -2.65008777e-01 8.01351249e-01 -9.21808556e-02
-6.97291076e-01 1.01816487e+00 -7.78485656e-01 8.84428322e-02
6.61508143e-01 -1.33676648e+00 5.11986196e-01 1.07822227e+00
1.25576064e-01 8.99843931e-01 8.21954727e-01 -4.82462853e-01
-1.58340764e+00 -1.28454506e+00 9.08467770e-01 -7.36725748e-01
8.34161162e-01 -3.97736013e-01 -9.83402550e-01 6.45082057e-01
2.30559096e-01 -2.62963269e-02 8.42438281e-01 5.39992034e-01
-5.88522732e-01 -3.85851800e-01 -1.00936508e+00 9.03425813e-01
5.64143717e-01 -7.30354786e-01 -8.58440876e-01 -7.97598064e-02
9.96224463e-01 -1.12931676e-01 -6.08023286e-01 8.71455014e-01
2.57320136e-01 -9.72233236e-01 7.77946770e-01 -6.66893601e-01
7.24327803e-01 -3.78840476e-01 -1.47100016e-01 -1.47911429e+00
1.78477615e-02 -5.41127980e-01 -1.32239744e-01 1.61501598e+00
7.86544502e-01 -6.79518223e-01 6.46895170e-01 6.25691056e-01
-1.08769558e-01 -1.39198387e+00 -6.09074950e-01 -5.25108457e-01
1.89984128e-01 -8.10462058e-01 7.61854947e-01 9.29078341e-01
-1.41099971e-02 8.57174277e-01 -2.58116961e-01 4.15605307e-01
6.71077520e-02 5.74866235e-01 9.16881740e-01 -7.30168521e-01
-4.34178770e-01 -5.11806548e-01 -2.29744822e-01 -5.82457662e-01
3.10565323e-01 -6.70117617e-01 -1.69322416e-01 -1.22640133e+00
2.20523909e-01 -2.05198303e-01 -4.21847671e-01 5.98476827e-01
-1.02813888e+00 1.86533019e-01 2.53516108e-01 -2.64976561e-01
-7.29364276e-01 8.28414202e-01 1.38764620e+00 -5.35382390e-01
-2.20480934e-01 4.49558169e-01 -1.31113851e+00 9.74462748e-01
8.81407440e-01 -3.82499456e-01 -4.82857525e-01 -6.17509708e-02
5.48946738e-01 -3.19341660e-01 -3.07759121e-02 -6.77985311e-01
-1.09668046e-01 -1.87010050e-01 -8.97153765e-02 -5.00367761e-01
3.83554608e-01 -8.03287268e-01 -4.71227109e-01 1.07008889e-01
-9.46780220e-02 3.89951542e-02 4.18687344e-01 4.20137048e-01
-8.07733893e-01 -1.19966835e-01 5.35275757e-01 1.79948807e-02
-5.48474610e-01 3.64452988e-01 -1.94260761e-01 2.81474888e-01
7.32555032e-01 2.10646212e-01 -3.07777733e-01 -4.58797932e-01
-3.74788076e-01 1.09033972e-01 3.42357308e-01 6.24321461e-01
4.94553357e-01 -1.16808987e+00 -5.80658317e-01 3.36697578e-01
4.18900847e-01 8.71467292e-02 2.38070622e-01 8.21247756e-01
2.85996478e-02 3.66102666e-01 2.50691414e-01 -3.58382314e-01
-1.26462972e+00 6.29336834e-01 1.70310274e-01 -7.64256954e-01
-1.71736047e-01 8.04225504e-01 6.21079803e-01 -7.98471391e-01
-2.08001554e-01 -3.20586950e-01 -5.29687107e-01 6.20499179e-02
6.08090401e-01 -2.60148436e-01 1.10259108e-01 -6.45253658e-01
-4.72134650e-01 7.02665210e-01 -3.13348025e-01 1.24493428e-01
1.24279273e+00 -2.33931486e-02 -1.15260012e-01 4.40312564e-01
1.01230466e+00 5.44364333e-01 -9.64503527e-01 -1.29741028e-01
-1.22991070e-01 -1.97184294e-01 -3.02238390e-02 -1.10418808e+00
-1.30341315e+00 9.01370227e-01 1.38245463e-01 7.76916444e-02
1.45656276e+00 -1.41383901e-01 8.40536773e-01 3.41977924e-01
2.07862720e-01 -7.89838135e-01 7.60766864e-02 7.60829628e-01
1.14908206e+00 -1.49426734e+00 3.07926685e-02 -6.34316504e-01
-1.16005290e+00 6.35420382e-01 8.78808975e-01 -1.22535460e-01
6.60308421e-01 3.17833513e-01 5.45733154e-01 -2.24693790e-01
-9.45247889e-01 -6.31529242e-02 6.72814608e-01 3.80248904e-01
3.49599838e-01 1.63463667e-01 -2.42449358e-01 1.38765466e+00
-5.33744931e-01 -5.40549576e-01 3.93908620e-01 9.63383019e-01
3.67848165e-02 -1.27861834e+00 -2.26550370e-01 4.11199927e-01
-8.09696972e-01 -5.35538018e-01 -5.63687027e-01 7.56226361e-01
-1.62947148e-01 1.02393496e+00 -5.22969186e-01 -7.67183006e-01
5.55151522e-01 1.20099947e-01 -3.64524722e-02 -6.16166413e-01
-9.99349117e-01 -1.92814153e-02 3.90264660e-01 -6.77944779e-01
-2.17438072e-01 -6.58802271e-01 -9.43515539e-01 4.63210121e-02
-4.06662852e-01 1.19623512e-01 5.95507205e-01 1.29182315e+00
3.38156998e-01 6.64931059e-01 9.98889208e-01 -3.04518938e-01
-1.25743136e-01 -9.96436000e-01 -2.31188387e-01 4.76451546e-01
2.99216568e-01 -2.67465949e-01 -6.06228888e-01 -2.69565433e-01]
|
[11.445094108581543, 6.720444202423096]
|
600bfa5a-7c7c-46c4-acec-6f4adacef537
|
3dcrowdnet-2d-human-pose-guided3d-crowd-human
|
2104.07300
| null |
https://arxiv.org/abs/2104.07300v3
|
https://arxiv.org/pdf/2104.07300v3.pdf
|
Learning to Estimate Robust 3D Human Mesh from In-the-Wild Crowded Scenes
|
We consider the problem of recovering a single person's 3D human mesh from in-the-wild crowded scenes. While much progress has been in 3D human mesh estimation, existing methods struggle when test input has crowded scenes. The first reason for the failure is a domain gap between training and testing data. A motion capture dataset, which provides accurate 3D labels for training, lacks crowd data and impedes a network from learning crowded scene-robust image features of a target person. The second reason is a feature processing that spatially averages the feature map of a localized bounding box containing multiple people. Averaging the whole feature map makes a target person's feature indistinguishable from others. We present 3DCrowdNet that firstly explicitly targets in-the-wild crowded scenes and estimates a robust 3D human mesh by addressing the above issues. First, we leverage 2D human pose estimation that does not require a motion capture dataset with 3D labels for training and does not suffer from the domain gap. Second, we propose a joint-based regressor that distinguishes a target person's feature from others. Our joint-based regressor preserves the spatial activation of a target by sampling features from the target's joint locations and regresses human model parameters. As a result, 3DCrowdNet learns target-focused features and effectively excludes the irrelevant features of nearby persons. We conduct experiments on various benchmarks and prove the robustness of 3DCrowdNet to the in-the-wild crowded scenes both quantitatively and qualitatively. The code is available at https://github.com/hongsukchoi/3DCrowdNet_RELEASE.
|
['Kyoung Mu Lee', 'JoonKyu Park', 'Gyeongsik Moon', 'Hongsuk Choi']
|
2021-04-15
| null |
http://openaccess.thecvf.com//content/CVPR2022/html/Choi_Learning_To_Estimate_Robust_3D_Human_Mesh_From_In-the-Wild_Crowded_CVPR_2022_paper.html
|
http://openaccess.thecvf.com//content/CVPR2022/papers/Choi_Learning_To_Estimate_Robust_3D_Human_Mesh_From_In-the-Wild_Crowded_CVPR_2022_paper.pdf
|
cvpr-2022-1
|
['2d-human-pose-estimation', '3d-multi-person-pose-estimation']
|
['computer-vision', 'computer-vision']
|
[-3.18781734e-01 -2.22325742e-01 7.03812763e-02 -2.58207202e-01
-7.85297215e-01 -2.89512575e-01 3.49817544e-01 -1.92408919e-01
-6.09959602e-01 6.92540169e-01 3.18318456e-01 4.03513074e-01
2.32223243e-01 -6.66200757e-01 -8.49759638e-01 -6.26075208e-01
-1.61028076e-02 8.88555706e-01 5.56647480e-01 -1.51679337e-01
-1.99717239e-01 6.26350224e-01 -1.66687405e+00 1.84115306e-01
3.59418422e-01 9.48030651e-01 -3.07643972e-02 7.42537677e-01
2.41494969e-01 5.21395326e-01 -7.34800696e-01 -7.94924647e-02
7.73733974e-01 -3.29386324e-01 -5.72624385e-01 3.21597815e-01
1.16286147e+00 -4.68358278e-01 -6.21076763e-01 7.32197881e-01
7.63188064e-01 1.94765896e-01 8.50671232e-01 -1.50819969e+00
-1.99131131e-01 -2.75271177e-01 -7.22891808e-01 8.49378332e-02
6.41727507e-01 5.19627333e-01 6.17206633e-01 -1.24031699e+00
9.08340752e-01 1.44603324e+00 9.25894737e-01 8.04126263e-01
-1.10296571e+00 -6.29456043e-01 8.75466838e-02 -1.88246980e-01
-1.76440597e+00 -3.32954645e-01 7.41644382e-01 -9.79768574e-01
6.71182632e-01 1.99608997e-01 1.08432853e+00 1.23123300e+00
1.63799495e-01 9.59927619e-01 8.99947286e-01 -1.44370481e-01
1.07158929e-01 -6.01239577e-02 -1.38240874e-01 1.04151189e+00
4.08987403e-01 2.56996006e-01 -5.98954022e-01 -2.27211520e-01
1.10497963e+00 2.60385066e-01 -2.82522351e-01 -8.80888104e-01
-1.30388069e+00 5.65583169e-01 5.57437062e-01 -7.55109172e-03
-3.84164572e-01 1.65052652e-01 2.36590207e-01 3.72406133e-02
5.12424290e-01 2.06826940e-01 -2.82040030e-01 1.85949683e-01
-9.50551331e-01 8.25389266e-01 5.81855595e-01 9.14504588e-01
9.21705306e-01 -4.66812588e-02 -2.88439155e-01 5.58437824e-01
-1.05106227e-01 9.17716026e-01 9.38316137e-02 -1.14247477e+00
4.44502562e-01 7.29210079e-01 3.33925128e-01 -1.09902763e+00
-4.14482772e-01 -1.05717503e-01 -7.30575562e-01 6.01891458e-01
1.06466901e+00 -4.32472914e-01 -9.93894517e-01 1.65079427e+00
7.52816558e-01 8.13039765e-03 -3.41858715e-01 1.34326839e+00
9.99641597e-01 3.16476852e-01 -1.63663462e-01 3.05663258e-01
1.20892239e+00 -9.69930947e-01 -4.89812344e-02 -5.05643845e-01
4.27322805e-01 -4.17082220e-01 1.01755774e+00 1.08940296e-01
-1.00930619e+00 -6.78466797e-01 -7.20284224e-01 -3.05173453e-02
-3.21314782e-01 -3.58397700e-02 4.47371483e-01 2.70450503e-01
-6.99608862e-01 4.11262214e-01 -6.55276477e-01 -5.42791486e-01
5.99876583e-01 4.24048543e-01 -8.81599784e-01 -2.15444863e-01
-9.35523868e-01 8.33399057e-01 -8.77048895e-02 -7.30401650e-02
-9.30332422e-01 -6.91143692e-01 -1.14291763e+00 -4.86100376e-01
4.90960449e-01 -1.11382592e+00 9.95607018e-01 -9.02054429e-01
-9.46692884e-01 1.23362434e+00 -2.48430967e-01 -2.96776909e-02
1.09662211e+00 -3.98097068e-01 1.03698811e-02 1.55146539e-01
6.06356382e-01 8.77644479e-01 9.34099376e-01 -1.37762201e+00
-7.45136261e-01 -5.73019803e-01 -2.39615738e-01 2.29965329e-01
2.17077434e-01 -2.24141479e-01 -7.28095949e-01 -7.04175413e-01
3.29051428e-02 -9.55425203e-01 -2.94718981e-01 4.33492184e-01
-6.14451051e-01 -2.09844902e-01 6.17843747e-01 -3.49659234e-01
6.60018206e-01 -1.95804989e+00 8.53208154e-02 2.38082811e-01
5.26669681e-01 2.19711568e-02 -1.30520210e-01 1.40997633e-01
1.58693671e-01 -1.86501801e-01 -6.49439543e-02 -5.68595171e-01
-2.25040808e-01 7.52573460e-02 9.89806578e-02 1.00134599e+00
3.37706745e-01 9.15050149e-01 -8.43822837e-01 -8.21602762e-01
3.17254633e-01 5.68610251e-01 -6.04077220e-01 4.86197650e-01
1.68868341e-02 7.69972920e-01 -5.06516695e-01 9.11106348e-01
6.15825295e-01 -1.71894357e-01 -4.29609329e-01 -8.32343325e-02
1.41217392e-02 -4.02186781e-01 -1.39725184e+00 1.70183969e+00
9.72356573e-02 5.29469788e-01 5.85980453e-02 -5.07486820e-01
7.79586852e-01 9.08258483e-02 7.73203194e-01 -5.24555266e-01
2.37679556e-01 2.12800056e-01 -4.23774868e-01 -6.31418467e-01
2.85073727e-01 -9.25533623e-02 -2.49701262e-01 1.78003415e-01
4.71741101e-03 -5.53775905e-03 -5.26732989e-02 -5.17570190e-02
1.26253927e+00 3.07959557e-01 3.18923771e-01 -1.55232668e-01
1.99133769e-01 2.53729641e-01 7.95824170e-01 8.08325171e-01
-5.38345218e-01 1.05442619e+00 1.65188879e-01 -9.04838741e-01
-1.15465605e+00 -1.39034474e+00 1.23093769e-01 9.97139692e-01
4.20404434e-01 -1.44376695e-01 -7.08034217e-01 -9.38552797e-01
4.94939178e-01 -8.50166008e-02 -1.07364845e+00 1.99405745e-01
-8.89582813e-01 -4.45724763e-02 6.48708701e-01 6.32091522e-01
5.03291309e-01 -7.52229333e-01 -9.52985227e-01 -2.17964306e-01
-2.94480056e-01 -9.81557727e-01 -7.75171340e-01 -1.70248538e-01
-3.35745305e-01 -1.38049483e+00 -1.21975100e+00 -7.55477607e-01
9.43457842e-01 4.03313339e-01 1.13835776e+00 1.56171307e-01
-5.62481403e-01 5.72044849e-01 -8.45990703e-02 -4.14543718e-01
1.46363854e-01 -1.29540265e-01 4.03843641e-01 -1.10944472e-02
4.73479450e-01 -3.33060890e-01 -7.64772475e-01 6.51235700e-01
-2.08124608e-01 -5.48304766e-02 3.90874505e-01 7.27989674e-01
7.15927303e-01 -7.55849257e-02 1.80359811e-01 -7.12328315e-01
-5.49155939e-03 -4.56823260e-01 -3.42296153e-01 -7.39305690e-02
2.17593178e-01 -2.35446677e-01 2.51370013e-01 -6.62530422e-01
-7.24893630e-01 5.94055355e-01 2.35599414e-01 -7.62640536e-01
-5.64645827e-01 -3.32464129e-01 -1.30381301e-01 2.02792659e-02
1.06495488e+00 -2.09651560e-01 7.66362026e-02 -3.23949873e-01
-1.38473436e-01 3.57548237e-01 7.17349410e-01 -7.00339913e-01
1.09392846e+00 8.46241951e-01 5.99297546e-02 -9.31842327e-01
-8.84492338e-01 -7.08362818e-01 -8.10255051e-01 -6.05749667e-01
1.04012394e+00 -1.26098049e+00 -7.33217955e-01 4.31930959e-01
-1.17621982e+00 -4.88462240e-01 -5.60912430e-01 5.45482099e-01
-5.62738776e-01 8.25089365e-02 -4.24158722e-01 -7.42403209e-01
6.57609839e-04 -1.00699914e+00 1.36646533e+00 3.59231457e-02
-6.68292940e-01 -7.00615287e-01 2.33194858e-01 4.25250560e-01
-1.06172368e-01 8.30533743e-01 2.54792035e-01 -4.95088458e-01
-4.30633813e-01 -5.03466368e-01 1.19544053e-03 1.17521226e-01
-3.63884307e-02 -2.87868321e-01 -1.04409814e+00 -2.92553127e-01
-3.07862520e-01 -4.99282569e-01 7.80584693e-01 6.64776504e-01
6.50796473e-01 -1.22008584e-01 -5.14673829e-01 5.62203407e-01
1.05188668e+00 -4.96742904e-01 3.52773935e-01 2.27953196e-01
1.05767369e+00 7.87103176e-01 7.06687152e-01 4.55186278e-01
5.01367807e-01 7.19217002e-01 2.96560347e-01 -4.64518607e-01
-3.56769085e-01 -4.95174110e-01 2.41027310e-01 1.01929776e-01
-5.10384560e-01 9.85880010e-03 -1.12614655e+00 7.02749729e-01
-1.89355397e+00 -9.37405109e-01 -9.50573459e-02 2.24749422e+00
5.49172163e-01 -4.33861315e-02 6.75287843e-01 -8.86451975e-02
8.44319344e-01 -2.72055186e-04 -6.45616531e-01 4.71661240e-01
-1.46275133e-01 -4.68053408e-02 5.77631176e-01 5.57282031e-01
-1.34280336e+00 9.56763327e-01 5.53813171e+00 5.79763532e-01
-8.62867892e-01 -3.25265042e-02 3.13057095e-01 -4.10755515e-01
2.37443894e-01 -2.88527906e-01 -8.82330298e-01 2.63674527e-01
-3.72984679e-03 2.82672614e-01 6.21370599e-02 1.02920663e+00
2.27161884e-01 -2.59982377e-01 -1.38454950e+00 1.29523265e+00
1.84370294e-01 -1.05356240e+00 -1.81020442e-02 3.10055524e-01
8.24056864e-01 -7.59468675e-02 -1.13068826e-01 1.71031609e-01
3.11995029e-01 -1.20469630e+00 1.05636215e+00 7.11293221e-01
8.54361534e-01 -6.08452916e-01 7.78065979e-01 5.30470550e-01
-1.20832288e+00 1.10467523e-01 -3.53397191e-01 -3.09840024e-01
-3.52598564e-03 5.02096474e-01 -9.42946017e-01 1.04413480e-01
9.60920691e-01 5.68160713e-01 -6.36798203e-01 1.04741156e+00
1.02526685e-02 7.40231574e-02 -2.96206176e-01 1.46652699e-01
-7.50774816e-02 3.06175232e-01 7.51231790e-01 1.23313284e+00
2.76300907e-01 2.40257401e-02 7.50473320e-01 7.12140501e-01
4.84564267e-02 4.13288325e-02 -9.84812796e-01 5.82426071e-01
5.39082944e-01 8.86368275e-01 -5.50436616e-01 -1.82151362e-01
-1.60138085e-01 1.02461994e+00 4.46732819e-01 5.03543735e-01
-7.78211296e-01 -1.46104276e-01 8.83595049e-01 8.29931915e-01
2.64315903e-01 -3.13999265e-01 -1.29657075e-01 -1.20321274e+00
2.50102103e-01 -7.84379900e-01 3.30782920e-01 -5.98634541e-01
-1.42724550e+00 5.00154793e-01 8.84251446e-02 -1.42403769e+00
-3.19597244e-01 -6.08705342e-01 -4.15841967e-01 6.73479199e-01
-8.57357264e-01 -1.42515528e+00 -7.46580899e-01 9.75511730e-01
5.60426235e-01 -1.75224632e-01 5.75150311e-01 1.86478749e-01
-3.53416294e-01 5.55179834e-01 -5.48347652e-01 6.22591019e-01
9.43354726e-01 -1.13186991e+00 2.92906761e-01 6.84380114e-01
-1.56864807e-01 3.70467335e-01 6.69356644e-01 -1.00689530e+00
-1.16704512e+00 -1.15034866e+00 8.00306141e-01 -1.01420653e+00
5.57649843e-02 -5.70226431e-01 -6.48829043e-01 7.02281892e-01
-3.81617218e-01 5.63693702e-01 3.44753981e-01 -1.60570517e-01
-3.94210309e-01 -2.41465662e-02 -1.29778755e+00 5.94520926e-01
1.38012171e+00 -3.10680836e-01 -5.08576214e-01 2.17834562e-01
3.48120093e-01 -6.43283188e-01 -7.22529054e-01 4.30943191e-01
7.04308271e-01 -1.07617116e+00 1.18369067e+00 -6.36289239e-01
3.12692851e-01 -5.50195336e-01 -4.11313236e-01 -1.21029615e+00
-2.53233194e-01 -3.14505488e-01 -5.23320772e-02 7.95291662e-01
2.89113112e-02 -2.80402392e-01 1.18994784e+00 7.61817813e-01
2.33785495e-01 -7.73592234e-01 -1.01149464e+00 -9.21978831e-01
1.42420635e-01 -3.72274965e-01 4.59721982e-01 7.38308489e-01
-3.89793605e-01 1.88341647e-01 -5.41035831e-01 2.02682719e-01
9.36976016e-01 -1.64700836e-01 1.56762302e+00 -1.33824122e+00
-8.67645219e-02 -4.94290702e-02 -7.44623303e-01 -1.09221601e+00
1.41862735e-01 -5.46278000e-01 2.36691296e-01 -1.34034610e+00
2.52967983e-01 -4.01642442e-01 3.04647177e-01 4.82463539e-01
-8.79549682e-02 4.61400270e-01 1.98947310e-01 3.22331458e-01
-6.55659795e-01 2.98192859e-01 1.46680427e+00 -1.54555127e-01
-1.45747662e-01 -6.32325839e-03 -2.22858220e-01 1.07854557e+00
6.76807582e-01 -5.10908604e-01 -6.35847300e-02 -3.49860042e-01
-1.58972949e-01 -1.17687099e-01 1.03414690e+00 -1.34794343e+00
3.59743655e-01 -3.17320198e-01 9.70048010e-01 -6.01125836e-01
7.77190804e-01 -7.12216496e-01 1.66700929e-01 4.60838079e-01
3.39703001e-02 -9.58398283e-02 -1.42859697e-01 5.00715315e-01
1.77105621e-01 2.05918819e-01 8.88827085e-01 -5.19598007e-01
-7.91797698e-01 5.82335234e-01 -1.68198615e-01 5.71022570e-01
1.05271947e+00 -6.18364513e-01 -1.08837306e-01 -3.36954683e-01
-6.75008178e-01 2.62300670e-01 9.21838224e-01 2.55925208e-01
8.91445041e-01 -1.46268976e+00 -9.96622324e-01 3.97550344e-01
2.41899863e-01 4.17091876e-01 2.63372779e-01 7.03051805e-01
-5.21790385e-01 -1.55423153e-02 -2.97167659e-01 -8.83506954e-01
-1.45306373e+00 3.54991287e-01 6.14035428e-01 2.69813314e-02
-7.48444021e-01 8.43029678e-01 5.37250340e-01 -6.09629810e-01
3.60134214e-01 -1.33616030e-01 1.23941004e-01 -1.52673945e-01
5.20452797e-01 6.55595064e-01 -3.61757755e-01 -1.14753294e+00
-5.76217055e-01 9.82958674e-01 2.26195619e-01 -7.51742348e-02
1.14794087e+00 4.63996120e-02 2.48836979e-01 4.61673766e-01
1.23195350e+00 2.31496185e-01 -1.71986210e+00 -3.19471151e-01
-3.68364722e-01 -8.79724562e-01 -3.84145200e-01 -3.99531782e-01
-9.76956904e-01 6.80603027e-01 5.86046159e-01 -3.63945961e-01
5.93015432e-01 3.69216859e-01 7.11789906e-01 3.05066943e-01
6.09123707e-01 -1.21240485e+00 4.07228976e-01 4.76426333e-01
9.61863399e-01 -1.45373511e+00 1.52694538e-01 -4.09428895e-01
-7.55032599e-01 7.48316944e-01 1.04137981e+00 -4.31863725e-01
5.74112773e-01 2.44759396e-01 2.03529924e-01 -5.33714354e-01
-1.98839337e-01 -4.64190453e-01 4.05909836e-01 9.99363005e-01
3.23122069e-02 2.24005014e-01 3.55994344e-01 5.31377137e-01
-3.63341749e-01 -8.97150710e-02 7.15053454e-02 9.09674883e-01
-4.80315357e-01 -5.70073307e-01 -8.30772579e-01 2.62335241e-01
-1.60433263e-01 4.46476638e-01 -5.25609076e-01 1.11499405e+00
7.55436182e-01 7.14467943e-01 1.75939381e-01 -6.22279346e-01
7.92019367e-01 -1.74832135e-01 5.52385867e-01 -6.38244987e-01
-4.00560081e-01 4.62658191e-03 3.70876640e-02 -7.85419822e-01
-3.22650015e-01 -7.13845372e-01 -1.37005019e+00 -4.95499969e-01
1.57076158e-02 -1.10321492e-01 2.14030832e-01 7.80762970e-01
2.11891428e-01 7.10892454e-02 4.30471420e-01 -1.48231876e+00
-1.05798401e-01 -6.52924955e-01 -5.85827172e-01 6.40285432e-01
6.09920025e-01 -1.13663065e+00 -3.11838329e-01 -1.62445346e-03]
|
[7.049046516418457, -0.8750078678131104]
|
24bcebf7-7af8-4130-8e1e-adf591f3c696
|
joint-hierarchical-priors-and-adaptive
|
2307.02273
| null |
https://arxiv.org/abs/2307.02273v1
|
https://arxiv.org/pdf/2307.02273v1.pdf
|
Joint Hierarchical Priors and Adaptive Spatial Resolution for Efficient Neural Image Compression
|
Recently, the performance of neural image compression (NIC) has steadily improved thanks to the last line of study, reaching or outperforming state-of-the-art conventional codecs. Despite significant progress, current NIC methods still rely on ConvNet-based entropy coding, limited in modeling long-range dependencies due to their local connectivity and the increasing number of architectural biases and priors, resulting in complex underperforming models with high decoding latency. Motivated by the efficiency investigation of the Tranformer-based transform coding framework, namely SwinT-ChARM, we propose to enhance the latter, as first, with a more straightforward yet effective Tranformer-based channel-wise auto-regressive prior model, resulting in an absolute image compression transformer (ICT). Through the proposed ICT, we can capture both global and local contexts from the latent representations and better parameterize the distribution of the quantized latents. Further, we leverage a learnable scaling module with a sandwich ConvNeXt-based pre-/post-processor to accurately extract more compact latent codes while reconstructing higher-quality images. Extensive experimental results on benchmark datasets showed that the proposed framework significantly improves the trade-off between coding efficiency and decoder complexity over the versatile video coding (VVC) reference encoder (VTM-18.0) and the neural codec SwinT-ChARM. Moreover, we provide model scaling studies to verify the computational efficiency of our approach and conduct several objective and subjective analyses to bring to the fore the performance gap between the adaptive image compression transformer (AICT) and the neural codec SwinT-ChARM.
|
['Luce Morin', 'Wassim Hamidouche', 'Ahmed Ghorbel']
|
2023-07-05
| null | null | null | null |
['image-compression']
|
['computer-vision']
|
[ 4.94054407e-01 -1.05352208e-01 -8.31487551e-02 -2.22961962e-01
-5.09093583e-01 1.64389163e-02 5.39388299e-01 -2.45489538e-01
-6.09643221e-01 4.14224625e-01 3.32156003e-01 -3.11658561e-01
-2.44952127e-01 -5.82671106e-01 -9.25083041e-01 -7.91563332e-01
-1.80535674e-01 3.88250686e-02 -5.05615259e-03 9.79164317e-02
7.22999424e-02 1.02637559e-01 -1.50360525e+00 3.84875238e-01
8.94925594e-01 1.49991369e+00 6.50977850e-01 5.49662471e-01
2.58997262e-01 9.70505655e-01 -1.40050277e-01 -6.13323033e-01
3.08123410e-01 -4.13941473e-01 -3.73382866e-01 -4.40361649e-02
1.03722289e-01 -7.60427535e-01 -8.56279671e-01 1.09454823e+00
4.26195413e-01 -2.67291665e-01 4.49533105e-01 -8.87440681e-01
-5.38457930e-01 7.13749111e-01 -3.98843408e-01 4.74676006e-02
-1.27532065e-01 2.20226511e-01 9.18682098e-01 -8.08994055e-01
5.55894315e-01 1.03067720e+00 6.87794387e-01 2.99993604e-01
-1.21288347e+00 -9.27035987e-01 -6.42744154e-02 6.38812363e-01
-1.55504274e+00 -6.19455516e-01 6.53876126e-01 -3.62201966e-02
1.23359227e+00 5.15357628e-02 9.16910350e-01 1.31812525e+00
4.31761563e-01 6.95833743e-01 8.72059047e-01 -2.09761411e-01
4.54058945e-01 -1.83567628e-01 -4.14657265e-01 6.62746787e-01
2.58374393e-01 2.69120008e-01 -7.02165127e-01 3.20925355e-01
9.17737246e-01 1.16261197e-02 -6.08621299e-01 -3.52399617e-01
-9.29919243e-01 6.72926664e-01 5.29518127e-01 3.22189629e-01
-6.01047039e-01 5.31310380e-01 5.12586594e-01 1.57409295e-01
1.93417370e-01 -5.05108293e-03 -2.79226571e-01 -5.15599549e-01
-1.56135046e+00 -1.41562760e-01 6.94460154e-01 1.07999313e+00
5.09048343e-01 3.75379473e-01 -5.86194843e-02 7.52213776e-01
3.72797400e-01 4.66039807e-01 6.85443819e-01 -1.02905846e+00
7.20477879e-01 3.00836176e-01 -5.39918602e-01 -1.03026187e+00
-1.05373278e-01 -8.87166381e-01 -1.43159986e+00 -6.95946291e-02
-2.72350639e-01 2.17202723e-01 -8.14745665e-01 1.67528999e+00
-3.22054744e-01 3.03611457e-01 1.37884766e-01 7.18341172e-01
3.00766230e-01 7.28767991e-01 4.71357144e-02 -3.19862157e-01
1.22790551e+00 -8.44209671e-01 -7.69851148e-01 -1.61245629e-01
5.58510303e-01 -5.15551329e-01 6.43422723e-01 5.36246598e-01
-1.30258846e+00 -7.54447281e-01 -1.43866038e+00 -2.30458871e-01
9.14613679e-02 2.00645894e-01 5.56526601e-01 5.53097963e-01
-1.24677312e+00 7.81511307e-01 -1.21358728e+00 4.06633280e-02
6.86687589e-01 4.08970118e-01 -2.65219450e-01 -2.53581405e-01
-1.02501547e+00 6.35009229e-01 6.83421910e-01 5.22213504e-02
-9.55203950e-01 -7.38921165e-01 -8.07304263e-01 6.00206971e-01
2.54870430e-02 -7.55037189e-01 7.22488582e-01 -8.77087116e-01
-1.56429887e+00 2.13570073e-01 -7.09893331e-02 -1.12483609e+00
3.98206830e-01 -1.98843211e-01 -3.07369083e-01 6.56672597e-01
-3.44723940e-01 9.55193281e-01 9.95674789e-01 -9.68979955e-01
-4.06337619e-01 -3.69639955e-02 -2.72604585e-01 3.64174247e-02
-7.95156360e-01 -2.56466746e-01 -7.70703852e-01 -8.89789045e-01
1.00318648e-01 -8.21372688e-01 2.89058238e-02 1.03376642e-01
-9.59522277e-02 3.88342381e-01 1.01328492e+00 -8.71873617e-01
1.44683516e+00 -2.34148622e+00 2.94042498e-01 6.20077271e-03
3.80388618e-01 4.95947838e-01 -2.70831078e-01 2.82601357e-01
-2.15796337e-01 1.85909886e-02 -3.22128683e-01 -7.14843154e-01
-7.44414404e-02 3.80882353e-01 -3.61066818e-01 4.15677041e-01
1.94193348e-01 1.11940110e+00 -5.08050203e-01 -4.44808781e-01
4.19794619e-01 1.07009244e+00 -9.73053455e-01 1.77353337e-01
7.11196363e-02 1.39776006e-01 -6.80187419e-02 4.31048036e-01
7.67061591e-01 -5.38970828e-01 3.05650115e-01 -5.81152737e-01
-9.78249460e-02 2.90688992e-01 -8.38889778e-01 1.92384815e+00
-6.84013367e-01 8.25108767e-01 1.50800467e-01 -1.19970679e+00
7.19086826e-01 3.92619848e-01 5.35691023e-01 -1.13247240e+00
4.01611030e-01 3.58220518e-01 -1.07545415e-02 -2.60957122e-01
5.69099844e-01 1.63409710e-01 3.79980683e-01 1.54857431e-02
4.88942951e-01 2.40360409e-01 5.12614064e-02 3.13395083e-01
1.12530339e+00 1.08175322e-01 1.59819365e-01 -1.92794323e-01
4.32603478e-01 -7.91237056e-01 4.53000695e-01 5.93799770e-01
-1.01005837e-01 6.64677501e-01 3.50419134e-01 -1.71367317e-01
-1.28014481e+00 -8.81909907e-01 -3.15811634e-01 4.88073438e-01
1.08277090e-01 -7.46798754e-01 -8.35911453e-01 -1.89429700e-01
-5.61102688e-01 5.34832060e-01 -3.12058449e-01 -3.83160651e-01
-6.24520242e-01 -5.36872685e-01 7.27017879e-01 4.80735660e-01
1.04048932e+00 -5.73074579e-01 -1.02139258e+00 4.57977384e-01
-4.92045492e-01 -1.48688364e+00 -1.95973739e-01 4.80106324e-01
-1.03974915e+00 -5.96293330e-01 -7.23857760e-01 -3.71213466e-01
2.95642942e-01 -3.56776938e-02 8.86111438e-01 1.42718688e-01
-4.99268174e-02 9.86706540e-02 -4.49642688e-01 1.47247732e-01
-4.23986614e-01 1.01449393e-01 -2.07486346e-01 9.28057283e-02
-9.66984928e-02 -1.07396066e+00 -9.31933522e-01 7.12996647e-02
-1.22313440e+00 5.83904624e-01 1.01449013e+00 9.51486707e-01
4.95986044e-01 2.35674009e-01 1.70347303e-01 -3.00095439e-01
1.94702908e-01 -3.79062653e-01 -5.48994064e-01 1.11846194e-01
-9.79904115e-01 4.48700130e-01 7.98329830e-01 -3.49310249e-01
-9.53576028e-01 -3.81718874e-02 -4.27304685e-01 -7.22389162e-01
3.11244845e-01 6.28057897e-01 -2.16572806e-01 -1.30826861e-01
1.82609141e-01 7.71197259e-01 -1.33546650e-01 -2.18325227e-01
4.45504533e-03 5.73615313e-01 7.38187134e-01 -3.36577326e-01
7.13688731e-01 4.36532974e-01 1.14306733e-01 -6.60172880e-01
-3.44346523e-01 -6.84257969e-02 -3.76595616e-01 -1.65779218e-01
8.47086787e-01 -1.39591444e+00 -6.35589004e-01 5.58383226e-01
-1.20751429e+00 -2.69392520e-01 -8.92369449e-02 6.70694888e-01
-6.50030494e-01 6.33052111e-01 -8.98382604e-01 -5.62331080e-01
-4.78694677e-01 -1.54080772e+00 1.04438972e+00 -1.50229827e-01
1.98700354e-01 -8.72619271e-01 -4.17928308e-01 2.69724369e-01
8.46337259e-01 1.08729042e-02 9.61983621e-01 -1.44405633e-01
-1.05060971e+00 1.05673168e-03 -4.91781414e-01 6.81415498e-01
-4.65127468e-01 -3.71874392e-01 -9.99788821e-01 -5.02560377e-01
3.29013735e-01 -2.30815113e-01 1.24418688e+00 4.24457312e-01
1.46257830e+00 -4.34086651e-01 5.21705560e-02 1.43691087e+00
1.75108123e+00 1.00665823e-01 1.23861623e+00 1.53509185e-01
5.16032100e-01 -4.74433601e-02 9.29864589e-03 7.70704627e-01
3.69817168e-01 7.14602470e-01 7.31061637e-01 9.05111954e-02
-3.88047129e-01 -4.52019811e-01 5.15688479e-01 1.52461076e+00
-2.54945695e-01 -3.63522708e-01 -4.76994932e-01 1.43679902e-01
-1.55018187e+00 -8.70892286e-01 3.56687456e-01 2.00585151e+00
8.53848755e-01 2.08269313e-01 -6.26981676e-01 3.99513304e-01
2.32832164e-01 3.47078741e-01 -5.36025107e-01 -1.35484681e-01
-1.55339569e-01 3.84581029e-01 8.00613940e-01 2.57389188e-01
-9.24232483e-01 6.16090596e-01 5.33330393e+00 1.26020217e+00
-1.31053424e+00 1.01329625e-01 6.75108075e-01 -1.35638312e-01
-2.34983042e-01 7.02529028e-02 -7.04002380e-01 5.68843484e-01
1.49729764e+00 1.75026312e-01 9.25548255e-01 7.37255573e-01
1.20901585e-01 5.42353988e-02 -1.14289367e+00 1.34162891e+00
1.33977741e-01 -1.49277258e+00 2.11934015e-01 4.33150589e-01
5.90112686e-01 1.88154921e-01 2.66614705e-01 4.13150936e-01
-2.12371677e-01 -9.12953198e-01 1.17124808e+00 4.45947468e-01
1.23472106e+00 -5.82556903e-01 7.50740647e-01 2.38287345e-01
-1.46239781e+00 -4.00425524e-01 -4.94504184e-01 -6.13967739e-02
1.52670607e-01 6.68714345e-01 -3.66725117e-01 6.37580216e-01
8.17542672e-01 1.02872920e+00 -5.44809282e-01 7.87185788e-01
4.92064767e-02 7.29751766e-01 -2.88094908e-01 4.50119883e-01
2.02568173e-01 -5.08143082e-02 2.53956378e-01 1.20136428e+00
7.14204550e-01 1.67663035e-03 -4.16072726e-01 9.86078501e-01
-3.34074378e-01 -2.26940706e-01 -1.67208746e-01 -1.80364288e-02
4.42514330e-01 9.65333343e-01 -6.00435793e-01 -2.53081769e-01
-3.57295692e-01 1.18596888e+00 2.03367367e-01 4.92503315e-01
-1.13790870e+00 -1.08064987e-01 4.12364393e-01 2.47588232e-02
8.09475660e-01 -3.47502112e-01 -3.19819063e-01 -1.35873830e+00
1.37857005e-01 -9.26826954e-01 -1.39218614e-01 -7.00080156e-01
-5.81741333e-01 7.85206318e-01 4.05615121e-02 -1.35533583e+00
-2.85153359e-01 -4.70213562e-01 -4.18705530e-02 5.49918592e-01
-1.83519220e+00 -1.07864761e+00 -1.99534252e-01 6.23571277e-01
5.32388985e-01 -2.58478850e-01 7.03418434e-01 7.07943499e-01
-5.09375155e-01 9.79768097e-01 3.99140149e-01 -4.25653607e-02
1.71737030e-01 -5.33863008e-01 8.81170556e-02 1.16257715e+00
4.05651405e-02 6.41598642e-01 4.44417894e-01 -3.99826467e-01
-1.67326689e+00 -1.14901567e+00 4.78564203e-01 3.35160553e-01
4.64115947e-01 -5.89509249e-01 -8.08000028e-01 3.98719251e-01
3.56080949e-01 2.02166326e-02 3.00760955e-01 -4.49179500e-01
-5.62716901e-01 -2.97821373e-01 -8.24271858e-01 4.24129516e-01
9.90317643e-01 -6.79573536e-01 -6.91294298e-03 -1.87952682e-01
7.85486639e-01 -2.43336186e-01 -8.47684801e-01 5.20781577e-01
6.59706593e-01 -1.41735268e+00 1.07227314e+00 4.43401784e-01
1.04778850e+00 -6.08649617e-03 -5.54120421e-01 -8.70105267e-01
-3.20809543e-01 -4.63770241e-01 -5.30480802e-01 1.06926239e+00
5.42619415e-02 -3.94897819e-01 5.73780835e-01 4.41533506e-01
-3.54844332e-01 -1.02905929e+00 -1.33702981e+00 -7.69936383e-01
-2.10434198e-01 -8.32499206e-01 4.74245518e-01 4.11053121e-01
-2.44381785e-01 -1.16215479e-02 -7.66682446e-01 8.03831592e-02
5.88978469e-01 -3.70016456e-01 2.34341979e-01 -8.08485091e-01
-7.86105871e-01 -4.10297543e-01 -6.33820891e-01 -1.59529757e+00
-2.45471783e-02 -9.10838604e-01 -6.91338070e-03 -1.15059805e+00
3.79370570e-01 -1.95037290e-01 -3.50229651e-01 2.81038791e-01
2.56580889e-01 3.10331613e-01 4.59880531e-01 4.57173914e-01
-6.74103081e-01 1.10199797e+00 9.90219176e-01 -1.63193062e-01
2.01877818e-01 -6.59729362e-01 -4.66374010e-01 4.70153242e-01
3.54105532e-01 -4.07861769e-01 -5.27235270e-01 -7.25917697e-01
3.15531969e-01 2.48299703e-01 5.68770945e-01 -1.62529945e+00
3.81348729e-01 4.73835140e-01 3.59086305e-01 -4.74617541e-01
5.09851813e-01 -1.17352033e+00 2.66271055e-01 6.22208953e-01
-4.06036377e-01 -1.02597535e-01 9.04924944e-02 6.25456929e-01
-4.72513288e-01 -1.20182239e-01 8.99898589e-01 2.47886449e-01
-5.42607188e-01 4.07941222e-01 -2.59832978e-01 -2.00567052e-01
5.63903451e-01 -4.13205028e-01 -1.51340798e-01 -5.17102778e-01
-1.74430996e-01 -3.08432162e-01 3.67329329e-01 1.63646713e-01
9.80459750e-01 -1.14406931e+00 -6.12385213e-01 6.09836102e-01
-1.37527108e-01 -2.96641707e-01 5.11747658e-01 9.46126461e-01
-6.41161323e-01 6.49902821e-01 -2.25227758e-01 -6.69904709e-01
-9.22926128e-01 5.23321092e-01 1.41260922e-01 -5.83706558e-01
-8.25491130e-01 5.90471864e-01 2.68175364e-01 2.96566367e-01
3.31429809e-01 -5.31560302e-01 6.80988207e-02 -4.26535070e-01
4.47818071e-01 2.45443746e-01 9.61112827e-02 -6.02800310e-01
-5.69687784e-03 3.71731132e-01 2.10581888e-02 2.27219593e-02
1.53185415e+00 -3.77892047e-01 -8.71164072e-03 -7.91242495e-02
1.67281950e+00 -5.76367855e-01 -1.68160665e+00 -1.04189448e-01
-3.12644243e-01 -4.04556602e-01 5.32969773e-01 -5.03400207e-01
-1.52052176e+00 1.04966712e+00 9.86884594e-01 -2.91273981e-01
1.68644583e+00 -3.67260247e-01 1.05643451e+00 2.65821189e-01
4.73004639e-01 -9.41510558e-01 9.29049104e-02 5.17224550e-01
8.43109965e-01 -8.58468711e-01 1.33444309e-01 -1.49453029e-01
-3.21702749e-01 1.18918502e+00 1.10344805e-01 4.64312360e-02
7.21047342e-01 6.52563572e-01 -4.99422491e-01 7.69375786e-02
-1.01471543e+00 3.57404739e-01 1.81716949e-01 3.28061074e-01
2.71603078e-01 -1.61043659e-01 -1.21310100e-01 6.18274212e-01
-2.53816634e-01 2.96649039e-01 3.82478893e-01 5.78041852e-01
-1.92608908e-01 -9.59283710e-01 1.04350159e-02 3.89355153e-01
-3.29764247e-01 -6.01113617e-01 5.82884133e-01 5.86070359e-01
1.72452644e-01 7.67628670e-01 6.77062795e-02 -7.74137437e-01
-1.12879820e-01 -1.48299128e-01 3.32923383e-01 1.76596612e-01
-3.48811030e-01 3.41026336e-01 -3.25682402e-01 -8.67187858e-01
-4.18030828e-01 -5.06651163e-01 -1.00210690e+00 -3.28870863e-01
-3.38203847e-01 -2.83739567e-01 8.37702513e-01 8.39767396e-01
6.10309303e-01 7.35406458e-01 4.73659843e-01 -9.90018189e-01
-5.83932400e-01 -8.58442485e-01 -4.34882581e-01 2.52792299e-01
2.07247630e-01 -3.75511706e-01 -2.57965595e-01 3.22754979e-01]
|
[11.314224243164062, -1.5987895727157593]
|
816a0142-6c8f-405e-849e-ab9d95a0e1bb
|
cosda-continual-source-free-domain-adaptation
|
2304.06627
| null |
https://arxiv.org/abs/2304.06627v1
|
https://arxiv.org/pdf/2304.06627v1.pdf
|
CoSDA: Continual Source-Free Domain Adaptation
|
Without access to the source data, source-free domain adaptation (SFDA) transfers knowledge from a source-domain trained model to target domains. Recently, SFDA has gained popularity due to the need to protect the data privacy of the source domain, but it suffers from catastrophic forgetting on the source domain due to the lack of data. To systematically investigate the mechanism of catastrophic forgetting, we first reimplement previous SFDA approaches within a unified framework and evaluate them on four benchmarks. We observe that there is a trade-off between adaptation gain and forgetting loss, which motivates us to design a consistency regularization to mitigate forgetting. In particular, we propose a continual source-free domain adaptation approach named CoSDA, which employs a dual-speed optimized teacher-student model pair and is equipped with consistency learning capability. Our experiments demonstrate that CoSDA outperforms state-of-the-art approaches in continuous adaptation. Notably, our CoSDA can also be integrated with other SFDA methods to alleviate forgetting.
|
['Shuicheng Yan', 'Wei Chen', 'Minfeng Zhu', 'Chao Du', 'Tianyu Pang', 'Hesun Chen', 'Zhaorui Yang', 'Haozhe Feng']
|
2023-04-13
| null | null | null | null |
['source-free-domain-adaptation']
|
['computer-vision']
|
[-9.48347375e-02 -2.66722664e-02 -2.36111611e-01 -3.73418897e-01
-7.05794394e-01 -5.43573081e-01 5.38888514e-01 -5.07912189e-02
-3.85299027e-01 1.02109194e+00 1.02767929e-01 -2.22143799e-01
1.14748947e-01 -6.45471573e-01 -9.24192131e-01 -7.40137517e-01
5.31805694e-01 3.72586459e-01 3.68954629e-01 -1.99014425e-01
-2.84791529e-01 2.22093299e-01 -1.15140486e+00 -1.85732752e-01
1.41822064e+00 7.74489462e-01 7.61229992e-02 1.62570607e-02
-1.61305532e-01 4.89490390e-01 -7.94825852e-01 -6.47437632e-01
1.39001906e-01 -3.83127093e-01 -6.38008475e-01 6.98806047e-02
4.52898860e-01 -4.76800710e-01 -4.41997856e-01 9.95812476e-01
5.71189642e-01 2.68301785e-01 5.53220332e-01 -1.28668046e+00
-1.25929022e+00 3.44112039e-01 -3.43513787e-01 8.61127526e-02
2.48938967e-02 4.72428352e-02 6.22639120e-01 -1.00143421e+00
5.15286982e-01 1.08060086e+00 5.96408427e-01 8.73746753e-01
-1.32763481e+00 -9.51707244e-01 5.84582925e-01 -5.27989632e-03
-1.14365387e+00 -5.51468790e-01 8.20389330e-01 -2.90251493e-01
5.63798785e-01 -2.18765363e-01 3.71062875e-01 1.62002087e+00
-5.92202023e-02 9.42239881e-01 9.48583364e-01 -4.71390277e-01
5.63258588e-01 4.74313021e-01 1.71428397e-01 4.77467805e-01
4.66293186e-01 4.55866642e-02 -7.21576571e-01 -5.31703591e-01
6.22736454e-01 2.24297687e-01 -3.12342972e-01 -9.49430943e-01
-8.07960153e-01 7.71606803e-01 6.84217736e-02 3.29532996e-02
2.12995335e-02 -2.57513225e-01 3.60194862e-01 7.06759870e-01
7.08653569e-01 2.03229517e-01 -8.51139843e-01 1.55192256e-01
-6.06956244e-01 2.00911283e-01 7.72988617e-01 1.24354208e+00
7.61477590e-01 1.05757207e-01 -1.60128549e-01 9.59417164e-01
1.19349211e-01 4.32997912e-01 7.36183822e-01 -6.65628254e-01
4.52625722e-01 5.56827128e-01 2.16907725e-01 -3.60658526e-01
7.72097334e-02 -5.15145898e-01 -7.15404749e-01 1.70297459e-01
4.09675896e-01 -1.46417901e-01 -9.06652927e-01 2.17780828e+00
5.60326755e-01 4.77526695e-01 2.53972530e-01 7.48198867e-01
4.11378890e-01 2.61936367e-01 2.62298346e-01 -1.73739642e-01
1.01813865e+00 -1.11945391e+00 -7.39089012e-01 -1.83606520e-01
3.53586227e-01 -3.95325989e-01 1.61810672e+00 3.88161153e-01
-7.97265828e-01 -3.88567805e-01 -1.18835664e+00 -2.09798828e-01
-4.18771684e-01 5.76593429e-02 3.55298668e-01 6.49872541e-01
-6.04507148e-01 4.05618101e-01 -9.53280270e-01 -3.08385074e-01
6.25150740e-01 1.84206590e-01 -4.27693576e-01 -1.18992552e-01
-1.44391274e+00 7.82098293e-01 2.59360522e-01 -6.56327784e-01
-1.18236649e+00 -1.16597855e+00 -7.58526504e-01 1.18329629e-01
3.85684103e-01 -1.07241213e+00 1.41120291e+00 -9.73301411e-01
-1.88255608e+00 6.68184280e-01 -1.85173109e-01 -5.60985625e-01
6.81740105e-01 -5.24781883e-01 -7.00669050e-01 -2.75839865e-01
-4.68714088e-02 2.35163018e-01 1.31532979e+00 -1.16352916e+00
-6.32935464e-01 -3.44690651e-01 -9.63530242e-02 1.03170410e-01
-9.77436125e-01 -3.52944672e-01 -4.23958570e-01 -9.61646199e-01
-3.88581038e-01 -7.89225578e-01 8.52777511e-02 1.23875052e-01
-6.28017336e-02 -3.20352167e-01 1.06965232e+00 -4.42421854e-01
1.59977913e+00 -2.39184546e+00 8.77379626e-02 -1.76629007e-01
2.07419813e-01 6.70826375e-01 -2.80525476e-01 1.97955996e-01
8.43373984e-02 -2.13096648e-01 -6.14060700e-01 -7.60285437e-01
2.76014153e-02 4.04924065e-01 -7.94997454e-01 3.26586634e-01
2.05124512e-01 5.60639918e-01 -9.80246544e-01 -1.34016126e-01
-7.88826048e-02 5.90236723e-01 -7.28379250e-01 4.86611605e-01
-4.10899103e-01 4.92059886e-01 -4.31932688e-01 6.12524807e-01
1.01504540e+00 -2.72809446e-01 1.22410022e-01 3.12536597e-01
2.31619015e-01 5.23881078e-01 -8.62234592e-01 2.07369590e+00
-4.82901722e-01 1.62961721e-01 4.80919071e-02 -7.58113742e-01
1.07206798e+00 2.46987909e-01 1.95292681e-01 -5.94180644e-01
-2.92251468e-01 2.87018865e-01 -6.04781091e-01 -8.47356841e-02
3.20869654e-01 -1.34922251e-01 -8.75898376e-02 5.35409749e-01
4.16684479e-01 2.02438593e-01 -3.37585151e-01 2.34468475e-01
1.09263527e+00 1.93193913e-01 3.44131887e-01 -1.56299993e-01
5.60131192e-01 -1.27028421e-01 9.12879407e-01 7.15866327e-01
-3.60404521e-01 5.00105500e-01 1.77292958e-01 -4.74816531e-01
-8.58264685e-01 -1.37862563e+00 -5.65498788e-03 1.43615222e+00
9.60715488e-02 -2.53578246e-01 -8.08913589e-01 -1.31640124e+00
4.54969674e-01 8.87637317e-01 -6.05735242e-01 -6.49946630e-01
-4.64235693e-01 -5.02381504e-01 5.33606052e-01 4.20621186e-01
4.94501799e-01 -5.46282887e-01 -1.23717241e-01 2.52321988e-01
3.39766145e-02 -7.83122063e-01 -9.08523798e-01 1.77396238e-01
-1.00493205e+00 -8.58544827e-01 -7.69591391e-01 -6.45794392e-01
7.25465953e-01 5.13926208e-01 1.20744503e+00 -2.38377064e-01
2.64522403e-01 5.28262138e-01 -2.11899027e-01 -5.12256801e-01
-5.17082810e-01 6.56373262e-01 4.58775640e-01 -7.77813420e-02
7.83041716e-01 -9.68207955e-01 -4.34214979e-01 3.03090185e-01
-1.08569193e+00 -4.17421520e-01 4.80098933e-01 1.09960854e+00
5.92412591e-01 -1.85919568e-01 1.01799691e+00 -1.28718495e+00
8.02959859e-01 -7.58971751e-01 -8.14208984e-01 5.13562322e-01
-1.10624778e+00 2.80708939e-01 8.57387304e-01 -8.73565614e-01
-1.34139466e+00 1.75854623e-01 1.06175646e-01 -7.45256960e-01
-9.82723683e-02 1.23499177e-01 -5.50591648e-01 -5.98856993e-03
7.64026821e-01 3.49978536e-01 6.40192106e-02 -1.02651107e+00
3.91559124e-01 6.40567958e-01 6.33623242e-01 -8.40358675e-01
8.76005173e-01 4.28191662e-01 -6.09132648e-01 -4.30327892e-01
-1.16757154e+00 -2.68612027e-01 -5.27115405e-01 4.79107648e-01
2.62677014e-01 -1.37809646e+00 -2.34018967e-01 6.48773491e-01
-1.01967669e+00 -4.02357012e-01 -5.10673106e-01 2.63207257e-01
-4.58044887e-01 4.11576033e-01 -3.52252394e-01 -4.13793772e-01
-3.91293079e-01 -6.76241398e-01 6.55793846e-01 1.98463812e-01
5.34930825e-02 -1.14182615e+00 3.79192978e-01 6.70327768e-02
5.54269612e-01 -2.44850382e-01 9.87792373e-01 -9.84513342e-01
-2.85079628e-01 1.70336857e-01 3.33267003e-02 6.27070308e-01
4.18230325e-01 -4.74718720e-01 -1.03585911e+00 -6.62459433e-01
2.40646660e-01 -3.45858902e-01 9.93266702e-01 -8.03259537e-02
1.11393166e+00 -5.21340728e-01 -2.67826110e-01 8.01247180e-01
1.23442852e+00 -3.53707261e-02 4.15874958e-01 4.83329713e-01
5.09973407e-01 1.04057610e-01 5.38381457e-01 6.49858773e-01
6.26675785e-01 5.62877774e-01 2.15269104e-01 7.07355812e-02
-3.13260615e-01 -8.08455408e-01 6.91162646e-01 8.00918341e-01
6.03498280e-01 -1.19838901e-01 -5.82691908e-01 7.77645051e-01
-1.83714008e+00 -4.89514560e-01 4.18439776e-01 2.49340105e+00
1.36566484e+00 -7.44279125e-04 2.75610626e-01 -3.04663926e-01
5.39348841e-01 3.60886566e-02 -1.09682143e+00 -1.45454764e-01
-1.23627841e-01 1.78644419e-01 4.94923860e-01 4.41267371e-01
-1.13198316e+00 1.11589313e+00 6.61811829e+00 9.25221264e-01
-1.04654336e+00 4.80492234e-01 3.21728826e-01 -9.59142298e-02
-5.30641556e-01 -2.70513836e-02 -1.13014901e+00 6.97641790e-01
8.39670181e-01 -3.56789142e-01 5.65570176e-01 1.27999365e+00
-3.09588283e-01 3.22487235e-01 -1.07489824e+00 5.29470742e-01
-8.46713558e-02 -9.95463550e-01 1.74206227e-01 -1.66523725e-01
8.56170774e-01 -1.05009042e-01 2.57802308e-01 6.38011158e-01
7.84768581e-01 -5.29097557e-01 4.89431739e-01 2.75489032e-01
8.51434290e-01 -7.64101267e-01 2.87272453e-01 4.94835407e-01
-7.11623847e-01 -1.26235589e-01 -6.97515786e-01 8.97524878e-02
-1.46180168e-01 7.89797008e-01 -7.01830149e-01 4.21222359e-01
6.66452825e-01 6.69809878e-01 -5.37372470e-01 8.78494680e-01
-3.79036218e-01 8.74450982e-01 -2.34221563e-01 4.66011345e-01
-2.42390156e-01 -1.06877826e-01 4.97079015e-01 9.84832525e-01
6.04002595e-01 -5.94465099e-02 4.83943373e-02 9.12401497e-01
-3.53956729e-01 -1.68258667e-01 -5.75705051e-01 4.95577492e-02
1.08741891e+00 7.21564293e-01 3.60732764e-01 -3.09688210e-01
-6.57311440e-01 1.28504527e+00 8.71303201e-01 5.07876396e-01
-7.05617607e-01 -3.05016100e-01 1.25062442e+00 2.52646595e-01
4.90215659e-01 -3.19224656e-01 -2.48819143e-01 -1.61798120e+00
1.15778022e-01 -9.84952271e-01 7.21474826e-01 -2.77206123e-01
-1.75141251e+00 4.08036232e-01 -2.31196895e-01 -1.24016762e+00
5.02442606e-02 -8.04544464e-02 -4.85114783e-01 7.42408335e-01
-1.91934526e+00 -1.02333820e+00 -6.18524551e-02 1.08094597e+00
4.78355348e-01 -4.95447189e-01 9.54544723e-01 4.56227899e-01
-6.44675076e-01 1.32161498e+00 4.55981791e-01 -3.95576842e-02
1.37617266e+00 -1.09346521e+00 5.17596960e-01 8.78986776e-01
-8.16479251e-02 8.29095423e-01 4.99382377e-01 -6.45728111e-01
-1.31596029e+00 -1.39678717e+00 9.85355735e-01 -6.76670909e-01
5.39733052e-01 -4.61586624e-01 -1.47643971e+00 8.96836758e-01
9.70630273e-02 2.23759457e-01 7.55860806e-01 1.29243523e-01
-8.95281732e-01 -4.45837855e-01 -1.33950913e+00 4.06964570e-01
1.02187133e+00 -4.98811454e-01 -7.93827713e-01 -2.52072010e-02
1.26642263e+00 -4.23901349e-01 -8.73658538e-01 1.47175878e-01
2.24630445e-01 -8.15868318e-01 9.98757362e-01 -8.35538268e-01
8.14613029e-02 -3.20392460e-01 1.21778198e-01 -1.64402139e+00
-4.24501300e-01 -7.43722796e-01 -7.21559823e-01 1.66454017e+00
1.63010463e-01 -8.79288614e-01 7.17978299e-01 7.54194260e-01
-4.53348346e-02 -2.61824846e-01 -1.10573840e+00 -1.22905636e+00
5.22579014e-01 -8.09066743e-02 1.06283796e+00 1.16191185e+00
-2.15880021e-01 3.12378377e-01 -7.50193715e-01 4.35341984e-01
7.18127787e-01 -3.01173963e-02 9.28302228e-01 -1.30784953e+00
-3.38661939e-01 -7.75020719e-02 2.59496182e-01 -1.36685562e+00
3.52307260e-01 -6.65772498e-01 -2.18163073e-01 -8.30215454e-01
-6.06048629e-02 -6.74654126e-01 -7.87539661e-01 7.61402071e-01
-3.91152978e-01 -4.22101259e-01 -6.45690691e-03 3.44820917e-01
-6.55454695e-01 1.04806781e+00 1.07062125e+00 4.95820958e-03
-3.96340936e-01 1.57388538e-01 -1.02102768e+00 3.45927238e-01
7.00517356e-01 -7.64772117e-01 -6.26954138e-01 -7.48029828e-01
-1.33177996e-01 -3.21803868e-01 8.44789967e-02 -9.80190694e-01
2.94689596e-01 -3.52913111e-01 1.14382640e-01 -5.66958189e-02
4.56660688e-02 -1.05209553e+00 5.73058911e-02 1.78216845e-01
-3.65309179e-01 -2.20454052e-01 8.26950073e-02 8.85659397e-01
-2.95101225e-01 4.13205847e-03 1.07879364e+00 1.22550912e-01
-5.94948947e-01 4.38087732e-01 -8.46669525e-02 2.57064193e-01
8.84259105e-01 2.33256936e-01 -4.93266106e-01 -8.87462497e-02
-5.62280476e-01 2.36771926e-01 8.14513445e-01 5.87053239e-01
4.57770020e-01 -1.62546611e+00 -5.06200731e-01 6.75065279e-01
3.35260808e-01 1.53901726e-01 1.03904411e-01 3.95151705e-01
2.45755151e-01 2.09101081e-01 -9.61476490e-02 -2.09094226e-01
-8.68369401e-01 9.46906626e-01 1.18781134e-01 -3.51109624e-01
-5.88230550e-01 9.05903637e-01 3.91148239e-01 -7.25684822e-01
5.51143587e-01 -1.90106198e-01 2.96447664e-01 -1.96889892e-01
6.71713352e-01 3.53750437e-01 5.02647609e-02 3.92383561e-02
-3.41188908e-01 1.90483034e-01 -5.36229789e-01 1.56721875e-01
1.24926865e+00 -4.73945409e-01 3.14756215e-01 1.92519844e-01
8.92223775e-01 1.08316131e-01 -1.86310995e+00 -8.76916468e-01
7.66288489e-03 -6.18342936e-01 -1.99869797e-01 -1.04097915e+00
-1.00492024e+00 8.28187466e-01 4.87297714e-01 -1.87014908e-01
1.31733990e+00 -2.11466536e-01 1.16126084e+00 4.03214097e-01
4.02740479e-01 -1.31359935e+00 9.93359312e-02 6.67937815e-01
6.33401811e-01 -1.36435902e+00 -2.71375384e-02 -1.32666349e-01
-8.71190131e-01 7.30757654e-01 8.46922219e-01 -6.92991912e-02
6.86270237e-01 1.12082481e-01 1.67145953e-02 4.00550365e-01
-1.05113590e+00 1.15034178e-01 1.58666626e-01 9.16558146e-01
7.38865063e-02 -1.17439769e-01 -1.00332983e-01 1.27571511e+00
2.09332436e-01 4.02572215e-01 2.45090738e-01 8.45435858e-01
-4.13904160e-01 -1.59589827e+00 -1.97151795e-01 -6.70537204e-02
-3.16389740e-01 6.01227209e-02 -3.81965339e-01 6.30789459e-01
8.06956813e-02 6.56536043e-01 -2.96105713e-01 -2.70938039e-01
4.80184168e-01 4.08447921e-01 1.64354831e-01 -5.86625457e-01
-5.59060395e-01 -2.40537032e-01 -3.37975889e-01 -3.68158221e-01
-3.88450809e-02 -5.57636082e-01 -9.41895425e-01 -3.20749193e-01
-2.09690630e-01 2.81689942e-01 3.49034101e-01 6.62872910e-01
9.69659925e-01 2.05544412e-01 6.08182728e-01 -1.11045614e-01
-1.07855034e+00 -6.59834862e-01 -5.81757188e-01 4.47131991e-01
7.27957249e-01 -8.26144934e-01 -4.56462055e-01 7.35526010e-02]
|
[10.389276504516602, 3.2277822494506836]
|
c7b67039-8e59-4c67-93e8-21a7319a7c00
|
190600434
|
1906.00434
| null |
https://arxiv.org/abs/1906.00434v1
|
https://arxiv.org/pdf/1906.00434v1.pdf
|
Deep Unknown Intent Detection with Margin Loss
|
Identifying the unknown (novel) user intents that have never appeared in the training set is a challenging task in the dialogue system. In this paper, we present a two-stage method for detecting unknown intents. We use bidirectional long short-term memory (BiLSTM) network with the margin loss as the feature extractor. With margin loss, we can learn discriminative deep features by forcing the network to maximize inter-class variance and to minimize intra-class variance. Then, we feed the feature vectors to the density-based novelty detection algorithm, local outlier factor (LOF), to detect unknown intents. Experiments on two benchmark datasets show that our method can yield consistent improvements compared with the baseline methods.
|
['Ting-En Lin', 'Hua Xu']
|
2019-06-02
|
deep-unknown-intent-detection-with-margin
|
https://aclanthology.org/P19-1548
|
https://aclanthology.org/P19-1548.pdf
|
acl-2019-7
|
['open-intent-detection']
|
['natural-language-processing']
|
[-1.81744486e-01 1.91713236e-02 -3.49062532e-01 -7.45490074e-01
-7.30650902e-01 -3.05659503e-01 5.38827717e-01 -3.76285389e-02
-7.29750574e-01 8.28552365e-01 2.13153154e-01 -2.75472760e-01
2.22550020e-01 -2.82205820e-01 -5.84522247e-01 -3.00002247e-01
-1.76495999e-01 2.25094289e-01 5.15629388e-02 -1.09573379e-01
4.82305497e-01 -3.79384942e-02 -8.86335909e-01 6.25119507e-01
9.74146783e-01 1.09984934e+00 -9.71485898e-02 9.04338539e-01
-1.43313006e-01 1.00357270e+00 -8.54445517e-01 -2.41736785e-01
3.12325382e-03 -5.02278805e-01 -9.20523405e-01 -2.62395352e-01
2.84239054e-01 -9.13561285e-01 -6.43373370e-01 1.01084566e+00
4.88221318e-01 6.60322130e-01 6.21580124e-01 -1.34040487e+00
-9.30657029e-01 3.89939159e-01 -3.71996611e-01 5.58194518e-01
5.19847691e-01 1.33353382e-01 1.15542424e+00 -1.48381221e+00
1.02071248e-01 1.23733652e+00 7.31895328e-01 7.09703386e-01
-1.01027977e+00 -4.15756911e-01 3.10509503e-01 4.59839702e-01
-1.39048588e+00 -6.17953122e-01 7.43374527e-01 -1.24840520e-01
1.32356119e+00 1.11447483e-01 9.48627442e-02 1.50644112e+00
4.01982754e-01 1.31441689e+00 4.51654583e-01 -1.94513708e-01
3.37051183e-01 2.36064121e-01 3.85674030e-01 9.58274782e-01
-3.28537494e-01 -8.67676511e-02 -7.38910198e-01 -6.90573633e-01
1.06006928e-01 2.64561504e-01 -3.02078277e-01 2.07306683e-01
-7.21540689e-01 1.14285398e+00 3.51607114e-01 3.69842947e-01
-1.95089012e-01 -3.84845920e-02 5.96315324e-01 6.11931503e-01
7.73081839e-01 1.42735422e-01 -8.44663262e-01 -5.98201692e-01
-7.22696245e-01 1.83030710e-01 1.06375945e+00 7.88752258e-01
5.11556506e-01 -2.33468916e-02 -6.06805861e-01 1.00300789e+00
5.20343781e-01 9.98905227e-02 1.07073712e+00 -6.26666486e-01
5.02628207e-01 6.13726914e-01 1.03624769e-01 -1.03054833e+00
-3.86586040e-01 -2.75141805e-01 -8.72341216e-01 -1.74066320e-01
3.94249707e-01 -4.31265771e-01 -7.85979509e-01 1.61502171e+00
1.72392666e-01 3.42711240e-01 1.19439699e-01 7.62379348e-01
7.36436963e-01 7.89916337e-01 -3.38912398e-01 -1.78732902e-01
1.00457132e+00 -1.06149888e+00 -9.65077817e-01 -2.39908308e-01
8.74858499e-01 -7.02227056e-01 1.48107660e+00 3.25317770e-01
-6.16570055e-01 -3.14927816e-01 -7.94583261e-01 -1.05020173e-01
-3.26194763e-01 2.70641178e-01 1.87443778e-01 5.49327075e-01
-8.05259764e-01 6.60956025e-01 -6.48880661e-01 -1.25768065e-01
3.04604471e-01 2.50864476e-01 -9.96696204e-02 2.44786650e-01
-1.54976535e+00 7.34745920e-01 4.16529536e-01 2.35989034e-01
-7.40208685e-01 -5.21373570e-01 -1.05871093e+00 1.17883854e-01
4.16691452e-01 -5.33761144e-01 1.69476867e+00 -8.91077161e-01
-1.85822856e+00 4.24638212e-01 -4.73731518e-01 -7.33632684e-01
5.36062300e-01 -4.62993532e-01 -2.04458475e-01 -3.28065634e-01
7.19769299e-02 2.56176025e-01 9.21206951e-01 -7.69104123e-01
-7.76103258e-01 -2.65500307e-01 -4.20638993e-02 2.70456046e-01
-5.98042309e-01 -4.08451930e-02 -1.59660056e-01 -6.63516164e-01
-6.52332380e-02 -7.20234454e-01 7.11711720e-02 -1.44283354e-01
-6.64675593e-01 -8.37614655e-01 1.20549607e+00 -9.41913664e-01
1.40441513e+00 -2.22199082e+00 -2.21710205e-01 -4.02786843e-02
2.39821628e-01 1.97378784e-01 6.70858994e-02 2.09066927e-01
4.90523726e-01 6.03345670e-02 -1.26680538e-01 -6.43777072e-01
8.25351551e-02 2.23344266e-01 -5.56084037e-01 4.10277605e-01
2.97458738e-01 8.94242048e-01 -8.98143053e-01 -1.13606960e-01
1.32316165e-02 2.06899405e-01 -3.71481478e-01 5.57986259e-01
-3.04674149e-01 2.51803309e-01 -2.78056741e-01 4.80555207e-01
4.98507261e-01 -2.15710267e-01 -2.55196840e-01 8.67784470e-02
2.93622524e-01 7.14952052e-01 -8.26049089e-01 1.65524566e+00
-6.89553857e-01 6.86756611e-01 -1.73969135e-01 -7.01301813e-01
8.24065089e-01 2.40203068e-01 -7.30893984e-02 -4.99091089e-01
4.93845455e-02 2.33962730e-01 5.04939444e-02 -4.91431653e-01
6.37968421e-01 -3.98805216e-02 -1.94718242e-01 4.94529784e-01
4.38767731e-01 5.36876619e-01 -1.95266083e-01 2.84115851e-01
1.08314216e+00 -3.28070074e-01 1.12080567e-01 1.16323918e-01
4.83890831e-01 -5.47986448e-01 6.99071527e-01 1.10501242e+00
-5.01506448e-01 5.10231853e-01 6.47579432e-01 -6.28216028e-01
-7.66369939e-01 -9.87256587e-01 6.84166998e-02 1.45325124e+00
-1.39164358e-01 -2.47948989e-01 -3.35108995e-01 -1.39153886e+00
3.24314348e-02 1.00966334e+00 -5.23189127e-01 -6.08832121e-01
-4.45753425e-01 -6.22054517e-01 6.72466755e-01 4.82314527e-01
5.70822954e-01 -1.02435577e+00 -1.08920276e-01 2.21308351e-01
-2.92198420e-01 -7.64930129e-01 -9.84863520e-01 1.97025225e-01
-6.41603887e-01 -7.74133027e-01 -7.21157193e-01 -7.11355865e-01
5.22529781e-01 8.88836849e-03 7.65611827e-01 -5.08491835e-03
-9.02596265e-02 -2.84237824e-02 -2.41897523e-01 -8.78928751e-02
-1.86562672e-01 2.21626073e-01 3.88288677e-01 1.32915571e-01
6.91124201e-01 -1.03700496e-01 -3.96980047e-01 1.20972194e-01
-4.54940915e-01 -4.35963690e-01 2.00896829e-01 1.35936379e+00
9.99392048e-02 -1.03212491e-01 8.56225073e-01 -6.34953439e-01
1.18700957e+00 -7.37121224e-01 -1.44984171e-01 8.48024860e-02
-5.83215535e-01 1.62715092e-01 1.03090549e+00 -5.16190410e-01
-1.03849697e+00 -2.50685543e-01 -3.50383461e-01 -3.92435580e-01
-1.34531379e-01 5.68353057e-01 -1.12209976e-01 2.60039955e-01
2.90018380e-01 4.76453274e-01 -1.89170197e-01 -6.93276942e-01
3.12194765e-01 1.10959244e+00 2.25750580e-01 -1.18184514e-01
2.99386919e-01 1.30646393e-01 -9.21483934e-01 -9.64177072e-01
-1.18166625e+00 -5.75999260e-01 -5.26082575e-01 -1.59536265e-02
4.89191204e-01 -7.45907903e-01 -7.88118899e-01 6.40831470e-01
-1.50981772e+00 -2.36278772e-01 -2.07547441e-01 5.91041744e-01
-2.17515886e-01 4.93400246e-01 -8.01611245e-01 -1.13241220e+00
-6.02176189e-01 -8.34474146e-01 7.55445123e-01 4.37865108e-01
-5.45839667e-01 -1.21320438e+00 1.08601823e-01 1.28348798e-01
5.00358939e-01 -3.34356427e-01 6.83783352e-01 -1.57410192e+00
-6.26177043e-02 -4.06242907e-01 -1.29449561e-01 7.11806118e-01
4.05905515e-01 -2.51967728e-01 -1.03945422e+00 -4.43401307e-01
4.36147213e-01 -5.12634993e-01 1.10535610e+00 2.17835903e-01
1.12844789e+00 -9.61074293e-01 -2.18137845e-01 4.21368361e-01
6.35079920e-01 1.59732267e-01 1.99318528e-01 1.43724829e-01
6.99695468e-01 1.97904363e-01 5.70080638e-01 6.91649854e-01
6.85380161e-01 5.59595883e-01 7.05250651e-02 1.08689800e-01
4.38058108e-01 -4.38711703e-01 7.99014151e-01 8.64364564e-01
7.46312261e-01 -4.23387468e-01 -8.09580028e-01 4.88801718e-01
-2.17412782e+00 -9.48858321e-01 1.48750141e-01 2.06924438e+00
9.35590208e-01 4.47543800e-01 1.01734631e-01 -2.12560385e-01
7.52035201e-01 2.60048807e-01 -8.00542772e-01 -7.14615345e-01
1.02047451e-01 -3.99307549e-01 1.57038003e-01 7.67709672e-01
-1.28627193e+00 9.22651410e-01 6.25218821e+00 7.94946730e-01
-9.19206500e-01 2.88931757e-01 9.72552359e-01 -1.21443152e-01
-7.75664672e-02 -9.76821408e-02 -9.24030364e-01 7.89931118e-01
1.12699521e+00 -1.18182600e-01 2.38945454e-01 1.10986328e+00
3.11376691e-01 -2.61900835e-02 -9.95962441e-01 8.43290567e-01
4.39190626e-01 -8.46155822e-01 -6.68454766e-02 -2.42802054e-01
4.19989169e-01 2.07263842e-01 3.93580720e-02 8.55784655e-01
3.00105333e-01 -8.79627407e-01 1.89526141e-01 7.50184655e-01
2.27609202e-01 -9.48367178e-01 1.09897232e+00 7.76296556e-01
-7.49559939e-01 -2.05155492e-01 -5.36856174e-01 -1.83152765e-01
1.36017740e-01 7.42629766e-01 -1.10971951e+00 5.92898354e-02
6.36695087e-01 6.90366030e-01 -3.38400781e-01 8.21424365e-01
-5.76159321e-02 7.38325417e-01 -3.67201716e-01 -4.46517676e-01
3.85108054e-01 -9.68602523e-02 7.20816135e-01 1.11014342e+00
2.28833873e-02 -4.12761927e-01 3.55487168e-01 1.00626636e+00
-5.02054691e-01 9.30174068e-02 -7.68516064e-01 -1.43464776e-02
3.09474766e-01 9.74378467e-01 -1.98544517e-01 -3.92140299e-01
-5.37670851e-01 1.63539958e+00 5.66444159e-01 2.78036058e-01
-7.42335320e-01 -8.40868413e-01 4.69585866e-01 -3.46628815e-01
1.27825052e-01 -1.00558929e-01 -2.00939551e-01 -1.50248206e+00
3.68649811e-01 -6.41884148e-01 6.23059094e-01 -2.45925322e-01
-1.80336976e+00 4.99967486e-01 -4.57632959e-01 -8.90123904e-01
-4.88681018e-01 -3.61838937e-01 -1.11009073e+00 9.34519589e-01
-1.42158663e+00 -8.44486237e-01 1.05637677e-01 3.54275465e-01
1.10472870e+00 -4.76443917e-01 6.09147787e-01 2.30202779e-01
-6.51164651e-01 1.00263667e+00 4.90072995e-01 3.97898823e-01
8.22411239e-01 -1.37844837e+00 6.20015979e-01 6.35732174e-01
1.15570724e-01 7.88442910e-01 4.41835761e-01 -6.77050173e-01
-9.97969031e-01 -1.17586970e+00 1.23594904e+00 -4.72017795e-01
6.17767990e-01 -6.65646911e-01 -1.25567400e+00 8.07508886e-01
1.68171749e-01 -1.13868350e-02 9.32073534e-01 2.12483078e-01
-3.02717447e-01 2.77733058e-01 -1.22243190e+00 5.06867707e-01
5.60990751e-01 -7.06295550e-01 -8.91852856e-01 4.36133355e-01
1.09263611e+00 -3.75262976e-01 -5.06017566e-01 3.53053510e-01
5.84968269e-01 -7.17605770e-01 5.90617776e-01 -1.01602554e+00
-3.19659971e-02 1.32564455e-01 -5.29755233e-03 -1.26776433e+00
-1.02183111e-01 -7.87421942e-01 -7.37532437e-01 1.11383522e+00
5.69280267e-01 -6.80652499e-01 7.21142650e-01 8.42182934e-01
-1.90557912e-01 -1.00261772e+00 -1.24436903e+00 -7.75219023e-01
-3.56221362e-03 -2.56069273e-01 8.07095990e-02 8.90597522e-01
2.06049949e-01 6.17168367e-01 -7.85804570e-01 -4.22035977e-02
1.95750028e-01 -1.72043711e-01 6.17783070e-01 -8.87677014e-01
-3.27372313e-01 -2.16787800e-01 -2.04417422e-01 -1.60218823e+00
5.06989062e-01 -7.41636753e-01 2.56676823e-01 -1.28454566e+00
2.54512966e-01 1.73042774e-01 -5.39234996e-01 5.83667159e-01
-5.89806199e-01 -1.17435694e-01 -1.59663916e-01 1.35043412e-01
-1.11878538e+00 1.29326236e+00 4.25148219e-01 -2.64198154e-01
-6.58220828e-01 4.14264113e-01 -3.61780196e-01 1.01407409e+00
1.03807306e+00 -8.02378595e-01 -5.48308864e-02 -1.29512608e-01
4.05117534e-02 -2.25985765e-01 2.30320841e-01 -7.31709540e-01
2.38541231e-01 -7.97199830e-03 4.24869686e-01 -9.26006019e-01
3.35288107e-01 -4.28360730e-01 -1.02409136e+00 6.81560218e-01
-7.40691006e-01 6.64645657e-02 1.05180055e-01 7.85665214e-01
-2.44518936e-01 -4.77317542e-01 6.31812036e-01 3.41767073e-02
-4.71323162e-01 1.71601236e-01 -7.71407425e-01 3.49850565e-01
6.78645909e-01 2.64625937e-01 -1.61585361e-01 -6.54717505e-01
-5.96591592e-01 5.28673649e-01 2.36787014e-02 8.27030182e-01
1.05706763e+00 -1.36264777e+00 -6.24121845e-01 3.28955114e-01
1.82692409e-01 -2.30663225e-01 2.71368504e-01 7.27399409e-01
-4.49263752e-02 3.23767096e-01 4.90576595e-01 -5.67324281e-01
-1.20306528e+00 2.40349650e-01 5.40488064e-01 -2.20908746e-01
-4.10788655e-01 1.14530289e+00 -1.13724232e-01 -1.09218562e+00
5.78426898e-01 -1.74826846e-01 -1.05694950e-01 -1.33990243e-01
6.24945581e-01 3.69974226e-01 5.12348162e-03 -5.56549489e-01
-5.05443037e-01 -2.88632333e-01 -7.05545843e-01 -5.03301322e-02
1.02662516e+00 -3.80248994e-01 5.65834455e-02 1.15037763e+00
1.61496568e+00 -3.16868842e-01 -1.09938633e+00 -5.42739987e-01
4.05563772e-01 -5.66301107e-01 -4.75930460e-02 -6.20589674e-01
-5.10259688e-01 9.02090251e-01 7.73595989e-01 2.69814759e-01
5.07881939e-01 -1.87141076e-01 1.28661251e+00 9.24266100e-01
-1.73417449e-01 -1.44051421e+00 4.63238329e-01 1.21171355e+00
7.88013339e-01 -1.58277380e+00 -4.95592803e-01 4.26896602e-01
-8.27087641e-01 1.09738576e+00 9.34077203e-01 -1.14139743e-01
6.85646713e-01 -9.63845104e-02 1.05333611e-01 8.27716440e-02
-1.14907944e+00 2.72741228e-01 4.10085171e-01 1.45944253e-01
5.11448205e-01 -3.38566840e-01 -2.25548953e-01 9.19400632e-01
5.88368252e-02 -1.57610744e-01 5.58933139e-01 8.34718704e-01
-6.45077825e-01 -7.48573959e-01 3.67436856e-02 9.79824066e-01
-6.46443963e-01 -3.76718819e-01 -5.59267581e-01 1.73698202e-01
-2.84641713e-01 1.08114028e+00 1.09312579e-01 -6.46213949e-01
3.13732475e-01 5.03768623e-01 -2.55184829e-01 -7.90636480e-01
-5.50805926e-01 -1.96548656e-01 -2.06353832e-02 -4.97647107e-01
3.30808192e-01 -1.92711338e-01 -1.31543934e+00 -1.82197597e-02
-8.72098804e-01 1.92424670e-01 5.31192541e-01 1.22583163e+00
7.11035848e-01 3.75931978e-01 1.04178596e+00 -4.61376518e-01
-1.12529194e+00 -1.36816144e+00 -3.63387048e-01 3.80706102e-01
8.20396602e-01 -4.84371305e-01 -7.85741448e-01 -2.48671904e-01]
|
[12.120235443115234, 7.560116767883301]
|
02537be3-bd08-4824-be9e-269f1f73892f
|
on-the-effects-of-video-grounding-on-language
| null | null |
https://aclanthology.org/2022.mmmpie-1.1
|
https://aclanthology.org/2022.mmmpie-1.1.pdf
|
On the Effects of Video Grounding on Language Models
|
Transformer-based models trained on text and vision modalities try to improve the performance on multimodal downstream tasks or tackle the problem Transformer-based models trained on text and vision modalities try to improve the performance on multimodal downstream tasks or tackle the problem of lack of grounding, e.g., addressing issues like models’ insufficient commonsense knowledge. While it is more straightforward to evaluate the effects of such models on multimodal tasks, such as visual question answering or image captioning, it is not as well-understood how these tasks affect the model itself, and its internal linguistic representations. In this work, we experiment with language models grounded in videos and measure the models’ performance on predicting masked words chosen based on their imageability. The results show that the smaller model benefits from video grounding in predicting highly imageable words, while the results for the larger model seem harder to interpret.of lack of grounding, e.g., addressing issues like models’ insufficient commonsense knowledge. While it is more straightforward to evaluate the effects of such models on multimodal tasks, such as visual question answering or image captioning, it is not as well-understood how these tasks affect the model itself, and its internal linguistic representations. In this work, we experiment with language models grounded in videos and measure the models’ performance on predicting masked words chosen based on their imageability. The results show that the smaller model benefits from video grounding in predicting highly imageable words, while the results for the larger model seem harder to interpret.
|
['Marco Kuhlmann', 'Ehsan Doostmohammadi']
| null | null | null | null |
mmmpie-coling-2022-10
|
['video-grounding']
|
['computer-vision']
|
[ 1.32464111e-01 3.98002118e-01 1.86495706e-01 -1.78180918e-01
-6.42467260e-01 -5.88159919e-01 9.42303717e-01 4.06959280e-02
-6.37878001e-01 5.13624012e-01 4.08411860e-01 -6.54194236e-01
7.32812285e-02 -8.05734873e-01 -9.33718026e-01 -6.35448575e-01
2.39231080e-01 3.32429022e-01 1.38332754e-01 -3.12183082e-01
2.47174114e-01 9.72556919e-02 -1.78957677e+00 7.83314645e-01
4.66853619e-01 9.83662903e-01 3.77360940e-01 6.26349747e-01
-2.90177017e-01 1.02341294e+00 -3.51171613e-01 -4.54008400e-01
-1.11006282e-01 -5.72398663e-01 -8.64094079e-01 2.54128903e-01
8.70305300e-01 -3.43701422e-01 -2.56388187e-01 1.06421816e+00
3.44858885e-01 -1.41369700e-01 8.19196701e-01 -1.40967011e+00
-1.07540369e+00 5.64972878e-01 4.76338305e-02 1.36185691e-01
5.93168914e-01 5.13860762e-01 1.06841862e+00 -7.70864964e-01
6.93278372e-01 1.67473650e+00 5.29665172e-01 7.09191382e-01
-1.40529299e+00 -3.55648905e-01 3.80689174e-01 2.06699073e-01
-1.20146930e+00 -4.53686386e-01 2.37581834e-01 -6.30195975e-01
9.82594192e-01 4.30293262e-01 4.99300897e-01 1.19093370e+00
8.63630623e-02 7.19149411e-01 1.34452033e+00 -3.95344079e-01
1.60680890e-01 4.89103228e-01 -1.14028517e-03 8.42618883e-01
2.12707333e-02 2.61494160e-01 -6.43517196e-01 2.47050911e-01
5.46934128e-01 -3.50853443e-01 -6.21673226e-01 -3.86825912e-02
-1.35817552e+00 9.06297982e-01 4.81291115e-01 6.33476675e-01
-3.77300620e-01 3.65367115e-01 2.31168941e-02 2.51356333e-01
2.64492303e-01 6.29702926e-01 -3.77880603e-01 2.44639851e-02
-8.89696002e-01 1.10168017e-01 3.88688207e-01 8.51337492e-01
9.56590056e-01 -1.99779093e-01 -5.68683028e-01 5.30337811e-01
4.12361592e-01 6.33407295e-01 4.29432333e-01 -1.13256931e+00
4.73598957e-01 6.59303844e-01 7.33313113e-02 -1.11699319e+00
-4.93128002e-01 4.25273292e-02 -5.34036100e-01 4.14393306e-01
9.70616579e-01 -5.27380146e-02 -1.05779004e+00 2.02510262e+00
-4.45280880e-01 -3.20211321e-01 2.25221172e-01 1.25217319e+00
1.35370457e+00 9.08921003e-01 6.80447519e-01 -1.08801365e-01
1.98303866e+00 -7.32070148e-01 -6.37682676e-01 -4.63130325e-01
9.92066324e-01 -8.34311962e-01 1.46456146e+00 1.16756566e-01
-1.27926481e+00 -7.27834344e-01 -7.38416195e-01 -3.42566937e-01
-5.05997419e-01 9.42220464e-02 3.46938193e-01 5.97405612e-01
-1.37577009e+00 3.74278903e-01 -2.94388443e-01 -7.55190134e-01
2.96105742e-01 1.29511446e-01 -6.20928884e-01 -1.83271617e-01
-1.27723932e+00 1.43576992e+00 1.91334113e-01 -6.64159581e-02
-8.20591688e-01 -4.58117008e-01 -1.08323503e+00 1.11636713e-01
2.09592536e-01 -1.01564991e+00 1.07597589e+00 -1.52250540e+00
-9.23625052e-01 1.18942237e+00 -2.08240181e-01 -4.73825604e-01
4.75559801e-01 2.45301515e-01 -1.18980981e-01 4.02143538e-01
1.44561201e-01 1.38006806e+00 6.68665707e-01 -1.43033576e+00
-5.54717779e-01 -3.80224198e-01 6.02664888e-01 3.25570494e-01
-3.84426951e-01 -1.70276240e-01 -1.91014200e-01 -3.18982035e-01
-2.60668397e-01 -9.16792750e-01 5.40033728e-03 2.36450747e-01
-1.82262167e-01 -2.94260979e-01 6.25255466e-01 -8.19535792e-01
9.30349112e-01 -2.09775114e+00 2.27958277e-01 -2.62082219e-01
2.53359079e-01 -1.20952753e-02 -5.84251761e-01 3.29984188e-01
-1.74763259e-02 6.54337287e-01 3.67214307e-02 -2.00041398e-01
1.27003361e-02 2.91648924e-01 -2.33823091e-01 8.92781019e-02
7.21787274e-01 1.04067004e+00 -6.89595759e-01 -5.48761189e-01
2.89847076e-01 5.18981516e-01 -5.04032433e-01 9.84564573e-02
-5.56201935e-01 4.24662322e-01 -1.60552710e-01 1.70389265e-01
3.94146085e-01 -1.99807242e-01 -1.82872981e-01 -7.26980448e-01
-1.51852831e-01 2.65165329e-01 -5.29428542e-01 1.42150784e+00
-5.75425804e-01 1.11823833e+00 -8.04907456e-02 -1.06675828e+00
4.59112763e-01 4.57849145e-01 -7.42013901e-02 -1.17130101e+00
4.39404435e-02 -1.40229957e-02 2.14318767e-01 -8.45152020e-01
2.94359297e-01 -8.16173494e-01 1.05354905e-01 4.20223564e-01
3.46343704e-02 -2.25964159e-01 1.51381209e-01 4.95551705e-01
9.96282458e-01 2.73149461e-01 -2.03811109e-01 -3.29500660e-02
7.26931632e-01 3.23971331e-01 -2.43787333e-01 6.43790066e-01
5.22729792e-02 6.76636279e-01 6.03662550e-01 -2.58607715e-01
-9.11203802e-01 -9.61465061e-01 -4.03522737e-02 1.22222745e+00
2.99730569e-01 -5.53366542e-01 -7.69007027e-01 -5.22206247e-01
-2.96828985e-01 1.22006750e+00 -6.95955634e-01 -4.19839352e-01
3.14089619e-02 -3.63657743e-01 5.80205560e-01 2.94072121e-01
4.44786519e-01 -1.29811704e+00 -8.14511180e-01 -1.01906955e-01
-7.15364814e-01 -1.55857098e+00 -1.69393077e-01 -2.52129853e-01
-8.42327058e-01 -8.00853372e-01 -8.78320456e-01 -5.97762585e-01
8.18585634e-01 2.35942766e-01 1.40989590e+00 4.57444996e-01
-1.06700078e-01 8.46109629e-01 -4.62313443e-01 -3.02285194e-01
-7.80186474e-01 -5.17310143e-01 -4.40403253e-01 1.47109091e-01
2.67331541e-01 2.21256673e-01 -5.45349419e-01 5.73251545e-01
-1.20706713e+00 3.21658462e-01 5.53168595e-01 5.88366091e-01
4.35732275e-01 -3.96792680e-01 1.97448567e-01 -2.73990214e-01
5.32884955e-01 -2.02679679e-01 -2.60387152e-01 4.75199789e-01
-3.74775469e-01 3.67143631e-01 1.57931194e-01 -6.26125515e-01
-8.97941232e-01 -2.62695581e-01 -1.86612699e-02 -4.02556360e-01
-3.80224913e-01 4.92727846e-01 -6.32299036e-02 2.21579954e-01
8.08890820e-01 1.49765313e-01 3.02263331e-02 -1.11778736e-01
6.00872815e-01 5.61811984e-01 1.60980940e-01 -5.67287445e-01
7.45146334e-01 4.88145322e-01 -2.46147290e-02 -7.94463933e-01
-9.81208324e-01 -2.27735326e-01 7.38309510e-03 -5.00946581e-01
1.58318913e+00 -9.57391262e-01 -5.80676138e-01 5.73704280e-02
-1.31918359e+00 -3.17516923e-01 -7.61401728e-02 4.00193125e-01
-9.43293452e-01 2.62658417e-01 -2.64451027e-01 -8.28770041e-01
-8.65904540e-02 -1.38037467e+00 1.30595922e+00 1.20418876e-01
-4.55543905e-01 -1.13782036e+00 -5.64628243e-01 9.45954442e-01
3.18000466e-01 1.13928458e-02 1.42683339e+00 -6.03353322e-01
-5.91118693e-01 5.90761863e-02 -5.79075456e-01 4.40717399e-01
-1.46153167e-01 -2.17212334e-01 -1.04126370e+00 4.55290042e-02
-1.48810998e-01 -5.19443274e-01 8.78004909e-01 1.08797558e-01
8.81982744e-01 -2.61214912e-01 -3.11962198e-02 9.94023867e-03
1.24703610e+00 -1.34335741e-01 9.80622649e-01 3.71732414e-01
3.73403013e-01 1.23312545e+00 6.47380173e-01 -3.73737931e-01
4.98577237e-01 8.77186477e-01 6.65114045e-01 -2.14994013e-01
-5.39471567e-01 -4.24546897e-01 6.42241895e-01 9.27979648e-02
-1.17302068e-01 -2.19821677e-01 -1.03088021e+00 5.67110181e-01
-1.80552721e+00 -1.04594374e+00 -5.44532180e-01 2.01762843e+00
5.88325679e-01 -7.31495693e-02 -7.31001515e-03 -2.32752129e-01
6.93015575e-01 -1.29693687e-01 -8.38727131e-02 -7.48719573e-01
-4.04359072e-01 -2.16252342e-01 -1.28044873e-01 5.23876786e-01
-6.44924164e-01 9.36049104e-01 4.49685812e+00 7.86519349e-01
-9.25930619e-01 3.34278226e-01 9.33772445e-01 1.79900993e-02
-6.23211086e-01 3.66666198e-01 -5.23056328e-01 4.08926398e-01
9.54654276e-01 2.77560651e-01 4.25690889e-01 2.74664193e-01
4.38603014e-01 -6.80381417e-01 -1.58748353e+00 1.16759813e+00
3.88660043e-01 -1.31894100e+00 4.58095431e-01 7.18792677e-02
4.55147624e-01 -3.24189246e-01 -7.57405534e-02 5.09246826e-01
-3.08366477e-01 -1.37144458e+00 1.21217096e+00 7.54892647e-01
6.74542725e-01 -1.75392523e-01 7.90238440e-01 3.84242833e-01
-6.96032286e-01 4.71047238e-02 -1.79985002e-01 -2.78923243e-01
1.82887211e-01 1.37287825e-01 -5.91758013e-01 9.01774019e-02
5.55410564e-01 9.33673084e-02 -8.08532417e-01 7.14263201e-01
-4.01826262e-01 4.48053211e-01 -1.21369343e-02 -1.65619180e-01
4.10189390e-01 2.19421349e-02 4.19161618e-01 1.27429557e+00
4.17629480e-01 1.97709724e-02 -8.68101716e-02 1.16921461e+00
6.97541833e-02 1.30433947e-01 -8.48177135e-01 -3.05349350e-01
-1.40916035e-01 1.09155917e+00 -5.13358533e-01 -4.75820541e-01
-5.68992913e-01 7.90510476e-01 1.03633240e-01 6.86891556e-01
-9.90756750e-01 6.74228892e-02 4.85063136e-01 5.89306235e-01
8.31743032e-02 8.49243999e-02 -2.73800343e-01 -1.11519969e+00
-9.52688232e-02 -9.81888413e-01 2.91138232e-01 -1.71634352e+00
-1.30281937e+00 4.29893494e-01 1.04534188e-02 -1.06667268e+00
-2.64631927e-01 -9.12645936e-01 -4.16050166e-01 1.01909268e+00
-1.43339443e+00 -1.35604763e+00 -3.00307840e-01 7.64820993e-01
4.36711729e-01 1.29057184e-01 6.91710949e-01 -1.49774015e-01
2.91688610e-02 2.66708910e-01 -6.45228565e-01 1.00753397e-01
6.82424009e-01 -8.18623364e-01 -2.77225673e-01 4.38535452e-01
3.42754126e-01 5.24814188e-01 8.59041691e-01 -2.56915659e-01
-1.30628872e+00 -7.51974881e-01 1.10966849e+00 -5.75164676e-01
6.77467585e-01 -7.69893527e-02 -8.26210737e-01 3.56655151e-01
4.20887321e-01 -3.75518590e-01 6.69261158e-01 -2.61382163e-02
-6.65114641e-01 1.69619665e-01 -1.15438879e+00 9.54687595e-01
1.00170803e+00 -8.44847918e-01 -9.69414413e-01 5.44850647e-01
4.00095075e-01 1.90776303e-01 -6.93090200e-01 4.04069245e-01
2.75140703e-01 -1.02717781e+00 9.41549778e-01 -9.23763216e-01
1.02683210e+00 -3.06657016e-01 -4.41467345e-01 -1.09135008e+00
-1.32632956e-01 7.54949898e-02 3.30808491e-01 1.18814528e+00
8.56492400e-01 -6.96112633e-01 4.02940959e-01 8.49661648e-01
1.84317663e-01 -5.05240917e-01 -9.63216066e-01 -7.04611361e-01
3.71383756e-01 -5.86001456e-01 -1.24025993e-01 8.88317406e-01
-7.15876073e-02 7.10271955e-01 -3.21529731e-02 6.20406270e-02
2.52652287e-01 1.36323750e-01 6.79233909e-01 -9.10579085e-01
-1.71315730e-01 -6.58079505e-01 -5.07480383e-01 -7.87887633e-01
2.70305455e-01 -8.18080783e-01 1.21542439e-01 -1.95073748e+00
3.23599905e-01 -7.70065784e-02 8.57656915e-03 7.31205583e-01
-1.44327544e-02 6.07636571e-01 8.45275283e-01 1.37547031e-01
-5.83364725e-01 1.63610145e-01 1.38811612e+00 -4.20925528e-01
3.38738054e-01 -4.88179535e-01 -8.77043605e-01 8.20568800e-01
5.43819427e-01 -3.51689070e-01 -4.44570810e-01 -8.07134688e-01
5.97571015e-01 3.81741047e-01 1.04819000e+00 -7.12389052e-01
-1.35216834e-02 -3.53737473e-02 3.93643409e-01 2.26898715e-02
4.84745800e-01 -6.65105641e-01 -1.72721326e-01 6.30391061e-01
-3.49706203e-01 -5.22090727e-03 6.03157222e-01 3.06349576e-01
-2.25372568e-01 -3.29782367e-01 5.30100584e-01 -2.88432688e-01
-5.65271795e-01 -2.13213205e-01 -8.20455313e-01 1.03846490e-01
8.14091623e-01 -4.58499223e-01 -4.76346076e-01 -8.55554521e-01
-1.14515591e+00 1.59433082e-01 6.43090129e-01 8.55033219e-01
6.57899082e-01 -1.25756764e+00 -7.98251867e-01 -8.73310640e-02
4.71787840e-01 -4.18322682e-01 8.92739967e-02 1.18642497e+00
-3.05691600e-01 6.34768903e-01 -1.78381741e-01 -7.23927617e-01
-1.37628794e+00 5.86454630e-01 5.03658354e-01 6.70224950e-02
1.87821630e-02 8.46448541e-01 7.94994891e-01 -1.31438032e-01
4.49603535e-02 -4.70704764e-01 -2.28562400e-01 3.54977101e-01
3.64801288e-01 -4.73369211e-02 -2.07880631e-01 -9.39987481e-01
-3.48634124e-01 8.38276207e-01 3.45785648e-01 -3.73733133e-01
9.69677806e-01 -2.82626152e-01 -2.68246889e-01 4.13232148e-01
1.09653389e+00 -2.73760617e-01 -9.32789147e-01 1.64816171e-01
-1.55179039e-01 -2.49556154e-01 1.23769999e-01 -1.19650435e+00
-7.39203036e-01 1.39177561e+00 7.56943524e-01 2.71637648e-01
1.16349983e+00 3.95734638e-01 3.13118875e-01 2.53358036e-01
3.95092726e-01 -8.33543360e-01 3.09396893e-01 3.75454009e-01
1.22773528e+00 -1.45307887e+00 -3.01276207e-01 -2.67813802e-01
-8.10878098e-01 1.04793394e+00 6.62216365e-01 3.27617824e-01
3.81385870e-02 -3.00377935e-01 3.15893590e-01 -2.09803060e-01
-7.58937478e-01 -5.93360543e-01 4.96468723e-01 6.50383770e-01
5.93789458e-01 -1.34869525e-02 -2.91858137e-01 4.23252791e-01
-1.29820555e-01 -1.89816639e-01 4.06896591e-01 2.69135118e-01
-3.18106055e-01 -8.32243860e-01 -7.71950424e-01 2.78151065e-01
-1.97466493e-01 -3.48908603e-01 -8.16607118e-01 6.27527475e-01
4.96963292e-01 1.54960239e+00 2.45098263e-01 -2.13326797e-01
2.42262200e-01 3.04353118e-01 6.95177436e-01 -6.58956468e-01
-5.36778569e-01 -2.13626966e-01 6.58674598e-01 -4.00655508e-01
-5.43671250e-01 -6.17807984e-01 -1.26755774e+00 -2.37689018e-01
-3.50853413e-01 -4.09616716e-02 7.89382756e-01 1.34100616e+00
2.27541998e-01 2.87428021e-01 -8.74471515e-02 -8.18155944e-01
-4.26258862e-01 -7.95837998e-01 4.86229360e-02 8.72580111e-01
2.17559472e-01 -5.36085844e-01 -4.78505045e-01 2.16612473e-01]
|
[10.82286548614502, 1.6919746398925781]
|
6f9e18db-6c9e-429f-a652-8f8f4043764d
|
laughter-during-cooperative-and-competitive
| null | null |
https://aclanthology.org/2022.smila-1.10
|
https://aclanthology.org/2022.smila-1.10.pdf
|
Laughter During Cooperative and Competitive Games
|
This exploratory study investigates the extent to which social context influences the frequency of laughter. In a within-subjects design, dyads of strangers played two simple laughter-inducing games in a cooperative and competitive setting, ostensibly to earn money individually and as a team. We examined the frequency of laughs produced in both settings. The analysis revealed that, the effects of cooperative versus competitive framing interacted with the game. Specifically, when playing a general knowledge quiz, participants tended to laugh more in the cooperative than in the competitive setting. However, the opposite was true when participants were asked to find a specific number of poker chips under time pressure. During this task participants laughed more in a competitive than in the cooperative setting. Further analyses revealed that familiarity with the task affected the amount of laughter differently for each of the two tasks. Playing the second round of the poker chips task was associated with a significant decreases in laughter frequency compared to the first round. This effect was less marked for the general knowledge quiz, where increased familiarity with the task in the second round led to more laughs in the cooperative, but not competitive setting. Together, the results highlight the flexibility of laughter as an interaction signal and illustrate the challenges of studying laughter in naturalistic settings.
|
['William Curran', 'Ian Sneddon', 'Gary McKeown', 'Magdalena Rychlowska']
| null | null | null | null |
smila-lrec-2022-6
|
['general-knowledge']
|
['miscellaneous']
|
[-9.32272598e-02 1.18484668e-01 4.11159366e-01 2.29808941e-01
-2.33691797e-01 -6.85332954e-01 4.32557344e-01 2.23949607e-02
-6.50463879e-01 6.38412058e-01 3.02045286e-01 -1.39027676e-02
-1.39261127e-01 -7.70730376e-01 -1.96146265e-01 -7.20589221e-01
1.08068317e-01 -8.01977888e-02 -6.82200193e-02 -3.76329869e-01
5.76979518e-01 2.11219508e-02 -1.58473802e+00 -6.24966361e-02
8.00956488e-02 3.41003120e-01 5.70485353e-01 5.46132326e-01
2.99489588e-01 1.09663689e+00 -9.04655099e-01 -3.48307937e-01
2.33839065e-01 -8.34862292e-01 -4.86250997e-01 -2.04133037e-02
1.35392249e-01 -3.35208863e-01 1.40379176e-01 4.91910696e-01
5.78363836e-01 7.14503467e-01 -4.38428484e-02 -1.19035137e+00
-3.07730824e-01 8.19358349e-01 -4.71076310e-01 3.46781820e-01
9.84945953e-01 3.39514792e-01 1.01857400e+00 -5.89570701e-01
7.28713572e-01 1.14471030e+00 5.89856565e-01 2.67497927e-01
-1.55429208e+00 -1.11500955e+00 -1.62814841e-01 -8.13884661e-03
-1.22643769e+00 -4.86969441e-01 6.67103171e-01 -5.33091664e-01
5.38604379e-01 3.86731356e-01 1.17108452e+00 1.03728807e+00
5.42845950e-02 -1.21743955e-01 1.48804414e+00 -1.52994588e-01
2.85183728e-01 3.51761907e-01 1.56101480e-01 -7.91868940e-02
3.16509992e-01 3.20257872e-01 -9.63398516e-01 -5.02597511e-01
8.99589300e-01 -1.69579744e-01 -2.71465749e-01 1.06348336e-01
-1.14300990e+00 1.05772281e+00 1.64622515e-01 8.00499618e-01
-3.09101343e-01 1.70978472e-01 5.32096028e-01 5.83122253e-01
3.98208618e-01 1.04602158e+00 1.61426082e-01 -9.52677488e-01
-5.27627409e-01 5.70594192e-01 1.06942010e+00 -1.15669124e-01
4.51485932e-01 -3.92067939e-01 -2.35638127e-01 6.52814209e-01
4.04334106e-02 -3.05283964e-02 3.12766023e-02 -7.62954056e-01
2.59798527e-01 2.27947593e-01 6.92284703e-01 -1.45097077e+00
-5.86573184e-01 -1.79484263e-01 5.34205809e-02 3.99365455e-01
9.30197179e-01 -3.83052826e-01 3.95866692e-01 1.89924109e+00
2.53538668e-01 4.37374525e-02 -3.58649939e-01 1.11968815e+00
4.44724113e-01 4.23235089e-01 1.82250321e-01 -7.88652897e-01
1.56614077e+00 -2.02468112e-01 -9.83441353e-01 -2.86453515e-01
7.61839747e-01 -1.17406487e+00 1.53459311e+00 1.58467829e-01
-8.76599908e-01 -5.64281523e-01 -9.92699444e-01 3.62017661e-01
1.18675247e-01 -6.16322100e-01 4.35614377e-01 1.21003866e+00
-7.66909003e-01 7.13954389e-01 -3.70712161e-01 -5.70494473e-01
-7.98254758e-02 -2.18231395e-01 -2.29004711e-01 2.25018993e-01
-1.28873992e+00 1.16323686e+00 1.76405877e-01 -2.00235620e-01
-3.72562945e-01 -6.87166274e-01 -8.65635335e-01 2.30798244e-01
6.25423014e-01 -2.07834840e-01 1.26392508e+00 -1.18565309e+00
-1.57940638e+00 1.02436829e+00 1.72223210e-01 -4.85797897e-02
4.41610545e-01 -2.18019441e-01 -3.69866788e-02 9.50819552e-02
5.71386039e-01 2.26050913e-01 4.30749893e-01 -9.76427794e-01
1.71260685e-01 -2.53447920e-01 4.44411933e-01 6.89092338e-01
2.38887891e-01 3.92553657e-01 8.40261638e-01 -3.81129980e-01
-1.58200502e-01 -7.20438719e-01 2.48331383e-01 -5.58646321e-01
1.05248243e-01 -3.31807882e-01 7.08553433e-01 -1.29962087e-01
1.46920860e+00 -2.72872472e+00 -4.81315583e-01 9.69857648e-02
5.72518647e-01 -2.27349222e-01 2.47586191e-01 1.08167291e+00
-2.22455218e-01 2.28438348e-01 5.14179170e-01 -4.53924909e-02
3.32342803e-01 -8.22078362e-02 3.17035973e-01 6.32556677e-01
-2.99480349e-01 7.58088708e-01 -1.13077569e+00 -1.92446098e-01
-1.92701086e-01 8.76867920e-02 -4.96222019e-01 2.83982068e-01
4.03637052e-01 4.38813031e-01 4.49409448e-02 7.60551021e-02
6.57797098e-01 -2.64495075e-01 3.44074219e-01 5.65829813e-01
-6.26986921e-01 6.72602713e-01 -8.87690842e-01 8.32934439e-01
-3.99966866e-01 1.12629926e+00 3.95765424e-01 -5.38741112e-01
9.34620976e-01 4.30276960e-01 -5.42130880e-03 -6.10612154e-01
4.58786190e-01 4.65481132e-02 6.62299275e-01 -4.15705502e-01
8.05779040e-01 -9.80806470e-01 -2.89173067e-01 1.34524930e+00
-5.22739291e-01 -5.47766864e-01 1.85873359e-01 3.46322745e-01
1.00647712e+00 -4.59116906e-01 6.06372118e-01 -3.31003577e-01
-2.72683322e-01 -3.64860237e-01 3.89191091e-01 1.17728579e+00
-5.12114942e-01 1.51050881e-01 9.04629886e-01 -3.63707751e-01
-4.93823975e-01 -9.26686764e-01 2.09250271e-01 1.45434880e+00
4.30625558e-01 -4.28385884e-01 -7.12396383e-01 -1.84003294e-01
-1.34219259e-01 1.04407430e+00 -8.00952375e-01 -4.65217650e-01
-3.05905104e-01 -3.35719585e-01 2.74128705e-01 -1.59222454e-01
7.88640797e-01 -1.20416462e+00 -1.34474075e+00 6.08584546e-02
-5.25071502e-01 -8.61226439e-01 -9.28799808e-01 -1.35114551e-01
-4.20608908e-01 -1.06328344e+00 -9.60687622e-02 -4.00969177e-01
8.17563981e-02 7.26501882e-01 9.11638856e-01 5.30981302e-01
-2.69416511e-01 6.86119556e-01 -6.94460690e-01 -2.69470245e-01
-2.09645838e-01 -5.73567092e-01 -2.86482610e-02 -1.09448954e-01
4.68494058e-01 -9.01342511e-01 -4.61596489e-01 5.94413161e-01
-7.75082946e-01 3.24309468e-02 -8.69575050e-03 5.77447355e-01
-4.76698577e-01 -4.27881218e-02 7.45896459e-01 -5.66396356e-01
1.20006800e+00 -6.27399504e-01 2.95644347e-02 -6.02972865e-01
-1.66645851e-02 -7.68245220e-01 3.14969629e-01 -7.26564109e-01
-7.98831344e-01 -5.26892066e-01 6.33059978e-01 2.07316920e-01
-4.02648151e-02 6.03272021e-01 3.59262735e-01 -8.55057463e-02
1.13775826e+00 -4.00144130e-01 4.07095075e-01 -2.05174014e-02
-2.46211961e-01 4.58730400e-01 -2.09514156e-01 -5.85030317e-01
6.44063652e-01 8.45722109e-02 -3.50503176e-01 -1.03681231e+00
-6.63233936e-01 -2.38465622e-01 -1.32692263e-01 -7.46311426e-01
8.25558364e-01 -8.92058730e-01 -1.52565265e+00 6.15839183e-01
-6.65015340e-01 -8.56508315e-01 -1.78382605e-01 7.67339230e-01
-6.23405635e-01 2.36731127e-01 -4.18811291e-01 -1.06129801e+00
2.78178394e-01 -8.93918574e-01 3.12879950e-01 4.16767627e-01
-9.69090879e-01 -6.44238353e-01 1.22817263e-01 6.06718421e-01
3.11887830e-01 3.38827610e-01 5.20187795e-01 -7.11026371e-01
-2.00550526e-01 -2.22041115e-01 -1.74919982e-02 -3.30911160e-01
4.96374249e-01 -4.70344990e-01 -5.85329056e-01 2.63606533e-02
5.39756358e-01 -7.30206788e-01 -2.47582588e-02 1.65073127e-01
-1.26336485e-01 -2.35951319e-01 -2.23250687e-02 -2.96503514e-01
8.55752409e-01 3.79422903e-01 6.05532348e-01 2.82602727e-01
1.07435894e-03 5.03390670e-01 8.10448766e-01 4.98316020e-01
5.96556515e-02 8.45979452e-01 6.59915581e-02 5.36409974e-01
4.13541943e-01 -2.58821934e-01 6.39824688e-01 -3.68708521e-02
6.04518168e-02 1.20991133e-01 -5.33866167e-01 2.93714941e-01
-1.48697209e+00 -1.44397449e+00 -1.80219904e-01 2.53558922e+00
6.30352378e-01 7.44783133e-02 5.23251235e-01 2.28602082e-01
9.30844903e-01 5.55514038e-01 2.34044846e-02 -5.20503402e-01
1.94205791e-01 1.31053180e-01 -8.09340104e-02 5.84894300e-01
-4.71943259e-01 6.67851388e-01 6.64571285e+00 5.49375832e-01
-6.83672547e-01 2.69136310e-01 6.41725540e-01 -6.80324316e-01
4.75925347e-03 4.56673242e-02 -1.80529431e-01 4.16893810e-01
8.81660581e-01 -5.33367276e-01 6.01347148e-01 5.74302495e-01
1.00489700e+00 -1.29777730e+00 -9.64478314e-01 8.81751001e-01
-7.23596662e-02 -5.83281338e-01 -1.08445251e+00 5.74227907e-02
6.57079071e-02 -8.33000541e-01 -1.13691412e-01 2.56719291e-01
2.75483012e-01 -1.06611812e+00 7.94244409e-01 1.92757566e-02
5.52926481e-01 -4.78119671e-01 5.17461479e-01 5.02802670e-01
-7.81124413e-01 9.98235196e-02 -7.44939456e-03 -1.74175560e+00
4.29616034e-01 2.08315328e-01 -1.08366597e+00 -2.45556459e-01
8.08392942e-01 1.14782527e-01 3.00065912e-02 5.31910777e-01
-2.70873755e-01 6.56577229e-01 -3.62095743e-01 -4.71168220e-01
2.35042535e-02 -2.50902236e-01 7.13022351e-01 3.61271024e-01
3.29898261e-02 6.49271965e-01 -2.36088172e-01 1.08370388e+00
2.42434844e-01 8.06053281e-02 -6.62394583e-01 -3.80675614e-01
8.77618372e-01 1.33823538e+00 -1.22785854e+00 -9.47960839e-02
-1.81815729e-01 7.13442326e-01 1.92847133e-01 1.84075817e-01
-6.67943180e-01 -4.30778056e-01 7.27198243e-01 6.74228549e-01
-7.51749724e-02 -2.96618462e-01 -4.48910147e-01 -8.67094457e-01
-1.12684235e-01 -6.47362411e-01 -3.26914899e-03 -8.77457976e-01
-1.07109427e+00 -5.69354631e-02 3.44713181e-01 -7.65987813e-01
7.05937594e-02 2.38412976e-01 -9.02724564e-01 1.01714265e+00
-1.84131518e-01 -2.53216445e-01 -4.08069253e-01 1.82074577e-01
2.76520222e-01 5.84571540e-01 4.87076998e-01 -1.63857490e-01
-3.18168223e-01 3.00974667e-01 -4.05485123e-01 -1.98141515e-01
1.05675995e+00 -7.10970402e-01 -9.68724191e-02 2.40862831e-01
-2.05440640e-01 9.43964183e-01 9.63565946e-01 -6.55844927e-01
-7.77762115e-01 1.43824324e-01 8.58045518e-01 -3.81227642e-01
7.97267675e-01 -6.63113892e-01 -5.01832664e-01 4.56095636e-01
2.18824223e-01 -4.48422521e-01 1.17872691e+00 4.87214416e-01
-1.50119513e-01 5.60633838e-01 -1.20758843e+00 8.46751332e-01
1.25161660e+00 -7.92872488e-01 -7.50442803e-01 5.37464442e-03
6.66329443e-01 -2.67092854e-01 -4.79029536e-01 -5.50384879e-01
7.13280618e-01 -1.60259771e+00 5.85488975e-01 -2.17502132e-01
6.47460878e-01 4.80347797e-02 2.21867472e-01 -1.40112555e+00
-5.64905345e-01 -9.56935763e-01 7.51710176e-01 9.96100724e-01
-1.39061347e-01 -8.22689712e-01 5.58711886e-01 8.77252877e-01
8.24029297e-02 -9.66585129e-02 -1.07547081e+00 -5.11616051e-01
-1.22600302e-01 -1.78709850e-01 1.24039710e-01 9.10410643e-01
8.99583340e-01 4.04498398e-01 -4.43324804e-01 -3.79925668e-01
5.88218234e-02 -1.91457286e-01 7.61563718e-01 -1.06547499e+00
-5.79818785e-01 -2.74400324e-01 -3.75650823e-01 -5.37888169e-01
-7.85642266e-02 -3.40276420e-01 -7.68710077e-02 -8.24063122e-01
3.00023615e-01 -1.79731280e-01 4.32320744e-01 1.56218439e-01
-5.38065016e-01 2.12036833e-01 9.03275907e-01 2.88291983e-02
-3.63853872e-01 1.62778527e-01 1.17039645e+00 6.81801140e-01
-1.02593565e+00 7.07427636e-02 -1.23337138e+00 3.98727924e-01
8.83624315e-01 -2.92012155e-01 -4.51952398e-01 3.32038671e-01
7.05788910e-01 5.27963996e-01 5.20857215e-01 -7.12196350e-01
-9.06885117e-02 -2.66627729e-01 -2.03389108e-01 2.38324374e-01
5.40793717e-01 -5.75908661e-01 6.01520538e-01 4.27008957e-01
-4.04098332e-02 -1.85862288e-01 4.91337299e-01 2.81121075e-01
1.88047573e-01 -3.86484116e-01 7.05688357e-01 -1.13017499e-01
8.56521130e-02 -8.88608217e-01 -1.13424885e+00 1.47948772e-01
1.29210639e+00 -8.56495976e-01 -1.83506370e-01 -1.13384175e+00
-9.36985016e-01 -2.40153804e-01 6.08693004e-01 3.58684510e-01
1.23698726e-01 -1.09026670e+00 -4.90008593e-01 -7.52472132e-02
-2.36380890e-01 -5.19253671e-01 6.61362231e-01 1.44291413e+00
-3.79468262e-01 -9.97233912e-02 -1.58884287e-01 1.66760925e-02
-1.37477708e+00 3.45854163e-01 1.96723759e-01 -3.89405280e-01
-2.32514173e-01 8.59486401e-01 5.72081745e-01 6.70196712e-02
-2.17922062e-01 3.27753812e-01 -2.17786297e-01 6.80939615e-01
8.30256641e-01 7.83767462e-01 -2.55588800e-01 -5.93979776e-01
-1.50621593e-01 6.57536387e-02 -7.08334744e-02 -2.96655238e-01
9.52873290e-01 -3.79682779e-01 -2.19911009e-01 9.57872927e-01
7.01326609e-01 2.75375366e-01 -1.08759487e+00 1.80154279e-01
-5.17019570e-01 -1.00566828e+00 -2.92308450e-01 -5.34819663e-01
-2.92481482e-01 9.05156210e-02 -3.99037898e-02 8.02066267e-01
6.82488739e-01 -3.44959348e-02 1.72216043e-01 6.32288754e-02
8.30742538e-01 -9.93398964e-01 4.64139372e-01 4.63122219e-01
8.89778137e-01 -7.57009327e-01 -1.60874009e-01 -4.56263483e-01
-8.67348433e-01 6.72469556e-01 5.78094482e-01 -1.92197785e-01
4.48963493e-01 1.12193517e-01 -1.05959848e-01 -6.17937148e-01
-6.51409864e-01 -3.31472829e-02 -6.15451574e-01 6.48443699e-01
7.15971351e-01 2.19593212e-01 -1.07853138e+00 1.00497353e+00
-8.17058802e-01 3.05092558e-02 1.01095331e+00 1.00859869e+00
-6.56458199e-01 -6.15309060e-01 -1.07080674e+00 5.11721253e-01
-3.13849300e-01 -7.78855458e-02 -9.82265592e-01 9.67479765e-01
7.50236511e-02 1.41319597e+00 3.90679300e-01 -5.69772780e-01
2.64944971e-01 2.12099910e-01 3.11845541e-01 -9.03502882e-01
-1.30795407e+00 -2.49041524e-02 5.07282436e-01 -5.34077406e-01
-3.92095208e-01 -1.07762635e+00 -7.05747545e-01 -8.38236094e-01
-4.21614319e-01 5.91159403e-01 -2.84307420e-01 8.45676243e-01
1.25952154e-01 -1.00975730e-01 5.02026141e-01 -1.21889865e+00
-8.42907056e-02 -1.26138163e+00 -1.26920485e+00 4.71966177e-01
-1.81440599e-02 -9.51090336e-01 -9.64463234e-01 -2.87700653e-01]
|
[12.515290260314941, 7.762846946716309]
|
3e89cec5-bae0-4083-9079-053b38b0300a
|
recurrent-neural-network-training-with
|
1606.04449
| null |
http://arxiv.org/abs/1606.04449v2
|
http://arxiv.org/pdf/1606.04449v2.pdf
|
Recurrent neural network training with preconditioned stochastic gradient descent
|
This paper studies the performance of a recently proposed preconditioned
stochastic gradient descent (PSGD) algorithm on recurrent neural network (RNN)
training. PSGD adaptively estimates a preconditioner to accelerate gradient
descent, and is designed to be simple, general and easy to use, as stochastic
gradient descent (SGD). RNNs, especially the ones requiring extremely long term
memories, are difficult to train. We have tested PSGD on a set of synthetic
pathological RNN learning problems and the real world MNIST handwritten digit
recognition task. Experimental results suggest that PSGD is able to achieve
highly competitive performance without using any trick like preprocessing,
pretraining or parameter tweaking.
|
['Xi-Lin Li']
|
2016-06-14
| null | null | null | null |
['handwritten-digit-recognition']
|
['computer-vision']
|
[ 7.15624616e-02 -1.59571901e-01 1.13587894e-01 -4.99682933e-01
-1.96392432e-01 1.25107430e-02 5.24126053e-01 -5.51731944e-01
-9.12951350e-01 1.05947971e+00 3.18948925e-02 -8.42037976e-01
2.82462448e-01 -1.98199227e-01 -6.41364992e-01 -8.18223476e-01
1.46392286e-02 3.90236586e-01 1.06637537e-01 -4.33786422e-01
1.91945687e-01 7.35505998e-01 -1.17765939e+00 3.29112314e-04
4.69159007e-01 7.55901456e-01 5.66617787e-01 9.16412055e-01
1.01787604e-01 1.17413890e+00 -5.85684121e-01 -1.69855565e-01
2.09025502e-01 -6.65710688e-01 -8.48562121e-01 -1.37877345e-01
-3.47378589e-02 -2.68924594e-01 -5.98514438e-01 9.29937124e-01
8.40987027e-01 3.50110590e-01 4.31271285e-01 -6.81129932e-01
-4.52662975e-01 9.37572777e-01 -1.52563587e-01 8.83673728e-01
-2.29221568e-01 4.90884017e-03 6.33886635e-01 -1.25849271e+00
7.26218641e-01 1.16589713e+00 8.50539267e-01 1.10426807e+00
-9.12570596e-01 -5.92248440e-01 3.56950194e-01 1.55892760e-01
-1.04046774e+00 -5.10038018e-01 7.55449653e-01 -3.64158228e-02
1.34434307e+00 8.51807892e-02 4.84397203e-01 1.50160849e+00
2.22484380e-01 1.08608198e+00 8.62928927e-01 -4.12812054e-01
4.44156319e-01 1.77190453e-02 3.94996136e-01 7.97096014e-01
-3.48371044e-02 2.14206368e-01 -3.48669618e-01 -1.83133677e-01
8.27100933e-01 6.09486401e-02 -2.89255530e-01 -4.68103476e-02
-9.65979099e-01 7.69853115e-01 3.14036965e-01 4.04584557e-01
-5.29832721e-01 4.08804417e-01 8.26716483e-01 6.50789738e-01
5.49561679e-01 1.89152017e-01 -5.91805100e-01 -4.37311769e-01
-9.54040647e-01 -1.27074122e-01 7.80098081e-01 3.88758153e-01
4.56303895e-01 8.33446026e-01 1.08280405e-01 1.13849413e+00
2.42790669e-01 3.81045967e-01 1.38494062e+00 -6.12488270e-01
5.61622083e-01 6.17016964e-02 -7.01995194e-02 -8.27264309e-01
-6.94806039e-01 -8.20245564e-01 -1.23798358e+00 2.33049303e-01
4.07759435e-02 -5.85713744e-01 -1.10517359e+00 1.63457954e+00
-7.63551742e-02 4.68075007e-01 2.82911927e-01 1.01088727e+00
7.42395103e-01 8.54714990e-01 -2.18911305e-01 -4.19062793e-01
4.72723842e-01 -1.01400089e+00 -4.77430612e-01 -4.03819203e-01
8.84376287e-01 -3.18920732e-01 1.15714216e+00 5.01179457e-01
-1.01925075e+00 -4.11125451e-01 -1.24739659e+00 1.58261493e-01
-1.37404472e-01 5.13436913e-01 7.74896979e-01 6.74655437e-01
-1.14680612e+00 9.43473577e-01 -1.33382726e+00 -1.54174522e-01
6.67784065e-02 4.66144115e-01 -1.12916313e-01 1.86646044e-01
-1.26588953e+00 1.03528440e+00 6.00758076e-01 6.57944322e-01
-1.17368221e+00 -3.01918715e-01 -5.76021194e-01 -2.07204208e-01
-1.33088708e-01 -6.76015615e-01 1.26087523e+00 -9.84077930e-01
-2.35106850e+00 8.16817880e-01 -2.80894160e-01 -1.04631317e+00
6.18856192e-01 -3.33531320e-01 -3.73811096e-01 -1.34265825e-01
-6.36863947e-01 2.75729805e-01 9.90649283e-01 -7.32882440e-01
-1.56479642e-01 -1.07731186e-01 -5.17936468e-01 2.41591930e-01
-2.86853611e-01 1.98501870e-01 -5.48751932e-03 -6.92061126e-01
3.35937262e-01 -9.84085977e-01 -6.17887616e-01 -7.13307440e-01
-4.73674208e-01 -2.88215011e-01 8.65892172e-01 -5.21252692e-01
1.15494800e+00 -1.99344146e+00 1.15701847e-01 2.90649265e-01
1.72129422e-02 7.33197808e-01 -1.56299934e-01 1.41566083e-01
-3.46871614e-01 -2.09974632e-01 -1.80621594e-01 -5.30746102e-01
-2.15222269e-01 4.02314454e-01 -7.22223282e-01 5.96461117e-01
-1.84387296e-01 8.50673676e-01 -8.91918659e-01 9.12984014e-02
8.48554000e-02 4.48913902e-01 -2.12767437e-01 9.26519856e-02
-1.05132125e-02 1.74916074e-01 -1.83376208e-01 1.47507444e-01
1.18991643e-01 -2.91370213e-01 3.82344015e-02 1.87590629e-01
2.45320663e-01 5.21050036e-01 -1.04289293e+00 1.30633950e+00
-6.61281466e-01 1.00594902e+00 -2.30086118e-01 -1.31896520e+00
1.29466295e+00 2.87475348e-01 -1.23115860e-01 -3.23569596e-01
2.80777097e-01 3.87215108e-01 -1.43609136e-01 -1.12897441e-01
3.00564766e-01 -2.23402888e-01 4.34235215e-01 6.85960889e-01
1.87815398e-01 -2.51145363e-02 8.09748694e-02 2.46168450e-01
1.21584380e+00 -3.60192537e-01 2.09990703e-02 -2.42187217e-01
5.75528443e-01 -4.32603955e-01 6.62846923e-01 9.61764753e-01
9.52266380e-02 5.52309990e-01 2.79157519e-01 -7.62639344e-01
-9.12362933e-01 -8.60561490e-01 3.25675398e-01 1.27078056e+00
-3.65597129e-01 -3.41918133e-02 -5.84388375e-01 -4.73776430e-01
-5.31157017e-01 7.77406275e-01 -4.10727531e-01 -4.28072840e-01
-8.07007730e-01 -1.14160991e+00 8.82823169e-01 4.95879889e-01
6.24537647e-01 -1.41714096e+00 -4.93550122e-01 4.55640793e-01
2.50969321e-01 -9.70984042e-01 -3.05548340e-01 6.27267301e-01
-1.45800781e+00 -6.30290091e-01 -9.42458868e-01 -1.04076099e+00
8.29933882e-01 -7.08537996e-02 1.07217026e+00 9.72482562e-02
-1.58562079e-01 5.27737066e-02 -1.48398668e-01 -9.70365331e-02
-5.78262448e-01 3.07185471e-01 6.39042258e-01 -1.75575405e-01
1.05083480e-01 -8.37866366e-01 -2.88027704e-01 3.26098025e-01
-6.96892560e-01 -9.15080607e-02 6.21020615e-01 1.25351608e+00
5.04420221e-01 -4.19228375e-02 5.59847891e-01 -1.19485188e+00
1.06656754e+00 -2.24935994e-01 -7.37428248e-01 1.61593139e-01
-7.59517670e-01 3.62381130e-01 1.01052153e+00 -6.56550169e-01
-1.07131279e+00 1.01200230e-01 -5.98324656e-01 -5.40309429e-01
1.20189413e-01 8.01017165e-01 3.97700429e-01 -3.30484472e-02
8.87633264e-01 7.42297113e-01 -1.99710950e-01 -6.44084275e-01
2.30230108e-01 3.56073081e-01 7.50326037e-01 -2.57358253e-01
5.03359199e-01 2.87653208e-01 -2.14789346e-01 -8.08740377e-01
-7.69762218e-01 -1.30843043e-01 -3.27528894e-01 -3.16842250e-03
2.61171609e-01 -7.66221106e-01 -5.46736419e-01 8.68976235e-01
-9.91283178e-01 -9.40371335e-01 -1.28837749e-01 8.54581177e-01
-7.02729464e-01 3.47724855e-01 -1.37100542e+00 -6.19242191e-01
-7.92341352e-01 -7.73948312e-01 2.04345584e-01 7.38555193e-02
1.26138702e-02 -1.24430060e+00 3.18863630e-01 -2.64750332e-01
4.97647613e-01 -1.34007841e-01 6.95027769e-01 -7.82169759e-01
4.86633554e-02 -2.46937871e-02 7.64799714e-02 8.77324879e-01
5.58177643e-02 2.98717786e-02 -6.84636056e-01 -4.67581034e-01
5.24573863e-01 -6.44676685e-01 1.26225173e+00 3.79603118e-01
9.70127583e-01 -3.25848371e-01 -1.37172922e-01 8.23944747e-01
1.13201833e+00 5.29386699e-02 6.34202898e-01 3.09512854e-01
6.99055731e-01 -6.96032271e-02 1.66413516e-01 3.95120621e-01
-4.88144457e-02 5.91352321e-02 2.13560224e-01 -4.39493284e-02
8.57704133e-03 -1.97812483e-01 6.81093693e-01 1.26972294e+00
-8.97665769e-02 2.42123455e-02 -9.91792023e-01 2.14547545e-01
-1.93513465e+00 -8.94525170e-01 2.71206856e-01 1.74116445e+00
1.04856062e+00 5.57788670e-01 -1.05241679e-01 3.77378523e-01
7.86301017e-01 1.19185604e-01 -8.95827293e-01 -7.52969861e-01
-2.86153466e-01 2.33999848e-01 5.72694004e-01 5.31729102e-01
-8.00900400e-01 1.26093006e+00 7.58130264e+00 7.14293540e-01
-1.53047383e+00 3.48026827e-02 8.31371963e-01 -2.14219555e-01
1.00553833e-01 -4.67487544e-01 -1.04307294e+00 3.19230795e-01
1.38281786e+00 1.18018389e-01 5.59186459e-01 1.10845780e+00
3.29452515e-01 2.52887100e-01 -8.97817373e-01 1.37787914e+00
-1.09840550e-01 -1.46233165e+00 -1.16529606e-01 -5.60676217e-01
8.06572199e-01 8.63843858e-01 3.62169266e-01 5.10797799e-01
8.11953843e-01 -1.05412579e+00 3.84544492e-01 3.29118013e-01
2.43812680e-01 -9.66865540e-01 7.76944995e-01 4.68298584e-01
-6.45056248e-01 -2.08306476e-01 -8.00405204e-01 -2.07099602e-01
-3.55094373e-02 1.02946758e+00 -1.02673650e+00 -1.58846468e-01
4.83034402e-01 7.40724027e-01 -3.72767478e-01 9.16076541e-01
-5.60398102e-01 1.10417688e+00 -4.36017096e-01 -4.99038011e-01
6.57810509e-01 -1.33736894e-01 5.55166483e-01 1.45790875e+00
6.41183257e-02 -1.09133497e-01 -2.41748199e-01 5.40318847e-01
-2.22965941e-01 -5.11589460e-03 -3.41774225e-01 -1.30854636e-01
2.68635631e-01 9.45229709e-01 -1.00445652e+00 -3.11005443e-01
1.47909954e-01 1.35101008e+00 6.92877293e-01 7.50766397e-01
-6.61715984e-01 -6.89688265e-01 5.69520473e-01 -5.00267386e-01
6.85632348e-01 -5.21015167e-01 -1.57962203e-01 -1.22874486e+00
-6.21980503e-02 -7.44653761e-01 1.66302711e-01 -7.06802845e-01
-9.86515045e-01 1.07108939e+00 -6.32961929e-01 -7.43340015e-01
-7.82583892e-01 -8.06038141e-01 -6.75978363e-01 7.00251639e-01
-1.37477946e+00 -3.00218195e-01 2.72377729e-01 8.83125305e-01
5.28963685e-01 -3.95009816e-01 6.27110004e-01 1.40663236e-01
-1.05245495e+00 8.30758274e-01 5.89146376e-01 2.59251148e-01
1.49597391e-01 -1.02811325e+00 7.08474278e-01 9.20724630e-01
2.86585450e-01 7.10438371e-01 1.02534819e+00 -3.77268344e-01
-1.24969661e+00 -1.14677072e+00 7.57225037e-01 1.03298128e-01
6.42640948e-01 -3.95269454e-01 -8.42908919e-01 7.18562007e-01
-3.15731108e-01 1.73646450e-01 3.64318728e-01 1.13763094e-01
-2.19731122e-01 -5.12468405e-02 -9.29357767e-01 7.70244539e-01
1.02579653e+00 -7.21757531e-01 -7.49017060e-01 4.89966154e-01
3.92722189e-01 -5.24646401e-01 -3.82293373e-01 1.34743020e-01
9.77235511e-02 -8.77845228e-01 7.15829313e-01 -7.21137285e-01
4.99093868e-02 -8.68143365e-02 7.08452910e-02 -1.63921642e+00
-1.07223600e-01 -1.04343367e+00 -4.48160082e-01 8.63562942e-01
6.53557599e-01 -9.08796012e-01 1.06173706e+00 4.15921777e-01
-1.90015376e-01 -9.37025607e-01 -9.49575782e-01 -1.07377374e+00
-7.89814144e-02 -4.93986011e-01 9.39059779e-02 6.30180836e-01
6.67224526e-02 5.14346540e-01 -6.32338226e-01 -1.15262382e-01
1.06311105e-01 -2.22047925e-01 3.45956743e-01 -9.62148309e-01
-4.42498028e-01 -4.61323619e-01 -3.61555219e-01 -1.37548769e+00
4.67605472e-01 -8.21460724e-01 1.59682691e-01 -1.36359715e+00
-3.68465692e-01 -2.41459027e-01 -7.17419803e-01 4.58237499e-01
-1.57241940e-01 2.49529228e-01 -1.13351621e-01 2.34924957e-01
-5.94775677e-01 7.28534877e-01 6.17567360e-01 8.16553310e-02
-5.68637967e-01 5.13111115e-01 -2.93863744e-01 9.18328583e-01
9.94953394e-01 -7.19254673e-01 -5.60117483e-01 -2.85678089e-01
3.51170182e-01 -1.26628548e-01 -7.13598207e-02 -9.78180826e-01
3.40160996e-01 3.80580276e-01 4.25502867e-01 -6.94951355e-01
2.82887489e-01 -3.42559367e-01 -1.79270908e-01 9.71446455e-01
-4.86566603e-01 2.70368457e-01 7.41170123e-02 3.86193871e-01
-1.11305758e-01 -4.93045032e-01 8.77982616e-01 -2.47069553e-01
-7.11229801e-01 1.76087499e-01 -8.89004469e-01 -5.33881038e-02
4.41888809e-01 -6.15490973e-02 2.28397459e-01 -3.95955414e-01
-8.32485080e-01 5.50075062e-02 2.43937224e-02 1.60598531e-01
1.10413599e+00 -1.06093764e+00 -7.73620009e-01 1.85377941e-01
-6.37876987e-01 -3.70861441e-01 -1.52470455e-01 6.15208924e-01
-5.62804699e-01 5.85746646e-01 8.22698325e-03 -3.11330557e-01
-1.19061291e+00 5.93670666e-01 7.06141949e-01 -6.82760715e-01
-1.04527104e+00 1.43028617e+00 -3.51727962e-01 -6.71323419e-01
6.05618060e-01 -6.60283625e-01 -1.02250665e-01 -1.71171546e-01
8.32132876e-01 4.48676705e-01 2.86192477e-01 -8.42439681e-02
-3.09188396e-01 1.87053934e-01 -4.74420786e-01 -1.46868512e-01
1.43955886e+00 -1.32923499e-01 -3.13951708e-02 6.73725247e-01
1.36973190e+00 -5.54801404e-01 -1.21738589e+00 -4.67982709e-01
3.37977439e-01 3.24230969e-01 2.66319573e-01 -4.84463871e-01
-1.34964287e+00 6.95193112e-01 6.87814355e-01 -2.41991952e-02
1.04074633e+00 -4.26227242e-01 1.08797026e+00 1.41480315e+00
2.74137825e-01 -1.30577457e+00 -4.92715240e-02 1.19697797e+00
7.17119098e-01 -8.91334057e-01 -7.41356090e-02 4.87776875e-01
-6.90304279e-01 1.39547181e+00 2.85564721e-01 -7.00305164e-01
9.04800713e-01 1.71180815e-01 2.90966541e-01 4.43895198e-02
-1.12351632e+00 2.65504569e-01 -1.32403016e-01 3.27706158e-01
1.40043855e-01 -3.17238092e-01 -1.89889789e-01 2.63697058e-01
-4.10540193e-01 2.39061892e-01 5.43905258e-01 7.49759018e-01
-3.89043033e-01 -6.92249894e-01 6.15471601e-02 5.35664082e-01
-7.32750058e-01 -3.75744790e-01 -1.45556495e-01 2.57014543e-01
-7.12242603e-01 5.95491648e-01 -2.14386448e-01 -4.91888374e-01
-9.08155367e-02 1.01170033e-01 3.10165524e-01 -5.49993753e-01
-7.30028152e-01 -4.18743014e-01 2.03157496e-02 -4.90224600e-01
-2.30434656e-01 -3.92892122e-01 -1.53811741e+00 -1.66261554e-01
-3.36424232e-01 3.66153270e-01 9.00049210e-01 1.20220912e+00
2.31167421e-01 4.35106814e-01 8.09173942e-01 -9.37451601e-01
-9.01688397e-01 -9.97159302e-01 -7.24586785e-01 -1.74234267e-02
5.39152265e-01 -2.89140970e-01 -5.99309385e-01 -1.43401027e-01]
|
[10.854975700378418, 6.36912202835083]
|
77b9a8b7-5074-4ff1-aa94-ae80e8e2f02e
|
towards-unseen-triples-effective-text-image
|
2306.13420
| null |
https://arxiv.org/abs/2306.13420v1
|
https://arxiv.org/pdf/2306.13420v1.pdf
|
Towards Unseen Triples: Effective Text-Image-joint Learning for Scene Graph Generation
|
Scene Graph Generation (SGG) aims to structurally and comprehensively represent objects and their connections in images, it can significantly benefit scene understanding and other related downstream tasks. Existing SGG models often struggle to solve the long-tailed problem caused by biased datasets. However, even if these models can fit specific datasets better, it may be hard for them to resolve the unseen triples which are not included in the training set. Most methods tend to feed a whole triple and learn the overall features based on statistical machine learning. Such models have difficulty predicting unseen triples because the objects and predicates in the training set are combined differently as novel triples in the test set. In this work, we propose a Text-Image-joint Scene Graph Generation (TISGG) model to resolve the unseen triples and improve the generalisation capability of the SGG models. We propose a Joint Fearture Learning (JFL) module and a Factual Knowledge based Refinement (FKR) module to learn object and predicate categories separately at the feature level and align them with corresponding visual features so that the model is no longer limited to triples matching. Besides, since we observe the long-tailed problem also affects the generalization ability, we design a novel balanced learning strategy, including a Charater Guided Sampling (CGS) and an Informative Re-weighting (IR) module, to provide tailor-made learning methods for each predicate according to their characters. Extensive experiments show that our model achieves state-of-the-art performance. In more detail, TISGG boosts the performances by 11.7% of zR@20(zero-shot recall) on the PredCls sub-task on the Visual Genome dataset.
|
['Hanzi Wang', 'Ying Shan', 'Tianxiang Hou', 'Zhongang Qi', 'Wenxi Ma', 'Qianji Di']
|
2023-06-23
| null | null | null | null |
['scene-graph-generation', 'scene-understanding']
|
['computer-vision', 'computer-vision']
|
[ 4.15865272e-01 2.84678638e-01 -2.72085607e-01 -4.86955702e-01
-6.87678754e-01 -3.53972912e-01 6.23242855e-01 9.94124264e-02
-1.00225858e-01 6.71378970e-01 1.67259619e-01 -2.14552861e-02
-1.33093745e-02 -1.01381242e+00 -1.16580355e+00 -6.66903675e-01
3.30208570e-01 7.26901233e-01 6.42774820e-01 -2.74799556e-01
2.28126999e-02 1.03449434e-01 -2.00625515e+00 7.41457164e-01
1.18019438e+00 9.55929756e-01 4.83009160e-01 2.37909377e-01
-2.85058975e-01 8.06099713e-01 -4.36912149e-01 -6.21623993e-01
1.34218186e-01 -4.25560921e-01 -8.01727533e-01 1.74273580e-01
7.30208218e-01 5.21884998e-04 -1.69899896e-01 1.01538420e+00
3.91942084e-01 -1.44986706e-02 7.29747534e-01 -1.42160463e+00
-7.37834036e-01 5.87818146e-01 -7.96791852e-01 -1.55747952e-02
2.02411160e-01 8.51325169e-02 1.19610262e+00 -1.01583755e+00
9.07956421e-01 1.33254290e+00 4.81024414e-01 5.47727466e-01
-1.12983882e+00 -6.27846479e-01 3.44577372e-01 6.46987021e-01
-1.35082281e+00 -2.37196445e-01 7.04653800e-01 -3.14220190e-01
7.86730230e-01 3.48554254e-01 6.54720962e-01 9.47429657e-01
-1.34315416e-01 9.94559109e-01 7.48241246e-01 -3.35587293e-01
4.24227156e-02 1.46242063e-02 6.64911494e-02 9.02124047e-01
4.58675891e-01 -1.13889769e-01 -7.20605195e-01 2.69769371e-01
3.74604344e-01 -1.20778009e-01 -3.73643100e-01 -8.21254432e-01
-1.12902462e+00 6.74785674e-01 8.63606215e-01 -5.26286848e-03
-2.98798263e-01 -2.37032212e-02 1.31249458e-01 -1.57602102e-01
3.38599056e-01 5.29513597e-01 -6.46423757e-01 3.67535949e-01
-5.52476704e-01 3.10349166e-01 4.08387393e-01 1.14661252e+00
1.14509881e+00 -3.30160260e-01 -8.02206755e-01 9.19522047e-01
3.98816839e-02 2.62667239e-01 3.32138360e-01 -4.74168152e-01
5.58973908e-01 1.09126294e+00 -2.72178859e-01 -9.38899517e-01
-3.73678118e-01 -5.36890507e-01 -7.73219585e-01 -2.00949371e-01
4.53887045e-01 2.07183033e-01 -1.34278023e+00 1.88766670e+00
6.62960052e-01 2.73372442e-01 -7.15769976e-02 8.87818336e-01
1.31029427e+00 5.91550887e-01 2.37375587e-01 -4.88301776e-02
1.38388252e+00 -1.04108143e+00 -3.35215300e-01 -4.03953463e-01
7.61126399e-01 -4.68407601e-01 1.20710301e+00 1.51116282e-01
-6.15719378e-01 -6.12568617e-01 -8.99582088e-01 -2.92964280e-01
-4.64278460e-01 1.57123014e-01 7.32399821e-01 2.46263966e-01
-6.61464036e-01 4.36854482e-01 -4.63572621e-01 -5.05206287e-01
7.15919375e-01 1.46112442e-01 -3.65155339e-01 -5.32200992e-01
-1.01834416e+00 6.57990694e-01 8.52572262e-01 -7.52794147e-02
-8.87204349e-01 -1.00230420e+00 -8.30628693e-01 2.13506892e-01
8.30256402e-01 -9.73787606e-01 7.12483644e-01 -7.82328367e-01
-8.43910098e-01 9.61718619e-01 -2.40255103e-01 -2.46763349e-01
4.34286684e-01 -6.73856959e-02 -1.94679990e-01 -4.08786461e-02
3.51513952e-01 9.48638558e-01 7.62448668e-01 -1.46565950e+00
-8.43510151e-01 -4.50227529e-01 6.61146939e-02 4.01798546e-01
-1.77427396e-01 -4.14582372e-01 -8.40649188e-01 -5.08726180e-01
3.56418461e-01 -7.40721941e-01 3.08598415e-03 -1.30280361e-01
-6.10465169e-01 -3.51747304e-01 6.72275901e-01 -3.68679225e-01
9.28562880e-01 -2.09327412e+00 2.96197355e-01 -3.19299437e-02
2.82740712e-01 3.84405494e-01 -3.56901586e-01 1.53918147e-01
-2.15670288e-01 1.21170375e-02 -1.08225808e-01 -1.15525909e-01
-1.49048522e-01 3.18148553e-01 -1.43471003e-01 7.59666413e-02
5.39226890e-01 1.18875957e+00 -1.03097725e+00 -5.22940457e-01
1.10686786e-01 1.43217713e-01 -6.82361662e-01 2.76600659e-01
-5.81179023e-01 3.19477439e-01 -3.52351308e-01 7.02589989e-01
7.95716226e-01 -6.35178626e-01 1.93906307e-01 -6.78501010e-01
2.35571697e-01 6.68696268e-03 -1.09113204e+00 1.59512842e+00
-2.12241560e-02 1.01499058e-01 -5.38876295e-01 -9.93599176e-01
8.43167663e-01 -2.25600570e-01 1.68908879e-01 -7.97689676e-01
-1.57951534e-01 8.37277621e-02 -5.61986491e-02 -6.37947857e-01
3.81869406e-01 -1.62483975e-01 -7.09249005e-02 -1.81027681e-01
3.22405189e-01 -4.10439670e-02 3.80510747e-01 4.10969019e-01
1.03683102e+00 4.06353593e-01 3.86744440e-01 -3.22600342e-02
4.07808363e-01 2.71113366e-01 1.00446975e+00 7.01150537e-01
2.94287745e-02 8.25063944e-01 6.28974497e-01 -3.76491427e-01
-8.67740095e-01 -9.30674076e-01 -6.55932948e-02 1.18125820e+00
6.08964682e-01 -6.19997621e-01 -5.51505089e-01 -1.15246439e+00
6.93560615e-02 8.28908980e-01 -7.07650423e-01 -4.53617483e-01
-3.17620665e-01 -1.04295468e+00 2.31014356e-01 5.07322013e-01
6.14618242e-01 -9.87345397e-01 -2.23977134e-01 -1.24338202e-01
-4.33760345e-01 -1.12130988e+00 -1.47464141e-01 1.64405406e-01
-4.47536796e-01 -1.29788792e+00 -4.07287419e-01 -7.32937932e-01
7.62830615e-01 4.18969929e-01 1.21300900e+00 2.02756643e-01
-3.09427112e-01 -4.89847101e-02 -6.15881860e-01 -3.95811796e-01
-1.52247742e-01 2.07627583e-02 -2.68926114e-01 2.36970440e-01
4.64158833e-01 -1.83729053e-01 -4.07421499e-01 2.88805455e-01
-7.02361107e-01 5.98173261e-01 5.88138223e-01 1.06175506e+00
9.11972404e-01 7.80442506e-02 5.57722330e-01 -1.20829892e+00
4.83471202e-03 -4.49964821e-01 -5.66063166e-01 6.88440323e-01
-6.11994982e-01 2.29199991e-01 4.57409292e-01 -2.85342038e-01
-1.21309304e+00 2.03886017e-01 1.36243880e-01 -4.19023424e-01
-1.17478155e-01 4.34589863e-01 -6.45785689e-01 1.24215722e-01
6.99209094e-01 2.00035632e-01 -3.40286553e-01 -3.78684282e-01
5.16526818e-01 1.90177649e-01 4.88384843e-01 -5.29812038e-01
9.22036767e-01 2.43446529e-01 1.65476233e-01 -5.82305014e-01
-1.45208371e+00 -4.48991418e-01 -5.44022143e-01 -1.71849340e-01
8.65956187e-01 -1.12595952e+00 -4.29233134e-01 3.80875796e-01
-9.29766715e-01 -1.84926912e-01 -3.40975314e-01 1.72370672e-01
-4.75352764e-01 3.05564761e-01 -1.59835085e-01 -5.22300601e-01
1.85478013e-02 -1.02819014e+00 1.28389215e+00 4.23507333e-01
1.32028863e-01 -5.38746536e-01 -2.05151230e-01 5.82746565e-01
-5.76975942e-03 3.02875549e-01 1.28586817e+00 -7.27879107e-01
-1.02430034e+00 1.64539441e-01 -6.61065102e-01 9.76459906e-02
-7.87053443e-03 -2.89730191e-01 -1.03961587e+00 -1.63311675e-01
-4.25374717e-01 -5.53024054e-01 1.38583219e+00 1.91344947e-01
1.39254737e+00 -1.00872822e-01 -5.79701483e-01 7.88348436e-01
1.42526114e+00 -1.45204142e-01 7.73730993e-01 2.14563876e-01
1.19546282e+00 6.76458895e-01 9.31133687e-01 2.25562066e-01
6.49800718e-01 7.31156349e-01 7.32754350e-01 -1.97147265e-01
-5.65569103e-01 -7.07823515e-01 2.90187658e-03 1.88659474e-01
9.19463784e-02 -4.19934452e-01 -7.98943341e-01 6.01228714e-01
-2.11864448e+00 -8.97453725e-01 -3.83581340e-01 2.09968901e+00
8.11209321e-01 2.20227651e-02 -1.64355393e-02 5.86669482e-02
8.21778774e-01 -8.80188774e-03 -6.14362657e-01 2.94882983e-01
-3.76885802e-01 -9.18887556e-03 4.03004795e-01 1.77433893e-01
-1.07220078e+00 1.27127373e+00 5.12286472e+00 1.18254602e+00
-8.27365994e-01 -6.58748224e-02 7.26025522e-01 1.42452568e-01
-5.18359780e-01 1.22857593e-01 -1.11053419e+00 3.31260651e-01
2.73318201e-01 1.63394306e-02 3.78004730e-01 7.56491244e-01
-2.58015096e-01 -2.01703593e-01 -1.09439182e+00 1.04484558e+00
2.17797264e-01 -1.51038671e+00 5.26294887e-01 -1.27173111e-01
7.61104763e-01 -7.61597380e-02 -1.45214543e-01 6.05958998e-01
1.88317433e-01 -8.72107089e-01 6.75847530e-01 5.52785516e-01
8.70735228e-01 -5.69333792e-01 6.40919983e-01 4.24471349e-01
-1.19534528e+00 -1.81600198e-01 -5.48120856e-01 1.28605336e-01
-3.41254584e-02 6.86303496e-01 -1.12374902e+00 9.48819876e-01
7.46274292e-01 8.20237339e-01 -1.01936960e+00 1.17232740e+00
-4.34713006e-01 4.34453726e-01 -9.37349200e-02 1.19012110e-02
2.48012990e-02 6.80698976e-02 4.54262972e-01 8.94165397e-01
2.42234394e-01 1.99066326e-02 1.83288500e-01 8.60414624e-01
-1.87005863e-01 -9.24207736e-03 -4.56237048e-01 8.82561579e-02
3.31455797e-01 1.40876186e+00 -7.75605857e-01 -3.40473235e-01
-2.84105837e-01 8.35878849e-01 7.90337980e-01 3.99940461e-01
-8.72609615e-01 -1.33758083e-01 4.38909441e-01 2.83636808e-01
6.26410425e-01 4.72405314e-01 -2.32856885e-01 -1.41960335e+00
1.73629299e-02 -7.90930688e-01 8.08586061e-01 -8.72099221e-01
-1.42706025e+00 3.67099732e-01 1.11257233e-01 -1.06107759e+00
-2.81628873e-02 -5.20367801e-01 -2.44953051e-01 4.86541718e-01
-1.68521273e+00 -1.49063098e+00 -5.99757850e-01 5.59548914e-01
4.38228309e-01 -1.43212825e-02 5.53857446e-01 2.64892012e-01
-6.74016654e-01 5.86150348e-01 -3.39062661e-01 -2.90975589e-02
7.23867953e-01 -1.35076320e+00 1.11517556e-01 1.00824130e+00
2.18811855e-01 2.84175813e-01 6.79137230e-01 -8.16251159e-01
-1.26127887e+00 -1.54934323e+00 9.43382144e-01 -5.04217565e-01
2.15596944e-01 -4.95765239e-01 -1.05772412e+00 6.93999469e-01
-2.38263398e-01 2.47150972e-01 4.11814451e-01 3.94497067e-01
-6.05698228e-01 -2.85246640e-01 -1.09786451e+00 5.58628917e-01
1.53911841e+00 -1.89703003e-01 -5.89487731e-01 4.50522363e-01
8.31645072e-01 -4.20980275e-01 -4.15890664e-01 8.52098167e-01
1.57137528e-01 -9.96547163e-01 9.10842121e-01 -8.52869034e-01
5.02234936e-01 -7.06066430e-01 -2.89821237e-01 -1.27903855e+00
-6.23267531e-01 -1.31302595e-01 -1.61378577e-01 1.49895525e+00
4.76152569e-01 -3.11676055e-01 8.61087024e-01 3.50374490e-01
-3.04268867e-01 -7.21020818e-01 -6.91169560e-01 -7.00111687e-01
-3.79742414e-01 -2.02653319e-01 8.30902338e-01 9.36752915e-01
-2.31921241e-01 8.45210910e-01 -3.54934871e-01 2.60923475e-01
5.98910809e-01 4.17757094e-01 9.62405920e-01 -1.28086221e+00
-2.47184888e-01 -1.44292668e-01 -5.10490358e-01 -9.29881573e-01
-1.80272944e-02 -1.20339966e+00 1.69130102e-01 -1.71330786e+00
7.23217607e-01 -5.42958856e-01 -2.27398366e-01 6.81575477e-01
-6.94730818e-01 2.23599702e-01 2.59598553e-01 -1.40573233e-01
-1.00176251e+00 7.85770774e-01 1.40189803e+00 -1.98869064e-01
8.97074025e-03 -2.79217601e-01 -9.62288678e-01 6.34170830e-01
4.34770823e-01 -3.80476266e-01 -5.67519546e-01 -4.54951137e-01
3.21465820e-01 -2.13269085e-01 5.64384937e-01 -8.83057892e-01
1.21074498e-01 -2.15443820e-01 6.35947108e-01 -7.90327132e-01
2.81318337e-01 -6.05955899e-01 2.64935941e-01 2.26632506e-01
-1.25215203e-01 -3.99689347e-01 1.58844460e-02 8.62404823e-01
1.14571559e-03 -5.69957979e-02 6.99709952e-01 -1.41834766e-01
-1.09701765e+00 4.53401446e-01 3.05795401e-01 3.04769725e-01
1.08674669e+00 -2.15291202e-01 -7.43229926e-01 -6.79087192e-02
-5.57115614e-01 4.39631045e-01 5.89663744e-01 4.59694028e-01
6.54997170e-01 -1.33619344e+00 -6.98320806e-01 3.47580492e-01
8.35544169e-01 3.73211026e-01 4.21940684e-01 6.05577767e-01
-8.28717574e-02 1.91505432e-01 -1.11149624e-01 -7.97909379e-01
-1.18341827e+00 7.85187840e-01 1.97563067e-01 -2.69329906e-01
-4.81863111e-01 1.21527708e+00 7.71510243e-01 -4.27725136e-01
7.38764778e-02 -1.54223859e-01 -3.31125438e-01 2.17833534e-01
4.00013536e-01 1.04508206e-01 7.88493454e-02 -5.73335052e-01
-4.41663563e-01 4.94953662e-01 -2.27165580e-01 6.05277479e-01
1.27252769e+00 5.70046827e-02 -5.01965880e-02 2.00901672e-01
8.80871594e-01 -2.03975171e-01 -1.27368307e+00 -3.75413805e-01
4.27373536e-02 -5.62097907e-01 -1.73221692e-01 -9.55811799e-01
-1.00322258e+00 6.22630119e-01 3.82770389e-01 -9.60965641e-03
1.22156775e+00 4.80025351e-01 4.62126404e-01 2.05312744e-01
3.40234041e-01 -8.19042027e-01 1.96025953e-01 4.76372510e-01
7.75205195e-01 -1.29876161e+00 -5.62163293e-02 -1.01384664e+00
-7.71426022e-01 7.34754920e-01 9.62986827e-01 8.62728134e-02
1.73063532e-01 -1.94767490e-01 -3.21548343e-01 -3.90213102e-01
-8.74284148e-01 -6.69688463e-01 5.85563183e-01 7.79830337e-01
1.49338007e-01 6.97502792e-02 -2.14320451e-01 5.80672979e-01
-7.64500126e-02 -1.60896555e-01 2.49805212e-01 5.03048003e-01
-5.43066740e-01 -1.01824200e+00 -3.30154561e-02 6.68796659e-01
-9.33368690e-03 -1.58084720e-01 -3.49455327e-01 7.19113469e-01
5.03565311e-01 8.03059161e-01 -7.09202364e-02 -5.53992391e-01
4.28462058e-01 1.55011145e-02 6.31427705e-01 -9.04077351e-01
-2.43783712e-01 -1.34003863e-01 2.91963726e-01 -6.39382899e-01
-3.65151972e-01 -5.89456201e-01 -1.23448038e+00 -4.85421275e-04
-4.62734014e-01 -1.00616366e-01 1.35047093e-01 9.95013058e-01
5.66562951e-01 8.02062809e-01 3.63736331e-01 -5.05490959e-01
-2.68768668e-01 -6.97127461e-01 -4.55625892e-01 7.48647213e-01
-4.35658656e-02 -1.00202680e+00 -2.11931784e-02 -3.53396609e-02]
|
[10.318424224853516, 1.7435222864151]
|
b70ca5d8-b508-4c50-82ee-1fa9bada9d0e
|
learning-a-high-fidelity-pose-invariant-model
|
1806.08472
| null |
http://arxiv.org/abs/1806.08472v2
|
http://arxiv.org/pdf/1806.08472v2.pdf
|
Learning a High Fidelity Pose Invariant Model for High-resolution Face Frontalization
|
Face frontalization refers to the process of synthesizing the frontal view of
a face from a given profile. Due to self-occlusion and appearance distortion in
the wild, it is extremely challenging to recover faithful results and preserve
texture details in a high-resolution. This paper proposes a High Fidelity Pose
Invariant Model (HF-PIM) to produce photographic and identity-preserving
results. HF-PIM frontalizes the profiles through a novel texture warping
procedure and leverages a dense correspondence field to bind the 2D and 3D
surface spaces. We decompose the prerequisite of warping into dense
correspondence field estimation and facial texture map recovering, which are
both well addressed by deep networks. Different from those reconstruction
methods relying on 3D data, we also propose Adversarial Residual Dictionary
Learning (ARDL) to supervise facial texture map recovering with only monocular
images. Exhaustive experiments on both controlled and uncontrolled environments
demonstrate that the proposed method not only boosts the performance of
pose-invariant face recognition but also dramatically improves high-resolution
frontalization appearances.
|
['Yibo Hu', 'Zhenan Sun', 'Jie Cao', 'Ran He', 'Hongwen Zhang']
|
2018-06-22
|
learning-a-high-fidelity-pose-invariant-model-1
|
http://papers.nips.cc/paper/7551-learning-a-high-fidelity-pose-invariant-model-for-high-resolution-face-frontalization
|
http://papers.nips.cc/paper/7551-learning-a-high-fidelity-pose-invariant-model-for-high-resolution-face-frontalization.pdf
|
neurips-2018-12
|
['robust-face-recognition']
|
['computer-vision']
|
[ 5.48148572e-01 2.98092160e-02 1.79214239e-01 -5.12247503e-01
-7.15301752e-01 -4.69586194e-01 6.32078469e-01 -9.95057583e-01
1.82173103e-01 3.71192783e-01 3.44646305e-01 3.54948670e-01
7.65185058e-02 -8.86632740e-01 -1.01849699e+00 -8.32299829e-01
5.41251779e-01 3.58136207e-01 -3.86943758e-01 -2.01147243e-01
-1.35028008e-02 1.10303617e+00 -1.86213040e+00 3.61204177e-01
4.56532568e-01 1.10914290e+00 -2.64670551e-01 1.22257158e-01
1.86419129e-01 4.19881761e-01 -1.05812766e-01 -6.60704672e-01
7.24631965e-01 -2.18534127e-01 -5.58144689e-01 5.13505936e-01
1.29356301e+00 -6.87057257e-01 -7.56038070e-01 1.00612319e+00
3.86864722e-01 -4.13157754e-02 5.97524583e-01 -1.04959512e+00
-7.79262722e-01 -4.20824260e-01 -7.70842910e-01 -4.14301157e-01
6.46374166e-01 -8.87776315e-02 4.05128092e-01 -1.34481823e+00
1.06714392e+00 1.52099359e+00 7.39833713e-01 7.24620938e-01
-1.43851960e+00 -7.85314798e-01 -1.45325959e-01 -4.35480028e-02
-1.49184918e+00 -1.03825128e+00 1.08871961e+00 -3.40121686e-01
2.82274455e-01 3.98905277e-01 5.23397923e-01 1.33407736e+00
3.04385990e-01 4.23183411e-01 1.42888725e+00 -2.96755463e-01
-1.37298822e-01 -2.81240553e-01 -5.34681916e-01 9.00319517e-01
8.53919238e-02 3.73357207e-01 -8.52639854e-01 -1.54009750e-02
1.34817111e+00 5.12446582e-01 -6.04807913e-01 -5.85546613e-01
-8.52107882e-01 4.66025978e-01 1.96720526e-01 -2.02392563e-01
-4.64672297e-01 -1.69686779e-01 -2.07530215e-01 3.60347062e-01
7.17079461e-01 1.79477781e-01 -8.99386406e-03 4.04617280e-01
-1.10851550e+00 2.53805280e-01 5.42698026e-01 8.94365013e-01
1.06445825e+00 3.96409929e-01 -6.77506551e-02 8.22788954e-01
1.94457620e-01 1.10809863e+00 5.40988967e-02 -1.09624732e+00
1.14182249e-01 3.77942562e-01 1.07814342e-01 -1.22135878e+00
9.37691629e-02 -8.58808160e-02 -8.37718427e-01 3.68358016e-01
2.84702748e-01 2.57364422e-01 -1.00853384e+00 1.82602131e+00
6.82697058e-01 2.84865379e-01 -2.52359688e-01 8.98575842e-01
7.92894363e-01 3.47856969e-01 -3.38164568e-01 -2.60654449e-01
1.24693060e+00 -6.43600821e-01 -9.21098769e-01 -1.44919187e-01
-4.05664980e-01 -1.02685666e+00 9.08154607e-01 3.21732074e-01
-1.26715279e+00 -5.01467884e-01 -1.06193495e+00 -5.12920260e-01
6.38182387e-02 -1.23401932e-01 3.80841732e-01 4.39405560e-01
-1.03547168e+00 6.37802005e-01 -9.24814045e-01 -1.49295494e-01
7.49706686e-01 3.44030440e-01 -1.19646800e+00 -6.59039676e-01
-7.93703496e-01 6.85736954e-01 -5.08927107e-01 2.70175636e-01
-1.08055639e+00 -9.29054022e-01 -1.14044654e+00 -2.11487621e-01
1.30466357e-01 -7.90335536e-01 6.61602974e-01 -9.66029882e-01
-1.76818371e+00 1.22229469e+00 -4.88989472e-01 2.73148805e-01
5.98013937e-01 -2.68158942e-01 -3.87207836e-01 3.38863969e-01
8.34237337e-02 4.26980823e-01 1.64532936e+00 -1.34763157e+00
-2.95829810e-02 -8.61038446e-01 -1.77267462e-01 2.28646070e-01
-2.91939914e-01 -1.51232779e-01 -6.63554311e-01 -6.80638552e-01
3.53588283e-01 -7.56119967e-01 3.30227077e-01 5.76314151e-01
-2.46169820e-01 5.40556192e-01 1.01660860e+00 -1.21621835e+00
5.53021073e-01 -2.10630345e+00 4.91043985e-01 1.36159480e-01
3.20230126e-01 4.18054350e-02 -4.22850192e-01 2.12248400e-01
-2.29927003e-01 -4.05223876e-01 -1.96077690e-01 -7.16801822e-01
1.60925183e-02 3.41578245e-01 -4.51843917e-01 9.57556367e-01
3.45186532e-01 9.69740450e-01 -4.94862258e-01 -8.24357867e-02
3.44540030e-01 1.15568221e+00 -7.24707007e-01 5.21134019e-01
8.06188658e-02 7.38233030e-01 -2.90517330e-01 1.01468539e+00
1.33896542e+00 6.75296411e-02 1.94622189e-01 -7.51767278e-01
5.11096679e-02 -2.94767737e-01 -9.49251294e-01 2.01367688e+00
-5.21304309e-01 2.79128373e-01 5.37925899e-01 -7.46251285e-01
8.41661155e-01 2.37327918e-01 5.45044005e-01 -8.52568567e-01
2.03235015e-01 1.31275490e-01 -7.60964096e-01 -3.58697683e-01
1.57271668e-01 -3.97936344e-01 3.53335559e-01 2.36237958e-01
2.08824232e-01 -2.32122734e-01 -6.17238224e-01 -4.06488895e-01
8.13645124e-01 6.03080690e-01 -6.73268214e-02 -2.65753031e-01
6.06047451e-01 -6.87315166e-01 6.79443955e-01 -7.25897849e-02
1.73207417e-01 1.14009559e+00 1.48539886e-01 -7.03362286e-01
-1.20474803e+00 -1.28231525e+00 -3.52374166e-01 5.75532973e-01
1.59529865e-01 -1.00873351e-01 -1.02522516e+00 -4.53506827e-01
2.43329018e-01 1.36694551e-01 -1.02466166e+00 -2.46808618e-01
-6.78459525e-01 -2.91677028e-01 2.46919394e-01 1.52960390e-01
6.92082644e-01 -7.10444689e-01 7.52514554e-03 -2.19996244e-01
-3.09794158e-01 -1.36716747e+00 -8.25227261e-01 -4.62077767e-01
-5.22331774e-01 -1.21266472e+00 -8.18761349e-01 -8.08647931e-01
9.81001377e-01 5.27078271e-01 8.39542329e-01 -1.22831129e-01
-4.11639452e-01 4.75770593e-01 1.79107022e-02 -3.61439697e-02
-1.72986060e-01 -5.66599011e-01 2.73717493e-01 1.01043403e+00
8.31577480e-02 -9.33313429e-01 -7.39001036e-01 4.60092157e-01
-1.01984763e+00 1.46840349e-01 4.48146194e-01 8.30937564e-01
8.69268060e-01 -1.46598965e-01 5.02967089e-02 -8.97507668e-01
-1.59992315e-02 -1.28084108e-01 -5.35892427e-01 2.74702519e-01
-3.45832676e-01 -4.11446802e-02 5.57902873e-01 -3.52122724e-01
-1.29190218e+00 1.98832810e-01 -3.09942007e-01 -1.07594192e+00
-2.60609724e-02 -2.47419596e-01 -7.05695331e-01 -6.47598326e-01
4.88751054e-01 3.94121438e-01 5.98420382e-01 -6.10857308e-01
2.28838488e-01 1.96594179e-01 7.50395298e-01 -5.55168808e-01
1.38085365e+00 1.14909673e+00 3.03896189e-01 -8.89634192e-01
-8.52380097e-01 3.53669003e-02 -8.95033240e-01 -2.61508405e-01
8.54722679e-01 -1.05859268e+00 -7.24197447e-01 5.99066794e-01
-9.31181848e-01 -1.60880372e-01 -1.42529458e-01 6.51175529e-02
-6.38304234e-01 4.68872905e-01 -3.66121650e-01 -3.11178565e-01
-4.29164618e-01 -1.03982198e+00 1.60793686e+00 -9.78274271e-03
2.44445831e-01 -7.29455411e-01 -5.76108471e-02 6.92821801e-01
4.45916563e-01 7.71300495e-01 5.39145052e-01 3.62638116e-01
-7.52917767e-01 -2.21985549e-01 -4.28801179e-02 4.38052148e-01
3.28549474e-01 -2.34591201e-01 -1.39434004e+00 -6.20222569e-01
2.85468072e-01 -2.52550870e-01 5.50742149e-01 1.68755233e-01
9.77167428e-01 -4.66904551e-01 -5.08086756e-02 1.45687687e+00
1.33762813e+00 -1.61228746e-01 9.19180036e-01 -1.32683776e-02
1.06218433e+00 9.06104505e-01 5.07719517e-01 3.32382679e-01
1.28330037e-01 1.08407247e+00 4.44657177e-01 -2.91313887e-01
-5.48750877e-01 -6.02818370e-01 4.69968230e-01 5.49616814e-01
-2.83859700e-01 2.50260264e-01 -4.75682467e-01 1.50969580e-01
-1.45705736e+00 -8.89665008e-01 4.82126355e-01 2.32044673e+00
7.86614120e-01 -6.39024436e-01 -4.03374612e-01 6.41983226e-02
6.38538599e-01 1.74618900e-01 -6.03293896e-01 1.27697647e-01
-3.17477643e-01 6.13683343e-01 1.28216192e-01 8.38547766e-01
-9.57497656e-01 9.61350799e-01 5.42423105e+00 7.49349236e-01
-1.37970483e+00 2.76434898e-01 5.55788815e-01 -1.58779502e-01
-4.80593652e-01 -2.68836945e-01 -5.32396436e-01 2.08672345e-01
5.00647426e-01 1.11286439e-01 7.63311625e-01 5.81539571e-01
-5.87877110e-02 4.64450359e-01 -1.05467498e+00 1.41196275e+00
5.62470376e-01 -1.34166348e+00 4.71497238e-01 3.67826998e-01
8.52018237e-01 -4.92079347e-01 4.13399071e-01 -2.33005494e-01
-6.79475814e-02 -1.31245494e+00 7.67841876e-01 9.12292182e-01
1.57255876e+00 -7.96428859e-01 2.31108457e-01 -1.40410841e-01
-1.06082320e+00 1.39383048e-01 -5.05234957e-01 3.32176507e-01
2.47235112e-02 4.22513694e-01 -2.42744371e-01 7.77194500e-01
6.73997581e-01 8.87348473e-01 -2.12970138e-01 2.18946189e-01
-2.77343601e-01 -1.24871219e-03 -1.01976737e-01 1.07548654e+00
-3.96490037e-01 -5.15852094e-01 4.04972136e-01 5.61674118e-01
3.60697210e-01 1.71787918e-01 -1.54149979e-01 1.00598562e+00
-2.99167573e-01 -3.54526825e-02 -8.38275433e-01 2.66217202e-01
2.97639161e-01 1.43334568e+00 -2.00557962e-01 8.03590342e-02
-3.76919121e-01 1.36387312e+00 3.01306874e-01 4.38673705e-01
-6.77388668e-01 1.76025689e-01 1.08268380e+00 6.83356583e-01
2.31011853e-01 -1.49983525e-01 9.16119814e-02 -1.50535393e+00
2.26383537e-01 -1.04968941e+00 -2.35109389e-01 -7.99060225e-01
-1.25908077e+00 7.62846589e-01 -3.48180979e-01 -1.21239948e+00
1.40285064e-02 -6.50726736e-01 -3.28075588e-01 1.07571125e+00
-1.82292461e+00 -1.76023531e+00 -7.80944407e-01 1.11457813e+00
2.99840868e-01 -2.24002734e-01 9.99822557e-01 4.98908132e-01
-5.23906767e-01 8.68205070e-01 -1.21200318e-03 1.06952824e-02
9.59090292e-01 -6.63654268e-01 4.23375726e-01 8.33899915e-01
-2.60882787e-02 7.50504136e-01 4.08194691e-01 -4.62061048e-01
-2.09980059e+00 -1.30444980e+00 5.75039983e-01 -5.96493661e-01
-1.63977705e-02 -5.46224177e-01 -8.96977425e-01 7.20009208e-01
-4.40373197e-02 4.80987281e-01 4.63260919e-01 -4.30677146e-01
-8.05697024e-01 -4.05369788e-01 -1.39738142e+00 5.20602643e-01
1.34252214e+00 -8.97718430e-01 -3.03040147e-01 3.17892879e-01
2.29657695e-01 -5.83378613e-01 -1.15754926e+00 4.90644246e-01
9.06582475e-01 -1.05193496e+00 1.24456060e+00 -4.55035985e-01
5.09876311e-01 -2.29196638e-01 -4.85783249e-01 -1.01146293e+00
-3.52687031e-01 -8.74719620e-01 -3.79316136e-02 1.22567070e+00
-2.97194272e-01 -6.52606905e-01 8.19695771e-01 5.47604501e-01
-1.34130135e-01 -4.95631307e-01 -9.29321289e-01 -5.17638445e-01
-1.70799438e-02 2.15419114e-01 9.20922101e-01 1.13880599e+00
-6.72962606e-01 -3.23532745e-02 -6.43612981e-01 3.79898220e-01
1.00982952e+00 3.75832409e-01 9.68700171e-01 -1.14789367e+00
-7.39763118e-03 2.24297270e-02 -5.00853360e-01 -8.94155324e-01
5.85182309e-01 -7.79901028e-01 -1.63874939e-01 -8.78002107e-01
1.69727907e-01 -2.33577769e-02 1.81986928e-01 4.55061316e-01
1.99421018e-01 8.97497833e-01 -1.75464958e-01 3.06695253e-01
1.00576103e-01 9.14330244e-01 1.72975779e+00 5.51815182e-02
3.66816014e-01 -2.93079466e-01 -7.18138218e-01 6.46195531e-01
1.99214965e-01 -6.49236292e-02 -4.55815792e-01 -6.65135980e-01
-1.36520654e-01 2.55843729e-01 4.89076555e-01 -8.21887910e-01
4.52347146e-03 -3.02967846e-01 8.54864478e-01 -9.98063609e-02
7.72568405e-01 -9.33481097e-01 6.18527651e-01 -3.88266258e-02
4.65565473e-02 -1.88582502e-02 1.69324279e-01 5.85484385e-01
-2.78357089e-01 5.59324563e-01 1.20741260e+00 6.85326159e-02
-4.09839362e-01 1.04159963e+00 3.57396305e-01 -1.65497929e-01
8.36098373e-01 -3.76043618e-01 -8.42362344e-02 -3.83764744e-01
-6.71941459e-01 -5.94985127e-01 1.06291175e+00 4.90775198e-01
9.76441443e-01 -1.69399559e+00 -7.59885907e-01 1.22920358e+00
2.16561593e-02 2.88647506e-02 5.16790926e-01 6.38830066e-01
-7.81470895e-01 8.78498405e-02 -7.51680195e-01 -4.69948024e-01
-1.29910219e+00 4.76633281e-01 5.62089443e-01 1.90099582e-01
-8.68795514e-01 6.97424471e-01 7.82907367e-01 -5.26285350e-01
8.58435556e-02 3.38753223e-01 7.56954551e-02 -2.24165902e-01
7.35190570e-01 7.90443569e-02 2.88627505e-01 -1.09531474e+00
-1.49746403e-01 1.28215551e+00 -7.40658268e-02 -1.14987038e-01
1.53889012e+00 -1.23502329e-01 -3.85129064e-01 -2.54897445e-01
1.51723397e+00 2.62396306e-01 -1.73470557e+00 -3.53453308e-01
-6.59727097e-01 -1.14066088e+00 1.64718017e-01 -2.67749518e-01
-1.46090961e+00 9.08430398e-01 4.90190268e-01 -6.61718965e-01
1.18564594e+00 -4.62573051e-01 7.88234174e-01 -1.11549266e-01
5.23848176e-01 -6.87259078e-01 1.90932795e-01 3.01536262e-01
1.41565824e+00 -9.58872914e-01 8.84975493e-02 -7.40655065e-01
-2.22495273e-01 1.11772394e+00 5.10730028e-01 -2.73372591e-01
8.19866359e-01 2.11450130e-01 1.52962562e-02 -3.05870831e-01
-3.72696072e-01 2.48180777e-01 5.50934553e-01 7.43494749e-01
1.34461045e-01 -1.56000882e-01 3.74249876e-01 2.49177858e-01
-2.44737640e-01 -2.64052361e-01 1.48735613e-01 6.62865460e-01
1.26907364e-01 -1.08810997e+00 -5.89040101e-01 2.85317730e-02
-2.80488104e-01 7.55408481e-02 -3.24717790e-01 6.72743976e-01
5.69375269e-02 7.43259370e-01 1.19497754e-01 -4.64942604e-01
4.53725785e-01 6.26333356e-02 1.01686954e+00 -4.84291703e-01
-1.01902582e-01 1.44766256e-01 -2.42329881e-01 -1.15734482e+00
-3.31417441e-01 -5.85590065e-01 -6.46413743e-01 -9.07089651e-01
1.63050875e-01 -4.02018309e-01 4.44769442e-01 6.50438845e-01
5.54134846e-01 2.26012290e-01 1.00189662e+00 -1.24696946e+00
-4.43984032e-01 -6.46575034e-01 -9.71290350e-01 8.37579787e-01
5.13248980e-01 -1.15476942e+00 -3.80622417e-01 2.32449487e-01]
|
[12.90057373046875, -0.05809829756617546]
|
f11b4491-b212-4537-b3d8-f03fff41864f
|
godel-large-scale-pre-training-for-goal
|
2206.11309
| null |
https://arxiv.org/abs/2206.11309v1
|
https://arxiv.org/pdf/2206.11309v1.pdf
|
GODEL: Large-Scale Pre-Training for Goal-Directed Dialog
|
We introduce GODEL (Grounded Open Dialogue Language Model), a large pre-trained language model for dialog. In contrast with earlier models such as DialoGPT, GODEL leverages a new phase of grounded pre-training designed to better support adapting GODEL to a wide range of downstream dialog tasks that require information external to the current conversation (e.g., a database or document) to produce good responses. Experiments against an array of benchmarks that encompass task-oriented dialog, conversational QA, and grounded open-domain dialog show that GODEL outperforms state-of-the-art pre-trained dialog models in few-shot fine-tuning setups, in terms of both human and automatic evaluation. A novel feature of our evaluation methodology is the introduction of a notion of utility that assesses the usefulness of responses (extrinsic evaluation) in addition to their communicative features (intrinsic evaluation). We show that extrinsic evaluation offers improved inter-annotator agreement and correlation with automated metrics. Code and data processing scripts are publicly available.
|
['Jianfeng Gao', 'Bill Dolan', 'Zhou Yu', 'Elnaz Nouri', 'Lars Liden', 'Chris Brockett', 'Pengcheng He', 'Michel Galley', 'Baolin Peng']
|
2022-06-22
| null | null | null | null |
['open-domain-dialog']
|
['natural-language-processing']
|
[-1.59161985e-01 6.41465902e-01 -4.05216664e-02 -7.35159039e-01
-1.16216588e+00 -9.83787775e-01 1.09419835e+00 2.18994319e-01
-4.97977734e-01 9.99301910e-01 9.84716713e-01 -8.17003548e-02
2.25625739e-01 -5.22836864e-01 1.89997286e-01 -7.65961334e-02
1.98389158e-01 1.17018628e+00 3.73482317e-01 -1.04751706e+00
2.64241278e-01 1.96370780e-02 -6.71873987e-01 5.10934830e-01
6.67961359e-01 5.75705588e-01 -3.74158099e-02 8.22728395e-01
-4.46898311e-01 1.22461271e+00 -8.05620551e-01 -7.17187583e-01
4.71027941e-02 -6.01795673e-01 -1.61356568e+00 -1.79514792e-02
2.44857538e-02 -7.42052436e-01 -1.56823590e-01 4.79111224e-01
7.08904922e-01 6.59821451e-01 5.09623349e-01 -1.07827175e+00
-6.06759131e-01 8.81643236e-01 3.28920513e-01 -5.33275828e-02
6.84046328e-01 8.30673456e-01 1.29809332e+00 -9.62347209e-01
8.21398377e-01 1.74842453e+00 5.30657053e-01 9.18012261e-01
-1.51200509e+00 -3.81269865e-02 -2.50042945e-01 -3.51145536e-01
-5.72382092e-01 -9.11440670e-01 2.52459586e-01 -5.65341473e-01
1.21920013e+00 9.28526223e-02 -1.78053886e-01 1.33107281e+00
-1.19550191e-02 5.17399728e-01 1.04562795e+00 -5.23641706e-01
3.20792824e-01 5.01884162e-01 6.52712762e-01 3.84104401e-01
-6.02426589e-01 -1.47318259e-01 -6.91138089e-01 -5.36370456e-01
2.49297217e-01 -4.72921848e-01 -1.32091820e-01 -2.21844569e-01
-1.06043470e+00 1.29124916e+00 1.58072054e-01 1.65709570e-01
-2.24324375e-01 -3.07801813e-01 6.67230427e-01 5.55505335e-01
3.96441042e-01 1.03975558e+00 -4.71571416e-01 -6.75506175e-01
-2.39386976e-01 3.73007596e-01 1.70560884e+00 8.39793503e-01
8.28545272e-01 -2.76123255e-01 -8.77934992e-01 1.27250719e+00
9.48105305e-02 1.36592507e-01 4.87517864e-01 -1.36797976e+00
6.44869149e-01 7.59225607e-01 4.10400331e-01 -2.46347398e-01
-3.31164300e-01 2.51753032e-01 -1.93493605e-01 6.99046478e-02
7.05023885e-01 -5.62550843e-01 -2.63090193e-01 1.80171919e+00
2.96884298e-01 -6.23125315e-01 4.83880162e-01 7.98755765e-01
1.30039942e+00 6.24948263e-01 1.91776887e-01 -1.44522667e-01
1.24750233e+00 -1.33971465e+00 -7.89187372e-01 -3.72519612e-01
7.96539426e-01 -7.22201049e-01 1.49328721e+00 1.04069896e-01
-1.17049539e+00 -3.62387657e-01 -6.05486810e-01 -3.71114820e-01
-2.34927207e-01 -4.14293230e-01 4.53806996e-01 5.45864582e-01
-1.13946474e+00 2.69690931e-01 -3.65651965e-01 -6.56693876e-01
-3.19667578e-01 1.62264526e-01 -2.13072360e-01 2.26480827e-01
-1.40990627e+00 1.46040201e+00 8.69862288e-02 -4.34538871e-01
-1.08062506e+00 -3.18408251e-01 -9.44020748e-01 2.64983296e-01
6.12859249e-01 -5.33622265e-01 2.27118993e+00 -4.34209108e-01
-2.00911069e+00 7.97227502e-01 1.42477021e-01 -6.12065136e-01
4.94603693e-01 -2.29048058e-01 -7.47775799e-03 2.97076553e-02
8.78140554e-02 8.92667770e-01 2.85843313e-01 -9.40814912e-01
-4.54093367e-01 2.40928847e-02 7.77395666e-01 5.32949924e-01
-4.03899461e-01 2.00665012e-01 -1.58097103e-01 -5.12031429e-02
-7.10817516e-01 -9.88917708e-01 -4.04009819e-01 -3.15593541e-01
-1.40485674e-01 -6.86658323e-01 6.37918949e-01 -5.53170443e-01
1.10313094e+00 -1.85192132e+00 1.42155424e-01 -3.45860273e-01
3.84687871e-01 1.46034926e-01 -2.71024078e-01 1.22688746e+00
5.28039753e-01 -7.99143314e-03 -2.24778876e-01 -6.43583000e-01
3.77231479e-01 2.63979107e-01 -2.77546436e-01 -1.01314597e-01
1.34671092e-01 9.72782314e-01 -9.73664165e-01 -5.43970108e-01
2.95896083e-01 -1.03723384e-01 -6.37821496e-01 8.10543776e-01
-7.57736564e-01 7.78709292e-01 -4.12906080e-01 1.03313752e-01
-1.37812207e-02 -2.48944119e-01 9.81421620e-02 2.90150255e-01
-1.08040676e-01 1.03201330e+00 -5.50051153e-01 1.97106230e+00
-7.00163305e-01 3.88917744e-01 3.70853037e-01 -1.25503257e-01
1.08408284e+00 6.76502764e-01 1.48203507e-01 -3.97500366e-01
2.74580270e-01 -1.66336909e-01 1.22088492e-01 -4.27727520e-01
9.85750377e-01 -2.59279832e-02 -5.83677769e-01 1.11982405e+00
6.36158943e-01 -3.28633994e-01 2.30943382e-01 8.51837873e-01
1.25646222e+00 -2.50334501e-01 5.92632651e-01 -1.80113211e-01
5.40638745e-01 2.95662910e-01 2.15022892e-01 9.26008940e-01
-4.02267277e-01 1.67378455e-01 6.17017150e-01 -5.82268164e-02
-8.03475618e-01 -7.55342066e-01 9.08841118e-02 1.78462350e+00
-1.10306785e-01 -5.58866501e-01 -8.75931025e-01 -7.77108371e-01
-1.19606577e-01 1.12063837e+00 -3.93377781e-01 -7.41774067e-02
-3.75151783e-01 -1.69535756e-01 7.98678219e-01 4.04505342e-01
4.20029342e-01 -1.39321005e+00 -5.39282858e-01 3.86095583e-01
-4.63645190e-01 -1.20982814e+00 -6.95856929e-01 1.42432645e-01
-4.72397655e-01 -8.09825242e-01 -3.42499971e-01 -4.61437047e-01
-9.56261829e-02 1.11352801e-01 1.59482288e+00 -1.51886493e-01
1.73660815e-01 8.83223593e-01 -7.14606225e-01 -1.39817044e-01
-1.13399076e+00 1.99645698e-01 -2.17219621e-01 -4.12048638e-01
5.93369365e-01 -1.20269060e-01 -2.45636463e-01 6.22960031e-01
-4.27470267e-01 -6.69355914e-02 2.70539671e-01 1.21724546e+00
-4.49686050e-01 -8.36226165e-01 8.70107114e-01 -1.01920807e+00
1.59412050e+00 -4.47858125e-01 -3.25186998e-01 3.67643446e-01
-5.16123712e-01 1.68770701e-01 4.03261483e-01 -2.47992828e-01
-1.54948008e+00 -3.15702468e-01 -2.91281849e-01 1.17134556e-01
-3.01036865e-01 5.12330115e-01 5.31829596e-02 1.72466874e-01
1.12810528e+00 -1.95560664e-01 2.61744559e-01 -5.29273629e-01
8.22693229e-01 1.08109105e+00 5.62090158e-01 -9.80908811e-01
3.07785600e-01 -8.53436291e-02 -7.06602097e-01 -5.97991824e-01
-7.40785778e-01 -7.43649542e-01 -6.89295888e-01 -1.64055184e-01
9.02229786e-01 -7.92707980e-01 -9.41360712e-01 -5.80636179e-03
-1.36127961e+00 -9.15114999e-01 -4.21947092e-01 1.59505799e-01
-5.54526448e-01 2.64901459e-01 -1.09655344e+00 -1.10271907e+00
-6.12007082e-01 -1.12643600e+00 9.56219673e-01 1.05917484e-01
-8.67301822e-01 -1.29115534e+00 4.40605670e-01 6.70491993e-01
5.19408345e-01 -1.76065460e-01 7.60469139e-01 -1.43146133e+00
-1.84903756e-01 -5.50424084e-02 4.72154878e-02 4.10990626e-01
1.48173710e-02 -4.93448436e-01 -1.18399000e+00 -2.33068958e-01
1.74153879e-01 -1.24494290e+00 3.68795097e-01 -1.69581801e-01
1.01283956e-02 -6.35021687e-01 1.12151273e-01 -3.03844035e-01
7.10564435e-01 2.87142932e-01 2.39115879e-01 9.95668620e-02
9.11214277e-02 1.00115693e+00 8.33408952e-01 6.76295042e-01
5.60959995e-01 9.03189898e-01 6.39825985e-02 8.31270069e-02
4.06984538e-02 -1.72989339e-01 4.58896548e-01 6.03994310e-01
3.79951000e-01 -4.98915732e-01 -1.02435410e+00 5.31509995e-01
-1.96614349e+00 -8.31936300e-01 4.54797782e-02 2.04336524e+00
1.22699559e+00 1.10272631e-01 5.40893614e-01 -6.82357728e-01
5.17961383e-01 1.44188523e-01 -4.39750046e-01 -8.82980108e-01
1.39287725e-01 2.70163268e-01 -2.90240049e-02 1.08844483e+00
-6.74355268e-01 1.20967770e+00 6.35690212e+00 3.04795265e-01
-5.18283248e-01 4.98641968e-01 5.34698248e-01 1.96686313e-01
-1.85528040e-01 3.41241121e-01 -9.30936038e-01 1.25623986e-01
1.20857024e+00 -4.31069702e-01 5.18408895e-01 9.49821889e-01
1.47570133e-01 1.58032926e-04 -1.52621853e+00 3.13504905e-01
-4.40986454e-02 -1.23666787e+00 -2.60357738e-01 3.80621776e-02
3.49529713e-01 1.79498903e-02 -3.97836864e-01 9.72463906e-01
1.27047956e+00 -9.14514124e-01 2.57578224e-01 2.00104073e-01
5.98667800e-01 -2.43668064e-01 8.01814437e-01 6.69973314e-01
-6.52972102e-01 -5.73942885e-02 -1.62321106e-01 -2.18018457e-01
4.82284218e-01 -2.99457818e-01 -1.73602843e+00 2.64770687e-01
2.38066226e-01 2.70152628e-01 -3.54263365e-01 2.77583867e-01
-3.44091147e-01 5.57003319e-01 4.82193865e-02 -1.63245484e-01
6.05719388e-01 -6.43206388e-02 6.62208617e-01 1.47526538e+00
-3.06752563e-01 4.72269624e-01 6.40463114e-01 9.63329732e-01
-2.08861887e-01 1.25853181e-01 -5.96801877e-01 -1.32123381e-01
8.47835064e-01 1.48906159e+00 -1.10960444e-02 -4.76173371e-01
-4.53350484e-01 6.73431396e-01 4.07656431e-01 1.17817499e-01
-3.84439468e-01 -2.03154922e-01 4.74238664e-01 -2.04359815e-01
-2.22982764e-01 -1.73703551e-01 -1.24731727e-01 -9.14630830e-01
-5.03671527e-01 -1.21175885e+00 6.02645159e-01 -6.93745494e-01
-1.57572353e+00 8.54575217e-01 4.52941656e-02 -8.73460829e-01
-1.08261740e+00 -3.80844593e-01 -8.27613413e-01 1.09662998e+00
-7.49166787e-01 -9.98463333e-01 -4.59908515e-01 5.73609710e-01
1.13339567e+00 -4.12562251e-01 1.32290459e+00 -2.10354313e-01
-2.28399530e-01 5.43397546e-01 -3.02086294e-01 3.74986708e-01
1.34073496e+00 -1.34825861e+00 6.08448029e-01 3.25371772e-01
-2.56056577e-01 7.26374209e-01 9.00563419e-01 -4.53448355e-01
-1.06303394e+00 -6.88466668e-01 8.80126655e-01 -9.87294793e-01
8.30490112e-01 -5.05888104e-01 -9.23786104e-01 8.29583049e-01
9.15727317e-01 -6.91652179e-01 8.11423838e-01 6.37767553e-01
-2.48930514e-01 3.41816276e-01 -1.28682435e+00 5.30658424e-01
7.52420664e-01 -6.83859408e-01 -1.18066084e+00 3.67793858e-01
1.05363250e+00 -5.60891867e-01 -1.09138811e+00 1.46130711e-01
3.29978645e-01 -9.96003687e-01 7.86710441e-01 -8.72132599e-01
4.54714209e-01 4.24628407e-01 -2.33719647e-01 -1.47556436e+00
-1.96936220e-01 -1.09239185e+00 1.10063285e-01 1.48122096e+00
5.12453079e-01 -5.30442774e-01 3.43720615e-01 1.25552297e+00
-4.72365677e-01 -3.63153011e-01 -6.26220763e-01 -6.12996042e-01
1.86634898e-01 1.23273678e-01 3.83905709e-01 9.67165589e-01
8.45594823e-01 1.37376988e+00 -4.80730772e-01 -4.81371760e-01
1.05940774e-01 -7.03323632e-02 1.44767857e+00 -1.29597962e+00
-5.21968365e-01 -3.31213057e-01 1.29438281e-01 -1.39145434e+00
2.98734099e-01 -6.22921407e-01 3.67506713e-01 -1.64631307e+00
9.25245136e-02 -4.37946439e-01 2.26111487e-01 6.57168925e-01
-1.56012237e-01 -3.13836694e-01 2.97147274e-01 3.58178020e-01
-8.70612979e-01 7.89927483e-01 9.92890716e-01 -3.05798892e-02
-6.63352191e-01 5.81013411e-02 -8.42168272e-01 5.17988920e-01
6.01381063e-01 -3.16218436e-01 -4.11735535e-01 -2.12121636e-01
-4.32107419e-01 7.84034610e-01 1.10402018e-01 -6.74890339e-01
4.10363615e-01 -2.89974093e-01 -5.10598540e-01 -2.48306692e-01
9.67848420e-01 -2.56256521e-01 -4.81490493e-01 9.23661515e-02
-1.09778810e+00 -1.03277399e-03 1.69003248e-01 3.16230655e-01
-2.25117832e-01 -5.81320345e-01 7.14038253e-01 -4.28039491e-01
-6.54465675e-01 -1.66926876e-01 -4.61258233e-01 5.48445523e-01
7.21961558e-01 1.12368979e-01 -9.31812108e-01 -1.15361309e+00
-7.48933017e-01 5.77823997e-01 4.25689042e-01 4.84315485e-01
2.04590887e-01 -7.97904134e-01 -8.25763643e-01 -4.16972786e-01
3.77657831e-01 -2.71833450e-01 -5.89129403e-02 6.12789512e-01
-2.09751725e-02 6.71526790e-01 -2.05768347e-01 -5.38475454e-01
-1.22423053e+00 1.76751807e-01 2.64966697e-01 -7.40792632e-01
-4.24594969e-01 7.42450535e-01 3.26562196e-01 -9.82631207e-01
3.81848603e-01 -2.02679709e-02 -1.02697365e-01 2.89172828e-02
4.80726093e-01 1.67633682e-01 -1.72425117e-02 -4.69779909e-01
-1.43920546e-02 -5.62685370e-01 -2.05751985e-01 -8.50727975e-01
1.12333441e+00 -3.08809698e-01 -1.42496508e-02 7.16446221e-01
6.30610049e-01 1.05704658e-03 -1.20877874e+00 -6.98758781e-01
2.40882769e-01 -2.23338529e-01 -4.49385017e-01 -1.29580986e+00
-2.06614733e-02 9.19580877e-01 1.11835986e-01 4.08296376e-01
4.71181899e-01 1.50083974e-01 7.30154455e-01 1.07985806e+00
5.04459321e-01 -1.14544129e+00 6.16408408e-01 1.16146052e+00
1.02313912e+00 -1.47940016e+00 -2.68849432e-01 -1.83126464e-01
-1.51209176e+00 9.74736512e-01 1.09253693e+00 3.81191611e-01
1.30532235e-01 2.11944997e-01 4.66250479e-01 -2.30946302e-01
-1.29575193e+00 -1.66038141e-01 -1.62394289e-02 3.51830572e-01
8.28357100e-01 -1.14337713e-01 -2.82885224e-01 6.91069961e-01
-3.66667569e-01 -2.77457446e-01 6.55111134e-01 9.47982430e-01
-7.34293699e-01 -1.22902453e+00 -1.28151238e-01 1.60656229e-01
-8.40731114e-02 -1.76073194e-01 -1.09329295e+00 7.28555560e-01
-6.96304858e-01 1.65797698e+00 -9.03307423e-02 -3.88000816e-01
3.43847632e-01 4.88743007e-01 4.79995124e-02 -1.29991066e+00
-1.39361966e+00 -7.94251859e-02 9.62228835e-01 -3.56493533e-01
-2.26649612e-01 -2.42184356e-01 -1.08677495e+00 -4.19780135e-01
-4.66614723e-01 7.28504717e-01 2.93706417e-01 9.89398658e-01
2.80806661e-01 1.67035177e-01 5.82933724e-01 -6.29343688e-01
-1.25881398e+00 -1.64648080e+00 -4.09039296e-02 5.73663771e-01
1.75665282e-02 -5.52673936e-01 -1.65799499e-01 -7.53807947e-02]
|
[12.790836334228516, 8.024829864501953]
|
1aecadcb-05cf-45d6-9732-b7c45df48fde
|
robustnet-improving-domain-generalization-in
|
2103.15597
| null |
https://arxiv.org/abs/2103.15597v2
|
https://arxiv.org/pdf/2103.15597v2.pdf
|
RobustNet: Improving Domain Generalization in Urban-Scene Segmentation via Instance Selective Whitening
|
Enhancing the generalization capability of deep neural networks to unseen domains is crucial for safety-critical applications in the real world such as autonomous driving. To address this issue, this paper proposes a novel instance selective whitening loss to improve the robustness of the segmentation networks for unseen domains. Our approach disentangles the domain-specific style and domain-invariant content encoded in higher-order statistics (i.e., feature covariance) of the feature representations and selectively removes only the style information causing domain shift. As shown in Fig. 1, our method provides reasonable predictions for (a) low-illuminated, (b) rainy, and (c) unseen structures. These types of images are not included in the training dataset, where the baseline shows a significant performance drop, contrary to ours. Being simple yet effective, our approach improves the robustness of various backbone networks without additional computational cost. We conduct extensive experiments in urban-scene segmentation and show the superiority of our approach to existing work. Our code is available at https://github.com/shachoi/RobustNet.
|
['Jaegul Choo', 'Seungryong Kim', 'Joanne Kim', 'Huiwon Yun', 'Sanghun Jung', 'Sungha Choi']
|
2021-03-29
| null |
http://openaccess.thecvf.com//content/CVPR2021/html/Choi_RobustNet_Improving_Domain_Generalization_in_Urban-Scene_Segmentation_via_Instance_Selective_CVPR_2021_paper.html
|
http://openaccess.thecvf.com//content/CVPR2021/papers/Choi_RobustNet_Improving_Domain_Generalization_in_Urban-Scene_Segmentation_via_Instance_Selective_CVPR_2021_paper.pdf
|
cvpr-2021-1
|
['scene-segmentation']
|
['computer-vision']
|
[ 3.56386036e-01 -8.26810598e-02 1.33773714e-01 -5.43866754e-01
-3.93152982e-01 -7.73317754e-01 5.63180208e-01 -1.96968675e-01
-5.02225876e-01 8.48821938e-01 -1.12321004e-01 -3.34215015e-01
9.07274242e-03 -5.71627736e-01 -7.63356090e-01 -1.00979638e+00
1.14914417e-01 8.20844546e-02 7.25817204e-01 -3.23954314e-01
7.22752884e-02 6.79692090e-01 -1.62394631e+00 1.59362763e-01
1.09491515e+00 9.65248227e-01 2.55622685e-01 4.05893952e-01
2.11020559e-01 4.54885066e-01 -5.93785822e-01 -1.28809452e-01
5.78482985e-01 -3.02564174e-01 -5.30288815e-01 1.38437197e-01
8.66817355e-01 -3.27438504e-01 -3.77689302e-01 1.30076516e+00
4.55320686e-01 3.11988354e-01 4.96085554e-01 -1.18736780e+00
-4.22771990e-01 2.44771078e-01 -5.65472662e-01 3.98017347e-01
-3.56489152e-01 3.10405582e-01 7.21181393e-01 -7.48035133e-01
6.32390857e-01 1.25031722e+00 5.03719032e-01 7.17225492e-01
-1.32562220e+00 -9.36131537e-01 4.21961844e-01 2.51486331e-01
-1.25756145e+00 -6.13642395e-01 9.45282400e-01 -2.73650438e-01
5.80925763e-01 7.31228217e-02 2.60359228e-01 1.24984896e+00
1.83954746e-01 7.89911270e-01 1.21414375e+00 -8.97064246e-03
4.00119334e-01 1.37617752e-01 2.40084022e-01 3.29932958e-01
5.89651525e-01 3.76482099e-01 -2.72047549e-01 1.33922026e-01
4.89224970e-01 -1.55819744e-01 -2.61689484e-01 -6.40675366e-01
-9.70847666e-01 7.94854939e-01 7.55381465e-01 1.67799398e-01
-8.06811005e-02 -3.25867422e-02 1.35589749e-01 1.99702606e-01
6.61844075e-01 2.78939247e-01 -6.33131921e-01 1.62053660e-01
-8.78465772e-01 2.81800479e-01 5.11656821e-01 9.49704170e-01
8.14646780e-01 4.12886173e-01 1.01564609e-01 9.48619902e-01
2.52574682e-02 7.37709880e-01 2.82236755e-01 -9.76367414e-01
4.36314911e-01 3.58760208e-01 -1.95945222e-02 -8.38746071e-01
-6.00729108e-01 -6.41511977e-01 -6.61322892e-01 3.52853358e-01
6.69347227e-01 -4.19638842e-01 -1.36042404e+00 1.81099105e+00
2.23426551e-01 3.43409538e-01 1.28882259e-01 1.01017928e+00
9.41424370e-01 3.27821165e-01 4.97106388e-02 3.41487348e-01
1.16717207e+00 -8.72098088e-01 -3.38962466e-01 -7.44848669e-01
5.03697574e-01 -6.27635360e-01 9.90470827e-01 1.44042060e-01
-7.06181407e-01 -5.89708626e-01 -1.11753643e+00 7.82847926e-02
-4.96735096e-01 -1.25451470e-02 3.93661648e-01 6.77877605e-01
-1.01495516e+00 5.37612617e-01 -8.15393448e-01 -6.31999075e-01
7.54397929e-01 3.36760312e-01 -2.66363889e-01 -2.24540785e-01
-1.25743389e+00 7.79419780e-01 6.17387295e-01 -3.21739949e-02
-8.42652142e-01 -6.46073937e-01 -9.09135044e-01 -1.54325113e-01
4.18509215e-01 -3.75075758e-01 1.05843151e+00 -1.13416016e+00
-1.22074044e+00 7.95090497e-01 -3.46364647e-01 -5.55226386e-01
6.02294266e-01 -5.00689089e-01 -3.72591704e-01 1.57000571e-01
5.19395210e-02 1.10242116e+00 1.02648616e+00 -1.61495078e+00
-7.12739289e-01 -4.44810241e-01 -2.57352274e-02 1.73898950e-01
-1.41752169e-01 -1.92772597e-01 -2.82909691e-01 -7.61563301e-01
2.88769811e-01 -1.18105149e+00 -1.92039102e-01 -1.71736740e-02
-4.70080316e-01 1.93591148e-01 1.09778702e+00 -5.42432249e-01
8.01785648e-01 -2.45758700e+00 -2.75672287e-01 1.79498687e-01
1.37123838e-01 4.37136918e-01 -3.32400858e-01 9.11206659e-03
-1.38288468e-01 9.25470591e-02 -6.91290975e-01 1.10376678e-01
-1.51130661e-01 4.85621601e-01 -3.89980644e-01 6.41280472e-01
5.41639209e-01 7.25011230e-01 -7.31232762e-01 -2.43449211e-01
2.57817894e-01 3.95928830e-01 -3.79693985e-01 -1.62210017e-01
-9.58595499e-02 7.82968521e-01 -4.72128302e-01 6.16439521e-01
1.10809171e+00 4.66799811e-02 -1.58481047e-01 -1.05886057e-01
1.91382226e-02 1.92650408e-01 -1.05096710e+00 1.43115807e+00
-1.04531489e-01 9.36818659e-01 2.18172595e-01 -8.93186569e-01
9.43076432e-01 -1.65664822e-01 9.25238058e-02 -1.01904356e+00
8.74867439e-02 1.74555779e-01 3.56592655e-01 -2.48453170e-01
5.37241161e-01 -9.74128023e-02 1.61051340e-02 1.73646305e-02
-7.70468190e-02 -6.28889427e-02 1.47019073e-01 6.76230788e-02
9.82272446e-01 1.44565910e-01 -3.45854871e-02 -5.68380058e-01
3.52392673e-01 1.39652580e-01 7.46178746e-01 8.70380402e-01
-6.14946663e-01 7.68293798e-01 3.45471889e-01 -2.39325076e-01
-9.90289032e-01 -1.17330265e+00 -4.72322404e-01 1.13065886e+00
5.97074807e-01 2.73726076e-01 -7.85569072e-01 -6.84819579e-01
2.09434614e-01 8.89747143e-01 -5.58066487e-01 -3.54605198e-01
-6.18018031e-01 -7.71259427e-01 7.11099327e-01 6.25131309e-01
8.86938870e-01 -9.04969275e-01 -5.41572452e-01 -1.80010051e-01
-2.04915449e-01 -1.45037210e+00 -1.36015102e-01 3.59155238e-01
-1.01471341e+00 -8.18388820e-01 -5.74971318e-01 -6.90496147e-01
5.63482344e-01 5.02481282e-01 1.06648409e+00 -7.18205944e-02
-1.35588869e-01 9.08987522e-02 -2.91289657e-01 -5.47921598e-01
-1.80640668e-01 1.98434636e-01 -7.26359636e-02 -3.76068763e-02
4.31230128e-01 -3.95804971e-01 -5.96005023e-01 6.11308455e-01
-1.00108755e+00 -3.79314601e-01 5.05298078e-01 8.41938972e-01
4.37415689e-01 2.08923966e-01 5.38448691e-01 -8.79320443e-01
4.24772054e-01 -4.52128708e-01 -6.85528457e-01 -1.98755980e-01
-3.37312460e-01 4.03920896e-02 5.83167493e-01 -4.35756266e-01
-1.21312058e+00 1.85004666e-01 -7.94620141e-02 -2.57845670e-01
-7.22651064e-01 2.75480878e-02 -2.65971750e-01 -2.45760411e-01
7.42253184e-01 1.06514916e-01 -2.57740885e-01 -5.17846942e-01
2.88937867e-01 4.13470209e-01 5.83683014e-01 -4.81932461e-01
1.23039949e+00 8.55428636e-01 -4.93069030e-02 -1.18338370e+00
-7.45052695e-01 -4.82566625e-01 -7.94700325e-01 -4.08471934e-02
7.62192607e-01 -1.04747486e+00 -3.66174206e-02 5.81031382e-01
-9.50466990e-01 -5.80567241e-01 -2.01085672e-01 3.75258476e-01
-3.33568007e-01 2.22405180e-01 -3.25563431e-01 -5.92190444e-01
1.00694306e-01 -1.17023861e+00 1.01854897e+00 3.82694572e-01
-1.15797579e-01 -1.03120887e+00 -2.26481199e-01 2.93064296e-01
3.52688640e-01 3.82713646e-01 7.87935138e-01 -8.47024500e-01
-4.44680423e-01 3.14111747e-02 -5.34628272e-01 5.00910461e-01
-4.42398787e-02 3.04741203e-03 -1.47990716e+00 -3.57286930e-01
-7.85731897e-02 -2.54319668e-01 1.43298662e+00 3.95916939e-01
1.10884118e+00 1.00986920e-01 -2.65129089e-01 7.85108626e-01
1.22447479e+00 2.07704067e-01 7.50265718e-01 4.54933643e-01
6.77262723e-01 7.52110362e-01 6.81236148e-01 1.75451145e-01
1.57731846e-01 4.19254690e-01 5.75667799e-01 -3.95479947e-01
-3.16717356e-01 6.67416826e-02 4.44624871e-01 7.86804631e-02
2.11826339e-01 -3.59871924e-01 -9.74409401e-01 8.06422353e-01
-1.62455881e+00 -9.25972521e-01 -2.52188355e-01 2.11708832e+00
5.77934206e-01 3.95930827e-01 -1.14156846e-02 -2.44815312e-02
7.67888725e-01 3.51747334e-01 -1.02179050e+00 -3.82933140e-01
-3.50262702e-01 2.35113725e-01 8.39761555e-01 3.85854453e-01
-1.49225843e+00 1.21397102e+00 5.57579422e+00 8.77055347e-01
-1.18932045e+00 -3.11511848e-02 5.78632355e-01 -1.69430092e-01
-1.06362641e-01 -1.99911207e-01 -7.89210141e-01 4.92130280e-01
8.42129588e-01 1.96621731e-01 1.27601117e-01 9.33612943e-01
1.69688359e-01 -1.76653832e-01 -8.44512284e-01 5.73135674e-01
-1.39721753e-02 -9.11591947e-01 -1.23999856e-01 2.52952836e-02
7.60058522e-01 6.19270205e-01 3.36477250e-01 2.53054351e-01
4.93544996e-01 -8.90959322e-01 7.23430097e-01 4.26519513e-02
5.49382508e-01 -6.68936849e-01 7.74017334e-01 2.80419320e-01
-7.70498157e-01 -1.38029471e-01 -5.57271302e-01 6.65263087e-02
-2.11304456e-01 6.50329471e-01 -7.21029758e-01 3.11182737e-01
9.75874960e-01 7.47697830e-01 -8.36870730e-01 1.08830702e+00
-3.73244286e-01 8.71688783e-01 -3.93481195e-01 3.19233954e-01
4.45918322e-01 -8.01684335e-02 8.61645341e-01 1.20202601e+00
1.04228593e-01 -9.50364321e-02 7.64123052e-02 8.87233734e-01
-5.14872447e-02 -3.96308631e-01 -7.54426897e-01 1.49762377e-01
3.13298762e-01 1.11340022e+00 -1.02835441e+00 -2.32706919e-01
-3.27005863e-01 8.46720994e-01 4.35680412e-02 8.13476503e-01
-8.44576836e-01 -5.27331293e-01 1.02235389e+00 5.85555024e-02
6.26798928e-01 -2.72969514e-01 -6.80108428e-01 -1.08545935e+00
5.42709865e-02 -7.35186219e-01 1.39812037e-01 -5.46250641e-01
-1.24436188e+00 6.21249795e-01 1.58426240e-01 -1.28181434e+00
1.34230465e-01 -8.20825696e-01 -5.12839496e-01 7.79518247e-01
-1.89252019e+00 -9.05510426e-01 -5.05079627e-01 5.90889752e-01
5.75842500e-01 -7.86675587e-02 3.33017409e-01 2.16962963e-01
-7.40704417e-01 5.79351723e-01 4.00483608e-01 2.30581209e-01
9.48510587e-01 -1.09117401e+00 6.99684739e-01 1.21175539e+00
-1.15345851e-01 4.21329707e-01 8.66608024e-01 -6.00261509e-01
-7.08037078e-01 -1.42759430e+00 4.94491726e-01 -4.70230967e-01
6.67136848e-01 -4.68010485e-01 -1.19886291e+00 4.22951072e-01
1.53446198e-01 1.46121606e-01 4.35640663e-01 -1.45632997e-01
-4.37869579e-01 -2.29835302e-01 -1.24120772e+00 7.00692356e-01
1.05055988e+00 -1.70488968e-01 -5.30483902e-01 6.92687556e-02
7.22730279e-01 -4.74222571e-01 -3.48500878e-01 6.11816406e-01
4.30241734e-01 -1.13638711e+00 9.57893968e-01 -5.01553237e-01
2.98145711e-01 -2.99917638e-01 -2.57157445e-01 -1.35851836e+00
-3.89144242e-01 -1.41351357e-01 2.50715882e-01 8.83336365e-01
5.15859067e-01 -1.02311730e+00 8.07613909e-01 5.32558799e-01
-2.69035071e-01 -3.82103384e-01 -9.05873954e-01 -1.10223699e+00
4.32076037e-01 -3.73158306e-01 4.70631272e-01 7.89570272e-01
-7.32190311e-01 1.56964213e-01 -1.04471602e-01 5.09298265e-01
6.32775187e-01 1.23684190e-01 6.22349739e-01 -1.43770003e+00
4.34939004e-02 -3.98019731e-01 -4.05946881e-01 -9.69100893e-01
2.04891697e-01 -7.61776924e-01 2.44490415e-01 -1.33626711e+00
-6.21331334e-02 -3.65698010e-01 -3.89317364e-01 6.08393490e-01
-1.56236306e-01 4.57263470e-01 3.84494543e-01 7.50528499e-02
-5.70531666e-01 4.72248107e-01 1.32309091e+00 -2.10690334e-01
-1.53795213e-01 4.42185774e-02 -7.73763716e-01 9.47635174e-01
1.37266839e+00 -6.11337602e-01 -5.22916317e-01 -6.52614415e-01
-2.49043435e-01 -6.26226246e-01 6.30546093e-01 -1.28594482e+00
-2.46984750e-01 -2.72230059e-01 6.11735225e-01 -3.98299068e-01
2.00098425e-01 -9.37339365e-01 -4.85167652e-01 4.44570214e-01
-2.05982119e-01 -2.16331244e-01 7.51452267e-01 5.93732476e-01
-1.81121841e-01 -5.47789447e-02 1.23896432e+00 5.85306026e-02
-1.18228233e+00 1.15679964e-01 -2.94597954e-01 4.36954111e-01
9.38861191e-01 -5.35015941e-01 -6.76013947e-01 -3.22574139e-01
-5.31523585e-01 2.95891792e-01 6.16883814e-01 5.13265073e-01
5.58060229e-01 -9.05229092e-01 -6.97394073e-01 4.95422751e-01
2.32570991e-01 1.67649612e-01 2.84712106e-01 6.69490814e-01
-4.56021786e-01 2.97023445e-01 -3.14617157e-01 -7.72208214e-01
-1.21456587e+00 2.15321675e-01 4.27786797e-01 9.29149762e-02
-5.25202870e-01 9.71056938e-01 6.50775015e-01 -4.92493331e-01
1.49079204e-01 -5.02340138e-01 -1.15905114e-01 -3.37419584e-02
2.69690782e-01 3.29355806e-01 2.45197430e-01 -8.54628503e-01
-5.60393393e-01 4.65372145e-01 -2.87103653e-01 1.62387028e-01
1.18425310e+00 -1.52140588e-01 1.72176287e-01 2.34746724e-01
1.15894675e+00 -2.28192806e-01 -1.71025407e+00 -2.36686900e-01
-4.46030423e-02 -3.91638756e-01 2.00512379e-01 -8.76097679e-01
-1.25861466e+00 1.12323058e+00 8.73390019e-01 -5.02540804e-02
1.24245775e+00 -9.89222601e-02 9.16935444e-01 4.20018196e-01
1.78771257e-01 -1.16800058e+00 -1.50945768e-01 7.75689125e-01
7.13833392e-01 -1.45527542e+00 -9.25392136e-02 -3.29416811e-01
-7.88359940e-01 7.96264112e-01 7.94641197e-01 -2.71320879e-01
5.70214748e-01 1.90103322e-01 4.00558859e-01 -1.28081935e-02
-5.32558501e-01 -6.29672706e-01 3.18265647e-01 8.38209569e-01
3.52051631e-02 2.98137777e-02 5.47510758e-02 2.25425825e-01
-2.28565082e-01 -4.20644850e-01 4.73369032e-01 8.96560848e-01
-6.82655334e-01 -7.79629111e-01 -4.72580433e-01 3.09891075e-01
-2.70256698e-01 -1.17270216e-01 -6.64363325e-01 9.31180537e-01
3.30118090e-01 1.04385448e+00 1.93469629e-01 -4.31774288e-01
4.88868237e-01 8.64456072e-02 8.92013833e-02 -4.97787684e-01
-3.07064950e-01 6.74664676e-02 4.75042462e-02 -5.99094510e-01
-3.61677796e-01 -7.02936590e-01 -1.30790412e+00 -1.89100251e-01
-9.58539844e-02 -2.91092575e-01 4.74999011e-01 7.22813189e-01
4.85775322e-01 5.56421161e-01 3.34669262e-01 -8.41893673e-01
-4.00438458e-01 -6.70155823e-01 -5.53372324e-01 6.28037870e-01
7.02164114e-01 -9.20310676e-01 -4.74782526e-01 2.03152318e-02]
|
[9.022075653076172, -0.6366678476333618]
|
def3f7e9-8a2b-4b4c-a3a7-e0d590bc7f21
|
filter-sharing-efficient-learning-of
|
1612.02575
| null |
http://arxiv.org/abs/1612.02575v1
|
http://arxiv.org/pdf/1612.02575v1.pdf
|
Filter sharing: Efficient learning of parameters for volumetric convolutions
|
Typical convolutional neural networks (CNNs) have several millions of
parameters and require a large amount of annotated data to train them. In
medical applications where training data is hard to come by, these
sophisticated machine learning models are difficult to train. In this paper, we
propose a method to reduce the inherent complexity of CNNs during training by
exploiting the significant redundancy that is noticed in the learnt CNN
filters. Our method relies on finding a small set of filters and mixing
coefficients to derive every filter in each convolutional layer at the time of
training itself, thereby reducing the number of parameters to be trained. We
consider the problem of 3D lung nodule segmentation in CT images and
demonstrate the effectiveness of our method in achieving good results with only
few training examples.
|
['Sheshadri Thiruvenkadam', 'Rahul Venkataramani', 'Hariharan Ravishankar', 'Vivek Vaidya', 'Prasad Sudhakar']
|
2016-12-08
| null | null | null | null |
['lung-nodule-segmentation']
|
['medical']
|
[ 3.13513875e-01 2.82603264e-01 1.17555276e-01 -4.15951371e-01
-3.06653857e-01 -4.44681138e-01 1.32834375e-01 1.37727499e-01
-7.30487168e-01 3.91308188e-01 -2.71598727e-01 -6.70214713e-01
8.80967006e-02 -6.94957614e-01 -8.02007973e-01 -5.75914741e-01
-5.27124777e-02 2.09964111e-01 4.43559170e-01 -4.42012027e-02
-3.27859789e-01 7.13028014e-01 -1.07625425e+00 3.85450244e-01
6.62572265e-01 1.07053041e+00 3.34907949e-01 9.68329966e-01
-1.03598580e-01 9.83316839e-01 -5.66409111e-01 -1.55057371e-01
5.12985051e-01 -4.37859297e-01 -8.02130222e-01 3.11994791e-01
2.70993710e-01 -2.99097151e-01 -5.13410211e-01 8.39161396e-01
6.03208482e-01 5.13862148e-02 3.92466247e-01 -6.77633345e-01
9.85525623e-02 4.70687956e-01 -1.30008653e-01 3.71236652e-01
-5.43875039e-01 1.96285844e-01 6.59904003e-01 -6.43487215e-01
3.20412666e-01 7.95273125e-01 9.32675958e-01 7.47868776e-01
-1.08216250e+00 -4.83233631e-01 -1.57094821e-01 -8.56140405e-02
-1.26933634e+00 -2.99978465e-01 5.37931085e-01 -3.89359027e-01
7.59783089e-01 1.73630655e-01 8.36111367e-01 6.30255640e-01
-3.25268717e-03 5.94615698e-01 6.13131821e-01 -3.17135155e-01
1.69336349e-01 1.86203107e-01 -4.70228679e-03 1.15785480e+00
3.84855598e-01 -4.95262854e-02 -1.02984101e-01 -5.18762209e-02
1.29459977e+00 1.38778880e-01 -3.00664425e-01 -4.59904894e-02
-9.30556834e-01 9.10721838e-01 8.76346052e-01 5.53957343e-01
-3.40997815e-01 3.37888986e-01 2.96097308e-01 2.03372031e-01
2.79112995e-01 5.43303907e-01 -6.62514329e-01 3.93539101e-01
-9.07854199e-01 -1.23138405e-01 9.06386435e-01 8.88729870e-01
6.66927993e-01 -1.37001008e-01 -8.52110609e-02 8.14105570e-01
-5.74225746e-02 -3.70585956e-02 4.92807657e-01 -5.86654603e-01
3.78940672e-01 7.64098942e-01 -2.38796517e-01 -7.14874089e-01
-6.28209710e-01 -7.81012356e-01 -1.22221541e+00 8.15955773e-02
5.58434427e-01 -3.27594370e-01 -1.22132218e+00 1.27339327e+00
3.07524711e-01 3.80793214e-01 -5.53237647e-02 6.74096167e-01
8.93598735e-01 4.77513820e-01 -1.60913929e-01 2.32848134e-02
1.20804870e+00 -1.11511087e+00 -2.88292527e-01 -4.88471061e-01
7.66315758e-01 -6.65769637e-01 7.94886708e-01 3.02054763e-01
-1.06834126e+00 -6.62610352e-01 -1.00527549e+00 -1.07599653e-01
-1.82062119e-01 3.21406901e-01 7.22955823e-01 4.47274268e-01
-7.99313664e-01 8.22483838e-01 -1.01645303e+00 2.37468723e-02
8.35515320e-01 8.56345296e-01 -2.15253800e-01 -1.97954893e-01
-8.00465524e-01 7.57601261e-01 6.18088841e-01 5.01130164e-01
-8.73709500e-01 -7.69162416e-01 -5.88236272e-01 3.72174233e-01
3.38779628e-01 -8.45326662e-01 1.41517282e+00 -1.04499090e+00
-1.27237558e+00 5.01112580e-01 1.92154586e-01 -5.76272905e-01
6.39789104e-01 6.99051917e-02 1.37747452e-01 2.20042363e-01
-4.42637384e-01 6.57278061e-01 8.76270056e-01 -8.80111873e-01
-7.09712625e-01 1.05661508e-02 1.15009576e-01 -8.97459835e-02
-5.49975991e-01 -2.94129193e-01 -7.99789131e-01 -5.01503170e-01
3.15950513e-01 -1.02089930e+00 -1.01410925e+00 1.67598858e-01
-3.95478070e-01 -4.21592630e-02 6.37989104e-01 -3.38578105e-01
9.80696738e-01 -2.18668747e+00 -8.25338215e-02 4.17290807e-01
5.10649204e-01 5.56608379e-01 -3.18245851e-02 -1.82189062e-01
9.29915067e-03 1.63896635e-01 -3.23499292e-01 -1.10019021e-01
-5.37577808e-01 6.66558921e-01 3.14644307e-01 3.30004543e-01
5.57618022e-01 9.65131044e-01 -7.81986237e-01 -6.80640876e-01
1.60175845e-01 6.39152706e-01 -7.55997598e-01 4.23541754e-01
-1.65637001e-01 5.68364620e-01 -6.42645895e-01 2.57369757e-01
4.74670917e-01 -7.49728262e-01 3.02934200e-01 -3.65684420e-01
1.21119700e-01 2.86136240e-01 -1.06299973e+00 1.28188848e+00
-6.33602738e-01 6.73504770e-01 -5.25126383e-02 -1.16658676e+00
5.20494461e-01 4.75505978e-01 7.05714226e-01 -2.05482289e-01
6.10049903e-01 3.58499616e-01 4.18096364e-01 -6.13488376e-01
-1.53664365e-01 -1.35406882e-01 1.26848266e-01 3.35237414e-01
2.44747922e-01 -2.82970518e-01 2.44988352e-01 -1.17830604e-01
1.37508118e+00 -7.18212366e-01 2.61100084e-01 -2.02189296e-01
4.74112868e-01 9.01672095e-02 4.87538517e-01 7.16470003e-01
1.48027286e-01 5.19861639e-01 4.78648722e-01 -8.74752581e-01
-1.22072685e+00 -5.37141681e-01 -2.38031745e-02 4.91059214e-01
-4.13006127e-01 -7.50535429e-02 -9.31304514e-01 -8.10738444e-01
-2.25242719e-01 -8.01081508e-02 -7.01162100e-01 -6.21435530e-02
-9.60184455e-01 -8.00513864e-01 7.01944292e-01 7.04405010e-01
6.45799875e-01 -8.87141705e-01 -8.65626633e-01 3.89873773e-01
1.37548164e-01 -1.33133829e+00 -3.73074323e-01 5.86085200e-01
-1.50357664e+00 -1.32467341e+00 -6.88490510e-01 -9.85138774e-01
1.38979423e+00 1.88708976e-01 9.87110794e-01 7.94111788e-01
-5.86377144e-01 -2.61226058e-01 -1.23133771e-01 -5.83726048e-01
-3.50312173e-01 4.23139513e-01 -4.87429202e-01 -1.96088046e-01
-6.52236193e-02 -3.43725473e-01 -6.27292395e-01 1.61588028e-01
-1.05777407e+00 2.19343156e-01 9.79252338e-01 1.17614532e+00
5.79980135e-01 3.73917550e-01 6.74720034e-02 -1.29706454e+00
3.63553703e-01 -2.12942839e-01 -6.04209483e-01 4.50731106e-02
-2.67425239e-01 2.00662035e-02 8.26601744e-01 -7.19790459e-01
-6.34345889e-01 6.92023695e-01 -2.36282349e-01 -4.48405564e-01
-2.46399194e-02 4.65854645e-01 3.61940861e-01 -5.14507055e-01
7.30305612e-01 -1.55938640e-01 5.78918830e-02 -5.27312398e-01
1.37225434e-01 2.65433133e-01 4.36990976e-01 -3.53709348e-02
8.73176873e-01 4.14714307e-01 1.60608903e-01 -7.35635042e-01
-1.12160516e+00 -5.34491181e-01 -8.38319778e-01 -1.40997097e-01
7.85519123e-01 -7.62429893e-01 -8.13987672e-01 3.18088621e-01
-1.06481218e+00 -4.76612657e-01 -4.92764324e-01 7.49612927e-01
-2.48940602e-01 1.50742307e-02 -7.93694735e-01 -2.52450168e-01
-3.56829792e-01 -1.16037273e+00 4.50462401e-01 3.05319071e-01
3.72832902e-02 -1.01600182e+00 -3.27371269e-01 6.22896701e-02
5.57312965e-01 2.03448460e-01 1.03237736e+00 -6.08038306e-01
-7.10191667e-01 -4.72001493e-01 -2.44426325e-01 6.91567302e-01
2.33717978e-01 -1.54752910e-01 -9.42156196e-01 -3.43929440e-01
2.65284330e-01 -1.60793692e-01 9.44078982e-01 5.79328299e-01
1.65167379e+00 -2.96897113e-01 -2.67327279e-01 7.55399644e-01
1.45945358e+00 3.96371298e-02 5.58957338e-01 -3.18828262e-02
8.07613254e-01 3.41517091e-01 -6.49996521e-03 2.11425155e-01
-2.65637606e-01 1.46956071e-01 4.96801645e-01 -7.20054507e-01
-3.49460207e-02 1.00495368e-01 -2.77777940e-01 1.05357063e+00
-2.92159259e-01 -9.03020725e-02 -8.93153489e-01 5.12972534e-01
-1.64773810e+00 -4.73827422e-01 -1.46685451e-01 1.82015646e+00
8.89105856e-01 3.07994753e-01 -2.37949848e-01 2.78377235e-01
4.00684774e-01 -2.46092886e-01 -4.48493183e-01 -1.06859639e-01
2.58309633e-01 7.25094259e-01 8.07964683e-01 3.19525361e-01
-1.13215041e+00 8.59322608e-01 7.16629887e+00 5.78589857e-01
-1.22113490e+00 -5.94696589e-03 6.91147029e-01 -1.37370139e-01
1.33367673e-01 -2.47447714e-01 -4.54783738e-01 4.09549251e-02
8.66311014e-01 2.70714879e-01 1.69792101e-01 8.43643665e-01
8.36838409e-02 -8.64662901e-02 -1.11359799e+00 8.13741565e-01
-3.40282351e-01 -1.61843932e+00 -1.51482016e-01 4.88693677e-02
7.77121782e-01 1.33994475e-01 -3.59801762e-02 -9.78654101e-02
3.03933471e-01 -1.27461648e+00 2.89192915e-01 2.10876450e-01
5.29731631e-01 -7.08183110e-01 8.56786311e-01 4.38843191e-01
-9.93588567e-01 -2.48727113e-01 -6.81436658e-01 1.29099429e-01
-1.78018242e-01 8.30008626e-01 -1.36394691e+00 2.52059847e-01
6.17965996e-01 4.30791169e-01 -5.56172252e-01 1.42446804e+00
-2.16876373e-01 8.38744581e-01 -4.60004985e-01 -1.33638784e-01
5.78227460e-01 2.65741259e-01 -3.49356122e-02 1.13636327e+00
3.02212417e-01 3.60253185e-01 1.25055790e-01 6.01493239e-01
-3.71080309e-01 -6.87495386e-03 -2.86048084e-01 -2.58144699e-02
-1.14275841e-02 1.38793766e+00 -1.02358341e+00 -4.44393426e-01
-5.01564443e-01 7.23336101e-01 2.79053867e-01 1.18684389e-01
-7.42190540e-01 -1.90782592e-01 3.24857682e-01 6.17676526e-02
4.95142311e-01 -3.53255779e-01 -4.04589385e-01 -8.72059941e-01
-1.03867769e-01 -6.89189553e-01 3.19039434e-01 -2.86447853e-01
-1.08425725e+00 8.17427039e-01 -4.43232328e-01 -1.29296947e+00
3.81646119e-02 -9.71295774e-01 -5.45743883e-01 7.47616231e-01
-1.49176455e+00 -8.93144548e-01 -5.19706428e-01 7.99717665e-01
4.67987329e-01 -8.84802546e-03 6.78414106e-01 5.80564737e-01
-5.11502564e-01 3.57272089e-01 -7.92319104e-02 5.78576148e-01
3.89027953e-01 -9.77523983e-01 3.40471804e-01 6.42350554e-01
1.94018576e-02 4.18513060e-01 3.76627624e-01 -2.61097223e-01
-1.16049802e+00 -1.15109396e+00 8.73450220e-01 3.96171808e-02
3.85405332e-01 -2.70855099e-01 -8.88572156e-01 5.86215436e-01
-1.23822771e-01 4.87630218e-01 7.29795933e-01 -1.12458497e-01
-1.53814450e-01 -3.15518044e-02 -1.08632457e+00 3.24471384e-01
7.11064637e-01 -2.70188332e-01 -1.06880635e-01 5.52063942e-01
6.05255425e-01 -6.69227302e-01 -7.99684465e-01 3.96607220e-01
4.31028396e-01 -5.39083719e-01 9.75107193e-01 -8.19724679e-01
4.60977823e-01 -1.88209832e-01 3.57955337e-01 -1.16903222e+00
-3.37793827e-01 -3.24874699e-01 1.10931255e-01 3.99139643e-01
6.52985930e-01 -4.37448740e-01 1.08276963e+00 5.21495581e-01
-2.57035673e-01 -1.00676525e+00 -7.71533370e-01 -5.30451179e-01
-2.87134387e-02 -3.02718699e-01 3.75445187e-01 8.22162747e-01
-5.35395324e-01 1.05574973e-01 -1.92386925e-01 1.55011013e-01
3.05826753e-01 -2.71990001e-01 6.05487287e-01 -1.36295700e+00
-4.73997384e-01 -2.49343410e-01 -2.85145015e-01 -9.23725784e-01
-2.92142600e-01 -7.61388361e-01 2.62320191e-01 -1.35353649e+00
-4.77124844e-03 -7.54193306e-01 -2.70214349e-01 7.20937371e-01
-1.77881017e-01 3.55602741e-01 1.52662858e-01 1.26839504e-01
-1.48056448e-01 -1.14023335e-01 1.65780640e+00 -2.88622171e-01
-1.58787340e-01 4.14041787e-01 -5.48448741e-01 8.85400176e-01
9.35620308e-01 -8.46523285e-01 -3.77486408e-01 -6.58102810e-01
3.02325152e-02 -1.16544202e-01 4.03246373e-01 -1.14524233e+00
5.38251340e-01 1.57234799e-02 7.64273047e-01 -4.14888054e-01
1.61475509e-01 -1.03165364e+00 1.87297016e-01 9.85723853e-01
-3.47758979e-01 -1.98864669e-01 2.71538287e-01 3.06433707e-01
-2.70610660e-01 -6.44983590e-01 1.08238792e+00 -5.64502001e-01
-3.82834315e-01 6.32856429e-01 -4.09898996e-01 -2.21694797e-01
8.05595279e-01 9.28719565e-02 9.41867605e-02 5.14687859e-02
-9.47653174e-01 1.43318042e-01 -1.43332053e-02 -4.12952304e-02
5.90641618e-01 -1.13140762e+00 -4.65880632e-01 2.83183575e-01
-3.35308582e-01 6.89716339e-01 3.08989257e-01 8.00383389e-01
-8.73294055e-01 2.39918157e-01 -2.54966617e-01 -5.72294414e-01
-1.30638754e+00 4.84384865e-01 6.79143548e-01 -6.58573806e-01
-8.43029916e-01 1.21141362e+00 -2.03753784e-02 -1.81256995e-01
3.45204413e-01 -6.27908945e-01 -8.90378430e-02 -8.80741328e-02
4.55982983e-01 3.20311449e-02 3.04283917e-01 -1.58937767e-01
-1.39855385e-01 2.97105104e-01 -1.35117576e-01 2.77406275e-01
1.55542028e+00 4.88265038e-01 -5.96764125e-02 -1.13743842e-02
1.44251561e+00 -4.93968755e-01 -1.25601649e+00 -3.65232229e-01
5.53132631e-02 -2.91581929e-01 2.27017522e-01 -3.44666809e-01
-1.55171096e+00 1.19249392e+00 6.21276677e-01 1.99267775e-01
1.27160668e+00 -4.15295586e-02 9.82131124e-01 7.95446157e-01
7.21109146e-03 -1.10382116e+00 2.60450035e-01 5.35013795e-01
4.65009958e-01 -1.18108046e+00 -7.68781826e-02 -6.88695669e-01
-1.53988108e-01 1.50949478e+00 4.47252512e-01 -4.62779909e-01
1.01414299e+00 5.58321714e-01 8.85221735e-02 -2.77666360e-01
-6.25548482e-01 -3.88517946e-01 3.44811857e-01 1.62619367e-01
4.01913345e-01 -1.64147869e-01 -9.58736837e-02 3.58291656e-01
-3.48812342e-02 1.24796323e-01 4.82204169e-01 1.04072189e+00
-5.73991179e-01 -1.13177288e+00 -6.94956630e-02 6.34668052e-01
-6.93315268e-01 -1.58766687e-01 -1.26599625e-01 7.15141416e-01
2.87858993e-01 6.05416417e-01 6.46274537e-02 -2.92868406e-01
4.22983080e-01 -1.72496930e-01 5.55924058e-01 -8.79696786e-01
-9.67661560e-01 2.85808623e-01 -1.11709759e-01 -3.80877912e-01
-5.16833067e-01 -1.33152634e-01 -1.27953637e+00 -1.79862753e-01
-5.15960574e-01 -9.37316287e-03 7.04561770e-01 8.75251174e-01
-6.29739985e-02 1.08638155e+00 6.97490633e-01 -7.48592019e-01
-5.74840188e-01 -7.88311720e-01 -3.21396708e-01 1.98704317e-01
4.50324297e-01 -2.36071706e-01 -1.38876393e-01 3.49472135e-01]
|
[14.767833709716797, -2.6654231548309326]
|
5607c8e9-53e5-4097-a14e-033dd293c2ce
|
dynamic-bicycle-dispatching-of-dockless
|
2101.07437
| null |
https://arxiv.org/abs/2101.07437v1
|
https://arxiv.org/pdf/2101.07437v1.pdf
|
Dynamic Bicycle Dispatching of Dockless Public Bicycle-sharing Systems using Multi-objective Reinforcement Learning
|
As a new generation of Public Bicycle-sharing Systems (PBS), the dockless PBS (DL-PBS) is an important application of cyber-physical systems and intelligent transportation. How to use AI to provide efficient bicycle dispatching solutions based on dynamic bicycle rental demand is an essential issue for DL-PBS. In this paper, we propose a dynamic bicycle dispatching algorithm based on multi-objective reinforcement learning (MORL-BD) to provide the optimal bicycle dispatching solution for DL-PBS. We model the DL-PBS system from the perspective of CPS and use deep learning to predict the layout of bicycle parking spots and the dynamic demand of bicycle dispatching. We define the multi-route bicycle dispatching problem as a multi-objective optimization problem by considering the optimization objectives of dispatching costs, dispatch truck's initial load, workload balance among the trucks, and the dynamic balance of bicycle supply and demand. On this basis, the collaborative multi-route bicycle dispatching problem among multiple dispatch trucks is modeled as a multi-agent MORL model. All dispatch paths between parking spots are defined as state spaces, and the reciprocal of dispatching costs is defined as a reward. Each dispatch truck is equipped with an agent to learn the optimal dispatch path in the dynamic DL-PBS network. We create an elite list to store the Pareto optimal solutions of bicycle dispatch paths found in each action, and finally, get the Pareto frontier. Experimental results on the actual DL-PBS systems show that compared with existing methods, MORL-BD can find a higher quality Pareto frontier with less execution time.
|
['Zeng Zeng', 'Philip S. Yu', 'Keqin Li', 'Kenli Li', 'Jianguo Chen']
|
2021-01-19
| null | null | null | null |
['multi-objective-reinforcement-learning']
|
['methodology']
|
[-7.66253352e-01 -2.16924384e-01 -5.64606011e-01 3.31995115e-02
-4.84237492e-01 -2.92806774e-01 -1.89893723e-01 -4.62650001e-01
-3.80583704e-01 9.63686168e-01 -2.78961621e-02 -4.87239599e-01
-6.44795239e-01 -1.22851777e+00 -6.24397576e-01 -9.16736364e-01
-6.07808605e-02 1.17517769e+00 2.22481117e-01 -6.74878657e-01
7.82632306e-02 6.32217109e-01 -1.15400195e+00 -4.15829942e-02
1.05565512e+00 9.30324733e-01 9.08078134e-01 2.45583832e-01
-2.32169583e-01 2.35271037e-01 -3.92420053e-01 -4.06291481e-04
3.57431054e-01 1.88339889e-01 -6.33131206e-01 -8.13846365e-02
-9.56461072e-01 -2.86425561e-01 -4.38124686e-01 5.46377003e-01
3.44042987e-01 5.20035684e-01 5.88582575e-01 -2.39163041e+00
-4.46959078e-01 3.90430480e-01 -3.72819126e-01 5.51959723e-02
-1.84688985e-01 6.52152061e-01 9.42369461e-01 -4.28692758e-01
1.24534473e-01 1.38395083e+00 2.78155535e-01 1.83087453e-01
-1.15376198e+00 -6.45064056e-01 3.93168271e-01 7.57394314e-01
-1.26150823e+00 1.90957859e-01 4.96419579e-01 -2.72117287e-01
1.11875236e+00 1.12313621e-01 9.28933144e-01 4.32432473e-01
7.13270843e-01 1.04420352e+00 6.55046105e-01 -1.66830812e-02
3.55488628e-01 -1.81766704e-01 -1.51078731e-01 5.62097490e-01
3.99050042e-02 2.49412119e-01 3.04374009e-01 6.78856224e-02
1.65692151e-01 1.47322088e-01 4.28686917e-01 -3.43713909e-01
-9.51281965e-01 9.08913970e-01 6.30481184e-01 -1.64449409e-01
-6.51521146e-01 1.92120492e-01 3.02129298e-01 -1.15309237e-02
-7.07322508e-02 3.96705389e-01 -4.33056265e-01 3.66572812e-02
-2.46366620e-01 3.76428485e-01 7.96611786e-01 1.07618928e+00
9.65817034e-01 1.80156529e-01 -1.91776782e-01 6.39201105e-01
3.78925562e-01 8.43397856e-01 -1.48505252e-02 -1.19709611e+00
7.20592916e-01 4.60346431e-01 6.56190991e-01 -1.06694400e+00
-8.26609194e-01 2.04135939e-01 -6.77325785e-01 1.03528529e-01
-1.38113776e-03 -4.88217294e-01 -3.33603591e-01 1.47225440e+00
2.66666174e-01 1.33738682e-01 6.89683855e-03 1.12545025e+00
1.26359284e-01 1.21329558e+00 2.53782012e-02 -1.82318822e-01
1.00762045e+00 -1.12411416e+00 -5.75654745e-01 -1.23865873e-01
4.80060279e-01 -4.79782522e-01 8.05872560e-01 2.04588864e-02
-8.28429997e-01 -2.94107735e-01 -8.50439847e-01 5.06751716e-01
-6.19018555e-01 3.20677757e-01 4.72027212e-01 2.84917265e-01
-6.91505373e-01 2.36821637e-01 -5.99318206e-01 4.82170321e-02
2.02243805e-01 7.39369750e-01 1.48133531e-01 -1.43731609e-01
-1.35484385e+00 1.34098327e+00 5.69900215e-01 2.61486411e-01
-1.06047213e+00 -5.07655084e-01 -6.48971796e-01 2.96557754e-01
9.47520375e-01 -3.87751132e-01 1.24789798e+00 -4.44283485e-01
-1.69816136e+00 6.49106801e-02 1.60131201e-01 -4.36617360e-02
-1.70041174e-02 5.22471547e-01 -6.67096734e-01 -3.72132629e-01
5.72584569e-01 4.54454720e-01 3.66980463e-01 -1.28192031e+00
-1.22809660e+00 -5.07447543e-03 1.86387002e-01 4.42832172e-01
1.69269979e-01 -3.81678909e-01 -4.30536270e-01 8.00654441e-02
-5.05710423e-01 -1.43406105e+00 -4.92955476e-01 -7.20856309e-01
-4.35589969e-01 -7.39002168e-01 8.06060433e-01 -4.86175001e-01
1.16727161e+00 -1.84873629e+00 2.14537889e-01 4.06119615e-01
-2.61015326e-01 9.15484279e-02 -4.58500117e-01 6.97097301e-01
3.44148099e-01 -2.71446377e-01 2.13750675e-01 -1.71833754e-01
5.18545210e-01 8.47703278e-01 -1.41154602e-01 1.93516880e-01
1.26465529e-01 1.04729378e+00 -9.45699811e-01 -3.71091992e-01
4.32714134e-01 -3.93281072e-01 -3.03978294e-01 1.63238555e-01
-5.35942376e-01 1.28584504e-01 -8.78697217e-01 7.17699707e-01
1.01416802e+00 8.23426992e-02 1.90360576e-01 -7.87575915e-02
-4.24302608e-01 -2.40026504e-01 -1.26303840e+00 1.15776479e+00
-7.55935133e-01 4.85582918e-01 2.07900226e-01 -1.34191418e+00
8.07917595e-01 -3.22107077e-01 9.38051343e-01 -1.42029881e+00
-1.23502277e-02 2.94466436e-01 -5.20107290e-03 -6.56915367e-01
1.00073314e+00 -2.41958275e-02 -7.05621958e-01 7.54130304e-01
-3.88598382e-01 -4.45881821e-02 4.05976892e-01 -1.69441566e-01
7.10282028e-01 3.80872786e-02 -4.16627884e-01 -2.55133152e-01
5.76220214e-01 4.55694228e-01 8.99333000e-01 3.97915423e-01
-3.26660514e-01 -3.64469230e-01 5.35162807e-01 -5.60982585e-01
-9.50833142e-01 -1.02930880e+00 5.35524368e-01 1.24664974e+00
8.43589664e-01 4.52605277e-01 -3.99855554e-01 -4.05334115e-01
4.37992334e-01 9.41047907e-01 -2.02628776e-01 -2.50929058e-01
-8.66307795e-01 -7.88030922e-01 1.32402703e-01 3.31632823e-01
2.91442037e-01 -1.11196220e+00 -2.83939034e-01 6.52459502e-01
-3.67946565e-01 -9.23591495e-01 -8.44462156e-01 4.09937613e-02
-1.86637379e-02 -1.09050560e+00 -5.62720180e-01 -1.01305938e+00
4.62634474e-01 5.12032390e-01 6.35147870e-01 -1.07766800e-01
-9.97651517e-02 2.00567320e-01 1.99767902e-01 -2.48264000e-01
-2.97292918e-01 3.14738512e-01 1.77800566e-01 1.60543278e-01
1.16788603e-01 -1.27582133e-01 -6.89179003e-01 1.05964828e+00
-5.36660612e-01 5.64481430e-02 5.17690241e-01 7.66716659e-01
8.65500271e-01 5.38291156e-01 1.12115812e+00 7.30863214e-02
9.85732913e-01 -9.54249203e-01 -1.11185575e+00 3.92366320e-01
-6.81819499e-01 -1.18508535e-02 1.14522457e+00 -1.06580369e-01
-8.83169234e-01 -7.87167549e-02 2.78549820e-01 -2.83702165e-01
2.39644259e-01 6.10764265e-01 -4.67602015e-01 3.01326185e-01
-1.84236363e-01 4.80559975e-01 2.15161130e-01 -1.51173204e-01
2.33229682e-01 6.37724698e-01 2.50413507e-01 -7.52721429e-01
3.74416679e-01 4.37324308e-02 3.92481536e-01 -7.21161664e-02
-1.71192303e-01 -3.47039789e-01 -9.70637649e-02 -4.96817499e-01
7.85044253e-01 -5.32012641e-01 -1.89844882e+00 3.39009315e-01
-1.06024766e+00 -6.75112188e-01 -3.60040963e-02 3.82904470e-01
-1.05150914e+00 4.98463959e-03 -2.17296124e-01 -7.76735425e-01
1.51972368e-01 -1.62531137e+00 6.49304807e-01 4.23756182e-01
5.35703301e-01 -7.40210414e-01 2.75872111e-01 4.78993535e-01
2.85014272e-01 -1.93563756e-02 1.24305940e+00 -2.08748002e-02
-7.88339555e-01 7.09068105e-02 -2.05183953e-01 1.58909075e-02
1.58310711e-01 -7.61999562e-02 -6.73353821e-02 -4.25150692e-01
-7.78050125e-01 1.48823336e-01 -1.74814444e-02 5.99883854e-01
1.12652230e+00 -5.69876373e-01 -7.82578170e-01 1.04227737e-01
1.59301937e+00 8.42206419e-01 5.79908490e-01 7.54413545e-01
3.62344563e-01 6.22435868e-01 1.32808828e+00 4.48112935e-01
1.33628047e+00 8.45537722e-01 1.03038859e+00 1.07854933e-01
5.36975980e-01 -2.08097905e-01 3.83157462e-01 6.22095406e-01
9.97458324e-02 -6.40073299e-01 -9.72208381e-01 6.78295732e-01
-2.35136032e+00 -9.33614194e-01 -2.37406548e-02 2.00333309e+00
1.58948183e-01 -4.06810343e-02 4.48689520e-01 -3.44700396e-01
9.41778123e-01 -2.91733146e-01 -9.88084972e-01 -9.68112230e-01
8.91934261e-02 -4.06128198e-01 8.39836359e-01 4.20527399e-01
-6.45482659e-01 8.39098930e-01 4.99944687e+00 1.01197219e+00
-9.67491627e-01 2.72347741e-02 6.06855094e-01 -3.38506401e-02
-1.18685052e-01 -9.65253860e-02 -8.20553541e-01 1.28914177e+00
7.96883762e-01 -5.89727044e-01 1.24733341e+00 8.36438179e-01
8.65427434e-01 -3.25388044e-01 -8.14183772e-01 7.38250196e-01
-6.16234601e-01 -1.56177008e+00 -3.88050199e-01 1.96577564e-01
8.12917352e-01 3.49755108e-01 -2.21152101e-02 7.38564014e-01
6.91153049e-01 -8.99954796e-01 6.33116603e-01 8.34194422e-01
3.84095520e-01 -1.39609098e+00 6.75673604e-01 6.26639843e-01
-1.52093029e+00 -8.17281783e-01 -1.50357962e-01 2.19969511e-01
5.82795024e-01 -6.93975687e-02 -7.88846612e-01 8.86190295e-01
5.74870288e-01 3.40728670e-01 2.87421674e-01 1.07584989e+00
3.15029383e-01 -2.69432396e-01 -3.85506183e-01 -5.85464776e-01
8.20239961e-01 -6.81706786e-01 3.93330276e-01 6.65654182e-01
3.82629633e-01 -6.05070591e-02 8.88480723e-01 9.41966057e-01
3.11428338e-01 -2.26528332e-01 -4.22536194e-01 4.53273878e-02
7.01690376e-01 1.32596004e+00 -5.28097749e-01 9.01692808e-02
-2.41848201e-01 3.01382393e-01 3.03527098e-02 4.11461234e-01
-1.09404254e+00 -5.52384973e-01 9.74245906e-01 -1.49693653e-01
2.91501999e-01 -3.30000401e-01 9.49390456e-02 -4.28954899e-01
-2.39432216e-01 -4.78378475e-01 2.71376938e-01 -1.14272857e+00
-1.36438978e+00 2.31116727e-01 -4.72982489e-02 -1.43082666e+00
-2.61446208e-01 -6.68221176e-01 -1.01353514e+00 9.94583964e-01
-1.85158145e+00 -9.25653338e-01 -1.45232692e-01 7.53404081e-01
4.84737068e-01 -5.26457071e-01 4.38580096e-01 5.28495133e-01
-9.58869636e-01 1.25539035e-01 5.89981139e-01 -3.21644157e-01
2.49770898e-02 -8.40571046e-01 1.62305366e-02 3.05056512e-01
-8.91546011e-01 -1.23560078e-01 3.25687349e-01 -5.27221084e-01
-2.06707811e+00 -1.00647879e+00 4.01200622e-01 2.05253646e-01
8.88459265e-01 1.01238534e-01 -2.86319971e-01 4.50930893e-01
1.54247597e-01 -3.08978498e-01 3.48880976e-01 -3.10458452e-01
7.34989643e-01 -5.76448739e-01 -1.07896161e+00 8.07164550e-01
5.57876945e-01 5.98382540e-02 -1.02102533e-01 6.05470240e-01
8.74409854e-01 -1.79216877e-01 -5.27087748e-01 1.96359009e-02
1.51825771e-01 -2.37533271e-01 7.87708938e-01 -6.95130646e-01
-4.29254305e-03 -5.88073134e-01 -1.26101851e-01 -1.89911187e+00
-7.16907501e-01 -5.36542237e-01 1.90279007e-01 8.49135995e-01
2.28178978e-01 -8.75884652e-01 7.43790567e-01 6.64893627e-01
-6.90522909e-01 -9.42139208e-01 -1.56581199e+00 -9.55946028e-01
-1.16477050e-01 -1.91065401e-01 1.38867426e+00 5.42962432e-01
-2.63827592e-01 1.40042335e-01 -3.52073669e-01 3.74210984e-01
4.71210778e-01 4.88185048e-01 6.82157516e-01 -7.71673143e-01
2.88492758e-02 -6.17545009e-01 1.52877435e-01 -7.53464997e-01
6.17086768e-01 -1.11337435e+00 3.07404995e-01 -1.92985857e+00
-1.59584537e-01 -1.20710266e+00 -4.13221210e-01 5.00301301e-01
6.05822623e-01 -6.76705301e-01 3.64990294e-01 9.69371200e-02
-7.83363760e-01 8.84448946e-01 1.52822733e+00 -5.82295239e-01
-4.43531513e-01 4.52486247e-01 -3.57421070e-01 2.82768846e-01
9.44271982e-01 -3.98081213e-01 -4.23505038e-01 -6.00996196e-01
2.88106024e-01 6.92127347e-01 1.38931349e-01 -4.30985600e-01
5.24680078e-01 -1.18538940e+00 -1.90804437e-01 -1.02114379e+00
4.83034223e-01 -1.18538332e+00 2.68296599e-01 6.95788085e-01
1.84498996e-01 4.79530841e-01 3.34699243e-01 6.96544349e-01
1.63465992e-01 -1.46796867e-01 3.67606491e-01 4.69296984e-03
-1.14185095e+00 6.69940412e-01 -9.64659929e-01 -2.13389203e-01
1.40591860e+00 -1.42203376e-01 -4.89602000e-01 3.69623955e-03
-7.67676055e-01 1.39930201e+00 -1.54539138e-01 7.13870049e-01
4.80225980e-01 -1.57937646e+00 -3.44489217e-01 -3.38691920e-02
-3.88556570e-01 -2.27546468e-01 5.43492913e-01 7.43202746e-01
-4.45224106e-01 6.93182588e-01 -6.82393849e-01 -4.49603796e-01
-4.91115272e-01 6.20974123e-01 5.24469018e-01 -5.50685704e-01
-4.21491563e-02 1.00679444e-02 -1.37996212e-01 -8.12417746e-01
-1.77889422e-01 -1.11517377e-01 -1.37105957e-01 1.80909365e-01
-1.40039027e-01 1.02738702e+00 -9.94928777e-02 -5.37887752e-01
-6.46969914e-01 3.04493159e-01 3.09525013e-01 1.43636197e-01
1.53393960e+00 -3.92158836e-01 -6.21625176e-03 -1.59798726e-01
9.19812679e-01 -6.92211688e-01 -1.24795282e+00 6.76194206e-02
-1.15185060e-01 -3.34267169e-01 5.75799271e-02 -8.71042609e-01
-1.37383032e+00 2.78428435e-01 6.01118624e-01 4.16353613e-01
9.39067900e-01 -2.64334470e-01 1.18676305e+00 5.05866051e-01
7.50868559e-01 -1.82827556e+00 2.57885903e-01 8.01952600e-01
7.62695014e-01 -1.09372866e+00 -3.39134961e-01 1.99560389e-01
-1.02295744e+00 9.57645535e-01 1.03350413e+00 -2.31231138e-01
6.11096621e-01 2.54095852e-01 -4.73177373e-01 9.52796489e-02
-6.95224464e-01 -2.63238996e-01 -1.50782287e-01 6.28007114e-01
-8.48362148e-01 7.48796582e-01 -1.94411039e-01 8.44905198e-01
1.66217089e-01 1.34502485e-01 3.88981760e-01 8.48280668e-01
-5.23879886e-01 -1.06859016e+00 -2.42095128e-01 2.56740034e-01
5.85516989e-01 5.48788309e-01 5.13391256e-01 7.52111554e-01
3.95268381e-01 9.61873233e-01 4.61383760e-01 -3.75883818e-01
6.11716866e-01 -4.46140945e-01 1.04649067e-01 -3.67163777e-01
-2.79288620e-01 -3.33212435e-01 -5.74661931e-03 -5.46908438e-01
6.71660975e-02 -4.38451707e-01 -1.66126442e+00 -8.15582097e-01
-2.19677106e-01 1.45228431e-01 7.83995688e-01 1.04798353e+00
8.39869618e-01 6.97382987e-01 1.33280241e+00 -1.02562749e+00
-3.43874544e-01 -2.06579193e-01 -6.55878127e-01 -4.19323087e-01
6.17878623e-02 -9.96312380e-01 1.30353555e-01 -7.53035486e-01]
|
[5.53166389465332, 1.7908034324645996]
|
e12cd517-7477-4949-a6e0-fa8fafe4a5b0
|
deep-manta-a-coarse-to-fine-many-task-network
|
1703.07570
| null |
http://arxiv.org/abs/1703.07570v1
|
http://arxiv.org/pdf/1703.07570v1.pdf
|
Deep MANTA: A Coarse-to-fine Many-Task Network for joint 2D and 3D vehicle analysis from monocular image
|
In this paper, we present a novel approach, called Deep MANTA (Deep
Many-Tasks), for many-task vehicle analysis from a given image. A robust
convolutional network is introduced for simultaneous vehicle detection, part
localization, visibility characterization and 3D dimension estimation. Its
architecture is based on a new coarse-to-fine object proposal that boosts the
vehicle detection. Moreover, the Deep MANTA network is able to localize vehicle
parts even if these parts are not visible. In the inference, the network's
outputs are used by a real time robust pose estimation algorithm for fine
orientation estimation and 3D vehicle localization. We show in experiments that
our method outperforms monocular state-of-the-art approaches on vehicle
detection, orientation and 3D location tasks on the very challenging KITTI
benchmark.
|
['Céline Teulière', 'Jaonary Rabarisoa', 'Florian Chabot', 'Mohamed Chaouch', 'Thierry Chateau']
|
2017-03-22
|
deep-manta-a-coarse-to-fine-many-task-network-1
|
http://openaccess.thecvf.com/content_cvpr_2017/html/Chabot_Deep_MANTA_A_CVPR_2017_paper.html
|
http://openaccess.thecvf.com/content_cvpr_2017/papers/Chabot_Deep_MANTA_A_CVPR_2017_paper.pdf
|
cvpr-2017-7
|
['vehicle-pose-estimation']
|
['computer-vision']
|
[-3.72700632e-01 -2.30558097e-01 -5.80015630e-02 -2.49826312e-01
-7.56341994e-01 -8.12976122e-01 9.05015886e-01 -7.19934329e-02
-5.05874932e-01 1.77386999e-01 -4.70754862e-01 -7.58667737e-02
3.94519538e-01 -6.27445400e-01 -1.33995438e+00 -7.11213231e-01
-6.59212396e-02 9.84789610e-01 5.57722986e-01 -1.08738199e-01
1.32788420e-01 1.16409957e+00 -1.58141470e+00 -1.62526276e-02
2.16468945e-01 1.24147069e+00 2.88388729e-01 9.06467915e-01
3.49169195e-01 6.47416174e-01 -3.64958167e-01 -1.48790136e-01
4.49597150e-01 4.99240041e-01 -3.01384568e-01 2.12400127e-02
1.11255419e+00 -7.62379348e-01 -2.84486413e-01 1.01039863e+00
1.18027754e-01 -2.14504004e-01 9.53388572e-01 -1.52637446e+00
7.97759816e-02 4.51013036e-02 -5.99209368e-01 2.82340676e-01
9.48334411e-02 4.09535855e-01 5.50256252e-01 -1.40158403e+00
6.94895744e-01 1.61809433e+00 8.67551208e-01 7.59737566e-02
-9.83567417e-01 -6.75579727e-01 1.35716856e-01 4.97252166e-01
-1.79501772e+00 -4.32426095e-01 6.37796283e-01 -6.77384734e-01
1.17132938e+00 -5.49856899e-03 2.69946247e-01 9.72133160e-01
2.32386425e-01 8.67165446e-01 7.86921918e-01 -1.68297607e-02
2.45543625e-02 1.68049738e-01 1.09359808e-01 9.68854964e-01
5.47796369e-01 4.83428717e-01 -7.70767480e-02 9.95960087e-02
5.08484006e-01 9.77643505e-02 2.62903571e-01 -8.48931313e-01
-1.14245284e+00 8.43300223e-01 5.08964658e-01 -1.16871871e-01
-4.70286310e-01 5.55729091e-01 3.14022928e-01 7.74097517e-02
4.32927012e-01 -2.19084546e-02 -5.26636183e-01 3.62992108e-01
-8.99169385e-01 7.53949106e-01 4.52024132e-01 9.89831388e-01
1.10390151e+00 1.19762748e-01 -3.84558409e-01 1.54808313e-01
6.61266088e-01 1.25000286e+00 -2.62704343e-01 -9.23476636e-01
4.00541812e-01 5.64414620e-01 1.84745714e-01 -8.67329657e-01
-7.85180449e-01 -5.13784826e-01 -6.95435107e-01 7.86824226e-01
4.78807300e-01 7.81900734e-02 -1.18626189e+00 1.40842474e+00
6.41364455e-01 2.67853409e-01 -8.51364806e-02 1.13530207e+00
1.09168863e+00 5.51442325e-01 -9.84920859e-02 4.38225299e-01
1.72729123e+00 -9.38558340e-01 -2.58144468e-01 -9.03032720e-01
5.14962614e-01 -7.57294536e-01 9.18526426e-02 6.83434084e-02
-7.40003765e-01 -7.91109800e-01 -1.09474969e+00 -1.61232859e-01
-6.27950728e-01 6.00262642e-01 4.08346593e-01 6.11271083e-01
-1.08133101e+00 1.14632115e-01 -5.97009778e-01 -3.43210906e-01
5.45272112e-01 2.11008161e-01 -7.49595702e-01 -1.90333769e-01
-8.40936065e-01 1.29340053e+00 2.75128335e-01 3.77826929e-01
-1.89140952e+00 -4.76479053e-01 -1.27159989e+00 7.90802762e-02
4.11239803e-01 -7.82627106e-01 1.19648707e+00 -4.14457470e-01
-9.91642475e-01 1.20527947e+00 -1.14363551e-01 -8.56002629e-01
5.97326636e-01 -1.37485906e-01 -6.29755436e-03 8.15309808e-02
2.48618007e-01 9.96384025e-01 1.30055189e+00 -1.20519328e+00
-9.17047262e-01 -7.03838170e-01 -1.26853183e-01 -1.49291316e-02
5.12148499e-01 7.45051280e-02 -7.87125945e-01 -2.49885485e-01
-4.00563404e-02 -9.09218788e-01 -1.84758425e-01 1.81324333e-01
-4.62285012e-01 -3.29385847e-01 1.16025758e+00 -4.67787385e-01
1.68657631e-01 -2.18426919e+00 3.46834734e-02 2.54209101e-01
5.21643102e-01 3.38170648e-01 -4.60046858e-01 -6.11109100e-02
9.39903259e-02 -4.72445875e-01 2.68656343e-01 -7.36239851e-01
3.31075460e-01 -1.73521057e-01 -1.68802664e-01 1.14199173e+00
2.27623224e-01 1.21074772e+00 -3.81603122e-01 -3.43539089e-01
8.64311576e-01 7.34785557e-01 -7.95946717e-02 1.63002804e-01
-1.64965004e-01 -2.98545999e-03 -1.82741418e-01 1.18744946e+00
1.38564801e+00 1.61316827e-01 -3.36189747e-01 -4.07775879e-01
-2.64459550e-01 -2.59233057e-01 -1.08150518e+00 1.04267526e+00
-3.92977357e-01 1.00826848e+00 4.73357975e-01 -8.04942966e-01
9.22099531e-01 -4.27829362e-02 9.19768438e-02 -8.33584130e-01
4.08366740e-01 2.29849651e-01 -2.75931984e-01 -3.56054515e-01
5.87774098e-01 4.10026073e-01 -2.82140449e-02 -2.73957461e-01
-7.39153549e-02 -1.28033757e-01 2.24604443e-01 1.67916059e-01
9.79465485e-01 1.38933337e-05 2.23835647e-01 -2.86933362e-01
6.89283967e-01 1.74117982e-01 4.35542047e-01 5.71981311e-01
-3.79104435e-01 4.30399209e-01 3.96642774e-01 -7.37042546e-01
-1.26724160e+00 -1.11042094e+00 -1.60018995e-01 9.36190546e-01
4.81013238e-01 -1.58661187e-01 -5.75523317e-01 -6.44797444e-01
5.46859860e-01 3.42022508e-01 -8.74990046e-01 1.27584159e-01
-7.85643041e-01 -3.37357998e-01 5.47772527e-01 5.96306562e-01
2.90070742e-01 -7.85335720e-01 -7.27999985e-01 1.06523663e-01
1.44291326e-01 -1.60821009e+00 -2.15263784e-01 2.51866192e-01
-4.23131436e-01 -1.09322679e+00 -7.20568240e-01 -8.01349759e-01
4.19885486e-01 6.13308907e-01 1.13195252e+00 -2.58029997e-01
-2.50090390e-01 5.79008237e-02 7.21644089e-02 -4.83020961e-01
-4.34136689e-01 -1.54349238e-01 5.74891679e-02 1.39075160e-01
4.27886218e-01 -4.82696556e-02 -7.51624048e-01 5.28271794e-01
-2.78047889e-01 -1.14379711e-01 1.04184282e+00 3.71584892e-01
6.80964231e-01 -2.10881442e-01 7.88790658e-02 -2.13744074e-01
-2.72459447e-01 -3.06685001e-01 -1.40464306e+00 -2.11066216e-01
-1.72915593e-01 7.44355693e-02 3.92183393e-01 -1.94267035e-01
-5.38759649e-01 6.91374362e-01 -3.19853455e-01 -7.78246462e-01
-4.23687309e-01 -1.57607377e-01 -2.11414129e-01 -6.46210432e-01
5.62926531e-01 5.85855059e-02 6.39145402e-03 -4.38405305e-01
3.65883142e-01 3.34413677e-01 6.14585340e-01 2.25096866e-01
1.23769939e+00 7.93408275e-01 4.32360291e-01 -8.87805760e-01
-4.38454419e-01 -7.20681071e-01 -5.78303099e-01 -4.48258013e-01
1.20972812e+00 -1.65092397e+00 -1.03624642e+00 6.09331906e-01
-1.57293165e+00 -3.94094378e-01 2.26522863e-01 4.11198229e-01
-3.80777687e-01 9.46121588e-02 -2.82695204e-01 -7.72858500e-01
-3.40278894e-01 -1.50584519e+00 1.98306680e+00 3.86184938e-02
2.99752504e-01 -6.85233116e-01 2.08016410e-02 1.94940805e-01
3.74254495e-01 2.67873436e-01 2.52420008e-01 -3.62418443e-01
-1.38275766e+00 -6.47303402e-01 -7.27861464e-01 1.10330649e-01
-7.54601896e-01 -1.15522243e-01 -1.11966515e+00 -4.26894397e-01
-5.35621166e-01 -1.61142647e-01 1.39664686e+00 8.37907374e-01
6.82143688e-01 1.29684746e-01 -9.15833056e-01 1.01719713e+00
1.37040436e+00 -3.70656222e-01 4.77223277e-01 3.74608666e-01
1.02913404e+00 2.54436523e-01 8.61276329e-01 9.81556848e-02
6.09297454e-01 1.08817649e+00 1.29047287e+00 -2.62352258e-01
-4.01655793e-01 7.36413151e-02 4.37842488e-01 -3.06850791e-01
6.28142357e-02 -8.44076499e-02 -7.55591393e-01 6.89530909e-01
-1.79896522e+00 -9.09663856e-01 -4.30912167e-01 1.88260376e+00
-2.50589848e-01 4.30860579e-01 1.83060855e-01 -2.59930342e-01
6.29361510e-01 2.11955562e-01 -3.21231067e-01 -1.73563227e-01
-3.46365064e-01 -4.66600567e-01 1.23586416e+00 5.00948966e-01
-1.73174226e+00 1.03991985e+00 5.84324551e+00 9.27919507e-01
-9.29536939e-01 2.84289986e-01 4.46313441e-01 2.22627968e-01
7.56127760e-02 -3.07572484e-01 -1.48288357e+00 8.38837028e-02
5.68044782e-01 4.75721776e-01 1.01613060e-01 1.31382525e+00
7.76775479e-02 -1.72486976e-01 -9.83011544e-01 1.16332197e+00
3.18249345e-01 -1.54261720e+00 -2.55070120e-01 2.18480930e-01
4.49466616e-01 7.43854702e-01 9.68195722e-02 4.26384121e-01
-4.73346263e-02 -1.05644131e+00 1.18313420e+00 2.44956553e-01
7.02919185e-01 -9.76725519e-01 9.55157220e-01 4.50100183e-01
-1.66272533e+00 -2.65730202e-01 -4.68261212e-01 1.52835384e-01
2.50135034e-01 4.32646960e-01 -1.24699426e+00 2.55117267e-01
6.05216146e-01 4.79854852e-01 -9.45044160e-01 1.19390512e+00
-1.20658249e-01 8.11253488e-02 -4.40546840e-01 2.87138056e-02
4.72677290e-01 2.52991110e-01 8.91294718e-01 1.41400421e+00
3.94582570e-01 -5.54021716e-01 3.47652942e-01 9.77266550e-01
-4.54464033e-02 -3.33699912e-01 -9.87182975e-01 5.51805794e-01
3.77349108e-01 1.81181443e+00 -8.15447390e-01 -3.30711514e-01
-1.67638615e-01 7.51805842e-01 1.77521184e-01 2.42498368e-01
-1.10418844e+00 -2.68245101e-01 9.77627814e-01 2.26073682e-01
9.89084363e-01 -2.74655581e-01 -5.02634570e-02 -7.64412582e-01
-4.82571833e-02 -5.18495560e-01 -2.53148545e-02 -8.91553283e-01
-8.07974637e-01 6.36019707e-01 4.70617414e-02 -1.09526169e+00
-4.73813117e-01 -1.22067022e+00 -5.45544505e-01 7.63380826e-01
-1.78100646e+00 -1.70704186e+00 -4.69891608e-01 4.24544811e-01
5.80099761e-01 -3.57584059e-01 2.82021224e-01 6.00401759e-01
-2.69113958e-01 4.79135364e-01 -4.47002500e-02 1.72132671e-01
4.37592298e-01 -1.10840285e+00 8.97147417e-01 9.98283029e-01
4.72011864e-02 -1.56103298e-01 8.39203596e-01 -6.08251810e-01
-1.80433285e+00 -1.54953551e+00 9.27541792e-01 -7.88789868e-01
3.57595533e-01 -7.60657549e-01 -5.41079342e-01 7.64455497e-01
-3.71693112e-02 3.53847653e-01 -3.50296080e-01 -2.53699750e-01
-4.88129139e-01 -3.75399202e-01 -1.01992095e+00 1.38916925e-01
7.47161984e-01 -4.43629056e-01 -1.85139865e-01 5.42679131e-01
5.21909833e-01 -5.61574876e-01 -8.08226913e-02 5.54935813e-01
5.69195986e-01 -8.01120579e-01 1.45800710e+00 1.03715807e-02
-2.41128311e-01 -8.01591456e-01 -2.49951899e-01 -9.59235609e-01
-5.46246588e-01 -1.53782025e-01 -4.29559290e-01 6.31879032e-01
2.86687911e-01 -3.54057997e-01 8.17203224e-01 -4.54594880e-01
-4.70533520e-01 -2.54103601e-01 -1.19518423e+00 -9.20333564e-01
-3.31358731e-01 -6.17226660e-01 6.07650101e-01 1.87800840e-01
-9.24342990e-01 6.07402980e-01 -3.56325001e-01 9.33270693e-01
1.25732291e+00 3.26153457e-01 1.15843475e+00 -1.46089995e+00
1.29488602e-01 -4.87972617e-01 -1.07427955e+00 -1.16441679e+00
3.16869587e-01 -7.09278226e-01 2.11110085e-01 -1.30510449e+00
2.55300105e-01 1.86046168e-01 2.59484589e-01 1.54103518e-01
1.55358016e-01 9.05886233e-01 2.61638612e-01 -1.82853639e-01
-9.66948330e-01 3.32682461e-01 9.31743145e-01 -3.99624527e-01
3.01173508e-01 2.48386711e-01 -9.00385454e-02 7.79189527e-01
5.37029088e-01 -3.11604947e-01 2.63454705e-01 -4.15376157e-01
2.68841118e-01 -1.19065374e-01 1.30937076e+00 -1.11557364e+00
2.21876755e-01 2.87994176e-01 7.93919623e-01 -1.66370523e+00
7.58584917e-01 -9.98823583e-01 5.32755964e-02 6.08347833e-01
3.51718694e-01 4.86250482e-02 3.31502199e-01 5.01893401e-01
-1.17319919e-01 1.03223383e-01 9.40495908e-01 1.61927804e-01
-1.40165067e+00 5.42016208e-01 -4.70522285e-01 -3.73961389e-01
1.27682793e+00 -2.66765267e-01 -3.83188188e-01 -1.02938652e-01
-4.56413090e-01 4.17432547e-01 3.76259804e-01 5.63195467e-01
6.78416908e-01 -1.50616050e+00 -1.04031503e+00 3.57332736e-01
4.89961296e-01 -4.88711521e-02 4.97977942e-01 9.33875918e-01
-8.55546653e-01 8.06444764e-01 -1.99437231e-01 -1.13882625e+00
-1.59462309e+00 8.10818672e-01 5.26239932e-01 4.38399240e-02
-5.07194519e-01 6.87835515e-01 5.52008986e-01 -5.34132123e-01
3.43908548e-01 -5.45928121e-01 -2.28821635e-01 -1.10528851e-02
7.60613322e-01 4.13273007e-01 2.51625270e-01 -1.15674436e+00
-9.33338046e-01 1.00955200e+00 1.24865890e-01 1.33949146e-01
1.16671491e+00 -3.26192588e-01 2.85006166e-01 -7.53953606e-02
1.27452254e+00 -2.57270396e-01 -1.61891317e+00 -9.79460478e-02
-3.81309897e-01 -3.75898302e-01 3.73495996e-01 -5.86864114e-01
-1.02767646e+00 9.48673368e-01 1.02251327e+00 2.65914779e-02
4.64928329e-01 4.53319073e-01 4.36686546e-01 6.22470617e-01
1.19918965e-01 -8.47219884e-01 -2.63518810e-01 8.27396750e-01
7.81738281e-01 -1.74038553e+00 -3.26390006e-02 -2.73800135e-01
-2.90928096e-01 1.06520677e+00 6.75240874e-01 -2.13358104e-01
4.34145898e-01 4.24953878e-01 5.69518358e-02 -4.51045245e-01
-4.82956201e-01 -6.80496693e-01 5.84345162e-01 7.97022939e-01
-2.62928575e-01 1.47776425e-01 5.08612633e-01 7.18651786e-02
2.13470697e-01 -4.70230192e-01 1.69443265e-01 3.30658793e-01
-7.62945175e-01 -3.26227933e-01 -8.87909889e-01 -1.39134645e-01
3.66621651e-02 1.60079315e-01 -3.26354533e-01 9.71312106e-01
5.14521003e-01 6.08593941e-01 3.62428516e-01 -1.92176759e-01
2.33012617e-01 -3.59143525e-01 4.43272203e-01 -2.57689089e-01
-1.87559515e-01 1.40069157e-01 1.67318135e-01 -7.90692985e-01
-3.31669562e-02 -7.12512672e-01 -9.18279707e-01 -3.95900726e-01
-1.32265329e-01 -3.65510941e-01 1.13460267e+00 1.00634801e+00
2.84546912e-01 3.90482754e-01 6.33186102e-01 -1.72873390e+00
-3.98476243e-01 -8.99969697e-01 -4.50773418e-01 -4.18630876e-02
9.45740938e-01 -8.01301420e-01 -2.45558396e-01 -3.45954835e-01]
|
[7.906476020812988, -2.12926983833313]
|
5c22d43e-36de-40a3-9609-d540f7b9fae6
|
progressive-self-training-with-discriminator
| null | null |
https://aclanthology.org/2021.emnlp-main.23
|
https://aclanthology.org/2021.emnlp-main.23.pdf
|
Progressive Self-Training with Discriminator for Aspect Term Extraction
|
Aspect term extraction aims to extract aspect terms from a review sentence that users have expressed opinions on. One of the remaining challenges for aspect term extraction resides in the lack of sufficient annotated data. While self-training is potentially an effective method to address this issue, the pseudo-labels it yields on unlabeled data could induce noise. In this paper, we use two means to alleviate the noise in the pseudo-labels. One is that inspired by the curriculum learning, we refine the conventional self-training to progressive self-training. Specifically, the base model infers pseudo-labels on a progressive subset at each iteration, where samples in the subset become harder and more numerous as the iteration proceeds. The other is that we use a discriminator to filter the noisy pseudo-labels. Experimental results on four SemEval datasets show that our model significantly outperforms the previous baselines and achieves state-of-the-art performance.
|
['Ruifeng Xu', 'Min Yang', 'Qin Zhao', 'Zhiyuan Wen', 'Qianlong Wang']
| null | null | null | null |
emnlp-2021-11
|
['term-extraction', 'extract-aspect']
|
['natural-language-processing', 'natural-language-processing']
|
[ 2.49194011e-01 3.51248801e-01 -6.47541165e-01 -5.56338310e-01
-1.10042787e+00 -6.91392183e-01 5.86100817e-01 2.84436166e-01
-3.89876902e-01 7.40501463e-01 2.89585352e-01 -2.23219305e-01
3.69367689e-01 -6.10260248e-01 -5.02681673e-01 -5.42595327e-01
4.50578094e-01 4.10575539e-01 1.01104863e-01 -2.10106373e-01
9.54581574e-02 -3.00823271e-01 -1.46646774e+00 4.03712004e-01
1.08346546e+00 8.36026549e-01 -1.41754240e-01 1.98537424e-01
-5.72184324e-01 5.30902326e-01 -7.12289989e-01 -6.29666865e-01
8.19669217e-02 -4.51563150e-01 -6.75418139e-01 4.33389395e-01
2.72735953e-01 -1.45555034e-01 1.84476361e-01 1.18525445e+00
1.19776130e-01 -1.04589965e-02 8.55192304e-01 -1.03211832e+00
-6.12400532e-01 8.69023979e-01 -7.45344996e-01 -1.21374488e-01
1.59694150e-01 -1.27817214e-01 1.47435760e+00 -1.09388554e+00
5.69702327e-01 9.32005942e-01 5.26851714e-01 5.94163299e-01
-1.23001230e+00 -7.04896986e-01 4.98002708e-01 -1.40232041e-01
-1.29715407e+00 -4.04237658e-01 9.05499578e-01 -3.30573946e-01
6.25665188e-01 1.45588294e-01 7.10204244e-01 1.07012618e+00
-1.99508980e-01 1.23315847e+00 1.27136075e+00 -6.30211830e-01
4.38989997e-01 5.37395298e-01 4.96541858e-01 5.15526175e-01
4.13385987e-01 -2.42219731e-01 -5.31532705e-01 -3.36347669e-01
2.44872063e-01 -5.00604063e-02 -2.79933542e-01 -3.10670733e-01
-6.41525865e-01 8.96489799e-01 -7.65614733e-02 2.83025742e-01
-3.37679684e-01 -2.06727698e-01 3.27310920e-01 2.85525471e-01
9.76018369e-01 5.99647820e-01 -9.11266565e-01 -1.45279378e-01
-1.10621834e+00 1.70901060e-01 9.01703119e-01 9.51584935e-01
1.08309174e+00 -1.20792612e-01 -3.14803123e-01 9.66753185e-01
3.50331515e-01 3.12900484e-01 4.51222152e-01 -6.01064503e-01
4.22770411e-01 9.83832538e-01 -5.03782593e-02 -4.50730205e-01
-5.03069200e-02 -7.64874756e-01 -4.97660905e-01 1.37344878e-02
1.26053378e-01 -3.99603099e-01 -1.18377471e+00 1.65126932e+00
4.75024939e-01 -4.96872142e-02 -5.25251850e-02 5.70382297e-01
9.56826508e-01 3.81828278e-01 1.85014591e-01 -3.58726948e-01
1.65837502e+00 -1.23344660e+00 -1.06001163e+00 -6.04342163e-01
4.46120977e-01 -8.51405978e-01 1.09745717e+00 2.86500812e-01
-6.81937575e-01 -2.66193211e-01 -1.15329063e+00 2.26299912e-01
-3.37886840e-01 2.56944329e-01 6.92802131e-01 6.62079513e-01
-6.45638049e-01 4.26027507e-01 -7.19233751e-01 -3.64300720e-02
5.83981156e-01 3.71258743e-02 -7.05792159e-02 9.70385373e-02
-1.16092360e+00 4.52697337e-01 1.73231587e-01 -1.37598023e-01
-4.85599428e-01 -7.75557339e-01 -1.00006330e+00 6.50237650e-02
7.12440312e-01 -5.96381187e-01 1.54788804e+00 -1.13362229e+00
-1.22144461e+00 9.22242284e-01 -6.08438671e-01 -1.46193251e-01
2.11838022e-01 -4.28119570e-01 -3.45577478e-01 -9.64962617e-02
3.46421480e-01 4.69043970e-01 1.14916039e+00 -1.61191261e+00
-6.57873690e-01 -2.22443476e-01 1.01513751e-01 3.47965807e-01
-6.07096672e-01 -1.51987687e-01 -8.43786120e-01 -7.81342447e-01
8.43313858e-02 -1.03737426e+00 -2.93252677e-01 -4.24529672e-01
-5.21250129e-01 -5.08207560e-01 7.29490757e-01 -2.70506114e-01
1.39419222e+00 -2.01111245e+00 -1.93984449e-01 1.30402580e-01
4.00669575e-01 4.11723852e-01 -2.38315955e-01 2.54620224e-01
3.53416242e-02 3.23949665e-01 -2.89883107e-01 -7.00078547e-01
-1.59196228e-01 1.05791047e-01 -4.30526197e-01 -2.75970828e-02
4.70605135e-01 8.38090241e-01 -1.08859968e+00 -6.11679494e-01
-3.21209848e-01 3.84074569e-01 -5.23729742e-01 2.43199751e-01
-5.07341146e-01 4.23252821e-01 -6.00233912e-01 6.74537003e-01
6.99881315e-01 -4.18232471e-01 1.11749038e-01 -2.51401871e-01
1.24186397e-01 8.60773861e-01 -1.07347226e+00 1.47719812e+00
-4.02307451e-01 4.83357996e-01 -1.59723178e-01 -6.50935411e-01
7.48613656e-01 3.55256379e-01 4.77856457e-01 -3.73085558e-01
1.55073121e-01 1.22103684e-01 -1.99323133e-01 -4.43017900e-01
7.45274544e-01 -2.81702697e-01 -2.26364657e-02 7.66711354e-01
1.70574456e-01 -1.54495418e-01 5.47362566e-01 2.17000499e-01
9.88560021e-01 2.77145714e-01 3.33206356e-01 -4.04995903e-02
5.65904379e-01 5.04337810e-02 9.14946795e-01 5.45733094e-01
-1.95091948e-01 6.94494724e-01 6.18230522e-01 -2.34822556e-01
-7.20200419e-01 -8.89468968e-01 -6.56857640e-02 1.02565336e+00
-1.29058525e-01 -8.88430536e-01 -6.78838253e-01 -1.29689896e+00
-2.21275821e-01 8.38782549e-01 -7.20342994e-01 -8.71198103e-02
-4.86046284e-01 -8.15436304e-01 5.02451919e-02 4.46074188e-01
3.04228127e-01 -1.02930546e+00 -2.07168207e-01 2.89446175e-01
-4.40828353e-01 -1.19120979e+00 -5.41441858e-01 4.08190161e-01
-8.55824411e-01 -9.40510690e-01 -4.27894801e-01 -7.79007137e-01
9.66353536e-01 4.43358600e-01 1.57329166e+00 7.03667551e-02
9.74768773e-02 9.70488191e-02 -6.46495283e-01 -6.79935753e-01
-2.41868407e-01 2.93824315e-01 -3.02203625e-01 5.17419493e-03
1.05925119e+00 -4.03814644e-01 -5.93357503e-01 -1.97064877e-02
-8.73645365e-01 -1.31293491e-01 6.33634031e-01 8.55491996e-01
8.38305175e-01 1.07308075e-01 6.39897525e-01 -1.67116797e+00
7.65057862e-01 -3.95175010e-01 -2.80507267e-01 1.60246834e-01
-1.11460221e+00 1.44883797e-01 4.66562510e-01 -5.56433916e-01
-1.06447887e+00 -3.33055668e-02 -1.13156438e-01 -9.59645435e-02
-1.66268069e-02 6.85109496e-01 -1.16888493e-01 4.37503368e-01
4.56513613e-01 1.09104998e-01 -3.12052011e-01 -4.40479875e-01
5.32782257e-01 7.51132607e-01 2.15586592e-02 -4.87258911e-01
9.98957455e-01 3.42031747e-01 -5.11562884e-01 -7.26176798e-01
-1.65793836e+00 -6.97787821e-01 -3.10561687e-01 -4.89758998e-02
5.84853828e-01 -1.11138475e+00 7.67624052e-03 1.86883211e-01
-9.43319619e-01 1.01616811e-02 -6.11203730e-01 1.80866331e-01
1.09497253e-02 2.97950834e-01 -5.82279623e-01 -9.57585990e-01
-6.16352916e-01 -1.00866210e+00 1.25488806e+00 5.32395363e-01
-5.88268697e-01 -8.74255717e-01 2.47173160e-01 4.37612623e-01
3.55748117e-01 -1.89189389e-01 8.69102359e-01 -8.77328932e-01
-4.97605741e-01 -3.90444279e-01 5.34165837e-02 5.16437948e-01
5.13403893e-01 -1.32865354e-01 -1.17265821e+00 -1.91566139e-01
3.44417870e-01 -5.51325977e-01 9.79550719e-01 1.32300407e-01
1.00861323e+00 -3.67485583e-01 -2.75306523e-01 2.50646800e-01
1.33317471e+00 -4.18542847e-02 3.23628634e-01 4.18566972e-01
6.80219054e-01 6.04758978e-01 7.38908410e-01 2.01354548e-01
7.17543542e-01 2.83946693e-01 3.56914401e-02 -1.71874762e-01
-1.35281682e-01 -4.28901225e-01 3.06074858e-01 1.12779486e+00
2.87122637e-01 -1.44914746e-01 -4.23986405e-01 7.63746619e-01
-1.64017832e+00 -5.54409325e-01 -3.17453332e-02 2.07065988e+00
1.29998958e+00 4.78210270e-01 5.63243926e-02 1.58638433e-01
4.26040560e-01 3.80481124e-01 -4.42206264e-01 -1.54089704e-01
-1.69244304e-01 3.14895183e-01 1.18944630e-01 3.85576069e-01
-1.28842974e+00 1.05398524e+00 6.17685604e+00 7.03707814e-01
-8.14711630e-01 2.70762086e-01 7.44681716e-01 -7.59526044e-02
-6.89623535e-01 3.15242559e-01 -1.06798148e+00 3.22968453e-01
6.95286930e-01 -8.35449174e-02 -2.75751520e-02 9.31268871e-01
-1.05202921e-01 -1.81555465e-01 -1.06475925e+00 6.31378531e-01
3.95160645e-01 -8.58627558e-01 5.30279204e-02 -1.06159911e-01
1.10992384e+00 -1.09312981e-02 -2.24187337e-02 6.74010694e-01
3.86212826e-01 -6.66020989e-01 5.04071712e-01 5.36933951e-02
6.66191041e-01 -6.19772673e-01 8.11958015e-01 3.73253763e-01
-1.19098234e+00 3.07992637e-01 -2.20435828e-01 7.19135553e-02
1.77470356e-01 1.04176176e+00 -7.22584188e-01 2.95143574e-01
4.66015667e-01 5.49816370e-01 -6.56007648e-01 8.89652908e-01
-8.47141922e-01 9.36439157e-01 -1.69142857e-02 -2.48728156e-01
1.07757002e-01 -4.28533107e-01 5.33233583e-01 1.17353678e+00
-7.09319413e-02 -2.07358688e-01 2.76782393e-01 7.88485408e-01
-2.34636992e-01 2.55695730e-01 -4.62022364e-01 -1.94532305e-01
4.15558100e-01 1.64498818e+00 -6.05960190e-01 -6.03625238e-01
-6.85503900e-01 7.84896791e-01 5.12109637e-01 5.60731649e-01
-4.00785536e-01 -2.36485079e-01 6.27318978e-01 -4.69418280e-02
6.90426469e-01 9.96920988e-02 -6.12802625e-01 -1.43797672e+00
3.42829347e-01 -1.15442193e+00 4.42714304e-01 -3.42813402e-01
-1.43728554e+00 7.47365177e-01 -2.42235482e-01 -1.36780119e+00
-2.90599287e-01 -3.54249626e-01 -5.85657418e-01 7.30617046e-01
-1.84589171e+00 -1.17665386e+00 -4.92393970e-02 -1.35491967e-01
8.64320338e-01 -1.32310197e-01 8.95426631e-01 2.62200832e-01
-4.21666354e-01 7.19340205e-01 -2.30552793e-01 2.22833291e-01
8.29091907e-01 -1.54248273e+00 4.48122501e-01 9.45880115e-01
4.43900764e-01 9.85626221e-01 8.10974360e-01 -7.85933197e-01
-1.04299676e+00 -9.18576598e-01 1.20652735e+00 -6.07201159e-01
6.48969710e-01 -4.53953922e-01 -1.00869966e+00 6.30600452e-01
4.14818645e-01 -2.79255688e-01 1.20885229e+00 6.47287309e-01
-6.87183022e-01 -6.62808791e-02 -8.20552349e-01 6.48148119e-01
6.25177026e-01 -5.90628445e-01 -8.63119841e-01 2.06326544e-01
8.46576750e-01 -2.83371091e-01 -5.25416851e-01 4.75447327e-01
3.25976610e-01 -6.19813800e-01 5.32076597e-01 -5.33913553e-01
5.68401217e-01 -4.60821122e-01 1.65489584e-01 -1.57040071e+00
-9.39421058e-02 -6.83102429e-01 -4.91221815e-01 1.89866054e+00
8.47288549e-01 -3.31451327e-01 1.07467175e+00 6.31097913e-01
1.97547555e-01 -1.00884330e+00 -6.10792637e-01 -5.76526880e-01
-6.13311976e-02 -4.72425848e-01 5.83274603e-01 8.61442089e-01
7.70885944e-02 1.08082330e+00 -2.57773310e-01 -1.53208345e-01
7.29265034e-01 3.49341840e-01 7.32034147e-01 -1.16030681e+00
-3.68518859e-01 -2.53695428e-01 2.36164048e-01 -1.34358978e+00
2.65468508e-01 -6.90072656e-01 2.95461506e-01 -1.48948443e+00
5.09047687e-01 -5.10956526e-01 -4.05820906e-01 5.63401043e-01
-8.59835267e-01 2.08920017e-01 -1.66127831e-01 1.54221073e-01
-8.75772834e-01 7.64503598e-01 1.23374832e+00 -2.95578778e-01
-3.91924977e-01 3.32266867e-01 -1.32486057e+00 7.85048008e-01
7.20731497e-01 -8.26159120e-01 -5.64265549e-01 -2.74985641e-01
4.40856367e-01 -5.59758008e-01 -4.23025727e-01 -4.70240980e-01
8.22947100e-02 1.48825854e-01 8.64407420e-02 -8.13374102e-01
1.97377965e-01 -6.99383438e-01 -5.54797232e-01 4.87773158e-02
-2.97742903e-01 -3.86351757e-02 -3.82284401e-03 5.87791979e-01
-3.54821205e-01 -5.88401973e-01 4.69744116e-01 -1.69830576e-01
-2.51117498e-01 1.92451000e-01 -3.91325593e-01 5.02270460e-01
5.43383241e-01 2.24646613e-01 -3.35774302e-01 -4.40013498e-01
-2.77391642e-01 4.38754857e-01 4.14175868e-01 5.52985549e-01
4.38157529e-01 -1.24992335e+00 -5.25686979e-01 2.52773732e-01
5.27984858e-01 2.41719961e-01 -1.58370480e-01 4.34700966e-01
1.15637124e-01 1.97622746e-01 5.64245462e-01 -4.66091871e-01
-1.23835313e+00 3.48027855e-01 -7.92069286e-02 -6.79595351e-01
-4.40882504e-01 9.10766959e-01 1.98113278e-01 -5.34849942e-01
2.99294114e-01 -1.26520127e-01 -6.29025578e-01 3.70026290e-01
6.95851922e-01 -1.34826630e-01 1.69998482e-01 -4.41404462e-01
-1.90821543e-01 4.47528452e-01 -6.02752864e-01 -3.06205124e-01
1.44838607e+00 -2.23328158e-01 -2.90585577e-01 7.27088034e-01
1.03535771e+00 3.12008470e-01 -1.10359585e+00 -6.40806198e-01
2.54619420e-01 -2.55961239e-01 2.83548944e-02 -7.58531809e-01
-1.00108516e+00 3.82995397e-01 2.17720017e-01 3.18941265e-01
1.06294823e+00 6.70310184e-02 8.75331581e-01 3.56700420e-01
-3.50459553e-02 -1.29151559e+00 1.13788202e-01 7.37019897e-01
4.33500290e-01 -1.47295320e+00 2.83625782e-01 -7.57517457e-01
-7.61200786e-01 6.79746807e-01 8.31421494e-01 1.65529083e-02
8.10878873e-01 3.18312287e-01 5.88490307e-01 -3.19882303e-01
-9.61796582e-01 -4.29391056e-01 4.94175106e-01 3.71759981e-01
7.72836745e-01 -5.31798713e-02 -7.84023225e-01 8.26309144e-01
-2.17277110e-01 -1.67779431e-01 3.77167076e-01 1.11368597e+00
-2.28927866e-01 -1.57538271e+00 -5.80777228e-02 7.92867780e-01
-7.78599262e-01 -4.13277924e-01 -5.16834557e-01 3.15035641e-01
8.33189860e-02 1.16093087e+00 -2.99453825e-01 -2.11133838e-01
3.52493286e-01 3.53328556e-01 1.33713886e-01 -1.11712039e+00
-6.67922258e-01 5.49735844e-01 4.15036261e-01 -1.75987333e-01
-6.17705941e-01 -6.85509741e-01 -1.07484233e+00 3.76310676e-01
-6.98934138e-01 5.51580787e-01 7.07380891e-01 1.21709633e+00
2.01282814e-01 7.31472969e-01 6.76319242e-01 -2.95805722e-01
-6.12625062e-01 -1.15849149e+00 -3.92361134e-01 6.18683875e-01
4.53426600e-01 -5.05665600e-01 -5.53152025e-01 1.00110069e-01]
|
[11.376204490661621, 6.729957103729248]
|
be03a056-706e-45a1-94e4-753e9a849a2e
|
contextual-argument-component-classification
|
2102.10290
| null |
https://arxiv.org/abs/2102.10290v1
|
https://arxiv.org/pdf/2102.10290v1.pdf
|
Contextual Argument Component Classification for Class Discussions
|
Argument mining systems often consider contextual information, i.e. information outside of an argumentative discourse unit, when trained to accomplish tasks such as argument component identification, classification, and relation extraction. However, prior work has not carefully analyzed the utility of different contextual properties in context-aware models. In this work, we show how two different types of contextual information, local discourse context and speaker context, can be incorporated into a computational model for classifying argument components in multi-party classroom discussions. We find that both context types can improve performance, although the improvements are dependent on context size and position.
|
['Diane Litman', 'Luca Lugini']
|
2021-02-20
| null |
https://aclanthology.org/2020.coling-main.128
|
https://aclanthology.org/2020.coling-main.128.pdf
|
coling-2020-8
|
['component-classification']
|
['natural-language-processing']
|
[ 5.46334088e-01 4.10045624e-01 -6.22485399e-01 -5.40725768e-01
-8.20084691e-01 -9.00425494e-01 7.64859557e-01 1.10973573e+00
-2.71446496e-01 7.81204462e-01 9.08964396e-01 -9.35708463e-01
-1.68987736e-01 -6.41030014e-01 -5.33135355e-01 -1.83839798e-01
4.91916060e-01 5.60555235e-02 5.03800809e-01 -2.12865531e-01
4.89982784e-01 1.97797209e-01 -1.96411288e+00 6.76232636e-01
1.06464028e+00 5.08963168e-01 -3.66603695e-02 8.87604654e-01
-4.71045226e-01 1.26249063e+00 -1.10270894e+00 -5.32490984e-02
-5.11880636e-01 -6.64622128e-01 -1.37756503e+00 -1.99039221e-01
5.61218381e-01 -4.99559566e-02 4.18480605e-01 7.04920173e-01
3.99858862e-01 1.36293277e-01 4.72840101e-01 -9.36005890e-01
-9.59693417e-02 1.35224116e+00 3.93194668e-02 4.65678960e-01
7.09469080e-01 -3.49932760e-01 1.24915397e+00 -3.73234719e-01
8.66026998e-01 1.23766661e+00 6.21379077e-01 3.76746446e-01
-1.14861655e+00 -4.06961650e-01 7.38000691e-01 2.54007071e-01
-3.34338576e-01 -3.76740307e-01 1.04951525e+00 -5.35534143e-01
8.93299103e-01 6.73659205e-01 7.31984377e-01 1.14812517e+00
-3.31348091e-01 8.63794208e-01 1.35928237e+00 -9.66229260e-01
2.17713058e-01 8.50762799e-02 9.19936895e-01 2.38288075e-01
2.30807178e-02 -5.17190993e-01 -6.36865377e-01 -4.45106059e-01
-3.07766218e-02 -5.43152392e-01 -1.28494859e-01 9.74868461e-02
-8.39978397e-01 9.78534877e-01 -1.27258599e-01 5.79062045e-01
1.07524864e-01 -1.27906606e-01 5.40502310e-01 4.53539014e-01
6.46814346e-01 7.87896454e-01 -6.04109824e-01 -7.23738611e-01
-4.43284810e-01 3.95265698e-01 1.13109517e+00 6.56770170e-01
4.34681565e-01 -3.95489633e-01 -1.23775862e-01 1.01933289e+00
1.94647983e-01 1.34914234e-01 3.99249911e-01 -9.59946990e-01
5.92229843e-01 9.89256918e-01 -1.91746682e-01 -4.75232661e-01
-4.03744191e-01 9.37453210e-02 4.99025017e-01 -1.94443285e-01
8.06237400e-01 -5.14598310e-01 -6.03432357e-01 1.80639338e+00
9.40322936e-01 2.10308079e-02 1.70008749e-01 4.99104708e-01
1.38143361e+00 1.53295517e-01 4.97129977e-01 -2.07681119e-01
1.71468866e+00 -7.98618197e-01 -9.95375216e-01 -4.24566716e-01
1.09309590e+00 -1.45356512e+00 9.71813262e-01 -2.77397670e-02
-1.20962656e+00 5.31978756e-02 -9.27063644e-01 -4.05073732e-01
-4.04080182e-01 -2.20189974e-01 7.32961357e-01 7.80788422e-01
-4.07347679e-01 3.88621211e-01 -5.03191948e-01 -1.89697310e-01
1.67061910e-01 1.97198004e-01 -8.96400139e-02 2.24776104e-01
-9.79526877e-01 8.02479625e-01 1.03503773e-02 -5.06408095e-01
-1.21779777e-01 -6.50756896e-01 -8.95842612e-01 3.28007601e-02
6.85102761e-01 -2.91118652e-01 1.60755789e+00 -7.89364219e-01
-1.46908903e+00 7.72156656e-01 -3.86161983e-01 -3.00626427e-01
5.78172989e-02 -3.25327367e-01 -9.08133686e-02 1.57690421e-01
2.87747141e-02 1.13795914e-01 4.86429751e-01 -9.88807321e-01
-7.04422116e-01 -3.70013088e-01 6.21071696e-01 5.56834161e-01
-2.39061102e-01 6.41439974e-01 9.69704464e-02 -5.50740778e-01
2.20273688e-01 -9.39912558e-01 4.08202484e-02 -6.23913944e-01
-4.65581834e-01 -7.43777692e-01 1.14303052e+00 -4.79010701e-01
1.45661008e+00 -1.76229656e+00 -4.02192846e-02 1.57102048e-01
2.70612799e-02 2.03888506e-01 2.68658549e-01 4.38904971e-01
-9.04064178e-02 4.37783867e-01 2.64742881e-01 1.59567129e-02
-2.14051325e-02 2.26493672e-01 -3.00340593e-01 -4.42356206e-02
-3.75083759e-02 5.83204150e-01 -9.34572220e-01 -7.51644969e-01
7.77713209e-02 3.57732981e-01 -4.87756163e-01 7.26554589e-03
-5.32479048e-01 2.59529173e-01 -8.10604870e-01 5.80826461e-01
8.18923302e-03 -2.04979032e-01 7.08972573e-01 2.69701391e-01
-3.36664259e-01 1.58982003e+00 -1.01961029e+00 1.24022365e+00
-5.36790729e-01 9.65986490e-01 3.63953084e-01 -9.10293400e-01
4.37841088e-01 6.37209296e-01 1.67495012e-01 -3.93389016e-01
2.86630571e-01 2.24008366e-01 4.15739626e-01 -4.65035051e-01
5.48837066e-01 2.56446689e-01 6.30971193e-02 8.91861796e-01
-2.80315578e-01 1.51003199e-03 4.57872838e-01 1.34846255e-01
1.15405095e+00 -2.13697180e-02 2.71084487e-01 -4.62005168e-01
4.82409477e-01 2.72019953e-01 5.09771228e-01 5.03943324e-01
-1.49738535e-01 2.70963490e-01 8.80052805e-01 -1.11107841e-01
-5.56495368e-01 -4.25786078e-01 -5.68869174e-01 1.65517247e+00
-1.38266981e-01 -7.12001443e-01 -1.01697171e+00 -1.09523702e+00
-8.76634121e-02 6.66919470e-01 -3.45840544e-01 2.46799186e-01
-1.23276722e+00 -6.16621017e-01 3.47434312e-01 5.78634918e-01
-3.88360396e-02 -9.88318145e-01 -8.80491495e-01 2.21016154e-01
-3.42517912e-01 -1.13879025e+00 -1.78248599e-01 6.39821351e-01
-9.40409005e-01 -1.44773877e+00 2.16221362e-01 -9.07572925e-01
4.53327149e-01 3.45694005e-01 1.20019174e+00 6.11277401e-01
9.52076614e-02 6.81513071e-01 -4.74587470e-01 -7.27724731e-01
-7.41146088e-01 3.19053233e-01 -4.91349548e-01 -8.21760535e-01
5.61969936e-01 -5.87024927e-01 -2.33262002e-01 1.87367141e-01
-5.23141563e-01 5.68192676e-02 3.30425100e-03 8.31582785e-01
2.87765235e-01 -3.38563263e-01 6.75149143e-01 -1.40481591e+00
1.12564540e+00 -2.37290949e-01 -1.80028737e-01 3.92397106e-01
-6.30556285e-01 3.95974144e-02 2.56240070e-01 -7.32507467e-01
-1.18807638e+00 -4.14566517e-01 -2.57668495e-01 4.84423250e-01
-4.69347566e-01 6.34037375e-01 -2.17414707e-01 -2.16030795e-02
5.90730131e-01 -6.01506531e-01 -1.07857257e-01 -4.66707855e-01
2.99382955e-01 6.12470031e-01 -3.05902604e-02 -1.20396447e+00
2.65568167e-01 -3.79786007e-02 -1.15683921e-01 -7.60704279e-01
-1.19179511e+00 -4.81040955e-01 -5.37982881e-01 3.11133061e-02
9.17378724e-01 -8.50422561e-01 -6.91103697e-01 -1.26121625e-01
-9.64624047e-01 -4.89212424e-01 -4.41406816e-01 6.38388634e-01
-2.55183190e-01 1.26541555e-01 -8.77557814e-01 -8.02816153e-01
-3.74758020e-02 -1.18699992e+00 7.00627923e-01 4.36285019e-01
-7.37662971e-01 -1.19286263e+00 -2.08276045e-02 1.01824903e+00
2.41754483e-02 1.73480973e-01 1.28477931e+00 -1.32554245e+00
-3.21975976e-01 9.32954326e-02 2.51985669e-01 1.86769590e-02
1.16500631e-01 1.97843373e-01 -1.32405818e+00 3.68422836e-01
6.55664578e-02 -4.32653695e-01 4.85370725e-01 2.87351608e-01
8.87789011e-01 -7.20895112e-01 -3.43325794e-01 -4.63870540e-02
6.70454383e-01 2.92024076e-01 -3.96108627e-02 6.10445678e-01
5.93620181e-01 8.34064603e-01 7.31962323e-01 -5.16756214e-02
5.95917106e-01 5.33589661e-01 1.10971227e-01 2.25377232e-01
-3.48820627e-01 -1.33088619e-01 1.06808074e-01 7.81175375e-01
4.14974689e-02 -1.85632199e-01 -1.09289694e+00 7.18908429e-01
-1.91058218e+00 -1.05517352e+00 -2.84426957e-01 1.69484961e+00
1.39830267e+00 2.24826813e-01 7.41272941e-02 4.56619382e-01
6.25106156e-01 2.62025386e-01 -1.32885650e-01 -8.15289140e-01
-5.06090000e-02 2.17670351e-01 -2.34145038e-02 7.04197705e-01
-1.21059394e+00 7.11311996e-01 6.87557459e+00 6.72882080e-01
-8.61649334e-01 2.46741474e-02 6.38864398e-01 -8.38759914e-02
-7.16849327e-01 2.49380276e-01 -8.41980100e-01 1.11811034e-01
9.73504305e-01 -4.03276011e-02 4.30802144e-02 8.24930310e-01
-8.41632187e-02 -2.63367236e-01 -1.38816988e+00 9.15874317e-02
-8.62866566e-02 -1.46205389e+00 -5.17921031e-01 1.01869181e-01
8.44184518e-01 -2.91000813e-01 -1.09585971e-01 2.88704425e-01
6.90079808e-01 -7.27276027e-01 5.43304086e-01 -2.54493028e-01
-6.03368990e-02 -6.36657834e-01 4.76509988e-01 4.11075264e-01
-1.00683403e+00 -2.27314204e-01 3.71888548e-01 -4.42633063e-01
-2.03887865e-01 3.15957397e-01 -1.15652418e+00 1.22241065e-01
4.64058191e-01 2.42692292e-01 -4.22142804e-01 4.92978424e-01
-6.82638109e-01 1.21820605e+00 -5.25504410e-01 -2.40022898e-01
2.67714769e-01 5.65419234e-02 7.62738228e-01 1.01045883e+00
-2.47382462e-01 4.28376019e-01 3.35489273e-01 3.75549436e-01
-6.10356890e-02 3.82866025e-01 -4.92912948e-01 7.01396614e-02
9.72203493e-01 1.22465587e+00 -7.70219564e-01 -3.40899318e-01
-5.40442169e-01 -1.92792505e-01 3.33475232e-01 2.14980707e-01
-4.40567851e-01 5.95015138e-02 8.76796722e-01 2.74735034e-01
2.49588236e-01 -2.69965753e-02 -6.09929979e-01 -8.79510343e-01
1.26298502e-01 -1.19247413e+00 6.36897504e-01 -3.96707654e-01
-8.84514153e-01 5.96099086e-02 3.21015239e-01 -6.84425354e-01
-6.10328019e-01 -4.40298349e-01 -1.13984144e+00 6.18281305e-01
-1.35676444e+00 -1.15761495e+00 1.23773843e-01 1.29089177e-01
5.94248354e-01 1.84452847e-01 8.43706429e-01 9.38341208e-03
-5.86829007e-01 4.61823106e-01 -2.90009230e-01 3.72883350e-01
7.56683111e-01 -1.51874471e+00 -6.29673013e-03 7.96712041e-01
4.04323727e-01 9.03464258e-01 8.07427585e-01 -6.35921240e-01
-1.16245866e+00 -4.82392699e-01 1.43969834e+00 -7.76649058e-01
7.82470822e-01 -2.13948116e-01 -9.92210388e-01 7.28015363e-01
5.35454631e-01 -2.82023460e-01 1.24894905e+00 1.10071099e+00
-5.75487673e-01 4.29729313e-01 -1.00412798e+00 6.38366103e-01
1.04036570e+00 -8.38395536e-01 -1.11942184e+00 2.12673604e-01
1.02690756e+00 -8.75631332e-01 -1.13523865e+00 3.06534320e-01
4.95995462e-01 -6.38104260e-01 1.10065973e+00 -7.43050516e-01
5.29561460e-01 1.06227204e-01 -2.14257408e-02 -9.04145002e-01
4.00414914e-01 -4.92141277e-01 -4.21449721e-01 1.63256764e+00
6.94836676e-01 -5.34201682e-01 6.26031816e-01 1.04096425e+00
-4.01060015e-01 -9.43967938e-01 -8.78108740e-01 -2.12153301e-01
4.08758074e-01 -4.35233295e-01 9.76578534e-01 1.24180949e+00
5.76154530e-01 9.29745853e-01 4.35261637e-01 -1.00334547e-01
3.85501189e-03 7.78605700e-01 7.28424311e-01 -1.54472375e+00
6.48464169e-03 -6.94178402e-01 1.18953697e-01 -8.49290013e-01
4.56494361e-01 -6.22216284e-01 -2.46611774e-01 -1.60182583e+00
-9.78867486e-02 -8.56059492e-01 4.92760018e-02 5.42971015e-01
-6.61491632e-01 -4.04958159e-01 1.97246879e-01 -3.74326706e-02
-4.83988315e-01 8.30694214e-02 1.12790656e+00 -6.08447418e-02
-4.32100922e-01 -3.06032244e-02 -8.09595346e-01 1.23948359e+00
9.09764111e-01 -4.73075122e-01 -6.63954496e-01 -3.67391050e-01
1.92789584e-01 1.25409672e-02 -3.23160295e-03 -6.18265927e-01
1.94713518e-01 -6.13052487e-01 1.40395924e-01 -6.59313917e-01
2.95074046e-01 -5.89775622e-01 -5.07235348e-01 -7.93142617e-02
-8.57418418e-01 4.89406586e-02 2.53698945e-01 2.64574170e-01
-3.86811703e-01 -6.82865798e-01 5.26414700e-02 -5.63900685e-03
-2.11329177e-01 -6.21240318e-01 -3.80249113e-01 4.67920482e-01
9.30439889e-01 -1.86023146e-01 -9.53333557e-01 -1.39962405e-01
-7.04952121e-01 1.72639504e-01 2.81342715e-01 3.55183214e-01
1.31273329e-01 -9.43758249e-01 -4.19442743e-01 -4.61226478e-02
-1.79466665e-01 1.78618371e-01 -1.57661825e-01 7.84464836e-01
-9.25521702e-02 3.23234290e-01 3.11895519e-01 -6.54248178e-01
-2.23252010e+00 1.56704336e-01 -1.91221341e-01 -5.04596472e-01
-5.89181840e-01 6.75530136e-01 -2.40991026e-01 -4.41731572e-01
2.97417969e-01 -6.92760289e-01 -7.80146539e-01 5.38079619e-01
3.97370309e-01 1.86220422e-01 8.41049701e-02 -5.25438488e-01
-2.63405055e-01 3.35663736e-01 -2.19162200e-02 -1.67587057e-01
1.23447382e+00 -1.63140371e-01 3.22632641e-02 6.72696710e-01
7.18491733e-01 7.24953830e-01 -7.21522272e-01 -2.06975430e-01
4.29814845e-01 -1.79731980e-01 9.51808840e-02 -7.97398508e-01
-4.66772646e-01 6.08431399e-01 4.77821603e-02 7.89042413e-01
7.90516436e-01 2.12119058e-01 5.50078928e-01 4.13103491e-01
-8.52979645e-02 -1.43935168e+00 -1.09539116e-02 7.65258014e-01
3.13534886e-01 -1.30742717e+00 4.46433902e-01 -7.89812148e-01
-5.28841138e-01 1.05370390e+00 7.49032855e-01 4.59401906e-01
6.90582395e-01 4.44313616e-01 1.87438339e-01 -3.60194147e-01
-1.25882936e+00 -3.63479227e-01 4.84871745e-01 3.01919997e-01
1.17403114e+00 2.04041868e-01 -6.54135644e-01 5.74298501e-01
-5.40299296e-01 -7.00619340e-01 5.91787457e-01 1.46056616e+00
-5.96352339e-01 -1.63058829e+00 -5.09779751e-01 3.79881442e-01
-9.38963234e-01 -2.03126401e-01 -9.19263959e-01 8.50298584e-01
1.58949345e-01 1.42550504e+00 7.75368512e-02 8.77983496e-02
2.18546197e-01 5.45618832e-01 3.76973420e-01 -9.56093848e-01
-1.45645475e+00 7.56772533e-02 1.09371006e+00 -2.97881126e-01
-7.89395928e-01 -9.28689301e-01 -1.38712621e+00 -2.87822247e-01
-7.67976582e-01 5.04632175e-01 7.63681591e-01 1.23649573e+00
2.21249580e-01 6.55273199e-01 1.92892715e-01 -9.25689191e-02
-2.47461036e-01 -9.77792680e-01 2.81582147e-01 4.29998070e-01
4.22720313e-01 -5.44999063e-01 -4.18509513e-01 -1.31157637e-02]
|
[10.506486892700195, 9.424967765808105]
|
ef047aae-892b-4f59-840b-0ce7f027b032
|
3han-a-deep-neural-network-for-fake-news
|
2306.12014
| null |
https://arxiv.org/abs/2306.12014v1
|
https://arxiv.org/pdf/2306.12014v1.pdf
|
3HAN: A Deep Neural Network for Fake News Detection
|
The rapid spread of fake news is a serious problem calling for AI solutions. We employ a deep learning based automated detector through a three level hierarchical attention network (3HAN) for fast, accurate detection of fake news. 3HAN has three levels, one each for words, sentences, and the headline, and constructs a news vector: an effective representation of an input news article, by processing an article in an hierarchical bottom-up manner. The headline is known to be a distinguishing feature of fake news, and furthermore, relatively few words and sentences in an article are more important than the rest. 3HAN gives a differential importance to parts of an article, on account of its three layers of attention. By experiments on a large real-world data set, we observe the effectiveness of 3HAN with an accuracy of 96.77%. Unlike some other deep learning models, 3HAN provides an understandable output through the attention weights given to different parts of an article, which can be visualized through a heatmap to enable further manual fact checking.
|
['Shrisha Rao', 'Nigel Fernandez', 'Sneha Singhania']
|
2023-06-21
| null | null | null | null |
['fake-news-detection']
|
['natural-language-processing']
|
[-3.00359219e-01 1.94472983e-01 -3.40622127e-01 -1.38961524e-01
-4.81622726e-01 -3.87979954e-01 7.56162822e-01 2.77842611e-01
-1.58673942e-01 4.21832830e-01 8.04846048e-01 -3.40138167e-01
4.33957398e-01 -8.86283219e-01 -9.74328041e-01 -4.46808964e-01
9.89902914e-02 2.44208932e-01 2.46382847e-01 -5.75501263e-01
6.72521174e-01 1.71393096e-01 -1.01175237e+00 7.97540963e-01
6.31478429e-01 9.38697457e-01 -2.42871746e-01 4.70275283e-01
-1.65865377e-01 1.20659292e+00 -1.37095082e+00 -8.08849096e-01
-2.82740640e-03 -5.31508923e-01 -7.13630795e-01 -3.64404097e-02
4.22537327e-01 -6.72857463e-01 -7.16792107e-01 1.46652818e+00
1.30039856e-01 -3.85628879e-01 6.09846711e-01 -1.27539480e+00
-1.28425395e+00 9.86820281e-01 -6.38775110e-01 5.57527721e-01
3.09198499e-01 1.50046304e-01 1.16964614e+00 -9.94149625e-01
6.28063738e-01 1.43574452e+00 7.50252128e-01 4.65313643e-02
-6.45279229e-01 -6.20505452e-01 1.04458250e-01 2.25981787e-01
-1.03900707e+00 -1.17746122e-01 8.30238283e-01 -5.62813282e-01
7.70916641e-01 7.26332515e-02 9.01352048e-01 1.51348019e+00
9.37102556e-01 1.10094380e+00 9.23633873e-01 6.18661381e-02
2.85738893e-02 2.15683058e-01 5.49002171e-01 7.33976603e-01
5.52231669e-01 -1.14988290e-01 -4.65204149e-01 -3.74562442e-01
6.33267999e-01 2.89948255e-01 -3.46843332e-01 2.18788713e-01
-1.10476410e+00 1.21514237e+00 9.94637609e-01 4.23282325e-01
-3.89663130e-01 2.55031198e-01 4.54007834e-01 5.02038777e-01
6.94957852e-01 6.56752944e-01 -1.79352954e-01 8.64327848e-02
-7.11937964e-01 2.14891344e-01 8.31028759e-01 6.96045041e-01
3.92960757e-01 -1.97196230e-02 -4.93931979e-01 4.97604162e-01
8.08912590e-02 5.64747453e-01 6.84893191e-01 -1.70119211e-01
5.96544921e-01 8.24884176e-01 1.97092518e-01 -1.82696307e+00
-2.80215979e-01 -8.19814622e-01 -9.45890903e-01 -1.40117124e-01
6.56917021e-02 -3.43417935e-02 -9.18815911e-01 1.27441359e+00
-1.72844250e-02 -9.52525809e-02 -2.14245811e-01 9.50698435e-01
1.05779219e+00 1.06946731e+00 -2.15659708e-01 1.61463797e-01
1.62473166e+00 -1.10974181e+00 -9.63202775e-01 -4.62155044e-01
5.03373682e-01 -5.43765008e-01 1.01848519e+00 2.32660711e-01
-6.74274206e-01 -1.44347474e-01 -1.25639725e+00 -4.37782496e-01
-8.12235475e-01 1.89470634e-01 4.02815074e-01 2.19642967e-01
-7.03672171e-01 4.65668589e-01 -3.47622514e-01 -6.48215786e-02
6.99408531e-01 -1.08116724e-01 -2.22324118e-01 5.07535487e-02
-1.69544089e+00 1.22687793e+00 3.09912562e-01 1.21840611e-01
-7.99443305e-01 -1.94265023e-01 -9.46000159e-01 4.02591646e-01
1.63540706e-01 -4.72268105e-01 1.00320041e+00 -9.77592170e-01
-1.22839594e+00 7.46875823e-01 2.00499520e-01 -5.87477267e-01
5.53888559e-01 -4.54991668e-01 -5.22104084e-01 2.10247681e-01
2.36885220e-01 1.56436816e-01 1.49042547e+00 -1.25824368e+00
-3.34446520e-01 -3.19771677e-01 1.59791455e-01 -1.35241628e-01
-4.62420732e-01 1.92590982e-01 -3.83623421e-01 -1.04571223e+00
-1.00754239e-02 -5.40358305e-01 1.78714752e-01 -2.61074811e-01
-8.78307998e-01 -1.71399921e-01 9.81464207e-01 -1.04602182e+00
1.55262113e+00 -2.10073256e+00 -2.26181243e-02 -8.88059735e-02
8.64040971e-01 2.28491902e-01 -1.16940834e-01 4.08728451e-01
9.34457332e-02 3.14285994e-01 -2.35990778e-01 5.89828528e-02
-5.30762188e-02 -4.02438641e-01 -6.36532784e-01 8.71324599e-01
4.44986939e-01 1.19713128e+00 -1.11207116e+00 -1.98532909e-01
-2.30319619e-01 4.02131826e-01 -4.24354196e-01 -6.52725548e-02
-1.62558973e-01 -2.23173484e-01 -4.88366574e-01 7.02929616e-01
5.88580906e-01 -5.29017270e-01 -2.10847825e-01 -1.81308076e-01
1.14433274e-01 4.62089241e-01 -4.45583016e-01 6.37386501e-01
-6.91683963e-02 1.23160219e+00 -5.41942865e-02 -9.24275756e-01
7.54950106e-01 -1.41087882e-02 -2.27952719e-01 -8.26423407e-01
7.35968173e-01 -7.36353127e-03 -5.69587238e-02 -5.39443672e-01
7.70640433e-01 5.10138795e-02 -5.97931325e-01 7.39171088e-01
-1.33013139e-02 -4.80710305e-02 -1.88263774e-01 6.83013320e-01
1.13223171e+00 -5.90312719e-01 4.21198726e-01 -2.29469180e-01
1.88716471e-01 2.96468660e-02 2.38293052e-01 1.12818539e+00
-2.95354068e-01 4.60534483e-01 1.04747367e+00 -8.75037193e-01
-1.01870894e+00 -4.23656136e-01 1.43032268e-01 1.06542122e+00
5.15160561e-01 -4.91169155e-01 -7.15541422e-01 -8.69900644e-01
3.08939278e-01 8.38568568e-01 -1.12937903e+00 -5.33125401e-01
-4.25160378e-01 -7.35246241e-01 5.01544356e-01 1.74087420e-01
6.56826854e-01 -1.35621786e+00 -5.49976707e-01 2.30537772e-01
-2.48022646e-01 -1.01399863e+00 -5.22101283e-01 -1.25799865e-01
-3.64012182e-01 -1.05936348e+00 -5.51461577e-01 -5.97522199e-01
7.26530015e-01 5.17674506e-01 1.05900633e+00 5.34559131e-01
-8.75883922e-02 -2.38068953e-01 -4.23114985e-01 -5.13416171e-01
-6.47105396e-01 -4.34323214e-02 -3.20461169e-02 2.43438974e-01
2.96225250e-01 1.27369724e-02 -4.34136927e-01 1.15233315e-02
-9.77224767e-01 -4.52262349e-02 7.48541772e-01 1.04747748e+00
-1.59986913e-01 9.57811624e-02 3.51284891e-01 -1.16843605e+00
1.19307721e+00 -9.18866575e-01 -4.69482064e-01 3.39267291e-02
-1.78736746e-01 -5.56408726e-02 7.78005302e-01 -1.94943592e-01
-4.62724239e-01 -5.64710259e-01 -4.85544093e-03 -1.78185806e-01
3.21559876e-01 7.31806219e-01 1.63924962e-01 1.89393356e-01
6.94061995e-01 4.03818607e-01 -4.68641520e-02 -2.85496414e-01
2.66567498e-01 1.00385916e+00 3.31273288e-01 2.15264574e-01
8.48919392e-01 3.79963100e-01 -7.01368690e-01 -8.36189508e-01
-1.24317813e+00 -2.12309718e-01 -3.13367754e-01 -1.86730966e-01
5.45954704e-01 -7.59228706e-01 -6.63838029e-01 7.40875304e-01
-1.48852956e+00 3.14594875e-03 3.36700737e-01 1.53109863e-01
8.93617049e-02 2.65978426e-01 -1.18136120e+00 -5.58466017e-01
-3.90539795e-01 -1.02220130e+00 1.01984763e+00 -9.78410318e-02
-9.28022414e-02 -8.31769288e-01 -3.15652579e-01 4.22145456e-01
4.82114881e-01 2.11737663e-01 9.33687270e-01 -9.69292879e-01
-3.29399168e-01 -6.75786555e-01 -5.76292574e-01 2.58939952e-01
-1.13511205e-01 1.26723349e-01 -8.29661548e-01 -3.28689635e-01
2.11250454e-01 -2.86547780e-01 1.30669534e+00 1.36271968e-01
1.04196978e+00 -1.07443726e+00 -3.19051951e-01 1.32714435e-01
1.11269510e+00 -7.04814494e-02 7.01808810e-01 5.27902484e-01
8.35971594e-01 2.58936405e-01 3.08445513e-01 3.97852957e-01
3.62217933e-01 3.12463313e-01 7.43000805e-01 -2.01734170e-01
1.75598711e-01 -2.45816216e-01 6.35099888e-01 8.86335552e-01
4.53824043e-01 -6.90624356e-01 -7.97705650e-01 5.29774487e-01
-1.64129412e+00 -1.29031169e+00 -2.43027106e-01 1.90877378e+00
5.68043470e-01 5.38731813e-01 1.57578513e-02 -2.98212804e-02
1.08066034e+00 5.66946089e-01 -4.99757558e-01 -4.16935563e-01
-2.85720885e-01 -5.71521282e-01 5.64010024e-01 4.96163994e-01
-1.38550532e+00 1.16444850e+00 6.75503159e+00 6.97896183e-01
-1.34608412e+00 1.02317326e-01 8.26375663e-01 1.09999664e-01
-3.71516496e-01 -3.32583338e-01 -4.99957472e-01 9.80244696e-01
7.05992222e-01 -1.44672319e-01 2.08299980e-01 8.92576754e-01
1.16303623e-01 1.00348212e-01 -6.11387730e-01 7.74130285e-01
4.38258171e-01 -1.76229835e+00 4.35932845e-01 -3.66300053e-04
6.00229263e-01 2.25263104e-01 1.27741128e-01 4.98395115e-01
3.82005900e-01 -1.00489461e+00 1.06102788e+00 2.33420074e-01
4.74118114e-01 -7.01709092e-01 1.20642543e+00 5.37201643e-01
-4.28427756e-01 -2.82511324e-01 -4.96480554e-01 -2.50425756e-01
-2.15653088e-02 8.02111626e-01 -7.62170434e-01 4.46385369e-02
6.69003427e-01 8.24644625e-01 -6.11816883e-01 8.12621355e-01
-5.36229908e-01 6.24470413e-01 5.74105084e-02 -5.11072218e-01
6.20493948e-01 3.13694142e-02 6.58986449e-01 1.48259985e+00
1.15935266e-01 2.00775757e-01 -1.72112547e-02 9.24799204e-01
-8.25614452e-01 -1.01156823e-01 -6.76701546e-01 -5.30715823e-01
5.02394140e-01 1.07368767e+00 -7.07141757e-01 -7.57175267e-01
-1.54489413e-01 1.22471488e+00 5.88970125e-01 1.89939335e-01
-1.04264712e+00 -7.30013251e-01 4.41407353e-01 -3.51346098e-03
4.33206946e-01 5.21765985e-02 -3.01925600e-01 -1.39628100e+00
-5.73861040e-02 -9.30002749e-01 2.60046601e-01 -8.22436333e-01
-1.38953495e+00 9.34509993e-01 -6.50360823e-01 -1.22980511e+00
4.04529184e-01 -6.23396873e-01 -6.74739063e-01 4.59290653e-01
-1.25202131e+00 -7.73233831e-01 -2.90492445e-01 2.79830903e-01
5.68325162e-01 -6.63600937e-02 4.93404508e-01 -2.35472508e-02
-8.04736972e-01 5.19584060e-01 2.28831351e-01 6.95388913e-01
4.21116263e-01 -1.11124575e+00 8.77455294e-01 8.60146761e-01
9.61208865e-02 7.43042588e-01 8.21749687e-01 -9.06270266e-01
-1.30275750e+00 -1.05824649e+00 1.21722269e+00 -7.53759980e-01
1.05585051e+00 -6.09943867e-01 -1.04300714e+00 7.28222728e-01
3.34226280e-01 -1.57367319e-01 3.55102032e-01 -1.85729608e-01
-7.96749294e-01 2.97352970e-01 -1.14362442e+00 5.46242952e-01
6.14334822e-01 -6.64926291e-01 -1.08616257e+00 7.20856011e-01
9.97186661e-01 -4.30894643e-01 -3.44599038e-01 -1.47659302e-01
5.50165892e-01 -1.06809235e+00 6.15873754e-01 -8.67786467e-01
1.10075963e+00 -1.21633537e-01 4.65520561e-01 -1.57290936e+00
-8.02244544e-01 -3.59409451e-01 -3.21929932e-01 6.46312475e-01
5.31011045e-01 -7.03453541e-01 3.44971567e-01 -9.79884490e-02
-1.50522605e-01 -6.60074592e-01 -6.17680907e-01 -4.91591185e-01
-1.29705653e-01 -2.62365222e-01 5.91816425e-01 1.35231292e+00
2.35571325e-01 6.63109481e-01 -7.45771468e-01 3.43715072e-01
2.94200867e-01 1.90129220e-01 4.74710196e-01 -1.17252147e+00
1.34152591e-01 -7.61065066e-01 -4.99466151e-01 -1.18829513e+00
1.66464206e-02 -6.70845985e-01 1.23948909e-01 -1.63731670e+00
3.74533117e-01 2.57053941e-01 -1.52534932e-01 4.39184427e-01
-3.35917324e-01 4.48384076e-01 1.03222653e-01 4.94925171e-01
-6.48540616e-01 6.97350919e-01 1.23411751e+00 -6.38164937e-01
2.06197217e-01 -2.62154847e-01 -9.83891189e-01 9.27914500e-01
6.24362409e-01 -7.21374273e-01 2.05482021e-01 -6.22688890e-01
3.41253579e-01 -1.46990016e-01 4.26615536e-01 -5.10866642e-01
2.57006586e-01 1.31609708e-01 5.48349023e-01 -6.72841072e-01
1.07619286e-01 -5.41129053e-01 -4.72101361e-01 7.35532403e-01
-3.65169734e-01 1.43213451e-01 -7.94798359e-02 7.29769766e-01
-3.89436960e-01 1.18300498e-01 8.06529105e-01 -2.76880383e-01
-5.24186254e-01 1.46521091e-01 -7.07308769e-01 -5.47093153e-02
6.81685090e-01 4.92832251e-02 -9.20725644e-01 -8.59527707e-01
-3.01857293e-01 1.04429655e-01 2.89746404e-01 6.84213161e-01
7.62856364e-01 -1.29754698e+00 -8.63497138e-01 1.96895450e-01
3.29560608e-01 -5.32165945e-01 -8.18812754e-03 7.25034535e-01
-7.57294118e-01 4.24963683e-01 -1.75305694e-01 -1.96926668e-01
-7.95996130e-01 7.23272026e-01 1.12765588e-01 -1.39637649e-01
-7.75490999e-01 9.39002037e-01 5.33851981e-02 -1.09997936e-01
2.18738522e-02 -4.20286983e-01 -4.72194850e-01 2.39080667e-01
9.27663624e-01 1.40290186e-01 7.05247894e-02 -9.68455791e-01
-4.11890417e-01 7.60609135e-02 -3.76869857e-01 3.80378574e-01
1.39451814e+00 7.97836035e-02 -3.74639839e-01 2.30539307e-01
1.33069980e+00 2.36115649e-01 -8.57129455e-01 -2.35043898e-01
8.65967385e-03 -5.37528753e-01 3.31907600e-01 -7.99210608e-01
-1.08987963e+00 6.05130494e-01 -2.13892177e-01 9.61451471e-01
5.35032213e-01 8.83987546e-02 1.06385148e+00 4.12094980e-01
1.67212978e-01 -9.30438042e-01 2.26369619e-01 9.15525615e-01
1.39702308e+00 -1.43420124e+00 2.60130465e-02 -1.70122519e-01
-7.26811349e-01 1.11111367e+00 3.48915577e-01 -3.73334795e-01
4.37087864e-01 3.63583602e-02 1.11161418e-01 -7.70332932e-01
-5.34744382e-01 2.63000697e-01 2.44625881e-01 -1.43283056e-02
1.71987608e-01 5.27725965e-02 -4.20361817e-01 7.40492344e-01
-2.72585154e-01 -4.52076495e-01 8.25916171e-01 7.26214349e-01
-7.56217897e-01 -1.26243174e-01 -4.60205913e-01 6.53679013e-01
-5.79585612e-01 -4.10859436e-01 -9.40598130e-01 7.32764244e-01
-2.01277778e-01 8.73971701e-01 1.28257498e-01 -6.91324949e-01
8.01184699e-02 -2.63368219e-01 -2.78067082e-01 -5.76213896e-01
-8.31357598e-01 -3.10059667e-01 3.70337740e-02 -6.39434516e-01
1.96687683e-01 -8.05285051e-02 -9.32582796e-01 -6.43420935e-01
-4.33395952e-01 2.90855855e-01 5.73946178e-01 1.11078894e+00
6.51137948e-01 4.34331506e-01 8.14140260e-01 -7.32211947e-01
-6.12639546e-01 -1.22550011e+00 -4.36347395e-01 5.95645726e-01
8.53716612e-01 -5.97351551e-01 -8.08947027e-01 -3.06493461e-01]
|
[8.11937427520752, 10.239374160766602]
|
462c0f41-37a9-4e58-a90b-6f7e17970725
|
joint-modelling-of-emotion-and-abusive
|
2005.14028
| null |
https://arxiv.org/abs/2005.14028v1
|
https://arxiv.org/pdf/2005.14028v1.pdf
|
Joint Modelling of Emotion and Abusive Language Detection
|
The rise of online communication platforms has been accompanied by some undesirable effects, such as the proliferation of aggressive and abusive behaviour online. Aiming to tackle this problem, the natural language processing (NLP) community has experimented with a range of techniques for abuse detection. While achieving substantial success, these methods have so far only focused on modelling the linguistic properties of the comments and the online communities of users, disregarding the emotional state of the users and how this might affect their language. The latter is, however, inextricably linked to abusive behaviour. In this paper, we present the first joint model of emotion and abusive language detection, experimenting in a multi-task learning framework that allows one task to inform the other. Our results demonstrate that incorporating affective features leads to significant improvements in abuse detection performance across datasets.
|
['Pushkar Mishra', 'Ekaterina Shutova', 'Santhosh Rajamanickam', 'Helen Yannakoudakis']
|
2020-05-28
|
joint-modelling-of-emotion-and-abusive-1
|
https://aclanthology.org/2020.acl-main.394
|
https://aclanthology.org/2020.acl-main.394.pdf
|
acl-2020-6
|
['abuse-detection']
|
['natural-language-processing']
|
[-8.69236663e-02 -4.03229520e-02 -6.22418933e-02 -3.53248179e-01
-4.39836770e-01 -4.67533708e-01 7.58992434e-01 4.19902951e-01
-6.00371420e-01 5.51313698e-01 3.34649771e-01 -7.62753487e-02
1.37029871e-01 -4.35360312e-01 -1.12889986e-02 -5.17849505e-01
-2.85428226e-01 2.53169656e-01 -2.71599561e-01 -1.38233930e-01
2.38884896e-01 4.45184618e-01 -1.24663174e+00 3.93750459e-01
5.61242342e-01 7.44912148e-01 -5.37238955e-01 4.21727985e-01
-3.53165686e-01 9.25036848e-01 -6.74751341e-01 -1.02442098e+00
-1.14438839e-01 -2.07061812e-01 -7.82885253e-01 9.06810388e-02
1.23102076e-01 -2.27158949e-01 -1.61331266e-01 1.09288859e+00
3.39892596e-01 -8.72695297e-02 5.15588462e-01 -1.30616283e+00
-2.56891340e-01 6.32231414e-01 -7.87077069e-01 2.59761274e-01
7.43255734e-01 4.83838320e-02 1.22210491e+00 -6.47350848e-01
6.93159759e-01 1.26560104e+00 5.87331653e-01 4.15149361e-01
-1.27879262e+00 -6.59076333e-01 -1.40984088e-01 1.81633011e-01
-1.04971337e+00 -5.07187188e-01 1.04349732e+00 -6.75001979e-01
9.45524693e-01 2.46946067e-01 6.26959860e-01 1.66200840e+00
-6.60750717e-02 7.78479457e-01 1.26444697e+00 -4.59526151e-01
1.79028958e-01 5.17322779e-01 3.19883436e-01 4.52511936e-01
-3.56356688e-02 -3.51002455e-01 -4.83446389e-01 -6.47793055e-01
-9.87215191e-02 -2.14541510e-01 1.36189740e-02 -8.66780877e-02
-3.62903088e-01 1.20522022e+00 1.46446005e-01 8.91005754e-01
-5.39548397e-01 -1.60834968e-01 9.77197945e-01 3.66590858e-01
8.21981549e-01 5.45344055e-01 -1.38135552e-01 -6.37916863e-01
-8.66729200e-01 2.92242855e-01 1.23632646e+00 1.47808075e-01
5.27810037e-01 -3.49758387e-01 1.34681270e-01 1.23048961e+00
3.24187517e-01 5.99009097e-02 4.68585074e-01 -5.54526925e-01
2.32219145e-01 6.13102078e-01 2.50409190e-02 -1.54568696e+00
-4.77038980e-01 -2.11656347e-01 -5.41067302e-01 5.00785299e-02
5.05308926e-01 -4.61793810e-01 -4.32610750e-01 1.87048662e+00
3.50747645e-01 -3.44044417e-01 -3.36491764e-01 7.79786766e-01
8.79890770e-02 4.38550800e-01 4.90625024e-01 -4.47718978e-01
1.15472579e+00 -5.31831861e-01 -9.42664564e-01 -4.83164281e-01
9.79625881e-01 -8.92844498e-01 8.37052405e-01 7.85058439e-01
-9.49229836e-01 1.78399295e-01 -7.82785058e-01 2.26153936e-02
-3.73631716e-01 -4.51116025e-01 7.47270286e-01 9.89384413e-01
-6.84897542e-01 5.20774186e-01 -4.72924650e-01 -5.39431989e-01
5.05462527e-01 1.05666980e-01 -4.91036654e-01 4.08600122e-02
-1.22637904e+00 1.30232894e+00 7.39342719e-02 -7.05666281e-03
-2.53624082e-01 -2.89106309e-01 -6.21321738e-01 -1.28448933e-01
4.88228172e-01 -1.42469749e-01 1.03805184e+00 -1.53670561e+00
-1.28734720e+00 1.18176973e+00 1.94843695e-01 -4.70471799e-01
7.17250347e-01 -5.05054772e-01 -4.24965084e-01 -1.88554630e-01
-1.82273120e-01 1.80401742e-01 9.35640812e-01 -9.67883766e-01
-2.61105627e-01 -6.54774189e-01 -4.42311242e-02 9.78015289e-02
-8.45981598e-01 8.52315366e-01 4.32392694e-02 -4.89392877e-01
-3.99448156e-01 -7.63856411e-01 -1.16360620e-01 3.40203345e-02
-3.65049899e-01 -3.90215456e-01 7.66936898e-01 -7.59769797e-01
1.33776665e+00 -2.43692827e+00 2.57969320e-01 2.53622904e-02
2.85703450e-01 6.19657993e-01 2.10485771e-01 8.45772505e-01
2.48798393e-02 4.81563658e-01 -2.19419077e-01 -9.64063525e-01
2.10031327e-02 4.74367946e-01 -3.01002800e-01 6.79785848e-01
1.28503904e-01 5.49202323e-01 -8.02026629e-01 -5.98621428e-01
-2.35713739e-02 5.58040798e-01 -4.74075943e-01 3.76378179e-01
3.32838297e-02 2.56063581e-01 -4.70671594e-01 3.77541840e-01
5.23080170e-01 2.57913500e-01 2.81630248e-01 4.80868191e-01
-7.92420432e-02 3.87418747e-01 -6.55557156e-01 1.31986415e+00
-5.78735948e-01 5.61518013e-01 5.89002550e-01 -8.02502215e-01
7.48862863e-01 3.58346879e-01 5.94667971e-01 -6.12880826e-01
4.29173976e-01 2.55283207e-01 3.22507441e-01 -9.31138515e-01
2.00827122e-01 -7.40957916e-01 -2.86594659e-01 5.81884682e-01
-2.93143354e-02 7.89629146e-02 5.85764367e-03 2.98899114e-01
1.14976084e+00 -6.56511635e-02 5.68182170e-01 1.66103721e-01
4.11933094e-01 -2.14118734e-01 5.48245549e-01 3.69675159e-01
-6.56146288e-01 7.23972023e-02 8.75951111e-01 -2.59250015e-01
-9.13539052e-01 -7.09736168e-01 -1.24467216e-01 1.13533044e+00
-4.26868886e-01 -4.11441445e-01 -6.72676206e-01 -6.23153806e-01
5.82715906e-02 8.34025800e-01 -5.52943707e-01 -3.27262968e-01
-4.35039371e-01 -7.75032401e-01 8.94105434e-01 -2.42655054e-02
1.87335089e-01 -9.87942278e-01 -7.32678115e-01 3.18094492e-01
-1.68846130e-01 -1.30242324e+00 9.86108780e-02 6.46936297e-02
-5.36049128e-01 -8.20195675e-01 -1.73603788e-01 -3.15714292e-02
4.64517325e-02 -2.97212362e-01 8.88648093e-01 3.53236377e-01
-5.42791069e-01 2.03696072e-01 -6.99902415e-01 -4.89495218e-01
-7.16394007e-01 2.70705372e-02 8.57312754e-02 4.50366169e-01
7.39875972e-01 -9.16250885e-01 -5.15685081e-02 -2.70663530e-01
-1.07193875e+00 -2.93311536e-01 3.75984997e-01 4.13340658e-01
-4.42343533e-01 -3.68534774e-01 5.38772464e-01 -1.12117219e+00
9.81234252e-01 -9.78668034e-01 9.83151123e-02 -2.53066927e-01
-4.97346878e-01 -3.24480742e-01 6.29212260e-01 -4.41087097e-01
-1.04808140e+00 -2.16802910e-01 -5.89572072e-01 -2.28849679e-01
-6.99987471e-01 6.66401863e-01 4.24245978e-03 3.18755247e-02
5.40179193e-01 -2.60504782e-01 9.24625248e-02 -6.66994691e-01
1.95236087e-01 1.09811318e+00 4.18003686e-02 -2.66742945e-01
6.99641407e-01 3.72058302e-01 -2.77470112e-01 -1.30461967e+00
-1.03147912e+00 -7.36945271e-01 -6.55821443e-01 -3.81777704e-01
7.60407329e-01 -5.37399411e-01 -3.86431068e-01 6.79623485e-01
-1.25048363e+00 2.89452411e-02 1.40365556e-01 1.73043311e-01
-2.90167063e-01 7.40463018e-01 -8.57927203e-01 -1.38199317e+00
-2.21017569e-01 -6.54944837e-01 5.38288534e-01 -2.58329213e-01
-8.78702044e-01 -9.70629036e-01 4.11052704e-01 6.05689704e-01
5.43943524e-01 5.30183136e-01 1.08755910e+00 -1.25012529e+00
3.54092181e-01 -5.69362938e-01 -4.22766283e-02 5.05457044e-01
-4.32709157e-02 1.65514469e-01 -1.21280503e+00 -1.15309609e-02
4.36526150e-01 -7.13121533e-01 4.91188496e-01 -2.86705822e-01
5.82634270e-01 -5.42656600e-01 -1.87739581e-01 -2.56061973e-03
1.26555681e+00 -1.13517419e-01 5.04732132e-01 2.28201047e-01
4.26717311e-01 1.00943542e+00 3.40374529e-01 8.14919412e-01
-1.07585385e-01 7.15021372e-01 6.16089165e-01 1.41318992e-01
4.19355750e-01 -7.41860494e-02 4.82975155e-01 5.00484109e-01
2.43072286e-01 -2.73276746e-01 -9.77263868e-01 5.28928280e-01
-1.77526331e+00 -8.64340127e-01 -4.14802969e-01 1.80772257e+00
8.34138036e-01 9.44417119e-02 4.23993915e-01 3.84231567e-01
2.99058467e-01 4.39151496e-01 -1.09575666e-01 -1.22127986e+00
6.96807951e-02 -3.51559147e-02 -1.15042016e-01 5.54280400e-01
-1.00552225e+00 6.95884943e-01 5.94625187e+00 5.10623991e-01
-1.11150527e+00 2.53477365e-01 5.15148044e-01 -3.05813909e-01
-6.77007809e-02 -1.73643455e-01 -3.02692473e-01 6.15971327e-01
9.79244232e-01 -8.05287510e-02 4.29787040e-01 7.78180599e-01
2.97771513e-01 -2.07762420e-01 -9.52485025e-01 7.38105059e-01
4.53732640e-01 -3.58053833e-01 -3.66971135e-01 2.63954848e-01
2.32674219e-02 1.12877302e-01 -6.33601844e-02 3.99610430e-01
-9.83086452e-02 -1.08781087e+00 5.73109090e-01 2.42509529e-01
-2.42727139e-04 -6.48063004e-01 8.32370400e-01 7.97490239e-01
-1.22385822e-01 -3.09557885e-01 1.51415050e-01 -7.26415455e-01
4.53323513e-01 8.44399571e-01 -7.23129153e-01 1.91133976e-01
6.06577575e-01 5.65207601e-01 -4.49253529e-01 8.50855410e-01
-1.82925433e-01 7.10084200e-01 -3.70623291e-01 -1.54252008e-01
4.04970586e-01 -4.77501243e-01 8.39412868e-01 1.51107204e+00
-1.54927462e-01 -7.47397617e-02 3.22070569e-02 7.91790187e-01
-7.51653537e-02 5.53223073e-01 -9.48381543e-01 -5.86591482e-01
-4.48671244e-02 1.56403792e+00 -2.88587153e-01 -4.13052365e-02
-6.22320831e-01 9.66278791e-01 8.36764932e-01 -7.17178732e-02
-6.97196305e-01 3.80102582e-02 8.61679554e-01 2.33608931e-01
-1.65845051e-01 -2.58498281e-01 -2.66757578e-01 -1.19883800e+00
6.31150827e-02 -9.83117163e-01 3.67087483e-01 -2.56688058e-01
-1.75310552e+00 2.65403390e-01 -7.64668286e-02 -3.14546227e-01
-3.38877022e-01 -5.82369685e-01 -5.65470636e-01 7.09176838e-01
-1.27629530e+00 -1.02869618e+00 1.93331212e-01 1.75017267e-01
2.86936402e-01 3.60788971e-01 9.08229709e-01 5.97413182e-01
-7.50382423e-01 2.02108815e-01 -4.01205063e-01 2.15342835e-01
7.24977970e-01 -1.09434032e+00 -1.35265619e-01 6.30415976e-01
1.23625077e-01 4.35758144e-01 1.22723627e+00 -4.12436336e-01
-1.02094352e+00 -4.46222305e-01 1.35492754e+00 -4.55046326e-01
1.33020115e+00 -7.04466343e-01 -1.20428753e+00 6.24696016e-01
3.70937079e-01 -2.99593270e-01 1.13307059e+00 5.05538106e-01
-7.05213428e-01 2.68821895e-01 -1.17694581e+00 5.55291891e-01
9.14554238e-01 -6.62523270e-01 -6.52426839e-01 4.09718275e-01
6.83451444e-02 4.71476614e-02 -7.49380589e-01 9.11927894e-02
4.64663774e-01 -1.30009592e+00 5.60246587e-01 -9.17072296e-01
6.46061122e-01 6.44219935e-01 8.86930823e-02 -1.21230817e+00
-1.29919842e-01 -5.26810348e-01 -3.84165794e-02 1.60951710e+00
2.43046165e-01 -5.12023628e-01 4.40184087e-01 1.15920997e+00
1.82473809e-01 -7.99284101e-01 -1.05349791e+00 -3.15600067e-01
2.95309305e-01 -6.06839418e-01 -6.88552260e-02 1.21059978e+00
6.29077435e-01 4.81650293e-01 -7.32599616e-01 -2.42949620e-01
1.89907044e-01 -1.02510326e-01 6.25710785e-01 -1.41129899e+00
-3.27140391e-01 -7.22740591e-01 -3.67423236e-01 -3.68260592e-01
4.00484443e-01 -6.67219460e-01 -1.24680869e-01 -9.41346467e-01
3.84684414e-01 -1.03869997e-01 2.23503821e-02 5.14873683e-01
3.96837387e-03 2.28653833e-01 2.34244049e-01 4.58212681e-02
-4.39878464e-01 3.99997771e-01 4.86822307e-01 2.16337889e-01
-9.11843553e-02 -1.19618878e-01 -5.70048630e-01 9.88235116e-01
7.94942558e-01 -5.98425865e-01 8.46347399e-03 -2.10860223e-01
7.90427864e-01 -1.89590812e-01 2.45141298e-01 -6.21820807e-01
-6.38488233e-02 7.59585109e-03 -1.38684720e-01 1.51434392e-01
6.18247390e-01 -8.91824841e-01 -1.76973552e-01 3.22770476e-01
-4.33276415e-01 -6.51222691e-02 4.86174189e-02 5.66535711e-01
-1.97193652e-01 -6.32211983e-01 8.46173584e-01 -6.68735281e-02
-1.79974571e-01 6.11835858e-03 -8.17854047e-01 9.24373865e-02
1.15847206e+00 5.55883832e-02 1.45448372e-01 -4.89179075e-01
-7.04965770e-01 8.45206231e-02 4.70762312e-01 5.43086171e-01
3.07281733e-01 -6.62511468e-01 -7.01181054e-01 -1.03588840e-02
3.59034352e-02 -6.13273442e-01 1.76805660e-01 1.04617667e+00
-1.24563081e-02 1.53270602e-01 -1.35119483e-01 -9.07510146e-02
-1.58665979e+00 7.00687528e-01 3.66928011e-01 -5.00979006e-01
-3.27615947e-01 6.99327648e-01 -4.25045192e-01 -1.72467396e-01
1.95969760e-01 3.46215874e-01 -4.68857318e-01 6.71641767e-01
3.55497092e-01 4.70973343e-01 3.90226878e-02 -8.23406756e-01
-2.93592393e-01 -3.39382440e-02 -3.99907827e-01 -9.00047347e-02
1.48359919e+00 -1.47521898e-01 -5.45233607e-01 8.94263685e-01
1.39766526e+00 1.74679130e-01 -5.81317127e-01 -9.18771699e-02
5.91526031e-01 -7.00637221e-01 5.60886227e-02 -7.54284322e-01
-6.74810946e-01 9.15150762e-01 5.91605715e-02 7.32650578e-01
7.96757221e-01 -3.06718182e-02 8.99898529e-01 7.95129240e-02
4.58132476e-01 -1.08624208e+00 7.59268180e-02 4.00623828e-01
7.08015144e-01 -1.27585018e+00 -8.12585950e-02 -4.78072613e-01
-7.19589591e-01 9.60429072e-01 3.62158269e-01 -1.41663894e-01
6.37510300e-01 2.90209532e-01 2.05011770e-01 -3.23947102e-01
-7.05794215e-01 4.68307696e-02 -1.47638202e-01 3.36294234e-01
7.77842641e-01 -1.68566164e-02 -8.46001208e-01 5.38234532e-01
3.09448987e-02 -9.93335396e-02 5.57672858e-01 8.31472039e-01
-2.13564560e-01 -1.31871998e+00 -1.61823437e-01 4.17323709e-01
-1.07668400e+00 7.19081238e-02 -1.20847011e+00 6.95432544e-01
5.95182590e-02 1.08382976e+00 -1.71934202e-01 -2.01371700e-01
1.90128163e-01 5.84988356e-01 9.98844281e-02 -6.99272752e-01
-1.16966462e+00 -1.77313939e-01 6.12670958e-01 -5.52398980e-01
-3.94774675e-01 -9.05245304e-01 -8.79089177e-01 -3.39038223e-01
-2.55571246e-01 8.76313895e-02 6.55678868e-01 1.28878367e+00
1.84978306e-01 -6.38251081e-02 6.01603925e-01 -5.02789438e-01
-8.29921722e-01 -1.16769361e+00 -8.11888218e-01 9.59803879e-01
2.89920658e-01 -4.88698542e-01 -6.74479604e-01 -5.01059234e-01]
|
[8.72294807434082, 10.409008026123047]
|
9a13e58a-a98d-4590-9574-38218d59d9d0
|
on-label-granularity-and-object-localization
|
2207.10225
| null |
https://arxiv.org/abs/2207.10225v1
|
https://arxiv.org/pdf/2207.10225v1.pdf
|
On Label Granularity and Object Localization
|
Weakly supervised object localization (WSOL) aims to learn representations that encode object location using only image-level category labels. However, many objects can be labeled at different levels of granularity. Is it an animal, a bird, or a great horned owl? Which image-level labels should we use? In this paper we study the role of label granularity in WSOL. To facilitate this investigation we introduce iNatLoc500, a new large-scale fine-grained benchmark dataset for WSOL. Surprisingly, we find that choosing the right training label granularity provides a much larger performance boost than choosing the best WSOL algorithm. We also show that changing the label granularity can significantly improve data efficiency.
|
['Oisin Mac Aodha', 'Andrew Howard', 'Serge Belongie', 'Pietro Perona', 'Marco Fornoni', 'Xuan Yang', 'Grant van Horn', 'Kimberly Wilber', 'Elijah Cole']
|
2022-07-20
| null | null | null | null |
['weakly-supervised-object-localization']
|
['computer-vision']
|
[-9.85739529e-02 -6.37602881e-02 -5.09361506e-01 -4.78221416e-01
-8.68962049e-01 -8.76958013e-01 5.57073891e-01 5.46919048e-01
-7.18772590e-01 7.43523419e-01 1.91620901e-01 -1.58604439e-02
-7.06669912e-02 -8.02285731e-01 -9.05341804e-01 -7.01909900e-01
-2.14529499e-01 4.65805590e-01 5.82413137e-01 6.58950135e-02
1.67587653e-01 6.06253088e-01 -1.53261864e+00 5.67658842e-01
3.32639754e-01 1.18042350e+00 1.48414940e-01 2.86861360e-01
-1.09360129e-01 9.00690377e-01 -6.33373976e-01 -8.80766138e-02
2.67909050e-01 -3.48251939e-01 -1.16912234e+00 -3.27313272e-03
8.69390905e-01 -1.92295574e-02 4.05521058e-02 1.12271440e+00
2.14820281e-01 1.68095678e-01 7.25812018e-01 -1.25351918e+00
-6.56031489e-01 6.15525484e-01 -3.85058135e-01 5.71354270e-01
5.55230007e-02 7.75795132e-02 1.21729171e+00 -6.94796622e-01
8.17068934e-01 1.17492580e+00 6.63240016e-01 2.29750410e-01
-1.40498388e+00 -5.32739043e-01 2.93378055e-01 7.91449994e-02
-1.66023874e+00 -3.24460030e-01 4.96553332e-01 -4.91345048e-01
6.67870879e-01 1.51233256e-01 2.14194417e-01 8.19991529e-01
-4.05582339e-02 6.13616049e-01 1.41123903e+00 -4.60082352e-01
5.50384283e-01 1.03527103e-02 4.16006505e-01 9.11271870e-01
4.82230455e-01 -3.56982313e-02 -4.88211244e-01 -1.78819790e-01
6.52064025e-01 1.47928178e-01 5.35103157e-02 -3.80030066e-01
-1.50493181e+00 8.77503097e-01 1.07776272e+00 4.55118746e-01
-9.25800130e-02 6.83351099e-01 3.37190390e-01 1.82465941e-01
4.04710084e-01 7.64934897e-01 -4.43210453e-01 1.79011896e-01
-7.84570098e-01 6.44320399e-02 5.12192070e-01 9.44952726e-01
8.64858627e-01 -2.54148841e-01 -5.01164317e-01 8.47872734e-01
-1.21612651e-02 2.52726525e-01 4.19938624e-01 -1.03235984e+00
1.78954244e-01 5.57231426e-01 4.60385978e-01 -5.35244226e-01
-6.42717838e-01 -5.41373134e-01 -3.20620894e-01 1.61196813e-01
5.50663471e-01 2.91104689e-02 -1.00738454e+00 1.75763190e+00
2.19669417e-01 1.97939351e-01 -1.57959327e-01 1.01229429e+00
9.32869077e-01 4.86460626e-01 5.50702512e-01 3.53374749e-01
1.61955678e+00 -1.15869021e+00 -3.15931201e-01 -3.03977638e-01
8.83917272e-01 -4.11342531e-01 1.19859362e+00 -6.83260784e-02
-7.51335561e-01 -4.69365895e-01 -9.31090534e-01 -3.32757801e-01
-8.33790839e-01 3.70723784e-01 8.52846682e-01 3.99070591e-01
-1.07319915e+00 3.74209881e-01 -6.56245530e-01 -5.74437499e-01
6.05702877e-01 1.71564698e-01 -4.91995096e-01 -8.51282254e-02
-8.90837133e-01 9.90112305e-01 4.93712991e-01 -4.70552891e-01
-9.87040102e-01 -4.76557493e-01 -8.21335137e-01 2.13136077e-01
5.24116516e-01 -1.96135402e-01 1.17350602e+00 -8.88573468e-01
-7.82018661e-01 1.22785461e+00 -2.22847402e-01 -6.69091463e-01
1.75398901e-01 1.02398925e-01 -2.31815010e-01 7.71129355e-02
4.97324705e-01 1.07486486e+00 5.13284206e-01 -1.30204558e+00
-1.04530013e+00 -3.11652124e-01 5.08211553e-01 3.13178115e-02
4.85534407e-02 -1.10289514e-01 -4.10147339e-01 -7.24637449e-01
2.34472677e-01 -1.01427579e+00 -2.67959893e-01 2.85520494e-01
-1.42406732e-01 -6.42687798e-01 4.08224821e-01 2.69577708e-02
6.01572692e-01 -2.44594264e+00 -2.91851789e-01 -1.16405807e-01
2.25468010e-01 -2.43389830e-01 -3.16786617e-01 1.02189444e-01
5.50217107e-02 1.12813868e-01 2.87229754e-03 -1.81033134e-01
1.71377122e-01 2.77719200e-01 -1.99453250e-01 6.51327074e-01
-1.19730877e-02 1.16858256e+00 -9.51188445e-01 -5.78380704e-01
-3.87604050e-02 6.84117302e-02 -4.29903746e-01 -1.09585389e-01
-2.82995880e-01 6.66872114e-02 -3.21591586e-01 7.93177724e-01
4.59564745e-01 -8.66010129e-01 -1.33375570e-01 -3.49648476e-01
-1.21827677e-01 2.79785216e-01 -8.82022142e-01 1.61399364e+00
-5.17973721e-01 7.36100554e-01 -2.64220834e-01 -7.89825439e-01
5.47861576e-01 -1.34184482e-02 3.60966802e-01 -8.61738801e-01
-2.44546384e-02 6.48919120e-02 -4.26083952e-02 -1.63585857e-01
3.22836220e-01 -1.69405848e-01 -3.46313059e-01 4.95340049e-01
1.60086289e-01 1.07552759e-01 3.04605544e-01 7.04832599e-02
1.14932024e+00 -1.72825575e-01 4.60532993e-01 -6.62658751e-01
-1.92266027e-03 2.90796101e-01 4.94361550e-01 1.11146331e+00
-2.95103967e-01 5.01182854e-01 5.15281975e-01 -6.82966709e-01
-8.21129739e-01 -1.13932228e+00 -4.65544969e-01 1.95732999e+00
4.82091993e-01 -2.59998232e-01 -6.93194449e-01 -9.67023373e-01
2.70653993e-01 3.88914943e-01 -8.81292462e-01 -2.49223039e-02
-1.30037278e-01 -6.12035215e-01 6.23186767e-01 7.22070098e-01
4.81608480e-01 -1.14254320e+00 -8.43547642e-01 -3.00086830e-02
-1.41178994e-02 -1.24693942e+00 -5.71891189e-01 7.93332815e-01
-5.51944315e-01 -9.78399336e-01 -5.20588279e-01 -9.93170500e-01
1.02742386e+00 2.53824472e-01 1.38671172e+00 1.15785904e-01
-2.02385560e-01 8.45497772e-02 -5.17915845e-01 -2.55399078e-01
1.46740168e-01 2.08295196e-01 5.76265715e-02 -2.23653048e-01
4.54991788e-01 -4.64611538e-02 -5.17941356e-01 3.58469039e-01
-9.39025462e-01 -1.54934391e-01 4.83404666e-01 7.97518551e-01
7.06997752e-01 1.99854359e-01 5.18321276e-01 -1.25644553e+00
3.14408660e-01 -4.59208310e-01 -8.19643021e-01 3.23681027e-01
-4.00631368e-01 3.53830904e-01 4.54077750e-01 -3.85449499e-01
-5.76592445e-01 2.62688100e-01 -9.00555030e-02 -2.49706078e-02
-4.78301257e-01 3.02102596e-01 1.58941373e-02 -2.48349771e-01
7.64425278e-01 3.41668352e-03 -6.97159588e-01 -5.69498539e-01
3.93345714e-01 4.62495297e-01 4.59926128e-01 -7.65281558e-01
4.93489772e-01 6.08368039e-01 -7.71667734e-02 -4.40396786e-01
-1.38206947e+00 -4.68245000e-01 -6.81633115e-01 2.03241184e-01
9.76621628e-01 -1.00389457e+00 -5.98056734e-01 -2.56519653e-02
-9.03115690e-01 -7.34888554e-01 -5.78182042e-01 3.35743934e-01
-4.80379611e-01 -2.94865727e-01 -4.98418987e-01 -1.17867887e-01
1.42040282e-01 -1.12435973e+00 1.30478191e+00 2.71964278e-02
-1.06725387e-01 -7.95265675e-01 -3.02811582e-02 1.65364474e-01
3.76862347e-01 1.39716804e-01 8.96872103e-01 -7.64958799e-01
-6.89657748e-01 -9.51497406e-02 -6.59750581e-01 -9.25362296e-03
1.88912436e-01 -4.88786519e-01 -8.89005959e-01 -4.15576518e-01
-4.99397933e-01 -5.87890744e-01 1.07563019e+00 2.64061570e-01
1.45726609e+00 -3.92221570e-01 -3.85289311e-01 7.22752750e-01
1.55300820e+00 -1.13696116e-03 5.31686805e-02 4.76182431e-01
6.79836869e-01 2.24540174e-01 6.60346866e-01 2.98007548e-01
3.99574131e-01 7.84706771e-01 4.35421914e-01 -1.49155408e-01
-4.02811736e-01 -1.75016746e-01 -3.70096900e-02 7.53388032e-02
1.32184580e-01 -3.13142866e-01 -1.12293673e+00 6.29238963e-01
-1.65206563e+00 -7.39127159e-01 2.34170184e-01 2.02905178e+00
7.46084094e-01 1.95195511e-01 5.41336462e-02 -2.35579804e-01
5.65398514e-01 2.85588235e-01 -4.34864998e-01 -2.43217852e-02
-1.36672566e-02 6.14245385e-02 1.02148056e+00 3.07036668e-01
-1.38840318e+00 1.17247176e+00 6.90935040e+00 8.73652697e-01
-1.23897076e+00 5.20496726e-01 6.26099527e-01 -1.95882127e-01
-6.45567896e-03 -2.48180494e-01 -9.53618228e-01 6.35687590e-01
6.62644088e-01 1.81754529e-01 3.03282768e-01 1.17267549e+00
-2.54954219e-01 -2.30776027e-01 -1.39436173e+00 1.08916259e+00
3.27006876e-02 -1.52301180e+00 -1.42009258e-01 -2.34779180e-03
8.53247225e-01 4.41168278e-01 -3.47711518e-02 3.52648824e-01
6.49171114e-01 -1.16451323e+00 9.87493932e-01 2.58120626e-01
7.47244537e-01 -5.81135631e-01 6.81240916e-01 3.74131173e-01
-1.24544561e+00 -2.08640441e-01 -5.45320988e-01 9.75392461e-02
-3.06481630e-01 3.21161538e-01 -6.74203038e-01 -1.93015337e-01
9.45689678e-01 3.68895441e-01 -1.13794637e+00 1.20374227e+00
-3.16189706e-01 6.55719042e-01 -4.34729666e-01 -1.55105554e-02
6.03106380e-01 3.65440995e-01 2.27015987e-02 1.43572903e+00
1.57616958e-01 8.59972611e-02 5.20479202e-01 5.93563914e-01
-4.21543032e-01 -2.46433794e-01 -5.40166497e-01 5.31890504e-02
6.23993456e-01 1.08823538e+00 -1.17769682e+00 -4.47779536e-01
-3.19197744e-01 1.01481998e+00 5.83089292e-01 3.54927927e-01
-7.00545430e-01 -1.32217467e-01 6.65806651e-01 3.58934224e-01
3.03335816e-01 -1.41243681e-01 -1.70671478e-01 -1.07833505e+00
-2.93459564e-01 -4.02796477e-01 5.12944102e-01 -8.02144706e-01
-1.18408906e+00 5.39148331e-01 1.32905375e-02 -1.00926745e+00
8.18679854e-02 -6.95895314e-01 -1.91007808e-01 4.25334334e-01
-1.47208095e+00 -1.20769560e+00 -4.39647257e-01 4.04667705e-01
5.51131964e-01 1.25932872e-01 6.74839854e-01 2.47799695e-01
-1.54862761e-01 4.74799901e-01 2.55877674e-01 3.76408160e-01
6.10365629e-01 -1.40681589e+00 2.44361281e-01 4.18950140e-01
6.52697802e-01 5.17951548e-01 5.53097248e-01 -2.97689110e-01
-9.39949453e-01 -1.38570905e+00 6.92638159e-01 -6.95782304e-01
5.34774184e-01 -4.89913017e-01 -6.53734684e-01 9.02036548e-01
-6.16510287e-02 9.24439549e-01 5.20257652e-01 3.10175180e-01
-6.32192910e-01 -9.75045711e-02 -1.33600068e+00 4.12446350e-01
1.13053632e+00 -7.52553403e-01 -5.61931133e-01 5.83110750e-01
8.98785472e-01 -3.14069748e-01 -7.69818008e-01 4.32590038e-01
2.88800299e-01 -5.82016826e-01 1.00619626e+00 -6.01315558e-01
9.22151506e-02 -5.53562224e-01 -3.76445800e-01 -1.35853088e+00
-7.04608262e-01 3.61530632e-01 1.56941354e-01 1.05737674e+00
2.95957297e-01 -4.92234737e-01 8.78532588e-01 4.11001205e-01
1.51641518e-01 -5.90521276e-01 -9.06664968e-01 -1.02591431e+00
-6.54189009e-03 -1.46310776e-01 8.40308547e-01 9.83682632e-01
-1.85250208e-01 3.38905036e-01 -9.69226065e-04 2.88343817e-01
7.06197679e-01 5.60088873e-01 3.89670372e-01 -1.29657686e+00
3.92347947e-02 -4.28411663e-01 -7.03808129e-01 -9.54037070e-01
2.36542717e-01 -1.05438209e+00 3.92787814e-01 -1.56483698e+00
2.48975694e-01 -8.17602694e-01 -7.61007428e-01 9.58267212e-01
2.50572730e-02 8.94053400e-01 1.89292178e-01 1.14399612e-01
-1.27349746e+00 3.81889157e-02 9.26198602e-01 -2.80348212e-01
3.67691845e-01 -2.19855964e-01 -5.99851847e-01 7.04342246e-01
6.46248937e-01 -7.71267951e-01 -2.92212278e-01 -6.06121659e-01
1.42866269e-01 -2.16667563e-01 4.56133872e-01 -1.03139079e+00
2.72349179e-01 -2.96737343e-01 4.85990942e-01 -4.15171057e-01
2.79913098e-01 -9.35841322e-01 -1.46628290e-01 3.71231407e-01
-7.32990324e-01 -2.58479059e-01 4.87297028e-02 4.23601389e-01
-1.91930130e-01 -3.46997738e-01 1.11412668e+00 -3.15696388e-01
-1.18989074e+00 3.32937002e-01 -3.22863519e-01 1.83887184e-01
1.06184220e+00 2.85543889e-01 -4.23224866e-01 1.26866490e-01
-6.86753213e-01 3.40536125e-02 7.76550531e-01 3.90763819e-01
8.16300362e-02 -1.40716720e+00 -4.29731995e-01 -4.56736609e-02
7.04739511e-01 -5.36427535e-02 -3.07922959e-01 3.53782654e-01
-4.94786233e-01 6.22769237e-01 -1.51597798e-01 -5.96709073e-01
-8.82346809e-01 5.92403948e-01 2.25005150e-01 -6.09035697e-03
-5.07070541e-01 1.15234268e+00 5.15647233e-01 -5.60529172e-01
3.79183561e-01 -6.01803958e-01 -1.46118537e-01 1.64641082e-01
4.51249301e-01 1.71796396e-01 1.44044906e-01 -9.43509817e-01
-5.87893724e-01 4.20236915e-01 1.11821480e-01 1.35564670e-01
1.18111980e+00 -3.01116146e-02 -6.12219833e-02 4.83601958e-01
1.18336916e+00 -1.14522077e-01 -1.42694795e+00 -4.13437098e-01
3.54195416e-01 -4.81570542e-01 1.87883139e-01 -8.80221605e-01
-1.03837287e+00 5.40047884e-01 8.83686483e-01 3.40537012e-01
7.96704650e-01 5.69566309e-01 3.92274410e-01 7.05781877e-01
9.86369789e-01 -9.58917737e-01 4.64466996e-02 4.45933104e-01
7.09212720e-01 -1.59442735e+00 -6.73288852e-02 -1.03228085e-01
-5.98835170e-01 4.41609114e-01 7.43096709e-01 -3.61077011e-01
5.94657063e-01 4.17226404e-01 -7.03666285e-02 -3.41007739e-01
-6.30887449e-01 -7.39068389e-01 4.30136442e-01 4.62961823e-01
4.31806207e-01 2.76666880e-01 1.18934205e-02 4.93173838e-01
-6.78174943e-02 -2.42558494e-03 7.14947805e-02 9.01043534e-01
-8.19941521e-01 -7.65031636e-01 -3.87634367e-01 7.22289979e-01
-4.10503536e-01 -1.21232338e-01 -3.12696964e-01 6.16083384e-01
6.25780642e-01 7.23027408e-01 3.30443054e-01 1.20365992e-01
1.63681999e-01 -5.11237457e-02 5.23819208e-01 -1.00370216e+00
-3.11901510e-01 -4.37980235e-01 -2.17170000e-01 -5.69912314e-01
-4.17121738e-01 -3.00867170e-01 -1.52107668e+00 -5.93521968e-02
-5.63676469e-02 3.53018403e-01 5.76103866e-01 9.95404661e-01
1.75353065e-01 2.97666460e-01 2.38843620e-01 -6.96164072e-01
-1.08816519e-01 -8.19202125e-01 -7.87984908e-01 4.64657903e-01
6.70632958e-01 -8.98003578e-01 -3.11076850e-01 1.10461831e-01]
|
[9.642200469970703, 2.0068936347961426]
|
20a5d8d9-dc25-49ff-8f0c-7f3d11abd350
|
selinet-sentiment-enriched-lightweight
|
2307.02773
| null |
https://arxiv.org/abs/2307.02773v1
|
https://arxiv.org/pdf/2307.02773v1.pdf
|
SeLiNet: Sentiment enriched Lightweight Network for Emotion Recognition in Images
|
In this paper, we propose a sentiment-enriched lightweight network SeLiNet and an end-to-end on-device pipeline for contextual emotion recognition in images. SeLiNet model consists of body feature extractor, image aesthetics feature extractor, and learning-based fusion network which jointly estimates discrete emotion and human sentiments tasks. On the EMOTIC dataset, the proposed approach achieves an Average Precision (AP) score of 27.17 in comparison to the baseline AP score of 27.38 while reducing the model size by >85%. In addition, we report an on-device AP score of 26.42 with reduction in model size by >93% when compared to the baseline.
|
['Barath Raj KR', 'Sumit Kumar', 'Shwetank Choudhary', 'Tuneer Khargonkar']
|
2023-07-06
| null | null | null | null |
['emotion-recognition']
|
['computer-vision']
|
[ 1.97823599e-01 3.11542243e-01 2.69664437e-01 -5.50286174e-01
-9.24605250e-01 -3.56953979e-01 1.20112076e-01 9.03039873e-02
-4.90727574e-01 1.25366226e-01 2.73836285e-01 4.74268824e-01
4.10439402e-01 -2.41234675e-01 -2.89452732e-01 -2.30953321e-01
1.69462264e-01 -2.89396346e-01 -4.13401753e-01 9.41807181e-02
-3.46606791e-01 -1.66516796e-01 -1.29812241e+00 6.65261686e-01
5.39630592e-01 1.93667257e+00 -4.13589865e-01 9.07640576e-01
2.78249443e-01 7.31132984e-01 -6.81756318e-01 -1.03098357e+00
9.40446258e-02 -1.70577630e-01 -3.28773648e-01 6.49024621e-02
4.24842507e-01 -2.29419634e-01 5.35310665e-03 8.59240830e-01
9.93849814e-01 -1.25687659e-01 4.38154876e-01 -1.30395007e+00
-4.16598350e-01 2.16937542e-01 -7.93004811e-01 -3.84989321e-01
3.79032910e-01 1.80706412e-01 7.52997816e-01 -9.85855758e-01
6.71802223e-01 9.27008927e-01 7.88951576e-01 7.15111673e-01
-9.09612179e-01 -7.71020710e-01 -1.33144185e-01 -3.54816973e-01
-1.21339941e+00 -5.21309257e-01 9.37827766e-01 -1.35878816e-01
1.00700462e+00 3.43286246e-01 7.99790859e-01 1.16549134e+00
3.72224182e-01 1.16916144e+00 1.27379930e+00 -1.45699665e-01
3.58696073e-01 5.12535870e-01 -6.72912449e-02 8.23153675e-01
-4.98176217e-02 -3.63187522e-01 -7.71145463e-01 -1.30256265e-01
1.29735500e-01 -9.30341035e-02 3.72123182e-01 9.77177843e-02
-7.34771430e-01 5.30975103e-01 3.19864482e-01 -1.36027575e-01
-8.03338826e-01 4.51017678e-01 1.03564179e+00 -1.12057537e-01
7.09642708e-01 3.43507588e-01 -6.21391833e-01 -5.75788796e-01
-1.03276300e+00 -2.38046423e-02 6.97380066e-01 8.19420516e-01
1.24040224e-01 1.08995132e-01 -3.57443959e-01 9.76688147e-01
3.12872529e-01 9.66983676e-01 2.38663048e-01 -9.82094884e-01
1.47032857e-01 7.44214118e-01 8.45914930e-02 -1.15056074e+00
-6.80949092e-01 -5.60472786e-01 -9.47665215e-01 1.42950803e-01
-2.34517857e-01 -7.55132556e-01 -5.68227291e-01 1.73615110e+00
4.75808978e-01 -7.71878958e-02 9.93403494e-02 9.57934558e-01
1.41303802e+00 4.92898554e-01 5.41179776e-01 -3.93565260e-02
1.72227347e+00 -1.19356120e+00 -9.02124107e-01 -2.03101397e-01
4.39536244e-01 -8.80014360e-01 1.22964001e+00 8.56403589e-01
-1.44227910e+00 -5.73915303e-01 -1.17870808e+00 -5.06381579e-02
-1.83791608e-01 8.36581767e-01 7.44116545e-01 1.03413284e+00
-9.17684197e-01 4.49780107e-01 -6.41049922e-01 -3.52256417e-01
6.22461140e-01 7.09478676e-01 -3.38113934e-01 5.20069003e-01
-7.01877832e-01 5.04968166e-01 -9.37605351e-02 1.38334343e-02
-4.91523892e-01 -6.10830784e-01 -8.72333229e-01 6.74353838e-02
9.15968120e-02 -7.98057735e-01 1.18500113e+00 -1.37662184e+00
-2.00153494e+00 1.02476323e+00 -8.32411200e-02 -2.58219242e-01
2.47358099e-01 -6.62085176e-01 -8.20930839e-01 4.96498972e-01
-2.76634693e-01 8.71959746e-01 7.91622281e-01 -7.75323868e-01
-5.14900506e-01 -3.90567720e-01 -4.00281325e-02 2.16416925e-01
-7.43374348e-01 4.31057483e-01 -6.79692149e-01 -3.35443705e-01
-4.16503757e-01 -1.23532367e+00 -1.34357855e-01 1.69698358e-01
-4.24845159e-01 7.74074793e-02 7.18621373e-01 -9.37305331e-01
1.29265285e+00 -2.43395042e+00 -1.02248929e-01 1.97769061e-01
5.93248963e-01 1.25967428e-01 -1.87193543e-01 -2.01080907e-02
1.56427212e-02 -1.13381818e-01 1.67358160e-01 -9.54970121e-01
2.31976911e-01 -2.84807593e-01 2.97955871e-01 4.62601960e-01
2.90958405e-01 1.17767251e+00 -6.47792816e-01 -4.88603473e-01
3.49720955e-01 8.45943511e-01 -5.17866850e-01 1.69788897e-01
2.04608038e-01 1.22679017e-01 -1.75643891e-01 9.02831852e-01
6.92338347e-01 -1.21505290e-01 1.57246888e-01 -6.12120509e-01
4.23040599e-01 -1.83419660e-01 -1.14251077e+00 1.84857392e+00
-8.51875842e-01 4.70786184e-01 3.68717343e-01 -1.93626538e-01
1.05203807e+00 5.17348766e-01 7.00013101e-01 -7.76119471e-01
7.54560888e-01 -1.00499645e-01 -3.53764117e-01 -5.93331277e-01
5.68087757e-01 -3.99408937e-02 -6.74361885e-01 2.83169508e-01
2.56372452e-01 2.90994972e-01 -2.08969593e-01 2.59595901e-01
1.10415459e+00 1.90988511e-01 8.99465010e-02 -1.01093367e-01
3.94411325e-01 -4.74447101e-01 4.78661239e-01 1.85699865e-01
-4.65062946e-01 4.92983013e-01 4.78628218e-01 -3.13921452e-01
-7.25721955e-01 -8.41370821e-01 2.66938061e-01 1.13952506e+00
8.67922083e-02 -1.00243568e+00 -1.06412137e+00 -8.41923535e-01
-3.90664935e-01 2.60504842e-01 -7.85245717e-01 -2.49085188e-01
4.96771485e-02 -6.10484123e-01 6.94160223e-01 7.24524379e-01
6.65545464e-01 -1.10714066e+00 -1.01718199e+00 -8.56742561e-02
-2.51156330e-01 -1.41157472e+00 -4.12110180e-01 -9.60843861e-02
-5.08924842e-01 -6.77916646e-01 -5.10013759e-01 -2.99051285e-01
5.26969373e-01 -3.70824099e-01 1.11971629e+00 -2.84095496e-01
-3.35167855e-01 6.96985602e-01 -3.17605108e-01 -7.10784018e-01
1.89946175e-01 -1.34827256e-01 9.74101480e-03 2.95537770e-01
4.05222744e-01 -2.75416523e-01 -1.27843416e+00 -1.26019458e-03
-5.78751862e-01 7.52149448e-02 5.73546469e-01 6.84285581e-01
6.26407206e-01 -4.10646588e-01 5.20316303e-01 -5.27045786e-01
5.06846845e-01 -4.00453955e-01 -4.86778393e-02 5.31488620e-02
-5.10156870e-01 -4.71952379e-01 3.95439386e-01 -5.43752968e-01
-1.21437287e+00 3.82455379e-01 -2.72766441e-01 -4.00889724e-01
-5.03998436e-02 2.25688860e-01 -1.95307076e-01 1.14135593e-01
4.49207157e-01 -2.51973420e-01 -2.04002798e-01 -1.94100454e-01
4.21740115e-01 9.17736173e-01 6.23835683e-01 -2.68699646e-01
9.88863111e-02 5.11944115e-01 -1.44178882e-01 -5.67313492e-01
-8.32156479e-01 -4.99301791e-01 -3.19419116e-01 -6.13282681e-01
1.04897559e+00 -1.47560966e+00 -1.04554689e+00 4.01719600e-01
-8.43525350e-01 -2.34336793e-01 -2.19003707e-01 3.15976858e-01
-4.56416100e-01 1.56346872e-01 -7.16437876e-01 -1.15334773e+00
-1.48803294e+00 -9.61743593e-01 1.66732085e+00 3.61911595e-01
-8.89996529e-01 -6.48132443e-01 -2.25161269e-01 5.81086516e-01
4.35516864e-01 7.41838396e-01 -2.43122503e-02 -3.76275361e-01
5.32930255e-01 -6.16356850e-01 -4.07723844e-01 5.04904509e-01
-2.16711029e-01 -1.95145532e-01 -1.29838479e+00 -1.60578519e-01
-1.32398918e-01 -6.62333071e-01 4.80170816e-01 3.49804193e-01
1.05338728e+00 3.40909278e-03 -2.03276098e-01 5.42085946e-01
1.40126526e+00 7.87991509e-02 6.64119720e-01 -9.85188931e-02
4.76402491e-01 4.14260864e-01 6.60170257e-01 8.70971262e-01
2.30446249e-01 6.36994958e-01 3.04264933e-01 -4.66811031e-01
-2.49590933e-01 -1.82872508e-02 5.72461367e-01 6.88694954e-01
1.46125928e-01 -7.49600306e-02 -4.98618931e-01 3.85181308e-01
-1.87810862e+00 -5.93674898e-01 -2.32313722e-01 1.67830181e+00
4.38041866e-01 5.39410301e-02 4.11794454e-01 -1.83569148e-01
3.69450927e-01 -2.89666243e-02 -5.44992626e-01 -9.87216413e-01
2.25315336e-02 3.96700352e-01 1.01041593e-01 1.41108483e-01
-1.17447054e+00 7.72837877e-01 5.60100794e+00 8.18899572e-01
-1.20862591e+00 4.53341603e-01 9.40810740e-01 -7.67653286e-01
2.54928976e-01 -4.83531952e-01 -3.96736383e-01 5.43057263e-01
1.22167540e+00 3.24675322e-01 1.38262957e-01 1.28133821e+00
8.95711109e-02 -2.99900115e-01 -6.47691250e-01 1.68110597e+00
3.56169045e-01 -7.25707054e-01 -6.70064569e-01 -2.64301021e-02
5.76561868e-01 -2.39056814e-02 2.38143489e-01 4.05000359e-01
-1.15936182e-01 -8.12033236e-01 7.01476216e-01 3.81220013e-01
1.13575399e+00 -1.10724056e+00 8.71229947e-01 -2.17298180e-01
-1.19566703e+00 9.67322960e-02 4.73467596e-02 -1.58892825e-01
1.17248125e-01 8.40502381e-01 -4.64376658e-01 2.13492274e-01
1.02056670e+00 4.98273045e-01 -5.15219033e-01 5.08667350e-01
-8.25030878e-02 6.54506028e-01 -5.09781420e-01 -2.71832526e-01
1.58369988e-02 2.09873751e-01 2.56022930e-01 1.74312139e+00
1.68228298e-01 9.51558352e-02 -1.71539068e-01 4.60241705e-01
-3.87432247e-01 3.96984011e-01 -3.48212928e-01 4.46518622e-02
-6.72763437e-02 2.28180671e+00 -8.18459749e-01 -5.14828682e-01
-1.58143476e-01 1.67119598e+00 -1.74055006e-02 -3.08637079e-02
-1.39482391e+00 -5.75133324e-01 5.74479878e-01 -3.84129375e-01
2.09026396e-01 2.54298359e-01 -5.06270647e-01 -1.11223924e+00
1.30494356e-01 -7.57771134e-01 1.46607041e-01 -1.00752819e+00
-1.07177126e+00 8.50813925e-01 -5.92841089e-01 -1.07978499e+00
1.06212832e-01 -5.51132739e-01 -6.08086228e-01 5.22835314e-01
-8.26564252e-01 -1.55762982e+00 -5.46531916e-01 4.17992800e-01
4.27607208e-01 9.79627818e-02 1.09638286e+00 3.94063890e-01
-7.77339876e-01 1.07252467e+00 -1.08995944e-01 1.33202150e-01
7.74637520e-01 -1.17606604e+00 8.79664794e-02 4.12453175e-01
-2.59569943e-01 5.49958885e-01 6.03011012e-01 -5.05830169e-01
-1.53351152e+00 -1.14340472e+00 6.39203548e-01 -4.98973191e-01
4.42900419e-01 -8.13515246e-01 -7.00166374e-02 2.25842550e-01
6.89360142e-01 7.21374974e-02 1.16058743e+00 3.37999761e-01
-3.50806713e-01 -3.52512479e-01 -1.54974294e+00 7.23655045e-01
8.03846300e-01 -4.16448861e-01 7.51720220e-02 5.02687916e-02
3.88861924e-01 -3.83211434e-01 -1.42480361e+00 5.05084038e-01
1.25350952e+00 -8.46274912e-01 6.37551606e-01 -1.41032517e-01
8.21123123e-01 -5.42761125e-02 -3.05041611e-01 -1.05579567e+00
-1.26781821e-01 -7.30287492e-01 -4.27160978e-01 1.46656227e+00
4.54557270e-01 -2.47794520e-02 8.06844413e-01 1.12909818e+00
1.33702800e-01 -1.32846284e+00 -5.60369015e-01 -2.76540488e-01
-6.15957439e-01 -9.11804736e-01 2.24564224e-01 4.65424567e-01
4.99902666e-01 7.30946660e-01 -7.75830567e-01 -2.58126616e-01
4.84198332e-01 -5.98718785e-03 8.86090338e-01 -6.70610368e-01
-3.38892967e-01 -1.74556762e-01 -3.47192168e-01 -4.55333263e-01
2.05169339e-02 -4.76864070e-01 -1.07713155e-01 -1.35856211e+00
6.94746971e-01 2.42615014e-01 -6.14181340e-01 6.57648027e-01
1.89525336e-02 1.00243139e+00 3.94115776e-01 -5.44528306e-01
-1.51313913e+00 5.56250989e-01 7.40354180e-01 -2.01510966e-01
-2.09873348e-01 -4.92287338e-01 -9.67273295e-01 9.63424027e-01
9.67737257e-01 -2.54447877e-01 -2.56898671e-01 -6.46464974e-02
6.87648475e-01 -1.25888541e-01 2.49585479e-01 -1.07642627e+00
-1.04997300e-01 5.98606825e-01 8.96831512e-01 -4.60281491e-01
6.92102909e-01 -9.97188807e-01 7.14371130e-02 3.26004475e-01
-2.36136138e-01 -3.13565396e-02 6.62429392e-01 1.96316689e-01
4.58727218e-02 3.34578127e-01 5.74729025e-01 3.58459145e-01
-3.82722735e-01 -8.19639955e-03 -3.13685298e-01 -2.40001842e-01
9.23972964e-01 -1.13216139e-01 -1.36209443e-01 -5.24512589e-01
-9.44829345e-01 1.06866322e-01 6.23421744e-02 5.85958123e-01
7.25423992e-01 -1.41136384e+00 -4.57429111e-01 -3.09107210e-02
2.16689840e-01 -6.55710042e-01 6.92967415e-01 1.24514103e+00
-1.12002984e-01 3.06030177e-02 -1.55453950e-01 -3.48448157e-01
-1.82787347e+00 1.48499817e-01 1.28629252e-01 -4.16933686e-01
-3.01859409e-01 7.61615694e-01 4.59051318e-02 -2.85278589e-01
1.16790310e-01 -7.28686228e-02 -1.13458438e-02 -7.30956569e-02
6.73737407e-01 5.27232587e-01 1.96618121e-02 -7.20554650e-01
-4.68066633e-01 5.41199684e-01 2.11744145e-01 -1.69326752e-01
1.40492833e+00 -2.75692582e-01 3.56549211e-02 1.96221173e-01
1.47710204e+00 1.57707185e-01 -9.58110929e-01 2.12127253e-01
-4.38781947e-01 1.73334628e-02 2.06022650e-01 -1.54704523e+00
-1.28023422e+00 6.50933087e-01 1.32240641e+00 -4.20787841e-01
1.79391372e+00 9.70741957e-02 1.21439624e+00 1.41946897e-01
5.49948551e-02 -1.45745480e+00 2.69390553e-01 3.09781004e-02
7.84847498e-01 -1.38030767e+00 9.42250118e-02 -2.93373406e-01
-1.27701461e+00 7.45357454e-01 7.39991009e-01 -3.30922127e-01
7.02491105e-01 6.31124139e-01 1.24310970e-01 -6.34896040e-01
-7.15788782e-01 -6.79079443e-03 7.28100955e-01 2.58004665e-01
5.77362895e-01 1.83052853e-01 -3.81472141e-01 1.43805921e+00
-8.90606493e-02 2.84116864e-01 -1.95846334e-01 5.82822204e-01
3.19225974e-02 -5.89898884e-01 -9.66983214e-02 3.88352960e-01
-1.06819785e+00 -1.14245735e-01 -5.77153385e-01 3.12778652e-01
4.95189518e-01 1.32697105e+00 -7.57331625e-02 -1.02404559e+00
4.11980450e-01 6.27867728e-02 4.05327111e-01 -1.25104859e-01
-1.23420107e+00 5.30326545e-01 5.21223247e-01 -1.10815871e+00
-4.68653053e-01 -3.37028563e-01 -1.31966615e+00 5.48653714e-02
-2.45585039e-01 -2.14734212e-01 1.06418002e+00 7.31577635e-01
9.60333228e-01 6.84564948e-01 5.19959629e-01 -7.40444362e-01
1.87933788e-01 -1.03336048e+00 -5.54040551e-01 6.88999593e-01
-1.34667963e-01 -2.68272012e-01 4.69565615e-02 6.17430694e-02]
|
[13.297731399536133, 5.0336689949035645]
|
b0eb5181-f514-4701-88f1-10c24b168b01
|
low-rank-convex-sparse-thermal-matrix
|
2010.06784
| null |
https://arxiv.org/abs/2010.06784v1
|
https://arxiv.org/pdf/2010.06784v1.pdf
|
Low-rank Convex/Sparse Thermal Matrix Approximation for Infrared-based Diagnostic System
|
Active and passive thermography are two efficient techniques extensively used to measure heterogeneous thermal patterns leading to subsurface defects for diagnostic evaluations. This study conducts a comparative analysis on low-rank matrix approximation methods in thermography with applications of semi-, convex-, and sparse- non-negative matrix factorization (NMF) methods for detecting subsurface thermal patterns. These methods inherit the advantages of principal component thermography (PCT) and sparse PCT, whereas tackle negative bases in sparse PCT with non-negative constraints, and exhibit clustering property in processing data. The practicality and efficiency of these methods are demonstrated by the experimental results for subsurface defect detection in three specimens (for different depth and size defects) and preserving thermal heterogeneity for distinguishing breast abnormality in breast cancer screening dataset (accuracy of 74.1%, 75.8%, and 77.8%).
|
['Xavier P. V. Maldague', 'Clemente Ibarra Castanedo', 'Bardia Yousefi']
|
2020-10-14
| null | null | null | null |
['breast-cancer-detection', 'breast-cancer-detection']
|
['knowledge-base', 'medical']
|
[ 4.65703756e-01 9.66771785e-03 1.62077006e-02 -4.04568501e-02
-8.04998696e-01 -1.69127017e-01 -4.14024293e-02 -1.75610438e-01
-6.40649274e-02 1.32085815e-01 5.33975661e-01 -1.00050397e-01
-6.19837821e-01 -3.94241035e-01 -2.20300302e-01 -1.42365503e+00
-2.71308661e-01 4.58359480e-01 -7.00919107e-02 5.64795993e-02
4.93403286e-01 4.13904846e-01 -1.19746971e+00 6.38369262e-01
1.14584994e+00 8.83932233e-01 2.22780272e-01 5.63431203e-01
1.24161534e-01 8.16453993e-01 -9.38689634e-02 9.59198773e-02
1.53413966e-01 2.59150378e-02 -4.55297261e-01 2.89083838e-01
-1.11594379e-01 -1.23998098e-01 -3.12621504e-01 7.60520518e-01
5.59739292e-01 6.23743385e-02 1.08222067e+00 -1.00871694e+00
-5.85089207e-01 1.73614308e-01 -1.25393391e+00 5.93209304e-02
4.28543448e-01 -4.42362577e-01 4.19070214e-01 -9.95568395e-01
3.31927419e-01 1.08615446e+00 8.51659358e-01 3.45732480e-01
-1.04116428e+00 -3.39667708e-01 -5.07604837e-01 2.92138219e-01
-9.76292253e-01 -5.97132027e-01 9.63543117e-01 -7.59370148e-01
9.32334721e-01 1.10907781e+00 7.47830689e-01 9.06974614e-01
5.57957768e-01 8.73145342e-01 1.39143252e+00 -5.07180154e-01
1.70238391e-01 -2.55353123e-01 2.31697574e-01 8.94522905e-01
6.12657726e-01 1.42562449e-01 -7.06289589e-01 -9.56460655e-01
5.93479037e-01 3.93514894e-02 -4.37046140e-01 -4.13642585e-01
-1.06606281e+00 4.96852398e-01 -2.86636472e-01 3.55089158e-01
-4.86021936e-01 -1.07915765e-02 5.13472378e-01 -1.82776451e-01
6.73049152e-01 2.39595413e-01 -3.81211489e-02 4.30684276e-02
-9.66499329e-01 -4.73146319e-01 3.90155166e-01 5.37879884e-01
7.08769619e-01 2.17482895e-01 -1.59473810e-02 1.05366468e+00
7.51798034e-01 1.14335811e+00 7.58751810e-01 -8.18052232e-01
2.17532143e-01 7.32221305e-01 -1.92591578e-01 -1.46838057e+00
-4.71276671e-01 2.26148605e-01 -1.29590571e+00 -2.37639740e-01
-3.59465718e-01 -6.71739131e-03 -1.18666077e+00 7.20257342e-01
4.43000823e-01 4.99830604e-01 4.71614078e-02 8.94019186e-01
8.93203199e-01 7.00723946e-01 -2.13863909e-01 -7.44404435e-01
1.19544244e+00 -4.59244490e-01 -8.28713894e-01 -1.53858557e-01
5.90163589e-01 -9.62279201e-01 5.29299915e-01 6.11946404e-01
-7.25704551e-01 -4.97773848e-02 -9.52878356e-01 1.36731550e-01
3.76150310e-01 5.82112432e-01 9.39076543e-01 9.80111957e-01
-1.04845512e+00 2.93325096e-01 -1.58794701e+00 -1.60044730e-01
4.60221805e-02 4.01669472e-01 -5.37564754e-01 -2.84278870e-01
-6.59974277e-01 3.01664710e-01 -2.21855432e-01 6.94805741e-01
-7.23050177e-01 -9.43866313e-01 -8.40589046e-01 -5.51135600e-01
-1.21052630e-01 -4.62051272e-01 4.91258532e-01 -6.83704734e-01
-1.44131351e+00 7.62116432e-01 -4.71174210e-01 2.87108153e-01
3.80418986e-01 -2.99573302e-01 -4.23859745e-01 6.63896978e-01
2.10243985e-01 -2.96934694e-01 9.64317381e-01 -1.25304878e+00
-3.33965793e-02 -6.69021070e-01 -6.01131022e-01 5.89511357e-02
-5.79793155e-01 7.91061297e-03 -4.04861271e-01 -4.66670454e-01
1.44706082e+00 -1.14660323e+00 -6.41154170e-01 -3.59082609e-01
-6.21824503e-01 1.96513027e-01 1.09662890e+00 -1.13144767e+00
9.81567144e-01 -2.17082596e+00 3.26454163e-01 8.08667123e-01
2.41582006e-01 -2.37114087e-01 1.34007096e-01 5.44749320e-01
-2.80652195e-01 -8.40394273e-02 -5.93928874e-01 -2.40889534e-01
-4.18958843e-01 2.32832402e-01 4.06813286e-02 1.05383444e+00
-2.00929791e-01 5.82007289e-01 -8.73911500e-01 -6.72277391e-01
2.15823546e-01 3.47168565e-01 -3.20420682e-01 -8.66930038e-02
7.23138869e-01 2.71669239e-01 -7.09650815e-01 1.04192472e+00
9.91264701e-01 2.55162995e-02 3.76905441e-01 -6.65983498e-01
-7.28813633e-02 -5.32178581e-01 -1.21672809e+00 1.42141509e+00
-2.40583062e-01 4.32992548e-01 5.83162248e-01 -1.12472844e+00
8.30293775e-01 6.29879713e-01 1.37848783e+00 -7.26777613e-01
-1.66189909e-01 4.59869564e-01 -3.02111953e-01 -1.20733058e+00
7.37931550e-01 -3.43808979e-01 8.99306908e-02 5.26276350e-01
-4.41204071e-01 6.92614764e-02 8.83626491e-02 5.08767664e-01
1.17339802e+00 -3.04517657e-01 -1.91475064e-01 -9.05477047e-01
5.17473280e-01 -1.27983779e-01 9.15885508e-01 2.85379857e-01
5.61433956e-02 8.56025696e-01 5.68257451e-01 4.91462182e-03
-8.45365465e-01 -7.85608888e-01 -2.39408135e-01 4.94883329e-01
1.43169120e-01 -2.56592244e-01 -1.07328027e-01 -2.67811120e-01
7.10592270e-02 -6.51247427e-02 -1.00308073e+00 -2.63829399e-02
-6.51752532e-01 -1.47430599e+00 2.64731407e-01 5.14906883e-01
2.43815407e-01 -2.11207256e-01 -3.26157629e-01 -1.14855319e-01
-4.86360550e-01 -6.48863137e-01 2.23859787e-01 2.38588601e-01
-1.25633597e+00 -1.41875541e+00 -7.78317571e-01 -5.51152587e-01
1.33633375e+00 5.79213798e-01 6.56691611e-01 8.28846544e-02
-6.53745234e-01 7.99858689e-01 -5.79263866e-01 1.16711393e-01
-3.31927329e-01 -6.48488581e-01 1.34189829e-01 2.29703799e-01
-2.87459522e-01 -3.95374656e-01 -7.91888952e-01 5.62623382e-01
-9.70540404e-01 7.72611797e-02 6.60655022e-01 9.13960099e-01
6.18670046e-01 2.52488047e-01 -2.54336625e-01 -1.25197136e+00
2.83573359e-01 -5.49853921e-01 -2.33589813e-01 4.53753948e-01
-4.47316825e-01 -1.83380976e-01 -3.54535878e-02 -1.48606583e-01
-1.44921994e+00 2.45327041e-01 1.14924818e-01 -4.17919070e-01
3.75323981e-01 7.88845599e-01 1.39531465e-02 -3.70346159e-01
8.76202047e-01 2.91632324e-01 1.86795425e-02 -2.52112627e-01
-3.36160928e-01 3.76893252e-01 1.87213078e-01 -7.89461434e-01
9.14373100e-01 1.02262044e+00 3.11849564e-01 -1.16817629e+00
-2.80530483e-01 -1.11752474e+00 -5.90073228e-01 -4.75155503e-01
5.44154406e-01 -7.74158061e-01 -5.11796534e-01 8.25311184e-01
-3.95367473e-01 -1.27756417e-01 1.41983494e-01 7.70145774e-01
-2.43113667e-01 1.14423048e+00 -9.12105620e-01 -9.15450871e-01
-5.74849129e-01 -9.76390123e-01 1.05232012e+00 -1.76251724e-01
1.31380603e-01 -1.00233889e+00 2.24505380e-01 7.40551949e-01
4.07333106e-01 4.88430470e-01 1.14811194e+00 3.20015162e-01
-2.02433124e-01 -2.34773591e-01 8.13961104e-02 2.88148433e-01
2.50535041e-01 3.87209505e-01 -1.09949577e+00 -5.13599813e-01
4.40965623e-01 -2.36676633e-02 9.54605818e-01 7.19397366e-01
8.84448767e-01 -5.39160557e-02 -7.45940983e-01 5.51120162e-01
1.65229905e+00 2.39301875e-01 9.32006836e-01 -7.36968443e-02
9.22192752e-01 6.63478255e-01 6.90124571e-01 5.02757967e-01
-2.91443408e-01 -3.21988389e-02 1.35579363e-01 -3.34297925e-01
2.49572888e-01 4.38220799e-01 3.09904724e-01 1.59567308e+00
-2.71966547e-01 1.80474907e-01 -1.39640510e+00 6.06007218e-01
-1.97668314e+00 -7.16994524e-01 -8.16000640e-01 1.88816941e+00
2.44369060e-01 -3.37840974e-01 -4.77752596e-01 7.40101218e-01
6.13114536e-01 6.15731962e-02 -2.90457755e-01 -1.23062111e-01
-1.40150547e-01 -1.47694916e-01 7.98177481e-01 1.66482747e-01
-8.93032849e-01 1.10897377e-01 6.86293983e+00 9.08416748e-01
-1.29126418e+00 7.89815933e-02 6.97200954e-01 2.39572436e-01
-7.70680010e-01 3.08103114e-02 1.36531621e-01 4.30139946e-03
5.73321998e-01 2.40396947e-01 -1.24426268e-01 4.41150308e-01
3.57531309e-01 -7.02952564e-01 -7.72312701e-01 1.21960366e+00
1.45256951e-01 -1.07341015e+00 -3.41955900e-01 -8.89405832e-02
9.74519074e-01 -1.23372100e-01 2.45980084e-01 -3.02046567e-01
-2.22487077e-01 -6.64451063e-01 4.75095570e-01 5.46826124e-01
8.01862657e-01 -4.51158643e-01 5.90858221e-01 5.89771103e-03
-1.09885800e+00 -2.33209744e-01 -5.99435627e-01 2.70828277e-01
-4.02979888e-02 1.28037870e+00 -7.09321380e-01 9.74749923e-01
9.40221012e-01 8.32878470e-01 -3.13685954e-01 9.86358404e-01
8.91820565e-02 9.71004307e-01 -4.02218044e-01 1.55853137e-01
-5.34829199e-02 -5.21393239e-01 6.13984406e-01 1.07890677e+00
4.27732676e-01 3.38236719e-01 6.45169010e-03 3.15648079e-01
6.43285275e-01 3.21740568e-01 -3.67284507e-01 -3.18516567e-02
3.31711739e-01 1.48168480e+00 -1.05804491e+00 -3.67126130e-02
-2.40171298e-01 6.74728036e-01 -3.86513472e-01 4.91285056e-01
-4.10465240e-01 -6.61115721e-02 8.04721788e-02 2.41252854e-01
-3.34153503e-01 -4.22160923e-01 -6.13083303e-01 -1.04695916e+00
5.12202904e-02 -6.00878894e-01 3.34213734e-01 -7.14453220e-01
-1.22794151e+00 4.55881543e-02 7.44931176e-02 -1.25835598e+00
1.69303373e-01 -5.95295012e-01 -5.93290567e-01 5.82942545e-01
-9.76584315e-01 -1.23201203e+00 -4.56174463e-01 6.63560569e-01
-8.96836072e-02 -1.05986342e-01 5.83824217e-01 3.53212655e-01
-8.35961461e-01 2.42900342e-01 8.22533011e-01 1.54715255e-01
3.02702844e-01 -1.11301172e+00 -2.04848006e-01 9.55238938e-01
-5.19576967e-01 7.64233053e-01 6.02549076e-01 -9.34618771e-01
-2.31892228e+00 -7.05164790e-01 2.50616819e-01 -7.78653175e-02
6.47251129e-01 -2.78846711e-01 -6.87797785e-01 2.60990351e-01
-5.96207194e-02 -1.97423160e-01 1.04430926e+00 2.70641714e-01
-9.91225317e-02 -9.01798680e-02 -1.42114103e+00 1.64392412e-01
5.27298510e-01 -5.34983635e-01 -1.70826420e-01 5.80859900e-01
-7.10986257e-02 -4.72103775e-01 -1.20021486e+00 7.46244192e-01
5.50944507e-01 -9.37047720e-01 9.90686119e-01 1.40908182e-01
3.76564294e-01 -1.65417627e-01 -2.00547680e-01 -7.12668598e-01
-8.01193595e-01 -4.45747554e-01 2.75302738e-01 1.04493177e+00
4.04170781e-01 -6.41070008e-01 1.47982681e+00 7.42707908e-01
-6.86999381e-01 -6.08975589e-01 -1.12852514e+00 -2.26348564e-01
-3.23593408e-01 -3.44963104e-01 -1.88175932e-01 1.33052492e+00
3.24625671e-01 -2.86656410e-01 -3.86594474e-01 3.70143414e-01
7.13213980e-01 1.52231231e-01 8.63048285e-02 -1.02079153e+00
-1.46962121e-01 3.34737688e-01 -4.21573639e-01 -2.83070624e-01
-3.19618247e-02 -5.22569001e-01 7.91732743e-02 -1.52650344e+00
5.80882490e-01 -8.34653437e-01 -2.11596429e-01 5.25332451e-01
7.88193848e-03 3.01079303e-01 -5.32432914e-01 5.27758777e-01
-6.83079194e-03 3.85444820e-01 1.23954308e+00 -3.85571420e-01
-2.41177548e-02 -2.77094096e-01 -1.97421983e-01 5.04580855e-01
4.67442036e-01 -2.96556741e-01 -8.10757101e-01 -4.67398524e-01
5.86164892e-01 5.59471965e-01 1.27358800e-02 -8.97901952e-01
3.05902302e-01 -4.11654979e-01 5.32420695e-01 -7.26107955e-01
4.62371528e-01 -7.64437675e-01 5.89800298e-01 6.32770896e-01
2.34023109e-01 -1.15897156e-01 4.91004027e-02 8.07220995e-01
-3.67213219e-01 -1.68012753e-01 7.34744668e-01 -2.56771266e-01
-7.72182047e-01 8.09051916e-02 -7.71261513e-01 -4.97960776e-01
8.59747767e-01 -6.77763224e-01 -3.08995903e-01 1.80919498e-01
-6.77480340e-01 -5.68706654e-02 2.92413950e-01 7.36898258e-02
1.17619884e+00 -1.17751014e+00 -7.23375916e-01 3.21406484e-01
2.83569451e-02 -1.02229148e-01 1.13855410e+00 1.55423403e+00
-9.87235308e-01 1.29172787e-01 -2.27029756e-01 -1.12149453e+00
-1.44303966e+00 1.19520567e-01 2.72949725e-01 -1.02683492e-01
-4.09049898e-01 1.05500221e+00 2.87586093e-01 -2.47263223e-01
-3.87263626e-01 -3.49109143e-01 -1.45060003e-01 1.73013613e-01
3.13360132e-02 1.01035404e+00 6.50752783e-01 -9.14335728e-01
-3.98774892e-01 9.06522095e-01 2.10343957e-01 2.28783954e-02
1.42523050e+00 -1.57509118e-01 -8.22134495e-01 3.89839143e-01
1.36168134e+00 1.20623633e-01 -5.87802708e-01 1.06422987e-03
-2.43495572e-02 -4.72188413e-01 3.18384528e-01 -4.12873298e-01
-1.56877744e+00 5.13129473e-01 8.71761918e-01 -3.06532718e-02
1.22696519e+00 -3.77090931e-01 5.27006269e-01 2.97599912e-01
4.76340085e-01 -1.07648289e+00 2.72393554e-01 4.10728231e-02
8.77039373e-01 -8.56195927e-01 5.59661806e-01 -9.78394806e-01
-5.21256626e-01 1.17643714e+00 3.60453457e-01 -1.63134068e-01
8.56672406e-01 5.04495025e-01 2.00948939e-01 -4.94119763e-01
-6.19414508e-01 4.92245436e-01 3.85471970e-01 5.88948369e-01
4.67814326e-01 2.51200259e-01 -3.36023539e-01 -1.80811528e-02
2.09382296e-01 -4.64661986e-01 5.29231250e-01 1.43284583e+00
-3.25093091e-01 -6.56154037e-01 -9.77460563e-01 8.17919910e-01
-4.39734161e-01 2.03355297e-01 -2.66515166e-01 3.59134704e-01
-2.30771795e-01 1.18683040e+00 -3.56325060e-01 -6.26804173e-01
-1.81196127e-02 -4.57349747e-01 3.85278314e-01 -3.84935290e-01
-5.45021296e-01 6.67115271e-01 2.52414286e-01 -7.49951482e-01
-5.81116140e-01 -9.58068848e-01 -1.14801323e+00 -1.58157483e-01
-5.40527582e-01 4.41949636e-01 6.69914603e-01 6.19932115e-01
4.06788737e-02 2.36295551e-01 7.81946838e-01 -6.00730777e-01
1.04961418e-01 -1.03324139e+00 -1.17129290e+00 3.47441047e-01
2.51448095e-01 -5.95974803e-01 -4.63217884e-01 1.01916716e-01]
|
[12.22036361694336, 0.23648008704185486]
|
8de9d460-29fb-479a-abd9-1038cc92643e
|
hybrid-instance-aware-temporal-fusion-for
|
2112.01695
| null |
https://arxiv.org/abs/2112.01695v2
|
https://arxiv.org/pdf/2112.01695v2.pdf
|
Hybrid Instance-aware Temporal Fusion for Online Video Instance Segmentation
|
Recently, transformer-based image segmentation methods have achieved notable success against previous solutions. While for video domains, how to effectively model temporal context with the attention of object instances across frames remains an open problem. In this paper, we propose an online video instance segmentation framework with a novel instance-aware temporal fusion method. We first leverages the representation, i.e., a latent code in the global context (instance code) and CNN feature maps to represent instance- and pixel-level features. Based on this representation, we introduce a cropping-free temporal fusion approach to model the temporal consistency between video frames. Specifically, we encode global instance-specific information in the instance code and build up inter-frame contextual fusion with hybrid attentions between the instance codes and CNN feature maps. Inter-frame consistency between the instance codes are further enforced with order constraints. By leveraging the learned hybrid temporal consistency, we are able to directly retrieve and maintain instance identities across frames, eliminating the complicated frame-wise instance matching in prior methods. Extensive experiments have been conducted on popular VIS datasets, i.e. Youtube-VIS-19/21. Our model achieves the best performance among all online VIS methods. Notably, our model also eclipses all offline methods when using the ResNet-50 backbone.
|
['Yan Lu', 'Xiao Li', 'Jinglu Wang', 'Xiang Li']
|
2021-12-03
| null | null | null | null |
['video-instance-segmentation']
|
['computer-vision']
|
[ 2.84938991e-01 -1.88843146e-01 -4.81936216e-01 -4.50319260e-01
-8.20323825e-01 -5.38286984e-01 5.65078378e-01 7.40815401e-02
-3.23721558e-01 4.87612784e-01 -5.26028611e-02 6.01450130e-02
-9.35793445e-02 -5.34209609e-01 -9.97064114e-01 -6.06204987e-01
2.67580990e-02 -6.32683933e-02 6.45773053e-01 8.89150202e-02
1.90747410e-01 8.85878503e-02 -1.77841401e+00 4.67166513e-01
6.97306037e-01 1.49297094e+00 3.17848593e-01 5.50920665e-01
-2.07680121e-01 1.15923846e+00 -2.96294540e-01 -3.43252748e-01
4.60172594e-01 -3.56197357e-01 -9.77797508e-01 5.41748583e-01
9.01949346e-01 -4.86056209e-01 -5.36265135e-01 9.32744026e-01
8.62638280e-02 3.96117836e-01 -3.22492532e-02 -1.38707447e+00
-4.52423513e-01 4.06588703e-01 -7.41033077e-01 4.55628812e-01
3.40135753e-01 3.00007045e-01 1.04923141e+00 -7.67990470e-01
1.00400460e+00 9.67144370e-01 4.45698202e-01 2.50597566e-01
-1.12232983e+00 -5.73558867e-01 7.29932129e-01 6.99070096e-01
-1.31388938e+00 -4.31959063e-01 7.67769635e-01 -5.00381112e-01
8.54628980e-01 1.73723653e-01 1.02223980e+00 8.48669350e-01
-6.34519234e-02 9.89593029e-01 9.67753172e-01 -1.04585253e-01
1.27769157e-01 -3.54402453e-01 2.03766853e-01 7.98188150e-01
-3.45553845e-01 1.46295717e-02 -7.72526562e-01 2.63283402e-01
8.54107320e-01 3.25025558e-01 -3.61103863e-01 -3.26183766e-01
-1.35665071e+00 3.84163499e-01 5.35277009e-01 3.48507762e-01
-4.19101983e-01 4.30343300e-01 4.30212557e-01 1.05940111e-01
4.80024785e-01 -1.39139980e-01 -6.25293732e-01 -1.90719873e-01
-1.37299848e+00 2.51039177e-01 5.51701665e-01 1.26733267e+00
1.07983983e+00 -2.92971611e-01 -5.01166403e-01 7.32055902e-01
1.50932819e-02 1.56271666e-01 1.66574776e-01 -1.32339716e+00
4.38864976e-01 5.42481482e-01 -1.67835549e-01 -1.06002426e+00
-8.87302309e-03 -3.52137476e-01 -4.90748733e-01 -2.15164497e-01
3.49379152e-01 2.14333653e-01 -1.16693795e+00 1.86454642e+00
6.56866193e-01 1.18217921e+00 -3.09795171e-01 7.82995343e-01
6.47583306e-01 5.78167260e-01 2.17466190e-01 -2.86903262e-01
1.37671828e+00 -1.27664733e+00 -7.61488914e-01 -5.45943901e-02
3.90070736e-01 -5.91120005e-01 5.61390996e-01 1.42018333e-01
-1.35445654e+00 -6.81499481e-01 -9.01923954e-01 -3.30405176e-01
-2.68050164e-01 1.86208952e-02 5.39891005e-01 1.00090638e-01
-1.09290898e+00 5.68381965e-01 -1.06722093e+00 -2.87532628e-01
5.54498136e-01 3.80889505e-01 -4.48022217e-01 -2.67340928e-01
-9.88492310e-01 2.89140940e-01 6.06221557e-01 1.30349189e-01
-9.22389746e-01 -9.23114479e-01 -1.02060914e+00 -9.18301046e-02
8.24944496e-01 -7.36546159e-01 1.19544220e+00 -1.41323674e+00
-1.28247535e+00 7.62570262e-01 -5.84501028e-01 -6.50455415e-01
4.95374709e-01 -1.19974948e-01 -2.57307321e-01 5.05059004e-01
2.09249422e-01 8.55047822e-01 1.01564527e+00 -1.21025419e+00
-1.17315376e+00 -1.12051591e-01 6.42505705e-01 1.25585571e-01
-2.91541010e-01 -4.26315367e-02 -1.34633183e+00 -6.90206647e-01
2.22197056e-01 -9.80365336e-01 1.48022994e-02 2.30922818e-01
-2.36165747e-01 -1.58944160e-01 1.04637325e+00 -8.52998137e-01
1.48582470e+00 -2.21101165e+00 2.91295946e-01 1.65287226e-01
3.30048651e-01 1.81454718e-01 -1.72252476e-01 -3.05505563e-02
-2.17361283e-02 -1.35959670e-01 -3.72157425e-01 -5.63965201e-01
-1.48670286e-01 4.58783746e-01 -1.87797546e-01 3.28724682e-01
3.09647858e-01 1.09815979e+00 -1.02374876e+00 -6.89864397e-01
4.78549629e-01 5.96553862e-01 -8.52449715e-01 8.12792480e-02
-4.11695629e-01 5.22013783e-01 -3.69340569e-01 7.48616219e-01
6.37654126e-01 -4.17057246e-01 1.17289819e-01 -5.82109332e-01
-1.59749895e-01 9.86363143e-02 -1.03428340e+00 2.22650838e+00
-3.52769583e-01 7.00672328e-01 5.54380640e-02 -1.09337568e+00
3.48682761e-01 2.69246817e-01 9.13554430e-01 -7.98374891e-01
2.20680926e-02 3.81677672e-02 -4.32721496e-01 -4.96111929e-01
6.22265995e-01 4.36484367e-01 1.44604906e-01 -8.99609458e-03
2.91253358e-01 2.20412388e-01 5.69741309e-01 4.46009904e-01
9.48686540e-01 6.90332413e-01 -2.74976715e-04 -3.82309519e-02
7.09176421e-01 -1.04792388e-02 8.38746369e-01 6.36897385e-01
-3.53024453e-01 7.73211658e-01 3.80222738e-01 -4.76772964e-01
-7.51267731e-01 -7.14986444e-01 -6.10565655e-02 1.02865875e+00
6.35896802e-01 -7.35666990e-01 -8.78249645e-01 -7.03304589e-01
-1.55242473e-01 1.07865460e-01 -7.43909001e-01 1.49807855e-01
-8.31559837e-01 -2.77912706e-01 1.11765645e-01 6.69514477e-01
7.66312063e-01 -6.33978903e-01 -4.47319329e-01 5.22624969e-01
-5.11526465e-01 -1.65855753e+00 -8.26188505e-01 -5.88161908e-02
-7.31024265e-01 -1.24644554e+00 -5.17968714e-01 -7.25594223e-01
4.59140569e-01 4.89118844e-01 1.09742475e+00 3.94566149e-01
-2.63819993e-01 7.58806765e-01 -5.19393206e-01 2.05535382e-01
2.03373969e-01 5.65827359e-04 -3.95071238e-01 3.78180653e-01
1.37905344e-01 -4.86545831e-01 -8.95555377e-01 3.83784711e-01
-1.11659718e+00 3.56337965e-01 1.07802272e-01 7.43568182e-01
1.05205476e+00 -1.31680846e-01 2.37325683e-01 -7.28217900e-01
-2.75136828e-01 -4.62153554e-01 -6.01379931e-01 4.30753499e-01
-1.69377610e-01 -1.65686741e-01 1.53446853e-01 -4.49204892e-01
-1.03445828e+00 1.25391275e-01 1.44947097e-01 -9.22307849e-01
4.61941771e-02 5.09166002e-01 -1.87685058e-01 -1.49571285e-01
3.10503356e-02 2.75751352e-01 -3.18513244e-01 -1.84329152e-01
4.69918132e-01 2.30950475e-01 8.93872440e-01 -8.85095954e-01
6.68238342e-01 7.65015423e-01 -2.44872555e-01 -5.97611725e-01
-9.72980857e-01 -6.56656206e-01 -7.69301653e-01 -4.18876022e-01
1.20219588e+00 -1.21337843e+00 -4.87105846e-01 5.55110693e-01
-1.11841607e+00 -5.03165841e-01 -2.93030590e-01 2.38345847e-01
-7.70663500e-01 4.09730464e-01 -7.09379852e-01 -3.49454105e-01
6.51462823e-02 -1.39301026e+00 1.29486156e+00 2.52449304e-01
1.20139368e-01 -8.75214994e-01 -2.65174985e-01 4.26405936e-01
1.53704107e-01 4.46833968e-01 3.41003925e-01 -1.88344613e-01
-1.31173897e+00 1.10392407e-01 -4.14329052e-01 1.76013887e-01
1.04700878e-01 3.58039796e-01 -8.70692670e-01 -1.58849731e-01
-1.95695043e-01 4.51767556e-02 1.03946829e+00 5.60926437e-01
1.55929792e+00 -2.07097158e-01 -3.91393751e-01 9.54754412e-01
1.48645341e+00 1.95687443e-01 6.08450711e-01 3.47860157e-01
1.01964092e+00 4.44108754e-01 8.07889998e-01 5.15130341e-01
6.94875598e-01 9.54025984e-01 4.12649661e-01 5.43952323e-02
-3.73723328e-01 -1.06778383e-01 2.23884895e-01 6.58423662e-01
-2.83481389e-01 -5.45075573e-02 -6.28627062e-01 6.36915863e-01
-2.27456307e+00 -1.14219022e+00 -1.39397616e-03 1.87187254e+00
7.60306597e-01 -8.66069198e-02 2.22723093e-02 -1.18405327e-01
9.10112977e-01 3.51838052e-01 -5.64011574e-01 2.18620554e-01
-1.15206033e-01 1.05567783e-01 4.63754535e-01 4.40861702e-01
-1.41853094e+00 9.81780171e-01 5.11537313e+00 8.99262846e-01
-9.98543799e-01 3.63201767e-01 8.04551184e-01 -3.96790236e-01
-8.27789009e-02 2.93118596e-01 -6.73821032e-01 6.39826775e-01
5.71539223e-01 -4.38533314e-02 5.32264411e-01 5.58589339e-01
2.09717989e-01 -1.14277542e-01 -1.21491051e+00 1.05210471e+00
6.89346790e-02 -1.58723676e+00 -7.56667629e-02 1.47842476e-02
1.03530204e+00 -1.19802900e-01 -5.11272848e-02 2.31989384e-01
-6.85335621e-02 -5.08575201e-01 1.27074409e+00 6.60959780e-01
7.59578586e-01 -6.43434703e-01 3.08168799e-01 -7.97002912e-02
-1.82827556e+00 -1.07197464e-01 -1.45349065e-02 2.53589600e-01
3.35349172e-01 5.58790088e-01 -2.13483889e-02 9.28230703e-01
9.08240557e-01 1.50391006e+00 -4.68300164e-01 1.06244481e+00
-4.21278030e-02 3.66411746e-01 -3.75844181e-01 8.21793854e-01
3.91703069e-01 -1.31707549e-01 3.55707794e-01 1.14690602e+00
2.50246495e-01 2.70805001e-01 4.41493690e-01 7.06417143e-01
-1.26363590e-01 -1.12645246e-01 -1.59393787e-01 1.67299360e-01
4.03038889e-01 1.26436436e+00 -8.22793007e-01 -6.96919680e-01
-6.97448492e-01 1.08982456e+00 2.07013696e-01 5.67168832e-01
-1.28703296e+00 2.28834569e-01 8.83814752e-01 -9.17034503e-03
6.89511001e-01 -3.23568493e-01 4.08407189e-02 -1.45047438e+00
3.00908864e-01 -7.64865637e-01 5.32602251e-01 -7.09663451e-01
-1.02005172e+00 4.72082227e-01 2.23049223e-01 -1.36201847e+00
-1.76762938e-01 -3.17857265e-01 -4.03517216e-01 3.98242563e-01
-1.85592186e+00 -1.34454322e+00 -4.82981473e-01 1.10540760e+00
8.40454280e-01 3.09292883e-01 1.25327691e-01 6.20726287e-01
-7.63456404e-01 5.45623660e-01 -2.47830659e-01 2.01011926e-01
6.13372087e-01 -9.57576156e-01 1.13226898e-01 1.04404151e+00
3.11808437e-01 6.43754840e-01 2.68784583e-01 -5.97820044e-01
-1.38077807e+00 -1.34221780e+00 4.95307505e-01 -3.38391453e-01
6.79352045e-01 -2.46677548e-01 -1.01961696e+00 8.37991416e-01
2.22423002e-01 6.80368960e-01 2.56075680e-01 -3.41750294e-01
-5.19690096e-01 -1.82492584e-01 -8.63021076e-01 4.29604948e-01
1.47350717e+00 -7.82448471e-01 -2.08711237e-01 2.46217623e-01
1.04423845e+00 -7.58381486e-01 -1.02552617e+00 3.99351120e-01
5.60000896e-01 -9.64078963e-01 1.10089767e+00 -2.77237654e-01
5.47674537e-01 -6.40753567e-01 -2.82636493e-01 -5.70419669e-01
-1.43074065e-01 -6.02928340e-01 -4.77839112e-01 1.53640831e+00
-3.57361138e-02 -2.80555457e-01 6.00611567e-01 9.02957022e-01
-2.27617070e-01 -7.69355536e-01 -1.04031146e+00 -6.05094731e-01
-3.79078686e-01 -6.62171245e-01 6.12158060e-01 9.92276847e-01
-2.58382738e-01 -3.67816150e-01 -3.67727548e-01 1.73370734e-01
5.77653050e-01 3.44074547e-01 6.34790599e-01 -8.02774191e-01
-4.24788564e-01 -3.48967731e-01 -6.56374335e-01 -1.22268653e+00
3.87853831e-01 -8.40123057e-01 -1.44490832e-02 -1.31400585e+00
2.19617620e-01 -3.85194957e-01 -5.57656705e-01 7.54804730e-01
-2.92776257e-01 4.68705326e-01 4.76950735e-01 2.77781099e-01
-1.09405267e+00 3.11773449e-01 1.10821486e+00 -2.90804744e-01
-8.38282034e-02 -5.49920261e-01 -2.79955745e-01 5.18395424e-01
4.79451925e-01 -3.49427551e-01 -4.76186723e-01 -5.76041639e-01
-3.10867317e-02 2.51704752e-01 6.15177393e-01 -1.13299263e+00
5.43607116e-01 -2.16464087e-01 1.22583359e-01 -6.10285282e-01
2.69813418e-01 -1.06468451e+00 3.86810988e-01 2.86212899e-02
-2.67451346e-01 -2.75099743e-02 1.77740201e-01 8.95111561e-01
-5.07483304e-01 1.33569092e-01 7.26602793e-01 -3.84267941e-02
-1.35608232e+00 9.35865462e-01 6.36597127e-02 2.65167952e-02
1.30313754e+00 -4.06041622e-01 -1.61557436e-01 -1.04710804e-02
-8.29263330e-01 3.99421662e-01 5.94888449e-01 6.01585150e-01
5.29918492e-01 -1.37820482e+00 -3.67613256e-01 2.48265833e-01
2.30260417e-01 4.61455397e-02 6.53873563e-01 1.26001740e+00
-2.08463997e-01 1.17805205e-01 -8.73872358e-03 -1.02708149e+00
-1.27278519e+00 6.02664053e-01 3.40088844e-01 -9.07753333e-02
-7.22973168e-01 7.66694844e-01 4.35940236e-01 1.58990592e-01
1.88588679e-01 -5.17526627e-01 -3.09282541e-02 1.70180187e-01
6.43369377e-01 1.40415505e-01 6.04794882e-02 -9.29289460e-01
-2.51856923e-01 8.28293204e-01 -2.15686426e-01 -1.01682976e-01
1.28307533e+00 -4.66017991e-01 -1.66921541e-01 2.04103738e-01
1.45463836e+00 -3.82358760e-01 -1.89410377e+00 -4.84328777e-01
-1.17648304e-01 -8.34073365e-01 7.24766776e-02 -4.29091901e-01
-1.68853152e+00 5.62817395e-01 4.73141760e-01 -6.40852898e-02
1.50342298e+00 -5.56749515e-02 1.03441262e+00 -9.85611826e-02
5.26165068e-01 -1.07558990e+00 6.37952462e-02 4.26224500e-01
4.41622883e-01 -1.12311232e+00 -1.48342773e-01 -5.36912620e-01
-3.85833085e-01 9.75261986e-01 5.59235752e-01 -8.88108164e-02
6.17579818e-01 1.93022490e-01 -2.52378285e-01 -4.47064899e-02
-9.02179778e-01 -4.24097866e-01 4.14621353e-01 3.59814465e-01
2.25631014e-01 -2.78069228e-01 -2.66667753e-01 3.40825588e-01
4.37178820e-01 3.04811627e-01 2.08630130e-01 1.04102075e+00
-1.32100642e-01 -1.18271625e+00 -1.02871448e-01 2.65247524e-01
-4.64338064e-01 -1.27743006e-01 2.00961158e-01 6.27958238e-01
4.66102958e-01 8.95127118e-01 1.69716835e-01 -3.40195239e-01
1.75050810e-01 2.19814349e-02 5.27701855e-01 -3.54273766e-01
-6.46846473e-01 1.51969627e-01 -1.92269191e-01 -1.17883754e+00
-1.13880610e+00 -8.94035697e-01 -1.35949779e+00 -1.74861759e-01
-2.43602321e-01 -2.18587518e-01 4.83169258e-01 1.03692901e+00
4.89060491e-01 6.16481185e-01 4.68915135e-01 -1.04610682e+00
2.82684565e-01 -3.39579225e-01 -3.65083545e-01 6.68792665e-01
4.55400497e-01 -7.49238312e-01 -1.88950166e-01 5.78790784e-01]
|
[9.256998062133789, 0.04123135283589363]
|
d8e0a868-7151-49c0-b2db-769f0a3aceb2
|
extractive-summarization-via-chatgpt-for
|
2304.04193
| null |
https://arxiv.org/abs/2304.04193v1
|
https://arxiv.org/pdf/2304.04193v1.pdf
|
Extractive Summarization via ChatGPT for Faithful Summary Generation
|
Extractive summarization is a crucial task in natural language processing that aims to condense long documents into shorter versions by directly extracting sentences. The recent introduction of ChatGPT has attracted significant interest in the NLP community due to its remarkable performance on a wide range of downstream tasks. However, concerns regarding factuality and faithfulness have hindered its practical applications for summarization systems. This paper first presents a thorough evaluation of ChatGPT's performance on extractive summarization and compares it with traditional fine-tuning methods on various benchmark datasets. Our experimental analysis reveals that ChatGPT's extractive summarization performance is still inferior to existing supervised systems in terms of ROUGE scores. In addition, we explore the effectiveness of in-context learning and chain-of-thought reasoning for enhancing its performance. Furthermore, we find that applying an extract-then-generate pipeline with ChatGPT yields significant performance improvements over abstractive baselines in terms of summary faithfulness. These observations highlight potential directions for enhancing ChatGPT's capabilities for faithful text summarization tasks using two-stage approaches.
|
['Jiawei Zhang', 'Xiao Liu', 'Haopeng Zhang']
|
2023-04-09
| null | null | null | null |
['extractive-summarization']
|
['natural-language-processing']
|
[ 4.18586254e-01 5.89334786e-01 -2.83342153e-01 -2.22001508e-01
-1.52698123e+00 -7.11409688e-01 8.55375171e-01 6.93726480e-01
-3.65730733e-01 9.43965971e-01 1.15915453e+00 -4.12711680e-01
2.17153743e-01 -4.39554632e-01 -4.48166370e-01 -1.94088444e-01
2.48506218e-01 5.30138254e-01 2.09437653e-01 -5.70789874e-01
9.85409856e-01 -5.44869248e-03 -8.76391053e-01 7.35351026e-01
1.31415331e+00 1.13917373e-01 -1.77911445e-02 8.75755668e-01
-3.29459846e-01 1.00328803e+00 -1.02597213e+00 -6.58788860e-01
-1.16945080e-01 -7.20899105e-01 -1.46092927e+00 -1.17144808e-01
7.00164855e-01 -3.76985222e-01 -1.10263228e-01 7.95726776e-01
5.19736707e-01 2.02240095e-01 6.86920762e-01 -7.22964466e-01
-5.66306233e-01 1.17823112e+00 -4.77281481e-01 4.99855340e-01
6.07801497e-01 2.11789519e-01 1.61348593e+00 -5.10914564e-01
7.11799800e-01 1.24563611e+00 6.97320938e-01 5.24861813e-01
-1.08373165e+00 -9.44251418e-02 6.62539899e-02 -1.16112996e-02
-6.76330686e-01 -7.93956816e-01 4.64654565e-01 -1.45889772e-02
1.80524778e+00 4.98016834e-01 5.06881833e-01 8.42546284e-01
5.57059050e-01 1.25006354e+00 5.62485337e-01 -5.38371861e-01
1.19790114e-01 -2.48390004e-01 3.53257537e-01 4.82203811e-01
4.37009931e-01 -7.64007449e-01 -7.62322724e-01 -2.47255579e-01
-1.13534201e-02 -5.52043378e-01 -2.03312531e-01 5.83162844e-01
-1.22119057e+00 8.26702297e-01 7.81438053e-02 3.23967487e-01
-4.07180786e-01 1.31440356e-01 9.55864131e-01 2.14564368e-01
8.16562474e-01 1.20665550e+00 -3.44845653e-01 -4.68504250e-01
-1.37032056e+00 5.03739476e-01 1.16613686e+00 1.14039159e+00
3.21505070e-01 6.60732687e-02 -5.79916775e-01 7.33505905e-01
-2.95029134e-01 2.49036238e-01 6.68617845e-01 -1.21349633e+00
1.03301597e+00 4.97671336e-01 -1.16651610e-01 -7.38070130e-01
-1.32710755e-01 -3.06192428e-01 -5.57648301e-01 -6.70225501e-01
3.98394139e-03 -2.32035071e-01 -4.71582115e-01 1.19924533e+00
-1.34889200e-01 -3.41769367e-01 4.73467112e-01 4.26173329e-01
1.13050127e+00 1.04623497e+00 -5.74933551e-03 -5.53349376e-01
1.19580042e+00 -1.26102626e+00 -7.38226414e-01 -5.32461882e-01
8.35285664e-01 -9.07130718e-01 1.04109907e+00 3.42179924e-01
-1.52414799e+00 -6.82154521e-02 -9.46718514e-01 -6.64341509e-01
9.39851403e-02 5.77072576e-02 4.59693432e-01 3.01850885e-01
-1.13013101e+00 8.02510202e-01 -8.53454471e-01 -7.10976899e-01
4.65347081e-01 -1.03128240e-01 -1.10625476e-01 7.46154115e-02
-9.02963281e-01 1.02072310e+00 5.79193532e-01 -3.44386160e-01
-2.89197892e-01 -6.02897048e-01 -8.27516973e-01 3.04730922e-01
4.53257203e-01 -1.00497568e+00 2.00598860e+00 -2.14229733e-01
-1.48130560e+00 5.13201535e-01 -4.24416989e-01 -8.41441929e-01
3.35170209e-01 -5.35583317e-01 1.09871842e-01 4.98946995e-01
4.63149071e-01 5.40011585e-01 4.92846012e-01 -6.90094411e-01
-7.84850717e-01 -5.98374866e-02 -2.46054027e-02 5.34057081e-01
-3.28628570e-01 1.33638412e-01 -3.03555608e-01 -6.45876944e-01
-4.83298674e-02 -7.34890103e-01 -1.09593719e-01 -9.33273613e-01
-9.46999013e-01 -7.04746127e-01 5.67113578e-01 -9.36922371e-01
1.49745798e+00 -1.57660353e+00 1.15342915e-01 -3.92889798e-01
2.24889979e-01 4.12320673e-01 -1.99955106e-01 1.18282998e+00
3.67274463e-01 5.91720939e-01 -2.90755540e-01 -4.17943150e-01
3.54458988e-02 1.53681096e-02 -7.09555149e-01 -5.91740794e-02
4.38442826e-01 1.25524879e+00 -1.10924220e+00 -8.36738706e-01
-1.39677897e-01 -1.62763149e-01 -6.80655241e-01 2.10474551e-01
-5.63400567e-01 1.46942928e-01 -6.25767589e-01 3.24204654e-01
4.14998084e-02 -8.79475474e-02 2.09180310e-01 3.28915156e-02
-3.18561822e-01 1.27227104e+00 -2.04547077e-01 1.82435131e+00
-3.09597552e-01 8.20061564e-01 -3.57729137e-01 -8.04048002e-01
4.52264935e-01 2.95033067e-01 1.02947362e-01 -4.27561939e-01
5.15571050e-02 1.65040448e-01 -7.63728395e-02 -5.88535607e-01
1.39621544e+00 -2.43649498e-01 -5.09296954e-01 9.32325304e-01
1.38694182e-01 -8.71505380e-01 7.82334805e-01 1.07468820e+00
1.35847306e+00 2.72971652e-02 6.77947044e-01 -2.94420183e-01
2.19144627e-01 5.75906932e-01 3.83051574e-01 9.52856719e-01
7.59251490e-02 6.22705579e-01 8.94863129e-01 8.46515223e-02
-1.08070230e+00 -6.13352358e-01 2.47741431e-01 1.11652136e+00
-2.82463372e-01 -9.72719610e-01 -9.27900255e-01 -7.94685066e-01
-2.63768226e-01 1.46293318e+00 -2.46128678e-01 -2.31190428e-01
-9.59157467e-01 -6.35030150e-01 9.56206739e-01 5.47817886e-01
5.09941041e-01 -1.19283557e+00 -5.38067281e-01 2.81985015e-01
-8.12362015e-01 -1.12674296e+00 -6.55552566e-01 -1.12684369e-01
-1.19528270e+00 -8.03870618e-01 -4.55419600e-01 -6.72240198e-01
2.36713216e-01 4.96155888e-01 1.26160133e+00 3.17032672e-02
1.62163973e-01 3.31778616e-01 -6.53823674e-01 -6.11510694e-01
-1.01607823e+00 9.36903656e-01 -3.06919277e-01 -8.74847531e-01
2.78300047e-01 -4.62558895e-01 -3.67433548e-01 -4.59494561e-01
-9.90216076e-01 1.96332768e-01 8.05179060e-01 5.80773056e-01
9.12597105e-02 -1.28277436e-01 1.00193846e+00 -1.15716553e+00
1.39290392e+00 -2.68266588e-01 1.41476870e-01 3.68655294e-01
-5.00380874e-01 3.09478313e-01 5.63014209e-01 8.25483873e-02
-1.42811382e+00 -6.85092330e-01 -2.74597675e-01 5.24372995e-01
2.16587529e-01 7.66649485e-01 2.04865053e-01 6.92976594e-01
8.33500206e-01 3.47303867e-01 -7.09450170e-02 -3.73169363e-01
5.53834438e-01 7.00167656e-01 6.86272740e-01 -6.35053277e-01
4.99533892e-01 9.36828256e-02 -4.71550852e-01 -1.06383789e+00
-1.31143689e+00 -5.92159748e-01 -3.96221906e-01 1.47001430e-01
5.69539666e-01 -8.42634797e-01 -3.16460282e-01 1.17117502e-01
-1.35037410e+00 -2.85404384e-01 -4.86527443e-01 1.10337846e-01
-4.91581410e-01 8.13689530e-01 -1.13898325e+00 -4.16022390e-01
-1.32233751e+00 -6.06614351e-01 1.24762154e+00 3.17350298e-01
-9.49297965e-01 -9.68859553e-01 2.00894177e-01 7.12387860e-01
2.07687765e-01 -6.05036654e-02 1.03401172e+00 -1.05675220e+00
-2.91070461e-01 -3.06536138e-01 -6.48732334e-02 4.19606030e-01
1.26488060e-01 1.46857023e-01 -5.34826994e-01 -1.56016424e-01
-1.33643169e-02 -5.21178186e-01 1.24906027e+00 2.44147778e-01
4.79827434e-01 -9.23045635e-01 -1.21595278e-01 -2.39626039e-02
1.05872810e+00 -2.68022746e-01 5.15757322e-01 3.87975723e-01
5.01861453e-01 6.02494121e-01 6.22995138e-01 4.30425256e-01
4.99078453e-01 5.14386855e-02 -9.16240141e-02 3.61477643e-01
-3.41023356e-01 -5.06555438e-01 4.65001404e-01 1.37235260e+00
7.88346976e-02 -5.38348258e-01 -7.60233760e-01 7.18872070e-01
-1.99821854e+00 -1.28052795e+00 -2.96533018e-01 1.60157073e+00
1.21035695e+00 3.70609015e-01 5.50299659e-02 -9.69774649e-02
5.00724971e-01 4.89848256e-01 -4.10816222e-01 -1.06577885e+00
-1.04680508e-01 2.07476526e-01 1.23320043e-01 4.67480034e-01
-6.96804047e-01 1.30620468e+00 6.37896872e+00 7.45117366e-01
-7.88657010e-01 -1.12358093e-01 4.19210970e-01 -2.60030478e-01
-4.15719718e-01 1.84771106e-01 -8.84450555e-01 1.99045330e-01
9.78016257e-01 -7.37626851e-01 2.55776178e-02 5.36747634e-01
4.08916026e-01 -4.71907616e-01 -1.11283922e+00 2.29389489e-01
4.32068259e-01 -1.67919242e+00 4.37324613e-01 -2.18352720e-01
9.36287344e-01 1.82997286e-01 -3.20269763e-01 5.88827908e-01
4.02381271e-01 -6.91325963e-01 6.36661828e-01 2.25121975e-01
4.34246480e-01 -7.89916158e-01 7.96730757e-01 7.08918869e-01
-6.80838704e-01 2.59277165e-01 -4.13365513e-01 -2.97406793e-01
4.80095863e-01 6.12124920e-01 -1.42811882e+00 7.82631218e-01
1.84643641e-01 8.01612198e-01 -5.74791491e-01 7.02773452e-01
-6.34840727e-01 1.01469505e+00 -6.70603067e-02 -3.53199303e-01
4.48146522e-01 -6.62197918e-02 8.41049373e-01 1.81339598e+00
1.40380040e-01 3.20454627e-01 1.93991467e-01 5.28248787e-01
-4.14090067e-01 2.08078429e-01 -4.25135523e-01 -4.08572018e-01
4.43784088e-01 1.16002488e+00 -6.45895600e-01 -7.70885289e-01
-6.80134445e-02 9.13440764e-01 4.15894985e-01 1.59933344e-01
-3.74881625e-01 -7.16245115e-01 3.15378122e-02 -4.99124117e-02
2.68779099e-01 -2.98689097e-01 -4.79634911e-01 -1.36958635e+00
1.57040372e-01 -1.02825558e+00 4.65864658e-01 -7.25960910e-01
-1.05222225e+00 5.84590495e-01 1.95722535e-01 -7.34761417e-01
-5.88673353e-01 1.56066775e-01 -1.10040152e+00 6.06730044e-01
-1.28210068e+00 -1.11400282e+00 2.55148441e-01 -2.84409896e-02
1.29912925e+00 8.59051347e-02 5.46765029e-01 -4.61124033e-01
-6.82355523e-01 3.35190564e-01 -2.14637406e-02 -2.73821913e-02
9.12918270e-01 -1.40254331e+00 9.88420904e-01 1.18506014e+00
3.56761850e-02 8.43823671e-01 1.17335653e+00 -9.18526947e-01
-1.33781600e+00 -1.10869563e+00 1.52062154e+00 -5.99957407e-01
7.71077335e-01 5.70293814e-02 -9.24861729e-01 8.73025954e-01
9.79519665e-01 -1.01954949e+00 6.17325723e-01 2.69049525e-01
-1.98730096e-01 1.89717978e-01 -7.57994652e-01 7.98675835e-01
8.25440109e-01 -4.16981250e-01 -1.56570101e+00 5.34547806e-01
9.45034266e-01 -2.93667108e-01 -6.99941099e-01 1.26883104e-01
2.87259370e-01 -8.41670454e-01 4.55020368e-01 -7.05696523e-01
1.12056696e+00 1.32444173e-01 2.00608805e-01 -1.63855505e+00
-3.01649392e-01 -9.86482859e-01 -1.14478670e-01 1.60457993e+00
4.22721326e-01 -4.65359598e-01 6.15315020e-01 4.80061620e-01
-6.45943403e-01 -7.09541976e-01 -4.42166716e-01 -6.13869309e-01
4.88862723e-01 -1.19064860e-01 2.20812917e-01 5.90880573e-01
8.46589565e-01 1.27374923e+00 -1.34081885e-01 -4.25172538e-01
3.59091491e-01 3.24457705e-01 8.41728985e-01 -8.62897098e-01
-8.94560367e-02 -6.50925457e-01 2.21945316e-01 -1.18597054e+00
3.73216361e-01 -1.10070741e+00 3.30979556e-01 -2.30828166e+00
5.88026702e-01 2.90249825e-01 4.27350283e-01 4.81524438e-01
-4.76162314e-01 -1.45533094e-02 1.68090656e-01 4.59517002e-01
-1.04808152e+00 5.83641410e-01 1.16825664e+00 -2.23236516e-01
-4.46047425e-01 3.09245232e-02 -1.40387905e+00 6.17147565e-01
1.03419816e+00 -3.72329772e-01 -3.64912599e-01 -5.05064845e-01
2.58820951e-01 2.14704454e-01 -2.93425053e-01 -7.29808867e-01
4.19045955e-01 -9.98535007e-02 -1.72074199e-01 -8.39016616e-01
-7.92406052e-02 3.29873979e-01 -5.62513113e-01 3.56629163e-01
-8.20903301e-01 1.70486346e-01 2.77178854e-01 3.92962873e-01
-2.29205593e-01 -4.66871858e-01 4.82072711e-01 -3.81864876e-01
-3.39712501e-01 -2.80458391e-01 -8.61609519e-01 7.13707089e-01
3.95146042e-01 -1.24141313e-01 -7.53997684e-01 -5.09499490e-01
-4.14229035e-02 2.41399869e-01 4.47332233e-01 1.61194354e-01
4.36826438e-01 -6.15809143e-01 -1.22170734e+00 -4.71823633e-01
4.23217490e-02 1.66507497e-01 3.48957330e-02 8.34352911e-01
-6.28816664e-01 8.25627327e-01 8.80251005e-02 -3.06429863e-01
-1.33393800e+00 2.31769979e-01 -3.41759980e-01 -7.67557740e-01
-9.17223036e-01 6.35291994e-01 -1.71687484e-01 -1.58408031e-01
-1.34234905e-01 -4.98504877e-01 -3.80188860e-02 1.70017168e-01
6.26516700e-01 6.76427007e-01 2.74479747e-01 -2.21862316e-01
-7.44308829e-02 -8.43083709e-02 -6.72970474e-01 -3.42173517e-01
1.43929350e+00 -2.21166089e-01 -4.65160459e-01 2.78145611e-01
1.04586852e+00 2.40718991e-01 -7.70353734e-01 -2.18020141e-01
4.75208104e-01 3.56678627e-02 -1.92228109e-01 -8.56497705e-01
-2.50523478e-01 6.81061625e-01 -8.82760882e-01 3.92766416e-01
8.70524406e-01 1.92543596e-01 1.32238090e+00 1.04550743e+00
7.68266022e-02 -1.27673697e+00 3.04086596e-01 1.00770938e+00
1.07991862e+00 -1.00998437e+00 5.48821449e-01 -3.04015487e-01
-9.53786850e-01 1.11524665e+00 3.85110140e-01 -1.37472376e-01
-2.05742806e-01 -4.44781110e-02 -1.69450089e-01 -1.53459698e-01
-1.15681410e+00 4.54147272e-02 2.22767949e-01 1.01847306e-01
6.75248027e-01 -1.65221006e-01 -7.26327538e-01 3.42516094e-01
-8.16861093e-01 -3.96289527e-01 1.05415535e+00 1.11887980e+00
-9.39777076e-01 -8.82196784e-01 -6.94652721e-02 7.16068029e-01
-6.88833177e-01 -4.11340803e-01 -7.85520673e-01 6.42502308e-01
-7.38134980e-01 1.34518445e+00 -9.20130238e-02 7.98508450e-02
2.88135409e-01 1.40216097e-01 5.07781029e-01 -1.23255837e+00
-1.06209683e+00 2.58100361e-01 7.51686692e-01 -1.76485330e-01
-2.26386324e-01 -9.02974069e-01 -1.41872418e+00 -5.61591923e-01
-2.03998223e-01 6.61115587e-01 3.89069468e-01 1.16750646e+00
4.92487043e-01 5.51984012e-01 2.05967247e-01 -7.43929684e-01
-9.13815081e-01 -1.33600390e+00 -6.13247231e-02 1.17043398e-01
2.49117523e-01 3.28773886e-01 -1.79152444e-01 2.44818985e-01]
|
[12.46371841430664, 9.466949462890625]
|
13cc305a-fa84-463e-8d4d-2ec87c28aee4
|
machine-learning-can-guide-experimental
|
2211.00625
| null |
https://arxiv.org/abs/2211.00625v1
|
https://arxiv.org/pdf/2211.00625v1.pdf
|
Machine learning can guide experimental approaches for protein digestibility estimations
|
Food protein digestibility and bioavailability are critical aspects in addressing human nutritional demands, particularly when seeking sustainable alternatives to animal-based proteins. In this study, we propose a machine learning approach to predict the true ileal digestibility coefficient of food items. The model makes use of a unique curated dataset that combines nutritional information from different foods with FASTA sequences of some of their protein families. We extracted the biochemical properties of the proteins and combined these properties with embeddings from a Transformer-based protein Language Model (pLM). In addition, we used SHAP to identify features that contribute most to the model prediction and provide interpretability. This first AI-based model for predicting food protein digestibility has an accuracy of 90% compared to existing experimental techniques. With this accuracy, our model can eliminate the need for lengthy in-vivo or in-vitro experiments, making the process of creating new foods faster, cheaper, and more ethical.
|
['Ranveer Chandra', 'Swati Sharma', 'Maria Angels de Luis Balaguer', 'Anvita Bhagavathula', 'Sara Malvar']
|
2022-11-01
| null | null | null | null |
['protein-language-model']
|
['medical']
|
[ 4.36640196e-02 -4.17175563e-03 -6.51427746e-01 -3.12017888e-01
1.74896091e-01 -8.12357008e-01 -7.78503641e-02 1.19092035e+00
-1.34000599e-01 2.94981360e-01 5.11937976e-01 -4.35001612e-01
-1.00841038e-01 -9.70418096e-01 -9.58671093e-01 -5.36334276e-01
-3.67490321e-01 1.68815553e-01 -1.43801540e-01 -3.76848191e-01
-2.25317612e-01 9.75203812e-02 -1.53676367e+00 5.65471649e-01
9.53814745e-01 9.31495070e-01 5.66055894e-01 3.67235094e-01
-1.62479609e-01 7.63980269e-01 1.80113852e-01 -1.88406795e-01
3.41213793e-01 -5.40321529e-01 -4.17366445e-01 -4.01403844e-01
-3.68777573e-01 -3.15313250e-01 1.76336020e-01 8.44190717e-01
1.37861893e-01 7.86976069e-02 8.66329134e-01 -7.88306892e-01
-1.29997826e+00 7.83586204e-01 -2.58711129e-01 -1.05001062e-01
7.00777769e-01 2.04200968e-01 1.08257973e+00 -5.69489777e-01
2.07416072e-01 9.75592256e-01 8.64271045e-01 4.28653717e-01
-1.19157135e+00 -1.58676386e-01 2.37661472e-04 2.29407862e-01
-8.87365878e-01 1.50074707e-02 5.39480269e-01 -6.90353155e-01
1.19587278e+00 5.25995232e-02 1.27242327e+00 5.95988929e-01
5.59447765e-01 7.84148693e-01 8.21442783e-01 -4.88719493e-01
3.27202588e-01 1.50742009e-01 3.12283069e-01 7.26284087e-01
7.25066662e-01 4.58584040e-01 -2.20145613e-01 -9.83193591e-02
3.29962760e-01 2.62643039e-01 -2.48453766e-02 -7.22866476e-01
-9.59627986e-01 1.14930010e+00 5.51795959e-01 1.12328239e-01
-6.77895784e-01 -3.14037323e-01 3.75737369e-01 3.23910892e-01
4.56393778e-01 5.12837946e-01 -9.69276667e-01 4.61784601e-01
-5.11186361e-01 2.07200840e-01 1.04194582e+00 5.19673467e-01
4.70741272e-01 -3.13939005e-01 5.13169587e-01 8.29416335e-01
6.50960505e-01 6.49639845e-01 6.66621506e-01 -5.31239331e-01
-1.55473784e-01 7.41373777e-01 2.80508131e-01 -1.12307334e+00
-6.62105501e-01 1.59478232e-01 -3.18661600e-01 -9.05328523e-03
3.97601187e-01 1.14429496e-01 -5.91093004e-01 1.60608864e+00
5.44553339e-01 -6.56651735e-01 3.65979344e-01 1.03470469e+00
7.01562285e-01 8.54225576e-01 7.23949373e-01 2.85522882e-02
1.67372763e+00 -1.01698089e+00 -5.27410209e-01 -2.13495120e-02
7.92836845e-01 -5.71085095e-01 7.29779482e-01 2.34525964e-01
-7.25603878e-01 -4.49368328e-01 -1.48674989e+00 -2.73935080e-01
-6.84938371e-01 8.59165471e-03 9.52856004e-01 6.28712416e-01
-6.23600066e-01 8.20680499e-01 -9.98001575e-01 -5.02895534e-01
-4.90220673e-02 5.25227249e-01 -4.13176239e-01 -2.17892349e-01
-1.21551454e+00 1.36292994e+00 5.79501688e-01 -1.61597831e-03
-3.43097210e-01 -1.11839640e+00 -1.33330345e+00 1.33381471e-01
5.45445131e-03 -8.32250416e-01 9.98900056e-01 -8.91405821e-01
-1.69138062e+00 5.41998684e-01 8.79753605e-02 -6.78058088e-01
-2.36501485e-01 -5.26867926e-01 -4.88117516e-01 1.46960109e-01
-4.09151278e-02 7.40854144e-01 1.12719178e-01 -5.76987445e-01
-2.07465842e-01 -2.81994909e-01 7.87634775e-02 8.85253549e-02
-1.05346933e-01 1.70655141e-03 6.23648524e-01 -3.79572570e-01
2.75456067e-02 -7.55792320e-01 -3.65836352e-01 5.90071559e-01
2.56659985e-01 -1.14167295e-01 -9.74463392e-03 -1.08494353e+00
6.89636290e-01 -2.05501127e+00 1.84968427e-01 4.66556735e-02
1.74576551e-01 4.42320436e-01 -2.96840340e-01 6.39022231e-01
-2.07808062e-01 1.17574714e-01 2.29866081e-03 8.03671598e-01
1.71476886e-01 -2.40746383e-02 2.14821354e-01 3.87841821e-01
3.79634172e-01 9.29277122e-01 -1.15564716e+00 -1.69626132e-01
6.84592903e-01 6.30967617e-01 -8.47354650e-01 1.76409051e-01
-7.27442622e-01 -1.55578405e-01 -5.59487045e-01 5.29344380e-01
6.91007435e-01 2.32056249e-02 8.38535070e-01 -7.38945603e-01
-2.56484419e-01 4.67009366e-01 -6.83910608e-01 1.56118691e+00
-2.92244315e-01 -1.50010988e-01 -2.15791702e-01 -9.68181193e-01
8.53249252e-01 4.09308895e-02 8.48741412e-01 -8.38123500e-01
3.25943649e-01 1.84610188e-01 2.15268478e-01 -7.95290351e-01
8.59994069e-02 -2.39337757e-01 2.58475274e-01 3.46397281e-01
1.18095756e-01 4.60398406e-01 4.60402966e-01 -3.28080237e-01
5.64587355e-01 7.33142912e-01 9.16792691e-01 -9.84625638e-01
5.05731702e-01 2.14395970e-01 3.53061557e-01 -2.86996096e-01
-1.79629266e-01 2.16922932e-03 -4.08780463e-02 -8.54529202e-01
-1.23768616e+00 -1.02155542e+00 -3.25495750e-02 9.78043437e-01
4.73780371e-02 -4.44027513e-01 -5.65892220e-01 -1.20616101e-01
5.38326740e-01 5.52572727e-01 -6.82875097e-01 -4.11286533e-01
-4.48132455e-01 -8.75431001e-01 1.35919809e-01 6.46551549e-01
3.96638960e-02 -7.21600771e-01 -6.57215834e-01 5.77488542e-01
-2.95695484e-01 -6.70684338e-01 -6.19730830e-01 6.16482496e-01
-8.38918328e-01 -1.38294137e+00 -4.93038654e-01 -1.08045697e+00
3.91963303e-01 3.49087745e-01 9.14159715e-01 -9.01798680e-02
-3.37012202e-01 -3.04358125e-01 -5.99376023e-01 -6.33966386e-01
-7.22044945e-01 -2.42243662e-01 3.43924105e-01 -3.61558616e-01
1.17400146e+00 -2.83948928e-01 -8.77610624e-01 5.94711900e-02
-8.78779650e-01 1.41603693e-01 5.04637718e-01 7.31553733e-01
7.58397341e-01 -7.91359246e-02 6.11312270e-01 -6.20199621e-01
3.22156161e-01 -8.52860391e-01 -2.70913512e-01 2.78458953e-01
-6.89054787e-01 6.33522630e-01 8.06172013e-01 -7.02740133e-01
-4.68020648e-01 3.20446521e-01 -2.87811041e-01 3.94601732e-01
1.00390717e-01 6.66255713e-01 -1.29705325e-01 2.01040283e-01
7.69327998e-01 -1.76095068e-01 5.09828269e-01 -6.46305263e-01
5.79626441e-01 4.04704511e-01 1.22647390e-01 -6.60414994e-01
2.67725855e-01 -1.49437666e-01 -7.81181827e-02 -8.41367662e-01
-7.26972401e-01 -3.57433081e-01 -5.06049454e-01 2.88736522e-01
8.81862819e-01 -1.08740222e+00 -8.41259420e-01 1.26351714e-01
-3.50887626e-01 -2.85034865e-01 -6.37229234e-02 8.72373402e-01
-5.57033360e-01 4.80781972e-01 -1.00167227e+00 -5.07595956e-01
-6.76660776e-01 -8.19251776e-01 4.55366254e-01 4.29321751e-02
-5.33457279e-01 -1.02827036e+00 2.86145419e-01 1.95272848e-01
3.60028356e-01 5.38432956e-01 1.31230819e+00 -4.87802804e-01
1.14176042e-01 2.85961568e-01 -1.48137376e-01 2.90105551e-01
6.10886574e-01 -6.61735013e-02 -5.06675959e-01 -3.05799246e-02
1.76788718e-01 -1.92995757e-01 8.30014229e-01 6.52012050e-01
3.44229609e-01 -5.05515754e-01 -1.83258995e-01 4.65413034e-01
1.60851824e+00 2.48878270e-01 1.97746515e-01 4.68418717e-01
3.91269922e-01 9.14799690e-01 6.19304717e-01 3.95279676e-02
5.79680979e-01 4.12919074e-01 1.83501542e-02 -1.24215677e-01
-1.81187898e-01 -7.32858539e-01 7.81696498e-01 8.45053792e-01
2.82728732e-01 -2.22105771e-01 -4.05425429e-01 5.90930641e-01
-1.37840915e+00 -8.59348893e-01 -1.86732337e-01 1.97174609e+00
1.21399319e+00 -4.67021853e-01 3.95084232e-01 4.96493503e-02
3.85760128e-01 -4.43132728e-01 -6.20337546e-01 -7.74238348e-01
-1.72665611e-01 3.13433826e-01 6.48304164e-01 2.85542011e-01
-1.04369318e+00 4.53193665e-01 7.27225971e+00 -1.35735586e-01
-9.12650228e-01 -2.26925969e-01 2.34347612e-01 3.33614759e-02
-3.51519495e-01 -3.75359893e-01 -5.31664014e-01 3.78175288e-01
1.47047102e+00 -9.44066271e-02 5.60805976e-01 7.52393842e-01
1.84726343e-01 1.55839715e-02 -1.30383289e+00 2.84110576e-01
1.27594788e-02 -9.30426061e-01 -2.11867958e-01 -1.64890215e-01
2.65687078e-01 7.93031454e-02 -3.03832203e-01 6.18118942e-02
5.65141320e-01 -7.94341445e-01 7.20894814e-01 1.89171523e-01
4.00436401e-01 -5.93310475e-01 6.66343391e-01 1.62660211e-01
-1.23398173e+00 -2.27768019e-01 -8.73677194e-01 -3.13718349e-01
1.69936605e-02 6.84151530e-01 -6.51619911e-01 5.25511682e-01
7.18680680e-01 8.50626051e-01 -2.44146138e-01 7.79516637e-01
7.80203864e-02 3.13527465e-01 -7.88458824e-01 -3.17800671e-01
-6.67909682e-02 -3.44537169e-01 -2.19868109e-01 1.05328310e+00
3.07321697e-01 1.11320697e-01 1.93171158e-01 8.16136360e-01
3.03622097e-01 7.63530910e-01 -4.84713703e-01 -5.69078743e-01
-3.67856584e-02 8.46862972e-01 -5.58210075e-01 -6.32195398e-02
-6.09360218e-01 7.82598317e-01 2.14771509e-01 -1.52858391e-01
-6.73580885e-01 1.00619420e-01 9.98074174e-01 6.19586445e-02
4.88915533e-01 -1.07967429e-01 2.50952244e-01 -1.13048816e+00
-1.37192860e-01 -8.28100026e-01 3.31281126e-01 -5.65691710e-01
-1.49969351e+00 1.05352521e-01 5.85842021e-02 -9.07196045e-01
-1.92697093e-01 -1.03340089e+00 2.36719087e-01 9.24483359e-01
-1.56167781e+00 -1.18134928e+00 3.43055397e-01 -1.83567494e-01
2.74462551e-01 2.37086728e-01 1.27309644e+00 3.30698043e-01
-2.77540773e-01 3.82062584e-01 4.57727641e-01 -1.02813639e-01
4.54313189e-01 -8.50097179e-01 3.33134145e-01 4.25419658e-01
-1.78558066e-01 8.85506630e-01 9.05135512e-01 -7.24254668e-01
-1.63215053e+00 -1.07096267e+00 1.09282255e+00 -4.14469302e-01
5.11786222e-01 -2.18033075e-01 -8.16748261e-01 4.84054923e-01
-1.92884967e-01 -6.03033364e-01 1.37013662e+00 1.21346246e-02
-5.85417688e-01 -3.16402540e-02 -1.44773293e+00 3.09011161e-01
7.03269243e-01 -2.32244253e-01 -6.20542109e-01 2.41168529e-01
1.04280984e+00 -8.66434425e-02 -1.42307985e+00 2.64890641e-01
1.11462772e+00 -3.19823891e-01 1.34388697e+00 -7.63986528e-01
5.98388016e-01 -4.80497807e-01 -3.87445331e-01 -1.27529848e+00
-8.23344290e-01 2.91779786e-01 -2.80925721e-01 6.65504098e-01
5.10831237e-01 -4.88546401e-01 4.77964252e-01 6.70831800e-01
-2.87223868e-02 -7.11476088e-01 -8.64057615e-02 -6.61929667e-01
3.27108979e-01 2.51864493e-01 1.19305825e+00 7.76652098e-01
5.42385519e-01 8.09180513e-02 -4.20976192e-01 8.28236118e-02
4.44856852e-01 3.28864187e-01 1.09393798e-01 -1.26689267e+00
-3.53448659e-01 -3.21154982e-01 -3.31495613e-01 -7.58463979e-01
-8.50881264e-02 -1.27388847e+00 2.89722905e-02 -1.54461992e+00
3.36235255e-01 -1.04829535e-01 -6.05063081e-01 6.04419410e-01
-2.08427757e-01 -7.74386153e-02 1.34052709e-01 -1.01419866e-01
1.88066468e-01 3.06356788e-01 9.88931239e-01 -4.16292489e-01
-4.35719848e-01 -5.25045991e-01 -1.15587628e+00 2.74201810e-01
1.08924580e+00 -5.50509274e-01 -5.44993997e-01 -4.20884043e-01
3.14967960e-01 -1.08519003e-01 -5.90224080e-02 -6.29766941e-01
-4.19394463e-01 -3.48845959e-01 5.10134339e-01 -1.71488464e-01
-3.28261331e-02 -8.43208075e-01 6.54786050e-01 1.08634913e+00
-6.70294285e-01 -1.21337138e-01 3.63040805e-01 3.64523172e-01
2.89364338e-01 -2.51964390e-01 5.39291322e-01 -7.41970018e-02
-6.22128248e-01 2.24602353e-02 -5.32890260e-01 -5.86288393e-01
9.66760278e-01 -5.66716865e-02 -1.11608118e-01 4.47096407e-01
-6.30264580e-01 1.96986884e-01 6.96837485e-01 7.19312489e-01
2.92603463e-01 -1.29035509e+00 -6.80968523e-01 5.55693448e-01
2.16425508e-01 -6.88715696e-01 2.77254462e-01 3.96522760e-01
-1.07917166e+00 6.49466395e-01 -7.92368650e-01 -3.17939967e-01
-8.86537135e-01 1.46997309e+00 1.94859281e-01 -1.88510850e-01
-6.97973669e-01 1.92601785e-01 3.70314658e-01 -6.60946071e-01
-5.94444215e-01 -8.68416488e-01 -4.46171165e-01 -2.30831146e-01
6.71894372e-01 1.39169842e-01 -5.58773726e-02 -5.89049160e-01
-3.79226685e-01 4.35620725e-01 1.42877966e-01 7.93089867e-01
1.54807818e+00 -1.23847194e-01 -6.89931214e-02 3.73600006e-01
1.19136286e+00 -2.47521982e-01 -9.35583055e-01 -5.10357693e-02
1.49662187e-03 7.82857761e-02 -8.31132755e-02 -1.09606004e+00
-8.47689986e-01 5.39373398e-01 9.20943737e-01 -1.61441952e-01
1.12003279e+00 -2.70947307e-01 1.18141758e+00 2.09469289e-01
3.35863620e-01 -7.59625196e-01 -5.04812002e-01 -7.28580216e-03
3.49225670e-01 -1.12596476e+00 2.00578317e-01 -6.71229482e-01
-3.62551779e-01 1.14873385e+00 1.07141897e-01 -9.12749991e-02
8.48223090e-01 1.16849512e-01 1.64935157e-01 -5.03248759e-02
-6.64920211e-01 3.04884966e-02 2.59461343e-01 9.90662277e-01
8.09768677e-01 5.08394718e-01 -9.56121027e-01 9.64403927e-01
-2.87034720e-01 4.82904971e-01 1.15746476e-01 8.21764171e-01
-6.20808721e-01 -1.51760101e+00 8.17542449e-02 4.53153938e-01
-5.19256771e-01 -2.61924654e-01 8.74996558e-02 3.95118147e-01
1.28982604e-01 1.01434064e+00 -4.01847839e-01 -3.46689671e-01
1.56962588e-01 2.94860065e-01 6.28971815e-01 -3.08400810e-01
-6.37347877e-01 -1.00816652e-01 1.18874602e-01 -3.41407329e-01
-7.04859495e-01 -6.22441947e-01 -1.36393321e+00 -6.88345253e-01
-3.30209255e-01 2.60754943e-01 8.09076667e-01 7.57011354e-01
4.15058106e-01 2.60295331e-01 4.21727449e-01 -4.73518014e-01
-6.79173589e-01 -6.64350390e-01 -6.23293638e-01 7.39349067e-01
1.10174201e-01 -6.40435219e-01 2.15574522e-02 3.08491677e-01]
|
[11.529703140258789, 4.4971489906311035]
|
ca2928b3-e5b4-49a9-9237-d39465513179
|
learning-dynamic-graphs-from-all-contextual
|
2306.15927
| null |
https://arxiv.org/abs/2306.15927v1
|
https://arxiv.org/pdf/2306.15927v1.pdf
|
Learning Dynamic Graphs from All Contextual Information for Accurate Point-of-Interest Visit Forecasting
|
Forecasting the number of visits to Points-of-Interest (POI) in an urban area is critical for planning and decision-making for various application domains, from urban planning and transportation management to public health and social studies. Although this forecasting problem can be formulated as a multivariate time-series forecasting task, the current approaches cannot fully exploit the ever-changing multi-context correlations among POIs. Therefore, we propose Busyness Graph Neural Network (BysGNN), a temporal graph neural network designed to learn and uncover the underlying multi-context correlations between POIs for accurate visit forecasting. Unlike other approaches where only time-series data is used to learn a dynamic graph, BysGNN utilizes all contextual information and time-series data to learn an accurate dynamic graph representation. By incorporating all contextual, temporal, and spatial signals, we observe a significant improvement in our forecasting accuracy over state-of-the-art forecasting models in our experiments with real-world datasets across the United States.
|
['Cyrus Shahabi', 'Yao-Yi Chiang', 'Maria Despoina Siampou', 'Haoji Hu', 'Sina Shaham', 'Haowen Lin', 'Arash Hajisafi']
|
2023-06-28
| null | null | null | null |
['management', 'decision-making', 'multivariate-time-series-forecasting']
|
['miscellaneous', 'reasoning', 'time-series']
|
[ 4.98477332e-02 -1.11861803e-01 -5.90345263e-01 -5.57687819e-01
-2.02676922e-01 -1.42066255e-01 6.34991169e-01 5.49940825e-01
8.93738717e-02 6.34134889e-01 6.80605710e-01 -8.38197172e-01
-4.80849534e-01 -1.23808765e+00 -5.87398827e-01 -4.27490950e-01
-8.33275914e-01 4.06849146e-01 6.59895539e-02 -4.55505222e-01
-2.55424201e-01 4.98767704e-01 -1.19076574e+00 -9.55646262e-02
8.16636980e-01 9.61185932e-01 -1.65118709e-01 6.14580452e-01
-2.95679748e-01 7.25858629e-01 -9.17094648e-02 2.46893689e-02
-3.05538648e-03 -1.83897242e-01 -4.18534160e-01 5.94011089e-03
8.99403766e-02 2.68543977e-02 -9.45260108e-01 5.53943992e-01
1.40093252e-01 6.29116297e-01 7.13212132e-01 -1.46681046e+00
-5.75872183e-01 4.99270082e-01 -5.15962780e-01 6.51324213e-01
1.03367388e-01 -1.05592735e-01 1.01838458e+00 -2.09542125e-01
2.05629542e-01 1.19098771e+00 7.92272210e-01 -1.78925365e-01
-1.11466622e+00 -5.41965544e-01 7.91577101e-01 1.99158952e-01
-1.38860607e+00 6.10227808e-02 1.00660574e+00 -3.55337590e-01
1.09552908e+00 2.89435416e-01 7.60165274e-01 9.21560466e-01
5.70423305e-01 5.57139814e-01 4.89099592e-01 -1.03736743e-02
-5.37674762e-02 -5.85413337e-01 2.47950166e-01 6.32632375e-01
8.23702440e-02 2.02567369e-01 -2.61044830e-01 -1.90447584e-01
4.84495848e-01 9.77585852e-01 8.73938501e-02 3.12005132e-01
-1.00094271e+00 7.43643522e-01 9.14587080e-01 5.43838859e-01
-8.73311162e-01 4.83779877e-01 1.50048271e-01 1.70538694e-01
1.10268497e+00 -5.74299768e-02 -3.13691735e-01 -1.15216158e-01
-9.01059747e-01 -5.88212255e-03 6.94406748e-01 9.38174188e-01
8.04852664e-01 3.40487599e-01 -2.67203242e-01 5.51105857e-01
2.83402562e-01 7.84393787e-01 4.89130504e-02 -4.03698683e-01
7.44725347e-01 9.28324521e-01 -4.53861244e-03 -1.74570739e+00
-1.03403616e+00 -5.91211140e-01 -1.28688610e+00 -7.41719127e-01
2.62011677e-01 -4.26702082e-01 -8.93830717e-01 1.70879972e+00
2.93705285e-01 9.65486884e-01 -4.05820817e-01 4.50247198e-01
8.21656287e-01 9.26392615e-01 1.83297098e-01 -1.02053300e-01
9.33132946e-01 -9.19642627e-01 -7.63367891e-01 -2.45782360e-01
7.95762122e-01 -1.77511021e-01 6.74564064e-01 -3.63316566e-01
-6.47502005e-01 -3.94632488e-01 -5.35006344e-01 3.32537532e-01
-7.10137486e-01 -2.82648534e-01 9.60495591e-01 1.86805606e-01
-1.08676982e+00 4.56595838e-01 -9.47560728e-01 -5.94531715e-01
3.37028861e-01 1.90425262e-01 -1.28800318e-01 -2.84095317e-01
-1.28997695e+00 4.24960047e-01 -4.21815813e-02 3.89786571e-01
-3.37787330e-01 -6.64198101e-01 -8.58786106e-01 3.78664613e-01
3.51889849e-01 -5.49539506e-01 8.62081230e-01 -4.27818447e-01
-7.64346302e-01 3.25425193e-02 -6.90345585e-01 -4.46103334e-01
-3.71634662e-02 4.61202860e-01 -1.29428470e+00 -4.24952060e-01
2.03690112e-01 2.57911929e-03 5.41132510e-01 -9.94935751e-01
-8.25526118e-01 -3.58897328e-01 7.67930690e-03 -1.07514389e-01
-2.33794823e-01 -3.75087976e-01 -4.24474746e-01 -5.54782629e-01
2.80280381e-01 -1.17754126e+00 -6.04972720e-01 -5.18960655e-01
-3.71615767e-01 -4.40241456e-01 9.19250488e-01 -6.58058167e-01
1.86534643e+00 -1.99278140e+00 -4.53160644e-01 7.07026422e-01
3.90911520e-01 -5.59566207e-02 -3.21553260e-01 1.04536641e+00
2.30932161e-01 7.02558160e-02 6.97265789e-02 -4.55157697e-01
-7.19806254e-02 7.46468127e-01 -3.70053113e-01 3.90178502e-01
-1.10685758e-01 1.32781374e+00 -1.22381198e+00 -1.09434664e-01
2.34267250e-01 5.38007259e-01 -3.15023094e-01 -1.20194033e-01
-1.42290860e-01 5.17434478e-01 -7.94673920e-01 5.84027290e-01
3.80047083e-01 -8.51041794e-01 2.83294439e-01 2.21323431e-01
1.53513953e-01 3.59581381e-01 -7.97202289e-01 1.02285719e+00
-6.90156221e-01 9.29165959e-01 -5.18765688e-01 -1.21932960e+00
9.23725307e-01 2.12205067e-01 9.12847757e-01 -1.04575920e+00
-2.46255815e-01 -1.30279779e-01 -1.89808652e-01 -3.40943903e-01
6.72375321e-01 2.80481428e-01 -2.84312606e-01 4.48544085e-01
-5.16577125e-01 5.49942851e-01 5.67893982e-02 2.03115419e-01
1.25626254e+00 -5.93705893e-01 -6.91528153e-03 -1.89026341e-01
2.07644075e-01 -4.32336284e-03 4.29258466e-01 6.26971304e-01
-2.83069238e-02 2.37390116e-01 4.87673640e-01 -1.09265995e+00
-5.60383916e-01 -6.93076789e-01 2.72851050e-01 1.23354566e+00
-5.66942021e-02 -3.86363000e-01 -3.77085768e-02 -6.10928059e-01
3.77690196e-01 7.59908736e-01 -8.31054091e-01 6.43843189e-02
-5.92304289e-01 -7.58671105e-01 1.48505479e-01 6.23414159e-01
3.14306378e-01 -7.35004723e-01 6.11024201e-02 4.59791422e-01
-4.75301325e-01 -1.20245898e+00 -7.46375322e-01 -2.95582175e-01
-7.88474143e-01 -9.25828099e-01 -3.03572804e-01 -5.03089130e-01
7.45921731e-01 8.17090154e-01 1.40274465e+00 3.59665692e-01
2.73122266e-02 6.57568157e-01 -1.50288656e-01 -3.39909822e-01
1.68627985e-02 3.36128801e-01 -1.64254189e-01 2.66625226e-01
5.24670362e-01 -1.05289912e+00 -7.75398254e-01 4.83106047e-01
-7.65709102e-01 2.16008374e-03 6.39357567e-02 5.65310717e-01
6.36516571e-01 3.98576438e-01 7.27875650e-01 -8.86383057e-01
7.03282833e-01 -1.11023462e+00 -4.91275251e-01 2.38987774e-01
-8.29817832e-01 -7.53555298e-02 4.81117696e-01 -4.66932863e-01
-5.77543616e-01 -2.77925760e-01 1.26379669e-01 -1.98707581e-01
-3.05776969e-02 1.22802401e+00 3.59059602e-01 6.53658733e-02
3.00996661e-01 2.35901788e-01 -3.96994352e-01 -1.91684395e-01
2.65504390e-01 3.80304247e-01 1.55932516e-01 -1.30495727e-01
5.67130804e-01 5.81838429e-01 6.31190717e-01 -9.33083594e-01
-9.62099135e-01 -9.23155189e-01 -4.95765090e-01 -4.37057883e-01
5.53427815e-01 -8.80346775e-01 -9.56177235e-01 7.50916004e-02
-8.15566242e-01 -5.28636873e-01 2.18820991e-03 4.67214316e-01
-1.15998246e-01 -3.05907149e-02 -1.43839002e-01 -9.54297245e-01
-4.81019691e-02 -6.20635152e-01 1.06129575e+00 -3.96111570e-02
-5.71153983e-02 -1.78894627e+00 1.23895057e-01 4.34988253e-02
7.59830534e-01 8.07703555e-01 1.02935374e+00 -4.00900126e-01
-6.95782006e-01 -5.66955447e-01 -4.54602122e-01 -4.31551129e-01
5.54300725e-01 -1.63472518e-01 -4.47705984e-01 -1.34048954e-01
-6.82471454e-01 5.49081802e-01 7.83487320e-01 8.09857011e-01
1.26958263e+00 -8.40736151e-01 -8.59266818e-01 5.61691642e-01
1.19426942e+00 1.55895218e-01 3.49182963e-01 -1.08139955e-01
1.06499815e+00 7.19372034e-01 3.23365390e-01 5.72694540e-01
1.43395114e+00 5.31877756e-01 6.03356302e-01 -2.06838384e-01
8.27342197e-02 -6.98882759e-01 7.15440547e-05 8.26015115e-01
-3.01999748e-01 -6.96582615e-01 -1.48907912e+00 1.02511394e+00
-2.41150570e+00 -1.21476793e+00 -5.29892981e-01 1.97241271e+00
-2.25297630e-01 -1.11063138e-01 3.50372434e-01 -1.92922443e-01
4.78112340e-01 6.45247638e-01 -6.20566130e-01 -8.22103620e-02
1.16179742e-01 -4.62077975e-01 9.97432113e-01 5.77191412e-01
-9.45592284e-01 7.35882580e-01 6.51512957e+00 4.89468485e-01
-1.32772660e+00 5.34035414e-02 7.49423683e-01 2.19805520e-02
-6.97198927e-01 -2.14342728e-01 -7.22778916e-01 5.78712404e-01
1.49837279e+00 -9.46670994e-02 8.80666733e-01 5.25687695e-01
6.55695438e-01 1.42036855e-01 -7.74581730e-01 7.83661187e-01
-2.08242476e-01 -1.52259934e+00 -1.10358857e-01 4.31797981e-01
1.05451357e+00 7.22115338e-01 1.20657332e-01 4.46999907e-01
6.71768665e-01 -1.14170182e+00 -1.46588743e-01 9.74243343e-01
3.97946537e-01 -8.48043144e-01 5.20538390e-01 5.75590193e-01
-1.87259614e+00 -2.23292485e-01 5.36496080e-02 -2.95425326e-01
5.07349730e-01 7.37260461e-01 -8.89027834e-01 6.53426051e-01
6.11508131e-01 1.30005276e+00 -4.01186019e-01 7.80035079e-01
9.51279104e-02 1.14135599e+00 -5.42914033e-01 -1.56126887e-01
7.06873178e-01 -3.28250080e-01 3.85840952e-01 9.98933494e-01
5.37787914e-01 2.52872676e-01 5.81881940e-01 3.35455090e-01
-5.88977709e-02 -6.23721555e-02 -1.04792523e+00 -3.15925211e-01
2.89005488e-01 9.29554164e-01 -5.63993692e-01 -2.98636526e-01
-6.98531330e-01 3.86040300e-01 2.30949402e-01 9.43329155e-01
-6.73518419e-01 -3.17231454e-02 6.83694243e-01 6.06656849e-01
1.46372125e-01 -8.88451397e-01 -1.73056602e-01 -9.47192311e-01
-7.04446286e-02 -1.21989653e-01 6.18088007e-01 -5.14261663e-01
-1.47898221e+00 5.21483481e-01 1.89272314e-01 -1.41036594e+00
-5.84291816e-01 -1.61410317e-01 -1.04236829e+00 7.32528210e-01
-1.98995376e+00 -1.71060181e+00 -3.88734579e-01 7.94201195e-01
1.99809641e-01 -9.70032066e-03 5.28336346e-01 3.31564337e-01
-3.80777210e-01 2.71342874e-01 3.62409294e-01 1.01145647e-01
6.24129288e-02 -9.01472151e-01 1.03843868e+00 8.57553542e-01
2.90411860e-01 2.41128772e-01 3.29969764e-01 -8.51661384e-01
-1.39483953e+00 -1.67264783e+00 1.66663754e+00 -1.63215354e-01
1.03330505e+00 -9.54796523e-02 -8.71707439e-01 9.72913146e-01
-1.77581578e-01 5.09751558e-01 7.34489977e-01 7.43178546e-01
-5.90637103e-02 -2.88915187e-01 -6.93899214e-01 7.71648109e-01
1.36092770e+00 -5.22790909e-01 2.72626609e-01 7.13465929e-01
8.61247480e-01 -1.82062015e-01 -1.02071381e+00 1.81041226e-01
3.32165450e-01 -3.66720319e-01 8.43890548e-01 -7.05485582e-01
1.28285661e-01 -1.03720471e-01 -1.82007656e-01 -1.59181821e+00
-7.04231858e-01 -6.15798593e-01 -2.69904137e-01 6.72598362e-01
4.73632663e-01 -1.17724597e+00 5.80510497e-01 8.38680565e-01
-1.13092825e-01 -6.91615701e-01 -1.01897466e+00 -6.32714272e-01
-3.64461809e-01 -8.03723514e-01 1.39371037e+00 1.01766205e+00
-9.84137058e-02 2.59659320e-01 -6.78646266e-01 6.13205671e-01
3.68403465e-01 3.67268801e-01 8.38850081e-01 -1.42487431e+00
-4.17355932e-02 -4.41574872e-01 -5.54228604e-01 -1.18928075e+00
2.29567304e-01 -8.08681309e-01 -3.66228193e-01 -2.00586081e+00
-3.74187201e-01 -6.31267905e-01 -4.97953653e-01 6.45414114e-01
-1.32426813e-01 -2.25790232e-01 -1.63809121e-01 1.25396505e-01
-6.05281770e-01 6.18069291e-01 1.26355350e+00 -4.32850063e-01
-4.86751854e-01 6.68734908e-01 -5.13857901e-01 3.20654750e-01
7.23495305e-01 -5.03000975e-01 -7.55792677e-01 -4.31939840e-01
4.92060751e-01 5.28099358e-01 3.34587932e-01 -8.15660417e-01
4.53461885e-01 -6.85673833e-01 -3.58745568e-02 -1.03917646e+00
2.95950860e-01 -1.10737288e+00 4.24141377e-01 1.73965871e-01
-9.45147797e-02 6.24055088e-01 2.91339338e-01 1.16852021e+00
-2.80073553e-01 6.89121604e-01 -2.51647443e-01 1.63543552e-01
-7.99957514e-01 1.18323588e+00 -4.25896674e-01 -1.34714395e-01
7.96929598e-01 -9.66693610e-02 -3.44747305e-01 -9.73626852e-01
-6.66426420e-01 7.86721051e-01 2.02982519e-02 5.86603403e-01
3.92187685e-01 -1.50638187e+00 -5.84816158e-01 3.03546995e-01
3.60539407e-01 -1.70833394e-01 7.20581532e-01 8.10311794e-01
4.96592186e-02 5.62903643e-01 3.28163862e-01 -6.53243721e-01
-8.53124976e-01 4.70660627e-01 1.07290819e-01 -7.21321285e-01
-6.10543728e-01 2.12320521e-01 -1.66154116e-01 -5.42653680e-01
3.94321531e-02 -7.78511524e-01 -3.71300876e-01 -2.75821611e-02
2.53122956e-01 4.38034892e-01 9.55533981e-03 -9.15626884e-01
-1.30127311e-01 2.96439320e-01 5.40562987e-01 2.77105600e-01
1.66342103e+00 -5.01416266e-01 1.30871728e-01 7.73217618e-01
1.29316533e+00 -2.69274056e-01 -9.37564671e-01 -7.10329294e-01
-5.34479623e-04 -4.34137732e-01 3.58055294e-01 -5.09191513e-01
-1.20260739e+00 8.22639287e-01 2.00737178e-01 7.66382933e-01
1.06437695e+00 -1.48077771e-01 1.21593058e+00 4.50530469e-01
4.13562804e-01 -8.61265898e-01 -2.77443588e-01 6.75089955e-01
6.87197745e-01 -1.50822389e+00 -2.46208757e-01 -2.81046063e-01
-7.44196475e-01 7.87708640e-01 1.60410881e-01 -1.18139207e-01
1.53231287e+00 -3.44713062e-01 -2.49581218e-01 -4.60013360e-01
-8.82897496e-01 -5.81380069e-01 8.95535588e-01 5.27480304e-01
2.10737035e-01 7.18682170e-01 2.07749546e-01 2.27055594e-01
-6.47903383e-02 -3.75982039e-02 1.05251424e-01 6.34209871e-01
-7.06468672e-02 -8.59493852e-01 1.53140843e-01 9.44097042e-01
-1.67357307e-02 2.77903453e-02 2.67303467e-01 6.69986725e-01
-3.41027021e-01 1.19746280e+00 4.29853290e-01 -5.66315055e-01
3.98754567e-01 -2.13785097e-01 -3.02925467e-01 -4.16774571e-01
-4.06819433e-01 -2.47149527e-01 1.64146915e-01 -6.78815782e-01
-4.75927591e-01 -5.49994946e-01 -1.05678785e+00 -8.24649870e-01
2.06328817e-02 -4.72745951e-03 7.19219685e-01 1.08614218e+00
8.25878739e-01 9.51814950e-01 9.34239805e-01 -8.76436472e-01
3.31081390e-01 -8.03819478e-01 -7.42275238e-01 3.64473850e-01
8.03668976e-01 -6.71595514e-01 -9.91632715e-02 -3.85043979e-01]
|
[6.619620323181152, 2.432827949523926]
|
1fa352c5-79a8-4c3d-bf5f-a6f162438e5c
|
the-development-of-standard-perceptual
| null | null |
https://doi.org/10.6180/jase.202202_25(1).0022
|
http://jase.tku.edu.tw/articles/jase-202202-25-1-0022.pdf
|
The Development of Standard Perceptual Attributes in Indonesian for Soundscape Evaluation: Result from Initial Study
|
ISO 12913-1, 12913-2, and 12913-3 have standardized soundscape evaluation from different aspects such as definition and framework, data collection methods, and data analysis. Central to ISO 12913-2 is that an acoustic environment can be evaluated based on perceptual attributes standardized only in English. These perceptual attributes might be interpreted differently in a different country, resulting in incorrect soundscape evaluation. Thus, to overcome the problem, International collaboration was initiated to develop standard perceptual attributes for soundscape evaluation in 15 languages.
This study explains the development of soundscape perceptual attributes in Indonesian. A focus group discussion had been conducted to develop the attributes in Indonesian. Afterward, in-situ experiments were carried out to identify soundscape evaluation using two different perceptual attributes both in English and in Indonesian. The Wilkinson signed-rank test analysis shows that the rating score between English and Indonesian attributes is not significant on several attributes. Those attributes are pleasant, vibrant, calm, annoying, and monotonous. The other attributes (chaotic, uneventful, and eventful) are rated differently compared to the Indonesian version. It is interesting to note that using English attributes or using a straightforward translation might not be suitable for soundscape study in Indonesia.
|
['Ni Putu Amanda Nitidara', 'Sugeng Joko Sarwono', 'Winda Setiasari', 'Anugrah Sabdono Sudarsono']
|
2021-08-10
| null | null | null |
journal-of-applied-science-and-engineering
|
['soundscape-evaluation']
|
['audio']
|
[-2.59489566e-01 -7.59997070e-01 7.31163561e-01 -1.47329137e-01
-4.45817888e-01 -8.85196447e-01 2.02862039e-01 5.55911899e-01
-5.52696466e-01 5.08952737e-01 5.01432359e-01 -2.64731556e-01
-5.35937369e-01 -5.04604578e-01 -8.76397416e-02 -5.32355070e-01
6.45191520e-02 -2.22133100e-01 1.41148075e-01 -3.59322160e-01
4.50524628e-01 4.07455117e-01 -1.85123646e+00 -4.51582745e-02
3.61279279e-01 5.75833619e-01 3.91634613e-01 9.52528894e-01
-1.61040857e-01 6.08817339e-01 -1.18955719e+00 -1.55008938e-02
6.79569691e-02 -4.80200112e-01 -6.57628119e-01 -3.75977367e-01
7.72468746e-02 -1.93644181e-01 4.61277544e-01 1.10998285e+00
8.29469621e-01 5.38514495e-01 5.92426121e-01 -1.09314632e+00
-7.95236170e-01 3.33506107e-01 2.62365401e-01 -2.50600511e-04
5.80652893e-01 1.16124190e-01 8.95630836e-01 -6.52774096e-01
1.84222922e-01 9.99276161e-01 6.22050226e-01 3.76051813e-01
-1.13138652e+00 -9.17587876e-01 -2.93753892e-01 1.09017208e-01
-1.28570390e+00 -1.55518249e-01 8.23486090e-01 -5.06042123e-01
8.23906958e-01 8.23585331e-01 9.05518949e-01 8.23653579e-01
1.13802940e-01 -2.53038764e-01 1.83319652e+00 -4.80436772e-01
5.26489437e-01 5.27390540e-01 -2.41847023e-01 -3.00813496e-01
3.81505281e-01 1.21810682e-01 -2.73639053e-01 5.61226793e-02
3.17485303e-01 -5.52097738e-01 -9.40891951e-02 3.49458992e-01
-8.05436015e-01 5.29358149e-01 2.92709190e-02 6.51512682e-01
-3.07012141e-01 -1.38249144e-01 8.31702352e-01 4.32422519e-01
8.36876631e-02 8.71421337e-01 -3.71642560e-01 -9.80365396e-01
-4.59019542e-01 1.53853931e-02 8.98997307e-01 3.15131307e-01
4.22000408e-01 4.43118304e-01 5.24018705e-01 1.26147628e+00
4.72092032e-01 8.93098533e-01 3.31851631e-01 -1.08107758e+00
-7.26856962e-02 -1.65104926e-01 2.42653176e-01 -1.26680112e+00
-2.66815096e-01 -1.52529508e-01 -8.64649191e-02 5.09660900e-01
2.39779111e-02 -3.04574370e-01 -7.21473455e-01 1.56176114e+00
-8.92458409e-02 -4.82597679e-01 2.44968295e-01 9.22655523e-01
1.10186887e+00 9.14335907e-01 3.96530360e-01 -7.15355948e-02
1.04030287e+00 -4.90833580e-01 -1.11462545e+00 9.43262316e-03
3.17384243e-01 -1.36928523e+00 1.73368931e+00 4.36472297e-01
-7.50479817e-01 -8.15869749e-01 -1.14050066e+00 5.66174150e-01
-6.86429441e-01 -5.70415676e-01 1.40193820e-01 1.32671463e+00
-9.10721660e-01 1.83137476e-01 -4.39970374e-01 -5.49252152e-01
-6.35149717e-01 -5.82345854e-03 -3.45983148e-01 5.01332045e-01
-1.25740719e+00 1.03110707e+00 2.30035871e-01 2.18290851e-01
-5.84713340e-01 -3.26597482e-01 -8.81271124e-01 -3.45553428e-01
-6.12725057e-02 2.53732145e-01 1.17970967e+00 -8.04580331e-01
-1.78510356e+00 2.72398353e-01 3.31666619e-01 4.28827226e-01
-1.51027888e-01 -9.74777937e-02 -1.38329518e+00 1.75244987e-01
3.32708627e-01 1.76025078e-01 -9.49422941e-02 -1.57064199e+00
-6.99883223e-01 1.66245028e-01 -1.30482644e-01 2.77111351e-01
-3.11950296e-01 5.54682553e-01 1.21348701e-01 -6.27602220e-01
-3.02234031e-02 -8.59162331e-01 1.17047146e-01 -5.83941340e-01
-1.22937478e-01 1.92331448e-01 7.04974532e-01 -8.98555636e-01
1.42011237e+00 -2.42437840e+00 -8.36310506e-01 5.72756708e-01
-3.97479802e-01 2.09784359e-01 -1.22390851e-01 8.09911311e-01
-6.07813932e-02 4.11262691e-01 2.71091852e-02 2.99015105e-01
3.71622354e-01 3.78790587e-01 1.40780598e-01 2.00911731e-01
-1.50379673e-01 -1.94697663e-01 -1.11180389e+00 -4.06986773e-01
4.98355180e-01 5.12248695e-01 -2.98896879e-01 6.48847818e-02
6.61879420e-01 3.80217165e-01 -1.42307431e-01 7.77489781e-01
8.28458965e-01 8.89979184e-01 -2.97164559e-01 -1.50135905e-01
-8.05973291e-01 4.11898255e-01 -1.42454135e+00 9.93758440e-01
-8.20770919e-01 8.80070925e-01 2.51326352e-01 -4.65414256e-01
1.30395138e+00 8.20346951e-01 1.86149970e-01 -1.09970117e+00
-2.68196985e-02 8.17827165e-01 3.66598874e-01 -1.07233357e+00
8.43716800e-01 -5.89159787e-01 -2.43434563e-01 1.40584558e-01
-2.14746565e-01 -7.66278744e-01 3.17642957e-01 -3.43156219e-01
4.42609429e-01 -1.71862632e-01 9.95859224e-03 -3.68374497e-01
2.43729651e-01 -4.05705065e-01 5.24797857e-01 4.96509194e-01
-4.36006725e-01 4.89889890e-01 4.19365354e-02 1.10040374e-01
-9.37040508e-01 -1.21428716e+00 -3.86320055e-01 1.00236917e+00
-2.71414989e-03 -3.32978904e-01 -5.69615185e-01 1.16589867e-01
-4.33823735e-01 1.14638603e+00 -3.16734970e-01 -5.37340157e-03
-1.93808302e-01 -6.67362511e-02 6.72750533e-01 3.33381295e-01
4.37815905e-01 -1.36094391e+00 -8.92625868e-01 3.03618133e-01
-2.29363605e-01 -8.43920410e-01 -1.88501045e-01 3.66865158e-01
-3.33198369e-01 -3.46044272e-01 -3.95841181e-01 -8.37794960e-01
1.86055750e-01 1.24068074e-01 6.05241835e-01 -3.73470992e-01
9.71159060e-03 7.07636535e-01 -7.83650517e-01 -9.08858001e-01
-7.87492216e-01 -4.76979464e-01 1.39323696e-01 -4.77266788e-01
4.08743262e-01 -5.01329362e-01 -3.56833220e-01 3.76754850e-01
-1.04456329e+00 -6.08688116e-01 3.99548709e-01 1.65848002e-01
3.82138431e-01 2.97514439e-01 8.26035142e-01 -2.99049586e-01
1.03831768e+00 -2.93491900e-01 6.36234358e-02 -1.09922245e-01
-3.80509645e-01 -4.90236580e-01 8.38344216e-01 -3.34566981e-01
-1.06165588e+00 -6.12425327e-01 -6.43154383e-01 3.36182237e-01
-6.74208581e-01 6.87452376e-01 -2.09122419e-01 -9.73385572e-02
6.92655146e-01 -8.95669833e-02 -1.19107910e-01 -4.12317276e-01
-3.11144441e-01 1.23797750e+00 5.98602772e-01 -6.05911851e-01
6.99608982e-01 1.97045458e-03 -2.38946944e-01 -1.32906592e+00
-2.22498149e-01 -5.52838743e-01 -1.60007507e-01 -7.40591764e-01
1.00452244e+00 -6.10335112e-01 -4.29839015e-01 4.27072465e-01
-6.73775077e-01 -1.96863919e-01 -3.20802748e-01 1.19617665e+00
-2.62891412e-01 2.93274373e-01 -3.07913631e-01 -1.10568488e+00
-6.89401478e-02 -1.12752533e+00 4.28568929e-01 3.78074288e-01
-7.42334425e-01 -1.03037655e+00 4.24261093e-01 2.15232134e-01
7.48767972e-01 4.68455791e-01 7.42375791e-01 -6.08230174e-01
6.44745171e-01 -2.79192656e-01 3.15981358e-01 9.62993860e-01
7.36015856e-01 3.66269678e-01 -9.88692522e-01 4.60513271e-02
2.99282610e-01 -3.29394281e-01 6.63803145e-02 1.52865693e-01
4.85738873e-01 -2.59387404e-01 6.44784510e-01 8.98346305e-02
1.66520512e+00 1.02208853e+00 5.33004284e-01 7.19282806e-01
2.63675869e-01 9.12800074e-01 7.81073749e-01 3.10182989e-01
-4.52355202e-03 4.13749635e-01 3.18968087e-01 -7.80218467e-02
-1.54544711e-01 -1.84297159e-01 6.96468234e-01 1.47479296e+00
-1.69756606e-01 -5.12146726e-02 -8.05197477e-01 7.15797782e-01
-7.51950324e-01 -9.47000921e-01 -4.06904697e-01 2.34636188e+00
7.42116868e-01 9.81470942e-02 2.62178004e-01 6.76378846e-01
6.87500358e-01 9.25146788e-02 3.52197170e-01 -1.52232111e+00
-1.65799260e-01 4.93191123e-01 1.91122100e-01 6.45392120e-01
-8.74706924e-01 3.42730552e-01 6.41748524e+00 3.64180416e-01
-1.37022603e+00 2.90354341e-02 -2.62805521e-02 3.47956978e-02
-6.27529025e-01 4.37080488e-02 -1.04778104e-01 5.92784703e-01
1.15593016e+00 -2.10771933e-01 2.78118372e-01 3.93383890e-01
7.05938339e-01 -2.51349300e-01 -4.93168801e-01 6.58325434e-01
-1.06494702e-01 -5.69084644e-01 -1.57385811e-01 -2.10759565e-01
5.80497563e-01 -2.63131469e-01 2.31665239e-01 1.61369130e-01
-2.85438627e-01 -8.99250507e-01 1.04884851e+00 2.32690930e-01
6.10387027e-01 -8.45762432e-01 9.87309515e-01 -3.45313668e-01
-1.16151047e+00 1.36758432e-01 -3.56933028e-01 -4.65008140e-01
2.19121426e-01 1.35223702e-01 -6.30263507e-01 2.61094809e-01
1.01434851e+00 -2.08247095e-01 -4.31527406e-01 1.29408896e+00
-2.22543806e-01 1.12218523e+00 -3.67343724e-01 -5.53356469e-01
5.37458181e-01 -2.74948150e-01 7.87470818e-01 1.56375790e+00
7.51605809e-01 -2.89339930e-01 -4.28515017e-01 6.11217320e-01
7.42454052e-01 6.96463943e-01 -5.85126400e-01 -1.67914510e-01
8.53132367e-01 9.96866167e-01 -7.83685863e-01 1.53628767e-01
-5.60905933e-01 6.00583792e-01 -9.54483509e-01 4.65899229e-01
-7.66947448e-01 -1.03722668e+00 4.21609402e-01 -2.89194155e-02
-9.64496434e-02 -2.31162310e-01 -3.32720190e-01 3.79411550e-03
-4.44459207e-02 -8.35300803e-01 -1.22702226e-01 -8.19531620e-01
-8.88228118e-01 4.72548932e-01 3.73961300e-01 -1.33038139e+00
1.30874261e-01 -6.77285075e-01 -8.08849096e-01 1.02310431e+00
-7.80741692e-01 -5.89775264e-01 -3.51128399e-01 1.96330801e-01
3.42060506e-01 4.22416031e-02 1.24612486e+00 5.21082044e-01
-3.27812552e-01 3.84028345e-01 3.43219191e-01 -9.83703285e-02
9.04814363e-01 -1.40958595e+00 -9.03056860e-02 7.34848738e-01
-3.08211535e-01 6.33156061e-01 1.20660603e+00 -4.57512140e-01
-6.84049964e-01 -6.87977850e-01 1.00132561e+00 8.24530721e-02
7.69004464e-01 -2.54630353e-02 -7.34338284e-01 -7.03700781e-02
5.17718136e-01 -5.64812779e-01 1.26134682e+00 -5.26380613e-02
-1.32186159e-01 -2.83128113e-01 -1.01522315e+00 7.13975489e-01
2.21578255e-01 -6.43060386e-01 -5.31349719e-01 -2.30180189e-01
5.68410218e-01 1.63118795e-01 -1.13822496e+00 -2.95828246e-02
8.39830160e-01 -1.11551666e+00 5.62216282e-01 5.07735275e-02
1.15329616e-01 -5.29868066e-01 -4.77964133e-01 -1.52079916e+00
-2.25379989e-01 -6.90873146e-01 1.09555674e+00 1.65691686e+00
5.87327540e-01 -9.74757314e-01 3.63181935e-05 5.02054036e-01
-5.30368924e-01 -9.23847929e-02 -7.39841759e-01 -1.12751997e+00
9.84285623e-02 -7.54956961e-01 4.05681431e-01 7.73038566e-01
1.70780301e-01 2.10121274e-01 -2.31976137e-01 1.05266526e-01
1.78364873e-01 -3.68242741e-01 5.01007617e-01 -1.08067048e+00
-1.22139394e-01 -5.46337008e-01 -4.03828919e-01 7.32945427e-02
-3.43138486e-01 -2.53226757e-01 2.32120916e-01 -1.80414605e+00
-9.40747559e-02 -4.26057637e-01 -4.36005116e-01 1.82859480e-01
-3.04885712e-02 3.14206481e-01 4.63843465e-01 -1.43173471e-01
2.21493572e-01 1.24574661e-01 1.14195216e+00 1.30986422e-01
-3.23933512e-01 -5.94078451e-02 -6.57010734e-01 4.89081830e-01
1.09540498e+00 -5.19011974e-01 -5.64478695e-01 -1.81761697e-01
4.71300691e-01 -2.95332074e-01 2.82824785e-01 -1.28131330e+00
2.42852401e-02 -4.47001606e-01 -2.09653080e-01 -3.03847253e-01
3.82314980e-01 -1.05166531e+00 5.94470024e-01 3.91953051e-01
-3.85380447e-01 2.89085239e-01 4.15803909e-01 -1.00994244e-01
-5.26793957e-01 -7.32333958e-01 7.25859642e-01 5.91899753e-02
-8.68514299e-01 -5.80022275e-01 -7.80425847e-01 -4.83932793e-02
8.41691256e-01 -9.34732258e-01 -2.21706599e-01 -7.13836372e-01
-6.25459373e-01 -3.88167083e-01 5.56723893e-01 3.22527409e-01
5.47824860e-01 -1.51711845e+00 -7.15000808e-01 -9.54210684e-02
1.80481762e-01 -6.75053358e-01 5.54025471e-01 8.11568260e-01
-1.16707695e+00 3.15246105e-01 -7.63342679e-01 2.81432066e-02
-1.20547926e+00 3.36119682e-02 8.59240741e-02 5.19742250e-01
-4.65228930e-02 5.96284807e-01 -1.98956374e-02 -4.14387733e-01
1.01200871e-01 -2.48553649e-01 -3.80400419e-01 1.90244213e-01
2.06098497e-01 8.93164992e-01 8.38881806e-02 -1.01741362e+00
-5.61432719e-01 7.49286771e-01 6.77909672e-01 -8.72536302e-01
9.71689403e-01 -5.81731163e-02 -2.37104535e-01 1.08741474e+00
1.41773868e+00 8.62432539e-01 -4.69350576e-01 6.80790901e-01
-2.57709801e-01 -4.70809877e-01 -9.02498886e-02 -1.12803781e+00
-4.37259346e-01 7.05472410e-01 8.68356526e-01 7.36751676e-01
1.41185892e+00 -3.72543752e-01 5.16450465e-01 1.22516230e-01
-4.60083932e-02 -1.82624149e+00 -1.68942008e-02 4.67965961e-01
1.01196754e+00 -5.75108528e-01 -5.86893082e-01 -9.30198357e-02
-8.51930559e-01 1.16046906e+00 5.07271767e-01 3.42223793e-01
7.97699213e-01 4.29776818e-01 6.41273022e-01 4.43355665e-02
-2.47230873e-01 -5.76642603e-02 3.37481499e-04 1.05294669e+00
8.84343445e-01 5.03710985e-01 -7.09106982e-01 4.80548918e-01
-6.69115484e-01 -5.51892698e-01 9.74698365e-01 8.97547066e-01
-6.57239735e-01 -1.02804708e+00 -9.34166491e-01 1.83431834e-01
-7.61455417e-01 -2.48556249e-02 -4.53960359e-01 1.05597317e+00
3.56737763e-01 1.60018945e+00 3.83805968e-02 -8.35071921e-01
6.03647888e-01 -1.62904710e-01 7.65369609e-02 -4.68773037e-01
-7.78026998e-01 4.53392118e-01 5.02428949e-01 3.26547958e-02
-6.86110735e-01 -6.08559728e-01 -1.31397438e+00 -2.73357719e-01
-2.10706502e-01 7.24346638e-01 1.16810274e+00 5.55857122e-01
-2.25533083e-01 6.00228727e-01 8.84358406e-01 -2.39183441e-01
-2.02105522e-01 -1.20055485e+00 -8.98564816e-01 3.87077570e-01
2.26971045e-01 -2.70072401e-01 -8.03383708e-01 -1.37979731e-01]
|
[15.177106857299805, 5.605398654937744]
|
82d76d62-1d6b-4a35-b7a4-9e7e3d76fd63
|
refinement-of-predicted-missing-parts-enhance
|
2010.04278
| null |
https://arxiv.org/abs/2010.04278v1
|
https://arxiv.org/pdf/2010.04278v1.pdf
|
Refinement of Predicted Missing Parts Enhance Point Cloud Completion
|
Point cloud completion is the task of predicting complete geometry from partial observations using a point set representation for a 3D shape. Previous approaches propose neural networks to directly estimate the whole point cloud through encoder-decoder models fed by the incomplete point set. By predicting the complete model, the current methods compute redundant information because the output also contains the known incomplete input geometry. This paper proposes an end-to-end neural network architecture that focuses on computing the missing geometry and merging the known input and the predicted point cloud. Our method is composed of two neural networks: the missing part prediction network and the merging-refinement network. The first module focuses on extracting information from the incomplete input to infer the missing geometry. The second module merges both point clouds and improves the distribution of the points. Our experiments on ShapeNet dataset show that our method outperforms the state-of-the-art methods in point cloud completion. The code of our methods and experiments is available in \url{https://github.com/ivansipiran/Refinement-Point-Cloud-Completion}.
|
['Cristian Lopez', 'Ivan Sipiran', 'Alexander Apaza', 'Alexis Mendoza']
|
2020-10-08
| null | null | null | null |
['point-cloud-completion']
|
['computer-vision']
|
[-1.51807472e-01 3.87281984e-01 -6.19162526e-03 -6.14143670e-01
-1.03254211e+00 -2.80967444e-01 2.84243792e-01 9.25491303e-02
-8.49500969e-02 2.57417202e-01 -1.06818698e-01 -3.05152102e-03
2.36745059e-01 -9.34425235e-01 -1.38604403e+00 -2.15010643e-01
1.01655819e-01 1.12478817e+00 7.07089305e-02 -5.20079434e-02
2.28486657e-01 7.28994071e-01 -1.50972176e+00 3.71755153e-01
6.20435953e-01 1.10153162e+00 5.99762261e-01 6.00625098e-01
-2.58441657e-01 2.35058174e-01 -2.12997928e-01 -3.54063421e-01
6.19536877e-01 3.66176128e-01 -6.24167800e-01 -2.58576646e-02
7.49555290e-01 -7.51556814e-01 -5.11375844e-01 9.27059531e-01
4.15059328e-01 -1.77966073e-01 3.26877117e-01 -1.23575246e+00
-6.79584980e-01 5.01958132e-01 -5.12306631e-01 -5.13913453e-01
3.95971499e-02 -1.58169761e-01 7.69752204e-01 -1.50493944e+00
6.97516263e-01 1.05513740e+00 9.25767660e-01 3.99580598e-01
-9.54194725e-01 -7.96402931e-01 1.54316097e-01 3.21158096e-02
-1.88562429e+00 -4.58288908e-01 1.05140579e+00 -4.83659953e-01
1.03946865e+00 1.62277911e-02 6.65596962e-01 4.28773969e-01
-1.65558994e-01 7.31034875e-01 6.57917485e-02 -1.89429134e-01
1.50411308e-01 -9.64623392e-02 -2.64723629e-01 7.17935205e-01
8.80612135e-02 2.45306522e-01 -1.94404051e-01 -1.86139941e-01
1.09440577e+00 5.15918911e-01 -9.59838256e-02 -6.07394874e-01
-1.12726915e+00 4.55887824e-01 9.18808937e-01 -1.25765800e-01
-5.89851379e-01 3.84336650e-01 -1.24120332e-01 4.59058359e-02
8.20526123e-01 9.74096730e-02 -8.34839165e-01 1.34242609e-01
-1.22951257e+00 5.04464090e-01 6.64676070e-01 1.56791961e+00
1.13045108e+00 -3.82583588e-02 4.47095871e-01 6.28311515e-01
5.26406527e-01 6.83769941e-01 -1.22307681e-01 -1.18132663e+00
8.99267495e-01 8.95684004e-01 1.39988765e-01 -6.80934548e-01
-2.61276454e-01 -4.16235775e-01 -7.59037077e-01 5.61938643e-01
9.10614282e-02 -1.77059621e-01 -1.08463085e+00 1.44658923e+00
4.88462895e-01 6.55886471e-01 -2.68004179e-01 9.54852462e-01
1.14151943e+00 5.79255939e-01 -3.26707333e-01 5.19773662e-01
8.46965969e-01 -8.49962294e-01 -2.64015198e-01 -6.96525648e-02
4.09068406e-01 -7.78666079e-01 4.31384474e-01 2.05746695e-01
-1.46314585e+00 -8.70910048e-01 -1.15568936e+00 -6.45031512e-01
-1.62458301e-01 5.80986917e-01 2.98679650e-01 -1.86248228e-01
-1.21290088e+00 1.02621126e+00 -1.15850985e+00 1.45483360e-01
8.43379200e-01 6.68731153e-01 -4.67476577e-01 -2.07113817e-01
-4.50294793e-01 4.93639857e-01 2.57940590e-01 1.55711979e-01
-8.48653615e-01 -1.02821994e+00 -9.29238379e-01 2.08081916e-01
1.45023018e-01 -1.11778533e+00 1.52000952e+00 -6.51456118e-01
-1.13247764e+00 6.37513459e-01 -3.98404062e-01 -1.63425580e-01
4.40644503e-01 -4.02800888e-01 2.37415209e-01 -1.11621074e-01
1.55932441e-01 1.26194704e+00 7.54516304e-01 -1.64128637e+00
-7.52944231e-01 -7.35039115e-01 -1.75648987e-01 3.05192679e-01
5.20294189e-01 -6.00537121e-01 -8.59905422e-01 -3.34169686e-01
7.04176605e-01 -8.78905058e-01 -2.69604951e-01 4.34577107e-01
-4.97849405e-01 -2.06117108e-01 8.96727979e-01 -7.96130300e-01
4.54774767e-01 -2.15811563e+00 6.92694858e-02 1.81491613e-01
4.87938941e-01 -1.22273192e-01 -2.55983293e-01 5.12426794e-01
-3.70109409e-01 -2.54635117e-03 -3.68904054e-01 -1.14644551e+00
2.30526756e-02 8.99179578e-02 -5.31439543e-01 4.82395083e-01
3.24472189e-01 9.55484033e-01 -5.79961598e-01 -1.95401534e-01
4.75290775e-01 8.92916262e-01 -6.72394276e-01 2.03101262e-01
-4.81128991e-01 5.76858044e-01 -2.23603487e-01 8.47258747e-01
1.24660301e+00 -3.57441992e-01 -3.06668937e-01 -2.04332873e-01
-1.18387625e-01 2.02913940e-01 -1.25753355e+00 2.53603745e+00
-1.79806665e-01 4.29248869e-01 5.61407693e-02 -5.41316807e-01
1.16762733e+00 3.03516328e-01 7.80894518e-01 -3.49908650e-01
9.08408612e-02 2.93606907e-01 -3.09230536e-01 -2.04548270e-01
6.62966669e-01 1.24461323e-01 3.18715036e-01 7.39271492e-02
2.14143783e-01 -5.41903496e-01 -4.30295259e-01 4.81869280e-02
8.55588257e-01 7.14903831e-01 -3.78167257e-02 3.10393661e-01
2.95914590e-01 1.39675662e-01 6.41424537e-01 3.21560115e-01
2.17228428e-01 1.18229675e+00 3.58864032e-02 -7.74350643e-01
-1.58726740e+00 -1.19755971e+00 3.83800864e-02 4.90116954e-01
3.63195449e-01 -4.96330529e-01 -5.39572716e-01 -3.91726017e-01
2.66234010e-01 7.73736298e-01 -4.40132916e-01 2.29392797e-01
-6.39305234e-01 2.44567007e-01 2.27508232e-01 7.57957935e-01
3.16527337e-01 -7.67159343e-01 -3.04557592e-01 2.56653614e-02
-1.82403132e-01 -1.12038302e+00 -1.97915584e-01 -1.13637179e-01
-1.30753446e+00 -9.79713082e-01 -4.47371542e-01 -7.76555538e-01
1.00468540e+00 1.99125290e-01 1.25749731e+00 5.11756420e-01
5.86263947e-02 -6.11387601e-04 -1.46406248e-01 -8.33521366e-01
-1.13308161e-01 1.30967334e-01 -1.23940393e-01 -3.31464946e-01
5.07957995e-01 -9.65707421e-01 -5.38217843e-01 4.56483252e-02
-5.87139308e-01 4.83686268e-01 7.01991498e-01 3.38822186e-01
1.22622073e+00 -2.21666351e-01 9.73346457e-02 -6.40471458e-01
-8.07364061e-02 -4.76910472e-01 -9.04969096e-01 -1.71708226e-01
-1.72979534e-01 1.21837214e-01 4.61275339e-01 6.63378686e-02
-8.72606814e-01 7.42690742e-01 -5.99846601e-01 -1.31136608e+00
-4.63689238e-01 3.17908615e-01 -1.94916010e-01 1.21496290e-01
3.52933526e-01 1.22141294e-01 1.38451336e-02 -1.07969296e+00
3.68506610e-01 4.74369258e-01 7.60731220e-01 -3.71192873e-01
9.19504344e-01 7.82924473e-01 7.73872286e-02 -6.28427148e-01
-5.47203779e-01 -5.32848835e-01 -1.27446461e+00 -9.72895697e-02
6.04831338e-01 -1.32576263e+00 -6.48087800e-01 2.21633390e-01
-1.89559710e+00 -7.88604543e-02 -6.41637921e-01 5.16706526e-01
-7.50193954e-01 2.58482903e-01 -5.15120685e-01 -5.37278771e-01
-5.39292991e-01 -1.20907199e+00 1.73711300e+00 -8.54863226e-02
9.33763161e-02 -5.45655549e-01 1.96365371e-01 1.25053898e-01
-1.23503610e-01 2.15081021e-01 6.63271487e-01 -4.86046880e-01
-1.23413646e+00 -4.19162273e-01 -2.30639696e-01 1.45572871e-01
-2.69946426e-01 4.47895341e-02 -1.05274057e+00 -2.27620140e-01
7.69338235e-02 6.70888051e-02 7.10592866e-01 6.20679498e-01
1.37969005e+00 -8.37591365e-02 -5.25979280e-01 1.10521257e+00
1.71283150e+00 -3.01340334e-02 6.62669659e-01 -1.15102939e-01
1.05187201e+00 3.54081929e-01 3.77708375e-01 5.52349210e-01
7.13461459e-01 5.22019207e-01 1.12121964e+00 -7.97947794e-02
-2.61831313e-01 -7.21939921e-01 -1.44158065e-01 9.17983234e-01
-2.38563329e-01 1.02399290e-01 -1.09869289e+00 5.47653794e-01
-1.94354105e+00 -7.34134614e-01 -2.80779660e-01 2.18310905e+00
3.96522105e-01 -1.42713860e-01 -2.87481189e-01 -8.49059969e-02
5.86248219e-01 -1.15863748e-01 -7.93985963e-01 -1.02149218e-01
2.92698801e-01 2.10022956e-01 4.77842122e-01 7.64680326e-01
-9.30171371e-01 1.00704503e+00 5.20949221e+00 5.11092722e-01
-9.23303664e-01 6.00930005e-02 2.43615881e-01 -2.36828983e-01
-3.05042833e-01 1.71875507e-01 -8.60475898e-01 2.03734413e-01
6.15792572e-01 2.65721470e-01 2.60752350e-01 1.12123370e+00
1.11763962e-01 1.48227334e-01 -1.48441935e+00 1.24924982e+00
4.40073088e-02 -1.63564754e+00 5.67007065e-02 2.18030170e-01
6.31925821e-01 8.18770587e-01 -1.74891010e-01 1.43415317e-01
2.49482036e-01 -8.46361041e-01 1.08621562e+00 1.06657374e+00
1.17909861e+00 -7.20316529e-01 7.01890290e-01 7.93227434e-01
-1.15343797e+00 2.97558635e-01 -6.82206869e-01 -2.42016777e-01
2.97924906e-01 6.73942089e-01 -1.04263222e+00 7.61799335e-01
7.65796185e-01 9.82582510e-01 -4.49624717e-01 1.33604681e+00
-1.43308342e-01 1.97605059e-01 -6.52278125e-01 5.45746446e-01
-8.86052251e-02 -3.39040726e-01 7.44372129e-01 6.22180879e-01
5.52022278e-01 1.13642156e-01 1.55881360e-01 1.43411088e+00
-2.93559611e-01 -2.91484684e-01 -8.46695244e-01 4.96737540e-01
7.63936877e-01 1.01760387e+00 -4.21742469e-01 -4.12622958e-01
-3.99926335e-01 7.09478259e-01 7.16583133e-01 2.62320131e-01
-5.71392059e-01 -2.83992559e-01 7.60985136e-01 2.24392250e-01
5.01063466e-01 -5.65242648e-01 -9.92455065e-01 -9.96295035e-01
3.27423811e-01 -1.97767779e-01 -1.23381503e-01 -1.23964536e+00
-9.85113800e-01 5.98688841e-01 -1.42452801e-02 -1.53715444e+00
-1.74702957e-01 -4.32298362e-01 -4.84212488e-01 1.17963839e+00
-1.35626030e+00 -1.41795599e+00 -6.00978374e-01 3.25934172e-01
8.39240074e-01 -1.01704478e-01 7.24194407e-01 4.00373459e-01
-9.75519866e-02 2.19155118e-01 -5.44933751e-02 3.21173847e-01
1.92638844e-01 -1.07133055e+00 1.05932438e+00 7.07947195e-01
2.43744865e-01 3.16362798e-01 3.38642746e-01 -1.06803393e+00
-1.29755378e+00 -1.22042537e+00 9.02393818e-01 -6.71307921e-01
-1.26035005e-01 -5.33424318e-01 -1.01194692e+00 9.90808010e-01
-6.31481633e-02 6.30851388e-02 2.25796044e-01 -2.25603074e-01
-1.82423696e-01 1.33489128e-02 -1.02988183e+00 4.13928181e-01
1.14350343e+00 -3.11655134e-01 -5.83702922e-01 2.05617934e-01
1.13820815e+00 -9.66869414e-01 -8.69767368e-01 6.00267589e-01
3.96060884e-01 -7.99716890e-01 1.11193180e+00 -3.01973641e-01
8.50302279e-01 -4.73570049e-01 -4.00872976e-01 -1.23713017e+00
-3.74146581e-01 -6.69381022e-02 -3.16779733e-01 7.26771116e-01
4.98170644e-01 -1.07881755e-01 1.35813224e+00 6.90097511e-01
-6.22051656e-01 -1.02890766e+00 -9.09352362e-01 -3.33990306e-01
1.34466559e-01 -8.48098159e-01 1.25103700e+00 7.08087027e-01
-5.01706839e-01 2.43829757e-01 -8.67858529e-02 5.73413551e-01
7.47367978e-01 2.79219657e-01 1.11298740e+00 -1.46091449e+00
-1.00178588e-02 -3.15209739e-02 -4.31816906e-01 -1.42507660e+00
1.48053452e-01 -1.03255069e+00 4.36346093e-03 -1.92984843e+00
-1.70052037e-01 -6.53761983e-01 3.31552297e-01 6.48051620e-01
3.13357949e-01 6.85849339e-02 4.20561016e-01 4.62722182e-01
-3.35924506e-01 7.13221252e-01 1.30014873e+00 -2.05299988e-01
-2.90376455e-01 2.68242747e-01 -3.94653380e-01 8.54342222e-01
8.29571068e-01 -6.26473844e-01 -2.78441668e-01 -1.12645662e+00
2.79016048e-01 2.39896894e-01 6.06310189e-01 -1.32352555e+00
6.24869764e-01 1.22412078e-01 7.53582180e-01 -1.64727640e+00
9.81596231e-01 -1.28808153e+00 4.67662603e-01 2.95874983e-01
-3.36440280e-03 3.68770242e-01 2.15539187e-01 5.00382423e-01
-5.16349040e-02 -1.56730756e-01 3.42076182e-01 -3.41706961e-01
-4.88070846e-01 9.58496034e-01 4.64169770e-01 -4.26678568e-01
7.58864403e-01 -4.25417602e-01 4.58963253e-02 -4.09109414e-01
-7.83437133e-01 2.86747009e-01 6.01205170e-01 6.32764637e-01
1.20403945e+00 -1.39581156e+00 -1.03693366e+00 5.48854947e-01
-2.96233874e-02 1.09681976e+00 2.50179321e-01 4.61042672e-01
-8.65511000e-01 3.29103976e-01 -1.03387721e-02 -9.84477401e-01
-1.11966777e+00 5.64611614e-01 3.54224741e-01 2.52459973e-01
-8.54470253e-01 7.88092434e-01 4.90743630e-02 -9.91932452e-01
2.42784157e-01 -5.84237933e-01 2.14415044e-01 -5.03136635e-01
2.62062311e-01 3.64464909e-01 3.30353707e-01 -8.23547006e-01
-8.55901465e-02 7.17216790e-01 4.53668162e-02 -6.01803772e-02
1.65854359e+00 7.67772039e-03 -2.18969420e-01 2.91068554e-01
1.40286696e+00 -3.04861546e-01 -1.46370804e+00 -4.66865391e-01
-3.60218674e-01 -6.27565026e-01 6.44441112e-04 -6.02840781e-01
-1.25617552e+00 1.09215605e+00 3.26856554e-01 -4.19823289e-01
8.01442087e-01 2.27963980e-02 7.44109273e-01 3.78513813e-01
5.29390574e-01 -7.46967256e-01 -5.65300524e-01 6.90145612e-01
1.21892142e+00 -1.20211124e+00 1.39197782e-01 -6.49343729e-01
-2.63077617e-01 1.09014881e+00 7.66273439e-01 -4.29769218e-01
1.05836940e+00 3.75237942e-01 -1.57274440e-01 -4.74886507e-01
-8.25512111e-01 -1.22071616e-01 3.57178420e-01 6.25645876e-01
1.32981995e-02 1.62977740e-01 3.40162039e-01 5.09425819e-01
-6.84272766e-01 2.13629395e-01 2.70560384e-01 7.56389976e-01
-3.02419364e-01 -1.01438940e+00 -5.19019723e-01 5.87471724e-01
3.20845917e-02 -4.58631739e-02 -5.38357198e-01 6.75364017e-01
4.08884019e-01 4.92366791e-01 5.20048618e-01 -6.26489580e-01
5.02928913e-01 -7.57285878e-02 2.32518911e-01 -8.06602061e-01
-3.12915534e-01 5.72331510e-02 -3.30390096e-01 -7.38138735e-01
-6.71162903e-02 -5.99969506e-01 -1.51455748e+00 -4.60733652e-01
-2.50793844e-01 -2.82112900e-02 1.10192072e+00 6.00875556e-01
9.12540913e-01 3.47296894e-01 3.97811055e-01 -1.66607857e+00
-3.10649663e-01 -9.50454354e-01 -3.43994617e-01 3.92849594e-01
5.25973976e-01 -5.00142694e-01 1.06334135e-01 4.88136597e-02]
|
[8.315632820129395, -3.5398099422454834]
|
f6d00f27-4677-4079-b52a-60b8575faa2f
|
an-experimental-study-of-the-transferability
|
2012.10258
| null |
https://arxiv.org/abs/2012.10258v1
|
https://arxiv.org/pdf/2012.10258v1.pdf
|
An Experimental Study of the Transferability of Spectral Graph Networks
|
Spectral graph convolutional networks are generalizations of standard convolutional networks for graph-structured data using the Laplacian operator. A common misconception is the instability of spectral filters, i.e. the impossibility to transfer spectral filters between graphs of variable size and topology. This misbelief has limited the development of spectral networks for multi-graph tasks in favor of spatial graph networks. However, recent works have proved the stability of spectral filters under graph perturbation. Our work complements and emphasizes further the high quality of spectral transferability by benchmarking spectral graph networks on tasks involving graphs of different size and connectivity. Numerical experiments exhibit favorable performance on graph regression, graph classification, and node classification problems on two graph benchmarks. The implementation of our experiments is available on GitHub for reproducibility.
|
['Xavier Bresson', 'Axel Nilsson']
|
2020-12-18
| null | null | null | null |
['graph-regression']
|
['graphs']
|
[-1.46122072e-02 2.81117827e-01 2.15646829e-02 1.52073801e-01
-8.80779177e-02 -8.08159292e-01 2.73667246e-01 3.90960693e-01
-1.07728764e-01 7.62082696e-01 3.01021058e-02 -6.77362740e-01
-3.20379913e-01 -9.94219065e-01 -8.17357183e-01 -4.20209378e-01
-6.40396059e-01 7.93166980e-02 3.76767308e-01 -3.27835023e-01
-5.10126911e-02 6.26473427e-01 -8.70096505e-01 1.52777612e-01
7.27552772e-01 5.12652576e-01 -1.56663537e-01 9.79303300e-01
7.96551183e-02 5.24020016e-01 -2.40985528e-01 -3.80730301e-01
2.88060993e-01 -5.17993093e-01 -8.92149091e-01 -6.00817017e-02
9.21440482e-01 3.56513470e-01 -7.83456624e-01 1.53138030e+00
4.01924580e-01 7.53306374e-02 5.60015857e-01 -1.47280157e+00
-1.22508800e+00 7.93664455e-01 -4.21653599e-01 6.13951862e-01
3.85016292e-01 4.39682864e-02 1.37510705e+00 -6.44622922e-01
6.68400288e-01 1.24534261e+00 1.21026361e+00 1.84329674e-01
-1.65414453e+00 -6.86094344e-01 1.76273137e-02 -9.02731996e-03
-1.46448278e+00 -1.41547382e-01 6.51922405e-01 -5.27946591e-01
1.12961090e+00 1.15760654e-01 8.96495223e-01 1.04872930e+00
3.42195481e-01 2.39238858e-01 9.04009640e-01 -2.23389015e-01
-1.66996717e-01 -1.53799370e-01 2.51600705e-02 1.20246911e+00
5.64783335e-01 4.82660495e-02 -3.84142131e-01 -2.97739301e-02
1.05083382e+00 -3.62066060e-01 -5.12681425e-01 -5.49751341e-01
-9.84548867e-01 9.74802494e-01 8.54050577e-01 4.96665180e-01
3.18830982e-02 4.13627267e-01 5.10299444e-01 8.85236382e-01
6.41168952e-01 4.38827217e-01 -2.57566245e-03 3.73639256e-01
-4.81470317e-01 -1.55603504e-02 1.08101094e+00 1.03499269e+00
9.43416178e-01 2.16390997e-01 3.73714089e-01 4.93998379e-01
-7.36037567e-02 4.29869384e-01 7.33116940e-02 -4.96530205e-01
3.44912618e-01 7.38480806e-01 -4.49050069e-01 -1.48322463e+00
-8.90718937e-01 -7.84685731e-01 -1.27633047e+00 -2.82246284e-02
6.26965404e-01 -2.10005343e-01 -6.80794299e-01 1.60202861e+00
-1.92708462e-01 4.76061672e-01 -8.97974223e-02 7.50182688e-01
1.05934548e+00 3.23015362e-01 -2.47762293e-01 9.57547128e-02
7.70372689e-01 -4.88853902e-01 -3.92432660e-01 -1.36102363e-01
7.77718723e-01 -5.15826821e-01 1.17085433e+00 -3.12250443e-02
-8.96088898e-01 -2.55162627e-01 -1.20227706e+00 2.51508594e-01
-6.71517372e-01 -2.55748838e-01 9.29135323e-01 7.34041154e-01
-1.59364116e+00 9.39988732e-01 -6.52732611e-01 -8.69439423e-01
4.61978257e-01 4.59903836e-01 -7.13417172e-01 7.41049200e-02
-1.29564357e+00 6.47901773e-01 4.13176209e-01 -3.40935290e-02
-3.29168648e-01 -9.31350410e-01 -9.99668062e-01 2.98520982e-01
1.86844230e-01 -6.17621422e-01 5.85303605e-01 -1.36259258e+00
-1.06075001e+00 8.15909982e-01 5.10441959e-01 -7.94560492e-01
4.43603128e-01 3.11494797e-01 -6.34044349e-01 2.73113281e-01
-7.51814852e-03 1.39438242e-01 5.95619142e-01 -6.70886099e-01
-3.74316797e-02 -5.62533028e-02 1.89248890e-01 -2.99083795e-02
-4.71391737e-01 -3.62723440e-01 7.08537549e-02 -4.63735759e-01
1.42991066e-01 -9.89133894e-01 -2.29087681e-01 -1.85532942e-01
-4.68355685e-01 -8.62743929e-02 6.46927536e-01 -3.35813314e-01
9.75547075e-01 -2.20043993e+00 2.79461384e-01 4.90507066e-01
7.33911216e-01 2.32400354e-02 -5.27170181e-01 8.40057552e-01
-7.34082937e-01 3.89060766e-01 -9.55285504e-02 4.43323255e-01
-1.80530161e-01 -2.58923501e-01 -1.53095871e-01 1.06974208e+00
1.02779441e-01 1.24003279e+00 -1.14355552e+00 -1.72267660e-01
2.47679874e-01 4.56539929e-01 -4.95565832e-01 -3.60081047e-01
4.65061143e-02 -5.10282256e-02 -1.09360456e-01 1.93819493e-01
6.84913814e-01 -1.01588118e+00 3.80190164e-01 -3.65046412e-01
3.50401439e-02 1.64420977e-01 -1.22726715e+00 1.70872295e+00
-1.26104400e-01 8.54890406e-01 1.94609001e-01 -1.05918384e+00
5.12909830e-01 1.07918158e-01 6.73965454e-01 -4.73069757e-01
1.24516087e-02 6.64495081e-02 3.65651935e-01 3.83987874e-02
1.13096766e-01 -6.24648556e-02 5.78333102e-02 4.64311957e-01
3.94611567e-01 -2.61510424e-02 3.80510628e-01 6.63068593e-01
1.50351286e+00 -2.99965292e-01 1.45784199e-01 -1.02369010e+00
1.86571509e-01 -1.76782250e-01 -6.84994878e-03 6.49697006e-01
-1.24601193e-01 4.26110685e-01 8.78792465e-01 -4.70279634e-01
-9.81572807e-01 -1.44874001e+00 8.95642117e-02 1.01531076e+00
8.97835493e-02 -6.29619300e-01 -5.39617419e-01 -6.04652166e-01
3.58794183e-01 4.36691456e-02 -8.89574826e-01 -2.89239794e-01
-3.56090724e-01 -7.23276734e-01 8.17013621e-01 4.40273166e-01
2.84749925e-01 -7.65271902e-01 6.05976731e-02 6.37389794e-02
1.90621048e-01 -1.09127843e+00 -6.81816459e-01 1.64878428e-01
-5.64376831e-01 -1.59585106e+00 -2.91045636e-01 -8.71467233e-01
7.05491364e-01 4.17199850e-01 1.21738267e+00 4.99081910e-01
-3.64100099e-01 6.85848117e-01 -1.42615482e-01 8.38178396e-03
-4.82588589e-01 5.28627574e-01 6.00196794e-02 -2.22948775e-01
-2.64610313e-02 -1.05817342e+00 -5.52064717e-01 1.24012940e-01
-9.06658351e-01 -1.98701955e-02 2.72503853e-01 8.17190409e-01
4.75501269e-02 5.67187220e-02 6.11913204e-01 -1.08575869e+00
1.15573096e+00 -4.73615140e-01 -7.08251834e-01 1.07325718e-01
-5.73830366e-01 1.45720616e-01 8.85627985e-01 -2.53634065e-01
-2.22934693e-01 1.82172954e-02 1.48174167e-01 -3.90899241e-01
1.63171291e-01 7.72181273e-01 4.44405943e-01 -7.22250640e-01
1.07598782e+00 9.18040704e-03 2.31468648e-01 1.28565073e-01
5.50651371e-01 -1.73379824e-01 3.53404671e-01 -2.86173731e-01
9.61630046e-01 4.32277352e-01 4.40139085e-01 -1.26037240e+00
-4.91686165e-01 -4.24764127e-01 -6.96950018e-01 -2.85875887e-01
7.10652649e-01 -7.68286228e-01 -8.62318873e-01 2.62917072e-01
-8.50799441e-01 -4.26010907e-01 -3.41858491e-02 1.93633601e-01
-3.81081402e-01 5.98811924e-01 -8.90815198e-01 -1.78454041e-01
-1.99721843e-01 -7.03047991e-01 5.82846642e-01 -2.00222731e-02
-2.90994830e-02 -1.74901128e+00 1.30951524e-01 -3.32930386e-01
4.93826687e-01 4.32601839e-01 1.02284920e+00 -4.68427598e-01
-4.84554321e-01 -2.21401334e-01 -5.77339411e-01 3.01720407e-02
2.12053880e-01 -3.85624394e-02 -6.33634686e-01 -7.68138170e-01
-8.05521727e-01 -3.21316242e-01 1.11178505e+00 4.81784403e-01
9.27358150e-01 2.39696112e-02 -4.25754875e-01 8.40151310e-01
1.59945750e+00 -4.90661532e-01 4.02900070e-01 -2.34081000e-01
8.77330780e-01 1.61733046e-01 -4.74475235e-01 -6.49426132e-02
6.49422929e-02 2.12724552e-01 3.45362455e-01 -4.13980842e-01
-4.31832701e-01 -3.05945009e-01 1.25790432e-01 8.46260786e-01
-1.28812820e-01 -4.30058032e-01 -1.14547503e+00 3.92872423e-01
-1.84374452e+00 -9.12032664e-01 -4.28187370e-01 2.00664139e+00
1.17808215e-01 2.08519548e-01 2.73132235e-01 -2.05991820e-01
9.47099268e-01 4.11016762e-01 -3.93475831e-01 -1.35229081e-01
-4.09213245e-01 2.71272331e-01 1.02592134e+00 7.67230451e-01
-1.27135932e+00 1.04990649e+00 6.87772036e+00 3.87940109e-01
-1.17140353e+00 4.08415683e-02 3.49694639e-01 1.16353683e-01
-3.02942276e-01 -1.30962640e-01 -9.81542468e-02 7.64157027e-02
9.79152858e-01 -5.85976899e-01 8.37797284e-01 3.77746731e-01
-1.21876501e-01 3.01137924e-01 -1.12995446e+00 8.32649171e-01
-1.15862668e-01 -1.81806958e+00 -1.63650542e-01 7.13171586e-02
6.45885110e-01 5.91803789e-01 1.86889470e-01 5.58424555e-03
7.41627276e-01 -1.43981457e+00 3.00452948e-01 2.50867277e-01
9.61861670e-01 -7.38513768e-01 3.96552444e-01 -1.43405765e-01
-1.69139016e+00 1.61056787e-01 -3.90305996e-01 -9.51299518e-02
-2.98248917e-01 4.81929213e-01 -8.64963651e-01 7.29909420e-01
5.63811958e-01 1.07687235e+00 -8.89153659e-01 9.29151475e-01
-1.66680142e-02 6.28721476e-01 -2.28471786e-01 5.48468418e-02
3.50787431e-01 -4.44553554e-01 5.92767537e-01 1.29097998e+00
2.75043342e-02 -2.31651753e-01 3.60016853e-01 1.11548913e+00
-2.18148604e-01 2.95713902e-01 -1.22706437e+00 -6.04699552e-01
9.51094329e-02 1.36726010e+00 -1.30124688e+00 2.39460260e-01
-8.27877402e-01 9.76628661e-01 7.37929642e-01 6.96074665e-01
-6.38897419e-01 -5.56636810e-01 4.35827792e-01 3.53753120e-01
1.55437320e-01 -5.05650163e-01 -8.34107772e-02 -1.13485050e+00
-1.88745424e-01 -4.91558492e-01 6.30249619e-01 -5.38551033e-01
-1.54358006e+00 4.70082641e-01 -2.61278749e-01 -6.92152321e-01
2.12251574e-01 -8.46019506e-01 -6.27373338e-01 9.65231597e-01
-9.80656207e-01 -1.14437723e+00 -3.48261625e-01 7.33344078e-01
-1.98193178e-01 -1.28010541e-01 7.89836347e-01 2.26861715e-01
-2.48906389e-01 5.57201385e-01 -9.72832218e-02 4.94535774e-01
5.51178753e-01 -1.56049979e+00 8.62916529e-01 8.73913348e-01
3.49047273e-01 5.33956468e-01 6.93972588e-01 -7.46377349e-01
-1.62189186e+00 -1.18463588e+00 3.72487873e-01 -2.50525266e-01
1.41203964e+00 -7.20725000e-01 -1.06260157e+00 9.47074831e-01
4.98458087e-01 4.06808019e-01 4.27758992e-01 4.00017530e-01
-6.18242860e-01 1.63734257e-01 -7.18598306e-01 5.74178219e-01
1.48874199e+00 -8.59604180e-01 1.40580311e-01 6.74099267e-01
7.26612031e-01 -2.74587631e-01 -1.03064752e+00 2.33378410e-01
2.76413321e-01 -9.34082210e-01 8.74850750e-01 -7.30410814e-01
1.62588850e-01 -5.99740930e-02 1.33886993e-01 -1.74369049e+00
-7.30232120e-01 -8.03285062e-01 2.52800554e-01 4.59683478e-01
6.22883379e-01 -1.08077002e+00 8.52076769e-01 -2.56703436e-01
-4.41273861e-02 -4.09154475e-01 -8.84765506e-01 -8.98481011e-01
4.02976066e-01 -6.79901764e-02 2.13157400e-01 1.23152888e+00
1.61360398e-01 5.81341267e-01 1.28860295e-01 4.05969054e-01
5.26587784e-01 1.76510304e-01 4.73403424e-01 -1.58871639e+00
-1.33497700e-01 -8.97194326e-01 -8.94943655e-01 -5.00403821e-01
6.64316416e-01 -1.65206575e+00 -5.66506684e-01 -1.51549840e+00
-2.03206930e-02 -1.70941800e-02 -2.98927814e-01 2.53413975e-01
8.47882554e-02 3.31631690e-01 -5.82935428e-03 -2.69672573e-01
-6.78900540e-01 6.74471483e-02 1.24623728e+00 -1.54024005e-01
-1.62984475e-01 -2.02149078e-01 -5.83737552e-01 4.65835929e-01
8.71453881e-01 -8.48325044e-02 -7.03423679e-01 -1.49672758e-02
8.02265763e-01 1.99802592e-02 6.92077935e-01 -1.10437644e+00
2.44654939e-01 6.33265302e-02 3.26793790e-01 -2.11888343e-01
-9.43468213e-02 -6.37238145e-01 3.29316318e-01 5.99182844e-01
-2.29781240e-01 4.76167142e-01 4.38521266e-01 8.99514258e-01
5.55425137e-02 3.54940325e-01 9.18508410e-01 -2.43929639e-01
-5.66827357e-01 6.96052015e-01 -1.49767578e-01 3.69989067e-01
7.98129678e-01 -1.78892687e-01 -6.74930751e-01 -6.00190401e-01
-1.02269948e+00 1.37924533e-02 5.99348307e-01 3.01159084e-01
5.81462741e-01 -1.39350665e+00 -7.76879609e-01 3.01383287e-01
4.42143828e-02 -6.09544218e-01 4.71688174e-02 1.10149455e+00
-7.39197075e-01 3.05344254e-01 -3.10844868e-01 -5.01576364e-01
-1.16101897e+00 5.66626549e-01 9.27497029e-01 5.29881492e-02
-9.57731962e-01 6.93087757e-01 1.59684107e-01 -3.98487955e-01
-8.15695617e-03 -4.21264887e-01 3.08030367e-01 -1.03985980e-01
-1.07171771e-03 3.20187479e-01 1.80809915e-01 -4.29085821e-01
-4.27205056e-01 2.22878113e-01 2.08134383e-01 2.80780822e-01
1.35253239e+00 1.91726923e-01 -4.43579644e-01 4.77950215e-01
1.30620968e+00 6.45197481e-02 -8.95744979e-01 -1.48211747e-01
3.25702839e-02 -1.72729760e-01 -5.96342385e-02 -3.54391158e-01
-1.06819344e+00 6.45044386e-01 3.37476492e-01 1.09448326e+00
7.73475230e-01 8.77732635e-02 2.14009345e-01 4.05929625e-01
1.91522256e-01 -8.69793713e-01 1.44302517e-01 6.97843969e-01
1.03039908e+00 -9.54157889e-01 6.31126165e-02 -7.28141189e-01
5.56160361e-02 1.33437800e+00 6.40412152e-01 -6.26800895e-01
1.12820613e+00 3.84035319e-01 -3.05109978e-01 -6.75126672e-01
-5.83338022e-01 -2.36645952e-01 4.68028992e-01 6.50646389e-01
7.41168976e-01 3.02759528e-01 2.47342903e-02 5.03024571e-02
-2.51034349e-01 -4.94976908e-01 7.25365460e-01 3.38302314e-01
-3.50941658e-01 -6.88291132e-01 5.03902510e-02 6.59928381e-01
-1.93997949e-01 -3.23374152e-01 -8.65610242e-01 1.05150163e+00
-4.52668518e-01 6.80767238e-01 5.57600074e-02 -4.99596179e-01
1.52213171e-01 -7.24844187e-02 7.30511487e-01 -6.41737103e-01
-4.90001202e-01 -3.42168301e-01 1.48416698e-01 -6.95667565e-01
-3.53323996e-01 -4.00127232e-01 -1.11525404e+00 -6.75169289e-01
-2.64159322e-01 1.19485833e-01 1.53168529e-01 4.93484408e-01
4.13877964e-01 7.48960018e-01 9.74268392e-02 -4.64036524e-01
-2.61066645e-01 -9.24598634e-01 -8.95782053e-01 5.96792579e-01
5.09989679e-01 -4.44518358e-01 -5.55859149e-01 -3.67041260e-01]
|
[6.872424602508545, 6.1190948486328125]
|
22faaf79-b397-4648-8124-26f0ff1810dd
|
robust-brain-age-estimation-via-regression
|
2306.05514
| null |
https://arxiv.org/abs/2306.05514v1
|
https://arxiv.org/pdf/2306.05514v1.pdf
|
Robust Brain Age Estimation via Regression Models and MRI-derived Features
|
The determination of biological brain age is a crucial biomarker in the assessment of neurological disorders and understanding of the morphological changes that occur during aging. Various machine learning models have been proposed for estimating brain age through Magnetic Resonance Imaging (MRI) of healthy controls. However, developing a robust brain age estimation (BAE) framework has been challenging due to the selection of appropriate MRI-derived features and the high cost of MRI acquisition. In this study, we present a novel BAE framework using the Open Big Healthy Brain (OpenBHB) dataset, which is a new multi-site and publicly available benchmark dataset that includes region-wise feature metrics derived from T1-weighted (T1-w) brain MRI scans of 3965 healthy controls aged between 6 to 86 years. Our approach integrates three different MRI-derived region-wise features and different regression models, resulting in a highly accurate brain age estimation with a Mean Absolute Error (MAE) of 3.25 years, demonstrating the framework's robustness. We also analyze our model's regression-based performance on gender-wise (male and female) healthy test groups. The proposed BAE framework provides a new approach for estimating brain age, which has important implications for the understanding of neurological disorders and age-related brain changes.
|
['Imdad Ullah Khan', 'Murray Patterson', 'Shafiq Alam', 'Sarwan Ali', 'Usama Sardar', 'Mansoor Ahmed']
|
2023-06-08
| null | null | null | null |
['age-estimation', 'age-estimation']
|
['computer-vision', 'miscellaneous']
|
[-1.85866114e-02 -2.66175270e-01 1.49716094e-01 -5.55422843e-01
-6.84069037e-01 -4.08062674e-02 5.93094230e-01 5.71231663e-01
-1.02414489e+00 7.28730679e-01 6.43899888e-02 2.38615694e-03
-1.43484920e-01 -5.60214579e-01 -3.30451965e-01 -7.74953544e-01
-6.96865261e-01 4.07771021e-01 2.14906946e-01 1.48861259e-01
5.10873675e-01 4.27138388e-01 -1.43015134e+00 -3.89001936e-01
1.12641203e+00 1.09878600e+00 2.35937342e-01 5.47047913e-01
3.34804237e-01 1.43651888e-01 -4.06448603e-01 -3.84399354e-01
2.27823816e-02 -7.02296719e-02 -6.82598054e-01 -4.26189005e-01
6.06883466e-01 -3.15187603e-01 -9.91667807e-02 8.72012973e-01
9.64892745e-01 -1.50023252e-02 9.21561897e-01 -1.22059858e+00
-5.21367908e-01 3.92644614e-01 -7.37343609e-01 7.76521683e-01
-1.34850353e-01 3.73849943e-02 4.61031973e-01 -7.49128759e-01
3.40022713e-01 8.67401004e-01 7.48394489e-01 8.40315223e-01
-8.86718094e-01 -9.25445795e-01 -1.85915217e-01 8.46938372e-01
-1.23756850e+00 -3.88717175e-01 3.51905376e-01 -8.91713381e-01
5.26625752e-01 2.67194696e-02 8.37372124e-01 8.66966844e-01
8.77637267e-01 1.84130490e-01 1.70380104e+00 -2.18614802e-01
3.25023532e-01 -7.26679265e-01 4.30064291e-01 5.67984343e-01
4.58976686e-01 6.68129474e-02 -5.26473165e-01 -1.34995565e-01
5.82521617e-01 -1.95352256e-01 5.55130951e-02 -3.69189620e-01
-1.52746749e+00 6.06071293e-01 5.36645353e-01 3.42620403e-01
-4.17823374e-01 7.10839778e-02 5.22212863e-01 1.34465739e-01
5.73228061e-01 -2.49886373e-03 -4.79208708e-01 7.71323740e-02
-1.18935537e+00 3.73305827e-01 -4.62382212e-02 2.67124146e-01
1.72670737e-01 1.65847950e-02 -4.68686521e-02 1.02317560e+00
2.29209125e-01 4.83938992e-01 9.14534450e-01 -7.19374835e-01
8.89201909e-02 2.46183783e-01 -3.45101506e-01 -6.48422062e-01
-9.11493003e-01 -5.16114831e-01 -9.69219327e-01 4.06627506e-01
7.95829654e-01 8.33037794e-02 -8.37728441e-01 1.91487873e+00
3.67934823e-01 -2.84614980e-01 -4.37356561e-01 9.29439664e-01
5.74963152e-01 -3.62109356e-02 6.00732625e-01 -2.18190804e-01
1.78815126e+00 -5.99784553e-01 -1.95662171e-01 -3.65089029e-01
3.53999496e-01 -2.37124965e-01 6.30720496e-01 3.34178358e-01
-1.11181641e+00 -3.62815708e-01 -9.84554410e-01 -5.33028953e-02
-3.98355126e-01 -1.08876536e-02 5.90097010e-01 7.65016198e-01
-1.05083370e+00 5.44167161e-01 -1.04969740e+00 -4.24619168e-01
5.70677519e-01 3.72539192e-01 -8.67268384e-01 -6.95857182e-02
-9.91220772e-01 1.17768312e+00 1.43763989e-01 1.78990215e-02
-8.91748726e-01 -1.24942327e+00 -5.61133027e-01 -2.91698158e-01
-2.74861723e-01 -8.49190116e-01 6.88359439e-01 -3.41909260e-01
-6.51684284e-01 1.20235157e+00 -1.25570089e-01 -5.78303218e-01
6.36888802e-01 -2.84430146e-01 -5.05401611e-01 3.10089946e-01
3.21495831e-01 7.59629607e-01 7.43858099e-01 -5.85303724e-01
-2.09882125e-01 -1.35516167e+00 -5.38631260e-01 -2.39369556e-01
-2.93861032e-01 5.87382495e-01 4.62691247e-01 -7.98880875e-01
2.12080821e-01 -7.38399148e-01 -2.70671457e-01 2.36933380e-01
7.24384412e-02 -1.17818169e-01 2.31464714e-01 -1.54015541e+00
1.11224008e+00 -1.77422893e+00 1.76741585e-01 9.14648101e-02
7.94006050e-01 -1.94718882e-01 1.10040620e-01 -1.49279952e-01
-3.93270850e-01 6.72216490e-02 -3.52881700e-01 -1.35108773e-02
-2.97677845e-01 -6.07667863e-01 4.26307291e-01 7.07035244e-01
-9.59485993e-02 6.40372455e-01 -6.70824528e-01 -6.52721226e-01
-4.41545509e-02 4.82929200e-01 -1.78422570e-01 2.21551247e-02
6.27419770e-01 6.86433017e-01 -1.55825227e-01 7.10181296e-01
5.62069654e-01 2.74550945e-01 -2.11561695e-01 -2.50384450e-01
-1.36302173e-01 -4.83428866e-01 -4.41531837e-01 1.60827184e+00
-1.97307050e-01 4.26926374e-01 1.30304053e-01 -9.02619720e-01
1.01780570e+00 6.98316991e-02 6.43664241e-01 -8.35193157e-01
3.78163099e-01 3.65614057e-01 5.52198052e-01 -2.40612924e-01
-3.00243169e-01 -3.02863181e-01 8.95318761e-02 5.05297899e-01
1.60864756e-01 1.78416237e-01 4.87730265e-01 -5.31706586e-02
1.24963236e+00 -2.35350594e-01 3.70831847e-01 -6.69231713e-01
8.40020835e-01 -5.79852164e-01 7.51076996e-01 3.01558733e-01
-1.07842517e+00 4.30276334e-01 3.46474230e-01 -5.58569014e-01
-1.21874309e+00 -1.32437003e+00 -5.58201492e-01 8.38363528e-01
-6.02885425e-01 -2.77893662e-01 -9.79597807e-01 -4.94738996e-01
2.40045339e-02 2.88144618e-01 -8.97623241e-01 -2.30844930e-01
-5.20437956e-01 -1.29925418e+00 5.33984303e-01 7.55273044e-01
5.25527120e-01 -6.60751998e-01 -5.69630623e-01 -2.99636349e-02
-9.03567374e-02 -8.60198200e-01 -4.64793593e-01 -9.05058086e-02
-1.27133441e+00 -1.22225058e+00 -1.48822594e+00 -6.23587132e-01
8.20106924e-01 -2.75146335e-01 9.04656470e-01 8.07111338e-02
-7.54511297e-01 3.13268036e-01 -1.62469432e-01 -4.19432282e-01
-1.55053809e-01 2.30446458e-01 3.97538841e-01 -2.41958424e-01
1.94714993e-01 -9.21093166e-01 -1.10850549e+00 2.33309731e-01
-5.43467104e-01 -1.49300292e-01 8.32902193e-01 5.56232214e-01
3.34541529e-01 -4.18337792e-01 9.88135874e-01 -3.08549762e-01
4.82043535e-01 -4.40126866e-01 -2.90094137e-01 2.21476078e-01
-9.43921566e-01 1.60369948e-02 1.62954435e-01 -3.60963911e-01
-7.46113598e-01 -3.79203022e-01 -1.85346976e-01 3.33019048e-01
-1.75585285e-01 2.50135988e-01 -1.97212487e-01 -2.81233668e-01
5.42072058e-01 8.19993988e-02 4.61618155e-01 -5.70675075e-01
-3.35458294e-02 4.81249183e-01 8.34881663e-01 -7.27620065e-01
4.70005035e-01 1.81070685e-01 3.86670619e-01 -8.10583174e-01
-4.57318097e-01 -1.87071517e-01 -1.05472791e+00 -5.31162143e-01
9.55323815e-01 -8.90062511e-01 -4.03633058e-01 1.09469390e+00
-5.98917484e-01 -3.00539434e-01 5.33972204e-01 6.33012474e-01
-4.45449352e-01 6.37576699e-01 -6.36802197e-01 -4.59607691e-01
-1.12692940e+00 -1.04187119e+00 7.83261597e-01 1.44824445e-01
-3.66604179e-01 -9.03260112e-01 -1.83300711e-02 6.51698172e-01
6.26188755e-01 6.31937563e-01 1.38757837e+00 -4.30006206e-01
1.60173133e-01 -3.40513468e-01 -3.62913251e-01 2.88807660e-01
-2.22475663e-01 -1.79402485e-01 -3.72877210e-01 -4.10079211e-01
-1.24174826e-01 -6.01027869e-02 7.68498480e-01 6.71406329e-01
1.08283901e+00 3.48158062e-01 -4.47797775e-02 2.00195417e-01
1.29302204e+00 1.55729085e-01 6.57543421e-01 6.19723320e-01
4.79407430e-01 7.10131884e-01 3.86141688e-01 3.99860889e-01
5.90228498e-01 4.56177294e-01 2.68176556e-01 1.54751882e-01
-2.23964572e-01 5.55578470e-01 3.67929310e-01 8.82912099e-01
-4.50651020e-01 6.92432940e-01 -1.27373350e+00 7.21684635e-01
-1.20099962e+00 -7.97320843e-01 -5.00046313e-01 2.45725465e+00
8.75409901e-01 1.44322589e-01 5.33778727e-01 3.44920486e-01
1.00596869e+00 -1.17283165e-02 -6.33565843e-01 -3.30415279e-01
-3.40197794e-02 5.47530770e-01 4.61634099e-01 4.50921692e-02
-8.62500072e-01 1.59876391e-01 7.01200819e+00 3.56092274e-01
-1.17249095e+00 5.14328241e-01 7.26435184e-01 -1.84457436e-01
2.80669868e-01 -2.94782311e-01 -4.74922180e-01 5.22930980e-01
1.20682633e+00 -5.57467759e-01 2.82661676e-01 6.52859628e-01
2.39448756e-01 -3.74088794e-01 -8.74197543e-01 8.11303854e-01
2.49834880e-01 -6.44414723e-01 -2.59081066e-01 9.89290252e-02
2.62154698e-01 1.66010391e-02 -2.31724475e-02 1.05070986e-01
-4.01425153e-01 -8.85454714e-01 8.86805832e-01 9.40102637e-01
1.02361941e+00 -7.85807788e-01 6.34645164e-01 7.45151658e-03
-1.07408154e+00 -2.68149793e-01 -2.29548588e-02 -1.60226092e-01
1.09743752e-01 1.03435409e+00 -3.78219604e-01 1.88914865e-01
9.57365394e-01 4.15244758e-01 -1.26571894e+00 1.34237218e+00
5.91216274e-02 3.20528895e-01 1.39850661e-01 4.57477301e-01
-3.36396754e-01 -1.41407236e-01 2.86715627e-01 7.17719972e-01
5.58985412e-01 -2.24920750e-01 -3.69946867e-01 6.00198567e-01
3.07535738e-01 3.37410867e-01 3.59434895e-02 8.95453021e-02
3.29544067e-01 1.51911962e+00 -1.01144052e+00 -7.79206008e-02
-2.72454947e-01 4.36677098e-01 2.49410719e-01 -2.18597025e-01
-5.96731484e-01 -4.36201334e-01 5.58570206e-01 3.69123191e-01
-2.28794619e-01 -6.07796967e-01 -3.73185992e-01 -9.09553587e-01
3.18887122e-02 -6.79500937e-01 3.36836696e-01 -7.52135634e-01
-1.43683636e+00 3.36474568e-01 2.15226576e-01 -4.81730491e-01
-7.69767389e-02 -4.99996275e-01 -7.52313852e-01 8.75748158e-01
-1.13434863e+00 -1.17468178e+00 -4.68567431e-01 3.15325737e-01
1.86656535e-01 -2.44032130e-01 7.30949640e-01 5.91752410e-01
-8.08545470e-01 4.54870880e-01 -9.34034511e-02 1.73218548e-01
9.16672051e-01 -1.39581954e+00 2.96449780e-01 9.00833607e-01
-6.93710446e-01 8.37676227e-01 6.59694552e-01 -7.28473902e-01
-8.78079832e-01 -8.74361396e-01 1.03177893e+00 -4.59733129e-01
7.71900356e-01 -2.28112444e-01 -7.26233542e-01 3.02187234e-01
-2.57263333e-01 -2.95760762e-02 1.03326654e+00 4.85422164e-02
-3.93089771e-01 -3.13347071e-01 -1.41788912e+00 4.08802837e-01
1.07061327e+00 -1.79689839e-01 -7.17199802e-01 -5.17568290e-02
1.66730478e-01 1.76620334e-01 -1.64635324e+00 6.55827940e-01
1.19743538e+00 -1.03204799e+00 1.17894042e+00 -4.37015861e-01
3.03782254e-01 -9.47760195e-02 2.46258184e-01 -1.15464556e+00
-3.57004732e-01 2.43833721e-01 -1.44026965e-01 1.34410095e+00
-3.67839262e-02 -6.00329399e-01 4.80999559e-01 8.67726147e-01
-4.50429171e-02 -9.33112562e-01 -1.10119569e+00 -9.27244782e-01
6.29343987e-01 -2.64450729e-01 5.76500833e-01 4.89344090e-01
-3.55429083e-01 -1.49876669e-01 1.56529188e-01 2.48497389e-02
1.03619611e+00 -4.62797403e-01 3.10519040e-01 -1.76675248e+00
4.75567192e-01 -6.30205512e-01 -9.45861280e-01 3.89976442e-01
2.90477008e-01 -8.66458774e-01 -3.01840991e-01 -1.48422515e+00
6.21327937e-01 -2.94905812e-01 -6.43988907e-01 1.92372546e-01
-1.54002368e-01 4.03243482e-01 -3.68918516e-02 -2.57944502e-03
-1.48680136e-01 2.97168076e-01 9.35384870e-01 -1.28206506e-01
3.60688120e-01 -2.58410126e-01 -5.30475020e-01 5.77499449e-01
1.04735923e+00 -4.60460246e-01 -6.09086379e-02 -1.10299177e-01
-7.75748342e-02 -3.56567532e-01 5.49117208e-01 -1.52255476e+00
-1.32054716e-01 1.46873882e-02 9.58654523e-01 -3.54904473e-01
2.19024401e-02 -2.90216148e-01 1.59673527e-01 8.59746993e-01
-4.56896648e-02 4.72344816e-01 -2.28275150e-01 2.21292615e-01
1.52296439e-01 -1.79818407e-01 1.01738405e+00 -8.71984139e-02
-6.37504637e-01 5.94533026e-01 -4.53307480e-01 -4.77000372e-03
1.07443333e+00 -1.85101554e-01 -2.70390749e-01 2.38983929e-01
-9.15221572e-01 7.20315576e-02 5.19257784e-01 4.93442953e-01
3.78201991e-01 -1.31901312e+00 -8.30900550e-01 -4.32473682e-02
3.12529951e-01 -6.70434892e-01 5.47011852e-01 1.48652756e+00
-5.98788023e-01 2.12262124e-01 -1.08400261e+00 -3.72272968e-01
-1.51019144e+00 4.20641929e-01 1.97032213e-01 -2.72075266e-01
-4.03628737e-01 6.09254301e-01 -3.91475037e-02 -5.72615117e-02
-2.43743006e-02 -1.90783441e-01 -4.60064143e-01 3.75009984e-01
8.53116453e-01 9.23309326e-01 2.33035281e-01 -9.03136194e-01
-5.32690644e-01 6.32412851e-01 -2.55212814e-01 -1.00648880e-01
1.60327685e+00 -2.61666000e-01 -5.78403890e-01 4.65216726e-01
8.43226850e-01 -2.47072294e-01 -7.99687445e-01 7.18440786e-02
3.39322090e-01 -1.40253067e-01 2.43769944e-01 -7.47856438e-01
-1.30101919e+00 9.33118343e-01 1.32953942e+00 -3.81704211e-01
1.02078915e+00 -1.38537586e-01 1.05230665e+00 -1.84260592e-01
6.75435185e-01 -1.10795248e+00 -1.49365619e-01 8.68090391e-02
9.73029733e-01 -9.99795079e-01 2.16190726e-01 3.00613255e-03
-1.25831127e-01 1.10990250e+00 6.42970383e-01 4.41923514e-02
6.25381947e-01 -2.34392866e-01 -1.59301106e-02 -1.06438316e-01
-2.81200916e-01 -4.07846272e-02 3.36018145e-01 8.83400440e-01
6.65021837e-01 1.11816898e-01 -1.22331166e+00 1.11222386e+00
-5.18658638e-01 1.21967770e-01 2.29343519e-01 8.22983146e-01
-4.74910289e-01 -1.30745566e+00 -5.43855906e-01 1.02019131e+00
-7.66013086e-01 9.27854106e-02 -1.52544245e-01 5.72998047e-01
2.37152338e-01 5.86201906e-01 -1.59900054e-01 8.99813604e-03
-6.58497959e-02 5.08270204e-01 8.13997209e-01 -2.75184870e-01
-3.17563742e-01 -4.84225065e-01 -5.20143844e-02 -4.12724167e-01
-4.51654583e-01 -1.08942819e+00 -1.13835752e+00 -2.86039531e-01
5.02862558e-02 -1.84819207e-01 1.04368007e+00 9.52831984e-01
2.97881782e-01 4.37688142e-01 2.95292944e-01 -7.50309467e-01
-1.88684925e-01 -1.11252654e+00 -8.07227015e-01 2.65461564e-01
-4.27745171e-02 -1.13883650e+00 -1.54419407e-01 6.80187717e-02]
|
[14.08324146270752, -1.5273454189300537]
|
e53537d4-c9ef-4648-8407-627716d30dce
|
deep-snapshot-hdr-reconstruction-based-on-the
|
2105.05824
| null |
https://arxiv.org/abs/2105.05824v1
|
https://arxiv.org/pdf/2105.05824v1.pdf
|
Deep Snapshot HDR Reconstruction Based on the Polarization Camera
|
The recent development of the on-chip micro-polarizer technology has made it possible to acquire four spatially aligned and temporally synchronized polarization images with the same ease of operation as a conventional camera. In this paper, we investigate the use of this sensor technology in high-dynamic-range (HDR) imaging. Specifically, observing that natural light can be attenuated differently by varying the orientation of the polarization filter, we treat the multiple images captured by the polarization camera as a set captured under different exposure times. In our approach, we first study the relationship among polarizer orientation, degree and angle of polarization of light to the exposure time of a pixel in the polarization image. Subsequently, we propose a deep snapshot HDR reconstruction framework to recover an HDR image using the polarization images. A polarized HDR dataset is created to train and evaluate our approach. We demonstrate that our approach performs favorably against state-of-the-art HDR reconstruction algorithms.
|
['Hong Zhang', 'Kangkang Hu', 'Xuesong Wu', 'Juiwen Ting']
|
2021-05-12
| null | null | null | null |
['hdr-reconstruction']
|
['computer-vision']
|
[ 4.18093294e-01 -4.06521171e-01 3.31033349e-01 -2.14344189e-01
-3.97033125e-01 -7.08058417e-01 3.55869770e-01 -7.54623830e-01
-3.67842793e-01 6.56298876e-01 4.27404158e-02 -7.21249580e-02
1.03875250e-01 -6.61815405e-01 -7.34622538e-01 -1.19478166e+00
3.01200479e-01 3.18516076e-01 8.95240605e-02 -4.99419458e-02
2.33190760e-01 6.38033211e-01 -1.62915897e+00 -1.94132943e-02
4.48110670e-01 7.95085907e-01 2.49116614e-01 6.71587646e-01
5.81309378e-01 5.00360668e-01 -3.91403109e-01 -5.70232689e-04
4.55048561e-01 -3.79695803e-01 -5.10338962e-01 1.83666125e-01
6.56056166e-01 -6.79952681e-01 -6.92435324e-01 9.59957898e-01
5.53015172e-01 -1.27220556e-01 2.62162566e-01 -5.67351401e-01
-5.65479338e-01 -3.53766158e-02 -8.41776311e-01 1.99415699e-01
5.86917222e-01 3.33329201e-01 4.53765959e-01 -6.22429132e-01
1.02224076e+00 6.00086749e-01 3.56862634e-01 3.58779043e-01
-1.55494916e+00 -3.91784906e-01 -5.72493315e-01 9.08958614e-02
-1.13186848e+00 -6.57466590e-01 9.38662052e-01 -4.12906408e-01
6.91425800e-01 -4.48843203e-02 7.42722988e-01 9.71418381e-01
4.05553430e-01 -1.97609484e-01 1.96894753e+00 -5.19548714e-01
5.12200408e-04 -1.57864004e-01 1.03624403e-01 3.49877328e-01
3.00403655e-01 5.16121686e-01 -7.60455132e-01 1.29279524e-01
9.59696293e-01 -7.09554926e-02 -7.70710051e-01 -3.78321886e-01
-1.32130861e+00 1.08749889e-01 1.52287126e-01 1.72185853e-01
-3.42404932e-01 -7.30539635e-02 -3.59525174e-01 2.99561143e-01
3.06507021e-01 7.01568127e-01 5.68048656e-02 1.82532117e-01
-7.83675849e-01 2.75342148e-02 6.02303565e-01 4.38150764e-01
9.14480209e-01 -1.35287032e-01 3.26214842e-02 5.67564428e-01
3.25907916e-01 1.07357514e+00 6.38335645e-02 -1.25471771e+00
7.01959357e-02 -7.07529411e-02 3.86262655e-01 -8.71562183e-01
-2.60886788e-01 -1.34156138e-01 -6.81288719e-01 3.09859693e-01
3.99562210e-01 -2.08944872e-01 -9.57667410e-01 1.51124382e+00
2.80938506e-01 2.04431906e-01 2.64319450e-01 1.27123070e+00
3.15168142e-01 8.77374828e-01 -5.80158174e-01 -7.22913980e-01
1.28856468e+00 -4.86092657e-01 -8.43018532e-01 -1.49852991e-01
-1.61007151e-01 -1.02504671e+00 7.22164631e-01 5.09251475e-01
-1.01847374e+00 -1.34134904e-01 -1.19102669e+00 -7.60344490e-02
2.47257411e-01 -3.82092036e-02 2.67354399e-01 4.32697773e-01
-9.80068624e-01 5.25446951e-01 -7.34203398e-01 -3.74601841e-01
1.34764360e-02 1.26212522e-01 -4.76366848e-01 -5.57030618e-01
-8.50724518e-01 7.55606472e-01 4.66361130e-03 2.43731011e-02
-7.84243762e-01 -7.62258351e-01 -1.73045889e-01 -4.02653217e-01
-1.39708742e-02 -7.25123227e-01 8.66998971e-01 -4.67461854e-01
-2.11000323e+00 1.34797156e+00 -2.40284696e-01 -1.84141278e-01
2.11171925e-01 -1.69878662e-01 -5.74352384e-01 5.65572441e-01
-2.26019219e-01 1.31373420e-01 8.83394182e-01 -1.39276969e+00
-3.37646753e-01 -5.61037004e-01 -2.06616893e-01 1.39165791e-02
2.82824785e-01 7.75206611e-02 -4.14831638e-01 2.07842588e-01
3.99548531e-01 -1.24414933e+00 4.73194458e-02 -4.15942639e-01
-5.47428370e-01 9.02830422e-01 1.14324582e+00 -4.04303223e-01
7.03172028e-01 -2.21828675e+00 2.07782909e-01 -1.49948940e-01
2.30718419e-01 2.35126540e-01 5.98398328e-04 3.83621931e-01
-2.03181431e-01 -3.50491613e-01 -2.24859789e-01 -3.09870150e-02
-5.45281947e-01 2.47157421e-02 -6.03120983e-01 9.81218994e-01
-2.55848080e-01 5.47109544e-01 -5.39762557e-01 -2.04934571e-02
2.48318121e-01 8.08920205e-01 -4.66088772e-01 7.17881501e-01
5.25772572e-02 1.24278557e+00 5.35323732e-02 6.44400060e-01
1.22175133e+00 -2.91392565e-01 6.33315086e-01 -7.33907998e-01
-7.67747581e-01 1.66503266e-01 -8.98791015e-01 1.32417202e+00
-2.56760806e-01 9.97930944e-01 8.99214372e-02 -6.01430178e-01
9.14419472e-01 2.87024230e-01 5.99512994e-01 -1.27641237e+00
-6.80080876e-02 4.23763603e-01 -8.20167363e-02 -7.03635216e-01
6.70725822e-01 -5.18811047e-01 1.15938030e-01 4.52461511e-01
2.05437347e-04 -1.44151255e-01 1.14594258e-01 -3.24145138e-01
9.93687630e-01 1.69701409e-02 1.12342954e-01 -1.85892567e-01
1.93387344e-01 -1.83438465e-01 2.96381921e-01 6.12863302e-01
-2.85837222e-02 8.78086746e-01 3.21973622e-01 -5.14801025e-01
-1.53743398e+00 -1.21880555e+00 -4.72962230e-01 4.05092478e-01
5.26088059e-01 9.05684605e-02 -2.71755666e-01 2.20575407e-01
-1.90759420e-01 2.07659483e-01 -1.35569230e-01 3.41889381e-01
-9.89255130e-01 -1.16119993e+00 1.28369346e-01 -1.96656704e-01
7.54879653e-01 -7.00724602e-01 -9.63095009e-01 -5.71822859e-02
-2.16800317e-01 -1.36691320e+00 1.93479463e-01 9.72615480e-02
-5.99956572e-01 -1.15002739e+00 -5.49865186e-01 -3.00773531e-01
5.33276916e-01 6.41439676e-01 1.22306073e+00 -6.64999008e-01
-2.25613073e-01 5.76411188e-01 -1.88165307e-01 1.83023453e-01
-8.47656354e-02 -2.62106806e-01 2.14540605e-02 1.95818424e-01
4.39104550e-02 -9.78801250e-01 -9.37138379e-01 2.71829098e-01
-9.62081373e-01 2.59608716e-01 4.76195335e-01 4.90600735e-01
1.02391469e+00 -6.82817027e-02 -2.75860459e-01 -9.19583976e-01
1.72145545e-01 -2.33657181e-01 -1.02034080e+00 2.63072401e-02
-5.02852023e-01 -2.55132262e-02 5.56216419e-01 -1.91458598e-01
-1.25510442e+00 -1.57588348e-02 4.75418717e-02 -3.36284041e-01
-2.54487008e-01 1.97491691e-01 -1.10183708e-01 -3.32100362e-01
3.31323802e-01 2.80201495e-01 -5.20194955e-02 -3.08471054e-01
2.08762944e-01 5.03763199e-01 9.10991848e-01 -3.89849812e-01
9.57989693e-01 1.15107548e+00 4.42860216e-01 -1.10118544e+00
-8.08905840e-01 -3.29077393e-01 -2.82532245e-01 -3.93295318e-01
9.48948383e-01 -1.10140634e+00 -8.57559860e-01 1.04388916e+00
-1.04642093e+00 -2.99503922e-01 -1.87826771e-02 5.20318747e-01
-5.09516239e-01 6.38470873e-02 -5.96391857e-01 -4.52398986e-01
-1.12680353e-01 -1.11244667e+00 1.24096882e+00 5.77365041e-01
3.84749681e-01 -6.82323396e-01 6.39912903e-01 4.47199851e-01
5.18493652e-01 3.01210731e-01 5.83729327e-01 4.22244698e-01
-1.18591678e+00 1.19044840e-01 -1.97141811e-01 1.38051182e-01
-1.08298659e-01 1.41933024e-01 -1.17644548e+00 -3.63128692e-01
4.71805602e-01 -1.24999762e-01 7.45456219e-01 7.38991678e-01
9.15276110e-01 -1.38734952e-02 -3.06764811e-01 1.47752154e+00
2.09535432e+00 1.94758952e-01 1.28678286e+00 5.20501375e-01
8.82533133e-01 4.24106121e-01 5.46367347e-01 2.82470584e-01
1.28541082e-01 1.11463046e+00 2.51904160e-01 -1.90100566e-01
-8.40315744e-02 1.12204552e-01 3.76949877e-01 7.34516442e-01
-3.28250289e-01 -5.16892731e-01 -7.99870253e-01 1.77704096e-01
-1.29697335e+00 -1.14921820e+00 -4.24833536e-01 2.20953679e+00
6.47530675e-01 -5.53507507e-01 -2.86415726e-01 -3.13661575e-01
7.12571919e-01 6.63674891e-01 -3.41833025e-01 1.91344935e-02
-5.53938389e-01 2.37329260e-01 8.23885143e-01 6.02507532e-01
-8.93682182e-01 5.96085191e-01 6.84466410e+00 -8.65749642e-02
-2.02722955e+00 -8.31005052e-02 3.56118441e-01 -2.76370823e-01
-4.95219231e-01 1.56278476e-01 -7.94414580e-01 7.41554081e-01
7.73033798e-01 1.10114425e-01 7.46973515e-01 2.07998872e-01
3.19830537e-01 -3.92361283e-01 -1.01756418e+00 1.23302865e+00
1.35959744e-01 -1.10936773e+00 -2.71586210e-01 3.05427998e-01
7.47430801e-01 2.17230484e-01 5.81454575e-01 -5.29762030e-01
-8.40849280e-02 -7.31788456e-01 3.23330671e-01 7.07281530e-01
1.20933449e+00 -2.80458689e-01 3.77273470e-01 -9.77633968e-02
-5.90665221e-01 -6.28970005e-03 -2.24249542e-01 5.92297576e-02
4.56925064e-01 1.08624697e+00 -6.06183052e-01 4.34762955e-01
8.57519567e-01 6.59446836e-01 -1.12602748e-01 5.48369229e-01
-3.19556564e-01 4.41586107e-01 -1.31738245e-01 8.29019845e-01
-4.10296440e-01 -7.16564000e-01 8.26552927e-01 6.49778306e-01
6.89399302e-01 4.79166269e-01 -3.12205464e-01 8.95959377e-01
-1.16002284e-01 -7.90202558e-01 -7.82076061e-01 -1.48281455e-03
4.01429892e-01 1.40431917e+00 -3.22838277e-01 1.53507218e-01
-4.00873810e-01 9.30402339e-01 9.74608883e-02 7.41543949e-01
-7.13569820e-01 -1.04410134e-01 8.26665342e-01 5.08640826e-01
4.01761264e-01 -4.11856264e-01 3.74382921e-02 -1.57663822e+00
3.24302800e-02 -7.64021873e-01 -1.83396652e-01 -1.19837987e+00
-1.09538901e+00 4.40768957e-01 -1.95600271e-01 -1.17353141e+00
1.47197828e-01 -6.90938771e-01 -2.97045648e-01 1.10172868e+00
-2.03206944e+00 -8.35902512e-01 -5.56122065e-01 5.16451776e-01
-2.85711318e-01 1.99906170e-01 6.40399039e-01 3.71890754e-01
-5.72819471e-01 -2.94240087e-01 5.53937018e-01 -1.12490118e-01
1.20339096e+00 -9.39723969e-01 -1.19457811e-01 1.24068749e+00
-3.17329496e-01 8.75523150e-01 1.07016671e+00 -3.06116402e-01
-1.97903919e+00 -7.82550752e-01 4.18871164e-01 -2.84759492e-01
3.89585346e-01 -1.46527812e-01 -6.69491529e-01 6.91236317e-01
5.14489472e-01 2.96384245e-01 7.35148489e-01 -2.11780190e-01
-4.43369061e-01 -5.78184962e-01 -1.15832198e+00 2.33835146e-01
8.12389195e-01 -7.84228981e-01 -3.46570015e-01 1.60670370e-01
4.72144753e-01 -7.71627903e-01 -7.93726027e-01 4.29585218e-01
8.13122094e-01 -1.54962790e+00 1.00706041e+00 9.48087946e-02
7.47802079e-01 -7.31587648e-01 -2.89796919e-01 -1.32279575e+00
-3.67992491e-01 -6.58046007e-01 -2.91429665e-02 9.18746412e-01
-8.35409239e-02 -9.54121172e-01 5.67058027e-01 2.53526241e-01
-6.78813383e-02 -2.88390785e-01 -7.23572254e-01 -4.55233425e-01
-3.16732407e-01 3.62186670e-01 3.97975445e-01 8.72351110e-01
-3.63226026e-01 3.23581964e-01 -5.87431610e-01 7.62884617e-01
1.04309690e+00 6.89016640e-01 7.00743377e-01 -1.07644498e+00
-5.45706749e-01 3.90514702e-01 -2.01191381e-01 -1.10160112e+00
9.64320898e-02 -4.22189355e-01 1.38323128e-01 -1.01858401e+00
4.86032963e-01 -2.74156630e-01 2.29203999e-02 -1.43030748e-01
4.91457470e-02 5.37987709e-01 1.99323192e-01 7.17609704e-01
-4.14490700e-01 1.61830291e-01 1.24576652e+00 1.76647887e-01
-2.21083522e-01 -5.18313229e-01 -4.64840025e-01 3.19370478e-01
4.46943909e-01 -4.40045565e-01 -1.18272960e-01 -8.44566584e-01
6.32903636e-01 3.37425321e-01 5.64760625e-01 -1.22829080e+00
2.47899011e-01 3.07809059e-02 3.69281530e-01 -2.84172356e-01
2.70209432e-01 -8.49070847e-01 8.03682864e-01 1.28144845e-01
6.30397862e-03 -1.36511132e-01 -2.47918963e-01 3.89826536e-01
-3.00756603e-01 2.21564934e-01 1.31640518e+00 -1.72182634e-01
-5.60378730e-01 2.24387929e-01 -3.16748053e-01 -5.85409217e-02
9.86754417e-01 -5.19210659e-02 -1.15500164e+00 2.23877709e-02
-1.48518443e-01 -3.78561378e-01 1.20829690e+00 -2.24866033e-01
4.97244865e-01 -1.02150548e+00 -2.31662631e-01 3.19785446e-01
-9.33446288e-02 -1.47847563e-01 6.45881355e-01 1.08734632e+00
-1.10603416e+00 3.92947882e-01 -5.03556013e-01 -7.56228745e-01
-1.29217803e+00 3.53369027e-01 6.36237800e-01 -2.46472880e-01
-6.60746276e-01 4.75063324e-01 -3.93184945e-02 -1.23947971e-01
-6.70055211e-01 1.62851200e-01 -7.25898594e-02 -2.31123865e-01
5.31220555e-01 2.61175364e-01 -1.12763010e-02 -6.60937250e-01
-3.16923469e-01 1.08909118e+00 9.35125649e-02 -3.54170471e-01
1.63258946e+00 -4.84433174e-01 -4.90565777e-01 4.09029394e-01
1.29230034e+00 3.65980297e-01 -1.36801624e+00 1.22940183e-01
-7.72495329e-01 -7.68892944e-01 1.38291895e-01 -5.83780289e-01
-1.13163686e+00 7.86737144e-01 8.68366063e-01 1.67740181e-01
1.37328053e+00 -4.16818820e-02 7.88421154e-01 2.72963464e-01
5.38751423e-01 -8.62849891e-01 -1.59182057e-01 4.41256791e-01
1.70379177e-01 -1.11770332e+00 4.53636795e-02 -3.97533536e-01
-3.43784332e-01 1.21891177e+00 1.93755865e-01 -1.00779332e-01
3.92035723e-01 4.49842781e-01 2.81291991e-01 -4.51151341e-01
-5.44945657e-01 1.56153515e-01 -2.74681568e-01 7.59426594e-01
5.70627928e-01 -2.49713957e-02 -1.98895708e-01 -3.86054426e-01
2.52906885e-02 2.09060863e-01 1.01151538e+00 6.99267685e-01
-3.16988200e-01 -1.16031837e+00 -4.60018009e-01 -4.88576591e-02
-5.34569025e-01 7.69999996e-02 4.22932208e-04 3.71755987e-01
-6.04695715e-02 5.93360782e-01 1.91740677e-01 -2.02709630e-01
3.65316242e-01 -1.72064841e-01 9.06662166e-01 -3.66695076e-01
1.76004887e-01 -4.31847759e-02 -5.12925200e-02 -7.69912541e-01
-1.08863461e+00 -5.62245309e-01 -8.20189655e-01 -3.66440415e-01
2.52739877e-01 -2.49231458e-01 6.01999223e-01 7.54772663e-01
6.74109161e-01 1.51874930e-01 8.75120878e-01 -9.90405023e-01
-9.73565578e-02 -3.99466991e-01 -1.12148976e+00 4.42985773e-01
7.62063444e-01 -4.68904853e-01 -6.79739237e-01 1.90002158e-01]
|
[10.236690521240234, -2.570350408554077]
|
91759d28-8523-415d-9a9e-bf6c0fbefc4e
|
vq-ar-vector-quantized-autoregressive
|
2205.15894
| null |
https://arxiv.org/abs/2205.15894v1
|
https://arxiv.org/pdf/2205.15894v1.pdf
|
VQ-AR: Vector Quantized Autoregressive Probabilistic Time Series Forecasting
|
Time series models aim for accurate predictions of the future given the past, where the forecasts are used for important downstream tasks like business decision making. In practice, deep learning based time series models come in many forms, but at a high level learn some continuous representation of the past and use it to output point or probabilistic forecasts. In this paper, we introduce a novel autoregressive architecture, VQ-AR, which instead learns a \emph{discrete} set of representations that are used to predict the future. Extensive empirical comparison with other competitive deep learning models shows that surprisingly such a discrete set of representations gives state-of-the-art or equivalent results on a wide variety of time series datasets. We also highlight the shortcomings of this approach, explore its zero-shot generalization capabilities, and present an ablation study on the number of representations. The full source code of the method will be available at the time of publication with the hope that researchers can further investigate this important but overlooked inductive bias for the time series domain.
|
['Kyung-Min Kim', 'Max Nihlén Ramström', 'Young-Jin Park', 'Kashif Rasul']
|
2022-05-31
| null | null | null | null |
['probabilistic-time-series-forecasting']
|
['time-series']
|
[ 5.44158705e-02 2.26405282e-02 -3.05287540e-01 -5.97227275e-01
-5.84881485e-01 -5.57320893e-01 1.08719230e+00 7.73308575e-02
-7.97111094e-02 6.15918577e-01 3.21244776e-01 -7.92130589e-01
-2.21462891e-01 -9.43335652e-01 -6.40478551e-01 -6.81480229e-01
-4.49425638e-01 3.85829926e-01 -1.95160285e-01 -4.68635798e-01
2.27744520e-01 5.53680062e-01 -1.54583454e+00 2.43684009e-01
5.50892830e-01 1.32299566e+00 -1.53230295e-01 4.77710128e-01
-3.18912417e-01 1.16439450e+00 -6.84118748e-01 -4.07733619e-01
2.04311565e-01 -2.40402654e-01 -4.16059524e-01 -3.64557981e-01
-6.53931871e-02 -5.77810764e-01 -6.86707318e-01 7.01867104e-01
2.83063501e-01 3.85985643e-01 8.97602141e-01 -1.28744328e+00
-8.23695958e-01 7.25085676e-01 -3.17901373e-01 6.58140004e-01
-1.41097352e-01 1.76808296e-03 9.91548419e-01 -6.64238274e-01
2.34017536e-01 9.66658950e-01 9.17828262e-01 2.61033863e-01
-1.12500989e+00 -8.10919523e-01 3.85902703e-01 1.77236199e-01
-1.03993893e+00 -3.41440022e-01 1.16843224e+00 -5.70135534e-01
1.21376908e+00 4.61705513e-02 5.45869529e-01 1.58796537e+00
6.14134490e-01 6.35966897e-01 8.61424327e-01 -2.61585593e-01
5.59965014e-01 -1.35101840e-01 2.66897976e-01 2.64206175e-02
-1.00387357e-01 5.68549931e-01 -1.53116941e-01 -2.58670837e-01
6.59452379e-01 6.64784729e-01 -5.90101769e-03 -5.01307473e-02
-9.48413312e-01 1.22751331e+00 5.45515418e-01 6.46355629e-01
-8.13548803e-01 2.76162684e-01 4.97488946e-01 6.78240836e-01
1.19216180e+00 4.73652214e-01 -5.45361042e-01 -2.98886865e-01
-1.10980058e+00 2.31976062e-01 7.85721481e-01 5.84679306e-01
4.21047837e-01 7.71773338e-01 -1.46541014e-01 4.93317217e-01
-2.64651328e-01 3.72374445e-01 7.96143711e-01 -7.81344533e-01
1.89616919e-01 9.81246904e-02 3.06210250e-01 -9.44855332e-01
-4.98385340e-01 -7.07410336e-01 -1.02237260e+00 -4.77667898e-02
1.56298518e-01 -5.51830649e-01 -1.08467221e+00 1.51690888e+00
-2.10297391e-01 6.09555364e-01 1.86963230e-01 7.50571609e-01
6.47655845e-01 1.20821500e+00 1.63028911e-01 -3.99013042e-01
7.15774357e-01 -6.11823797e-01 -6.96264029e-01 -3.21349353e-01
5.22891164e-01 -4.31688130e-01 5.14880419e-01 2.67802954e-01
-6.59475982e-01 -7.23224938e-01 -8.34729016e-01 1.12905651e-01
-6.26203001e-01 9.77632850e-02 1.05756664e+00 3.87756497e-01
-1.06040180e+00 9.49274123e-01 -9.04699743e-01 -2.40996197e-01
1.39055341e-01 4.48911749e-02 1.65222049e-01 2.71569908e-01
-1.58903718e+00 9.09994245e-01 2.06790924e-01 2.91907266e-02
-8.66049826e-01 -8.15928638e-01 -8.42003345e-01 3.45921606e-01
-8.66078362e-02 -3.38085890e-01 1.59336686e+00 -1.21134603e+00
-1.30167615e+00 4.80823785e-01 -2.65692323e-01 -1.07410467e+00
4.23496246e-01 -2.06959739e-01 -8.48299086e-01 -2.49610960e-01
-8.50736797e-02 2.90052652e-01 9.39078689e-01 -6.38451159e-01
-4.85362679e-01 -3.26687694e-01 -1.71166733e-02 -4.21214312e-01
-2.42989182e-01 -1.58193395e-01 4.65801150e-01 -8.70128632e-01
6.59085885e-02 -6.99235201e-01 -5.54077506e-01 -5.22152185e-01
1.87897176e-01 -6.57190979e-01 6.44358933e-01 -6.35086894e-01
1.38507438e+00 -2.18931222e+00 -1.93256721e-01 6.76818006e-03
2.54394952e-02 -2.00643875e-02 -1.02072127e-01 9.46201682e-01
-5.18273473e-01 9.99155641e-02 -7.04537109e-02 -4.03386921e-01
3.83071639e-02 2.16086924e-01 -1.48158622e+00 5.26042163e-01
3.33068311e-01 1.00204098e+00 -9.46788490e-01 2.82056838e-01
4.58181947e-01 5.71737647e-01 4.14311476e-02 1.38996363e-01
-3.65295708e-01 3.80882114e-01 -4.96025354e-01 2.54664153e-01
3.82861406e-01 -3.62214327e-01 7.64069101e-03 1.45130143e-01
-2.88984835e-01 3.72841477e-01 -5.14451325e-01 1.34360552e+00
-4.30447221e-01 8.26057434e-01 -7.68652081e-01 -1.42735016e+00
1.38351905e+00 5.98694324e-01 6.45730436e-01 -6.58263385e-01
1.58179641e-01 2.96020389e-01 3.85510474e-02 -1.34815527e-02
6.42612994e-01 -4.33334798e-01 -2.33049020e-01 3.73052776e-01
1.07238702e-01 -1.04637101e-01 -1.83239877e-01 -7.85898194e-02
7.98940599e-01 8.95796791e-02 3.01201642e-01 3.83245572e-02
-2.97460910e-02 7.26841912e-02 5.28709829e-01 9.33497727e-01
4.30099070e-02 4.85707521e-01 6.16724372e-01 -1.13774407e+00
-1.21301186e+00 -8.33667159e-01 -1.66336223e-01 1.21520567e+00
-5.54996908e-01 -2.39599079e-01 1.59080058e-01 -2.56238818e-01
9.99076888e-02 1.49784923e+00 -8.93073142e-01 -2.14122817e-01
-4.29297060e-01 -6.32563710e-01 1.16298489e-01 9.24195111e-01
-1.20064810e-01 -1.14886379e+00 -6.50698066e-01 4.44617212e-01
1.86274290e-01 -6.75982773e-01 1.72102869e-01 2.84756601e-01
-1.05273318e+00 -4.90000248e-01 -9.62181151e-01 -3.07255715e-01
1.00833904e-02 6.64850250e-02 1.14583707e+00 -4.55754071e-01
1.76086232e-01 4.10853744e-01 -4.31473613e-01 -1.05740654e+00
-3.88941556e-01 2.64261030e-02 5.81248887e-02 1.15748672e-02
8.08849394e-01 -8.90260577e-01 -5.13719141e-01 -3.46676558e-01
-6.45907104e-01 -3.35723758e-01 1.71372175e-01 8.30656886e-01
2.98006356e-01 4.51904088e-02 9.99659538e-01 -7.48083889e-01
7.38995194e-01 -1.02035809e+00 -6.26098156e-01 -1.38434410e-01
-6.31515503e-01 -1.11878611e-01 9.00805533e-01 -4.80998427e-01
-8.68077219e-01 -3.10793847e-01 -2.29344591e-01 -9.82429206e-01
3.52656655e-02 8.96310627e-01 8.29256535e-01 7.03611553e-01
6.44484520e-01 6.07401729e-01 -1.44792527e-01 -5.74787199e-01
3.17300916e-01 4.88166988e-01 3.92158359e-01 -2.12778389e-01
2.91389227e-01 4.86768425e-01 -5.74591346e-02 -5.96058547e-01
-1.00910354e+00 -2.26304859e-01 -3.82406503e-01 -1.74101770e-01
3.08964819e-01 -1.08098137e+00 -2.50571668e-01 2.41422653e-01
-1.13119400e+00 -2.95748740e-01 -6.02061450e-01 6.64733887e-01
-7.21781254e-01 -2.07668647e-01 -5.73512018e-01 -1.32172048e+00
-3.72043997e-01 -5.45114994e-01 1.07143176e+00 8.29083323e-02
-3.49812180e-01 -1.34494615e+00 2.11987063e-01 -4.86090183e-01
6.27418637e-01 3.80904406e-01 8.66585732e-01 -1.16980278e+00
-1.93796493e-02 -6.08758748e-01 1.11899354e-01 2.79458374e-01
-1.09579027e-01 3.67972907e-03 -1.19668174e+00 -4.13323194e-01
2.24266291e-01 -1.17226191e-01 1.19218218e+00 8.18530202e-01
1.48882055e+00 -1.84926108e-01 -3.10270935e-01 5.71741879e-01
1.26754224e+00 6.22216702e-01 5.85335791e-01 1.33901164e-01
2.39483669e-01 5.43441117e-01 5.65429628e-01 9.16329920e-01
2.45938942e-01 1.54507518e-01 3.85670602e-01 1.31822705e-01
4.43208843e-01 -4.05779600e-01 2.55385011e-01 7.44722724e-01
-1.98364064e-01 -3.64821941e-01 -1.18785334e+00 8.73347878e-01
-1.97463691e+00 -1.38258708e+00 7.77914673e-02 2.02394271e+00
3.42707276e-01 2.00322762e-01 4.88205999e-02 1.55378804e-01
4.76040810e-01 6.32476807e-01 -8.70952845e-01 -5.95492184e-01
-3.61400433e-02 3.38649213e-01 3.80421311e-01 1.19098850e-01
-1.20421755e+00 7.38057256e-01 7.33269644e+00 4.70653623e-01
-1.64359093e+00 4.39314321e-02 9.35286343e-01 1.08467008e-03
-5.62498927e-01 -2.24284939e-02 -5.62121391e-01 4.67879087e-01
1.70004880e+00 -6.79190516e-01 1.60017610e-01 1.01495111e+00
2.78831571e-01 4.26542401e-01 -1.13483596e+00 9.50643361e-01
-2.87204564e-01 -1.55926418e+00 -4.80098650e-03 -7.32458904e-02
8.12328100e-01 4.77230608e-01 3.37490469e-01 9.13937271e-01
4.13365334e-01 -1.19507885e+00 5.60790241e-01 9.92342055e-01
7.60556757e-01 -8.81457627e-01 7.35079467e-01 4.19373155e-01
-9.30608273e-01 -6.00454092e-01 -5.15649140e-01 -5.98701775e-01
1.74853966e-01 6.41491890e-01 -6.61151350e-01 6.05011463e-01
6.19977415e-01 1.18796504e+00 -2.33464673e-01 7.16053963e-01
1.78298533e-01 1.11969543e+00 -1.91743895e-01 -8.47323611e-02
5.94116271e-01 -8.15179199e-03 4.59043682e-01 9.36206162e-01
9.17588472e-01 5.61149530e-02 1.68117374e-01 8.12865317e-01
1.87011436e-01 -1.88323990e-01 -1.04160786e+00 -5.55723071e-01
4.52070206e-01 7.10498273e-01 -2.90300697e-01 -5.12793958e-01
-4.70828176e-01 6.15588069e-01 1.79707766e-01 6.99425638e-01
-7.47593999e-01 -3.26444060e-01 5.60899198e-01 1.85293667e-02
5.04577696e-01 -2.54141629e-01 -2.18966112e-01 -1.16900349e+00
-1.44482568e-01 -5.83699107e-01 5.64859569e-01 -9.58454132e-01
-1.67713535e+00 6.48929238e-01 -2.04393510e-02 -1.44517124e+00
-1.16805637e+00 -6.04087174e-01 -9.26673889e-01 1.08274078e+00
-1.48173225e+00 -9.37131703e-01 1.37166858e-01 3.00525308e-01
7.68239021e-01 -4.36311901e-01 9.81475532e-01 -2.20906094e-01
-1.46243528e-01 1.16502501e-01 6.93389118e-01 2.30971143e-01
4.08896089e-01 -1.09900188e+00 8.94243598e-01 6.00888789e-01
7.89578781e-02 3.92904997e-01 9.26469922e-01 -5.02976418e-01
-1.11480868e+00 -1.29315901e+00 1.02594066e+00 -3.60720903e-01
1.05090332e+00 -4.18325700e-02 -1.07233059e+00 1.19563389e+00
1.52608141e-01 -2.61681620e-03 6.40374303e-01 3.99862856e-01
-2.74815887e-01 -2.75392830e-01 -7.16094434e-01 3.18948746e-01
5.81450820e-01 -5.49328685e-01 -7.89731205e-01 3.63399923e-01
8.78717899e-01 -6.67568818e-02 -9.60185230e-01 4.04723674e-01
4.88741785e-01 -1.06477594e+00 8.21652412e-01 -9.74687815e-01
6.91182852e-01 3.33566487e-01 -2.14139730e-01 -1.61530590e+00
-6.41771674e-01 -5.80402911e-01 -5.09206533e-01 9.59768236e-01
1.38334230e-01 -1.04333186e+00 4.15665686e-01 4.16036874e-01
-1.68345988e-01 -7.87341535e-01 -1.02498662e+00 -8.03382933e-01
5.65252781e-01 -6.14667833e-01 9.04102385e-01 1.13184404e+00
-1.53421789e-01 2.38725811e-01 -5.57114482e-01 4.26477231e-02
2.97177583e-01 5.91758251e-01 4.19455945e-01 -1.51884186e+00
-8.03472102e-02 -3.92903805e-01 -4.09164131e-01 -9.56721783e-01
3.81791383e-01 -7.11651146e-01 -3.17921132e-01 -1.31514192e+00
-3.89934987e-01 -3.14958394e-01 -7.26353467e-01 1.06002748e-01
2.40129024e-01 -2.91591734e-01 2.20607893e-04 3.38247329e-01
-8.14740136e-02 8.80134284e-01 8.71476710e-01 -1.94857985e-01
-1.66892812e-01 4.18118447e-01 -6.18109167e-01 5.39279878e-01
1.10018313e+00 -2.14864388e-01 -6.39567077e-01 -3.13637406e-01
4.53104116e-02 5.44976950e-01 5.07063568e-01 -8.56432378e-01
9.51658841e-03 -2.31140628e-01 7.45298743e-01 -8.28793406e-01
6.26173079e-01 -6.33065164e-01 2.58476853e-01 3.03223521e-01
-5.05747497e-01 4.09371436e-01 1.67969242e-01 8.40289295e-01
-4.34210986e-01 -7.74847418e-02 3.16393644e-01 -2.84642845e-01
-1.03900158e+00 4.45317328e-01 -4.33437258e-01 -2.44486243e-01
9.14351344e-01 -2.75303517e-02 -2.24423423e-01 -1.04234159e+00
-7.27077425e-01 5.29561639e-02 2.06275042e-02 7.87334919e-01
4.80334997e-01 -1.26765847e+00 -9.67299938e-01 2.18295828e-01
3.47242504e-02 -5.02654195e-01 2.99400151e-01 6.61624968e-01
1.49332415e-02 6.69452608e-01 -1.51410148e-01 -5.99796593e-01
-2.68234730e-01 9.74030316e-01 4.08323914e-01 -2.94938475e-01
-8.75349164e-01 4.74857658e-01 2.27538347e-01 -1.57474443e-01
3.79324928e-02 -4.31341946e-01 -3.88641804e-01 3.45509022e-01
7.40861177e-01 2.48435855e-01 -3.99342924e-02 -4.43021595e-01
-1.88379705e-01 -1.12530030e-02 8.80198146e-04 -6.50872216e-02
1.87540448e+00 -1.91713180e-02 1.79027766e-01 1.38839531e+00
1.09721911e+00 -7.00275898e-01 -1.33251500e+00 -3.18800360e-01
5.83084747e-02 -2.76816398e-01 2.08420813e-01 -5.04454136e-01
-7.66551137e-01 1.18513548e+00 5.39632022e-01 7.09877491e-01
1.15393937e+00 -1.18879378e-01 5.90647221e-01 2.65734911e-01
3.78352821e-01 -8.39697897e-01 -4.88277048e-01 7.28739858e-01
1.11308300e+00 -1.14739525e+00 -2.07782373e-01 3.88932616e-01
-6.82929993e-01 1.32384920e+00 1.84493423e-01 -4.68365580e-01
8.85569394e-01 7.20727742e-02 2.84277126e-02 -1.55151263e-02
-1.36775243e+00 -1.45505995e-01 2.11350501e-01 5.28931141e-01
8.77803266e-01 2.97944635e-01 -1.42282128e-01 8.72906685e-01
-4.61155027e-01 1.50133014e-01 4.29449260e-01 6.16209984e-01
-3.54734331e-01 -5.87615669e-01 -1.46050334e-01 7.94751406e-01
-4.79014903e-01 -7.11230934e-02 -4.31025997e-02 5.51866472e-01
-4.37339067e-01 1.04706848e+00 7.23906577e-01 -2.53097385e-01
1.14559077e-01 4.49228317e-01 7.45121762e-02 -2.73029298e-01
-4.89760518e-01 2.85133272e-02 -5.37507311e-02 -3.77277285e-01
-2.49002263e-01 -6.52842224e-01 -8.18335354e-01 -4.39232022e-01
1.65849268e-01 6.42608404e-02 4.85028803e-01 7.79968262e-01
4.89836454e-01 6.31894886e-01 8.80097806e-01 -9.75117326e-01
-1.02250886e+00 -1.19822681e+00 -6.84837103e-01 3.72686267e-01
7.34154642e-01 -5.34591436e-01 -3.60222995e-01 -1.05155692e-01]
|
[6.986021995544434, 3.1470634937286377]
|
1f61fb26-7124-423c-ba16-19204a7489ae
|
view-guided-point-cloud-completion
|
2104.05666
| null |
https://arxiv.org/abs/2104.05666v2
|
https://arxiv.org/pdf/2104.05666v2.pdf
|
View-Guided Point Cloud Completion
|
This paper presents a view-guided solution for the task of point cloud completion. Unlike most existing methods directly inferring the missing points using shape priors, we address this task by introducing ViPC (view-guided point cloud completion) that takes the missing crucial global structure information from an extra single-view image. By leveraging a framework that sequentially performs effective cross-modality and cross-level fusions, our method achieves significantly superior results over typical existing solutions on a new large-scale dataset we collect for the view-guided point cloud completion task.
|
['Yue Gao', 'Yandong Guo', 'Xibin Zhao', 'Hai Wan', 'Changqing Zou', 'Siqi Li', 'Yutong Feng', 'Xuancheng Zhang']
|
2021-04-12
| null |
http://openaccess.thecvf.com//content/CVPR2021/html/Zhang_View-Guided_Point_Cloud_Completion_CVPR_2021_paper.html
|
http://openaccess.thecvf.com//content/CVPR2021/papers/Zhang_View-Guided_Point_Cloud_Completion_CVPR_2021_paper.pdf
|
cvpr-2021-1
|
['point-cloud-completion']
|
['computer-vision']
|
[ 1.46014929e-01 -3.83070186e-02 7.44405091e-02 -3.43090594e-01
-1.32486033e+00 -8.45632076e-01 8.10924888e-01 -6.14982396e-02
4.82066721e-02 9.43510532e-02 1.11501269e-01 -2.87046265e-02
9.30278525e-02 -5.93752682e-01 -9.07210290e-01 -2.50944912e-01
3.35770130e-01 8.39729249e-01 4.35041815e-01 -6.11480065e-02
4.18851137e-01 7.56154180e-01 -1.45143044e+00 4.03158128e-01
7.28076160e-01 7.26906598e-01 5.04171073e-01 5.17647326e-01
4.97523919e-02 -1.29279271e-01 1.00936651e-01 -2.87517697e-01
5.70453465e-01 4.01695251e-01 -6.39916182e-01 5.00877857e-01
9.69350994e-01 -4.83030647e-01 -1.34251788e-01 7.75000274e-01
1.97117776e-01 2.99500730e-02 4.47911352e-01 -1.14856172e+00
-6.70627058e-02 -3.25944036e-01 -1.03666341e+00 -2.43094951e-01
7.65828848e-01 -8.37472975e-02 9.59955871e-01 -1.59221411e+00
9.88090217e-01 1.24364161e+00 7.85384655e-01 1.27848268e-01
-1.41009963e+00 -3.11582685e-01 3.73416513e-01 -1.82194456e-01
-1.34742570e+00 -4.72584575e-01 1.06870079e+00 -5.58946431e-01
1.00891495e+00 1.04053654e-01 5.73964417e-01 8.35566700e-01
-1.81779608e-01 4.95613843e-01 9.67370152e-01 -2.86753565e-01
1.21236354e-01 -3.98779720e-01 -2.82714546e-01 4.74527061e-01
-4.48619053e-02 2.19802663e-01 -6.02413893e-01 -6.08691454e-01
1.13735938e+00 5.09933293e-01 -2.43087396e-01 -1.04157293e+00
-1.66972589e+00 6.51327074e-01 5.02693832e-01 -3.03550482e-01
-5.41209161e-01 2.19271034e-01 9.03670304e-03 -3.49268802e-02
6.54962897e-01 6.45827353e-02 -6.86065972e-01 1.47411972e-01
-1.16641688e+00 2.14484453e-01 3.92638952e-01 1.39015937e+00
1.00935256e+00 -2.98049450e-01 2.59211332e-01 5.19301534e-01
4.66022521e-01 9.94954705e-01 -5.29664099e-01 -1.48797905e+00
7.43349075e-01 6.18251979e-01 3.32685858e-01 -9.67117727e-01
-2.72505015e-01 -2.17948660e-01 -4.80409294e-01 4.89042819e-01
1.86858047e-02 1.63729519e-01 -1.04123080e+00 1.16025198e+00
8.00210834e-01 2.46140301e-01 -1.97652251e-01 8.05241644e-01
8.50655198e-01 4.69007283e-01 -3.35845560e-01 3.55403274e-02
1.34926057e+00 -9.31754887e-01 -3.15671355e-01 -1.70169085e-01
-1.78199373e-02 -9.54852879e-01 8.20517719e-01 5.83980024e-01
-1.16784883e+00 -3.29187959e-01 -9.77513492e-01 -4.71921176e-01
-1.00078829e-01 -5.08031845e-02 6.26209319e-01 4.19195294e-02
-1.25748312e+00 2.20579222e-01 -1.01549566e+00 -2.04415917e-01
3.60403806e-01 2.76886135e-01 -8.98384631e-01 -6.46784306e-01
-6.40484095e-02 5.09445369e-01 -3.51249194e-03 -2.45956127e-02
-6.85205638e-01 -1.08457565e+00 -9.06713903e-01 -6.57262355e-02
6.40697420e-01 -1.36759651e+00 1.07529676e+00 -1.78272650e-01
-1.13046277e+00 1.00851738e+00 -7.35241175e-01 1.16405822e-01
5.08453965e-01 -3.46381694e-01 1.77805275e-01 6.20522559e-01
3.60396236e-01 9.60078478e-01 1.00132895e+00 -2.06137657e+00
-4.71551090e-01 -7.48967886e-01 6.74880669e-02 4.96393144e-01
6.09773934e-01 -2.12951705e-01 -1.17162704e+00 -5.04702330e-01
1.01680100e+00 -9.36197937e-01 -2.86196828e-01 3.40008020e-01
-3.07976037e-01 -7.70108849e-02 8.97905648e-01 -6.71793163e-01
3.42913538e-01 -2.11334848e+00 3.17266405e-01 3.19632381e-01
4.89853859e-01 -3.07923287e-01 -1.47423893e-02 6.36449158e-01
-5.18201059e-03 -2.04846144e-01 -3.62864643e-01 -1.11045325e+00
-2.10600495e-01 3.42826009e-01 -4.82190520e-01 4.91321355e-01
-1.05767019e-01 8.42946410e-01 -7.73685634e-01 -4.52597588e-01
6.22120202e-01 7.46582389e-01 -8.10288727e-01 4.71451767e-02
-2.16621295e-01 7.60507822e-01 -2.77391285e-01 8.43187749e-01
1.26284933e+00 -4.55813259e-01 -1.19779013e-01 -5.66145778e-01
-1.97021902e-01 -2.59823471e-01 -1.08028913e+00 2.58044910e+00
-5.87797463e-01 1.97741941e-01 5.74151158e-01 -3.26281130e-01
5.79669893e-01 4.28195298e-01 9.75659728e-01 -6.88127726e-02
-2.68352836e-01 3.11768893e-02 -8.22500288e-01 4.11575884e-02
5.57687938e-01 -2.43574291e-01 1.07427895e-01 1.70085937e-01
5.20820543e-02 -6.85820460e-01 -5.93814313e-01 7.10435450e-01
9.75143433e-01 5.66088080e-01 1.64558932e-01 1.02239229e-01
5.00487924e-01 2.40499243e-01 4.02785301e-01 3.39677185e-01
1.79616913e-01 1.56119108e+00 -6.05149828e-02 -2.55797744e-01
-1.11043334e+00 -1.54141951e+00 -6.10312857e-02 5.96017241e-01
2.68298656e-01 -6.61168218e-01 -2.34872401e-01 -6.29969716e-01
1.94075495e-01 4.13017750e-01 -3.72620136e-01 5.13678074e-01
-4.86279070e-01 -4.16868320e-03 -1.89410985e-01 5.97654164e-01
1.42847985e-01 -5.29322207e-01 -2.77932405e-01 -2.08504610e-02
-5.58444619e-01 -1.63445199e+00 -5.44513047e-01 -2.43647680e-01
-1.36744142e+00 -1.20678091e+00 -6.99281037e-01 -4.33457732e-01
1.04341567e+00 1.00907516e+00 1.42412865e+00 1.75300598e-01
1.24877162e-01 1.01011825e+00 -2.03873262e-01 -2.85115808e-01
2.24526897e-01 -2.53121853e-01 -7.99817815e-02 -7.15820789e-02
-1.00979812e-01 -9.93081570e-01 -7.65863717e-01 2.67640561e-01
-6.94320500e-01 2.72782236e-01 4.82051313e-01 4.82274145e-01
1.11302876e+00 -5.03106117e-01 1.06015921e-01 -7.92998552e-01
-2.01736405e-01 -1.95391506e-01 -7.82294333e-01 6.04499169e-02
-2.95635939e-01 -3.20499182e-01 1.29121110e-01 2.35148013e-01
-1.00997281e+00 5.35330415e-01 -2.16464967e-01 -1.08424890e+00
-1.60820484e-01 4.55392571e-03 -3.33655894e-01 -3.57661039e-01
1.18343800e-01 2.34960511e-01 2.24816706e-02 -8.80789340e-01
6.64361477e-01 1.36535406e-01 8.44078004e-01 -6.24937057e-01
1.12820220e+00 1.50119293e+00 2.31602266e-01 -7.55654693e-01
-7.62944281e-01 -9.99814034e-01 -1.29465520e+00 -6.44832477e-02
8.95132780e-01 -1.32230234e+00 -6.74853325e-01 6.76334500e-02
-1.48822379e+00 3.32690239e-01 -2.03824788e-01 4.17752653e-01
-9.24948394e-01 7.74825633e-01 -1.22277558e-01 -5.02789140e-01
-3.78987253e-01 -9.72436905e-01 2.00955868e+00 -4.73697722e-01
2.59839147e-01 -9.63512719e-01 2.02483878e-01 5.97252250e-01
-1.98762491e-01 5.42779744e-01 4.05792981e-01 -5.02256490e-02
-1.22264469e+00 -2.04086646e-01 -3.26196343e-01 -1.35293892e-02
-1.17177650e-01 3.09310500e-02 -9.42907035e-01 -4.20553267e-01
8.79111215e-02 8.31435472e-02 7.13011742e-01 5.40544987e-01
8.31556320e-01 3.50623399e-01 -6.27970397e-01 1.12562454e+00
1.67977250e+00 -3.77047598e-01 3.74001056e-01 3.14687937e-03
1.19412708e+00 4.84783679e-01 6.18269145e-01 4.42450941e-01
7.68130481e-01 8.16081643e-01 8.78602266e-01 -2.69494325e-01
-1.25320107e-01 -4.79946226e-01 -1.89107865e-01 7.53449917e-01
-3.43256950e-01 1.86415911e-01 -9.57435489e-01 5.94902635e-01
-1.84068036e+00 -7.10223794e-01 -4.40177470e-01 2.18244815e+00
1.44051658e-02 -2.63461322e-01 -1.18628375e-01 -2.24880457e-01
4.09448415e-01 1.94211766e-01 -4.26580817e-01 3.27854544e-01
1.43339271e-02 2.83189155e-02 4.85034943e-01 8.07820499e-01
-9.72081780e-01 7.98744678e-01 7.39169407e+00 5.38947284e-01
-6.90087557e-01 3.23092788e-01 -1.12324946e-01 -1.64208621e-01
-6.99009359e-01 4.24375027e-01 -5.43614686e-01 2.10421942e-02
2.15511069e-01 2.70316929e-01 2.63504773e-01 7.85385668e-01
1.65289044e-01 -3.48572671e-01 -1.33009803e+00 1.51956379e+00
2.62298971e-01 -1.47668636e+00 6.25097845e-03 4.88640487e-01
8.88553798e-01 6.53197229e-01 -1.66651830e-01 -1.91669673e-01
3.68431985e-01 -5.70323646e-01 5.96140087e-01 7.21918046e-01
9.23585474e-01 -5.82050085e-01 2.84749448e-01 4.93382156e-01
-1.22446012e+00 3.74570221e-01 -2.99417913e-01 2.03266963e-01
6.83044136e-01 9.04005289e-01 -5.62782705e-01 1.35484660e+00
6.64718986e-01 9.30391073e-01 -5.95458269e-01 9.80991185e-01
-4.76759709e-02 1.00032754e-01 -7.64776886e-01 1.24237251e+00
-4.86872345e-03 -3.31274092e-01 8.06994975e-01 4.89061981e-01
5.41889906e-01 8.62147808e-02 3.47599655e-01 7.35907018e-01
2.13619575e-01 -4.19921219e-01 -9.57820952e-01 8.20531607e-01
4.15330499e-01 1.41949809e+00 -7.34211683e-01 -5.32627404e-01
-6.21956885e-01 1.08738375e+00 3.00879806e-01 3.89710277e-01
-4.38939571e-01 3.63736957e-01 4.54727322e-01 2.69765735e-01
6.83490813e-01 -6.92834437e-01 -7.31810689e-01 -1.56962574e+00
2.73677677e-01 -3.53186607e-01 2.47236326e-01 -1.26167023e+00
-1.32430291e+00 3.46337855e-01 3.79158527e-01 -1.63434112e+00
-1.94129914e-01 -3.70983601e-01 -6.20484352e-01 1.14565325e+00
-1.56101930e+00 -1.76138699e+00 -4.58822638e-01 9.57430482e-01
6.04690850e-01 2.21378952e-01 5.21582186e-01 1.69524606e-02
2.91709214e-01 -8.77286047e-02 -8.42190012e-02 -3.33627731e-01
6.36502266e-01 -1.14586318e+00 7.08341837e-01 8.08820128e-01
1.98186994e-01 6.92997992e-01 4.74537969e-01 -8.44597340e-01
-1.66437876e+00 -7.41335392e-01 5.03014207e-01 -1.23456681e+00
1.37271598e-01 -4.79112148e-01 -6.82025850e-01 9.75723982e-01
5.20775653e-02 4.29456204e-01 3.66150588e-01 1.54988497e-01
-4.56609637e-01 7.08093122e-02 -1.26740491e+00 1.97086424e-01
1.31956828e+00 -6.10365987e-01 -7.06435204e-01 3.79210323e-01
8.17903697e-01 -6.95027411e-01 -9.13445950e-01 6.08979106e-01
3.80411565e-01 -1.05711925e+00 1.52362740e+00 -7.41629824e-02
2.93436557e-01 -5.97620308e-01 -5.26493788e-01 -1.17352331e+00
-2.63942331e-01 -6.04185104e-01 -3.01373661e-01 9.75200236e-01
-1.62343666e-01 -4.49004650e-01 1.02206838e+00 4.76183295e-01
-4.16608542e-01 -4.73876476e-01 -1.04720473e+00 -5.40980637e-01
-1.81936756e-01 -6.26551867e-01 6.73367143e-01 7.88416743e-01
-3.77216727e-01 2.27627352e-01 -2.17728376e-01 6.69714987e-01
1.16732526e+00 4.53350782e-01 1.26529479e+00 -1.43397486e+00
-1.48838222e-01 2.60118693e-01 -3.27733189e-01 -1.32250321e+00
-1.70998588e-01 -7.55330980e-01 -8.04264918e-02 -1.89170551e+00
6.73157573e-02 -2.47758061e-01 3.32248151e-01 1.50246367e-01
-1.86506733e-01 3.87181491e-01 5.05661726e-01 5.22665918e-01
-6.36324346e-01 5.22209227e-01 1.34744918e+00 2.88112819e-01
1.06002174e-01 1.16052248e-01 -5.75656772e-01 1.03799319e+00
8.40337276e-02 -2.84965217e-01 -4.37903523e-01 -7.60117769e-01
2.49086678e-01 5.41552901e-01 6.58821940e-01 -8.58811259e-01
2.84587175e-01 -1.71437174e-01 4.34540898e-01 -1.64967525e+00
1.07140315e+00 -1.34852815e+00 4.55912799e-01 -1.34894460e-01
4.66483533e-01 4.78567988e-01 4.64355526e-03 1.10224378e+00
-1.18191659e-01 2.13420615e-01 2.83843338e-01 -2.99747288e-01
-5.61736524e-01 6.91490948e-01 2.90554971e-01 -1.59193978e-01
1.02465713e+00 -4.40253526e-01 -4.00434658e-02 -3.59538764e-01
-1.10028708e+00 4.21413809e-01 1.18811679e+00 3.73896152e-01
9.67774153e-01 -1.42562962e+00 -6.22161865e-01 4.68404412e-01
4.81743604e-01 8.13064992e-01 5.34779370e-01 9.11545515e-01
-5.65256834e-01 3.06953490e-01 1.09047540e-01 -1.30688190e+00
-1.35569608e+00 7.18131542e-01 -6.69928938e-02 -1.22399502e-01
-1.28551996e+00 5.97136497e-01 7.16721296e-01 -8.10413539e-01
-2.98593819e-01 -2.15156585e-01 1.60623401e-01 -3.95889133e-01
2.54122823e-01 2.72651941e-01 3.93142074e-01 -8.55613768e-01
-4.27364945e-01 1.01076949e+00 8.83988887e-02 -5.29480278e-01
1.44614112e+00 -3.49996716e-01 -4.16674837e-02 3.74832541e-01
1.13868845e+00 3.84831905e-01 -1.72799349e+00 -3.17352235e-01
-3.93840700e-01 -8.88423383e-01 1.23899370e-01 -6.51826799e-01
-9.20717061e-01 9.03570890e-01 6.07167482e-02 -3.56371641e-01
9.23746765e-01 2.81867534e-01 6.44513726e-01 1.51823118e-01
8.66416752e-01 -6.66610718e-01 -2.28988647e-01 4.41521227e-01
1.21073830e+00 -1.40927517e+00 5.68916023e-01 -1.16089737e+00
-4.06966537e-01 8.39019358e-01 4.45677459e-01 -4.45665836e-01
8.51676822e-01 3.46083567e-02 -1.12109639e-01 -6.94914699e-01
-8.21339011e-01 -4.76829857e-02 4.89686102e-01 9.98335481e-01
-2.16476172e-01 -7.45434836e-02 1.61440194e-01 -5.27616292e-02
-1.01402126e-01 -1.10456243e-01 3.52165699e-01 1.02065384e+00
-1.19261846e-01 -1.10297167e+00 -8.56992424e-01 2.83105582e-01
-5.77832572e-03 -1.16271861e-01 -2.66137809e-01 6.71080589e-01
-1.70637593e-01 8.24742436e-01 3.05118620e-01 -3.04113239e-01
4.34156209e-01 -1.15380749e-01 6.74172878e-01 -8.64276707e-01
-2.47224376e-01 4.03659970e-01 -2.02938333e-01 -9.47063982e-01
-6.73292398e-01 -9.18480814e-01 -1.06838906e+00 -2.37158015e-01
-4.14028428e-02 -1.51673362e-01 7.97783673e-01 7.33985126e-01
8.18128586e-01 1.82396382e-01 7.33274460e-01 -1.70535493e+00
-2.18518659e-01 -6.11035407e-01 -3.37173939e-01 6.07528269e-01
4.47958708e-01 -9.94648218e-01 -3.12349886e-01 9.13732797e-02]
|
[8.383389472961426, -3.175233840942383]
|
7730b19d-2b49-4a3e-bf93-f1b6762d3c5d
|
building-manufacturing-deep-learning-models
|
2306.00202
| null |
https://arxiv.org/abs/2306.00202v1
|
https://arxiv.org/pdf/2306.00202v1.pdf
|
Building Manufacturing Deep Learning Models with Minimal and Imbalanced Training Data Using Domain Adaptation and Data Augmentation
|
Deep learning (DL) techniques are highly effective for defect detection from images. Training DL classification models, however, requires vast amounts of labeled data which is often expensive to collect. In many cases, not only the available training data is limited but may also imbalanced. In this paper, we propose a novel domain adaptation (DA) approach to address the problem of labeled training data scarcity for a target learning task by transferring knowledge gained from an existing source dataset used for a similar learning task. Our approach works for scenarios where the source dataset and the dataset available for the target learning task have same or different feature spaces. We combine our DA approach with an autoencoder-based data augmentation approach to address the problem of imbalanced target datasets. We evaluate our combined approach using image data for wafer defect prediction. The experiments show its superior performance against other algorithms when the number of labeled samples in the target dataset is significantly small and the target dataset is imbalanced.
|
['Ting-Yan Wu', 'Rih-Teng Wu', 'Elisa Bertino', 'Adrian Shuai Li']
|
2023-05-31
| null | null | null | null |
['defect-detection']
|
['computer-vision']
|
[ 3.46875519e-01 5.89617006e-02 -3.71135861e-01 -5.77074587e-01
-6.28489494e-01 -3.71566564e-02 2.22784579e-02 4.64955091e-01
-1.55170903e-01 9.03955936e-01 -1.64144218e-01 -3.85409035e-02
1.22622214e-01 -1.11632776e+00 -7.23147154e-01 -6.86938763e-01
4.80293036e-01 6.76574051e-01 -2.12048087e-02 -2.67438423e-02
1.56942740e-01 4.02965873e-01 -1.63394296e+00 5.98205030e-01
9.77706909e-01 1.27425659e+00 3.16610217e-01 3.15031230e-01
-4.83381033e-01 1.05548334e+00 -8.61876070e-01 1.51672706e-01
3.83223981e-01 -3.47711712e-01 -8.03903580e-01 4.65130240e-01
3.72834474e-01 -1.38307750e-01 -2.42233668e-02 8.20499897e-01
6.33888721e-01 -2.12699011e-01 6.22240841e-01 -1.44724500e+00
-3.92787457e-01 2.48383284e-01 -7.81921685e-01 3.47488284e-01
-2.25817800e-01 -2.47353837e-01 4.47529167e-01 -6.78999245e-01
6.42947912e-01 9.01302040e-01 8.04011524e-01 4.88905549e-01
-1.06752598e+00 -7.96261787e-01 -6.76286146e-02 1.25116378e-01
-9.52811956e-01 -3.01808029e-01 1.14976537e+00 -5.16636550e-01
6.48534119e-01 -2.38412753e-01 3.56586069e-01 8.26590359e-01
-1.48087621e-01 8.90095651e-01 1.09071147e+00 -6.73722923e-01
5.00831723e-01 4.30055171e-01 2.28886530e-01 3.95687252e-01
3.50155681e-01 1.02917112e-01 -5.10915518e-01 5.55938818e-02
5.92266381e-01 1.82017371e-01 1.34984002e-01 -4.23182368e-01
-8.69571686e-01 1.04734766e+00 3.44358295e-01 5.11768997e-01
-4.89241332e-01 -4.94744182e-01 6.54206097e-01 8.25785697e-01
7.61222184e-01 4.83176380e-01 -8.79967391e-01 2.37497926e-01
-9.39018965e-01 1.04251124e-01 6.88784420e-01 6.59168065e-01
9.14565027e-01 1.44976661e-01 1.15406431e-01 9.89942074e-01
8.94793496e-02 1.29266754e-01 9.30377305e-01 -4.02577162e-01
7.41472960e-01 1.18644643e+00 -3.04276720e-02 -9.53679264e-01
-4.45671439e-01 -4.52133030e-01 -8.54487002e-01 3.65542740e-01
5.35392106e-01 -2.32820913e-01 -1.17636740e+00 1.45268548e+00
5.56926429e-01 -4.71786596e-02 4.03300554e-01 5.25725186e-01
8.56516182e-01 5.85369825e-01 -7.28456816e-03 -2.11045250e-01
7.56751537e-01 -8.41827631e-01 -6.51892900e-01 -4.71290916e-01
9.44091618e-01 -6.11455202e-01 1.08678782e+00 5.20658493e-01
-6.44497395e-01 -7.08924949e-01 -1.32768893e+00 1.61929056e-01
-5.22210419e-01 6.01536691e-01 5.47233403e-01 6.16533816e-01
-5.62874019e-01 5.48126280e-01 -4.88699585e-01 -3.33321661e-01
9.00892437e-01 7.02767789e-01 -4.80033964e-01 -5.91848254e-01
-7.36048400e-01 5.66237271e-01 6.82860613e-01 -2.52912909e-01
-8.03440094e-01 -7.07697690e-01 -7.80204356e-01 -9.93135050e-02
2.23567054e-01 -4.16463427e-02 8.99744093e-01 -1.50201762e+00
-1.09830093e+00 1.02711165e+00 3.20316821e-01 -4.00126070e-01
2.37427890e-01 -9.67735946e-02 -4.24973100e-01 -1.66339219e-01
1.27350524e-01 4.39922363e-01 1.05273366e+00 -1.30192733e+00
-6.19487584e-01 -8.70051086e-01 -1.41267017e-01 9.90647301e-02
-7.26093471e-01 -2.57989854e-01 1.25061229e-01 -6.76992893e-01
2.78273374e-01 -6.42842114e-01 -1.90250576e-05 -1.64753184e-01
-4.40724224e-01 -1.11647815e-01 1.33843172e+00 -5.99575341e-01
9.73338306e-01 -2.24949455e+00 -3.54543477e-02 1.93375394e-01
7.97256604e-02 4.97212559e-01 -2.25597426e-01 2.12447271e-01
-5.18612385e-01 -1.58284560e-01 -3.50990981e-01 -1.47087112e-01
-4.72247720e-01 3.15668315e-01 1.64746597e-01 3.52889538e-01
4.92175430e-01 3.75398040e-01 -5.36601484e-01 -5.28153718e-01
1.11072540e-01 -6.73710927e-03 -4.40649599e-01 5.74934065e-01
-1.90645933e-01 6.46466315e-01 -3.95203978e-01 9.30164337e-01
8.60936642e-01 -1.48216560e-01 3.44031788e-02 -3.48955035e-01
2.81708628e-01 -1.87144782e-02 -1.37028229e+00 1.59472144e+00
-5.96144795e-01 5.25162518e-01 -2.62456566e-01 -1.69192946e+00
1.25467980e+00 2.17791721e-01 6.17729485e-01 -8.27604353e-01
9.67908204e-02 2.88298041e-01 1.33596286e-01 -6.69991195e-01
-1.11248344e-03 -5.39233506e-01 7.64819607e-02 4.67042953e-01
3.09181094e-01 2.77722776e-02 1.71641424e-01 -2.86762953e-01
1.28096831e+00 -9.46815535e-02 3.34583610e-01 -3.76741067e-02
5.21574795e-01 3.12586576e-01 9.50663984e-01 3.20470124e-01
-2.12785542e-01 4.27911907e-01 5.81336021e-01 -7.41510868e-01
-1.37331152e+00 -5.24624527e-01 -1.44121096e-01 7.47566700e-01
-8.33856687e-02 4.50304560e-02 -5.84711611e-01 -1.11300790e+00
2.07404375e-01 2.57003993e-01 -6.95235491e-01 -3.44192982e-01
-4.00448233e-01 -1.08519149e+00 2.84393400e-01 7.29491889e-01
5.60469806e-01 -1.18935716e+00 -5.18386185e-01 1.71013370e-01
3.71454172e-02 -8.91358078e-01 1.88353360e-01 4.56680626e-01
-1.26431620e+00 -1.17641819e+00 -4.90068674e-01 -1.06519914e+00
1.06520522e+00 -4.18080948e-02 1.23228776e+00 2.31497344e-02
-4.97057110e-01 -1.04410209e-01 -4.08694267e-01 -9.23606694e-01
-5.79217136e-01 8.37890729e-02 8.63539353e-02 6.87932000e-02
7.15013564e-01 -4.18271333e-01 -3.07142496e-01 1.09876998e-01
-1.13904452e+00 -1.19752705e-01 7.30573177e-01 1.17825663e+00
5.47975540e-01 7.45787203e-01 1.17195642e+00 -1.22902548e+00
4.68079090e-01 -6.27590597e-01 -3.59198242e-01 5.37986576e-04
-6.35129154e-01 8.01311955e-02 5.76302826e-01 -5.43571711e-01
-1.14914453e+00 1.72056317e-01 -6.14383332e-02 -3.71175677e-01
-4.78741527e-01 6.51988924e-01 -5.12608767e-01 -7.35517293e-02
7.19637096e-01 -2.74329543e-01 1.96779326e-01 -5.69312453e-01
-3.86241525e-01 9.58214581e-01 8.37234557e-02 -2.69065112e-01
5.52482843e-01 3.60064149e-01 -1.53647110e-01 -5.53178549e-01
-8.83217156e-01 -4.11295891e-01 -8.89822781e-01 9.62323025e-02
4.29046512e-01 -9.16269064e-01 -1.28419355e-01 8.26363862e-01
-8.07154953e-01 -2.97808379e-01 -6.45626485e-01 3.85043472e-01
-3.35157275e-01 9.82631817e-02 -3.73196036e-01 -6.14805996e-01
-1.90127909e-01 -1.04949749e+00 8.57309282e-01 1.44443482e-01
1.13835216e-01 -8.45587492e-01 1.50249852e-02 6.45709574e-01
2.23087713e-01 4.81066465e-01 1.32192814e+00 -1.14230788e+00
-1.15031265e-01 -5.79374135e-01 -3.28823328e-01 8.40144932e-01
4.97423083e-01 -5.44388473e-01 -8.80057335e-01 -3.34188938e-01
3.55114609e-01 -7.96336830e-01 6.76585495e-01 3.30431491e-01
1.22846067e+00 -1.70816824e-01 -1.82585552e-01 1.65254269e-02
1.79949510e+00 3.68609220e-01 4.84383345e-01 4.18586105e-01
7.40653694e-01 8.58111143e-01 1.08496439e+00 5.15368700e-01
9.87765491e-02 4.20431226e-01 3.78814310e-01 -4.06171829e-01
-8.51581767e-02 1.05678894e-01 4.17964756e-02 7.40235090e-01
4.86144811e-01 -1.72791198e-01 -1.12539470e+00 8.19575012e-01
-1.66384184e+00 -5.22492111e-01 2.09695205e-01 2.22680569e+00
9.07717407e-01 1.86064884e-01 1.04029186e-01 7.88372457e-01
6.54965460e-01 -2.71241158e-01 -8.53680313e-01 -1.87728360e-01
-2.90342029e-02 3.22342187e-01 4.05025929e-01 5.02413400e-02
-1.26683557e+00 5.91295183e-01 5.67575598e+00 7.38689125e-01
-1.18211091e+00 2.64418960e-01 9.35536087e-01 -5.94056360e-02
1.07855581e-01 -4.14603978e-01 -5.28436899e-01 5.08771241e-01
7.90780067e-01 2.20345259e-01 -9.73025337e-02 9.14054692e-01
-6.37062415e-02 -4.24072355e-01 -1.28459895e+00 1.11122191e+00
2.53314227e-01 -1.16059685e+00 -2.43534997e-01 1.37637615e-01
8.51006925e-01 -1.21030532e-01 7.99944401e-02 3.24606687e-01
2.15232391e-02 -1.07117856e+00 5.02392016e-02 1.85318395e-01
5.78311622e-01 -1.01932919e+00 1.20795834e+00 5.42855263e-01
-4.36735958e-01 -5.45601130e-01 -5.49406230e-01 -8.73557627e-02
-4.77362186e-01 1.11835146e+00 -9.53863859e-01 3.50091428e-01
6.33087218e-01 7.10429609e-01 -4.94815260e-01 8.74097526e-01
2.90332884e-01 5.75495005e-01 -2.39042893e-01 4.25698370e-01
1.47673190e-02 -2.48378073e-03 4.49398942e-02 6.18778944e-01
2.43203104e-01 -2.25762144e-01 4.36976045e-01 5.59456825e-01
-3.17081809e-01 3.11381668e-01 -9.99581993e-01 -9.76520777e-02
1.57781079e-01 9.78039563e-01 -6.35085523e-01 -1.76391125e-01
-4.54129577e-01 8.87471378e-01 2.74966240e-01 -6.50391653e-02
-3.56065780e-01 -2.31383637e-01 3.05619419e-01 6.98457956e-02
1.02669634e-01 1.68655276e-01 -5.07929385e-01 -9.43058968e-01
1.76718667e-01 -1.07023311e+00 5.37769318e-01 -4.15031314e-01
-1.49371433e+00 4.18385744e-01 -3.39445084e-01 -1.39123094e+00
-1.80714086e-01 -6.83362424e-01 -3.51097196e-01 6.08203709e-01
-1.65269709e+00 -1.35409415e+00 -5.03098071e-01 5.53844929e-01
7.42698491e-01 -7.08453238e-01 7.48833597e-01 6.66148543e-01
-7.17106640e-01 4.71905082e-01 2.85277098e-01 1.68557718e-01
7.96905518e-01 -1.12471592e+00 -3.10963839e-01 6.37705147e-01
-1.29592419e-02 -6.10818155e-02 4.66460705e-01 -8.00452232e-01
-1.17756236e+00 -1.28763878e+00 5.65970123e-01 -1.64410397e-01
1.19940817e-01 -1.12555720e-01 -1.04607677e+00 6.10114753e-01
-1.18023284e-01 2.77139753e-01 1.18502414e+00 1.48122713e-01
-5.32341488e-02 -3.93095404e-01 -1.72651994e+00 -1.60663918e-01
4.93414283e-01 -3.16187978e-01 -3.63995641e-01 6.20066762e-01
3.64704967e-01 -4.70146537e-01 -1.00653982e+00 5.78914165e-01
1.95557639e-01 -9.06758010e-01 7.19105065e-01 -8.49107325e-01
6.79323852e-01 -2.77632296e-01 -2.67239243e-01 -1.40605021e+00
1.17921509e-01 5.15241683e-01 -5.43651693e-02 1.35236681e+00
2.50835806e-01 -2.02291325e-01 1.22564518e+00 3.65356684e-01
9.82730910e-02 -8.19078267e-01 -6.86344862e-01 -6.84747040e-01
1.53446868e-01 -2.48696849e-01 6.27615631e-01 1.15606320e+00
-2.65658140e-01 1.82056189e-01 -3.12875330e-01 1.77622154e-01
4.21357989e-01 2.18555212e-01 5.59101999e-01 -1.64573443e+00
-1.15574569e-01 3.98257703e-01 -6.07245028e-01 -1.10683560e-01
3.10785741e-01 -8.42155457e-01 -1.01147234e-01 -1.25856686e+00
1.76941276e-01 -9.17617083e-01 -4.17866945e-01 7.10286915e-01
1.07509904e-01 2.66506582e-01 -6.97509125e-02 2.34353289e-01
-1.94249287e-01 4.64985460e-01 8.95379066e-01 -4.23529625e-01
-4.31009918e-01 7.47002065e-02 -6.34887636e-01 7.17590630e-01
9.88124192e-01 -6.63913131e-01 -6.19429410e-01 -4.76916909e-01
1.54503407e-02 -1.44987687e-01 1.14974380e-01 -1.46426141e+00
2.34053209e-02 -5.45244617e-03 9.47636664e-01 -5.80169499e-01
5.16606681e-02 -1.11131728e+00 -1.46941885e-01 4.43781048e-01
-2.70522743e-01 -2.70913571e-01 3.99607331e-01 5.34925520e-01
-5.84976673e-01 -4.90358233e-01 1.06075501e+00 -2.36748219e-01
-7.87997484e-01 2.97807485e-01 -5.29213483e-03 -2.04279244e-01
1.25487268e+00 -3.29328686e-01 -8.09519365e-02 -3.46024893e-02
-8.35572541e-01 6.01561777e-02 4.97269869e-01 4.78518963e-01
6.95826232e-01 -1.41655254e+00 -6.48701191e-01 4.26159650e-01
4.66477245e-01 2.56991059e-01 2.81331062e-01 4.60394979e-01
-3.78341109e-01 1.84492171e-01 -6.98486626e-01 -4.56032872e-01
-1.23014617e+00 8.56379747e-01 1.53677240e-01 -5.77316582e-01
-3.69502306e-01 7.74726093e-01 3.64927016e-02 -7.98540056e-01
1.09559536e-01 -1.27908379e-01 -3.10149133e-01 2.83634573e-01
4.50744748e-01 3.56415987e-01 5.60197949e-01 -4.21365827e-01
-5.00675030e-02 2.27414563e-01 -2.78020501e-01 4.16428089e-01
1.62243557e+00 3.30063663e-02 -1.32563129e-01 6.00696385e-01
1.32997310e+00 -3.38398784e-01 -9.73291337e-01 -4.05284196e-01
2.52857447e-01 -4.95067835e-01 1.17261983e-01 -8.87699783e-01
-1.57683945e+00 9.61402237e-01 1.22178614e+00 -1.05331965e-01
1.55449009e+00 -1.23002805e-01 7.09136724e-01 2.99247861e-01
2.36509517e-01 -1.50049341e+00 4.11004454e-01 8.56391415e-02
7.23262012e-01 -1.85628235e+00 -5.74824214e-02 -4.45585281e-01
-6.28824770e-01 1.11191928e+00 1.12270272e+00 -1.48991302e-01
6.49396837e-01 4.50729221e-01 1.51963204e-01 -3.16753596e-01
-4.61061150e-01 -4.41538282e-02 -2.01957226e-01 8.87966812e-01
3.42402935e-01 -2.18167305e-01 -1.06876075e-01 5.61713278e-01
1.93733975e-01 3.18359286e-01 5.30753314e-01 1.26045108e+00
-2.39575073e-01 -1.57352006e+00 -4.82841760e-01 9.83735859e-01
-5.52890122e-01 1.44339159e-01 -2.35573933e-01 6.45756781e-01
6.98283613e-01 8.65694046e-01 2.43394613e-01 -2.35935852e-01
2.63551652e-01 1.46322653e-01 3.46868366e-01 -8.46298933e-01
-3.98106307e-01 -2.78513312e-01 -7.94725493e-02 -2.91362703e-01
-6.13912880e-01 -5.74636400e-01 -1.14481914e+00 1.28291354e-01
-4.59006637e-01 -1.78413346e-01 6.61581814e-01 1.04687870e+00
4.37009186e-01 5.99430203e-01 8.32866848e-01 -4.58195299e-01
-3.63720745e-01 -1.06764328e+00 -9.18113708e-01 6.03342533e-01
4.58430290e-01 -8.12015295e-01 -1.55969188e-02 2.67996907e-01]
|
[7.599326133728027, 2.1553118228912354]
|
75fa74e8-5ab2-44ad-912f-dfc44b43f1ee
|
belfusion-latent-diffusion-for-behavior
|
2211.14304
| null |
https://arxiv.org/abs/2211.14304v2
|
https://arxiv.org/pdf/2211.14304v2.pdf
|
BeLFusion: Latent Diffusion for Behavior-Driven Human Motion Prediction
|
Stochastic human motion prediction (HMP) has generally been tackled with generative adversarial networks and variational autoencoders. Most prior works aim at predicting highly diverse movements in terms of the skeleton joints' dispersion. This has led to methods predicting fast and motion-divergent movements, which are often unrealistic and incoherent with past motion. Such methods also neglect contexts that need to anticipate diverse low-range behaviors, or actions, with subtle joint displacements. To address these issues, we present BeLFusion, a model that, for the first time, leverages latent diffusion models in HMP to sample from a latent space where behavior is disentangled from pose and motion. As a result, diversity is encouraged from a behavioral perspective. Thanks to our behavior coupler's ability to transfer sampled behavior to ongoing motion, BeLFusion's predictions display a variety of behaviors that are significantly more realistic than the state of the art. To support it, we introduce two metrics, the Area of the Cumulative Motion Distribution, and the Average Pairwise Distance Error, which are correlated to our definition of realism according to a qualitative study with 126 participants. Finally, we prove BeLFusion's generalization power in a new cross-dataset scenario for stochastic HMP.
|
['Cristina Palmero', 'Sergio Escalera', 'German Barquero']
|
2022-11-25
| null | null | null | null |
['human-pose-forecasting', 'stochastic-human-motion-prediction']
|
['computer-vision', 'computer-vision']
|
[-1.54417679e-01 1.90619320e-01 -3.38588178e-01 -1.02590509e-01
-4.36258525e-01 -3.90157014e-01 7.54390955e-01 -5.35905659e-01
-1.33205533e-01 7.71654129e-01 7.44527102e-01 3.25078696e-01
9.28963944e-02 -6.20653510e-01 -7.88381755e-01 -7.22561002e-01
-8.57884511e-02 3.45308244e-01 1.44873351e-01 -3.76314580e-01
-1.91863090e-01 4.59224969e-01 -1.49376166e+00 1.63611203e-01
7.02934563e-01 5.66435516e-01 -1.92130551e-01 6.77002966e-01
4.76819277e-01 9.46670234e-01 -4.84181523e-01 -5.45754194e-01
2.14978889e-01 -6.26651525e-01 -5.25760889e-01 -1.38734002e-03
2.92260259e-01 -5.91799021e-01 -6.47325814e-01 7.71018386e-01
4.15797025e-01 3.53906751e-01 8.82935703e-01 -1.48077643e+00
-6.45415962e-01 4.43574935e-01 -2.12907270e-01 -1.38665289e-01
6.49869144e-01 7.14202344e-01 1.02719760e+00 -4.96624619e-01
1.18075418e+00 1.31870520e+00 9.30083871e-01 9.44114745e-01
-1.50573635e+00 -3.40523869e-01 7.17684403e-02 1.13495529e-01
-1.15383327e+00 -1.14683799e-01 1.10596633e+00 -7.00112045e-01
5.84666014e-01 3.54113966e-01 1.06280005e+00 2.27310801e+00
4.84496802e-01 1.10215342e+00 7.75873005e-01 1.18667446e-01
5.16993165e-01 -2.42809623e-01 -4.14971948e-01 2.50797719e-01
2.60233041e-02 4.10237819e-01 -4.87754315e-01 7.85803888e-03
8.33609223e-01 8.80224779e-02 -3.38930845e-01 -6.78869367e-01
-1.37890506e+00 9.33311701e-01 3.71086985e-01 1.15805097e-01
-6.05995774e-01 4.03703362e-01 1.70153424e-01 -7.91403651e-02
4.48943675e-01 2.53361970e-01 -1.55606732e-01 -4.62609679e-01
-1.06354690e+00 7.98104882e-01 8.79207373e-01 7.01221943e-01
3.57326984e-01 3.10991466e-01 -3.37406129e-01 2.57241905e-01
3.69066387e-01 6.94053590e-01 6.06753826e-01 -1.33074200e+00
2.86312371e-01 3.53276432e-01 2.76319116e-01 -1.36102962e+00
-4.02790308e-01 -2.51829952e-01 -9.46485519e-01 3.93102348e-01
5.98818064e-01 -2.30610922e-01 -6.77790821e-01 2.18479753e+00
3.16142917e-01 2.28812546e-01 -1.79963391e-02 1.15840375e+00
3.43807548e-01 4.20240819e-01 1.86111838e-01 -4.87994812e-02
7.44491994e-01 -9.17667150e-01 -6.72883451e-01 -4.73863930e-02
4.35579270e-01 -3.49216849e-01 1.18433332e+00 3.61956924e-01
-1.15881383e+00 -6.77209318e-01 -9.26110625e-01 -1.71710607e-02
-7.17751728e-03 -1.80069134e-01 4.93908435e-01 5.58363318e-01
-9.37814116e-01 1.02985573e+00 -1.37585640e+00 -2.75126010e-01
3.10601413e-01 6.81699440e-02 -3.06192487e-01 2.40929872e-01
-1.31042838e+00 8.43627810e-01 -2.82189008e-02 -3.48273069e-02
-9.65799749e-01 -5.97169578e-01 -8.17152798e-01 -1.54935613e-01
1.59887388e-01 -1.09254372e+00 9.42213655e-01 -9.53166068e-01
-1.70057595e+00 3.68836045e-01 1.56920314e-01 -5.78876257e-01
1.24087620e+00 -4.25316304e-01 -2.65510827e-01 2.09899560e-01
6.54927939e-02 9.88635957e-01 8.78751338e-01 -1.15430057e+00
8.47048461e-02 -2.66509116e-01 -1.42776981e-01 2.25947559e-01
-1.22950360e-01 -7.44200170e-01 -2.27538317e-01 -9.04139221e-01
-1.75982535e-01 -1.28833640e+00 -3.59819174e-01 1.38542846e-01
-5.25965869e-01 1.00967132e-01 5.67419827e-01 -7.19598711e-01
1.02178872e+00 -2.00730443e+00 8.18600059e-01 -1.48361223e-02
3.65051121e-01 -8.65339562e-02 -2.79329997e-02 4.99746799e-01
1.58445701e-01 8.14027935e-02 -1.64714828e-01 -7.83153355e-01
4.17757064e-01 3.81777138e-01 -3.85446131e-01 6.01962149e-01
2.14764792e-02 1.25485563e+00 -9.79399085e-01 -4.23900187e-01
3.31610858e-01 6.75752640e-01 -8.93906176e-01 1.42513081e-01
-4.85073775e-01 1.16744435e+00 -3.93401742e-01 3.50767851e-01
2.71038145e-01 -2.43575141e-01 1.81355607e-02 -1.01832204e-01
2.03823492e-01 -2.22122043e-01 -9.70036745e-01 1.98397326e+00
-1.33582652e-01 5.90661168e-01 -4.46645349e-01 -5.80926538e-01
5.87659895e-01 2.63746649e-01 9.10763562e-01 -3.80264938e-01
1.76462129e-01 4.43211012e-03 -1.42987788e-01 -6.63385987e-01
4.50357825e-01 -2.43485942e-01 -1.05684511e-01 3.62711489e-01
-9.20730904e-02 -1.64208606e-01 -1.51656881e-01 -3.30210626e-02
1.05757082e+00 8.36708605e-01 1.08758844e-01 1.00633159e-01
1.02764055e-01 2.14047488e-02 5.78382492e-01 6.43650234e-01
-4.48407859e-01 9.79470372e-01 4.76761460e-01 -3.69218588e-01
-1.25435889e+00 -1.41001356e+00 3.08979869e-01 5.43589532e-01
2.56555915e-01 -2.07895055e-01 -8.53859901e-01 -5.78020036e-01
-1.63567029e-02 7.61088133e-01 -1.01764488e+00 -3.33552122e-01
-7.22102284e-01 -4.82945770e-01 6.69880450e-01 6.31790042e-01
2.18856230e-01 -1.14423728e+00 -8.89747739e-01 1.59261674e-01
-3.87953222e-01 -1.09456599e+00 -4.50234175e-01 -5.32010496e-01
-7.79007375e-01 -8.65970194e-01 -1.17550671e+00 -2.58355960e-02
1.42763093e-01 -3.20792705e-01 1.09994543e+00 -4.06711787e-01
-1.28887504e-01 4.90439534e-01 -4.10170436e-01 5.68627417e-02
-5.92587531e-01 -2.40157894e-03 3.07399482e-01 1.13843046e-01
1.54454529e-01 -1.05066586e+00 -1.01689327e+00 2.95542717e-01
-8.55256319e-01 1.81704417e-01 4.89093632e-01 7.84342885e-01
4.32752997e-01 -4.26010609e-01 2.29566544e-01 -4.72817510e-01
4.79536951e-01 -8.46184790e-01 -6.08332418e-02 -5.20663597e-02
-3.50257277e-01 8.36236328e-02 5.66258073e-01 -1.02306032e+00
-9.01436210e-01 3.50492708e-02 -1.89246267e-01 -8.86775792e-01
-2.53752291e-01 1.58529818e-01 -1.16686568e-01 3.16860020e-01
8.40853155e-01 3.72799367e-01 2.21213043e-01 -1.94848910e-01
7.34399974e-01 -1.11042380e-01 4.86049294e-01 -5.28382480e-01
6.61353648e-01 8.67048502e-01 1.19465537e-01 -7.11356878e-01
-4.32027191e-01 4.20168824e-02 -6.91623330e-01 -5.41152120e-01
1.26283431e+00 -7.89614022e-01 -9.63284552e-01 4.04330343e-01
-1.02794635e+00 -6.80139899e-01 -6.96704865e-01 7.70130575e-01
-1.10943878e+00 5.29780328e-01 -7.46513784e-01 -7.99835801e-01
8.87772217e-02 -1.19071281e+00 1.21154737e+00 -5.46183288e-02
-8.51770639e-01 -1.11302221e+00 4.72824156e-01 1.83323666e-01
3.09485078e-01 1.05973828e+00 4.93728757e-01 -4.48381990e-01
-4.57353622e-01 -1.26419231e-01 4.24158812e-01 2.57406026e-01
-1.08233653e-01 -7.42467167e-03 -6.17561162e-01 -1.17887028e-01
1.01624422e-01 -2.27405220e-01 6.01802826e-01 6.05798960e-01
7.41008043e-01 -4.00250942e-01 -2.23081753e-01 6.43495023e-01
1.02320468e+00 -1.20775007e-01 7.02259660e-01 1.29123732e-01
9.48023498e-01 7.73374736e-01 2.95965999e-01 5.72787046e-01
3.15401018e-01 9.39531386e-01 6.53067112e-01 3.03142011e-01
-2.04265028e-01 -7.42259920e-01 5.60852885e-01 7.06418574e-01
-4.08496976e-01 -3.86391878e-01 -7.33780861e-01 3.75381529e-01
-1.93742180e+00 -1.39007616e+00 -5.77895828e-02 1.85984421e+00
4.66585279e-01 1.57999575e-01 4.96525884e-01 -7.89588690e-02
3.99112016e-01 4.90101010e-01 -7.46844172e-01 -4.24486510e-02
-8.74659866e-02 -1.08303137e-01 2.50417501e-01 4.36183631e-01
-7.68790901e-01 7.45153189e-01 6.35360146e+00 6.42659366e-01
-1.07773006e+00 1.63665518e-01 4.59203333e-01 -3.68779451e-01
-5.89036703e-01 -1.48422331e-01 -3.56806934e-01 7.69703448e-01
8.42031300e-01 1.44410133e-01 2.60662645e-01 8.98380637e-01
3.89017016e-01 1.24551289e-01 -1.28048229e+00 7.65244961e-01
1.77818444e-02 -1.10560179e+00 1.23432942e-01 3.40681016e-01
8.38389039e-01 -1.05943926e-01 4.57208723e-01 4.53293830e-01
5.29551327e-01 -1.25820291e+00 1.02467513e+00 9.86075878e-01
4.53874856e-01 -5.37537098e-01 4.59579587e-01 6.43841088e-01
-7.86365092e-01 1.01934381e-01 -4.13201889e-03 -8.86070654e-02
6.45840704e-01 1.93888500e-01 -3.12328428e-01 5.20092666e-01
4.84682709e-01 8.70269537e-01 -2.30915576e-01 4.95074064e-01
-2.22747341e-01 5.42042792e-01 -3.22415799e-01 -1.52456025e-02
2.73205608e-01 -2.83153743e-01 8.83879900e-01 8.25800300e-01
4.90531832e-01 -1.65951490e-01 3.29771452e-02 1.16642678e+00
2.98486650e-01 -7.81176239e-02 -7.77615845e-01 2.41576321e-02
1.36379629e-01 6.28119171e-01 -4.34763938e-01 -8.65206346e-02
-1.41899168e-01 1.30877125e+00 1.46843508e-01 5.17097056e-01
-1.20077717e+00 3.96943301e-01 8.49900305e-01 1.42887995e-01
1.90059796e-01 -5.51562190e-01 -1.85945630e-01 -1.40982497e+00
1.57318607e-01 -7.06679761e-01 -6.71362951e-02 -7.59919345e-01
-1.25599933e+00 4.88227338e-01 1.93423897e-01 -1.66148353e+00
-8.33578348e-01 -3.24338317e-01 -4.12796259e-01 5.77397227e-01
-7.39718318e-01 -1.31677413e+00 -2.34544128e-01 5.97725153e-01
5.51494539e-01 1.27489999e-01 6.00766480e-01 1.73653036e-01
-4.42800134e-01 5.28110623e-01 -8.08909014e-02 5.94881997e-02
5.22467792e-01 -1.16832924e+00 4.68473971e-01 6.95248365e-01
2.79417485e-01 4.70601022e-01 1.20468009e+00 -7.47847974e-01
-1.23467493e+00 -9.02817249e-01 3.96853060e-01 -8.86387348e-01
6.59225404e-01 -2.30242625e-01 -7.55814373e-01 7.64976919e-01
-2.31166080e-01 1.83147833e-01 5.64404249e-01 -2.32027441e-01
-1.08430423e-01 4.26196724e-01 -9.84019578e-01 1.09153450e+00
1.33118248e+00 -3.13692570e-01 -4.65288103e-01 7.63554275e-02
6.90642834e-01 -4.67171311e-01 -1.06143296e+00 4.18861002e-01
9.77904320e-01 -1.29876411e+00 1.15532815e+00 -6.83374345e-01
8.54860127e-01 3.01002804e-02 -3.19986314e-01 -1.30720496e+00
-1.35138303e-01 -6.70131683e-01 -5.22185028e-01 8.50865245e-01
4.39372919e-02 -2.14739054e-01 1.15884852e+00 5.79115272e-01
1.50004297e-01 -7.73237407e-01 -9.40784872e-01 -8.79807532e-01
3.57452095e-01 -6.17229640e-01 5.93323350e-01 8.78890812e-01
-3.31297427e-01 1.80180930e-02 -1.07115650e+00 -7.05395192e-02
6.23322904e-01 -1.13247037e-01 1.06969666e+00 -9.97392058e-01
-7.04994500e-01 -5.76812327e-01 -6.51927650e-01 -1.30752158e+00
2.94907659e-01 -5.47192335e-01 7.92100728e-02 -1.18750346e+00
-9.40600187e-02 1.26243517e-01 1.77527651e-01 -3.67297865e-02
-6.11773459e-03 2.22546369e-01 2.47912809e-01 5.73930562e-01
-4.38751519e-01 1.12249231e+00 1.65945840e+00 -5.17851338e-02
-2.92991191e-01 1.24147289e-01 -1.05266906e-01 9.29800987e-01
6.04776382e-01 -3.56308162e-01 -6.47510707e-01 -1.67442128e-01
1.94749981e-01 4.64806050e-01 6.97983563e-01 -1.23570681e+00
2.13914625e-02 -3.84547681e-01 4.58175480e-01 -3.55573118e-01
7.90522993e-01 -6.45157397e-01 6.51747704e-01 6.69967651e-01
-4.75335598e-01 -2.49294490e-02 -1.35173440e-01 1.00686479e+00
-8.78108963e-02 3.45604211e-01 4.40153301e-01 -7.00424090e-02
-5.87066174e-01 4.93739069e-01 -4.10461843e-01 1.03595242e-01
1.08890867e+00 -4.30535674e-01 5.21971695e-02 -8.05158079e-01
-1.23957229e+00 -6.69820234e-02 7.50187099e-01 5.00178218e-01
4.68106031e-01 -1.60498536e+00 -6.63620412e-01 4.87836450e-03
-1.20007848e-04 -9.76082683e-02 7.08621740e-01 1.13216913e+00
-6.12879276e-01 3.47436517e-02 -4.09869552e-01 -8.20949197e-01
-7.70104587e-01 6.64381087e-01 2.04663992e-01 -3.04242611e-01
-9.16845977e-01 4.49564546e-01 9.38507169e-02 -2.12589756e-01
-8.13198686e-02 -2.92939693e-01 -4.92910109e-02 -1.45463675e-01
6.92661181e-02 5.16884923e-01 -6.48232937e-01 -1.01723909e+00
-1.64483592e-01 4.53906476e-01 4.04471189e-01 -4.82747406e-01
1.13624072e+00 -2.17804939e-01 5.00210404e-01 7.22810388e-01
1.18585873e+00 1.86285332e-01 -2.08682084e+00 2.35824242e-01
-1.71908036e-01 -3.12405139e-01 -5.67631602e-01 -4.69746977e-01
-9.67242360e-01 8.26225042e-01 5.02858877e-01 1.38241008e-01
6.81171596e-01 -8.73019695e-02 9.60699022e-01 -7.72355199e-02
3.68518353e-01 -9.10389364e-01 4.34838772e-01 1.28722966e-01
9.30095375e-01 -1.27642703e+00 -2.20066324e-01 -1.89429253e-01
-9.69960511e-01 8.12111855e-01 5.07005215e-01 -4.31987882e-01
5.96350372e-01 -3.02221403e-02 -5.99049143e-02 6.10249750e-02
-6.61875248e-01 -9.59376916e-02 4.50927109e-01 7.30609119e-01
9.10652652e-02 2.07691386e-01 -1.96885869e-01 6.39166296e-01
-4.78428215e-01 1.29542202e-01 3.08715403e-01 6.36203408e-01
-2.95194294e-02 -8.68477225e-01 -2.02451870e-01 -1.79846212e-01
-3.45193148e-01 4.06248748e-01 -2.03884616e-01 1.16731179e+00
1.41305536e-01 6.00374937e-01 -5.57113700e-02 -6.48151278e-01
1.76735416e-01 -2.04694550e-02 5.25785267e-01 -1.43596143e-01
-2.10854501e-01 -2.81958953e-02 -2.77800113e-01 -1.04749346e+00
-4.70503539e-01 -7.90666699e-01 -8.83441925e-01 -4.59639937e-01
3.60309154e-01 -2.81127453e-01 4.17568594e-01 9.02270257e-01
2.44410038e-01 5.29569507e-01 2.33004034e-01 -1.32611334e+00
-8.02349091e-01 -9.40083802e-01 -4.54561114e-01 1.13622379e+00
3.39500397e-01 -8.12475443e-01 -5.54531097e-01 9.97538567e-02]
|
[7.29479455947876, -0.12489165365695953]
|
cd547836-a6a0-4723-8489-efa067f67a4c
|
ots-a-one-shot-learning-approach-for-text
|
2304.00746
| null |
https://arxiv.org/abs/2304.00746v2
|
https://arxiv.org/pdf/2304.00746v2.pdf
|
OTS: A One-shot Learning Approach for Text Spotting in Historical Manuscripts
|
Historical manuscript processing poses challenges like limited annotated training data and novel class emergence. To address this, we propose a novel One-shot learning-based Text Spotting (OTS) approach that accurately and reliably spots novel characters with just one annotated support sample. Drawing inspiration from cognitive research, we introduce a spatial alignment module that finds, focuses on, and learns the most discriminative spatial regions in the query image based on one support image. Especially, since the low-resource spotting task often faces the problem of example imbalance, we propose a novel loss function called torus loss which can make the embedding space of distance metric more discriminative. Our approach is highly efficient and requires only a few training samples while exhibiting the remarkable ability to handle novel characters, and symbols. To enhance dataset diversity, a new manuscript dataset that contains the ancient Dongba hieroglyphics (DBH) is created. We conduct experiments on publicly available VML-HD, TKH, NC datasets, and the new proposed DBH dataset. The experimental results demonstrate that OTS outperforms the state-of-the-art methods in one-shot text spotting. Overall, our proposed method offers promising applications in the field of text spotting in historical manuscripts.
|
['Hongjian Zhan', 'WenBo Hu', 'Yue Lu', 'Bing Yin', 'Cong Liu']
|
2023-04-03
| null | null | null | null |
['text-spotting', 'one-shot-learning']
|
['computer-vision', 'methodology']
|
[ 2.96490014e-01 -4.54126894e-01 -1.16153039e-01 -3.84007514e-01
-9.33331549e-01 -4.08095688e-01 6.06558144e-01 2.49437302e-01
-5.95436931e-01 7.49841869e-01 1.75887626e-02 -1.73782520e-02
-3.70109409e-01 -7.56675422e-01 -6.87108815e-01 -7.08273888e-01
1.91076458e-01 5.22891462e-01 3.02730232e-01 -4.55207340e-02
9.10843432e-01 3.79783154e-01 -1.44796038e+00 2.43760049e-01
1.04123509e+00 7.38565564e-01 7.86481380e-01 5.16484439e-01
-4.21889514e-01 2.62967467e-01 -6.96284115e-01 -6.05323315e-01
2.64829576e-01 -1.90392449e-01 -4.66893315e-01 -5.57104126e-02
7.59007871e-01 -1.64660811e-01 -5.26311517e-01 7.80180395e-01
1.06071925e+00 3.67269903e-01 6.80056155e-01 -7.96483696e-01
-9.87415791e-01 6.07944369e-01 -7.85293698e-01 5.00522137e-01
1.98048860e-01 4.34522610e-03 1.09495699e+00 -1.60058296e+00
1.05979776e+00 9.75515902e-01 5.90359330e-01 3.52337062e-01
-8.84743750e-01 -5.39828777e-01 3.86081859e-02 6.46872044e-01
-1.72036874e+00 -4.75480437e-01 9.03895557e-01 -3.90384346e-01
4.10388172e-01 3.72916281e-01 4.72011805e-01 1.05599642e+00
7.70857036e-02 1.12261629e+00 8.14198732e-01 -5.80741584e-01
5.19828498e-01 -9.45688225e-03 1.56122983e-01 6.83834136e-01
2.00110897e-01 -4.75281984e-01 -1.09612489e+00 3.88429649e-02
5.03484309e-01 8.64008516e-02 -6.81302249e-02 -2.61617631e-01
-1.47616220e+00 5.32606721e-01 2.30911538e-01 2.93636501e-01
7.32404068e-02 -3.12975913e-01 4.40654427e-01 1.10936329e-01
3.57732236e-01 5.25820315e-01 7.32282102e-02 -3.40563923e-01
-1.49038506e+00 2.72937387e-01 3.47090006e-01 1.10910618e+00
5.78119457e-01 -1.84481859e-01 -5.91861665e-01 1.27027178e+00
-3.11810523e-01 5.35238624e-01 4.15868908e-01 -3.11546803e-01
6.99329674e-01 5.55103362e-01 -1.77867994e-01 -1.31345737e+00
-3.01395237e-01 -3.33559036e-01 -6.93672359e-01 -3.30446839e-01
1.35450259e-01 2.50969201e-01 -8.73244286e-01 1.16125441e+00
1.23945676e-01 1.06360599e-01 -2.95203954e-01 9.94946063e-01
7.92134345e-01 7.87443459e-01 -4.69437391e-01 -1.43247411e-01
1.24949586e+00 -9.57468092e-01 -8.91511381e-01 -2.50425786e-01
5.11578739e-01 -9.41478968e-01 1.52966619e+00 3.51455808e-01
-8.64495933e-01 -5.22418857e-01 -1.39791095e+00 -4.34765458e-01
-6.31167769e-01 5.02903521e-01 3.23061258e-01 4.66930926e-01
-4.13126081e-01 5.54611862e-01 -4.17533487e-01 -5.51613152e-01
7.46413648e-01 -4.16364521e-02 -1.23999722e-01 -2.94982791e-01
-8.24778914e-01 7.13477731e-01 5.07785618e-01 1.42101973e-01
-6.67611182e-01 -8.36253047e-01 -6.56542718e-01 9.27936360e-02
5.60773373e-01 -2.00353399e-01 5.63590109e-01 -3.64347696e-01
-1.29406166e+00 1.04421020e+00 -1.53560683e-01 -1.71389982e-01
6.47718906e-01 -3.22704256e-01 -4.65207607e-01 3.14736009e-01
3.24423850e-01 5.64944625e-01 8.74431908e-01 -8.42842638e-01
-4.09337103e-01 -5.20300090e-01 -6.28093421e-01 1.16650410e-01
-8.82933676e-01 -8.92637484e-03 -6.70127392e-01 -1.26479959e+00
4.63135839e-02 -6.08240843e-01 1.79827780e-01 5.98866820e-01
-6.33906245e-01 -1.36538580e-01 9.21251655e-01 -7.23707378e-01
1.49244952e+00 -2.18185115e+00 2.72342235e-01 2.40347385e-01
8.51781517e-02 1.64109662e-01 -1.68153390e-01 4.12721366e-01
3.57963026e-01 -8.30189735e-02 -5.15192091e-01 -3.33535075e-01
5.82504012e-02 -4.48804423e-02 -5.86043060e-01 4.96779919e-01
2.04236135e-02 9.25433517e-01 -7.09632695e-01 -1.03123856e+00
1.35880515e-01 -1.52869588e-02 -1.46014705e-01 -9.28859189e-02
-3.00332904e-01 1.13255039e-01 -3.11132967e-01 9.76914346e-01
8.62231135e-01 -2.59775609e-01 8.05948004e-02 -4.48016040e-02
-2.75684565e-01 -1.97054490e-01 -1.13225484e+00 2.36495161e+00
1.18495986e-01 1.02831602e+00 -6.36184633e-01 -8.84255767e-01
1.35806775e+00 -3.88594419e-01 1.57685295e-01 -1.12999225e+00
-1.35406956e-01 3.84266526e-01 -3.89212966e-01 -6.80772305e-01
8.63573670e-01 2.20808432e-01 -2.38509148e-01 4.55068797e-01
8.42669681e-02 2.30639666e-01 4.40072864e-01 4.10308838e-01
9.22629058e-01 1.42500073e-01 1.42988965e-01 -2.97353446e-01
3.49284679e-01 4.07054126e-02 7.15348065e-01 9.42600250e-01
-1.70159861e-01 7.67630994e-01 3.26023012e-01 -7.58635104e-01
-1.51440501e+00 -1.06110835e+00 -4.23128694e-01 1.28460300e+00
5.01046777e-01 -5.62960386e-01 -4.88518059e-01 -4.88001019e-01
1.21304393e-01 5.83071828e-01 -6.31848693e-01 -4.32801098e-02
-8.43487382e-01 -8.25491846e-01 7.94018090e-01 3.34826261e-01
5.21800280e-01 -9.30685937e-01 -4.20834541e-01 -1.05003566e-02
-7.82846361e-02 -8.91821861e-01 -7.30524123e-01 -1.50905192e-01
-6.13661408e-01 -7.54479349e-01 -1.27189744e+00 -1.18550932e+00
7.97732592e-01 2.83216327e-01 5.37937880e-01 -9.55003574e-02
-9.42564070e-01 -8.75014141e-02 -3.82211208e-01 -2.79094577e-01
3.79418999e-01 1.93802029e-01 -1.23770128e-03 9.67242420e-02
3.39742571e-01 -1.42684817e-01 -6.97395623e-01 2.90334016e-01
-8.32027137e-01 1.54580429e-01 6.69648826e-01 1.03934789e+00
6.76160932e-01 -2.60240465e-01 5.69786549e-01 -6.48199677e-01
5.24194002e-01 -2.72021919e-01 -3.91009748e-01 7.64530063e-01
-5.60018480e-01 1.76712796e-02 6.79988027e-01 -3.64985168e-01
-9.80382085e-01 -2.26576775e-01 4.25278395e-01 -2.54195422e-01
4.00410056e-01 4.69596326e-01 -1.73997983e-01 -1.53665751e-01
6.00563467e-01 7.14987457e-01 -2.02823967e-01 -7.38470197e-01
4.40959722e-01 8.54520917e-01 8.22981477e-01 -6.10058665e-01
7.93921649e-01 5.87531805e-01 -1.03826225e-01 -1.08366978e+00
-7.31405139e-01 -3.98703218e-01 -9.34387922e-01 -2.78665453e-01
4.59688812e-01 -6.29990399e-01 -4.16847080e-01 3.82109553e-01
-9.11712825e-01 1.11981273e-01 -1.69482306e-01 2.08140790e-01
-4.35124457e-01 7.83438921e-01 -4.88164872e-01 -6.86536908e-01
-5.47558486e-01 -7.50026822e-01 1.20093155e+00 3.59138757e-01
-8.66432339e-02 -6.25077724e-01 2.38995299e-01 2.62610048e-01
1.71015948e-01 1.56906024e-02 1.17721844e+00 -6.42492890e-01
-7.47178137e-01 -1.69384390e-01 -3.20031166e-01 -3.24879825e-01
-9.28320959e-02 -8.60027373e-02 -7.72485137e-01 -4.02119815e-01
-3.94042850e-01 -1.84310257e-01 1.13182294e+00 6.76397607e-02
1.52156639e+00 -1.32031187e-01 -4.09422040e-01 8.11372161e-01
1.11700511e+00 8.51853862e-02 7.25733221e-01 7.64573574e-01
6.79544926e-01 5.03987134e-01 1.00804865e+00 9.10002410e-01
1.65188834e-01 8.15199375e-01 -1.41971231e-01 -2.88745482e-02
-1.03723183e-01 -3.21941495e-01 2.54183672e-02 9.53483820e-01
4.37942684e-01 -3.38274837e-01 -8.89231324e-01 6.31334305e-01
-2.03774691e+00 -1.09747279e+00 2.55243599e-01 2.12665844e+00
1.01187205e+00 3.30944508e-02 -6.45831823e-02 2.77994066e-01
9.11313713e-01 4.00297165e-01 -6.30142868e-01 -7.46422932e-02
-8.10378075e-01 2.06013054e-01 3.00942004e-01 9.23742503e-02
-1.14317536e+00 1.15731061e+00 5.79133701e+00 1.42530084e+00
-9.88065302e-01 -6.55183941e-02 5.68162799e-01 -5.38341403e-01
-2.16960311e-01 -4.41031419e-02 -8.73663664e-01 9.69909847e-01
3.39852720e-01 -3.61319244e-01 1.93010643e-01 6.09360814e-01
1.72796726e-01 5.33485524e-02 -8.94555211e-01 1.37429869e+00
6.31636083e-01 -1.91933382e+00 4.14370775e-01 -2.60630101e-01
5.83943486e-01 -6.19863153e-01 1.66941196e-01 1.72728784e-02
-4.64104414e-01 -1.04187357e+00 7.62703598e-01 9.18428242e-01
1.07289827e+00 -8.27510774e-01 3.20502639e-01 1.41243860e-01
-9.25379813e-01 -7.01179206e-02 -7.42496490e-01 1.55715838e-01
-5.01057245e-02 7.49990225e-01 -8.16311836e-01 3.94899338e-01
5.66579223e-01 1.02660990e+00 -1.04710877e+00 1.48994136e+00
8.67032558e-02 2.20764697e-01 -1.21890031e-01 -5.70379138e-01
1.49801522e-02 -1.34550184e-01 6.93157196e-01 1.23872399e+00
5.32426238e-01 -1.47501677e-01 1.92271903e-01 8.87114584e-01
-2.93106586e-01 5.47075868e-01 -4.86860931e-01 -9.64671895e-02
8.83426905e-01 1.17951357e+00 -9.59283471e-01 -3.39506209e-01
1.85086369e-03 1.39340293e+00 4.05348182e-01 1.96318820e-01
-6.23831153e-01 -1.09378040e+00 2.19076663e-01 -1.19223773e-01
5.44924200e-01 -2.22818062e-01 -7.74218976e-01 -1.20974839e+00
6.10479832e-01 -6.41529620e-01 4.98392016e-01 -7.52490222e-01
-1.24943590e+00 9.20046642e-02 -4.41137314e-01 -1.41756773e+00
5.77024519e-01 -5.94062865e-01 -7.35154569e-01 5.00577629e-01
-1.19056594e+00 -1.22718000e+00 -3.89578611e-01 2.71626681e-01
9.88670290e-01 -6.57238424e-01 4.81581092e-01 5.01866162e-01
-1.04297972e+00 9.87435281e-01 7.28977561e-01 1.74609110e-01
1.15459311e+00 -1.00640988e+00 5.83802521e-01 1.02201700e+00
2.39933118e-01 7.59126544e-01 5.13978124e-01 -9.46820259e-01
-1.71789205e+00 -9.37138140e-01 8.56033623e-01 -2.50216454e-01
4.83216316e-01 -7.75730491e-01 -9.38748777e-01 2.98168510e-01
-1.93755358e-01 -1.76841646e-01 8.43567133e-01 -2.60776818e-01
-2.97932923e-01 -2.14158252e-01 -8.63954365e-01 8.45405817e-01
1.14622867e+00 -5.43039441e-01 -7.12557256e-01 5.60369551e-01
4.61014956e-01 -2.44182259e-01 -6.47379935e-01 8.24284554e-02
6.97407603e-01 -4.53556269e-01 9.71425176e-01 -3.82090837e-01
4.52435851e-01 -3.89937311e-01 -1.42878786e-01 -7.58605599e-01
-3.45265508e-01 -7.51857996e-01 -1.21510133e-01 1.23925543e+00
3.87241304e-01 -3.16755064e-02 8.67246926e-01 1.55449092e-01
-2.66489595e-01 -7.43131578e-01 -1.23097241e+00 -8.59546363e-01
-8.12132843e-04 -8.72922540e-02 4.84418482e-01 1.09037375e+00
1.52043059e-01 1.15860388e-01 -7.62535036e-01 -2.95400143e-01
7.93091357e-01 3.46449435e-01 8.02896976e-01 -1.24086511e+00
9.33950618e-02 -4.62066263e-01 -5.57949305e-01 -1.15922332e+00
-1.37578458e-01 -1.07495713e+00 1.35320470e-01 -1.24223936e+00
4.31267232e-01 -4.06880915e-01 -2.08257437e-01 1.97896004e-01
-3.68305326e-01 1.07962579e-01 1.60356060e-01 5.72226882e-01
-1.01200628e+00 9.27724242e-01 1.06326532e+00 -3.63048643e-01
5.43818623e-02 -5.64235806e-01 -4.38359201e-01 2.94987857e-01
4.28415000e-01 -3.62226248e-01 2.68279463e-02 -5.47633350e-01
3.77312571e-01 -4.61269706e-01 1.96165547e-01 -9.90346670e-01
7.74964392e-01 -8.27839002e-02 8.76179576e-01 -1.15591633e+00
2.75225282e-01 -2.75929719e-01 -4.16561723e-01 1.41413912e-01
-5.35667658e-01 -6.13556504e-02 7.99466204e-03 8.08631599e-01
2.37925351e-01 -3.52675170e-01 6.40920997e-01 6.48291036e-02
-1.17134798e+00 1.82026073e-01 3.81612778e-02 1.47880062e-01
1.12776315e+00 -5.69068253e-01 -7.47452736e-01 2.99586624e-01
-3.32950920e-01 4.09004718e-01 5.83066344e-01 5.73953629e-01
1.09282804e+00 -1.32516325e+00 -6.89788163e-01 2.68307149e-01
5.78052223e-01 -1.81610629e-01 3.10280561e-01 6.81994915e-01
-7.84412324e-01 3.90333980e-01 -4.22138423e-01 -5.70289671e-01
-1.06542575e+00 5.30448735e-01 -2.71694869e-01 2.81122804e-01
-1.09005773e+00 9.02768791e-01 -1.68296292e-01 -4.36915040e-01
5.21150529e-01 2.99601883e-01 -8.06092322e-02 2.82524288e-01
7.92626321e-01 7.64135003e-01 1.61350947e-02 -2.60010630e-01
-4.08164769e-01 8.52160215e-01 -4.37690079e-01 -8.40343311e-02
1.30163157e+00 -1.20152041e-01 -1.05947472e-01 6.57285511e-01
1.06334233e+00 7.46105835e-02 -9.31956291e-01 -6.08727038e-01
5.45892715e-01 -9.15159285e-01 -1.93957776e-01 -6.46363437e-01
-6.02078259e-01 8.78024638e-01 6.82264328e-01 -5.25295019e-01
8.13905895e-01 -2.28599325e-01 1.02964497e+00 7.42528856e-01
8.35488737e-02 -1.78964460e+00 5.03142655e-01 4.51095402e-01
8.49011242e-01 -1.18102336e+00 3.36912841e-01 -3.11565939e-02
-5.01296222e-01 1.42239988e+00 5.96555233e-01 5.48706613e-02
2.11480543e-01 1.24045506e-01 -2.40232453e-01 -3.40902925e-01
-4.55620944e-01 8.86568949e-02 2.25036815e-01 2.54780054e-01
5.96904829e-02 -1.57383859e-01 -4.00898010e-01 5.21774590e-01
-3.09019923e-01 -1.47563428e-01 4.27027702e-01 1.34560347e+00
-7.49753952e-01 -8.27750862e-01 -3.68836045e-01 6.08902276e-01
-1.30176634e-01 -2.57289141e-01 -6.41097307e-01 4.06896949e-01
-1.21809587e-01 4.23313826e-01 2.81813353e-01 -2.97668755e-01
5.28042726e-02 6.18275441e-02 3.02629411e-01 -4.31505829e-01
-3.80338356e-02 -4.19529416e-02 -2.73002177e-01 -7.94685706e-02
1.09255902e-01 -5.42183697e-01 -1.11620057e+00 -3.84290516e-01
-3.44776362e-01 -9.68699306e-02 6.64443672e-01 8.26525509e-01
6.06378734e-01 2.92773902e-01 7.35674202e-01 -5.88582039e-01
-3.06870610e-01 -9.10926044e-01 -6.26078844e-01 3.28757346e-01
9.41342637e-02 -6.26639366e-01 5.56123862e-03 7.16935098e-02]
|
[11.799678802490234, 2.285433530807495]
|
299edeb9-5c5b-4c99-bab2-774ed0c3e382
|
deepdeform-learning-non-rigid-rgb-d
|
1912.04302
| null |
https://arxiv.org/abs/1912.04302v2
|
https://arxiv.org/pdf/1912.04302v2.pdf
|
DeepDeform: Learning Non-rigid RGB-D Reconstruction with Semi-supervised Data
|
Applying data-driven approaches to non-rigid 3D reconstruction has been difficult, which we believe can be attributed to the lack of a large-scale training corpus. Unfortunately, this method fails for important cases such as highly non-rigid deformations. We first address this problem of lack of data by introducing a novel semi-supervised strategy to obtain dense inter-frame correspondences from a sparse set of annotations. This way, we obtain a large dataset of 400 scenes, over 390,000 RGB-D frames, and 5,533 densely aligned frame pairs; in addition, we provide a test set along with several metrics for evaluation. Based on this corpus, we introduce a data-driven non-rigid feature matching approach, which we integrate into an optimization-based reconstruction pipeline. Here, we propose a new neural network that operates on RGB-D frames, while maintaining robustness under large non-rigid deformations and producing accurate predictions. Our approach significantly outperforms existing non-rigid reconstruction methods that do not use learned data terms, as well as learning-based approaches that only use self-supervision.
|
['Matthias Nießner', 'Christian Theobalt', 'Michael Zollhöfer', 'Aljaž Božič']
|
2019-12-09
| null | null | null | null |
['rgb-d-reconstruction']
|
['computer-vision']
|
[ 2.93491453e-01 7.95990750e-02 4.18124013e-02 -6.15716219e-01
-1.22850025e+00 -5.61610579e-01 5.86820543e-01 -3.21664572e-01
-3.39649141e-01 4.45493281e-01 5.91507435e-01 1.92682639e-01
1.36440881e-02 -5.19216299e-01 -1.01208079e+00 -4.81096894e-01
3.38655710e-01 6.55574739e-01 5.32846868e-01 -1.99947417e-01
1.52624503e-01 6.98941588e-01 -1.52587903e+00 1.52974606e-01
3.25334758e-01 8.14538419e-01 7.86636621e-02 4.31652814e-01
3.84294271e-01 6.13958478e-01 -6.39318824e-02 -2.80575365e-01
4.01149362e-01 -3.68265867e-01 -9.16453123e-01 3.14675540e-01
7.27671802e-01 -6.44644380e-01 -4.30561393e-01 6.58674002e-01
6.95575476e-01 3.71747226e-01 3.51912975e-01 -5.71451306e-01
-3.10629338e-01 1.32583827e-01 -4.42716897e-01 -4.74894457e-02
8.16436231e-01 7.99371228e-02 9.60434020e-01 -1.01119149e+00
1.24534047e+00 1.16096103e+00 9.45427120e-01 7.51366735e-01
-1.39408565e+00 -2.46888667e-01 -1.85378522e-01 -5.97868264e-02
-1.29380786e+00 -8.41141284e-01 1.07898402e+00 -4.62002695e-01
1.13733077e+00 1.94564342e-01 7.26054728e-01 1.14891624e+00
-2.29213268e-01 5.15131891e-01 9.65627611e-01 -5.20506740e-01
6.15251735e-02 -3.19718689e-01 -4.04594809e-01 7.00198352e-01
-2.02927932e-01 3.79214436e-01 -4.98467773e-01 -9.61180031e-03
9.34015691e-01 -5.23413010e-02 -3.71143639e-01 -7.22107708e-01
-1.49271524e+00 5.01962245e-01 4.75229263e-01 2.90261686e-01
-2.06086352e-01 4.16323692e-02 1.97277054e-01 1.76316068e-01
5.52914321e-01 2.64105052e-01 -5.54319263e-01 -1.97465971e-01
-1.07949030e+00 3.38182420e-01 7.48016775e-01 7.59894550e-01
9.60735857e-01 -3.42775844e-02 2.11463422e-01 6.29738748e-01
4.07691419e-01 4.36739802e-01 4.16225612e-01 -1.28590608e+00
3.48168910e-01 4.28459823e-01 7.65827820e-02 -9.41639721e-01
-5.31630814e-01 1.38628542e-01 -5.36275804e-01 1.33885220e-01
3.86254221e-01 3.33278924e-01 -9.06980336e-01 1.69864190e+00
5.76164901e-01 1.54175416e-01 3.88246551e-02 1.18156874e+00
8.14757407e-01 1.43450379e-01 -3.29241484e-01 -9.69983488e-02
8.98117542e-01 -1.01548707e+00 -4.78630453e-01 2.02731956e-02
4.24619138e-01 -9.07203197e-01 1.09531951e+00 1.34457737e-01
-1.20921528e+00 -5.10619760e-01 -7.23839462e-01 -3.33187371e-01
7.37530291e-02 -1.29218444e-01 3.17040026e-01 2.89063215e-01
-1.08348191e+00 8.57789993e-01 -1.11086512e+00 -5.46305478e-01
3.98424298e-01 3.76777977e-01 -9.18340266e-01 -2.06414074e-01
-6.48004413e-01 1.05114090e+00 2.68804431e-01 -1.46858832e-02
-9.03153539e-01 -5.04127681e-01 -1.14829540e+00 -4.65180755e-01
3.51030171e-01 -7.36052334e-01 1.07682753e+00 -9.28170979e-01
-1.74759388e+00 1.23620939e+00 -2.04856664e-01 -1.27785102e-01
5.70563138e-01 -3.84492785e-01 -1.70744844e-02 2.88447261e-01
-7.31738331e-03 7.00248897e-01 8.12883973e-01 -1.53841484e+00
-1.40581772e-01 -3.12203348e-01 -2.95129716e-02 2.20703512e-01
-1.05467103e-01 1.57098264e-01 -6.74371839e-01 -7.96928883e-01
4.85197693e-01 -1.24796581e+00 -4.14632916e-01 7.09146708e-02
-5.39664686e-01 1.00581601e-01 6.60753429e-01 -6.11746013e-01
5.93421340e-01 -2.17163467e+00 5.21170855e-01 2.10210308e-01
4.03845683e-02 2.00551841e-03 -1.60358652e-01 1.30601764e-01
-2.28604391e-01 -2.22483054e-01 -3.88884008e-01 -7.02479482e-01
-5.81878498e-02 4.61107165e-01 -2.30684295e-01 6.26986265e-01
3.60971898e-01 8.91765237e-01 -7.73239911e-01 -5.30423760e-01
5.20776629e-01 6.94368243e-01 -8.41570854e-01 5.21555305e-01
-2.60238081e-01 9.32926118e-01 -3.62229168e-01 8.58419418e-01
2.81524301e-01 -2.64937580e-01 -2.07256731e-02 -5.96985519e-01
2.62058545e-02 3.12022924e-01 -1.13631511e+00 2.56077552e+00
-2.00116366e-01 4.46515292e-01 -1.89801633e-01 -9.99846518e-01
8.27964902e-01 3.00031930e-01 9.30097103e-01 -4.98125970e-01
1.57943502e-01 4.11897480e-01 -3.36299390e-01 -5.77814400e-01
3.81584048e-01 -2.92987913e-01 1.61372080e-01 6.39118135e-01
2.43837297e-01 -5.18711329e-01 -1.59634743e-02 -1.51990503e-01
1.35187912e+00 1.02355778e+00 1.59283713e-01 9.24644619e-02
4.72874701e-01 1.01424858e-01 6.32587612e-01 2.35573292e-01
-1.59726024e-01 1.33724129e+00 -5.06537519e-02 -8.45012367e-01
-1.24891651e+00 -8.98252189e-01 -2.23099172e-01 6.93110526e-01
1.48513243e-01 -4.54019070e-01 -7.04876840e-01 -4.54773992e-01
-6.34333342e-02 2.37232968e-01 -5.12345195e-01 -1.23498216e-03
-9.10752833e-01 -5.65273643e-01 2.33830526e-01 4.80122685e-01
2.11540416e-01 -1.09099627e+00 -5.34217477e-01 1.56342596e-01
-2.99125016e-01 -1.51649678e+00 -4.33383882e-01 1.33498877e-01
-1.08867800e+00 -1.16845322e+00 -5.93548298e-01 -5.37322283e-01
8.45202625e-01 2.44558454e-01 1.38908446e+00 6.74106181e-02
-2.53854036e-01 5.96182406e-01 -4.33569610e-01 8.89305547e-02
-5.05518615e-01 7.58936703e-02 2.11954653e-01 -2.24327415e-01
2.33490840e-01 -8.37408006e-01 -4.65129912e-01 5.69424689e-01
-8.21618676e-01 7.76255503e-02 4.93450552e-01 7.16031849e-01
1.11455119e+00 -4.65884596e-01 3.63719203e-02 -6.02286279e-01
-6.58790842e-02 -1.59729421e-01 -5.10470867e-01 1.04072049e-01
-2.76960999e-01 1.28743783e-01 2.62605220e-01 -4.20022577e-01
-9.23339307e-01 7.15421796e-01 -3.78343552e-01 -8.21356773e-01
-2.93234617e-01 1.86921939e-01 -1.37628317e-01 -3.72420490e-01
7.89727211e-01 1.06508315e-01 2.12859899e-01 -7.15882897e-01
3.21896136e-01 3.05375576e-01 7.33446956e-01 -7.51306534e-01
1.18838561e+00 7.06336439e-01 7.01729432e-02 -5.39109528e-01
-1.18234932e+00 -3.38474154e-01 -1.09823596e+00 -1.44275174e-01
8.03678036e-01 -1.16412270e+00 -3.09514850e-01 2.68125534e-01
-1.12938392e+00 -4.88175839e-01 -6.51770473e-01 7.69044220e-01
-1.13798952e+00 4.74830449e-01 -5.71460724e-01 -2.75103986e-01
-2.13217318e-01 -1.39825928e+00 1.53979051e+00 -1.37185916e-01
-2.59175837e-01 -7.80312538e-01 4.70401019e-01 6.31981552e-01
2.73748517e-01 6.40492201e-01 3.65685165e-01 -5.77329993e-01
-7.72634149e-01 -2.54929066e-01 -1.31330602e-02 2.88823813e-01
3.00339758e-01 -6.28103837e-02 -1.00012124e+00 -4.10681814e-01
4.16350476e-02 -7.66183794e-01 5.78723311e-01 1.78220674e-01
1.13317370e+00 1.05080336e-01 -4.83975746e-02 1.06138682e+00
1.35987163e+00 -5.50473869e-01 7.08689511e-01 5.68687677e-01
1.01864064e+00 6.16350710e-01 5.21548033e-01 2.17172444e-01
6.57400548e-01 1.05571830e+00 6.81935012e-01 -1.37049615e-01
-1.32282630e-01 -1.71906829e-01 2.86561608e-01 1.08729684e+00
-7.26045489e-01 1.49957940e-01 -7.53911138e-01 3.62825960e-01
-1.72110415e+00 -8.10279071e-01 1.12732999e-01 2.17538047e+00
1.12149155e+00 2.87909601e-02 8.44124109e-02 1.66537642e-01
3.64144683e-01 8.82724449e-02 -6.44479632e-01 1.02641083e-01
-2.06809506e-01 2.90340662e-01 2.41227731e-01 4.68571901e-01
-1.11210454e+00 1.00457656e+00 6.65888023e+00 3.72814447e-01
-1.04406929e+00 7.05461800e-02 2.48654678e-01 -2.01195836e-01
-4.59518224e-01 1.00405678e-01 -5.16130269e-01 2.26632088e-01
8.24868381e-01 2.71357745e-01 5.08183718e-01 8.31637442e-01
2.49844864e-01 1.00221187e-01 -1.32783961e+00 1.07789397e+00
3.86607677e-01 -1.43799210e+00 -9.35405940e-02 2.35812087e-02
7.80120730e-01 4.27222550e-01 -5.93921542e-01 -2.55192280e-01
3.11962843e-01 -8.76080394e-01 7.86986470e-01 6.92240238e-01
8.40396523e-01 -4.10290420e-01 5.30210853e-01 1.72115311e-01
-9.46002603e-01 3.71042460e-01 -3.23758096e-01 2.25586563e-01
4.20621127e-01 6.08394623e-01 -1.49705485e-01 7.58258402e-01
1.02876890e+00 1.13581812e+00 -3.55723828e-01 7.42954195e-01
-2.42679402e-01 2.04524472e-01 -5.13349891e-01 6.53677881e-01
-1.36162534e-01 -9.48893726e-02 6.36028826e-01 9.40940797e-01
4.06129837e-01 1.98522761e-01 1.87450424e-01 6.06864393e-01
-1.08195737e-01 -3.90261561e-02 -7.33003378e-01 4.78101581e-01
2.38060907e-01 1.29402030e+00 -5.40525734e-01 -1.36916742e-01
-5.89601696e-01 1.10074186e+00 5.37048697e-01 1.32410958e-01
-6.76690340e-01 1.85689643e-01 5.06744027e-01 2.31971428e-01
2.10631028e-01 -5.36892354e-01 1.29942775e-01 -1.53794098e+00
2.45482042e-01 -1.02981651e+00 2.07441077e-01 -9.60268915e-01
-1.20332313e+00 7.12549627e-01 -2.48419736e-02 -1.43256223e+00
-5.64977229e-01 -4.32608187e-01 -2.02665076e-01 5.82855463e-01
-1.75967574e+00 -1.37118542e+00 -6.45038843e-01 8.64963770e-01
4.59425777e-01 -1.16875738e-01 9.60780621e-01 5.29412806e-01
-4.64654684e-01 2.49915421e-01 -2.02160910e-01 8.17505792e-02
9.90848422e-01 -1.18838322e+00 3.52831632e-01 7.34947324e-01
3.02075744e-01 5.57390690e-01 5.60221851e-01 -3.08935374e-01
-1.78058302e+00 -9.55134988e-01 7.74954855e-01 -8.81120384e-01
4.59430218e-01 -2.49402031e-01 -9.54952002e-01 8.98390114e-01
-2.26966873e-01 7.97705829e-01 6.15527153e-01 -1.98630184e-01
-2.81793565e-01 6.73201680e-02 -1.18915284e+00 2.94348389e-01
1.44647086e+00 -6.30756319e-01 -7.53573954e-01 4.49534237e-01
7.17433989e-01 -1.04786003e+00 -1.29150224e+00 5.12796521e-01
6.22400165e-01 -1.10238636e+00 1.25490177e+00 -4.55409080e-01
6.62886024e-01 -3.32705349e-01 -5.25189519e-01 -1.03712583e+00
-1.43676639e-01 -7.03440070e-01 -2.17864916e-01 1.06133819e+00
-5.28027257e-03 -3.09032142e-01 8.54477406e-01 8.03810775e-01
-3.00908744e-01 -6.05694771e-01 -1.04303920e+00 -6.14635587e-01
-1.67140827e-01 -4.98330951e-01 4.71195161e-01 9.85807240e-01
-5.07935941e-01 -6.34415150e-02 -5.29726326e-01 -9.56904441e-02
6.54728711e-01 2.15871319e-01 1.14200389e+00 -1.05756271e+00
-1.52243912e-01 9.37521085e-03 -6.46961808e-01 -1.00576138e+00
4.16628003e-01 -7.83315003e-01 3.23085457e-01 -1.26033413e+00
1.70419604e-01 -5.97627819e-01 -4.66862433e-02 6.99303746e-01
8.84678662e-02 7.69908309e-01 -1.22999914e-01 7.40057826e-01
-6.02460861e-01 6.13678336e-01 1.35781813e+00 2.45735690e-01
-3.23394984e-02 -3.08214724e-01 -3.49573910e-01 9.65224743e-01
3.68828654e-01 -4.64210272e-01 -8.59611258e-02 -6.66392863e-01
1.00990713e-01 -1.41506210e-01 4.41268146e-01 -9.64254677e-01
-5.47044054e-02 -2.09825084e-01 4.17174667e-01 -4.58989888e-01
4.60254341e-01 -1.03906775e+00 4.67010379e-01 1.93623647e-01
-1.65810451e-01 1.25477254e-01 2.22329441e-02 4.42815989e-01
-2.69736618e-01 -3.51735801e-02 7.33275831e-01 -2.63026744e-01
-6.20441020e-01 3.85546356e-01 2.14589119e-01 1.10740975e-01
9.34568286e-01 -4.15865272e-01 6.84622750e-02 -2.08621114e-01
-7.52073944e-01 -1.76081821e-01 1.08564544e+00 4.50613439e-01
6.38948560e-01 -1.57018113e+00 -6.25032663e-01 3.77440333e-01
1.66504458e-01 6.10158682e-01 -1.37066860e-02 8.06497812e-01
-6.32419109e-01 1.88549012e-02 -2.22177729e-01 -9.42924440e-01
-1.12980056e+00 2.17758626e-01 3.93258840e-01 -1.10038549e-01
-1.09825480e+00 5.34231305e-01 -2.13151991e-01 -8.50647330e-01
6.89896718e-02 -2.13193282e-01 5.92027158e-02 -1.74506828e-01
2.19089940e-01 1.09785115e-02 2.16036648e-01 -1.22957361e+00
-4.40694243e-01 1.19423914e+00 3.33433181e-01 -6.88559785e-02
1.81731260e+00 -1.58091187e-01 -1.60465330e-01 3.26195359e-01
1.23726475e+00 -3.27069648e-02 -1.58814549e+00 -5.90664506e-01
-1.02786757e-01 -7.58649826e-01 -8.18917379e-02 -2.90988863e-01
-1.24925673e+00 5.45828044e-01 5.78940928e-01 -2.37717614e-01
1.13010442e+00 1.18204005e-01 8.85263503e-01 4.89634901e-01
4.65734601e-01 -9.58509326e-01 3.36898863e-01 4.59938735e-01
9.63658452e-01 -1.55306101e+00 2.93140978e-01 -3.20196331e-01
-2.52855510e-01 1.17731595e+00 4.23124343e-01 -2.68094331e-01
5.54987669e-01 2.19654053e-01 1.05925292e-01 -2.30357587e-01
-5.41551292e-01 -2.00354353e-01 4.15137500e-01 6.69690430e-01
4.64639127e-01 -5.05348206e-01 1.34819895e-01 6.78632036e-02
-3.04914296e-01 2.77862340e-01 2.46822119e-01 1.11725533e+00
-1.42864525e-01 -1.36597848e+00 -3.85924459e-01 1.11306384e-01
-4.21663225e-01 1.88081279e-01 -1.77819699e-01 7.94313550e-01
-1.49626002e-01 5.49619496e-01 4.03300934e-02 -4.43843722e-01
6.25506699e-01 -1.97807714e-01 8.02057981e-01 -5.83517194e-01
-2.97302634e-01 1.76996142e-01 1.02786228e-01 -1.04596245e+00
-1.28667831e+00 -8.75965536e-01 -1.11738181e+00 -2.40074605e-01
-2.87478328e-01 -3.20797026e-01 6.57110751e-01 1.08752036e+00
3.34438592e-01 2.67405659e-01 7.29347050e-01 -1.57618690e+00
-4.70675021e-01 -7.02976048e-01 -1.05800763e-01 9.63147461e-01
3.72348249e-01 -7.74050891e-01 -4.14999336e-01 5.32546163e-01]
|
[8.328749656677246, -2.467207193374634]
|
83c076aa-e81f-4dfe-a3e8-ee992ecdf8c3
|
a-distribution-dependent-mumford-shah-model
|
2203.15058
| null |
https://arxiv.org/abs/2203.15058v2
|
https://arxiv.org/pdf/2203.15058v2.pdf
|
A distribution-dependent Mumford-Shah model for unsupervised hyperspectral image segmentation
|
Hyperspectral images provide a rich representation of the underlying spectrum for each pixel, allowing for a pixel-wise classification/segmentation into different classes. As the acquisition of labeled training data is very time-consuming, unsupervised methods become crucial in hyperspectral image analysis. The spectral variability and noise in hyperspectral data make this task very challenging and define special requirements for such methods. Here, we present a novel unsupervised hyperspectral segmentation framework. It starts with a denoising and dimensionality reduction step by the well-established Minimum Noise Fraction (MNF) transform. Then, the Mumford-Shah (MS) segmentation functional is applied to segment the data. We equipped the MS functional with a novel robust distribution-dependent indicator function designed to handle the characteristic challenges of hyperspectral data. To optimize our objective function with respect to the parameters for which no closed form solution is available, we propose an efficient fixed point iteration scheme. Numerical experiments on four public benchmark datasets show that our method produces competitive results, which outperform three state-of-the-art methods substantially on three of these datasets.
|
['Benjamin Berkels', 'Chandrajit Bajaj', 'Jan-Christopher Cohrs']
|
2022-03-28
| null | null | null | null |
['hyperspectral-image-segmentation']
|
['computer-vision']
|
[ 1.06804383e+00 -4.87180889e-01 3.31276134e-02 -3.37562978e-01
-9.70830142e-01 -7.10485637e-01 2.75911510e-01 1.08133502e-01
-4.26228285e-01 6.32128775e-01 -4.58607763e-01 -2.47174621e-01
-5.38981080e-01 -8.35475922e-01 -3.56415361e-01 -1.27349484e+00
1.22139379e-01 4.28279072e-01 -1.49662390e-01 5.80657162e-02
2.45959178e-01 6.85918152e-01 -1.62642252e+00 -1.00549012e-01
1.52297473e+00 1.34885538e+00 5.17014205e-01 2.72047579e-01
-7.59056062e-02 4.86823767e-02 -2.77700305e-01 1.40683830e-01
6.10599875e-01 -4.71054822e-01 -8.76745045e-01 9.00242448e-01
2.59084821e-01 1.12182371e-01 1.04132950e-01 1.75186765e+00
3.83483440e-01 4.10797745e-01 9.35863733e-01 -1.01888466e+00
-2.31639236e-01 2.69161999e-01 -8.62087071e-01 -1.07688002e-01
-3.93028766e-01 -8.36864859e-02 9.51621771e-01 -7.56037533e-01
3.19818258e-01 6.15513921e-01 5.28944135e-01 3.15009393e-02
-1.42444527e+00 -3.84142607e-01 -5.38628958e-02 1.83491305e-01
-1.56457996e+00 -7.97277763e-02 8.58728468e-01 -5.87622881e-01
2.28291571e-01 5.06205857e-01 5.26344776e-01 4.10418421e-01
-4.13979769e-01 4.56187516e-01 1.45693576e+00 -2.38656655e-01
3.82051349e-01 -1.29870325e-01 2.21995786e-01 2.86958516e-01
3.48507375e-01 -2.01768890e-01 -3.09715085e-02 -1.30442679e-01
4.76163656e-01 2.08287805e-01 -6.04252577e-01 -4.04608309e-01
-8.80176783e-01 7.66695321e-01 4.61541533e-01 1.49685159e-01
-7.34695852e-01 -4.84489232e-01 9.20142978e-02 7.39635304e-02
7.30392039e-01 2.59800971e-01 -3.00403744e-01 4.85224038e-01
-1.17357397e+00 -2.08740965e-01 4.09890383e-01 4.17633832e-01
1.22487211e+00 -2.34757304e-01 4.73546796e-02 1.05053282e+00
2.21987382e-01 6.52707577e-01 2.53643692e-01 -8.10116410e-01
1.80147752e-01 6.35208070e-01 1.88727885e-01 -8.68104935e-01
-5.19419253e-01 -6.76526308e-01 -1.21800172e+00 1.12318687e-01
3.25871140e-01 -3.04701850e-02 -9.60518122e-01 1.18311167e+00
5.35835981e-01 2.91552305e-01 1.08304240e-01 1.01868522e+00
4.34767872e-01 8.86207402e-01 1.23768104e-02 -7.13149548e-01
1.06574142e+00 -6.45935655e-01 -4.90580261e-01 -1.34091109e-01
3.68724614e-01 -6.28461659e-01 9.96092498e-01 7.28368282e-01
-6.80230618e-01 -1.98931754e-01 -1.05195701e+00 4.23953116e-01
-3.15087646e-01 3.17165434e-01 6.15338981e-01 7.82988369e-01
-5.30988693e-01 6.13858521e-01 -7.29867578e-01 -1.51973203e-01
5.73852539e-01 2.65284956e-01 -3.07180315e-01 -5.12903631e-02
-7.67320156e-01 3.55721205e-01 6.38607025e-01 4.60016161e-01
-6.56872988e-01 -6.19496942e-01 -7.04840064e-01 -5.22832908e-02
6.65125251e-01 -2.03163505e-01 7.56992102e-01 -1.02955139e+00
-1.54685092e+00 7.81343877e-01 -8.62717032e-02 -1.77012593e-01
3.71587127e-01 1.22269988e-01 -2.33636871e-01 5.42923152e-01
1.69648100e-02 2.03561828e-01 1.05905759e+00 -1.49493885e+00
-5.79357505e-01 -6.58611536e-01 -2.58622736e-01 3.33431572e-01
-6.24566674e-01 -2.54322022e-01 -1.69694364e-01 -4.92212206e-01
7.13230789e-01 -8.71877909e-01 -4.30062950e-01 -1.80205241e-01
-6.36044681e-01 1.80930242e-01 8.66245866e-01 -7.00347841e-01
9.99507308e-01 -2.21747613e+00 2.03401119e-01 5.90168238e-01
8.03230554e-02 3.29617947e-01 -4.17799577e-02 2.34503523e-01
-3.04482758e-01 -7.02678338e-02 -1.19409108e+00 -1.18321203e-01
-1.15068659e-01 1.35179982e-02 -1.17445312e-01 9.88424838e-01
1.18799388e-01 3.91709059e-01 -8.09449375e-01 -2.51788378e-01
2.66992271e-01 5.23205042e-01 -2.02131897e-01 1.19850159e-01
-2.75686115e-01 6.19672179e-01 -4.18000907e-01 6.87178075e-01
1.20485187e+00 -4.96070921e-01 1.45797983e-01 -4.27947789e-01
-3.00820827e-01 -2.71600425e-01 -1.43824232e+00 1.57478249e+00
-5.09032942e-02 2.04310283e-01 5.16681731e-01 -1.61831748e+00
8.55486095e-01 -1.65617303e-03 7.85511315e-01 -2.19873101e-01
1.60812005e-01 4.00425076e-01 -3.60457063e-01 -4.26869959e-01
1.83618143e-01 -5.04266083e-01 3.48396361e-01 2.53485680e-01
-2.97566295e-01 -5.06920457e-01 2.81998098e-01 -2.30685502e-01
6.50835872e-01 -2.37707943e-01 1.66149631e-01 -5.87514520e-01
9.08967495e-01 2.09939703e-01 4.74981964e-01 4.81407702e-01
2.52528524e-04 6.90076530e-01 1.92242682e-01 2.34080087e-02
-6.10881448e-01 -6.85865223e-01 -4.85118419e-01 6.12264216e-01
2.74877161e-01 2.17089236e-01 -8.72513413e-01 -5.70797086e-01
-1.07793145e-01 3.89958233e-01 -3.19575191e-01 1.07409298e-01
1.47290334e-01 -1.44700265e+00 9.86584648e-02 3.45755704e-02
7.09533095e-01 -6.60504580e-01 -4.02288735e-01 6.59017935e-02
-2.69050777e-01 -1.13693118e+00 -2.26153880e-01 3.63491744e-01
-1.03450572e+00 -1.25442457e+00 -8.76111507e-01 -7.11637616e-01
8.65747511e-01 6.10453784e-01 5.68819523e-01 -1.76684797e-01
-4.64276135e-01 6.93147182e-02 -4.00080979e-01 -2.22629815e-01
-1.47051677e-01 6.52846172e-02 -3.05951774e-01 5.52545190e-01
8.06851387e-02 -5.50832272e-01 -6.80275202e-01 3.70383322e-01
-1.50844049e+00 -7.79172704e-02 6.07155740e-01 8.72258961e-01
1.07490492e+00 8.23672354e-01 3.51235807e-01 -1.03695357e+00
2.94044256e-01 -3.38877261e-01 -1.03197920e+00 2.66146541e-01
-6.35214150e-01 -2.87551641e-01 6.22632265e-01 -1.36824340e-01
-1.01572311e+00 6.20123267e-01 1.31576985e-01 -1.92559302e-01
-1.84825644e-01 9.10556793e-01 -3.53891015e-01 -4.06824648e-01
6.74890876e-01 4.94025797e-01 1.34930655e-01 -4.86408323e-01
2.27189243e-01 9.94062066e-01 6.08684123e-01 -3.83710891e-01
1.01334071e+00 8.50455940e-01 2.85490543e-01 -1.25326717e+00
-1.10137808e+00 -9.89511669e-01 -6.44940197e-01 -9.85007882e-02
7.30278134e-01 -6.86100960e-01 -4.89665449e-01 6.69829607e-01
-7.55728900e-01 -3.78730714e-01 -2.02308163e-01 4.64610189e-01
-4.69567895e-01 8.58113587e-01 -2.68099755e-01 -9.47697818e-01
-3.47327143e-01 -1.28061831e+00 8.61153841e-01 1.76159307e-01
5.70440829e-01 -8.54832470e-01 -8.70751664e-02 5.64917088e-01
2.10916400e-01 3.88765275e-01 8.52234721e-01 -3.44521105e-01
-3.39255035e-01 -7.77222142e-02 -6.00012660e-01 8.47786546e-01
4.33023572e-01 -3.90128642e-02 -9.69685733e-01 -3.21008712e-01
4.07787651e-01 -2.04415232e-01 1.13527930e+00 6.35094941e-01
1.67163599e+00 1.17872264e-02 -1.30580306e-01 1.06786001e+00
1.79362309e+00 6.73495233e-02 5.30029356e-01 2.96737969e-01
7.19477057e-01 7.72732139e-01 7.60244310e-01 6.53647184e-01
-9.86168087e-02 1.50162265e-01 7.17672348e-01 -1.93493053e-01
4.09504443e-01 4.24227327e-01 -1.41348943e-01 4.15396571e-01
-1.95883408e-01 -1.94556296e-01 -8.14501822e-01 3.25395882e-01
-1.68003213e+00 -8.27624500e-01 -4.93157476e-01 2.45232916e+00
8.69892716e-01 -2.65278995e-01 -6.02204353e-02 6.26437902e-01
9.66866434e-01 2.26407662e-01 -8.05304229e-01 5.67197144e-01
-3.92007738e-01 3.47965032e-01 8.68916333e-01 4.81511325e-01
-1.49640608e+00 6.45497143e-01 5.60631227e+00 9.54505503e-01
-1.33655703e+00 -2.38452647e-02 7.41516173e-01 4.58067805e-01
-1.94120049e-01 -1.51735082e-01 -3.60677809e-01 2.73401797e-01
4.85703230e-01 -7.03843608e-02 5.96885502e-01 4.37419206e-01
3.80573422e-01 -2.69533515e-01 -3.82211030e-01 1.27902842e+00
-1.12412475e-01 -1.00887764e+00 -1.00576319e-01 1.77319378e-01
1.02000690e+00 -5.69267944e-02 1.40758857e-01 -3.73935461e-01
-3.22502375e-01 -9.53380942e-01 3.38867784e-01 4.04343069e-01
8.79784465e-01 -8.32360625e-01 4.17550743e-01 4.17919219e-01
-1.07945657e+00 -1.52305499e-01 -5.99668801e-01 2.55960613e-01
-3.56868841e-02 1.27452075e+00 -2.39649341e-01 7.34956622e-01
4.24768537e-01 8.49264979e-01 -1.76138371e-01 1.34237063e+00
-9.84354541e-02 6.07438207e-01 -4.28866804e-01 4.31526780e-01
2.56459475e-01 -1.07246733e+00 5.82541406e-01 9.62471902e-01
5.40935695e-01 4.31409270e-01 2.91103721e-01 8.71413589e-01
3.69503326e-03 3.85773420e-01 -1.69032678e-01 -5.26152194e-01
1.10904939e-01 1.69162881e+00 -1.16622007e+00 -3.08536794e-02
-1.43440232e-01 1.05578232e+00 -1.67101592e-01 5.10528386e-01
-5.43675900e-01 -3.27576071e-01 5.57958364e-01 -5.12981415e-03
2.73619294e-01 -3.04212123e-01 -3.32694560e-01 -1.11666751e+00
-2.44623236e-02 -8.94928396e-01 5.40536046e-01 -5.03637910e-01
-1.33554542e+00 5.38394690e-01 -2.02093050e-01 -1.33386886e+00
3.39317352e-01 -9.10042763e-01 -3.15446228e-01 8.94310236e-01
-2.07257819e+00 -8.16395879e-01 -8.30899954e-01 6.58326209e-01
3.38899076e-01 1.76420808e-01 7.74179459e-01 3.29542726e-01
-8.30254912e-01 -6.39921427e-02 5.26934862e-01 -1.36188969e-01
4.45888191e-01 -1.18658829e+00 -3.63129020e-01 9.49245632e-01
-2.26156279e-01 1.29911110e-01 5.82228720e-01 -4.86007154e-01
-1.38760459e+00 -1.22097385e+00 1.12350620e-01 3.52067381e-01
7.62410581e-01 1.30397841e-01 -9.94419277e-01 1.63755104e-01
-2.59963959e-01 1.94014475e-01 1.00302625e+00 -4.54220146e-01
-2.86440197e-02 -2.66048759e-01 -1.16608644e+00 1.39884725e-01
8.33078980e-01 -3.82554084e-01 -2.20209491e-02 7.70891070e-01
2.27107331e-01 -1.24692716e-01 -8.82832229e-01 7.19457030e-01
1.38537046e-02 -9.43900764e-01 9.72955227e-01 -2.14769140e-01
2.23023400e-01 -4.83298928e-01 -2.59816080e-01 -1.41474164e+00
-1.57817110e-01 -7.13054001e-01 2.78062612e-01 8.95521164e-01
4.72837448e-01 -8.99205685e-01 8.44477057e-01 4.43054497e-01
-2.23219052e-01 -4.10662472e-01 -5.13586700e-01 -8.33230853e-01
-1.63409173e-01 -2.77235508e-01 4.79156196e-01 1.11588323e+00
-2.50615001e-01 -1.28253363e-02 6.63409755e-02 7.54740536e-01
1.14399111e+00 4.89437997e-01 3.45194399e-01 -1.71326292e+00
-5.56897372e-02 -6.06417894e-01 -2.20563129e-01 -9.77114499e-01
2.99376100e-01 -9.17346954e-01 2.12971538e-01 -1.64421725e+00
2.55344659e-01 -4.23330605e-01 -2.44602054e-01 2.75732756e-01
-2.98120767e-01 3.77083272e-01 -2.06857234e-01 3.32978636e-01
-2.02418804e-01 6.59723759e-01 1.28647399e+00 -4.48483527e-01
-3.71554911e-01 1.36348501e-01 -6.55849457e-01 7.40265667e-01
8.65344763e-01 -3.56510311e-01 -5.60846984e-01 -1.82884648e-01
1.18922427e-01 -1.64381295e-01 2.76342660e-01 -1.04854190e+00
1.45845234e-01 -5.45185983e-01 2.15026185e-01 -6.13951445e-01
3.23357016e-01 -1.00569963e+00 1.79743916e-01 3.17326099e-01
-1.04812514e-02 -7.38417685e-01 -1.52265215e-02 5.63853085e-01
-3.57484847e-01 -5.76472461e-01 1.16994059e+00 2.87749842e-02
-6.52217150e-01 6.16537273e-01 -1.60219312e-01 -1.65748239e-01
1.05238616e+00 -1.89304650e-01 -1.66607842e-01 -1.86238110e-01
-7.23408699e-01 4.24299724e-02 5.05104840e-01 -4.61127669e-01
5.03563523e-01 -9.04802561e-01 -7.76907146e-01 1.96964651e-01
1.51024058e-01 3.00239742e-01 3.83947194e-01 9.93214428e-01
-7.32576013e-01 2.51789056e-02 1.11949161e-01 -7.27511883e-01
-1.03417766e+00 3.28456372e-01 5.25734484e-01 -3.34171914e-02
-3.93852204e-01 7.89765060e-01 2.44327709e-01 -3.80189836e-01
-8.39512125e-02 -2.08596691e-01 -3.61270428e-01 3.90432596e-01
4.96201664e-01 5.17031848e-01 2.96185255e-01 -7.72753477e-01
-1.02343626e-01 6.01778388e-01 4.34411407e-01 -3.28850746e-02
1.54679155e+00 -1.57653093e-01 -5.02573848e-01 1.72571152e-01
1.19448864e+00 -2.37473071e-01 -1.36934674e+00 -2.49981865e-01
-1.60064012e-01 -4.99311149e-01 5.07203162e-01 -4.87684667e-01
-1.27887762e+00 8.10605049e-01 5.31758308e-01 6.56643033e-01
1.59778929e+00 -4.26573902e-01 6.28235579e-01 3.07331026e-01
2.00151607e-01 -1.26499033e+00 -3.53344351e-01 4.05612171e-01
5.75347304e-01 -1.36830091e+00 1.01737015e-01 -1.09687829e+00
-3.08947861e-01 1.03061306e+00 1.28509626e-01 1.41113937e-01
7.81675518e-01 -1.25973180e-01 1.26712784e-01 -2.20831811e-01
2.67987698e-01 -5.90169430e-01 4.03457999e-01 3.70583117e-01
9.86978486e-02 1.85507014e-01 -4.94699508e-01 4.03476596e-01
7.83591270e-02 -2.72868603e-01 2.71037340e-01 7.59623706e-01
-7.45969474e-01 -8.68779957e-01 -5.55098474e-01 6.03843510e-01
-2.79976696e-01 -1.11975268e-01 -1.85308486e-01 3.31236452e-01
-1.15595505e-01 1.17267120e+00 -2.62031436e-01 3.97045948e-02
1.08346738e-01 7.15390220e-02 2.85522729e-01 -7.46406019e-01
-5.89402132e-02 3.99010450e-01 -4.10823792e-01 -2.69373715e-01
-8.02725375e-01 -7.33401835e-01 -1.37053943e+00 1.70482114e-01
-6.06186152e-01 2.61203587e-01 9.27094638e-01 1.01668298e+00
3.43192834e-03 1.37633935e-01 1.02951813e+00 -1.10007954e+00
-8.60035598e-01 -7.20590830e-01 -1.25011492e+00 5.47283649e-01
2.16896117e-01 -5.52691877e-01 -7.38416851e-01 1.71456307e-01]
|
[10.024361610412598, -1.9564764499664307]
|
cd070148-2ae8-4e4c-9744-3e8953e8c039
|
vln-trans-translator-for-the-vision-and
|
2302.09230
| null |
https://arxiv.org/abs/2302.09230v1
|
https://arxiv.org/pdf/2302.09230v1.pdf
|
VLN-Trans: Translator for the Vision and Language Navigation Agent
|
Language understanding is essential for the navigation agent to follow instructions. We observe two kinds of issues in the instructions that can make the navigation task challenging: 1. The mentioned landmarks are not recognizable by the navigation agent due to the different vision abilities of the instructor and the modeled agent. 2. The mentioned landmarks are applicable to multiple targets, thus not distinctive for selecting the target among the candidate viewpoints. To deal with these issues, we design a translator module for the navigation agent to convert the original instructions into easy-to-follow sub-instruction representations at each step. The translator needs to focus on the recognizable and distinctive landmarks based on the agent's visual abilities and the observed visual environment. To achieve this goal, we create a new synthetic sub-instruction dataset and design specific tasks to train the translator and the navigation agent. We evaluate our approach on Room2Room~(R2R), Room4room~(R4R), and Room2Room Last (R2R-Last) datasets and achieve state-of-the-art results on multiple benchmarks.
|
['Parisa Kordjamshidi', 'Yue Zhang']
|
2023-02-18
| null | null | null | null |
['vision-and-language-navigation']
|
['robots']
|
[-1.95869192e-01 4.42806892e-02 2.39171997e-01 -5.49377084e-01
-4.34838593e-01 -7.17231214e-01 4.28621829e-01 3.66219617e-02
-5.35232484e-01 3.09887111e-01 -6.74319789e-02 -5.35461307e-01
1.38370186e-01 -5.62162161e-01 -8.00134480e-01 -3.57778013e-01
1.76920876e-01 6.66847169e-01 6.14191592e-01 -4.25170720e-01
4.43904519e-01 5.48683286e-01 -1.75719690e+00 2.56602526e-01
1.18508720e+00 5.33901036e-01 6.67479813e-01 5.96747041e-01
-1.32966712e-01 7.38792002e-01 -5.55993855e-01 7.24018440e-02
3.96197408e-01 -5.98747313e-01 -6.67596579e-01 -1.45174162e-02
7.19304979e-01 -3.30014855e-01 -1.58021525e-01 1.13799226e+00
2.79173851e-01 7.11540341e-01 7.77622640e-01 -1.12716436e+00
-5.38140237e-01 4.58424985e-01 -5.34913361e-01 -3.48717859e-03
7.52417624e-01 1.86769381e-01 6.50588810e-01 -8.41197610e-01
7.08797872e-01 1.26180196e+00 1.40473336e-01 6.88378036e-01
-7.98696876e-01 -5.39230645e-01 7.12928236e-01 3.30201387e-01
-1.17351270e+00 -3.57617915e-01 5.62566757e-01 -3.09438676e-01
5.38529992e-01 2.42091596e-01 2.85548002e-01 1.06771362e+00
1.96524486e-01 7.21532822e-01 1.10966253e+00 -4.27567244e-01
2.35463157e-01 4.70905662e-01 2.06089586e-01 1.11807549e+00
-5.89985177e-02 1.05839185e-01 -4.48276609e-01 4.81810570e-01
7.02929735e-01 8.65669325e-02 -6.87354684e-01 -7.38891065e-01
-1.27104199e+00 6.26654088e-01 7.80718327e-01 2.33123913e-01
-2.28110969e-01 -1.90260202e-01 9.05215368e-02 3.96829635e-01
-3.91171247e-01 6.97263002e-01 -2.66072065e-01 -7.21591040e-02
-4.01642919e-01 -1.44094244e-01 5.73456109e-01 1.18982553e+00
7.07769811e-01 -1.10541666e-02 1.41820349e-02 4.97111827e-01
4.73583549e-01 4.69815642e-01 4.92316097e-01 -7.78537154e-01
6.93689466e-01 6.98343754e-01 1.78456753e-01 -9.58474874e-01
-4.99010623e-01 -3.32601905e-01 -2.82120526e-01 7.97777414e-01
7.15158343e-01 3.79785523e-02 -1.06298673e+00 1.54518712e+00
5.01512945e-01 -3.30518857e-02 1.80133730e-01 1.00301456e+00
1.16343987e+00 6.31908178e-01 -3.46540310e-03 2.23059997e-01
1.45059419e+00 -1.71486437e+00 -5.14850080e-01 -7.13378787e-01
8.10994565e-01 -8.45844328e-01 1.68116784e+00 3.31128210e-01
-8.92230332e-01 -9.53732133e-01 -1.10149372e+00 -6.77439570e-02
-6.72265291e-01 5.27943075e-01 9.28741992e-02 3.01793665e-01
-8.83700788e-01 2.47438163e-01 -6.41532779e-01 -5.12243986e-01
-2.20357746e-01 2.27870613e-01 -6.05328381e-01 -3.81358206e-01
-7.45624483e-01 1.12316799e+00 5.99600784e-02 1.33370399e-01
-1.22351539e+00 -2.65901029e-01 -1.07237506e+00 -5.01467846e-03
4.98453081e-01 -4.45573449e-01 1.11780763e+00 -9.74348009e-01
-1.59125900e+00 7.75106788e-01 -1.52893960e-01 2.16840774e-01
7.02888846e-01 -1.41656548e-01 -2.13350847e-01 7.39300475e-02
2.70631969e-01 5.30732095e-01 4.82181221e-01 -1.73615849e+00
-1.03306568e+00 -4.17910308e-01 4.12720948e-01 6.94164038e-01
1.01341210e-01 -2.61277467e-01 -7.26067126e-01 -3.64010185e-01
2.20773950e-01 -1.14715409e+00 -2.05864057e-01 -2.24630266e-01
-6.42009079e-01 -8.62702876e-02 8.04819942e-01 -3.27130079e-01
7.51909196e-01 -2.19055843e+00 2.23694429e-01 1.06569022e-01
9.71021876e-02 1.45633996e-01 -5.57951212e-01 1.92296192e-01
-4.16740812e-02 -3.71698558e-01 4.66595322e-01 -5.94777763e-01
-1.97004884e-01 7.04952702e-02 -5.74272983e-02 2.35028088e-01
-5.30183733e-01 4.39881325e-01 -9.51855361e-01 -2.61905551e-01
3.20171207e-01 3.05980623e-01 -2.59316713e-01 6.69971824e-01
-4.77190651e-02 7.08396971e-01 -6.56747222e-01 4.49711472e-01
5.09438753e-01 7.26305228e-03 -5.66527955e-02 -1.49671495e-01
-1.94308624e-01 4.16497946e-01 -1.30519819e+00 1.66304088e+00
-7.24211693e-01 5.50046384e-01 1.18226796e-01 -4.09740627e-01
9.22988951e-01 -1.33446947e-01 -1.65613323e-01 -7.89185226e-01
7.51404762e-02 1.43008351e-01 2.07883552e-01 -5.81793964e-01
4.72051442e-01 2.78907746e-01 1.46918669e-01 4.80497092e-01
-2.15415195e-01 -1.54393562e-03 2.02450633e-01 8.51051230e-03
1.01418126e+00 2.90075332e-01 5.71805760e-02 -1.70362294e-01
8.67386639e-01 1.87800795e-01 2.74550617e-01 7.79389739e-01
-2.90703773e-01 5.69978595e-01 1.82534114e-01 -6.53116703e-01
-3.69004428e-01 -9.86083329e-01 6.35412812e-01 1.62269568e+00
6.36551559e-01 -1.80642202e-01 -5.85675955e-01 -1.17054784e+00
-4.63078052e-01 1.14601290e+00 -8.51702332e-01 -1.91802785e-01
-7.43978620e-01 9.26479325e-02 2.74550468e-02 6.33731425e-01
4.51253921e-01 -1.20089018e+00 -1.24368095e+00 -6.69897646e-02
-2.44684443e-01 -1.16456342e+00 -7.98788548e-01 3.68193954e-01
-4.20836300e-01 -1.29921114e+00 -3.22613358e-01 -1.13117969e+00
1.42500746e+00 7.28412330e-01 1.21945620e+00 2.89455861e-01
1.01263210e-01 5.91872275e-01 -4.81185585e-01 -3.00253630e-01
-4.88891870e-01 9.32262093e-02 -7.28123188e-02 -1.87827855e-01
1.72775507e-01 -5.41173816e-02 -6.61639273e-01 7.76147962e-01
-5.09005547e-01 2.59809643e-01 6.94668353e-01 6.41465604e-01
5.58337152e-01 8.37950334e-02 4.05427143e-02 -7.41832912e-01
6.89882696e-01 8.12695399e-02 -7.87381172e-01 5.53026736e-01
-4.33018446e-01 2.81120598e-01 1.08340275e+00 -7.12676823e-01
-1.03641248e+00 3.84112924e-01 9.98633504e-02 -3.19743395e-01
-4.88643289e-01 2.30607852e-01 -3.97953808e-01 -1.37996763e-01
6.38293505e-01 3.01360130e-01 -1.16605990e-01 -2.29941279e-01
3.71322870e-01 3.02397728e-01 5.56179345e-01 -5.53922892e-01
1.01703322e+00 2.05998689e-01 -1.64619341e-01 -4.40761745e-01
-6.75885677e-01 -2.67804295e-01 -7.13257372e-01 -2.42246658e-01
1.02098072e+00 -6.57429278e-01 -7.94153750e-01 4.84067649e-02
-1.02827013e+00 -5.92441618e-01 -1.32893294e-01 5.42750537e-01
-3.79672199e-01 3.25015411e-02 -9.95544419e-02 -4.08949614e-01
-9.02981758e-02 -2.05359268e+00 8.53030622e-01 6.75679982e-01
-1.29550025e-01 -8.12246978e-01 -3.08380947e-02 2.54843622e-01
3.95644724e-01 -2.54066754e-03 9.67965722e-01 -8.23846459e-01
-6.66749239e-01 -2.03815371e-01 -6.78570345e-02 -8.69793668e-02
1.83097944e-01 1.11008696e-01 -6.50066018e-01 -2.31232375e-01
-2.99323257e-02 -2.66811550e-01 4.15589273e-01 9.80846882e-02
6.72976792e-01 8.89052823e-02 -4.35539335e-01 5.82931519e-01
1.09454191e+00 7.83102691e-01 5.15797019e-01 7.71453202e-01
7.20459878e-01 6.98240876e-01 9.50655937e-01 -1.43656373e-01
7.49929368e-01 8.41664612e-01 6.68919325e-01 -9.06354338e-02
-2.18555555e-01 -3.52596611e-01 6.69435501e-01 5.63089550e-01
1.51021421e-01 -1.94919154e-01 -8.86137545e-01 3.91597241e-01
-1.71141648e+00 -5.77814758e-01 -7.16993213e-02 2.37298203e+00
4.48348314e-01 1.92293465e-01 -1.27113789e-01 -3.30342799e-01
4.44033951e-01 1.14344284e-01 -4.15616959e-01 -5.25778413e-01
4.78851944e-01 -2.39667341e-01 1.44350871e-01 6.25563025e-01
-9.20479119e-01 9.95227337e-01 5.51958179e+00 4.50399965e-01
-1.20999312e+00 -3.05152625e-01 4.28779632e-01 2.11316064e-01
-1.10994361e-01 -1.38949752e-01 -1.00368953e+00 3.64613354e-01
5.04878879e-01 1.81521311e-01 5.14797628e-01 1.01803768e+00
5.30891009e-02 -3.28226507e-01 -1.56094265e+00 7.92913318e-01
4.11183864e-01 -7.29164779e-01 5.62485680e-03 -4.04670060e-01
4.41975147e-01 -1.01263888e-01 3.80582124e-01 6.64631844e-01
3.47219825e-01 -1.11912405e+00 5.60715795e-01 1.58027574e-01
3.58609527e-01 -7.55425811e-01 6.42245293e-01 5.81488788e-01
-1.33439481e+00 -7.27677420e-02 -3.49716574e-01 2.27805704e-01
-8.03770591e-03 -3.08862031e-01 -1.08777583e+00 4.34200346e-01
7.68073916e-01 1.99320927e-01 -9.05504584e-01 9.61128175e-01
-7.82067001e-01 -2.30793118e-01 -1.92392081e-01 -2.79894412e-01
5.13087988e-01 -4.73070323e-01 3.38037640e-01 7.31686831e-01
4.47892636e-01 -8.82417634e-02 4.84813869e-01 6.32759929e-01
7.24050999e-02 1.28601059e-01 -6.66515887e-01 4.47192609e-01
4.54918504e-01 1.31483746e+00 -9.00157690e-01 -2.11970210e-01
-4.61469293e-01 1.07824481e+00 3.53749126e-01 4.28885788e-01
-9.16956782e-01 -6.42656565e-01 6.30577445e-01 1.57504991e-01
1.88091472e-01 -2.85169482e-01 1.78151190e-01 -1.00377011e+00
-9.10430774e-02 -1.27731729e+00 3.08298498e-01 -1.06878769e+00
-7.36931503e-01 1.25792336e+00 -1.64045691e-01 -1.35887444e+00
-1.70511499e-01 -8.82144451e-01 -9.45964992e-01 7.91138530e-01
-1.53460383e+00 -9.81486380e-01 -1.02513206e+00 6.99550092e-01
9.67155039e-01 -2.52112061e-01 8.90277267e-01 1.42344385e-01
-7.58363128e-01 6.56444728e-01 -2.23055273e-01 5.84200695e-02
9.74265754e-01 -1.43199837e+00 3.49502832e-01 8.85335803e-01
2.97420919e-01 7.31401861e-01 9.23722684e-01 -3.18772525e-01
-1.44207549e+00 -7.74534643e-01 4.68666732e-01 -7.20138609e-01
3.07306170e-01 -1.88190848e-01 -7.61528850e-01 9.14740741e-01
3.35541904e-01 -3.38973030e-02 5.49743474e-01 -1.42369226e-01
-3.98620278e-01 2.65316721e-02 -1.05467534e+00 1.13847184e+00
8.04605782e-01 -2.41260335e-01 -7.60785043e-01 3.50565404e-01
4.38541025e-01 -9.01708245e-01 -2.32008755e-01 2.18809649e-01
4.73298222e-01 -1.14020860e+00 9.68728900e-01 -3.77570808e-01
3.33806366e-01 -6.98986351e-01 1.51414737e-01 -1.67016435e+00
-1.65306866e-01 -2.54659742e-01 3.36166531e-01 9.66972589e-01
5.69870055e-01 -6.00275695e-01 6.50602698e-01 5.69113851e-01
-2.12232143e-01 -7.22264886e-01 -6.94932282e-01 -6.30593181e-01
-2.82366037e-01 1.53376997e-01 5.29727638e-01 6.74458444e-01
-1.49666741e-02 4.95578647e-01 -8.96949247e-02 5.04261017e-01
2.33478412e-01 4.67733353e-01 1.24085796e+00 -9.73746121e-01
-2.63029337e-01 -4.14831072e-01 -2.27570115e-03 -1.74188840e+00
1.54523417e-01 -4.97351587e-01 4.44131315e-01 -1.68325543e+00
-1.00645162e-01 -4.98184055e-01 -1.33681819e-01 4.60216701e-01
-3.36735785e-01 -2.58194834e-01 4.04000342e-01 -1.68720968e-02
-8.53493631e-01 3.87373358e-01 1.57183897e+00 -1.56962514e-01
-5.29874325e-01 2.34873012e-01 -6.09696031e-01 8.81351709e-01
5.95365107e-01 -2.60847598e-01 -7.94474661e-01 -8.49784076e-01
-1.37916896e-02 1.90854110e-02 -6.36140853e-02 -1.17257512e+00
5.11415124e-01 -3.12764227e-01 4.20836031e-01 -7.38931715e-01
4.29677695e-01 -1.20580614e+00 -1.74502596e-01 5.48832417e-01
-3.63285363e-01 7.25714147e-01 2.13893890e-01 2.92023361e-01
-1.76151156e-01 -5.19965887e-01 6.88875318e-01 -2.72856981e-01
-9.37369049e-01 -1.02464579e-01 -3.89732122e-01 -1.87081650e-01
1.28750062e+00 -3.73936892e-01 -5.58275044e-01 -5.79719424e-01
-5.73712766e-01 6.53953135e-01 6.55957878e-01 5.72584808e-01
9.64659154e-01 -9.96151507e-01 -3.79052877e-01 3.84765655e-01
3.27050716e-01 2.04277933e-01 7.94433802e-02 6.83294773e-01
-8.19601476e-01 2.53200471e-01 -4.67230409e-01 -5.13010442e-01
-1.44543564e+00 7.25455523e-01 5.91582000e-01 -3.50169986e-01
-6.11298680e-01 1.02702534e+00 6.74244165e-01 -8.70407283e-01
6.82692587e-01 -5.67839563e-01 -5.51190853e-01 -9.19806138e-02
5.91541588e-01 1.75858781e-01 -1.83267504e-01 -6.79262280e-01
-3.12425017e-01 8.63341451e-01 -3.37659150e-01 1.44654870e-01
9.55415547e-01 -3.68402451e-01 3.32875669e-01 3.22366118e-01
8.75499070e-01 3.27976525e-01 -1.08912718e+00 2.68034879e-02
-2.87669837e-01 -4.68720704e-01 -4.08026189e-01 -8.94552827e-01
-1.06575084e+00 9.04988408e-01 7.41885066e-01 4.52583889e-03
9.04616475e-01 -3.41191232e-01 3.70018721e-01 7.11557925e-01
5.91541827e-01 -8.91095579e-01 4.73846465e-01 6.70587897e-01
9.72498715e-01 -1.36009622e+00 -1.45932883e-01 -5.00246763e-01
-9.51205075e-01 1.22763646e+00 1.37255239e+00 1.25725225e-01
2.22663954e-01 -1.50272399e-01 6.96733415e-01 -3.47163826e-01
-4.79725093e-01 -1.00101821e-01 4.03813243e-01 6.49671137e-01
3.60785931e-01 -1.32188067e-01 2.09104016e-01 2.64578402e-01
-3.09208184e-01 -6.14525497e-01 7.14088500e-01 8.79401922e-01
-4.75259215e-01 -9.88678098e-01 -4.77211863e-01 -1.15113765e-01
1.56376436e-01 2.69971520e-01 -4.48259294e-01 9.82182860e-01
1.78026985e-02 1.08569372e+00 -2.05254361e-01 -3.40382814e-01
8.58364224e-01 -1.25796586e-01 1.46759242e-01 -7.97111690e-01
-5.93188226e-01 -7.82338977e-02 -1.46866456e-01 -7.36064255e-01
2.65223645e-02 -2.27458596e-01 -1.63703406e+00 5.98127246e-02
-3.26492101e-01 3.19877595e-01 6.59873545e-01 7.14378417e-01
2.55574405e-01 9.75574017e-01 5.74853957e-01 -8.94260168e-01
-5.82846880e-01 -7.57000685e-01 -1.74216196e-01 4.79817629e-01
2.06869975e-01 -7.88452029e-01 -4.46111262e-01 -2.33419508e-01]
|
[4.462834358215332, 0.46994268894195557]
|
7a2c8116-4cac-42e0-8128-5075e7f74543
|
efficient-mask-correction-for-click-based
| null | null |
http://openaccess.thecvf.com//content/CVPR2023/html/Du_Efficient_Mask_Correction_for_Click-Based_Interactive_Image_Segmentation_CVPR_2023_paper.html
|
http://openaccess.thecvf.com//content/CVPR2023/papers/Du_Efficient_Mask_Correction_for_Click-Based_Interactive_Image_Segmentation_CVPR_2023_paper.pdf
|
Efficient Mask Correction for Click-Based Interactive Image Segmentation
|
The goal of click-based interactive image segmentation is to extract target masks with the input of positive/negative clicks. Every time a new click is placed, existing methods run the whole segmentation network to obtain a corrected mask, which is inefficient since several clicks may be needed to reach satisfactory accuracy. To this end, we propose an efficient method to correct the mask with a lightweight mask correction network. The whole network remains a low computational cost from the second click, even if we have a large backbone. However, a simple correction network with limited capacity is not likely to achieve comparable performance with a classic segmentation network. Thus, we propose a click-guided self-attention module and a click-guided correlation module to effectively exploits the click information to boost performance. First, several templates are selected based on the semantic similarity with click features. Then the self-attention module propagates the template information to other pixels, while the correlation module directly uses the templates to obtain target outlines. With the efficient architecture and two click-guided modules, our method shows preferable performance and efficiency compared to existing methods. The code will be released at https://github.com/feiaxyt/EMC-Click.
|
['Fan Wang', 'Zhibin Wang', 'Jianlong Yuan', 'Fei Du']
|
2023-01-01
| null | null | null |
cvpr-2023-1
|
['semantic-textual-similarity', 'semantic-similarity']
|
['natural-language-processing', 'natural-language-processing']
|
[ 2.43679762e-01 -4.75623235e-02 -2.83232089e-02 -4.61590588e-01
-8.13366115e-01 -4.70962614e-01 1.09064430e-01 -1.15066938e-01
-5.72560430e-01 3.53419691e-01 -3.17383319e-01 -2.16517717e-01
3.47233683e-01 -6.95870459e-01 -5.95628798e-01 -4.30845469e-01
4.62119550e-01 2.65424073e-01 1.02973402e+00 1.45612331e-02
5.13538718e-01 2.38886744e-01 -1.46395373e+00 2.82952219e-01
1.29657745e+00 1.02988732e+00 7.86689818e-01 4.97617036e-01
-4.62802947e-01 3.02475840e-01 -3.56621981e-01 -4.17604864e-01
2.55076706e-01 -3.45836729e-01 -6.78390980e-01 -4.55080234e-02
3.95788342e-01 -4.03789401e-01 -4.04031947e-02 1.23889244e+00
5.17251849e-01 3.45330313e-02 3.24629486e-01 -1.01383877e+00
-1.61164805e-01 5.86765885e-01 -9.29538369e-01 1.62164658e-01
1.75454468e-01 2.24019557e-01 1.10865283e+00 -1.18314803e+00
5.54884315e-01 8.15920353e-01 5.13605833e-01 4.63721871e-01
-1.19774079e+00 -8.28744709e-01 3.45294803e-01 2.02265218e-01
-1.37676382e+00 -1.63493827e-01 7.73616493e-01 -2.90437728e-01
4.56098467e-01 4.58609730e-01 9.03363228e-01 5.89343071e-01
-1.47422865e-01 1.05809951e+00 9.05235350e-01 -1.88521624e-01
3.78767587e-03 3.13348144e-01 9.13522393e-02 7.47454762e-01
-9.39168129e-03 -1.90155447e-01 -5.74475527e-01 2.21282065e-01
7.44942069e-01 2.06144825e-01 -2.48099506e-01 -3.09408844e-01
-9.57442939e-01 8.32594931e-01 8.57343078e-01 4.41462725e-01
-3.68895203e-01 2.34644338e-01 -9.01714526e-03 -9.55638289e-02
4.61058885e-01 4.32931691e-01 -4.67340529e-01 5.44948019e-02
-1.48783636e+00 9.93342623e-02 5.09397864e-01 8.43331277e-01
1.16911292e+00 -3.04544151e-01 -3.11491162e-01 8.07651341e-01
2.28413180e-01 2.68308282e-01 2.41983101e-01 -8.39068294e-01
3.17582935e-01 6.86870992e-01 1.22370362e-01 -1.16914928e+00
-3.67752016e-01 -6.83619916e-01 -5.78874469e-01 2.50193536e-01
4.42246228e-01 -8.32422897e-02 -1.13332665e+00 1.32614040e+00
4.45071608e-01 6.27820343e-02 -7.49042988e-01 9.94867563e-01
8.00553858e-01 3.70162696e-01 -3.58778164e-02 -5.00738323e-02
1.26038122e+00 -1.31489813e+00 -7.29414701e-01 -5.24966359e-01
4.67421830e-01 -1.09545279e+00 1.45690215e+00 1.22865990e-01
-1.24925172e+00 -4.69502002e-01 -1.03973317e+00 -7.67383277e-02
-4.15306479e-01 2.79671192e-01 5.84946036e-01 4.24267948e-01
-8.50004733e-01 8.02637696e-01 -7.74326384e-01 -2.09120676e-01
6.06677830e-01 3.86522591e-01 3.60262394e-01 1.29944503e-01
-9.55691040e-01 4.21161622e-01 1.95244849e-01 1.66211426e-01
-4.85337555e-01 -8.18791866e-01 -4.23484385e-01 1.63682252e-01
5.86238325e-01 -5.27418554e-01 1.34524083e+00 -1.05881500e+00
-1.30463433e+00 5.68924546e-01 -4.21922863e-01 -2.27840006e-01
8.49048078e-01 -2.97450364e-01 1.40743017e-01 3.00759912e-01
4.24425870e-01 1.04659724e+00 9.32051837e-01 -1.24469173e+00
-8.48758340e-01 -1.76976383e-01 8.22583586e-03 2.07430601e-01
-4.61270005e-01 -2.16704477e-02 -1.24559772e+00 -7.46365309e-01
4.96455640e-01 -8.38029623e-01 -4.16774929e-01 1.50143325e-01
-6.51375175e-01 5.51986136e-02 7.72382557e-01 -5.83384514e-01
1.49804044e+00 -2.35974026e+00 -2.99732715e-01 4.36936766e-01
4.19824362e-01 3.04419607e-01 -7.92200044e-02 1.53664082e-01
-5.98024987e-02 2.47004822e-01 -2.86412984e-01 -2.64755219e-01
-2.85875171e-01 -2.91043758e-01 -2.75540631e-02 8.24165344e-02
1.11502111e-01 9.52944517e-01 -7.37520337e-01 -8.33980978e-01
2.08430871e-01 2.04362527e-01 -7.95729101e-01 1.24024689e-01
-3.17832649e-01 2.76803941e-01 -5.03903985e-01 5.46703398e-01
8.64228904e-01 -6.52170718e-01 -4.29946855e-02 -5.47764421e-01
-2.62267709e-01 1.32285923e-01 -1.20446849e+00 1.61574388e+00
-2.00140893e-01 5.82132041e-01 1.26739785e-01 -6.71800733e-01
8.29689741e-01 -2.89259069e-02 5.50929725e-01 -7.65615106e-01
2.01113448e-01 3.55807781e-01 -3.92972343e-02 -3.54474872e-01
3.53440434e-01 3.69315654e-01 3.39691132e-01 5.15651464e-01
-2.79524833e-01 -1.81944862e-01 3.35730255e-01 5.38595498e-01
9.41608787e-01 1.66281655e-01 -1.59963921e-01 1.09521272e-02
3.95729154e-01 1.78016201e-01 7.16072381e-01 7.96359479e-01
-2.14100748e-01 7.85910726e-01 4.18300450e-01 -2.10499823e-01
-8.26476276e-01 -8.21864665e-01 2.01037928e-01 1.20338118e+00
5.03042281e-01 -5.95331907e-01 -1.08939075e+00 -8.37058544e-01
-2.22297072e-01 4.66148138e-01 -4.36833858e-01 1.27030060e-01
-4.69819129e-01 -4.66487050e-01 8.09165537e-02 5.92042804e-01
8.80853772e-01 -1.05581582e+00 -4.45345372e-01 1.54398873e-01
-2.52711207e-01 -8.01124930e-01 -1.00992179e+00 7.87676796e-02
-8.68951917e-01 -1.10543001e+00 -7.64959872e-01 -9.11716878e-01
1.05216300e+00 5.53516328e-01 6.72961116e-01 5.08754194e-01
-4.11541611e-01 8.79148394e-03 -2.82499343e-01 -2.30566278e-01
2.04378381e-01 3.14405471e-01 -5.38833737e-01 1.35512426e-01
1.70307860e-01 -4.90518034e-01 -1.22821689e+00 5.32267094e-01
-7.73622990e-01 4.38680679e-01 8.00677478e-01 6.88200355e-01
6.81633353e-01 -2.05116734e-01 2.40861669e-01 -1.03382599e+00
4.53966647e-01 7.09874332e-02 -6.39630914e-01 6.11911602e-02
-5.82568407e-01 -1.98519111e-01 3.89136106e-01 -3.57878208e-01
-1.04091918e+00 5.45854449e-01 -1.98688820e-01 -2.51478523e-01
1.46013483e-01 2.98990130e-01 4.70860768e-03 -2.45927900e-01
4.94273484e-01 1.66931838e-01 -1.30489379e-01 -5.19385159e-01
3.19304913e-01 4.92969990e-01 2.80772984e-01 -4.27875221e-02
9.32802141e-01 4.31446880e-01 -5.01870811e-01 -4.53906804e-01
-1.02802742e+00 -5.03116012e-01 -7.38974273e-01 -4.22232747e-01
8.92797589e-01 -5.10001242e-01 -6.15947008e-01 3.59193444e-01
-1.21633565e+00 -4.03637558e-01 -1.89548191e-02 4.22665864e-01
-2.96464831e-01 2.53791064e-01 -6.01710439e-01 -5.10793924e-01
-3.72729421e-01 -1.28561842e+00 7.99027026e-01 6.30433261e-01
-2.58876771e-01 -6.12753272e-01 -4.66985554e-01 3.97833109e-01
5.17216742e-01 -1.71762705e-01 5.24266005e-01 -5.72405457e-01
-1.08866835e+00 -3.88639838e-01 -6.68848753e-01 1.26840621e-01
1.39095545e-01 1.43469796e-01 -8.57734144e-01 -8.23196173e-02
-2.28114828e-01 -6.66276412e-03 1.03136337e+00 5.09016335e-01
1.57376921e+00 -5.11786491e-02 -5.72326720e-01 6.28623784e-01
1.29435074e+00 1.99869841e-01 4.87441003e-01 2.37792253e-01
6.81053936e-01 3.63660961e-01 8.57375741e-01 2.34732985e-01
2.20380351e-01 5.38576722e-01 3.22220087e-01 -5.94713926e-01
-3.13439846e-01 -3.14629465e-01 -1.14700273e-01 7.60126352e-01
6.11912608e-02 1.02353707e-01 -8.90539467e-01 4.30699140e-01
-1.79503644e+00 -6.64452851e-01 -1.87966108e-01 2.14265966e+00
9.29942608e-01 4.55797285e-01 8.26699380e-03 -8.41368660e-02
9.49033916e-01 1.76564217e-01 -7.75084198e-01 1.15552731e-01
2.18255043e-01 1.57748953e-01 4.88316149e-01 5.13981283e-01
-9.41044927e-01 1.24079359e+00 6.02756929e+00 1.15890253e+00
-1.18011332e+00 1.41554028e-01 7.52417386e-01 -1.47628725e-01
-3.63768160e-01 1.57962546e-01 -9.31558251e-01 6.27144039e-01
2.01141477e-01 8.23618565e-03 2.74900079e-01 8.64747703e-01
5.51939428e-01 -5.41996241e-01 -7.23144650e-01 9.34850931e-01
-5.87083064e-02 -1.33847916e+00 -1.29807770e-01 -1.64443091e-01
5.35065293e-01 -9.13535804e-02 3.43094803e-02 -1.38392329e-01
1.32839918e-01 -5.73971808e-01 6.03241086e-01 4.27199692e-01
6.76930070e-01 -5.80270052e-01 4.66135174e-01 2.50376731e-01
-1.21565902e+00 1.05909064e-01 -1.14102580e-01 3.06131244e-01
2.09973618e-01 7.55188763e-01 -7.55249500e-01 2.07505614e-01
8.99146974e-01 4.66571212e-01 -7.40225554e-01 1.44960999e+00
-3.73534769e-01 5.79228699e-01 -3.89689803e-01 -1.93879738e-01
1.96078062e-01 -1.98598400e-01 4.47100669e-01 1.15805197e+00
1.63727120e-01 -1.41351014e-01 1.86519936e-01 9.24152792e-01
4.35738172e-03 4.18818116e-01 -7.50824213e-02 1.30641803e-01
4.86376971e-01 1.56012905e+00 -1.36204982e+00 -4.73976135e-01
-3.87846917e-01 1.21713805e+00 3.11382920e-01 4.78974134e-01
-1.08516455e+00 -7.93110073e-01 1.35624960e-01 4.55714107e-01
5.72700202e-01 -2.50683248e-01 -7.19494283e-01 -8.81751955e-01
8.68320316e-02 -5.51759541e-01 2.65817523e-01 -8.43750775e-01
-9.22670066e-01 5.95587254e-01 -5.01375139e-01 -1.10174012e+00
3.65901381e-01 -2.04936862e-01 -8.32304955e-01 7.10214019e-01
-1.36383128e+00 -9.33228552e-01 -7.42142618e-01 5.04490435e-01
7.29880273e-01 3.51972163e-01 3.14316094e-01 5.20055413e-01
-7.31240928e-01 7.76217222e-01 -2.60463238e-01 1.30148321e-01
9.24839854e-01 -1.22150671e+00 1.92399949e-01 8.26238096e-01
7.10049272e-02 6.02479577e-01 5.88618100e-01 -9.19681489e-01
-7.98693061e-01 -9.69103336e-01 6.10848069e-01 8.53136480e-02
5.27808368e-01 -3.54743659e-01 -9.81188297e-01 4.12609100e-01
2.48311698e-01 -6.84135333e-02 4.79483098e-01 -1.32096365e-01
-1.23023205e-02 -3.05222392e-01 -7.54700840e-01 9.26605225e-01
9.85608280e-01 -1.98165014e-01 -1.76580667e-01 4.12082255e-01
8.27877045e-01 -3.53203535e-01 -3.39611292e-01 3.39780807e-01
5.28920889e-01 -1.01892805e+00 7.41854191e-01 -3.59741822e-02
3.31254900e-01 -5.47275603e-01 2.58568048e-01 -1.05833340e+00
-3.72351140e-01 -6.94425702e-01 4.03048843e-01 1.23474610e+00
9.52334344e-01 -6.09636247e-01 1.12940276e+00 6.96529329e-01
-3.20500843e-02 -9.42658246e-01 -5.15321672e-01 -4.30151969e-01
-4.17339504e-01 -4.75535482e-01 5.21307707e-01 5.52629828e-01
-3.39335382e-01 1.88410148e-01 -2.06906930e-01 7.30914101e-02
6.26910746e-01 3.77099097e-01 7.73667872e-01 -9.57638383e-01
-1.13945045e-01 -5.23822606e-01 1.83317840e-01 -1.45019138e+00
-4.62120295e-01 -8.35230887e-01 2.07930103e-01 -1.68149424e+00
3.14041764e-01 -5.76712787e-01 -1.61267459e-01 4.93846118e-01
-4.93109852e-01 6.07649982e-01 3.22284192e-01 2.85959423e-01
-6.98278487e-01 3.98361355e-01 1.55022323e+00 -3.14838588e-02
-4.71177787e-01 2.01100081e-01 -7.07420945e-01 1.01562226e+00
1.07512200e+00 -5.52788019e-01 -1.67486697e-01 -3.00030679e-01
1.49936855e-01 -2.67143011e-01 2.20325276e-01 -8.63137305e-01
7.00062275e-01 -9.42263678e-02 5.81493497e-01 -1.07324481e+00
3.16402674e-01 -5.99091709e-01 -1.88432127e-01 6.24633014e-01
-2.77028292e-01 -2.27072313e-01 -4.45303172e-02 4.04826313e-01
-1.63088605e-01 -3.65942776e-01 8.19414794e-01 -2.41708428e-01
-5.32483459e-01 4.96390611e-01 -2.91676879e-01 -1.31045848e-01
1.02243900e+00 -4.49741662e-01 -3.63571309e-02 -3.45104635e-01
-9.64967847e-01 4.59620535e-01 4.46753979e-01 1.46102726e-01
7.06740320e-01 -1.09876239e+00 -2.39213854e-01 2.35139489e-01
-1.93008378e-01 7.46675134e-02 3.29222500e-01 1.08191144e+00
-5.15154958e-01 1.31148711e-01 6.50407150e-02 -6.24961972e-01
-1.38879144e+00 3.33767354e-01 2.42966637e-01 -1.13630846e-01
-5.35944402e-01 9.58700359e-01 4.12918240e-01 -1.65339366e-01
2.80347228e-01 -9.06039216e-03 -1.56127080e-01 2.59626061e-01
4.75814253e-01 2.07552850e-01 -8.09515491e-02 2.03420892e-02
-2.58091062e-01 6.74235344e-01 -3.98112267e-01 -1.54061526e-01
1.17509639e+00 -2.77886569e-01 -8.41529295e-03 1.52174994e-01
9.13394630e-01 1.33864552e-01 -1.29742277e+00 -1.58581838e-01
-7.65006095e-02 -8.06957841e-01 9.53680053e-02 -7.31989741e-01
-1.50978267e+00 8.46770465e-01 5.19953072e-01 2.16851279e-01
1.16861522e+00 2.87321154e-02 9.45413947e-01 -6.68004602e-02
4.02740762e-03 -1.45612872e+00 5.64016223e-01 2.84299940e-01
8.09078276e-01 -1.08257985e+00 -8.35540742e-02 -1.06979597e+00
-6.60050392e-01 9.38774705e-01 1.11174238e+00 -4.97076400e-02
7.11284041e-01 3.15290928e-01 6.39267564e-02 -2.31319293e-01
-3.77887160e-01 -4.47623342e-01 1.84756383e-01 2.26466089e-01
3.88859898e-01 -1.29400447e-01 -6.74936891e-01 5.81447721e-01
-1.69593498e-01 -3.34196426e-02 3.09239000e-01 7.10403204e-01
-7.58311272e-01 -1.07458544e+00 -2.19905362e-01 7.91410625e-01
-4.91094798e-01 -3.10005128e-01 -3.61602336e-01 5.04667819e-01
5.04478030e-02 8.09025884e-01 1.06258318e-01 -3.42011869e-01
2.11024895e-01 -1.37143657e-01 -2.48187371e-02 -7.03358412e-01
-5.72665989e-01 5.68476737e-01 -2.04572082e-01 -9.84437287e-01
-2.85257071e-01 -4.88700002e-01 -1.33672249e+00 -1.47306040e-01
-6.48686230e-01 2.89963894e-02 7.00747490e-01 6.02092922e-01
4.88801807e-01 6.20113075e-01 6.71887398e-01 -7.23399520e-01
-9.19217989e-02 -8.09144318e-01 -4.11875337e-01 2.47452021e-01
-2.00713113e-01 -5.86978376e-01 -4.18883026e-01 -3.25797498e-02]
|
[9.522497177124023, -0.07616961747407913]
|
e671bce3-a9cd-4c83-9520-182ffe22bc4c
|
dimsum-laysumm-20-bart-based-approach-for
|
2010.09252
| null |
https://arxiv.org/abs/2010.09252v1
|
https://arxiv.org/pdf/2010.09252v1.pdf
|
Dimsum @LaySumm 20: BART-based Approach for Scientific Document Summarization
|
Lay summarization aims to generate lay summaries of scientific papers automatically. It is an essential task that can increase the relevance of science for all of society. In this paper, we build a lay summary generation system based on the BART model. We leverage sentence labels as extra supervision signals to improve the performance of lay summarization. In the CL-LaySumm 2020 shared task, our model achieves 46.00\% Rouge1-F1 score.
|
['Pascale Fung', 'Wenliang Dai', 'Dan Su', 'Tiezheng Yu']
|
2020-10-19
| null | null | null | null |
['lay-summarization', 'scientific-article-summarization']
|
['natural-language-processing', 'natural-language-processing']
|
[ 3.91972065e-01 6.46580875e-01 -5.12063682e-01 -1.87288284e-01
-1.42242026e+00 -3.03127617e-01 7.86597073e-01 4.31355000e-01
-2.49563485e-01 1.60386133e+00 1.02960563e+00 -2.44081989e-01
-9.08533335e-02 -4.63727742e-01 -1.00088739e+00 -8.23888406e-02
4.93163049e-01 3.27124119e-01 -7.30647519e-02 8.63956809e-02
7.24656105e-01 -5.48594035e-02 -9.80049729e-01 8.35784495e-01
1.53691745e+00 3.80193412e-01 2.34953657e-01 9.98784900e-01
-1.92949563e-01 8.21334541e-01 -1.33714211e+00 -5.50291359e-01
-4.85551804e-01 -7.10957646e-01 -8.87209952e-01 -2.15604737e-01
6.85945630e-01 5.87504879e-02 -1.65848240e-01 1.04407716e+00
5.94671428e-01 -1.19662471e-01 1.00689697e+00 -7.68131733e-01
-7.21666873e-01 1.39897346e+00 -5.38669765e-01 3.80283833e-01
4.19715315e-01 -1.82586789e-01 1.18519151e+00 -7.93528140e-01
8.87516975e-01 9.93763149e-01 3.61815929e-01 6.23974144e-01
-1.11745465e+00 -5.72599888e-01 3.16122137e-02 -8.04371536e-02
-8.24384809e-01 -7.16344655e-01 6.08395576e-01 -2.11615041e-01
7.41611063e-01 4.89686191e-01 5.76966286e-01 1.15006113e+00
4.54713464e-01 1.08813691e+00 5.40585041e-01 -4.20676440e-01
7.63741657e-02 -1.55751362e-01 4.84515548e-01 3.33440036e-01
9.45000172e-01 -8.79908860e-01 -1.02351320e+00 -1.49899367e-02
1.45781502e-01 -2.59824395e-01 -3.18257600e-01 6.01543367e-01
-1.53938520e+00 6.95052147e-01 2.85677612e-01 1.01547204e-01
-4.26897973e-01 2.73098737e-01 4.43392158e-01 -9.94634107e-02
9.92006719e-01 1.16104090e+00 -9.39941034e-02 -1.14354461e-01
-1.47714496e+00 5.33490419e-01 7.81300426e-01 1.02408838e+00
8.06283429e-02 -1.44811258e-01 -9.33562577e-01 7.09220886e-01
-6.19571358e-02 6.53824031e-01 3.05233747e-01 -1.03606355e+00
7.14439094e-01 6.27543688e-01 -2.27825139e-02 -6.80389702e-01
-1.53096929e-01 -8.00192595e-01 -9.75658178e-01 -7.00676799e-01
-4.14431393e-01 -5.24861097e-01 -7.07540333e-01 1.14215779e+00
-1.58830553e-01 1.28218472e-01 2.54496098e-01 2.49151066e-01
1.78365445e+00 9.37703967e-01 -2.40446813e-02 -5.97071409e-01
9.96849716e-01 -1.15538955e+00 -1.38179207e+00 -1.56827718e-01
6.21997118e-01 -7.88211644e-01 6.17452145e-01 2.92954087e-01
-1.32952034e+00 -3.36278558e-01 -1.34817004e+00 -2.13311300e-01
-3.14778090e-02 6.13820374e-01 4.18285042e-01 8.68381746e-03
-9.83635664e-01 8.24822128e-01 -4.33401883e-01 1.88469309e-02
7.31314480e-01 -9.80309397e-02 -1.38838083e-01 1.00715756e-01
-1.18964875e+00 6.19426012e-01 4.94843513e-01 -1.17516123e-01
-5.15589058e-01 -9.33215201e-01 -7.18736351e-01 -1.31298853e-02
1.81288391e-01 -1.04003787e+00 1.35701811e+00 1.85636014e-01
-1.24369359e+00 7.66221046e-01 -5.49413264e-01 -9.21816885e-01
5.14614284e-01 -5.88829994e-01 -2.17601821e-01 2.02937588e-01
6.38089299e-01 7.13716209e-01 3.09471726e-01 -1.05631423e+00
-5.94632268e-01 8.10539648e-02 -3.94232810e-01 1.81793496e-01
-8.12560618e-02 -6.14328496e-02 -1.53362423e-01 -7.28601992e-01
-2.64751643e-01 -4.69877213e-01 -1.77326575e-01 -8.79572272e-01
-1.14788091e+00 -7.28648245e-01 4.93856072e-01 -8.00005794e-01
1.60132766e+00 -1.57543099e+00 3.29134986e-02 -3.59529853e-01
2.53872454e-01 4.13339853e-01 -6.74948543e-02 6.92824483e-01
1.40670881e-01 6.10216916e-01 -2.73046613e-01 -4.66542721e-01
-1.20853223e-01 -2.69501716e-01 -7.54378140e-01 -8.84663835e-02
5.24883509e-01 1.13498044e+00 -1.12475729e+00 -8.03761899e-01
-2.48988256e-01 -3.15236337e-02 -2.86423653e-01 2.41297454e-01
-4.88921642e-01 2.49618068e-01 -5.86119890e-01 2.69150943e-01
2.48337135e-01 -4.45094466e-01 -2.36327752e-01 -6.47467077e-02
-3.28714073e-01 7.87238657e-01 -2.85178155e-01 1.84130859e+00
-1.16638668e-01 8.73165727e-01 -3.80357623e-01 -7.24520922e-01
9.99970555e-01 4.23199505e-01 2.13360101e-01 -1.83084160e-01
-1.15868039e-01 3.34873557e-01 -2.57979184e-01 -2.77337611e-01
8.75560999e-01 2.49565497e-01 -2.79541433e-01 3.84372383e-01
1.61740258e-01 -5.86818397e-01 8.57381463e-01 8.28904331e-01
1.14600384e+00 9.28367861e-03 2.93913215e-01 -4.18584615e-01
4.93232608e-01 2.00887725e-01 4.22359228e-01 8.62185717e-01
2.24846169e-01 9.09674942e-01 8.60438645e-01 -6.38897195e-02
-9.30441976e-01 -7.73262441e-01 -3.17198902e-01 4.90637273e-01
-2.32084230e-01 -7.50760913e-01 -8.28198195e-01 -8.20716202e-01
-1.27344593e-01 1.35230017e+00 -4.98030484e-01 -3.36708456e-01
-5.43699801e-01 -7.60316133e-01 6.80159211e-01 1.36487588e-01
4.95479345e-01 -1.27373850e+00 -1.77296534e-01 2.65316755e-01
-6.70555830e-01 -9.94438589e-01 -7.48601556e-01 9.47266817e-02
-8.15113664e-01 -7.01505601e-01 -9.37307298e-01 -7.75626719e-01
4.66761082e-01 1.40806571e-01 1.23553121e+00 -3.08742911e-01
2.95681339e-02 -2.36029565e-01 -4.07968313e-01 -1.05350256e+00
-5.90876698e-01 7.14735806e-01 -2.28560880e-01 -5.93733013e-01
1.21099181e-01 -1.64334089e-01 -4.11590934e-01 -4.97312099e-01
-5.89131057e-01 5.48705518e-01 7.14596868e-01 7.82150865e-01
6.50671601e-01 -4.72281307e-01 1.37816489e+00 -1.13106310e+00
1.22500706e+00 -3.51804703e-01 5.68380244e-02 4.98597890e-01
-6.17123544e-01 2.59867013e-01 5.37455559e-01 4.25107181e-02
-1.07086420e+00 -4.94580299e-01 -6.56772852e-02 1.73325926e-01
2.11860105e-01 7.80536473e-01 -2.40461174e-02 6.46843135e-01
5.90912759e-01 3.40374172e-01 -3.38880569e-01 -5.74349523e-01
3.83100390e-01 1.04187548e+00 7.48990178e-01 -4.49554920e-01
2.17492491e-01 -1.27901018e-01 6.96309656e-02 -9.63945091e-01
-1.74870479e+00 -4.20392215e-01 -2.80674428e-01 -1.64694667e-01
8.22762191e-01 -9.83503282e-01 -3.20516497e-01 -3.85464393e-02
-1.62952614e+00 8.68326426e-02 -3.60596746e-01 2.80126035e-01
-2.84168661e-01 2.83260375e-01 -5.46627998e-01 -5.92520595e-01
-1.22161031e+00 -6.33697152e-01 1.28957725e+00 5.54448783e-01
-5.76852620e-01 -6.40662432e-01 3.05646956e-01 4.53961223e-01
-5.52495271e-02 4.26292181e-01 3.66985112e-01 -8.15113127e-01
-1.04287036e-01 -1.87619478e-01 -2.96612114e-01 3.73741001e-01
1.11317120e-01 2.04937607e-01 -6.44246757e-01 1.44078374e-01
-3.10930818e-01 -3.42955977e-01 1.62383544e+00 7.40736127e-01
1.31952202e+00 -6.80095732e-01 -3.78535330e-01 3.61008435e-01
8.59271049e-01 -2.37241201e-03 4.47635800e-01 2.29311064e-01
7.49312162e-01 5.40312588e-01 6.21761203e-01 3.54587793e-01
3.06908101e-01 1.14091538e-01 -2.11125672e-01 5.24706990e-02
-5.62584221e-01 -5.93569994e-01 3.50075901e-01 1.47420645e+00
1.57248154e-01 -5.24289489e-01 -8.19181383e-01 6.90068483e-01
-1.91678965e+00 -1.18112350e+00 -5.31850338e-01 1.71683228e+00
1.40030324e+00 4.35571700e-01 -3.12193453e-01 -1.00694790e-01
7.24032402e-01 4.68987316e-01 -2.59665072e-01 -6.04180217e-01
-3.21930051e-01 1.70677409e-01 3.01911622e-01 4.50707585e-01
-1.10575402e+00 9.04854000e-01 6.64357138e+00 1.04320347e+00
-6.52197897e-01 -2.81921268e-01 8.07485402e-01 -2.27845907e-01
-4.82434541e-01 -2.32857272e-01 -1.13023901e+00 7.33824730e-01
1.18375647e+00 -1.11348641e+00 -4.08696562e-01 4.53896374e-01
4.05342102e-01 -1.12929620e-01 -1.16537309e+00 6.43380404e-01
4.44738925e-01 -1.98495936e+00 5.85296571e-01 -1.36507273e-01
1.32005918e+00 -1.45685956e-01 -2.15124711e-01 1.99679524e-01
3.76691937e-01 -1.09298515e+00 4.88612473e-01 8.98085892e-01
9.94749606e-01 -7.80543685e-01 1.12255895e+00 5.33111393e-01
-4.73418057e-01 6.48612380e-01 -3.62769276e-01 1.45370379e-01
4.43635076e-01 1.16685367e+00 -1.02556252e+00 7.94546068e-01
1.06571227e-01 1.31847644e+00 -6.43284917e-01 1.14211690e+00
-6.87011063e-01 9.85851705e-01 1.53747350e-01 -4.28786188e-01
1.49275273e-01 -1.32328227e-01 7.13681400e-01 1.54991245e+00
3.02572161e-01 5.26436344e-02 6.56716351e-04 8.75513673e-01
-8.74768674e-01 2.71951780e-02 -6.63233280e-01 -3.64736915e-01
4.59564686e-01 1.11673021e+00 -4.13040489e-01 -7.24233568e-01
2.56471038e-01 7.39315987e-01 1.90184295e-01 6.01609312e-02
-5.76710582e-01 -7.21474826e-01 5.81998788e-02 -2.17676118e-01
-1.06052704e-01 1.93688318e-01 -6.75409138e-01 -1.34162223e+00
-1.05165511e-01 -6.25804782e-01 1.67389125e-01 -8.08644712e-01
-1.14247835e+00 3.66544485e-01 -1.37099847e-01 -9.97656465e-01
-1.23323157e-01 -1.00674428e-01 -9.85022843e-01 7.32808053e-01
-1.42034602e+00 -1.00274563e+00 2.40166765e-03 -4.88560587e-01
1.09601235e+00 -1.77664697e-01 5.85455179e-01 -2.57887304e-01
-8.00126493e-01 2.61311114e-01 2.65602320e-01 -1.17344903e-02
1.08624339e+00 -1.47954035e+00 6.96417630e-01 8.32631469e-01
1.29765838e-01 6.83729291e-01 9.37375009e-01 -1.06240451e+00
-8.65249455e-01 -1.38385403e+00 1.66177869e+00 -5.64088285e-01
5.91484666e-01 -4.64994740e-03 -7.46837676e-01 3.34815770e-01
6.05043113e-01 -7.64475286e-01 7.15565681e-01 1.75744563e-01
2.21358702e-01 -4.99390848e-02 -6.17192626e-01 5.82364619e-01
9.68936503e-01 -1.02026023e-01 -1.12054288e+00 6.45948350e-01
1.30602491e+00 -2.47307554e-01 -8.14189315e-01 2.65788794e-01
1.92856014e-01 -2.56114066e-01 8.53668213e-01 -6.57707036e-01
1.40209901e+00 -7.41196275e-02 4.46444780e-01 -1.57952487e+00
-3.69152725e-01 -6.68930650e-01 -3.47072273e-01 1.47383738e+00
7.44634748e-01 -2.32794464e-01 6.42731309e-01 1.68982580e-01
-6.58059716e-01 -6.70980752e-01 -6.91404760e-01 -5.59429109e-01
3.55825633e-01 1.08767681e-01 5.00910103e-01 5.20073771e-01
4.96290654e-01 1.11774492e+00 -3.29875529e-01 -4.79999065e-01
8.74022782e-01 1.55758455e-01 6.00681543e-01 -1.24388587e+00
2.07061723e-01 -7.28610575e-01 2.33956113e-01 -1.00296319e+00
3.41510892e-01 -1.08728993e+00 1.33548677e-01 -2.33104992e+00
8.34374547e-01 2.50095129e-01 -2.84697622e-01 3.11948478e-01
-7.80469716e-01 -1.01093791e-01 -3.99389453e-02 2.64425367e-01
-1.20866454e+00 7.29376674e-01 1.43499005e+00 -4.17249233e-01
-2.35302821e-02 -3.50267775e-02 -1.13193417e+00 3.76060575e-01
9.12743926e-01 -5.11380672e-01 -1.43757448e-01 -4.35577393e-01
3.95854503e-01 1.67953953e-01 -8.27615038e-02 -1.00769603e+00
3.69230390e-01 -1.83773816e-01 3.74996543e-01 -1.19272804e+00
-2.21338764e-01 2.61575520e-01 -1.73605934e-01 3.79391700e-01
-1.10113192e+00 -1.53144538e-01 8.09639245e-02 5.03895104e-01
-3.02709222e-01 -4.66713816e-01 1.23316564e-01 6.56061396e-02
1.66529000e-01 -1.16407007e-01 -1.63043320e-01 4.30944145e-01
6.54049218e-01 4.48399365e-01 -9.33114886e-01 -3.92046899e-01
-4.58064340e-02 6.60570741e-01 5.49629815e-02 2.08369404e-01
6.18446410e-01 -1.05107713e+00 -1.57003164e+00 -3.56474906e-01
6.22248948e-02 3.20317239e-01 -1.09401252e-02 5.98272145e-01
-5.89955866e-01 1.02784383e+00 9.94277596e-02 -1.82487473e-01
-1.17190480e+00 7.43014887e-02 -3.47938567e-01 -6.38236225e-01
-6.41216874e-01 8.67035508e-01 5.24349734e-02 -6.01731651e-02
4.52229232e-02 -1.25123233e-01 -6.65568709e-01 1.27265051e-01
9.14803267e-01 6.62743509e-01 9.78670120e-02 -2.05316916e-01
-1.59328535e-01 7.07676336e-02 -3.70455712e-01 -2.25675315e-01
1.53014994e+00 2.69516230e-01 -4.08301950e-01 6.96767092e-01
8.81289601e-01 1.98123157e-01 -8.27578783e-01 1.05165817e-01
3.32431555e-01 9.39397588e-02 1.98837787e-01 -1.03146899e+00
-3.14942271e-01 8.77531111e-01 -5.62134266e-01 1.83039159e-01
5.00819385e-01 9.80574936e-02 8.26749086e-01 4.68391955e-01
-8.60051736e-02 -9.87033546e-01 5.47861755e-02 4.93532777e-01
1.25641751e+00 -1.24154317e+00 5.43158412e-01 -2.69028515e-01
-6.63410783e-01 9.42434728e-01 3.41577411e-01 -1.10810786e-01
-1.17266759e-01 -6.99342787e-02 -3.15900445e-01 -2.43862927e-01
-1.22790360e+00 1.52291849e-01 8.43159437e-01 1.66219190e-01
9.19428229e-01 1.75808907e-01 -1.08063185e+00 9.63046134e-01
-4.18849081e-01 4.00674716e-02 8.77764285e-01 4.89663780e-01
-7.60699987e-01 -8.65909338e-01 -2.47967895e-02 1.02044356e+00
-8.51492226e-01 -3.35380971e-01 -8.19081187e-01 8.55730921e-02
-2.38917723e-01 1.20204532e+00 -2.02254415e-01 -3.75209982e-03
2.72670925e-01 5.32664657e-02 9.99090001e-02 -1.07377052e+00
-5.86506605e-01 -2.22786646e-02 6.31258607e-01 4.79451939e-02
-2.88361549e-01 -7.47037351e-01 -1.19433439e+00 -2.34681845e-01
-2.96340644e-01 8.34591329e-01 6.93515301e-01 6.12723827e-01
6.00292683e-01 1.15717351e+00 4.55465972e-01 -4.25460786e-01
-5.58125794e-01 -1.43409944e+00 -2.85464436e-01 1.69583559e-01
2.93028235e-01 -6.84904829e-02 -4.26940590e-01 3.79735947e-01]
|
[12.489481925964355, 9.552290916442871]
|
fb898fe4-5e50-4b4a-a2b5-f48bcf3a178e
|
volumetric-supervised-contrastive-learning
|
2206.08158
| null |
https://arxiv.org/abs/2206.08158v1
|
https://arxiv.org/pdf/2206.08158v1.pdf
|
Volumetric Supervised Contrastive Learning for Seismic Semantic Segmentation
|
In seismic interpretation, pixel-level labels of various rock structures can be time-consuming and expensive to obtain. As a result, there oftentimes exists a non-trivial quantity of unlabeled data that is left unused simply because traditional deep learning methods rely on access to fully labeled volumes. To rectify this problem, contrastive learning approaches have been proposed that use a self-supervised methodology in order to learn useful representations from unlabeled data. However, traditional contrastive learning approaches are based on assumptions from the domain of natural images that do not make use of seismic context. In order to incorporate this context within contrastive learning, we propose a novel positive pair selection strategy based on the position of slices within a seismic volume. We show that the learnt representations from our method out-perform a state of the art contrastive learning methodology in a semantic segmentation task.
|
['Ghassan AlRegib', 'Mohit Prabhushankar', 'Kiran Kokilepersaud']
|
2022-06-16
| null | null | null | null |
['seismic-interpretation']
|
['miscellaneous']
|
[ 3.72867733e-01 2.44345739e-01 2.54105255e-02 -7.14875281e-01
-9.18073714e-01 -5.54273725e-01 5.88508904e-01 3.58461708e-01
-5.72392702e-01 8.63568604e-01 -2.13759273e-01 -1.92006588e-01
-5.40935695e-02 -1.08049500e+00 -7.37773120e-01 -8.03536713e-01
1.55405030e-01 7.66435742e-01 5.57172596e-01 -1.52549505e-01
4.33189034e-01 4.62913215e-01 -1.48386765e+00 5.81606179e-02
8.38810921e-01 8.01119208e-01 2.86260664e-01 1.25699684e-01
-7.24823236e-01 7.93819547e-01 -3.51696581e-01 -1.45301539e-02
3.19111556e-01 -7.13464677e-01 -1.09978211e+00 3.61808956e-01
4.13748682e-01 -2.34038845e-01 3.28170434e-02 1.18233049e+00
1.71862558e-01 2.33114481e-01 8.67046475e-01 -9.80115235e-01
-1.71649009e-01 6.77819848e-01 -8.33903909e-01 3.41682702e-01
-1.96187943e-01 -1.70649752e-01 1.08105206e+00 -9.78878915e-01
5.17197549e-01 9.24771905e-01 5.40242076e-01 3.75703484e-01
-1.34381914e+00 -5.02358079e-01 2.09083959e-01 9.53984037e-02
-1.18646049e+00 -2.26518184e-01 1.29100573e+00 -5.19212246e-01
3.87893200e-01 2.21340910e-01 8.06671143e-01 6.66190386e-01
-3.91124964e-01 8.09654713e-01 1.44684601e+00 -6.91842854e-01
6.91445589e-01 3.51031795e-02 2.05726311e-01 6.00704789e-01
2.51530588e-01 -6.22441843e-02 -3.84632647e-01 8.05662572e-02
8.30264866e-01 -5.74980900e-02 -1.66916236e-01 -6.52443171e-01
-9.11593378e-01 9.01999593e-01 5.44420719e-01 5.64485312e-01
-1.49254844e-01 1.39522702e-01 3.56016576e-01 9.76615176e-02
7.83236086e-01 5.66850364e-01 -3.18497002e-01 2.41060466e-01
-1.42093146e+00 1.67652339e-01 7.29556501e-01 4.31126177e-01
1.28104305e+00 1.96454525e-01 4.08406526e-01 7.21948385e-01
2.91712344e-01 2.48438150e-01 4.72499698e-01 -6.93966925e-01
2.28797719e-01 7.06885695e-01 -4.01915610e-02 -7.97571778e-01
-2.52823561e-01 -3.75763297e-01 -5.07305622e-01 5.07531226e-01
5.99430144e-01 8.68959576e-02 -1.25584805e+00 1.41812217e+00
2.91141331e-01 1.69400141e-01 5.02836294e-02 8.87802303e-01
5.63080609e-01 3.68293732e-01 -3.95183265e-03 -1.68076172e-01
8.04604530e-01 -9.34766054e-01 -4.10064727e-01 -3.50966364e-01
5.51293612e-01 -4.70459789e-01 1.23113286e+00 3.04010391e-01
-8.65617156e-01 -3.32800925e-01 -1.28034949e+00 5.21935932e-02
-4.07381833e-01 -3.35515141e-01 4.22617465e-01 6.61800206e-01
-8.41528594e-01 8.87886882e-01 -9.18407798e-01 1.19311353e-02
5.54704547e-01 3.33016247e-01 -2.98144609e-01 5.25912419e-02
-8.67238045e-01 7.45026708e-01 5.95729232e-01 3.46768737e-01
-8.23981822e-01 -3.91850144e-01 -7.56007552e-01 3.92764360e-02
6.60300255e-01 -2.65373856e-01 1.22612786e+00 -1.42212987e+00
-1.17197740e+00 1.08553195e+00 2.06605151e-01 -4.23732072e-01
7.81258464e-01 -1.72361359e-01 1.55861646e-01 4.83764917e-01
1.40166491e-01 5.90589166e-01 9.19246972e-01 -1.73225379e+00
-5.48687518e-01 -4.09136683e-01 -4.75419424e-02 1.38270304e-01
-2.00939372e-01 -4.54483390e-01 -9.32566449e-02 -7.18851864e-01
5.73068321e-01 -7.36144423e-01 -4.88636523e-01 1.45870894e-01
-4.29126799e-01 2.05440223e-02 1.00322270e+00 -5.60696185e-01
6.24342024e-01 -1.99740970e+00 9.77609754e-02 4.62086767e-01
3.58435690e-01 3.60422164e-01 2.86428690e-01 7.27685392e-02
-3.07434909e-02 1.71947554e-01 -1.05171931e+00 -1.46917030e-01
-2.63386458e-01 3.99650663e-01 -1.46836162e-01 4.31068331e-01
2.30683342e-01 5.01824260e-01 -1.13710177e+00 -9.66831088e-01
3.56589288e-01 1.47371590e-01 -4.40834701e-01 1.49801701e-01
-4.03159976e-01 8.68383229e-01 -6.59317076e-01 5.08971870e-01
6.40599430e-01 -2.99532771e-01 2.53797531e-01 5.91644011e-02
-1.63972244e-01 2.14476585e-01 -1.19089925e+00 1.80155265e+00
-2.64115751e-01 4.58838522e-01 -8.84255618e-02 -1.63250482e+00
9.60968733e-01 1.71002403e-01 6.07129395e-01 -4.81009960e-01
1.63340017e-01 7.65895784e-01 -2.47222036e-02 -4.57406372e-01
4.83412385e-01 -6.04917645e-01 1.78437993e-01 5.50594389e-01
-3.95311648e-03 -5.22673249e-01 1.07910484e-01 -1.30911618e-01
9.37053859e-01 4.26669031e-01 2.42930636e-01 -4.99799073e-01
4.92094815e-01 2.80922174e-01 6.61225677e-01 5.36514103e-01
4.47284710e-03 9.29380894e-01 3.02408010e-01 -6.13823175e-01
-1.24018359e+00 -1.00851989e+00 -1.80688158e-01 6.99250698e-01
2.00429305e-01 1.49119362e-01 -9.18044448e-01 -8.67186069e-01
-3.26823533e-01 4.08502132e-01 -5.97016692e-01 7.76016712e-02
-9.29410040e-01 -8.00094604e-01 2.83819437e-01 6.89638495e-01
5.99529624e-01 -1.09273624e+00 -1.00220561e+00 4.10411537e-01
-5.56896776e-02 -8.35643113e-01 4.10865620e-02 6.28008723e-01
-1.28819084e+00 -1.06964529e+00 -9.08889532e-01 -8.85233462e-01
9.89625096e-01 2.04903528e-01 1.09162784e+00 2.22028926e-01
-2.86863912e-02 2.44820967e-01 -3.87142301e-01 -2.15764508e-01
-2.51468956e-01 2.55022883e-01 -4.88132000e-01 3.46095040e-02
2.05840886e-01 -6.53999805e-01 -4.76982862e-01 2.36381292e-02
-1.07488465e+00 7.10933581e-02 4.58189428e-01 1.05185187e+00
7.41539240e-01 -1.14738710e-01 7.00938523e-01 -1.28582108e+00
9.36248675e-02 -4.31959480e-01 -7.28721499e-01 3.06559563e-01
-6.23099208e-01 4.52626407e-01 3.80372465e-01 -2.72536457e-01
-1.18094039e+00 3.23201627e-01 -1.79006234e-01 -2.78460175e-01
-1.51251510e-01 6.78546906e-01 -7.85652921e-02 -1.12645097e-01
5.60203314e-01 -1.62523438e-03 -1.14040792e-01 -4.55764860e-01
2.66092300e-01 4.70116884e-01 4.96183097e-01 -6.02891624e-01
8.01577866e-01 7.11871624e-01 7.45683461e-02 -7.80705094e-01
-9.23886716e-01 -4.13603365e-01 -9.23823714e-01 -2.64559865e-01
7.81277180e-01 -5.75111687e-01 1.51660874e-01 3.65737468e-01
-7.08375692e-01 -4.25876677e-01 -5.52390814e-01 6.55139744e-01
-6.30132079e-01 6.43418193e-01 -3.92551750e-01 -9.71176207e-01
-8.03701133e-02 -1.12217104e+00 8.95672381e-01 2.06181258e-01
-7.49056190e-02 -9.86627638e-01 -3.96955535e-02 3.22778553e-01
1.23568915e-01 3.15719634e-01 1.03762305e+00 -8.30995262e-01
-6.23435378e-01 -4.72793542e-02 -1.81137383e-01 4.40011203e-01
2.96657979e-01 -2.29192853e-01 -1.19344902e+00 4.09900248e-02
2.90147036e-01 -4.02799934e-01 1.12159371e+00 3.35336268e-01
1.08631408e+00 8.83197486e-02 -1.97431073e-01 3.37452561e-01
1.62374079e+00 2.26142228e-01 5.51344156e-01 4.27767396e-01
7.94432878e-01 1.00704861e+00 5.43324888e-01 2.00121239e-01
9.13476646e-02 2.87307501e-01 5.82438529e-01 -4.14874494e-01
1.44438192e-01 -1.47750616e-01 -3.79103988e-01 5.23013890e-01
-2.13084266e-01 -2.16937602e-01 -1.16557395e+00 8.30522478e-01
-1.75745726e+00 -7.05686033e-01 -2.18345821e-01 2.26559639e+00
9.37377691e-01 3.53247166e-01 -7.80671090e-02 8.76412094e-01
6.71493888e-01 1.83915600e-01 -4.59990740e-01 1.22540921e-01
-7.56006390e-02 3.32319796e-01 6.33018553e-01 2.87181139e-01
-1.27701676e+00 8.81672442e-01 6.22338533e+00 5.80437183e-01
-1.29327977e+00 1.07423782e-01 8.07410598e-01 3.56198668e-01
-5.80191016e-01 1.72936216e-01 -2.30532691e-01 2.93476522e-01
5.87073803e-01 2.52530903e-01 -8.47169980e-02 8.56630087e-01
-1.32888146e-02 -3.78141016e-01 -1.13072109e+00 9.45688069e-01
-1.23584315e-01 -1.32930422e+00 -1.93530634e-01 5.81631623e-02
7.70160735e-01 -1.72986850e-01 -1.11528166e-01 -7.27183968e-02
2.02055737e-01 -9.44487631e-01 7.70140707e-01 3.58938903e-01
3.77402872e-01 -6.34656191e-01 7.20083535e-01 2.47536212e-01
-9.76955354e-01 -1.30990725e-02 -6.31895363e-02 -2.93166880e-02
3.40880960e-01 5.60033143e-01 -7.52655029e-01 4.46800321e-01
6.09753132e-01 5.14558315e-01 -4.07996923e-01 1.07239783e+00
-2.50217855e-01 7.23170519e-01 -2.42574751e-01 2.61508405e-01
6.49226069e-01 -2.43717596e-01 2.83678740e-01 8.40546548e-01
2.05415487e-01 -2.07048021e-02 3.70921463e-01 9.22654688e-01
-7.62886778e-02 2.70570666e-01 -5.99382937e-01 3.39876581e-03
1.77177638e-01 1.05595589e+00 -1.57008851e+00 -4.12121236e-01
-5.27283490e-01 7.31224835e-01 2.11007833e-01 7.61365816e-02
-4.35385525e-01 7.83389881e-02 -1.85589388e-01 3.10943902e-01
2.48290747e-01 -2.44189397e-01 -5.75221717e-01 -9.24897373e-01
1.28471464e-01 -5.16934156e-01 2.59809077e-01 -4.68984425e-01
-1.25379014e+00 5.46554744e-01 2.39681482e-01 -1.38476110e+00
-4.10338134e-01 -3.14086020e-01 -4.32646036e-01 4.98237044e-01
-1.75720668e+00 -1.10643029e+00 -2.66649067e-01 3.21437031e-01
7.38443911e-01 -6.67668087e-03 5.71349025e-01 3.62072587e-01
-8.79272521e-02 3.96106839e-02 6.55529574e-02 3.05938572e-01
5.18536329e-01 -1.51850724e+00 2.20120959e-02 9.25416648e-01
3.60609859e-01 1.86684072e-01 7.89973617e-01 -5.33816576e-01
-6.63355231e-01 -8.33156526e-01 6.43914998e-01 1.19019724e-01
4.90179420e-01 -1.08474068e-01 -1.42956257e+00 5.97389519e-01
4.79575694e-02 4.12037373e-01 6.59599960e-01 -1.15735531e-01
-2.15431198e-01 -1.94883533e-02 -1.15576530e+00 3.52708340e-01
6.79908991e-01 -5.62614501e-01 -8.64854932e-01 1.05130434e-01
2.61545032e-01 -1.25972748e-01 -4.12467927e-01 4.02850211e-01
2.36426815e-01 -7.81997621e-01 7.71296263e-01 -2.86858737e-01
5.10155141e-01 -4.06383544e-01 -7.59685636e-02 -1.07372606e+00
3.61284316e-01 -3.85910296e-03 4.22994435e-01 1.03265023e+00
4.15629208e-01 -4.40166175e-01 1.12140584e+00 6.09262168e-01
-3.45552802e-01 -4.21540111e-01 -9.25156951e-01 -7.72593617e-01
2.58836031e-01 -1.13271162e-01 3.63180071e-01 1.22621024e+00
-3.08706790e-01 2.99292594e-01 -2.51332134e-01 -1.29532769e-01
7.70026207e-01 3.04974198e-01 2.78551221e-01 -1.75064135e+00
-2.55128354e-01 -3.13445151e-01 -2.89148241e-01 -7.18717039e-01
2.98932344e-01 -7.84261644e-01 5.01135528e-01 -1.46805644e+00
1.21858865e-01 -9.06444550e-01 -2.80938268e-01 5.18353522e-01
-1.09044291e-01 6.39017165e-01 -1.18305489e-01 3.69374633e-01
-2.14909062e-01 5.05979180e-01 1.14827621e+00 -3.72383028e-01
-2.53994703e-01 -2.05723971e-01 -2.54553080e-01 1.06564188e+00
8.94484699e-01 -7.45530844e-01 -6.08938038e-01 -4.99276489e-01
1.97703764e-01 2.65044719e-02 3.84201318e-01 -9.12849903e-01
-5.96284168e-03 -1.85537785e-01 2.02157080e-01 -6.55675352e-01
3.41155887e-01 -8.82658064e-01 -3.10293324e-02 4.33916748e-01
-3.79020721e-01 -2.44323671e-01 -2.97203630e-01 6.65066600e-01
-5.34987628e-01 -8.32839727e-01 8.09162557e-01 -6.54970467e-01
-8.76958966e-01 2.04027854e-02 -3.49129260e-01 3.28897201e-02
8.50189567e-01 -3.87747586e-01 1.83368415e-01 -2.54468862e-02
-7.57715046e-01 -1.37750506e-01 5.56556761e-01 -5.05291000e-02
6.26531065e-01 -9.42041576e-01 -4.37861353e-01 -5.24345897e-02
-5.43385483e-02 5.52033603e-01 2.06261516e-01 6.73911750e-01
-7.70712972e-01 -1.44773334e-01 -3.51591796e-01 -6.73321426e-01
-9.22196209e-01 4.62767810e-01 4.19601411e-01 -3.03941786e-01
-9.27200258e-01 7.71779060e-01 2.01982349e-01 -2.81918347e-01
-2.52335519e-02 -2.69241780e-01 -3.63021374e-01 3.30879003e-01
1.97566539e-01 1.16750263e-01 1.91997349e-01 -6.84216440e-01
-1.28396079e-01 4.83502120e-01 1.61314502e-01 -4.69372571e-01
1.51801383e+00 -5.81156574e-02 4.79359142e-02 7.82780826e-01
9.56803679e-01 -1.73834547e-01 -1.46297455e+00 -3.57832134e-01
5.80649912e-01 -4.48361337e-01 1.81367889e-01 -3.60060871e-01
-1.45507729e+00 1.05148458e+00 6.45138919e-01 1.11294188e-01
1.03980041e+00 -7.97553211e-02 6.25034571e-01 4.18781489e-01
5.23998439e-01 -1.33676541e+00 2.74753422e-01 7.92524964e-03
5.25718927e-01 -1.69751573e+00 5.38019463e-02 -6.02209032e-01
-3.65413576e-01 1.20877814e+00 4.40018773e-01 -3.51588577e-01
6.89461708e-01 1.82899117e-01 1.90210238e-01 -2.72595435e-01
-8.11237171e-02 -3.67556274e-01 -3.15398700e-03 4.51420635e-01
5.16636908e-01 -1.80453196e-01 -5.83968580e-01 2.73769021e-01
1.17383748e-01 -1.80955768e-01 5.79262972e-01 1.60944164e+00
-5.35120130e-01 -1.35890734e+00 -4.28311467e-01 3.70825619e-01
-4.87789869e-01 -4.49602082e-02 -2.90244043e-01 8.69373322e-01
1.07149005e-01 7.21319139e-01 1.14720725e-01 -9.52781588e-02
-1.20809607e-01 6.99555203e-02 5.94315648e-01 -8.59228611e-01
-4.59400594e-01 4.31317240e-01 -6.93130791e-02 9.14705731e-03
-1.02107549e+00 -5.78403533e-01 -1.65645111e+00 2.68388569e-01
-4.06446189e-01 2.69443333e-01 5.24947703e-01 1.32757020e+00
-4.51582789e-01 4.79795277e-01 4.76724088e-01 -1.02300334e+00
-3.45393181e-01 -8.63434076e-01 -8.54601443e-01 3.97688061e-01
2.07368910e-01 -9.78523672e-01 -3.47166598e-01 3.10501546e-01]
|
[7.445321559906006, 2.0022830963134766]
|
c394931c-17df-489c-9ab3-9e086b80c6d4
|
investigating-eeg-based-functional
|
2004.01973
| null |
https://arxiv.org/abs/2004.01973v1
|
https://arxiv.org/pdf/2004.01973v1.pdf
|
Investigating EEG-Based Functional Connectivity Patterns for Multimodal Emotion Recognition
|
Compared with the rich studies on the motor brain-computer interface (BCI), the recently emerging affective BCI presents distinct challenges since the brain functional connectivity networks involving emotion are not well investigated. Previous studies on emotion recognition based on electroencephalography (EEG) signals mainly rely on single-channel-based feature extraction methods. In this paper, we propose a novel emotion-relevant critical subnetwork selection algorithm and investigate three EEG functional connectivity network features: strength, clustering coefficient, and eigenvector centrality. The discrimination ability of the EEG connectivity features in emotion recognition is evaluated on three public emotion EEG datasets: SEED, SEED-V, and DEAP. The strength feature achieves the best classification performance and outperforms the state-of-the-art differential entropy feature based on single-channel analysis. The experimental results reveal that distinct functional connectivity patterns are exhibited for the five emotions of disgust, fear, sadness, happiness, and neutrality. Furthermore, we construct a multimodal emotion recognition model by combining the functional connectivity features from EEG and the features from eye movements or physiological signals using deep canonical correlation analysis. The classification accuracies of multimodal emotion recognition are 95.08/6.42% on the SEED dataset, 84.51/5.11% on the SEED-V dataset, and 85.34/2.90% and 86.61/3.76% for arousal and valence on the DEAP dataset, respectively. The results demonstrate the complementary representation properties of the EEG connectivity features with eye movement data. In addition, we find that the brain networks constructed with 18 channels achieve comparable performance with that of the 62-channel network in multimodal emotion recognition and enable easier setups for BCI systems in real scenarios.
|
['Bao-liang Lu', 'Wei-Long Zheng', 'Xun Wu']
|
2020-04-04
| null | null | null | null |
['multimodal-emotion-recognition', 'multimodal-emotion-recognition']
|
['computer-vision', 'speech']
|
[-8.73723328e-02 -3.93848151e-01 2.20655531e-01 -3.65358084e-01
8.12415555e-02 -2.20222205e-01 2.21287176e-01 -1.55433506e-01
-6.06685460e-01 9.51061428e-01 -4.30499054e-02 1.58196360e-01
-5.55236876e-01 -4.06447828e-01 -9.05976743e-02 -8.63864899e-01
-5.40604949e-01 -3.32519144e-01 -5.78036368e-01 -2.39962995e-01
3.78972620e-01 3.04580808e-01 -1.51061022e+00 1.73473358e-01
1.12262416e+00 1.59858859e+00 -2.22506300e-01 2.52575934e-01
1.62758008e-01 2.80309673e-02 -7.71253288e-01 -3.56945664e-01
-1.64356664e-01 -5.18475294e-01 -5.60446024e-01 -3.66793007e-01
-5.46739697e-01 3.49167585e-01 -3.03213507e-01 1.11801064e+00
8.56999338e-01 1.97190288e-02 7.44007826e-01 -1.81405067e+00
-3.40502858e-01 3.53565753e-01 -7.09320128e-01 4.34634358e-01
6.57728791e-01 1.66731291e-02 7.93617070e-01 -8.10221672e-01
3.87140095e-01 6.34277463e-01 3.27220351e-01 4.63462651e-01
-9.93431985e-01 -1.35704637e+00 -6.62511727e-03 6.74518347e-01
-1.71042264e+00 -2.23423451e-01 8.08826983e-01 -4.07717139e-01
1.17870378e+00 3.34525764e-01 1.34942436e+00 1.34737217e+00
7.72768080e-01 3.54072869e-01 1.55464816e+00 1.30901068e-01
1.25027061e-01 1.88325703e-01 3.44247997e-01 3.41899514e-01
2.43641406e-01 -8.99928249e-03 -8.96991014e-01 -2.20524162e-01
3.20656478e-01 -8.03672522e-02 -8.91239703e-01 2.43824542e-01
-1.40395641e+00 4.84560847e-01 3.32798898e-01 4.28313106e-01
-6.18378401e-01 -1.84524357e-01 5.17947733e-01 4.97538984e-01
4.19527620e-01 5.88980854e-01 -6.19696200e-01 -6.47438526e-01
-7.01328397e-01 -4.23817515e-01 7.96352148e-01 7.22907960e-01
6.20082676e-01 -5.49259735e-03 6.29400089e-02 8.05344880e-01
2.18278393e-01 4.13850695e-01 7.11285710e-01 -3.27170253e-01
7.10651744e-03 6.44396186e-01 -4.58317816e-01 -1.20670497e+00
-8.94798636e-01 -3.45225453e-01 -1.30173528e+00 -4.17994224e-02
-1.65945321e-01 -5.39704680e-01 -1.85004830e-01 1.81559896e+00
-1.45026222e-01 3.75611894e-02 1.37091994e-01 9.26865339e-01
1.05925167e+00 2.31902137e-01 1.16408102e-01 -4.24015075e-01
1.37223101e+00 -2.46300846e-01 -9.55109775e-01 1.47235155e-01
2.86912411e-01 -2.75946856e-01 7.76476502e-01 6.98334694e-01
-7.32364893e-01 -2.69641668e-01 -1.02579200e+00 7.83361018e-01
-5.06087780e-01 2.53582031e-01 8.33987951e-01 7.73773909e-01
-1.06058526e+00 3.61999810e-01 -5.23720682e-01 -3.48342627e-01
5.09622514e-01 7.42352307e-01 -8.88862789e-01 3.32481146e-01
-1.39478910e+00 1.05027580e+00 2.89021909e-01 2.59416461e-01
-2.85974562e-01 -5.68256021e-01 -4.82062668e-01 2.02803373e-01
-2.65261710e-01 -3.57423812e-01 2.13209733e-01 -1.24230540e+00
-1.89583147e+00 5.41956723e-01 -1.08000822e-01 2.41397008e-01
-2.08728790e-01 9.21454951e-02 -8.84870291e-01 4.05899137e-01
-2.43972629e-01 5.98573267e-01 4.38053578e-01 -7.40363598e-01
-5.69830835e-02 -5.78067541e-01 -4.52448338e-01 7.36546591e-02
-7.75898218e-01 1.95025712e-01 1.16463862e-01 -2.74496615e-01
1.02762178e-01 -7.11978674e-01 2.91861653e-01 -4.54360962e-01
-5.66838503e-01 -1.50041729e-01 4.27543461e-01 -5.24280787e-01
1.32749760e+00 -2.15209937e+00 3.45208079e-01 8.48407388e-01
4.07591403e-01 -2.19680071e-01 -1.80995524e-01 2.57972836e-01
-6.65718734e-01 1.92190930e-01 -7.56374188e-03 4.22103167e-01
-4.26031053e-02 -3.78786117e-01 2.95599550e-01 6.76118076e-01
4.56139147e-01 1.03987205e+00 -5.90022624e-01 -1.68592632e-01
5.19347973e-02 6.15746617e-01 -4.50265139e-01 9.95646045e-02
9.50499415e-01 6.11619711e-01 -1.94617033e-01 7.19100416e-01
5.51426172e-01 6.27451297e-03 2.28707731e-01 -4.49136019e-01
-3.20976488e-02 -1.05365757e-02 -8.49122345e-01 1.62911105e+00
-1.95405066e-01 9.55721200e-01 3.01929843e-02 -1.09472108e+00
1.20935869e+00 5.98061264e-01 8.05563092e-01 -8.61210644e-01
6.65619850e-01 -1.51948677e-02 7.52322912e-01 -6.93757772e-01
-6.27420768e-02 2.56101638e-02 -2.19850361e-01 4.47161645e-01
4.79929358e-01 2.03801349e-01 -2.91721262e-02 1.39476001e-01
1.10785186e+00 -3.43763620e-01 3.60241801e-01 -6.06125236e-01
6.13855839e-01 -6.48438036e-01 4.54464316e-01 1.06287777e-01
-3.99750829e-01 1.16259694e-01 7.92843819e-01 2.27381244e-01
-1.88978553e-01 -7.34808564e-01 -5.06524920e-01 5.55866003e-01
2.92017162e-01 -6.89681113e-01 -7.23899662e-01 -2.36327842e-01
-2.18321860e-01 3.27046156e-01 -6.41575754e-01 -7.84506977e-01
3.42899054e-01 -1.12102807e+00 5.34897685e-01 3.84515733e-01
6.83006704e-01 -1.00103581e+00 -6.42514050e-01 1.32651329e-02
-2.88845092e-01 -8.59752417e-01 9.93772671e-02 3.12013179e-01
-6.15882099e-01 -1.20450723e+00 -3.60844105e-01 -4.74467516e-01
5.65206289e-01 -1.69097364e-01 7.67124116e-01 1.97842065e-02
-4.61614698e-01 5.63932478e-01 -4.33023632e-01 -3.27500165e-01
4.62182522e-01 -9.95681658e-02 4.47167158e-01 2.33023793e-01
7.48178899e-01 -9.80848074e-01 -7.51510441e-01 4.25584882e-01
-5.18596888e-01 -1.27300888e-01 5.99286437e-01 8.98412406e-01
3.07658315e-01 1.22419536e-01 1.00430834e+00 9.77961645e-02
1.13426137e+00 -8.12291324e-01 9.84509513e-02 1.60729200e-01
-6.14394426e-01 -3.61709028e-01 3.12099367e-01 -6.55625165e-01
-9.10277784e-01 -2.62418032e-01 -2.25813165e-02 -3.64821553e-02
-3.57097238e-01 7.38308728e-01 -3.65678400e-01 -3.64596754e-01
4.30287302e-01 1.78964481e-01 -7.70232528e-02 5.48290648e-02
-1.01968817e-01 1.05741763e+00 2.35103428e-01 -3.54179919e-01
4.97953407e-02 1.00013047e-01 -2.32861102e-01 -9.84900355e-01
1.16813660e-01 -2.93252856e-01 -5.43877840e-01 -5.60210645e-01
8.71956468e-01 -8.23277295e-01 -1.45710862e+00 5.52570164e-01
-9.79294538e-01 1.03518762e-01 3.46734524e-01 1.05501330e+00
-3.47293764e-01 2.68659234e-01 -4.66892511e-01 -7.98046350e-01
-7.37948596e-01 -1.25401425e+00 5.16980469e-01 2.33849853e-01
-6.30401909e-01 -5.16356826e-01 -9.34760422e-02 -1.98496729e-01
6.09557152e-01 3.44463438e-01 9.41483438e-01 -6.42021656e-01
2.19803080e-01 -2.13441208e-01 -2.48407945e-01 2.46131539e-01
1.63973585e-01 6.47465885e-02 -8.20366442e-01 3.23160999e-02
1.37446607e-02 -3.35035354e-01 4.90196347e-01 2.70016909e-01
1.33592415e+00 2.54777104e-01 -2.96776742e-01 5.91624975e-01
1.11514735e+00 6.34624243e-01 8.70721936e-01 -2.39666477e-02
2.27983832e-01 5.02995670e-01 6.42326176e-02 6.68541908e-01
2.23457739e-01 2.00800031e-01 3.79750282e-01 9.94387120e-02
6.35014772e-01 4.85315889e-01 4.57050413e-01 1.02057362e+00
-4.51695383e-01 -5.60859032e-02 -8.25689971e-01 2.09531233e-01
-1.55667257e+00 -8.73350263e-01 -2.66882330e-01 1.94888878e+00
7.04888880e-01 -2.19074443e-01 -8.64855200e-02 3.10443133e-01
5.47554612e-01 -4.10700947e-01 -5.61954081e-01 -4.65731859e-01
-3.48598391e-01 6.84794486e-01 -3.82896699e-02 -2.46296868e-01
-8.58296394e-01 5.10533512e-01 5.68516874e+00 4.22385335e-01
-1.24762678e+00 7.88951665e-02 5.92862546e-01 -5.65227389e-01
4.99899015e-02 -5.12044728e-01 -1.94341078e-01 6.46585941e-01
1.22000170e+00 -1.98556140e-01 7.91467190e-01 4.10095602e-01
1.47250190e-01 -4.52861577e-01 -1.04762876e+00 1.76298988e+00
2.08578959e-01 -7.22687006e-01 -5.04771471e-01 9.54739749e-02
3.33856553e-01 1.15088508e-01 -7.49172196e-02 1.53520241e-01
-4.29928929e-01 -1.27262437e+00 2.96173394e-01 1.11002743e+00
1.13101470e+00 -1.03007901e+00 8.26390803e-01 -7.68576935e-03
-1.01424479e+00 -1.59472391e-01 -2.05185875e-01 -3.09193075e-01
-1.35497630e-01 5.89298010e-01 1.91194355e-01 5.52382886e-01
1.09527981e+00 1.06186962e+00 -3.74617904e-01 1.14781260e+00
-3.09402376e-01 5.33488095e-01 -2.91433126e-01 -5.09873569e-01
-2.47669220e-01 -5.14135122e-01 3.54864746e-01 1.17172372e+00
5.28085291e-01 4.20246750e-01 -6.14622533e-01 1.09178770e+00
-6.51490837e-02 3.04390252e-01 -6.03661656e-01 -2.47329861e-01
2.79701591e-01 1.77031589e+00 -6.60658479e-01 -5.85823506e-02
-4.29186344e-01 1.10671389e+00 1.86999053e-01 5.94355643e-01
-9.74533081e-01 -1.00928831e+00 1.12679446e+00 -8.33603978e-01
-2.23698854e-01 -1.01650722e-01 -4.05880332e-01 -1.47286510e+00
-8.46265405e-02 -7.54044712e-01 -1.24804443e-02 -9.08958197e-01
-1.44117653e+00 9.77982819e-01 -5.96596375e-02 -9.82558668e-01
2.50360698e-01 -8.47303331e-01 -1.02146578e+00 1.00778878e+00
-1.18667150e+00 -2.43115142e-01 -6.63732111e-01 1.01746440e+00
-3.62262100e-01 -1.73323199e-01 1.10715532e+00 3.78729343e-01
-1.03506064e+00 5.66717803e-01 -1.44028515e-02 2.11828817e-02
5.61909318e-01 -7.28143215e-01 -5.84617078e-01 3.29898119e-01
-1.80374727e-01 6.96108162e-01 -6.16463497e-02 -2.67770290e-01
-1.53544009e+00 -4.80001569e-01 4.66292679e-01 -3.17671127e-03
6.87880337e-01 -5.19964993e-01 -6.37940586e-01 8.86676833e-02
6.56815171e-01 -2.68508077e-01 1.29440391e+00 1.39856219e-01
1.23888150e-01 -9.69202593e-02 -1.14347732e+00 4.60195929e-01
1.03906655e+00 -6.01320326e-01 -3.66060197e-01 6.44780546e-02
-5.32531627e-02 2.58298665e-01 -1.44038200e+00 4.29495513e-01
8.43561947e-01 -8.87756705e-01 5.81584573e-01 -5.96236408e-01
3.64200979e-01 1.40974283e-01 -7.08205476e-02 -1.74018812e+00
-4.35076833e-01 -4.03569520e-01 2.77496725e-01 1.05760145e+00
5.16901433e-01 -1.07498169e+00 1.67832434e-01 7.98847020e-01
5.65640815e-02 -1.09940350e+00 -1.00015044e+00 -4.29774761e-01
-1.27757668e-01 -5.68907440e-01 7.17784941e-01 1.03462946e+00
1.15758586e+00 3.93741608e-01 -3.40976119e-02 -3.50210816e-01
8.14038068e-02 -2.46008821e-02 2.74079680e-01 -1.56700599e+00
2.75125980e-01 -7.24587917e-01 -7.86774278e-01 -2.11839035e-01
5.70173383e-01 -1.22333956e+00 -3.10392022e-01 -1.26086569e+00
3.48431885e-01 4.49740626e-02 -8.97752285e-01 6.25833213e-01
2.10941546e-02 2.62317806e-01 -1.74166724e-01 -2.18205690e-01
-2.99722433e-01 9.62346017e-01 9.08780456e-01 -8.71702135e-02
-3.19387555e-01 -4.86937851e-01 -8.66584778e-01 5.55528522e-01
1.26405323e+00 -1.54150873e-01 -3.13486964e-01 1.99327379e-01
3.07379127e-01 -2.92400960e-02 2.89865017e-01 -1.00566280e+00
1.85700789e-01 2.59544402e-01 8.31360638e-01 -1.27341956e-01
4.01109874e-01 -8.01578701e-01 1.99830160e-01 2.98828155e-01
-1.75724062e-03 1.91867381e-01 3.51491094e-01 4.36275214e-01
-2.85192609e-01 3.16877484e-01 4.52208102e-01 2.71960825e-01
-6.96653247e-01 3.79129797e-01 -8.51660252e-01 -1.22968078e-01
1.14118838e+00 -4.61849689e-01 -3.22307765e-01 -4.03823316e-01
-8.11679304e-01 9.29267928e-02 -1.14969432e-01 4.26972955e-01
1.00553811e+00 -1.47949553e+00 -4.14422333e-01 5.86014450e-01
1.09075382e-01 -8.92860174e-01 3.15950572e-01 1.81928658e+00
2.27950603e-01 2.82053590e-01 -9.49959874e-01 -4.06520396e-01
-1.23089302e+00 4.33200188e-02 3.99181247e-01 2.72662997e-01
-2.82645702e-01 8.17773402e-01 2.64009628e-02 2.41638925e-02
6.08171262e-02 -2.04908416e-01 -5.72260380e-01 4.61723536e-01
2.77147949e-01 5.43113291e-01 2.11676881e-01 -7.94424355e-01
-8.25627506e-01 4.30549800e-01 4.34128284e-01 -1.40115336e-01
1.48889494e+00 4.16612849e-02 -6.87449515e-01 2.79515505e-01
1.34478736e+00 -4.10146922e-01 -5.01798391e-01 2.96263009e-01
-2.23974183e-01 -1.76369846e-01 1.94542602e-01 -1.18150878e+00
-1.46793509e+00 1.10307312e+00 8.08665514e-01 4.16952930e-03
1.54615569e+00 -2.48644054e-01 4.57808912e-01 4.31631655e-01
5.00013292e-01 -1.22308648e+00 -7.79283326e-03 3.77167463e-01
1.12615204e+00 -9.59546804e-01 -1.68532908e-01 3.51087451e-02
-8.97306859e-01 1.30009055e+00 7.82470882e-01 1.02537468e-01
1.13109767e+00 1.13527797e-01 -3.00596625e-01 -7.42484450e-01
-7.10696876e-01 8.26009586e-02 5.37570179e-01 4.39042151e-01
5.69295883e-01 2.00917304e-01 -7.06089616e-01 1.42141640e+00
-4.40952480e-01 -1.57594189e-01 2.60200411e-01 5.04745841e-01
-8.93512517e-02 -5.78855395e-01 -7.61859789e-02 1.00741661e+00
-3.99722755e-01 -3.34426820e-01 -7.28092432e-01 7.67957926e-01
-1.38843562e-02 1.28417981e+00 1.61117002e-01 -1.03351653e+00
2.35456824e-01 5.23369312e-01 4.30763602e-01 -1.33199826e-01
-8.35181832e-01 -9.20358226e-02 2.71753483e-02 -9.57801282e-01
-6.33653581e-01 -5.07770777e-01 -1.38794756e+00 -1.69090793e-01
-6.86257005e-01 3.10851842e-01 8.65193725e-01 7.84094274e-01
8.25721562e-01 5.30368745e-01 5.47683418e-01 -9.42373872e-01
3.04285944e-01 -1.21420765e+00 -1.05785668e+00 2.93923020e-01
-2.09415257e-01 -1.01618111e+00 -4.68366355e-01 -5.04659712e-01]
|
[13.143817901611328, 3.450885772705078]
|
dd2c7ea3-99d2-4f28-8d09-3244639cc516
|
infwide-image-and-feature-space-wiener
|
2207.08201
| null |
https://arxiv.org/abs/2207.08201v2
|
https://arxiv.org/pdf/2207.08201v2.pdf
|
INFWIDE: Image and Feature Space Wiener Deconvolution Network for Non-blind Image Deblurring in Low-Light Conditions
|
Under low-light environment, handheld photography suffers from severe camera shake under long exposure settings. Although existing deblurring algorithms have shown promising performance on well-exposed blurry images, they still cannot cope with low-light snapshots. Sophisticated noise and saturation regions are two dominating challenges in practical low-light deblurring. In this work, we propose a novel non-blind deblurring method dubbed image and feature space Wiener deconvolution network (INFWIDE) to tackle these problems systematically. In terms of algorithm design, INFWIDE proposes a two-branch architecture, which explicitly removes noise and hallucinates saturated regions in the image space and suppresses ringing artifacts in the feature space, and integrates the two complementary outputs with a subtle multi-scale fusion network for high quality night photograph deblurring. For effective network training, we design a set of loss functions integrating a forward imaging model and backward reconstruction to form a close-loop regularization to secure good convergence of the deep neural network. Further, to optimize INFWIDE's applicability in real low-light conditions, a physical-process-based low-light noise model is employed to synthesize realistic noisy night photographs for model training. Taking advantage of the traditional Wiener deconvolution algorithm's physically driven characteristics and arisen deep neural network's representation ability, INFWIDE can recover fine details while suppressing the unpleasant artifacts during deblurring. Extensive experiments on synthetic data and real data demonstrate the superior performance of the proposed approach.
|
['Qionghai Dai', 'Liheng Bian', 'Jinli Suo', 'Yuxiao Cheng', 'Zhihong Zhang']
|
2022-07-17
| null | null | null | null |
['blind-image-deblurring']
|
['computer-vision']
|
[ 2.41451234e-01 -5.80148101e-01 3.72074336e-01 -1.35752320e-01
-3.62397134e-01 -2.65354812e-01 4.36551243e-01 -1.22075152e+00
-1.10398516e-01 9.15073872e-01 4.88902748e-01 -1.06933847e-01
-3.25686544e-01 -2.75718987e-01 -4.99245793e-01 -1.33374882e+00
4.65843499e-01 -5.38818359e-01 -1.82426900e-01 -1.13311402e-01
5.79410605e-02 4.43459868e-01 -1.48321283e+00 -2.34838381e-01
1.27788150e+00 8.54660451e-01 5.60921133e-01 7.48070300e-01
2.52100080e-01 8.00180435e-01 -4.97489721e-01 -1.70331821e-02
5.52538335e-01 -5.92072546e-01 -2.13244244e-01 2.92878538e-01
5.80695808e-01 -9.24106121e-01 -9.29823637e-01 1.41636455e+00
8.35366726e-01 3.24669242e-01 4.54934478e-01 -4.84297603e-01
-1.21618271e+00 -1.72906846e-01 -7.45834351e-01 4.87249583e-01
-6.83248788e-02 7.04668760e-01 3.59025270e-01 -1.00736654e+00
1.74304530e-01 1.02863562e+00 9.38855231e-01 5.77958167e-01
-1.07157063e+00 -5.77103019e-01 -2.84866184e-01 2.38193661e-01
-1.35592031e+00 -8.30327988e-01 8.99301410e-01 -2.29723573e-01
5.31985283e-01 2.84462124e-01 4.44449186e-01 9.34055448e-01
2.87071317e-01 3.42909098e-01 1.52200782e+00 -2.45634377e-01
-9.91248563e-02 -1.88687652e-01 1.53226599e-01 4.04521793e-01
4.86121535e-01 5.49777746e-01 -2.39676371e-01 -1.97073147e-02
1.12094355e+00 2.09089652e-01 -1.20756328e+00 1.19720638e-01
-9.74660695e-01 2.35012606e-01 5.64661086e-01 2.46649742e-01
-5.29774010e-01 1.26038268e-01 -1.36173233e-01 2.23018721e-01
6.55875146e-01 3.98157179e-01 -2.75438905e-01 1.37550846e-01
-9.87521350e-01 1.81938931e-02 4.48756158e-01 6.36459649e-01
5.99825084e-01 4.06542659e-01 -3.51485521e-01 1.22987425e+00
3.53245497e-01 8.61627102e-01 6.18149519e-01 -1.05665565e+00
2.10144848e-01 -1.20423205e-01 5.16594112e-01 -9.37266886e-01
-1.64720237e-01 -7.84758687e-01 -1.47217691e+00 3.41987103e-01
1.40741095e-01 -2.53359288e-01 -8.89530957e-01 1.50852919e+00
1.30712509e-01 5.00330865e-01 2.48108909e-01 1.75361192e+00
8.09096813e-01 9.17455196e-01 -4.85892147e-01 -7.40441144e-01
1.14735925e+00 -9.29199934e-01 -1.16708398e+00 -3.05322289e-01
-2.11221442e-01 -8.59947383e-01 9.30211842e-01 5.16157150e-01
-1.25808215e+00 -6.89703107e-01 -1.14700222e+00 -2.33503848e-01
2.42705062e-01 2.78439999e-01 3.24305445e-01 4.76598531e-01
-1.12095082e+00 6.20782852e-01 -4.67216492e-01 -6.81017116e-02
4.00983632e-01 -1.08638685e-03 -1.29254274e-02 -4.27601755e-01
-1.24447274e+00 9.83293951e-01 -1.42484128e-01 8.88029516e-01
-9.59063351e-01 -4.52723145e-01 -4.98114735e-01 3.38679552e-02
2.13356793e-01 -1.10895216e+00 1.11031890e+00 -1.02113461e+00
-1.81357861e+00 4.95065480e-01 -2.67639875e-01 -8.09293687e-02
5.05077899e-01 -4.41602468e-01 -6.74642384e-01 2.55364239e-01
-3.12180668e-01 3.64500750e-03 1.59410822e+00 -1.49412537e+00
-3.51924961e-03 -2.63935387e-01 -2.13759601e-01 5.31068325e-01
-2.96337694e-01 -5.11174835e-02 -3.74156505e-01 -9.18477595e-01
7.95878395e-02 -5.26500821e-01 -1.16724841e-01 1.28962234e-01
-1.73031911e-01 5.10417581e-01 8.41199875e-01 -9.95692194e-01
1.16274881e+00 -2.11469722e+00 8.24569725e-03 -5.34366846e-01
3.92585665e-01 7.57142901e-01 -1.94126755e-01 9.86586511e-02
-2.19603285e-01 -4.60150778e-01 -4.27342504e-01 -2.80873686e-01
-4.38871592e-01 1.80511758e-01 -4.07654673e-01 8.63712907e-01
-4.78537381e-02 8.24340165e-01 -7.17968106e-01 5.93157895e-02
3.42115641e-01 7.29544699e-01 -1.54944971e-01 5.33754110e-01
1.62359759e-01 5.42347848e-01 -2.00415418e-01 5.07172048e-01
1.24852669e+00 -2.83530623e-01 -4.26483095e-01 -5.34323514e-01
-2.97838807e-01 -3.51195991e-01 -9.36716199e-01 1.62043798e+00
-6.40452921e-01 8.28763783e-01 7.83203423e-01 -5.05416036e-01
8.94200146e-01 1.47739440e-01 -4.45422418e-02 -5.33448994e-01
2.10327730e-01 2.22911984e-01 -2.51110703e-01 -1.14500546e+00
5.85857809e-01 -5.82262933e-01 6.74583256e-01 1.77722290e-01
-2.51210809e-01 -2.39516556e-01 -6.27781928e-01 -2.02185825e-01
8.44130695e-01 1.25890210e-01 1.01310678e-01 -2.25570083e-01
8.53200138e-01 -4.50876147e-01 5.55766702e-01 7.33280182e-01
-2.98959345e-01 1.16060269e+00 -3.21885198e-01 -3.58406574e-01
-9.39530015e-01 -7.99649060e-01 -2.10839719e-01 4.94100809e-01
6.97533786e-01 2.89237440e-01 -7.52665699e-01 -1.78123876e-01
-4.07341659e-01 5.43921232e-01 -2.34634653e-01 -3.53075385e-01
-5.77992976e-01 -1.17962456e+00 3.83702189e-01 -6.85037598e-02
1.08329594e+00 -7.03140259e-01 -1.81282163e-01 -6.82286024e-02
-3.65604013e-01 -8.96720707e-01 -7.29819834e-01 -2.79839128e-01
-6.75293148e-01 -1.07066202e+00 -1.30405545e+00 -6.55115426e-01
6.37984455e-01 1.17036664e+00 6.96720302e-01 -4.91330549e-02
-3.23887289e-01 1.37391970e-01 3.24794054e-02 8.02344754e-02
-3.31463248e-01 -5.73405325e-01 1.82142198e-01 3.53894472e-01
-4.44303080e-02 -7.85932064e-01 -1.10284746e+00 4.15371805e-01
-1.08896339e+00 1.34097144e-01 8.46582115e-01 1.28472435e+00
4.63071205e-02 4.13574815e-01 3.65315527e-01 -8.55958685e-02
1.00667727e+00 -2.05224872e-01 -6.63745880e-01 1.28310636e-01
-7.10281014e-01 -1.95233077e-01 7.51200080e-01 -6.44415915e-01
-1.71241403e+00 -4.87162232e-01 1.66060895e-01 -8.50283146e-01
-4.83385436e-02 1.87003732e-01 -1.96867719e-01 -4.58987802e-01
7.26055264e-01 7.40097761e-01 2.53922611e-01 -7.23287404e-01
3.31705540e-01 1.00262558e+00 9.00003731e-01 -1.69712290e-01
1.01095295e+00 5.36946654e-01 -4.46899049e-02 -1.11578608e+00
-8.51476669e-01 -3.47972602e-01 -9.68568847e-02 -3.76927882e-01
7.37759054e-01 -1.27826464e+00 -8.26626658e-01 1.26663709e+00
-1.22039282e+00 -2.88442552e-01 -4.80921194e-02 5.32446504e-01
-2.76196271e-01 8.27757657e-01 -7.80238211e-01 -9.45098102e-01
-5.30647933e-01 -1.03260899e+00 7.16916323e-01 7.15650678e-01
4.92403775e-01 -7.91858852e-01 1.41997617e-02 4.58392143e-01
8.32943261e-01 -2.63755143e-01 3.02020729e-01 4.56189811e-01
-7.41796792e-01 -6.01042667e-03 -7.37226009e-01 9.60318744e-01
3.06824625e-01 -3.96161973e-01 -1.38113499e+00 -4.12668467e-01
9.22810018e-01 -8.14464986e-02 1.02321792e+00 8.35359514e-01
1.09761083e+00 -4.66638714e-01 1.08876795e-01 1.27413845e+00
1.55379450e+00 1.75218433e-02 8.75216663e-01 2.20919684e-01
8.03041577e-01 3.51996183e-01 1.97030663e-01 3.12214702e-01
5.95507883e-02 5.16966343e-01 4.72043693e-01 -3.87320429e-01
-5.88341773e-01 1.75646186e-01 4.76021111e-01 8.81666899e-01
-1.58011466e-01 -3.39518189e-01 -3.17446858e-01 4.24615860e-01
-1.60662520e+00 -1.00995672e+00 -4.30273145e-01 2.10678244e+00
9.50117707e-01 -3.62273663e-01 -5.07766545e-01 -3.17302048e-01
9.16549087e-01 5.72304189e-01 -7.43532360e-01 1.46560103e-01
-5.95183372e-01 -1.96820885e-01 7.10052669e-01 5.84269047e-01
-9.04089272e-01 5.29323637e-01 5.91528177e+00 8.88601780e-01
-1.18923724e+00 2.45890319e-01 5.48615396e-01 -1.11511052e-01
-1.49632886e-01 -5.29152602e-02 -4.64153677e-01 7.22593009e-01
4.88468230e-01 1.58283770e-01 1.00399375e+00 2.89243937e-01
8.86970103e-01 -2.09468678e-01 -2.78747499e-01 1.45247746e+00
3.19540828e-01 -1.10930479e+00 -2.41615847e-01 -8.18633065e-02
8.94818366e-01 1.18376426e-02 1.31897628e-01 -3.04359198e-01
1.30194560e-01 -9.95908439e-01 4.09961373e-01 1.19627082e+00
1.03864574e+00 -2.95327157e-01 6.25210106e-01 4.72921580e-01
-4.37834382e-01 -2.41424382e-01 -4.50550973e-01 -2.61084050e-01
3.71492684e-01 9.68448043e-01 -4.50165160e-02 7.64353573e-01
5.72297037e-01 1.09790504e+00 -1.59901008e-01 1.13142192e+00
-4.35822189e-01 5.49027205e-01 3.53452936e-02 4.52155113e-01
1.04856215e-01 -6.76384866e-01 9.65636313e-01 9.38892543e-01
5.62231064e-01 6.14044309e-01 -4.68468428e-01 1.05380082e+00
3.97330560e-02 -6.91891849e-01 -3.85560393e-01 3.59809905e-01
2.66215444e-01 1.26827383e+00 9.65277106e-02 -2.29780674e-01
-3.10750395e-01 1.30972826e+00 -3.51107597e-01 9.01904285e-01
-7.60287225e-01 -4.97480184e-01 1.00875056e+00 2.48640347e-02
1.72965780e-01 -2.65274405e-01 -3.44432175e-01 -1.66804719e+00
1.01207204e-01 -7.88288593e-01 -2.41861850e-01 -1.61790514e+00
-1.57254827e+00 5.86353481e-01 -3.16000134e-01 -1.47643077e+00
4.85554785e-01 -5.00998735e-01 -8.29783082e-01 1.46454644e+00
-2.04335141e+00 -9.70935106e-01 -8.54241848e-01 6.10869586e-01
6.22671127e-01 1.65534429e-02 3.03862333e-01 2.78329760e-01
-9.55583453e-01 6.41887933e-02 6.40670419e-01 -3.97036999e-01
9.80296791e-01 -8.32701564e-01 1.41117424e-01 1.36380029e+00
-6.80631161e-01 7.14250386e-01 9.08603787e-01 -5.79093754e-01
-1.51062739e+00 -9.91692364e-01 2.47741744e-01 2.11099647e-02
5.83311856e-01 1.37598768e-01 -1.25251293e+00 1.12554826e-01
4.61741686e-01 2.47969270e-01 1.77009441e-02 -5.94992638e-01
-5.59233874e-02 -3.61622602e-01 -1.03733599e+00 4.90113497e-01
9.93217587e-01 -4.76236254e-01 -7.26240695e-01 2.63607681e-01
6.18621051e-01 -3.98039669e-01 -5.22639215e-01 3.04697901e-01
4.28282171e-01 -1.15555620e+00 1.17980278e+00 2.86732073e-04
5.51976562e-01 -6.07611835e-01 1.64092898e-01 -1.58186686e+00
-6.01848125e-01 -1.29037476e+00 -3.06163043e-01 1.21817219e+00
-2.99230367e-01 -9.90006685e-01 2.93070614e-01 3.51639926e-01
-4.34701025e-01 -4.04940575e-01 -6.10363483e-01 -7.08316326e-01
-2.77601749e-01 7.42762089e-02 2.67943025e-01 9.89247799e-01
-6.24517262e-01 3.06598663e-01 -1.07346809e+00 5.49138844e-01
1.02180159e+00 2.99564935e-03 6.57138467e-01 -8.60687375e-01
-3.40480179e-01 -3.03925574e-01 2.15620488e-01 -1.60020816e+00
-9.05409977e-02 -2.99280465e-01 3.76902282e-01 -1.50476956e+00
2.03557953e-01 -1.79069892e-01 -4.78139520e-02 -1.45116508e-01
-5.17481625e-01 2.55425662e-01 -9.50183645e-02 8.04309487e-01
-9.12853405e-02 9.00696576e-01 1.70435452e+00 5.36723249e-02
-8.14866275e-02 -4.78171222e-02 -7.36802936e-01 6.43713653e-01
3.63808244e-01 -1.06681205e-01 -4.40402478e-01 -8.89707267e-01
-2.76608001e-02 3.52963358e-01 6.84606671e-01 -9.01175261e-01
3.62060934e-01 -2.17272654e-01 4.70097780e-01 -2.97896951e-01
5.71576595e-01 -8.43976557e-01 3.14146340e-01 1.78620860e-01
-8.51936191e-02 -5.27148068e-01 -7.81540275e-02 8.06640446e-01
-3.07987124e-01 -1.91153541e-01 1.17148900e+00 -1.95453584e-01
-4.43865031e-01 2.25049466e-01 -2.39769861e-01 -1.99567646e-01
6.19999468e-01 -4.04826254e-01 -7.61021495e-01 -5.16023755e-01
-3.43471318e-01 4.97131981e-02 7.37506747e-01 5.69095053e-02
7.14683354e-01 -1.04026699e+00 -6.92044497e-01 4.93774712e-01
-3.52061957e-01 -2.48088911e-02 8.97422016e-01 1.05165064e+00
-6.63383067e-01 -1.73208807e-02 -4.78885286e-02 -2.68322349e-01
-1.04503405e+00 4.53357458e-01 7.35737026e-01 1.61743060e-01
-8.19486439e-01 8.35148454e-01 4.53304678e-01 5.52095473e-02
-8.49771649e-02 -1.32487774e-01 -3.36766355e-02 -2.90562749e-01
9.30220425e-01 6.26854658e-01 -1.03541993e-01 -5.82674503e-01
1.22416042e-01 7.33599365e-01 1.13328435e-01 1.57457560e-01
1.45033705e+00 -8.71125042e-01 -3.35786551e-01 -4.99967625e-03
1.12413502e+00 -2.57081110e-02 -1.88218868e+00 -3.38938236e-01
-7.30799615e-01 -8.84113789e-01 7.09733844e-01 -8.18272233e-01
-1.08438003e+00 7.96227157e-01 6.94996834e-01 1.41043961e-01
1.69096351e+00 -5.18615007e-01 1.06003046e+00 1.77107096e-01
-2.25264117e-01 -6.79200828e-01 7.89726898e-02 4.55659062e-01
9.36358392e-01 -1.18182421e+00 1.62700281e-01 -1.54567048e-01
-3.76855493e-01 1.18643665e+00 4.34051812e-01 -1.26247570e-01
3.74480814e-01 4.18182313e-02 2.62936592e-01 -1.11488491e-01
-4.22017097e-01 5.55168390e-02 2.22853258e-01 5.16502142e-01
-1.51856229e-01 -4.00862038e-01 -1.71887368e-01 5.91365695e-01
3.22347224e-01 2.97559559e-01 8.16689193e-01 3.71438891e-01
-5.19232810e-01 -3.35612118e-01 -8.98763657e-01 1.79791093e-01
-4.65713203e-01 -3.25303376e-01 1.70757800e-01 1.88752115e-01
1.25151008e-01 1.00526965e+00 -2.64748335e-01 -1.13354646e-01
1.62556365e-01 -3.36333573e-01 3.09060186e-01 -2.46100333e-02
-2.94829905e-01 4.39054728e-01 -1.43735334e-01 -3.43315631e-01
-6.39623404e-01 -3.13306212e-01 -4.31926519e-01 -2.97237396e-01
-7.36332953e-01 -1.32318199e-01 4.66967523e-01 8.30476820e-01
3.77917558e-01 4.58946526e-01 8.60003650e-01 -1.20757711e+00
-7.77215123e-01 -1.16973400e+00 -8.26732516e-01 1.31688550e-01
1.08088481e+00 -5.40187359e-01 -9.61404979e-01 1.43528968e-01]
|
[11.335759162902832, -2.7218456268310547]
|
89b356cb-ca1c-4276-856c-f646a3778356
|
transg-transformer-based-skeleton-graph
|
2303.06819
| null |
https://arxiv.org/abs/2303.06819v2
|
https://arxiv.org/pdf/2303.06819v2.pdf
|
TranSG: Transformer-Based Skeleton Graph Prototype Contrastive Learning with Structure-Trajectory Prompted Reconstruction for Person Re-Identification
|
Person re-identification (re-ID) via 3D skeleton data is an emerging topic with prominent advantages. Existing methods usually design skeleton descriptors with raw body joints or perform skeleton sequence representation learning. However, they typically cannot concurrently model different body-component relations, and rarely explore useful semantics from fine-grained representations of body joints. In this paper, we propose a generic Transformer-based Skeleton Graph prototype contrastive learning (TranSG) approach with structure-trajectory prompted reconstruction to fully capture skeletal relations and valuable spatial-temporal semantics from skeleton graphs for person re-ID. Specifically, we first devise the Skeleton Graph Transformer (SGT) to simultaneously learn body and motion relations within skeleton graphs, so as to aggregate key correlative node features into graph representations. Then, we propose the Graph Prototype Contrastive learning (GPC) to mine the most typical graph features (graph prototypes) of each identity, and contrast the inherent similarity between graph representations and different prototypes from both skeleton and sequence levels to learn discriminative graph representations. Last, a graph Structure-Trajectory Prompted Reconstruction (STPR) mechanism is proposed to exploit the spatial and temporal contexts of graph nodes to prompt skeleton graph reconstruction, which facilitates capturing more valuable patterns and graph semantics for person re-ID. Empirical evaluations demonstrate that TranSG significantly outperforms existing state-of-the-art methods. We further show its generality under different graph modeling, RGB-estimated skeletons, and unsupervised scenarios.
|
['Chunyan Miao', 'Haocong Rao']
|
2023-03-13
| null |
http://openaccess.thecvf.com//content/CVPR2023/html/Rao_TranSG_Transformer-Based_Skeleton_Graph_Prototype_Contrastive_Learning_With_Structure-Trajectory_Prompted_CVPR_2023_paper.html
|
http://openaccess.thecvf.com//content/CVPR2023/papers/Rao_TranSG_Transformer-Based_Skeleton_Graph_Prototype_Contrastive_Learning_With_Structure-Trajectory_Prompted_CVPR_2023_paper.pdf
|
cvpr-2023-1
|
['person-re-identification', 'graph-reconstruction']
|
['computer-vision', 'graphs']
|
[ 1.33991437e-02 -8.74101296e-02 -4.83702451e-01 -2.07815796e-01
-1.25535443e-01 -2.60537714e-01 7.41447151e-01 4.49701883e-02
6.64248019e-02 1.80993617e-01 6.28050089e-01 4.89592254e-01
-2.87105501e-01 -8.14100981e-01 -1.54879838e-01 -3.63002032e-01
-2.04045042e-01 6.37538552e-01 2.21081987e-01 -2.92889833e-01
-1.28138244e-01 5.05156577e-01 -1.36001456e+00 -1.21568888e-01
5.30845940e-01 5.46186566e-01 -2.29993761e-01 4.56917703e-01
-8.89624879e-02 5.78161120e-01 -1.59588546e-01 -7.51510322e-01
1.36645153e-01 -6.75405085e-01 -6.52612984e-01 4.99231994e-01
6.38776779e-01 -9.89849791e-02 -9.75170195e-01 1.05969954e+00
6.32710397e-01 5.83882809e-01 8.15608621e-01 -1.38396251e+00
-5.96561551e-01 3.99924159e-01 -9.33390081e-01 1.49082169e-01
1.06623626e+00 2.13926911e-01 1.05040526e+00 -6.42064214e-01
6.60621881e-01 1.65696299e+00 1.02428985e+00 7.00455964e-01
-1.06534255e+00 -5.47644138e-01 5.78159809e-01 4.32715237e-01
-1.49578106e+00 -1.56165093e-01 1.21688998e+00 -4.17502463e-01
8.66448283e-01 9.93060321e-02 1.40151834e+00 1.36241508e+00
-4.93537635e-02 8.00012231e-01 7.41448104e-01 -1.87181160e-01
-2.21196190e-01 -7.83101559e-01 3.51249278e-01 1.25776887e+00
4.21643704e-01 8.67017284e-02 -8.22935402e-01 -9.55260321e-02
1.04852343e+00 4.73858654e-01 -6.13737516e-02 -5.49310088e-01
-1.19922912e+00 4.97911274e-01 5.60873985e-01 1.20051868e-01
-3.10759485e-01 4.62071240e-01 6.02670550e-01 9.53689963e-02
1.12266935e-01 -1.97445914e-01 1.29850075e-01 7.91782793e-03
-7.59864569e-01 4.44186598e-01 5.05113006e-01 1.15871525e+00
8.35651159e-01 8.47830176e-02 -3.38713497e-01 8.63058746e-01
6.31104469e-01 5.71352482e-01 7.12892115e-01 -5.62017262e-01
4.64908957e-01 1.04892778e+00 -5.29098570e-01 -1.37142718e+00
-5.27412832e-01 -3.62890095e-01 -1.00782108e+00 -4.43196595e-01
3.31462801e-01 4.09284443e-01 -8.95053208e-01 1.48341906e+00
4.62807119e-01 6.69370115e-01 -3.82546574e-01 9.69774544e-01
1.11256993e+00 2.87281070e-02 3.32997292e-01 7.56766126e-02
1.64404964e+00 -9.68177497e-01 -3.84347707e-01 -2.15003341e-01
6.55388474e-01 -1.10567965e-01 1.03726196e+00 -2.17905547e-02
-7.33458102e-01 -9.74122763e-01 -7.75346935e-01 2.16600932e-02
-1.32504389e-01 -8.96365047e-02 5.20235598e-01 6.89381480e-01
-7.01030076e-01 7.00426996e-01 -7.86855280e-01 -7.38725543e-01
2.46423647e-01 2.79790968e-01 -6.66869164e-01 -2.47603849e-01
-1.14380085e+00 4.33243364e-01 4.22096372e-01 5.81909791e-02
-6.41530991e-01 -3.10114026e-01 -1.22412908e+00 -2.37935841e-01
3.64360213e-01 -1.18805563e+00 6.13982499e-01 -5.30179083e-01
-1.31205106e+00 1.06618810e+00 -2.12967128e-01 -1.87659249e-01
4.52867597e-01 -2.18662530e-01 -5.67932427e-01 3.60070944e-01
2.25650311e-01 3.40036720e-01 1.16554987e+00 -1.07401121e+00
-3.06354493e-01 -8.17669153e-01 -2.41667971e-01 4.40859795e-01
-4.30380404e-01 -1.73619941e-01 -9.65026021e-01 -1.17758179e+00
3.28928679e-01 -9.86916244e-01 -1.81514949e-01 -9.28465277e-03
-5.77889979e-01 -5.67841470e-01 4.95092213e-01 -8.90788257e-01
1.40083981e+00 -1.90452683e+00 5.20766914e-01 6.27748072e-01
5.19915700e-01 3.20420275e-03 -1.23983994e-01 4.62428004e-01
-2.35396087e-01 -2.04984963e-01 -1.12402417e-01 -5.49705744e-01
1.07730605e-01 4.12857234e-01 1.83386683e-01 6.52262628e-01
-2.34512672e-01 1.37066436e+00 -1.18254900e+00 -8.84805143e-01
3.38262528e-01 3.68906975e-01 -3.11647385e-01 6.73594624e-02
1.28042325e-01 4.70985264e-01 -7.21892357e-01 8.85096908e-01
2.72900522e-01 -3.99897426e-01 5.06047130e-01 -6.14586830e-01
5.08294523e-01 -2.93210179e-01 -1.28883684e+00 1.92854500e+00
1.82186067e-01 -7.08350688e-02 -4.66560036e-01 -1.21194148e+00
1.08944654e+00 -6.57738373e-02 7.48253942e-01 -7.24367440e-01
1.82041600e-02 -2.02814206e-01 -4.91827518e-01 -4.96811688e-01
2.00742006e-01 -1.32781230e-02 -9.04461592e-02 4.57690209e-01
1.84806272e-01 3.94212604e-01 3.33347842e-02 2.86579400e-01
9.90849733e-01 5.96548319e-01 6.01781785e-01 -4.40291986e-02
7.69639194e-01 -2.94254154e-01 5.59477329e-01 3.31626773e-01
-5.27230144e-01 6.40417874e-01 7.86035210e-02 -5.48339725e-01
-7.09174395e-01 -1.34386790e+00 5.77677608e-01 1.19012439e+00
6.93274319e-01 -1.04082775e+00 -4.69316244e-01 -9.49997485e-01
1.89216226e-01 1.21061668e-01 -5.76151788e-01 -3.75862181e-01
-9.78141069e-01 -4.37839448e-01 6.84154689e-01 8.04849923e-01
7.64632702e-01 -7.79304743e-01 -1.04283802e-01 7.97667354e-02
-3.77482325e-01 -1.04135787e+00 -9.55115020e-01 -7.35490859e-01
-1.00573778e+00 -1.40587568e+00 -9.62058723e-01 -9.11540627e-01
6.32716000e-01 5.99969685e-01 1.05524874e+00 3.08963120e-01
-3.70604008e-01 1.21552086e+00 -5.51096797e-01 4.04385984e-01
8.62431675e-02 -1.57545984e-01 4.07123893e-01 2.74064362e-01
3.74968290e-01 -9.12457943e-01 -7.01757789e-01 3.23597491e-01
-3.20293427e-01 -2.29139477e-02 4.09308225e-01 8.10883641e-01
8.66864026e-01 2.16607954e-02 2.20870495e-01 -6.91007853e-01
6.31318271e-01 -2.80748367e-01 1.18376717e-01 5.67369342e-01
-6.53897345e-01 1.01887614e-01 3.93609822e-01 -4.73812670e-01
-8.31479311e-01 1.82973430e-01 3.18430178e-03 -7.95444727e-01
-1.41196311e-01 4.42348123e-01 -3.58069181e-01 -8.05245936e-02
5.89173317e-01 7.31620908e-01 1.86938047e-01 -5.84686935e-01
6.68506384e-01 -2.26186849e-02 9.59190905e-01 -8.76301467e-01
1.12824023e+00 5.59131861e-01 3.60448927e-01 -7.81161606e-01
-5.30034125e-01 -8.72896433e-01 -1.06224084e+00 -5.88687658e-01
8.45454037e-01 -1.02593386e+00 -6.82963252e-01 4.49417531e-01
-7.24239171e-01 -1.73160329e-01 -4.09607798e-01 4.25157726e-01
-6.30629838e-01 1.37844074e+00 -5.71571708e-01 -6.45190001e-01
-4.20688510e-01 -5.95172942e-01 1.35999954e+00 3.14560711e-01
-4.26318914e-01 -1.22794116e+00 3.50547701e-01 3.67149442e-01
-4.07022238e-01 6.82682812e-01 5.20450771e-01 -4.04613167e-01
-1.76510826e-01 -2.71235645e-01 -2.88054734e-01 -2.37188995e-01
3.61236036e-01 -4.39753085e-01 -4.97833878e-01 -5.23421168e-01
-5.89451194e-01 -8.03935230e-02 8.67222607e-01 1.32592246e-01
9.05989945e-01 -3.38573724e-01 -7.69898355e-01 9.14387107e-01
9.70124006e-01 -5.23527801e-01 4.04234856e-01 2.69737780e-01
1.53746414e+00 9.01301146e-01 4.99819726e-01 5.73537827e-01
7.80529261e-01 9.13676023e-01 1.28493473e-01 2.67641634e-01
-7.62684405e-01 -9.22082961e-01 5.78479230e-01 8.98883164e-01
-7.93439448e-01 2.18594790e-01 -7.33933270e-01 3.33594531e-01
-2.09577870e+00 -1.10373199e+00 -3.20027769e-01 1.79552889e+00
3.17527562e-01 -8.45613182e-02 8.31124663e-01 6.66796267e-02
1.09125781e+00 3.37451041e-01 -5.95407486e-01 4.50157166e-01
-1.01480722e-01 2.80815046e-02 2.69892275e-01 1.81412086e-01
-1.01474559e+00 1.00366139e+00 5.31324768e+00 8.84934008e-01
-4.12294626e-01 6.34388477e-02 -1.22587338e-01 4.15780067e-01
-3.85161400e-01 -6.51819259e-02 -7.06554174e-01 3.60994995e-01
4.22041833e-01 -2.72716582e-01 3.00669312e-01 9.47914839e-01
-1.17183492e-01 5.92410743e-01 -1.07516122e+00 1.81964529e+00
3.47056627e-01 -9.71418500e-01 5.30676723e-01 1.43235326e-01
2.82185704e-01 -5.36358237e-01 -2.07987487e-01 2.06904054e-01
3.14797819e-01 -8.89961660e-01 6.64331198e-01 8.65434706e-01
8.90163243e-01 -6.11963511e-01 3.27817231e-01 2.74609998e-02
-2.28408575e+00 -5.97772934e-02 -2.72124827e-01 8.68193358e-02
1.98476374e-01 1.74307302e-01 -5.79699039e-01 1.22241020e+00
6.93699539e-01 1.55226612e+00 -1.00761747e+00 8.02810431e-01
-3.80395800e-01 4.86184627e-01 -1.51665747e-01 1.28184676e-01
-2.05117524e-01 -3.34913313e-01 6.90562010e-01 1.27635193e+00
1.84354335e-01 1.50331676e-01 5.31659901e-01 6.13248825e-01
2.61167645e-01 1.97908625e-01 -5.99084496e-01 7.47663826e-02
3.94603550e-01 1.03377497e+00 -9.10181642e-01 -3.12681705e-01
-4.94914860e-01 1.42608142e+00 4.25674975e-01 4.23521370e-01
-6.79069221e-01 6.92076012e-02 5.91823637e-01 2.64423519e-01
3.11652962e-02 -5.81008255e-01 1.27294928e-01 -1.44208729e+00
-9.03395265e-02 -6.75453663e-01 1.13810945e+00 -4.13577020e-01
-1.88063776e+00 2.05947086e-01 4.04553294e-01 -1.38292229e+00
-3.09405208e-01 -4.85506803e-01 -6.83190107e-01 5.42278469e-01
-1.15539098e+00 -1.96166945e+00 -7.74425507e-01 1.41246998e+00
5.11677980e-01 -3.80067945e-01 5.44947147e-01 2.17473581e-02
-6.35199785e-01 9.36395288e-01 -4.91142809e-01 6.06588781e-01
4.53687131e-01 -1.18314755e+00 7.59817600e-01 9.16808009e-01
4.76588517e-01 9.01000321e-01 3.04128140e-01 -1.24361336e+00
-1.70500016e+00 -1.34466505e+00 5.34911811e-01 -3.63775700e-01
4.67886209e-01 -1.71040729e-01 -7.06064165e-01 7.74967432e-01
-4.98678446e-01 2.47176841e-01 7.55313873e-01 2.91061729e-01
-7.61834264e-01 -6.47928342e-02 -7.51836061e-01 8.65281701e-01
2.16059923e+00 -7.23042309e-01 -8.53775144e-01 2.61589646e-01
5.25484025e-01 -7.55998269e-02 -1.04712677e+00 3.54499429e-01
8.54169130e-01 -7.84437835e-01 1.61441171e+00 -6.54692769e-01
-3.10343176e-01 -3.92515361e-01 -6.95679709e-02 -9.48172569e-01
-8.00542712e-01 -7.25479126e-01 -7.06154406e-01 1.26788461e+00
-4.55325752e-01 -5.46055853e-01 1.07048690e+00 1.98768944e-01
-1.98684819e-02 -4.24041182e-01 -7.83063471e-01 -1.05489731e+00
-4.54181463e-01 -4.52865928e-01 7.20320702e-01 1.07431281e+00
4.24108356e-02 3.87039006e-01 -5.63195586e-01 5.56705222e-02
1.20760810e+00 2.67167509e-01 1.07090366e+00 -1.40236533e+00
-3.99455756e-01 -5.52755892e-01 -1.10165679e+00 -1.24221790e+00
3.67490888e-01 -1.36776054e+00 -5.04333675e-01 -1.67389643e+00
3.56385320e-01 -1.18528232e-01 -1.92180395e-01 4.25605893e-01
-4.93432045e-01 1.58188388e-01 2.96854973e-01 6.51126266e-01
-7.59554327e-01 8.60179007e-01 1.41868842e+00 -5.15492380e-01
-8.89405683e-02 4.98073921e-02 -4.64270711e-01 7.73691893e-01
3.52579564e-01 3.40982974e-02 -7.60886431e-01 5.15533164e-02
-1.99062694e-02 -9.28455144e-02 9.69239235e-01 -1.15149844e+00
3.79141480e-01 -1.34760916e-01 6.49452746e-01 -6.03352249e-01
3.35795999e-01 -4.50717568e-01 4.53349650e-01 5.91198564e-01
-9.06666368e-02 6.11245260e-02 -4.37413454e-01 1.23414946e+00
2.77001653e-02 1.77405238e-01 4.49375927e-01 -4.94588763e-01
-1.28393722e+00 9.42322493e-01 1.28972590e-01 5.47800735e-02
8.43794405e-01 -8.86896431e-01 1.86215922e-01 -2.80421376e-01
-1.16560113e+00 1.95732728e-01 4.95948583e-01 5.09360313e-01
1.14405692e+00 -1.93126047e+00 -6.70921087e-01 1.36835232e-01
4.69580352e-01 -3.30389142e-01 4.59653258e-01 6.62964880e-01
-3.22604865e-01 -4.00480479e-02 -3.13901991e-01 -7.49174476e-01
-1.35635817e+00 6.73035502e-01 3.52483958e-01 -3.88014436e-01
-1.48838139e+00 8.77252221e-01 3.21082383e-01 -3.97779137e-01
1.56911425e-02 4.71633822e-02 -5.04377902e-01 6.05430864e-02
3.55847806e-01 7.48378158e-01 -4.50126588e-01 -1.13964176e+00
-5.40125310e-01 1.23038030e+00 2.28943914e-01 8.84919912e-02
8.81499410e-01 -3.46240968e-01 4.68204580e-02 1.92357376e-01
1.00132561e+00 -1.39526531e-01 -1.19313467e+00 -5.19803584e-01
1.12293281e-01 -5.91929078e-01 -6.45156741e-01 -2.67815739e-02
-1.01366127e+00 7.44405210e-01 5.50798655e-01 -1.03538230e-01
9.34009492e-01 1.90199450e-01 1.17854440e+00 2.31884584e-01
5.64282894e-01 -9.66365218e-01 6.92736208e-01 2.17426181e-01
8.27895463e-01 -8.42203140e-01 3.78678352e-01 -6.20046437e-01
-6.76256597e-01 1.18536246e+00 6.49005175e-01 -3.63092810e-01
6.93910062e-01 -5.48649490e-01 -5.04804611e-01 -6.33545518e-01
1.09558560e-01 -7.28785157e-01 7.30012059e-01 1.24767506e+00
1.06521033e-01 2.97533870e-01 -2.51308661e-02 8.53664339e-01
-2.62422711e-01 -3.41491789e-01 -2.93609887e-01 6.92308187e-01
-3.38373840e-01 -1.25238502e+00 -3.92028689e-01 3.32045317e-01
2.56205976e-01 2.78719664e-01 -6.19133592e-01 8.83774459e-01
1.76474199e-01 6.79163396e-01 -3.32336098e-01 -8.83950233e-01
4.82969940e-01 -5.04683740e-02 8.40117097e-01 -6.96343899e-01
-2.85753727e-01 2.12242398e-02 1.33211210e-01 -8.97370636e-01
-7.02594042e-01 -9.39204395e-01 -1.32349050e+00 -2.59879291e-01
8.82059485e-02 -9.84015539e-02 -2.26672262e-01 1.03022361e+00
2.00210676e-01 3.93755823e-01 2.52974331e-01 -1.04904151e+00
-3.44833612e-01 -7.56246150e-01 -8.89999151e-01 9.19409633e-01
-5.28954230e-02 -1.16608596e+00 1.05928510e-01 3.49348903e-01]
|
[14.525135040283203, 1.1826584339141846]
|
557bc8a9-fc77-4ee7-bb74-aa667dfc67f2
|
audio-visual-speech-enhancement-with-score
|
2306.01432
| null |
https://arxiv.org/abs/2306.01432v1
|
https://arxiv.org/pdf/2306.01432v1.pdf
|
Audio-Visual Speech Enhancement with Score-Based Generative Models
|
This paper introduces an audio-visual speech enhancement system that leverages score-based generative models, also known as diffusion models, conditioned on visual information. In particular, we exploit audio-visual embeddings obtained from a self-super\-vised learning model that has been fine-tuned on lipreading. The layer-wise features of its transformer-based encoder are aggregated, time-aligned, and incorporated into the noise conditional score network. Experimental evaluations show that the proposed audio-visual speech enhancement system yields improved speech quality and reduces generative artifacts such as phonetic confusions with respect to the audio-only equivalent. The latter is supported by the word error rate of a downstream automatic speech recognition model, which decreases noticeably, especially at low input signal-to-noise ratios.
|
['Timo Gerkmann', 'Simone Frintrop', 'Julius Richter']
|
2023-06-02
| null | null | null | null |
['lipreading', 'speech-enhancement']
|
['computer-vision', 'speech']
|
[ 1.83580235e-01 2.82499582e-01 -1.29969418e-01 -1.65979519e-01
-1.18451381e+00 -3.76501441e-01 6.41838133e-01 -2.45381668e-01
-2.06750423e-01 3.84475917e-01 8.24448705e-01 -1.85978681e-01
1.11552700e-02 -3.10114026e-01 -4.68632311e-01 -8.41117203e-01
1.79294020e-01 -3.58710915e-01 -5.02216555e-02 3.78481891e-05
-1.63655832e-01 7.73355365e-02 -1.72182465e+00 3.63507837e-01
9.14475739e-01 1.23613143e+00 3.52258950e-01 1.18421757e+00
1.75973345e-02 8.42885196e-01 -5.51139653e-01 -6.79452956e-01
-8.99915025e-02 -4.38752979e-01 -2.19055131e-01 1.96331993e-01
3.50619107e-01 -2.36467838e-01 -8.28131616e-01 1.25502038e+00
7.91409850e-01 1.65073797e-01 8.45305622e-01 -1.14096832e+00
-1.20144999e+00 6.72279835e-01 -2.14590371e-01 2.20427006e-01
7.87508488e-02 4.05725658e-01 1.21514881e+00 -1.30630958e+00
4.48882580e-01 1.40859210e+00 4.74468172e-01 6.99234486e-01
-1.30491006e+00 -5.81924975e-01 1.03575043e-01 4.26683903e-01
-1.25798941e+00 -9.23824906e-01 1.06154597e+00 -2.43244737e-01
1.11126256e+00 1.10255077e-01 4.63457137e-01 1.49187005e+00
1.67785555e-01 1.06120360e+00 7.99950957e-01 -3.08061630e-01
1.95831269e-01 3.60154420e-01 -3.45995456e-01 4.29913163e-01
-4.31978852e-01 6.51767492e-01 -1.12867332e+00 1.42670095e-01
4.25290942e-01 -4.68884617e-01 -2.97531366e-01 -4.48065579e-01
-5.99684119e-01 7.02006996e-01 2.97756970e-01 3.99321079e-01
-5.13518572e-01 2.04860196e-01 2.85066009e-01 3.94299895e-01
6.30137026e-01 -1.09506860e-01 -2.08092362e-01 -4.27876472e-01
-1.27194786e+00 -3.88194382e-01 2.81564742e-01 7.85416067e-01
4.38174099e-01 9.76099014e-01 -4.35793102e-01 1.23006856e+00
7.08151162e-01 7.57284582e-01 6.03785634e-01 -9.28785563e-01
3.83017063e-01 1.43285468e-02 -3.86574030e-01 -6.65776789e-01
8.64212960e-02 -8.16870332e-01 -8.37681472e-01 4.88422900e-01
-2.62548447e-01 -1.33541033e-01 -1.03503358e+00 1.98546469e+00
1.18142497e-02 3.22418094e-01 3.88388485e-01 7.26325929e-01
9.05561090e-01 8.01354468e-01 1.94179714e-01 -3.33924025e-01
1.11749387e+00 -1.05575228e+00 -1.14215064e+00 -1.30582243e-01
-1.68242846e-02 -7.97684491e-01 1.07486761e+00 3.20216417e-01
-1.48972106e+00 -9.43748534e-01 -1.08236098e+00 2.31432375e-02
-1.15231216e-01 1.57926232e-01 -5.29941879e-02 1.09674096e+00
-1.63220453e+00 4.68935430e-01 -5.80151379e-01 -1.40060177e-02
4.69513535e-01 -4.97279391e-02 -1.65578321e-01 1.83706418e-01
-1.18148530e+00 6.47326410e-01 -1.30418226e-01 -2.97036737e-01
-1.44392502e+00 -8.40494335e-01 -1.10466671e+00 2.86522210e-01
-1.22078378e-02 -7.27562070e-01 1.33958042e+00 -9.26734984e-01
-2.01688337e+00 3.85275602e-01 -4.22984123e-01 -4.65884268e-01
4.19349134e-01 -1.50578305e-01 -8.59115899e-01 2.70811349e-01
-2.20143870e-01 7.13218629e-01 1.49084437e+00 -1.42939317e+00
-5.68714440e-01 -1.50148809e-01 -3.57807904e-01 3.88740867e-01
-8.37285817e-01 -2.87426878e-02 -5.11565030e-01 -1.08708012e+00
-2.44218662e-01 -5.25017500e-01 -6.25782236e-02 6.28104955e-02
-3.17192912e-01 -1.98687054e-02 9.52012300e-01 -1.03122592e+00
1.23654294e+00 -2.70615482e+00 9.41898823e-02 1.12088092e-01
8.61193091e-02 4.99579310e-01 -4.53693062e-01 2.71200031e-01
-1.46362647e-01 7.84335509e-02 -2.66477522e-02 -7.51841187e-01
2.35055715e-01 -1.34666830e-01 -3.76142979e-01 1.79032460e-01
5.78968108e-01 8.26286376e-01 -7.81353116e-01 -3.71810377e-01
4.58700389e-01 1.27641094e+00 -7.67045677e-01 3.58363539e-01
7.67759830e-02 1.70217872e-01 2.16743082e-01 4.47680444e-01
6.04089797e-01 1.71973497e-01 -1.36106461e-01 -2.24670991e-01
5.51227629e-02 4.99048114e-01 -9.16401088e-01 1.80765605e+00
-7.27629066e-01 8.93914700e-01 2.24462345e-01 -4.38083321e-01
8.60820293e-01 6.96941495e-01 1.92685947e-01 -8.58712554e-01
-7.06166029e-02 9.47012380e-03 -2.83857018e-01 -3.63077879e-01
4.93360400e-01 -1.32649913e-01 3.44305396e-01 8.32520649e-02
6.29837036e-01 -4.54916544e-02 -2.95705855e-01 3.38370055e-01
9.63604510e-01 -6.35718778e-02 -4.29249778e-02 1.01625659e-01
5.46968400e-01 -7.60388315e-01 1.48780003e-01 5.36273956e-01
-4.14836705e-01 6.29023254e-01 9.18843597e-02 5.96446931e-01
-1.16397583e+00 -1.88676381e+00 -1.68690994e-01 1.38526142e+00
-1.80814102e-01 -4.11948532e-01 -8.89361560e-01 -3.09054673e-01
-9.37422514e-02 7.80471146e-01 -4.40168113e-01 -5.09750724e-01
8.45385194e-02 -1.58247247e-01 7.71093667e-01 7.48924315e-01
9.61285457e-02 -8.31360042e-01 4.17616256e-02 3.37773621e-01
-3.31620052e-02 -9.91933227e-01 -6.80105031e-01 2.62675285e-01
-6.00601852e-01 -3.60440105e-01 -1.22611153e+00 -7.13642120e-01
2.40327239e-01 1.20268904e-01 6.23762786e-01 -6.36080563e-01
-9.59371179e-02 7.32914031e-01 -1.60146877e-01 -2.54745752e-01
-8.54054272e-01 -3.21357042e-01 2.76414335e-01 3.89818847e-01
7.77112842e-02 -7.01960802e-01 -5.70929646e-01 4.40112874e-02
-6.78856134e-01 -3.39529335e-01 5.98343015e-01 1.08413017e+00
4.02293772e-01 -4.37412672e-02 9.38972235e-01 2.30462477e-01
8.93269718e-01 -4.82105874e-02 -2.77277738e-01 -2.57393178e-02
-6.76086545e-01 1.14712976e-01 2.63339460e-01 -6.71018362e-01
-1.53310847e+00 -1.56677678e-01 -4.47536439e-01 -9.62549567e-01
-6.59591407e-02 1.91783488e-01 -5.03193736e-01 1.43563271e-01
5.40470302e-01 4.42651749e-01 8.38563517e-02 -5.47510982e-01
8.01330030e-01 1.14122593e+00 8.47345233e-01 4.60676961e-02
8.89397085e-01 1.48900658e-01 -6.40045583e-01 -1.16036069e+00
-1.93695799e-01 -5.26207805e-01 -2.36664057e-01 -4.05038625e-01
8.12774241e-01 -1.16674793e+00 -5.52935660e-01 5.03892720e-01
-1.11166894e+00 -2.30200186e-01 -6.24016702e-01 5.65158963e-01
-7.37115979e-01 3.23709339e-01 -8.46058667e-01 -1.33241820e+00
-4.50360149e-01 -1.08274770e+00 1.05131853e+00 1.58524141e-01
-3.50062065e-02 -9.84827697e-01 6.74312338e-02 2.81164676e-01
7.71425486e-01 -5.47150493e-01 8.01587105e-01 -2.71284878e-01
-3.11377108e-01 1.58546478e-01 -1.69358224e-01 9.90159154e-01
1.52035877e-01 1.07032776e-01 -1.76966679e+00 -2.62179852e-01
-8.26762021e-02 -2.04825073e-01 9.83494341e-01 6.73441589e-01
8.02093506e-01 -3.89021814e-01 -5.00432700e-02 4.49230909e-01
1.04730070e+00 2.95273125e-01 5.46571255e-01 -3.77458364e-01
4.60697234e-01 3.88724357e-01 1.51643917e-01 4.77348328e-01
6.09871335e-02 8.27957809e-01 3.48055452e-01 -3.13250273e-01
-1.11386502e+00 -6.71999991e-01 9.20411348e-01 1.33718109e+00
1.74560711e-01 -2.95002878e-01 -3.97998720e-01 7.83538699e-01
-1.26825249e+00 -1.02290213e+00 3.19413275e-01 1.94402802e+00
1.04927146e+00 6.04632311e-02 1.53584301e-01 4.45070505e-01
7.91591942e-01 3.71194959e-01 -5.59989572e-01 -5.25377989e-01
-3.29824358e-01 4.32559609e-01 3.49893384e-02 8.52635264e-01
-9.43026006e-01 9.98522520e-01 6.69496107e+00 1.09149301e+00
-1.11261940e+00 4.85705763e-01 4.06099588e-01 -4.12889510e-01
-5.11073530e-01 -5.61146319e-01 -4.28366899e-01 2.10642919e-01
1.34670722e+00 -4.75073457e-01 4.39333737e-01 9.09940660e-01
5.32331645e-01 4.99889195e-01 -8.72559786e-01 1.00269973e+00
2.77712643e-01 -1.15681052e+00 7.09891841e-02 1.18028924e-01
4.87415135e-01 1.91619754e-01 9.21894848e-01 3.37798655e-01
3.27591926e-01 -9.08632517e-01 9.51978445e-01 2.54027575e-01
1.23285890e+00 -8.85629058e-01 3.61648083e-01 -1.08682327e-01
-1.26677179e+00 -2.37286314e-01 -8.31940249e-02 4.69429076e-01
3.82776886e-01 5.27553976e-01 -9.95490968e-01 2.47000113e-01
7.38762021e-01 5.92203915e-01 -2.49775618e-01 9.93814290e-01
-3.82063568e-01 8.77195299e-01 1.06450491e-01 1.84617832e-01
7.40610659e-02 3.91545445e-01 9.27146494e-01 1.39876091e+00
5.17663777e-01 -4.30032074e-01 -4.99484986e-01 9.28853929e-01
-2.28736758e-01 1.25066668e-01 -8.10935318e-01 -5.08815870e-02
4.03281778e-01 9.89185035e-01 1.65326908e-01 -1.88072771e-01
-2.53923267e-01 1.17961466e+00 -1.24047481e-01 7.38047957e-01
-8.93589079e-01 -4.18777466e-01 1.08771217e+00 1.01626525e-02
6.03408933e-01 2.46150959e-02 -1.13861427e-01 -8.45630825e-01
-2.56115288e-01 -8.52036476e-01 -1.77857101e-01 -1.09298956e+00
-1.19986916e+00 7.47759581e-01 -5.38208127e-01 -1.11572039e+00
-5.25662363e-01 -4.63875055e-01 -3.26872796e-01 8.47601533e-01
-1.59114397e+00 -1.00370109e+00 1.45814493e-01 8.07258248e-01
7.82382607e-01 -4.15613562e-01 7.48170972e-01 5.06307065e-01
-2.84692943e-01 1.32267654e+00 2.00989202e-01 1.10905580e-01
6.69174194e-01 -9.23204541e-01 4.47389334e-01 9.31627035e-01
4.86314803e-01 2.53605515e-01 6.57701552e-01 -3.95616710e-01
-1.02305615e+00 -1.34443402e+00 1.00999534e+00 -2.57161319e-01
6.79879963e-01 -4.67418313e-01 -8.55860412e-01 1.99609220e-01
6.84013486e-01 -9.15521979e-02 7.84993410e-01 -8.55604783e-02
-6.34187341e-01 -2.37833843e-01 -1.01911211e+00 6.74324930e-01
1.04005527e+00 -1.25055087e+00 -6.05619013e-01 -1.82582885e-01
1.15420485e+00 1.16873262e-02 -7.19827712e-01 1.74846038e-01
3.86030644e-01 -8.26648712e-01 1.13281202e+00 -6.41854465e-01
2.53903300e-01 -4.12868522e-02 -4.16551054e-01 -1.71083057e+00
-5.01092851e-01 -8.61234128e-01 -1.73069373e-01 1.60378873e+00
7.11397886e-01 -2.74054110e-01 4.89974499e-01 -2.11783633e-01
-3.88330877e-01 -1.82498619e-01 -1.39432943e+00 -1.08973956e+00
1.97945591e-02 -9.34479177e-01 1.51213914e-01 6.54302061e-01
2.52336740e-01 4.20263737e-01 -5.27154446e-01 3.65378082e-01
6.35707021e-01 -6.96105540e-01 3.11532080e-01 -9.36586142e-01
-3.28313261e-01 -6.47337139e-01 -5.28033316e-01 -1.02012372e+00
1.06874675e-01 -9.31467235e-01 2.66100556e-01 -1.28617203e+00
-3.09782699e-02 5.55638038e-02 -6.30872369e-01 2.80249953e-01
-5.36095584e-04 3.69556457e-01 4.05558169e-01 -2.54617512e-01
-4.10866320e-01 1.16045308e+00 9.81879771e-01 -4.81790006e-01
-2.87981689e-01 -6.31159618e-02 -6.33652151e-01 6.16498590e-01
5.05015850e-01 -3.34034026e-01 -6.12349689e-01 -3.13829750e-01
-2.14466900e-01 1.95250660e-01 2.48843729e-01 -1.21127450e+00
3.38871330e-01 2.11864397e-01 1.12792430e-02 -3.33439022e-01
8.13622713e-01 -6.33367717e-01 -1.44307688e-01 2.69700617e-01
-7.73776114e-01 -3.26212227e-01 2.79426008e-01 8.97586405e-01
-5.21434367e-01 2.33391464e-01 1.03780782e+00 5.49318075e-01
-2.85951465e-01 1.61785215e-01 -6.42254531e-01 -4.86501008e-02
8.11253071e-01 -1.04079276e-01 -2.80494004e-01 -9.50463533e-01
-9.73282337e-01 -4.58220422e-01 -1.17661068e-02 7.62037158e-01
9.45879340e-01 -1.69744682e+00 -8.98412049e-01 4.54750687e-01
2.25269943e-01 -7.41503060e-01 5.51942647e-01 8.00164223e-01
5.35591960e-01 4.71303225e-01 5.29343560e-02 -7.45427132e-01
-1.43635666e+00 6.69133365e-01 3.24169189e-01 1.84862658e-01
-4.46944565e-01 1.17908263e+00 4.01644200e-01 5.37848622e-02
7.25007832e-01 -3.08878213e-01 -1.04729973e-01 6.71490747e-03
6.60321355e-01 2.70708293e-01 1.90432817e-01 -8.04967523e-01
-3.38145107e-01 2.78199971e-01 1.58670664e-01 -6.58436954e-01
1.16705775e+00 -4.78171915e-01 6.89171970e-01 3.49115759e-01
1.25103748e+00 2.94898123e-01 -1.60368526e+00 -5.51151752e-01
-3.94080758e-01 -2.30882749e-01 6.23827219e-01 -1.04082942e+00
-1.27063763e+00 1.39523017e+00 1.18208075e+00 8.17144364e-02
1.00197911e+00 1.11766107e-01 6.65474474e-01 -2.59556144e-01
-8.41178298e-02 -1.31125176e+00 6.39118731e-01 3.50894600e-01
9.56798494e-01 -9.67541337e-01 -7.94232428e-01 -7.25056231e-02
-9.06665564e-01 6.35896146e-01 7.46815279e-02 2.62030393e-01
7.29738951e-01 7.34196246e-01 3.54718834e-01 3.93338829e-01
-1.06336427e+00 -4.38806415e-01 5.44357479e-01 1.33957517e+00
2.89426982e-01 5.75511716e-02 4.66807961e-01 7.60056973e-01
-2.74393350e-01 -2.10576877e-01 1.17603518e-01 2.80872852e-01
-5.43816030e-01 -7.79015601e-01 -2.60349602e-01 -1.07143454e-01
-3.21713686e-01 -6.45875931e-01 -2.85724580e-01 -5.15429489e-03
-2.30879575e-01 1.53338933e+00 8.99163410e-02 -7.53931820e-01
2.82619208e-01 3.42193365e-01 2.09790662e-01 -2.51422375e-01
-4.71483946e-01 5.42773128e-01 6.43820092e-02 -5.17569840e-01
-1.71309620e-01 -5.62825978e-01 -1.00434661e+00 -3.35290916e-02
-2.78231293e-01 -1.66881904e-01 9.05616343e-01 6.86704457e-01
6.31597757e-01 9.92284477e-01 8.31453919e-01 -7.52025545e-01
-6.85562968e-01 -1.06034005e+00 -5.20855784e-01 2.86355913e-01
7.30977714e-01 -4.82472718e-01 -6.79244161e-01 2.14826941e-01]
|
[14.904451370239258, 5.9462456703186035]
|
6c32e613-1177-45cf-9e11-e48da4c2bc03
|
self-discriminative-learning-for-unsupervised
| null | null |
https://aclanthology.org/N19-1255
|
https://aclanthology.org/N19-1255.pdf
|
Self-Discriminative Learning for Unsupervised Document Embedding
|
Unsupervised document representation learning is an important task providing pre-trained features for NLP applications. Unlike most previous work which learn the embedding based on self-prediction of the surface of text, we explicitly exploit the inter-document information and directly model the relations of documents in embedding space with a discriminative network and a novel objective. Extensive experiments on both small and large public datasets show the competitiveness of the proposed method. In evaluations on standard document classification, our model has errors that are 5 to 13{\%} lower than state-of-the-art unsupervised embedding models. The reduction in error is even more pronounced in scarce label setting.
|
['Shou-De Lin', 'Chin-Hua Hu', 'Hong-You Chen', 'Leila Wehbe']
|
2019-06-01
| null | null | null |
naacl-2019-6
|
['document-embedding']
|
['methodology']
|
[ 2.03804478e-01 3.95091027e-01 -6.51561558e-01 -6.38476968e-01
-5.39731920e-01 -5.46364725e-01 1.08611441e+00 6.71797872e-01
-5.70470870e-01 4.86060351e-01 5.13721347e-01 -1.99954808e-02
-3.42512667e-01 -6.63786173e-01 -2.95147985e-01 -6.46087646e-01
-1.05434731e-01 8.66798818e-01 1.27584368e-01 -6.39817044e-02
4.03729439e-01 3.46004039e-01 -1.36339855e+00 3.28710884e-01
7.23279715e-01 8.75806212e-01 -6.13644756e-02 5.79890668e-01
-5.28263390e-01 5.50256014e-01 -3.74557376e-01 -2.79688746e-01
5.15649058e-02 4.53976989e-02 -8.62649202e-01 1.16932146e-01
5.18541932e-01 -2.28196532e-02 -5.30843794e-01 9.47131038e-01
3.34642351e-01 3.02241087e-01 1.29111612e+00 -9.26625192e-01
-1.01694572e+00 6.14415586e-01 -4.53401387e-01 1.52298421e-01
3.89878571e-01 -7.16464996e-01 1.63081825e+00 -1.18847394e+00
7.22280502e-01 1.20119262e+00 6.20195150e-01 2.25812241e-01
-1.40877795e+00 -2.72368342e-01 6.13680184e-02 2.53915846e-01
-1.28151357e+00 -2.31127918e-01 9.65805590e-01 -5.94620883e-01
1.28567469e+00 9.88462418e-02 7.42290616e-02 1.14739931e+00
2.02280730e-01 6.80183709e-01 1.02386808e+00 -1.03388095e+00
-3.33319120e-02 6.57300174e-01 7.51824498e-01 6.52723610e-01
2.97938019e-01 -9.87604111e-02 -5.52839041e-01 -3.08479756e-01
3.23063284e-01 2.84799159e-01 -7.41722509e-02 -5.08227825e-01
-1.16300106e+00 1.16813862e+00 2.85502046e-01 7.11252868e-01
-1.07575260e-01 2.84930356e-02 4.90498096e-01 2.80703187e-01
1.00110149e+00 3.92065704e-01 -3.08580160e-01 -5.97047582e-02
-9.52938914e-01 -1.12838466e-02 9.83899355e-01 9.37629223e-01
6.98908389e-01 -2.04810247e-01 -1.51213169e-01 1.10229158e+00
5.49825728e-01 1.28501505e-01 8.86159241e-01 -3.75126630e-01
6.47437155e-01 9.35328305e-01 -2.06702009e-01 -1.37159503e+00
-3.80815715e-01 -4.61251915e-01 -9.15220499e-01 -2.41101757e-01
7.98486471e-02 3.14873338e-01 -6.44961953e-01 1.06247258e+00
2.21241042e-01 6.95433840e-02 1.99688226e-01 3.36013436e-01
7.37073779e-01 8.35200369e-01 -5.29640839e-02 -6.61666617e-02
1.43186390e+00 -1.38075531e+00 -9.61641133e-01 -2.10668996e-01
1.11972260e+00 -7.74961770e-01 8.84338737e-01 2.05133647e-01
-4.88490671e-01 -6.17147744e-01 -9.93553460e-01 -2.33509928e-01
-9.86547530e-01 4.22933072e-01 8.88451576e-01 7.19446838e-01
-8.16028118e-01 5.92609286e-01 -5.38585663e-01 -5.15887797e-01
5.20005584e-01 4.43661898e-01 -7.99996316e-01 -1.21312849e-01
-1.07873535e+00 8.32847178e-01 6.08484566e-01 -4.48475443e-02
-3.21173191e-01 -6.58531129e-01 -9.76207554e-01 3.13103274e-02
-3.59899923e-02 -1.82496205e-01 5.92802167e-01 -3.85105580e-01
-1.43511903e+00 9.15082932e-01 -2.29507282e-01 -4.66752261e-01
2.30257332e-01 -3.62688065e-01 -4.63742286e-01 2.70194173e-01
-7.74142370e-02 5.39357483e-01 7.88679242e-01 -1.26513600e+00
-2.93679118e-01 -3.96256685e-01 -9.57023576e-02 1.03066929e-01
-1.36587346e+00 -1.01734148e-02 -2.27927685e-01 -7.26134658e-01
6.66766465e-02 -8.02085042e-01 4.15398031e-02 4.72716317e-02
-3.63062412e-01 -7.92789280e-01 1.15125024e+00 -6.13417447e-01
1.27316642e+00 -2.04271221e+00 7.95180872e-02 2.51025826e-01
2.16481507e-01 1.88835353e-01 -1.42738283e-01 7.83068061e-01
-5.25678843e-02 -1.84441581e-02 -1.55943573e-01 -6.41255796e-01
3.96700293e-01 3.52392524e-01 -5.56683242e-01 5.90485632e-01
1.58026405e-02 7.94234097e-01 -8.57897282e-01 -7.59882450e-01
6.54201508e-02 7.33851612e-01 -5.76141000e-01 2.19507694e-01
3.68662961e-02 8.71426985e-02 -3.97258669e-01 4.29059923e-01
3.04025948e-01 -4.22158301e-01 2.29315937e-01 -6.90807328e-02
2.52435476e-01 5.88595510e-01 -1.07605219e+00 1.82530296e+00
-5.64816773e-01 9.68085766e-01 -6.68098390e-01 -1.48152649e+00
1.10393190e+00 3.92432511e-01 3.15562636e-01 -2.68783450e-01
-3.86361405e-02 1.14208926e-02 -1.78832173e-01 -4.37402695e-01
6.26252651e-01 1.83564842e-01 1.48855135e-01 7.17167199e-01
4.98792827e-01 1.69123650e-01 2.31785223e-01 3.79998147e-01
1.01983142e+00 -2.32687369e-02 1.86280996e-01 -5.57657540e-01
6.25612974e-01 -5.25436103e-01 -8.15833639e-03 5.92247605e-01
3.42681408e-02 4.02177840e-01 5.55164337e-01 -3.00458699e-01
-8.28213990e-01 -6.03970408e-01 -5.60105562e-01 1.43655396e+00
-1.14618361e-01 -8.63029957e-01 -4.27574724e-01 -1.24215174e+00
2.73466825e-01 7.63654768e-01 -1.02629519e+00 -3.25656943e-02
-2.77102798e-01 -7.60319948e-01 4.29018736e-01 8.06693017e-01
4.75385413e-02 -8.99635673e-01 6.37509599e-02 2.58925129e-02
1.82919607e-01 -1.26294684e+00 -3.19054663e-01 4.01055753e-01
-1.01396644e+00 -7.97812700e-01 -5.85487366e-01 -1.19194138e+00
1.02131402e+00 2.56057382e-02 9.01831448e-01 -1.62077192e-02
-3.77905577e-01 5.19973755e-01 -4.78266627e-01 -3.02266985e-01
-1.22223973e-01 3.75586897e-01 7.25547001e-02 2.50681993e-02
8.35555971e-01 -5.74377239e-01 -4.62302804e-01 1.29845515e-01
-8.42392683e-01 -4.37199533e-01 4.29586053e-01 1.21617448e+00
4.26079601e-01 3.05280417e-01 4.41967994e-01 -1.31901264e+00
7.60481179e-01 -5.12508750e-01 -3.60360712e-01 2.35609666e-01
-1.32709312e+00 3.39370370e-01 4.69007403e-01 -4.05450135e-01
-9.26396787e-01 -1.70483783e-01 3.01492333e-01 8.43187142e-03
-1.45721912e-01 5.18557668e-01 1.84755519e-01 1.65891767e-01
6.53819561e-01 2.98085690e-01 -2.81967223e-01 -7.91709781e-01
5.04455626e-01 8.71721864e-01 2.82507818e-02 -5.67195833e-01
9.17889357e-01 5.67943633e-01 -1.05749033e-02 -8.24727118e-01
-9.28851485e-01 -9.20351267e-01 -1.24282837e+00 2.62855560e-01
5.83841980e-01 -7.11631000e-01 -2.34522179e-01 -2.92470992e-01
-1.30404520e+00 3.02810613e-02 -2.82283276e-01 6.60201013e-01
-2.28932962e-01 6.42080128e-01 -5.08065522e-01 -6.85006797e-01
-4.74009722e-01 -7.86107242e-01 1.18896735e+00 -2.17402965e-01
-3.88635516e-01 -1.77409017e+00 6.51808858e-01 3.60062778e-01
2.96358794e-01 -6.29062504e-02 1.12351477e+00 -1.21757615e+00
2.50786189e-02 -6.01427734e-01 -5.00353217e-01 5.13867021e-01
2.79973447e-01 -1.02950633e-01 -1.10053480e+00 -3.47197801e-01
-3.84886980e-01 -4.22481865e-01 1.18651140e+00 -8.50286856e-02
1.15794945e+00 -3.74948889e-01 -6.40733719e-01 2.93791622e-01
1.47031307e+00 -1.71788633e-01 4.47496921e-01 4.85191613e-01
8.36984098e-01 9.01863754e-01 5.00498116e-01 4.74474758e-01
2.71248221e-01 7.05644011e-01 9.85245705e-02 2.43521169e-01
-9.81638283e-02 -3.14881593e-01 4.88485843e-01 1.19766808e+00
-7.71903470e-02 -3.30071211e-01 -9.66064692e-01 6.50070369e-01
-1.85176218e+00 -7.90810168e-01 -1.67842031e-01 1.69257545e+00
1.01168656e+00 2.73098797e-01 -2.68007487e-01 4.74311382e-01
5.19326270e-01 3.68752122e-01 9.49870870e-02 -5.23998499e-01
-4.24381755e-02 4.30951744e-01 4.14801061e-01 5.39104044e-01
-1.19819117e+00 9.91019249e-01 7.00113440e+00 7.89648235e-01
-6.54740572e-01 2.49444351e-01 1.51427805e-01 2.28702039e-01
-2.72910565e-01 5.06364480e-02 -1.16224968e+00 2.71462142e-01
1.06439257e+00 3.83968130e-02 -9.80620012e-02 1.00575507e+00
-3.59053493e-01 3.36774796e-01 -1.47278595e+00 8.94678175e-01
5.60572088e-01 -1.28062415e+00 2.11673662e-01 3.32266599e-01
9.38845515e-01 -1.47480816e-01 2.08044350e-01 3.58284861e-01
1.83340132e-01 -1.12198913e+00 1.87334254e-01 4.39742267e-01
6.22069359e-01 -7.15524018e-01 1.00269330e+00 1.26991004e-01
-1.06970942e+00 9.63546336e-02 -8.41416478e-01 -2.13495381e-02
-2.09549204e-01 7.41746187e-01 -9.78807867e-01 4.52858955e-01
3.77230078e-01 1.19566202e+00 -8.77093375e-01 5.47988534e-01
-3.83797348e-01 8.06434810e-01 -2.20571056e-01 -1.54693022e-01
3.69258672e-01 -1.78647980e-01 3.03724438e-01 1.72297978e+00
4.41929251e-02 -2.61648655e-01 -2.40727272e-02 6.13662839e-01
-3.58735144e-01 5.56762934e-01 -7.97879636e-01 -4.91222531e-01
2.01068625e-01 1.49397516e+00 -5.75125515e-01 -4.33708787e-01
-4.03162152e-01 1.15490091e+00 6.25579953e-01 1.98442936e-01
-3.58511835e-01 -7.60189056e-01 2.05863774e-01 -1.12676121e-01
5.66020131e-01 -2.36538336e-01 -1.61263883e-01 -1.12188911e+00
1.54965490e-01 -3.28741610e-01 3.15628529e-01 -1.75655514e-01
-1.64558971e+00 6.30934358e-01 -6.73009157e-02 -1.15217936e+00
-3.41756225e-01 -1.14238918e+00 -5.15676022e-01 5.67226708e-01
-1.83862925e+00 -1.23217893e+00 -2.94876061e-02 4.00774747e-01
4.59799200e-01 -6.49730504e-01 1.46098197e+00 3.10510069e-01
-5.58181643e-01 7.98467100e-01 7.84289479e-01 3.67412686e-01
9.47468162e-01 -1.56407380e+00 2.59514377e-02 2.50254422e-01
8.66251111e-01 7.66560316e-01 4.62013245e-01 -2.45549947e-01
-1.24202657e+00 -9.07797873e-01 1.51512623e+00 -8.69083047e-01
9.48198318e-01 -7.79841900e-01 -9.10436213e-01 7.58821845e-01
4.25029010e-01 4.06689972e-01 1.33458078e+00 5.22532701e-01
-9.92210507e-01 -3.87294777e-02 -9.04717743e-01 2.21637726e-01
7.58533657e-01 -8.90717566e-01 -1.02280653e+00 7.34305203e-01
6.03911996e-01 1.38879910e-01 -1.16098452e+00 4.76949103e-02
6.97086096e-01 -5.31688154e-01 1.08565688e+00 -8.76500666e-01
7.18822837e-01 2.04863548e-01 -3.31926376e-01 -1.22481370e+00
-3.61395806e-01 -3.31292868e-01 -5.85199535e-01 1.56273901e+00
3.81907254e-01 -6.82538748e-01 8.36907029e-01 3.16307187e-01
2.20847130e-01 -8.12882006e-01 -6.88082278e-01 -9.34502542e-01
1.81794629e-01 -3.27036500e-01 7.33244419e-02 1.32624626e+00
4.70871776e-01 5.56822240e-01 -1.49573267e-01 9.10941735e-02
7.28598535e-01 -4.76748422e-02 6.09857142e-01 -1.62579536e+00
-3.38219464e-01 -3.53241116e-01 -8.16976368e-01 -1.01565814e+00
9.13651764e-01 -1.45468628e+00 -2.80584693e-01 -1.64510965e+00
3.22660685e-01 -3.49190503e-01 -8.07985365e-01 4.96074110e-01
3.53737548e-02 2.28155792e-01 -1.94699541e-01 3.37319374e-01
-8.32965016e-01 8.43303204e-01 8.00689340e-01 -5.77744961e-01
9.26894322e-02 -5.40395021e-01 -4.58492726e-01 7.12466598e-01
4.97280747e-01 -7.68350422e-01 -3.82168323e-01 -5.08282423e-01
1.87838808e-01 -7.70861268e-01 5.30210733e-02 -6.18044734e-01
2.67803520e-01 1.86964288e-01 5.01881421e-01 -5.69511056e-01
3.20065945e-01 -1.08086801e+00 -7.26110160e-01 1.69409052e-01
-9.27026570e-01 -2.97748357e-01 -3.03365593e-03 1.06000769e+00
-3.65572482e-01 -7.08307385e-01 2.41482481e-01 3.59826952e-01
-4.18142974e-01 1.78733319e-01 -6.86533451e-02 -2.86646932e-02
8.17699075e-01 -5.07963821e-02 -2.19251513e-01 -7.57083446e-02
-5.35663784e-01 -9.56529751e-02 -2.24074647e-02 5.52009404e-01
5.96693158e-01 -1.57895720e+00 -5.60006082e-01 1.95701897e-01
5.82689881e-01 -2.03628063e-01 -2.29047880e-01 6.52623534e-01
-2.87998438e-01 9.96232033e-01 1.55003369e-01 -4.75778311e-01
-1.15141416e+00 3.75413716e-01 -2.43368194e-01 -7.83903122e-01
-7.95362055e-01 8.09434533e-01 7.92330205e-02 -5.85741043e-01
3.46209288e-01 -1.31232649e-01 -6.05283618e-01 4.10019189e-01
5.40194869e-01 4.27104861e-01 8.67411941e-02 -6.76478207e-01
-3.53494734e-01 5.50646067e-01 -6.34345710e-01 -3.37939747e-02
1.66376889e+00 5.53822704e-02 -3.05738121e-01 7.88176060e-01
1.81611717e+00 3.64011414e-02 -8.06684911e-01 -7.59317875e-01
3.79121780e-01 -4.73015606e-01 4.50329632e-01 -3.93967271e-01
-7.02925324e-01 1.11810327e+00 5.18251896e-01 2.57411927e-01
3.13091666e-01 3.89047638e-02 5.11441469e-01 8.95428360e-01
1.20229041e-02 -1.31691539e+00 3.98076773e-01 4.48496461e-01
9.32051718e-01 -1.41762030e+00 3.10513407e-01 -3.63514185e-01
-3.36946487e-01 1.45871282e+00 1.56414956e-01 -4.84766096e-01
1.10734975e+00 1.31004534e-04 -2.21774280e-01 -3.11999083e-01
-7.02903509e-01 5.67074642e-02 7.74369121e-01 5.50897956e-01
8.39793503e-01 -2.23601207e-01 -3.96875232e-01 5.62581062e-01
-1.26269147e-01 -3.88527811e-01 1.77938566e-02 1.03899634e+00
-2.68830806e-01 -1.48302996e+00 4.51728189e-03 4.62263048e-01
-3.54674578e-01 -2.17178613e-01 -4.39443856e-01 5.88066936e-01
-1.57570094e-01 8.18149209e-01 1.14725381e-01 -1.02136984e-01
1.82103999e-02 4.75904703e-01 4.33813214e-01 -9.51118827e-01
-3.38005513e-01 -1.74471498e-01 -6.04136176e-02 -2.80792058e-01
-5.82803428e-01 -4.69264060e-01 -1.17069221e+00 1.46199137e-01
-5.61957657e-01 1.67949349e-01 7.27156281e-01 1.00769997e+00
4.52851444e-01 4.83060271e-01 6.97509706e-01 -6.35496259e-01
-6.98658884e-01 -1.36381686e+00 -8.57604742e-01 6.26702368e-01
1.33746326e-01 -8.62064362e-01 -6.32130444e-01 3.30453843e-01]
|
[10.438349723815918, 8.277606964111328]
|
8a3a9867-2c54-4f1d-8663-a57e1197f079
|
learning-gradient-fields-for-shape-generation
|
2008.06520
| null |
https://arxiv.org/abs/2008.06520v2
|
https://arxiv.org/pdf/2008.06520v2.pdf
|
Learning Gradient Fields for Shape Generation
|
In this work, we propose a novel technique to generate shapes from point cloud data. A point cloud can be viewed as samples from a distribution of 3D points whose density is concentrated near the surface of the shape. Point cloud generation thus amounts to moving randomly sampled points to high-density areas. We generate point clouds by performing stochastic gradient ascent on an unnormalized probability density, thereby moving sampled points toward the high-likelihood regions. Our model directly predicts the gradient of the log density field and can be trained with a simple objective adapted from score-based generative models. We show that our method can reach state-of-the-art performance for point cloud auto-encoding and generation, while also allowing for extraction of a high-quality implicit surface. Code is available at https://github.com/RuojinCai/ShapeGF.
|
['Hadar Averbuch-Elor', 'Ruojin Cai', 'Serge Belongie', 'Guandao Yang', 'Zekun Hao', 'Noah Snavely', 'Bharath Hariharan']
|
2020-08-14
| null |
https://www.ecva.net/papers/eccv_2020/papers_ECCV/html/462_ECCV_2020_paper.php
|
https://www.ecva.net/papers/eccv_2020/papers_ECCV/papers/123480375.pdf
|
eccv-2020-8
|
['point-cloud-generation']
|
['computer-vision']
|
[ 8.98430198e-02 2.17893422e-01 1.71810798e-02 -1.17850840e-01
-1.23802495e+00 -5.31761587e-01 8.20915341e-01 -5.02053052e-02
6.18247129e-02 5.57525396e-01 1.56901162e-02 -2.20827349e-02
2.34471247e-01 -1.27045667e+00 -1.10461020e+00 -5.68804979e-01
1.41559793e-02 1.35487747e+00 1.68610930e-01 1.52895883e-01
2.65225559e-01 9.71580088e-01 -1.58272183e+00 8.92340615e-02
9.87890780e-01 5.29848695e-01 3.96965593e-01 8.39374840e-01
-3.17910731e-01 -2.57399887e-01 -4.58153725e-01 -3.75454068e-01
2.64238596e-01 -2.64792651e-01 -4.81996149e-01 1.18651606e-01
4.89962935e-01 -1.35031492e-01 2.82869413e-02 1.00107002e+00
2.81752616e-01 1.06018912e-02 1.28558075e+00 -9.95460093e-01
-7.02029228e-01 -3.63894589e-02 -5.12825727e-01 -2.64270812e-01
2.35651731e-01 2.16604084e-01 9.49908257e-01 -1.50766671e+00
7.02488601e-01 1.10367680e+00 4.63505119e-01 5.45075178e-01
-1.52304482e+00 -5.64672172e-01 -4.01338577e-01 -4.93553191e-01
-1.70790768e+00 -3.74823600e-01 8.53579700e-01 -5.78529716e-01
7.89812684e-01 1.94491506e-01 9.60201442e-01 7.36170113e-01
-4.85833315e-03 8.67879272e-01 5.88648915e-01 -4.07518804e-01
5.48828244e-01 -2.54429411e-04 -6.54710948e-01 4.91881579e-01
3.77955317e-01 4.35386933e-02 -2.94234931e-01 -7.69688964e-01
1.23609996e+00 4.72868010e-02 -8.10857192e-02 -7.18342304e-01
-1.01654470e+00 8.94251883e-01 4.56420779e-01 1.08326666e-01
-5.98502338e-01 4.08375144e-01 -3.58250827e-01 -4.38841075e-01
7.17149079e-01 2.54868895e-01 -1.23575561e-01 -3.07726800e-01
-1.16136670e+00 7.21719503e-01 6.31412864e-01 1.30986881e+00
1.10336351e+00 -3.89419682e-02 -1.41639218e-01 8.38967621e-01
5.80126524e-01 1.11350465e+00 -1.53656587e-01 -1.11174750e+00
2.63173729e-01 5.07643521e-01 3.33669960e-01 -5.53945601e-01
3.70874405e-01 -1.74818292e-01 -6.06202066e-01 5.68442464e-01
2.05299333e-01 -4.80682701e-02 -1.20826864e+00 1.31281340e+00
4.14714396e-01 4.45857197e-01 -2.07198143e-01 7.40122259e-01
3.00007164e-01 1.00611520e+00 -1.11838445e-01 1.74116060e-01
9.82483268e-01 -3.34362328e-01 -1.69579551e-01 -9.48568061e-02
1.86767623e-01 -7.40714133e-01 1.21982181e+00 2.18358651e-01
-1.38662076e+00 -1.96734637e-01 -6.82531536e-01 3.36747393e-02
1.48515999e-01 1.00817837e-01 3.35492343e-01 4.19643551e-01
-1.11414933e+00 7.50933468e-01 -1.19283938e+00 -7.59464968e-03
8.93821597e-01 1.67146981e-01 -7.99201503e-02 -8.56399462e-02
-4.31520522e-01 3.28463346e-01 7.67930076e-02 -3.43690962e-01
-8.09345067e-01 -8.34489584e-01 -7.26219594e-01 -1.48046520e-02
-1.56385347e-01 -1.05300081e+00 1.19811738e+00 -4.28588986e-01
-1.43994677e+00 7.98939705e-01 -6.39051259e-01 -1.59420237e-01
2.97809988e-01 -1.83008209e-01 1.74360007e-01 1.88506052e-01
1.96637675e-01 9.01163340e-01 8.62674057e-01 -1.75070202e+00
-3.24592590e-01 -2.60331184e-01 -6.23504817e-01 1.89868167e-01
1.63955450e-01 -3.67962599e-01 -5.29992342e-01 -5.65135300e-01
2.12726608e-01 -1.11157644e+00 -3.33841890e-01 1.56438857e-01
-3.83955330e-01 -2.78769791e-01 5.95414698e-01 -3.80831212e-01
8.09547305e-01 -1.93974769e+00 1.08644702e-01 5.42337537e-01
1.09434672e-01 -7.10178120e-03 2.13055890e-02 4.50249195e-01
1.35231167e-01 3.88112694e-01 -8.42253685e-01 -5.24058461e-01
6.90408573e-02 1.61919102e-01 -4.59745318e-01 3.57758909e-01
6.48423135e-01 9.40815866e-01 -1.08091271e+00 -3.47301930e-01
2.71973819e-01 7.49868274e-01 -8.30022514e-01 1.86004058e-01
-6.31598949e-01 4.01942134e-01 -5.30013144e-01 6.31933153e-01
1.00124621e+00 -4.22749907e-01 -2.34510362e-01 2.82049507e-01
1.28776163e-01 2.12918803e-01 -9.38711524e-01 1.76782990e+00
-3.33679020e-01 2.70107597e-01 -2.11752906e-01 -1.67531759e-01
1.21513820e+00 1.59932062e-01 5.71573019e-01 7.80663416e-02
-1.29861727e-01 2.55647421e-01 -2.68758029e-01 1.26212403e-01
6.72234416e-01 -4.63425219e-01 7.41836950e-02 2.93588161e-01
-7.30467886e-02 -1.04176247e+00 -2.45494649e-01 2.01137409e-01
8.53725851e-01 3.06981057e-01 -4.42834981e-02 -7.30942041e-02
1.10622076e-02 9.85054448e-02 1.94876313e-01 4.84146386e-01
5.07027566e-01 1.23005545e+00 2.62895584e-01 3.48686613e-02
-1.58786690e+00 -1.85696566e+00 -4.08359408e-01 1.25548348e-01
-8.89945850e-02 -5.10489106e-01 -6.43542588e-01 -4.05396819e-01
3.93401414e-01 1.06572485e+00 -3.54376167e-01 1.60054974e-02
-5.26002824e-01 -4.93730992e-01 1.33551925e-01 3.61647576e-01
-2.43475124e-01 -9.41232920e-01 -1.92882642e-01 -9.56809334e-03
-1.03347311e-02 -6.80255055e-01 -4.30493265e-01 -1.86927795e-01
-1.39827645e+00 -6.01270914e-01 -9.28445280e-01 -5.38184226e-01
1.09265804e+00 5.91844283e-02 1.47384405e+00 8.46521631e-02
-9.07011852e-02 2.72183478e-01 -1.32089585e-01 -6.72201693e-01
-6.37480974e-01 1.55048415e-01 -2.16775179e-01 -1.29957944e-01
4.15954113e-01 -8.09941053e-01 -5.64721525e-01 4.47984450e-02
-9.11921382e-01 6.26687054e-03 4.22004402e-01 6.43649697e-01
1.19975853e+00 -3.68185788e-01 2.23237738e-01 -4.93177235e-01
4.28704053e-01 -6.35407627e-01 -7.98371851e-01 -3.08611512e-01
-2.90652692e-01 2.28546798e-01 3.04560363e-01 -1.94553643e-01
-6.46676660e-01 3.22129518e-01 -3.25354844e-01 -9.66747582e-01
-4.14271951e-01 -2.28470378e-02 5.48848882e-02 1.64860606e-01
7.32366085e-01 3.00567269e-01 7.09869415e-02 -4.57372516e-01
5.91958106e-01 5.15954614e-01 3.07514250e-01 -9.36234713e-01
1.24889576e+00 8.15963209e-01 1.12771347e-01 -9.59499538e-01
-3.29462796e-01 -4.05127257e-01 -5.64925551e-01 -2.52368122e-01
6.84665263e-01 -7.86412954e-01 -3.15778136e-01 1.19932361e-01
-1.21320570e+00 -4.66224194e-01 -8.11357260e-01 2.71803051e-01
-8.38635504e-01 1.51015565e-01 -3.16480637e-01 -1.11084664e+00
-3.01915079e-01 -8.04008722e-01 1.77515936e+00 1.51763126e-01
-2.14008659e-01 -8.22546661e-01 4.60111141e-01 -1.38700545e-01
1.43555820e-01 2.31919572e-01 7.59391904e-01 -1.90156639e-01
-1.05239093e+00 -3.55535835e-01 -1.04442174e-02 2.32803389e-01
1.34312987e-01 3.67378980e-01 -9.68529582e-01 -1.66898504e-01
-3.09614569e-01 -1.78250586e-04 6.03441298e-01 7.30834723e-01
1.17246401e+00 -2.35219792e-01 -6.06164277e-01 7.30478942e-01
1.54362142e+00 -8.29402506e-02 9.26367521e-01 -1.10864326e-01
6.94432139e-01 1.81508690e-01 3.39812070e-01 5.90238690e-01
3.30905169e-01 6.90420985e-01 4.00876433e-01 1.67185798e-01
-7.73332193e-02 -7.36386716e-01 1.81960792e-03 7.43857622e-01
-2.54058570e-01 -1.96565464e-01 -1.23150182e+00 9.03999984e-01
-1.56779909e+00 -9.44900274e-01 -3.17009300e-01 2.47348261e+00
7.29487717e-01 1.92428175e-02 2.12975070e-01 -2.55236953e-01
6.50788546e-01 -8.84221420e-02 -4.35703695e-01 3.42205986e-02
8.24006498e-02 5.66151142e-01 4.69187438e-01 7.29796886e-01
-7.54480600e-01 9.63929415e-01 6.59818029e+00 9.46530402e-01
-9.00371432e-01 -1.07733078e-01 6.48494720e-01 -3.04573298e-01
-9.54402387e-01 8.13378990e-02 -8.53377759e-01 7.22918570e-01
8.26343715e-01 -3.82112354e-01 2.78962344e-01 9.39014018e-01
1.48992911e-01 -5.39680272e-02 -8.20203662e-01 1.03330016e+00
-1.66096568e-01 -1.54529750e+00 2.68770397e-01 4.98120278e-01
1.00711298e+00 3.75004053e-01 8.86108279e-02 -1.30229339e-01
5.89883208e-01 -1.02335274e+00 7.57161677e-01 8.29598904e-01
1.04301214e+00 -8.05446267e-01 2.61201918e-01 5.15725553e-01
-1.02471960e+00 5.85998774e-01 -6.38043404e-01 2.38449901e-01
4.05945003e-01 1.12018836e+00 -1.39205229e+00 2.89477378e-01
5.58314025e-01 4.76633549e-01 -1.59137979e-01 1.17816401e+00
-2.19266921e-01 7.75302470e-01 -8.27456117e-01 -1.10079005e-01
-7.55808800e-02 -7.83134580e-01 8.52463722e-01 1.00427282e+00
9.32597578e-01 7.73213282e-02 -4.89816293e-02 1.58063650e+00
-2.14524776e-01 -5.02251205e-04 -1.06035745e+00 1.30645260e-01
8.10319185e-01 1.10225260e+00 -7.41314709e-01 -3.38163972e-01
3.80613990e-02 8.28924358e-01 3.96336138e-01 2.62698680e-01
-5.91425776e-01 -2.00634152e-01 7.33147681e-01 6.70769036e-01
5.20887434e-01 -6.10589802e-01 -6.13302112e-01 -1.03675044e+00
1.60245165e-01 -2.03369692e-01 -2.69928426e-01 -1.11826217e+00
-1.37184322e+00 5.91485918e-01 1.75717235e-01 -1.47417307e+00
-4.24426824e-01 -3.72900903e-01 -7.51711309e-01 1.31664360e+00
-1.19104481e+00 -9.37825084e-01 -2.03408599e-01 1.00171708e-01
3.50680709e-01 -1.29705295e-03 8.48540723e-01 6.07309565e-02
1.25394896e-01 9.34019461e-02 1.82250604e-01 -1.59443066e-01
2.30283678e-01 -1.38963032e+00 1.10964417e+00 6.67702258e-01
2.69379586e-01 5.47302783e-01 6.23604894e-01 -1.06410134e+00
-1.10898709e+00 -1.16220844e+00 7.36348271e-01 -8.35983634e-01
2.17014074e-01 -6.09361112e-01 -1.10504341e+00 5.31530559e-01
-1.66614354e-01 2.34573218e-03 4.28415358e-01 -1.83707640e-01
1.67367235e-01 4.53512847e-01 -1.31148851e+00 6.55235112e-01
1.03311169e+00 -3.36391896e-01 -4.00874883e-01 5.15690207e-01
3.63629311e-01 -4.95769918e-01 -7.71529198e-01 2.28689983e-01
1.71847746e-01 -7.70056844e-01 1.21281993e+00 -3.13907117e-01
5.54173827e-01 -4.87793326e-01 -2.07630545e-01 -1.50563622e+00
-4.63654250e-01 -5.76728582e-01 -2.73422241e-01 9.38273728e-01
5.36827505e-01 -3.81055951e-01 1.41176665e+00 4.72234935e-01
-1.57733753e-01 -8.84875834e-01 -1.06023419e+00 -8.52972746e-01
4.18129355e-01 -4.89838809e-01 8.29188049e-01 4.97862101e-01
-3.64787251e-01 2.96540391e-02 8.79491568e-02 3.91836911e-01
9.01828945e-01 4.83303182e-02 1.01461267e+00 -1.34190619e+00
-2.16144279e-01 -3.80018473e-01 -3.47855538e-01 -1.37282765e+00
6.30735308e-02 -1.24737501e+00 1.85299203e-01 -1.86536658e+00
-1.00614810e-02 -9.00961637e-01 4.72554743e-01 4.75232638e-02
-1.09756447e-01 2.48280928e-01 2.03731522e-01 4.03014690e-01
1.52470469e-02 9.88061607e-01 1.36023092e+00 7.29111880e-02
-1.50190413e-01 1.64147705e-01 -5.03851652e-01 6.98257744e-01
8.45529974e-01 -6.74941123e-01 -3.02083075e-01 -3.96441817e-01
1.15343019e-01 1.00886121e-01 4.37025934e-01 -1.00933862e+00
-1.41097561e-01 -3.28591228e-01 4.95040625e-01 -9.09040153e-01
8.13279867e-01 -6.71634495e-01 4.75236118e-01 9.42867026e-02
1.67850599e-01 6.23385832e-02 2.01174796e-01 4.76616770e-01
8.67040604e-02 -4.53083545e-01 8.47552299e-01 -1.80421531e-01
1.08602732e-01 5.43184638e-01 -1.55367523e-01 6.46075308e-02
8.13311219e-01 -1.99319005e-01 5.28652966e-02 -4.48070496e-01
-5.75099945e-01 -1.36126772e-01 1.16495883e+00 1.29207969e-01
1.03717983e+00 -1.68693888e+00 -9.23171043e-01 4.71591979e-01
2.86974628e-02 7.59883046e-01 -1.08674336e-02 6.84043095e-02
-7.45654047e-01 1.67431664e-02 2.58112311e-01 -1.02848148e+00
-7.01332808e-01 2.22412214e-01 3.48453641e-01 1.28915280e-01
-7.41639495e-01 8.89698744e-01 1.43417299e-01 -6.29636467e-01
-4.67124581e-01 -3.80032480e-01 4.84427929e-01 -4.73744869e-01
1.65702522e-01 1.94548607e-01 6.64261356e-03 -5.63458204e-01
-2.51475036e-01 7.84051597e-01 4.02181357e-01 -5.45252860e-01
1.35409307e+00 4.83685911e-01 5.03361188e-02 4.15108114e-01
1.08844697e+00 3.79553348e-01 -1.45266485e+00 1.22338042e-01
-2.95680374e-01 -8.94774020e-01 -2.24221852e-02 -4.63763863e-01
-7.89050817e-01 7.84998596e-01 3.29524577e-01 1.33219540e-01
5.69491923e-01 5.07523000e-01 7.38887131e-01 -1.20758981e-01
5.95264614e-01 -6.00788772e-01 -7.02551305e-02 3.94897282e-01
1.03912532e+00 -9.55490828e-01 4.18338068e-02 -5.87704837e-01
-5.31251132e-01 8.86109114e-01 2.56799489e-01 -7.19311714e-01
9.23361301e-01 5.95662832e-01 -2.29191676e-01 -4.18955088e-01
-6.97259545e-01 1.44318473e-02 4.34736580e-01 8.42719555e-01
3.16414058e-01 2.69905061e-01 2.80646235e-01 6.28188401e-02
-4.95755017e-01 2.50297878e-02 3.95234376e-01 7.68465042e-01
-5.81179261e-01 -1.23296285e+00 -6.55262887e-01 7.49736667e-01
1.06696766e-02 -2.36549154e-01 -1.63732842e-01 2.95500040e-01
-2.36403376e-01 3.31292212e-01 7.05412745e-01 -1.26614600e-01
3.03280950e-01 9.63525176e-02 5.28480768e-01 -8.52369010e-01
1.19020514e-01 2.92750388e-01 -2.44631171e-01 -1.67189613e-01
-7.09332852e-03 -1.14740169e+00 -1.34970927e+00 -2.00681180e-01
-2.06196532e-01 2.64468282e-01 8.41550410e-01 2.54086971e-01
7.22867429e-01 1.59782991e-01 6.85751975e-01 -1.50642407e+00
-3.01962614e-01 -8.70817542e-01 -4.78718609e-01 3.21338624e-01
9.86192524e-02 -6.25218689e-01 -4.38498437e-01 1.17080383e-01]
|
[8.876984596252441, -3.664203643798828]
|
bc224593-0d4b-4988-8e60-1331b9beff60
|
learning-with-proper-partial-labels
|
2112.12303
| null |
https://arxiv.org/abs/2112.12303v2
|
https://arxiv.org/pdf/2112.12303v2.pdf
|
Learning with Proper Partial Labels
|
Partial-label learning is a kind of weakly-supervised learning with inexact labels, where for each training example, we are given a set of candidate labels instead of only one true label. Recently, various approaches on partial-label learning have been proposed under different generation models of candidate label sets. However, these methods require relatively strong distributional assumptions on the generation models. When the assumptions do not hold, the performance of the methods is not guaranteed theoretically. In this paper, we propose the notion of properness on partial labels. We show that this proper partial-label learning framework requires a weaker distributional assumption and includes many previous partial-label learning settings as special cases. We then derive a unified unbiased estimator of the classification risk. We prove that our estimator is risk-consistent, and we also establish an estimation error bound. Finally, we validate the effectiveness of our algorithm through experiments.
|
['Masashi Sugiyama', 'Jiaqi Lv', 'Zhenguo Wu']
|
2021-12-23
| null | null | null | null |
['partial-label-learning']
|
['methodology']
|
[ 1.58637270e-01 4.50639546e-01 -7.56471813e-01 -7.40747869e-01
-1.11779618e+00 -6.82248712e-01 3.59616786e-01 1.42683163e-01
-3.12524945e-01 9.59873676e-01 -1.44984081e-01 -2.52140462e-01
-3.12593400e-01 -6.61516249e-01 -7.74340451e-01 -7.80651987e-01
3.02783251e-01 5.17731428e-01 -6.13770485e-02 3.96706045e-01
-5.79570979e-02 8.33490565e-02 -1.40026772e+00 -1.33526266e-01
8.32187772e-01 1.02092993e+00 -3.11556578e-01 3.50787818e-01
6.98870867e-02 6.38138056e-01 -3.38790178e-01 -5.06358802e-01
5.09202898e-01 -5.67041039e-01 -9.86374140e-01 2.46611625e-01
5.59783876e-01 -3.22623909e-01 1.36671618e-01 1.50095260e+00
2.81780601e-01 -1.29142419e-01 1.06552172e+00 -1.56944931e+00
-4.01584715e-01 9.38680053e-01 -6.58603430e-01 -3.82733375e-01
1.10862933e-01 -3.86812180e-01 1.29213369e+00 -9.37899649e-01
3.48238736e-01 1.01027453e+00 8.83056283e-01 7.16849625e-01
-1.21931148e+00 -9.99060869e-01 2.05499664e-01 -2.12503672e-01
-1.43945634e+00 -2.06009984e-01 6.58317029e-01 -4.80688959e-01
-4.32265028e-02 1.69585243e-01 -2.81296000e-02 9.27823722e-01
-4.07784581e-01 7.90762067e-01 1.53268015e+00 -6.37115717e-01
3.43536168e-01 5.20579398e-01 7.51922846e-01 8.50540221e-01
5.72901011e-01 3.00022155e-01 -3.10419619e-01 -5.85453093e-01
3.90682310e-01 8.36723950e-03 -2.40143374e-01 -5.64895451e-01
-8.03518414e-01 9.63489473e-01 -5.27521456e-03 1.62967220e-01
1.02137767e-01 1.60601795e-01 2.83666253e-01 2.50447512e-01
8.21118712e-01 -3.52074462e-03 -5.28417766e-01 4.49972183e-01
-8.62674415e-01 6.22162558e-02 9.12098944e-01 1.21819568e+00
7.41671205e-01 -3.75019193e-01 -2.14870274e-01 8.88394475e-01
2.62261540e-01 5.93092024e-01 1.65981621e-01 -9.45713401e-01
2.04793572e-01 2.20248133e-01 4.68107104e-01 -5.54526567e-01
-5.16403377e-01 -6.26727819e-01 -6.67512596e-01 -1.06293261e-01
7.68430293e-01 -3.22356611e-01 -5.27312219e-01 2.32440925e+00
4.22274143e-01 3.65942627e-01 -1.23311661e-01 6.76308215e-01
1.78203106e-01 4.72440273e-01 1.89959660e-01 -8.61625254e-01
1.00566173e+00 -9.34993446e-01 -8.19694579e-01 2.70478547e-01
1.05548012e+00 -4.63819683e-01 1.00144827e+00 3.75370145e-01
-8.98239970e-01 -2.85882115e-01 -8.45695734e-01 1.41402587e-01
9.03071687e-02 2.74209172e-01 5.33697605e-01 1.01087940e+00
-7.05788791e-01 5.24727881e-01 -4.71379727e-01 -1.18336678e-01
2.29157820e-01 2.47940764e-01 -1.77330181e-01 -1.11788750e-01
-1.04719973e+00 5.76843441e-01 3.58018011e-01 -1.96264550e-01
-1.01399338e+00 -5.33492506e-01 -7.03919232e-01 -5.81934229e-02
6.58550203e-01 -5.12078881e-01 1.60792005e+00 -8.41346502e-01
-1.19015861e+00 9.83501136e-01 -1.55207247e-01 -1.68787256e-01
7.68641412e-01 -1.54164180e-01 -1.38928622e-01 -1.09231517e-01
3.96371722e-01 1.21796750e-01 5.86967707e-01 -1.44161916e+00
-1.01406670e+00 -2.96845645e-01 2.41755471e-01 8.77301097e-02
-4.11343187e-01 4.59008738e-02 6.41910359e-02 -5.03075123e-01
9.65627208e-02 -9.85953450e-01 -2.01870486e-01 9.81783047e-02
-6.54022634e-01 -6.39354646e-01 4.49869901e-01 -9.05153751e-02
1.14466119e+00 -2.16260338e+00 -4.05612230e-01 3.89124960e-01
5.75456470e-02 2.18173917e-02 2.03352213e-01 2.24572141e-02
-1.83893442e-01 5.03115714e-01 -4.49668080e-01 -5.97864449e-01
3.53437662e-01 9.23244208e-02 -5.55168867e-01 7.20896006e-01
-2.23103210e-01 6.40508473e-01 -1.02166116e+00 -6.72757506e-01
-2.12238505e-01 -9.76586267e-02 -3.77157569e-01 2.18878716e-01
-1.34193465e-01 1.04634121e-01 -5.57546020e-01 5.60122848e-01
7.82341063e-01 -4.38206226e-01 3.26497495e-01 9.15924013e-02
3.67753029e-01 3.55573893e-02 -1.39426875e+00 1.18545842e+00
-6.66805565e-01 -2.22751185e-01 8.04091543e-02 -1.12080157e+00
5.24252176e-01 4.43554312e-01 4.02538031e-01 1.28564343e-01
2.51148850e-01 5.26212275e-01 -5.58329105e-01 -2.41483390e-01
-6.30276138e-03 -9.54464674e-01 -3.22960615e-01 9.75046933e-01
6.56590164e-02 2.18860328e-01 1.11352727e-02 5.05458303e-02
6.70513213e-01 1.19129010e-01 5.13506830e-01 -3.64079088e-01
3.61359596e-01 -4.12060440e-01 8.19158316e-01 9.90012884e-01
-2.58034796e-01 4.87489074e-01 5.20242751e-01 -1.18712455e-01
-7.67348647e-01 -1.01592505e+00 -6.73412979e-01 1.23415112e+00
2.83122718e-01 -2.84053117e-01 -6.79923177e-01 -1.43641257e+00
-2.95445304e-02 6.62558317e-01 -6.58536434e-01 -7.50910193e-02
-7.82970861e-02 -1.01236582e+00 6.41930640e-01 4.09746975e-01
2.27989882e-01 -4.79666471e-01 1.49145806e-02 -5.53508885e-02
-2.09561467e-01 -9.03260112e-01 -6.65383339e-01 3.25775772e-01
-8.24367881e-01 -1.25270772e+00 -6.82524979e-01 -1.01283097e+00
9.38118994e-01 1.50052890e-01 8.57264161e-01 8.28390867e-02
4.21954036e-01 1.59187511e-01 -1.65670559e-01 -1.83198780e-01
-5.21504581e-01 -1.15056224e-02 4.53590125e-01 1.32702753e-01
2.11597919e-01 -3.60499531e-01 -1.06451638e-01 5.48770666e-01
-8.57010901e-01 -1.11407958e-01 3.41062665e-01 1.16900349e+00
9.37414289e-01 2.38663837e-01 1.02102339e+00 -1.62028515e+00
4.41335618e-01 -7.12121487e-01 -7.67452002e-01 5.73534012e-01
-8.75419855e-01 3.40380728e-01 6.90453410e-01 -5.36601126e-01
-1.24658465e+00 3.06009322e-01 1.23611428e-01 -2.07269266e-01
-2.13273913e-01 4.43509996e-01 -5.10794699e-01 6.88842386e-02
6.11260831e-01 -2.88026690e-01 -3.07665795e-01 -6.58202648e-01
3.78861040e-01 7.99059093e-01 3.41069549e-01 -9.43593562e-01
1.03434968e+00 5.26742578e-01 1.11711800e-01 -1.89823672e-01
-1.83475053e+00 -5.24599373e-01 -5.29097378e-01 -6.45566434e-02
3.45471770e-01 -7.35133588e-01 -4.64247197e-01 3.10705870e-01
-7.71948457e-01 -4.13449049e-01 -3.05549979e-01 5.36148131e-01
-6.68701530e-01 5.40127814e-01 -6.50148451e-01 -1.12064266e+00
6.59458758e-03 -9.10811782e-01 8.50536704e-01 -8.50559920e-02
-5.01224436e-02 -1.21131873e+00 6.82077482e-02 2.11344391e-01
-9.27227959e-02 1.07746899e-01 8.37044179e-01 -9.38386202e-01
-1.51147142e-01 -2.95089841e-01 -2.73702681e-01 5.64893723e-01
2.66062528e-01 -4.52528238e-01 -1.05551982e+00 -2.85538971e-01
2.84941256e-01 -7.35218108e-01 1.01409090e+00 3.69336516e-01
1.25147223e+00 -5.99147439e-01 -3.99275273e-01 4.57542896e-01
1.35706079e+00 -2.91298777e-01 -1.14143178e-01 -7.91159347e-02
5.78491569e-01 7.21913457e-01 9.93587971e-01 4.97884691e-01
1.97302878e-01 4.68859673e-01 1.26781151e-01 1.05984986e-01
2.22841501e-01 -4.48192209e-01 3.24474752e-01 6.53557122e-01
3.17017317e-01 -1.97125942e-01 -5.71855962e-01 4.98208672e-01
-1.88779783e+00 -7.52873600e-01 -1.65609881e-01 2.52049398e+00
1.38001621e+00 4.62209620e-02 2.48739228e-01 3.00250739e-01
1.15613568e+00 -3.06194901e-01 -5.91562688e-01 -4.18659821e-02
-1.32264853e-01 3.60702835e-02 6.45165861e-01 7.25279570e-01
-1.35150731e+00 5.99394917e-01 6.97664499e+00 1.03607774e+00
-5.85680962e-01 4.98531461e-01 8.11197877e-01 -4.04508933e-02
-4.05723065e-01 2.00670347e-01 -9.49966431e-01 4.54881310e-01
7.36389160e-01 -3.96886945e-01 -4.18959633e-02 1.21929693e+00
-3.24203074e-01 -5.34984050e-03 -1.38389301e+00 6.59057260e-01
3.67079414e-02 -6.50209129e-01 -3.50338489e-01 2.27314383e-01
1.11087668e+00 -5.11177480e-01 1.75701678e-01 3.55383575e-01
6.68083251e-01 -8.36199045e-01 7.76127338e-01 2.41039723e-01
1.20625353e+00 -1.01814651e+00 8.98088694e-01 8.68862629e-01
-8.25202525e-01 -2.60541946e-01 -4.81189758e-01 -1.01826549e-01
2.93956667e-01 1.04822600e+00 -4.32763100e-01 5.06616235e-01
5.41294627e-02 4.30487722e-01 -3.87280315e-01 9.49853301e-01
-5.50415397e-01 9.11421895e-01 -5.27580857e-01 1.57982320e-01
-4.89070229e-02 1.20911226e-01 1.70021072e-01 1.09338117e+00
3.53293449e-01 2.16577679e-01 5.52227497e-01 5.90141892e-01
-5.33804238e-01 2.80956835e-01 -5.68141878e-01 5.06609559e-01
6.27357423e-01 1.16170275e+00 -6.70341492e-01 -3.64087492e-01
-4.91664588e-01 6.16671443e-01 5.04472256e-01 1.83524072e-01
-8.02177072e-01 -1.82661787e-02 2.15007961e-01 -3.88614237e-02
-1.68802649e-01 3.67834032e-01 -2.53553420e-01 -1.23976099e+00
4.85766567e-02 -5.01727283e-01 8.08812320e-01 -2.72376239e-01
-1.76102924e+00 -1.93360094e-02 3.40523571e-01 -1.07892013e+00
-3.13343972e-01 -5.04294813e-01 -2.10154012e-01 6.58683717e-01
-1.47334826e+00 -1.05667663e+00 1.94802746e-01 3.81611943e-01
8.94189179e-02 1.50744885e-01 7.92885780e-01 1.73775092e-01
-5.25467932e-01 9.25222993e-01 2.91166812e-01 -8.43443722e-03
9.76814866e-01 -1.50506115e+00 -3.50410968e-01 7.75209129e-01
1.85905948e-01 4.61168736e-01 6.41641378e-01 -4.97316718e-01
-5.51127315e-01 -1.41742504e+00 1.08974278e+00 -5.34936965e-01
5.57572246e-01 -3.37100863e-01 -7.98354387e-01 8.90183687e-01
-2.44461268e-01 3.66717935e-01 1.08589709e+00 4.21735704e-01
-7.07194448e-01 -2.98309978e-02 -1.36363482e+00 1.97397470e-01
1.13907802e+00 -5.40640414e-01 -4.97563392e-01 6.43235803e-01
7.38577843e-01 -1.11412093e-01 -7.08074152e-01 5.64310908e-01
4.73655194e-01 -6.33345485e-01 5.62464416e-01 -6.91363275e-01
1.22404106e-01 -3.79312932e-01 -1.68168277e-01 -1.05334985e+00
-2.36596182e-01 -2.94371456e-01 -9.60319787e-02 1.44567633e+00
6.04842007e-01 -6.19861960e-01 7.31195152e-01 7.48565078e-01
1.48086220e-01 -6.87197804e-01 -7.96804905e-01 -1.19574964e+00
3.69974852e-01 -6.48551583e-01 4.49141413e-01 1.22594941e+00
3.45771134e-01 2.75559515e-01 -6.97613835e-01 2.56485403e-01
1.02801037e+00 3.23345482e-01 2.89901257e-01 -1.41371286e+00
-4.62467492e-01 -1.74860090e-01 6.16916642e-02 -9.87846792e-01
8.22010696e-01 -1.11250842e+00 2.79654682e-01 -1.14129603e+00
8.16682875e-01 -9.61264849e-01 -6.13482952e-01 6.98914289e-01
-4.29449439e-01 2.78625220e-01 -1.63012594e-01 2.91409492e-01
-8.68767619e-01 3.64973873e-01 9.43297327e-01 1.09470278e-01
1.75284788e-01 4.23436016e-01 -9.06867445e-01 9.69474733e-01
9.54296589e-01 -8.54052007e-01 -5.91714799e-01 3.64496857e-02
3.61223191e-01 -6.13796413e-02 1.89210698e-01 -6.53709769e-01
-2.38286369e-02 -2.54449397e-01 -1.01187050e-01 -3.30931634e-01
-1.71807081e-01 -7.19043672e-01 4.26228233e-02 4.05266881e-01
-1.02901220e+00 -7.10787237e-01 -5.09804785e-01 8.66293550e-01
1.31461903e-01 -7.08180904e-01 9.93947387e-01 1.01445936e-01
-7.77457058e-02 4.80711609e-01 7.70826414e-02 4.18205649e-01
1.05005372e+00 3.92790765e-01 -2.28491910e-02 -4.06574130e-01
-8.53553593e-01 3.26410174e-01 4.96922404e-01 -7.63623565e-02
1.62054330e-01 -1.60145998e+00 -7.33820260e-01 -2.25188911e-01
2.42326722e-01 6.14619814e-03 -5.51098771e-02 6.33475840e-01
2.26729140e-01 3.23518336e-01 5.03122330e-01 -2.87353337e-01
-1.07455301e+00 1.00458980e+00 1.69952497e-01 -3.70883137e-01
-1.70332342e-01 8.11697960e-01 6.76633596e-01 -6.13702059e-01
4.31857854e-01 -2.09413599e-02 -5.49923182e-02 9.86680947e-03
4.73800570e-01 3.55032593e-01 -2.04134658e-01 -6.39872074e-01
-1.16521291e-01 5.22777021e-01 2.22970620e-02 -3.07076305e-01
8.53992581e-01 -3.44283879e-01 2.09454969e-02 7.14967132e-01
1.27519488e+00 1.31780118e-01 -1.14733672e+00 -7.39476979e-01
2.95692712e-01 -4.20629323e-01 -1.45366341e-01 -6.49814308e-01
-9.95779932e-01 8.14531326e-01 3.45153242e-01 2.10379481e-01
1.05511951e+00 2.20143005e-01 4.44512963e-01 3.86015773e-01
7.70022452e-01 -1.15622985e+00 -1.39851108e-01 2.52547115e-01
3.39842856e-01 -1.28976166e+00 -4.57220078e-02 -7.26139009e-01
-4.08805430e-01 7.61052012e-01 5.30765295e-01 1.42048672e-01
9.27830875e-01 1.48357138e-01 -1.12178892e-01 2.08613023e-01
-7.48236477e-01 -2.27391452e-01 8.36709812e-02 5.32044172e-01
4.51649100e-01 2.78354138e-01 -7.25089967e-01 1.21343207e+00
-8.64061862e-02 -4.36507873e-02 5.01567602e-01 6.21912897e-01
-4.81986046e-01 -1.28171003e+00 -2.50565350e-01 4.47171748e-01
-8.01127493e-01 5.90015799e-02 -2.26379246e-01 5.40456295e-01
2.77928501e-01 1.05353403e+00 -4.68026668e-01 -1.85352385e-01
3.81734706e-02 4.02212560e-01 4.50198054e-01 -9.65465486e-01
2.25484017e-02 3.84294838e-02 1.09527223e-01 -1.33134753e-01
-5.89648187e-01 -6.57761097e-01 -1.16460121e+00 -1.24312721e-01
-1.01129735e+00 4.14578080e-01 2.31342509e-01 1.00073242e+00
-2.04606310e-01 -1.59270570e-01 1.02539754e+00 -2.53898233e-01
-1.16964674e+00 -8.58755410e-01 -1.09264374e+00 4.94324088e-01
2.17287809e-01 -8.31044912e-01 -8.27815533e-01 1.62958682e-01]
|
[9.190786361694336, 4.153933048248291]
|
a6a1e7bd-6759-4017-8675-7c90cbc005a7
|
auditory-neural-response-inspired-sound-event
|
2306.11427
| null |
https://arxiv.org/abs/2306.11427v1
|
https://arxiv.org/pdf/2306.11427v1.pdf
|
Auditory Neural Response Inspired Sound Event Detection Based on Spectro-temporal Receptive Field
|
Sound event detection (SED) is one of tasks to automate function by human auditory system which listens and understands auditory scenes. Therefore, we were inspired to make SED recognize sound events in the way human auditory system does. Spectro-temporal receptive field (STRF), an approach to describe the relationship between perceived sound at ear and transformed neural response in the auditory cortex, is closely related to recognition of sound. In this work, we utilized STRF as a kernel of the first convolutional layer in SED model to extract neural response from input sound to make SED model similar to human auditory system. In addition, we constructed two-branched SED model named as Two Branch STRFNet (TB-STRFNet) composed of STRF branch and baseline branch. While STRF branch extracts sound event information from auditory neural response, baseline branch extracts sound event information directly from the mel spectrogram just as conventional SED models do. TB-STRFNet outperformed the DCASE baseline by 4.3% in terms of threshold-independent macro F1 score, achieving 4th rank in DCASE Challenge 2023 Task 4b. We further improved TB-STRFNet by applying frequency dynamic convolution (FDYConv) which also leveraged domain knowledge on acoustics. As a result, two branch model applied with FDYConv on both branches outperformed the DCASE baseline by 6.2% in terms of the same metric.
|
['Yong-Hwa Park', 'Hyeonuk Nam', 'Deokki Min']
|
2023-06-20
| null | null | null | null |
['sound-event-detection']
|
['audio']
|
[-4.99312393e-02 -4.70529974e-01 8.22830021e-01 -1.47600859e-01
-6.62797570e-01 -6.45573437e-01 3.19257587e-01 6.52064458e-02
-7.22534478e-01 1.58730417e-01 4.47402626e-01 -1.92930773e-01
-1.40253862e-03 -6.52062595e-01 -7.16181278e-01 -3.20475638e-01
-1.14634432e-01 -6.00971043e-01 7.58191466e-01 -1.90419361e-01
1.41938284e-01 2.45404646e-01 -1.65163541e+00 8.05481195e-01
3.05253357e-01 1.47872722e+00 5.64949274e-01 9.22563374e-01
-7.13600218e-02 4.60504174e-01 -6.18893683e-01 -1.18348345e-01
3.44952419e-02 -4.86414641e-01 -6.20338798e-01 -6.23204231e-01
1.60395101e-01 -1.44131720e-01 -3.80999655e-01 9.55489218e-01
9.35947478e-01 3.32860529e-01 5.89451134e-01 -9.35035944e-01
-5.00168920e-01 7.74630845e-01 -1.08797304e-01 8.18707228e-01
4.03690279e-01 2.15895951e-01 1.01942384e+00 -1.38823891e+00
-1.74549714e-01 1.22810185e+00 6.64967775e-01 4.11518246e-01
-9.72215772e-01 -8.34950745e-01 2.52010673e-02 3.37787300e-01
-1.25097191e+00 -5.57151556e-01 5.50420046e-01 -5.13189316e-01
1.35410190e+00 4.48015243e-01 5.50358653e-01 9.26525116e-01
3.01983804e-01 5.76814294e-01 1.18642497e+00 -4.12381977e-01
2.21986875e-01 -1.68055534e-01 3.32338333e-01 9.88953710e-02
-4.68549997e-01 5.23828149e-01 -9.66569543e-01 -1.49152977e-02
6.56764925e-01 -3.45234573e-01 -4.54996318e-01 6.90368652e-01
-1.07571805e+00 4.43246901e-01 6.70013368e-01 2.79030859e-01
-3.96712929e-01 3.21161211e-01 5.36428213e-01 3.89408231e-01
7.24882446e-03 2.62045115e-01 -5.28144181e-01 -1.63507387e-01
-6.26090050e-01 2.09985510e-01 5.92223406e-01 2.67577589e-01
2.23316357e-01 5.86641192e-01 -5.10061145e-01 1.24739194e+00
3.22770357e-01 5.08733034e-01 8.45086157e-01 -6.49439692e-01
3.47262658e-02 -5.93195260e-02 -1.82373211e-01 -6.56828046e-01
-5.27398705e-01 -8.40815961e-01 -7.41558313e-01 1.19819157e-01
1.93524078e-01 1.30771503e-01 -1.10533285e+00 1.73681247e+00
1.29333492e-02 3.96221876e-01 -1.64178908e-02 9.94770586e-01
1.13569200e+00 7.89028704e-01 3.71650010e-02 1.42945260e-01
1.85825038e+00 -5.74230492e-01 -5.45479357e-01 -1.82588086e-01
7.07690641e-02 -1.00012350e+00 1.31386304e+00 5.91724515e-01
-9.63948667e-01 -1.12821245e+00 -9.40883756e-01 1.73393920e-01
-3.18300128e-01 1.29300386e-01 3.60399187e-01 5.11841655e-01
-1.01739931e+00 2.78061479e-01 -7.23170936e-01 -1.10921785e-01
2.37762004e-01 2.41368160e-01 -2.50391036e-01 5.41855872e-01
-1.55502260e+00 5.09056747e-01 5.77530265e-01 -5.03196195e-02
-1.28464234e+00 -9.24441397e-01 -6.12417400e-01 4.55624819e-01
1.85723677e-02 -4.71222401e-01 1.82630742e+00 -2.88767368e-01
-1.56795120e+00 5.22400975e-01 -4.72237244e-02 -9.18384552e-01
-1.75488770e-01 -4.12613928e-01 -9.43271101e-01 7.63095096e-02
-1.36484861e-01 6.27395391e-01 8.91943574e-01 -7.03539789e-01
-6.69984043e-01 3.65203135e-02 -3.35026115e-01 -1.76597401e-01
-1.94110692e-01 2.86087304e-01 9.01424736e-02 -1.00202799e+00
2.64103770e-01 -4.61318821e-01 2.13194460e-01 -3.62977594e-01
-2.48437047e-01 -4.67338681e-01 5.46235025e-01 -7.46586263e-01
1.55699146e+00 -2.81153679e+00 -6.39982462e-01 -3.66839245e-02
1.18251920e-01 4.69494998e-01 -3.84841830e-01 4.22637999e-01
-4.37481582e-01 -6.78240359e-02 -3.61803658e-02 5.92238046e-02
2.96695918e-01 -3.27308118e-01 -9.19126332e-01 -6.71280324e-02
4.26239222e-01 6.25439584e-01 -6.83424056e-01 9.13825780e-02
-8.48502219e-02 4.85829920e-01 -7.77575433e-01 2.40125626e-01
-4.24292274e-02 1.63835853e-01 -7.85617828e-02 2.45173693e-01
7.29693055e-01 3.18889827e-01 -7.02867806e-01 -3.28628004e-01
-3.01525831e-01 8.93271983e-01 -1.08327734e+00 1.46258712e+00
-5.76496899e-01 4.52026576e-01 -9.66999307e-02 -6.84009135e-01
1.18704593e+00 6.93913162e-01 1.09577172e-01 -9.99603152e-01
3.84160019e-02 3.64034921e-01 6.07684493e-01 -3.98403823e-01
-8.08559656e-02 -3.79499286e-01 -7.36366361e-02 3.70694131e-01
3.80935878e-01 -1.78002417e-01 -2.34576121e-01 4.68335301e-02
1.15771711e+00 -1.40057355e-01 2.04427764e-01 2.22362000e-02
5.98081112e-01 -8.13966334e-01 3.93743247e-01 8.96436334e-01
-2.98207134e-01 7.02100217e-01 9.11671892e-02 -2.77240902e-01
-5.20098507e-01 -1.60762107e+00 -3.50112498e-01 1.32578599e+00
-3.10933948e-01 -4.59070146e-01 -7.79492378e-01 -3.02249491e-01
-1.38944313e-01 9.13522840e-01 -2.55576491e-01 -4.41765964e-01
-5.09623766e-01 -3.89588356e-01 1.09491849e+00 6.43121123e-01
6.62510395e-01 -1.52802956e+00 -7.37639427e-01 6.03440344e-01
-1.84723794e-01 -1.16943598e+00 -7.03560650e-01 6.82844937e-01
-9.88362506e-02 -5.79429388e-01 -5.04729807e-01 -6.86081827e-01
-3.28400135e-01 1.18289202e-01 8.35684359e-01 -5.49364746e-01
-3.87713104e-01 1.94238782e-01 -4.94957060e-01 -9.67202306e-01
-1.13774367e-01 -2.67950982e-01 1.49887145e-01 2.73126215e-01
4.67978239e-01 -9.26789880e-01 -7.46892452e-01 2.92670012e-01
-9.87317085e-01 -2.31602982e-01 5.11242449e-01 6.64014995e-01
5.44117749e-01 9.34146494e-02 1.19658184e+00 1.31269097e-01
6.68600380e-01 -3.82306278e-01 -2.88841069e-01 -1.64001524e-01
-3.05656418e-02 -2.34800488e-01 7.38396406e-01 -6.62224829e-01
-8.79239976e-01 -2.24464182e-02 -7.25919664e-01 -2.60947943e-01
-3.10908109e-01 4.48843151e-01 3.48415598e-02 2.85121918e-01
8.95671010e-01 5.03108263e-01 -3.24228436e-01 -7.17272758e-01
-1.76005825e-01 1.07243586e+00 9.20952141e-01 -2.80963749e-01
2.82963783e-01 7.07284957e-02 -4.05617088e-01 -7.21869528e-01
-8.93578589e-01 -4.76283818e-01 -9.59803388e-02 -1.77711859e-01
9.74441707e-01 -8.54079545e-01 -9.22669947e-01 7.81128585e-01
-1.30036116e+00 -1.32353589e-01 -1.66624978e-01 1.04449213e+00
-4.13074702e-01 -8.67136568e-02 -6.58509314e-01 -1.05831265e+00
-4.43000078e-01 -7.35561192e-01 8.88873219e-01 7.53234001e-03
-2.30314910e-01 -3.46440375e-01 1.06984794e-01 -1.18999168e-01
7.64435053e-01 -1.31045520e-01 8.71955574e-01 -1.00722528e+00
5.84131889e-02 2.32601631e-02 -2.54909154e-02 8.57784510e-01
2.99623758e-02 -3.78027678e-01 -1.62466919e+00 -6.05492247e-03
4.82030928e-01 -1.15499020e-01 1.21171963e+00 5.45000732e-01
1.33899403e+00 -1.21214287e-02 2.27272421e-01 4.41614747e-01
1.01937997e+00 5.55573881e-01 7.32811391e-01 9.26100370e-03
5.41591905e-02 3.90039086e-01 2.92296082e-01 3.47922474e-01
4.86718491e-02 7.60537624e-01 3.33671689e-01 1.27746153e-03
-6.74583852e-01 -3.52004647e-01 8.13821435e-01 1.04669189e+00
2.36312181e-01 -1.14016972e-01 -8.95150602e-01 4.80812162e-01
-1.33289230e+00 -9.59216058e-01 -2.09126443e-01 2.12797260e+00
9.11562800e-01 4.94750023e-01 1.69257879e-01 6.76232576e-01
6.16738975e-01 -2.96932250e-01 -3.69750947e-01 -6.84720039e-01
-1.38664752e-01 9.52642262e-01 -2.05060244e-01 3.13909560e-01
-1.00605655e+00 7.50356674e-01 5.23125887e+00 1.06331742e+00
-1.34744751e+00 2.12730676e-01 1.46940365e-01 -1.64634600e-01
7.05074072e-02 -2.28276625e-01 -8.31857324e-01 4.81143296e-01
1.36685085e+00 6.57413602e-02 6.25119448e-01 3.11260104e-01
3.43913168e-01 -6.75506443e-02 -1.07976246e+00 1.01872146e+00
-3.39239597e-01 -8.41192782e-01 -1.44001869e-02 -1.48355708e-01
8.80622715e-02 1.61334455e-01 2.84345597e-01 6.19850278e-01
-1.97944250e-02 -1.20771706e+00 1.04905474e+00 5.85431457e-01
7.16227233e-01 -7.11979210e-01 4.97270942e-01 3.33676249e-01
-1.58862853e+00 -2.37874106e-01 -2.95818299e-01 -1.52125046e-01
1.14897802e-01 5.95579565e-01 -1.01995456e+00 3.41244370e-01
1.21764338e+00 2.27938756e-01 -4.24245983e-01 1.42233348e+00
6.65770620e-02 1.30806577e+00 -4.55326289e-01 9.88612473e-02
1.62948921e-01 5.84361792e-01 7.29871690e-01 1.45750988e+00
5.07660091e-01 1.05621098e-02 -1.17874838e-01 7.36322761e-01
1.17070779e-01 1.42652681e-02 -2.19938710e-01 -3.47211137e-02
6.95391774e-01 9.90714073e-01 -5.47407448e-01 4.67997305e-02
-1.95070490e-01 7.00328708e-01 -1.16594531e-01 4.44917828e-01
-9.17221725e-01 -8.57815385e-01 5.02526999e-01 1.71763226e-01
6.41126931e-01 -1.03352163e-02 1.48741364e-01 -5.00677407e-01
-1.47537783e-01 -7.74851143e-01 2.54895568e-01 -1.02702081e+00
-1.31722069e+00 1.08229959e+00 -1.52476832e-01 -1.32800794e+00
-5.04312366e-02 -7.45361567e-01 -7.10057318e-01 1.25622308e+00
-1.38470018e+00 -8.90830755e-01 -8.03063512e-02 9.28206742e-01
6.91048741e-01 -1.30662367e-01 1.05943859e+00 4.03733015e-01
-2.52996445e-01 8.10180366e-01 -3.91784936e-01 2.31908157e-01
7.46715248e-01 -1.05475938e+00 6.19672239e-01 6.90009058e-01
4.70709115e-01 5.93948126e-01 3.65328431e-01 -3.01866531e-01
-8.29675913e-01 -1.13817251e+00 7.10262835e-01 1.01928994e-01
7.09075987e-01 -5.83063841e-01 -1.09433055e+00 1.44401327e-01
1.22201286e-01 2.03952953e-01 8.03483725e-01 -2.60723382e-01
-6.30206048e-01 -4.30975139e-01 -8.02297294e-01 2.92093158e-01
7.48224974e-01 -9.24516737e-01 -8.72358859e-01 -1.46442324e-01
1.13493824e+00 -5.37390485e-02 -5.98993361e-01 5.40983975e-01
4.63863313e-01 -9.81473923e-01 1.04395223e+00 -3.96140516e-01
7.39688426e-02 -6.29464090e-01 -6.96786284e-01 -1.40589833e+00
-4.30661023e-01 -6.06257319e-01 -1.19333982e-01 1.27979279e+00
2.16492832e-01 -7.22134531e-01 -1.97833583e-01 -4.67153221e-01
-7.39030361e-01 -6.12116992e-01 -1.00720656e+00 -1.12266695e+00
7.01292306e-02 -1.07578456e+00 6.70799434e-01 3.04296166e-01
-3.00945640e-01 3.89783859e-01 1.21867843e-01 2.92352200e-01
9.93543789e-02 -3.78200322e-01 1.79600790e-01 -1.13505816e+00
-6.55790687e-01 -5.61924815e-01 -4.81140971e-01 -1.00606596e+00
-2.01826885e-01 -1.14497483e+00 2.37330854e-01 -1.12692142e+00
-1.64537922e-01 3.30627933e-02 -1.09559548e+00 5.32287955e-01
-3.82815041e-02 3.53808552e-01 1.99997842e-01 -2.04519451e-01
5.71982525e-02 4.70270544e-01 1.12141502e+00 1.35047838e-01
-1.81589097e-01 4.18061227e-01 -5.00591218e-01 7.81752527e-01
7.73131907e-01 -4.60181385e-01 -3.12920421e-01 -2.34189793e-01
7.47921467e-02 1.30297408e-01 9.03260946e-01 -1.46814132e+00
4.55144405e-01 2.55620211e-01 3.45560282e-01 -7.32038021e-01
4.93929714e-01 -3.97793233e-01 3.70095149e-02 5.50965130e-01
-3.91155839e-01 -1.60921022e-01 6.63892150e-01 3.54182810e-01
-6.32005632e-01 -1.65745318e-01 8.12033296e-01 5.87727055e-02
-6.55467629e-01 -8.95796642e-02 -6.28786802e-01 3.08149550e-02
4.91524130e-01 -1.46099910e-01 -1.60448372e-01 -1.38508022e-01
-9.07222450e-01 -4.17487860e-01 -8.19032013e-01 3.03929389e-01
9.47441518e-01 -1.38272476e+00 -9.18300331e-01 5.56633472e-01
-9.59091112e-02 -3.58012110e-01 4.56927538e-01 7.68979549e-01
-1.46940416e-02 4.38433081e-01 -2.00011164e-01 -6.21509135e-01
-9.39237535e-01 2.65848756e-01 6.63654149e-01 2.57214308e-01
-5.74643254e-01 1.44202220e+00 7.63823509e-01 -2.47981250e-01
6.05925798e-01 -7.48158872e-01 -3.37194651e-01 -7.55637810e-02
8.34179163e-01 2.27726862e-01 5.48382401e-01 -2.84767240e-01
-5.38397610e-01 2.69108504e-01 -5.63342422e-02 -6.66739941e-01
1.30793977e+00 5.16642556e-02 9.11415145e-02 6.30228996e-01
1.18903697e+00 1.35136575e-01 -9.41276193e-01 -2.76878297e-01
-3.12202293e-02 -3.27780447e-03 3.78031641e-01 -1.32224417e+00
-7.83587456e-01 1.32857537e+00 1.00739968e+00 4.15795326e-01
1.58722734e+00 -5.71759231e-02 1.03860760e+00 1.08026952e-01
1.58558294e-01 -6.72032118e-01 4.84567970e-01 8.83946359e-01
1.30984616e+00 -6.81319416e-01 -7.41393209e-01 -1.76137075e-01
-5.43574810e-01 1.21528196e+00 6.09380007e-01 -2.13286906e-01
1.10746157e+00 4.76527303e-01 -8.65096226e-02 6.21717714e-04
-8.78261864e-01 -2.94099748e-01 6.36344969e-01 4.31437135e-01
4.55516815e-01 -8.23614225e-02 6.51924461e-02 1.49257195e+00
-9.37902927e-01 -2.32895926e-01 7.37776831e-02 7.14476049e-01
-7.88421750e-01 -7.41737843e-01 -5.29428899e-01 3.52842718e-01
-7.10253000e-01 -6.37715697e-01 -1.66930005e-01 2.54119903e-01
3.29659224e-01 1.14745450e+00 2.92662144e-01 -7.91197538e-01
6.86978281e-01 3.08063239e-01 2.24771395e-01 -7.54372537e-01
-1.01774871e+00 7.37967670e-01 -3.41537535e-01 -3.53381604e-01
1.84763744e-01 -4.06031549e-01 -1.66732800e+00 3.11377466e-01
-3.09080064e-01 1.43623024e-01 6.23024046e-01 6.94118202e-01
2.70662755e-01 1.09212685e+00 7.78194547e-01 -4.42715466e-01
-7.89719641e-01 -1.34685695e+00 -7.05195189e-01 -3.95214371e-02
6.96964204e-01 -4.51244652e-01 -4.57637608e-01 1.70522422e-01]
|
[15.180243492126465, 5.223280429840088]
|
00091978-665e-412d-937b-6a662c78d342
|
learning-to-discriminate-information-for-1
|
2109.03393
| null |
https://arxiv.org/abs/2109.03393v3
|
https://arxiv.org/pdf/2109.03393v3.pdf
|
Learning to Discriminate Information for Online Action Detection: Analysis and Application
|
Online action detection, which aims to identify an ongoing action from a streaming video, is an important subject in real-world applications. For this task, previous methods use recurrent neural networks for modeling temporal relations in an input sequence. However, these methods overlook the fact that the input image sequence includes not only the action of interest but background and irrelevant actions. This would induce recurrent units to accumulate unnecessary information for encoding features on the action of interest. To overcome this problem, we propose a novel recurrent unit, named Information Discrimination Unit (IDU), which explicitly discriminates the information relevancy between an ongoing action and others to decide whether to accumulate the input information. This enables learning more discriminative representations for identifying an ongoing action. In this paper, we further present a new recurrent unit, called Information Integration Unit (IIU), for action anticipation. Our IIU exploits the outputs from IDU as pseudo action labels as well as RGB frames to learn enriched features of observed actions effectively. In experiments on TVSeries and THUMOS-14, the proposed methods outperform state-of-the-art methods by a significant margin in online action detection and action anticipation. Moreover, we demonstrate the effectiveness of the proposed units by conducting comprehensive ablation studies.
|
['Changick Kim', 'Chanho Jung', 'Yoonhyung Kim', 'Seokeon Choi', 'Jinyoung Moon', 'Hyunjun Eun', 'Sumin Lee']
|
2021-09-08
| null | null | null | null |
['action-anticipation', 'online-action-detection']
|
['computer-vision', 'computer-vision']
|
[ 8.47360253e-01 -1.72869280e-01 -4.83186334e-01 -1.84926957e-01
-5.46949744e-01 -2.74760872e-01 5.05517364e-01 -6.94071278e-02
-4.11705941e-01 4.59373474e-01 5.43689311e-01 1.36901259e-01
1.10138319e-01 -5.08205354e-01 -4.35854822e-01 -8.14510465e-01
-2.24886477e-01 -2.36815169e-01 5.17532408e-01 3.77727598e-02
3.45346004e-01 2.93039501e-01 -1.70574546e+00 6.85044587e-01
5.40716410e-01 1.26584148e+00 4.12195355e-01 4.78671253e-01
-1.06074261e-02 1.52823985e+00 -5.64412773e-01 3.85970771e-01
1.09068558e-01 -1.01145065e+00 -7.69559264e-01 3.40743512e-01
5.41603230e-02 -6.77101910e-01 -6.30355537e-01 7.64956057e-01
5.50355792e-01 4.29309368e-01 3.67056191e-01 -1.21654344e+00
-2.38354817e-01 6.12215817e-01 -3.87345821e-01 5.94935834e-01
6.21365368e-01 2.36740574e-01 1.08465302e+00 -6.92850173e-01
6.94721282e-01 1.04199338e+00 2.18828231e-01 5.41141927e-01
-5.90415955e-01 -5.79308629e-01 5.58483064e-01 7.47569501e-01
-1.06655645e+00 -5.23029506e-01 1.05577135e+00 -2.83576787e-01
8.69982302e-01 1.49748236e-01 9.04604495e-01 1.16266334e+00
1.46767244e-01 1.60959828e+00 8.05695415e-01 -2.29443550e-01
2.37179413e-01 -5.30290663e-01 -4.01985794e-02 5.09978473e-01
-3.51253510e-01 1.17534921e-02 -7.70265937e-01 2.39822119e-01
8.48200023e-01 3.38476032e-01 -3.39675069e-01 2.43535843e-02
-1.40198684e+00 4.18611854e-01 4.38560218e-01 4.55449909e-01
-6.55108809e-01 3.62650424e-01 6.24988317e-01 1.38559252e-01
3.55085462e-01 9.84651744e-02 -1.37092441e-01 -5.00251234e-01
-5.54170072e-01 -2.71351308e-01 1.80531502e-01 6.30310774e-01
6.52961135e-01 9.74848941e-02 -6.67808950e-01 6.20871007e-01
8.85933936e-02 2.96979416e-02 7.84278631e-01 -1.08909249e+00
5.36542177e-01 9.29041088e-01 1.26818880e-01 -8.78345549e-01
-3.29044551e-01 -2.66984850e-01 -6.59921408e-01 8.44485834e-02
4.81911078e-02 -2.22456828e-02 -9.01849985e-01 1.67466724e+00
2.87283063e-01 5.61412573e-01 2.40483299e-01 1.02310038e+00
7.21398413e-01 6.20481372e-01 7.53988624e-02 -4.66821820e-01
9.99170601e-01 -1.11035883e+00 -9.36038077e-01 -4.03677642e-01
7.29809821e-01 -4.76085842e-01 8.51422608e-01 2.07581773e-01
-9.45910037e-01 -7.84260571e-01 -9.75983024e-01 5.83431497e-02
5.83114428e-03 3.90951008e-01 7.89096534e-01 -4.75226678e-02
-6.34566605e-01 7.14152753e-01 -1.07262337e+00 -2.74745226e-01
4.26243961e-01 2.04915389e-01 -2.69158840e-01 1.30785391e-01
-1.35874748e+00 4.11270738e-01 5.24517477e-01 3.90799969e-01
-1.17359531e+00 -1.35884941e-01 -9.71779287e-01 -2.56790165e-02
6.98182702e-01 -1.20155856e-01 1.25573933e+00 -1.54318237e+00
-1.57199049e+00 4.99013990e-01 -2.01598406e-01 -6.42086923e-01
3.69504422e-01 -3.73558402e-01 -4.70835298e-01 4.79808927e-01
3.55770476e-02 5.87236881e-01 1.02604508e+00 -7.08703041e-01
-8.77212226e-01 -2.12053478e-01 3.90533924e-01 5.33105493e-01
-3.18790913e-01 9.52393338e-02 -5.06947815e-01 -8.40004981e-01
3.84779871e-01 -8.28793347e-01 -1.09443210e-01 1.00486711e-01
-1.78285927e-01 -3.78770560e-01 1.01528156e+00 -5.48287630e-01
1.40177691e+00 -2.36162066e+00 5.41464351e-02 -2.90047020e-01
2.16509053e-03 2.94155240e-01 -2.16292292e-01 3.49393070e-01
-1.23697273e-01 -2.78643906e-01 -1.80892944e-01 -2.28722543e-02
-3.14308405e-01 3.19321305e-01 -2.90193170e-01 4.57291037e-01
3.68608505e-01 7.89843380e-01 -1.05567050e+00 -5.96703172e-01
3.00509751e-01 3.58038068e-01 -2.74325550e-01 3.80573869e-01
-1.47320703e-01 8.06436181e-01 -7.02058554e-01 8.11771095e-01
1.18968613e-01 -1.80368215e-01 1.04565203e-01 -1.33377165e-01
-1.59613580e-01 3.88195872e-01 -1.20873809e+00 1.87893033e+00
-3.12217206e-01 7.11671472e-01 -4.53074545e-01 -1.00829899e+00
7.27590024e-01 4.18942124e-01 8.21818590e-01 -9.45075274e-01
1.26275301e-01 7.57253021e-02 1.96846396e-01 -7.03208566e-01
3.39134336e-01 2.08106503e-01 6.45928457e-02 6.78227127e-01
3.33100744e-03 6.02486670e-01 3.26279700e-01 1.52215287e-01
1.43108094e+00 5.82439005e-01 4.75670576e-01 3.75020713e-01
7.27942824e-01 -4.61002886e-01 1.03777862e+00 6.11972570e-01
-6.25798404e-01 6.88136995e-01 3.95863295e-01 -5.28564751e-01
-4.03833061e-01 -7.18650401e-01 3.58620733e-01 1.25160468e+00
4.29921031e-01 -4.39083070e-01 -4.41440225e-01 -9.64613974e-01
-3.93761039e-01 4.73618776e-01 -8.41655195e-01 -3.57688844e-01
-7.64338493e-01 -3.42233509e-01 2.76422888e-01 9.77072716e-01
9.32726681e-01 -1.57062685e+00 -1.18312407e+00 2.30857670e-01
-7.17643261e-01 -1.18609715e+00 -5.81685007e-01 4.49486859e-02
-9.49963987e-01 -1.13809741e+00 -6.12366676e-01 -4.87294286e-01
7.05910087e-01 5.32134950e-01 8.54741991e-01 -9.61284805e-03
-1.91257238e-01 5.19081891e-01 -7.93355882e-01 -3.63273025e-02
-1.74861744e-01 -3.42881322e-01 -1.03421256e-01 5.57328403e-01
2.65156090e-01 -4.31386620e-01 -9.57306445e-01 5.69056153e-01
-9.94879127e-01 2.92879313e-01 8.51931870e-01 6.86864197e-01
6.54350519e-01 8.24908316e-02 4.79037702e-01 -4.99406725e-01
1.64218515e-01 -3.09330881e-01 -7.80447870e-02 3.44482064e-01
2.50362828e-02 1.43291965e-01 5.02295136e-01 -5.14288843e-01
-1.15495634e+00 4.12504315e-01 2.09137946e-02 -5.73776662e-01
-1.96226332e-02 4.80038673e-01 -2.44529694e-01 2.11751625e-01
2.98651248e-01 4.98009831e-01 -3.24816316e-01 -2.12356955e-01
1.13623023e-01 5.70326209e-01 4.68047887e-01 -1.07637800e-01
9.37887579e-02 6.46097004e-01 -1.26994997e-01 -6.62340105e-01
-9.12905991e-01 -6.95676923e-01 -7.88376927e-01 -6.62315845e-01
8.70356143e-01 -9.56570327e-01 -6.99627876e-01 5.52813947e-01
-1.11810791e+00 -2.73889452e-01 -3.58821064e-01 7.71081626e-01
-7.39123285e-01 5.14886320e-01 -4.86202270e-01 -1.06909847e+00
-1.62769064e-01 -1.04063654e+00 1.12544537e+00 2.20924124e-01
-1.62549362e-01 -6.30240083e-01 -3.15664634e-02 2.82504320e-01
2.00167838e-02 3.03049386e-01 2.69999921e-01 -4.48586226e-01
-6.92070007e-01 -2.86285102e-01 -6.49318323e-02 4.17087317e-01
4.43134755e-01 -1.30087659e-01 -8.79924893e-01 -1.22079924e-01
1.68564394e-01 -5.24840117e-01 1.26376140e+00 1.86957896e-01
1.16109502e+00 -2.29479358e-01 -2.39727750e-01 3.31627727e-01
1.03566980e+00 6.04906261e-01 9.23377872e-01 1.44053146e-01
6.71934485e-01 4.48586047e-01 1.09602618e+00 7.44306505e-01
-1.09899614e-04 7.15899765e-01 6.29722595e-01 7.14149773e-02
-1.32206127e-01 -2.32180148e-01 9.59514260e-01 7.15124071e-01
-4.32676017e-01 -2.84525543e-01 -6.04915380e-01 4.40238774e-01
-2.18400788e+00 -1.13350272e+00 2.44087473e-01 2.20199466e+00
6.94156468e-01 2.55692989e-01 6.81109652e-02 3.34432393e-01
7.53465295e-01 5.65441191e-01 -8.84403765e-01 -8.90741684e-03
-1.42226174e-01 -1.00738317e-01 1.67702049e-01 8.84636864e-02
-1.42217052e+00 8.49803329e-01 5.48149347e+00 8.02747548e-01
-1.08226693e+00 8.75866860e-02 6.12988651e-01 1.22548062e-02
1.11335807e-01 4.69343318e-03 -6.06985807e-01 4.34399933e-01
6.27601266e-01 1.90811872e-01 1.12718873e-01 7.12956965e-01
4.93806750e-01 -4.21320677e-01 -1.07976830e+00 8.43387187e-01
2.68457294e-01 -9.53276694e-01 6.04445003e-02 -3.57077271e-01
5.13312221e-01 -2.26601228e-01 -1.69909745e-01 3.01555693e-01
-1.17299054e-02 -6.42119288e-01 5.21356761e-01 6.93667948e-01
5.64202726e-01 -5.64157784e-01 6.10961497e-01 2.00750396e-01
-1.59325612e+00 -4.35995579e-01 -1.17916942e-01 -3.63703042e-01
1.73208266e-01 2.98707664e-01 -5.76189995e-01 5.82393825e-01
4.42269206e-01 1.46703255e+00 -6.47144616e-01 9.33783531e-01
-6.74648702e-01 5.05931497e-01 -8.98980498e-02 1.09277241e-01
4.27121252e-01 2.00178381e-02 4.26415026e-01 8.87785375e-01
2.58591831e-01 2.36540154e-01 3.14700902e-01 3.97016585e-01
1.75559863e-01 -8.99044052e-02 -5.34511089e-01 -3.23554784e-01
1.03531070e-02 1.00621617e+00 -8.81005824e-01 -3.91765296e-01
-3.81173313e-01 1.40934694e+00 1.45792052e-01 4.82122064e-01
-9.30033147e-01 -3.34195882e-01 5.02284169e-01 -1.91135556e-01
3.30053300e-01 -2.68564641e-01 2.91734308e-01 -1.24239004e+00
2.21902996e-01 -6.50113523e-01 6.16097331e-01 -8.33677113e-01
-6.53239787e-01 5.11587262e-01 -1.65936038e-01 -1.99832261e+00
-5.69917440e-01 -2.45143920e-01 -7.11642802e-01 3.13433200e-01
-1.19030106e+00 -1.02620089e+00 -5.34446657e-01 5.28543532e-01
1.00793624e+00 3.21822707e-03 4.43830103e-01 2.52152890e-01
-7.27908432e-01 2.63005018e-01 -1.47557363e-01 2.96377718e-01
5.14577448e-01 -9.24266160e-01 1.64022624e-01 1.05619156e+00
4.75137174e-01 2.48504341e-01 4.02596176e-01 -8.01927030e-01
-1.22698832e+00 -1.04154694e+00 6.40292466e-01 -7.65716732e-02
5.95751047e-01 -1.97413981e-01 -7.89373100e-01 7.31162727e-01
7.01532364e-02 2.05893680e-01 6.16270602e-01 -3.06168288e-01
-8.78450572e-02 -1.40162811e-01 -6.45386696e-01 6.29827976e-01
1.42224371e+00 -6.26274049e-01 -5.27886629e-01 3.30614567e-01
7.38718271e-01 -2.35497043e-01 -4.24669951e-01 5.55622816e-01
8.05068552e-01 -1.17941797e+00 8.66189003e-01 -5.57988346e-01
6.56292975e-01 -4.19811696e-01 -1.37701362e-01 -8.65154505e-01
-1.95252612e-01 -6.29808307e-01 -3.86985004e-01 1.05377150e+00
-1.26599923e-01 -2.54153430e-01 7.32274115e-01 2.36472785e-01
-1.81184039e-01 -8.50957453e-01 -9.44516480e-01 -7.41269588e-01
-8.25586081e-01 -5.42371392e-01 2.07462862e-01 6.86097443e-01
4.63501737e-02 2.26310208e-01 -6.02436423e-01 -1.33894742e-01
1.71777695e-01 2.86734819e-01 5.42988420e-01 -7.73860216e-01
-4.11346257e-01 -2.83910364e-01 -7.87262201e-01 -1.38525808e+00
2.41775900e-01 -5.40469468e-01 3.01403075e-01 -1.53018200e+00
1.33993074e-01 -5.55746369e-02 -8.69658709e-01 7.16394842e-01
-2.99438536e-01 2.55510688e-01 3.05574626e-01 4.16507751e-01
-1.09891939e+00 8.35887551e-01 1.32254267e+00 -2.47937053e-01
-3.36404055e-01 2.01269120e-01 -1.93860978e-01 8.31060112e-01
7.24980891e-01 -1.78246036e-01 -6.54300153e-01 -2.87815511e-01
1.93049386e-02 1.99975744e-01 3.73571277e-01 -1.33581078e+00
2.20240995e-01 -2.32083365e-01 3.22178543e-01 -8.49022150e-01
4.09339666e-01 -6.68294132e-01 -1.12073295e-01 5.40867507e-01
-6.91315711e-01 -1.31103903e-01 -3.91672887e-02 7.74235487e-01
-5.51680744e-01 -2.17049345e-01 3.74100208e-01 -1.95680007e-01
-1.40599358e+00 3.18100393e-01 -5.02425730e-01 -1.33554697e-01
1.10886180e+00 -3.28027278e-01 -1.32597789e-01 -5.26620090e-01
-8.34043145e-01 1.98401168e-01 3.38886417e-02 4.95538265e-01
9.68576550e-01 -1.46386576e+00 -5.47569633e-01 2.59871781e-01
2.48132974e-01 -3.08790326e-01 5.98461688e-01 1.08365560e+00
-2.09282130e-01 1.82795063e-01 -3.67485762e-01 -5.24778605e-01
-1.31952989e+00 4.90962446e-01 1.18269131e-01 -2.81952441e-01
-8.42158675e-01 6.82771742e-01 4.70723659e-01 3.49587023e-01
3.81232172e-01 -4.10249323e-01 -6.09427094e-01 4.07303691e-01
7.40649045e-01 3.80822062e-01 -2.95246124e-01 -7.35060632e-01
-3.78200889e-01 3.97790283e-01 6.83696521e-03 1.32723199e-02
9.65427935e-01 -2.26148173e-01 6.51488081e-02 8.28024566e-01
1.18130672e+00 -4.68705803e-01 -1.69354057e+00 -3.31601411e-01
3.36772809e-03 -5.77247679e-01 -8.32378715e-02 -5.69773674e-01
-1.24785328e+00 1.00422096e+00 6.67908430e-01 -1.49085268e-01
1.62920785e+00 -1.43122926e-01 9.07639980e-01 4.79018480e-01
2.72598892e-01 -1.31249690e+00 6.18911564e-01 4.53723609e-01
1.06937099e+00 -1.40978742e+00 9.65195447e-02 -2.36139849e-01
-8.53293061e-01 1.18511546e+00 6.52824402e-01 -8.53047334e-03
2.87464380e-01 -1.80309251e-01 -1.13996267e-02 9.18615423e-03
-8.02840889e-01 -6.53879702e-01 1.43923834e-01 2.43562579e-01
2.04762802e-01 -1.98376969e-01 -3.75814646e-01 3.86396557e-01
6.66650176e-01 2.25659758e-01 3.79280150e-01 1.32232404e+00
-4.16243613e-01 -8.30002785e-01 -4.63905223e-02 3.50203484e-01
-3.32318902e-01 3.97161059e-02 -4.97698992e-01 4.91600394e-01
1.00646488e-01 8.23054731e-01 2.53952533e-01 -6.80915713e-01
8.34405050e-03 -8.03055540e-02 3.03051114e-01 -4.52712178e-01
-4.46561426e-01 2.78937697e-01 -2.43796944e-03 -9.72920120e-01
-1.05925727e+00 -7.55880594e-01 -1.48655391e+00 5.33682227e-01
-1.80978388e-01 -1.23374730e-01 1.75952017e-01 1.08214259e+00
3.86925399e-01 8.53131950e-01 9.91841435e-01 -9.02581334e-01
-1.43215030e-01 -9.51012611e-01 -4.91693556e-01 4.57706630e-01
2.84455568e-01 -9.21003699e-01 -3.63503605e-01 1.88712433e-01]
|
[8.367554664611816, 0.4986291527748108]
|
a652736a-0166-4721-ac00-58adaa409413
|
shape-constraint-recurrent-flow-for-6d-object-1
|
2306.13266
| null |
https://arxiv.org/abs/2306.13266v1
|
https://arxiv.org/pdf/2306.13266v1.pdf
|
Shape-Constraint Recurrent Flow for 6D Object Pose Estimation
|
Most recent 6D object pose methods use 2D optical flow to refine their results. However, the general optical flow methods typically do not consider the target's 3D shape information during matching, making them less effective in 6D object pose estimation. In this work, we propose a shape-constraint recurrent matching framework for 6D object pose estimation. We first compute a pose-induced flow based on the displacement of 2D reprojection between the initial pose and the currently estimated pose, which embeds the target's 3D shape implicitly. Then we use this pose-induced flow to construct the correlation map for the following matching iterations, which reduces the matching space significantly and is much easier to learn. Furthermore, we use networks to learn the object pose based on the current estimated flow, which facilitates the computation of the pose-induced flow for the next iteration and yields an end-to-end system for object pose. Finally, we optimize the optical flow and object pose simultaneously in a recurrent manner. We evaluate our method on three challenging 6D object pose datasets and show that it outperforms the state of the art significantly in both accuracy and efficiency.
|
['Yinlin Hu', 'Jiaojiao Li', 'Rui Song', 'Yang Hai']
|
2023-06-23
|
shape-constraint-recurrent-flow-for-6d-object
|
http://openaccess.thecvf.com//content/CVPR2023/html/Hai_Shape-Constraint_Recurrent_Flow_for_6D_Object_Pose_Estimation_CVPR_2023_paper.html
|
http://openaccess.thecvf.com//content/CVPR2023/papers/Hai_Shape-Constraint_Recurrent_Flow_for_6D_Object_Pose_Estimation_CVPR_2023_paper.pdf
|
cvpr-2023-1
|
['pose-estimation', 'optical-flow-estimation', '6d-pose-estimation']
|
['computer-vision', 'computer-vision', 'computer-vision']
|
[-2.28806451e-01 -3.65841836e-01 -1.57477230e-01 -3.67264062e-01
-3.68358940e-01 -5.50749242e-01 3.10844153e-01 5.34734279e-02
-3.42460603e-01 5.53019606e-02 1.65148869e-01 2.27776036e-01
-2.92619094e-02 -5.36602497e-01 -7.23549545e-01 -3.94279569e-01
1.17799580e-01 7.78470397e-01 2.16503993e-01 1.48496419e-01
4.50438231e-01 8.42133641e-01 -1.21350682e+00 -3.59984428e-01
7.26406991e-01 1.21182430e+00 2.63613939e-01 4.90509242e-01
9.47450027e-02 4.22737062e-01 -3.50027919e-01 -2.22760379e-01
5.95893383e-01 -1.49163112e-01 -6.09547019e-01 5.27737260e-01
1.03361142e+00 -6.00642323e-01 -5.51936328e-01 1.03931129e+00
5.99489570e-01 2.41773903e-01 3.28244776e-01 -1.05131483e+00
-3.36841941e-01 -6.91190511e-02 -6.14540517e-01 6.42278194e-02
4.97394860e-01 3.22136045e-01 8.49279344e-01 -1.17808580e+00
8.36638391e-01 1.42000997e+00 4.85775203e-01 4.30758893e-01
-1.18259215e+00 -6.90760791e-01 2.94617414e-01 1.04282666e-02
-1.28279793e+00 -3.20446968e-01 1.09476876e+00 -5.44249296e-01
6.99106812e-01 -2.76815742e-01 1.05544269e+00 4.45536047e-01
-1.60947070e-01 8.48326325e-01 4.61719066e-01 -9.26500484e-02
-1.03385955e-01 -3.07254970e-01 -2.58183807e-01 8.39606106e-01
9.83755887e-02 1.46553203e-01 -5.78670979e-01 -7.24456757e-02
1.16837561e+00 2.87399083e-01 -2.97439188e-01 -9.98608410e-01
-1.40922260e+00 6.85533166e-01 8.33601236e-01 -4.71677519e-02
-3.77239883e-01 2.81856745e-01 1.31101087e-01 -3.25730518e-02
4.87575918e-01 4.22219455e-01 -2.19407335e-01 -1.47432819e-01
-4.62982804e-01 4.91762608e-01 6.96177781e-01 9.31007922e-01
9.87649560e-01 -3.30745935e-01 -2.35703439e-01 7.14741111e-01
7.87939250e-01 4.60719526e-01 -8.75125676e-02 -1.28849506e+00
7.26739347e-01 9.02014136e-01 4.56758946e-01 -1.23483813e+00
-2.90441573e-01 -5.95114529e-01 -4.27506387e-01 -2.36480823e-03
6.16458893e-01 1.71264276e-01 -8.97279382e-01 1.67313981e+00
7.94770062e-01 6.38188183e-01 -2.00366706e-01 1.42259765e+00
5.23836255e-01 4.33708161e-01 -3.33707869e-01 -1.40110552e-02
9.04766083e-01 -1.14771521e+00 -4.89547014e-01 -2.05334499e-01
5.71710467e-01 -9.45052028e-01 5.75933278e-01 -1.71365496e-02
-1.21160650e+00 -6.21002018e-01 -1.00419116e+00 -5.94198368e-02
2.32205182e-01 1.84357464e-01 4.02856529e-01 1.91788360e-01
-6.46171331e-01 7.27204084e-01 -9.90928531e-01 -9.04357061e-03
4.94229138e-01 4.82878745e-01 -2.51857579e-01 -9.65723544e-02
-8.23304296e-01 1.03996217e+00 2.78024882e-01 5.60180664e-01
-7.30939507e-01 -9.75534916e-01 -9.83121037e-01 -1.52113855e-01
5.60617924e-01 -9.23099399e-01 1.11892879e+00 -2.08288699e-01
-1.56159794e+00 9.46544766e-01 -2.38358587e-01 -1.43427387e-01
7.22590625e-01 -4.73933518e-01 3.81047010e-01 1.63446754e-01
7.67820999e-02 8.44150424e-01 7.85296977e-01 -1.05141711e+00
-4.63750273e-01 -7.07203567e-01 1.75913617e-01 5.41984797e-01
-8.26247856e-02 -4.37201083e-01 -1.02534986e+00 -3.33560586e-01
6.36051118e-01 -1.15019727e+00 -2.35493019e-01 8.77709031e-01
-1.19595282e-01 -4.31072652e-01 1.03757906e+00 -3.32069814e-01
8.59665155e-01 -2.09762049e+00 5.21085501e-01 1.56581223e-01
3.14611644e-01 2.65991926e-01 -2.82374829e-01 -9.98942927e-02
1.52046397e-01 -2.86354929e-01 -5.67303859e-02 -6.90028489e-01
-1.93088070e-01 1.98072963e-03 -2.20952686e-02 7.43556201e-01
4.42218870e-01 1.26100445e+00 -1.15048885e+00 -3.47150147e-01
4.87340569e-01 7.50284374e-01 -8.77228200e-01 5.66613972e-01
-1.34318843e-01 6.31344616e-01 -6.87098742e-01 3.73359382e-01
8.17441583e-01 -2.40295857e-01 -1.25438541e-01 -7.21371889e-01
-9.14985836e-02 3.26615721e-01 -1.33424747e+00 2.29143262e+00
-4.46196109e-01 3.88669580e-01 -1.01217605e-01 -6.04912996e-01
1.21047139e+00 2.21983977e-02 8.34954381e-01 -3.07166368e-01
2.64881790e-01 2.83733100e-01 -1.74701020e-01 -2.16546848e-01
1.03678159e-01 1.66893110e-01 2.43257359e-01 3.98764044e-01
-6.20573722e-02 -5.12331605e-01 -2.89867865e-03 -3.76115590e-02
6.53971732e-01 4.49877292e-01 -1.67258140e-02 1.73455104e-01
7.24170685e-01 -3.78651291e-01 6.20731354e-01 2.96454400e-01
-2.73499668e-01 6.74724221e-01 3.35067779e-01 -6.40644491e-01
-1.06068742e+00 -1.11720514e+00 -9.25727859e-02 1.99187055e-01
7.31833816e-01 -2.47231945e-01 -3.96961480e-01 -8.23578954e-01
5.13663173e-01 9.30980407e-03 -3.60891849e-01 -2.35675156e-01
-9.57216740e-01 -2.02037632e-01 -1.00993730e-01 5.01969576e-01
5.22058547e-01 -7.87281156e-01 -6.80205882e-01 3.98509085e-01
-3.57472003e-01 -1.24348974e+00 -1.08463883e+00 -3.38746756e-01
-1.16672122e+00 -9.70511138e-01 -7.48055100e-01 -7.64600098e-01
8.14121127e-01 3.68760705e-01 6.83659077e-01 2.85643160e-01
-2.76351482e-01 9.89553332e-02 -3.55453789e-02 2.39014365e-02
-2.03561168e-02 -7.27624670e-02 -4.87848818e-02 1.84078351e-01
1.54371217e-01 -5.03563046e-01 -9.55151498e-01 5.90948820e-01
-4.89163607e-01 -5.61754964e-02 3.63546044e-01 6.98066950e-01
6.04363918e-01 -4.54575568e-01 1.66127220e-01 -3.59936923e-01
8.04092810e-02 1.18384406e-01 -8.83437335e-01 2.59706602e-02
-3.34369659e-01 2.72292733e-01 2.34926753e-02 -6.47863805e-01
-7.87056506e-01 5.10498941e-01 -5.54320253e-02 -1.10772252e+00
2.03794643e-01 2.85830408e-01 -1.02851093e-01 -3.96285653e-01
2.88168579e-01 -1.52157456e-01 3.38598639e-01 -6.46229208e-01
1.74835742e-01 3.16753447e-01 6.24576688e-01 -5.13615310e-01
1.14220440e+00 4.78607833e-01 1.59929290e-01 -2.85578132e-01
-1.20262039e+00 -7.25583673e-01 -8.76447737e-01 -4.47565526e-01
7.86122918e-01 -8.39201808e-01 -1.21226418e+00 7.05873430e-01
-1.49345100e+00 2.51882058e-02 -1.81498140e-01 8.09900463e-01
-6.96411490e-01 3.65771383e-01 -4.60089654e-01 -5.73383868e-01
-3.48708034e-01 -1.34086025e+00 1.54884422e+00 1.20681956e-01
-4.23403457e-02 -6.85243487e-01 1.22900316e-02 3.11239362e-01
1.32950217e-01 3.69785041e-01 4.75366771e-01 1.34681240e-01
-1.12012565e+00 -3.63820583e-01 -4.54939425e-01 3.42168286e-02
2.12105259e-01 -2.80854136e-01 -5.58947802e-01 -4.69518334e-01
5.68492664e-03 -1.05691381e-01 6.66957736e-01 2.83426166e-01
9.22276318e-01 3.40451375e-02 -2.19583869e-01 8.92406881e-01
1.35434163e+00 3.15911397e-02 3.88429314e-01 1.02885388e-01
1.02313316e+00 6.17593229e-01 8.93843830e-01 4.99105185e-01
3.31952363e-01 1.05219531e+00 7.63938189e-01 2.48282596e-01
-1.53862387e-01 -5.52686453e-01 1.08015515e-01 8.71138096e-01
-4.95211557e-02 4.46768887e-02 -6.52236402e-01 3.81979436e-01
-1.94480038e+00 -4.35285360e-01 9.61574018e-02 2.36014533e+00
6.17042303e-01 2.38883466e-01 -6.47829622e-02 -2.64111638e-01
7.21965194e-01 2.15386197e-01 -8.70439589e-01 3.31334531e-01
3.79572898e-01 -3.48591357e-02 1.71637550e-01 6.94253445e-01
-9.38190222e-01 1.02105165e+00 4.91967678e+00 2.66497999e-01
-1.12055075e+00 -1.82476297e-01 5.50267696e-02 -1.29624218e-01
-1.80423796e-01 1.15031622e-01 -9.08630252e-01 2.35185683e-01
1.01963848e-01 7.61524867e-03 2.20720619e-01 6.19108796e-01
7.49899298e-02 -7.91909620e-02 -1.37019622e+00 1.21038747e+00
6.52859285e-02 -1.30991316e+00 4.83842716e-02 5.04487641e-02
6.56188548e-01 -1.58006310e-01 -9.74014923e-02 -1.77888811e-01
-2.75303751e-01 -4.99216527e-01 8.11622858e-01 4.49616313e-01
4.70678121e-01 -7.62292504e-01 5.75096250e-01 2.80090868e-01
-1.43269825e+00 6.46925122e-02 -3.29696208e-01 3.06510143e-02
4.11556453e-01 3.69528860e-01 -7.32095361e-01 3.43024760e-01
6.24086440e-01 1.19563317e+00 -2.87515372e-01 1.41226912e+00
-2.34313309e-01 -1.71129748e-01 -4.55144763e-01 -6.94104359e-02
1.62093952e-01 -3.06315005e-01 9.83592868e-01 5.49653292e-01
2.58433282e-01 -9.88282487e-02 5.60723543e-01 9.96607900e-01
-7.69008920e-02 -1.84037551e-01 -2.37249374e-01 1.78204015e-01
5.29102325e-01 1.11102653e+00 -7.10447073e-01 -7.80786499e-02
-1.22671619e-01 9.78321612e-01 4.38284248e-01 2.33403891e-01
-5.73591113e-01 -1.77331820e-01 8.99734616e-01 -4.86098751e-02
3.58848780e-01 -3.97327632e-01 5.78944944e-02 -1.29264522e+00
3.80548269e-01 -3.73028487e-01 1.07614845e-01 -7.97765911e-01
-1.21594727e+00 2.37434655e-01 -1.78892836e-01 -1.47355342e+00
-2.30184406e-01 -4.51802075e-01 -2.11793676e-01 9.36627090e-01
-1.49358845e+00 -8.51240575e-01 -5.83898365e-01 2.98584282e-01
3.51428002e-01 3.75426382e-01 3.90553474e-01 3.92376065e-01
-2.44166225e-01 4.30180281e-01 -6.64743781e-01 3.41780573e-01
8.75011861e-01 -8.81754637e-01 4.98249978e-01 5.40240228e-01
1.03781391e-02 5.36489725e-01 4.78765398e-01 -6.14387512e-01
-1.69262064e+00 -1.01672029e+00 9.77526605e-01 -6.78538203e-01
4.88336176e-01 -4.54676449e-01 -8.69220734e-01 4.68827248e-01
-4.88405734e-01 4.57973868e-01 -1.80668816e-01 -2.26724014e-01
-4.93286580e-01 -4.01082873e-01 -9.81391668e-01 4.95525032e-01
1.45690513e+00 -5.25142968e-01 -5.04096746e-01 1.64062798e-01
9.03005719e-01 -9.71469998e-01 -9.82106388e-01 6.36196494e-01
7.57942319e-01 -6.16110206e-01 1.08362198e+00 -3.52756143e-01
1.21128611e-01 -7.69389570e-01 1.14810608e-01 -1.06657720e+00
-2.62549698e-01 -6.70971394e-01 -2.99148977e-01 8.68761122e-01
1.58204868e-01 -4.54282433e-01 1.09507596e+00 4.33554351e-01
-4.43582870e-02 -9.03477609e-01 -8.92240942e-01 -5.08078933e-01
-3.86351019e-01 -3.42031032e-01 5.88297546e-01 6.40222013e-01
-5.89797318e-01 2.96407163e-01 -2.91615516e-01 2.22067341e-01
9.37829971e-01 6.08119786e-01 1.03866529e+00 -1.31847346e+00
-1.21040970e-01 -3.71592313e-01 -8.92269790e-01 -1.86488342e+00
2.32739031e-01 -8.36087525e-01 2.98374057e-01 -1.19512188e+00
2.98081577e-01 -4.99254674e-01 -7.87957758e-02 2.42283762e-01
-3.08483422e-01 3.65601629e-01 4.74823207e-01 2.25817084e-01
-3.79501551e-01 6.87930167e-01 1.98907185e+00 -1.30919367e-01
-5.34890711e-01 9.14575458e-02 -2.45461494e-01 3.95501554e-01
3.32122236e-01 -5.35797715e-01 -1.59946904e-01 -5.12600482e-01
-5.90211600e-02 1.54020205e-01 6.18101478e-01 -6.81949615e-01
3.43755364e-01 4.18943055e-02 5.04790604e-01 -1.05131984e+00
5.75372756e-01 -9.13500965e-01 5.34286238e-02 6.08311117e-01
-3.21851015e-01 -7.79519603e-02 -1.50367513e-01 6.24320447e-01
-1.33226767e-01 -1.52958930e-01 7.78050244e-01 7.92325959e-02
-4.49031264e-01 1.10559738e+00 4.84690756e-01 3.89203010e-03
8.80930960e-01 -3.92953873e-01 2.56033272e-01 -2.50579327e-01
-6.29887938e-01 5.16875029e-01 5.43806851e-01 7.35720932e-01
7.95957088e-01 -1.57366204e+00 -7.07194090e-01 3.04456711e-01
1.85213387e-01 6.63406610e-01 4.54959124e-02 9.39382672e-01
-5.14944792e-01 8.31419379e-02 -7.07102492e-02 -1.32432294e+00
-1.04914260e+00 3.56365085e-01 5.97110569e-01 -1.88810542e-01
-5.87743402e-01 8.37570786e-01 2.54860044e-01 -5.38154066e-01
4.78495181e-01 -1.56435639e-01 1.71294883e-02 6.52069002e-02
2.47855723e-01 3.25216293e-01 -1.26636162e-01 -6.87444627e-01
-4.46085513e-01 1.33688664e+00 -6.60618246e-02 1.09343104e-01
1.42513680e+00 -1.48321599e-01 -1.29434928e-01 2.55261749e-01
1.82961953e+00 -4.32801604e-01 -1.84444857e+00 -5.11708498e-01
-1.23263486e-01 -1.12730145e+00 -7.87932351e-02 -1.78683549e-01
-1.45218050e+00 9.06060815e-01 5.78155398e-01 -5.59416890e-01
8.65782320e-01 1.10193707e-01 9.05781388e-01 3.29829901e-01
4.64866042e-01 -8.06476176e-01 5.62268615e-01 6.03037357e-01
9.47014928e-01 -1.11487174e+00 9.98081872e-04 -7.42244899e-01
-1.11392453e-01 1.24944186e+00 8.32757890e-01 -4.38665926e-01
6.63561106e-01 5.75616099e-02 4.07476984e-02 -5.81558160e-02
-4.26408976e-01 -1.13019332e-01 6.78080261e-01 3.65915418e-01
5.35183176e-02 -3.11880499e-01 -6.41302322e-04 -3.49131405e-01
5.74262664e-02 9.03141871e-03 -3.20869237e-02 8.83551002e-01
-3.42695564e-01 -1.06841552e+00 -2.67377168e-01 1.43497318e-01
-1.17375299e-01 4.43218231e-01 -3.29117209e-01 4.84596997e-01
4.34080064e-02 5.35908580e-01 2.94940442e-01 -1.89445034e-01
6.61550045e-01 -3.10938835e-01 9.88245368e-01 -7.13505447e-01
-3.39243442e-01 3.45533580e-01 -2.92090327e-01 -8.45486701e-01
-7.00223267e-01 -6.25796735e-01 -1.23621535e+00 -1.55129865e-01
-6.59568429e-01 -5.51847070e-02 5.85774660e-01 8.67444277e-01
3.28086257e-01 7.97103904e-03 8.31603408e-01 -1.37187362e+00
-5.65953314e-01 -5.87635279e-01 -9.32921320e-02 6.13326907e-01
6.52939022e-01 -9.63167131e-01 -4.32211220e-01 -2.06830084e-01]
|
[7.49861478805542, -2.6279945373535156]
|
8c3336ad-88b6-4a65-acd4-2077502e1516
|
simple-yet-powerful-an-overlooked
| null | null |
https://openreview.net/forum?id=cL4tgY1ZxS
|
https://openreview.net/pdf?id=cL4tgY1ZxS
|
Simple yet Powerful: An Overlooked Architecture for Nested Named Entity Recognition
|
Named Entity Recognition (NER) is an important task in Natural Language Processing that aims to identify text spans belonging to predefined categories. Traditional NER research ignores nested entities, which are entities contained in other entity mentions. Although several methods have been proposed to address this case, most of them rely on complex task-specific structures and ignore potentially useful baselines for the task. We argue that this creates an overly optimistic impression of their performance. This paper revisits the Multiple LSTM-CRF (MLC) model, a simple, overlooked, yet powerful approach based on training independent sequence labeling models for each entity type. Extensive experiments with three nested NER corpora show that, regardless of the simplicity of this model, its performance is better or at least as well as more sophisticated methods. Furthermore, we show that the MLC architecture achieves state-of-the-art results in the Chilean Waiting List corpus by including pre-trained language models. In addition, we propose new task-specific metrics that adequately measure the ability of models to detect nestings. The results show that standard NER metrics do not measure well the ability of a model to detect nested entities, while our task-specific metrics provide new evidence on how existing approaches handle the task.
|
['Anonymous']
|
2021-11-16
| null | null | null |
acl-arr-november-2021-11
|
['nested-named-entity-recognition']
|
['natural-language-processing']
|
[-2.94272929e-01 2.74314135e-01 -1.28039107e-01 -4.59932923e-01
-9.11745787e-01 -7.26724207e-01 8.02598000e-01 5.32360435e-01
-1.11125052e+00 8.84420156e-01 5.98083496e-01 -5.33462882e-01
9.27585289e-02 -6.06055498e-01 -4.69943613e-01 -1.86892271e-01
-1.83436245e-01 5.96321881e-01 2.45507047e-01 -3.08609217e-01
2.61722833e-01 4.73576635e-01 -1.10047090e+00 2.98919588e-01
9.07323360e-01 3.40646297e-01 1.40325218e-01 5.56018531e-01
-5.42575777e-01 1.03715050e+00 -8.47058713e-01 -6.91226006e-01
-1.73832089e-01 -2.18399451e-03 -1.23947930e+00 -5.76241314e-01
3.01712096e-01 9.90977231e-03 -8.19481313e-02 7.43854284e-01
4.30788338e-01 2.52175987e-01 6.06801212e-01 -7.34499633e-01
-5.17028093e-01 1.05652058e+00 -4.15858664e-02 4.35960114e-01
2.83515900e-01 -3.13982636e-01 1.16116345e+00 -8.17745268e-01
8.09343219e-01 9.79425550e-01 1.09923589e+00 6.07532978e-01
-1.00535488e+00 -4.96659309e-01 9.89664271e-02 -1.95511803e-01
-1.21316123e+00 -5.73803425e-01 1.33241620e-02 -3.29084128e-01
1.46450114e+00 1.10920735e-01 -9.40209478e-02 9.33386207e-01
3.08479726e-01 6.42283976e-01 1.03606784e+00 -6.76649630e-01
-2.53289193e-02 2.42124915e-01 5.64235389e-01 5.08657157e-01
5.21260917e-01 -1.72266081e-01 -1.89975545e-01 -1.99361995e-01
2.82470554e-01 -4.18906331e-01 -1.23313434e-01 9.07051936e-02
-1.28693235e+00 7.26221919e-01 2.99568474e-01 1.16117239e+00
-4.97088999e-01 -1.04454868e-01 6.34850502e-01 3.03863198e-03
4.93440449e-01 8.96999300e-01 -8.98204565e-01 -1.89323515e-01
-1.18947184e+00 4.67456179e-03 1.26162732e+00 8.47614586e-01
4.57605541e-01 1.94102209e-02 -3.91357720e-01 7.11401343e-01
6.08552583e-02 1.62355766e-01 5.75735033e-01 -5.40343165e-01
5.22214115e-01 5.73040068e-01 2.81572908e-01 -9.18555975e-01
-9.95177865e-01 -4.32371408e-01 -5.55154383e-01 -2.22779304e-01
4.83451992e-01 -4.28736210e-01 -9.00554895e-01 1.76432717e+00
-8.09916109e-03 -1.57024950e-01 1.88475028e-01 4.68648434e-01
1.03418326e+00 5.72449625e-01 8.06134343e-01 4.08340618e-02
1.55274796e+00 -9.97319102e-01 -8.06520522e-01 -4.99332309e-01
1.11734807e+00 -7.77626097e-01 6.35445952e-01 -1.08971022e-01
-9.53984261e-01 -4.50120330e-01 -7.45420516e-01 -2.51454949e-01
-9.43998396e-01 3.19123536e-01 8.02293122e-01 1.00322318e+00
-1.07564270e+00 6.21736228e-01 -7.68062294e-01 -5.91533184e-01
-2.91496187e-01 3.01839352e-01 -5.17995059e-01 1.97291791e-01
-1.46093643e+00 1.49032021e+00 7.46268451e-01 2.19690487e-01
-4.33968484e-01 -3.62067282e-01 -1.05857933e+00 3.02770734e-01
2.60873884e-01 -3.69497269e-01 1.41187143e+00 -5.03494084e-01
-8.03162456e-01 9.00393188e-01 -4.34828669e-01 -6.82340920e-01
2.02517867e-01 -3.33381414e-01 -8.10449421e-01 -2.15100676e-01
3.99073094e-01 5.86659729e-01 -5.85671999e-02 -1.20360720e+00
-8.68631959e-01 -4.00449373e-02 1.26633018e-01 1.10905685e-01
-2.71241039e-01 5.65471947e-01 -1.06983431e-01 -5.39215446e-01
-2.81726509e-01 -7.86518574e-01 -5.41672826e-01 -8.89716387e-01
-4.98460144e-01 -7.20021427e-01 1.96216747e-01 -8.56408238e-01
1.61615217e+00 -1.92944109e+00 -4.36398327e-01 -1.21960476e-01
2.52537757e-01 4.81243551e-01 -1.47797793e-01 7.43295491e-01
-7.74053261e-02 8.38663995e-01 -1.39883071e-01 -4.41087753e-01
2.18459770e-01 1.11783147e-01 -1.72029316e-01 1.47774369e-01
3.38468283e-01 9.69304919e-01 -8.37590098e-01 -5.58779061e-01
-9.31268036e-02 6.22516870e-01 -1.43991923e-02 -1.54975921e-01
2.08419248e-01 8.97694454e-02 -2.28996456e-01 2.45033056e-01
3.72364283e-01 -2.83630461e-01 3.88567328e-01 4.04834673e-02
-6.06582522e-01 9.33837414e-01 -1.08934045e+00 1.32719564e+00
-4.10760343e-01 6.98608339e-01 -1.29819050e-01 -7.59691238e-01
7.31900990e-01 6.82798088e-01 4.37871879e-03 -5.68419695e-01
-1.29001020e-02 5.00621378e-01 1.83969960e-01 -3.96744877e-01
1.03055501e+00 -2.02543348e-01 -3.87179375e-01 2.52703041e-01
2.06176266e-01 5.60386002e-01 5.84734321e-01 1.75757751e-01
1.23579741e+00 -4.59542777e-03 6.57253265e-01 -3.68539184e-01
4.24300969e-01 1.67925492e-01 7.76311874e-01 1.05312049e+00
-2.37800926e-01 3.06078225e-01 1.36025250e-01 -3.53392959e-01
-9.71485436e-01 -6.74438894e-01 -2.68150002e-01 1.33009171e+00
-1.76046178e-01 -4.64677662e-01 -7.46091783e-01 -1.07060647e+00
-4.33119535e-01 1.10411835e+00 -6.03635371e-01 3.10885876e-01
-9.38629866e-01 -7.82405019e-01 1.07087123e+00 6.69873893e-01
4.00768936e-01 -1.42948008e+00 -6.62520826e-01 5.11579633e-01
-5.85592210e-01 -1.31769860e+00 -2.81341136e-01 4.89879966e-01
-7.09256351e-01 -8.84713233e-01 -7.41304576e-01 -1.07576847e+00
3.52537006e-01 5.22364490e-02 1.55173254e+00 1.40335679e-01
8.36237073e-02 2.40473926e-01 -5.52736163e-01 -3.39851350e-01
-5.02920926e-01 6.60590649e-01 -1.33059278e-01 -5.92361748e-01
8.58356297e-01 -5.49317934e-02 -1.18779428e-01 7.32122809e-02
-7.43837297e-01 -3.40842217e-01 9.01713312e-01 6.66509271e-01
3.52012180e-02 -5.05621471e-02 7.45140135e-01 -1.27529752e+00
7.35175371e-01 -4.31367189e-01 -1.00393921e-01 5.43215334e-01
-6.26593828e-01 2.59254724e-01 5.68055034e-01 -2.21358582e-01
-1.28569603e+00 -1.15859307e-01 -4.62008327e-01 3.14659387e-01
-6.87859893e-01 8.73960435e-01 -4.93054790e-03 2.70445675e-01
7.01403201e-01 2.97953933e-02 -8.14541101e-01 -6.14792466e-01
2.38515243e-01 5.89256525e-01 4.77035373e-01 -4.75464284e-01
5.70758104e-01 8.45257416e-02 -3.54900748e-01 -8.82557034e-01
-1.07592118e+00 -7.69787967e-01 -8.65444958e-01 1.33186504e-01
1.10132694e+00 -9.43295717e-01 -4.73188072e-01 -4.61744517e-03
-1.39378285e+00 -2.72879928e-01 -4.26716208e-02 5.69401324e-01
4.96435203e-02 4.11051691e-01 -1.05597687e+00 -1.00498927e+00
-4.53195035e-01 -7.18451560e-01 9.12884831e-01 2.74388045e-01
-4.56525832e-01 -1.36293721e+00 2.18881220e-01 5.78267239e-02
5.16676486e-01 1.40831202e-01 9.02159929e-01 -1.43206298e+00
-6.59404043e-03 -3.76674354e-01 -1.09281562e-01 6.57026749e-03
-2.07281768e-01 -1.68890387e-01 -8.63785207e-01 -6.32349625e-02
-1.68132976e-01 -1.56632140e-01 9.52836454e-01 1.63320601e-01
3.35730642e-01 -3.46293122e-01 -6.04360580e-01 -1.90009493e-02
1.41797209e+00 2.72067159e-01 7.11905479e-01 7.90787160e-01
4.31547314e-01 8.84408057e-01 3.14844847e-01 1.12867076e-03
8.06098878e-01 3.31011534e-01 -4.05775644e-02 -3.30725431e-01
3.36161926e-02 -3.39531958e-01 3.60707372e-01 9.02834594e-01
7.90819824e-02 -5.54096103e-01 -1.25411332e+00 7.35254884e-01
-1.60613525e+00 -1.00121880e+00 -6.30517602e-01 2.03504610e+00
7.48420000e-01 3.02256048e-01 -1.51344889e-03 -1.16265334e-01
9.43353057e-01 6.97389916e-02 -8.90977681e-03 -6.05252206e-01
-2.84423977e-01 3.20844889e-01 6.48564041e-01 2.86919624e-01
-1.50591195e+00 1.15326703e+00 7.05293274e+00 6.40598774e-01
-7.95987546e-01 2.03871399e-01 3.18569720e-01 5.93518019e-01
-1.04724392e-01 1.21095300e-01 -1.27577066e+00 2.10814029e-01
1.50633311e+00 -1.18415251e-01 -1.98135316e-01 7.99063861e-01
2.32303217e-02 -5.65320626e-02 -1.02635539e+00 3.24404687e-01
8.72203335e-02 -1.05638146e+00 -1.56363308e-01 1.26224637e-01
4.52399373e-01 2.97089905e-01 -4.20337170e-01 9.39999163e-01
6.82943881e-01 -1.03669918e+00 6.12256765e-01 3.38562995e-01
3.88860375e-01 -5.13023317e-01 1.14883280e+00 4.96191025e-01
-1.20417738e+00 3.84305716e-02 -3.47046375e-01 1.11027919e-02
5.10154545e-01 4.91332829e-01 -8.73319209e-01 6.80644274e-01
6.11474097e-01 2.49917492e-01 -8.67924273e-01 1.32640362e+00
-3.63961697e-01 8.18033397e-01 -2.40986034e-01 -1.61208764e-01
5.50639868e-01 2.98450112e-01 3.66469830e-01 1.95467782e+00
1.41071603e-01 -5.33411726e-02 1.20162129e-01 5.13546169e-01
-2.83272356e-01 4.55262035e-01 -6.09421968e-01 -1.36633277e-01
5.45084059e-01 1.42749882e+00 -9.96141911e-01 -4.53144759e-01
-6.65936351e-01 7.93441474e-01 7.29264677e-01 1.98078573e-01
-6.24266505e-01 -7.42636442e-01 1.54016539e-02 -2.42965177e-01
2.95267463e-01 -4.24639165e-01 -2.42705867e-01 -1.22390497e+00
-6.97213560e-02 -6.02151573e-01 6.59986496e-01 -5.20911813e-01
-1.34328079e+00 9.55137432e-01 -2.75442064e-01 -8.04627657e-01
-3.18957239e-01 -6.30907416e-01 -5.35199881e-01 7.04532921e-01
-1.56460071e+00 -1.01233459e+00 1.11068860e-01 1.29599720e-01
4.62666601e-01 2.18994781e-01 1.08310628e+00 4.87497449e-01
-5.87134361e-01 5.37047982e-01 9.68842357e-02 8.10531914e-01
8.97687554e-01 -1.49163222e+00 7.52939522e-01 1.03678322e+00
4.71250206e-01 9.82823372e-01 7.06070185e-01 -7.13106215e-01
-5.47188818e-01 -9.32390213e-01 2.01000118e+00 -8.09152186e-01
5.84072649e-01 -3.23553085e-01 -9.99241590e-01 9.92908478e-01
5.84703207e-01 -3.98342192e-01 8.95116746e-01 7.33518898e-01
-4.17093515e-01 5.34479976e-01 -9.76662695e-01 4.18911874e-01
8.15379143e-01 -4.01569307e-01 -1.21528649e+00 1.94834113e-01
6.39572561e-01 -1.42063826e-01 -1.05693698e+00 4.42479044e-01
3.44855517e-01 -8.33084285e-01 6.62000895e-01 -8.98047805e-01
1.85332194e-01 -4.58634906e-02 3.58822010e-02 -1.10343790e+00
-4.90248829e-01 -1.76002041e-01 2.05142811e-01 1.60423851e+00
9.34806108e-01 -5.72171748e-01 4.96762812e-01 9.16953743e-01
-2.88268805e-01 -2.70246297e-01 -7.78516650e-01 -9.72470164e-01
4.64711219e-01 -2.90751100e-01 3.31888914e-01 1.36150301e+00
1.91641703e-01 9.26865160e-01 -1.11948781e-01 1.82263926e-01
2.20698193e-01 -3.04076612e-01 3.46735924e-01 -1.52515376e+00
2.24974588e-01 -5.00801206e-01 -6.86773434e-02 -8.60192299e-01
5.19661307e-01 -9.05885279e-01 3.12161118e-01 -2.09847021e+00
1.01877086e-01 -5.91422081e-01 -2.74355620e-01 7.95128226e-01
-3.40939105e-01 -3.73160839e-02 2.51880199e-01 3.11699659e-01
-8.52054656e-01 1.02324232e-01 4.69901651e-01 1.25375882e-01
-1.82609469e-01 -1.49831682e-01 -6.69716895e-01 7.15922356e-01
7.56422698e-01 -7.27166474e-01 3.38769674e-01 -5.46870112e-01
4.13253158e-01 -9.19449627e-02 1.13667235e-01 -9.75740433e-01
5.22457600e-01 2.25747719e-01 4.29228932e-01 -5.43278635e-01
-8.24127346e-02 -5.54860353e-01 -1.58860117e-01 4.13000464e-01
-5.76304436e-01 3.77349228e-01 3.21697950e-01 3.52810472e-01
-2.96753794e-01 -7.61992514e-01 3.58355373e-01 -3.06134164e-01
-9.42404628e-01 -2.06797570e-01 -7.74902582e-01 3.87195885e-01
6.19480491e-01 1.66088883e-02 -4.12641376e-01 -1.42352849e-01
-7.41024792e-01 2.83214092e-01 1.63166672e-01 4.89728451e-01
-1.64876226e-02 -8.95742893e-01 -8.04005682e-01 -3.59333515e-01
1.32636530e-02 -3.28690559e-01 -1.63896065e-02 6.91719651e-01
-2.54714936e-01 9.63679850e-01 1.37055993e-01 -6.31000325e-02
-1.01570582e+00 5.90815485e-01 3.04505616e-01 -9.74321604e-01
-4.64986175e-01 5.73515475e-01 6.99355453e-02 -8.56258929e-01
1.11904144e-01 -1.65103078e-01 -7.57629216e-01 2.08439291e-01
4.49692011e-01 3.00000876e-01 2.87923396e-01 -9.22109067e-01
-5.07901967e-01 1.58690009e-02 -1.85870066e-01 -9.77428406e-02
1.26229334e+00 -1.19109273e-01 -1.04239486e-01 6.16681755e-01
8.16196561e-01 2.61340827e-01 -2.46857405e-01 -8.68933052e-02
9.41142559e-01 2.68856227e-01 -1.94153771e-01 -1.04514158e+00
-4.42266226e-01 7.75525749e-01 2.46880010e-01 5.11636555e-01
6.89140975e-01 -1.79375648e-01 6.82906926e-01 7.76404977e-01
4.44204330e-01 -1.06486428e+00 -6.43137515e-01 1.09523141e+00
3.20958048e-01 -1.12585640e+00 -2.07870737e-01 -2.55765080e-01
-6.44425511e-01 1.12299097e+00 6.08984351e-01 1.36335179e-01
4.26762670e-01 2.48781994e-01 1.29958734e-01 -9.78073925e-02
-6.33553624e-01 -5.99164724e-01 2.24493742e-01 4.34094548e-01
1.09248698e+00 -7.59068504e-02 -8.56500983e-01 7.14261174e-01
-2.05856442e-01 -2.21211106e-01 5.36327124e-01 8.56521606e-01
-5.35498857e-01 -1.09011590e+00 -1.77563712e-01 4.42269027e-01
-1.14342260e+00 -5.06441057e-01 -4.62587357e-01 1.26443398e+00
-2.26214938e-02 1.24006629e+00 -2.59728953e-02 -1.05543785e-01
5.72111845e-01 4.78933513e-01 -3.68354023e-02 -7.73220122e-01
-1.26244485e+00 -1.47109881e-01 6.81373656e-01 -1.66262716e-01
-6.15871012e-01 -4.82098132e-01 -1.38456845e+00 -9.69974846e-02
-5.81837296e-01 8.42619121e-01 6.77142978e-01 1.08485937e+00
3.78203481e-01 3.77300560e-01 9.48838741e-02 -5.88910818e-01
-4.24623519e-01 -1.23111022e+00 -3.90127867e-01 2.65204996e-01
1.49321452e-01 -4.72621530e-01 -3.29499751e-01 -1.49676919e-01]
|
[9.768868446350098, 9.584970474243164]
|
9e2bc2fd-f73a-4057-a4ff-64923c0fdb91
|
fusion-aware-point-convolution-for-online
|
2003.06233
| null |
https://arxiv.org/abs/2003.06233v4
|
https://arxiv.org/pdf/2003.06233v4.pdf
|
Fusion-Aware Point Convolution for Online Semantic 3D Scene Segmentation
|
Online semantic 3D segmentation in company with real-time RGB-D reconstruction poses special challenges such as how to perform 3D convolution directly over the progressively fused 3D geometric data, and how to smartly fuse information from frame to frame. We propose a novel fusion-aware 3D point convolution which operates directly on the geometric surface being reconstructed and exploits effectively the inter-frame correlation for high quality 3D feature learning. This is enabled by a dedicated dynamic data structure which organizes the online acquired point cloud with global-local trees. Globally, we compile the online reconstructed 3D points into an incrementally growing coordinate interval tree, enabling fast point insertion and neighborhood query. Locally, we maintain the neighborhood information for each point using an octree whose construction benefits from the fast query of the global tree.Both levels of trees update dynamically and help the 3D convolution effectively exploits the temporal coherence for effective information fusion across RGB-D frames.
|
['Chenyang Zhu', 'Lintao Zheng', 'Kai Xu', 'Jiazhao Zhang']
|
2020-03-13
|
fusion-aware-point-convolution-for-online-1
|
http://openaccess.thecvf.com/content_CVPR_2020/html/Zhang_Fusion-Aware_Point_Convolution_for_Online_Semantic_3D_Scene_Segmentation_CVPR_2020_paper.html
|
http://openaccess.thecvf.com/content_CVPR_2020/papers/Zhang_Fusion-Aware_Point_Convolution_for_Online_Semantic_3D_Scene_Segmentation_CVPR_2020_paper.pdf
|
cvpr-2020-6
|
['rgb-d-reconstruction']
|
['computer-vision']
|
[-7.12668225e-02 -2.61908192e-02 2.45034412e-01 -3.25283885e-01
-1.05858612e+00 -6.64211035e-01 3.23141158e-01 3.78371954e-01
-3.69113833e-01 1.67115346e-01 -3.49305004e-01 -1.92280725e-01
-1.19618796e-01 -1.16777229e+00 -8.04368854e-01 -5.39765716e-01
-1.82289049e-01 6.89668596e-01 7.40726352e-01 -9.84986722e-02
1.07116006e-01 1.19972873e+00 -1.93324268e+00 1.44613232e-03
7.70467579e-01 1.66861010e+00 3.58599305e-01 8.99039924e-01
-6.32311702e-01 1.33019432e-01 -2.27965921e-01 6.94314688e-02
5.21370649e-01 1.92113608e-01 -7.23783135e-01 5.03142834e-01
3.69244605e-01 -5.65611005e-01 -1.79750144e-01 6.89331710e-01
4.30265814e-01 2.78867960e-01 -1.11043409e-01 -9.66593683e-01
2.96838701e-01 -1.70778513e-01 -4.32807654e-01 -5.71161620e-02
5.67356586e-01 1.65273666e-01 5.40577948e-01 -9.90207791e-01
8.72898221e-01 1.17798126e+00 8.18459630e-01 5.37061878e-02
-9.33016837e-01 -3.67115319e-01 1.34342059e-01 1.16657846e-01
-1.22293329e+00 -2.27444425e-01 1.08791494e+00 -1.47491217e-01
1.02925873e+00 4.35500085e-01 1.28092575e+00 2.62781650e-01
-8.45229253e-02 3.98023248e-01 8.23996246e-01 -1.71501949e-01
4.70599979e-01 -5.50507307e-01 -5.71230315e-02 8.95148695e-01
-3.34749371e-01 2.38689482e-01 -8.46633673e-01 -9.32974145e-02
1.31931150e+00 2.64335185e-01 -1.03079863e-01 -7.60669470e-01
-1.42568088e+00 4.10606742e-01 5.79591036e-01 8.07380676e-02
-4.93745506e-01 5.35160959e-01 1.85101062e-01 3.46760511e-01
7.43865788e-01 -1.70435846e-01 -9.42659795e-01 -2.47004911e-01
-8.64231825e-01 -5.10821678e-02 5.63272119e-01 1.13377643e+00
1.59611654e+00 -3.29894543e-01 3.36516142e-01 3.50557923e-01
2.61717588e-01 6.62153840e-01 -3.60277109e-02 -1.58763325e+00
2.28191718e-01 9.38288391e-01 7.99489319e-02 -8.06601882e-01
-5.46797812e-01 -1.93529174e-01 -4.91604894e-01 6.31564438e-01
2.41313592e-01 2.83342421e-01 -8.65782976e-01 1.13365459e+00
1.30071747e+00 4.03209388e-01 -1.79226592e-01 6.63876951e-01
8.10819864e-01 3.59342784e-01 -3.24011564e-01 -1.60232678e-01
1.22841537e+00 -4.46427226e-01 -3.32914919e-01 2.11770773e-01
6.26053274e-01 -6.85195863e-01 7.52786815e-01 3.33931521e-02
-1.29394138e+00 -6.60690308e-01 -9.47775245e-01 -5.76323986e-01
-3.44615549e-01 -3.88001621e-01 7.67452359e-01 2.39246592e-01
-1.35478318e+00 6.47124112e-01 -1.26468778e+00 -1.94582194e-01
6.09086037e-01 5.89537501e-01 -5.74913323e-01 -2.92314529e-01
-6.36532307e-01 5.87245941e-01 4.31685030e-01 5.16514573e-03
-4.87865150e-01 -1.05189836e+00 -9.57896888e-01 -2.28233978e-01
5.37215948e-01 -1.12559867e+00 1.20626295e+00 -4.67378050e-01
-1.46174908e+00 1.04112399e+00 -3.39093983e-01 -2.91553855e-01
3.75936359e-01 -2.10809842e-01 1.36295393e-01 5.88458240e-01
1.28248990e-01 6.88445210e-01 7.85919130e-01 -1.33420217e+00
-8.37000847e-01 -9.92085040e-01 2.36569002e-01 4.77914870e-01
3.07239294e-01 -5.13423681e-01 -8.75398159e-01 -1.86521515e-01
1.03693068e+00 -4.28166837e-01 -4.67651457e-01 7.86827624e-01
-1.59477014e-02 -1.70884475e-01 1.21568489e+00 -4.48750824e-01
6.07111573e-01 -2.21620488e+00 -6.60746768e-02 3.19761485e-01
4.24142838e-01 -2.54002720e-01 2.51752585e-01 -3.96809280e-02
1.62329480e-01 -3.85758668e-01 -3.21091533e-01 -7.22048163e-01
-2.28488818e-01 6.93656504e-01 1.67393219e-02 6.12671852e-01
-2.45656688e-02 1.03377795e+00 -9.91707921e-01 -7.00654507e-01
8.44415426e-01 7.50082016e-01 -5.29523075e-01 1.47686854e-01
-3.88636023e-01 6.55173540e-01 -9.55196142e-01 8.85983825e-01
1.03849554e+00 -2.65421331e-01 -4.09196556e-01 -5.16131997e-01
-5.08637786e-01 2.50560850e-01 -1.17634654e+00 2.85492945e+00
-4.56380218e-01 6.66413680e-02 3.57607484e-01 -6.45485759e-01
9.07023251e-01 1.62953958e-01 1.10528350e+00 -7.01816797e-01
9.40258056e-02 3.05502564e-01 -9.46815133e-01 3.83144058e-02
6.34793103e-01 9.70445350e-02 -6.22301362e-02 2.85833985e-01
4.31555463e-03 -9.73117471e-01 -4.98506546e-01 1.19262695e-01
1.15413427e+00 5.62032163e-01 7.07126316e-03 1.10304028e-01
4.63755697e-01 3.45161021e-01 2.29901627e-01 4.97373641e-01
1.01184785e-01 5.04950762e-01 -5.93783855e-02 -5.82535625e-01
-9.49012160e-01 -1.41710556e+00 -1.47408098e-01 4.72648978e-01
5.97666323e-01 -3.16656977e-01 -6.12068176e-01 -5.01300573e-01
2.29088858e-01 2.47568235e-01 -4.37222362e-01 1.43746957e-01
-7.02606440e-01 2.89100111e-02 -2.53526121e-01 4.18970674e-01
8.16900671e-01 -5.29426277e-01 -1.28676331e+00 4.86527473e-01
5.36698550e-02 -1.04692888e+00 -3.59465569e-01 4.68694061e-01
-1.56603789e+00 -1.06111908e+00 -1.76931590e-01 -3.05877239e-01
5.58584630e-01 6.13948584e-01 1.14868474e+00 1.11014523e-01
-1.22937903e-01 9.09470320e-01 -3.03852975e-01 1.49679603e-02
-1.77172467e-01 -1.80386499e-01 -3.56839299e-01 -2.39415050e-01
-2.05785837e-02 -1.01665986e+00 -7.55643666e-01 3.08012933e-01
-7.63351202e-01 3.65141660e-01 9.53101069e-02 3.56536537e-01
1.39679015e+00 1.35641336e-01 -1.47263572e-01 -3.68826658e-01
-3.90166819e-01 -4.74758148e-02 -1.06034684e+00 7.96613321e-02
-1.44105386e-02 -2.21090348e-04 -9.21497121e-02 6.17839806e-02
-8.23529899e-01 6.59970701e-01 -3.92204314e-01 -8.83275926e-01
-1.30412698e-01 5.88374250e-02 -2.68220812e-01 -3.20896417e-01
6.55818284e-02 1.29790604e-01 7.80604035e-02 -7.15155721e-01
8.49614680e-01 1.97021082e-01 8.39240491e-01 -6.77145481e-01
8.24504077e-01 1.07670152e+00 3.43343675e-01 -6.20791078e-01
-7.99433172e-01 -6.55594170e-01 -1.32746720e+00 -4.32197124e-01
1.08491504e+00 -7.84000039e-01 -8.54621768e-01 5.73777378e-01
-1.49769175e+00 -4.24884528e-01 -9.96479154e-01 2.54903048e-01
-9.59825456e-01 3.36351633e-01 -3.54801148e-01 -6.12018883e-01
-3.07544202e-01 -1.10295534e+00 1.72230101e+00 -6.44070283e-02
1.45029217e-01 -8.54915261e-01 -1.23772882e-01 1.48978218e-01
1.95580535e-02 7.59258628e-01 5.17290711e-01 1.13515168e-01
-1.31530261e+00 -1.79540604e-01 -2.36272633e-01 4.19303589e-02
2.38422498e-01 -8.40100870e-02 -9.67109323e-01 9.94600952e-02
4.34730738e-01 2.75505960e-01 4.07509536e-01 3.93937677e-01
1.19872093e+00 2.48380303e-01 -4.39991504e-01 1.13670599e+00
1.39151812e+00 -1.54068936e-02 2.73580492e-01 6.13745116e-02
8.93448293e-01 3.98251802e-01 8.03181648e-01 5.24503291e-01
7.03531086e-01 6.42714918e-01 8.53524923e-01 -2.48566568e-01
-1.75402358e-01 -1.84264168e-01 -8.38383734e-02 9.69516993e-01
-9.72404107e-02 2.83419102e-01 -6.54053807e-01 6.72439411e-02
-1.75343120e+00 -5.35573006e-01 -2.74622887e-01 2.38051438e+00
6.50263309e-01 4.21687728e-03 -2.48652801e-01 2.93966681e-01
3.08615535e-01 1.23786047e-01 -8.22249472e-01 8.28009397e-02
-9.24990326e-02 7.28794873e-01 7.25382268e-01 6.78390265e-01
-9.48573351e-01 8.60876918e-01 5.64524698e+00 8.03497851e-01
-1.00168300e+00 3.58432114e-01 2.66477585e-01 -1.31495506e-01
-4.68945622e-01 2.26844907e-01 -5.58182597e-01 7.81104639e-02
5.19595325e-01 8.15603957e-02 2.48522043e-01 7.82489359e-01
8.58763978e-02 -5.12841702e-01 -1.02789164e+00 1.38534284e+00
-3.97890240e-01 -1.66396856e+00 -2.21939981e-01 3.16675276e-01
5.71203053e-01 3.53326023e-01 -3.17004174e-01 -4.05220836e-01
2.34820649e-01 -4.28882688e-01 9.01527345e-01 7.40425885e-01
8.70705605e-01 -8.87029231e-01 6.26569092e-02 4.14378613e-01
-1.55539262e+00 2.07533240e-01 -2.13009641e-01 2.03063399e-01
4.57131892e-01 1.02560210e+00 -4.26881909e-01 9.78747487e-01
8.98860633e-01 9.24750328e-01 -2.89374501e-01 7.54578710e-01
1.90055683e-01 -2.48397872e-01 -9.97300386e-01 5.71950436e-01
1.23059087e-01 -3.17009151e-01 6.29693151e-01 5.50194085e-01
6.68270767e-01 4.12133664e-01 2.00002238e-01 7.72836983e-01
2.19675899e-01 -2.86717921e-01 -4.75988090e-01 5.97381175e-01
6.14510894e-01 1.22615707e+00 -1.04977953e+00 -4.26098377e-01
-4.73119736e-01 1.21042728e+00 8.29276964e-02 1.26326948e-01
-5.01085937e-01 3.01859155e-02 9.32494283e-01 2.19160810e-01
5.64077079e-01 -9.36262012e-01 -5.67045569e-01 -8.90473247e-01
9.46165323e-02 -3.80879343e-02 3.96584392e-01 -1.00738764e+00
-7.71228790e-01 4.79682475e-01 -9.43102539e-02 -1.32781172e+00
-6.20813817e-02 -2.95631230e-01 -1.57832682e-01 7.47462213e-01
-1.58698058e+00 -1.01107192e+00 -7.13084221e-01 1.15432608e+00
4.04386938e-01 6.65624857e-01 6.58187211e-01 1.26938879e-01
1.92559019e-01 -5.95438108e-02 -2.28974596e-01 -3.72469634e-01
1.32657238e-03 -1.04595232e+00 4.81018722e-01 5.50859332e-01
6.91477060e-02 1.51894242e-01 1.00126058e-01 -7.06364036e-01
-1.82040167e+00 -8.16303551e-01 5.43972433e-01 -6.54043794e-01
3.27810913e-01 -3.47591907e-01 -9.03722703e-01 3.60923737e-01
-5.62243819e-01 4.98501867e-01 1.51347190e-01 -4.20154095e-01
-1.79729477e-01 -3.26032937e-01 -1.55872297e+00 1.38152376e-01
1.57921278e+00 -6.21605635e-01 -2.91258991e-01 3.18395525e-01
1.41736007e+00 -1.10040224e+00 -1.31494832e+00 4.49651569e-01
3.17477465e-01 -1.09271181e+00 1.42081904e+00 9.39240605e-02
-3.95128019e-02 -7.27706432e-01 -5.00985086e-01 -7.02313960e-01
1.92529187e-01 -7.37092674e-01 -3.29997838e-01 8.27131152e-01
-2.91915983e-01 -5.09841681e-01 9.03659880e-01 7.02471852e-01
-5.57889163e-01 -7.29819834e-01 -1.66568530e+00 -3.46596211e-01
-5.08723319e-01 -1.26256776e+00 1.02213883e+00 6.93681180e-01
-6.76854670e-01 -3.19080740e-01 4.37524587e-01 4.60042447e-01
9.19712782e-01 4.28784817e-01 9.51715708e-01 -1.40455329e+00
3.91169302e-02 -1.88876152e-01 -7.76450098e-01 -1.61373341e+00
-3.06003362e-01 -8.06684375e-01 -1.46336570e-01 -1.57074928e+00
-5.59964836e-01 -8.94825101e-01 6.16789460e-02 3.02360743e-01
1.80560201e-01 2.02710271e-01 3.07391379e-02 9.54005122e-02
-6.47064805e-01 5.93926728e-01 1.59393632e+00 3.53845090e-01
-3.16256374e-01 6.00530067e-04 -5.88194840e-02 6.46274328e-01
2.60059327e-01 -2.91932642e-01 -2.76607990e-01 -6.28886461e-01
1.80278514e-02 5.73453844e-01 7.64881194e-01 -1.03753304e+00
3.59098405e-01 1.35149136e-02 4.28415537e-01 -1.55462873e+00
9.78842735e-01 -1.30694890e+00 5.60727179e-01 1.59625590e-01
3.73657107e-01 3.19495834e-02 1.39924958e-01 4.80611503e-01
4.16099317e-02 3.63620371e-01 7.73578525e-01 -4.50150669e-01
-8.00448477e-01 9.67286229e-01 3.61097217e-01 -3.66153210e-01
1.14178050e+00 -7.89551497e-01 4.95904595e-01 -4.62879241e-02
-1.16816521e+00 2.70951241e-01 1.03828096e+00 8.08292329e-02
8.71332467e-01 -1.31624007e+00 -1.90633222e-01 5.33185899e-01
-1.09404474e-01 1.14254797e+00 5.83615363e-01 7.85969019e-01
-6.71254694e-01 -2.06356216e-02 1.49895057e-01 -1.34465265e+00
-1.02067196e+00 4.04097468e-01 4.94992524e-01 3.26862000e-02
-1.16966558e+00 8.55136871e-01 2.83834897e-02 -4.60300088e-01
8.99934620e-02 -8.70786905e-01 4.80463088e-01 -6.46531731e-02
2.73173243e-01 4.14095759e-01 4.17395771e-01 -6.53947592e-01
-3.03832114e-01 1.17223239e+00 4.77008671e-01 -2.18485609e-01
1.48839617e+00 -5.52125692e-01 -4.14769322e-01 5.19554496e-01
1.46901476e+00 -2.47213840e-01 -1.77141798e+00 -4.80377227e-01
-1.51695073e-01 -8.19859684e-01 3.58033836e-01 -3.04039091e-01
-1.21739221e+00 7.48682976e-01 5.77935815e-01 6.60149306e-02
1.24298084e+00 5.38450837e-01 1.07095873e+00 1.26224965e-01
1.05892313e+00 -8.13055694e-01 -2.12271333e-01 5.15748441e-01
5.35416603e-01 -7.37057328e-01 6.99344799e-02 -7.50206828e-01
1.48899928e-01 1.29442823e+00 4.76768725e-02 -3.43243591e-02
9.85625267e-01 3.75232428e-01 -1.09010302e-01 -5.23914099e-01
-4.08843666e-01 -3.65439802e-01 3.44991013e-02 8.38224351e-01
-3.40042591e-01 -1.52384669e-01 3.56481969e-01 -3.40974443e-02
-4.53763366e-01 6.36459738e-02 -1.59087211e-01 1.04193175e+00
-5.52136481e-01 -1.22336698e+00 -5.20079792e-01 2.52433300e-01
3.82038057e-01 2.93798864e-01 1.86155871e-01 6.64378345e-01
3.27589959e-01 5.34279168e-01 6.19423330e-01 -2.54964501e-01
4.40989733e-01 3.88118438e-02 8.37854147e-01 -4.43310052e-01
-3.91390562e-01 2.86235869e-01 -3.24956834e-01 -1.37818217e+00
-6.70397460e-01 -8.43353689e-01 -1.82535303e+00 -2.61698365e-01
-1.37362912e-01 -1.35281280e-01 1.09796274e+00 8.26917708e-01
5.75180233e-01 3.34808379e-01 9.17419970e-01 -1.38273478e+00
1.13642171e-01 -2.77533889e-01 -4.90905851e-01 -3.08025647e-02
5.47169685e-01 -7.63777494e-01 -2.56185144e-01 -8.79278257e-02]
|
[8.354684829711914, -2.9504592418670654]
|
37e3c5e9-6539-4b47-afb1-ec379a7d736e
|
semi-automated-segmentation-of-geoscientific
|
2303.11404
| null |
https://arxiv.org/abs/2303.11404v1
|
https://arxiv.org/pdf/2303.11404v1.pdf
|
Semi-Automated Segmentation of Geoscientific Data Using Superpixels
|
Geological processes determine the distribution of resources such as critical minerals, water, and geothermal energy. However, direct observation of geology is often prevented by surface cover such as overburden or vegetation. In such cases, remote and in-situ surveys are frequently conducted to collect physical measurements of the earth indicative of the geology. Developing a geological segmentation based on these measurements is challenging since individual datasets can differ in properties (e.g. units, dynamic ranges, textures) and because the data does not uniquely constrain the geology. Further, as the number of datasets grows the information to constrain geology increases while simultaneously becoming harder to make sense of. Inspired by the concept of superpixels, we propose a deep-learning based approach to segment rasterized survey data into regions with similar characteristics. We demonstrate its use for semi-automated geoscientific mapping with datasets arising from independent sensors and with diverse properties. In addition, we introduce a new loss function for superpixels including a novel regularization parameter penalizing image segmentation with non-connected component superpixels. This improves integration of prior knowledge by allowing better control over the number of superpixels generated.
|
['Eldad Haber', 'Conrad P. Koziol']
|
2023-03-20
| null | null | null | null |
['superpixels']
|
['computer-vision']
|
[ 2.56205827e-01 7.92670473e-02 3.28033008e-02 -4.15702909e-01
-4.48339939e-01 -5.92327595e-01 4.66268182e-01 4.04667675e-01
-3.93592179e-01 1.00433123e+00 9.85171720e-02 -2.41989866e-01
6.15270995e-02 -1.31789207e+00 -8.10162008e-01 -7.43816614e-01
-8.02785456e-02 5.25997281e-01 3.82398009e-01 -1.16890721e-01
1.99379697e-01 6.49742723e-01 -1.35571516e+00 -3.95063579e-01
1.22721136e+00 8.24546635e-01 6.72377825e-01 2.74943471e-01
-2.02725101e-02 1.40653178e-01 -2.91932613e-01 3.47468823e-01
4.70952690e-01 -3.29131573e-01 -6.48389339e-01 4.35489595e-01
5.62385321e-01 -3.31835330e-01 1.53389629e-02 1.17400944e+00
7.67117739e-02 2.13401720e-01 7.82633901e-01 -4.87992674e-01
-3.16651672e-01 5.60452759e-01 -8.34015787e-01 -2.14776039e-01
-3.73205781e-01 4.61444221e-02 1.02193189e+00 -4.24156040e-01
3.90382975e-01 9.83411729e-01 6.38274133e-01 -2.23719865e-01
-1.40244544e+00 -1.69554263e-01 3.86353701e-01 -8.70132744e-02
-1.42755306e+00 -3.00325602e-01 5.75050294e-01 -7.25799799e-01
5.15646577e-01 2.07708001e-01 7.83460438e-01 2.43764728e-01
-9.70222652e-02 4.74149704e-01 1.12422311e+00 -3.01011294e-01
5.28592348e-01 -1.92346632e-01 -1.70249209e-01 4.50556099e-01
6.44879460e-01 -2.26497054e-01 -1.81040421e-01 -2.19063926e-02
1.02374935e+00 4.00186218e-02 -4.36134785e-01 -2.12355450e-01
-8.92521322e-01 6.54946566e-01 4.29767817e-01 1.31851867e-01
-4.99781698e-01 2.97756821e-01 -1.18135408e-01 -3.58104557e-02
7.69753933e-01 5.06193161e-01 -2.50512034e-01 1.54580414e-01
-1.44135630e+00 2.23871395e-01 7.26313472e-01 5.82156360e-01
1.40482664e+00 2.82090425e-01 4.86042291e-01 8.68761599e-01
2.97654867e-01 8.26367855e-01 2.42514491e-01 -1.08859241e+00
1.63257197e-01 5.81823409e-01 3.81545931e-01 -8.53223145e-01
-3.27573508e-01 -1.40770212e-01 -6.29243731e-01 3.22120696e-01
4.80740666e-01 -1.32585302e-01 -1.23214948e+00 1.45024931e+00
3.27581644e-01 -2.60015745e-02 -2.41611019e-01 9.67614472e-01
1.80513203e-01 5.83046019e-01 1.64378151e-01 2.12080881e-01
1.22127080e+00 -3.09633076e-01 -2.00619787e-01 -6.56795979e-01
3.37815285e-01 -2.36582145e-01 9.20217574e-01 9.17142257e-02
-5.49009800e-01 -1.48027048e-01 -1.00235164e+00 1.64531052e-01
-4.59146231e-01 8.38183835e-02 7.52792656e-01 5.37122309e-01
-9.15549159e-01 8.79793882e-01 -1.35613835e+00 -2.88020819e-01
3.75589609e-01 1.13549910e-01 -2.21858770e-01 2.84843028e-01
-1.04547215e+00 8.48706603e-01 4.52747732e-01 3.14785779e-01
-8.29808176e-01 -6.11270308e-01 -8.52139473e-01 1.13122448e-01
2.86925882e-01 -7.62085170e-02 6.40956998e-01 -7.09000826e-01
-1.16712189e+00 8.57972443e-01 2.17247769e-01 -3.73813957e-01
6.32201314e-01 -2.02606156e-01 -8.73586088e-02 1.08348258e-01
2.97663271e-01 6.72439337e-01 5.41304648e-01 -1.17019057e+00
-6.31016552e-01 -3.95126224e-01 -9.98897478e-02 2.03921124e-01
-1.77321434e-01 -3.48946661e-01 -3.31028759e-01 -4.15751815e-01
4.97249186e-01 -7.86995888e-01 -4.65470999e-01 4.03175876e-02
-4.48205531e-01 3.65959257e-01 6.53805256e-01 -8.93595815e-01
7.58211553e-01 -1.87737203e+00 1.55834794e-01 2.95844823e-01
1.93388700e-01 -1.39666542e-01 1.83702320e-01 1.93784952e-01
3.61961246e-01 3.42491508e-01 -1.03236246e+00 8.72216299e-02
-8.30216333e-02 5.34386098e-01 3.36640067e-02 6.83561087e-01
1.28857285e-01 3.28075022e-01 -8.19296658e-01 -3.79684478e-01
3.90780717e-01 2.92706162e-01 -1.78324983e-01 -1.55355223e-02
-4.42502528e-01 5.58234692e-01 -7.27154315e-01 7.67370582e-01
1.00942862e+00 -2.25871295e-01 2.61572272e-01 8.80448073e-02
-7.16619551e-01 2.99253821e-01 -1.15319002e+00 1.44138479e+00
-5.56410909e-01 7.85196602e-01 4.95011479e-01 -1.26516509e+00
1.02991474e+00 -7.40535706e-02 3.91029209e-01 -5.06466329e-01
-1.58320934e-01 3.51478487e-01 -2.26726592e-01 -5.23234665e-01
7.84265816e-01 -3.69109720e-01 1.18168071e-01 3.52890074e-01
-3.89674038e-01 -6.36240244e-01 1.69860914e-01 -6.96314424e-02
6.58113241e-01 3.04495335e-01 7.94057772e-02 -8.21197808e-01
1.18229121e-01 2.77130634e-01 7.04355419e-01 6.42806053e-01
2.18887284e-01 7.16439247e-01 4.10096526e-01 -2.68852025e-01
-1.34917808e+00 -1.02204466e+00 -5.32847941e-01 6.51238203e-01
3.07120949e-01 3.50700468e-01 -6.25103712e-01 -1.27046287e-01
3.44647497e-01 4.15109038e-01 -5.29182017e-01 1.62246317e-01
-4.28354442e-01 -1.28276026e+00 4.39490169e-01 6.28051341e-01
7.06254244e-01 -6.96327507e-01 -8.83482456e-01 4.81544793e-01
-2.63090134e-01 -9.93891895e-01 1.03446849e-01 3.21376413e-01
-1.12112236e+00 -9.71017301e-01 -7.19689131e-01 -3.70992064e-01
8.84666383e-01 6.85732365e-02 1.07663655e+00 -1.95434734e-01
-1.31681174e-01 1.16345592e-01 -2.67839491e-01 -2.57792085e-01
-2.07914382e-01 1.46428809e-01 -2.75489211e-01 4.17386144e-02
1.60507374e-02 -7.77707636e-01 -7.50043154e-01 2.59402722e-01
-8.71594191e-01 2.52178580e-01 4.74596322e-01 4.29753900e-01
7.80546546e-01 1.03400178e-01 3.33637953e-01 -1.06135798e+00
1.64135545e-02 -7.54405677e-01 -9.87483263e-01 3.23423505e-01
-4.85769689e-01 8.24502259e-02 3.88794154e-01 -1.76756918e-01
-1.23793280e+00 2.26890892e-02 2.17426598e-01 1.31384492e-01
-3.24884742e-01 9.69966173e-01 -1.14024639e-01 -1.54806733e-01
5.39260149e-01 1.79305777e-01 -1.93843901e-01 -6.16735101e-01
2.66741604e-01 6.59555197e-01 8.87612402e-01 -6.90987289e-01
6.04894638e-01 9.30493355e-01 3.72106470e-02 -1.43791187e+00
-5.20560980e-01 -4.39591825e-01 -7.46965826e-01 -1.85596362e-01
7.95974314e-01 -1.07157528e+00 -1.65795550e-01 5.93021393e-01
-6.60801053e-01 -7.06130028e-01 -1.07346997e-01 6.99324071e-01
-1.73291102e-01 5.98024607e-01 -4.04328883e-01 -7.94001520e-01
-4.97313514e-02 -8.90290737e-01 8.64328086e-01 2.63201952e-01
9.87918815e-04 -1.19675231e+00 1.78409263e-01 2.31575817e-01
4.12124634e-01 8.48407626e-01 5.45107186e-01 1.15904994e-02
-7.25955844e-01 -6.60046712e-02 -4.02969062e-01 4.79068696e-01
5.71728587e-01 7.64813423e-02 -8.80397856e-01 -2.54121684e-02
-2.74587572e-02 -2.70789699e-03 1.21506107e+00 7.07874298e-01
8.35066438e-01 -1.83991089e-01 -2.44223163e-01 8.06978881e-01
1.64986682e+00 2.04690807e-02 5.75801551e-01 5.41960716e-01
1.01632857e+00 7.49048054e-01 3.44688952e-01 6.54165983e-01
2.69628584e-01 3.27991635e-01 5.04490554e-01 -1.52380928e-01
9.93492007e-02 8.05560648e-02 -1.76214986e-02 2.39819124e-01
-3.67695451e-01 -1.38993829e-01 -1.22758150e+00 1.02130163e+00
-1.56135011e+00 -6.75940454e-01 -2.89818764e-01 2.40584350e+00
7.72559106e-01 -7.95504227e-02 -2.66612619e-01 -1.71633378e-01
9.42600012e-01 3.39262903e-01 -8.02014172e-01 1.73537627e-01
-4.30944473e-01 4.73066941e-02 1.37999105e+00 7.78331280e-01
-1.11235499e+00 9.89753783e-01 5.53601265e+00 3.54372084e-01
-1.41432631e+00 -1.86836392e-01 6.92280352e-01 2.24775746e-01
-6.71457052e-01 3.65453064e-01 -7.24214733e-01 5.32757163e-01
6.12622321e-01 1.11341149e-01 3.94188911e-01 6.43427253e-01
6.33383930e-01 -7.71509767e-01 -6.59858882e-01 5.72372139e-01
-4.52357262e-01 -1.28051043e+00 -2.95918614e-01 2.12637231e-01
8.41982126e-01 5.04308581e-01 -3.20830226e-01 -4.15169626e-01
6.40647471e-01 -7.20329225e-01 9.71884608e-01 4.98344868e-01
8.69305015e-01 -3.25868428e-01 2.50383705e-01 5.71405888e-02
-1.27576303e+00 1.71915993e-01 -4.41975296e-01 -1.95893899e-01
2.98159570e-01 1.11495209e+00 -6.39523029e-01 4.52606410e-01
5.97726107e-01 6.65990829e-01 -3.45529139e-01 1.00532162e+00
-4.84385163e-01 9.07963574e-01 -9.95088577e-01 3.27043325e-01
3.68044734e-01 -8.75289202e-01 3.49652529e-01 9.88641441e-01
3.72936934e-01 1.19158655e-01 2.18667895e-01 1.18320882e+00
-4.34037223e-02 -1.37999505e-01 -3.98694068e-01 -9.52124000e-02
7.79880047e-01 1.14821005e+00 -1.19548643e+00 -2.23509252e-01
-3.15909296e-01 6.35895908e-01 9.85667557e-02 3.21692646e-01
-4.31661367e-01 -2.06078693e-01 5.64277530e-01 3.60867679e-01
3.12163860e-01 -5.62748492e-01 -5.97608387e-01 -1.03973651e+00
1.49098217e-01 -2.95025498e-01 6.39740899e-02 -5.07403076e-01
-7.91767955e-01 5.87716512e-02 2.03619942e-01 -1.05784154e+00
-1.26935467e-02 -3.21108371e-01 -7.49389529e-01 1.13975573e+00
-1.78245509e+00 -1.03345132e+00 -6.07755363e-01 -7.35991299e-02
1.31001458e-01 3.69307250e-01 3.99963200e-01 1.25947744e-01
-4.67997968e-01 -3.14533412e-01 8.41960132e-01 7.19834417e-02
4.15235639e-01 -1.46176231e+00 5.53846121e-01 1.06156743e+00
-2.96639472e-01 8.67153555e-02 7.01020539e-01 -8.59348953e-01
-1.03275323e+00 -1.18597257e+00 3.82655263e-01 1.95133075e-01
8.36707056e-01 -1.39148489e-01 -1.21817672e+00 5.57655156e-01
-5.27311087e-01 -5.98415807e-02 3.49209517e-01 -3.27409245e-02
2.00617407e-02 -2.87154049e-01 -1.12063527e+00 3.13266724e-01
5.89914918e-01 -4.68650341e-01 -2.05200702e-01 3.67973298e-01
4.35694866e-02 -1.76488847e-01 -8.25645030e-01 3.35296005e-01
5.48645616e-01 -7.81556070e-01 7.38280118e-01 8.15801173e-02
4.75617498e-01 -3.90433431e-01 -2.93628156e-01 -1.21900487e+00
-2.50306249e-01 -1.35468215e-01 5.21622598e-01 1.16927326e+00
3.45192492e-01 -5.55927336e-01 1.11298144e+00 8.81002367e-01
-3.25511098e-01 -1.98936373e-01 -7.67299354e-01 -9.70169127e-01
2.82752991e-01 -1.81739703e-01 5.53997099e-01 9.34728146e-01
-4.47797894e-01 -3.49996209e-01 -3.28656495e-01 5.47103941e-01
8.50701988e-01 3.03255230e-01 5.22349358e-01 -1.47824395e+00
-1.54266611e-01 -3.66157472e-01 -3.03579301e-01 -1.00814569e+00
-1.67091638e-01 -4.43002641e-01 3.75125319e-01 -1.58588922e+00
-1.11542672e-01 -9.40832675e-01 9.40724369e-03 6.69625759e-01
-9.16608050e-02 2.28664815e-01 -2.22034782e-01 5.03677964e-01
1.69858336e-01 5.38194656e-01 9.47837174e-01 -3.61124277e-01
-3.94317329e-01 -3.80402565e-01 -2.24941716e-01 7.50088871e-01
1.00554109e+00 -4.33078915e-01 -2.43219316e-01 -8.88173997e-01
2.75009483e-01 5.00976667e-02 3.39605480e-01 -1.11294472e+00
-6.26296178e-02 -5.81714153e-01 1.96859986e-01 -5.69359958e-01
1.41497418e-01 -6.20264947e-01 3.54663700e-01 3.65904897e-01
5.47302589e-02 -8.26964259e-01 -2.96993852e-02 4.62224036e-01
-2.51891226e-01 -4.11749423e-01 9.67027724e-01 -4.64260846e-01
-8.84354472e-01 3.63506615e-01 -4.86590087e-01 -1.00865573e-01
5.29957831e-01 -5.00830352e-01 -1.59621507e-01 -9.94313806e-02
-4.68166173e-01 4.39468384e-01 9.17857528e-01 -1.41352862e-01
3.46829295e-01 -7.39393890e-01 -9.50538039e-01 3.78525481e-02
-4.77848761e-02 6.00937486e-01 3.26771557e-01 6.31777704e-01
-1.17572224e+00 -7.69921839e-02 -3.53786498e-01 -5.72262168e-01
-7.17677951e-01 -1.55044407e-01 4.84193832e-01 -1.21606765e-02
-7.83001363e-01 7.77985454e-01 3.43999118e-01 -2.96438128e-01
-4.05307114e-01 -5.23477495e-01 -1.26351789e-01 3.83960247e-01
1.72116235e-01 5.20885050e-01 -1.61332078e-02 -5.27249157e-01
-1.59455568e-01 4.13350254e-01 1.37536094e-01 -2.47701973e-01
1.62796903e+00 -3.36192369e-01 -2.84522146e-01 4.88466203e-01
8.80539060e-01 -1.05512172e-01 -1.83140814e+00 -1.56258494e-01
1.55329004e-01 -5.15804589e-01 3.87534440e-01 -3.88785541e-01
-1.33408689e+00 7.29323447e-01 3.24275613e-01 1.82808563e-01
8.82523596e-01 5.21904416e-03 4.17336702e-01 4.05873716e-01
2.83389449e-01 -1.43601394e+00 -5.76816082e-01 2.64058411e-01
6.84864640e-01 -1.33570766e+00 3.52907449e-01 -2.56030530e-01
-2.01323822e-01 1.17471075e+00 3.05069238e-01 -1.49257824e-01
7.11246252e-01 2.86739141e-01 9.39603969e-02 -2.47824952e-01
1.08360820e-01 -2.76051044e-01 -1.70530856e-01 4.16457504e-01
1.94820940e-01 4.15747315e-01 -2.51907259e-01 -2.18023472e-02
-2.61505425e-01 -2.54052818e-01 7.30376184e-01 1.05831337e+00
-8.63816261e-01 -7.33201861e-01 -7.16588020e-01 6.14127040e-01
-3.02610666e-01 -1.27888262e-01 -7.14198798e-02 5.09415209e-01
4.70183156e-02 6.69827402e-01 4.53644991e-01 7.10998550e-02
-2.11025402e-01 -3.93605053e-01 2.86290139e-01 -7.19336331e-01
2.30999961e-01 2.46811062e-01 1.49113163e-01 3.00883297e-02
-6.12878263e-01 -9.56049860e-01 -1.21633983e+00 -2.09538981e-01
-3.24137717e-01 2.06322774e-01 9.08754587e-01 9.26538825e-01
-1.60975561e-01 1.55841425e-01 5.62807500e-01 -1.10410118e+00
-1.73795283e-01 -1.10783136e+00 -1.23524332e+00 2.15403244e-01
2.36603752e-01 -6.79581285e-01 -3.57475758e-01 3.70215893e-01]
|
[9.609047889709473, -1.3933149576187134]
|
ebb2b3b6-4903-4744-931f-cc4a53e6bb8f
|
distributed-representations-for-biological
|
1608.05949
| null |
http://arxiv.org/abs/1608.05949v2
|
http://arxiv.org/pdf/1608.05949v2.pdf
|
Distributed Representations for Biological Sequence Analysis
|
Biological sequence comparison is a key step in inferring the relatedness of
various organisms and the functional similarity of their components. Thanks to
the Next Generation Sequencing efforts, an abundance of sequence data is now
available to be processed for a range of bioinformatics applications. Embedding
a biological sequence over a nucleotide or amino acid alphabet in a lower
dimensional vector space makes the data more amenable for use by current
machine learning tools, provided the quality of embedding is high and it
captures the most meaningful information of the original sequences. Motivated
by recent advances in the text document embedding literature, we present a new
method, called seq2vec, to represent a complete biological sequence in an
Euclidean space. The new representation has the potential to capture the
contextual information of the original sequence necessary for sequence
comparison tasks. We test our embeddings with protein sequence classification
and retrieval tasks and demonstrate encouraging outcomes.
|
['James M. Hogan', 'Dhananjay Kimothi', 'Akshay Soni', 'Pravesh Biyani']
|
2016-08-21
| null | null | null | null |
['document-embedding']
|
['methodology']
|
[ 6.38046741e-01 -3.88590604e-01 -1.40939072e-01 -2.76366949e-01
-5.02926052e-01 -8.18869174e-01 5.25148213e-01 7.79456139e-01
-6.54976010e-01 6.39300883e-01 5.07283509e-01 -4.50960070e-01
3.25477533e-02 -3.78042996e-01 -3.45507026e-01 -9.41431880e-01
-9.89217237e-02 2.81539857e-01 -1.42141148e-01 -1.34735510e-01
2.34698772e-01 6.70302987e-01 -1.31688404e+00 1.68665573e-01
3.93481523e-01 5.99751949e-01 5.92705250e-01 8.74849558e-01
-4.21446800e-01 7.86662251e-02 -4.67198700e-01 -2.68472373e-01
2.00418234e-01 -5.16203821e-01 -8.38504076e-01 -3.17817599e-01
1.54755771e-01 7.11956248e-03 -5.66408217e-01 8.80178392e-01
5.98427474e-01 2.26271912e-01 7.64222920e-01 -8.83885384e-01
-7.51031637e-01 -1.35373607e-01 -3.43949437e-01 2.97520101e-01
4.50501472e-01 2.43994355e-01 1.40497208e+00 -7.32123673e-01
1.03772390e+00 1.18282247e+00 5.80322027e-01 4.01337981e-01
-1.69213140e+00 9.56756820e-04 -2.65066504e-01 2.90400535e-01
-1.03841746e+00 8.44287034e-03 3.17605138e-01 -7.37385750e-01
1.35350406e+00 3.10342133e-01 6.76342010e-01 1.01135576e+00
6.20494246e-01 3.51100415e-01 4.69678253e-01 -2.89405137e-01
1.30287960e-01 -2.74344504e-01 2.80168504e-01 7.19518065e-01
1.80071458e-01 -2.05076471e-01 -4.29863036e-01 -6.86000824e-01
2.51905143e-01 5.39539337e-01 -4.56264615e-01 -5.47488630e-01
-1.50013471e+00 1.00899255e+00 2.79070169e-01 4.79461581e-01
-4.36540961e-01 1.65130824e-01 8.81641805e-01 3.38029355e-01
3.52489084e-01 6.87066495e-01 -4.04710203e-01 -3.44292015e-01
-4.81297106e-01 9.76765305e-02 8.11273515e-01 3.49372059e-01
4.96621758e-01 -3.68670493e-01 2.42125139e-01 7.91869104e-01
1.66428268e-01 1.71168089e-01 9.64939356e-01 -6.06023312e-01
-6.43902123e-02 7.85422444e-01 4.44702618e-02 -1.02918994e+00
-1.27816126e-01 6.77186996e-02 -6.03856742e-01 1.36583507e-01
2.55413949e-01 4.21287179e-01 -5.80027938e-01 1.58715463e+00
4.95735347e-01 8.75024498e-02 2.28884861e-01 8.06284547e-01
3.50483507e-01 9.25223529e-01 -7.06012398e-02 1.54060334e-01
1.59855294e+00 -3.25621307e-01 -4.16941851e-01 1.51432812e-01
7.33583629e-01 -7.65708268e-01 9.17857528e-01 1.52421966e-01
-4.84337091e-01 -3.40742141e-01 -1.38711679e+00 -4.36920434e-01
-6.89064145e-01 -6.34244621e-01 6.84331298e-01 4.05625552e-01
-7.46951997e-01 8.37097704e-01 -6.43850863e-01 -6.72846079e-01
1.12421475e-01 1.66709691e-01 -9.56728220e-01 -4.98244703e-01
-9.99217570e-01 1.01445401e+00 5.51941276e-01 -2.68958807e-01
-4.46349919e-01 -7.85770833e-01 -9.61345077e-01 2.47729644e-01
-2.81146079e-01 -6.20724440e-01 7.97686875e-01 -5.12674451e-01
-9.45687354e-01 8.54916751e-01 -3.43643695e-01 -3.22604388e-01
9.11398034e-04 6.55861273e-02 -1.75101027e-01 2.28385895e-01
-3.20443183e-01 4.61443514e-01 4.55137998e-01 -5.57006538e-01
-3.81913364e-01 -5.01253128e-01 -1.82501391e-01 7.78372735e-02
-2.11022094e-01 -1.04553722e-01 -2.40225881e-01 -6.06133461e-01
-1.36342004e-01 -9.41357970e-01 -3.92589778e-01 4.40767288e-01
1.44367144e-01 -2.19246477e-01 7.60957420e-01 -7.59371758e-01
7.76309133e-01 -2.39524126e+00 6.23289287e-01 7.43512288e-02
5.01436055e-01 4.33147728e-01 -4.75891918e-01 1.04162920e+00
-4.23791200e-01 1.30736202e-01 -5.10512650e-01 2.32961327e-01
-5.62518202e-02 3.16302925e-01 -1.65077001e-01 8.21873248e-01
3.41385543e-01 8.02183449e-01 -1.16885209e+00 -1.27032921e-02
1.06525593e-01 9.37289536e-01 -3.85108083e-01 3.37769806e-01
-7.36126080e-02 1.27981827e-01 -4.44778085e-01 1.61847219e-01
1.79514930e-01 -3.54645401e-01 6.01023316e-01 -1.73456773e-01
1.39183715e-01 3.69772524e-01 -6.97279751e-01 1.76938534e+00
-2.13988677e-01 1.02872384e+00 -2.71803916e-01 -1.04076827e+00
8.55905414e-01 3.06829602e-01 8.62295926e-01 -2.10522279e-01
-7.77458446e-03 -2.61168368e-02 4.37928140e-01 -4.74636972e-01
4.73997712e-01 -2.61711866e-01 5.04016876e-02 7.11952329e-01
7.25712404e-02 -5.46656130e-03 7.89247304e-02 3.75034660e-01
1.49916673e+00 -3.76913697e-02 7.51495600e-01 -3.81269932e-01
4.05248880e-01 3.35585363e-02 5.10529280e-01 2.51652420e-01
-2.73314834e-01 3.87456447e-01 4.69879329e-01 -6.72946930e-01
-1.62629056e+00 -8.41875970e-01 -1.02018684e-01 1.00824952e+00
-1.12980157e-01 -4.19629067e-01 -3.34434956e-01 -4.78105634e-01
4.76836383e-01 7.09348992e-02 -6.95329130e-01 -4.25977856e-01
-3.53248745e-01 -6.41870022e-01 5.66690326e-01 5.03486633e-01
-4.30681229e-01 -6.76463902e-01 -5.86789966e-01 4.82627720e-01
-5.99702783e-02 -7.11383700e-01 -7.49445081e-01 4.66853589e-01
-8.14068198e-01 -1.28998458e+00 -7.30012178e-01 -8.98775995e-01
4.13720757e-01 3.37761998e-01 9.22259331e-01 4.69529182e-02
-1.03663433e+00 2.58839041e-01 -2.93903530e-01 -2.55843699e-01
-4.73265260e-01 -3.90612632e-01 2.58497655e-01 -1.93707138e-01
7.34782040e-01 -4.85733181e-01 -7.02855110e-01 -1.28671274e-01
-1.39706993e+00 -2.11170629e-01 2.84215659e-01 1.26456118e+00
6.34494841e-01 -4.38988686e-01 4.26780164e-01 -6.61391020e-01
8.33448589e-01 -6.97543621e-01 -3.56370062e-01 2.88835049e-01
-3.67659539e-01 5.72483480e-01 6.35137677e-01 -3.25361431e-01
-2.78764635e-01 -5.09541407e-02 -2.32890159e-01 -5.73008396e-02
3.14854123e-02 6.98702216e-01 -1.79568101e-02 1.05211273e-01
5.38240314e-01 4.78185713e-01 4.32844847e-01 -5.65204382e-01
4.66131687e-01 9.17594969e-01 3.80228788e-01 -3.78590554e-01
1.55411810e-01 4.06710029e-01 7.97139630e-02 -1.06199920e+00
-1.21883415e-01 -8.96946847e-01 -7.59016335e-01 3.75094026e-01
8.77856195e-01 -5.35948217e-01 -9.35878575e-01 -1.40462995e-01
-1.01737916e+00 1.91377625e-01 -2.90923789e-02 4.87512410e-01
-5.10708213e-01 9.82915819e-01 -5.95979929e-01 -3.52029741e-01
-3.16072971e-01 -1.30630565e+00 9.27217901e-01 -2.54348606e-01
-5.58605492e-01 -9.98097599e-01 6.66264296e-01 2.67834235e-02
2.18552127e-02 3.65153223e-01 1.63434470e+00 -9.18809712e-01
-3.22099060e-01 -4.49856609e-01 -5.30667230e-02 2.47811183e-01
3.87574822e-01 1.31187951e-02 -6.63506031e-01 -5.34439802e-01
-2.09797919e-01 -2.64026791e-01 9.70288515e-01 -5.89282699e-02
1.01285923e+00 5.72212646e-03 -4.25716460e-01 5.45681417e-01
1.56326497e+00 1.45702437e-01 4.18183863e-01 1.79029882e-01
4.26377147e-01 6.07617736e-01 3.32268029e-01 3.82583797e-01
-2.46551439e-01 6.45592988e-01 2.21071005e-01 1.50820121e-01
2.05320463e-01 -1.99149221e-01 1.65574457e-02 8.54857206e-01
3.46140355e-01 -3.49119127e-01 -9.40615594e-01 5.49906373e-01
-1.63497150e+00 -1.05128610e+00 4.01160866e-01 2.22904205e+00
8.35848451e-01 -3.51283133e-01 -1.60661414e-01 1.35818139e-01
5.81882417e-01 2.52877355e-01 -8.96520495e-01 -8.64421427e-01
-3.39667983e-02 2.07599416e-01 4.25072134e-01 2.89743274e-01
-7.15351880e-01 4.35164988e-01 6.93204927e+00 4.67785567e-01
-1.03271985e+00 -3.14497679e-01 3.64752859e-01 1.05949111e-01
-5.25988221e-01 -2.80485630e-01 -3.63570809e-01 4.99553889e-01
1.23160958e+00 -5.45473635e-01 2.61222363e-01 5.72775006e-01
7.41351619e-02 7.97730684e-02 -1.48395371e+00 1.15527105e+00
2.05049478e-02 -1.65350521e+00 1.20873719e-01 2.68348038e-01
2.67935395e-01 3.25871438e-01 3.03190611e-02 -5.47647402e-02
4.70835000e-01 -1.19849253e+00 -1.45970449e-01 2.81782210e-01
1.00891364e+00 -7.68842518e-01 7.87124276e-01 1.12392180e-01
-9.20887411e-01 1.28761113e-01 -8.11345398e-01 1.65631667e-01
9.02749896e-02 4.42674756e-01 -1.28507793e+00 4.05410916e-01
3.62363249e-01 8.91395450e-01 -1.71402603e-01 8.99257123e-01
3.03889841e-01 2.97889769e-01 -2.16569081e-02 -2.77557880e-01
4.19475585e-01 -4.49882567e-01 5.34780264e-01 1.35884082e+00
2.71896776e-02 1.56053558e-01 1.13944195e-01 4.36976701e-01
-2.31877342e-01 3.29931498e-01 -8.55006576e-01 -8.04512143e-01
3.22059482e-01 1.04172218e+00 -4.90575939e-01 -2.52678424e-01
-5.41934311e-01 1.24005508e+00 5.52255154e-01 2.08154887e-01
-4.32016999e-01 -7.30250418e-01 1.80102575e+00 -4.48531926e-01
3.42573494e-01 -6.11679196e-01 3.05229068e-01 -9.43801701e-01
-1.05908863e-01 -1.03026092e+00 3.47235203e-01 -5.74634969e-01
-1.43955529e+00 3.84051949e-01 -4.75890309e-01 -9.05215859e-01
-2.95235395e-01 -1.01507950e+00 -1.35736167e-01 1.21941447e+00
-1.24218535e+00 -4.42051262e-01 2.24390328e-01 6.38762489e-02
5.34054279e-01 -2.71566540e-01 1.27865005e+00 1.32008985e-01
-3.23815644e-01 3.40746105e-01 1.23226130e+00 2.77176332e-02
7.13416934e-01 -1.13270545e+00 8.71447027e-01 3.28903794e-01
2.80081898e-01 1.08157456e+00 8.61564040e-01 -4.37423706e-01
-1.84775341e+00 -8.15923452e-01 9.87915039e-01 -4.09465760e-01
7.84355044e-01 -5.59495986e-01 -1.22073388e+00 4.90822345e-01
-8.35544255e-04 -1.58139735e-01 1.52440715e+00 -6.62453696e-02
-9.36125100e-01 2.77607799e-01 -1.10262775e+00 6.65047228e-01
7.72176802e-01 -9.42048013e-01 -8.16287875e-01 2.01109231e-01
8.59477043e-01 2.06211165e-01 -1.18238628e+00 9.72555727e-02
9.90403831e-01 -4.80297118e-01 1.10702229e+00 -1.33865857e+00
3.77070278e-01 -2.59437412e-01 -5.51060498e-01 -1.50500739e+00
-5.87413430e-01 -2.82540590e-01 1.92544386e-01 5.73908508e-01
2.22013101e-01 -6.48864329e-01 5.70844769e-01 4.07338709e-01
1.35273874e-01 -6.80494130e-01 -7.78801024e-01 -7.64270127e-01
4.98335361e-02 1.09354056e-01 6.39774024e-01 1.03626001e+00
3.57473552e-01 1.73601791e-01 -1.42146379e-01 -1.67640507e-01
1.61181867e-01 3.75413209e-01 5.21882355e-01 -1.31999409e+00
-3.55265290e-01 -4.33688015e-01 -1.26330805e+00 -1.06725299e+00
2.47397274e-01 -1.30390072e+00 -1.13221705e-02 -1.37150896e+00
5.37816107e-01 1.98767394e-01 -7.28014767e-01 1.62796631e-01
-4.74481225e-01 2.16283873e-01 -3.13972309e-02 1.71267644e-01
-1.50266066e-01 5.81285894e-01 6.27869844e-01 -2.87609190e-01
2.66744435e-01 -7.59812057e-01 -4.88805562e-01 1.64961219e-01
6.73527360e-01 -3.43539536e-01 -1.84367254e-01 -3.59565914e-01
3.31135094e-02 -1.15057655e-01 -8.35474655e-02 -5.13596475e-01
-2.34356448e-01 -1.63754135e-01 5.34421444e-01 -3.06254685e-01
5.40929019e-01 -8.48172069e-01 2.75049716e-01 7.66702414e-01
-6.20151222e-01 3.99132758e-01 1.63644105e-01 1.08935308e+00
-1.97972447e-01 -1.28524393e-01 6.59532785e-01 1.63161047e-02
-7.55782247e-01 2.45052651e-01 -5.72691977e-01 -1.65388733e-01
8.98101807e-01 -3.49078119e-01 -1.46721244e-01 -1.18046835e-01
-3.61453503e-01 -1.77895173e-01 7.60095894e-01 6.03606641e-01
7.29648411e-01 -1.13932610e+00 -6.88607097e-01 2.14373797e-01
5.86309373e-01 -8.50189745e-01 1.82511751e-02 3.55916679e-01
-8.58778179e-01 6.22991264e-01 -4.95456815e-01 -4.62383330e-01
-1.61475372e+00 7.50292838e-01 -4.26316336e-02 -1.33369574e-02
-7.15077758e-01 6.45637989e-01 3.05892050e-01 -5.18092334e-01
-8.41786116e-02 1.43511564e-01 -2.99936216e-02 -6.08397834e-02
8.65306973e-01 1.75762430e-01 1.35777909e-02 -6.15410089e-01
-4.16477531e-01 2.09873274e-01 -2.35906526e-01 1.04251273e-01
1.52853847e+00 1.11262374e-01 -2.10473672e-01 6.35970414e-01
1.85916603e+00 -2.64608115e-01 -9.53040302e-01 -2.64005661e-01
2.95881242e-01 -6.68010712e-01 -3.30354005e-01 -2.89487809e-01
-3.19434166e-01 9.63880897e-01 5.12445927e-01 -1.62492037e-01
5.16538024e-01 -1.15007728e-01 8.49985898e-01 7.39333451e-01
1.76619664e-01 -6.41674042e-01 -1.32464215e-01 5.66720307e-01
4.84600455e-01 -1.19521713e+00 -4.19609956e-02 1.25667527e-01
-4.19547886e-01 1.17561948e+00 -1.79467946e-01 1.02795236e-01
3.43377411e-01 1.54173553e-01 -6.75393827e-03 -1.21518075e-01
-9.11104798e-01 2.00498961e-02 2.28723556e-01 6.43297732e-01
9.42354977e-01 -1.85904410e-02 -5.12932837e-01 -9.93724465e-02
2.23991394e-01 -5.14183700e-01 3.07277262e-01 1.03883564e+00
-7.77822733e-01 -1.44930530e+00 2.70058811e-02 5.28554618e-01
-5.54200113e-01 -2.90793777e-01 -6.04609132e-01 3.08461636e-01
-5.05786896e-01 6.39982045e-01 -1.44940214e-02 -1.35426015e-01
1.43776918e-02 6.35888278e-01 3.54851395e-01 -7.09497511e-01
-4.12331074e-01 -2.55697161e-01 -2.23457858e-01 -5.61398923e-01
2.43286602e-02 -6.52728021e-01 -1.23206794e+00 -3.26947182e-01
-1.16914570e-01 4.10583138e-01 1.06879306e+00 5.73133767e-01
6.96624398e-01 3.19451362e-01 5.00009835e-01 -6.51911020e-01
-7.17071056e-01 -6.74258709e-01 -5.58403015e-01 8.30775499e-01
4.13626999e-01 -3.80155653e-01 -2.70114932e-02 2.61948556e-01]
|
[4.791141033172607, 5.5687947273254395]
|
1f928458-5cc1-4c2b-934d-12f87bbbe80a
|
alleviating-matthew-effect-of-offline
|
2307.04571
| null |
https://arxiv.org/abs/2307.04571v1
|
https://arxiv.org/pdf/2307.04571v1.pdf
|
Alleviating Matthew Effect of Offline Reinforcement Learning in Interactive Recommendation
|
Offline reinforcement learning (RL), a technology that offline learns a policy from logged data without the need to interact with online environments, has become a favorable choice in decision-making processes like interactive recommendation. Offline RL faces the value overestimation problem. To address it, existing methods employ conservatism, e.g., by constraining the learned policy to be close to behavior policies or punishing the rarely visited state-action pairs. However, when applying such offline RL to recommendation, it will cause a severe Matthew effect, i.e., the rich get richer and the poor get poorer, by promoting popular items or categories while suppressing the less popular ones. It is a notorious issue that needs to be addressed in practical recommender systems. In this paper, we aim to alleviate the Matthew effect in offline RL-based recommendation. Through theoretical analyses, we find that the conservatism of existing methods fails in pursuing users' long-term satisfaction. It inspires us to add a penalty term to relax the pessimism on states with high entropy of the logging policy and indirectly penalizes actions leading to less diverse states. This leads to the main technical contribution of the work: Debiased model-based Offline RL (DORL) method. Experiments show that DORL not only captures user interests well but also alleviates the Matthew effect. The implementation is available via https://github.com/chongminggao/DORL-codes.
|
['Xiangnan He', 'Zhong Zhang', 'Shiqi Wang', 'Peng Jiang', 'Biao Li', 'Yuan Zhang', 'Jiawei Chen', 'Kexin Huang', 'Chongming Gao']
|
2023-07-10
| null | null | null | null |
['reinforcement-learning-1', 'recommendation-systems', 'offline-rl', 'decision-making']
|
['methodology', 'miscellaneous', 'playing-games', 'reasoning']
|
[-3.47159624e-01 2.63760716e-01 -8.33954453e-01 -1.88734576e-01
-4.83269542e-01 -6.83139145e-01 1.13312423e-01 -1.42932683e-01
-4.16615129e-01 9.92584229e-01 4.27060783e-01 -6.26077354e-01
-1.83777973e-01 -6.77885354e-01 -5.96540928e-01 -8.40947628e-01
-1.26293406e-01 2.76851237e-01 -2.28043735e-01 -4.24480408e-01
3.18142205e-01 1.64730400e-01 -1.32867861e+00 8.24076012e-02
9.70729828e-01 8.54221344e-01 2.54360288e-01 2.49491796e-01
1.00348443e-01 8.54643762e-01 -1.79265589e-01 -2.86234617e-01
6.74849927e-01 -4.73324090e-01 -5.70383072e-01 -2.70788997e-01
-9.99992713e-02 -7.71421313e-01 -3.86244208e-01 1.20764565e+00
4.69798028e-01 5.60344338e-01 1.77056894e-01 -1.31455100e+00
-6.66900039e-01 1.02786398e+00 -4.36665177e-01 -3.54536576e-03
3.08363229e-01 3.92542034e-01 1.23990607e+00 -3.65970224e-01
3.45625907e-01 1.20858800e+00 3.28466952e-01 5.87623239e-01
-1.06129396e+00 -7.01458573e-01 7.29003489e-01 1.47498190e-01
-1.10547090e+00 -3.40954632e-01 7.33341575e-01 -3.44670117e-02
6.30920529e-01 5.00779629e-01 7.87798762e-01 1.02081490e+00
1.05013765e-01 1.03157854e+00 1.34794497e+00 -1.46024659e-01
4.60940570e-01 4.24620181e-01 1.12536497e-01 5.70849776e-01
4.74130780e-01 5.20830095e-01 -3.59382361e-01 -4.15097147e-01
6.86302066e-01 2.53206551e-01 -1.79829419e-01 -2.21197322e-01
-5.97631872e-01 9.75157976e-01 2.75154740e-01 -6.44918457e-02
-6.76525772e-01 1.91702917e-02 1.78226858e-01 6.58822119e-01
4.20858234e-01 6.57508075e-01 -6.13559008e-01 -4.76617813e-01
-6.02511287e-01 3.35690737e-01 9.38131213e-01 8.04240108e-01
6.40455484e-01 2.57717110e-02 -2.17217207e-01 5.99102378e-01
3.82564843e-01 4.44414794e-01 3.86912286e-01 -1.21474481e+00
3.24966013e-01 2.98903346e-01 7.01276660e-01 -9.39368665e-01
-2.56488144e-01 -4.58075404e-01 -6.01980686e-01 -4.69185933e-02
5.22230029e-01 -5.55597007e-01 -2.93772787e-01 1.94887197e+00
2.50991940e-01 4.34656590e-02 -4.95732799e-02 1.13822377e+00
9.77368206e-02 7.51234710e-01 -5.73854074e-02 -8.93977106e-01
8.56940567e-01 -8.62557411e-01 -8.80511284e-01 -7.48312920e-02
4.89743441e-01 -3.67341638e-01 1.52095926e+00 7.33172297e-01
-1.17421019e+00 -4.71626073e-02 -6.43454313e-01 4.49451417e-01
4.04430600e-03 -1.90751385e-02 7.73270547e-01 5.70975363e-01
-9.05940294e-01 7.64578104e-01 -8.47449481e-01 -2.48669341e-01
1.95850566e-01 4.66939390e-01 2.89997727e-01 1.73068628e-01
-1.40335035e+00 8.59576643e-01 5.63836582e-02 3.88296507e-02
-8.00077558e-01 -4.25747424e-01 -3.31002086e-01 3.03814709e-01
1.02751625e+00 -2.18353450e-01 1.62839031e+00 -1.25306714e+00
-2.07237124e+00 2.60739438e-02 1.58626601e-01 -5.41494489e-01
6.20554268e-01 -5.23660362e-01 -2.69487172e-01 -2.64581621e-01
-3.19099218e-01 1.16312452e-01 7.33557165e-01 -1.29483581e+00
-6.07092679e-01 -1.89370602e-01 2.73756027e-01 5.58574617e-01
-6.45420015e-01 -1.89485893e-01 -2.24489629e-01 -4.86085176e-01
-2.94529229e-01 -1.15290558e+00 -6.00989461e-01 -4.34482485e-01
-2.47606650e-01 -3.35178137e-01 4.65278894e-01 -5.25187373e-01
1.67577505e+00 -1.94901311e+00 -1.91400155e-01 4.34405446e-01
-5.32416403e-02 2.25778341e-01 -1.54911295e-01 6.12576365e-01
3.81738961e-01 1.94631636e-01 3.11856002e-01 -1.94870327e-02
1.21224836e-01 3.03219438e-01 -5.61298370e-01 6.80794418e-01
-4.87962723e-01 6.01246119e-01 -1.02369249e+00 -1.24583825e-01
7.72465095e-02 -9.04856622e-02 -1.05850804e+00 4.54271197e-01
-4.11268055e-01 3.89498144e-01 -7.12421656e-01 4.62744027e-01
5.20717382e-01 -2.57821202e-01 6.57024145e-01 1.71580706e-02
-4.04471964e-01 6.38576150e-01 -1.20949042e+00 1.13941336e+00
-4.48909372e-01 -1.43900871e-01 2.86363333e-01 -9.54857051e-01
6.26570225e-01 1.79546237e-01 7.58188307e-01 -7.04782188e-01
1.87561125e-01 -1.50752952e-03 7.65503990e-03 -5.48549056e-01
5.58045328e-01 -6.14802465e-02 -8.28645527e-02 7.11248636e-01
-4.26018894e-01 2.91604102e-01 -1.72868773e-01 3.13729078e-01
8.76934409e-01 1.65906847e-01 4.86013234e-01 -3.47453535e-01
5.71385846e-02 -1.20805942e-01 9.01537001e-01 9.74048793e-01
-4.61891532e-01 -3.28474164e-01 6.06257915e-01 -1.18363321e-01
-6.21648788e-01 -6.62502408e-01 3.40296686e-01 1.32759798e+00
4.01652336e-01 -3.03817719e-01 -3.77068937e-01 -8.89678240e-01
2.67565131e-01 1.08331156e+00 -3.50323290e-01 -5.50483763e-01
-4.25767452e-01 -4.56662059e-01 1.18821062e-01 1.86308503e-01
3.36045176e-01 -1.00200880e+00 -5.54933608e-01 2.82017857e-01
-2.30336010e-01 -4.80152220e-01 -8.00833046e-01 7.91103467e-02
-9.91575301e-01 -6.28218889e-01 -4.30613279e-01 -1.94156900e-01
5.98591745e-01 4.55829501e-01 8.34369719e-01 4.60141711e-02
5.41333377e-01 3.65595847e-01 -4.60128158e-01 -2.05205232e-01
-1.68710575e-01 -5.16899303e-02 6.13099456e-01 -2.73031648e-02
1.68533072e-01 -5.60568273e-01 -9.21196878e-01 4.82067078e-01
-4.29632515e-01 -7.20946938e-02 4.33786690e-01 9.16634142e-01
3.81674916e-01 4.59221378e-02 9.63981092e-01 -9.66469586e-01
9.78096902e-01 -7.87702084e-01 -8.08047652e-01 2.60527492e-01
-1.29361761e+00 1.18784063e-01 1.12777877e+00 -7.79192030e-01
-1.15051782e+00 -4.66955081e-02 -8.34597349e-02 -3.69557887e-01
1.74955711e-01 5.13860583e-01 -5.21655604e-02 2.70606101e-01
4.60861117e-01 1.48429617e-01 1.88491955e-01 -5.57265699e-01
4.65511501e-01 8.36326420e-01 4.95204367e-02 -7.42167115e-01
4.22066867e-01 3.13112557e-01 -5.18528283e-01 -4.51024354e-01
-1.09960687e+00 -4.96768683e-01 2.33155891e-01 -4.68274921e-01
1.33524895e-01 -7.98153996e-01 -1.17855942e+00 -1.48684412e-01
-4.92512643e-01 -7.19083667e-01 -3.83879989e-01 5.83223939e-01
-6.83659852e-01 2.48844683e-01 -7.10615456e-01 -1.39095032e+00
-4.12280917e-01 -8.12360942e-01 1.45001918e-01 5.82800269e-01
-7.75574297e-02 -6.44984603e-01 1.94608003e-01 2.45577052e-01
5.32321215e-01 -4.29401368e-01 4.67353195e-01 -5.99731863e-01
-2.98148841e-01 1.11451589e-01 2.42648125e-01 2.69542605e-01
1.01630099e-01 -7.53447488e-02 -6.73122048e-01 -7.37689257e-01
3.37859727e-02 -4.07249182e-01 5.26875794e-01 4.39218313e-01
1.35793793e+00 -1.10388601e+00 7.31759965e-02 2.87546396e-01
1.28124058e+00 4.83977258e-01 4.90546465e-01 2.12815017e-01
2.31519610e-01 4.36918110e-01 1.16616344e+00 1.12845254e+00
4.35273647e-01 5.04026592e-01 5.74317753e-01 1.38499573e-01
4.32936639e-01 -7.96637058e-01 9.68588591e-01 7.91277528e-01
-1.57621592e-01 -4.10120875e-01 -3.17181557e-01 2.28279755e-01
-2.33153296e+00 -1.02388465e+00 2.91415155e-01 2.61135340e+00
1.08720732e+00 1.33677661e-01 4.97511685e-01 -3.71440440e-01
4.19911534e-01 -1.44946605e-01 -1.01794600e+00 -6.09982252e-01
2.97463149e-01 -1.97814181e-01 8.69807482e-01 5.44546068e-01
-5.58946848e-01 1.08709168e+00 5.54120636e+00 7.22460985e-01
-1.11572230e+00 2.07530022e-01 5.68226397e-01 -4.02786911e-01
-3.58998865e-01 2.58242369e-01 -7.16529191e-01 6.46940708e-01
1.06951559e+00 -3.09531093e-01 1.07853734e+00 1.01980221e+00
8.95616055e-01 -2.72142410e-01 -7.63265848e-01 7.67482400e-01
-3.58892769e-01 -9.99610960e-01 -3.33142608e-01 1.88061908e-01
7.32496321e-01 -4.83962968e-02 8.28823596e-02 7.04653144e-01
8.64745259e-01 -6.16742074e-01 7.06879079e-01 7.10326433e-01
5.26197076e-01 -9.11968648e-01 4.55872297e-01 8.40434074e-01
-7.22850680e-01 -6.18622959e-01 -4.85782892e-01 -4.95259017e-01
1.74415365e-01 5.82250834e-01 -5.40921211e-01 1.48557454e-01
4.45367128e-01 6.77585542e-01 3.68148312e-02 8.41093719e-01
-1.91260397e-01 1.07110047e+00 -2.32990116e-01 -3.43326390e-01
1.40718415e-01 -7.28781760e-01 4.50215757e-01 6.49591208e-01
3.12053680e-01 3.59873086e-01 5.19199848e-01 7.28907824e-01
-7.90957734e-02 1.81450680e-01 -5.83019257e-01 -2.78053254e-01
5.81055701e-01 1.16432726e+00 -3.01777571e-01 -3.14416438e-01
-2.14409649e-01 7.27084041e-01 3.38041872e-01 4.85079110e-01
-8.32606077e-01 1.02698132e-01 7.11856306e-01 1.31250218e-01
-6.30261377e-03 -1.33490205e-01 -2.57311463e-01 -1.27491009e+00
-4.16762114e-01 -1.21905375e+00 4.43781197e-01 -3.06057543e-01
-1.28782320e+00 -9.18677747e-02 -1.67856172e-01 -1.21839702e+00
-5.30548207e-03 -5.20497635e-02 -5.49246192e-01 4.46720719e-01
-1.37016702e+00 -3.80139202e-01 2.08291620e-01 5.94871342e-01
4.89186943e-01 1.43246368e-01 4.87924159e-01 3.83945435e-01
-8.30047011e-01 7.64754713e-01 4.48347658e-01 -5.65520525e-01
8.90598357e-01 -1.03552771e+00 -2.72139639e-01 5.46841323e-01
-1.62720621e-01 8.79135847e-01 9.54162419e-01 -6.34784579e-01
-1.81193638e+00 -8.46308053e-01 4.86442626e-01 -1.08923845e-01
6.07602000e-01 -1.65181622e-01 -6.45950496e-01 5.56311965e-01
1.54176047e-02 -3.30027938e-01 5.22599936e-01 2.38885000e-01
1.21968221e-02 -2.15630889e-01 -1.18778777e+00 9.26781178e-01
9.77174699e-01 -2.54398257e-01 -1.85906738e-01 4.69288677e-01
6.86223507e-01 -1.23242229e-01 -7.77602196e-01 9.40190628e-02
6.58722341e-01 -7.89257705e-01 6.48613513e-01 -8.44036460e-01
5.01930490e-02 -7.71315396e-02 1.06951231e-02 -1.44472528e+00
-4.28823650e-01 -1.08706689e+00 -6.96208537e-01 9.53061700e-01
3.57076555e-01 -8.18756998e-01 7.22822368e-01 9.01678741e-01
4.68305275e-02 -9.63562846e-01 -5.71874201e-01 -9.00212228e-01
1.16272494e-01 -9.90269780e-02 4.83629018e-01 9.32845771e-01
5.40820360e-01 6.21908419e-02 -9.52902496e-01 1.19790018e-01
4.48555976e-01 3.35659295e-01 6.69881225e-01 -7.83484042e-01
-6.74896240e-01 -2.75215238e-01 7.71597803e-01 -1.54782939e+00
2.74257250e-02 -5.51421881e-01 1.75739974e-01 -1.25006592e+00
1.87097147e-01 -8.07583570e-01 -7.03442693e-01 6.48593664e-01
1.22454189e-01 -4.03500497e-01 2.09948525e-01 3.66028458e-01
-1.00221789e+00 6.91594601e-01 1.42465019e+00 3.49995881e-01
-7.21603990e-01 4.90649223e-01 -1.14597213e+00 6.57763600e-01
1.36122942e+00 -5.81057668e-01 -5.60034573e-01 2.65637070e-01
6.03266060e-01 4.83262032e-01 -4.26100194e-02 -2.27834046e-01
1.39241377e-02 -9.43859756e-01 -3.03313404e-01 -3.82410437e-01
5.89725636e-02 -8.12186122e-01 1.10285701e-02 8.14646780e-01
-7.68632174e-01 -9.62959751e-02 -2.71755099e-01 8.02739799e-01
4.27026361e-01 -5.19202426e-02 6.85580969e-01 -1.44498885e-01
-3.20163161e-01 2.95382321e-01 -5.82800925e-01 4.64329543e-03
7.83358097e-01 2.60697335e-01 -3.00863683e-01 -9.43543077e-01
-5.56659937e-01 6.50894523e-01 3.74504775e-01 3.13841224e-01
3.25175226e-01 -1.10496199e+00 -2.27609441e-01 -1.94184668e-02
-3.11073869e-01 -6.98849261e-01 2.52126157e-01 1.12833667e+00
1.45779803e-01 3.11265230e-01 -2.39572730e-02 1.88840181e-01
-9.94273245e-01 5.95327735e-01 3.12307417e-01 -4.80578572e-01
-4.83348936e-01 5.70291400e-01 -2.41382420e-01 -3.40920269e-01
5.70832729e-01 -3.57109696e-01 -2.68945009e-01 4.90215980e-03
3.37581366e-01 6.32134080e-01 -3.12125981e-01 6.98071793e-02
-1.71952188e-01 -1.89403072e-01 -2.16568887e-01 -1.48302495e-01
1.27990103e+00 -5.06047368e-01 1.88476846e-01 3.63237351e-01
7.24418581e-01 2.15134814e-01 -1.57560980e+00 -2.12940037e-01
-7.39619210e-02 -6.19416595e-01 3.50783825e-01 -1.00801122e+00
-1.03906441e+00 3.21158171e-01 6.12951279e-01 4.80946064e-01
1.00781727e+00 -2.66591638e-01 9.45540071e-01 6.00241601e-01
6.47807300e-01 -1.63611317e+00 5.24064116e-02 4.78921980e-01
6.44495547e-01 -1.33414900e+00 9.02397260e-02 8.54251757e-02
-9.86804545e-01 6.80868685e-01 7.75965750e-01 -2.41853848e-01
6.98719919e-01 -3.35042179e-03 -1.88265145e-01 8.75361562e-02
-1.19012308e+00 -1.47369757e-01 -1.52836800e-01 2.98141725e-02
2.83445209e-01 5.21547854e-01 -8.64186406e-01 1.05413759e+00
5.57047762e-02 1.68122739e-01 7.24050701e-01 9.53929186e-01
-6.03182912e-01 -1.23662078e+00 -2.52294183e-01 8.22201431e-01
-4.01300520e-01 -7.03985989e-02 -2.85186112e-01 3.82134765e-01
-1.83805600e-01 1.19982874e+00 -2.08531812e-01 -5.43522596e-01
2.10871309e-01 -1.70742661e-01 6.86071888e-02 -3.58731508e-01
-7.94471562e-01 2.92852461e-01 1.87154755e-01 -9.43918943e-01
-7.93690905e-02 -6.29083395e-01 -1.26024902e+00 -5.40403843e-01
-6.22528970e-01 3.62850666e-01 3.27347100e-01 6.53646350e-01
5.74605823e-01 1.72875956e-01 1.17886424e+00 -4.20499831e-01
-1.56587636e+00 -6.72171354e-01 -7.61276066e-01 2.12311924e-01
2.93814421e-01 -8.04969609e-01 -6.85732245e-01 -5.63984215e-01]
|
[4.153491020202637, 2.41606068611145]
|
7a861b7a-3ead-4cc3-bd06-6d51c87005d6
|
referring-expression-comprehension-using
|
2306.04451
| null |
https://arxiv.org/abs/2306.04451v1
|
https://arxiv.org/pdf/2306.04451v1.pdf
|
Referring Expression Comprehension Using Language Adaptive Inference
|
Different from universal object detection, referring expression comprehension (REC) aims to locate specific objects referred to by natural language expressions. The expression provides high-level concepts of relevant visual and contextual patterns, which vary significantly with different expressions and account for only a few of those encoded in the REC model. This leads us to a question: do we really need the entire network with a fixed structure for various referring expressions? Ideally, given an expression, only expression-relevant components of the REC model are required. These components should be small in number as each expression only contains very few visual and contextual clues. This paper explores the adaptation between expressions and REC models for dynamic inference. Concretely, we propose a neat yet efficient framework named Language Adaptive Dynamic Subnets (LADS), which can extract language-adaptive subnets from the REC model conditioned on the referring expressions. By using the compact subnet, the inference can be more economical and efficient. Extensive experiments on RefCOCO, RefCOCO+, RefCOCOg, and Referit show that the proposed method achieves faster inference speed and higher accuracy against state-of-the-art approaches.
|
['Xi Li', 'Yongjian Fu', 'Huanzhang Dou', 'Peihan Miao', 'Wei Su']
|
2023-06-06
| null | null | null | null |
['referring-expression']
|
['computer-vision']
|
[ 1.70107156e-01 1.63913146e-01 -2.20554218e-01 -4.73218620e-01
-2.69284040e-01 -7.47389495e-01 5.25864840e-01 -1.65387150e-02
-3.25184494e-01 4.34690267e-01 1.46885529e-01 -3.46044838e-01
-1.46814436e-01 -8.13772738e-01 -4.77116138e-01 -5.59140146e-01
8.72150734e-02 3.32726449e-01 4.30519670e-01 -3.86518657e-01
1.60813436e-01 7.95163095e-01 -1.75505233e+00 4.80756760e-01
7.13067651e-01 1.05786216e+00 4.34209079e-01 5.98876297e-01
-7.98374295e-01 1.23648202e+00 -6.81631744e-01 -4.41260010e-01
-2.65036047e-01 -6.76356971e-01 -9.59611893e-01 1.07824981e-01
6.60479903e-01 1.06582614e-02 -1.03862688e-01 1.33379376e+00
2.89475530e-01 4.02224243e-01 6.18225336e-01 -1.12824059e+00
-9.10732031e-01 7.14549363e-01 -4.86703485e-01 4.45584923e-01
6.53505623e-01 -6.91058040e-02 1.13302231e+00 -1.09149194e+00
9.57104683e-01 1.62491417e+00 2.38931715e-01 6.34635508e-01
-1.12059307e+00 -4.32319611e-01 8.61094713e-01 5.67789376e-01
-1.57750785e+00 -3.29801828e-01 8.69129121e-01 -1.68473542e-01
1.07689047e+00 4.17013735e-01 8.15446615e-01 8.81311297e-01
-2.23570719e-01 1.00274301e+00 9.61798847e-01 -6.32235229e-01
1.48855031e-01 1.90394670e-02 4.98509258e-01 1.15763056e+00
-2.57649571e-01 -3.47953379e-01 -6.05708122e-01 5.38178310e-02
7.26773739e-01 -3.98840792e-02 -4.29391295e-01 -2.07154185e-01
-1.05667031e+00 5.69156349e-01 6.19233191e-01 6.40682459e-01
-3.99789989e-01 3.35070997e-01 3.89000684e-01 2.36833930e-01
1.85340226e-01 2.76650071e-01 -1.96939081e-01 -5.42538948e-02
-4.75376159e-01 1.09612420e-01 6.35976613e-01 1.35221195e+00
9.72930491e-01 9.60200373e-03 -4.32365388e-01 9.41124797e-01
-2.25585885e-03 4.38351065e-01 6.30670860e-02 -1.07315040e+00
2.17993841e-01 1.01846623e+00 -1.00686878e-01 -1.40183306e+00
-4.95112509e-01 -3.55865866e-01 -7.90474653e-01 -1.35238349e-01
3.85254025e-01 3.22736800e-02 -7.26959109e-01 2.03196812e+00
3.55418533e-01 -1.43067822e-01 -1.97788730e-01 8.03492963e-01
1.03378940e+00 6.53139651e-01 3.56438160e-01 -1.27097458e-01
1.67994368e+00 -9.98395920e-01 -9.44218516e-01 -4.03551906e-01
6.71487629e-01 -4.76213336e-01 1.43833160e+00 1.25849172e-01
-1.07736814e+00 -6.65635526e-01 -8.61656070e-01 -3.74380857e-01
-7.70544350e-01 2.15718165e-01 9.08644140e-01 2.26516277e-01
-1.09578478e+00 -6.03335090e-02 -2.87524521e-01 -5.13862252e-01
3.65745723e-01 2.49880508e-01 -2.29816183e-01 -2.02187046e-01
-1.24396884e+00 1.05386019e+00 4.54916298e-01 4.48196083e-01
-6.78330481e-01 -3.07127059e-01 -9.35974598e-01 2.80895561e-01
8.78125489e-01 -7.87470758e-01 1.13831937e+00 -1.40756738e+00
-1.23063076e+00 1.07004523e+00 -8.43034625e-01 -1.84168220e-01
1.51377127e-01 -3.18244368e-01 -3.26534927e-01 7.26752162e-01
-2.51407046e-02 7.62941420e-01 7.13022232e-01 -1.30725360e+00
-6.33130074e-01 -2.42970780e-01 6.32055223e-01 1.74889639e-01
-9.60626304e-02 4.53272372e-01 -1.10438895e+00 -3.93681109e-01
2.22681552e-01 -6.39155924e-01 -5.28200008e-02 3.71326178e-01
-3.51274192e-01 -7.87995815e-01 8.57213557e-01 -3.80332232e-01
1.39070153e+00 -2.15494227e+00 3.56162041e-01 2.48402491e-01
4.51863229e-01 5.80877736e-02 -1.99386284e-01 1.58211738e-01
-6.89728409e-02 1.75713852e-01 -3.73803638e-02 1.73474569e-02
4.82289046e-02 5.22319078e-01 -2.45580792e-01 1.75030425e-01
1.77858323e-01 1.30709398e+00 -1.12482512e+00 -8.95393550e-01
2.02618554e-01 1.98448136e-01 -3.50029290e-01 1.64113939e-01
-5.28019786e-01 1.01482764e-01 -6.55438662e-01 9.58611965e-01
5.91412425e-01 -3.90281916e-01 1.73529685e-01 -3.49741548e-01
7.95831233e-02 -1.72779128e-01 -9.98652577e-01 1.45069480e+00
-5.32466292e-01 8.48561287e-01 7.50494823e-02 -1.13151193e+00
1.06316495e+00 -3.10368389e-02 -2.52383202e-02 -9.58551705e-01
1.12238288e-01 -1.46032378e-01 -9.06805918e-02 -9.61809516e-01
5.10312974e-01 -1.67034101e-02 -2.10877493e-01 2.37183407e-01
1.01665087e-01 6.91316053e-02 3.95209461e-01 5.51224113e-01
8.15316498e-01 1.98494017e-01 7.08077729e-01 -2.04579294e-01
9.17907715e-01 -1.84355229e-01 5.20309448e-01 1.02900100e+00
-1.61631033e-01 2.15387553e-01 8.42793584e-01 -3.83472264e-01
-4.30730283e-01 -9.66945350e-01 1.38942435e-01 1.65642869e+00
3.78605634e-01 -5.38379371e-01 -6.32652462e-01 -6.57510161e-01
-4.77955729e-01 8.47447932e-01 -8.05178821e-01 -3.02200243e-02
-7.72671461e-01 -3.69858921e-01 4.44548398e-01 8.51480603e-01
8.29833329e-01 -1.38665676e+00 -7.80087173e-01 -3.95494178e-02
-1.78474903e-01 -1.34193933e+00 -2.33002141e-01 1.36633053e-01
-6.05864942e-01 -1.10830116e+00 -3.58670115e-01 -9.13272619e-01
1.03190958e+00 2.09580094e-01 1.51732922e+00 5.93319058e-01
-9.68279466e-02 6.51346564e-01 -4.82670605e-01 -1.77076265e-01
-1.37878090e-01 -3.77720259e-02 -5.01477778e-01 5.87381870e-02
6.91154718e-01 -5.06844401e-01 -2.57798851e-01 1.74918979e-01
-7.59664834e-01 2.47527450e-01 5.00631094e-01 7.12883830e-01
6.81632519e-01 -3.50064218e-01 3.62783521e-01 -9.71343219e-01
6.90176129e-01 -2.09611341e-01 -4.98976946e-01 7.88816512e-01
-1.62314743e-01 1.27965704e-01 6.62975729e-01 -5.92537582e-01
-1.32222581e+00 3.05151604e-02 4.46457975e-02 -3.41561586e-01
-2.65370429e-01 6.33980095e-01 -3.67110163e-01 3.42358351e-02
6.13064051e-01 1.94366693e-01 -5.32724798e-01 -3.79800498e-01
8.00177872e-01 2.41621241e-01 7.93210268e-01 -9.31787372e-01
4.51504827e-01 4.78533298e-01 1.66043177e-01 -9.28736031e-01
-1.24100685e+00 -3.27705890e-01 -9.26644981e-01 -4.61813629e-01
9.04907703e-01 -5.53851247e-01 -8.15577328e-01 1.44953817e-01
-1.54278243e+00 -2.07669854e-01 -2.40676090e-01 -2.11022317e-01
-3.24963748e-01 3.72566909e-01 -3.22948188e-01 -1.05122530e+00
-1.73943728e-01 -8.49832535e-01 1.01539445e+00 4.36436415e-01
-2.43535891e-01 -1.11658537e+00 -4.34087247e-01 -1.44955397e-01
3.68067145e-01 2.87384242e-01 1.27621067e+00 -3.91987711e-01
-7.03205526e-01 -3.27994302e-02 -6.69714987e-01 2.60890555e-02
-5.03019765e-02 1.75217777e-01 -9.98513162e-01 1.65193602e-01
-1.53459132e-01 -2.54972368e-01 9.00621712e-01 1.13560766e-01
1.57154942e+00 -2.94823408e-01 -4.90188867e-01 6.40048146e-01
1.37884569e+00 3.38407099e-01 5.86410701e-01 2.83454265e-02
7.31832802e-01 7.31211603e-01 6.29987836e-01 1.51186168e-01
4.76432890e-01 7.74023950e-01 3.62704486e-01 -1.86240539e-01
-1.40886948e-01 -2.08966732e-01 2.86709875e-01 5.75620830e-01
-2.79652596e-01 -2.68434256e-01 -9.38971698e-01 7.22468615e-01
-1.95902491e+00 -1.00471997e+00 -6.61156997e-02 1.52294290e+00
6.32963419e-01 -1.49114475e-01 -2.90737897e-01 -2.02656344e-01
6.84557021e-01 2.35928982e-01 -5.70851505e-01 -4.84118253e-01
-5.83556473e-01 2.51163363e-01 -1.74237832e-01 5.31422019e-01
-7.18344688e-01 1.29765391e+00 6.64699078e+00 8.97838533e-01
-9.34841156e-01 1.07205175e-01 4.46709365e-01 1.79118291e-01
-3.22963327e-01 1.84752300e-01 -8.22637379e-01 -1.57962833e-03
6.39239848e-01 -2.09935173e-01 3.05598021e-01 9.29406226e-01
3.20043899e-02 -5.68164170e-01 -1.18055260e+00 1.12343240e+00
3.63573402e-01 -1.33474565e+00 3.33623976e-01 -4.81535941e-01
3.90104562e-01 -3.28297049e-01 -1.75610036e-01 4.01340038e-01
1.42570689e-01 -9.89601612e-01 7.36928463e-01 9.80969608e-01
5.29552877e-01 -7.82661259e-01 6.92732394e-01 1.76588640e-01
-1.44772136e+00 -1.21200187e-02 -4.46664721e-01 2.06945790e-03
1.29164442e-01 1.22661613e-01 -3.01270157e-01 4.13563639e-01
7.76659787e-01 6.63102329e-01 -9.97558832e-01 5.38162291e-01
-7.69182503e-01 3.96358132e-01 -2.03046903e-01 -4.66805398e-01
2.62169540e-01 -7.75087848e-02 4.54683363e-01 1.52853644e+00
5.48674306e-03 3.58336508e-01 8.11487883e-02 1.24545276e+00
-5.07950746e-02 1.75968125e-01 -5.49963057e-01 1.00546502e-01
5.39034843e-01 1.30212963e+00 -7.33600199e-01 -7.20585823e-01
-5.63109159e-01 9.84662890e-01 6.99676394e-01 7.69007385e-01
-7.27344275e-01 -4.84513134e-01 3.53068441e-01 -4.18664590e-02
2.85545081e-01 -1.43355072e-01 1.20333498e-02 -1.10740972e+00
-2.96328422e-02 -7.62977660e-01 7.25608349e-01 -1.28413260e+00
-1.21457911e+00 7.25238264e-01 4.26191747e-01 -8.58632088e-01
-3.78168762e-01 -9.36919808e-01 -5.15498459e-01 7.92689204e-01
-1.47891700e+00 -1.19410169e+00 -7.31941521e-01 8.99290681e-01
5.10948539e-01 7.20097348e-02 8.77205968e-01 -7.43930787e-02
-5.40179789e-01 4.74055737e-01 -6.11063659e-01 1.69207633e-01
1.59733921e-01 -1.18739510e+00 -6.86991960e-02 9.79207039e-01
4.01170582e-01 1.12749660e+00 6.00335479e-01 -1.58832088e-01
-1.22278202e+00 -8.18740904e-01 1.00868869e+00 -3.26327115e-01
5.77877164e-01 -4.75640684e-01 -1.16760862e+00 7.99317539e-01
3.29492390e-01 2.21481428e-01 3.98029119e-01 2.76372045e-01
-4.05856311e-01 -4.76891547e-03 -8.99995208e-01 9.96199310e-01
1.44224894e+00 -8.97149026e-01 -7.95607030e-01 1.92838237e-01
7.69041955e-01 -4.36123043e-01 -3.41641247e-01 8.35246518e-02
3.01594585e-01 -9.79331672e-01 9.38472211e-01 -7.62836516e-01
2.02063486e-01 -4.27703530e-01 -2.62761742e-01 -8.09311986e-01
-2.86305606e-01 -5.11368215e-01 -3.66352946e-01 1.32523561e+00
2.92897612e-01 -4.34826761e-01 4.14756119e-01 6.17770791e-01
1.25598475e-01 -7.17290521e-01 -8.27303410e-01 -6.03998303e-01
-2.76105732e-01 -7.25783288e-01 5.61916888e-01 8.68874252e-01
5.03719114e-02 5.98556280e-01 -1.63627684e-01 4.46433388e-02
1.47235274e-01 5.31481802e-01 7.17306018e-01 -9.88978505e-01
-1.95549712e-01 -4.32124555e-01 -3.98438543e-01 -1.62314951e+00
5.94934523e-01 -9.37105119e-01 -1.28753230e-01 -1.59987199e+00
4.35536563e-01 -1.78161591e-01 -2.40155920e-01 8.22840750e-01
-3.50350529e-01 -3.08417473e-02 4.36395645e-01 -5.22145368e-02
-1.00009537e+00 4.91966307e-01 1.36546147e+00 -2.29576141e-01
-8.72367993e-02 -3.99807096e-01 -7.31150866e-01 1.03876233e+00
5.63811302e-01 -2.29242697e-01 -4.64444965e-01 -5.15048921e-01
4.70590234e-01 -1.70038670e-01 6.19585752e-01 -6.26471043e-01
4.64969903e-01 -3.02981943e-01 3.10603529e-01 -7.75833309e-01
3.52553666e-01 -7.77916610e-01 -1.59810513e-01 -1.16576970e-01
-3.94780606e-01 3.51766795e-01 2.49881268e-01 3.37067127e-01
-3.94561648e-01 -4.42094296e-01 5.96609712e-01 -5.11067331e-01
-1.38479316e+00 -8.33377391e-02 -5.30366778e-01 2.00887546e-01
7.17090666e-01 -3.00377011e-01 -4.09697831e-01 -5.55801034e-01
-9.39591408e-01 2.35441014e-01 6.55849427e-02 3.58326107e-01
8.81454408e-01 -1.32620704e+00 -1.72895819e-01 -2.35758677e-01
3.83884966e-01 1.72049820e-01 2.14277759e-01 7.94156671e-01
-1.80299312e-01 4.05633152e-01 -9.41081271e-02 -7.25167394e-01
-1.41688335e+00 8.52120697e-01 6.14566863e-01 -7.97595531e-02
-6.21043444e-01 9.95273650e-01 5.24970353e-01 -2.56915271e-01
1.95214331e-01 -4.54489708e-01 -3.52757007e-01 1.98507860e-01
6.56029105e-01 9.79708806e-02 -4.90853786e-01 -9.04168010e-01
-3.40900153e-01 9.11551178e-01 7.12624043e-02 8.11641514e-02
9.84235883e-01 -3.18342566e-01 -5.21160245e-01 5.07115245e-01
9.98632669e-01 3.70006412e-02 -9.22291756e-01 -5.84073961e-01
2.17580453e-01 -3.61415237e-01 -1.23527117e-01 -7.25256085e-01
-1.15712965e+00 9.86370623e-01 3.92761022e-01 8.52177292e-02
1.58658373e+00 5.17979562e-01 4.48500738e-02 8.72189879e-01
3.46101493e-01 -9.60162461e-01 2.52120256e-01 6.43247008e-01
1.28498423e+00 -7.73683310e-01 -5.88146830e-03 -5.85925221e-01
-6.16374910e-01 1.29111278e+00 9.52325463e-01 9.34925675e-02
2.82899588e-01 2.05046743e-01 3.92045313e-03 -6.60199821e-01
-9.29728746e-01 -5.24082303e-01 3.22166800e-01 7.03566372e-01
3.87149751e-01 -1.50185257e-01 -1.03284076e-01 4.42673117e-01
-3.50129604e-02 -1.80350617e-01 7.66211376e-02 8.22265208e-01
-5.59279859e-01 -8.45438480e-01 -2.90773928e-01 1.55778155e-01
-2.12632447e-01 -2.11738691e-01 -7.15376675e-01 1.06914508e+00
2.45382890e-01 7.47305036e-01 1.25118747e-01 -1.64919794e-02
5.53099632e-01 2.51664966e-01 5.80888510e-01 -3.79093945e-01
-4.08363432e-01 -1.76248521e-01 2.58766830e-01 -8.35223019e-01
-8.82104278e-01 -4.58129346e-01 -1.56448734e+00 -6.83352053e-02
-2.46053889e-01 -1.04440257e-01 5.80585860e-02 1.12987053e+00
2.20325649e-01 5.82659900e-01 6.30871579e-02 -4.68208879e-01
1.13122642e-01 -8.06421161e-01 -3.00658345e-01 4.40008372e-01
2.33638167e-01 -7.61171937e-01 -1.65653601e-01 1.35387599e-01]
|
[10.463554382324219, 1.5058119297027588]
|
813178da-86d9-4be0-930e-5a46941eda9b
|
marginal-utility-for-planning-in-continuous
|
2006.06054
| null |
https://arxiv.org/abs/2006.06054v2
|
https://arxiv.org/pdf/2006.06054v2.pdf
|
Marginal Utility for Planning in Continuous or Large Discrete Action Spaces
|
Sample-based planning is a powerful family of algorithms for generating intelligent behavior from a model of the environment. Generating good candidate actions is critical to the success of sample-based planners, particularly in continuous or large action spaces. Typically, candidate action generation exhausts the action space, uses domain knowledge, or more recently, involves learning a stochastic policy to provide such search guidance. In this paper we explore explicitly learning a candidate action generator by optimizing a novel objective, marginal utility. The marginal utility of an action generator measures the increase in value of an action over previously generated actions. We validate our approach in both curling, a challenging stochastic domain with continuous state and action spaces, and a location game with a discrete but large action space. We show that a generator trained with the marginal utility objective outperforms hand-coded schemes built on substantial domain knowledge, trained stochastic policies, and other natural objectives for generating actions for sampled-based planners.
|
['Michael Bowling', 'Levi H. S. Lelis', 'Zaheen Farraz Ahmad']
|
2020-06-10
| null |
http://proceedings.neurips.cc/paper/2020/hash/14da15db887a4b50efe5c1bc66537089-Abstract.html
|
http://proceedings.neurips.cc/paper/2020/file/14da15db887a4b50efe5c1bc66537089-Paper.pdf
|
neurips-2020-12
|
['action-generation']
|
['computer-vision']
|
[ 6.47519052e-01 8.28486681e-01 -3.75762612e-01 -5.55282570e-02
-1.28829896e+00 -6.33876920e-01 1.09633183e+00 1.23679623e-01
-4.93082196e-01 1.46248519e+00 5.87094426e-01 -3.40458095e-01
-2.14844882e-01 -8.41438830e-01 -7.19099820e-01 -5.48847318e-01
-2.47902125e-01 9.20987070e-01 2.88409412e-01 -2.07692653e-01
4.67267811e-01 2.76033044e-01 -1.48269653e+00 3.64414193e-02
9.71599102e-01 6.22814834e-01 5.11974275e-01 8.71588171e-01
1.05378911e-01 8.50558817e-01 -7.41007805e-01 3.24826449e-01
3.88300627e-01 -8.06277335e-01 -1.05585992e+00 1.14050850e-01
-3.33130300e-01 -4.30557400e-01 -3.12904976e-02 7.80280709e-01
4.52669293e-01 6.83867097e-01 7.34099686e-01 -1.31171048e+00
1.52192377e-02 8.59510481e-01 2.65948474e-01 -1.02301061e-01
7.74419904e-01 8.49496663e-01 8.40770125e-01 -3.44593748e-02
8.10398221e-01 1.51047170e+00 8.46821591e-02 9.49063480e-01
-1.35495400e+00 -8.66839811e-02 1.25595674e-01 -2.89635509e-01
-9.65831280e-01 -3.11883420e-01 3.53975356e-01 -4.10483599e-01
1.08732522e+00 4.83239777e-02 7.46265769e-01 1.11780512e+00
2.20463097e-01 9.38242137e-01 1.09551895e+00 -3.52829695e-01
9.17296529e-01 -3.64360422e-01 -9.26680446e-01 6.50832176e-01
-7.98861980e-02 6.30376875e-01 -3.03968549e-01 -4.72647786e-01
8.26765180e-01 -2.13145241e-01 3.33164744e-02 -4.18429017e-01
-1.25737631e+00 7.40245342e-01 -5.85822947e-03 -5.32366522e-02
-7.77094841e-01 6.01707876e-01 1.56660184e-01 2.46178925e-01
2.83349212e-02 1.27540517e+00 -4.50716645e-01 -8.08082223e-01
-6.80904388e-01 9.72523630e-01 8.81305993e-01 9.73098040e-01
7.55455375e-01 2.19450697e-01 -6.33471191e-01 2.45972008e-01
1.30412942e-02 4.31832016e-01 5.26082397e-01 -1.63946366e+00
6.31666899e-01 6.14049017e-01 9.28663790e-01 -1.10164471e-01
-1.54420063e-01 7.53925294e-02 5.58012910e-02 7.30347574e-01
5.59515119e-01 -6.07391417e-01 -9.88573194e-01 1.87677550e+00
3.85865062e-01 -4.68294472e-02 3.56913120e-01 4.88151699e-01
-1.61700770e-01 7.09791422e-01 6.14148490e-02 -3.18876982e-01
5.87149560e-01 -9.60896313e-01 -4.32346523e-01 -4.80007261e-01
7.28094518e-01 -1.06347889e-01 1.13681817e+00 2.75766253e-01
-1.43780422e+00 -1.94131076e-01 -7.36851871e-01 4.60377157e-01
-4.17799614e-02 -1.61882773e-01 6.91600978e-01 3.24147731e-01
-1.03881562e+00 1.13153458e+00 -1.14216208e+00 -1.78153634e-01
4.23459560e-01 3.85239482e-01 -6.28758222e-02 -5.74184768e-02
-9.51686323e-01 1.00021255e+00 6.55070901e-01 -4.57558364e-01
-1.67552710e+00 -1.32122278e-01 -9.57046092e-01 5.09694517e-02
9.33387637e-01 -4.21342313e-01 1.88442457e+00 -5.55441141e-01
-1.99909127e+00 6.04135320e-02 1.29714891e-01 -7.02389598e-01
6.61941290e-01 -1.53664807e-02 1.44667521e-01 -1.09849557e-01
3.18808138e-01 7.11259007e-01 8.23218286e-01 -9.42625523e-01
-7.91105509e-01 8.88289735e-02 5.81126928e-01 5.94833672e-01
3.38029325e-01 -1.72097802e-01 3.07392329e-01 -1.95936903e-01
-2.97078937e-01 -1.10100925e+00 -9.81051505e-01 -3.97593558e-01
-2.96002418e-01 -3.80060107e-01 2.14602560e-01 -3.65419269e-01
1.03129923e+00 -1.64439332e+00 3.05863649e-01 2.38277495e-01
-4.19497669e-01 1.20667011e-01 -3.59066784e-01 7.26146460e-01
3.86222988e-01 -1.71785839e-02 -4.86884832e-01 8.53746086e-02
3.79608482e-01 3.83146524e-01 -1.84482247e-01 2.92005744e-02
2.26997152e-01 8.91536176e-01 -1.54166794e+00 -2.46501565e-01
-1.38166696e-02 -1.65708065e-01 -7.77982771e-01 3.08986336e-01
-1.08261180e+00 6.97439849e-01 -1.06671643e+00 5.15223861e-01
-2.45212093e-01 8.84535238e-02 3.71405989e-01 7.96056330e-01
-1.14536703e-01 6.74803853e-01 -1.22250032e+00 1.83749151e+00
-4.91034091e-01 2.02396885e-01 -5.78381084e-02 -6.57220006e-01
8.44319642e-01 3.70247424e-01 4.47144538e-01 -3.67398381e-01
-4.81771231e-02 2.86568910e-01 7.12646171e-02 -3.60026926e-01
5.68342924e-01 -2.25701690e-01 -4.03189272e-01 9.95345533e-01
-7.62781426e-02 -8.04531217e-01 4.49010968e-01 1.47040769e-01
1.69987595e+00 7.60907948e-01 5.69201589e-01 -3.64433005e-02
1.01254851e-01 3.60709995e-01 5.50356567e-01 1.18418169e+00
-1.96907744e-01 3.58846992e-01 7.63336360e-01 -2.78828293e-01
-9.51577246e-01 -9.34819996e-01 5.27215421e-01 9.31755602e-01
-1.12799264e-01 -3.84433240e-01 -7.96823978e-01 -8.79159749e-01
-1.60778388e-01 1.12453866e+00 -5.00333190e-01 -1.82955503e-01
-8.27923417e-01 -2.68434286e-01 3.11784774e-01 5.34439325e-01
2.22924814e-01 -1.63169098e+00 -1.20175767e+00 6.55619442e-01
-6.41037077e-02 -5.95560014e-01 -6.57945037e-01 2.84724891e-01
-9.23908770e-01 -1.08070719e+00 -5.39322555e-01 -3.16367775e-01
6.49213552e-01 -4.98010397e-01 1.04572392e+00 -2.22023621e-01
-2.46121645e-01 4.42638487e-01 -2.64057755e-01 -3.65789711e-01
-9.32245016e-01 3.21759358e-02 1.27605036e-01 -5.60558438e-01
-1.87222317e-01 -5.81527174e-01 -3.50614429e-01 1.05119571e-01
-7.80431688e-01 1.74550861e-01 6.18675053e-01 1.02418125e+00
3.41669083e-01 2.72118956e-01 5.34629643e-01 -4.66737658e-01
1.12991071e+00 -3.46850187e-01 -7.27309465e-01 1.00234076e-01
-4.01606202e-01 8.55680287e-01 5.57584226e-01 -6.30170941e-01
-1.21184373e+00 3.67487103e-01 2.20170781e-01 1.43345043e-01
-2.75182217e-01 3.30712736e-01 -2.93762565e-01 5.35420597e-01
1.06222677e+00 2.23436788e-01 1.17495686e-01 -2.40787208e-01
4.94619131e-01 2.26730064e-01 3.78871500e-01 -1.35709357e+00
5.84608555e-01 1.54761299e-01 8.12636241e-02 -2.95934021e-01
-3.36487919e-01 2.17442028e-02 -3.04808974e-01 -1.00804739e-01
6.97625160e-01 -4.48062062e-01 -6.79400980e-01 4.02254127e-02
-8.68445814e-01 -1.27815866e+00 -9.29560721e-01 4.04629111e-01
-1.43466365e+00 -3.75057250e-01 -1.78510293e-01 -1.21867788e+00
2.42804632e-01 -1.41502845e+00 1.11435986e+00 3.21152717e-01
-5.59569597e-01 -7.66567826e-01 3.10680568e-01 -1.65298328e-01
3.44565749e-01 6.27657115e-01 6.14506900e-01 -5.65781534e-01
-7.75349140e-01 -1.15375258e-01 4.90373164e-01 -2.59721335e-02
3.37432414e-01 -3.61422598e-01 -4.03780609e-01 -2.39829332e-01
-2.72448450e-01 -6.91668630e-01 3.46343577e-01 3.15343171e-01
9.39355850e-01 -7.05127299e-01 -4.35999304e-01 5.08128628e-02
1.06194878e+00 7.16777265e-01 4.69439298e-01 4.54351604e-01
3.44120637e-02 4.29748446e-01 1.08532107e+00 8.06712329e-01
1.70014173e-01 7.05638409e-01 3.90587389e-01 5.10621786e-01
1.67782813e-01 -7.64231026e-01 5.85225880e-01 -3.76647949e-01
-3.54470789e-01 -1.20271452e-01 -8.38555396e-01 7.33785212e-01
-2.03310037e+00 -1.22118449e+00 8.11933100e-01 2.22063375e+00
1.19508028e+00 2.99593896e-01 5.52977443e-01 -8.43738988e-02
5.04298627e-01 9.35229883e-02 -1.03179443e+00 -4.16573107e-01
3.60282481e-01 3.11022997e-01 5.40736675e-01 8.27548087e-01
-7.57309198e-01 1.01095128e+00 6.67810059e+00 7.43983030e-01
-5.04803956e-01 -2.11978912e-01 4.36992377e-01 -3.43566328e-01
-3.93483639e-01 3.01544905e-01 -6.52149737e-01 4.12697196e-01
9.49490070e-01 -4.53622729e-01 9.11537051e-01 1.09719849e+00
6.08178973e-01 -4.63976383e-01 -1.30993247e+00 1.26987532e-01
-5.34405351e-01 -1.37542582e+00 -4.07295793e-01 3.88636142e-01
9.57055688e-01 -2.53217012e-01 -1.17211744e-01 5.72742343e-01
1.36234438e+00 -1.13491762e+00 8.60958219e-01 3.85105342e-01
6.58991575e-01 -8.67768645e-01 2.88849235e-01 8.41414273e-01
-8.41930091e-01 -3.02541912e-01 -5.93776405e-02 -4.88308579e-01
4.78852034e-01 -1.33995354e-01 -1.35044348e+00 2.03932244e-02
-4.81104478e-02 1.19058132e-01 -1.18551947e-01 8.77913058e-01
-5.42606890e-01 5.19692183e-01 -3.74963909e-01 -4.30506021e-01
6.90525115e-01 -2.08647385e-01 6.10254467e-01 6.80346429e-01
4.59471494e-01 3.35268974e-01 6.79446280e-01 1.01998067e+00
5.13350844e-01 -5.38375795e-01 -9.17843878e-01 -4.57528442e-01
5.76196671e-01 6.51892841e-01 -6.36128604e-01 -3.96045655e-01
3.21111381e-01 6.93212986e-01 1.71011388e-01 4.63552654e-01
-7.43737936e-01 -2.49501362e-01 6.56901538e-01 9.72544178e-02
4.19170372e-02 -3.04985464e-01 8.74375850e-02 -6.59227133e-01
-1.61326736e-01 -9.85340893e-01 1.24906063e-01 -6.67802811e-01
-5.22764564e-01 3.76529872e-01 5.18333375e-01 -1.12114882e+00
-1.36563158e+00 -2.29557544e-01 -6.96599305e-01 8.71288180e-01
-8.46740901e-01 -5.85598230e-01 1.72912523e-01 2.26822063e-01
8.50680709e-01 -3.54373723e-01 8.03508520e-01 -6.78796172e-01
-1.24342643e-01 9.83758569e-02 -7.47405663e-02 -3.33620936e-01
1.09205879e-01 -1.56135714e+00 6.90478444e-01 7.01153636e-01
-2.40904748e-01 3.98221463e-01 1.00615096e+00 -8.80787671e-01
-1.32411706e+00 -8.68166029e-01 4.74551857e-01 -5.39264083e-01
4.67393100e-01 -1.45007610e-01 -3.19357097e-01 6.66728973e-01
-9.43233073e-02 -3.47330570e-01 -3.72757874e-02 -3.01685601e-01
3.16984296e-01 4.58074331e-01 -1.52360833e+00 9.04021978e-01
1.30014575e+00 -2.10017145e-01 -6.17901981e-01 4.83500481e-01
7.92066813e-01 -6.36145473e-01 -4.93995458e-01 1.57435052e-02
3.10938388e-01 -8.77855003e-01 7.71933556e-01 -9.39639807e-01
4.16029990e-01 -3.25045526e-01 -4.58006412e-02 -1.85626006e+00
-1.53548688e-01 -1.31907868e+00 -1.80797160e-01 7.43788362e-01
5.41978538e-01 -4.41837460e-01 1.03121281e+00 1.07022631e+00
-2.69354433e-01 -8.64806473e-01 -7.43606985e-01 -1.06883669e+00
3.19495983e-02 -3.79416823e-01 1.10569572e+00 1.33499131e-01
3.45262557e-01 -6.62022550e-03 -2.09530577e-01 -1.79479301e-01
4.47001010e-01 -2.42876634e-03 8.72767687e-01 -7.03476846e-01
-6.25055850e-01 -4.14167792e-01 -2.60952134e-02 -1.11391997e+00
3.04717869e-01 -4.31883782e-01 7.71710694e-01 -1.74130642e+00
-2.99108654e-01 -4.98134196e-01 2.60456115e-01 5.84429860e-01
-7.99916610e-02 -6.56033576e-01 1.68481246e-01 -1.65576354e-01
-6.97242916e-01 7.66840219e-01 1.50360119e+00 -5.55839948e-02
-6.54300570e-01 3.47548068e-01 -5.88254094e-01 7.87266076e-01
8.73307645e-01 -3.12959105e-01 -8.34593117e-01 1.87637862e-02
1.85571089e-01 6.73021138e-01 -2.82077529e-02 -9.20515358e-01
-5.24767116e-02 -1.17011976e+00 -1.09575517e-01 -5.81734963e-02
3.70552838e-01 -3.79013896e-01 5.88611737e-02 7.17396021e-01
-7.90408850e-01 1.69902503e-01 -1.32310957e-01 7.56902218e-01
2.04951782e-02 -4.12098080e-01 4.67582822e-01 -7.06711531e-01
-7.75914609e-01 1.85236409e-01 -6.78927481e-01 4.49014843e-01
1.07159340e+00 -2.40562648e-01 4.61170897e-02 -6.83597922e-01
-8.60883117e-01 4.01006550e-01 5.98844111e-01 4.26498204e-02
6.23446822e-01 -1.35187364e+00 -5.46790123e-01 5.71507262e-03
-4.62044813e-02 2.81576067e-01 -4.82336789e-01 2.60019183e-01
-1.67315900e-01 2.48796985e-01 -1.99977040e-01 3.74736730e-03
-6.93398118e-01 3.06454599e-01 3.01949263e-01 -5.40861249e-01
-5.79569519e-01 7.22345829e-01 -1.84568301e-01 -4.60104853e-01
1.66383535e-01 -5.13474703e-01 5.77686317e-02 -4.10731643e-01
5.65130055e-01 6.25711620e-01 -4.97403651e-01 -1.64952368e-01
-1.14422537e-01 -2.18338862e-01 4.81964320e-01 -1.02523494e+00
1.31870389e+00 2.04202920e-01 3.77401769e-01 2.22065359e-01
4.48583543e-01 -3.36565495e-01 -2.07124829e+00 -1.91247405e-03
1.77361578e-01 -6.15482807e-01 -4.02471304e-01 -9.13909912e-01
-2.41724312e-01 2.29850441e-01 -3.48022543e-02 2.49021545e-01
7.62124658e-01 -8.86877552e-02 5.64008832e-01 6.45072162e-01
1.10610688e+00 -1.37146032e+00 4.46193486e-01 6.37206972e-01
9.12930429e-01 -1.03656220e+00 -1.54386163e-01 2.97067523e-01
-1.00289202e+00 7.21698105e-01 6.87358558e-01 -1.57984450e-01
3.60333249e-02 2.93072581e-01 -3.63361537e-01 -8.02352745e-03
-1.04298937e+00 -3.77045482e-01 -6.90735057e-02 9.79230106e-01
-6.36832928e-03 2.04873711e-01 -1.80577710e-01 4.80788685e-02
-4.37825143e-01 2.25348577e-01 7.04614520e-01 1.42516959e+00
-7.78905094e-01 -1.45364964e+00 -3.39427412e-01 5.91371179e-01
-5.47857117e-03 1.84400126e-01 -3.74106258e-01 5.88268280e-01
-2.65110657e-02 1.00265872e+00 -6.76132068e-02 -1.08329952e-01
9.67978090e-02 2.24574476e-01 7.01315403e-01 -1.13985443e+00
-3.22751284e-01 -1.02108590e-01 6.11508131e-01 -1.00494754e+00
-1.58820301e-01 -1.06870019e+00 -1.54969501e+00 1.36935279e-01
-1.39532406e-02 4.43071812e-01 5.66460371e-01 1.15182257e+00
1.55830592e-01 4.82552916e-01 5.15837252e-01 -1.11386931e+00
-1.10973132e+00 -8.74332190e-01 -4.51927722e-01 2.36167789e-01
3.33209723e-01 -7.83146024e-01 -1.29917592e-01 -1.16448253e-01]
|
[4.141141891479492, 1.7359895706176758]
|
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.