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
|
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
b1e41cbe-2193-49c3-b0c1-ca3e40469022
|
answer-me-multi-task-open-vocabulary-visual
|
2205.00949
| null |
https://arxiv.org/abs/2205.00949v2
|
https://arxiv.org/pdf/2205.00949v2.pdf
|
Answer-Me: Multi-Task Open-Vocabulary Visual Question Answering
|
We present Answer-Me, a task-aware multi-task framework which unifies a variety of question answering tasks, such as, visual question answering, visual entailment, visual reasoning. In contrast to previous works using contrastive or generative captioning training, we propose a novel and simple recipe to pre-train a vision-language joint model, which is multi-task as well. The pre-training uses only noisy image captioning data, and is formulated to use the entire architecture end-to-end with both a strong language encoder and decoder. Our results show state-of-the-art performance, zero-shot generalization, robustness to forgetting, and competitive single-task results across a variety of question answering tasks. Our multi-task mixture training learns from tasks of various question intents and thus generalizes better, including on zero-shot vision-language tasks. We conduct experiments in the challenging multi-task and open-vocabulary settings and across a variety of datasets and tasks, such as VQA2.0, SNLI-VE, NLVR2, GQA. We observe that the proposed approach is able to generalize to unseen tasks and that more diverse mixtures lead to higher accuracy in both known and novel tasks.
|
['Anelia Angelova', 'Fred Bertsch', 'Mohammad Saffar', 'Weicheng Kuo', 'Wei Li', 'AJ Piergiovanni']
|
2022-05-02
| null | null | null | null |
['visual-entailment']
|
['reasoning']
|
[ 7.46440738e-02 -3.67168710e-02 1.15748584e-01 -3.13152164e-01
-1.25442660e+00 -5.18544495e-01 9.13396239e-01 -1.54613718e-01
-3.88272882e-01 5.95013320e-01 2.80211151e-01 -4.10510331e-01
1.92940950e-01 -4.10692066e-01 -1.04749823e+00 -4.65070158e-01
7.10850835e-01 7.11191297e-01 3.48155916e-01 -2.87988126e-01
-1.01839878e-01 -1.94034889e-01 -1.54362082e+00 7.50913799e-01
7.07664609e-01 9.74634171e-01 4.29793537e-01 1.03190291e+00
-3.61501515e-01 1.32009804e+00 -5.94827831e-01 -9.32790756e-01
-1.15611777e-01 -4.31935340e-01 -9.59945261e-01 2.94621825e-01
1.09012592e+00 -3.01538974e-01 -4.31839168e-01 6.58437192e-01
6.91910565e-01 3.40685695e-01 7.90347576e-01 -1.42000759e+00
-1.73666537e+00 -1.00183927e-01 -5.17701745e-01 2.28465468e-01
4.10722405e-01 6.14028096e-01 1.05958974e+00 -1.33153629e+00
6.81592643e-01 1.71306384e+00 3.61229032e-01 8.10820460e-01
-1.27095461e+00 -2.65910059e-01 2.22729519e-01 6.05686605e-01
-1.13566923e+00 -4.16425526e-01 4.99715775e-01 -5.23046613e-01
1.05656910e+00 9.35863927e-02 -4.17566486e-02 1.79076827e+00
3.00041437e-01 1.18692911e+00 1.18586230e+00 -2.78718203e-01
1.49840310e-01 2.06641689e-01 1.84981689e-01 7.84306407e-01
-1.75333694e-02 -4.36137825e-01 -3.59988123e-01 -1.31320819e-01
4.13654268e-01 1.81337029e-01 -3.31304938e-01 -4.65568990e-01
-1.34076929e+00 1.01101053e+00 3.06404948e-01 -2.87629068e-02
-3.82378370e-01 4.38302785e-01 4.29583579e-01 3.94452989e-01
6.19094312e-01 1.03879131e-01 -3.02159697e-01 3.32960308e-01
-7.87197709e-01 3.96661431e-01 7.06145465e-01 1.03093588e+00
8.94147813e-01 2.53611326e-01 -1.07784033e+00 1.00991333e+00
3.60973775e-01 7.47971714e-01 3.92865032e-01 -1.00841796e+00
6.38868809e-01 2.65867054e-01 9.67066512e-02 -4.09677207e-01
-2.36730710e-01 -2.87116140e-01 -9.45524573e-01 1.82834417e-01
2.99516797e-01 -1.23429619e-01 -1.69197273e+00 1.74672759e+00
3.12733114e-01 4.31299545e-02 3.75577897e-01 8.91768157e-01
1.51212966e+00 9.17963207e-01 4.30679679e-01 2.81802982e-01
1.72901654e+00 -1.81048226e+00 -8.52877259e-01 -6.89876676e-01
1.97164193e-01 -6.26720011e-01 1.57656777e+00 9.87441689e-02
-1.15480673e+00 -1.04105842e+00 -7.63879061e-01 -7.13179410e-01
-5.21500945e-01 4.23272066e-02 2.96537459e-01 3.16440314e-01
-1.26003551e+00 -2.13788360e-01 -3.23951721e-01 -3.89734179e-01
5.56599200e-01 -3.47058594e-01 -3.56668562e-01 -7.01140046e-01
-1.02636600e+00 1.25697243e+00 3.24641258e-01 -1.83466509e-01
-1.65987098e+00 -6.62942648e-01 -1.17443192e+00 6.34972230e-02
6.25045478e-01 -1.54297173e+00 1.43443465e+00 -8.93907547e-01
-1.15463150e+00 1.07199812e+00 -4.04957771e-01 -4.79765058e-01
4.49311614e-01 -2.38032728e-01 -3.31898481e-01 2.65459001e-01
2.83273667e-01 9.26400185e-01 1.19268274e+00 -1.41395819e+00
-3.50272000e-01 -3.04698139e-01 3.87876987e-01 1.91373765e-01
8.22781771e-02 -1.93621367e-01 -5.89319110e-01 -4.54066336e-01
-6.26677752e-01 -6.70978367e-01 -4.73444499e-02 1.47167474e-01
-1.64040148e-01 -4.91126448e-01 8.88233304e-01 -8.64885330e-01
5.39194047e-01 -1.84229732e+00 3.84770989e-01 -7.37205923e-01
3.59420836e-01 2.61517733e-01 -6.25696898e-01 4.66754943e-01
9.63220447e-02 -1.07180797e-01 -3.38793933e-01 -8.36620986e-01
2.08502486e-01 5.19126415e-01 -5.32548845e-01 2.79428601e-01
4.37686384e-01 1.68493211e+00 -8.13198805e-01 -5.81175387e-01
2.46194243e-01 4.02609259e-01 -2.64812678e-01 5.25043905e-01
-8.23205709e-01 3.23922664e-01 -1.10951483e-01 8.10139120e-01
5.39871395e-01 -8.88503134e-01 -4.47927386e-01 -4.97038811e-02
4.44233954e-01 -2.75757641e-01 -7.44352579e-01 2.12230206e+00
-7.44628787e-01 7.95544922e-01 9.78255793e-02 -9.30724978e-01
7.06782579e-01 4.71608251e-01 -3.02848548e-01 -9.81638074e-01
-1.07088596e-01 -2.54018754e-01 -4.01374936e-01 -1.04474974e+00
4.87222522e-01 -1.40842691e-01 -8.02954659e-02 2.19790220e-01
8.97187710e-01 -8.30131546e-02 2.59576410e-01 4.39738601e-01
7.75181413e-01 3.52807283e-01 4.07660395e-01 6.70882612e-02
4.43734676e-01 2.91199740e-02 -8.98400024e-02 8.94474268e-01
-1.53363988e-01 8.03498685e-01 3.04692745e-01 -1.84857577e-01
-1.07132375e+00 -1.43457222e+00 2.68544734e-01 1.77281260e+00
1.56191736e-01 -6.34815767e-02 -2.66816050e-01 -6.88824892e-01
1.89369619e-02 1.18389130e+00 -9.05988395e-01 -7.69085586e-02
-1.35491222e-01 -5.22841632e-01 5.01338363e-01 4.94642913e-01
5.59501112e-01 -1.24178779e+00 -3.80034626e-01 3.53277847e-02
-3.05264950e-01 -1.53005755e+00 -4.06011492e-01 -1.31664529e-01
-4.05410677e-01 -8.73912811e-01 -1.34448135e+00 -9.74192798e-01
2.41438374e-01 4.65058386e-01 1.56733072e+00 -2.24549681e-01
-3.13987553e-01 1.00409377e+00 -2.59700328e-01 -2.78335214e-01
-4.08166349e-01 -2.86776066e-01 -6.81881487e-01 2.09712401e-01
2.12298706e-01 -1.51824087e-01 -5.80547988e-01 6.82729110e-02
-1.08914578e+00 1.25274330e-01 8.04961920e-01 1.06888127e+00
6.31811917e-01 -1.05399430e+00 8.55141461e-01 -6.90784037e-01
8.32118273e-01 -8.62553120e-01 -1.48283571e-01 9.04058039e-01
-2.36652449e-01 2.79462814e-01 5.18003583e-01 -6.67797983e-01
-1.27388465e+00 -2.78645605e-01 -1.77985489e-01 -1.01826692e+00
-4.15172756e-01 1.10584766e-01 -1.56361252e-01 1.87888131e-01
7.79472828e-01 5.41417956e-01 -1.58685088e-01 -4.01168406e-01
1.31559479e+00 4.55281585e-01 8.33469152e-01 -3.96039367e-01
5.44349134e-01 6.06517434e-01 -2.43729904e-01 -8.63359034e-01
-1.05994892e+00 -7.18917131e-01 -2.29163498e-01 -1.80927008e-01
1.51353908e+00 -1.34043443e+00 -6.99017346e-01 3.29492599e-01
-1.62634313e+00 -4.29001600e-01 -7.96614587e-02 3.10096936e-03
-5.39283812e-01 5.62889397e-01 -5.00022233e-01 -8.32969785e-01
-7.74784088e-01 -1.14611483e+00 1.35379875e+00 2.70958126e-01
1.44155785e-01 -1.17590976e+00 2.43254066e-01 8.60638797e-01
4.71643925e-01 2.34882429e-01 9.96811092e-01 -7.69088209e-01
-7.31843591e-01 4.31703568e-01 -6.35354340e-01 5.09446323e-01
-3.46794575e-01 -4.54190314e-01 -1.32808971e+00 -2.51375794e-01
6.77673221e-02 -1.16875124e+00 1.46562314e+00 2.03981042e-01
1.04308939e+00 -2.05365375e-01 -1.36877656e-01 5.01340151e-01
1.47973251e+00 -1.68047220e-01 6.99842334e-01 1.37523636e-02
9.98777628e-01 6.37438178e-01 3.01058650e-01 2.17047520e-02
9.98521745e-01 4.12562877e-01 6.89200997e-01 -2.23522395e-01
-6.03850484e-01 -2.41213486e-01 3.73826087e-01 4.26929533e-01
3.17980647e-01 -6.30589247e-01 -8.76855135e-01 1.03513193e+00
-2.02367806e+00 -9.95425403e-01 -3.11867535e-01 1.58085954e+00
6.90462887e-01 -3.61755818e-01 8.34923238e-02 -6.26832426e-01
5.15905797e-01 5.67815661e-01 -7.29858935e-01 -4.41191316e-01
-2.65059501e-01 2.27864400e-01 2.03950942e-01 4.62471575e-01
-1.05100739e+00 1.03062296e+00 5.90643692e+00 8.22939515e-01
-6.64557576e-01 9.27244723e-01 5.95075071e-01 -1.01716429e-01
-5.76159894e-01 -1.29586428e-01 -6.37358189e-01 7.55290538e-02
9.09812570e-01 4.24150825e-02 2.46452495e-01 8.93406272e-01
-5.33182383e-01 -8.24119002e-02 -1.12329364e+00 1.27171886e+00
9.83780921e-01 -1.41490579e+00 4.52507615e-01 -5.75582325e-01
8.46760571e-01 2.87664741e-01 2.42615879e-01 9.28680062e-01
3.54544133e-01 -1.15674520e+00 8.18940282e-01 6.26620829e-01
8.57095540e-01 -2.45934948e-01 4.45052952e-01 3.88397574e-01
-8.98535848e-01 2.96510365e-02 -3.63100380e-01 1.46684870e-01
5.79127848e-01 1.62593022e-01 -6.34302557e-01 7.06966162e-01
6.28256023e-01 2.68436998e-01 -9.71017957e-01 9.50007319e-01
-3.55683625e-01 2.76439965e-01 3.14284116e-01 -1.63717270e-01
3.64500016e-01 2.07267165e-01 3.29337984e-01 1.28739130e+00
6.04399145e-02 -2.48837352e-01 1.18607134e-01 1.08344865e+00
-2.23509416e-01 -3.44918594e-02 -5.29994249e-01 2.21218184e-01
-1.07854344e-01 1.25725436e+00 -2.32808620e-01 -6.91052914e-01
-8.32687557e-01 1.56237388e+00 5.62474310e-01 9.39639926e-01
-1.12453759e+00 -2.34269187e-01 4.47618544e-01 -3.03082883e-01
6.26846492e-01 -2.19100416e-01 -2.34234836e-02 -1.46511471e+00
6.75032958e-02 -8.72653842e-01 5.56351542e-01 -1.32232058e+00
-1.65685534e+00 7.38959908e-01 1.84656307e-01 -6.96779728e-01
-5.36857188e-01 -9.44989860e-01 -5.68288088e-01 9.38599467e-01
-1.94986117e+00 -1.68424141e+00 -6.27151251e-01 8.90814900e-01
1.05168605e+00 -1.23252869e-01 7.24697828e-01 2.49652252e-01
-2.17600480e-01 4.37484562e-01 -3.97422910e-02 -5.52866906e-02
1.06149948e+00 -1.32785094e+00 6.27479017e-01 7.94157982e-01
5.06866455e-01 4.89770249e-02 6.44128919e-01 -3.31282020e-01
-1.45517981e+00 -1.31914556e+00 5.94078898e-01 -9.70548213e-01
6.57997608e-01 -7.37994790e-01 -1.13752413e+00 9.17749524e-01
9.35646832e-01 1.17673792e-01 5.44551671e-01 -1.28647059e-01
-7.19929218e-01 2.12644368e-01 -7.80710578e-01 6.23921633e-01
8.55538070e-01 -9.04464543e-01 -1.09581351e+00 6.15646422e-01
1.16330147e+00 -3.34359050e-01 -5.45415401e-01 1.42881393e-01
2.27624893e-01 -7.35424519e-01 1.30857253e+00 -9.04286563e-01
6.89486623e-01 -7.65679628e-02 -3.43813002e-01 -1.23853827e+00
-4.23647195e-01 -2.68716007e-01 -4.66986328e-01 1.11151195e+00
5.30128777e-01 -3.15280348e-01 2.56613553e-01 2.39577204e-01
-2.37969890e-01 -6.60938025e-01 -9.26743746e-01 -5.81422091e-01
9.27575827e-02 -3.40202838e-01 7.94142485e-02 5.86266160e-01
-7.79177070e-01 1.22701585e+00 -9.11639154e-01 7.76341483e-02
7.67379820e-01 3.07317466e-01 1.02730250e+00 -8.72233152e-01
-6.33947134e-01 -2.19031930e-01 5.13507426e-02 -1.28263438e+00
2.99757659e-01 -9.33393478e-01 2.81980932e-02 -2.27734423e+00
3.58740717e-01 5.40099561e-01 -9.42848846e-02 4.57442790e-01
-6.33440435e-01 2.54337698e-01 3.78201216e-01 -5.57546131e-03
-1.35841596e+00 9.68004167e-01 1.53148127e+00 -5.16614974e-01
1.58450812e-01 -3.93866509e-01 -6.26121879e-01 4.75417882e-01
3.45205337e-01 -2.23915502e-01 -5.98829687e-01 -9.45576787e-01
1.09015837e-01 1.88819587e-01 9.24592614e-01 -6.46566212e-01
1.64683685e-01 4.53252904e-02 4.39387530e-01 -6.39091492e-01
7.82502174e-01 -3.55533689e-01 -2.72962511e-01 1.32863745e-01
-2.77785689e-01 1.11739501e-01 3.16094428e-01 9.13580537e-01
-1.46762729e-01 -1.96321175e-01 5.39613128e-01 -2.15489298e-01
-1.30704486e+00 3.62027228e-01 -1.59667563e-02 6.36113524e-01
1.10764420e+00 8.66543055e-02 -8.71165752e-01 -6.38137996e-01
-6.78892434e-01 6.53910637e-01 9.75551978e-02 9.06312227e-01
8.16530168e-01 -1.28966379e+00 -1.21929908e+00 -2.68163025e-01
7.47316360e-01 -1.18898645e-01 8.10967207e-01 6.41268909e-01
-2.00010702e-01 5.12091577e-01 -2.34145895e-01 -9.16012883e-01
-9.96672273e-01 1.14780927e+00 1.99596807e-01 -2.42663279e-01
-4.32319492e-01 9.48697805e-01 7.49617517e-01 -3.02610755e-01
1.66186765e-01 -3.55847210e-01 -2.06284255e-01 1.36930779e-01
6.71939373e-01 1.52920529e-01 -1.94923714e-01 -4.72657949e-01
-2.59825647e-01 4.32826608e-01 -7.22629055e-02 -2.16448709e-01
9.14786339e-01 -3.56452852e-01 1.71907306e-01 7.49206901e-01
1.35483277e+00 -5.73891342e-01 -1.40037942e+00 -3.38241935e-01
-2.20889717e-01 -1.47167712e-01 -2.37694308e-01 -8.91933322e-01
-6.37185812e-01 1.31774998e+00 4.97364610e-01 -3.40347737e-02
8.75528991e-01 5.75396538e-01 6.61457241e-01 4.92156357e-01
1.58502888e-02 -6.53404593e-01 6.75997138e-01 5.45186579e-01
1.26912475e+00 -1.75182760e+00 -4.36230123e-01 2.26937771e-01
-1.26391184e+00 6.34280264e-01 8.08200419e-01 -8.24820027e-02
1.42605528e-01 -3.02427560e-01 9.44580138e-02 -2.96432912e-01
-1.23845291e+00 -6.56301558e-01 6.90490007e-01 7.75130093e-01
5.23721240e-02 -1.92365333e-01 1.80984482e-01 5.40095747e-01
4.06259328e-01 -5.46015427e-02 2.35774532e-01 5.64391792e-01
-3.77992928e-01 -5.23184121e-01 -3.27177137e-01 2.20986724e-01
-3.94634426e-01 -3.18477243e-01 -1.53149411e-01 8.14051867e-01
-3.33012417e-02 1.13965654e+00 3.39358523e-02 -1.91011205e-02
3.30482721e-01 5.75523317e-01 6.46610200e-01 -7.09267855e-01
-3.64852160e-01 -3.15451086e-01 1.66815430e-01 -4.94836897e-01
-4.43415970e-01 -1.77355371e-02 -7.52615631e-01 2.37613872e-01
-7.97280446e-02 -2.70070821e-01 5.29185534e-01 1.12919486e+00
4.55886543e-01 8.06231797e-01 -1.47105172e-01 -4.97152835e-01
-6.46895170e-01 -1.23423684e+00 -9.02564824e-02 7.35519290e-01
6.78784430e-01 -6.68115795e-01 -1.98590964e-01 1.77970767e-01]
|
[10.851680755615234, 1.6480261087417603]
|
4f89ba03-4126-4308-a72d-820f05c6c338
|
weakly-supervised-3d-human-pose-and-shape
|
2003.10350
| null |
https://arxiv.org/abs/2003.10350v2
|
https://arxiv.org/pdf/2003.10350v2.pdf
|
Weakly Supervised 3D Human Pose and Shape Reconstruction with Normalizing Flows
|
Monocular 3D human pose and shape estimation is challenging due to the many degrees of freedom of the human body and thedifficulty to acquire training data for large-scale supervised learning in complex visual scenes. In this paper we present practical semi-supervised and self-supervised models that support training and good generalization in real-world images and video. Our formulation is based on kinematic latent normalizing flow representations and dynamics, as well as differentiable, semantic body part alignment loss functions that support self-supervised learning. In extensive experiments using 3D motion capture datasets like CMU, Human3.6M, 3DPW, or AMASS, as well as image repositories like COCO, we show that the proposed methods outperform the state of the art, supporting the practical construction of an accurate family of models based on large-scale training with diverse and incompletely labeled image and video data.
|
['Rahul Sukthankar', 'Eduard Gabriel Bazavan', 'Hongyi Xu', 'Bill Freeman', 'Cristian Sminchisescu', 'Andrei Zanfir']
|
2020-03-23
| null |
https://www.ecva.net/papers/eccv_2020/papers_ECCV/html/6296_ECCV_2020_paper.php
|
https://www.ecva.net/papers/eccv_2020/papers_ECCV/papers/123510460.pdf
|
eccv-2020-8
|
['3d-human-pose-and-shape-estimation']
|
['computer-vision']
|
[-1.51105821e-01 6.39638007e-02 -7.34867036e-01 -2.15583548e-01
-4.66545820e-01 -5.37743986e-01 5.55488944e-01 -5.30425787e-01
-6.23403013e-01 7.23250091e-01 4.23759341e-01 3.66898388e-01
4.59531583e-02 2.52776016e-02 -8.35940659e-01 -4.01367068e-01
-4.36281651e-01 8.07808220e-01 1.47008851e-01 1.26627088e-02
-2.79422313e-01 3.37665200e-01 -1.45107090e+00 -1.63986936e-01
2.81290382e-01 8.22627544e-01 -5.39677367e-02 8.99328291e-01
3.92333895e-01 7.38671064e-01 -6.85779303e-02 -2.42137030e-01
6.37474179e-01 -3.71825606e-01 -8.05920660e-01 7.85090029e-01
1.04906034e+00 -4.25260246e-01 -4.65262771e-01 5.23308694e-01
6.62964582e-01 2.15008914e-01 7.81617224e-01 -1.60680234e+00
-9.63441506e-02 -5.57418168e-02 -4.29812968e-01 2.12157264e-01
6.19685590e-01 4.03144628e-01 7.25651741e-01 -7.68798649e-01
1.20413721e+00 1.21147931e+00 8.61043692e-01 7.96085954e-01
-1.07906067e+00 -4.36777681e-01 1.33619364e-02 1.38186440e-01
-1.16868651e+00 -6.24533713e-01 7.02109814e-01 -6.75178826e-01
7.94063807e-01 -2.05946237e-01 9.90950108e-01 1.50155127e+00
1.75370172e-01 1.18219900e+00 6.74381375e-01 -2.77230769e-01
1.01201028e-01 -3.43639344e-01 -2.14578003e-01 1.04321718e+00
3.49519193e-01 3.57251287e-01 -7.93865979e-01 -1.39844954e-01
1.22868049e+00 -4.78122234e-02 -1.08425781e-01 -1.46026456e+00
-1.40541804e+00 5.58372259e-01 2.44807199e-01 -7.81743601e-02
-2.27735251e-01 4.08799529e-01 5.09166300e-01 2.28146985e-01
3.24948192e-01 4.89378572e-02 -5.72258234e-01 -3.69566083e-01
-9.48102057e-01 5.82733631e-01 6.82944894e-01 1.29146230e+00
3.88335288e-01 1.41185403e-01 3.13720912e-01 5.76357722e-01
3.86775434e-01 8.23073566e-01 6.74877465e-01 -1.46380007e+00
6.27678216e-01 4.42938983e-01 2.53925949e-01 -9.38778400e-01
-7.12831736e-01 -3.41972053e-01 -7.06377089e-01 3.20732594e-01
7.18205690e-01 -1.55123115e-01 -9.38339829e-01 1.70389044e+00
5.85009754e-01 1.94678515e-01 -1.28887519e-01 1.21214175e+00
5.21606565e-01 1.29394531e-01 6.66170493e-02 -1.89306438e-02
8.80249977e-01 -1.17829669e+00 -4.75997239e-01 -5.84655762e-01
8.38560641e-01 -3.94772887e-01 7.66372502e-01 2.67165035e-01
-1.15676236e+00 -9.19014037e-01 -8.67645621e-01 -8.53915513e-02
1.93987325e-01 2.68757880e-01 6.86303437e-01 4.54043001e-01
-6.74253821e-01 4.56790954e-01 -1.36108983e+00 -6.87858462e-01
3.10991287e-01 2.55418748e-01 -1.05547023e+00 -1.34089768e-01
-7.68714607e-01 9.60126340e-01 2.68286139e-01 1.11320794e-01
-1.12197554e+00 -5.14413774e-01 -1.37833035e+00 -5.99605441e-01
4.43114460e-01 -1.11974299e+00 1.12485933e+00 -8.04108799e-01
-1.29931688e+00 1.27713001e+00 3.15916866e-01 -5.98571122e-01
9.60231483e-01 -8.08691025e-01 1.10624284e-01 5.59463620e-01
9.53254923e-02 9.18125749e-01 9.33958948e-01 -9.80418265e-01
-3.18462282e-01 -6.84173048e-01 -2.07206175e-01 6.59194589e-01
-7.90519826e-03 -2.61333406e-01 -6.22173965e-01 -6.29907250e-01
1.23867437e-01 -1.32028544e+00 -3.51232201e-01 4.87001300e-01
-2.20230684e-01 1.65240571e-01 6.18569911e-01 -7.65800774e-01
6.53492570e-01 -1.93689692e+00 7.53779769e-01 -7.19204322e-02
1.30146975e-02 1.16433434e-01 2.16281805e-02 1.61243360e-02
-2.82163545e-02 -3.89422536e-01 -9.74666849e-02 -5.69485486e-01
1.49473688e-02 5.62579930e-01 2.02248171e-01 1.08668172e+00
-1.00618210e-02 1.13064826e+00 -9.96020257e-01 -8.14593971e-01
5.32155156e-01 1.69580325e-01 -7.03197539e-01 4.25023347e-01
-8.52304623e-02 8.60793471e-01 -2.97581285e-01 7.80239880e-01
5.51831275e-02 -4.37096953e-01 2.41604730e-01 -2.21039608e-01
3.65568608e-01 -2.46386260e-01 -1.47031462e+00 2.45446229e+00
-2.76977178e-02 2.97602117e-01 2.02759236e-01 -1.07441461e+00
4.60691601e-01 3.34636420e-01 9.98478472e-01 -3.28491271e-01
2.02940196e-01 8.31853375e-02 -3.95918787e-01 -6.29000485e-01
1.56129092e-01 -8.71617198e-02 -1.73729062e-01 3.21018845e-01
6.50173783e-01 -3.35345184e-03 1.48971498e-01 1.17984042e-02
7.87107646e-01 9.12665367e-01 4.50594038e-01 -1.89261198e-01
5.12108147e-01 1.79429010e-01 5.50554931e-01 4.19732064e-01
-5.48841357e-01 6.98090255e-01 -4.92139943e-02 -5.54017901e-01
-1.24453115e+00 -1.10188270e+00 4.44920361e-02 8.81310463e-01
9.00296867e-02 -3.98615241e-01 -5.45673192e-01 -5.60273468e-01
1.62560344e-01 -1.53256372e-01 -4.95939165e-01 -1.42286241e-01
-8.81897509e-01 -2.83444554e-01 5.25090337e-01 8.91926706e-01
3.77001077e-01 -8.56066167e-01 -9.35955346e-01 -2.98847165e-02
-3.12748730e-01 -1.57776880e+00 -6.70311153e-01 -2.49770880e-01
-1.05099738e+00 -1.49044001e+00 -1.04611754e+00 -8.81993830e-01
6.08141482e-01 9.42611471e-02 9.77950454e-01 -2.67177850e-01
-6.21507585e-01 9.52504516e-01 -2.63534456e-01 1.14303187e-01
-2.53903300e-01 1.14227766e-02 6.04131162e-01 -2.48697042e-01
-1.02834798e-01 -5.65218270e-01 -5.40240765e-01 5.75509548e-01
-4.69013393e-01 9.88928527e-02 3.98337156e-01 8.66688311e-01
6.32244766e-01 -4.33278441e-01 -9.03446451e-02 -5.05797863e-01
-1.96155712e-01 -2.00869128e-01 -3.88293386e-01 2.31180396e-02
-3.54758888e-01 -3.54963988e-02 1.14254683e-01 -5.17356396e-01
-9.37370241e-01 5.27541399e-01 1.57094792e-01 -8.58017921e-01
-3.25081199e-01 1.29411280e-01 -3.80109213e-02 -1.22127958e-01
8.88684988e-01 8.41712877e-02 5.15027463e-01 -4.84100640e-01
6.70870841e-01 1.11467414e-01 9.28694546e-01 -6.22522712e-01
1.09392309e+00 7.78854430e-01 1.36670068e-01 -7.88599133e-01
-7.81858981e-01 -8.16793263e-01 -1.32742167e+00 -3.61531526e-01
1.04186475e+00 -1.36691403e+00 -3.08975458e-01 5.83475351e-01
-7.49864459e-01 -4.25917685e-01 -3.27246547e-01 9.76752341e-01
-1.29397213e+00 7.45491087e-01 -7.75914371e-01 -6.77282214e-01
-2.13314205e-01 -6.97004020e-01 1.30757785e+00 -1.65421039e-01
-4.66565698e-01 -1.10742903e+00 4.23412383e-01 7.08619595e-01
-2.03568742e-01 5.65890431e-01 2.50624090e-01 -2.37035573e-01
-4.14600909e-01 -3.17608505e-01 2.86800206e-01 4.73608404e-01
1.55623972e-01 -4.17672276e-01 -4.31219637e-01 -7.29744673e-01
-2.41843149e-01 -1.01871848e+00 5.59189737e-01 5.82774639e-01
5.76543987e-01 -8.84741023e-02 -2.43632630e-01 8.99413705e-01
9.43702757e-01 -6.19784296e-01 2.65428960e-01 2.64908910e-01
1.02473819e+00 6.36098683e-01 8.91929865e-01 3.79459441e-01
4.43992138e-01 7.57101893e-01 2.65157878e-01 -1.38825953e-01
-2.37109855e-01 -5.72816849e-01 4.70761269e-01 7.38000095e-01
-4.23405707e-01 4.34186876e-01 -8.83117020e-01 4.67749953e-01
-2.07294774e+00 -1.04802632e+00 1.89631715e-01 2.17021608e+00
6.92396522e-01 1.88322589e-01 5.94487846e-01 1.50178848e-02
3.80313516e-01 1.37292802e-01 -6.11732244e-01 3.89441460e-01
-7.05556273e-02 -1.82203710e-01 5.64938903e-01 3.47696126e-01
-1.43636620e+00 8.89719427e-01 6.80338764e+00 2.50516146e-01
-7.39898026e-01 6.50223866e-02 2.68948138e-01 -3.63966078e-01
4.35321182e-01 -1.95069000e-01 -7.34446406e-01 5.68478331e-02
6.31927848e-01 7.28945360e-02 2.52253003e-02 9.32511985e-01
6.15651160e-02 1.05039567e-01 -1.22699690e+00 1.37834537e+00
4.43589151e-01 -1.00754726e+00 -2.30262622e-01 9.89155248e-02
7.77196646e-01 8.61546919e-02 -2.88770139e-01 1.51905656e-01
1.99616566e-01 -8.21670473e-01 7.34311223e-01 4.06828314e-01
6.87361240e-01 -3.74104679e-01 4.09732372e-01 7.18468368e-01
-1.05772495e+00 -6.75761402e-02 -2.32936561e-01 -5.43091856e-02
3.62202138e-01 7.39235571e-03 -3.14741641e-01 5.42449415e-01
8.09976578e-01 1.15948486e+00 -4.98353511e-01 9.17904973e-01
-2.48440295e-01 3.87053877e-01 -4.57310587e-01 4.99809742e-01
2.98086256e-02 -3.81979533e-02 5.53109109e-01 9.20509636e-01
-1.14724599e-01 -1.40586615e-01 7.68273056e-01 2.79325515e-01
1.80818245e-01 7.38639832e-02 -6.67522669e-01 5.47084212e-02
-2.68757015e-01 1.15786040e+00 -5.15717387e-01 -2.26328328e-01
-4.30548698e-01 1.10288417e+00 3.13365698e-01 2.70765901e-01
-7.89807618e-01 4.63946968e-01 5.90993285e-01 1.99198544e-01
2.55269617e-01 -8.80300999e-01 2.25923970e-01 -1.65479302e+00
9.32733193e-02 -8.57221484e-01 6.16047502e-01 -6.47785962e-01
-1.31312478e+00 2.37346068e-01 5.05597949e-01 -1.53013325e+00
-9.38069105e-01 -8.05367947e-01 -5.58284558e-02 2.56490409e-01
-1.27231741e+00 -1.43155575e+00 -3.18642020e-01 9.00375903e-01
5.18776000e-01 -2.05293968e-01 6.42368615e-01 2.57498026e-01
-1.77308202e-01 3.91464055e-01 -3.43395680e-01 4.70895618e-01
8.60951662e-01 -1.07373476e+00 4.56232220e-01 6.65912747e-01
4.49093819e-01 2.83768803e-01 6.45815313e-01 -6.50622547e-01
-1.64714861e+00 -8.44166398e-01 5.70267081e-01 -9.53592122e-01
5.60986638e-01 -5.70196807e-01 -4.28244978e-01 1.00692141e+00
-2.91749477e-01 5.12811482e-01 6.89222693e-01 2.74603907e-02
-1.76459685e-01 2.37031206e-01 -8.64088714e-01 3.69213074e-01
1.81726754e+00 -2.61009097e-01 -9.62625265e-01 5.21940351e-01
3.19330484e-01 -7.99190581e-01 -8.19099724e-01 5.07102668e-01
8.45781207e-01 -7.15755343e-01 1.35445690e+00 -1.07783973e+00
2.44170532e-01 -3.32888253e-02 -1.65370345e-01 -9.45390999e-01
-2.31727093e-01 -5.65618098e-01 -6.56260967e-01 5.49869239e-01
-7.45373964e-02 6.29695579e-02 1.28407431e+00 4.27218467e-01
2.12533280e-01 -4.41565067e-01 -1.04350233e+00 -1.01207638e+00
9.51166004e-02 -3.44902247e-01 -1.70614123e-01 8.16118538e-01
-8.46202970e-02 1.79168910e-01 -9.27676916e-01 -9.97312949e-04
1.17501962e+00 1.16190448e-01 1.40014172e+00 -1.09650016e+00
-5.81137359e-01 2.19887290e-02 -1.07581973e+00 -1.33589232e+00
4.09652919e-01 -6.66400492e-01 -7.92195871e-02 -1.11173046e+00
2.79240906e-01 1.32405106e-02 1.30690157e-01 5.05635023e-01
3.64712663e-02 4.35775131e-01 2.17040822e-01 4.80030984e-01
-8.75418484e-01 7.28379250e-01 1.38396382e+00 3.21547166e-02
-1.06534950e-01 2.05023866e-02 6.56904578e-02 1.00476813e+00
3.03019166e-01 -2.66613871e-01 -6.11328661e-01 -3.70569319e-01
-7.00370148e-02 2.55066544e-01 8.88425171e-01 -1.18829036e+00
1.87725216e-01 -1.86169669e-01 7.68058240e-01 -5.22373438e-01
4.81343687e-01 -7.52011240e-01 1.18263796e-01 6.35010004e-01
-2.75052875e-01 9.24523026e-02 -1.32844700e-02 7.56674826e-01
-7.82017112e-02 2.25868762e-01 6.18091822e-01 -5.81567883e-01
-1.07861817e+00 6.91643119e-01 6.78032786e-02 6.21260762e-01
9.20831501e-01 -5.02486944e-01 2.57796258e-01 -6.14394069e-01
-1.10621130e+00 5.05313694e-01 6.73389375e-01 7.01581538e-01
5.68883121e-01 -1.57228935e+00 -7.16068745e-01 2.96088696e-01
3.38030964e-01 1.24938423e-02 2.40552336e-01 8.31975341e-01
-5.67989588e-01 4.21209782e-01 -6.40716016e-01 -1.06032348e+00
-1.19595718e+00 3.72736216e-01 3.83886158e-01 -1.33213058e-01
-1.04280949e+00 4.47883219e-01 1.39334230e-02 -7.66003847e-01
4.13830459e-01 -1.80459395e-01 5.04426323e-02 -3.60973626e-01
1.88177839e-01 3.56780797e-01 -3.23176384e-01 -1.19880664e+00
-3.59797418e-01 8.40848267e-01 4.10012364e-01 -8.83721337e-02
1.14899290e+00 -2.96278447e-01 4.76484209e-01 4.54360873e-01
1.30024183e+00 -3.97183597e-01 -1.70721591e+00 -3.50259691e-01
-2.13528439e-01 -5.31908691e-01 -3.68760020e-01 -2.86959559e-01
-8.53241026e-01 6.91717029e-01 6.18092060e-01 -7.26977825e-01
6.28714383e-01 2.44914874e-01 6.81855738e-01 5.14580369e-01
8.07324052e-01 -1.31549978e+00 4.95736748e-01 3.19171160e-01
7.57475436e-01 -1.34527993e+00 3.92712355e-01 -4.66130257e-01
-7.34855592e-01 7.13176370e-01 8.44605327e-01 -3.64165068e-01
6.24960899e-01 -5.46465628e-02 2.61363059e-01 -7.82349780e-02
-5.28551757e-01 -1.29026815e-01 6.12976909e-01 7.08356142e-01
1.66647181e-01 -2.78064638e-01 -1.16146386e-01 3.24152827e-01
-2.16613382e-01 1.98245466e-01 4.75887172e-02 1.41266465e+00
-1.75763920e-01 -1.02735257e+00 -2.82900989e-01 2.47111991e-02
-9.30483714e-02 6.02510870e-01 -1.52227283e-01 1.12434602e+00
1.66655853e-02 4.77630436e-01 -3.43849435e-02 -1.13032110e-01
4.33620006e-01 1.05017968e-01 1.15455604e+00 -6.45330846e-01
-4.24527638e-02 2.55171657e-01 2.17474565e-01 -1.02055228e+00
-1.02077496e+00 -9.11787570e-01 -1.29991889e+00 1.88946556e-02
-5.00969477e-02 -2.65905708e-01 2.97933578e-01 1.14146686e+00
2.17386037e-01 -7.42046610e-02 1.58573762e-01 -1.27040422e+00
-8.48541677e-01 -8.38717997e-01 -6.92593634e-01 1.15217459e+00
2.94563919e-01 -1.16580403e+00 -1.32168949e-01 5.15967190e-01]
|
[7.0708136558532715, -0.8341258764266968]
|
29bfb591-0346-4784-9774-62c177c29461
|
dwa-differential-wavelet-amplifier-for-image-1
|
2307.04593
| null |
https://arxiv.org/abs/2307.04593v1
|
https://arxiv.org/pdf/2307.04593v1.pdf
|
DWA: Differential Wavelet Amplifier for Image Super-Resolution
|
This work introduces Differential Wavelet Amplifier (DWA), a drop-in module for wavelet-based image Super-Resolution (SR). DWA invigorates an approach recently receiving less attention, namely Discrete Wavelet Transformation (DWT). DWT enables an efficient image representation for SR and reduces the spatial area of its input by a factor of 4, the overall model size, and computation cost, framing it as an attractive approach for sustainable ML. Our proposed DWA model improves wavelet-based SR models by leveraging the difference between two convolutional filters to refine relevant feature extraction in the wavelet domain, emphasizing local contrasts and suppressing common noise in the input signals. We show its effectiveness by integrating it into existing SR models, e.g., DWSR and MWCNN, and demonstrate a clear improvement in classical SR tasks. Moreover, DWA enables a direct application of DWSR and MWCNN to input image space, reducing the DWT representation channel-wise since it omits traditional DWT.
|
['Andreas Dengel', 'Sebastian Palacio', 'Federico Raue', 'Stanislav Frolov', 'Brian B. Moser']
|
2023-07-10
| null | null | null | null |
['image-super-resolution', 'super-resolution']
|
['computer-vision', 'computer-vision']
|
[ 8.65807116e-01 1.42765855e-02 -9.38807428e-02 -3.87037918e-02
-9.21370447e-01 -2.80926198e-01 4.24999028e-01 -2.44153842e-01
-4.59314734e-01 4.27991003e-01 4.90987033e-01 -1.07381158e-01
-1.04279831e-01 -9.28480625e-01 -4.97518748e-01 -8.34669828e-01
9.04784352e-02 -6.09167695e-01 5.11947513e-01 -5.18868744e-01
1.56442508e-01 8.45305264e-01 -1.49351096e+00 7.28870749e-01
6.26614630e-01 1.05372965e+00 5.42475045e-01 6.53525889e-01
-1.49722800e-01 8.29656959e-01 -6.12870991e-01 -1.84671115e-02
3.47425938e-01 -5.04926860e-01 -5.81165731e-01 -1.56440839e-01
6.56576812e-01 -2.10106477e-01 -7.02227533e-01 9.55229223e-01
7.81313181e-01 1.02996551e-01 7.35751614e-02 -6.43917918e-01
-1.04486763e+00 7.22817957e-01 -9.20213878e-01 8.71580601e-01
2.26850346e-01 7.59637356e-02 8.00275505e-01 -1.19360161e+00
8.82067502e-01 1.43256104e+00 9.59883749e-01 4.18324262e-01
-1.72042298e+00 -5.56518495e-01 -1.63690284e-01 4.49425012e-01
-1.34273934e+00 -5.93387425e-01 1.09360027e+00 1.35740086e-01
9.01073992e-01 4.23192948e-01 6.39292061e-01 1.06042993e+00
2.87473619e-01 7.44732320e-01 1.41803086e+00 -6.26785517e-01
3.08485597e-01 -3.79482478e-01 -1.11050248e-01 2.45733902e-01
-1.10516280e-01 8.30633268e-02 -9.76935804e-01 1.06067568e-01
1.32397175e+00 -1.43677592e-01 -4.75268126e-01 -3.77433561e-02
-1.12539411e+00 6.63623571e-01 4.81189013e-01 5.61835945e-01
-3.22951019e-01 1.48447633e-01 2.26227731e-01 6.74645662e-01
7.40765572e-01 3.05569053e-01 -2.61166632e-01 2.06006452e-01
-1.27897787e+00 2.06855848e-01 -1.25757620e-01 5.12718558e-01
3.15678000e-01 3.72231901e-01 -4.87473577e-01 8.04355562e-01
-3.64056647e-01 2.75473833e-01 4.80520904e-01 -1.19044554e+00
8.25169012e-02 2.63399214e-01 -1.40351906e-01 -1.02230835e+00
-3.99237037e-01 -7.11335421e-01 -9.99500930e-01 6.41643286e-01
1.71595261e-01 4.76470828e-01 -8.09076846e-01 1.50718081e+00
7.56447539e-02 5.55928290e-01 1.35307133e-01 8.43598723e-01
1.00217509e+00 5.06727755e-01 -4.09916453e-02 -3.40374768e-01
1.46695387e+00 -7.18131781e-01 -7.66561091e-01 1.11475075e-02
3.36482213e-03 -8.49059999e-01 1.14471853e+00 5.64911962e-01
-1.39354599e+00 -1.09557712e+00 -1.07853043e+00 -5.61842084e-01
-1.28957465e-01 2.14996636e-01 4.14822549e-01 3.13607663e-01
-1.49965084e+00 7.79104590e-01 -8.57839465e-01 -2.97696870e-02
8.53568912e-01 2.34624997e-01 -3.57741714e-01 -3.16489458e-01
-1.02574074e+00 1.04251444e+00 -1.68082304e-02 2.56412979e-02
-5.97718894e-01 -1.17909312e+00 -9.81058121e-01 1.86833501e-01
-3.43160843e-03 -4.85183716e-01 8.99385452e-01 -5.51435232e-01
-1.66936958e+00 7.24121094e-01 -3.74032408e-01 -7.47151494e-01
2.97941834e-01 -6.05349801e-02 -7.09153831e-01 5.52108645e-01
-2.19336867e-01 4.95659143e-01 1.67305982e+00 -9.79059517e-01
-5.88056743e-01 -3.07075530e-01 -1.78973064e-01 -3.18518765e-02
-3.12364161e-01 3.05310190e-01 5.88275567e-02 -1.22093022e+00
3.72777253e-01 -5.27895130e-02 -3.09691459e-01 3.62707347e-01
1.59596786e-01 1.26426108e-02 1.22859812e+00 -1.00596666e+00
1.28992534e+00 -2.56158233e+00 3.68135333e-01 -1.23460993e-01
7.34041035e-01 3.86494339e-01 -6.22025788e-01 1.92072242e-01
-2.76884735e-01 -1.71747893e-01 -4.82727677e-01 -2.94573665e-01
-4.46329743e-01 3.48200276e-02 -4.90522921e-01 4.92424756e-01
6.68506503e-01 9.96953845e-01 -7.05744386e-01 -3.59877646e-01
2.90169567e-01 1.06016397e+00 -2.29412094e-01 -1.81580424e-01
3.39037627e-01 5.57538688e-01 -1.56168044e-01 4.05680239e-01
1.11219716e+00 1.04932785e-01 -1.46955192e-01 -5.63310921e-01
-4.40275460e-01 4.99289073e-02 -1.15571487e+00 1.82615924e+00
-8.08882475e-01 8.65117729e-01 1.84187219e-01 -1.12461722e+00
1.22886908e+00 7.17510805e-02 4.08710599e-01 -1.12934148e+00
-1.93749875e-01 2.25472122e-01 -2.72422045e-01 -2.94583559e-01
3.83467525e-01 -1.20341428e-01 4.25707191e-01 3.98081467e-02
2.01644152e-01 -2.12437510e-01 -1.03189453e-01 -2.21014302e-02
1.31334853e+00 2.55856931e-01 3.37050050e-01 -5.25241315e-01
7.09054589e-01 -2.35457927e-01 3.86888772e-01 6.50127828e-01
1.60866112e-01 9.60321724e-01 -2.38179229e-02 -6.60039485e-01
-1.10560918e+00 -1.18031287e+00 -3.23466092e-01 8.96737158e-01
1.72105044e-01 -2.12635040e-01 -4.81709957e-01 -3.11051786e-01
-1.18586764e-01 5.86759627e-01 -5.91048419e-01 -2.57565439e-01
-1.02028298e+00 -5.47873557e-01 3.79244983e-01 4.52380478e-01
7.84672558e-01 -9.54900324e-01 -1.09285080e+00 3.13439459e-01
-6.89016283e-02 -1.35991287e+00 -4.49286968e-01 1.37978792e-01
-1.04464436e+00 -5.76876044e-01 -1.02818310e+00 -8.02527368e-01
4.53045785e-01 6.96765065e-01 9.51535344e-01 -1.12269251e-02
-7.81232059e-01 7.32955784e-02 -4.78168070e-01 -7.07313195e-02
-2.57075757e-01 -2.93625355e-01 -1.20654725e-01 4.62673753e-02
4.27314732e-03 -1.00322986e+00 -8.40401649e-01 -1.75555935e-04
-1.00923705e+00 1.89696893e-01 6.77230179e-01 8.94444764e-01
6.16023004e-01 4.55160320e-01 6.38542295e-01 -5.37002683e-01
5.58813870e-01 6.84808046e-02 -4.30962116e-01 -1.07970543e-01
-5.37969530e-01 -9.93693843e-02 6.78989291e-01 -6.62598848e-01
-1.22636557e+00 -4.94307093e-02 -3.27426016e-01 -4.22742158e-01
2.06888229e-01 2.13290468e-01 2.13781521e-01 -6.27498806e-01
7.34275937e-01 7.12790251e-01 2.90522814e-01 -7.61848807e-01
4.17477071e-01 4.28521544e-01 9.69179273e-01 -9.99504775e-02
1.18877697e+00 8.93947721e-01 1.31784379e-01 -1.02877867e+00
-6.17818117e-01 -3.23205203e-01 -7.97838748e-01 8.98949131e-02
5.58903635e-01 -1.10142374e+00 -4.20781672e-01 3.40141505e-01
-1.06825173e+00 -1.74669564e-01 -9.88738775e-01 2.93651283e-01
-4.84430403e-01 5.26955009e-01 -7.62998581e-01 -6.00628853e-01
-4.46803421e-01 -8.75505030e-01 1.13141263e+00 1.60774395e-01
-1.48392424e-01 -6.60392821e-01 -4.21562016e-01 1.11945719e-02
1.13081622e+00 3.23392838e-01 8.49748790e-01 2.21626367e-02
-1.88216180e-01 1.97095364e-01 -4.61218894e-01 3.93061668e-01
1.45742819e-01 -4.74087477e-01 -1.01288426e+00 -4.64991778e-01
3.18948030e-01 2.94567961e-02 1.24188828e+00 6.36617959e-01
1.12457204e+00 -6.43536076e-02 1.31696150e-01 7.61461437e-01
1.49802291e+00 -2.29351491e-01 9.65498269e-01 4.25223649e-01
2.16381788e-01 2.70447195e-01 3.46344948e-01 3.57298851e-01
-5.28369844e-02 1.06138146e+00 2.18461081e-01 -8.63072157e-01
-1.10389638e+00 2.44666450e-02 5.54806828e-01 5.88360965e-01
-3.55698705e-01 5.39917409e-01 -2.61709929e-01 3.89195353e-01
-1.35756552e+00 -9.21228349e-01 -1.06945701e-01 1.95499587e+00
1.04519725e+00 2.92045623e-02 -3.16190496e-02 5.07861137e-01
4.11744714e-01 4.45954323e-01 -4.75061744e-01 -2.66787291e-01
-6.80940211e-01 1.22997093e+00 3.72907370e-01 5.56145608e-01
-8.62132967e-01 8.62599075e-01 6.70205641e+00 1.19157147e+00
-1.14093578e+00 2.39880830e-01 4.18119550e-01 4.15702872e-02
-2.46727303e-01 -2.66039550e-01 -5.08102894e-01 1.45498782e-01
5.50119579e-01 -2.35812515e-01 6.01635695e-01 5.36804259e-01
5.19748271e-01 2.80763879e-02 -6.37822568e-01 1.00756693e+00
-3.28490771e-02 -1.69950938e+00 2.37191543e-01 -2.39103779e-01
3.24478775e-01 -3.12294453e-01 2.68262357e-01 1.51238427e-01
-2.40978077e-02 -1.14921308e+00 7.18217671e-01 3.25169891e-01
1.17724633e+00 -8.66917074e-01 6.99789405e-01 8.02102219e-03
-1.41892517e+00 -2.80731171e-01 -6.61070287e-01 4.55625309e-03
2.74835210e-02 7.83239543e-01 -9.15668979e-02 7.80418813e-01
1.01332259e+00 7.86016285e-01 -6.25392258e-01 6.52294099e-01
-1.22553438e-01 5.08223891e-01 -1.59442306e-01 9.52277243e-01
-1.07924379e-01 -1.97059214e-02 7.70272434e-01 1.36435533e+00
6.03115857e-01 3.36666226e-01 -1.32656664e-01 9.56079662e-01
7.62562081e-02 -1.62602484e-01 -2.92972386e-01 4.24402297e-01
4.15584594e-01 1.30931473e+00 -7.38470137e-01 -1.28507867e-01
-4.13077801e-01 1.25346804e+00 -1.68074816e-01 3.87514025e-01
-4.92567569e-01 -4.85434383e-01 5.44451714e-01 2.71455884e-01
6.11664951e-01 -2.82288373e-01 -4.24410909e-01 -9.13855553e-01
4.60835248e-02 -9.69897509e-01 2.40247369e-01 -7.00908184e-01
-9.93610859e-01 8.78579140e-01 -1.25277266e-01 -1.40596640e+00
2.96637714e-01 -2.90992111e-01 -4.90925759e-01 1.12137020e+00
-2.27685213e+00 -1.32053506e+00 -3.63030255e-01 7.08687961e-01
9.42402363e-01 -5.70731573e-02 6.60971880e-01 2.23897547e-01
-2.26330280e-01 5.38769305e-01 -1.21000253e-01 -1.67890996e-01
4.43864465e-01 -8.95704150e-01 7.50105023e-01 1.16755021e+00
6.75125271e-02 4.40599978e-01 8.99091303e-01 -3.59156340e-01
-1.61510611e+00 -9.43473756e-01 7.49071777e-01 -1.56349484e-02
5.74574053e-01 -3.63339722e-01 -1.08313835e+00 1.90275133e-01
1.00518644e-01 5.58203220e-01 1.64280713e-01 -4.96497780e-01
-5.78171015e-01 -1.72804669e-01 -1.47260690e+00 4.85178888e-01
1.25387597e+00 -4.80823994e-01 -5.13465405e-01 -6.98985755e-02
1.06399429e+00 -3.98128182e-01 -1.05244875e+00 3.94480795e-01
4.80138212e-01 -1.02887821e+00 1.69393444e+00 -2.14156255e-01
6.55810595e-01 -4.88295704e-01 -1.73901573e-01 -1.20709836e+00
-8.02255273e-01 -8.04435432e-01 -3.78555447e-01 8.24162841e-01
5.17612100e-02 -5.02644539e-01 4.76299018e-01 -1.63036242e-01
-4.63940613e-02 -5.40066123e-01 -1.41337216e+00 -9.49255943e-01
-1.72985286e-01 -2.11426571e-01 4.49687004e-01 9.44844782e-01
-2.88751900e-01 -1.00817896e-01 -2.24897161e-01 1.92596436e-01
9.00712848e-01 6.69018403e-02 2.20561981e-01 -1.17769158e+00
-3.19662452e-01 -4.98542488e-01 -4.43335146e-01 -9.96382773e-01
-2.34162822e-01 -8.57809842e-01 -2.80992091e-01 -1.44307017e+00
-8.96589383e-02 2.00013906e-01 -3.97660255e-01 4.56577241e-01
-3.48041840e-02 8.88518095e-01 3.45490932e-01 2.77947038e-01
-1.55742839e-02 4.26147014e-01 1.55692863e+00 -1.32496506e-01
-2.29804739e-01 -2.20002443e-01 -7.10806072e-01 6.89339817e-01
6.33635640e-01 -3.37463439e-01 -1.88268587e-01 -3.77874732e-01
-2.12296054e-01 6.35281112e-03 5.19533098e-01 -1.05712593e+00
1.72350794e-01 2.56269742e-02 6.98098719e-01 -3.75330001e-01
3.93909931e-01 -7.25453258e-01 1.49247468e-01 5.80165684e-01
-4.17554528e-01 -1.58008501e-01 3.44708353e-01 4.03914779e-01
-4.23483163e-01 1.93997905e-01 1.29143727e+00 -4.56268750e-02
-8.35633516e-01 -7.32091069e-02 -4.22657579e-01 -4.72603738e-01
6.48682475e-01 -6.69770479e-01 -2.57152200e-01 -1.15132451e-01
-8.49122584e-01 -3.15312803e-01 1.17642291e-01 1.99797496e-01
1.19575942e+00 -1.17566824e+00 -1.20039284e+00 5.25551081e-01
-2.41521463e-01 -1.46524176e-01 5.28763413e-01 9.86326337e-01
-2.55959421e-01 -5.71886189e-02 -5.53557873e-01 -3.55520517e-01
-1.50931406e+00 5.26934206e-01 -4.99248952e-02 -2.93697000e-01
-1.75015366e+00 8.95092726e-01 1.07233617e-02 2.11892843e-01
7.55554140e-02 -3.52971792e-01 -5.32706976e-01 -6.66281208e-02
1.18795657e+00 4.89989042e-01 2.99464986e-02 -4.67358202e-01
-8.85926932e-02 8.85200262e-01 -8.26902092e-02 1.28773257e-01
1.90476441e+00 -3.07025552e-01 -2.80539215e-01 9.32434946e-03
8.98966968e-01 1.93013310e-01 -1.28370559e+00 -7.13127732e-01
1.13714457e-01 -6.42940044e-01 6.99118018e-01 -7.21358240e-01
-1.19581103e+00 7.72949278e-01 7.32769430e-01 8.09414014e-02
1.87994885e+00 -1.03642285e-01 1.04485536e+00 -2.14731082e-01
4.21882480e-01 -8.29842865e-01 9.11967158e-02 8.48139599e-02
1.34848928e+00 -7.54369915e-01 3.28460276e-01 -7.49514520e-01
-2.54034787e-01 1.60182297e+00 6.22914592e-03 -3.70180607e-01
4.50751305e-01 7.66230464e-01 -9.12741795e-02 -5.56300208e-03
-4.79517817e-01 -2.15885073e-01 2.09424525e-01 8.96542907e-01
1.14763096e-01 -2.56235391e-01 -3.62521052e-01 5.02404869e-01
-5.61340638e-02 3.15549880e-01 4.94876146e-01 7.25287557e-01
-4.16763961e-01 -1.00071573e+00 -3.76132756e-01 1.33726597e-01
-4.36543286e-01 -3.96455824e-01 7.89284036e-02 5.58311701e-01
3.12030297e-02 7.68430531e-01 -6.63440302e-02 -3.02989185e-01
4.36381072e-01 -2.39556327e-01 5.99948347e-01 -3.04741263e-01
-5.06464899e-01 1.89191431e-01 -2.48452753e-01 -7.43635774e-01
-5.08272350e-01 -4.31752324e-01 -9.63170767e-01 -1.19663700e-01
4.14326675e-02 -4.46201593e-01 4.97144520e-01 6.22999191e-01
4.93992001e-01 8.43125999e-01 6.73756003e-01 -1.07461011e+00
-5.78624964e-01 -8.41405213e-01 -6.78204536e-01 1.55218422e-01
7.17611670e-01 -4.31144565e-01 -3.39179486e-01 3.19654822e-01]
|
[11.106490135192871, -1.93467378616333]
|
ea09e366-8ae1-4764-9704-94ffce7dd53f
|
landmine-detection-using-autoencoders-on
|
1810.01316
| null |
http://arxiv.org/abs/1810.01316v1
|
http://arxiv.org/pdf/1810.01316v1.pdf
|
Landmine Detection Using Autoencoders on Multi-polarization GPR Volumetric Data
|
Buried landmines and unexploded remnants of war are a constant threat for the
population of many countries that have been hit by wars in the past years. The
huge amount of human lives lost due to this phenomenon has been a strong
motivation for the research community toward the development of safe and robust
techniques designed for landmine clearance. Nonetheless, being able to detect
and localize buried landmines with high precision in an automatic fashion is
still considered a challenging task due to the many different boundary
conditions that characterize this problem (e.g., several kinds of objects to
detect, different soils and meteorological conditions, etc.). In this paper, we
propose a novel technique for buried object detection tailored to unexploded
landmine discovery. The proposed solution exploits a specific kind of
convolutional neural network (CNN) known as autoencoder to analyze volumetric
data acquired with ground penetrating radar (GPR) using different
polarizations. This method works in an anomaly detection framework, indeed we
only train the autoencoder on GPR data acquired on landmine-free areas. The
system then recognizes landmines as objects that are dissimilar to the soil
used during the training step. Experiments conducted on real data show that the
proposed technique requires little training and no ad-hoc data pre-processing
to achieve accuracy higher than 93% on challenging datasets.
|
['Federico Lombardi', 'Paolo Bestagini', 'Francesco Picetti', 'Maurizio Lualdi', 'Stefano Tubaro']
|
2018-10-02
| null | null | null | null |
['landmine']
|
['computer-vision']
|
[ 3.48507673e-01 2.17001345e-02 5.89979589e-01 -3.73676568e-01
-3.06362003e-01 -3.53175290e-02 4.16158408e-01 3.22702438e-01
-7.55171299e-01 7.58900821e-01 -2.47339487e-01 -4.22344178e-01
-2.40529448e-01 -1.34490335e+00 -5.94845891e-01 -9.23684716e-01
-3.21910352e-01 5.30704618e-01 2.53694355e-01 -8.11855912e-01
2.37788990e-01 1.09147966e+00 -1.91934705e+00 1.30176947e-01
9.26590979e-01 1.09990537e+00 3.87021571e-01 2.01526076e-01
5.10841645e-02 2.97369003e-01 -4.94482517e-01 1.08121760e-01
3.04076970e-01 2.82406118e-02 -5.66538811e-01 4.03058343e-02
2.26887077e-01 -4.69663978e-01 -1.92925096e-01 1.21398580e+00
4.08738106e-01 2.55987138e-01 8.00514638e-01 -6.33749127e-01
2.49519572e-03 2.11816654e-01 -6.45152271e-01 4.75687265e-01
-1.38487563e-01 -2.28106827e-01 3.24009508e-01 -7.83503234e-01
1.03789359e-01 8.10714245e-01 8.10204864e-01 1.33762464e-01
-8.02686334e-01 -4.25601929e-01 -2.42621854e-01 5.31167209e-01
-1.33082914e+00 -2.32520923e-01 6.82589889e-01 -5.18873572e-01
7.05154181e-01 1.34688899e-01 7.10477412e-01 8.73557627e-01
4.42594707e-01 3.76119107e-01 1.06988525e+00 -4.30912614e-01
3.70339751e-01 8.24095681e-02 1.00644805e-01 4.32327658e-01
9.06662881e-01 1.31712541e-01 1.16566442e-01 -1.20076105e-01
2.81015426e-01 7.15014860e-02 -3.15976143e-01 -2.25333706e-01
-5.02610862e-01 1.01413679e+00 4.68764275e-01 8.15717995e-01
-1.01296175e+00 -2.06773371e-01 3.62089247e-01 1.99533597e-01
3.81449491e-01 2.19628841e-01 -2.38323152e-01 2.15049148e-01
-1.13838673e+00 4.25754309e-01 5.59021890e-01 1.54483989e-01
9.26429987e-01 5.33084512e-01 4.78632182e-01 6.78405881e-01
2.22403169e-01 8.53484511e-01 5.09210467e-01 4.89256869e-04
2.71858841e-01 6.28608882e-01 8.09322894e-02 -1.60376227e+00
-6.08121574e-01 -6.21598125e-01 -1.00571012e+00 4.89299804e-01
1.16003990e-01 -1.87484354e-01 -1.05887580e+00 1.18687284e+00
3.00803393e-01 -2.70247966e-01 4.30707961e-01 8.12192440e-01
4.78206724e-01 8.04157674e-01 1.18030146e-01 2.17999443e-01
1.43463182e+00 -1.44059315e-01 -5.45503020e-01 -5.11164367e-01
5.78765213e-01 -2.57233053e-01 5.45081377e-01 4.19474870e-01
-3.90759408e-01 -2.58888721e-01 -1.24383295e+00 7.36159563e-01
-7.06409216e-01 2.31166929e-01 6.50657356e-01 7.90903032e-01
-6.61984563e-01 7.27378905e-01 -9.06228244e-01 -6.40595853e-01
4.56846416e-01 2.63020217e-01 -4.71799850e-01 -1.20060787e-01
-1.33665216e+00 1.04414594e+00 6.41946554e-01 8.61552954e-01
-7.38781214e-01 -1.03556104e-01 -8.09085131e-01 -6.42716791e-03
8.09719861e-02 -5.75073762e-03 5.06019771e-01 -8.71461332e-01
-8.48026812e-01 8.13673437e-01 5.36447644e-01 -8.10830474e-01
4.32383627e-01 -4.41297531e-01 -7.45022535e-01 1.85739368e-01
9.43929479e-02 -8.61024112e-02 7.25879848e-01 -1.17937016e+00
-7.95033753e-01 -6.77657783e-01 -3.39037389e-01 -5.17756620e-04
-4.72419649e-01 -1.06129438e-01 4.23257291e-01 -4.13072616e-01
5.78125238e-01 -7.03793049e-01 -3.43944103e-01 -4.60917801e-01
-2.42353424e-01 1.17936902e-01 1.17314637e+00 -9.37966883e-01
7.30839074e-01 -1.99369872e+00 -1.92984208e-01 5.38943291e-01
-1.83342174e-01 5.16553700e-01 2.41132751e-01 3.55562210e-01
-2.85678916e-02 -3.97703618e-01 -1.01099324e+00 3.34039271e-01
-3.06577533e-01 2.75290251e-01 -1.55677706e-01 7.79779553e-01
2.30434567e-01 1.20347127e-01 -5.92264950e-01 -1.84797585e-01
1.99070156e-01 5.24399996e-01 -2.45152950e-01 1.33219868e-01
6.21002950e-02 3.34445179e-01 -7.19062448e-01 8.06559145e-01
1.25549150e+00 6.23826921e-01 -8.28818679e-02 -3.48096043e-01
-4.19497699e-01 -4.11663473e-01 -1.14024758e+00 9.38399136e-01
-4.19340700e-01 6.55862808e-01 2.69953340e-01 -1.67928064e+00
1.36080837e+00 2.27738157e-01 4.95105147e-01 -9.34569657e-01
4.21595693e-01 6.88048601e-01 -7.13626295e-02 -8.96135926e-01
6.67054057e-01 -5.09605110e-01 -2.12403294e-02 1.04509555e-01
-2.04922348e-01 7.07642809e-02 -1.13728167e-02 -4.26924229e-01
1.03020823e+00 -1.56780422e-01 1.20830797e-01 -4.96414125e-01
7.00603485e-01 2.79776752e-01 4.87777263e-01 6.04231298e-01
-2.64521018e-02 1.92571685e-01 -7.57015944e-02 -8.39305937e-01
-8.01869690e-01 -8.40335667e-01 -4.70292002e-01 6.34792149e-01
8.30104873e-02 7.06890047e-01 -5.74917555e-01 -2.29929566e-01
2.11516488e-03 6.96275890e-01 -6.19962335e-01 -2.87589639e-01
-6.03419840e-01 -1.20007885e+00 8.81775916e-01 3.07849437e-01
1.02728248e+00 -1.43704128e+00 -1.28948748e+00 4.72434193e-01
-1.14120133e-01 -1.00941122e+00 9.31917548e-01 4.37823474e-01
-1.04763520e+00 -1.13404620e+00 -7.55939126e-01 -6.59881115e-01
5.12829185e-01 7.75089413e-02 7.92603135e-01 -4.04217429e-02
-6.41874075e-01 2.12493539e-01 -5.29682875e-01 -6.30024552e-01
-3.60100657e-01 -1.18834153e-02 1.41860068e-01 3.81149411e-01
5.21018267e-01 -7.77632058e-01 -4.50878978e-01 1.07163548e-01
-1.09854150e+00 -5.32992184e-01 1.04755390e+00 6.70145988e-01
3.08063924e-01 6.73988283e-01 5.42155325e-01 -5.63234627e-01
5.33653259e-01 -8.33209276e-01 -6.38963878e-01 -7.77514875e-02
-1.10849909e-01 -1.46523342e-01 2.98997760e-01 3.82875241e-02
-1.06985438e+00 -1.59181327e-01 -5.35007894e-01 7.16230646e-02
-7.50546098e-01 7.69319177e-01 -1.74532637e-01 -1.71111956e-01
8.49876165e-01 4.21657532e-01 -1.07285765e-03 -5.37848651e-01
-3.01474839e-01 8.55059028e-01 7.21747279e-01 -3.89677823e-01
9.73134875e-01 7.69661963e-01 1.88784987e-01 -1.54070807e+00
-4.68462944e-01 -3.90208781e-01 -4.49210703e-01 -2.25933895e-01
8.06084454e-01 -6.48562789e-01 -2.77553529e-01 7.56357431e-01
-1.01049054e+00 3.62961628e-02 2.35191896e-03 6.42372310e-01
-2.30170771e-01 3.84605914e-01 -2.12239604e-02 -1.09645534e+00
-7.09994555e-01 -6.31419361e-01 4.91367549e-01 2.10904852e-01
2.24801704e-01 -7.35761583e-01 1.68574512e-01 4.96896617e-02
7.27755427e-01 7.45867729e-01 7.49359190e-01 -6.49736881e-01
-1.37680486e-01 -6.03279352e-01 -1.53003022e-01 5.75529337e-01
2.96037793e-01 -3.68631661e-01 -8.66062045e-01 -3.19561392e-01
2.91026115e-01 -1.46652222e-01 1.14044881e+00 3.15552890e-01
7.94070780e-01 -1.33331105e-01 -3.76528054e-01 4.17544931e-01
1.63250911e+00 3.56487483e-01 9.79877651e-01 8.83048415e-01
4.23214525e-01 8.17873657e-01 8.78947914e-01 5.62157333e-01
-1.98645979e-01 3.35786104e-01 9.09138203e-01 -2.47880325e-01
6.42631590e-01 4.07980263e-01 2.10891604e-01 1.55575231e-01
-4.41301584e-01 -1.40166124e-02 -1.06046104e+00 8.52856636e-01
-1.50906098e+00 -1.23755956e+00 -1.13530152e-01 2.11151481e+00
1.08648628e-01 6.15153797e-02 -1.80093750e-01 6.58252656e-01
8.02157044e-01 5.35232089e-02 -1.78404912e-01 -4.52056020e-01
-2.77949184e-01 4.85281348e-01 7.30767012e-01 1.93894625e-01
-1.23530519e+00 7.27509916e-01 4.67621994e+00 4.39947873e-01
-1.52800643e+00 -2.45888442e-01 -8.66762623e-02 6.31633580e-01
-1.33181766e-01 -1.92669660e-01 -4.72534180e-01 2.23626986e-01
6.93844557e-01 2.80634522e-01 -1.63701773e-01 9.41829383e-01
2.23353386e-01 -3.44027668e-01 -1.54052466e-01 6.33021176e-01
2.57531907e-02 -8.56238544e-01 7.41017386e-02 1.08023835e-02
1.19397745e-01 1.81607589e-01 -6.63793087e-02 1.32561713e-01
-2.19307154e-01 -8.58995080e-01 5.26451170e-01 5.27566552e-01
2.26115957e-01 -9.98310745e-01 1.09843087e+00 2.75468588e-01
-8.74154925e-01 -3.17484617e-01 -6.22935951e-01 -1.15311086e-01
1.67797789e-01 9.89099920e-01 -8.51499498e-01 6.53504670e-01
9.91769671e-01 2.41045535e-01 -2.78497428e-01 1.09157622e+00
5.03443107e-02 6.36807084e-01 -5.60619056e-01 1.00791901e-01
5.89702964e-01 -3.27541828e-01 7.68796206e-01 1.10810018e+00
8.09566975e-01 1.45461529e-01 -1.40437111e-01 6.66662693e-01
5.07006228e-01 2.33529717e-01 -9.97708082e-01 -2.93333270e-02
1.06818095e-01 1.17469060e+00 -6.57636285e-01 1.40312454e-02
-9.16010514e-02 6.53764427e-01 -2.85675824e-01 6.21728189e-02
-6.72061086e-01 -7.16710448e-01 4.30500180e-01 5.98251104e-01
3.04367423e-01 -3.00109357e-01 2.84293711e-01 -6.93921566e-01
1.37858186e-02 -6.25572562e-01 3.14669281e-01 -4.26758260e-01
-9.80717957e-01 9.01438177e-01 -2.01256964e-02 -1.15665245e+00
-5.97217120e-03 -7.89340734e-01 -6.52007878e-01 6.61437571e-01
-1.66625512e+00 -1.00517952e+00 -8.20795178e-01 6.23274624e-01
1.20807558e-01 -3.29126298e-01 9.83867764e-01 4.68413800e-01
-3.08189005e-01 1.27651636e-02 2.36114055e-01 1.90398559e-01
2.01994732e-01 -7.79322088e-01 2.60495394e-02 8.76070321e-01
-1.62709579e-01 9.21931416e-02 9.13244128e-01 -7.36057341e-01
-1.08024073e+00 -1.12145650e+00 5.41546583e-01 4.86106813e-01
3.93068492e-01 1.22708187e-01 -1.18181705e+00 5.47190964e-01
-2.14273140e-01 -3.42247970e-02 4.11023676e-01 -6.87329099e-02
1.60041630e-01 -2.37057701e-01 -1.47963977e+00 2.38661423e-01
3.72878939e-01 -5.30258520e-03 -7.56581485e-01 1.05494969e-01
-1.10819861e-01 -5.07594794e-02 -4.82958674e-01 7.25569189e-01
5.10129511e-01 -1.22615337e+00 6.32506311e-01 -3.27039182e-01
2.48651236e-01 -2.47244641e-01 -5.00745714e-01 -1.14910591e+00
-5.74606806e-02 2.11432159e-01 2.44639009e-01 7.37334847e-01
1.50147289e-01 -7.56424427e-01 9.99445498e-01 -2.70208698e-02
-3.78691077e-01 -5.63429415e-01 -9.14812684e-01 -7.69591391e-01
-3.73740457e-02 -3.10741007e-01 4.27147150e-01 8.61546636e-01
-5.75870037e-01 -3.22034627e-01 -2.22932056e-01 7.95787334e-01
9.21045959e-01 8.57148692e-02 4.72864836e-01 -1.71487916e+00
9.98702273e-02 9.56342649e-03 -9.54670787e-01 -2.14544863e-01
1.43977143e-02 -5.84824622e-01 2.12680042e-01 -1.58504772e+00
-3.80483717e-01 -7.22610354e-01 -3.57082635e-01 5.54820240e-01
2.76442349e-01 3.02953541e-01 -4.35967594e-01 -1.56143447e-02
2.25343511e-01 8.04850638e-01 5.77079117e-01 -4.50463593e-01
-1.75581619e-01 2.33105108e-01 -3.56667638e-01 9.29506540e-01
9.88501549e-01 -6.04405820e-01 7.65670836e-02 -4.62430120e-01
2.11547285e-01 -1.01059489e-01 5.67235172e-01 -1.82263875e+00
7.16349408e-02 -3.10276579e-02 3.47288013e-01 -9.62533116e-01
2.75757521e-01 -1.08836257e+00 -1.19746625e-02 7.96892881e-01
5.49121380e-01 -1.66698873e-01 5.61536074e-01 5.45310915e-01
-4.22417015e-01 -4.72027153e-01 8.95246387e-01 -9.29016471e-02
-1.17144632e+00 1.72714636e-01 -9.17833090e-01 -5.11507273e-01
1.03437042e+00 -4.97817129e-01 2.46150821e-01 -1.50049523e-01
-7.26914346e-01 -9.32883248e-02 7.48282820e-02 2.34470665e-01
6.68454766e-01 -9.18744326e-01 -8.35377157e-01 2.28651091e-01
1.67888299e-01 5.62673211e-02 6.21571720e-01 6.22251153e-01
-8.61620665e-01 1.48637801e-01 -8.74345601e-01 -5.22931993e-01
-1.06588888e+00 7.08246455e-02 6.07944548e-01 -1.13184460e-01
-8.10362279e-01 4.40310061e-01 -3.57611984e-01 -3.46674651e-01
-2.46279091e-01 -4.41596471e-02 -7.21768320e-01 4.05746013e-01
3.71357381e-01 4.12096173e-01 6.66202486e-01 -8.65642011e-01
-5.10069191e-01 5.89877903e-01 -3.94842327e-02 3.72368656e-02
1.78013599e+00 3.07441175e-01 -5.26369922e-02 9.18917656e-02
8.37495387e-01 -1.42491996e-01 -6.64840937e-01 -1.94789413e-02
7.58192539e-02 -2.40951940e-01 3.97968203e-01 -4.70709711e-01
-1.25260520e+00 8.60969961e-01 1.03187716e+00 1.56765774e-01
1.04116440e+00 -3.60453069e-01 7.19761610e-01 1.01895285e+00
5.45483828e-01 -1.25439727e+00 -2.40590096e-01 4.09816116e-01
1.00482380e+00 -1.00249553e+00 -2.03913148e-03 2.65928209e-02
-2.24760383e-01 1.16681695e+00 3.53311121e-01 -2.95799702e-01
7.08021998e-01 1.76872686e-01 1.23117201e-01 -5.52554607e-01
2.82201827e-01 -3.90347689e-01 -3.62059981e-01 7.03091264e-01
-4.73814597e-03 2.45135482e-02 -3.93762916e-01 5.97614467e-01
-4.09230664e-02 -6.73019662e-02 6.52372420e-01 1.41943979e+00
-1.11546624e+00 -7.17875302e-01 -8.07016492e-01 5.61921537e-01
-4.37080771e-01 2.14062363e-01 1.58952683e-01 9.65608776e-01
3.18058848e-01 7.94934928e-01 2.35651255e-01 -1.98492497e-01
5.40375650e-01 -1.76709339e-01 3.15541834e-01 -3.09918225e-01
-2.76403219e-01 -5.64233780e-01 -2.55296752e-02 -1.76542044e-01
-3.81468564e-01 -5.41222036e-01 -1.37090194e+00 -9.18338448e-02
-3.46263856e-01 4.77499306e-01 1.07509911e+00 1.01369619e+00
2.49266364e-02 3.72924328e-01 6.83815181e-01 -9.72910285e-01
-5.95210254e-01 -1.14418685e+00 -1.20035291e+00 2.82806307e-01
2.14257568e-01 -1.13914096e+00 -5.68196833e-01 -4.07250851e-01]
|
[6.9623332023620605, 1.2477383613586426]
|
e4436145-11f9-4000-ae5e-bc69d967dfc7
|
unsupervised-dependency-graph-network
| null | null |
https://openreview.net/forum?id=yYJhaF4-dZ9
|
https://openreview.net/pdf?id=yYJhaF4-dZ9
|
Unsupervised Dependency Graph Network
|
Recent work has identified properties of pretrained self-attention models that mirror those of dependency parse structures. In particular, some self-attention heads correspond well to individual dependency types. Inspired by these developments, we propose a new competitive mechanism that encourages these attention heads to model different dependency relations. We introduce a new model, the Unsupervised Dependency Graph Network (UDGN), that can induce dependency structures from raw corpora and the masked language modeling task. Experiment results show that UDGN achieves very strong unsupervised dependency parsing performance without gold POS tags and any other external information. The competitive gated heads show a strong correlation with human-annotated dependency types. Furthermore, the UDGN can also achieve competitive performance on masked language modeling and sentence textual similarity tasks.
|
['Anonymous']
|
2021-11-16
| null | null | null |
acl-arr-november-2021-11
|
['unsupervised-dependency-parsing']
|
['natural-language-processing']
|
[-3.92604381e-01 4.93705481e-01 -2.53278583e-01 -7.63647079e-01
-7.92255402e-01 -4.37932491e-01 4.11575615e-01 8.09625089e-02
-3.69922668e-01 7.79433250e-01 8.04966271e-01 -4.37639505e-01
3.99237514e-01 -7.45896339e-01 -5.66205740e-01 -5.36838710e-01
-7.88444839e-03 8.10504973e-01 3.18764091e-01 -5.69407940e-01
-1.28202289e-01 3.90952975e-01 -9.35539067e-01 5.18586934e-01
9.82406437e-01 3.22407484e-01 5.51931918e-01 5.75334132e-01
-6.43788099e-01 1.21086609e+00 -6.76688731e-01 -8.35154474e-01
-2.32062578e-01 -6.12025440e-01 -1.15636384e+00 -3.84100765e-01
2.42564157e-02 7.13865682e-02 -3.81182075e-01 1.13433039e+00
6.14786804e-01 -5.63257141e-03 5.16436577e-01 -6.61350071e-01
-1.40081549e+00 1.58935380e+00 -3.66832227e-01 9.16288018e-01
4.03582811e-01 -2.53030546e-02 1.45526516e+00 -8.87064636e-01
4.84138072e-01 1.60181880e+00 4.68985856e-01 6.15618527e-01
-1.15368235e+00 -7.19171643e-01 4.21810269e-01 1.80224448e-01
-9.06630933e-01 -9.33411792e-02 6.51889801e-01 -1.89699799e-01
1.42780387e+00 -5.60325161e-02 -1.06612131e-01 1.25394785e+00
2.56982327e-01 7.04720914e-01 1.07739091e+00 -7.23657787e-01
-1.57199532e-01 1.78482607e-02 1.02613330e+00 5.98822355e-01
1.73719749e-01 -5.45949414e-02 -3.97920310e-01 1.74162298e-01
5.28939605e-01 -5.98739982e-01 -8.73358995e-02 3.49715292e-01
-7.11055219e-01 1.10477364e+00 7.16266215e-01 8.87460232e-01
-9.63066816e-02 -1.83913022e-01 3.71972531e-01 3.17534477e-01
6.11960471e-01 3.95444423e-01 -6.11774564e-01 6.22746944e-01
-1.93180606e-01 -4.24533367e-01 7.55047500e-01 1.17859173e+00
8.99197757e-01 5.71298525e-02 -5.07134914e-01 7.62528181e-01
2.39276469e-01 4.02855635e-01 7.13329971e-01 -3.07953805e-01
8.32510173e-01 5.91055095e-01 -3.73109311e-01 -7.16281593e-01
-7.69070923e-01 -2.76176661e-01 -8.96964848e-01 -4.88434553e-01
1.12490781e-01 -3.21813941e-01 -8.59541297e-01 1.83577502e+00
9.93867293e-02 -2.01501310e-01 3.58375013e-01 7.01817393e-01
1.39334500e+00 6.24737680e-01 6.10306859e-01 -3.73515010e-01
1.51708686e+00 -9.70757246e-01 -1.00768566e+00 -5.61956882e-01
9.94622529e-01 -6.33961082e-01 1.22522175e+00 -3.21242422e-01
-1.16688430e+00 -1.00906336e+00 -7.68680751e-01 -2.08473340e-01
-2.26675183e-01 -2.03570694e-01 7.54101753e-01 6.08224034e-01
-1.05283237e+00 6.53351784e-01 -5.46668291e-01 -5.49517214e-01
2.08193719e-01 5.55628061e-01 -3.48626912e-01 2.26153396e-02
-1.69543314e+00 1.12479520e+00 6.17352068e-01 1.17563315e-01
-4.65779275e-01 -4.57859963e-01 -1.09940898e+00 1.42685771e-01
1.94423217e-02 -4.79667962e-01 1.22178900e+00 -1.12057185e+00
-1.15597892e+00 1.22762692e+00 -1.75602287e-01 -6.02424920e-01
-3.25023949e-01 -4.34135020e-01 -6.46966696e-01 -1.09409317e-01
3.89018238e-01 3.75967383e-01 4.92477000e-01 -1.17333388e+00
-2.84198672e-01 -3.71142805e-01 1.34507507e-01 6.52438253e-02
-4.35642540e-01 8.61343801e-01 -4.04985428e-01 -6.99938059e-01
-2.46814247e-02 -7.67005503e-01 -5.20744205e-01 -1.31582332e+00
-7.71721482e-01 -7.36846268e-01 3.87728125e-01 -5.82845330e-01
1.39633214e+00 -1.82847250e+00 2.02054352e-01 -2.17270881e-01
2.56389901e-02 3.06092739e-01 -4.55834568e-01 4.09346104e-01
-5.14410853e-01 3.42906475e-01 -4.23718184e-01 -3.75983238e-01
-1.23805523e-01 7.86060929e-01 -2.34878123e-01 -1.75922569e-02
5.39779842e-01 1.29838824e+00 -9.11688030e-01 -7.66932487e-01
-3.01379651e-01 6.72475547e-02 -3.63792986e-01 7.93667197e-01
-1.56028926e-01 6.59112751e-01 -3.19774985e-01 2.64420331e-01
6.74017549e-01 -4.13226157e-01 7.14771450e-01 -3.23266625e-01
-1.02717720e-01 7.85409570e-01 -5.02869964e-01 1.49851108e+00
-5.08090317e-01 2.49003157e-01 -3.26427184e-02 -1.03945506e+00
1.10656607e+00 3.64003301e-01 -1.93608612e-01 -7.28466213e-01
4.89893615e-01 -1.36088580e-02 5.18702149e-01 -7.00403810e-01
2.12955654e-01 -3.28085512e-01 -4.85577047e-01 5.22863328e-01
6.05174899e-01 9.48721021e-02 3.85119975e-01 5.18342257e-01
1.08831012e+00 -2.55980581e-01 2.54978657e-01 -6.15949392e-01
7.29068220e-01 -4.79980297e-02 6.44164205e-01 6.20786190e-01
1.33888006e-01 6.06528580e-01 5.20794511e-01 -3.79078776e-01
-6.23485267e-01 -1.02308142e+00 4.91050771e-03 1.76874602e+00
-2.14680266e-02 -2.97720194e-01 -9.32940423e-01 -1.10300565e+00
-3.01732451e-01 7.51399577e-01 -6.61776543e-01 -3.70930992e-02
-1.15727174e+00 -1.06044745e+00 5.12694061e-01 1.12707174e+00
2.70093709e-01 -1.91709864e+00 1.28717378e-01 2.92689234e-01
-2.21806988e-01 -1.64401674e+00 -4.53765571e-01 8.62478316e-01
-7.52411962e-01 -8.53290796e-01 -3.50492984e-01 -1.52440846e+00
6.92678809e-01 -1.08682670e-01 1.78762090e+00 3.00913334e-01
4.00385819e-02 -1.40920699e-01 -7.88014293e-01 -1.85967162e-01
-7.81660557e-01 2.82152355e-01 1.79251313e-01 -2.42179692e-01
7.51343191e-01 -5.42095482e-01 4.98958305e-02 1.23084299e-02
-6.24896288e-01 -3.16046596e-01 7.22465992e-01 8.08194697e-01
2.82349169e-01 -4.30614144e-01 5.48539042e-01 -1.44393742e+00
7.48087108e-01 -6.80745125e-01 -2.73442298e-01 1.76520705e-01
-2.96645671e-01 4.73876745e-01 9.38294053e-01 -2.50906795e-01
-1.44586337e+00 1.73967719e-01 -6.11730397e-01 -3.88909206e-02
-3.37829024e-01 5.44911385e-01 -5.73725879e-01 3.92826498e-01
6.37303591e-01 -1.62227452e-01 -5.57801843e-01 -8.60652685e-01
5.98905981e-01 4.29053038e-01 8.75444174e-01 -6.46827519e-01
7.72399306e-01 6.54001310e-02 -4.99866188e-01 -5.80529571e-01
-1.51663911e+00 -2.34608397e-01 -1.26682782e+00 5.53903401e-01
1.38822818e+00 -7.23655105e-01 -1.96462825e-01 9.39329043e-02
-1.72933829e+00 -3.30499351e-01 -3.31552364e-02 4.00172979e-01
-1.59426019e-01 6.08442545e-01 -1.29423404e+00 -5.90621233e-01
-6.03689551e-01 -7.32993126e-01 6.02580726e-01 2.83611387e-01
-1.47039518e-01 -1.15233755e+00 3.61190021e-01 1.27206340e-01
4.42132503e-02 -2.73984432e-01 1.27726555e+00 -1.14062345e+00
-5.38462028e-02 7.28537068e-02 -3.75875324e-01 4.73044008e-01
1.16052940e-01 -1.42643750e-01 -9.87546146e-01 -1.46762989e-02
5.09076826e-02 -4.16215122e-01 9.26959872e-01 3.08131337e-01
6.06500864e-01 -1.09242983e-01 -3.30645084e-01 3.80990624e-01
1.13576031e+00 1.48224220e-01 4.89837021e-01 -3.46124060e-02
1.01309586e+00 8.62414181e-01 3.19180101e-01 5.39049394e-02
5.26685417e-01 3.55835170e-01 2.65977234e-01 -4.67265368e-01
-2.01107562e-01 -1.58178926e-01 5.22476196e-01 1.52335942e+00
-1.19069122e-01 -3.43562990e-01 -9.32276547e-01 5.18112361e-01
-1.71261406e+00 -6.91050053e-01 -7.57814467e-01 1.68158340e+00
9.35603678e-01 4.35668737e-01 3.41199542e-04 -2.29314655e-01
1.30781150e+00 2.63921004e-02 -1.13142401e-01 -8.77433956e-01
-6.10599518e-01 9.13371325e-01 3.14614177e-01 6.15516186e-01
-1.13689995e+00 1.51632178e+00 6.79588461e+00 2.48652041e-01
-6.04225695e-01 4.03307319e-01 5.73892117e-01 6.11875713e-01
-3.75638962e-01 3.69663872e-02 -1.20627403e+00 2.58224726e-01
1.28362131e+00 -1.78937703e-01 -2.42085874e-01 6.78059936e-01
-2.41886854e-01 4.06700164e-01 -1.05033374e+00 3.88740987e-01
2.59867068e-02 -1.11063409e+00 1.47584766e-01 -2.94064224e-01
5.99509597e-01 3.19012940e-01 -3.42354238e-01 5.81920624e-01
1.26903796e+00 -1.07966077e+00 2.59807855e-01 -1.33150712e-01
4.89908695e-01 -7.65172720e-01 9.97906029e-01 3.26967984e-01
-1.33980167e+00 -1.99271664e-01 -9.03847873e-01 -2.63738155e-01
4.32827562e-01 4.01973486e-01 -4.94124174e-01 6.95816457e-01
7.84496009e-01 4.58421767e-01 -5.84501266e-01 3.94438863e-01
-1.03941321e+00 8.14658821e-01 8.83780196e-02 -1.61553532e-01
2.89574802e-01 -1.41650112e-02 1.44465849e-01 1.74651110e+00
-1.49458051e-01 3.08892846e-01 1.38931587e-01 7.47320414e-01
-1.29947379e-01 4.36886311e-01 -4.81106102e-01 1.18498676e-01
4.93496448e-01 1.33776665e+00 -6.55323803e-01 -4.72028255e-01
-7.22298145e-01 1.07799506e+00 1.06141126e+00 1.59401655e-01
-9.53761339e-01 -1.40701056e-01 6.31206214e-01 -1.25646755e-01
4.68067259e-01 -1.24816947e-01 -2.35769302e-01 -9.89199460e-01
-3.35369617e-01 -5.77326298e-01 9.02343631e-01 -5.99959135e-01
-1.94504631e+00 1.45380521e+00 -1.68179840e-01 -6.62566543e-01
-9.12266672e-02 -6.40121996e-01 -1.01407146e+00 1.03964484e+00
-1.61539805e+00 -1.05014992e+00 3.58091220e-02 7.22242713e-01
5.95310390e-01 -1.14498204e-02 1.13162875e+00 1.26140177e-01
-9.79696512e-01 7.57246971e-01 -3.92062485e-01 8.19447100e-01
6.05143905e-01 -1.41291153e+00 1.14228666e+00 9.34602499e-01
4.92078692e-01 7.66614020e-01 4.50859159e-01 -6.29601061e-01
-8.64268780e-01 -1.15564239e+00 1.57791674e+00 -6.64625466e-01
9.22806323e-01 -4.42712992e-01 -1.33951485e+00 8.10107052e-01
8.03748906e-01 9.21262279e-02 8.20908189e-01 5.17282486e-01
-6.32714629e-01 1.07903697e-01 -6.72942281e-01 2.53813386e-01
1.39510465e+00 -4.90764350e-01 -1.21533346e+00 5.65620661e-01
1.19748068e+00 -1.23887748e-01 -6.75998092e-01 6.21284187e-01
-2.37391442e-01 -1.01706052e+00 8.07040155e-01 -1.23655999e+00
6.99705899e-01 2.01448604e-01 8.04495998e-03 -1.31585944e+00
-1.07363927e+00 -5.50683796e-01 1.00847743e-01 1.75094318e+00
7.96926856e-01 -3.85399550e-01 5.52308559e-01 6.30370259e-01
-4.97613430e-01 -2.02507168e-01 -6.21632814e-01 -9.98407960e-01
5.90263307e-01 -2.46612698e-01 5.43061316e-01 1.10250247e+00
2.59774804e-01 1.14048970e+00 -1.31832138e-01 3.99686098e-01
3.31209093e-01 2.45198488e-01 3.52000922e-01 -1.29367232e+00
-4.51555818e-01 -1.70417920e-01 5.77434618e-03 -1.08781445e+00
7.32575417e-01 -1.29361057e+00 3.18929017e-01 -1.58950591e+00
1.58288434e-01 -3.57385427e-01 -6.45267189e-01 4.70936954e-01
-7.93675244e-01 1.93380579e-01 8.03309530e-02 2.18387574e-01
-5.60153246e-01 3.03436637e-01 1.13351274e+00 3.81192490e-02
-2.04415917e-01 -1.98156103e-01 -9.54953969e-01 8.85458648e-01
9.58850980e-01 -8.60237718e-01 -5.77059761e-02 -8.46739411e-01
-8.59786123e-02 5.30663244e-02 -3.68105650e-01 -7.15379059e-01
1.13257602e-01 1.32195204e-01 9.71100852e-02 -5.62914789e-01
-2.79713809e-01 -3.39782506e-01 -4.03668076e-01 3.46701741e-01
-3.87176603e-01 4.87870336e-01 6.65019453e-02 2.53933132e-01
-2.26173431e-01 -4.80811447e-01 7.62543321e-01 -5.05088091e-01
-7.87414789e-01 1.53853804e-01 -4.98915374e-01 4.84250963e-01
4.93799865e-01 3.47228259e-01 -4.06478405e-01 -1.60447881e-02
-1.02682304e+00 2.05195740e-01 -1.42351896e-01 6.57492757e-01
2.08268881e-01 -1.21006274e+00 -1.04403090e+00 2.05417529e-01
2.37782467e-02 -1.26455948e-01 -6.52953535e-02 4.16377157e-01
-7.15938509e-02 4.74637002e-01 -9.34510231e-02 -2.08423823e-01
-1.28492117e+00 1.09255731e+00 -1.63836125e-02 -6.83524907e-01
-5.73633373e-01 1.45223939e+00 6.16273940e-01 -5.67945361e-01
3.89205813e-02 -3.57870847e-01 -6.08889103e-01 7.08013251e-02
5.09477556e-01 -2.42948115e-01 5.64379059e-02 -9.05614913e-01
-5.71079433e-01 4.86220598e-01 -2.86769360e-01 3.25025618e-01
1.25617266e+00 -4.58618514e-02 -4.60999519e-01 1.68288261e-01
1.06939888e+00 7.82195777e-02 -7.69396007e-01 -3.61766100e-01
6.95343316e-01 1.03582934e-01 -5.02751291e-01 -5.01494825e-01
-1.05653417e+00 8.90373409e-01 6.62916675e-02 5.33197165e-01
8.62883389e-01 8.07435036e-01 8.62787902e-01 3.82996380e-01
9.27766785e-02 -8.22113276e-01 2.08780393e-01 1.17420852e+00
7.65779257e-01 -1.28527796e+00 -4.60756510e-01 -8.29303622e-01
-8.86917531e-01 8.30454230e-01 1.10691464e+00 -4.52463567e-01
6.12749279e-01 8.36302400e-01 4.61927414e-01 -1.75621241e-01
-6.75553679e-01 -9.60566223e-01 1.62595987e-01 8.56840074e-01
1.05278027e+00 -9.87168029e-02 -6.51150763e-01 1.33101141e+00
-2.31257901e-01 -6.92173123e-01 2.08102703e-01 4.92462367e-01
-5.11810422e-01 -1.51766467e+00 -2.45522603e-01 -3.70847364e-03
-6.51056707e-01 -8.12204957e-01 -6.94953322e-01 6.09874547e-01
-9.91142467e-02 1.11024153e+00 2.43188009e-01 -4.23919708e-01
4.27241176e-01 2.96032071e-01 4.69571501e-01 -1.34108913e+00
-1.19418156e+00 -4.78618108e-02 3.84874910e-01 -3.21026146e-01
-4.58654404e-01 -4.86128852e-02 -1.75070608e+00 1.12699345e-01
-4.74556893e-01 5.70187330e-01 -8.98168422e-03 1.04647040e+00
1.68634579e-01 5.10858297e-01 8.42823982e-01 -6.34028494e-01
-1.86334074e-01 -1.24917948e+00 -3.91693771e-01 7.69956112e-01
-2.06182763e-01 -3.19756836e-01 -1.97392672e-01 8.21928233e-02]
|
[10.310148239135742, 9.62943172454834]
|
04636ee4-6e51-4eff-8070-71fde2767c1d
|
owq-lessons-learned-from-activation-outliers
|
2306.02272
| null |
https://arxiv.org/abs/2306.02272v2
|
https://arxiv.org/pdf/2306.02272v2.pdf
|
OWQ: Lessons learned from activation outliers for weight quantization in large language models
|
Large language models (LLMs) with hundreds of billions of parameters show impressive results across various language tasks using simple prompt tuning and few-shot examples, without the need for task-specific fine-tuning. However, their enormous size requires multiple server-grade GPUs even for inference, creating a significant cost barrier. To address this limitation, we introduce a novel post-training quantization method for weights with minimal quality degradation. While activation outliers are known to be problematic in activation quantization, our theoretical analysis suggests that we can identify factors contributing to weight quantization errors by considering activation outliers. We propose an innovative PTQ scheme called outlier-aware weight quantization (OWQ), which identifies vulnerable weights and allocates high-precision to them. Our extensive experiments demonstrate that the 3.01-bit models produced by OWQ exhibit comparable quality to the 4-bit models generated by OPTQ.
|
['Eunhyeok Park', 'HyungJun Kim', 'Taesu Kim', 'Jungyu Jin', 'Changhun Lee']
|
2023-06-04
| null | null | null | null |
['quantization']
|
['methodology']
|
[-3.62069234e-02 -3.80671442e-01 -3.93543273e-01 -3.80566955e-01
-1.18285179e+00 -1.98036470e-02 3.59546602e-01 4.15775239e-01
-7.72425652e-01 4.73895997e-01 9.33816805e-02 -6.71301544e-01
3.85930657e-01 -6.19275331e-01 -7.43658662e-01 -5.53490639e-01
-1.31432727e-01 4.33544725e-01 5.45897007e-01 -5.77400997e-02
6.60453737e-01 1.62307233e-01 -1.29441011e+00 3.99060011e-01
1.17365932e+00 8.73576581e-01 6.62720427e-02 8.82451832e-01
-4.56806511e-01 4.57063824e-01 -8.41986775e-01 -6.33648336e-01
2.59309739e-01 1.60857961e-01 -4.37139750e-01 -4.08239007e-01
6.31291509e-01 -6.38038576e-01 -3.09477389e-01 1.16252792e+00
8.22705984e-01 -1.11396857e-01 5.44499516e-01 -1.20006704e+00
-4.40710157e-01 9.38120604e-01 -7.29124367e-01 6.28038883e-01
-1.09431751e-01 3.16192925e-01 1.06473374e+00 -1.04255033e+00
7.30547160e-02 1.34652138e+00 7.32058287e-01 5.48100829e-01
-1.33696187e+00 -1.01330030e+00 -1.26001731e-01 1.79251686e-01
-1.74415088e+00 -9.02902246e-01 1.02501385e-01 -1.40309811e-01
1.60941494e+00 1.89301386e-01 2.91339725e-01 9.16762710e-01
4.01055008e-01 7.49245584e-01 6.43134117e-01 -5.35969555e-01
6.00501478e-01 -2.91150272e-01 4.04368043e-01 7.89270937e-01
8.62947941e-01 -1.83432937e-01 -1.03257930e+00 -1.01434767e+00
2.82852352e-01 -2.55308688e-01 1.96590368e-02 1.48177460e-01
-1.03631353e+00 8.54799926e-01 -1.29215628e-01 -9.67200473e-02
-1.04181610e-01 1.05251086e+00 7.43943691e-01 2.07836211e-01
5.57518482e-01 2.38122866e-01 -3.69239420e-01 -6.66814804e-01
-1.37589359e+00 1.23681352e-01 4.50403601e-01 9.73456383e-01
6.14502609e-01 5.91221213e-01 -4.96710420e-01 7.52872825e-01
4.77659941e-01 5.24822831e-01 6.83589458e-01 -8.86496246e-01
7.57934988e-01 8.44800547e-02 1.36933357e-01 -5.26006639e-01
-2.15877295e-01 -2.74560183e-01 -7.57343650e-01 -6.40960410e-02
4.46580082e-01 1.34212777e-01 -9.44119751e-01 1.57262719e+00
7.28081092e-02 2.45308667e-01 -2.65912652e-01 5.17398179e-01
-1.81598961e-02 8.40561330e-01 2.89886355e-01 -1.31603897e-01
1.38902295e+00 -8.58111143e-01 -7.49413013e-01 -2.52344012e-01
1.13690853e+00 -8.42324436e-01 1.62722790e+00 7.39273667e-01
-1.30139470e+00 -1.84094548e-01 -1.28670931e+00 -1.54595360e-01
4.66743559e-02 -1.32203668e-01 8.26030135e-01 1.27790809e+00
-1.03927922e+00 5.50198078e-01 -1.08949065e+00 5.89433173e-03
5.83566964e-01 5.03128350e-01 3.67293030e-01 4.41030003e-02
-1.06970882e+00 6.47590160e-01 4.60885584e-01 -4.22841966e-01
-6.85921013e-01 -8.59291136e-01 -5.20763695e-01 2.85017401e-01
2.21387133e-01 -3.66724461e-01 1.46771240e+00 -5.47572017e-01
-1.44675040e+00 4.59187716e-01 -6.25045061e-01 -7.16098726e-01
4.59248036e-01 -3.50377738e-01 -4.29836839e-01 -7.85746872e-02
-2.48589024e-01 4.42047715e-01 1.16331601e+00 -7.64466166e-01
-4.46800202e-01 -1.38627827e-01 -3.89286309e-01 -2.30609164e-01
-8.91629696e-01 3.20419937e-01 -7.56848752e-01 -9.65727925e-01
-2.06089951e-02 -5.05637586e-01 -3.66555840e-01 -5.74856624e-03
-4.54910964e-01 -2.76778102e-01 2.29160920e-01 -3.33854795e-01
2.03249264e+00 -2.14859796e+00 -6.88155890e-01 5.48385680e-01
3.78270328e-01 4.09907669e-01 -1.88061640e-01 6.08015396e-02
4.60950404e-01 4.84269649e-01 -2.94120945e-02 -6.10738397e-01
4.20513034e-01 2.33097956e-01 -6.56357527e-01 3.86282057e-01
-4.38341312e-02 8.31854284e-01 -9.51698363e-01 -5.34756303e-01
-2.31695190e-01 9.97562110e-02 -1.03925562e+00 -6.27288297e-02
-2.55668968e-01 -5.53099811e-01 -2.43352368e-01 7.54233778e-01
8.96757066e-01 -3.57081354e-01 -1.28575489e-01 -1.15869969e-01
1.20776981e-01 5.35977721e-01 -1.10013449e+00 1.46696007e+00
-4.22386795e-01 5.11188686e-01 -2.37292945e-01 -5.39399028e-01
6.78287506e-01 1.20416500e-01 1.70212686e-02 -7.81105578e-01
3.56419459e-02 6.24163806e-01 1.36001790e-02 -1.33880720e-01
1.16613305e+00 2.03731522e-01 -2.25970939e-01 8.13597977e-01
-1.81089610e-01 -5.68880588e-02 2.15645865e-01 4.73571807e-01
1.21741557e+00 -6.81387007e-01 6.98247328e-02 -2.48085335e-01
2.81203054e-02 -2.47293651e-01 5.23388147e-01 1.12537348e+00
-4.38948721e-01 4.11456913e-01 3.91494244e-01 -3.71270210e-01
-1.43993413e+00 -1.13123631e+00 -1.70331180e-01 1.37407899e+00
-2.76425332e-01 -1.03452885e+00 -8.54784310e-01 -1.50038868e-01
3.35490890e-02 9.08622444e-01 1.10629149e-01 -2.69716114e-01
-5.69127619e-01 -1.26436377e+00 1.17272663e+00 4.60488170e-01
3.11262667e-01 -3.92780542e-01 -5.89967608e-01 4.18817878e-01
2.86910795e-02 -1.07045078e+00 -9.80877817e-01 3.52013916e-01
-1.09354460e+00 -4.26575899e-01 -4.56264943e-01 -1.85047165e-01
6.09436035e-01 1.44854039e-01 1.26692748e+00 4.35382009e-01
-2.66740143e-01 -8.84948373e-02 4.38195281e-03 -3.28970492e-01
-5.73054373e-01 1.61627918e-01 6.40365660e-01 -4.03052419e-01
6.80948377e-01 -4.69588637e-01 -6.47685230e-01 -2.79579703e-02
-7.54858613e-01 -2.38550112e-01 4.88252908e-01 8.35697949e-01
6.71203792e-01 -1.30675867e-01 4.68902171e-01 -7.27227926e-01
9.47061539e-01 -3.12456399e-01 -8.32412660e-01 1.65596172e-01
-9.23170865e-01 4.15140241e-01 8.98473501e-01 -6.66379869e-01
-5.88189006e-01 -2.70788103e-01 -3.29302102e-01 -3.73752773e-01
3.97462547e-01 -2.27416903e-02 1.33072898e-01 -4.89700027e-02
9.26587820e-01 1.68272600e-01 -4.59241569e-01 -2.85420358e-01
4.69217479e-01 8.36711526e-01 3.61305445e-01 -9.68368351e-01
6.16306305e-01 3.27945560e-01 -4.90841657e-01 -6.72370255e-01
-4.65816319e-01 -2.64684677e-01 -2.71639284e-02 3.37610751e-01
4.35840726e-01 -1.03206635e+00 -5.21713972e-01 3.15620154e-01
-1.11633778e+00 -6.73003376e-01 -1.66777402e-01 2.43322238e-01
-3.91181231e-01 7.75759816e-01 -8.39617312e-01 -9.89583373e-01
-8.36873651e-01 -1.25919175e+00 1.03876531e+00 -1.94983497e-01
-5.81932247e-01 -5.50277710e-01 -2.26826862e-01 1.90052874e-02
5.40660024e-01 -6.23390913e-01 1.23605716e+00 -4.93809015e-01
-6.71189010e-01 -9.59697515e-02 -3.97666931e-01 3.56004275e-02
-3.53837967e-01 9.91510749e-02 -1.04665828e+00 -5.05070210e-01
-2.18175754e-01 -5.22048414e-01 8.85998666e-01 2.41514146e-01
1.80330944e+00 -3.17836314e-01 -2.63227224e-01 7.69498885e-01
1.32984579e+00 -3.45268518e-01 5.03910661e-01 1.01488039e-01
4.94605124e-01 -3.80443543e-01 3.01008552e-01 8.97521138e-01
1.89321548e-01 7.98853755e-01 8.87561664e-02 3.27623129e-01
8.28932300e-02 -1.53139129e-01 7.80033231e-01 1.27839243e+00
2.67852157e-01 -4.08138156e-01 -1.03494060e+00 4.65024471e-01
-1.53252065e+00 -6.00863039e-01 -2.51953989e-01 2.57343960e+00
1.13699710e+00 7.59916842e-01 -1.27472833e-01 3.14527035e-01
4.77067143e-01 5.63311167e-02 -6.25898421e-01 -1.15342057e+00
8.18739608e-02 5.57854950e-01 1.10470295e+00 5.51276684e-01
-6.38021708e-01 1.07398450e+00 7.63922977e+00 1.71125805e+00
-8.74541402e-01 4.90364969e-01 8.05264533e-01 -5.86021900e-01
-6.39088571e-01 -2.58324951e-01 -1.22711170e+00 8.77438605e-01
1.60949159e+00 -2.53898174e-01 5.72221100e-01 6.88938320e-01
3.88082922e-01 -1.26103729e-01 -9.32273567e-01 1.20114148e+00
-1.47062212e-01 -1.36695123e+00 3.12601060e-01 8.15262571e-02
7.73901403e-01 4.00538474e-01 2.23044962e-01 3.11719000e-01
5.49578071e-01 -8.84870410e-01 8.31811249e-01 2.39058509e-01
1.08945799e+00 -8.70270967e-01 3.43941301e-01 3.02762836e-01
-9.72577751e-01 -9.64788049e-02 -9.40194607e-01 -2.33090650e-02
1.00882970e-01 1.07830679e+00 -6.62295222e-01 -2.04747200e-01
5.55758238e-01 -2.30230272e-01 -6.16919816e-01 1.09624183e+00
1.10921271e-01 1.14500797e+00 -7.58083403e-01 -1.72411963e-01
1.91945404e-01 2.83088952e-01 2.13364720e-01 1.40172267e+00
6.48036182e-01 -1.49275288e-01 -5.57994433e-02 8.09866130e-01
-3.31495017e-01 1.32385433e-01 -5.85355721e-02 -8.00768137e-02
1.05490768e+00 7.34918714e-01 -6.44344151e-01 -7.81498313e-01
-3.23964745e-01 1.14201343e+00 4.31041539e-01 1.99681506e-01
-9.42633271e-01 -4.23402339e-01 8.88038158e-01 -5.70538007e-02
1.45059422e-01 -5.59234679e-01 -7.14565873e-01 -1.29068398e+00
-8.08389261e-02 -8.93607855e-01 1.96279466e-01 -4.39358830e-01
-1.12523115e+00 9.65977982e-02 -2.61634678e-01 -8.34183753e-01
-6.66881874e-02 -4.75774676e-01 -6.46680593e-01 7.62633502e-01
-1.55189526e+00 -3.50166231e-01 7.68076554e-02 3.54316205e-01
5.83540320e-01 -1.17938571e-01 8.40391219e-01 5.44271529e-01
-5.07849932e-01 1.58463085e+00 5.69268107e-01 -2.83291489e-01
8.04168880e-01 -1.13055408e+00 1.14166009e+00 8.19509745e-01
1.32873533e-02 7.59579957e-01 6.26462698e-01 -5.79263151e-01
-1.60773385e+00 -1.11067665e+00 1.20640397e+00 -2.77749568e-01
6.96987569e-01 -6.33453012e-01 -9.95478511e-01 1.41561329e-01
-1.90594554e-01 2.04096317e-01 6.77648842e-01 1.35595985e-02
-6.73325717e-01 4.93746698e-02 -1.06727219e+00 9.88439023e-01
9.47333574e-01 -5.47665238e-01 -1.43429339e-01 4.41622227e-01
9.35717285e-01 -4.06503856e-01 -5.73384106e-01 -4.12231199e-02
4.03613627e-01 -8.60646009e-01 1.04765618e+00 -3.18529218e-01
4.06950302e-02 -7.19368756e-02 -4.03261244e-01 -9.69083130e-01
-1.73601374e-01 -9.01502788e-01 -7.63847709e-01 9.12509918e-01
3.98318231e-01 -5.82971096e-01 9.69253480e-01 7.02408791e-01
-1.82653844e-01 -6.56490505e-01 -1.14020216e+00 -1.00398481e+00
2.35995024e-01 -1.17853701e+00 7.82194018e-01 6.16187394e-01
-1.96742918e-02 -1.08654223e-01 -3.67322326e-01 1.52617842e-01
7.36588895e-01 -5.61430514e-01 6.52950108e-01 -6.60252035e-01
-3.14235061e-01 -6.22281313e-01 -1.14679493e-01 -1.28348577e+00
1.56230843e-02 -9.21350658e-01 -5.40548563e-02 -9.09243166e-01
1.11620128e-01 -7.59132743e-01 -3.34412634e-01 4.73046333e-01
-4.98678952e-01 6.21854901e-01 7.62103649e-04 3.63564849e-01
-8.78137708e-01 5.38132250e-01 6.49486184e-01 -2.02678978e-01
4.53109145e-02 -2.42817834e-01 -4.50186789e-01 7.36947536e-01
8.73521686e-01 -6.77525938e-01 -2.76923686e-01 -8.20527315e-01
7.63694406e-01 -4.85932261e-01 -1.98299453e-01 -1.11051416e+00
3.48407060e-01 -9.69841611e-03 1.87842637e-01 -5.95641434e-01
3.42413634e-01 -4.28199649e-01 -4.40332413e-01 8.01543772e-01
-2.70678312e-01 4.29120451e-01 3.35208178e-01 5.30461848e-01
2.61255711e-01 -4.81518835e-01 8.09042990e-01 1.90072805e-01
-6.22099578e-01 3.02746296e-01 -7.83468723e-01 3.63365889e-01
5.00167489e-01 -2.55786061e-01 -2.39176318e-01 -2.17291787e-01
-2.21758887e-01 1.73039630e-01 5.23780286e-01 -6.70675263e-02
6.35123372e-01 -1.39890385e+00 -5.50395489e-01 4.10761744e-01
1.45720214e-01 -2.14765757e-01 -1.38515625e-02 5.03999650e-01
-6.20154679e-01 4.50402975e-01 3.90805304e-01 -6.44187450e-01
-1.10579896e+00 3.23034674e-01 5.73809855e-02 -4.27325189e-01
-5.23945510e-01 1.10924923e+00 -3.94815058e-01 -6.25166148e-02
4.64625448e-01 -6.89767599e-01 7.19464719e-01 -2.45927930e-01
8.70518684e-01 5.19537032e-01 4.12905812e-01 -1.34301558e-01
-8.09830651e-02 1.78717658e-01 -3.23814780e-01 -1.24870524e-01
6.66058779e-01 -5.98365329e-02 1.66342929e-02 4.03367668e-01
1.00081050e+00 2.07747683e-01 -1.04030359e+00 -4.84443277e-01
4.10138339e-01 -6.56722367e-01 2.17577934e-01 -4.03361559e-01
-6.85590446e-01 1.19525635e+00 7.52254367e-01 -7.61838406e-02
8.89472067e-01 -5.84732771e-01 1.51017928e+00 6.83633924e-01
5.99151075e-01 -1.50038767e+00 3.04691851e-01 9.08823192e-01
9.78535041e-02 -1.00937521e+00 1.07661217e-01 -1.20550528e-01
-2.42645860e-01 9.60405886e-01 5.56895077e-01 7.68672302e-02
4.21504438e-01 6.16310358e-01 -1.11836694e-01 1.51721328e-01
-1.17555404e+00 1.32873341e-01 -1.29707083e-01 3.15471232e-01
3.96051407e-01 6.43059909e-02 -3.66305232e-01 6.52453899e-01
-3.02058369e-01 -3.99772562e-02 4.82763112e-01 8.13609183e-01
-9.64645147e-01 -1.20118415e+00 -5.53337693e-01 8.67276132e-01
-7.74918020e-01 -7.83610821e-01 2.77979046e-01 6.49037734e-02
-9.23679918e-02 7.02786267e-01 4.18664426e-01 -3.29508215e-01
-1.08451612e-01 3.91456038e-01 4.18088675e-01 -5.21983325e-01
-6.89044476e-01 3.92355360e-02 5.08980043e-02 -6.85826421e-01
4.62937772e-01 -1.44428566e-01 -1.42922950e+00 -9.46002483e-01
-5.46915591e-01 5.21761812e-02 6.97440743e-01 7.05551505e-01
6.33818150e-01 3.90459836e-01 3.40111226e-01 -6.23020172e-01
-1.17834544e+00 -7.84657061e-01 -4.86529559e-01 2.02773795e-01
1.64973482e-01 -2.48083919e-01 -3.38779300e-01 -3.42263699e-01]
|
[8.663995742797852, 3.45025897026062]
|
ad9a24ee-10fa-492d-9f67-c3cb174f0325
|
question-generation-based-on-grammar
| null | null |
https://aclanthology.org/2022.coling-1.562
|
https://aclanthology.org/2022.coling-1.562.pdf
|
Question Generation Based on Grammar Knowledge and Fine-grained Classification
|
Question generation is the task of automatically generating questions based on given context and answers, and there are problems that the types of questions and answers do not match. In minority languages such as Tibetan, since the grammar rules are complex and the training data is small, the related research on question generation is still in its infancy. To solve the above problems, this paper constructs a question type classifier and a question generator. We perform fine-grained division of question types and integrate grammatical knowledge into question type classifiers to improve the accuracy of question types. Then, the types predicted by the question type classifier are fed into the question generator. Our model improves the accuracy of interrogative words in generated questions, and the BLEU-4 on SQuAD reaches 17.52, the BLEU-4 on HotpotQA reaches 19.31, the BLEU-4 on TibetanQA reaches 25.58.
|
['Xiaobing Zhao', 'Zhengcuo Dan', 'Sisi Liu', 'Yuan Sun']
| null | null | null | null |
coling-2022-10
|
['question-generation']
|
['natural-language-processing']
|
[-1.79579318e-01 3.31778288e-01 1.01653649e-03 -4.23036575e-01
-1.00687134e+00 -8.27422917e-01 5.37850976e-01 3.48068774e-02
-2.84400433e-01 1.09780455e+00 3.35705400e-01 -6.56426549e-01
1.10529497e-01 -1.25047374e+00 -2.91212708e-01 -2.76247654e-02
4.98155087e-01 6.28482461e-01 5.68062663e-01 -9.27315891e-01
5.59018731e-01 -3.19147646e-01 -1.66529930e+00 6.84080720e-01
1.66992962e+00 1.13053322e+00 1.91765800e-01 9.45852935e-01
-1.06584477e+00 1.18658876e+00 -1.08953440e+00 -6.81480944e-01
-1.88705072e-01 -1.25012410e+00 -1.46206796e+00 -3.75278145e-01
4.21636999e-01 -8.01783875e-02 -9.18484852e-02 8.73356998e-01
2.48688072e-01 2.58135181e-02 7.17515290e-01 -1.17047381e+00
-9.35686886e-01 6.03217900e-01 2.50598162e-01 1.83119029e-01
8.09729040e-01 -7.40961581e-02 1.30105078e+00 -7.26991236e-01
4.66056377e-01 1.24819219e+00 4.00030971e-01 8.88728917e-01
-5.16323864e-01 -5.44997573e-01 -1.76012572e-02 4.43161994e-01
-1.00405061e+00 -1.26008123e-01 3.59303474e-01 -3.86049271e-01
5.94380736e-01 5.74046493e-01 5.70056498e-01 6.50243104e-01
2.60144949e-01 5.62994480e-01 1.28593481e+00 -6.70654058e-01
2.43687004e-01 8.89092088e-02 3.94283533e-01 6.59249663e-01
1.93131536e-01 -2.52947509e-01 -3.89621645e-01 -2.27144331e-01
4.78841364e-01 -5.25338113e-01 -8.90488103e-02 7.66018689e-01
-9.47588623e-01 1.25940251e+00 3.52425963e-01 4.67702955e-01
-4.00562622e-02 -2.27581248e-01 1.08089976e-01 7.00908482e-01
3.14156234e-01 9.75885332e-01 -6.38734460e-01 -1.91255838e-01
-7.95329869e-01 9.94729400e-01 1.33779621e+00 1.16387451e+00
8.68444085e-01 -2.28099406e-01 -6.81067169e-01 1.02902520e+00
2.35764042e-01 5.75291336e-01 7.72003531e-01 -9.28330302e-01
4.99488086e-01 1.10440063e+00 1.28190741e-01 -7.30215669e-01
-2.13507399e-01 -1.14286520e-01 -4.25812185e-01 -3.43308479e-01
8.68089855e-01 -4.89764839e-01 -8.18882227e-01 1.53520429e+00
3.31576556e-01 -5.17205060e-01 6.25444874e-02 7.02271521e-01
1.64683926e+00 9.17290211e-01 3.49279791e-02 2.41226368e-02
1.93658674e+00 -1.01000476e+00 -9.34323192e-01 -3.96486849e-01
7.24945426e-01 -8.83542120e-01 1.36074626e+00 -1.44557476e-01
-1.04612923e+00 -6.56522512e-01 -5.03080189e-01 -3.00981134e-01
-3.67804199e-01 3.19883265e-02 3.52259636e-01 8.94608974e-01
-7.26235449e-01 -4.45769262e-03 1.55011490e-01 -3.54869008e-01
1.47945154e-02 -6.80772588e-02 1.53308466e-01 -1.02130234e-01
-1.59164894e+00 9.55063343e-01 5.50267577e-01 -2.41053700e-01
-5.24763107e-01 -6.11553252e-01 -7.53344655e-01 -1.02991208e-01
4.35038239e-01 -7.59611785e-01 1.62766230e+00 -7.39963353e-01
-1.42713428e+00 7.57630885e-01 -2.46792540e-01 -3.48821208e-02
1.19303018e-01 1.60660058e-01 -4.82407272e-01 1.08988427e-01
5.07161260e-01 6.76694930e-01 3.05198640e-01 -9.99795914e-01
-1.17771482e+00 -2.47733444e-01 5.14826715e-01 2.57002294e-01
-3.38672698e-02 1.14534326e-01 -1.31372269e-03 -4.72089887e-01
1.43051282e-01 -6.11396730e-01 -2.10866064e-01 -5.03083587e-01
-4.57118973e-02 -1.05215943e+00 4.48947996e-01 -1.03423941e+00
1.58553362e+00 -1.39979208e+00 -6.23377323e-01 -1.14745256e-02
8.90655257e-03 1.17938749e-01 -1.24653183e-01 4.51983541e-01
3.99428576e-01 2.86990702e-01 -1.15608223e-01 7.00070441e-01
1.04774497e-01 4.25991982e-01 -4.36180532e-01 -5.80778539e-01
2.40517035e-01 1.19076908e+00 -1.02393329e+00 -6.68251812e-01
-5.85510910e-01 -5.87118924e-01 -7.84718215e-01 8.01262915e-01
-7.71708608e-01 1.63301587e-01 -7.64788449e-01 6.18202031e-01
3.71231824e-01 -9.81361866e-02 1.96570046e-02 3.46848667e-01
1.13862995e-02 1.07586336e+00 -8.52374375e-01 1.23320186e+00
-6.11375451e-01 3.13812464e-01 -2.28858382e-01 -5.72810948e-01
1.39682472e+00 3.14003378e-01 -4.05821681e-01 -8.90184164e-01
1.19045258e-01 5.97831547e-01 1.98970467e-01 -1.09308755e+00
8.98935378e-01 -5.31770885e-01 -4.67863172e-01 4.49041694e-01
1.18039481e-01 -6.73434734e-01 6.80464745e-01 -1.47780776e-02
1.03363669e+00 3.48825343e-02 1.99988157e-01 -4.35970694e-01
7.09154904e-01 4.66744095e-01 4.16312426e-01 8.79732490e-01
1.75912753e-01 6.91232443e-01 7.02183902e-01 -2.45774463e-01
-8.37041259e-01 -9.10382450e-01 -2.03740299e-02 1.40227175e+00
-5.41102923e-02 -3.96393061e-01 -1.13374960e+00 -1.01135480e+00
-2.91058093e-01 1.04281330e+00 -4.71042305e-01 -1.59467608e-02
-7.91510105e-01 -4.84221905e-01 7.58475125e-01 2.59074956e-01
6.62955284e-01 -1.48566186e+00 -1.76477373e-01 3.85910988e-01
-1.02973580e+00 -8.78078938e-01 -4.99142349e-01 -4.39668894e-01
-7.94739366e-01 -1.12099195e+00 -5.29980540e-01 -1.24253285e+00
5.67321539e-01 -1.40223771e-01 1.61240017e+00 3.59197617e-01
2.43371159e-01 -2.54195537e-02 -8.53506029e-01 -5.60095549e-01
-6.97279394e-01 5.51916122e-01 -6.40593529e-01 -2.97914326e-01
4.06874359e-01 -9.65630934e-02 -5.01305044e-01 2.91882277e-01
-8.60293269e-01 -2.56879061e-01 1.67580128e-01 7.81317651e-01
9.01288539e-02 -4.24244165e-01 1.30959201e+00 -1.02011263e+00
1.11837995e+00 -7.23998666e-01 -3.62231135e-01 4.87417221e-01
-5.59998155e-01 -2.31554196e-03 6.42948091e-01 -8.55724514e-02
-1.32647073e+00 -8.19449723e-01 -7.23582864e-01 8.63059163e-01
-1.26212806e-01 5.74401081e-01 -2.91912735e-01 2.38071561e-01
9.52872396e-01 2.33980715e-01 -6.83281198e-02 -4.32490647e-01
3.18201125e-01 9.52322066e-01 3.50574404e-01 -6.50103271e-01
6.82821095e-01 -1.66754127e-01 -5.66942632e-01 -7.44829357e-01
-1.28876483e+00 -7.51974359e-02 -1.38380587e-01 -3.05170357e-01
9.54333007e-01 -5.72501957e-01 -5.78403473e-01 1.96235642e-01
-1.26601481e+00 -2.45285794e-01 -4.20264870e-01 9.79405791e-02
-4.46322143e-01 3.70422035e-01 -5.61709464e-01 -7.60492921e-01
-5.09995997e-01 -6.44010127e-01 6.74390435e-01 6.69124007e-01
-6.97142899e-01 -8.77363265e-01 1.33437410e-01 8.66494715e-01
3.13208103e-01 -1.77859306e-01 1.56392705e+00 -9.38974202e-01
-4.32224363e-01 -2.09026057e-02 -1.17157198e-01 4.14875686e-01
1.89716011e-01 -3.78264815e-01 -5.72161138e-01 3.95321846e-01
3.18064123e-01 -3.82623851e-01 4.28304493e-01 -7.76915066e-03
9.85792518e-01 -6.67141855e-01 2.53849983e-01 -1.89015809e-02
1.06216431e+00 3.53463829e-01 7.60141015e-01 2.88600653e-01
3.43430728e-01 1.10540998e+00 7.87517726e-01 4.46418673e-02
1.17292392e+00 2.95400769e-01 8.96546096e-02 5.77232063e-01
-2.44352907e-01 -5.85002303e-01 3.52563202e-01 1.13819897e+00
2.86862701e-01 -2.36898676e-01 -9.63453472e-01 9.05266285e-01
-1.40175545e+00 -8.66254032e-01 -7.43610561e-01 1.88825893e+00
1.29243863e+00 -1.28434658e-01 1.24823868e-01 1.32807925e-01
6.22354567e-01 2.01387797e-02 2.78832503e-02 -6.60385907e-01
-8.59219059e-02 5.71630716e-01 -1.58455953e-01 5.31009138e-01
-4.54654068e-01 9.85861301e-01 6.57617044e+00 8.65586162e-01
-4.75152522e-01 -4.22144216e-03 4.69977498e-01 3.64951134e-01
-8.04513156e-01 3.44670177e-01 -9.60828364e-01 5.74365258e-01
1.08913791e+00 -4.28535372e-01 1.40871003e-01 4.41899300e-01
-5.06723896e-02 -3.65628242e-01 -8.57150376e-01 6.59390986e-01
3.35736930e-01 -1.35134113e+00 3.79462868e-01 -2.52127945e-01
7.08806992e-01 -5.69812417e-01 -3.17022294e-01 8.21626246e-01
3.88536781e-01 -1.12932086e+00 6.65899575e-01 4.78202045e-01
4.54168200e-01 -7.06977844e-01 8.52415979e-01 6.41416728e-01
-7.53600717e-01 -1.42149463e-01 -5.83413541e-01 -4.12872821e-01
5.90014346e-02 5.31682670e-01 -8.77779603e-01 6.02082193e-01
5.60312450e-01 -2.82356590e-01 -8.62814426e-01 9.58952188e-01
-7.23396301e-01 1.06975567e+00 -2.89137028e-02 -7.86889255e-01
3.05848956e-01 -2.13231787e-01 8.61372352e-02 8.00872564e-01
5.04917085e-01 4.63072330e-01 1.01099350e-01 7.99069107e-01
-2.89149314e-01 5.30262053e-01 -1.10996902e-01 2.20373988e-01
4.84629124e-01 1.18637121e+00 -3.05780321e-01 -4.54220563e-01
-3.24044138e-01 5.39434493e-01 2.18415350e-01 5.07651679e-02
-5.93823493e-01 -8.28289628e-01 2.27354586e-01 2.49129280e-01
1.13664577e-02 9.71800014e-02 -5.64841151e-01 -1.20849073e+00
2.81425357e-01 -1.32411492e+00 5.68497837e-01 -8.05178344e-01
-1.36787999e+00 5.93250453e-01 -5.82757518e-02 -8.00384164e-01
-6.08363211e-01 -5.55417001e-01 -8.70954096e-01 8.87582183e-01
-1.49342811e+00 -9.70278203e-01 -3.54871780e-01 8.73312131e-02
6.41333103e-01 -1.83863446e-01 8.05238068e-01 2.46236518e-01
-1.59995273e-01 8.25234592e-01 -2.15999603e-01 3.83214861e-01
5.20444512e-01 -1.30127347e+00 5.02712846e-01 4.41331685e-01
-1.35606259e-01 4.17748332e-01 5.26846707e-01 -6.05901003e-01
-1.05513239e+00 -1.11641061e+00 1.84980929e+00 -6.00826204e-01
6.05279565e-01 -2.96830952e-01 -1.15174353e+00 2.57870972e-01
4.13590640e-01 -6.20900750e-01 9.24364746e-01 9.25576761e-02
-2.57367432e-01 4.40538041e-02 -1.15676963e+00 6.86554790e-01
7.35099018e-01 -5.49043536e-01 -1.22360599e+00 3.99695575e-01
7.44145751e-01 -2.34031215e-01 -8.05802047e-01 6.61723167e-02
3.28898162e-01 -7.65233338e-01 2.76859999e-01 -8.98365378e-01
6.01544738e-01 -1.98667422e-01 -1.23322241e-01 -1.16430187e+00
-2.28410780e-01 -4.96377379e-01 1.10506907e-01 1.50641859e+00
9.18936908e-01 -5.66238403e-01 9.00849700e-01 5.53581834e-01
-1.63610145e-01 -8.02297533e-01 -9.98729110e-01 -5.81656635e-01
7.04554081e-01 1.97794754e-02 1.00017917e+00 7.31004119e-01
-3.58063690e-02 9.05280709e-01 5.64715862e-02 -3.05973083e-01
1.22491606e-01 3.24269056e-01 6.49773240e-01 -1.12873185e+00
-8.62058997e-02 -2.45520383e-01 2.80450910e-01 -1.37815106e+00
8.08674693e-02 -1.03155017e+00 1.54332429e-01 -1.83228946e+00
-1.04936041e-01 -4.59725559e-01 3.92432958e-01 1.76579863e-01
-9.29074347e-01 8.74922052e-03 1.76342010e-01 -1.99781418e-01
-3.66751254e-01 5.70267081e-01 1.54929698e+00 1.16340779e-01
9.59042981e-02 1.46556750e-01 -1.23464704e+00 5.97764909e-01
1.03052914e+00 -4.38187063e-01 -2.37972721e-01 -4.90375698e-01
6.01658106e-01 3.21750671e-01 -2.55301464e-02 -8.31299484e-01
6.01995364e-02 -4.33735758e-01 6.78362921e-02 -7.08037972e-01
-2.06694067e-01 -2.75677919e-01 -5.16189814e-01 5.09706259e-01
-5.44727802e-01 3.03726912e-01 -2.80738384e-01 2.54020840e-02
-4.10120726e-01 -9.95103896e-01 6.00670040e-01 -4.42220926e-01
-4.35191929e-01 1.18322931e-01 -7.38495529e-01 1.07951367e+00
6.82863653e-01 -3.36546451e-01 -3.88205320e-01 -6.51480675e-01
-2.02614665e-01 5.95028818e-01 2.05661100e-03 6.83859408e-01
5.07705867e-01 -1.56848907e+00 -1.14786899e+00 1.12471525e-02
2.37983003e-01 8.25245455e-02 1.52074903e-01 1.50203466e-01
-7.86273777e-01 4.01084572e-01 -2.37272121e-02 -7.84577336e-03
-9.49594855e-01 -1.46282628e-01 3.98059458e-01 -5.55706620e-01
1.90804601e-01 8.61068308e-01 1.28138021e-01 -9.55467403e-01
-3.06696177e-01 -3.59839350e-01 -7.41369069e-01 3.10034931e-01
6.53881848e-01 5.61991751e-01 7.46741220e-02 -3.07229996e-01
3.42336670e-02 3.20401490e-01 -1.36656919e-02 -2.07517624e-01
8.71570349e-01 -1.77329667e-02 -6.37193739e-01 3.17780674e-01
9.93051589e-01 9.09358636e-02 -3.09449315e-01 -1.23123743e-01
4.20972943e-01 -2.14670718e-01 -5.75358748e-01 -8.89335036e-01
-4.26383108e-01 7.20287800e-01 1.40291572e-01 7.58579910e-01
9.32723343e-01 2.56082684e-01 1.29204738e+00 4.87671286e-01
2.25098237e-01 -1.28834116e+00 3.00597489e-01 1.48393023e+00
9.93474185e-01 -9.17188406e-01 -5.22524655e-01 -4.70080674e-01
-5.23933053e-01 9.45023358e-01 1.09175372e+00 -1.13947734e-01
2.26574734e-01 -2.28498250e-01 4.50811803e-01 -1.00986905e-01
-8.77365708e-01 -2.81125456e-01 3.17911148e-01 5.80227256e-01
7.40161419e-01 4.82380390e-02 -1.01874113e+00 9.82158840e-01
-1.16182411e+00 -2.63664722e-01 7.36925364e-01 8.94345820e-01
-1.01393962e+00 -1.46394813e+00 -5.68494141e-01 9.87729967e-01
-8.03246379e-01 -2.81999260e-01 -5.09353936e-01 4.58448082e-01
2.13476270e-01 1.54043412e+00 3.07018813e-02 -3.01287353e-01
5.03466785e-01 4.18255031e-01 4.92891550e-01 -1.13442457e+00
-1.11113977e+00 -6.06974959e-01 6.55359864e-01 5.06680533e-02
8.48630071e-02 -3.55082542e-01 -1.19587791e+00 -2.60865033e-01
-5.40640950e-01 8.99025202e-01 2.48362124e-01 1.15498340e+00
1.34710848e-01 2.32822612e-01 7.86980033e-01 4.26932216e-01
-7.92445540e-01 -1.16482341e+00 -3.71071309e-01 3.27409923e-01
7.91047066e-02 -7.94229209e-02 -1.74242601e-01 -1.43692285e-01]
|
[11.511557579040527, 8.160409927368164]
|
8e4f9de1-ff24-41f3-99a7-d817bad57c2c
|
prediction-of-cytochrome-p450-mediated
|
1811.09366
| null |
http://arxiv.org/abs/1811.09366v1
|
http://arxiv.org/pdf/1811.09366v1.pdf
|
Prediction of Cytochrome P450-Mediated Metabolism Using a Combination of QSAR Derived Reactivity and Induced Fit Docking
|
Prediction of metabolism in cytochrome P450s remains to be a crucial yet
challenging topic in discovering and designing drugs, agrochemicals and
nutritional supplements. The problem is challenging because the rate of P450
metabolism depends upon both the intrinsic chemical reactivity of the site and
the protein-ligand geometry that is energetically accessible in the active site
of a given P450 isozyme. We have addressed this problem using a two-level
screening system. The first level implements an empirical QSAR-based scoring
function employing the local chemical motifs to characterize the intrinsic
reactivity. The second level uses molecular docking and molecular mechanics to
account for the geometrical effects, including induced-fit effects in the
protein which can be very important in P450 interactions with ligands. This
approach has achieved high accuracy for both the P450 3A4 and 2D6 isoforms. In
identifying at least one metabolic site in the top two ranked positions, the
prediction rate can reach as high as 92.7% for the test set of isoform 3A4. For
the 2D6 isoform, 100% accuracy is achieved on this basic evaluation metric,
and, because this active site is considerably smaller and more selective than
3A4, very high precision is attained for full prediction of all metabolic
sites. The method also requires considerably less CPU time than our previous
efforts, which involved a large number of expensive simulations for each ligand
to be evaluated. After screening using the empirical score function, only a few
best candidates are left for each ligand, making the number of necessary
estimations in the second level very small, which significantly reduces the
computation time.
|
[]
|
2018-11-23
| null | null | null | null |
['molecular-docking']
|
['medical']
|
[ 1.90550685e-01 4.34312038e-02 -2.41684079e-01 4.59074713e-02
-6.31291866e-01 -7.89409041e-01 2.50331819e-01 5.27123213e-01
-2.39434898e-01 1.24379110e+00 -2.73214608e-01 -3.97424877e-01
-9.17213410e-03 -6.49166405e-01 -7.15122342e-01 -1.14740562e+00
-1.84222832e-01 4.26616460e-01 3.95279795e-01 -3.20455432e-02
3.16062182e-01 8.19947898e-01 -1.45241916e+00 -2.17031091e-02
1.21997881e+00 5.71872115e-01 3.41918230e-01 9.52251852e-02
6.14148472e-03 1.18595511e-01 -4.66449916e-01 -3.49252939e-01
-8.73116702e-02 -7.48072326e-01 -7.42480516e-01 -2.62842804e-01
1.39496475e-01 3.59432787e-01 5.09809196e-01 9.53854918e-01
7.64302790e-01 1.27847567e-02 1.02080667e+00 -3.63342792e-01
1.25294849e-01 3.17219310e-02 -4.65739995e-01 -3.35696861e-02
7.00602412e-01 1.84650570e-01 9.53724027e-01 -1.11228287e+00
5.75413048e-01 7.94455409e-01 3.93323392e-01 2.92240202e-01
-1.52559257e+00 -5.54756045e-01 -3.48228216e-01 3.91389757e-01
-1.74659145e+00 -2.78302372e-01 2.88224548e-01 -5.78234494e-01
1.06397557e+00 3.29146713e-01 9.09141898e-01 4.67700779e-01
3.66982520e-01 -1.17918458e-02 1.03816342e+00 -2.89584726e-01
7.23368049e-01 1.17350802e-01 -1.91696063e-01 5.88667333e-01
4.22896296e-01 -2.00180694e-01 -5.04692793e-01 -4.77233499e-01
5.19107401e-01 -3.19379300e-01 -2.93432921e-01 -5.49212277e-01
-7.56239057e-01 9.01632071e-01 2.36981779e-01 2.78773546e-01
-6.19142950e-01 -5.58885694e-01 1.64633572e-01 -2.41042659e-01
-4.94051389e-02 7.81426311e-01 -7.68123686e-01 -1.53705865e-01
-6.92813218e-01 1.40004352e-01 1.05954564e+00 2.20062494e-01
7.46296823e-01 -3.72094750e-01 2.44926497e-01 8.12113643e-01
3.77614349e-01 2.10196391e-01 9.21028331e-02 -5.17416775e-01
5.15690260e-02 7.32178390e-01 2.47002751e-01 -4.73408967e-01
-5.34858286e-01 -1.05515338e-01 -5.39520085e-01 4.03777122e-01
9.26134527e-01 -3.35789889e-01 -4.43632990e-01 1.59613073e+00
6.20354235e-01 -3.35387141e-01 1.19105563e-01 6.79399550e-01
3.76732558e-01 9.40908492e-01 3.53851646e-01 -9.53101754e-01
1.64435196e+00 -6.29293501e-01 -2.81330943e-01 3.79258752e-01
4.87913489e-01 -9.15805757e-01 6.03406012e-01 5.54956555e-01
-1.13471401e+00 -1.85742930e-01 -1.20480347e+00 3.65922779e-01
-3.72630328e-01 3.81722935e-02 7.25953400e-01 7.55239129e-01
-6.87192678e-01 8.42965484e-01 -7.07001925e-01 -4.03456628e-01
9.88850892e-02 9.19783592e-01 -5.78844249e-01 1.99587390e-01
-1.23576021e+00 1.33255756e+00 4.05951828e-01 -1.17173813e-01
-7.44491637e-01 -7.01255679e-01 -5.23764729e-01 2.86471128e-01
1.48503738e-03 -3.52722824e-01 7.74602234e-01 -7.39565909e-01
-1.86461234e+00 3.10588568e-01 -3.85278523e-01 -3.61902937e-02
2.50057012e-01 3.31137240e-01 -1.73473954e-01 2.30472654e-01
3.17591354e-02 3.15670133e-01 1.59455568e-01 -7.01537490e-01
-1.94595233e-01 -4.95160222e-01 -2.41617024e-01 4.59069103e-01
1.91949874e-01 2.58942366e-01 -8.73445198e-02 -8.13907944e-03
1.19200073e-01 -7.23872662e-01 -4.40014720e-01 -2.97227353e-01
-1.04741067e-01 -3.88373405e-01 3.81755717e-02 -4.02364969e-01
9.00447905e-01 -1.68264830e+00 1.63314745e-01 6.71123087e-01
-2.11938396e-01 4.27998155e-01 1.84127778e-01 8.88003349e-01
-5.81557095e-01 1.41548693e-01 -7.72935003e-02 8.97301853e-01
-5.48668683e-01 -4.77501243e-01 2.20025331e-01 7.29996383e-01
2.28296623e-01 3.19513053e-01 -7.31200337e-01 -2.01816201e-01
1.08397968e-01 7.66271055e-01 -5.69402754e-01 -9.62778181e-02
-1.27249464e-01 4.52770025e-01 -7.04411447e-01 5.64921737e-01
7.19304860e-01 -1.73098385e-01 9.78092194e-01 -2.09650815e-01
-6.13343477e-01 5.63891232e-01 -1.23100924e+00 1.33133256e+00
-2.01372594e-01 -6.15412891e-02 -4.66716111e-01 -5.54286182e-01
8.37025404e-01 6.50036693e-01 7.79850185e-01 -6.42372251e-01
-1.76886767e-01 4.29865628e-01 4.35321003e-01 -1.77193314e-01
-3.35197926e-01 -3.42593223e-01 4.97049779e-01 -2.77318843e-02
1.46782279e-01 2.07310319e-01 6.50774598e-01 -1.55572742e-01
7.97086656e-01 4.81897414e-01 7.90518522e-01 -7.01626956e-01
9.95545864e-01 1.54295014e-02 6.76496565e-01 2.60031112e-02
1.87471375e-01 2.42263556e-01 8.46687555e-01 -4.85679299e-01
-9.36679959e-01 -6.25684261e-01 -7.03725755e-01 5.69554985e-01
2.54609823e-01 -4.74897891e-01 -1.15677154e+00 -3.45916688e-01
-3.29155207e-01 4.54640031e-01 -2.37330601e-01 -1.85967572e-02
-1.96319491e-01 -1.35203028e+00 2.32354894e-01 -1.83238834e-01
-1.66006573e-02 -7.74108887e-01 -1.95850015e-01 5.80397725e-01
1.30689517e-01 -3.11264634e-01 -1.70534223e-01 6.47647381e-01
-9.19482708e-01 -1.59139228e+00 -5.87307453e-01 -4.23417568e-01
6.34531558e-01 -1.09003365e-01 6.92639709e-01 -2.50088304e-01
-3.02686661e-01 -6.67989969e-01 -5.15609467e-03 -4.15026784e-01
-3.24027330e-01 -2.77193427e-01 1.09130032e-01 -1.31183907e-01
4.55375105e-01 -7.38666534e-01 -6.60445511e-01 6.22614563e-01
-3.79318237e-01 -2.22037598e-01 7.28191257e-01 7.52465129e-01
8.01001489e-01 1.24742612e-01 5.05198538e-01 -6.51988089e-01
1.63087025e-01 -5.28454006e-01 -8.95369053e-01 1.50398806e-01
-5.93013346e-01 2.18164936e-01 6.76037848e-01 -3.48902613e-01
-9.76535082e-01 6.86299801e-01 -3.70118678e-01 5.70745289e-01
-1.06204629e-01 5.76036394e-01 -7.07731664e-01 -3.40336919e-01
8.33874226e-01 2.42502227e-01 1.65198192e-01 -6.50050044e-01
-2.32411131e-01 2.01515973e-01 -1.63437009e-01 -5.18643200e-01
6.15029991e-01 3.34183685e-02 5.61194003e-01 -1.27428436e+00
-4.12616670e-01 -4.40843940e-01 -5.18677711e-01 3.19130644e-02
8.70912731e-01 -8.59589696e-01 -1.37343764e+00 -1.54708885e-02
-1.16234410e+00 1.79507613e-01 2.22796753e-01 7.88598895e-01
-4.34831053e-01 7.46053278e-01 -2.89458811e-01 -6.46033049e-01
-2.16497168e-01 -1.70922434e+00 5.43265879e-01 1.10538639e-01
-2.68240094e-01 -7.73365855e-01 3.91665727e-01 2.87439257e-01
-5.44513538e-02 4.28842455e-01 1.25068641e+00 -7.99055040e-01
-5.14708340e-01 -1.95273444e-01 1.85067639e-01 -1.00260563e-02
7.26184845e-02 2.91242748e-01 -7.21571028e-01 -1.39583185e-01
-2.15326875e-01 -1.73208818e-01 5.03689826e-01 3.31717938e-01
6.90986872e-01 -1.24377869e-01 -2.78354526e-01 3.48161370e-01
1.48030818e+00 6.46985233e-01 9.21372712e-01 2.64717042e-01
3.49341899e-01 5.86987138e-01 8.07721436e-01 3.09801340e-01
-1.23487458e-01 9.97850060e-01 5.40643215e-01 -4.76141453e-01
4.32053089e-01 -6.20508334e-03 6.75925851e-01 1.28751293e-01
-7.08779752e-01 3.68032865e-02 -6.57695532e-01 -4.24195491e-02
-1.47294950e+00 -1.08548558e+00 -7.24487424e-01 2.77572870e+00
1.22011065e+00 -2.26857170e-01 4.72313941e-01 1.15739238e-02
6.52164459e-01 -4.30535764e-01 -5.38890421e-01 -5.22269189e-01
-3.38284001e-02 5.31387508e-01 4.74254400e-01 6.48971081e-01
-7.18945682e-01 5.93622208e-01 7.03366661e+00 8.19470108e-01
-9.69562352e-01 -3.38099480e-01 4.13858086e-01 5.19869477e-02
1.21776484e-01 4.49039191e-01 -9.66349185e-01 6.16078258e-01
1.18516970e+00 -2.72144526e-01 2.56312758e-01 8.21968973e-01
6.53158724e-01 -4.48938608e-01 -1.08792770e+00 6.52670026e-01
-4.57543761e-01 -1.34514689e+00 7.75316358e-02 7.38860309e-01
5.22891760e-01 -1.20885551e-01 -5.29151380e-01 -3.61290187e-01
-2.62631357e-01 -9.20653343e-01 3.32708836e-01 1.89608648e-01
6.94678128e-01 -1.15048110e+00 6.12302899e-01 2.70359337e-01
-9.64596927e-01 1.91210195e-01 -7.07215965e-01 1.61697462e-01
-1.05885029e-01 6.55153334e-01 -7.58434594e-01 3.76865178e-01
1.07482672e-01 2.36806720e-01 -2.89227754e-01 1.22178984e+00
-3.09577316e-01 2.73003817e-01 -3.49805087e-01 -2.50939578e-01
8.86586532e-02 -8.09467435e-01 2.80507743e-01 8.83856595e-01
1.81852743e-01 1.98605478e-01 -4.33043800e-02 7.43256927e-01
6.29492924e-02 7.87317753e-01 -3.52352023e-01 7.38814399e-02
2.25054860e-01 1.27480543e+00 -5.33048332e-01 -1.41693473e-01
-4.85022396e-01 7.59058058e-01 4.38916720e-02 1.60914809e-01
-7.01945364e-01 -3.45222503e-01 1.03674221e+00 2.65344352e-01
1.90611213e-01 2.37533569e-01 4.26323473e-01 -9.13174212e-01
-8.63044262e-02 -8.31678867e-01 3.50168765e-01 -2.02172786e-01
-6.55057788e-01 3.77970874e-01 -1.94591790e-01 -9.34799492e-01
-9.22146812e-02 -8.20924282e-01 -5.94667912e-01 1.30643070e+00
-1.30276811e+00 -6.49355590e-01 2.50851244e-01 3.23459893e-01
3.49604964e-01 1.12633474e-01 1.19996500e+00 3.26656640e-01
-7.51053691e-01 2.08479226e-01 4.05742466e-01 -7.17000902e-01
7.50112534e-01 -1.45442414e+00 -1.62879989e-01 5.81506014e-01
-2.78248161e-01 9.74117517e-01 7.20685840e-01 -5.46084702e-01
-1.38520014e+00 -6.89381957e-01 1.00817204e+00 -3.62701476e-01
4.02976573e-01 -2.46513441e-01 -7.49537528e-01 -1.55200049e-01
-1.41222373e-01 -4.91889894e-01 1.35175312e+00 -9.64108482e-03
-1.24405339e-01 1.73543274e-01 -1.07058561e+00 4.44149822e-01
4.15848196e-01 -1.28097132e-01 -4.67794873e-02 7.69574642e-01
1.04093276e-01 -1.76880240e-01 -1.33878970e+00 8.98155496e-02
6.57285094e-01 -9.96399045e-01 9.47441220e-01 -4.33093101e-01
6.18302152e-02 -8.45895648e-01 2.71396369e-01 -9.26771045e-01
-4.55892324e-01 -6.78533614e-01 1.01925999e-01 7.63124108e-01
9.58364606e-01 -5.01938343e-01 7.74940431e-01 5.33359170e-01
1.71667963e-01 -8.78349423e-01 -8.61643374e-01 -6.44298971e-01
1.34917779e-03 4.13950175e-01 4.86589313e-01 7.62565672e-01
3.80626112e-01 7.48540223e-01 -1.73723519e-01 1.75269723e-01
4.26597774e-01 2.01747734e-02 3.99715304e-01 -1.45623147e+00
-2.71939218e-01 -2.26769641e-01 -2.93965995e-01 -6.35007381e-01
-1.34830832e-01 -7.66538262e-01 -4.69339728e-01 -1.24502397e+00
3.92142296e-01 -2.09665280e-02 -1.91480055e-01 4.28197086e-01
-8.66806321e-03 2.15313777e-01 -4.20723915e-01 6.68935105e-02
-1.62191883e-01 1.71297625e-01 9.35031891e-01 2.28753667e-02
-5.46923101e-01 1.51265591e-01 -8.01233292e-01 6.95075929e-01
6.38664603e-01 -4.04421926e-01 -2.60430008e-01 5.77903807e-01
3.77979487e-01 5.86757287e-02 -1.95138976e-01 -5.28786719e-01
-1.54629633e-01 -4.53006327e-01 7.07216680e-01 -3.40956837e-01
3.22569102e-01 -7.42078483e-01 7.82757163e-01 7.92308152e-01
2.14027062e-01 -4.80799228e-01 4.58424948e-02 2.94389784e-01
-1.00783976e-02 -4.08931047e-01 1.34573328e+00 -1.39368340e-01
-2.96303213e-01 2.93799102e-01 -8.38443756e-01 -5.92235148e-01
1.06729567e+00 -6.02070093e-01 1.64777189e-01 1.54841870e-01
-9.32228148e-01 -2.04603210e-01 7.57788777e-01 -1.66300714e-01
1.47357389e-01 -9.01498079e-01 -3.31986189e-01 -1.08129270e-01
2.68938720e-01 -3.04006189e-01 1.14182457e-01 1.02531457e+00
-9.89566624e-01 5.63955545e-01 -2.63489723e-01 -3.72075021e-01
-1.43548083e+00 5.78627288e-01 7.21053660e-01 -2.06070408e-01
-8.83298868e-04 3.22193921e-01 3.31244081e-01 1.73082799e-01
-1.81165516e-01 1.62902445e-01 -6.48291051e-01 1.93190753e-01
7.68333972e-01 6.52991474e-01 4.01319057e-01 -8.10047448e-01
-6.50137663e-01 6.09072149e-01 7.25379661e-02 5.68101346e-01
1.43268096e+00 3.39496434e-01 -4.44224894e-01 -1.66353181e-01
1.06461763e+00 2.49198630e-01 -9.73253906e-01 4.99726683e-01
5.90604879e-02 -2.79358715e-01 3.60873365e-03 -9.74305928e-01
-4.22964424e-01 6.23485088e-01 5.46302557e-01 -1.39750198e-01
8.08972716e-01 -1.97868779e-01 1.31916836e-01 4.17057157e-01
4.71571773e-01 -8.88872802e-01 -3.52005690e-01 6.69201091e-02
7.01117396e-01 -7.76149929e-01 2.74607748e-01 -7.04206288e-01
-4.55891937e-01 1.13205779e+00 4.66395348e-01 3.65077138e-01
1.67452648e-01 -3.97614390e-01 -2.15466484e-01 -2.81710654e-01
-6.09017730e-01 -6.42018812e-03 1.34941667e-01 4.93649662e-01
8.78554225e-01 -1.30837858e-01 -1.17754245e+00 1.55736282e-01
2.65333742e-01 -4.94550079e-01 4.86067683e-01 5.59038460e-01
-4.70125109e-01 -1.90635800e+00 -2.66933620e-01 -5.96424602e-02
-8.76979411e-01 -1.40711486e-01 -7.05975533e-01 6.73315287e-01
1.45431325e-01 8.29446673e-01 -5.52298963e-01 3.42878640e-01
4.70634878e-01 1.41722843e-01 5.61464012e-01 -5.86963892e-01
-5.88113725e-01 5.87717891e-01 1.48809820e-01 -7.38268256e-01
-2.63209134e-01 -6.10240459e-01 -1.41375697e+00 -1.22081265e-01
-9.71679330e-01 1.00464833e+00 9.88362610e-01 7.79853880e-01
3.99900675e-01 -7.10906610e-02 9.09500957e-01 -7.82845140e-01
-3.30840528e-01 -6.38804674e-01 -1.03166509e+00 -5.55136390e-02
-1.62407070e-01 -8.03511202e-01 -1.29853994e-01 -6.46309287e-04]
|
[4.813024997711182, 5.361324787139893]
|
fe36f5ae-3657-4945-b804-f6576ff64213
|
point-transformer
|
2011.00931
| null |
https://arxiv.org/abs/2011.00931v2
|
https://arxiv.org/pdf/2011.00931v2.pdf
|
Point Transformer
|
In this work, we present Point Transformer, a deep neural network that operates directly on unordered and unstructured point sets. We design Point Transformer to extract local and global features and relate both representations by introducing the local-global attention mechanism, which aims to capture spatial point relations and shape information. For that purpose, we propose SortNet, as part of the Point Transformer, which induces input permutation invariance by selecting points based on a learned score. The output of Point Transformer is a sorted and permutation invariant feature list that can directly be incorporated into common computer vision applications. We evaluate our approach on standard classification and part segmentation benchmarks to demonstrate competitive results compared to the prior work. Code is publicly available at: https://github.com/engelnico/point-transformer
|
['Klaus Dietmayer', 'Vasileios Belagiannis', 'Nico Engel']
|
2020-11-02
| null | null | null | null |
['3d-object-classification', '3d-part-segmentation']
|
['computer-vision', 'computer-vision']
|
[ 5.29533476e-02 -1.37557313e-01 -2.49710232e-01 -5.26995242e-01
-5.56234598e-01 -6.71730161e-01 5.78156412e-01 8.41238275e-02
-1.90430731e-01 2.68062770e-01 1.69338629e-01 -1.41218796e-01
-3.37568641e-01 -1.10711753e+00 -1.13735378e+00 -3.20153624e-01
4.27342989e-02 6.06835008e-01 4.50905770e-01 -1.30239531e-01
4.64591593e-01 7.81062305e-01 -1.61423361e+00 4.50944453e-01
6.80779397e-01 1.12277436e+00 3.08714151e-01 3.47054154e-01
-3.81721109e-02 3.50991637e-01 -2.57553637e-01 -1.63538605e-01
5.74976087e-01 -1.53499320e-01 -1.01824987e+00 -1.30822301e-01
6.61057651e-01 -2.61835098e-01 -3.30653965e-01 1.07660949e+00
3.47679108e-01 2.96565026e-01 4.90543067e-01 -1.21707785e+00
-1.04279780e+00 6.54553354e-01 -4.44588155e-01 2.83165008e-01
1.51357040e-01 2.78811723e-01 1.46547616e+00 -1.11861932e+00
5.70204794e-01 1.22661042e+00 6.66465402e-01 1.47224352e-01
-1.12723529e+00 -5.32823801e-01 8.23501050e-02 3.47562909e-01
-1.32497609e+00 -3.26886863e-01 1.02047074e+00 -1.55450687e-01
1.00297415e+00 4.27524745e-01 8.31617236e-01 6.87216282e-01
1.86384872e-01 1.10863721e+00 7.51561344e-01 6.44812128e-04
2.26113141e-01 -5.77658176e-01 4.51717675e-01 7.64533341e-01
2.12337039e-02 7.00557977e-03 -4.22365516e-01 -3.23532969e-02
1.15196836e+00 4.53933626e-01 -2.77115852e-01 -8.26480150e-01
-1.29379857e+00 6.48690999e-01 1.22973144e+00 1.20086670e-01
-5.50466299e-01 4.50589448e-01 2.83350587e-01 1.55207917e-01
1.82411551e-01 5.07829905e-01 -5.44630885e-01 -6.93999082e-02
-6.28401279e-01 2.44886115e-01 4.34791118e-01 1.07913578e+00
9.64502275e-01 -4.30734992e-01 -5.56950986e-01 9.63810384e-01
3.01173598e-01 4.74415749e-01 3.81489038e-01 -9.83705342e-01
3.59724104e-01 1.08171260e+00 -3.02639931e-01 -9.58659470e-01
-4.13980603e-01 -4.33661163e-01 -6.16662979e-01 1.20809823e-01
-8.02653581e-02 4.95837003e-01 -1.38429320e+00 1.51320648e+00
1.70895815e-01 3.62812966e-01 -3.99718732e-01 1.17692506e+00
9.57097709e-01 4.33927178e-01 -8.03223774e-02 5.88525236e-01
1.39083183e+00 -1.13658118e+00 -2.02122241e-01 -5.20303175e-02
1.10360354e-01 -5.03807604e-01 1.35156631e+00 1.87918115e-02
-1.25998676e+00 -5.90955317e-01 -9.64026749e-01 -6.37873888e-01
-5.59962511e-01 1.15569815e-01 5.43211818e-01 -1.50224706e-02
-1.20699036e+00 9.01000977e-01 -1.29671097e+00 -2.84154862e-01
9.00373638e-01 6.76793277e-01 -4.34581518e-01 -2.21785228e-03
-8.58823478e-01 2.67192304e-01 1.81758612e-01 1.14839926e-01
-5.10870874e-01 -8.04889202e-01 -1.09363532e+00 4.64232236e-01
4.85172495e-02 -1.06546807e+00 1.34936070e+00 -6.01462126e-01
-1.51608777e+00 8.16520333e-01 -1.75069690e-01 -4.34074849e-01
2.54463434e-01 -4.36885446e-01 2.54146904e-01 3.32699955e-01
4.10636276e-01 9.03261185e-01 6.09079897e-01 -1.14205587e+00
-4.53592092e-01 -6.06800437e-01 1.05158865e-01 2.36591771e-01
5.79307042e-02 -2.02036798e-01 -9.78806019e-01 -6.85831010e-01
3.89517844e-01 -8.47347617e-01 -2.10197225e-01 1.67795971e-01
-7.14491308e-01 -5.71584582e-01 9.37081814e-01 -3.78220558e-01
6.65609837e-01 -2.20191836e+00 2.04166099e-01 4.32549596e-01
3.04065406e-01 1.78750068e-01 -3.57705384e-01 3.52182060e-01
-2.34027043e-01 1.44469872e-01 -4.96486217e-01 -2.60990083e-01
2.96799570e-01 1.67118177e-01 -1.52834356e-01 4.56784219e-01
5.09946465e-01 1.41043305e+00 -7.92854965e-01 -1.16365492e-01
4.95379299e-01 4.54806179e-01 -8.08708012e-01 -9.32182837e-03
-2.57978499e-01 4.05717120e-02 -6.61777794e-01 7.95656919e-01
7.98947811e-01 -4.39547330e-01 -2.75201917e-01 -3.60246599e-01
-1.38344793e-02 5.97146928e-01 -8.25218499e-01 2.06952953e+00
-1.71521455e-01 3.23952824e-01 -2.00768426e-01 -9.18379068e-01
8.44046116e-01 -2.23516777e-01 5.63299417e-01 -6.84657753e-01
2.68366665e-01 3.80910264e-04 -1.05093844e-01 3.78202721e-02
4.17932242e-01 3.78195882e-01 -1.87767789e-01 2.72159487e-01
2.20969841e-01 -1.04595631e-01 6.04742989e-02 1.57781571e-01
1.34939110e+00 2.63077110e-01 2.69928366e-01 -4.45671171e-01
4.04590666e-01 7.08075613e-02 4.51068699e-01 7.19708264e-01
-5.46021536e-02 9.00107265e-01 4.50441778e-01 -6.43137872e-01
-8.69335234e-01 -1.54244614e+00 -2.20215812e-01 1.11903405e+00
4.84720290e-01 -4.16490346e-01 -4.05400544e-01 -7.74365544e-01
3.20750624e-01 3.19444209e-01 -6.90888584e-01 -1.22657932e-01
-7.75375605e-01 -3.81070338e-02 1.63904205e-01 1.03082216e+00
6.02208078e-01 -1.33089161e+00 -6.05098784e-01 2.52420511e-02
2.18728468e-01 -7.98752129e-01 -7.42472291e-01 3.71981740e-01
-7.61833966e-01 -1.09697080e+00 -5.23619533e-01 -9.86317813e-01
7.71444976e-01 2.93081462e-01 1.21015751e+00 1.54718548e-01
-1.68707058e-01 3.65678072e-01 -4.26375568e-01 -3.35931033e-01
3.67196530e-01 5.24409711e-01 -2.89227754e-01 -1.47290453e-01
5.05830407e-01 -7.87421703e-01 -1.00230074e+00 4.16886777e-01
-9.01064992e-01 7.37980101e-03 6.85698509e-01 7.88694322e-01
9.50785279e-01 -3.85412037e-01 2.17642099e-01 -6.77031100e-01
4.46878761e-01 -2.69773185e-01 -7.04231381e-01 -1.32348999e-01
5.13734035e-02 1.47084177e-01 5.21289766e-01 1.21079870e-01
-3.25423509e-01 3.47639889e-01 -3.85456920e-01 -7.76146472e-01
-2.25030005e-01 4.75337535e-01 -3.09012204e-01 -3.85313854e-02
1.85122550e-01 2.50719339e-01 1.41643360e-02 -5.65581083e-01
5.38220942e-01 3.70168328e-01 6.98354483e-01 -6.21360421e-01
9.00014877e-01 6.83269143e-01 -1.24602377e-01 -5.38398445e-01
-6.79264724e-01 -5.91361821e-01 -9.21835303e-01 2.60695189e-01
7.40681946e-01 -6.37003064e-01 -6.41079843e-01 4.07209575e-01
-1.09858990e+00 -5.97153962e-01 -6.04894400e-01 1.55499458e-01
-7.95654833e-01 9.47345048e-02 -6.21335864e-01 -1.31068364e-01
-4.90753740e-01 -1.19959223e+00 1.66207266e+00 1.99950516e-01
-1.04892589e-01 -7.64979064e-01 1.19764004e-02 -1.57791190e-02
2.43188068e-01 6.76245540e-02 7.73881972e-01 -8.30094337e-01
-1.13454378e+00 -3.52220312e-02 -5.33874214e-01 2.27444038e-01
1.80929437e-01 -1.58870190e-01 -5.80058992e-01 -2.74122059e-01
-3.76203150e-01 -1.31494299e-01 1.16842580e+00 4.23378378e-01
1.76535559e+00 -2.84369737e-01 -5.74125588e-01 1.02100646e+00
1.44012296e+00 -1.07880933e-02 7.29094326e-01 3.55045319e-01
9.40829694e-01 1.27525449e-01 2.73872823e-01 2.76537567e-01
5.41578472e-01 6.29703939e-01 5.41953683e-01 -1.11736998e-01
-2.17746735e-01 -4.56619203e-01 -1.13442428e-02 7.03493893e-01
-3.80022340e-02 -1.50380313e-01 -1.03738737e+00 6.41669571e-01
-1.96000111e+00 -9.32497442e-01 7.31025189e-02 1.83534801e+00
5.08083165e-01 1.21295815e-02 -1.13156419e-02 6.70630718e-04
6.54031038e-01 2.06525683e-01 -5.38608909e-01 -2.85667747e-01
1.81067109e-01 6.60605729e-01 7.01840580e-01 3.27553958e-01
-1.36858857e+00 1.03574049e+00 5.69041443e+00 6.89474642e-01
-1.21369195e+00 -7.20833018e-02 4.23601031e-01 -1.10250358e-02
-3.83301318e-01 -2.38050804e-01 -4.81753677e-01 3.96420091e-01
2.37338603e-01 -6.65854961e-02 3.25077683e-01 7.52410889e-01
1.04497690e-02 4.37954277e-01 -1.24566865e+00 8.33577871e-01
-1.03880085e-01 -1.42565846e+00 3.02515298e-01 1.41488820e-01
4.81973082e-01 5.52938998e-01 1.98570862e-01 3.13845694e-01
3.88230443e-01 -1.00634694e+00 7.00687528e-01 4.46755677e-01
6.21445894e-01 -7.57600427e-01 5.55291951e-01 -1.46987140e-02
-1.41593254e+00 -7.43150935e-02 -4.99841243e-01 8.00534934e-02
8.10297355e-02 4.38199967e-01 -6.32783651e-01 5.55510342e-01
1.01843381e+00 1.24132121e+00 -5.94032407e-01 1.36049020e+00
-3.48986179e-01 4.97977585e-01 -6.27271473e-01 1.37688026e-01
3.47354233e-01 -2.07556501e-01 6.33188546e-01 9.59905505e-01
3.11750382e-01 7.36721754e-02 2.37899944e-01 1.21601081e+00
-3.97272557e-01 -1.50345303e-02 -7.50113785e-01 2.02797487e-01
5.48012733e-01 1.29680657e+00 -9.84544992e-01 -1.52054012e-01
-3.09035003e-01 1.12579119e+00 6.40916944e-01 1.71655849e-01
-8.49568605e-01 -6.17047489e-01 7.97695816e-01 8.15971047e-02
8.07261407e-01 -2.88007915e-01 -4.72756207e-01 -9.08545792e-01
2.15233281e-01 -4.76598233e-01 3.29732329e-01 -7.78387070e-01
-1.46076965e+00 5.47661841e-01 -9.74363834e-02 -1.19465411e+00
1.71669796e-01 -8.21122468e-01 -8.93877447e-01 6.75412059e-01
-1.40844452e+00 -1.34081674e+00 -4.38082963e-01 6.41347349e-01
5.87949991e-01 2.72064246e-02 4.48698878e-01 2.13065520e-01
-3.57011050e-01 5.98437130e-01 -1.19084470e-01 4.38159317e-01
3.01726818e-01 -1.43632841e+00 9.30302501e-01 5.40761709e-01
2.68061519e-01 9.86402392e-01 1.13822602e-01 -5.73070884e-01
-1.41652656e+00 -1.40062702e+00 4.69871372e-01 -7.01579154e-01
5.84930778e-01 -5.87738574e-01 -9.14446294e-01 9.24301744e-01
1.47914782e-01 4.56737846e-01 3.92877668e-01 -5.53402584e-03
-2.89161861e-01 -5.22069559e-02 -9.73794460e-01 5.39826810e-01
1.62493968e+00 -3.79706085e-01 -8.18615675e-01 1.67050779e-01
8.63422811e-01 -6.16332412e-01 -8.61932278e-01 6.67007208e-01
4.77903932e-01 -9.51584995e-01 1.24518085e+00 -3.40659261e-01
4.83444780e-01 -3.99531364e-01 -1.48569673e-01 -1.40191185e+00
-8.18005919e-01 -2.60320842e-01 7.48044029e-02 9.62526858e-01
4.52061385e-01 -8.35334778e-01 9.58051085e-01 1.28108293e-01
-5.17667949e-01 -1.05375242e+00 -9.58130836e-01 -7.98496604e-01
7.52754211e-02 -1.64567798e-01 1.13698709e+00 6.61979020e-01
-2.96957582e-01 3.09515446e-01 2.72057772e-01 3.62645209e-01
3.76923829e-01 6.41169846e-01 6.52999043e-01 -1.24458802e+00
-9.28205997e-02 -7.76572704e-01 -8.12964499e-01 -1.39883852e+00
2.46367395e-01 -1.30267978e+00 1.50495768e-01 -1.76777029e+00
1.09471768e-01 -4.25306886e-01 -6.32071733e-01 8.55231345e-01
-4.80637737e-02 5.80579996e-01 3.35203916e-01 3.15820485e-01
-7.96954453e-01 8.54957461e-01 1.31033158e+00 -3.48701149e-01
-2.90840000e-01 -7.20264344e-03 -8.03531706e-01 6.91661775e-01
1.04676974e+00 -2.41759717e-01 -2.74569750e-01 -7.89021313e-01
-1.73936725e-01 -4.06540364e-01 7.36143768e-01 -1.28670394e+00
2.53271669e-01 4.69439849e-02 5.19294620e-01 -9.22303975e-01
3.60749125e-01 -9.00291085e-01 -2.15291321e-01 2.11201131e-01
-3.27147961e-01 3.40256125e-01 1.83851004e-01 3.77714515e-01
-2.45166197e-01 1.71148524e-01 5.09514749e-01 -1.31495848e-01
-8.13565612e-01 8.46207738e-01 2.95808554e-01 -1.70728073e-01
9.81231511e-01 -1.35930672e-01 -3.89907956e-01 2.79797670e-02
-5.50476074e-01 3.38251442e-01 7.26723433e-01 6.45136297e-01
8.86574864e-01 -1.63995671e+00 -4.62647587e-01 5.25146365e-01
2.03439876e-01 4.16208327e-01 -8.95029213e-03 7.09827304e-01
-8.10775936e-01 5.30874252e-01 -4.59967881e-01 -8.49151731e-01
-9.69909668e-01 5.61835587e-01 2.32846051e-01 6.97975010e-02
-9.48969305e-01 8.38821709e-01 3.46368998e-01 -7.39640117e-01
-1.97353847e-02 -7.88449347e-01 -9.90304649e-02 -4.65505332e-01
1.71147630e-01 1.43751159e-01 1.44814596e-01 -6.27788603e-01
-6.28202081e-01 8.17493677e-01 -8.10564123e-03 3.88206504e-02
1.51930356e+00 2.01041669e-01 -4.55049455e-01 1.72334149e-01
1.42335689e+00 -3.82979750e-03 -1.26343834e+00 -1.80449143e-01
-7.07575306e-02 -5.28088748e-01 -5.37078716e-02 -5.59765756e-01
-1.20297492e+00 7.56449163e-01 4.10088062e-01 2.72393137e-01
1.14676845e+00 4.11686391e-01 9.65459108e-01 2.67704576e-01
3.76487732e-01 -6.26032889e-01 -4.25978638e-02 6.83466494e-01
1.21652901e+00 -9.47202921e-01 -1.61312640e-01 -4.91622925e-01
-2.81368703e-01 8.90093267e-01 7.62834311e-01 -7.59814322e-01
8.30075145e-01 1.46366596e-01 -1.22651778e-01 -5.28185129e-01
-5.32633245e-01 -4.42788869e-01 6.18211567e-01 4.71840739e-01
1.96064100e-01 1.04175173e-01 -2.10176006e-01 6.94603145e-01
-4.79710966e-01 3.29430960e-02 -1.19055495e-01 1.04776907e+00
-4.09804434e-01 -1.15270114e+00 -1.63494021e-01 6.62270546e-01
-1.69407874e-01 -3.01329698e-02 -4.70107079e-01 6.77818418e-01
1.46224737e-01 2.31508926e-01 5.87034225e-01 -3.40703219e-01
5.37534595e-01 -2.33537316e-01 1.68490559e-01 -8.12018871e-01
-4.69974577e-01 -1.93814516e-01 -3.71396035e-01 -9.78536844e-01
-2.27480590e-01 -6.15489602e-01 -1.43921769e+00 9.93969217e-02
1.07667431e-01 1.63522977e-02 3.31562877e-01 5.27081490e-01
8.24291945e-01 6.21622503e-01 5.06468952e-01 -1.27041984e+00
-3.11603725e-01 -6.88884199e-01 -3.31113428e-01 5.92401981e-01
4.11792219e-01 -7.26860225e-01 -1.88491086e-03 -3.23728770e-01]
|
[7.928971290588379, -3.580961227416992]
|
49c1a602-b444-460e-9b85-781cf688c4c6
|
scaling-distributed-training-of-flood-filling
|
1905.06236
| null |
https://arxiv.org/abs/1905.06236v4
|
https://arxiv.org/pdf/1905.06236v4.pdf
|
Scaling Distributed Training of Flood-Filling Networks on HPC Infrastructure for Brain Mapping
|
Mapping all the neurons in the brain requires automatic reconstruction of entire cells from volume electron microscopy data. The flood-filling network (FFN) architecture has demonstrated leading performance for segmenting structures from this data. However, the training of the network is computationally expensive. In order to reduce the training time, we implemented synchronous and data-parallel distributed training using the Horovod library, which is different from the asynchronous training scheme used in the published FFN code. We demonstrated that our distributed training scaled well up to 2048 Intel Knights Landing (KNL) nodes on the Theta supercomputer. Our trained models achieved similar level of inference performance, but took less training time compared to previous methods. Our study on the effects of different batch sizes on FFN training suggests ways to further improve training efficiency. Our findings on optimal learning rate and batch sizes agree with previous works.
|
['Peter Littlewood', 'Narayanan Kasthuri', 'Samuel Flender', 'Murat Keceli', 'Wushi Dong', 'Tom Uram', 'Rafael Vescovi', 'Hanyu Li', 'Elise Jennings', 'Corey Adams', 'Venkatram Vishwanath', 'Nicola Ferrier']
|
2019-05-13
| null | null | null | null |
['2048']
|
['playing-games']
|
[-1.49588943e-01 6.00894392e-02 4.22446221e-01 -6.01378798e-01
-5.81874788e-01 -3.53684992e-01 2.64465362e-01 8.17121863e-02
-9.73735988e-01 1.13341784e+00 -2.95769721e-01 -5.28629899e-01
4.19756994e-02 -7.84824014e-01 -8.95283937e-01 -7.24464655e-01
-1.65770262e-01 1.06205583e+00 6.17045343e-01 3.45433205e-01
5.28207958e-01 8.23768079e-01 -1.36195517e+00 3.39151293e-01
3.90100509e-01 7.34380960e-01 6.90744400e-01 8.00691247e-01
-6.62177801e-02 7.31966197e-01 -4.60456640e-01 -6.39899448e-03
2.44744450e-01 -1.80419043e-01 -1.16536868e+00 -5.08970559e-01
3.98402423e-01 -4.57680732e-01 -3.55417165e-03 5.52681029e-01
7.57308543e-01 5.97903058e-02 3.52456033e-01 -8.04521978e-01
8.56280327e-02 5.01142442e-01 -2.85269231e-01 7.95536041e-01
-4.14049923e-01 -2.42489576e-01 2.72280395e-01 -9.19522583e-01
9.55139160e-01 7.16137111e-01 9.51930761e-01 5.49713671e-01
-1.49825191e+00 -7.84444690e-01 -4.79353487e-01 -4.80002165e-02
-1.44809186e+00 -5.77661514e-01 6.69204369e-02 -4.31838632e-01
1.44466150e+00 -6.78924844e-02 1.01241577e+00 4.76214230e-01
6.04514956e-01 3.94094065e-02 1.29381096e+00 -3.58987242e-01
4.98028338e-01 -3.93991619e-02 -1.83325186e-02 7.29129076e-01
3.28395426e-01 -2.21844986e-01 -5.80858827e-01 -4.84267682e-01
1.41323125e+00 -1.30301416e-01 1.56656489e-01 4.53797467e-02
-1.10720968e+00 6.65021241e-01 2.28234649e-01 3.86939794e-01
-1.06021494e-01 3.71857733e-01 7.93378174e-01 4.25926894e-02
6.56695187e-01 5.69137990e-01 -8.55570495e-01 -7.58700147e-02
-1.25996995e+00 1.32113695e-01 7.70454347e-01 6.02808058e-01
9.20024812e-01 -1.12038255e-01 2.76225388e-01 7.35542774e-01
1.83099687e-01 1.41627848e-01 2.98069745e-01 -1.52187181e+00
1.38445348e-01 3.97031784e-01 -2.16806576e-01 -5.58929801e-01
-7.57066011e-01 -2.45559886e-01 -7.25391686e-01 4.39401954e-01
7.13726342e-01 -3.58972251e-01 -7.24510610e-01 1.18888938e+00
4.61932302e-01 1.85529262e-01 -3.25743616e-01 7.05748200e-01
5.83085299e-01 5.42065322e-01 1.03037119e-01 -1.87062621e-01
1.03856194e+00 -5.79252183e-01 -2.81582892e-01 -8.37091058e-02
1.10293126e+00 -6.84370995e-01 5.63852906e-01 3.13666701e-01
-1.23812866e+00 -2.87966788e-01 -7.06640899e-01 -2.32697293e-01
-3.16715389e-01 -2.08480211e-04 1.15949750e+00 5.25453448e-01
-1.57021523e+00 9.08467412e-01 -1.32247531e+00 -3.27261716e-01
8.10086012e-01 7.88976133e-01 -4.48551238e-01 2.37779155e-01
-5.11538327e-01 9.71110523e-01 3.68455887e-01 1.02612272e-01
-9.29026365e-01 -1.00306189e+00 -4.10888761e-01 -1.50728419e-01
-1.66252851e-01 -6.27729595e-01 1.13510168e+00 -5.01136065e-01
-1.37153447e+00 1.03146219e+00 -2.29656696e-01 -7.11475432e-01
1.76975772e-01 3.81599694e-01 3.26903164e-01 2.67454028e-01
-9.26000327e-02 1.13152182e+00 1.08398937e-01 -8.35094154e-01
-1.64182633e-01 -5.00677526e-01 -3.35088313e-01 5.18495478e-02
-1.25905827e-01 5.14291115e-02 -2.81020124e-02 -8.68402869e-02
2.07125783e-01 -7.08730519e-01 -3.41594756e-01 -7.95225874e-02
1.60083562e-01 -6.18033074e-02 5.64164877e-01 -6.98933899e-01
4.96383309e-01 -1.79592311e+00 -2.04443395e-01 3.05170685e-01
4.37450677e-01 -7.64855137e-03 2.65200675e-01 3.53379399e-01
7.03842491e-02 -8.10490474e-02 -1.74939260e-01 -4.85904068e-01
-5.04268944e-01 5.17521381e-01 -1.66749746e-01 7.48536527e-01
-4.00525033e-01 6.18337035e-01 -5.30889750e-01 -8.74038279e-01
-1.70220196e-01 5.33757806e-01 -8.08937013e-01 -8.94086286e-02
7.55644590e-02 7.38776445e-01 -2.64051612e-02 3.89328599e-01
5.74414730e-01 -5.57865143e-01 4.54293162e-01 3.85241434e-02
-3.52037966e-01 3.97359163e-01 -7.59321749e-01 1.72127795e+00
-5.03102541e-01 7.88380623e-01 3.65041345e-01 -1.19110692e+00
8.92774165e-01 4.00635719e-01 6.40213370e-01 -6.42236292e-01
1.00003779e-01 4.02208149e-01 6.79024681e-02 -3.09889298e-02
1.26120210e-01 -2.13637576e-01 5.55571735e-01 6.04092062e-01
4.33424085e-01 -4.26168591e-02 2.89170653e-01 2.12348059e-01
1.21385610e+00 1.99255362e-01 -7.26852491e-02 -8.80626678e-01
-9.70218256e-02 2.59176850e-01 5.27309418e-01 6.79050505e-01
-1.39756985e-02 5.09906948e-01 6.16379023e-01 -8.48305583e-01
-1.53701484e+00 -6.07360065e-01 -7.54811585e-01 1.26051891e+00
-2.02985018e-01 -5.20476282e-01 -1.04469466e+00 -1.10269859e-01
-3.28515917e-01 2.03365609e-01 -4.62040305e-01 5.18710613e-01
-7.66749561e-01 -1.00371742e+00 6.92200840e-01 6.24446571e-01
3.42835665e-01 -1.34480774e+00 -1.04693961e+00 3.17992568e-01
2.14678422e-01 -1.00722075e+00 1.17516123e-01 7.52208471e-01
-1.51900518e+00 -8.94311666e-01 -5.44884443e-01 -9.19548452e-01
1.12356639e+00 -1.94765598e-01 1.09569740e+00 4.02215034e-01
-7.02640533e-01 6.06099404e-02 1.87065229e-01 -3.74018192e-01
-5.66778667e-02 3.03045779e-01 -6.80196751e-03 -8.99865448e-01
1.11904144e-01 -8.66361558e-01 -6.34106398e-01 2.24282175e-01
-6.94323957e-01 5.17043829e-01 3.70428354e-01 8.54934096e-01
9.57011759e-01 -1.20452367e-01 5.03674448e-01 -1.14045024e+00
1.94122717e-01 -2.58039713e-01 -9.21135604e-01 -1.21591017e-01
-8.69551897e-01 1.13650769e-01 7.32975066e-01 -2.40463346e-01
-8.53608370e-01 2.00320750e-01 -3.93948436e-01 -7.61302710e-02
-1.99304849e-01 1.88258633e-01 6.51960671e-01 -5.85930586e-01
5.96998572e-01 -1.66098252e-02 2.80907810e-01 -3.00535679e-01
-4.24740463e-01 1.62685633e-01 9.18327272e-02 -8.12837601e-01
4.26297598e-02 6.59967959e-01 2.96562344e-01 -7.17823207e-01
-1.80874899e-01 -1.95014268e-01 -7.46269643e-01 -1.66247442e-01
7.72448719e-01 -6.41537488e-01 -9.74615037e-01 3.64753067e-01
-1.27037311e+00 -9.74126756e-01 -1.79026127e-01 4.91602331e-01
-6.07580841e-01 -4.35022563e-02 -1.02686810e+00 -4.84031171e-01
-5.65991104e-01 -1.04829037e+00 1.03815448e+00 1.90615892e-01
-1.87770441e-01 -1.20392394e+00 5.12868047e-01 1.95804492e-01
6.51302040e-01 -1.27743647e-01 7.26516366e-01 -5.02084792e-01
-4.52079087e-01 1.83153361e-01 -2.72320300e-01 -1.16215035e-01
-5.31514049e-01 -1.04484580e-01 -8.56740117e-01 -4.06255782e-01
5.92466891e-02 -5.56981981e-01 6.66101635e-01 6.56272054e-01
1.46083748e+00 -1.14250250e-01 -6.27685547e-01 8.75852942e-01
1.58169198e+00 -2.37037074e-02 8.25908840e-01 4.62153196e-01
4.76113260e-01 4.79274899e-01 -1.11880049e-01 3.72027218e-01
9.25669074e-02 2.18909740e-01 1.05335265e-01 -2.62370825e-01
-1.87625189e-03 2.06057340e-01 -4.43726359e-03 7.84633338e-01
-4.27316785e-01 4.29442286e-01 -1.31993651e+00 3.20404559e-01
-1.54040051e+00 -8.66819978e-01 -1.94699094e-01 1.93588579e+00
9.93922412e-01 7.70213753e-02 8.70506316e-02 -2.04610810e-01
5.01403809e-01 -2.78307885e-01 -4.59285110e-01 -4.21703398e-01
1.91306114e-01 5.72807610e-01 9.26599860e-01 4.85306770e-01
-5.07267773e-01 7.52627373e-01 8.05546093e+00 6.64640844e-01
-9.83027577e-01 5.99217176e-01 1.07970870e+00 -5.33270121e-01
6.32902980e-02 2.15293199e-01 -1.13253009e+00 3.36253107e-01
1.47886944e+00 1.40019581e-01 7.43888199e-01 8.25547218e-01
-3.67742330e-02 -5.46322763e-01 -8.65894735e-01 8.46894145e-01
-2.87463337e-01 -1.88710570e+00 -6.00409925e-01 4.78485197e-01
6.06649756e-01 7.13380694e-01 -4.79276806e-01 1.61709651e-01
2.20442176e-01 -1.06205416e+00 3.02394867e-01 4.50229108e-01
7.87506104e-01 -8.00292909e-01 7.78277218e-01 4.88930166e-01
-9.99513984e-01 2.46183857e-01 -8.35155785e-01 -2.87158757e-01
-8.01049825e-03 7.57125676e-01 -1.07590365e+00 -1.58038870e-01
9.94804621e-01 7.62316063e-02 -7.07992971e-01 9.33296680e-01
4.47441876e-01 7.83022106e-01 -7.84939349e-01 8.50939006e-02
-8.97147059e-02 -2.00791791e-01 -1.44271761e-01 1.14554727e+00
1.79195225e-01 1.01077773e-01 -1.31064773e-01 9.16877508e-01
-1.07104480e-02 -4.61131707e-02 -4.86612916e-01 1.36420324e-01
4.96185660e-01 1.59425092e+00 -1.62940180e+00 -3.95230949e-01
-1.47468463e-01 5.34275889e-01 8.35386872e-01 4.78743427e-02
-5.19075155e-01 -2.48608038e-01 1.67177528e-01 4.92179394e-01
2.91653007e-01 -4.76986647e-01 -7.33561516e-01 -6.68997884e-01
-3.29477578e-01 -2.84378380e-01 3.43708880e-02 -7.16063619e-01
-9.55542505e-01 5.21938920e-01 9.39672142e-02 -2.86238194e-01
-8.76037329e-02 -7.23170578e-01 -5.95948756e-01 9.09883201e-01
-7.95526683e-01 -7.91735172e-01 -8.16618204e-02 3.37983221e-01
1.31438911e-01 1.80118419e-02 9.88913178e-01 2.63864487e-01
-4.22082067e-01 1.80023134e-01 2.00269535e-01 1.07988425e-01
4.58477348e-01 -1.02162564e+00 3.65506113e-01 3.59790266e-01
-1.76240001e-02 9.05765414e-01 5.35279095e-01 -8.71015370e-01
-1.23965287e+00 -7.53496587e-01 9.53942776e-01 -2.75682271e-01
4.81638074e-01 -5.59183598e-01 -1.04894125e+00 8.74046862e-01
2.64434934e-01 3.90604168e-01 7.94886887e-01 2.86995709e-01
1.85683295e-01 -7.21202865e-02 -1.18624103e+00 1.61014408e-01
1.00553572e+00 -4.83308077e-01 -9.25828218e-02 5.51401973e-01
1.63713634e-01 -5.68486333e-01 -1.17006874e+00 4.28185910e-02
5.36318660e-01 -9.79374349e-01 8.12764645e-01 -1.18230574e-01
2.11300954e-01 -7.96654634e-03 1.02125995e-01 -8.79918098e-01
-3.27667147e-01 -1.86679199e-01 1.09951280e-01 8.78375530e-01
2.84954995e-01 -8.45331073e-01 1.07306445e+00 6.79651141e-01
-3.42313498e-01 -1.05145323e+00 -1.23430479e+00 -4.57467645e-01
2.62891084e-01 -1.89476758e-02 2.91727394e-01 5.89878142e-01
2.17976677e-03 1.81951925e-01 2.87364751e-01 -1.78786427e-01
7.34453261e-01 1.20615531e-02 3.46449643e-01 -1.30063450e+00
-2.71676391e-01 -1.54409900e-01 -3.13641220e-01 -5.36557794e-01
2.60663182e-01 -1.21662700e+00 -3.38357575e-02 -1.62671483e+00
4.28577542e-01 -7.44970798e-01 1.52691796e-01 8.53483975e-01
3.83180916e-01 5.70756912e-01 -2.48702243e-01 4.21204001e-01
-5.56220889e-01 -1.93492472e-02 1.12257755e+00 4.61966336e-01
7.77976140e-02 -6.32267177e-01 -3.08826566e-01 7.56474197e-01
9.81890857e-01 -8.61974001e-01 -2.86821485e-01 -7.40998089e-01
5.39473295e-01 6.44184723e-02 4.49032426e-01 -1.12794471e+00
7.09986031e-01 7.74414390e-02 1.00814962e+00 -6.95014656e-01
2.19133869e-01 -5.93602777e-01 3.71664971e-01 6.26672089e-01
-8.64983499e-02 2.72635728e-01 3.24511290e-01 -4.17962410e-02
2.21413374e-01 -4.18041438e-01 9.06021118e-01 -4.79340285e-01
-2.68410385e-01 1.92437097e-01 -7.50628948e-01 -1.92254454e-01
8.96290421e-01 -2.44360864e-01 -4.48330641e-01 2.54388571e-01
-6.65513098e-01 -7.16091320e-02 7.02889502e-01 -7.28535473e-01
4.44739908e-01 -9.69388962e-01 -3.11056674e-01 3.90404940e-01
-6.88234568e-01 3.49461049e-01 2.01386124e-01 1.19124126e+00
-1.25576138e+00 5.22864580e-01 -5.78016698e-01 -6.09545588e-01
-9.49892879e-01 1.49309710e-01 4.26120818e-01 -2.26423204e-01
-8.45381916e-01 9.34854805e-01 -1.74292490e-01 -5.82034528e-01
-4.54117730e-02 -9.13638175e-02 -2.14341711e-02 -2.34103724e-01
4.20293093e-01 6.74320459e-01 4.63368654e-01 -8.93908087e-03
-5.27909279e-01 2.27724865e-01 -1.93290472e-01 -2.20772550e-01
1.68159401e+00 1.66158915e-01 -7.26650715e-01 4.93567258e-01
1.08204484e+00 -3.58603001e-01 -1.17930555e+00 2.93077320e-01
-1.46025941e-01 1.25610679e-01 1.94166586e-01 -3.23631197e-01
-9.41632986e-01 6.97287679e-01 5.68223119e-01 -2.81578917e-02
7.24490941e-01 1.16415121e-01 7.88884938e-01 7.35563993e-01
8.56018901e-01 -1.28970873e+00 -3.05111051e-01 8.21179152e-01
4.26573724e-01 -9.80463862e-01 3.89403611e-01 2.45896168e-03
-1.09909751e-01 1.21634579e+00 9.08102453e-01 -5.42718768e-01
6.11852467e-01 9.46370780e-01 -2.05200970e-01 -5.86587667e-01
-9.41491604e-01 5.13992727e-01 -4.04080272e-01 3.56560022e-01
5.79733610e-01 -3.05087835e-01 -4.18229669e-01 1.34255797e-01
-3.55796933e-01 3.16640496e-01 3.89849812e-01 1.19889116e+00
-5.81753552e-01 -1.04100597e+00 -3.53574395e-01 8.89960170e-01
-5.77861965e-01 -1.82666451e-01 1.02408165e-02 5.47824144e-01
2.68365443e-01 1.94073841e-01 6.25305235e-01 -2.86899181e-03
-3.90147835e-01 2.59773135e-01 1.03018689e+00 -7.63719618e-01
-8.16704810e-01 3.66767421e-02 -1.04490094e-01 -6.33328736e-01
-3.02415103e-01 -5.59668958e-01 -1.79750109e+00 -6.56590164e-01
-3.03054869e-01 2.32780114e-01 9.39231217e-01 1.12691557e+00
5.15740395e-01 5.62907934e-01 1.50514878e-02 -1.42202437e+00
2.07765028e-03 -9.31639314e-01 -8.02004695e-01 -2.53794014e-01
-2.62150973e-01 -4.85575706e-01 -3.09573472e-01 1.82514623e-01]
|
[14.255128860473633, -3.1185030937194824]
|
45f807e2-066f-4c30-a6ca-273fab7a8340
|
refin-a-refinement-approach-for-video-frame
| null | null |
https://openreview.net/forum?id=4_cgHrh0BpN
|
https://openreview.net/pdf?id=4_cgHrh0BpN
|
ReFIn: A Refinement Approach for Video Frame Interpolation
|
Video Frame Interpolation is an important video enhancement problem which aims to generate one or multiple frames between consecutive frames in video. Optical flow-based frame interpolation approaches estimate intermediate optical flow from interpolated frame to input frames and warped frames are fused to generate interpolated frame. However, intermediate flow estimates can itself be erroneous leading to inaccurate interpolation results. In this work, we improve an flow-based intrtpolation algorithm, Super-SloMo by residual refinement. Specifically, we feed intermediate flowmaps, visibility map, warped input frames and intermediate interpolation estimate to a refinement network to predict a frame residual. We have also experimented with different architecture choices to be used in different modules to further improve the results. We found out that GridNet with four pyramid levels achieves the best results whereas UNet++ performs moderately well with significantly less number of parameters.
|
['Anurag Mittal', 'Saikat Dutta']
|
2021-10-19
| null | null | null |
neurips-workshop-deep-invers-2021-12
|
['video-enhancement']
|
['computer-vision']
|
[ 1.51959524e-01 -1.80806786e-01 -3.91387828e-02 -2.62542754e-01
-3.16087902e-01 -2.08367795e-01 3.79086435e-01 -3.64097685e-01
-3.40606481e-01 1.13374615e+00 2.57071793e-01 -1.27572939e-01
5.41128933e-01 -5.79434693e-01 -6.71702445e-01 -2.97804505e-01
-1.77993804e-01 -2.02323854e-01 7.08852232e-01 -1.61768109e-01
4.91446376e-01 5.29761553e-01 -1.42354751e+00 5.94633043e-01
7.16662765e-01 8.77950728e-01 1.73140585e-01 1.11907184e+00
-6.89003170e-02 1.37942386e+00 -5.03359973e-01 -6.83496743e-02
5.49579024e-01 -6.56584263e-01 -8.80100310e-01 1.34347826e-01
7.46380806e-01 -8.08814466e-01 -3.04882407e-01 7.00035155e-01
1.55432016e-01 3.33767176e-01 2.28962988e-01 -1.25626862e+00
-4.48752016e-01 3.20694685e-01 -7.13033736e-01 5.94031811e-01
8.15797687e-01 4.97987181e-01 3.26911181e-01 -1.23692095e+00
9.60886061e-01 1.41080654e+00 7.87344992e-01 5.33745706e-01
-1.38150322e+00 -6.05646431e-01 -1.42538115e-01 3.23671818e-01
-1.29655409e+00 -5.41068435e-01 5.55198431e-01 -2.67020196e-01
9.32525277e-01 3.17324430e-01 8.60237300e-01 6.04994416e-01
4.39176947e-01 3.54300797e-01 8.91983211e-01 -2.58407563e-01
5.42158261e-02 -1.58698320e-01 -2.73070127e-01 6.87649310e-01
-1.44962162e-01 2.58296579e-01 -6.17918372e-01 1.84751838e-01
1.42579114e+00 -2.68251568e-01 -6.02375090e-01 1.70845762e-01
-1.49428511e+00 5.16164005e-01 6.87508583e-01 1.69067442e-01
-6.24721766e-01 4.49028850e-01 1.49461925e-01 3.65644634e-01
3.30990851e-01 1.58447936e-01 -1.87183380e-01 3.33041027e-02
-1.46679151e+00 4.92795646e-01 5.24427354e-01 7.88455904e-01
1.29156089e+00 4.28609639e-01 -4.45521683e-01 4.22646642e-01
3.20302248e-01 -6.71287701e-02 1.72214195e-01 -1.90909529e+00
4.53981698e-01 1.68713525e-01 3.38983238e-01 -1.16351438e+00
-2.21509799e-01 7.50417262e-02 -7.86048532e-01 8.21543396e-01
6.28053844e-01 -3.66363913e-01 -9.59335864e-01 1.20972085e+00
2.85448045e-01 9.34730709e-01 -1.52025163e-01 1.32879603e+00
5.96597612e-01 9.71098423e-01 2.46699482e-01 -3.40771914e-01
1.06329668e+00 -1.27015853e+00 -8.01399946e-01 -8.06217566e-02
3.36454183e-01 -1.16424227e+00 7.12023616e-01 3.56285989e-01
-1.62843931e+00 -1.07378459e+00 -1.10313010e+00 -3.33727181e-01
2.59341806e-01 -2.40330979e-01 4.07849461e-01 4.14121062e-01
-1.55978215e+00 8.16290200e-01 -8.03096414e-01 5.40827997e-02
3.81461829e-01 4.72168237e-01 -4.79828089e-01 7.43232220e-02
-9.95995700e-01 7.62369931e-01 6.38346553e-01 6.29327595e-02
-6.99632347e-01 -1.04983687e+00 -8.64128530e-01 -1.74603447e-01
-1.35934919e-01 -1.08130038e+00 9.22694027e-01 -1.29809797e+00
-1.46614003e+00 2.33954877e-01 -5.53746939e-01 -6.92752600e-01
6.90922379e-01 -6.42844150e-03 -2.20944077e-01 1.19376294e-01
-1.05927937e-01 1.45911896e+00 9.82767165e-01 -1.06722844e+00
-1.11405563e+00 1.74549863e-01 2.38968119e-01 2.36116499e-01
2.33881712e-01 7.19076470e-02 -2.89962411e-01 -7.77115464e-01
9.60615650e-03 -6.87859178e-01 -4.82975096e-01 2.86466897e-01
5.79845756e-02 1.81468025e-01 1.15075386e+00 -8.49870563e-01
1.26153409e+00 -1.83441556e+00 6.91002607e-02 -1.54209703e-01
1.93564683e-01 3.05957407e-01 -1.10263295e-01 -1.66257933e-01
-2.10607931e-01 6.81213588e-02 -7.33171925e-02 -3.26327711e-01
-6.91739261e-01 1.80913568e-01 -9.22888815e-02 3.11155915e-01
3.66003931e-01 6.43453538e-01 -8.90125692e-01 -7.68686414e-01
6.73390746e-01 1.05818582e+00 -9.74766195e-01 2.50483125e-01
-1.51047751e-01 1.02598643e+00 7.57080764e-02 2.73948491e-01
9.54454184e-01 6.65828632e-03 -2.76670367e-01 -5.14409482e-01
-4.92333889e-01 -5.23635857e-02 -1.37181258e+00 1.86243737e+00
-5.26788533e-01 1.23050034e+00 -4.16544564e-02 -3.29336196e-01
7.94759691e-01 4.80427623e-01 6.14552259e-01 -2.60892570e-01
1.39997527e-01 -2.60940054e-03 -1.07049279e-01 -4.65964437e-01
1.06145489e+00 1.92558914e-01 6.74613178e-01 2.11537462e-02
-1.28110396e-02 2.75458675e-02 4.52232718e-01 6.95971539e-03
9.12123919e-01 6.59222424e-01 1.14682131e-01 -2.01543391e-01
1.00507200e+00 5.01516536e-02 6.40457988e-01 4.28578943e-01
-2.95215756e-01 1.19352591e+00 3.10292721e-01 -9.86693263e-01
-1.41671264e+00 -8.65045786e-01 -1.63806695e-02 8.36005747e-01
2.76433200e-01 -4.28657711e-01 -8.46118867e-01 -3.15066576e-01
-4.63865101e-01 4.88038719e-01 -3.09628904e-01 6.38641417e-01
-1.16906476e+00 -3.17841142e-01 3.10060591e-01 4.53676879e-01
8.41969132e-01 -1.07576907e+00 -7.05571473e-01 5.00426650e-01
-2.78857738e-01 -1.22686124e+00 -7.86084652e-01 -4.24180061e-01
-9.55085576e-01 -9.10559833e-01 -8.26567352e-01 -8.55972886e-01
5.97624063e-01 2.68344700e-01 1.15822744e+00 3.98979366e-01
-2.65925284e-02 -3.23642612e-01 -7.28839785e-02 1.48418009e-01
-6.69499397e-01 -2.49445543e-01 -2.59651572e-01 5.38028553e-02
-1.84363313e-02 -5.60332835e-01 -1.10338056e+00 3.56021821e-01
-9.64582205e-01 5.10840952e-01 -8.01818147e-02 6.79277837e-01
4.74913120e-01 -4.00310695e-01 1.36188745e-01 -5.66972375e-01
4.48208570e-01 -1.83350742e-01 -6.77302003e-01 -8.62640291e-02
-1.44701824e-01 1.02794901e-01 6.71799302e-01 -2.08517164e-01
-1.21985817e+00 2.85478741e-01 -3.97352606e-01 -6.87507570e-01
-1.35060787e-01 -6.17361180e-02 4.18275177e-01 -3.77802640e-01
8.60628724e-01 -1.54100955e-01 -5.86218722e-02 3.72575074e-02
2.52175212e-01 3.29265594e-01 7.88463950e-01 -1.28743887e-01
5.43487966e-01 4.93117362e-01 1.75503381e-02 -6.53904021e-01
-3.15731972e-01 -4.14659493e-02 -6.53718948e-01 -5.30489922e-01
1.05766976e+00 -1.14833486e+00 -6.29805207e-01 1.83253884e-01
-1.58513308e+00 -4.28731263e-01 -1.43912509e-01 5.77367663e-01
-5.48243880e-01 5.09238005e-01 -8.53516877e-01 -4.18013602e-01
-2.70579159e-01 -1.54595876e+00 7.26523578e-01 5.24793863e-01
-2.81388223e-01 -1.12347019e+00 3.04024741e-02 1.22740500e-01
6.56601727e-01 4.41292971e-01 1.44543365e-01 4.03611213e-01
-1.00043881e+00 3.34894478e-01 -4.35835361e-01 2.34652609e-01
2.92043775e-01 3.84364784e-01 -9.65865016e-01 -1.47546887e-01
-9.03190374e-02 2.10491076e-01 7.54013956e-01 6.52355433e-01
1.13448882e+00 -1.89444631e-01 7.97464326e-02 1.01547742e+00
1.56410587e+00 1.45538822e-01 1.22029996e+00 3.52144599e-01
8.81329834e-01 2.33544677e-01 5.57956636e-01 3.95133048e-01
2.82598019e-01 6.97830915e-01 2.19012946e-01 -1.37904271e-01
-7.84213781e-01 -8.89681429e-02 5.37265718e-01 4.79634315e-01
-7.87137926e-01 -1.75216183e-01 -6.27569616e-01 3.03724587e-01
-1.76596916e+00 -1.21880424e+00 -4.24580723e-01 1.96440625e+00
7.41685152e-01 9.64315310e-02 8.77078325e-02 9.33579281e-02
8.97676170e-01 1.92059532e-01 -1.78633243e-01 -4.80336487e-01
4.81176712e-02 3.33186001e-01 6.42911315e-01 1.13237000e+00
-8.42989981e-01 1.06913841e+00 6.62513494e+00 5.30255377e-01
-1.27065670e+00 1.25728533e-01 8.73459756e-01 1.29519954e-01
-3.19929242e-01 3.37379366e-01 -6.76862299e-01 5.87155879e-01
9.30833340e-01 -3.43480520e-02 4.95229155e-01 2.96210319e-01
7.70467520e-01 -3.84323031e-01 -9.33378160e-01 1.07495201e+00
-2.18787700e-01 -1.95740592e+00 1.53224543e-01 -2.15758041e-01
1.01467049e+00 -7.70915747e-02 -2.09058359e-01 -1.72262996e-01
2.53334522e-01 -9.86731350e-01 6.50826812e-01 4.93579775e-01
7.82575727e-01 -7.23082840e-01 4.61015999e-01 1.76319554e-02
-1.39778829e+00 2.35696107e-01 -4.60926503e-01 -4.49247330e-01
6.65402770e-01 1.46478161e-01 -8.74482453e-01 3.77417684e-01
6.14852130e-01 8.12674761e-01 -5.37039578e-01 1.22526872e+00
-8.92101973e-02 2.23564684e-01 -1.03937291e-01 6.02445185e-01
2.24240407e-01 -2.05859825e-01 3.73581469e-01 1.29645252e+00
4.18285400e-01 -4.53846082e-02 1.78950831e-01 8.84082556e-01
1.37364283e-01 -1.48465693e-01 -4.56789762e-01 8.64876688e-01
2.35967919e-01 1.24154484e+00 -7.64528394e-01 -6.11443877e-01
-5.36218405e-01 1.13970935e+00 -3.65849473e-02 6.04567766e-01
-1.12796760e+00 -1.67592183e-01 9.42190230e-01 3.29435855e-01
5.99710569e-02 -3.08689952e-01 -2.91726589e-01 -1.27025926e+00
-3.35125357e-01 -4.43312883e-01 1.11946993e-01 -1.16576576e+00
-5.18571019e-01 9.95637655e-01 -1.70086980e-01 -1.46297264e+00
-6.74105704e-01 -2.11634576e-01 -5.94578505e-01 1.21904862e+00
-1.61273110e+00 -7.04410613e-01 -7.37827897e-01 8.50608051e-01
1.06566954e+00 7.77775869e-02 3.42473865e-01 4.17850971e-01
-3.18502635e-01 2.27762237e-01 -5.19489527e-01 1.12633020e-01
6.92700446e-01 -9.73971546e-01 7.67275631e-01 1.23142254e+00
-1.22340858e-01 1.34948865e-01 9.70392108e-01 -6.47682726e-01
-1.06499481e+00 -1.27641213e+00 7.13977814e-01 -2.42233992e-01
3.91097277e-01 2.57735312e-01 -9.71133232e-01 7.23113418e-01
9.27128613e-01 6.29941106e-01 -2.95653287e-02 -1.01236832e+00
1.78739116e-01 3.33185047e-02 -1.31248915e+00 7.38199174e-01
7.85694063e-01 -1.94415554e-01 -1.26895934e-01 -6.04225583e-02
7.34561861e-01 -8.04877639e-01 -1.03671205e+00 1.12815224e-01
5.36005914e-01 -1.51888704e+00 1.14455140e+00 -1.94035083e-01
7.07212508e-01 -8.23850513e-01 1.22669287e-01 -1.16017950e+00
-5.19784629e-01 -8.52565169e-01 -4.06062268e-02 8.96700680e-01
2.07546145e-01 -3.63700837e-01 1.08348322e+00 7.99887478e-01
-4.78449799e-02 -2.54678190e-01 -6.47689462e-01 -3.17395180e-01
-2.22512797e-01 -3.16018134e-01 5.09135246e-01 8.34932089e-01
-2.17975214e-01 9.97725204e-02 -6.24837577e-01 4.80096079e-02
6.39726758e-01 -3.10517997e-01 8.05199564e-01 -6.65148437e-01
-4.51782951e-03 -1.56971782e-01 -5.80790699e-01 -1.17450178e+00
2.01704782e-02 -4.51999784e-01 1.49606811e-02 -1.33347356e+00
-4.41851765e-01 -4.03526843e-01 4.01213951e-02 1.89204708e-01
-3.55808824e-01 1.00240970e+00 5.16981721e-01 1.68701679e-01
-2.26994574e-01 -1.97293758e-02 1.56074667e+00 1.39921769e-01
-6.24216378e-01 -3.08989644e-01 -8.32355842e-02 7.80030549e-01
8.13768327e-01 -1.60396680e-01 -4.54275131e-01 -7.46521056e-01
-2.14301929e-01 7.49840021e-01 4.77906108e-01 -1.49749792e+00
3.29390496e-01 -3.23033839e-01 8.70035946e-01 -4.48161781e-01
3.99371207e-01 -7.18983293e-01 7.10709691e-01 5.18867016e-01
-3.50028008e-01 5.59312582e-01 8.91405791e-02 5.40749878e-02
-2.67427146e-01 -1.93367720e-01 9.81292188e-01 -1.97545826e-01
-1.16159570e+00 4.02554184e-01 -4.40277457e-01 -2.53458679e-01
9.73003149e-01 -6.85057342e-01 -6.33506030e-02 -3.14953506e-01
-8.00956905e-01 -9.31991339e-02 7.03901589e-01 3.91743600e-01
9.33566689e-01 -1.34551573e+00 -1.01009107e+00 4.77236867e-01
-5.98458529e-01 3.67095158e-03 2.02378631e-01 8.01427066e-01
-1.61217022e+00 2.08129242e-01 -6.92532122e-01 -7.86655009e-01
-1.20200193e+00 5.07315993e-01 4.17676002e-01 1.54570073e-01
-6.63381994e-01 6.28791869e-01 3.29718702e-02 3.36756557e-01
1.75501721e-03 -5.29384375e-01 -2.88381785e-01 -2.31255308e-01
9.52326596e-01 6.53337240e-01 -1.03365213e-01 -8.76848102e-01
-9.94322449e-02 6.29882395e-01 1.14678830e-01 -1.86670899e-01
8.84855926e-01 -4.43222314e-01 -4.14432175e-02 -1.20031714e-01
1.16405475e+00 -1.93945304e-01 -2.04563618e+00 2.29566529e-01
-2.22914055e-01 -9.93085027e-01 2.03592658e-01 -9.18863565e-02
-1.55189228e+00 4.66654658e-01 7.23970294e-01 7.44683221e-02
1.28950489e+00 -6.51834905e-01 8.63341331e-01 -2.91487128e-01
4.88152087e-01 -5.77224672e-01 -2.39871025e-01 3.25301886e-01
5.96009612e-01 -1.15456903e+00 6.90503418e-02 -5.82978547e-01
-4.22999173e-01 1.45383811e+00 7.01662183e-01 -5.39016545e-01
4.44030970e-01 5.43415964e-01 1.47153184e-01 5.22630274e-01
-9.46325958e-01 1.61102340e-02 7.52604678e-02 7.47852087e-01
9.11514521e-01 -5.33020675e-01 -4.16494846e-01 -7.17010856e-01
-3.76230553e-02 6.77945971e-01 9.54551876e-01 5.20338833e-01
-4.30946797e-01 -1.10624015e+00 -7.53513277e-01 -2.13324409e-02
-5.79439163e-01 -1.21202499e-01 4.02610064e-01 6.13302469e-01
2.58338273e-01 8.70513678e-01 3.18860471e-01 -3.06082517e-01
3.02543696e-02 -3.48159999e-01 6.29365563e-01 -7.52370432e-02
-8.80867243e-01 1.36291787e-01 -4.49221171e-02 -1.00856197e+00
-8.93952847e-01 -4.30976689e-01 -1.36874914e+00 -7.40107715e-01
1.44969448e-01 4.04895330e-03 3.39419216e-01 5.96971810e-01
2.70828456e-01 6.14626706e-01 4.47640389e-01 -1.44959891e+00
4.62410688e-01 -7.28262305e-01 -3.51711251e-02 3.66325766e-01
7.22135127e-01 -1.50595576e-01 -2.02762753e-01 8.28322411e-01]
|
[10.713141441345215, -1.4371494054794312]
|
ec676819-bdc6-4f5d-add9-6b194d6783e5
|
a-one-covariate-at-a-time-method-for
|
2204.12023
| null |
https://arxiv.org/abs/2204.12023v1
|
https://arxiv.org/pdf/2204.12023v1.pdf
|
A One-Covariate-at-a-Time Method for Nonparametric Additive Models
|
This paper proposes a one-covariate-at-a-time multiple testing (OCMT) approach to choose significant variables in high-dimensional nonparametric additive regression models. Similarly to Chudik, Kapetanios and Pesaran (2018), we consider the statistical significance of individual nonparametric additive components one at a time and take into account the multiple testing nature of the problem. One-stage and multiple-stage procedures are both considered. The former works well in terms of the true positive rate only if the marginal effects of all signals are strong enough; the latter helps to pick up hidden signals that have weak marginal effects. Simulations demonstrate the good finite sample performance of the proposed procedures. As an empirical application, we use the OCMT procedure on a dataset we extracted from the Longitudinal Survey on Rural Urban Migration in China. We find that our procedure works well in terms of the out-of-sample forecast root mean square errors, compared with competing methods.
|
['Qiankun Zhou', 'Yonghui Zhang', 'Thomas Tao Yang', 'Liangjun Su']
|
2022-04-26
| null | null | null | null |
['additive-models']
|
['methodology']
|
[ 1.35182485e-01 -1.66027144e-01 -4.19135690e-01 -4.75666434e-01
-9.65743005e-01 -1.39961615e-01 5.14435053e-01 3.92391011e-02
-4.55510825e-01 1.18650782e+00 1.41613930e-01 -4.10790682e-01
-4.54744577e-01 -6.72766149e-01 -8.62902522e-01 -7.73693025e-01
-3.90700191e-01 3.31292123e-01 -2.13796631e-01 3.70303363e-01
4.16705459e-02 -1.52695537e-01 -1.17289555e+00 -3.06457520e-01
1.41267180e+00 3.54812622e-01 -4.98835463e-03 5.33337951e-01
3.81390929e-01 3.78783554e-01 -3.48838538e-01 -6.01183772e-01
2.00738579e-01 -4.46215391e-01 7.03511983e-02 1.28814340e-01
3.85995179e-01 -2.43795007e-01 1.35911003e-01 1.16014659e+00
5.54418445e-01 8.77250135e-02 9.98952150e-01 -1.13239455e+00
-7.32149482e-01 6.93290293e-01 -1.05694497e+00 -6.33564666e-02
-1.35186717e-01 -6.26847446e-02 1.00548947e+00 -1.35643470e+00
2.41628423e-01 1.39771736e+00 5.44664383e-01 -3.58793885e-02
-1.70570266e+00 -9.06587183e-01 1.77400187e-01 -3.73160481e-01
-1.37455583e+00 -4.51778829e-01 3.68141025e-01 -7.93069303e-01
2.97136962e-01 1.57198817e-01 9.25721452e-02 9.97862995e-01
2.93097526e-01 6.29447877e-01 1.46036160e+00 -3.99179280e-01
3.79853874e-01 1.15711883e-01 4.77789909e-01 4.29840565e-01
6.57344341e-01 3.23610216e-01 -2.89783806e-01 -8.03656042e-01
9.63435352e-01 -1.16628468e-01 1.31682500e-01 -2.06870630e-01
-9.55788195e-01 1.14412045e+00 -3.31148624e-01 1.11749895e-01
-6.80312693e-01 7.05444589e-02 1.02996209e-03 4.44578350e-01
8.91289115e-01 -5.59313484e-02 -6.51768088e-01 6.86279088e-02
-1.10567939e+00 4.48652864e-01 3.90681356e-01 5.52624524e-01
5.41993439e-01 1.13520391e-01 -5.13087273e-01 8.66761684e-01
2.87404895e-01 7.29475558e-01 2.06220403e-01 -7.04400897e-01
5.24556637e-01 2.26373866e-01 3.02875906e-01 -6.49457037e-01
-2.32840419e-01 -4.89927530e-01 -1.13763368e+00 5.30046597e-02
6.00391984e-01 -7.30493248e-01 -7.22717822e-01 2.05228257e+00
3.65483493e-01 6.03324831e-01 -2.29849741e-01 7.08581746e-01
1.98660120e-01 4.66282040e-01 2.87703425e-01 -6.17819190e-01
1.27382851e+00 -5.39508760e-01 -7.12120235e-01 -2.67198712e-01
5.21438122e-01 -2.67923355e-01 8.19173336e-01 4.73652810e-01
-1.15353096e+00 -4.93265986e-01 -6.60571635e-01 3.46275240e-01
1.51009932e-01 3.39151204e-01 7.17956722e-01 7.67941415e-01
-9.96019840e-01 3.29568744e-01 -6.30826890e-01 1.37205228e-01
1.90659553e-01 4.88214284e-01 -1.38689786e-01 -5.49343154e-02
-1.13409865e+00 3.84024888e-01 -2.80960053e-01 2.34722793e-01
-4.91148084e-01 -8.56020451e-01 -6.78662717e-01 1.81772485e-01
4.53695238e-01 -5.69741666e-01 9.41446662e-01 -1.15602422e+00
-1.11773932e+00 4.55174744e-01 -5.50758183e-01 -2.44594261e-01
8.32140744e-01 -1.86618432e-01 -2.90522695e-01 -4.02922899e-01
5.84421396e-01 1.78576112e-01 5.88000834e-01 -9.09184813e-01
-6.29023492e-01 -4.34833199e-01 -5.39375365e-01 -1.27824694e-01
-9.61008221e-02 3.50425035e-01 -2.72368312e-01 -9.70936179e-01
-1.39480487e-01 -9.34794784e-01 -5.06974399e-01 -3.47143501e-01
-4.23204571e-01 -4.90471534e-02 1.03557788e-01 -9.15237427e-01
1.29107559e+00 -2.10930181e+00 2.17684567e-01 4.58329111e-01
1.89313572e-02 -4.02885348e-01 -1.36554286e-01 8.83802697e-02
-1.77638784e-01 1.17406368e-01 -4.93129671e-01 -5.20536900e-01
3.69172581e-02 -4.75160778e-02 -9.03837848e-03 7.11898863e-01
4.84771729e-01 8.01613629e-01 -5.13963640e-01 -2.96842158e-01
-1.68922469e-01 1.98414817e-01 -5.33384383e-01 -1.50148615e-01
3.48842710e-01 5.27456939e-01 -3.36762607e-01 6.32966518e-01
1.04149508e+00 -6.51172176e-02 3.95057172e-01 5.56253672e-01
-3.76921147e-01 7.44119585e-02 -1.48538601e+00 6.81175888e-01
-1.43306563e-02 4.30541754e-01 3.15186769e-01 -1.20944905e+00
7.03084230e-01 3.48762006e-01 1.48909017e-01 -7.48143196e-01
-2.31080294e-01 4.56227273e-01 1.62340000e-01 -2.21446514e-01
2.69765079e-01 -3.64383876e-01 -3.02560568e-01 -1.06921727e-02
-2.12366357e-01 4.63573009e-01 1.02259465e-01 -3.34152728e-01
9.25181985e-01 3.42283659e-02 7.27838099e-01 -6.32414103e-01
2.64798999e-01 -4.21160370e-01 1.07788634e+00 1.15299428e+00
1.80136293e-01 4.34034944e-01 1.01270604e+00 4.56710428e-01
-7.19362736e-01 -1.28866482e+00 -5.02508998e-01 1.24392831e+00
-3.77061427e-01 3.42763364e-01 -3.02103877e-01 -3.24223697e-01
6.63746417e-01 7.01030970e-01 -8.73889029e-01 2.41053164e-01
-3.29989344e-01 -1.11739743e+00 2.05485329e-01 4.38508987e-01
1.11814193e-01 -5.61317503e-01 -1.70200586e-01 2.12993249e-01
2.37897947e-01 -6.71945572e-01 -2.52300769e-01 2.07196876e-01
-9.85974729e-01 -8.41231048e-01 -1.04210162e+00 -4.01564538e-01
7.25067317e-01 2.66774833e-01 7.52050221e-01 -3.06553394e-01
3.16769242e-01 -7.13082179e-02 -1.03887163e-01 -2.74209768e-01
-1.42384008e-01 -4.49697793e-01 1.69763312e-01 1.93723455e-01
3.25420082e-01 -5.47559679e-01 -1.85471848e-01 1.99238241e-01
-5.31520486e-01 -1.58910900e-01 7.64682055e-01 9.63776827e-01
5.06714165e-01 1.08434647e-01 9.62479770e-01 -9.38964307e-01
6.09663308e-01 -9.56384301e-01 -9.58500206e-01 2.58157194e-01
-5.61880291e-01 -1.47709846e-01 2.69155115e-01 -9.67120409e-01
-9.05497789e-01 -3.69118720e-01 4.38629955e-01 -2.37123102e-01
3.63394842e-02 9.47909176e-01 -2.16749653e-01 1.75410032e-01
1.37868419e-01 -7.10791871e-02 -2.21836656e-01 -5.36202073e-01
-1.65146872e-01 4.70551729e-01 2.54724860e-01 -5.54282129e-01
7.45640397e-01 6.87514842e-02 2.73528188e-01 -9.19864237e-01
-3.23734969e-01 -3.60234588e-01 -6.71081543e-01 8.16006437e-02
6.40062869e-01 -1.27302420e+00 -5.57174146e-01 4.09424067e-01
-6.78398311e-01 -4.76758420e-01 9.46812928e-02 9.16544080e-01
-2.25284383e-01 1.58076406e-01 -3.83634031e-01 -1.47360599e+00
2.29640827e-01 -1.10675192e+00 9.23207521e-01 8.60999674e-02
-8.34997594e-02 -1.03951073e+00 1.79909408e-01 6.17639273e-02
-1.19063690e-01 4.29130673e-01 1.05338943e+00 -7.20403671e-01
-1.56417742e-01 -1.83099374e-01 -8.08287933e-02 -4.70167138e-02
8.54268670e-02 2.32059017e-01 -7.18731344e-01 -3.22893173e-01
-6.83846772e-02 1.82941213e-01 9.17247057e-01 1.32528484e+00
1.04164457e+00 -4.08831418e-01 -7.45754540e-02 1.58578128e-01
1.28678489e+00 1.10721469e-01 5.19321024e-01 -4.83861901e-02
4.83204216e-01 8.86393249e-01 7.09506333e-01 7.32264698e-01
5.15866756e-01 7.25435972e-01 -1.06758356e-01 -2.48349875e-01
6.82766557e-01 -2.02292889e-01 5.37396789e-01 4.04774487e-01
1.63475782e-01 -1.90387964e-01 -6.57544136e-01 6.02217674e-01
-2.12162113e+00 -1.05885983e+00 -7.33648896e-01 2.73093200e+00
8.00786614e-01 1.15551904e-01 6.63325787e-01 -7.24744275e-02
8.73573899e-01 -2.26133436e-01 -4.01334524e-01 -3.76399100e-01
-5.35313308e-01 2.16785207e-01 7.37978101e-01 6.62893176e-01
-1.13257360e+00 3.51910919e-01 6.20552778e+00 6.52733684e-01
-6.23270869e-01 2.15241626e-01 9.94247019e-01 -5.44839597e-04
-3.51457506e-01 1.15247160e-01 -6.33573413e-01 5.90280712e-01
1.22160816e+00 -7.83847123e-02 9.83254984e-02 4.56720054e-01
7.32656896e-01 -6.73340797e-01 -7.88641095e-01 5.06050050e-01
-3.20697963e-01 -6.84689224e-01 -5.85455120e-01 6.54433072e-01
9.37379122e-01 -2.47645527e-01 2.96474844e-01 3.66255164e-01
4.73742545e-01 -1.03304088e+00 8.94717813e-01 5.61193168e-01
6.80680215e-01 -9.76702213e-01 7.70922542e-01 6.66387618e-01
-8.44530940e-01 -2.51178443e-01 -3.20402741e-01 -3.43686402e-01
1.13846045e-02 1.07866228e+00 -4.42805946e-01 4.50433612e-01
5.08527219e-01 2.70768493e-01 -4.97946262e-01 1.13759625e+00
-2.43285656e-01 1.14681494e+00 -1.86964914e-01 2.08757576e-02
-7.95510709e-02 -6.86710000e-01 3.65871936e-01 1.11158156e+00
6.54104352e-01 5.90735115e-02 -3.68549861e-02 9.67997432e-01
4.16427143e-02 3.50521654e-02 -5.42510688e-01 -4.06664647e-02
5.08736610e-01 8.70188177e-01 -5.32300949e-01 -2.02504307e-01
-7.97574401e-01 6.05937839e-01 -3.25935967e-02 6.63754463e-01
-7.13980377e-01 1.20450556e-01 5.34259319e-01 -4.03362177e-02
4.03682262e-01 -3.85283828e-01 -5.55299520e-01 -1.11256158e+00
7.92442169e-03 -7.94555485e-01 3.06215882e-01 -2.41455123e-01
-1.44892943e+00 -7.09745288e-01 2.74150133e-01 -7.95805752e-01
-2.90568978e-01 -5.15091658e-01 -6.14199281e-01 1.20433247e+00
-1.06023741e+00 -7.70163417e-01 3.03333312e-01 1.76071048e-01
4.22890544e-01 3.06647569e-01 4.63599652e-01 3.29457581e-01
-8.47149014e-01 5.74101925e-01 6.68385625e-01 -9.75996032e-02
6.89869881e-01 -1.24985719e+00 3.72996390e-01 1.13598680e+00
-1.91091076e-01 7.95457840e-01 7.62832224e-01 -1.15679634e+00
-1.16678786e+00 -6.61756814e-01 9.60699379e-01 -2.28566483e-01
8.05264711e-01 -7.23917842e-01 -9.83205676e-01 8.31467152e-01
-2.46816531e-01 -4.04120713e-01 8.77264202e-01 6.14785135e-01
-2.71701422e-02 1.94549158e-01 -1.16715145e+00 6.36963069e-01
5.52087128e-01 -5.05066402e-02 -1.12988785e-01 -3.00726737e-03
4.88916665e-01 1.90951347e-01 -6.98443532e-01 5.61002135e-01
7.46316731e-01 -9.65523362e-01 7.50397444e-01 -5.66388130e-01
3.66974831e-01 -1.04242966e-01 -2.94075429e-01 -1.00599146e+00
-4.61148590e-01 -3.29749554e-01 1.61493048e-01 1.36344600e+00
6.84302628e-01 -8.22990060e-01 3.68617386e-01 7.59879291e-01
2.92423517e-01 -4.42282140e-01 -1.27726722e+00 -8.07142675e-01
4.69866663e-01 -4.66265440e-01 3.71392846e-01 1.05119622e+00
-4.23424482e-01 6.22144006e-02 -8.06819618e-01 2.21219987e-01
9.63835537e-01 -5.68161421e-02 8.69714558e-01 -1.77477002e+00
-6.75335169e-01 -2.22653091e-01 9.22212843e-03 -6.20217443e-01
2.58692890e-01 -2.98905671e-01 7.85569921e-02 -9.30767477e-01
7.71769166e-01 -1.17566973e-01 -2.88642198e-01 2.69470334e-01
-5.97761691e-01 -1.84252113e-01 -2.40839347e-01 -1.18343174e-01
-7.04686567e-02 4.13024068e-01 8.04910243e-01 1.05803721e-01
-5.03381193e-01 4.89150375e-01 -9.88433361e-01 5.06171763e-01
5.06797731e-01 -6.22976422e-01 -2.49291718e-01 -7.06114620e-02
1.42847195e-01 4.67157871e-01 5.85197747e-01 -2.71037787e-01
-9.49052572e-02 -7.80091941e-01 4.94114578e-01 -6.05306029e-01
1.19233206e-01 -5.33821881e-01 4.47736204e-01 5.12410402e-01
-2.60396183e-01 2.18817636e-01 1.06398292e-01 5.20538688e-01
1.63117170e-01 -8.40950832e-02 6.22744560e-01 3.31167161e-01
-8.71691406e-02 -6.87891096e-02 -5.21383703e-01 -7.64647201e-02
6.57584548e-01 -9.12274495e-02 -3.10845554e-01 -5.01291275e-01
-4.02179688e-01 3.98319781e-01 1.46465048e-01 1.02577783e-01
4.70939189e-01 -1.29957879e+00 -1.10370743e+00 3.58413965e-01
-8.32172334e-02 -4.92568403e-01 1.44268498e-01 1.56179821e+00
3.07293713e-01 2.95156598e-01 1.65942654e-01 -3.23028356e-01
-1.25961196e+00 2.59847045e-01 -3.99504006e-02 -1.95837662e-01
-1.62348673e-01 6.80456281e-01 7.20983624e-01 -3.84791255e-01
-1.97629392e-01 -2.43829399e-01 3.10642598e-03 2.53320754e-01
4.25079823e-01 7.74160266e-01 -4.92076665e-01 -5.06442428e-01
-3.87086928e-01 4.07683313e-01 3.50129187e-01 -6.88157856e-01
1.39890254e+00 -2.69619882e-01 -3.00552875e-01 1.10061026e+00
9.90962207e-01 6.05962694e-01 -1.16503239e+00 -8.16500634e-02
4.28722829e-01 -5.33529520e-01 -1.09548926e-01 -6.15620852e-01
-4.94918615e-01 4.87951517e-01 3.78114522e-01 2.19249502e-01
1.03989458e+00 -3.60350639e-01 -2.51516938e-01 -1.88468322e-01
2.67235518e-01 -9.61387932e-01 -3.84190679e-01 2.47622237e-01
5.45084119e-01 -1.41871798e+00 2.07000136e-01 -2.68718928e-01
-7.47628510e-01 7.10114062e-01 2.80814797e-01 -2.13466823e-01
6.37683392e-01 2.29176730e-01 -4.64387327e-01 2.94875592e-01
-8.48291814e-01 -1.56049803e-01 6.36355519e-01 4.24930960e-01
4.96663302e-01 5.43419719e-01 -8.20848703e-01 8.65104735e-01
1.04700536e-01 -1.77537158e-01 4.82627273e-01 5.14499485e-01
-2.27209434e-01 -9.58317101e-01 -5.79723835e-01 8.18865538e-01
-8.09405208e-01 -1.01390287e-01 -4.32307661e-01 1.10555661e+00
-1.85375780e-01 1.18757617e+00 8.81225392e-02 -1.46889389e-01
1.55926883e-01 -1.27978092e-02 9.87889767e-02 -4.68656778e-01
-3.92162293e-01 8.15517306e-01 2.21347243e-01 1.40389509e-03
-4.13631946e-01 -1.16527879e+00 -7.48960018e-01 -3.31891119e-01
-4.72967207e-01 2.43534029e-01 2.81243622e-01 9.82395172e-01
6.12982810e-02 1.90309197e-01 9.66045499e-01 -7.17864335e-01
-5.69645584e-01 -1.00181019e+00 -9.00565684e-01 -1.47156507e-01
6.53209984e-01 -8.89120758e-01 -7.19327152e-01 -4.61302698e-01]
|
[7.685286998748779, 4.957160949707031]
|
efed939f-9d25-4573-9b9d-3c8dc8c6d28c
|
studying-the-impact-of-filling-information
| null | null |
https://aclanthology.org/2020.inlg-1.6
|
https://aclanthology.org/2020.inlg-1.6.pdf
|
Studying the Impact of Filling Information Gaps on the Output Quality of Neural Data-to-Text
|
It is unfair to expect neural data-to-text to produce high quality output when there are gaps between system input data and information contained in the training text. Thomson et al. (2020) identify and narrow information gaps in Rotowire, a popular data-to-text dataset. In this paper, we describe a study which finds that a state-of-the-art neural data-to-text system produces higher quality output, according to the information extraction (IE) based metrics, when additional input data is carefully selected from this newly available source. It remains to be shown, however, whether IE metrics used in this study correlate well with humans in judging text quality.
|
['Somayajulu Sripada', 'Zhijie Zhao', 'Craig Thomson']
| null | null | null | null |
inlg-acl-2020-12
|
['data-to-text-generation']
|
['natural-language-processing']
|
[ 4.49013151e-02 1.56854033e-01 -4.05379564e-01 -5.90988040e-01
-6.02240682e-01 -6.50380552e-01 6.66358531e-01 6.38894737e-01
-6.68476164e-01 7.82732904e-01 4.69145447e-01 -4.97219235e-01
-3.97395432e-01 -8.25532913e-01 -5.51641583e-01 2.89684325e-01
5.26768565e-01 4.78997469e-01 -1.78617761e-01 -3.40426207e-01
8.68014693e-01 2.82889515e-01 -1.70925534e+00 6.60563946e-01
1.13878536e+00 7.76629627e-01 1.29963979e-01 8.08869779e-01
-5.02666533e-01 8.71304393e-01 -1.14034653e+00 -5.39851665e-01
2.40691140e-01 -5.29351413e-01 -1.10631287e+00 -4.22986239e-01
7.07810938e-01 -4.32161152e-01 -5.25610596e-02 8.09480667e-01
5.05554616e-01 -1.24562569e-01 7.55772591e-01 -1.13193893e+00
-1.16860354e+00 1.03833616e+00 8.57884437e-02 3.92621696e-01
4.84151602e-01 9.11434069e-02 1.04871595e+00 -1.15309632e+00
6.35704696e-01 9.82604027e-01 6.22524559e-01 4.65843260e-01
-1.06896591e+00 -5.99015653e-01 -2.48666793e-01 -1.12888768e-01
-7.33811736e-01 -6.28881335e-01 5.13233244e-01 -8.04621100e-01
1.32632387e+00 1.39022559e-01 6.60669565e-01 9.48813021e-01
5.99299133e-01 4.33277726e-01 9.42917645e-01 -7.26570249e-01
2.23504007e-01 5.23459554e-01 5.25192142e-01 3.23601216e-01
8.20132732e-01 2.92419136e-01 -9.93883908e-01 2.84757793e-01
3.44093055e-01 -4.54689682e-01 -1.04639694e-01 4.53297138e-01
-8.09264004e-01 6.17572725e-01 2.72853464e-01 7.85721362e-01
-5.31935692e-01 -1.71148539e-01 4.28351402e-01 8.11426997e-01
5.81753790e-01 1.13310194e+00 -6.61458075e-01 -4.81911987e-01
-1.43394196e+00 1.48955479e-01 1.06375718e+00 9.41662669e-01
3.74353439e-01 1.78666875e-01 -2.68376768e-01 6.10485971e-01
4.88331258e-01 5.29960513e-01 6.81307197e-01 -8.47933173e-01
8.78133237e-01 9.66984272e-01 2.23930031e-01 -1.27456629e+00
-5.07607281e-01 -3.79570335e-01 -6.43245041e-01 4.09071743e-01
5.93879402e-01 -4.31690395e-01 -7.83047557e-01 1.24290812e+00
-3.35838377e-01 -1.27823198e+00 2.61582643e-01 6.35509789e-01
1.10887933e+00 5.73202014e-01 -2.83560306e-01 -1.39449248e-02
9.58929598e-01 -5.38051724e-01 -1.14376736e+00 -1.96563900e-01
4.41268265e-01 -7.27110982e-01 1.36634254e+00 6.88331723e-01
-1.35824013e+00 -8.70828152e-01 -1.25244129e+00 -2.96688944e-01
-8.63699913e-01 3.59145075e-01 1.31916448e-01 8.73067200e-01
-1.09557331e+00 9.28765297e-01 -3.09102833e-01 -3.16985399e-01
4.37028170e-01 2.25694358e-01 -3.71023864e-01 3.82928491e-01
-1.36845231e+00 1.35043287e+00 6.06504977e-01 -1.70708939e-01
-5.38347065e-01 -6.22473061e-01 -5.81845045e-01 -2.22738367e-04
2.75576085e-01 -4.83442932e-01 1.49021041e+00 -9.74412084e-01
-1.23059618e+00 5.98302066e-01 -8.87073390e-03 -3.10310632e-01
2.35246256e-01 -3.71636659e-01 -5.64914703e-01 4.37126756e-02
6.46350905e-02 3.89915735e-01 4.83641177e-01 -1.39358926e+00
-7.92599559e-01 -5.25293946e-01 -5.69577143e-02 1.45691484e-01
-5.45594633e-01 1.91527650e-01 -3.90742831e-02 -5.17315984e-01
-3.87185901e-01 -2.47842923e-01 3.16611558e-01 5.42410323e-03
-3.85927707e-01 -5.51621258e-01 2.18023300e-01 -1.04410887e+00
1.88843417e+00 -1.71206367e+00 -3.39617938e-01 2.32200682e-01
4.32756811e-01 2.37602443e-01 -3.23251821e-02 6.56760395e-01
2.14381188e-01 8.07782590e-01 1.45673960e-01 1.58662379e-01
1.99637756e-01 -1.18025243e-01 -7.62235895e-02 -4.77002710e-02
3.19929600e-01 7.87837029e-01 -7.08038032e-01 -6.84612215e-01
-1.46887422e-01 -7.57117867e-02 -2.26020575e-01 1.35304138e-01
-9.08401459e-02 -2.68028110e-01 -3.88749510e-01 5.05623519e-01
1.28714472e-01 -1.42075941e-01 -3.85658324e-01 -8.26316997e-02
-3.08878481e-01 4.69350576e-01 -6.88525140e-01 1.51317966e+00
-2.99366921e-01 1.26907408e+00 -3.66388798e-01 -4.68663007e-01
1.12947774e+00 4.58826631e-01 9.94083881e-02 -1.09044242e+00
6.12512827e-01 3.90231848e-01 2.63206065e-01 -7.41734982e-01
1.21263862e+00 -3.81700252e-03 -4.67738099e-02 6.11159146e-01
2.69012988e-01 -3.82924080e-01 6.11654282e-01 2.84134179e-01
7.98361301e-01 -7.80109465e-02 2.47484207e-01 -4.80121493e-01
-4.76512052e-02 5.26932299e-01 2.48971544e-02 9.56663549e-01
6.49184138e-02 3.59065652e-01 4.99914348e-01 -1.92156628e-01
-1.65545738e+00 -5.91069400e-01 -2.65639931e-01 8.55529249e-01
-4.79970217e-01 -4.38835770e-01 -1.08401358e+00 -7.48494625e-01
7.24430606e-02 1.41792333e+00 -6.89077318e-01 -1.10993795e-01
1.47291988e-01 -2.07431391e-01 6.10202551e-01 5.74518204e-01
1.99798152e-01 -1.23888862e+00 -9.56008494e-01 3.80059689e-01
-5.27756512e-02 -4.71481532e-01 -2.07584724e-01 5.40034533e-01
-1.13945389e+00 -8.09601665e-01 -5.17784953e-01 -3.38032156e-01
5.14483690e-01 -1.28798202e-01 1.45287716e+00 1.87740579e-01
1.56383172e-01 8.32625255e-02 -6.12347186e-01 -8.28848422e-01
-8.99836838e-01 1.71564698e-01 -1.19902760e-01 -8.64683211e-01
7.70835459e-01 2.53308564e-01 -2.27156803e-02 2.51667023e-01
-1.12662268e+00 8.30064267e-02 6.77697539e-01 7.47987032e-01
9.59606543e-02 3.73069882e-01 8.76347482e-01 -8.84301901e-01
1.44199407e+00 -5.05618989e-01 -2.00332299e-01 1.94045916e-01
-1.42555261e+00 1.62462518e-01 8.05840492e-01 -1.80641904e-01
-8.40169549e-01 -5.09650409e-01 3.02829384e-03 2.08968282e-01
-1.76531047e-01 1.02685380e+00 7.79460892e-02 4.47114676e-01
1.39245951e+00 -1.79324895e-01 2.86495257e-02 -3.86545479e-01
2.75642186e-01 1.35103917e+00 4.10544068e-01 -5.76167226e-01
5.26925981e-01 -2.47468352e-01 -5.48706353e-01 -4.11550611e-01
-8.22540820e-01 -2.40109727e-01 -8.08663726e-01 -5.93709409e-01
5.64212322e-01 -5.20709276e-01 -3.38371336e-01 4.02626008e-01
-1.25359988e+00 -3.76089215e-01 -6.08547807e-01 3.81922305e-01
-3.85489404e-01 -2.03772858e-01 -2.16843873e-01 -8.62854421e-01
-6.66287243e-01 -7.39112079e-01 5.34834445e-01 4.56691682e-01
-7.17022419e-01 -7.30351388e-01 -2.20628276e-01 2.78692335e-01
4.35615361e-01 -2.94702556e-02 9.49663818e-01 -1.14796579e+00
2.28032202e-01 -4.12680656e-01 -1.94008917e-01 5.64900815e-01
2.18081549e-01 4.60999727e-01 -7.92746663e-01 -1.10679358e-01
8.41028616e-03 -4.68475342e-01 5.36427140e-01 1.69225812e-01
9.07331645e-01 -7.28102684e-01 1.84397310e-01 -1.79224595e-01
1.35853922e+00 6.42524421e-01 4.59984571e-01 4.85240132e-01
2.15143457e-01 1.06806672e+00 5.84881127e-01 3.10770333e-01
1.18023053e-01 6.36584386e-02 -5.87712154e-02 9.20808390e-02
-8.98649022e-02 -4.72099960e-01 2.04300880e-01 9.66028750e-01
2.34501585e-01 -6.76356316e-01 -1.16064835e+00 7.44147480e-01
-1.35523641e+00 -8.10701847e-01 -3.32653105e-01 1.98893631e+00
1.39168799e+00 7.46235490e-01 7.29119182e-02 4.98869777e-01
5.31324685e-01 -3.77011776e-01 -5.92142105e-01 -6.84968650e-01
-4.30170856e-02 1.59238279e-01 4.12422955e-01 3.64913344e-01
-4.93094444e-01 4.94105369e-01 7.33580160e+00 5.98209083e-01
-7.21884489e-01 -2.72438735e-01 6.51402891e-01 -2.06183702e-01
-6.38263226e-01 -2.33427450e-01 -8.59617233e-01 6.04527414e-01
1.58117008e+00 -6.36348188e-01 2.74554878e-01 5.86389482e-01
3.94400209e-01 -4.05008912e-01 -1.22679031e+00 5.81684947e-01
2.18500972e-01 -1.25464141e+00 2.19891757e-01 -9.98964384e-02
6.77835405e-01 -2.41111323e-01 -1.12665288e-01 1.16407745e-01
5.30619383e-01 -1.23639607e+00 1.05810750e+00 9.05977726e-01
8.21914256e-01 -6.59014940e-01 9.95983720e-01 4.12155479e-01
-3.62215489e-01 2.05196831e-02 -4.53149438e-01 -1.13189546e-02
-1.50309086e-01 8.26900780e-01 -7.90517449e-01 4.64928746e-01
8.21898580e-01 4.89065886e-01 -1.00712669e+00 7.97324181e-01
-3.72413546e-02 4.94417101e-01 -6.09224029e-02 -7.53035724e-01
5.37091829e-02 1.98715240e-01 3.14928710e-01 1.13976920e+00
3.59846532e-01 -1.60941169e-01 -5.13387680e-01 1.11013532e+00
-1.77991569e-01 3.95930082e-01 -1.02656448e+00 -6.04776740e-01
5.80274642e-01 8.11570168e-01 -4.63106602e-01 -5.24278581e-01
-2.89213449e-01 5.09531021e-01 -4.59837317e-02 1.66811123e-01
4.53979000e-02 -9.70747173e-01 -7.05482289e-02 1.40528426e-01
-2.85794020e-01 -1.10410593e-01 -1.00585914e+00 -6.53931081e-01
2.32772723e-01 -1.21867156e+00 1.78900167e-01 -1.15300345e+00
-1.46801186e+00 7.58219481e-01 1.34096920e-01 -1.01868725e+00
-5.01016319e-01 -7.12161899e-01 -5.46204746e-02 1.20770967e+00
-9.14275706e-01 -5.04494727e-01 -3.14681619e-01 1.84880942e-01
6.04353368e-01 -5.23828328e-01 4.98336256e-01 -6.86278120e-02
-2.49881685e-01 7.22678065e-01 3.97377938e-01 3.87135893e-01
6.81917012e-01 -1.53307807e+00 4.93557632e-01 7.46136725e-01
6.09244630e-02 9.12226915e-01 7.65278637e-01 -1.06535196e+00
-1.19426405e+00 -6.72607303e-01 1.29861045e+00 -8.42702568e-01
4.90338296e-01 1.14146970e-01 -9.03372765e-01 2.82559037e-01
9.32646871e-01 -9.96000290e-01 8.94093394e-01 3.31833139e-02
-2.60942072e-01 -1.21923946e-01 -1.34824061e+00 4.80040550e-01
5.45404196e-01 -7.10327804e-01 -1.25640440e+00 -6.49607852e-02
8.13412845e-01 -2.75282204e-01 -1.28348780e+00 1.01518996e-01
7.33508229e-01 -7.22942173e-01 2.74498999e-01 -8.41120601e-01
1.06622934e+00 1.21579207e-02 -5.16268574e-02 -1.71283090e+00
-2.36286111e-02 -3.48478019e-01 1.44822642e-01 1.31943870e+00
1.12451017e+00 -3.73862267e-01 4.92449701e-01 8.88834715e-01
-1.57108784e-01 -5.22552133e-01 -5.95862567e-01 -7.12044060e-01
6.06813431e-01 -4.23119813e-01 6.06342316e-01 9.32112813e-01
4.61146027e-01 6.30257428e-01 -7.80933350e-03 -5.59524119e-01
1.66080698e-01 -3.68939847e-01 5.38079977e-01 -1.64206827e+00
2.22437695e-01 -8.34384203e-01 1.14176065e-01 -7.60299265e-01
-3.28646421e-01 -7.79760361e-01 3.09286952e-01 -2.05858254e+00
-2.80102063e-02 -1.16754614e-01 -1.43944383e-01 6.00690782e-01
-4.89219353e-02 -2.80239761e-01 3.50647867e-01 2.70137042e-01
-1.66700426e-02 1.59844726e-01 1.39114964e+00 -7.77178705e-02
-2.40835294e-01 -1.90245688e-01 -1.35637772e+00 3.16326052e-01
1.21185803e+00 -7.14071810e-01 -2.99499989e-01 -5.77134907e-01
8.20447624e-01 1.26233578e-01 -8.88931081e-02 -1.08912420e+00
3.22907835e-01 -3.45924795e-01 8.23486030e-01 -9.55326140e-01
-2.28981227e-01 -1.02330744e+00 -3.07679754e-02 2.25791693e-01
-1.03449881e+00 5.75366199e-01 5.63290775e-01 -8.63763690e-03
-9.95451584e-02 -1.00286055e+00 5.57997763e-01 -1.07582454e-02
-1.98981687e-01 -2.18942240e-01 -7.28484452e-01 3.52186739e-01
3.48139435e-01 -3.03962499e-01 -9.34378743e-01 -4.50211257e-01
-2.29384676e-01 1.30634502e-01 4.45407927e-01 6.41810477e-01
6.00594342e-01 -1.23195267e+00 -9.22309637e-01 1.01026855e-01
2.04050899e-01 -1.68305472e-01 -3.26884747e-01 3.03372115e-01
-6.28063500e-01 6.87961400e-01 -6.25159442e-01 -3.86211388e-02
-9.19689119e-01 3.66807789e-01 1.34283468e-01 2.55998466e-02
-3.86061132e-01 3.77809286e-01 -8.22765470e-01 -2.93548793e-01
4.06527936e-01 -6.84591174e-01 -5.11840641e-01 5.18802345e-01
6.07100129e-01 4.80052829e-01 2.75314063e-01 -2.14519694e-01
2.53735155e-01 1.35005489e-01 -2.95240302e-02 -6.33739769e-01
1.06003392e+00 4.48592007e-02 1.43366456e-01 9.72063959e-01
8.34620595e-01 -1.39611691e-01 -6.48166060e-01 -2.91595459e-01
2.92037427e-01 -5.91956317e-01 3.88660967e-01 -1.48928022e+00
-7.15575278e-01 9.24989700e-01 5.59394419e-01 8.02485168e-01
1.10952032e+00 -3.64678681e-01 5.30760944e-01 8.06145072e-01
7.64369965e-02 -1.83684039e+00 2.39075080e-01 4.76631045e-01
1.23825622e+00 -1.02254629e+00 3.97602320e-02 3.34180236e-01
-6.91187203e-01 1.35998809e+00 8.11733305e-01 2.56214380e-01
5.21502614e-01 3.68204862e-01 1.35442615e-01 -2.68345594e-01
-8.60327840e-01 -4.11973859e-04 4.96575624e-01 5.92924416e-01
7.50923038e-01 -2.13950917e-01 -5.88367760e-01 7.09385633e-01
-7.22021699e-01 3.66346180e-01 7.92899132e-01 9.27820861e-01
-8.95380020e-01 -8.70015740e-01 -3.11769426e-01 1.33698606e+00
-4.97104496e-01 -3.88919234e-01 -9.66215432e-01 8.99240136e-01
-1.40115738e-01 1.50848830e+00 9.36751440e-02 -6.15729153e-01
6.02182865e-01 2.68600404e-01 7.53188059e-02 -6.46354795e-01
-1.25666666e+00 -2.71320730e-01 5.19327283e-01 -1.17462724e-01
-3.57965231e-01 -4.99868244e-01 -1.24458051e+00 -4.39337909e-01
-5.22199392e-01 3.80766958e-01 1.03732061e+00 7.87542760e-01
2.66544998e-01 6.48288012e-01 4.57381994e-01 -1.90999106e-01
-6.36166334e-01 -1.46363533e+00 -3.65345150e-01 2.77103156e-01
3.41049612e-01 -2.46779084e-01 -2.66522318e-01 2.68524319e-01]
|
[11.7523193359375, 9.150795936584473]
|
e38ad972-d61c-4353-9778-dba266c67819
|
deltanet-conditional-medical-report
| null | null |
https://aclanthology.org/2022.coling-1.261
|
https://aclanthology.org/2022.coling-1.261.pdf
|
DeltaNet: Conditional Medical Report Generation for COVID-19 Diagnosis
|
Fast screening and diagnosis are critical in COVID-19 patient treatment. In addition to the gold standard RT-PCR, radiological imaging like X-ray and CT also works as an important means in patient screening and follow-up. However, due to the excessive number of patients, writing reports becomes a heavy burden for radiologists. To reduce the workload of radiologists, we propose DeltaNet to generate medical reports automatically. Different from typical image captioning approaches that generate reports with an encoder and a decoder, DeltaNet applies a conditional generation process. In particular, given a medical image, DeltaNet employs three steps to generate a report: 1) first retrieving related medical reports, i.e., the historical reports from the same or similar patients; 2) then comparing retrieved images and current image to find the differences; 3) finally generating a new report to accommodate identified differences based on the conditional report. We evaluate DeltaNet on a COVID-19 dataset, where DeltaNet outperforms state-of-the-art approaches. Besides COVID-19, the proposed DeltaNet can be applied to other diseases as well. We validate its generalization capabilities on the public IU-Xray and MIMIC-CXR datasets for chest-related diseases.
|
['Li Xiao', 'S. Kevin Zhou', 'Yefeng Zheng', 'Xingwang Wu', 'Yangtian Yan', 'Shen Ge', 'Zhaopeng Qiu', 'Shuxin Yang', 'Xian Wu']
| null | null | null | null |
coling-2022-10
|
['covid-19-detection', 'medical-report-generation']
|
['medical', 'medical']
|
[ 4.11981940e-01 1.66315794e-01 -2.17563361e-01 -2.25250915e-01
-1.22989321e+00 -3.74015629e-01 3.59660536e-01 5.14628053e-01
-3.70531768e-01 9.73661065e-01 4.21184868e-01 -6.20621383e-01
-1.93529606e-01 -9.11934137e-01 -5.26207387e-01 -5.29525578e-01
2.39656165e-01 7.14476109e-01 3.50439250e-01 3.02735567e-01
-6.26274664e-03 3.12792748e-01 -9.60627437e-01 3.86039555e-01
6.77145898e-01 8.59171152e-01 5.82391322e-01 7.83070922e-01
-2.38739438e-02 1.04447222e+00 -7.42464244e-01 -3.76384884e-01
-1.00731418e-01 -9.86001194e-01 -7.11320579e-01 3.08400691e-01
-2.84065548e-02 -6.51982248e-01 -4.16666716e-01 7.45587647e-01
7.00877309e-01 -1.05397157e-01 6.97800636e-01 -6.28548443e-01
-7.60485053e-01 5.45017958e-01 -7.70508170e-01 4.58679199e-01
4.91141528e-01 2.39000946e-01 4.84729439e-01 -8.06266010e-01
8.69819522e-01 8.14049542e-01 3.29930246e-01 6.88669205e-01
-7.59013176e-01 -7.12723553e-01 -1.68770477e-01 -3.38849686e-02
-1.28707886e+00 1.11355960e-01 1.33251831e-01 -2.90980786e-01
5.42071760e-01 4.47921991e-01 7.42472291e-01 1.01108134e+00
5.07071972e-01 5.25955796e-01 7.80939639e-01 -1.36807039e-01
1.34551659e-01 -2.35400144e-02 -1.78694487e-01 7.88559675e-01
3.73595268e-01 -1.89880431e-01 -2.76985168e-01 -2.45381042e-01
7.85079122e-01 2.62660593e-01 -5.44643998e-01 2.40920320e-01
-1.59871733e+00 8.23656619e-01 4.17532057e-01 2.88080633e-01
-6.78101361e-01 -3.43684629e-02 4.98040795e-01 -1.69482753e-01
2.23839551e-01 1.17821567e-01 5.57053229e-03 7.84564838e-02
-1.00485921e+00 1.72322452e-01 5.33470273e-01 7.70832539e-01
1.37693420e-01 -4.00715947e-01 -7.26444185e-01 6.37604833e-01
1.79585442e-01 4.00918841e-01 8.61005425e-01 -3.54404569e-01
7.64416039e-01 5.25649607e-01 -6.88354895e-02 -8.94577146e-01
-3.33012670e-01 -5.46667695e-01 -1.25716710e+00 -5.36951900e-01
-2.64896490e-02 -1.51005760e-01 -1.20197189e+00 1.27737844e+00
2.32917026e-01 3.17981809e-01 1.46591872e-01 9.81672108e-01
1.28242862e+00 7.01819956e-01 1.95103865e-02 -5.75404167e-01
1.67863107e+00 -7.55693793e-01 -7.28359103e-01 1.73886016e-03
6.53883934e-01 -9.28697050e-01 6.30047321e-01 2.27306426e-01
-1.15986753e+00 -3.67431998e-01 -8.83892953e-01 1.84222147e-01
2.04550046e-02 5.44362426e-01 3.64120513e-01 1.93988040e-01
-9.16376829e-01 1.10649385e-01 -7.07755625e-01 -3.82507890e-01
2.39059836e-01 2.32364222e-01 -3.06102216e-01 -3.46036792e-01
-1.07724512e+00 7.69544661e-01 7.85454690e-01 3.73983681e-02
-8.99556875e-01 -7.56153226e-01 -9.35865760e-01 -1.13288529e-01
5.53944588e-01 -1.05253983e+00 1.48894846e+00 -1.41876861e-01
-1.00834012e+00 8.64105463e-01 -1.02123976e-01 -5.65220892e-01
7.01094985e-01 1.64203614e-01 -4.71819103e-01 4.56115425e-01
4.34288889e-01 6.70561910e-01 5.18282115e-01 -9.71942842e-01
-5.11312664e-01 -4.75640446e-02 -1.66381598e-01 2.23168701e-01
1.80461314e-02 -1.05839008e-02 -9.78657901e-01 -8.15366089e-01
1.48222387e-01 -9.35779750e-01 -4.36013162e-01 2.98668556e-02
-7.95473993e-01 -1.28577054e-02 8.51878405e-01 -7.74431646e-01
1.42936158e+00 -2.10992956e+00 -3.31434518e-01 1.92353994e-01
4.37679023e-01 2.50261098e-01 2.07669958e-01 9.66342911e-02
-2.63185918e-01 1.57604128e-01 -4.50888932e-01 -6.38512000e-02
-5.61975837e-01 1.40538767e-01 -6.21177740e-02 1.77056581e-01
3.54032040e-01 8.90220821e-01 -8.97140741e-01 -1.10742378e+00
1.34335697e-01 4.08309996e-01 -1.88743293e-01 2.88750380e-01
-1.19324923e-01 6.55700684e-01 -5.82004249e-01 5.46159863e-01
5.52505374e-01 -1.03316021e+00 2.00327247e-01 -1.01060502e-01
1.13829300e-01 2.84040242e-01 -9.19883788e-01 1.54100120e+00
-5.41263640e-01 3.36203307e-01 -3.00736487e-01 -7.26367176e-01
7.55134761e-01 7.09600687e-01 5.01848400e-01 -6.16927862e-01
8.30071419e-02 2.18956843e-01 8.87889788e-02 -5.99214733e-01
4.48917836e-01 -1.25489026e-01 -1.05073601e-01 9.09017503e-01
-3.43835801e-01 -2.89030463e-01 4.10361886e-01 5.26861310e-01
1.30180955e+00 -3.65350038e-01 6.95289612e-01 4.29118931e-01
6.06805086e-01 1.97137177e-01 4.99964774e-01 8.35329831e-01
1.93332344e-01 1.02177441e+00 2.61300474e-01 -2.87695736e-01
-8.05237710e-01 -1.20216477e+00 -1.36595309e-01 2.45441735e-01
8.23056772e-02 -4.03179824e-01 -3.78847092e-01 -9.77856576e-01
-2.55201548e-01 6.19877338e-01 -6.97832644e-01 -3.18966322e-02
-5.33406079e-01 -8.91489625e-01 4.29063946e-01 7.18084812e-01
5.66016674e-01 -1.09356821e+00 -7.84822464e-01 5.54222822e-01
-3.88562053e-01 -1.02800071e+00 -7.76269019e-01 2.99872402e-02
-7.76427150e-01 -1.15132964e+00 -1.20898628e+00 -6.72554314e-01
8.62290502e-01 1.58766508e-01 1.14066112e+00 4.22718704e-01
-5.81716299e-01 1.14208348e-01 -2.65772402e-01 -3.99147034e-01
-8.12821388e-01 4.86014970e-02 -3.81354988e-01 -3.66811961e-01
-2.31655911e-01 -3.91827300e-02 -8.98153305e-01 1.15270995e-01
-1.23310590e+00 4.71147507e-01 9.54993188e-01 9.11279380e-01
9.80163813e-01 -7.80380741e-02 4.40266222e-01 -9.41473365e-01
8.45924437e-01 -6.34202540e-01 -3.53885680e-01 3.89451593e-01
-4.07407105e-01 -1.81569420e-02 4.76711839e-01 -2.77237952e-01
-1.24588680e+00 1.93633717e-02 -3.11854720e-01 -2.71414191e-01
7.91470036e-02 6.50032699e-01 5.72705925e-01 4.68142599e-01
5.47832966e-01 5.18442333e-01 -9.11246985e-02 -8.46280828e-02
-3.39441299e-02 5.72036028e-01 7.96656728e-01 3.27247055e-03
5.34854233e-01 3.96781564e-01 7.78478533e-02 -3.56048226e-01
-1.03309059e+00 -4.72089618e-01 -2.27278516e-01 -1.63256809e-01
1.28722596e+00 -7.36329734e-01 -5.06260157e-01 -9.52908620e-02
-1.46363294e+00 3.89667720e-01 -1.93416402e-01 8.63620520e-01
-2.68506050e-01 2.27735460e-01 -8.90661716e-01 -3.31812054e-01
-7.65046000e-01 -1.60826421e+00 1.07794476e+00 4.14163113e-01
-2.22905219e-01 -8.04473221e-01 -2.02332586e-02 4.40907218e-02
5.65217622e-02 3.79332334e-01 1.00723553e+00 -6.41939402e-01
-6.02825105e-01 -1.14411704e-01 -4.64690864e-01 5.02301119e-02
5.53498805e-01 -5.27397357e-02 -4.73252982e-01 -4.70930040e-02
-7.06081167e-02 -8.52281973e-02 9.00324464e-01 4.21305656e-01
1.36573124e+00 -2.22694427e-01 -5.86145163e-01 4.21386182e-01
1.26047242e+00 7.47942388e-01 5.04407406e-01 1.04206301e-01
4.60632443e-01 7.33564273e-02 7.72990346e-01 5.35711467e-01
3.07984501e-01 2.96540916e-01 4.75455493e-01 -3.20883662e-01
-1.77987918e-01 -2.19305143e-01 -1.97749883e-01 9.76750791e-01
9.55047458e-02 -5.30177951e-01 -1.06757390e+00 5.37159324e-01
-1.53291559e+00 -7.10292339e-01 -1.73574403e-01 1.94508004e+00
9.37604666e-01 2.92979889e-02 -2.76076585e-01 -1.10599957e-01
8.45347822e-01 -1.99298523e-02 -5.47784567e-01 -1.32497787e-01
1.57890096e-01 4.78870183e-01 3.21186513e-01 -1.34309605e-01
-7.51640022e-01 2.10552379e-01 5.94759703e+00 5.93835711e-01
-1.24678242e+00 2.72864252e-01 9.63664949e-01 8.08086246e-02
-2.08740726e-01 -3.72086257e-01 -7.29914784e-01 5.52012861e-01
6.45302892e-01 -1.52424261e-01 -4.32856649e-01 6.86126173e-01
1.61367670e-01 -4.74194169e-01 -9.97533441e-01 9.43028092e-01
1.56224340e-01 -1.82416856e+00 2.85799921e-01 1.01257525e-01
6.60330117e-01 -2.24614084e-01 2.57364772e-02 -3.80786892e-04
2.61067182e-01 -7.77385235e-01 3.76841784e-01 4.78894889e-01
1.09541798e+00 -5.53876758e-01 1.02323079e+00 1.62449121e-01
-1.13761592e+00 4.10737395e-01 -1.87667847e-01 6.90915763e-01
4.13109690e-01 6.86335802e-01 -1.54081142e+00 1.02734327e+00
5.85620403e-01 4.35007453e-01 -5.34241319e-01 1.15900683e+00
-3.39612484e-01 6.25801623e-01 -9.77580324e-02 7.28348568e-02
1.68745622e-01 6.37866408e-02 2.60869801e-01 1.24617267e+00
7.89053738e-01 2.85005838e-01 1.04106978e-01 7.07309544e-01
-2.17687860e-01 2.02144712e-01 -6.43060327e-01 1.14100590e-01
2.48253226e-01 1.06278729e+00 -1.10640037e+00 -7.61525273e-01
-4.79074627e-01 9.36420143e-01 -2.54266471e-01 -3.29121612e-02
-1.10695016e+00 -1.29040211e-01 -3.52411531e-02 1.78843260e-01
2.56182075e-01 1.73727110e-01 3.77217382e-02 -8.26726794e-01
5.64549863e-02 -6.76580310e-01 7.67344952e-01 -1.09906948e+00
-1.00579619e+00 8.45506608e-01 3.19226272e-02 -1.65570521e+00
-7.06891835e-01 -7.09091499e-02 -5.70446551e-01 8.76655698e-01
-1.29327953e+00 -7.56909072e-01 -5.93166530e-01 3.57624114e-01
6.13376498e-01 1.11437291e-01 8.18024635e-01 2.06030980e-01
-4.08128232e-01 2.99381703e-01 -1.69695273e-01 3.03371131e-01
7.33185172e-01 -1.01621401e+00 3.93790960e-01 6.97706580e-01
-2.68661857e-01 5.18358290e-01 2.65119970e-01 -8.93100560e-01
-8.87004972e-01 -1.40585756e+00 7.38217056e-01 1.29229218e-01
3.41431558e-01 1.84780329e-01 -9.42223012e-01 5.33153832e-01
2.86132336e-01 1.58785716e-01 6.14721179e-01 -8.42922390e-01
1.73403203e-01 5.55778891e-02 -1.15825999e+00 5.16013563e-01
7.40869999e-01 -2.05000669e-01 -4.96992290e-01 5.29459417e-01
1.02368259e+00 -9.45547283e-01 -9.13888931e-01 3.41830105e-01
3.09528083e-01 -7.12362885e-01 8.14965904e-01 -1.63776562e-01
7.81712055e-01 -4.27488446e-01 3.23911369e-01 -1.20654297e+00
-3.25445496e-02 -2.59161979e-01 2.54186541e-01 9.70707238e-01
5.88174045e-01 -5.10795355e-01 5.81720710e-01 1.57670200e-01
-1.30829334e-01 -8.76879036e-01 -6.16120994e-01 -3.65155071e-01
-5.44305742e-01 -5.38364947e-01 6.60116613e-01 7.03458250e-01
-4.62731212e-01 9.21659693e-02 -6.09772764e-02 6.41184375e-02
2.63652921e-01 9.33514237e-02 4.57163751e-01 -8.18720520e-01
-4.65210885e-01 -6.80368394e-02 -1.50249243e-01 -7.17520535e-01
-3.69759351e-01 -8.99081051e-01 1.58813357e-01 -2.06922555e+00
5.81880569e-01 -3.35686833e-01 -9.10215825e-02 3.90491515e-01
-4.43660915e-01 4.35670912e-01 1.38985813e-01 3.15557778e-01
-4.66666937e-01 1.22543452e-02 1.74612653e+00 -2.84312874e-01
-1.69124350e-01 1.17317155e-01 -5.83331943e-01 6.19324088e-01
6.34502828e-01 -7.97160923e-01 -4.36559767e-01 -1.94431230e-01
2.89263278e-01 8.64603579e-01 3.64197880e-01 -9.56351459e-01
1.50445402e-01 -1.00345716e-01 6.16733372e-01 -1.25530303e+00
6.85455352e-02 -5.24491429e-01 3.13815057e-01 9.24098849e-01
-2.03200534e-01 6.16537273e-01 2.19316006e-01 4.98219311e-01
-4.75519717e-01 -4.48778927e-01 6.38632357e-01 -6.02179110e-01
-3.56176585e-01 3.84944439e-01 -6.05445802e-01 5.94585985e-02
1.19149756e+00 -4.54718247e-02 -4.74315494e-01 -3.78266305e-01
-7.19949603e-01 2.23206624e-01 7.72073939e-02 1.79351807e-01
1.07250023e+00 -1.15231109e+00 -8.99840772e-01 -6.14491440e-02
3.87658864e-01 5.77426493e-01 4.57571417e-01 1.10401106e+00
-8.90790880e-01 4.48415160e-01 3.06252062e-01 -8.80692303e-01
-1.38855696e+00 4.25602227e-01 2.11554140e-01 -8.84347975e-01
-7.40488946e-01 6.19303405e-01 6.00800037e-01 6.19204119e-02
-7.39490017e-02 -5.49522817e-01 -1.62595343e-02 -8.88487399e-02
6.38678849e-01 -2.21794188e-01 4.01175290e-01 -3.00467372e-01
-3.65988344e-01 2.61041373e-01 -6.67405665e-01 -2.04945073e-01
1.21403420e+00 1.17496796e-01 -4.73637544e-02 6.54286519e-02
9.02939916e-01 -2.27140754e-01 -5.97202063e-01 -1.84011295e-01
-1.61557555e-01 -2.21726969e-01 -1.54427841e-01 -8.50887239e-01
-1.16409397e+00 5.65518856e-01 5.11988640e-01 -9.37838107e-02
1.35602987e+00 3.45946431e-01 7.89495826e-01 6.42940402e-02
1.27796635e-01 -6.38691604e-01 3.48806947e-01 7.16420710e-02
8.77137840e-01 -1.05566847e+00 2.61449456e-01 -3.49390239e-01
-7.73728251e-01 9.94893372e-01 5.68534315e-01 4.07391787e-01
3.76484245e-01 3.88495922e-01 7.72322789e-02 -2.82218516e-01
-8.50195706e-01 3.70718762e-02 2.43535176e-01 3.39783758e-01
6.28680289e-01 1.27592469e-02 -5.32367408e-01 3.76916111e-01
-9.82094035e-02 2.56507397e-01 7.77910769e-01 1.22697401e+00
1.10083884e-02 -8.49515259e-01 -7.53136456e-01 9.91400540e-01
-7.93050826e-01 -1.56999245e-01 -1.56419307e-01 1.04830098e+00
1.05117247e-01 7.21542001e-01 7.27077723e-02 -2.10809454e-01
3.49761248e-01 -3.18638623e-01 3.02684814e-01 -9.90263283e-01
-6.34016514e-01 -6.87049553e-02 -3.42580862e-02 -1.80673823e-01
-4.00094092e-01 -6.06670678e-01 -1.59958470e+00 2.23980714e-02
-3.26914281e-01 1.56025261e-01 6.26024663e-01 8.09717715e-01
1.45035639e-01 1.01345944e+00 3.25462699e-01 -1.46596685e-01
-3.12276661e-01 -9.39888775e-01 -3.50454539e-01 2.98079431e-01
2.70429581e-01 -4.69561517e-01 1.64021105e-01 3.52875203e-01]
|
[15.05380916595459, -1.3825308084487915]
|
16dc8c0b-45cd-43cc-a3f9-7dbbc2d0cec3
|
graph-transformer-for-graph-to-sequence
|
1911.07470
| null |
https://arxiv.org/abs/1911.07470v2
|
https://arxiv.org/pdf/1911.07470v2.pdf
|
Graph Transformer for Graph-to-Sequence Learning
|
The dominant graph-to-sequence transduction models employ graph neural networks for graph representation learning, where the structural information is reflected by the receptive field of neurons. Unlike graph neural networks that restrict the information exchange between immediate neighborhood, we propose a new model, known as Graph Transformer, that uses explicit relation encoding and allows direct communication between two distant nodes. It provides a more efficient way for global graph structure modeling. Experiments on the applications of text generation from Abstract Meaning Representation (AMR) and syntax-based neural machine translation show the superiority of our proposed model. Specifically, our model achieves 27.4 BLEU on LDC2015E86 and 29.7 BLEU on LDC2017T10 for AMR-to-text generation, outperforming the state-of-the-art results by up to 2.2 points. On the syntax-based translation tasks, our model establishes new single-model state-of-the-art BLEU scores, 21.3 for English-to-German and 14.1 for English-to-Czech, improving over the existing best results, including ensembles, by over 1 BLEU.
|
['Deng Cai', 'Wai Lam']
|
2019-11-18
| null | null | null | null |
['graph-to-sequence']
|
['natural-language-processing']
|
[ 5.69161534e-01 6.25374615e-01 -3.09545547e-01 -4.37120013e-02
-7.13780761e-01 -5.77198744e-01 8.87752295e-01 3.62875879e-01
-2.09947318e-01 9.57028091e-01 5.89563966e-01 -8.50219309e-01
1.83443934e-01 -1.27371395e+00 -9.85932350e-01 -3.23604733e-01
7.54368082e-02 6.74873114e-01 -1.37293741e-01 -7.67074764e-01
-1.30193830e-01 1.77464206e-02 -6.86604261e-01 5.39438367e-01
1.00710881e+00 6.18269086e-01 2.08284855e-01 5.16161323e-01
-5.27860880e-01 7.94135749e-01 -5.54248750e-01 -7.65194058e-01
-2.33134910e-01 -8.73708129e-01 -9.11795437e-01 -5.49313843e-01
1.22914426e-01 1.97945654e-01 -5.87078571e-01 1.08761561e+00
4.99628663e-01 -3.42201404e-02 8.09113443e-01 -7.62180269e-01
-1.66830409e+00 1.42735744e+00 -2.22777173e-01 -2.40106285e-01
5.93050599e-01 -1.54323220e-01 1.34702337e+00 -6.94535077e-01
8.68762255e-01 1.37982166e+00 4.97705877e-01 9.11971271e-01
-1.31708395e+00 -5.19356310e-01 1.69172823e-01 6.91510662e-02
-1.31251717e+00 -2.12727785e-01 7.19271839e-01 -3.90558615e-02
1.61081862e+00 1.83278129e-01 6.49539351e-01 1.34529078e+00
5.08381903e-01 5.18807352e-01 8.89977872e-01 -6.29415989e-01
1.16628699e-01 -5.22143066e-01 -2.96285748e-02 8.26426864e-01
5.03620505e-01 -1.30052902e-02 -5.38620114e-01 -8.21906626e-02
8.07057738e-01 -3.45058531e-01 -4.64806825e-01 1.86282933e-01
-1.37845230e+00 9.36969519e-01 8.40565145e-01 2.58859307e-01
-3.53432685e-01 4.95625705e-01 2.76098609e-01 6.81311488e-01
4.16469663e-01 4.69403237e-01 -2.66895056e-01 3.79359759e-02
-2.64075220e-01 8.16788599e-02 8.73115778e-01 1.26946008e+00
6.36082470e-01 3.11065733e-01 -4.19746131e-01 7.80292809e-01
4.21692163e-01 7.60332227e-01 5.60953557e-01 -4.47079837e-01
9.15394604e-01 8.51778328e-01 -4.95182008e-01 -8.93329501e-01
-2.05050752e-01 -6.34555757e-01 -1.48238444e+00 -5.10270536e-01
3.49330530e-02 -2.23075226e-01 -1.05966043e+00 1.95568240e+00
-3.38760316e-01 -1.73597023e-01 4.97705668e-01 4.41618294e-01
1.33102667e+00 9.35534894e-01 5.09034358e-02 -1.49762467e-01
1.15943742e+00 -9.24917996e-01 -8.27831626e-01 -4.87981141e-01
1.03133345e+00 -5.08459508e-01 1.00741673e+00 -1.61639795e-01
-1.15466177e+00 -5.12326002e-01 -8.55064571e-01 8.24627280e-02
-5.25897741e-01 -1.79713875e-01 6.14948153e-01 4.52588379e-01
-1.62776494e+00 6.09094203e-01 -4.52894211e-01 -5.99703193e-01
9.35994461e-03 3.71433288e-01 -4.63498682e-01 -1.10236190e-01
-1.69112647e+00 8.88083398e-01 6.29641414e-01 -3.26721072e-02
-5.61997175e-01 -3.42919946e-01 -1.10046697e+00 9.81563181e-02
9.51520652e-02 -1.16932464e+00 8.29348624e-01 -6.00539684e-01
-1.73354042e+00 6.21290684e-01 -1.23121202e-01 -6.65546358e-01
-5.61406203e-02 1.27583504e-01 -5.29813409e-01 -1.24152988e-01
-2.12517500e-01 8.53828490e-01 3.80351573e-01 -9.30640757e-01
-1.55322598e-02 -1.05183557e-01 1.39027283e-01 3.88215542e-01
-1.84981868e-01 -1.43355429e-01 -3.63190264e-01 -8.06377172e-01
1.81975798e-03 -9.78374839e-01 -2.40765750e-01 -7.25217819e-01
-4.34700608e-01 -4.68776494e-01 3.35121363e-01 -8.18149149e-01
1.41424084e+00 -1.46932626e+00 4.72274840e-01 -3.89565602e-02
1.98861122e-01 2.54012972e-01 -7.51915574e-01 9.55399930e-01
-1.46138206e-01 4.13345605e-01 -3.29182506e-01 -1.77813619e-01
-3.29025351e-02 2.62443304e-01 -2.26826385e-01 -4.59938496e-03
3.83041173e-01 1.74317408e+00 -9.09336746e-01 -2.05882743e-01
-4.71004173e-02 4.27866608e-01 -4.62784201e-01 2.44255617e-01
-2.21352413e-01 2.85821229e-01 -4.20654178e-01 3.51495266e-01
3.64158750e-01 -5.30519426e-01 5.54856658e-01 2.17004865e-01
5.70682883e-01 7.76317835e-01 -4.90509808e-01 2.11429477e+00
-5.49062490e-01 5.09030700e-01 -3.01040322e-01 -9.48761404e-01
1.21453643e+00 4.00752634e-01 -2.24947125e-01 -7.92952418e-01
9.17005935e-04 1.59558237e-01 2.74810940e-01 1.98359877e-01
5.63618124e-01 -4.53708731e-02 -2.84842372e-01 4.80370432e-01
3.49335760e-01 -2.51974404e-01 2.00790897e-01 4.33636725e-01
1.30534136e+00 1.38105959e-01 4.55741256e-01 -3.52162540e-01
4.92439181e-01 -3.97681713e-01 2.28834808e-01 6.97585940e-01
1.80765137e-01 5.04093468e-01 5.14066756e-01 -2.38553420e-01
-7.55948305e-01 -1.07565022e+00 7.07152784e-01 9.31819022e-01
-1.56873658e-01 -8.24025929e-01 -9.37535167e-01 -8.26423168e-01
-3.67096990e-01 1.16253054e+00 -5.59690177e-01 -5.64710915e-01
-7.74150550e-01 -5.81529081e-01 8.83814573e-01 5.01858532e-01
3.40833396e-01 -1.33660257e+00 3.40131253e-01 3.72291237e-01
-6.55918121e-01 -1.13133883e+00 -6.21215224e-01 -7.41103962e-02
-1.09991980e+00 -4.37945455e-01 -8.37276459e-01 -1.08613944e+00
5.96436083e-01 -6.66221902e-02 1.52977753e+00 1.14878669e-01
3.39212388e-01 -4.52044755e-02 -4.95057195e-01 -6.71187341e-02
-9.20578182e-01 5.89750350e-01 -3.20805967e-01 -3.06427062e-01
1.06047299e-02 -6.75652742e-01 -2.93487251e-01 -8.49259645e-02
-6.54364288e-01 4.88708884e-01 8.58944595e-01 8.58850956e-01
4.83340263e-01 -6.14459753e-01 8.00368607e-01 -9.12912726e-01
1.07387757e+00 -2.60498494e-01 -3.23813051e-01 5.92572153e-01
-8.29563856e-01 4.40634310e-01 9.12502408e-01 -1.96142077e-01
-8.61076593e-01 -2.09277436e-01 -3.05448491e-02 2.28993651e-02
1.11985222e-01 8.76334906e-01 -1.68213993e-01 1.06230147e-01
8.19818795e-01 5.64594805e-01 -2.42634997e-01 -1.99703932e-01
8.76237154e-01 5.51010311e-01 4.23140913e-01 -5.60196757e-01
6.18464470e-01 -1.22810923e-01 6.49297386e-02 -5.31282902e-01
-3.93589407e-01 6.99954554e-02 -4.62306947e-01 9.94297341e-02
1.01242733e+00 -9.59131539e-01 -3.66536349e-01 3.73234004e-01
-1.66748703e+00 -6.16318226e-01 -9.31863114e-02 3.04798990e-01
-5.89125991e-01 3.77917349e-01 -9.05819237e-01 -4.04694498e-01
-9.42806721e-01 -9.01772141e-01 1.01501775e+00 -2.13014856e-01
-2.22378418e-01 -1.42119980e+00 1.11713886e-01 2.18246698e-01
4.78332400e-01 1.68759957e-01 1.43945968e+00 -5.99444628e-01
-5.00776529e-01 1.22869853e-02 -3.15690726e-01 1.26384243e-01
2.14678168e-01 -4.88705516e-01 -4.17543739e-01 -3.62783909e-01
-6.72615111e-01 -2.48312399e-01 9.50105608e-01 2.27718413e-01
7.32185721e-01 -4.40522432e-01 -3.10809493e-01 5.29973984e-01
1.35770571e+00 1.32225052e-01 8.75516772e-01 -1.46617457e-01
1.00368166e+00 4.41325635e-01 -1.00630321e-01 -2.22806975e-01
6.64996684e-01 4.22711849e-01 4.47687984e-01 -7.80371949e-02
-7.15490401e-01 -7.59116352e-01 6.56460464e-01 1.62399161e+00
-3.57146889e-01 -9.12150323e-01 -9.14909124e-01 4.04612690e-01
-1.92593396e+00 -5.29511094e-01 -2.34214470e-01 2.04699254e+00
8.53359222e-01 2.01619819e-01 -4.33523417e-01 -3.69540811e-01
9.20068920e-01 4.10323471e-01 1.10059921e-02 -7.07096040e-01
-2.21790954e-01 5.55120885e-01 6.08046055e-01 7.83752263e-01
-6.24794126e-01 1.47641242e+00 6.27813530e+00 8.96484673e-01
-7.85332620e-01 -2.54213680e-02 5.41458011e-01 4.86522883e-01
-8.85448277e-01 1.23059534e-01 -4.81035620e-01 7.98477381e-02
1.38518512e+00 -1.68747738e-01 9.68937755e-01 3.28762531e-01
-1.18997425e-01 6.48952365e-01 -1.14512050e+00 1.03357446e+00
1.57078758e-01 -1.76291049e+00 8.67833853e-01 1.90661252e-01
9.36601758e-01 2.01821744e-01 -2.94444263e-01 5.94193280e-01
8.04621339e-01 -1.40315998e+00 4.27796781e-01 5.31293035e-01
1.10389698e+00 -6.89381421e-01 7.65836120e-01 2.60150701e-01
-1.38826978e+00 4.23521847e-01 -6.33491158e-01 -2.67680943e-01
1.69962451e-01 4.22346383e-01 -7.94712126e-01 1.13791978e+00
8.09379891e-02 8.61219406e-01 -5.77185214e-01 2.30561614e-01
-8.03391457e-01 8.27672482e-01 1.25326097e-01 -6.42935038e-01
3.33025932e-01 -3.82966936e-01 4.68900979e-01 1.23659432e+00
6.39724672e-01 -1.50713816e-01 -7.40723312e-02 1.02840912e+00
-7.26700008e-01 3.19915324e-01 -1.08445728e+00 -3.97392750e-01
3.82702500e-01 1.00535822e+00 -4.39810932e-01 -3.97960573e-01
-2.82041460e-01 1.30236840e+00 8.36977780e-01 7.33075917e-01
-7.48238385e-01 -7.04455197e-01 1.61328822e-01 -1.51411816e-01
2.19143301e-01 -2.74054587e-01 -1.35371715e-01 -1.07868052e+00
-1.19290195e-01 -9.83779609e-01 2.49648869e-01 -8.59207749e-01
-1.16096485e+00 9.71663356e-01 -2.51568556e-01 -8.28610897e-01
-5.99075794e-01 -6.79017186e-01 -7.48865783e-01 1.16138327e+00
-1.39888835e+00 -1.54615569e+00 1.60844922e-01 4.48187202e-01
2.77177662e-01 -5.48798859e-01 1.38498855e+00 -6.08528107e-02
-1.61969736e-01 8.39554489e-01 4.84632775e-02 4.64103281e-01
4.66726810e-01 -1.21042776e+00 1.39828265e+00 8.48293185e-01
5.71025252e-01 8.17314088e-01 2.45671242e-01 -7.93253243e-01
-1.56508648e+00 -1.45048976e+00 1.56218100e+00 -3.96288246e-01
7.56325543e-01 -6.45169854e-01 -7.32095897e-01 8.25110555e-01
7.24175572e-01 -3.34321111e-01 4.80735153e-01 2.66111493e-01
-4.91251141e-01 1.20538466e-01 -6.18325293e-01 9.32137311e-01
1.53310215e+00 -7.12172866e-01 -7.56109953e-01 2.67671734e-01
1.38389349e+00 -2.71498710e-01 -9.09579754e-01 2.98020512e-01
2.18513295e-01 -4.76123720e-01 7.34662116e-01 -8.94723117e-01
5.16034424e-01 -6.21811114e-02 -4.30450410e-01 -1.82533681e+00
-7.35525787e-01 -1.09232271e+00 -3.01024675e-01 1.04005861e+00
1.02814901e+00 -8.51976931e-01 5.63749731e-01 -1.43924534e-01
-2.94912636e-01 -6.50639653e-01 -9.34678733e-01 -7.95156896e-01
4.85636503e-01 -3.61515731e-01 7.84518778e-01 9.11224127e-01
1.14979342e-01 1.19872475e+00 -2.66204000e-01 -2.89094061e-01
4.00979161e-01 -1.78947859e-02 4.73326862e-01 -1.13049889e+00
-4.39406693e-01 -5.19540727e-01 -2.68262386e-01 -1.23344576e+00
5.62926054e-01 -1.70834601e+00 -1.46132365e-01 -2.42229366e+00
1.97336093e-01 1.49208575e-01 -4.35516745e-01 5.50760627e-01
-3.20233852e-01 1.73675939e-01 2.89438546e-01 -3.28216478e-02
-4.31860477e-01 8.17070425e-01 1.57334006e+00 -4.56231505e-01
-1.05517190e-02 -2.00016692e-01 -9.68723834e-01 2.90447503e-01
9.66826856e-01 -4.21417117e-01 -6.07118964e-01 -7.84409165e-01
5.85061848e-01 1.93625063e-01 -1.13300299e-02 -7.10896611e-01
-3.66248265e-02 -9.10314769e-02 -1.13009345e-02 -4.20554906e-01
8.43778327e-02 -2.65638351e-01 5.55790029e-03 6.80624485e-01
-6.69384062e-01 3.43788058e-01 6.29623458e-02 7.15127289e-01
-1.03397287e-01 2.99698580e-02 3.00937712e-01 -3.17883730e-01
-3.60597283e-01 3.34928185e-01 -5.10049403e-01 3.22082430e-01
3.65578502e-01 1.73126906e-01 -5.97647667e-01 -9.50332880e-01
-5.60481489e-01 -7.37914070e-02 1.65629774e-01 6.16125524e-01
8.02594781e-01 -1.66115189e+00 -1.26993322e+00 7.28100315e-02
2.62021929e-01 -2.66739428e-01 -2.58110195e-01 4.70217168e-01
-3.32689941e-01 6.91042840e-01 -9.27797798e-03 -2.82197833e-01
-8.98584723e-01 4.22489703e-01 1.29295528e-01 -7.79290438e-01
-4.68977779e-01 7.45057762e-01 4.07747895e-01 -9.10767555e-01
-1.02292433e-01 -6.71909869e-01 -1.07493863e-01 -4.68073666e-01
1.94231778e-01 2.15166897e-01 2.29536191e-01 -5.17515302e-01
-3.00655305e-01 4.21499103e-01 -6.88015949e-03 -1.57185704e-01
9.76286590e-01 5.04701398e-02 -5.21031082e-01 1.82835266e-01
1.17814434e+00 -1.45760357e-01 -4.17948902e-01 -2.99807250e-01
-7.23999217e-02 2.87122339e-01 -1.11594692e-01 -1.02101564e+00
-8.05800438e-01 9.88160372e-01 1.91376418e-01 1.37049511e-01
8.97215724e-01 7.22559988e-02 1.00777590e+00 8.19565952e-01
3.64188462e-01 -7.81073332e-01 2.55842470e-02 1.12399352e+00
1.20551479e+00 -8.82675231e-01 -3.74122202e-01 -3.29262882e-01
-5.60321927e-01 9.61497426e-01 2.43697956e-01 -1.06675334e-01
3.45796794e-01 1.36769991e-02 -1.43796597e-02 2.26088390e-02
-1.00359738e+00 -1.37491643e-01 6.97223008e-01 9.75462139e-01
8.15916598e-01 5.38652062e-01 -4.81717110e-01 4.70640898e-01
-3.56548131e-01 -3.02165210e-01 4.27748054e-01 4.35553223e-01
-1.62867695e-01 -1.42977273e+00 1.10318691e-01 2.83365071e-01
-2.76528865e-01 -8.96487057e-01 -9.43245173e-01 6.75389588e-01
-3.71196032e-01 1.15115082e+00 -1.35013238e-01 -5.89747488e-01
2.34623507e-01 2.01222613e-01 5.29674768e-01 -7.83308446e-01
-6.65499687e-01 -6.83443397e-02 5.20045519e-01 -3.71431053e-01
-2.36965865e-01 -3.55504826e-02 -1.38634872e+00 -3.39917153e-01
-2.58355379e-01 1.65698946e-01 4.16350245e-01 6.01493716e-01
6.69656932e-01 8.78316700e-01 2.96357006e-01 -3.26555938e-01
-4.51782227e-01 -1.35611153e+00 -3.17202806e-01 3.45236480e-01
-4.55009798e-03 6.05090298e-02 8.24067742e-02 -2.43035611e-02]
|
[10.275025367736816, 8.365957260131836]
|
2b5b01a6-1a03-4566-8e94-f3714899124b
|
robust-cross-view-gait-identification-with
|
1811.10493
| null |
https://arxiv.org/abs/1811.10493v3
|
https://arxiv.org/pdf/1811.10493v3.pdf
|
Robust Cross-View Gait Recognition with Evidence: A Discriminant Gait GAN (DiGGAN) Approach
|
Gait as a biometric trait has attracted much attention in many security and privacy applications such as identity recognition and authentication, during the last few decades. Because of its nature as a long-distance biometric trait, gait can be easily collected and used to identify individuals non-intrusively through CCTV cameras. However, it is very difficult to develop robust automated gait recognition systems, since gait may be affected by many covariate factors such as clothing, walking speed, camera view angle etc. Out of them, large view angle changes has been deemed as the most challenging factor as it can alter the overall gait appearance substantially. Existing works on gait recognition are far from enough to provide satisfying performances because of such view changes. Furthermore, very few works have considered evidences -- the demonstrable information revealing the reliabilities of decisions, which are regarded as important demands in machine learning-based recognition/authentication applications. To address these issues, in this paper we propose a Discriminant Gait Generative Adversarial Network, namely DiGGAN, which can effectively extract view-invariant features for cross-view gait recognition; and more importantly, to transfer gait images to different views -- serving as evidences and showing how the decisions have been made. Quantitative experiments have been conducted on the two most popular cross-view gait datasets, the OU-MVLP and CASIA-B, where the proposed DiGGAN has outperformed state-of-the-art methods. Qualitative analysis has also been provided and demonstrates the proposed DiGGAN's capability in providing evidences.
|
['Yan Gao', 'Yu Guan', 'Thomas Ploetz', 'BingZhang Hu', 'Nicholas Lane', 'Yang Long']
|
2018-11-26
| null | null | null | null |
['gait-identification']
|
['computer-vision']
|
[ 4.25832011e-02 -6.03010595e-01 -1.62689552e-01 -1.19988203e-01
-1.57847837e-01 -5.32267392e-01 3.50258619e-01 -4.53266293e-01
-1.25690416e-01 7.38333523e-01 4.04805019e-02 1.00989550e-01
-5.11906072e-02 -8.49812508e-01 -2.28968576e-01 -1.01699519e+00
-1.50533214e-01 1.57636534e-02 -3.07101551e-02 -2.94088960e-01
3.10707986e-01 5.02191246e-01 -1.47673094e+00 -2.71969706e-01
6.21981502e-01 7.85453618e-01 -6.45938277e-01 3.31351638e-01
4.57222521e-01 8.94878879e-02 -5.66020191e-01 -9.92633462e-01
2.02579051e-01 -3.62496585e-01 -1.82812706e-01 1.68252677e-01
1.32585302e-01 -3.73865575e-01 -5.55157006e-01 1.01523077e+00
8.40944827e-01 -2.96191752e-01 6.66400790e-01 -1.35077453e+00
-8.10692728e-01 -3.92435640e-02 -9.75975335e-01 3.58619429e-02
8.05838704e-01 3.60152602e-01 6.11624658e-01 -6.51843309e-01
3.73064399e-01 1.10119319e+00 7.08157778e-01 6.59217775e-01
-9.03580964e-01 -8.11642885e-01 -3.79771560e-01 5.91497183e-01
-1.34567010e+00 -2.52158046e-01 1.24383903e+00 -3.37556303e-01
4.55779463e-01 2.88228512e-01 8.24018002e-01 1.44100535e+00
5.21493554e-01 6.53244257e-01 1.43658888e+00 -2.86920488e-01
-4.19393927e-02 -6.79416731e-02 -1.71633571e-01 7.18203127e-01
7.83757567e-01 1.27129361e-01 -3.92595202e-01 -1.44437216e-02
6.68551385e-01 3.49542238e-02 -5.42666018e-01 -3.27300519e-01
-8.98947656e-01 4.05672193e-01 6.65554628e-02 2.78866947e-01
-1.71634704e-01 -1.94124617e-02 5.33207357e-01 2.99151629e-01
2.07125112e-01 -1.74087346e-01 -4.40879688e-02 -4.05369967e-01
-6.14141166e-01 8.03314820e-02 6.17764771e-01 5.46819627e-01
3.45342636e-01 4.37994868e-01 1.82122707e-01 5.89640975e-01
3.73775721e-01 1.04167700e+00 5.84822536e-01 -2.33545139e-01
6.36170864e-01 6.80404425e-01 -1.89376231e-02 -1.82354188e+00
-1.76984653e-01 -1.94073617e-01 -1.27295160e+00 8.64828974e-02
4.61041749e-01 3.81394736e-02 -5.94014108e-01 1.57540822e+00
1.54524073e-01 1.68950960e-01 -6.37221262e-02 9.21348631e-01
7.72655964e-01 1.31525859e-01 -1.28895149e-01 -7.64544010e-02
1.60155785e+00 -2.22080946e-01 -5.99955857e-01 -1.62935540e-01
1.17175266e-01 -7.27107167e-01 9.13590968e-01 4.80811834e-01
-7.21654058e-01 -6.02745831e-01 -1.27897143e+00 4.77818042e-01
-4.32370931e-01 4.03836936e-01 5.13156295e-01 1.46097219e+00
-8.17608953e-01 3.93659502e-01 -7.90684462e-01 -6.03685915e-01
3.12420636e-01 4.05864626e-01 -7.38958001e-01 -4.16631579e-01
-1.33894718e+00 7.05014527e-01 6.56746700e-02 5.14514387e-01
-3.43247682e-01 2.46364102e-02 -8.16340327e-01 -1.11547478e-01
1.55100390e-01 -6.93717718e-01 2.82384545e-01 -3.56806338e-01
-1.30938852e+00 9.81738627e-01 -1.35902120e-02 -1.37228891e-01
7.62160778e-01 1.30934253e-01 -9.84907985e-01 1.51200980e-01
7.72752836e-02 -4.86293435e-02 1.16873622e+00 -9.58036900e-01
-8.96438062e-02 -9.83830690e-01 -2.17747346e-01 2.39892751e-02
-4.76492703e-01 -1.37393147e-01 -4.41829771e-01 -7.34278321e-01
1.74391061e-01 -1.11261404e+00 3.12669873e-01 -2.52751380e-01
-5.76717675e-01 1.14726305e-01 1.19280124e+00 -1.02292871e+00
1.19411635e+00 -2.04249096e+00 -1.04658501e-02 2.76433885e-01
2.89273970e-02 4.51571941e-01 2.98891366e-01 4.80582565e-01
2.21404172e-02 2.51882106e-01 -4.84987460e-02 -9.00204107e-02
3.53293866e-02 1.48595899e-01 2.22172812e-02 8.34856868e-01
-2.44864705e-03 8.28134358e-01 -5.49415171e-01 -3.09213847e-01
4.31458414e-01 6.71924174e-01 -5.56199998e-02 -5.73859029e-02
6.50238335e-01 5.87064981e-01 -5.87599576e-01 1.07235420e+00
9.23626304e-01 2.04553604e-01 1.21746682e-01 -4.30736333e-01
3.44177902e-01 -6.21493399e-01 -1.11572587e+00 1.14378536e+00
-1.75335575e-02 3.66462231e-01 -2.55673915e-01 -1.17673528e+00
1.13360858e+00 4.59694028e-01 5.77783465e-01 -6.04969680e-01
3.28776091e-01 1.37605891e-01 -1.17228150e-01 -8.25199723e-01
3.56145561e-01 6.40740916e-02 -3.67704928e-01 1.65724039e-01
-2.73105949e-01 1.98126405e-01 9.77271050e-02 -2.91552991e-01
8.08330059e-01 1.59053385e-01 4.53686297e-01 2.19533518e-01
8.94476235e-01 -4.60982531e-01 7.22543240e-01 1.90441415e-01
-5.14562249e-01 6.16411924e-01 3.08612049e-01 -3.61336350e-01
-8.74641240e-01 -9.82699454e-01 9.94505063e-02 5.14582694e-02
2.32566223e-01 -4.90734130e-02 -6.27772212e-01 -6.69282496e-01
1.26854137e-01 3.64309410e-03 -6.60999596e-01 -3.53375584e-01
-5.45911014e-01 -1.00506687e+00 1.18724883e+00 6.00206614e-01
1.25742650e+00 -7.28054941e-01 -6.26812279e-01 4.73587476e-02
-3.33088726e-01 -1.33898962e+00 -3.84824842e-01 -6.12712383e-01
-6.95557594e-01 -1.38665652e+00 -9.03293192e-01 -4.43580717e-01
6.51583195e-01 3.31181198e-01 6.30657434e-01 2.18832836e-01
-5.19243777e-01 5.01193285e-01 -3.62871081e-01 -9.55381617e-02
-8.45816731e-02 -1.55587345e-01 5.23592174e-01 5.20343721e-01
6.21637464e-01 -8.54615152e-01 -8.15921366e-01 7.06124306e-01
-6.68189645e-01 -3.88491571e-01 5.40653110e-01 9.62417066e-01
1.63343456e-02 2.71544844e-01 6.48367882e-01 -5.83265543e-01
6.14726841e-01 -1.76680133e-01 -2.74965584e-01 2.94025987e-01
-4.90920663e-01 -3.73303324e-01 6.05749965e-01 -1.54898822e-01
-8.20326388e-01 -3.68372351e-01 -1.64537057e-01 -3.19432765e-01
-3.33085239e-01 4.69290942e-01 -6.14206970e-01 -4.73944932e-01
1.86668888e-01 5.52055061e-01 1.68410599e-01 -1.01688571e-01
3.94734554e-02 6.73078954e-01 6.33741498e-01 -5.60764909e-01
1.24061131e+00 5.46691120e-01 3.53428870e-01 -9.91174400e-01
3.10447719e-02 -4.21236187e-01 -6.74799979e-01 -5.49162507e-01
9.27403688e-01 -7.18329072e-01 -1.14292634e+00 1.13134468e+00
-6.98630750e-01 5.87229729e-01 3.41609001e-01 3.90882641e-01
-3.16127777e-01 9.70558405e-01 -6.32959723e-01 -8.19283187e-01
-3.97115469e-01 -1.21893167e+00 7.36149430e-01 5.64247787e-01
-2.55850668e-04 -1.08948481e+00 -1.67591617e-01 6.34898663e-01
4.24879521e-01 7.78274596e-01 6.92306042e-01 -2.93474644e-01
-2.83534467e-01 -7.11544752e-01 -3.52858119e-02 1.40165284e-01
6.60561204e-01 8.24128985e-02 -1.07754612e+00 -6.15855157e-01
-6.05393350e-02 -1.34961471e-01 5.09253502e-01 2.60444731e-01
6.81459606e-01 -2.01593548e-01 -3.14122468e-01 6.11879647e-01
1.38199830e+00 3.97164345e-01 1.12700033e+00 3.56263518e-01
6.86726093e-01 4.22839165e-01 6.01115227e-01 4.70497608e-01
2.93312043e-01 8.69535863e-01 3.65567714e-01 -3.97258438e-02
1.26722634e-01 -2.84739971e-01 4.22501475e-01 9.81862485e-01
-7.30401874e-01 -2.59602219e-01 -5.56899250e-01 2.52289712e-01
-1.68018663e+00 -1.24277282e+00 -8.47879872e-02 2.24523664e+00
7.39750490e-02 -2.59409249e-02 2.71930367e-01 7.56780148e-01
7.33914495e-01 2.40439892e-01 -4.68798637e-01 -9.18876529e-02
-1.77407309e-01 1.49671882e-01 3.96820366e-01 -2.43938074e-01
-1.10084474e+00 5.53641558e-01 5.04398966e+00 5.81647694e-01
-1.35898113e+00 -1.21298410e-01 4.66108620e-01 4.98529524e-01
1.24738500e-01 -3.30348521e-01 -4.92695421e-01 6.90855145e-01
4.22420889e-01 -1.45638091e-02 6.32337332e-02 6.79107010e-01
3.11710477e-01 -6.95081893e-03 -5.69060445e-01 1.28661144e+00
4.56224084e-01 -5.89519620e-01 3.59555744e-02 5.94093621e-01
3.02903831e-01 -9.26882923e-01 4.15964901e-01 2.26272438e-02
-1.67248234e-01 -8.65956545e-01 8.52776989e-02 2.32106775e-01
1.05851996e+00 -8.79937172e-01 1.07403207e+00 6.22923262e-02
-1.58327568e+00 8.89955685e-02 -3.76849443e-01 8.60205144e-02
1.81669325e-01 5.00638783e-01 -5.67007720e-01 1.19123399e+00
5.88508666e-01 7.96617985e-01 -6.78267956e-01 8.31259429e-01
-3.26862156e-01 7.77507186e-01 -6.74054697e-02 -1.47337019e-02
-2.07322836e-01 -4.35790360e-01 6.81039870e-01 8.12372983e-01
7.01725066e-01 3.05951033e-02 -1.03345200e-01 5.95681310e-01
9.72070470e-02 3.02873813e-02 -9.90346193e-01 -2.56685372e-02
3.73536527e-01 1.12896192e+00 -8.07100415e-01 8.29476193e-02
-4.06745136e-01 1.08130503e+00 -3.38325411e-01 1.87672600e-01
-8.65894496e-01 -2.49477729e-01 6.91005647e-01 8.84931684e-02
2.91338027e-01 -3.17493051e-01 -1.10950910e-01 -1.72904217e+00
2.32916400e-01 -1.01978672e+00 3.13704401e-01 -4.68695045e-01
-1.26841891e+00 2.60757625e-01 -2.72059679e-01 -1.71495533e+00
-2.44536340e-01 -6.56648040e-01 -5.86217642e-01 7.14977860e-01
-1.19034410e+00 -1.68960989e+00 -4.40636367e-01 7.85826921e-01
1.97969988e-01 -4.80501324e-01 8.93624961e-01 6.78053737e-01
-7.41708696e-01 1.01456106e+00 -1.32597059e-01 4.83717710e-01
7.65941441e-01 -8.66240323e-01 4.42576736e-01 1.20863235e+00
-7.55412653e-02 7.15995908e-01 6.32563591e-01 -7.40975142e-01
-1.84356785e+00 -6.31951571e-01 4.51317519e-01 -2.62623250e-01
3.33588690e-01 -7.89210722e-02 -7.13235080e-01 4.09935325e-01
-1.28287658e-01 5.74827977e-02 7.26225078e-01 -3.02826345e-01
-2.95524120e-01 -2.49635875e-01 -1.39448345e+00 4.90143120e-01
9.94873941e-01 -4.26068008e-01 -3.62330019e-01 -4.15855497e-01
-1.92825347e-01 -2.17949972e-01 -9.31173503e-01 4.70869273e-01
1.02873611e+00 -1.22007549e+00 1.18816411e+00 -2.51172245e-01
1.39703095e-01 -3.54246944e-01 -2.04366133e-01 -1.21659160e+00
-5.67219481e-02 -2.43757486e-01 1.53205339e-02 1.70866656e+00
-2.64378399e-01 -1.12733889e+00 1.08640897e+00 4.69408005e-01
4.17234182e-01 -5.22126555e-01 -1.03945816e+00 -8.01149487e-01
-4.76534188e-01 -1.06017880e-01 6.34302497e-01 1.08656287e+00
-4.23358232e-02 -1.50878102e-01 -9.00493741e-01 2.41707370e-01
9.63721752e-01 -1.89442977e-01 1.11680698e+00 -1.17696023e+00
-1.85158551e-01 -1.61236078e-01 -1.34206355e+00 -4.95383590e-01
-1.64496526e-01 -4.38556284e-01 -4.91938382e-01 -1.10724795e+00
1.73000902e-01 -2.06190392e-01 -1.56480595e-01 2.62623161e-01
-2.03523755e-01 6.53710485e-01 1.25490382e-01 2.12621912e-01
1.67284310e-01 4.64426070e-01 1.31041956e+00 -3.43280494e-01
1.72026411e-01 2.68264025e-01 -5.32044947e-01 6.24760568e-01
7.12335467e-01 -2.49387436e-02 -3.86036873e-01 6.24605827e-02
-4.42248322e-02 2.96330631e-01 4.32686120e-01 -1.29981112e+00
3.84658650e-02 5.99191822e-02 6.30244315e-01 -4.53433603e-01
5.26585519e-01 -8.54748011e-01 3.08361858e-01 6.84415281e-01
5.31160355e-01 4.81289566e-01 5.54347858e-02 6.70481086e-01
-3.42959762e-01 1.94711521e-01 4.85148221e-01 -5.31221665e-02
-7.72345126e-01 2.91734576e-01 -3.17917138e-01 -9.36369076e-02
1.00505173e+00 -8.50295067e-01 -3.09055418e-01 -5.42904317e-01
-6.27765536e-01 -2.17489034e-01 6.16612673e-01 3.98807168e-01
8.51327419e-01 -1.40398586e+00 -6.20767653e-01 5.02421439e-01
1.37804568e-01 -6.21827960e-01 5.34754574e-01 8.33878517e-01
-4.98509079e-01 2.82933503e-01 -6.24701083e-01 -6.73130751e-01
-1.68392980e+00 3.28481138e-01 7.45760128e-02 -4.12507415e-01
-4.47494507e-01 2.64158994e-01 -3.80584449e-01 -2.32548907e-01
-2.00744241e-01 6.71784282e-02 -3.82588416e-01 8.62102285e-02
3.31612259e-01 4.79699165e-01 1.62662968e-01 -9.69433665e-01
-4.64825213e-01 1.30995440e+00 2.35903651e-01 6.98168129e-02
1.07249892e+00 -3.09506863e-01 9.94266942e-02 3.79527621e-02
9.08961117e-01 1.32490233e-01 -8.08537662e-01 8.92814323e-02
-2.51876533e-01 -8.10244918e-01 -5.75599611e-01 -4.92084861e-01
-1.24906397e+00 1.08173919e+00 8.75680864e-01 1.84785768e-01
1.15828109e+00 -7.83595800e-01 9.44114268e-01 1.29082811e-03
9.79119539e-01 -7.71651745e-01 2.27726102e-01 -2.88394410e-02
4.91457254e-01 -1.11976695e+00 1.09343335e-01 -5.23693025e-01
-4.54156846e-01 1.26235414e+00 5.41212201e-01 -5.85416611e-03
4.58521813e-01 1.74152583e-01 -1.90515409e-03 4.14359421e-02
3.11316526e-03 1.68852448e-01 3.57911676e-01 8.91064048e-01
2.00077355e-01 1.99431792e-01 -3.70282203e-01 4.58764195e-01
-3.17604184e-01 1.71826258e-01 4.62130606e-01 8.61266792e-01
1.21831335e-01 -1.24487484e+00 -7.51222372e-01 3.79936635e-01
-7.20758498e-01 5.03791928e-01 -2.72161186e-01 1.03842998e+00
7.06141815e-02 9.39551353e-01 -5.45636117e-01 -9.93643463e-01
2.95983344e-01 -2.51304079e-02 6.07298434e-01 9.17415023e-02
-2.52063096e-01 -2.54298300e-01 7.33358338e-02 -2.94618219e-01
-6.94564879e-01 -7.69769251e-01 -6.56485081e-01 -8.20188761e-01
-2.90491611e-01 -1.65414419e-02 4.26007211e-01 1.05003858e+00
8.97420123e-02 2.60023117e-01 7.17996299e-01 -8.51911306e-01
-4.84313667e-01 -7.00572312e-01 -7.91196883e-01 6.86316133e-01
-8.15697387e-02 -9.16486561e-01 -3.25501651e-01 3.99477668e-02]
|
[14.241421699523926, 1.4074974060058594]
|
01fba422-96e8-4c79-9356-23357585a138
|
a-topological-view-of-rule-learning-in
|
2110.02510
| null |
https://arxiv.org/abs/2110.02510v3
|
https://arxiv.org/pdf/2110.02510v3.pdf
|
Cycle Representation Learning for Inductive Relation Prediction
|
In recent years, algebraic topology and its modern development, the theory of persistent homology, has shown great potential in graph representation learning. In this paper, based on the mathematics of algebraic topology, we propose a novel solution for inductive relation prediction, an important learning task for knowledge graph completion. To predict the relation between two entities, one can use the existence of rules, namely a sequence of relations. Previous works view rules as paths and primarily focus on the searching of paths between entities. The space of rules is huge, and one has to sacrifice either efficiency or accuracy. In this paper, we consider rules as cycles and show that the space of cycles has a unique structure based on the mathematics of algebraic topology. By exploring the linear structure of the cycle space, we can improve the searching efficiency of rules. We propose to collect cycle bases that span the space of cycles. We build a novel GNN framework on the collected cycles to learn the representations of cycles, and to predict the existence/non-existence of a relation. Our method achieves state-of-the-art performance on benchmarks.
|
['Chao Chen', 'Zhi Tang', 'Liangcai Gao', 'Tengfei Ma', 'Zuoyu Yan']
|
2021-10-06
| null | null | null | null |
['inductive-relation-prediction']
|
['graphs']
|
[ 8.42515081e-02 2.18388006e-01 -4.99652147e-01 9.20001939e-02
1.71212003e-01 -6.83562934e-01 4.52863485e-01 2.22555578e-01
1.01268806e-01 3.11795831e-01 5.32536209e-02 -8.36722553e-01
-3.63672048e-01 -1.48523724e+00 -9.33112383e-01 -3.70025069e-01
-5.71664035e-01 4.02854741e-01 2.75854439e-01 -3.22417945e-01
1.40561312e-01 4.08025861e-01 -1.42090809e+00 5.94353303e-02
8.89444530e-01 5.91750920e-01 -7.73367360e-02 4.57587630e-01
-3.44035029e-01 8.95583093e-01 1.58376187e-01 -4.11238790e-01
2.40821078e-01 -5.09163082e-01 -1.20684266e+00 -2.24678248e-01
1.68105498e-01 2.55705684e-01 -9.23082352e-01 1.00482643e+00
-1.09959021e-01 1.76505167e-02 4.84257817e-01 -1.23596489e+00
-5.12905598e-01 7.35936642e-01 -2.37243861e-01 4.63903099e-02
5.21806359e-01 -3.51837903e-01 1.50287426e+00 -8.57482731e-01
8.49737227e-01 9.99486327e-01 7.32672632e-01 3.16085160e-01
-1.30790877e+00 -4.80789334e-01 9.23962668e-02 6.28292322e-01
-1.72100163e+00 -4.44768332e-02 7.71982074e-01 -5.71177304e-01
7.27453709e-01 4.32608217e-01 9.66408014e-01 3.49218518e-01
9.77639928e-02 4.43263203e-01 6.33832991e-01 -7.55072892e-01
-2.75880378e-02 -3.45667303e-01 4.07230526e-01 1.44390666e+00
4.23949361e-01 -2.11692862e-02 -3.66675258e-01 -1.26685277e-02
1.00495803e+00 -2.53598951e-02 -4.89733607e-01 -8.01780760e-01
-1.08010435e+00 1.05273879e+00 8.75258207e-01 4.34764594e-01
1.13936774e-01 6.44353554e-02 3.69383954e-02 3.05592537e-01
-2.10480411e-02 3.92430842e-01 -1.86972365e-01 4.43177998e-01
-3.33562404e-01 1.77700091e-02 1.43250299e+00 8.53801250e-01
1.22896171e+00 -4.82584924e-01 3.06068391e-01 4.73870814e-01
2.72065610e-01 4.65726368e-02 1.45011380e-01 -4.71853554e-01
3.14107358e-01 1.13550329e+00 -3.58186334e-01 -1.78395712e+00
-5.95813751e-01 -4.64978397e-01 -9.47673082e-01 -3.47636044e-01
5.52137613e-01 4.09421474e-01 -7.48537719e-01 1.80098057e+00
4.99090165e-01 5.37269354e-01 -4.26478535e-02 5.91810107e-01
4.80421722e-01 6.37801290e-01 -1.78117588e-01 -2.35991865e-01
1.11611187e+00 -6.66403949e-01 -3.38773966e-01 1.43513247e-01
1.18153632e+00 -3.06348681e-01 8.83892238e-01 1.44632325e-01
-6.87328100e-01 -2.31204748e-01 -1.03489459e+00 -2.38228440e-02
-5.00150442e-01 -3.51421721e-02 1.09326923e+00 3.44649404e-01
-8.57599616e-01 8.66441548e-01 -8.42831016e-01 -4.69173104e-01
2.70920008e-01 5.00839353e-01 -4.88932401e-01 -2.63132662e-01
-1.41208172e+00 8.59889090e-01 7.38442957e-01 8.28793459e-03
-4.71733749e-01 -5.28567493e-01 -1.14708400e+00 7.27542862e-02
7.94959724e-01 -6.21877968e-01 6.32956386e-01 -5.41057229e-01
-7.58122325e-01 7.77761459e-01 -6.75992146e-02 -5.40169954e-01
2.91913059e-02 6.72200173e-02 -5.86398005e-01 1.72568098e-01
-1.66578785e-01 2.37711612e-02 4.03893411e-01 -9.10571694e-01
-4.85064328e-01 -3.56970668e-01 5.25974810e-01 -7.82219097e-02
-4.69899565e-01 -4.85693604e-01 -6.19396389e-01 -3.09244007e-01
6.21950746e-01 -1.17914629e+00 -3.22850883e-01 -1.76194608e-01
-4.59256798e-01 -7.60799289e-01 5.90752482e-01 -4.79186743e-01
1.64660263e+00 -1.88373792e+00 2.01397449e-01 8.19573283e-01
6.36263013e-01 2.46451706e-01 9.70896930e-02 4.39382255e-01
-2.05708221e-01 4.77581203e-01 -1.39136910e-01 4.31977689e-01
-1.51097551e-01 4.35781002e-01 -5.65230072e-01 4.90017742e-01
5.94909899e-02 8.51996124e-01 -9.06591773e-01 -5.51278234e-01
-1.03440814e-01 2.60292560e-01 -5.99218488e-01 5.87890609e-05
-4.16463673e-01 6.57521784e-02 -5.85818350e-01 1.69948816e-01
5.77692389e-01 -6.12080038e-01 8.01050246e-01 -2.57914692e-01
8.28971267e-02 4.97062594e-01 -1.33890975e+00 1.56803095e+00
-3.22172910e-01 2.57624120e-01 -4.98024702e-01 -1.32170665e+00
1.07827628e+00 3.52489203e-02 4.30112541e-01 -4.09832150e-01
-2.31482714e-01 2.82977521e-01 1.06572181e-01 -3.94680917e-01
7.76136890e-02 1.69135079e-01 2.25107267e-01 3.06811988e-01
-4.87561636e-02 3.61782998e-01 4.99336421e-01 5.33325434e-01
1.27716482e+00 1.69041008e-01 3.49940240e-01 -2.65791923e-01
8.66762877e-01 -1.09152179e-02 7.36515880e-01 3.60839874e-01
5.53180754e-01 1.59933344e-02 8.75566721e-01 -8.61310720e-01
-8.93183768e-01 -1.02750301e+00 1.83997378e-01 8.40997338e-01
3.23857427e-01 -1.18127668e+00 -5.45317411e-01 -6.86627686e-01
-1.12262674e-01 2.20882952e-01 -6.85431719e-01 -4.72765237e-01
-1.01124763e+00 -3.36308897e-01 4.56466049e-01 2.36775324e-01
3.93939435e-01 -7.46119559e-01 -2.22835630e-01 1.00101419e-01
-3.14863563e-01 -1.13706803e+00 -4.79992539e-01 -2.66473681e-01
-9.64499235e-01 -1.70389378e+00 1.65986456e-02 -1.04933321e+00
7.33614147e-01 2.38547444e-01 1.19073904e+00 6.92552924e-01
-2.91969150e-01 4.02736336e-01 -2.73895025e-01 7.67876282e-02
-2.76497573e-01 2.76986659e-01 1.45945564e-01 1.48747027e-01
1.51110157e-01 -9.02530015e-01 -4.64963287e-01 3.45952988e-01
-7.94593990e-01 2.52518475e-01 4.31955040e-01 6.02841914e-01
9.09026384e-01 3.48667502e-01 3.20393413e-01 -9.28285599e-01
7.00239241e-02 -4.99110222e-01 -6.12878144e-01 6.08001530e-01
-7.12151587e-01 5.72569430e-01 7.18541324e-01 -3.22680235e-01
-4.10074651e-01 2.46796131e-01 3.59666079e-01 -4.61593032e-01
3.71240944e-01 1.17861426e+00 -3.16191375e-01 -3.27352732e-01
5.86686492e-01 1.94571689e-01 -8.52055028e-02 -4.38998163e-01
7.02857018e-01 -1.15132853e-02 4.48351800e-01 -6.08730793e-01
8.17574322e-01 4.50876653e-01 6.89651430e-01 -8.41896594e-01
-8.67519319e-01 -3.68141204e-01 -6.80270255e-01 -6.93726838e-02
4.50008005e-01 -6.28965437e-01 -9.32770908e-01 -1.67988509e-01
-1.10149598e+00 2.10576076e-02 8.10893718e-03 4.86878961e-01
-5.85346758e-01 6.35595798e-01 -2.89388567e-01 -5.10368705e-01
-1.77041665e-01 -5.14468908e-01 1.91567019e-01 1.86942182e-02
-1.17054127e-01 -1.00740433e+00 4.20499086e-01 6.22046329e-02
-7.51376227e-02 3.12555104e-01 1.46972716e+00 -7.06586659e-01
-1.04882371e+00 -1.04939044e-01 -4.58119325e-02 -2.28379428e-01
-1.82700399e-02 -1.37917355e-01 -1.46832317e-01 -1.19398385e-01
-2.90557355e-01 1.93998143e-02 1.14643371e+00 -2.10329741e-01
1.24295580e+00 -5.80017626e-01 -6.93086982e-01 7.50083566e-01
1.46862984e+00 -2.12015629e-01 8.10616016e-01 1.06098503e-01
1.15097797e+00 6.40873432e-01 3.04073274e-01 9.00664926e-02
6.81346238e-01 6.17209017e-01 3.32352579e-01 3.34244132e-01
-1.10521376e-01 -7.65523672e-01 1.98316723e-01 1.07070518e+00
-4.73352253e-01 3.43219072e-01 -1.21670544e+00 5.31879902e-01
-1.99357843e+00 -9.09134209e-01 -3.59714359e-01 2.38291478e+00
8.44520688e-01 4.89481278e-02 2.04559430e-01 3.72469604e-01
8.67033422e-01 -1.73931345e-01 -2.25915536e-01 -2.18613401e-01
-6.33795979e-03 3.88445765e-01 4.48466778e-01 5.93537807e-01
-9.86711860e-01 9.80812728e-01 5.48271704e+00 7.31959045e-01
-9.60473239e-01 -2.64092088e-01 1.14686459e-01 5.68150043e-01
-4.71571982e-01 4.97851551e-01 -6.50834322e-01 9.87643972e-02
7.67796934e-01 -4.96401757e-01 6.72115564e-01 7.21343398e-01
-3.80491912e-01 2.50696480e-01 -1.27833462e+00 8.83221269e-01
-1.03357779e-02 -1.56513286e+00 2.61760712e-01 2.79405206e-01
6.15150988e-01 -3.23370874e-01 -2.99806505e-01 2.95675814e-01
2.72501230e-01 -1.29896867e+00 2.21696436e-01 7.02446282e-01
6.93071306e-01 -8.90760720e-01 2.83162117e-01 4.13676828e-01
-1.76381290e+00 -8.89981389e-02 -4.41306472e-01 -2.41538242e-01
-2.82414913e-01 5.97763658e-01 -8.37870717e-01 1.07858419e+00
2.47495338e-01 9.77425814e-01 -5.62437534e-01 1.10149026e+00
-4.80206162e-01 7.17936397e-01 -5.00747204e-01 6.38790205e-02
-2.75775880e-01 -4.98146296e-01 3.11745822e-01 1.03910196e+00
3.34838867e-01 5.61672449e-01 5.30049324e-01 7.93610573e-01
-3.93285692e-01 2.57924974e-01 -1.08039284e+00 -2.25580171e-01
4.00633395e-01 1.09857643e+00 -8.32076848e-01 -2.55738292e-02
-4.91526008e-01 4.10666317e-01 7.00338483e-01 8.71495530e-02
-5.38080692e-01 -4.50829566e-01 4.31206405e-01 2.23078504e-01
6.05799198e-01 -5.27994573e-01 -2.21556410e-01 -1.24853516e+00
1.25980198e-01 -7.18940258e-01 7.21491516e-01 -1.59066215e-01
-9.22508419e-01 3.03268522e-01 -3.29412669e-01 -9.90352571e-01
7.49169439e-02 -5.72227836e-01 -7.33239889e-01 4.83340800e-01
-1.18112159e+00 -1.15408468e+00 -1.54347569e-01 9.09819305e-01
-1.89506859e-01 1.70763388e-01 1.04443324e+00 1.92602396e-01
-4.33395416e-01 4.78094876e-01 -2.45627940e-01 5.13298750e-01
7.17672184e-02 -1.10391140e+00 2.48811916e-01 6.75406754e-01
7.33049512e-01 7.75596678e-01 3.92299116e-01 -6.01258576e-01
-1.88802505e+00 -1.07183754e+00 1.10959196e+00 -2.77089745e-01
9.21144664e-01 -2.68610984e-01 -1.17537379e+00 8.31407905e-01
-5.16536951e-01 1.54353291e-01 6.14451408e-01 5.76760769e-01
-8.62323463e-01 -2.59317696e-01 -5.14112890e-01 7.28782117e-01
1.51478958e+00 -6.58989668e-01 -5.76984644e-01 3.53612602e-01
9.07396197e-01 -1.21572226e-01 -8.84352446e-01 5.92718363e-01
5.09694219e-01 -5.57427466e-01 1.08060813e+00 -9.32752013e-01
2.96991378e-01 -4.74029541e-01 -1.02308681e-02 -9.32253301e-01
-3.74678820e-01 -7.32104540e-01 -8.07953715e-01 9.02544141e-01
5.07585526e-01 -5.14779806e-01 9.39764798e-01 1.48903981e-01
1.84541658e-01 -1.11253440e+00 -8.26298296e-01 -8.19671571e-01
7.03024566e-02 -3.43031526e-01 6.41952693e-01 1.13273549e+00
4.54575151e-01 6.97288990e-01 -3.69471073e-01 3.07329684e-01
6.05512440e-01 5.88494837e-01 8.31831038e-01 -1.70655787e+00
-2.31847048e-01 -3.13920975e-01 -8.49054337e-01 -9.01767075e-01
2.27031127e-01 -1.53482807e+00 -3.55868250e-01 -1.65919125e+00
1.32102415e-01 -6.25704110e-01 -3.91867518e-01 4.93007958e-01
7.75524303e-02 -1.20629303e-01 2.60649681e-01 4.03181732e-01
-6.88892126e-01 4.55719650e-01 1.13050222e+00 4.98064123e-02
-2.50227749e-01 -4.02315445e-02 -6.28557086e-01 7.84821272e-01
7.64883995e-01 -2.61863470e-01 -2.66533345e-01 -1.34910882e-01
8.32200706e-01 -1.93828605e-02 3.36749256e-01 -9.54363286e-01
5.36198378e-01 -1.01775229e-01 4.19203155e-02 -3.85861486e-01
7.89272785e-02 -8.35369527e-01 2.29869649e-01 6.92880630e-01
-2.17679530e-01 -2.19805449e-01 -2.69014835e-01 9.58481729e-01
-3.85152772e-02 -3.76373194e-02 4.05920655e-01 4.33916934e-02
-7.38784730e-01 6.53656960e-01 2.04894572e-01 -1.41396252e-02
9.42419291e-01 4.77793179e-02 -1.99070454e-01 -1.90293625e-01
-9.86824572e-01 3.69839132e-01 3.51423055e-01 2.21892089e-01
6.95711732e-01 -1.59804046e+00 -3.94036740e-01 6.80198595e-02
1.07257083e-01 6.29871245e-03 -3.95749956e-02 8.71892333e-01
-5.11415601e-01 4.82340515e-01 8.93614218e-02 -3.95357907e-01
-1.21717334e+00 8.95022690e-01 3.21433008e-01 -6.34874344e-01
-7.82122791e-01 4.94343311e-01 2.59912819e-01 -3.87176245e-01
1.57754093e-01 -1.49548143e-01 -4.90633458e-01 -1.82390288e-01
3.41696829e-01 3.97618771e-01 -8.51553008e-02 -4.61150467e-01
-2.95935422e-01 8.81140292e-01 -1.14347085e-01 3.66218597e-01
1.18890488e+00 1.67392805e-01 -7.19784498e-01 3.56799841e-01
1.16827643e+00 -3.89661491e-02 -5.83555222e-01 -6.63152158e-01
4.99717057e-01 -3.90616566e-01 -3.23761195e-01 -2.71258444e-01
-9.04879034e-01 6.36948466e-01 8.68818015e-02 4.93651241e-01
1.03532279e+00 4.12267208e-01 6.88556075e-01 7.97655344e-01
4.66731519e-01 -7.54920661e-01 2.26216968e-02 9.49414551e-01
5.66395104e-01 -7.29880273e-01 -7.85293654e-02 -1.10943663e+00
-1.53645128e-01 1.35718441e+00 3.32484186e-01 -2.33396709e-01
9.34459507e-01 -4.54857111e-01 -7.45239198e-01 -3.56488883e-01
-7.11201131e-01 -5.55035710e-01 5.19384682e-01 3.84480685e-01
2.24187642e-01 3.19257647e-01 -6.48516476e-01 5.95804155e-01
-4.31711465e-01 2.33987775e-02 4.51016098e-01 6.67833924e-01
-7.10102260e-01 -1.35488188e+00 -1.80828497e-01 4.50697184e-01
-1.71337098e-01 -1.14817470e-01 -5.65003932e-01 6.99260712e-01
2.18280539e-01 8.07290673e-01 -1.14966154e-01 -6.45283163e-01
3.36256683e-01 1.37897089e-01 6.97310805e-01 -6.37645900e-01
3.08285832e-01 -5.70027471e-01 1.09970912e-01 -5.30001462e-01
-3.63320887e-01 -4.10829812e-01 -1.45894504e+00 -6.13652408e-01
-4.98481810e-01 4.42271024e-01 3.92410427e-01 1.07935941e+00
4.14183527e-01 1.47429675e-01 8.64320159e-01 1.94057834e-03
-3.05586159e-01 -5.54241478e-01 -5.93534052e-01 2.84770250e-01
6.78534433e-02 -7.06638753e-01 -2.33797282e-01 -1.41397002e-03]
|
[8.64362621307373, 7.701328277587891]
|
aaac113d-cfc3-4e02-9973-9a5956c97bbf
|
190807654
|
1908.07654
| null |
https://arxiv.org/abs/1908.07654v2
|
https://arxiv.org/pdf/1908.07654v2.pdf
|
FusionNet: Incorporating Shape and Texture for Abnormality Detection in 3D Abdominal CT Scans
|
Automatic abnormality detection in abdominal CT scans can help doctors improve the accuracy and efficiency in diagnosis. In this paper we aim at detecting pancreatic ductal adenocarcinoma (PDAC), the most common pancreatic cancer. Taking the fact that the existence of tumor can affect both the shape and the texture of pancreas, we design a system to extract the shape and texture feature at the same time for detecting PDAC. In this paper we propose a two-stage method for this 3D classification task. First, we segment the pancreas into a binary mask. Second, a FusionNet is proposed to take both the binary mask and CT image as input and perform a binary classification. The optimal architecture of the FusionNet is obtained by searching a pre-defined functional space. We show that the classification results using either shape or texture information are complementary, and by fusing them with the optimized architecture, the performance improves by a large margin. Our method achieves a specificity of 97% and a sensitivity of 92% on 200 normal scans and 136 scans with PDAC.
|
['Fengze Liu', 'Yuyin Zhou', 'Elliot Fishman', 'Alan Yuille']
|
2019-08-21
| null | null | null | null |
['3d-classification']
|
['computer-vision']
|
[ 9.96664613e-02 6.61010249e-03 -2.86191255e-01 -3.07103604e-01
-5.68189204e-01 -5.96323073e-01 1.74472839e-01 3.97569478e-01
-2.81356126e-01 1.31110877e-01 3.34134288e-02 -4.21461582e-01
-4.64430116e-02 -8.08056116e-01 -3.62006456e-01 -1.08297265e+00
-7.48867840e-02 6.03541315e-01 5.08627057e-01 4.54041988e-01
3.80861908e-02 7.84194529e-01 -1.02670753e+00 3.49743783e-01
1.10134578e+00 1.16604018e+00 2.54606724e-01 6.51558816e-01
-3.37498009e-01 2.58846760e-01 -3.09860885e-01 -2.79258370e-01
4.83501881e-01 -7.10717738e-01 -5.10735631e-01 3.71526271e-01
-8.44493210e-02 4.01663221e-02 -7.63380900e-02 1.37025630e+00
2.80731648e-01 -4.22544837e-01 8.30295444e-01 -6.87408745e-01
9.03870910e-02 5.89388132e-01 -4.10712302e-01 4.74575572e-02
1.04598671e-01 1.65281698e-01 4.11982208e-01 -4.79207724e-01
3.83923501e-01 7.38274038e-01 6.54868186e-01 2.97758371e-01
-9.34283793e-01 -2.78727174e-01 1.42432768e-02 5.26169874e-02
-1.20950794e+00 1.94314439e-02 7.09439874e-01 -3.29520971e-01
3.16695064e-01 6.24316752e-01 1.15623784e+00 4.44440186e-01
6.60424173e-01 6.39813185e-01 1.16051304e+00 -6.63064599e-01
1.01747938e-01 7.79303238e-02 -7.20300851e-03 1.07985127e+00
7.42316365e-01 5.05319908e-02 3.34588051e-01 -1.41053259e-01
9.19754028e-01 3.67676944e-01 -4.21098918e-01 -5.70747972e-01
-1.26212370e+00 7.34563768e-01 4.09523368e-01 6.22342169e-01
-4.52266872e-01 -2.09498629e-01 1.18104622e-01 2.12597519e-01
-2.19950274e-01 4.17228729e-01 -1.85253918e-01 4.02784050e-01
-4.65537727e-01 -2.77292341e-01 1.21404338e+00 5.54467559e-01
9.64761227e-02 -3.89586806e-01 2.77279690e-02 7.00831056e-01
5.75541615e-01 7.31780291e-01 7.91033506e-01 -4.69768167e-01
1.13637261e-01 9.22664523e-01 -2.39452064e-01 -6.30880892e-01
-6.75078154e-01 -3.59701186e-01 -7.37781525e-01 3.45078588e-01
7.54009962e-01 -1.13557063e-01 -1.23267567e+00 1.04611075e+00
4.10190105e-01 -2.29326665e-01 1.27517372e-01 1.10045362e+00
8.28934312e-01 3.32230628e-01 1.72881693e-01 -2.88766682e-01
1.87775993e+00 -9.54643965e-01 -5.31791866e-01 1.75275430e-01
7.80424535e-01 -8.59796643e-01 4.41327840e-01 4.73035395e-01
-8.64029050e-01 4.86053005e-02 -1.14267361e+00 5.09628057e-01
-1.46216318e-01 4.71104562e-01 4.81536239e-01 7.99510360e-01
-5.39094210e-01 2.81293333e-01 -1.11589599e+00 -6.31847322e-01
1.07558720e-01 5.76356292e-01 -4.05140996e-01 -1.58661991e-01
-5.41161656e-01 1.09093392e+00 4.34033990e-01 -1.68329105e-01
-3.41143876e-01 -2.37717181e-01 -6.93658769e-01 7.68931210e-02
2.60294706e-01 -6.59084976e-01 1.00674844e+00 -1.16447628e+00
-1.57507491e+00 8.47370684e-01 1.54372111e-01 -2.78635442e-01
5.99076986e-01 4.84586984e-01 -4.58121151e-01 3.89739931e-01
-2.27800071e-01 3.68301839e-01 5.15676796e-01 -1.14154387e+00
-1.04477131e+00 -5.97225010e-01 -3.92420113e-01 2.74178356e-01
3.70294228e-02 -3.28575790e-01 -6.60163879e-01 -7.49684393e-01
7.58392930e-01 -1.05008841e+00 -5.28496981e-01 1.84639767e-01
-2.66431212e-01 6.82800710e-02 6.99183702e-01 -7.82433271e-01
8.85968626e-01 -2.19586730e+00 -3.40341739e-02 8.44506264e-01
2.96315044e-01 2.54459590e-01 1.43982798e-01 -4.28369224e-01
8.24344158e-02 -1.00515671e-01 -2.33634502e-01 2.94234395e-01
-3.17389488e-01 4.09081936e-01 4.90138680e-01 7.37054288e-01
8.53549540e-02 8.17857027e-01 -5.09678304e-01 -1.00191057e+00
2.93645203e-01 1.78009272e-01 -3.94146085e-01 1.28249004e-01
1.66963432e-02 3.27469230e-01 -6.79642558e-01 9.74778235e-01
6.62860990e-01 -2.67273545e-01 7.85041213e-01 -2.89280474e-01
-1.16073251e-01 -1.49277791e-01 -1.22654557e+00 1.19479072e+00
-9.26725343e-02 1.20414272e-01 2.45261043e-01 -9.68144119e-01
1.15209162e+00 4.87496704e-01 9.03778136e-01 -7.37966239e-01
3.60224724e-01 6.10001326e-01 4.55355495e-01 -7.26060271e-01
-2.84319758e-01 -2.69993275e-01 6.78279400e-02 2.23797157e-01
-2.14095175e-01 5.00611477e-02 1.23984866e-01 -4.92704690e-01
1.09160340e+00 -3.54453415e-01 7.46021986e-01 -5.07743597e-01
7.83431411e-01 4.07058507e-01 5.39405346e-01 5.06650031e-01
-2.12915868e-01 4.66406494e-01 6.34691477e-01 -5.53321779e-01
-8.03437173e-01 -9.42278564e-01 -4.03888077e-01 3.24408382e-01
4.69989657e-01 2.45845571e-01 -5.39682984e-01 -1.13500154e+00
2.58646697e-01 2.21163094e-01 -5.46742141e-01 -8.14322680e-02
-8.62455964e-01 -9.63745236e-01 2.28255376e-01 4.49592739e-01
5.36575556e-01 -6.70531034e-01 -7.74010599e-01 2.80161023e-01
7.13673085e-02 -6.98602796e-01 -3.98505419e-01 5.58691084e-01
-1.28593397e+00 -1.35591245e+00 -8.63246024e-01 -1.26936543e+00
1.11536801e+00 1.54850110e-01 6.67689979e-01 3.84042382e-01
-3.47725660e-01 8.06734189e-02 -5.20836115e-01 -1.81690603e-01
-6.48986697e-01 -5.98597750e-02 -4.04128402e-01 -2.59906016e-02
2.00978622e-01 -1.43215552e-01 -6.57387674e-01 4.01673377e-01
-7.77332544e-01 1.63831748e-02 1.21908295e+00 6.77268922e-01
6.95483744e-01 2.40593284e-01 1.41102478e-01 -6.28574669e-01
1.77164450e-01 -3.19490731e-01 -7.25515127e-01 3.11991990e-01
-7.07785666e-01 2.42065176e-01 4.58774716e-01 -6.20083869e-01
-9.17436957e-01 6.49599671e-01 -9.93234739e-02 -1.91587865e-01
-2.52516806e-01 4.11181897e-01 -2.02894315e-01 -4.36496526e-01
3.55913669e-01 1.72002092e-01 4.28164810e-01 -5.80771983e-01
-1.72277853e-01 4.88668710e-01 4.63954717e-01 -1.61022291e-01
3.34776074e-01 4.16358203e-01 2.86986798e-01 -4.23645347e-01
-2.52593458e-01 -5.89659870e-01 -5.90417743e-01 -2.09485427e-01
9.87392843e-01 -3.92777950e-01 -5.15502393e-01 1.95566192e-01
-6.83493018e-01 -7.04879388e-02 -2.34626874e-01 1.09253633e+00
-1.76558822e-01 4.05075133e-01 -5.32497942e-01 -3.36190820e-01
-4.58486617e-01 -1.45566952e+00 6.04002833e-01 3.39926511e-01
1.42364085e-01 -1.04274094e+00 -4.30832729e-02 -1.46484405e-01
2.71292925e-01 3.30180675e-01 1.27328467e+00 -1.11767375e+00
-5.16348004e-01 -5.21327615e-01 -2.70097047e-01 -1.05853923e-01
1.67175218e-01 -9.36862156e-02 -2.79516041e-01 -3.52850929e-02
3.10060292e-01 5.21603942e-01 8.86535525e-01 6.34591162e-01
9.09752190e-01 -1.15130864e-01 -7.32337654e-01 8.74375343e-01
1.68485522e+00 8.73820603e-01 4.60489124e-01 3.59308213e-01
4.03550595e-01 2.44145885e-01 4.31461602e-01 2.88352609e-01
2.79757660e-02 3.77802998e-01 6.19826853e-01 -3.58054847e-01
-2.83482462e-01 1.91687182e-01 3.08855233e-05 6.26883566e-01
-8.38592798e-02 -1.87900171e-01 -1.00917494e+00 3.23950022e-01
-1.52847111e+00 -3.77402961e-01 -3.10738117e-01 2.29105139e+00
7.51089513e-01 5.59615642e-02 -4.59575802e-02 3.57520990e-02
8.24480355e-01 -5.30534685e-01 -2.54805446e-01 -1.62513331e-01
5.05110882e-02 4.91716303e-02 8.06601644e-01 4.38749760e-01
-1.30369020e+00 1.25189632e-01 6.49637318e+00 3.60121101e-01
-1.63195682e+00 -3.40568990e-01 5.22025228e-01 1.80084690e-01
-1.64124817e-01 -1.18782610e-01 -4.62764144e-01 4.79998380e-01
3.65223885e-01 1.91313207e-01 3.86440344e-02 7.18621612e-01
-3.13561380e-01 -3.57931882e-01 -1.01138413e+00 8.60121191e-01
9.07774791e-02 -9.86968935e-01 -1.04463831e-01 2.72645205e-01
5.07232010e-01 -2.89382398e-01 -2.64221847e-01 4.52621244e-02
8.90812725e-02 -8.69080186e-01 4.66117889e-01 6.12134337e-01
6.25077248e-01 -3.92285347e-01 1.27397525e+00 2.47789562e-01
-1.11740625e+00 -2.31889952e-02 1.17994420e-01 5.38823307e-01
-1.79124773e-01 4.93255585e-01 -1.04324615e+00 3.87468308e-01
3.40118289e-01 2.84889698e-01 -5.14886498e-01 1.70855010e+00
-1.07440236e-03 3.82334739e-01 -6.25057220e-01 -2.06638709e-01
5.83116449e-02 -3.82713169e-01 7.74671137e-01 1.32554567e+00
6.04567826e-01 3.57256085e-01 3.92443866e-01 5.14606833e-01
2.34570026e-01 5.03214478e-01 -1.96943760e-01 2.09173605e-01
1.57443151e-01 1.27935386e+00 -1.27110088e+00 -5.18846214e-01
-4.68121231e-01 7.10244119e-01 -3.86630595e-01 -8.72328654e-02
-8.07926536e-01 -1.99881002e-01 3.95563926e-04 -1.35800973e-01
4.43819702e-01 2.75638700e-01 -6.38664186e-01 -1.07042205e+00
1.87546480e-02 -7.13632166e-01 7.13124037e-01 4.50642593e-03
-9.95566308e-01 4.66519475e-01 -2.93469727e-01 -1.40791643e+00
-7.52586648e-02 -9.52245831e-01 -7.07285941e-01 6.69343591e-01
-1.31866133e+00 -1.05980134e+00 -5.37107050e-01 4.33812737e-01
2.29179695e-01 1.21812839e-02 8.59104276e-01 1.40935406e-01
-4.61402953e-01 3.51227611e-01 1.52066499e-01 2.60883868e-01
6.10370219e-01 -1.47283804e+00 -5.00413656e-01 5.39338648e-01
-4.36123043e-01 1.06663227e-01 4.15169775e-01 -7.25674510e-01
-1.58161628e+00 -8.49443316e-01 6.11435771e-01 1.07901558e-01
2.38052994e-01 1.89137861e-01 -7.17240393e-01 4.80643243e-01
-1.68849304e-01 1.59827307e-01 6.76601708e-01 -4.55046445e-01
5.53512666e-03 -1.54957116e-01 -1.60729718e+00 4.16520447e-01
3.75632226e-01 3.55262578e-01 -6.47377312e-01 1.52480736e-01
2.85027653e-01 -7.08425820e-01 -1.27012837e+00 5.97381592e-01
9.04722929e-01 -6.73279762e-01 7.81159759e-01 -1.07250370e-01
2.94451066e-03 -3.31073314e-01 -1.32073775e-01 -1.14054692e+00
-4.52073693e-01 5.25884591e-02 2.02619225e-01 5.04125416e-01
4.68291909e-01 -8.36759031e-01 8.20857108e-01 3.63527596e-01
-2.40000471e-01 -8.63284707e-01 -7.05532908e-01 -5.64728975e-01
-1.18734604e-02 1.59938365e-01 5.10403514e-01 8.32556367e-01
1.26291811e-01 -2.56523401e-01 5.07389843e-01 3.20516795e-01
5.45084596e-01 4.23747957e-01 2.83531874e-01 -1.49724507e+00
-4.95067202e-02 -7.62790680e-01 -3.69510055e-01 -7.20583558e-01
-6.00187659e-01 -9.55432892e-01 1.93532594e-02 -1.59831738e+00
3.35647851e-01 -7.24686682e-01 -2.81014860e-01 4.63505387e-01
-6.98819384e-02 1.35089234e-01 1.73016489e-01 4.17213798e-01
-3.37233841e-01 -1.00918859e-01 1.56148446e+00 1.94512475e-02
-5.05219102e-01 2.63958633e-01 -5.57905018e-01 8.38921964e-01
8.36822629e-01 -4.45020646e-01 2.51336396e-01 -6.14266321e-02
-4.99495149e-01 5.22834182e-01 2.53731608e-01 -9.43872571e-01
2.48084038e-01 -2.92880207e-01 8.55067015e-01 -5.51218629e-01
9.18817893e-03 -1.38817596e+00 3.31583679e-01 1.35369241e+00
-6.13849349e-02 -1.89193428e-01 -6.17120508e-03 3.36657465e-01
-1.49092987e-01 -6.39319420e-01 1.04348314e+00 -2.71379173e-01
-4.53564852e-01 1.38358464e-02 -5.07414758e-01 -6.01162612e-01
1.29949188e+00 -2.50403583e-01 -1.87690228e-01 3.63056734e-02
-9.56281781e-01 1.73650518e-01 5.36594987e-01 -5.14320172e-02
4.48644131e-01 -1.36213255e+00 -6.26510680e-01 6.22925818e-01
5.78886271e-02 7.15324804e-02 -5.72949052e-02 1.46313691e+00
-9.84498084e-01 5.22200465e-01 -3.13954055e-01 -8.35896730e-01
-1.49056411e+00 4.42187250e-01 6.58761322e-01 -5.48174202e-01
-8.50978434e-01 4.58741605e-01 9.98355821e-02 -7.63757750e-02
2.24504575e-01 -7.71614373e-01 -3.49906862e-01 -1.98024496e-01
3.39365363e-01 6.03345595e-02 1.76919252e-01 -3.48529041e-01
-3.38809371e-01 6.13474131e-01 5.83198816e-02 1.37534231e-01
1.01985276e+00 9.69135016e-02 -1.69660226e-01 -1.87076345e-01
1.03600454e+00 1.77351922e-01 -8.80695462e-01 -1.98673427e-01
-3.65268029e-02 -4.01791275e-01 2.17222854e-01 -1.04913163e+00
-1.41456664e+00 2.42670596e-01 9.31581914e-01 3.75742942e-01
1.36066175e+00 1.70555517e-01 6.46561027e-01 6.15582466e-02
3.64023484e-02 -5.66290617e-01 -2.24879995e-01 2.18455762e-01
5.38904130e-01 -1.18195391e+00 1.94427036e-02 -5.98270774e-01
-5.75043261e-01 1.59615219e+00 4.22608912e-01 -2.70264864e-01
6.08455300e-01 5.74020028e-01 3.27869624e-01 -1.51179448e-01
-5.34398198e-01 -2.49229208e-01 3.72348458e-01 4.85240191e-01
2.12870523e-01 4.93962437e-01 -7.25156069e-01 7.97336698e-01
1.80677891e-01 -2.02863932e-01 1.74155205e-01 1.07241488e+00
-9.89281714e-01 -1.10661781e+00 -9.50820208e-01 7.43354857e-01
-6.14747047e-01 3.46974880e-01 -2.28390038e-01 9.44787860e-01
2.50530660e-01 3.97074997e-01 2.21675426e-01 -1.31100252e-01
2.70142406e-01 9.20521095e-02 6.23503983e-01 -1.73489720e-01
-6.69781387e-01 7.31692612e-01 3.02696577e-03 -2.89248645e-01
-8.68643224e-02 -7.53152549e-01 -1.63628590e+00 4.92904373e-02
-4.25500572e-01 1.96793467e-01 9.75790560e-01 7.86576211e-01
-1.70386463e-01 5.57201684e-01 6.01354003e-01 -3.51060420e-01
-7.02962458e-01 -7.85066247e-01 -6.72769964e-01 3.84013385e-01
2.84137398e-01 -5.13858438e-01 -2.79157221e-01 9.30747986e-02]
|
[14.750021934509277, -2.6617166996002197]
|
1ac14fd6-30f4-457f-8473-b3cbb63b8c56
|
image-morphing-with-perceptual-constraints
|
2004.14071
| null |
https://arxiv.org/abs/2004.14071v1
|
https://arxiv.org/pdf/2004.14071v1.pdf
|
Image Morphing with Perceptual Constraints and STN Alignment
|
In image morphing, a sequence of plausible frames are synthesized and composited together to form a smooth transformation between given instances. Intermediates must remain faithful to the input, stand on their own as members of the set, and maintain a well-paced visual transition from one to the next. In this paper, we propose a conditional GAN morphing framework operating on a pair of input images. The network is trained to synthesize frames corresponding to temporal samples along the transformation, and learns a proper shape prior that enhances the plausibility of intermediate frames. While individual frame plausibility is boosted by the adversarial setup, a special training protocol producing sequences of frames, combined with a perceptual similarity loss, promote smooth transformation over time. Explicit stating of correspondences is replaced with a grid-based freeform deformation spatial transformer that predicts the geometric warp between the inputs, instituting the smooth geometric effect by bringing the shapes into an initial alignment. We provide comparisons to classic as well as latent space morphing techniques, and demonstrate that, given a set of images for self-supervision, our network learns to generate visually pleasing morphing effects featuring believable in-betweens, with robustness to changes in shape and texture, requiring no correspondence annotation.
|
['Daniel Cohen-Or', 'Noa Fish', 'Lilach Perry', 'Connelly Barnes', 'Richard Zhang', 'Eli Shechtman']
|
2020-04-29
| null | null | null | null |
['image-morphing']
|
['computer-vision']
|
[ 5.45209587e-01 7.10295677e-01 -3.68348323e-03 -3.44366401e-01
-6.37789190e-01 -7.94905484e-01 1.04539108e+00 -3.61672014e-01
1.49325170e-02 6.78950131e-01 2.50539660e-01 1.92498162e-01
1.78586274e-01 -1.07484031e+00 -1.26982641e+00 -6.41251445e-01
1.88960046e-01 5.00326216e-01 2.21595570e-01 -3.84680897e-01
-2.59925844e-03 6.12031817e-01 -1.38352895e+00 2.87430346e-01
7.25795925e-01 7.79908359e-01 -4.12709303e-02 8.89706433e-01
3.00475657e-01 6.88105941e-01 -4.46716845e-01 -7.69667625e-01
7.17249453e-01 -8.31076026e-01 -9.72814798e-01 6.17846668e-01
7.46390641e-01 -4.55304742e-01 -4.88322884e-01 8.26175570e-01
-1.20210806e-02 3.04361314e-01 6.44647121e-01 -1.28267455e+00
-9.77793694e-01 2.63596624e-01 -3.40325296e-01 -3.95394236e-01
5.52562773e-01 6.69548631e-01 9.93907988e-01 -6.52859330e-01
1.11150074e+00 1.32638288e+00 6.84372842e-01 7.08457232e-01
-1.64768553e+00 -2.27302670e-01 -4.14607264e-02 -1.95937753e-01
-9.77188289e-01 -6.09908700e-01 1.04879200e+00 -4.99416828e-01
5.07256269e-01 2.77538210e-01 9.74448740e-01 1.21012211e+00
2.80857623e-01 2.62181073e-01 9.37624872e-01 -6.06971264e-01
3.29182982e-01 -1.36504218e-01 -6.86917126e-01 7.94274032e-01
-1.98278442e-01 4.75110233e-01 -5.62945843e-01 1.28065512e-01
1.31388426e+00 2.76052505e-02 -3.59927475e-01 -7.21617818e-01
-1.27542019e+00 6.38645113e-01 5.77140510e-01 1.35240719e-01
-4.80630785e-01 2.59756893e-01 -1.21395230e-01 2.97393680e-01
4.13762391e-01 6.49004281e-01 -4.45190966e-02 1.87798068e-01
-9.48282719e-01 3.97549808e-01 3.95110250e-01 9.31709230e-01
1.09098291e+00 2.52582908e-01 -2.54742473e-01 3.34382087e-01
-8.08866173e-02 3.37808579e-01 2.93519825e-01 -1.48611903e+00
1.87765390e-01 4.90247905e-01 7.08826110e-02 -1.00879323e+00
3.45825493e-01 -2.15045940e-02 -8.30404878e-01 6.74332380e-01
4.84683841e-01 1.76589698e-01 -1.09964228e+00 2.10610747e+00
3.48028690e-01 2.84706116e-01 -2.17724200e-02 6.88968599e-01
9.70474258e-02 6.33732796e-01 -4.96751107e-02 -1.18051246e-01
1.07906687e+00 -8.91695678e-01 -5.44849634e-01 -3.10827404e-01
6.07691817e-02 -7.69742310e-01 1.25085509e+00 3.20234410e-02
-1.84846091e+00 -7.96207905e-01 -9.78313625e-01 -3.12609076e-01
-3.77979269e-03 -4.43581372e-01 3.01508754e-01 2.54220039e-01
-1.41745698e+00 9.81947303e-01 -9.32434976e-01 -2.44532436e-01
4.37518120e-01 1.73491910e-01 -7.48493552e-01 1.48618922e-01
-9.99189317e-01 9.20255363e-01 1.59769937e-01 -3.75817828e-02
-1.06876564e+00 -7.80308127e-01 -9.76436794e-01 -2.50577778e-01
-3.92497331e-02 -1.29266405e+00 9.97303426e-01 -1.60331261e+00
-1.74330199e+00 1.17266798e+00 -1.19693512e-02 -3.21950138e-01
8.99568379e-01 2.12536633e-01 -9.09570605e-02 2.28611052e-01
2.12375909e-01 1.37687552e+00 1.27797604e+00 -1.51016939e+00
-2.85365909e-01 -4.34327833e-02 2.86947340e-01 2.66919076e-01
7.75163695e-02 -2.51396418e-01 -4.31889176e-01 -9.86344576e-01
7.38097206e-02 -9.17452037e-01 -3.06228966e-01 4.83024240e-01
-3.97760570e-01 4.21580851e-01 8.63379538e-01 -8.47795188e-01
5.19813657e-01 -2.06138802e+00 6.94650650e-01 1.15139879e-01
7.92599618e-02 -2.71671087e-01 -4.03744221e-01 3.69584948e-01
-4.09844369e-01 2.97213458e-02 -6.63851559e-01 -6.88915789e-01
-2.38643825e-01 4.73683447e-01 -6.10690415e-01 5.94146252e-01
4.76824701e-01 1.19245505e+00 -9.56074774e-01 -2.48250097e-01
2.83180177e-01 5.52664995e-01 -7.84746349e-01 7.27413058e-01
-4.01975781e-01 9.57661510e-01 -5.14580980e-02 1.71003193e-01
3.97098213e-01 -1.10813446e-01 -1.06784299e-01 -2.96343535e-01
5.33996522e-02 1.11450285e-01 -9.10007775e-01 2.02093983e+00
-4.24443364e-01 4.97681767e-01 -1.12739904e-02 -9.58448589e-01
7.98911333e-01 3.02624345e-01 3.35532159e-01 -4.41897452e-01
5.76890633e-02 4.17880248e-03 -2.00481951e-01 -3.08366954e-01
4.26425427e-01 -5.47146499e-01 7.00246915e-02 5.75221539e-01
1.52494818e-01 -8.23485494e-01 -1.04524590e-01 1.31396085e-01
8.05207491e-01 5.86610377e-01 2.34098643e-01 4.57866956e-03
2.48054519e-01 -1.91098258e-01 4.07800913e-01 2.60233819e-01
1.75352171e-01 1.16220808e+00 3.48152101e-01 -7.28190303e-01
-1.62991714e+00 -1.31083906e+00 9.77893472e-02 5.49567401e-01
2.06292018e-01 2.23710518e-02 -1.09686327e+00 -6.01812363e-01
-3.27377319e-01 7.07941175e-01 -9.68652844e-01 -5.15736043e-01
-8.46112192e-01 -2.15032846e-02 4.28282112e-01 4.26134586e-01
5.06491959e-01 -1.25620949e+00 -5.99582374e-01 2.05578208e-01
-2.83287376e-01 -9.87643421e-01 -8.84285986e-01 5.81167452e-03
-8.79244387e-01 -9.42167461e-01 -8.34854484e-01 -9.88984942e-01
1.06111717e+00 -1.30720913e-01 1.21717560e+00 9.20295045e-02
-2.38497928e-01 3.22654933e-01 -5.25951851e-03 1.96755141e-01
-1.04524362e+00 -4.42882091e-01 -1.36919290e-01 3.95049304e-01
-7.44828343e-01 -1.11490762e+00 -6.63401842e-01 2.52335727e-01
-1.21707654e+00 5.96502364e-01 2.78136164e-01 9.39623058e-01
6.90519035e-01 -1.93941072e-01 -2.07263380e-02 -6.21389627e-01
-1.42966700e-03 -1.29710048e-01 -3.21238965e-01 1.71609819e-01
-3.66759390e-01 2.40833953e-01 8.13640118e-01 -6.28677070e-01
-1.07976389e+00 2.21936435e-01 4.41674665e-02 -7.19604552e-01
-2.23385885e-01 -1.93626329e-01 -4.60434288e-01 -8.51311460e-02
7.96393335e-01 1.99675977e-01 2.18174577e-01 -9.03841108e-02
9.63690937e-01 -1.11399122e-01 1.08829832e+00 -7.96125233e-01
1.26065445e+00 7.42071569e-01 -1.08298939e-02 -3.11677307e-01
-5.47341466e-01 5.99508286e-01 -9.68266368e-01 -2.41407543e-01
1.06203783e+00 -7.39075541e-01 -1.25950456e-01 5.04424036e-01
-1.34931231e+00 -7.51472414e-01 -1.07966149e+00 4.75036167e-03
-1.04652560e+00 3.01035404e-01 -5.39826274e-01 -2.66427547e-01
-3.95190120e-02 -1.16953397e+00 1.11932445e+00 -3.59034538e-02
-4.67328340e-01 -1.05807686e+00 1.98764771e-01 1.97706655e-01
2.00367123e-01 8.59895527e-01 9.17987168e-01 8.09473321e-02
-9.04141724e-01 1.14569478e-02 2.71283180e-01 4.50299531e-01
5.30578196e-01 2.74654984e-01 -8.35375607e-01 -3.30496162e-01
1.13633871e-02 -2.47760758e-01 5.38170755e-01 3.94716501e-01
8.47774565e-01 -8.03085923e-01 -8.62176940e-02 9.92139041e-01
1.11019194e+00 1.60684168e-01 9.00041163e-01 1.11268625e-01
9.24805582e-01 8.15682232e-01 6.25733985e-03 3.03601567e-02
1.79945886e-01 7.93477535e-01 5.96080601e-01 -4.12387669e-01
-6.52235389e-01 -7.03258038e-01 3.94675434e-01 4.80538696e-01
-2.93352425e-01 -7.87917003e-02 -4.93492693e-01 4.08804774e-01
-1.48747385e+00 -1.12448502e+00 2.91347355e-01 2.25481701e+00
1.15263522e+00 9.06039253e-02 1.39894942e-02 3.86164896e-02
7.86462188e-01 3.82766008e-01 -5.02892017e-01 -2.34046489e-01
-3.31378669e-01 4.03451234e-01 2.15269253e-01 9.02763188e-01
-7.38592505e-01 1.02093351e+00 6.58243084e+00 6.08099461e-01
-1.31366110e+00 7.12425262e-02 1.07525754e+00 1.25502199e-01
-9.13867831e-01 3.90734166e-01 9.54218134e-02 3.47720593e-01
5.43574572e-01 -1.55993640e-01 8.23825061e-01 4.76495862e-01
1.97096035e-01 2.63922244e-01 -1.40274537e+00 6.28062308e-01
3.16877291e-02 -1.82010078e+00 4.39571679e-01 -9.81320515e-02
1.11565888e+00 -5.38081586e-01 3.05702299e-01 -2.89341480e-01
5.54744661e-01 -1.11103833e+00 1.41064596e+00 6.64685726e-01
1.21439064e+00 -6.73156500e-01 3.79504636e-02 1.56608149e-01
-1.01397955e+00 2.93515235e-01 -5.03735952e-02 -9.18308571e-02
5.15470505e-01 1.91356421e-01 -4.02991533e-01 5.94001353e-01
3.44827116e-01 6.63392723e-01 -3.40723574e-01 2.18218356e-01
-3.69163424e-01 1.27639443e-01 4.85559106e-02 8.41181815e-01
1.37601905e-02 -5.05048811e-01 6.43228412e-01 7.20559180e-01
4.07964230e-01 8.72870460e-02 -7.43317651e-03 1.25572097e+00
-2.12607995e-01 -2.62684315e-01 -8.12818766e-01 1.91012144e-01
3.51674795e-01 1.03936493e+00 -7.41082430e-01 -3.29199582e-01
-2.90901326e-02 1.45799994e+00 2.96830028e-01 4.49922264e-01
-8.26842189e-01 5.85865509e-03 6.08317316e-01 4.40941691e-01
1.48696467e-01 -3.14571783e-02 -4.85279053e-01 -1.06402707e+00
8.39812905e-02 -8.48284006e-01 -3.34194861e-02 -1.03075421e+00
-1.15749705e+00 8.07233155e-01 -1.31263763e-01 -1.37418771e+00
-5.19105673e-01 -6.76757842e-02 -1.02582967e+00 8.72691512e-01
-1.12243021e+00 -1.49798441e+00 -3.04026067e-01 7.31310606e-01
5.26491702e-01 7.24098273e-03 7.57644892e-01 -1.83019742e-01
-1.40523463e-01 5.88773787e-01 -5.09261668e-01 5.72447516e-02
7.32902348e-01 -1.11022091e+00 8.11964869e-01 1.16282022e+00
2.74690956e-01 5.57563007e-01 8.33927989e-01 -5.82908452e-01
-1.10312319e+00 -1.29838574e+00 7.03456521e-01 -6.31294787e-01
3.34019065e-01 -2.55894899e-01 -9.18848038e-01 1.01918209e+00
4.94559377e-01 2.69991219e-01 1.18219323e-01 -6.10732317e-01
-4.97718930e-01 1.87781136e-02 -1.22618091e+00 9.07498062e-01
1.26439071e+00 -7.58564949e-01 -6.39974833e-01 1.02015935e-01
8.25288832e-01 -6.76665604e-01 -7.73603737e-01 1.61658779e-01
4.82891113e-01 -1.12737298e+00 1.12139332e+00 -6.81542635e-01
8.85679781e-01 -4.19803828e-01 1.75170824e-02 -1.47465360e+00
-3.33290756e-01 -1.24578595e+00 1.91167891e-01 1.16392934e+00
9.16884020e-02 -2.76636511e-01 7.80329406e-01 7.42439926e-01
-3.79034460e-01 -4.89067882e-01 -8.77139211e-01 -7.34388649e-01
2.70250499e-01 -1.89776331e-01 7.78327167e-01 1.06730640e+00
-2.34880358e-01 2.30168030e-01 -5.32698572e-01 1.36826545e-01
5.84036648e-01 4.46387194e-02 9.46737409e-01 -7.66373396e-01
-6.51100397e-01 -4.16240007e-01 -3.45649868e-01 -9.85596538e-01
2.44541302e-01 -7.45630682e-01 1.90426335e-02 -1.18910038e+00
2.74757612e-02 -2.83000648e-01 2.90337443e-01 5.96284151e-01
-1.28097817e-01 4.83542174e-01 2.73336291e-01 4.34177756e-01
-7.91067723e-03 7.10834980e-01 1.72615850e+00 -8.21618214e-02
-1.96595505e-01 -2.93156862e-01 -4.97705579e-01 7.33304024e-01
4.05022651e-01 -2.86526620e-01 -6.01132572e-01 -5.22070825e-01
-1.21039236e-02 4.29368943e-01 6.72778487e-01 -8.01054835e-01
-4.63774577e-02 -3.49047422e-01 5.29643416e-01 1.34552851e-01
4.89686340e-01 -6.96234107e-01 8.85228813e-01 4.10050929e-01
-6.58678651e-01 1.56083271e-01 -8.27669129e-02 4.72979844e-01
-1.12986371e-01 1.73200816e-01 1.12782180e+00 -1.13534681e-01
-2.66944885e-01 6.34711921e-01 1.12005323e-01 1.82883918e-01
1.08968258e+00 -6.80288434e-01 8.47226083e-02 -5.96464932e-01
-8.20912182e-01 -3.87733400e-01 1.30053055e+00 3.49279344e-01
4.79380906e-01 -1.64404416e+00 -6.36435151e-01 5.60783088e-01
-1.30111352e-01 3.69794548e-01 3.54889452e-01 3.33317518e-01
-6.21101797e-01 -2.79465497e-01 -6.47738993e-01 -5.44879079e-01
-7.30888665e-01 6.49961948e-01 5.27368188e-01 -9.83830690e-02
-8.67401958e-01 7.10721195e-01 6.21115088e-01 -1.88996717e-01
3.01502421e-02 -3.51839900e-01 2.93870538e-01 -2.80692756e-01
2.36241832e-01 -1.50460241e-04 -9.42599922e-02 -7.47525513e-01
9.25491005e-02 7.50071526e-01 4.22180563e-01 -5.32461286e-01
1.31780171e+00 -1.47158206e-01 -1.28191337e-01 2.43137419e-01
1.18151510e+00 2.02583358e-01 -2.22403669e+00 -6.05777279e-02
-4.66215640e-01 -6.47191167e-01 -3.92247856e-01 -5.68177879e-01
-1.30602765e+00 5.95818281e-01 1.66697055e-01 -1.59141142e-02
1.06910992e+00 6.45215064e-02 8.49668205e-01 -2.48215407e-01
3.04539561e-01 -5.73188901e-01 5.28198183e-01 1.05029300e-01
1.26743734e+00 -9.15545821e-01 -3.60439450e-01 -2.42142290e-01
-6.30486667e-01 1.00230503e+00 4.44458991e-01 -4.73804504e-01
2.33708650e-01 2.45578200e-01 -4.88849692e-02 2.76603810e-02
-5.17517090e-01 3.12501162e-01 3.55205387e-01 8.80367041e-01
1.02622539e-01 -9.04552788e-02 1.80980384e-01 -3.13920006e-02
-5.75973749e-01 -2.10355133e-01 4.21876699e-01 6.43068612e-01
-6.65805340e-02 -1.27480268e+00 -4.51199085e-01 -2.73326665e-01
4.98186201e-02 1.37721999e-02 -2.29434639e-01 8.27012837e-01
4.40187246e-01 5.06331980e-01 4.43779498e-01 -2.05168918e-01
3.38518202e-01 -4.75063100e-02 7.80126095e-01 -5.61814964e-01
-3.67516309e-01 -9.90420505e-02 -1.50177240e-01 -7.80900240e-01
-3.97571981e-01 -7.13691175e-01 -1.12252510e+00 -4.62065548e-01
2.17381224e-01 -1.90746889e-01 1.35856777e-01 8.67615461e-01
2.58684069e-01 4.46833611e-01 8.17703724e-01 -1.34495163e+00
-2.71214664e-01 -4.98830438e-01 -2.39732444e-01 9.51617062e-01
5.53175807e-01 -4.53526556e-01 -3.53083014e-01 9.49696302e-01]
|
[11.631582260131836, -0.6190052032470703]
|
f64d572c-0b95-4538-bc08-638369d8a354
|
deepfake-mnist-a-deepfake-facial-animation
|
2108.07949
| null |
https://arxiv.org/abs/2108.07949v1
|
https://arxiv.org/pdf/2108.07949v1.pdf
|
DeepFake MNIST+: A DeepFake Facial Animation Dataset
|
The DeepFakes, which are the facial manipulation techniques, is the emerging threat to digital society. Various DeepFake detection methods and datasets are proposed for detecting such data, especially for face-swapping. However, recent researches less consider facial animation, which is also important in the DeepFake attack side. It tries to animate a face image with actions provided by a driving video, which also leads to a concern about the security of recent payment systems that reply on liveness detection to authenticate real users via recognising a sequence of user facial actions. However, our experiments show that the existed datasets are not sufficient to develop reliable detection methods. While the current liveness detector cannot defend such videos as the attack. As a response, we propose a new human face animation dataset, called DeepFake MNIST+, generated by a SOTA image animation generator. It includes 10,000 facial animation videos in ten different actions, which can spoof the recent liveness detectors. A baseline detection method and a comprehensive analysis of the method is also included in this paper. In addition, we analyze the proposed dataset's properties and reveal the difficulty and importance of detecting animation datasets under different types of motion and compression quality.
|
['Chang Xu', 'Pei Du', 'Bo Du', 'Xueyu Wang', 'Jiajun Huang']
|
2021-08-18
| null | null | null | null |
['image-animation']
|
['computer-vision']
|
[ 2.21161142e-01 -1.94136843e-01 -7.13079944e-02 -9.42540616e-02
-6.37444779e-02 -6.47368610e-01 7.10243940e-01 -7.70192742e-01
-6.35505766e-02 2.33924687e-01 4.13844138e-02 -2.02890515e-01
1.63819820e-01 -6.46953106e-01 -3.65498841e-01 -8.76413167e-01
-1.60509482e-01 -1.75427064e-01 2.13058740e-01 -3.56323421e-01
4.01650488e-01 7.06508338e-01 -1.74376881e+00 3.77300769e-01
8.10340345e-02 9.55421925e-01 -2.40538880e-01 7.68709242e-01
1.12329453e-01 6.32384896e-01 -8.98504376e-01 -8.69280040e-01
4.46298748e-01 -6.59281433e-01 -5.18174827e-01 4.37339395e-02
6.13186598e-01 -1.05926394e+00 -4.95745301e-01 1.43119204e+00
7.57892907e-01 -5.36778510e-01 2.78840214e-01 -2.01182485e+00
-3.79868299e-01 3.92756671e-01 -9.02451396e-01 3.88650030e-01
6.58893704e-01 3.95741761e-01 1.86273888e-01 -6.86484635e-01
6.73955679e-01 1.81939805e+00 4.70545501e-01 1.01083910e+00
-4.80598450e-01 -1.52006972e+00 -4.32986051e-01 4.78699416e-01
-1.39023864e+00 -8.04939151e-01 9.14588630e-01 -2.60090232e-01
1.89721018e-01 3.25894564e-01 5.66825032e-01 1.51122236e+00
1.40204072e-01 4.44260389e-01 9.61549819e-01 -3.12994868e-01
-2.05619752e-01 2.77097262e-02 -6.15427457e-03 8.63109350e-01
3.50175351e-01 3.93935263e-01 -5.08525133e-01 -5.09524047e-01
6.58341289e-01 -3.33945274e-01 -2.71834105e-01 2.69738734e-01
-7.97163785e-01 9.04848218e-01 -3.43403220e-01 1.40138373e-01
-7.11015537e-02 -9.93141811e-03 6.92840457e-01 3.83394688e-01
-1.62329957e-01 -2.56193697e-01 1.02006458e-01 -2.49755874e-01
-7.50979960e-01 1.40916437e-01 5.97378969e-01 6.87278032e-01
4.11592126e-01 2.68562019e-01 1.79946586e-01 2.70881772e-01
4.69889551e-01 7.87476063e-01 4.31913465e-01 -1.00982559e+00
3.19323331e-01 3.68942350e-01 -2.17273071e-01 -1.87710154e+00
-9.02655497e-02 4.20858324e-01 -7.22191215e-01 3.78330380e-01
4.35819477e-01 -2.26208061e-01 -5.02578437e-01 1.60022640e+00
6.49248660e-01 6.31955743e-01 -1.87705263e-01 8.06879401e-01
9.51884627e-01 5.05723357e-01 -1.26905292e-01 -5.07836461e-01
1.64100575e+00 -4.50273633e-01 -9.91672158e-01 4.83734488e-01
6.84897244e-01 -1.06625974e+00 8.66350889e-01 7.73438513e-01
-7.13440955e-01 -6.03207767e-01 -9.05160904e-01 4.53566134e-01
-2.18026578e-01 7.51074329e-02 3.07011276e-01 1.80785263e+00
-8.94756913e-01 1.33340463e-01 -3.27005893e-01 -2.85019666e-01
6.17630184e-01 3.35533321e-01 -5.44255316e-01 1.09685749e-01
-1.46264768e+00 5.80048800e-01 7.32312649e-02 1.39880642e-01
-1.12654972e+00 -1.81994185e-01 -5.82858443e-01 -2.50194341e-01
2.14328319e-01 4.02506292e-02 9.58965421e-01 -1.35943770e+00
-1.44530094e+00 1.08809066e+00 -1.27731096e-02 -2.76977122e-01
6.43607199e-01 1.58277936e-02 -1.10316539e+00 5.58334231e-01
-2.59293944e-01 5.04369915e-01 1.59553325e+00 -9.00906980e-01
-3.09430778e-01 -3.56286198e-01 1.40951006e-02 -3.60427409e-01
-5.91707289e-01 8.15439880e-01 -1.58360109e-01 -5.83415151e-01
-4.80073571e-01 -1.01260686e+00 4.62390095e-01 1.62355974e-01
-4.59518641e-01 -2.64095306e-01 1.85227931e+00 -7.89400280e-01
1.46050882e+00 -2.24120259e+00 -5.22825480e-01 4.18731242e-01
2.17926234e-01 7.99603403e-01 -4.27913368e-01 3.52664709e-01
-2.84353524e-01 3.71162653e-01 2.59068519e-01 1.28429666e-01
-2.85462260e-01 4.65345383e-02 -5.19327104e-01 9.32197213e-01
3.50770690e-02 5.33730268e-01 -5.79065919e-01 -5.37218273e-01
6.87732846e-02 4.99134421e-01 -4.37265784e-01 4.23535518e-02
3.58919829e-01 2.45101646e-01 -3.94402593e-01 8.51021171e-01
1.31243908e+00 2.51839042e-01 1.68090448e-01 -3.58094245e-01
2.30583712e-01 -2.65738934e-01 -1.33778727e+00 8.70357394e-01
2.91809708e-01 7.96357751e-01 2.88348287e-01 -7.76499391e-01
8.62181485e-01 4.22508180e-01 3.90568525e-01 -5.95115006e-01
7.07404196e-01 7.78692737e-02 1.60358921e-01 -9.20618117e-01
3.46288741e-01 2.09227815e-01 7.48723969e-02 6.77271962e-01
-1.89012185e-01 4.48843867e-01 3.63519341e-02 3.34311992e-01
1.18840969e+00 5.06586023e-03 9.02147815e-02 -6.58986345e-03
9.40232694e-01 -5.52815616e-01 4.40705895e-01 4.15148497e-01
-7.42061973e-01 2.63552308e-01 6.65000021e-01 -5.04713953e-01
-7.93665111e-01 -5.02363801e-01 -2.45309304e-02 7.79046834e-01
4.22821820e-01 -6.88611984e-01 -1.21315169e+00 -1.07004297e+00
6.17534742e-02 3.19191217e-02 -4.58603650e-01 -3.95938307e-01
-6.67344153e-01 -5.62301219e-01 1.56486499e+00 -1.25134699e-02
9.68273997e-01 -1.17380702e+00 -7.32796729e-01 -2.59126216e-01
-3.64961594e-01 -1.26591516e+00 -4.77632016e-01 -1.07381213e+00
-2.69219011e-01 -1.70364237e+00 -6.35808885e-01 -7.10029423e-01
4.67937052e-01 5.23804545e-01 5.48487604e-01 5.63119531e-01
-4.58680689e-01 2.93532103e-01 -4.00463492e-01 -2.18248159e-01
-9.60324228e-01 -6.46172643e-01 5.45788944e-01 4.00779307e-01
6.77197874e-01 -9.31450352e-02 -4.31661218e-01 7.23578691e-01
-9.79118109e-01 -3.46045464e-01 1.94307446e-01 3.96002233e-01
-4.63822573e-01 2.67436296e-01 5.39579749e-01 -6.95466816e-01
6.09441161e-01 -4.33577508e-01 -4.51695114e-01 2.58749574e-01
-1.78997874e-01 -2.54640311e-01 3.62710714e-01 -7.40793288e-01
-9.18182135e-01 -2.45086521e-01 -2.51765072e-01 -6.70855463e-01
-3.95362228e-01 -2.34503046e-01 -5.20377398e-01 -6.00587904e-01
5.15729427e-01 2.55257279e-01 5.37167430e-01 -1.16825104e-01
1.41978681e-01 8.44337881e-01 3.79972041e-01 -1.69667244e-01
1.01981509e+00 5.23926258e-01 1.01260610e-01 -1.20881820e+00
3.10118616e-01 -7.12401196e-02 -5.51839024e-02 -7.31508017e-01
6.06619239e-01 -6.33757770e-01 -1.35881639e+00 1.26601851e+00
-1.42348754e+00 3.46265495e-01 5.28582156e-01 2.45854288e-01
-3.46877873e-01 1.25372565e+00 -7.47176528e-01 -9.11668420e-01
-2.82688141e-01 -1.21975148e+00 9.34698761e-01 8.25902969e-02
-1.19206339e-01 -4.68670726e-01 -4.13062647e-02 2.83696264e-01
3.30505431e-01 2.69415945e-01 4.84869599e-01 -3.25648695e-01
-3.13641578e-01 -1.34169549e-01 -2.28487089e-01 3.65272284e-01
4.08951968e-01 6.35330856e-01 -1.01869667e+00 -3.62412870e-01
1.34187117e-01 -3.57705325e-01 4.42338198e-01 3.16757336e-02
1.14230561e+00 -7.30343699e-01 -1.98163286e-01 4.84332979e-01
9.27183568e-01 4.55207795e-01 1.16451263e+00 7.80443996e-02
4.47893232e-01 8.30550611e-01 6.75420821e-01 6.54936612e-01
-2.20384568e-01 6.87917173e-01 6.61971629e-01 1.29878029e-01
-1.13580257e-01 -6.92213923e-02 8.97853851e-01 3.11633736e-01
-1.96015239e-01 -3.83928716e-01 -5.74962020e-01 -1.73828498e-01
-1.23215830e+00 -1.48348761e+00 -3.11680257e-01 1.77726376e+00
4.50798064e-01 -5.86100072e-02 5.08696914e-01 5.59131205e-01
1.26588786e+00 2.32626766e-01 -2.78829068e-01 -4.08031523e-01
-1.78967729e-01 1.68909013e-01 3.47552806e-01 1.08783536e-01
-1.02904534e+00 9.34883237e-01 5.93137169e+00 9.31832433e-01
-1.31112349e+00 7.55738840e-02 5.24711728e-01 2.91012973e-01
7.91529119e-02 -2.68783540e-01 -1.00439727e+00 9.91954803e-01
9.34360683e-01 1.25307199e-02 2.31805369e-02 7.51207590e-01
4.65588778e-01 1.43675864e-01 -4.89851952e-01 1.40550840e+00
5.74331164e-01 -1.07229900e+00 3.11622679e-01 3.47416162e-01
3.47222537e-02 -7.58127511e-01 2.08126515e-01 2.96527473e-03
-2.28165407e-02 -8.44676793e-01 4.94044483e-01 -4.90651019e-02
1.03107989e+00 -8.29608202e-01 6.72842205e-01 2.31929913e-01
-1.23360717e+00 -2.19880417e-01 -2.73802817e-01 1.33710608e-01
2.75604844e-01 9.68815107e-03 -6.16117895e-01 1.75827861e-01
6.82080805e-01 6.65599346e-01 -4.86288428e-01 6.20712519e-01
-2.45353561e-02 6.81961536e-01 -3.18195224e-01 9.97096151e-02
-1.45264044e-01 6.01267591e-02 5.72910070e-01 1.09294415e+00
6.56756878e-01 1.06454544e-01 -3.73693585e-01 3.48811775e-01
-1.33182675e-01 -3.73215303e-02 -9.92664874e-01 -2.56997198e-01
4.77406561e-01 1.29340625e+00 -5.03149569e-01 -1.78023517e-01
-3.28338057e-01 1.02518165e+00 -5.03319800e-01 4.04374637e-02
-1.22984254e+00 -3.28414917e-01 9.53665197e-01 2.34415367e-01
-3.88094448e-02 -1.46201968e-01 4.75784302e-01 -1.03318369e+00
-2.72962630e-01 -1.37967336e+00 4.98348236e-01 -6.30290389e-01
-9.22625780e-01 3.69776189e-01 -1.13975808e-01 -1.38786387e+00
-2.01010585e-01 -6.61894441e-01 -5.60659111e-01 2.76480526e-01
-1.21636450e+00 -9.23640609e-01 -6.04151309e-01 1.07772136e+00
4.59802777e-01 -6.53334737e-01 5.93174458e-01 6.05926037e-01
-6.52648032e-01 1.07119155e+00 -6.37334645e-01 5.20286322e-01
8.57707381e-01 -4.40975688e-02 4.46472377e-01 9.63145375e-01
9.92519967e-03 5.01964390e-01 5.47055602e-01 -7.73780823e-01
-1.58993733e+00 -6.25257432e-01 5.62717676e-01 -1.48793623e-01
4.58001345e-01 -1.89244643e-01 -7.17874527e-01 3.53434175e-01
2.29910105e-01 -5.67417070e-02 5.34152865e-01 -9.20436144e-01
-4.42621887e-01 -8.17210302e-02 -1.62912798e+00 4.56972063e-01
1.07635725e+00 -4.73140895e-01 -2.57306635e-01 6.74154097e-03
1.99700400e-01 7.30112866e-02 -3.03697765e-01 3.49936932e-01
9.67143178e-01 -1.22667718e+00 8.50668371e-01 -7.06417918e-01
2.24869266e-01 -1.66009605e-01 9.49290581e-03 -5.06386876e-01
9.05178562e-02 -1.07504654e+00 -3.50768626e-01 1.42357361e+00
-4.54746306e-01 -6.43644512e-01 9.33768511e-01 1.70041397e-01
6.78804040e-01 1.02595344e-01 -1.03179312e+00 -8.28827024e-01
-4.74504590e-01 -2.85387158e-01 7.13308096e-01 1.30614471e+00
-1.22684419e-01 -9.50139090e-02 -9.86248851e-01 2.49164373e-01
8.59995604e-01 -5.71357608e-01 1.21873474e+00 -1.00055814e+00
2.68926769e-02 -4.06284839e-01 -9.80660975e-01 -8.43281806e-01
1.15425445e-01 -2.83459753e-01 -5.08984983e-01 -3.47039849e-01
9.01159570e-02 -9.98618081e-02 1.97522596e-01 3.12098086e-01
2.26375580e-01 6.34599626e-01 2.97453463e-01 2.24583969e-01
7.28233829e-02 2.72471577e-01 9.92277622e-01 -6.08987771e-02
2.48817071e-01 -8.89309794e-02 -2.80417293e-01 8.47691357e-01
7.79982030e-01 -3.89699697e-01 -4.37561363e-01 1.23025917e-01
9.83948261e-02 2.06406280e-01 5.40930390e-01 -9.07764912e-01
-2.51893289e-02 -2.06408679e-01 1.50838718e-01 -4.20785666e-01
2.49039829e-01 -8.82626653e-01 2.31237441e-01 9.25722122e-01
-5.52401394e-02 4.02153611e-01 5.96332317e-03 4.25596207e-01
-1.68977574e-01 -1.53069645e-01 1.03678751e+00 -8.00906681e-03
-9.43441808e-01 3.61180156e-01 -4.93217826e-01 -2.07121149e-01
1.38961923e+00 -5.31957805e-01 -7.64214396e-01 -6.28308117e-01
-2.89774597e-01 -2.70153403e-01 3.62344623e-01 5.99482894e-01
1.08466089e+00 -1.34725094e+00 -8.09875011e-01 7.46753097e-01
-2.03442931e-01 -9.67117429e-01 2.71122217e-01 6.51966870e-01
-8.25964451e-01 -1.00941934e-01 -7.41611540e-01 -4.67048585e-01
-2.22238398e+00 6.97472930e-01 1.96330070e-01 3.81259203e-01
-4.10812199e-01 6.09870374e-01 1.72931001e-01 2.06637740e-01
1.20878130e-01 3.42956930e-01 -4.79019910e-01 2.25015670e-01
1.04454029e+00 5.93648255e-01 -1.75617740e-01 -1.16820979e+00
-6.01651430e-01 5.02841890e-01 -2.55544204e-03 1.31865993e-01
5.22250831e-01 -2.56439626e-01 2.81678885e-02 -4.10570323e-01
1.37672222e+00 3.31145734e-01 -7.69875288e-01 1.88841239e-01
-2.09030032e-01 -9.46865201e-01 -3.11272830e-01 -1.42118573e-01
-1.40867341e+00 9.24423099e-01 9.60538089e-01 3.82369637e-01
1.06312025e+00 -2.73219317e-01 1.30586839e+00 4.52846549e-02
5.47477722e-01 -7.69805491e-01 4.51534152e-01 1.89095959e-01
6.12375677e-01 -1.00369608e+00 -2.29204640e-01 -6.75392985e-01
-5.07031202e-01 1.45050824e+00 5.76775968e-01 4.77039907e-03
7.82557309e-01 4.90612358e-01 1.97150782e-01 -7.24693015e-02
-3.08990300e-01 2.08292693e-01 -2.75530249e-01 7.90276289e-01
-2.51228392e-01 -2.50157803e-01 -3.58304471e-01 2.63487250e-01
-1.48182690e-01 1.82333604e-01 8.88044178e-01 8.04050088e-01
-3.30301046e-01 -1.16217768e+00 -5.91568351e-01 5.70881926e-03
-6.62783325e-01 3.47486615e-01 -5.90581357e-01 8.22484791e-01
2.19984308e-01 1.06898355e+00 -1.20374598e-01 -9.00160313e-01
-1.13743953e-01 -1.22765541e-01 3.26474220e-01 1.86538734e-02
-3.74795765e-01 -2.06562132e-01 9.88088399e-02 -7.78110743e-01
-7.26195395e-01 -4.75915283e-01 -8.33202243e-01 -1.08308613e+00
-3.22177380e-01 -1.02257997e-01 4.50261176e-01 6.46043360e-01
1.40155852e-01 -3.81025910e-01 1.15538681e+00 -7.38395572e-01
-5.66025019e-01 -6.45517409e-01 -6.99559748e-01 7.01775849e-01
3.95882905e-01 -8.01837564e-01 -6.94797635e-01 2.11409166e-01]
|
[12.887554168701172, 1.1060163974761963]
|
c60e0362-beb6-4324-9b31-1af18a9ad4a3
|
pats-patch-area-transportation-with
|
2303.07700
| null |
https://arxiv.org/abs/2303.07700v2
|
https://arxiv.org/pdf/2303.07700v2.pdf
|
PATS: Patch Area Transportation with Subdivision for Local Feature Matching
|
Local feature matching aims at establishing sparse correspondences between a pair of images. Recently, detector-free methods present generally better performance but are not satisfactory in image pairs with large scale differences. In this paper, we propose Patch Area Transportation with Subdivision (PATS) to tackle this issue. Instead of building an expensive image pyramid, we start by splitting the original image pair into equal-sized patches and gradually resizing and subdividing them into smaller patches with the same scale. However, estimating scale differences between these patches is non-trivial since the scale differences are determined by both relative camera poses and scene structures, and thus spatially varying over image pairs. Moreover, it is hard to obtain the ground truth for real scenes. To this end, we propose patch area transportation, which enables learning scale differences in a self-supervised manner. In contrast to bipartite graph matching, which only handles one-to-one matching, our patch area transportation can deal with many-to-many relationships. PATS improves both matching accuracy and coverage, and shows superior performance in downstream tasks, such as relative pose estimation, visual localization, and optical flow estimation. The source code is available at \url{https://zju3dv.github.io/pats/}.
|
['Guofeng Zhang', 'Zhaopeng Cui', 'Hujun Bao', 'Hongsheng Li', 'Zhaoyang Huang', 'Yijin Li', 'Junjie Ni']
|
2023-03-14
| null |
http://openaccess.thecvf.com//content/CVPR2023/html/Ni_PATS_Patch_Area_Transportation_With_Subdivision_for_Local_Feature_Matching_CVPR_2023_paper.html
|
http://openaccess.thecvf.com//content/CVPR2023/papers/Ni_PATS_Patch_Area_Transportation_With_Subdivision_for_Local_Feature_Matching_CVPR_2023_paper.pdf
|
cvpr-2023-1
|
['visual-localization', 'graph-matching']
|
['computer-vision', 'graphs']
|
[-4.25661132e-02 -2.74819940e-01 -2.13297531e-01 -6.12601358e-03
-7.60476589e-01 -7.26047516e-01 2.88344681e-01 2.43382663e-01
-1.54624164e-01 3.96253139e-01 -6.65898994e-02 -4.48876731e-02
-1.10590179e-02 -9.14292097e-01 -7.11433589e-01 -5.24949789e-01
1.21829472e-01 3.57973099e-01 5.38229287e-01 -6.03433065e-02
2.20633358e-01 6.48317695e-01 -1.26439238e+00 -3.57525311e-02
8.85490358e-01 9.87903595e-01 3.64035875e-01 4.36511189e-01
5.67216203e-02 4.69748080e-01 -4.68164645e-02 -2.59747714e-01
5.61200321e-01 -6.00366890e-01 -7.97340572e-01 3.06341112e-01
1.11926699e+00 -2.27971122e-01 -6.89965963e-01 1.29389012e+00
2.48024642e-01 1.40218318e-01 3.33100379e-01 -1.34853697e+00
-3.68904650e-01 -3.33203450e-02 -1.04384637e+00 1.97720498e-01
4.99697775e-01 8.81082267e-02 1.10155630e+00 -9.05712605e-01
6.41182005e-01 1.14627624e+00 7.45602608e-01 6.72249869e-02
-1.42951882e+00 -6.97188139e-01 5.76227345e-02 2.27802992e-01
-1.65212643e+00 -5.07153749e-01 8.47510874e-01 -4.72590536e-01
4.15514290e-01 2.25160137e-01 7.65580654e-01 4.70730156e-01
-5.93918823e-02 5.75969398e-01 9.53145027e-01 -2.43191913e-01
-7.36333430e-02 -2.95458406e-01 -1.43404439e-01 8.64461482e-01
2.36285716e-01 1.17206154e-02 -5.21054804e-01 -1.45829856e-01
1.33972931e+00 2.96533793e-01 -6.35122299e-01 -8.73320699e-01
-1.68410575e+00 7.09509790e-01 9.39371049e-01 3.52643460e-01
-2.82830805e-01 1.80482849e-01 1.10894829e-01 2.83328772e-01
1.83411941e-01 3.38922322e-01 -1.71907455e-01 2.01304615e-01
-8.44660580e-01 3.98817174e-02 5.17458141e-01 1.11327744e+00
1.43350935e+00 -3.51151109e-01 8.46162960e-02 6.67934418e-01
3.78567837e-02 4.86541212e-01 1.18372507e-01 -1.11697066e+00
4.79501307e-01 6.86836958e-01 7.09170699e-02 -1.72072601e+00
-3.97455215e-01 -2.63092577e-01 -1.19928145e+00 -5.52061712e-03
7.33185232e-01 1.67214528e-01 -5.88313520e-01 1.67765296e+00
4.71164554e-01 3.90448451e-01 -3.72007906e-01 1.06862247e+00
6.85021818e-01 5.19863307e-01 -2.92497575e-01 -7.95144439e-02
1.34239483e+00 -1.17090666e+00 -3.10567379e-01 -3.55894595e-01
4.48549837e-01 -9.72162306e-01 7.77894080e-01 -4.04395349e-03
-1.01529849e+00 -5.65449297e-01 -7.10558295e-01 -2.61963516e-01
-1.12149924e-01 1.28335401e-01 4.50540811e-01 1.21845268e-01
-1.17493021e+00 4.86174405e-01 -8.35897267e-01 -3.95428210e-01
3.16465706e-01 2.64486134e-01 -6.66376114e-01 -4.31877315e-01
-9.59060967e-01 5.48248053e-01 -7.06979781e-02 5.92600144e-02
-1.88694000e-01 -6.55506194e-01 -1.13426673e+00 -1.19956821e-01
5.17156303e-01 -8.78708839e-01 7.25421309e-01 -7.56981254e-01
-1.03422463e+00 8.91799808e-01 -4.12409425e-01 -5.75980097e-02
6.02756798e-01 2.61594176e-01 -5.93129219e-03 2.90852696e-01
4.56670672e-01 8.30020070e-01 8.43556225e-01 -1.07707059e+00
-5.96642435e-01 -4.90076095e-01 1.69422254e-01 2.69190222e-01
-1.31085768e-01 -2.12840348e-01 -9.26349759e-01 -5.22129118e-01
5.83420873e-01 -9.33976293e-01 -3.98384482e-01 5.93455970e-01
-2.41008818e-01 2.24209316e-02 6.53210938e-01 -5.72073042e-01
1.01727843e+00 -2.17917418e+00 1.48225501e-01 3.11298579e-01
5.22619545e-01 -1.48920028e-03 -3.13564003e-01 2.53995389e-01
8.65853578e-02 -3.07903886e-01 -2.38068283e-01 -3.00354123e-01
-4.09646630e-01 2.93551460e-02 1.40507787e-01 8.60966265e-01
-4.21924777e-02 8.88283432e-01 -1.13734138e+00 -6.77579045e-01
4.13539529e-01 3.46915662e-01 -5.70317030e-01 7.08927661e-02
2.66963273e-01 7.61901379e-01 -4.36431020e-01 6.50474608e-01
1.01551068e+00 -4.58710194e-01 -1.82133131e-02 -6.46839678e-01
-2.15262398e-01 -5.43872900e-02 -1.56839132e+00 2.00002146e+00
-5.05040586e-01 7.59601653e-01 1.39477044e-01 -1.18414342e+00
9.08688068e-01 -2.44750530e-02 7.08262801e-01 -7.24270999e-01
-1.16639223e-03 3.12345833e-01 -2.56091624e-01 -2.05529407e-01
3.04431140e-01 1.39628172e-01 1.60552617e-02 4.27843407e-02
-2.24582165e-01 -2.46143699e-01 2.36355364e-01 1.84674636e-01
1.05124474e+00 -8.73884037e-02 6.01608574e-01 -1.65562391e-01
7.02024460e-01 2.24069357e-02 8.63297164e-01 4.63815898e-01
-3.33326459e-01 1.03956485e+00 3.96868259e-01 -4.29715097e-01
-9.25649583e-01 -1.02024364e+00 -2.09128276e-01 4.15925235e-01
1.06565809e+00 -4.99916732e-01 -5.43825150e-01 -5.50442994e-01
1.75321296e-01 -1.31116614e-01 -4.79716510e-01 9.74019691e-02
-6.16495192e-01 -1.72474623e-01 1.35427117e-01 3.93985808e-01
7.16240585e-01 -6.31703138e-01 -3.54172230e-01 1.02155633e-01
-6.07556462e-01 -1.25365269e+00 -1.19564629e+00 -3.24181855e-01
-7.58292854e-01 -1.36832500e+00 -8.36161613e-01 -1.01582837e+00
1.03481698e+00 1.02566576e+00 1.02422726e+00 2.18629107e-01
-4.96271938e-01 3.11466515e-01 -1.86486751e-01 3.03407073e-01
-2.21395474e-02 -7.09915310e-02 -9.21877846e-02 2.57330239e-01
6.13273643e-02 -6.93529487e-01 -1.07202971e+00 7.85719097e-01
-7.12006688e-01 1.14322610e-01 6.23890340e-01 8.75598669e-01
8.52595329e-01 -9.02396515e-02 -9.78207774e-03 -4.97313499e-01
-4.87167314e-02 -2.20380530e-01 -8.22121203e-01 3.61309201e-01
-2.78510958e-01 -1.59520343e-01 5.54619253e-01 -2.90063560e-01
-5.10684848e-01 4.32726413e-01 2.67579883e-01 -7.40744174e-01
-1.73564285e-01 3.11106481e-02 -1.35081381e-01 -5.70027292e-01
4.55820322e-01 1.87814191e-01 2.25384489e-01 -3.49303842e-01
3.19072336e-01 3.39345545e-01 5.46522558e-01 -2.67559469e-01
1.06073070e+00 7.08334029e-01 2.00312182e-01 -8.68073463e-01
-6.54107153e-01 -9.51177776e-01 -9.39387321e-01 -1.29126847e-01
7.32324541e-01 -9.64656532e-01 -5.58208823e-01 4.93516862e-01
-1.15090907e+00 -1.78026870e-01 -2.14187458e-01 5.47737420e-01
-4.98007953e-01 7.16512501e-01 -5.04303098e-01 -1.68938622e-01
-1.32692292e-01 -1.19691586e+00 1.25160336e+00 1.66670576e-01
3.01230624e-02 -9.89811122e-01 9.40561518e-02 4.00306731e-01
3.13091785e-01 2.04522982e-01 4.12844747e-01 8.00667703e-02
-9.59938586e-01 -2.08443910e-01 -6.12434149e-01 4.42917012e-02
2.89114475e-01 -3.00145179e-01 -5.45240521e-01 -5.44753611e-01
-3.48502070e-01 2.92600151e-02 7.68862724e-01 3.27446461e-01
8.97525609e-01 -8.78319964e-02 -4.25725758e-01 8.65797102e-01
1.46024835e+00 -1.62077144e-01 5.74315846e-01 3.47339779e-01
9.57723975e-01 6.98347807e-01 7.02961922e-01 2.16505840e-01
4.61687356e-01 1.05725431e+00 4.13254321e-01 -3.76320124e-01
-4.49327707e-01 -3.27633262e-01 6.07578456e-02 5.69141924e-01
1.00253718e-02 1.24329619e-01 -8.70833337e-01 5.34269512e-01
-1.92773700e+00 -8.69302094e-01 -3.80226523e-01 2.41612172e+00
6.08217955e-01 -3.28188539e-01 1.06962703e-01 -2.52851903e-01
1.01860750e+00 2.89385766e-01 -5.01844406e-01 4.36604321e-01
-1.16673149e-01 -2.90137470e-01 6.60751700e-01 6.90417886e-01
-1.02163780e+00 9.25135374e-01 4.58033752e+00 8.98718536e-01
-9.86595452e-01 1.21814609e-01 5.56400955e-01 1.40371427e-01
-2.30295047e-01 4.34905291e-01 -5.98988831e-01 3.82858217e-01
3.07650445e-03 -9.76890028e-02 4.85618889e-01 5.02019346e-01
3.10298055e-02 -2.15456322e-01 -9.31133807e-01 1.38422430e+00
1.85283870e-01 -1.34560251e+00 -7.19670113e-03 5.75999394e-02
7.75850058e-01 6.93949983e-02 -2.55077988e-01 -2.51267314e-01
-1.93452641e-01 -5.40623009e-01 4.97012109e-01 2.94373274e-01
7.68872023e-01 -5.33761621e-01 4.19102758e-01 3.49503458e-01
-1.76844847e+00 2.13642061e-01 -6.07211828e-01 1.38440773e-01
4.51399051e-02 7.86147892e-01 -1.97273493e-01 7.17304647e-01
7.27907538e-01 1.05981040e+00 -6.06496513e-01 1.43151736e+00
-2.33743545e-02 -1.33809790e-01 -3.76074910e-01 4.50339496e-01
1.59654114e-02 -7.59263098e-01 6.19343162e-01 9.35008585e-01
4.00516063e-01 -2.43747607e-03 5.41622758e-01 7.21455216e-01
-6.07589521e-02 2.08154187e-01 -8.11760485e-01 5.48267245e-01
5.12216866e-01 1.36983240e+00 -8.92210364e-01 -1.27524763e-01
-7.86320269e-01 1.24878478e+00 4.60758448e-01 3.28054011e-01
-6.96967900e-01 -4.89400238e-01 7.82109618e-01 4.19415683e-01
3.41116488e-01 -3.49511862e-01 4.96500246e-02 -1.52849936e+00
2.26920590e-01 -7.61413813e-01 3.79013956e-01 -5.25452137e-01
-1.16722250e+00 4.82436955e-01 -1.03477150e-01 -1.82690120e+00
1.16517708e-01 -3.12425435e-01 -4.86792386e-01 7.45612741e-01
-1.68897545e+00 -1.19477475e+00 -7.88147211e-01 8.88157010e-01
3.49139899e-01 2.99538791e-01 4.31367517e-01 4.26039249e-01
-4.64418977e-01 5.93608379e-01 9.95134190e-02 3.61205161e-01
9.35960352e-01 -9.78164315e-01 3.73093843e-01 9.42346215e-01
2.98001140e-01 4.46455330e-01 2.46592149e-01 -4.06924009e-01
-1.47134137e+00 -1.12342513e+00 8.76997530e-01 -2.03853428e-01
7.59347141e-01 -3.70024562e-01 -9.19641554e-01 4.10744011e-01
5.27675869e-03 5.77057123e-01 1.25246078e-01 -2.48168439e-01
-5.03759921e-01 -3.34672391e-01 -9.34563458e-01 5.06926596e-01
1.42918265e+00 -5.90816200e-01 -5.64106219e-02 3.97443831e-01
2.77790397e-01 -6.25944734e-01 -8.62661481e-01 2.72822559e-01
3.04613858e-01 -1.08252573e+00 1.18724620e+00 -1.04136614e-03
1.37732595e-01 -5.97450137e-01 -7.41134211e-02 -1.19921649e+00
-4.28066760e-01 -7.29723454e-01 3.14087212e-01 1.19081104e+00
1.08680978e-01 -8.90759349e-01 8.81886542e-01 3.43928605e-01
1.72390893e-01 -5.60602725e-01 -9.36701953e-01 -9.54133987e-01
-1.68289781e-01 -4.57500294e-03 4.17053640e-01 1.05371225e+00
-9.25893113e-02 3.18166137e-01 -2.82430679e-01 4.12275612e-01
9.50827897e-01 7.87282526e-01 9.58618581e-01 -1.08282924e+00
-3.09497625e-01 -5.29112577e-01 -9.01848257e-01 -1.56500697e+00
9.90618914e-02 -7.27066398e-01 2.78974082e-02 -1.46877515e+00
2.36166626e-01 -6.83458805e-01 7.31001720e-02 4.36247587e-01
-2.13642448e-01 8.10255349e-01 1.85103133e-01 6.12927973e-01
-6.05820358e-01 4.46623504e-01 1.53260839e+00 -1.14734635e-01
-1.74937427e-01 3.52612287e-02 -4.57060874e-01 6.27625465e-01
8.31406593e-01 -2.69235611e-01 -2.49557942e-01 -4.47802186e-01
-1.29111102e-02 3.47432494e-01 5.95944405e-01 -9.89354312e-01
5.40465236e-01 -2.54721940e-01 1.53738484e-01 -4.62668926e-01
2.81639606e-01 -8.95641446e-01 2.13795453e-01 4.73674208e-01
-5.05577028e-02 1.51107997e-01 -1.35195807e-01 7.46102035e-01
-5.81058383e-01 1.43207721e-02 9.84049439e-01 -6.27243668e-02
-8.18394482e-01 8.77410412e-01 1.62380219e-01 1.02476671e-01
1.15335023e+00 -3.73828620e-01 -2.35479549e-01 -5.03947437e-01
-4.89841759e-01 3.71116042e-01 8.61112952e-01 3.74325037e-01
6.08937740e-01 -1.47339571e+00 -6.89834118e-01 3.89278322e-01
4.67218637e-01 3.04796040e-01 4.26073134e-01 1.33505845e+00
-5.93906999e-01 2.42960617e-01 -1.47118330e-01 -9.17786658e-01
-1.38017297e+00 5.50233126e-01 5.58532357e-01 1.07695572e-02
-7.78479159e-01 6.54197395e-01 6.39251888e-01 -4.13278550e-01
1.78724065e-01 -1.28942624e-01 -5.51806800e-02 8.59554037e-02
3.59953821e-01 3.91525298e-01 5.95385246e-02 -1.02085876e+00
-4.22784775e-01 1.38982129e+00 1.20045021e-01 3.26488584e-01
9.23103034e-01 -4.27146047e-01 -3.54936421e-01 1.14200814e-02
1.68150198e+00 1.11008666e-01 -1.35332906e+00 -6.00376070e-01
-3.76143128e-01 -1.11793590e+00 1.05645023e-01 7.37736970e-02
-1.51717544e+00 7.26202548e-01 3.99067551e-01 6.82801828e-02
1.20026994e+00 2.46704400e-01 9.49357629e-01 1.73329413e-01
5.41953385e-01 -5.83179057e-01 9.36835557e-02 2.99791217e-01
8.40818167e-01 -1.49266458e+00 6.60856962e-02 -7.61430204e-01
-3.25080574e-01 1.13239920e+00 6.63359284e-01 -2.33814105e-01
5.96833110e-01 -4.43016402e-02 5.54256663e-02 -1.57518074e-01
-1.18512884e-01 -3.82241547e-01 4.82244253e-01 5.85877657e-01
1.62099093e-01 -1.12605162e-01 -1.86341360e-01 -3.75505328e-01
9.02487785e-02 -3.00185710e-01 2.34515965e-01 7.16002643e-01
-2.40319267e-01 -1.27034938e+00 -4.53263879e-01 3.26167971e-01
1.58326611e-01 -2.57889573e-02 -2.55523324e-01 7.50579953e-01
-9.49746929e-03 9.06795859e-01 2.90434182e-01 -2.33872727e-01
5.21960378e-01 -6.32736921e-01 5.40126204e-01 -3.99226397e-01
-1.11966006e-01 1.16241336e-01 -1.85474873e-01 -9.29692328e-01
-5.53140402e-01 -7.59657979e-01 -1.01456404e+00 -4.88202184e-01
-3.71011943e-01 4.68841977e-02 2.37031624e-01 4.99760658e-01
5.24262846e-01 4.44044592e-03 9.94134247e-01 -1.04850733e+00
-4.05450463e-01 -3.66600275e-01 -5.92221737e-01 6.37783051e-01
5.52100599e-01 -6.73302591e-01 -5.13858974e-01 -1.46073803e-01]
|
[8.532251358032227, -2.2230277061462402]
|
753821ca-2e3b-4e28-b9ca-c74a1c1873ca
|
image-denoising-by-gaussian-patch-mixture
|
2011.10290
| null |
https://arxiv.org/abs/2011.10290v1
|
https://arxiv.org/pdf/2011.10290v1.pdf
|
Image Denoising by Gaussian Patch Mixture Model and Low Rank Patches
|
Non-local self-similarity based low rank algorithms are the state-of-the-art methods for image denoising. In this paper, a new method is proposed by solving two issues: how to improve similar patches matching accuracy and build an appropriate low rank matrix approximation model for Gaussian noise. For the first issue, similar patches can be found locally or globally. Local patch matching is to find similar patches in a large neighborhood which can alleviate noise effect, but the number of patches may be insufficient. Global patch matching is to determine enough similar patches but the error rate of patch matching may be higher. Based on this, we first use local patch matching method to reduce noise and then use Gaussian patch mixture model to achieve global patch matching. The second issue is that there is no low rank matrix approximation model to adapt to Gaussian noise. We build a new model according to the characteristics of Gaussian noise, then prove that there is a globally optimal solution of the model. By solving the two issues, experimental results are reported to show that the proposed approach outperforms the state-of-the-art denoising methods includes several deep learning ones in both PSNR / SSIM values and visual quality.
|
['Michael Kwok-Po Ng', 'Qiyu Jin', 'Chen Luo', 'Shuping Wang', 'Jing Guo']
|
2020-11-20
| null | null | null | null |
['patch-matching']
|
['computer-vision']
|
[-8.04650411e-02 -5.95183313e-01 3.65297288e-01 1.24965884e-01
-8.29867482e-01 -5.14513664e-02 -3.71276848e-02 2.71265917e-02
-1.99252620e-01 3.78979385e-01 1.97803557e-01 4.28507656e-01
-5.07489860e-01 -8.79655659e-01 -7.12558448e-01 -1.09951639e+00
-1.15092075e-03 -8.73222873e-02 5.90329945e-01 -4.10134405e-01
3.29096317e-01 2.40696758e-01 -1.53598464e+00 3.96140277e-01
9.92330670e-01 1.06842065e+00 4.07665193e-01 2.70454586e-01
-1.04594775e-01 6.03382945e-01 -2.92602628e-01 6.16964996e-02
5.22828162e-01 -5.46931267e-01 -4.64625210e-01 1.68795437e-01
4.64263856e-01 -2.12596402e-01 -3.40487570e-01 1.63932562e+00
8.43983531e-01 2.46484920e-01 5.35186231e-01 -8.51012707e-01
-6.47745609e-01 2.66358614e-01 -9.19630289e-01 2.68255562e-01
1.71626642e-01 -2.57617235e-01 5.15777469e-01 -9.00255263e-01
4.63503987e-01 1.52823806e+00 1.04426503e+00 -2.00598184e-02
-1.17186534e+00 -5.24063408e-01 1.84796408e-01 5.59145868e-01
-1.76409733e+00 -1.20835111e-01 9.35497224e-01 -2.63823211e-01
2.53962874e-01 3.65954340e-01 5.03501236e-01 5.51731586e-01
4.23927277e-01 4.87252474e-01 1.38272774e+00 -2.82377869e-01
1.61227956e-01 -2.70587504e-01 1.93682045e-01 5.63537598e-01
2.15697870e-01 -4.86120358e-02 -2.71597445e-01 -3.82195562e-01
9.53230143e-01 1.90317303e-01 -5.45978487e-01 -1.23059094e-01
-1.09637058e+00 7.10957170e-01 5.97890019e-01 8.87719631e-01
-5.57331324e-01 -8.12489390e-02 3.90836954e-01 4.35115218e-01
4.56192136e-01 6.15358772e-03 -1.34934753e-01 2.81936496e-01
-8.52005899e-01 2.63319850e-01 6.55986011e-01 6.13378108e-01
1.17787540e+00 -2.35006753e-02 -1.32011905e-01 1.25223267e+00
4.12811577e-01 5.10998189e-01 5.27367949e-01 -1.15088809e+00
2.81569541e-01 3.11895579e-01 -9.91830677e-02 -1.90469182e+00
-1.47374615e-01 -6.17092311e-01 -1.59505618e+00 2.29730427e-01
1.62304372e-01 1.44926563e-01 -6.24004662e-01 1.40566218e+00
3.26105654e-01 7.12283373e-01 -1.09942980e-01 1.01535952e+00
9.89853978e-01 1.08617449e+00 -4.42112297e-01 -6.63331509e-01
1.28666270e+00 -7.07615733e-01 -1.02606058e+00 -5.13615943e-02
1.56399608e-01 -1.25618243e+00 7.06658959e-01 6.58886254e-01
-9.13066089e-01 -8.21138442e-01 -1.01331985e+00 2.60211885e-01
6.93028048e-02 1.90558419e-01 9.72555354e-02 3.19988638e-01
-1.00028586e+00 9.54686403e-01 -4.59731668e-01 -3.25444549e-01
5.22539355e-02 6.23249151e-02 -4.66635168e-01 -5.00868440e-01
-1.22276831e+00 6.22320294e-01 1.06518697e-02 5.60884714e-01
-6.24974132e-01 -4.95779455e-01 -6.07686877e-01 -5.65818213e-02
1.20906144e-01 -6.27056599e-01 5.69017887e-01 -1.12094259e+00
-1.12649012e+00 3.85592580e-01 -2.67907858e-01 7.37588182e-02
3.76346201e-01 4.49363999e-02 -4.60893571e-01 3.32430750e-01
2.63643295e-01 -5.53298146e-02 1.14080679e+00 -1.65078318e+00
-4.74825263e-01 -3.65809321e-01 -4.96939242e-01 2.38803014e-01
-2.78132617e-01 -3.98709700e-02 -5.65819621e-01 -8.00083756e-01
9.02744353e-01 -6.69400930e-01 -7.22886145e-01 6.12538233e-02
-1.17335282e-02 -1.41046196e-01 7.97747970e-01 -1.03007603e+00
1.39779162e+00 -2.41463089e+00 1.06288329e-01 6.63896501e-01
1.17427178e-01 9.92600843e-02 -2.74084568e-01 3.27284932e-01
-1.87325612e-01 8.63889884e-03 -2.47714862e-01 6.72498811e-03
-4.33774710e-01 2.88864970e-01 -3.16090882e-02 5.53930640e-01
-2.46893078e-01 6.74446449e-02 -5.76602101e-01 -6.12303913e-01
-1.19036818e-02 6.17085099e-01 -4.05016273e-01 2.00260743e-01
4.81707633e-01 2.79270619e-01 -4.80693460e-01 3.74792248e-01
1.34236228e+00 2.20036611e-01 -2.97632694e-01 -1.06797707e+00
-1.24760710e-01 -4.83656913e-01 -2.23150992e+00 1.53692985e+00
-1.39499158e-01 2.19105482e-01 6.47862315e-01 -1.27191973e+00
1.11163330e+00 1.74085662e-01 4.97846991e-01 -5.08991420e-01
1.26902878e-01 4.89921838e-01 -2.30188057e-01 -7.68127739e-01
-1.02205031e-01 -1.92774057e-01 5.85224986e-01 -7.68779069e-02
-2.32587680e-01 8.61971155e-02 1.76131979e-01 -3.24026402e-03
1.09682178e+00 -2.03200281e-01 -4.94080409e-02 -6.07741654e-01
9.16833699e-01 -2.58827895e-01 1.03134251e+00 7.31815517e-01
2.38952320e-02 1.12200022e+00 1.93877235e-01 -4.70846206e-01
-9.48419869e-01 -7.93102980e-01 -1.70828581e-01 5.38331747e-01
5.67300081e-01 -3.51426154e-01 -7.75003731e-01 -5.69134243e-02
-4.34740871e-01 -4.68940660e-02 -4.64311361e-01 -2.51436271e-02
-7.79056728e-01 -8.89727592e-01 1.97157003e-02 1.46254495e-01
1.00049031e+00 -6.63478792e-01 3.18630189e-01 3.85007262e-01
-5.12618840e-01 -7.10371912e-01 -5.86547971e-01 -2.40905657e-01
-1.02987027e+00 -1.05371749e+00 -9.11846757e-01 -1.22571194e+00
9.64806020e-01 7.53785193e-01 7.88715780e-01 4.60003138e-01
-2.15939894e-01 2.40881354e-01 -4.77573127e-01 2.63985038e-01
-2.80438453e-01 -5.09499609e-01 -2.90196817e-02 4.63601321e-01
-1.07069679e-01 -9.54978824e-01 -8.39189172e-01 6.80067420e-01
-1.18392789e+00 -3.11505586e-01 8.28387380e-01 9.10232127e-01
1.19711185e+00 9.03750300e-01 2.36533403e-01 -4.35355932e-01
8.28894675e-01 -2.30248675e-01 -4.21503931e-01 1.30093992e-01
-4.70157951e-01 -1.32452011e-01 7.29551375e-01 -6.52807713e-01
-8.25098991e-01 1.04471460e-01 -2.96669751e-01 -5.05645871e-01
2.37123929e-02 7.15547502e-01 -4.21422422e-01 -4.06407505e-01
6.82632506e-01 5.85796297e-01 3.46273780e-01 -8.06555808e-01
1.12245493e-01 4.80265319e-01 4.92729515e-01 -5.59720933e-01
1.07994211e+00 5.18568277e-01 3.28572035e-01 -1.01408517e+00
-3.88679445e-01 -7.40554810e-01 -3.77765656e-01 -1.35811269e-01
7.89808393e-01 -9.81921732e-01 -5.47724605e-01 6.54085517e-01
-1.11202145e+00 1.98941216e-01 9.80646908e-02 5.26026785e-01
-3.37317884e-01 1.05647671e+00 -8.39762807e-01 -5.71473062e-01
-3.56022924e-01 -1.25942135e+00 6.71307027e-01 1.23073846e-01
3.78602058e-01 -6.83883429e-01 9.96391401e-02 3.95031795e-02
4.08634722e-01 1.41480789e-01 6.20471060e-01 -6.20865971e-02
-5.99999130e-01 -2.80357301e-01 -2.58067578e-01 7.32635677e-01
4.24826555e-02 -1.63895607e-01 -3.49301100e-01 -4.45060611e-01
6.14626527e-01 2.71751612e-01 8.75718117e-01 5.83190978e-01
1.02263522e+00 -5.19783258e-01 -3.03223640e-01 5.71569979e-01
1.75932360e+00 5.93509823e-02 1.21020663e+00 5.07764578e-01
7.18183100e-01 5.84389448e-01 7.28002250e-01 1.20379351e-01
-4.45540249e-02 6.20922208e-01 3.04610491e-01 -2.57437408e-01
-1.94947898e-01 -9.67648104e-02 2.00407922e-01 1.25865424e+00
-4.89760190e-02 2.34813079e-01 -5.83665609e-01 5.25917947e-01
-2.28360009e+00 -1.06455231e+00 -7.17727721e-01 2.23497558e+00
6.26120090e-01 -1.84610710e-01 -1.66881204e-01 4.54319835e-01
9.85847235e-01 3.49285640e-02 7.83714950e-02 -4.58609499e-02
-3.97294164e-01 -1.06255941e-01 4.62374151e-01 6.78462088e-01
-1.07541370e+00 4.17867631e-01 5.62661934e+00 1.47252333e+00
-8.80150378e-01 2.02557445e-01 6.31716251e-01 6.13674045e-01
-3.01318735e-01 2.38579378e-01 -5.41406810e-01 5.29538572e-01
1.27850160e-01 1.15569875e-01 5.60515046e-01 5.51815867e-01
4.23584849e-01 -1.67150691e-01 -4.85705733e-01 1.38349164e+00
2.69994199e-01 -1.05182350e+00 1.49832189e-01 -1.43791795e-01
8.80493701e-01 -3.44286829e-01 -2.01231897e-01 -4.69688252e-02
-3.13929409e-01 -8.36697102e-01 2.55076647e-01 8.84834945e-01
-6.24795705e-02 -7.77770758e-01 1.02061439e+00 5.11074185e-01
-1.38999176e+00 4.86401692e-02 -1.02707660e+00 1.84428200e-01
-3.67954560e-02 1.08442283e+00 3.85300845e-01 7.09283650e-01
1.22282171e+00 8.17135334e-01 -4.84612674e-01 1.61382663e+00
-1.46498473e-03 4.66881692e-01 -5.15710950e-01 2.58675456e-01
1.28936306e-01 -7.83656120e-01 7.52996147e-01 9.95034039e-01
7.45500088e-01 4.30002540e-01 3.83506894e-01 6.49414241e-01
3.81394684e-01 7.29008317e-01 -4.43697751e-01 7.87548125e-01
2.12070331e-01 1.29177558e+00 -5.45260310e-01 -1.41307533e-01
-3.79512876e-01 1.07132447e+00 -1.66259035e-01 5.07106543e-01
-3.21208179e-01 -4.66533452e-01 3.37665677e-01 3.57551605e-01
2.62649536e-01 -2.30836511e-01 -2.39547983e-01 -1.02599490e+00
1.87363803e-01 -1.27164781e+00 2.94330001e-01 -8.06009948e-01
-1.58773649e+00 4.68286216e-01 -2.57462829e-01 -1.70490992e+00
4.42249805e-01 -3.69519770e-01 -8.04713905e-01 8.31807017e-01
-1.23288763e+00 -1.04921460e+00 -5.32961249e-01 8.84998322e-01
2.96005458e-01 -2.04155087e-01 4.20508623e-01 6.65605068e-01
-3.70773673e-01 2.69121945e-01 5.89538634e-01 -3.22066210e-02
7.64034033e-01 -7.66904652e-01 -4.73293811e-01 1.05151558e+00
-1.66244015e-01 6.73392713e-01 8.24277401e-01 -7.78450251e-01
-1.34060156e+00 -7.25536942e-01 5.92948258e-01 2.18964234e-01
4.40588683e-01 1.45770773e-01 -1.29435670e+00 3.87007855e-02
9.52937827e-02 2.29044646e-01 3.80531102e-01 -1.09593213e-01
-2.40851894e-01 -6.41767859e-01 -1.14167237e+00 5.64293027e-01
7.07596898e-01 -1.73069596e-01 -3.66826087e-01 4.45051342e-01
4.15884316e-01 -1.73534825e-01 -9.79203284e-01 7.13757932e-01
2.63138354e-01 -1.10385799e+00 1.21306980e+00 4.92726490e-02
2.40166768e-01 -1.05241477e+00 -3.77334833e-01 -1.23885930e+00
-9.49262798e-01 -4.92058843e-01 3.92317355e-01 1.47581089e+00
1.22749507e-01 -5.00767291e-01 4.45651114e-01 8.72576535e-02
-1.39430076e-01 -7.31995642e-01 -1.08527958e+00 -9.20296907e-01
8.35481938e-03 -5.77554815e-02 2.28141487e-01 8.98921251e-01
-4.22164977e-01 2.04196587e-01 -5.90682805e-01 6.60461247e-01
9.61103916e-01 1.00789927e-01 7.04246223e-01 -1.19568419e+00
-3.39711726e-01 -3.78292739e-01 -5.90304613e-01 -1.19797707e+00
-1.52393952e-01 -4.01925445e-01 2.03875333e-01 -1.97872508e+00
3.43566179e-01 -3.29953790e-01 -4.28196341e-01 1.55398101e-01
-2.98850387e-01 3.83820415e-01 3.73000428e-02 5.48466444e-01
-4.43750411e-01 5.00694036e-01 1.40119052e+00 -3.96180183e-01
-1.95917681e-01 5.49512655e-02 -6.49850130e-01 6.28324807e-01
4.52596128e-01 -4.23629820e-01 -1.39568359e-01 -4.25611585e-01
2.10118145e-01 -2.39168927e-02 2.47542277e-01 -1.27096915e+00
5.36415339e-01 -6.04372956e-02 4.39997196e-01 -5.30260384e-01
1.44812822e-01 -1.21602535e+00 3.37909102e-01 4.42795277e-01
1.67829946e-01 -1.11405246e-01 5.90404915e-03 7.58296669e-01
-6.36867046e-01 -4.72026885e-01 1.13199937e+00 -2.34378785e-01
-5.75604916e-01 4.00172591e-01 -5.30767083e-01 -3.21515113e-01
7.23391116e-01 -4.17804360e-01 1.53602157e-02 -6.22555077e-01
-7.91352689e-01 8.92683640e-02 2.24874794e-01 1.88495666e-01
7.69982040e-01 -1.73308933e+00 -1.07507741e+00 1.97035059e-01
-1.32351741e-01 -1.65424883e-01 6.16686046e-01 9.66307044e-01
-7.33076692e-01 -4.13806707e-01 -4.44243550e-02 -6.98762178e-01
-1.46342123e+00 5.52087188e-01 4.39672559e-01 -1.04428418e-01
-5.00363290e-01 8.09132874e-01 2.70568520e-01 -3.14203680e-01
1.09517165e-01 -1.24857947e-02 -4.00400072e-01 -1.30358934e-01
6.56592131e-01 5.45170784e-01 1.17893610e-02 -9.55040634e-01
-1.27594009e-01 1.31591427e+00 2.34868929e-01 1.38361171e-01
1.34695780e+00 -3.00984651e-01 -9.08194304e-01 -4.91241086e-03
1.51725876e+00 1.54483184e-01 -7.14110136e-01 -3.05693924e-01
-5.08742511e-01 -7.99967885e-01 4.13329005e-01 -1.96370170e-01
-1.28715038e+00 6.69799030e-01 1.29777896e+00 2.15949982e-01
1.44998372e+00 -5.11891782e-01 9.21799123e-01 3.00579131e-01
2.27998838e-01 -1.19722164e+00 2.55817652e-01 4.56340104e-01
1.08338869e+00 -1.02052653e+00 2.23028481e-01 -8.92959177e-01
-1.15258925e-01 1.11133730e+00 6.32610202e-01 -5.92700005e-01
1.08071244e+00 6.71240836e-02 1.66852534e-01 -6.45794393e-03
-3.30102853e-02 -1.24505274e-01 3.63819718e-01 6.19181871e-01
2.45782554e-01 -4.13506985e-01 -1.08042574e+00 7.21928060e-01
2.32856914e-01 -3.12587857e-01 2.62694299e-01 5.47385633e-01
-9.12832499e-01 -1.28590000e+00 -1.07764447e+00 3.43255073e-01
-4.48370099e-01 -4.36617061e-02 1.70004457e-01 1.84010774e-01
3.82986426e-01 1.27935135e+00 -3.30020368e-01 -3.88721257e-01
5.31296074e-01 -5.66552520e-01 2.80348063e-01 -1.57713652e-01
-3.03836942e-01 6.61977470e-01 -2.93072790e-01 -5.22237122e-01
-3.77988786e-01 -4.34332788e-01 -8.92227173e-01 -1.47405595e-01
-3.64700973e-01 3.77359867e-01 4.27266181e-01 6.40507758e-01
1.45598784e-01 2.28221267e-01 7.98153877e-01 -8.66241693e-01
-4.52692270e-01 -8.42872918e-01 -8.53951693e-01 5.95601499e-01
1.33854717e-01 -3.61489385e-01 -7.10662782e-01 -2.75885146e-02]
|
[11.326695442199707, -2.4338443279266357]
|
3bc7f9a1-8e60-4aec-a60b-eb4b87617961
|
question-relevance-in-vqa-identifying-non
|
1606.06622
| null |
http://arxiv.org/abs/1606.06622v3
|
http://arxiv.org/pdf/1606.06622v3.pdf
|
Question Relevance in VQA: Identifying Non-Visual And False-Premise Questions
|
Visual Question Answering (VQA) is the task of answering natural-language
questions about images. We introduce the novel problem of determining the
relevance of questions to images in VQA. Current VQA models do not reason about
whether a question is even related to the given image (e.g. What is the capital
of Argentina?) or if it requires information from external resources to answer
correctly. This can break the continuity of a dialogue in human-machine
interaction. Our approaches for determining relevance are composed of two
stages. Given an image and a question, (1) we first determine whether the
question is visual or not, (2) if visual, we determine whether the question is
relevant to the given image or not. Our approaches, based on LSTM-RNNs, VQA
model uncertainty, and caption-question similarity, are able to outperform
strong baselines on both relevance tasks. We also present human studies showing
that VQA models augmented with such question relevance reasoning are perceived
as more intelligent, reasonable, and human-like.
|
['Arijit Ray', 'Mohit Bansal', 'Dhruv Batra', 'Gordon Christie', 'Devi Parikh']
|
2016-06-21
|
question-relevance-in-vqa-identifying-non-1
|
https://aclanthology.org/D16-1090
|
https://aclanthology.org/D16-1090.pdf
|
emnlp-2016-11
|
['question-similarity']
|
['natural-language-processing']
|
[ 2.50341952e-01 5.30913174e-01 1.79358989e-01 -6.16838813e-01
-1.13240421e+00 -7.69247055e-01 7.71501124e-01 4.55376357e-01
-4.17231441e-01 6.08626604e-01 5.44038355e-01 -7.23566055e-01
1.31051019e-01 -5.96339047e-01 -8.39747190e-01 -1.10604048e-01
5.85974574e-01 7.21639276e-01 3.44879895e-01 -3.53279591e-01
3.81076723e-01 2.45045289e-01 -1.37059677e+00 6.15747809e-01
5.81815898e-01 1.26208556e+00 3.66182089e-01 9.07851398e-01
-3.13836634e-01 1.47667980e+00 -5.73080897e-01 -8.54969501e-01
-8.49763975e-02 -6.03585184e-01 -1.83720541e+00 1.93479568e-01
9.04603720e-01 -7.00351894e-01 -1.37398168e-01 1.07720149e+00
1.48585998e-02 2.65953541e-01 8.39081943e-01 -1.36946404e+00
-1.44481754e+00 1.75647587e-01 -1.34331554e-01 4.30306524e-01
6.71011984e-01 5.19612491e-01 1.28534627e+00 -9.44841802e-01
7.31562376e-01 1.56974399e+00 4.72279042e-02 7.74667323e-01
-8.65568161e-01 4.91677821e-02 3.22295099e-01 6.77068293e-01
-9.23042417e-01 -4.11199808e-01 4.91666824e-01 -4.35933918e-01
9.18600023e-01 5.85857093e-01 2.91589916e-01 1.06237912e+00
2.86408305e-01 7.59122133e-01 9.13425386e-01 -3.20583612e-01
3.12779427e-01 1.92900747e-01 1.98139846e-01 4.93372411e-01
-5.52883446e-02 -1.69223219e-01 -2.53394514e-01 -9.68168378e-02
4.96685386e-01 -5.46449661e-01 -3.00443709e-01 -2.49147877e-01
-1.29088473e+00 1.05049324e+00 8.16099763e-01 2.14227036e-01
-5.35772324e-01 3.78887296e-01 2.97371060e-01 3.54066521e-01
1.44929141e-02 9.91492510e-01 -1.80508181e-01 1.47849888e-01
-4.88496900e-01 2.72468388e-01 7.22080171e-01 8.13872159e-01
6.86015785e-01 -1.93606049e-01 -7.21140683e-01 2.99232006e-01
5.40246189e-01 6.01691127e-01 1.67044014e-01 -1.57643163e+00
3.94806772e-01 4.93002832e-01 5.19311309e-01 -1.08553219e+00
-7.94693902e-02 2.42543355e-01 -3.53322387e-01 2.04305440e-01
6.97840393e-01 1.15606263e-02 -1.00947118e+00 1.76804423e+00
3.37463439e-01 -5.39557874e-01 3.21179211e-01 1.33588398e+00
1.44851089e+00 8.13903928e-01 5.34813523e-01 -1.18986913e-03
2.02447534e+00 -1.03436482e+00 -7.99430072e-01 -7.26364374e-01
2.78075207e-02 -9.15941596e-01 1.29694843e+00 -1.18992344e-01
-1.33275437e+00 -7.10623443e-01 -7.12747276e-01 -8.31452489e-01
-4.01172101e-01 -1.23991907e-01 2.34024808e-01 3.46549809e-01
-1.51155996e+00 8.21933746e-02 2.60620452e-02 -6.30798578e-01
2.07229704e-01 -6.93429708e-02 -3.21523517e-01 -2.24688038e-01
-1.54946744e+00 1.29855621e+00 -3.37749865e-04 2.25446627e-01
-1.30677080e+00 -8.31786394e-02 -1.13913596e+00 1.75039575e-01
4.20281589e-01 -1.23425770e+00 1.71250796e+00 -1.47122693e+00
-8.77835155e-01 1.27444589e+00 -5.25918782e-01 -4.73597080e-01
3.43331605e-01 -1.99800760e-01 -2.86733836e-01 7.71120310e-01
4.73801613e-01 1.20120215e+00 1.06864715e+00 -1.55047190e+00
-7.07575440e-01 -3.69513482e-01 9.35020685e-01 4.97235596e-01
2.76979029e-01 7.06121251e-02 -5.36079943e-01 -9.59531441e-02
-1.56290382e-01 -6.27435029e-01 -1.70046583e-01 5.31623125e-01
-5.64376973e-02 -5.72169721e-01 6.50306582e-01 -1.06474829e+00
5.52553654e-01 -1.71455479e+00 -1.86220288e-01 -1.67590022e-01
3.45411003e-01 1.95533350e-01 -4.42098767e-01 2.84857124e-01
1.40090019e-01 4.34320599e-01 1.25856265e-01 1.23415872e-01
1.39201194e-01 3.98946971e-01 -5.34864008e-01 1.93650588e-01
5.88941336e-01 1.38957584e+00 -1.21610498e+00 -8.28702509e-01
1.34275019e-01 3.23546290e-01 8.96385685e-02 4.82125193e-01
-7.35264063e-01 3.13122541e-01 -3.75456542e-01 3.74336988e-01
4.75417376e-01 -7.10924506e-01 -1.86120883e-01 -5.03942966e-01
3.22973460e-01 2.49953121e-01 -5.44579268e-01 1.23121023e+00
-4.08143789e-01 1.03806043e+00 1.66687444e-01 -5.78682005e-01
7.66295373e-01 4.16070551e-01 -3.42170328e-01 -1.12348962e+00
1.70910105e-01 -7.23654777e-02 -1.69085279e-01 -1.03995097e+00
6.14453435e-01 -2.77463924e-02 7.46832937e-02 5.31218767e-01
-1.10830210e-01 -5.52992880e-01 1.31405622e-01 5.96349418e-01
7.07321525e-01 1.45499051e-01 3.42805296e-01 -1.44323677e-01
8.17302763e-01 1.94816738e-01 -7.86215514e-02 9.21407402e-01
-8.00189912e-01 7.32840657e-01 7.39937723e-01 -3.88826519e-01
-1.04137194e+00 -1.26825893e+00 4.08529937e-01 1.22349393e+00
6.54217541e-01 5.59413955e-02 -7.41587281e-01 -8.11214507e-01
-1.63864508e-01 1.08879685e+00 -1.04081273e+00 -4.39606085e-02
-1.90594062e-01 3.79666865e-01 3.21786031e-02 4.86628443e-01
4.95949268e-01 -1.41646075e+00 -8.96348178e-01 -3.04614872e-01
-8.41743946e-01 -1.22885585e+00 -7.10045099e-01 -2.85236061e-01
-4.79103893e-01 -1.07056892e+00 -7.28902936e-01 -9.37162697e-01
5.85465312e-01 3.44956756e-01 1.83620620e+00 3.11256289e-01
1.11898430e-01 1.11238980e+00 -3.58882606e-01 -2.92283177e-01
-5.26466966e-01 -4.17987108e-01 -5.61831415e-01 -1.43291250e-01
4.05450284e-01 2.74028629e-01 -8.95676136e-01 3.27329069e-01
-9.43737209e-01 -1.77978605e-01 4.45436954e-01 5.65564215e-01
4.69068944e-01 -6.29791915e-01 6.55455053e-01 -3.66907120e-01
1.07373834e+00 -4.97033358e-01 -1.82286337e-01 8.61774802e-01
-2.65237391e-01 2.79130578e-01 1.46280527e-01 -1.30907610e-01
-1.26265597e+00 -3.14119793e-02 -3.54727149e-01 -2.43390203e-01
-3.40284497e-01 2.65266389e-01 -1.38938010e-01 1.16044223e-01
8.93765330e-01 1.42302085e-02 -5.43770306e-02 3.62020940e-01
8.00454378e-01 4.23843503e-01 8.57325256e-01 -5.64642787e-01
5.96613586e-01 6.01101875e-01 -1.78640142e-01 -6.41689181e-01
-1.07328272e+00 -6.47301972e-01 -2.89238006e-01 -5.68021476e-01
1.50251830e+00 -7.53855467e-01 -1.11861897e+00 -2.94904053e-01
-1.72994995e+00 -1.21014372e-01 -3.97771269e-01 6.02085702e-02
-6.32089019e-01 6.40365660e-01 -4.40670371e-01 -9.63535786e-01
-6.00526094e-01 -1.12443662e+00 1.06632602e+00 3.75988215e-01
-4.00188208e-01 -1.01970065e+00 -2.94335783e-01 1.02915573e+00
5.28609335e-01 3.30214500e-02 1.07436180e+00 -5.39611936e-01
-6.81611240e-01 1.51721686e-01 -7.83452988e-01 1.94748253e-01
1.03838004e-01 -2.23658442e-01 -9.80854154e-01 4.70714197e-02
1.67307645e-01 -8.63195717e-01 5.94877362e-01 2.92951226e-01
7.94764161e-01 -4.87650931e-01 5.73899411e-02 -2.27752715e-01
1.16482091e+00 1.31252363e-01 8.13442767e-01 2.36865088e-01
3.56662333e-01 1.06281066e+00 9.59826291e-01 -8.54275841e-03
9.91284549e-01 4.03690934e-01 1.07166350e+00 -3.48803997e-01
-2.86019862e-01 -2.61728466e-01 1.71994030e-01 -7.47718988e-03
1.78714380e-01 -4.89915937e-01 -9.51314747e-01 1.09794092e+00
-1.87900198e+00 -9.85978663e-01 -5.28958619e-01 1.85653090e+00
8.28600347e-01 -3.25400501e-01 -5.29845171e-02 -5.73942304e-01
7.48386502e-01 1.17211185e-01 -7.84895897e-01 -8.48937690e-01
-1.73961334e-02 -2.58456647e-01 -8.87726620e-02 7.42718160e-01
-8.84788394e-01 9.19799507e-01 6.28918362e+00 2.75363952e-01
-6.96932018e-01 1.59815416e-01 1.16040218e+00 3.60669285e-01
-7.81729102e-01 8.88412446e-02 -2.98486531e-01 -1.53591767e-01
5.55132985e-01 -2.00702846e-02 2.23307297e-01 5.92450559e-01
3.33561525e-02 -4.57488298e-01 -1.35727060e+00 9.63427365e-01
6.62971556e-01 -1.21060979e+00 5.49868882e-01 -4.21978056e-01
4.74535435e-01 -3.30464214e-01 2.57101148e-01 3.22387397e-01
4.62481022e-01 -1.41259217e+00 9.62630212e-01 6.73229516e-01
4.63719040e-01 -4.75997716e-01 9.03487682e-01 1.76092550e-01
-9.71876562e-01 1.98285356e-01 -5.67759931e-01 -9.07771364e-02
3.84287566e-01 -2.75759697e-02 -1.02594805e+00 4.30744708e-01
8.91314149e-01 -4.96379770e-02 -1.09717095e+00 7.24133849e-01
-7.21419215e-01 1.19033009e-01 2.61902124e-01 -2.49737740e-01
4.94593650e-01 1.24821305e-01 3.08627158e-01 6.65347040e-01
-1.60559528e-02 2.27407917e-01 -2.16599807e-01 1.13856077e+00
-1.38096780e-01 1.31239429e-01 -4.90824401e-01 3.50487232e-02
-7.55101023e-03 1.15161633e+00 -4.29333955e-01 -4.65143532e-01
-4.20652956e-01 1.21125782e+00 3.53735834e-01 6.35098338e-01
-6.38132870e-01 -2.35297278e-01 3.29590470e-01 -5.83242886e-02
2.45407075e-01 2.39672095e-01 -1.61393397e-02 -8.03403497e-01
1.70571387e-01 -1.05707800e+00 7.42675781e-01 -1.80107892e+00
-1.39411235e+00 9.73574519e-01 -3.04052621e-01 -6.85523450e-01
-5.11866629e-01 -7.46901572e-01 -5.01478672e-01 1.02062368e+00
-1.78234434e+00 -1.41691828e+00 -5.36110640e-01 6.66928828e-01
6.09750748e-01 3.21543604e-01 6.35770679e-01 -3.85088444e-01
3.04180413e-01 7.37847909e-02 -5.68241954e-01 9.86694619e-02
8.63308430e-01 -1.32844007e+00 6.86572134e-01 8.17462087e-01
1.65126026e-01 5.04653811e-01 9.43282008e-01 -6.15826786e-01
-1.09067953e+00 -8.38474393e-01 1.27649117e+00 -1.11007464e+00
4.67861950e-01 -1.09747134e-01 -1.05177701e+00 6.10979617e-01
1.00353312e+00 -1.13055944e-01 4.11433220e-01 -1.39247447e-01
-6.01466060e-01 1.49206221e-01 -1.28763247e+00 7.82399237e-01
5.37034452e-01 -9.66987193e-01 -1.28271890e+00 4.96463180e-01
1.04599082e+00 -9.50835198e-02 -5.20815194e-01 1.72761783e-01
4.15311396e-01 -8.60683441e-01 1.11075866e+00 -8.75423491e-01
6.95894420e-01 -3.96194518e-01 -3.02351445e-01 -9.90804613e-01
-2.90712804e-01 -3.12111080e-01 1.30868420e-01 1.07948077e+00
5.99375129e-01 -6.34883121e-02 4.29850549e-01 1.16230083e+00
2.59223849e-01 -2.64534801e-01 -1.00756776e+00 -3.63401562e-01
2.09107995e-01 -2.77271569e-01 5.76240659e-01 7.28937089e-01
-3.38100381e-02 8.83529186e-01 -3.40650320e-01 1.93149135e-01
2.91636139e-01 1.84990689e-01 5.18182158e-01 -1.00779819e+00
-3.46435681e-02 -3.28004092e-01 -6.05598651e-02 -1.06048059e+00
1.54256389e-01 -4.64743346e-01 4.02912110e-01 -2.51967311e+00
2.38606289e-01 3.75069499e-01 1.41977191e-01 3.44554424e-01
-6.10168457e-01 5.57520054e-02 4.78648365e-01 5.75260036e-02
-1.06866288e+00 3.64400208e-01 1.58114529e+00 -4.89269316e-01
2.64794260e-01 -2.50347733e-01 -9.22772348e-01 5.72364688e-01
6.91021502e-01 1.08105607e-01 -6.40153646e-01 -8.12723160e-01
7.50048578e-01 5.46359241e-01 6.67537868e-01 -5.80287158e-01
3.08928639e-01 -2.83406496e-01 4.37038988e-01 -3.78604323e-01
5.04718125e-01 -9.26568568e-01 -3.84022504e-01 4.40912694e-01
-6.48813546e-01 4.19571370e-01 3.17553133e-01 5.77156365e-01
-2.15372711e-01 -6.17005408e-01 4.98782128e-01 -4.18347359e-01
-1.09567177e+00 1.16390571e-01 -5.66808045e-01 4.08895046e-01
9.58858371e-01 -2.66023353e-02 -7.68480361e-01 -1.24042332e+00
-6.76260293e-01 6.86542869e-01 4.12684739e-01 6.00830853e-01
1.08653784e+00 -1.25218165e+00 -9.34758842e-01 -6.40133262e-01
5.96166015e-01 -4.24508542e-01 4.18623894e-01 2.78225929e-01
-5.12887716e-01 6.94329441e-01 -1.86791003e-01 -5.27858019e-01
-1.24935055e+00 9.55792189e-01 3.91729563e-01 -3.13819274e-02
6.36534914e-02 9.89444613e-01 6.59050047e-01 -3.26289982e-01
1.66936263e-01 -2.61196196e-01 -7.67062962e-01 1.13689005e-01
6.96856022e-01 -1.80417493e-01 -3.18416357e-01 -1.02401137e+00
-4.54987884e-01 4.61469471e-01 3.67854275e-02 -3.95270109e-01
5.46285689e-01 -5.36736310e-01 -1.92937076e-01 3.12346369e-01
9.40924466e-01 -3.85391802e-01 -1.10188079e+00 -1.78407520e-01
-1.15228727e-01 -3.92632097e-01 -2.32353434e-01 -1.08914196e+00
-8.33267093e-01 1.17275500e+00 5.86429119e-01 2.86723971e-01
9.34204519e-01 6.84281051e-01 5.23647845e-01 5.68579435e-01
3.83800939e-02 -9.00900602e-01 5.68873525e-01 4.20734137e-01
1.58207870e+00 -1.64411938e+00 -9.50359330e-02 -4.13392447e-02
-1.25543904e+00 1.01386881e+00 9.49189544e-01 3.66655439e-01
3.18513811e-02 -6.81419551e-01 5.42493045e-01 -5.99708557e-01
-1.01974034e+00 -6.12768888e-01 6.87569082e-01 8.54573548e-01
2.13720277e-01 -1.75583258e-01 3.71579751e-02 2.22818479e-01
-2.90297829e-02 -2.95680791e-01 6.25393212e-01 5.16484439e-01
-6.51024342e-01 -4.30735290e-01 -5.32145321e-01 2.89854467e-01
-3.20875585e-01 -2.31046125e-01 -8.60998213e-01 5.77374220e-01
-2.37836033e-01 1.57227015e+00 1.58794299e-01 -7.17191212e-03
3.12183022e-01 -1.02132782e-01 4.04547095e-01 -5.17053604e-01
-6.98339403e-01 -4.86748993e-01 3.19657236e-01 -4.94343400e-01
-4.54572052e-01 -4.11894679e-01 -1.10257125e+00 -1.28206797e-02
-1.51879832e-01 2.93474525e-01 4.26044762e-01 1.06867123e+00
2.22703665e-01 2.63980955e-01 1.16719566e-01 -1.17919572e-01
-4.14342225e-01 -6.03133559e-01 -3.09157092e-02 7.97266543e-01
6.14670753e-01 -3.76093350e-02 -4.23593462e-01 2.39153966e-01]
|
[10.965899467468262, 1.6846423149108887]
|
294f98b5-12fe-45e6-8f1b-094daa605f61
|
nonlinear-supervised-dimensionality-reduction
|
1710.07120
| null |
http://arxiv.org/abs/1710.07120v2
|
http://arxiv.org/pdf/1710.07120v2.pdf
|
Nonlinear Supervised Dimensionality Reduction via Smooth Regular Embeddings
|
The recovery of the intrinsic geometric structures of data collections is an
important problem in data analysis. Supervised extensions of several manifold
learning approaches have been proposed in the recent years. Meanwhile, existing
methods primarily focus on the embedding of the training data, and the
generalization of the embedding to initially unseen test data is rather
ignored. In this work, we build on recent theoretical results on the
generalization performance of supervised manifold learning algorithms.
Motivated by these performance bounds, we propose a supervised manifold
learning method that computes a nonlinear embedding while constructing a smooth
and regular interpolation function that extends the embedding to the whole data
space in order to achieve satisfactory generalization. The embedding and the
interpolator are jointly learnt such that the Lipschitz regularity of the
interpolator is imposed while ensuring the separation between different
classes. Experimental results on several image data sets show that the proposed
method outperforms traditional classifiers and the supervised dimensionality
reduction algorithms in comparison in terms of classification accuracy in most
settings.
|
['Elif Vural', 'Cem Ornek']
|
2017-10-19
| null | null | null | null |
['supervised-dimensionality-reduction']
|
['computer-vision']
|
[-9.04375017e-02 1.10893875e-01 -3.98362517e-01 -4.27675933e-01
-4.59623218e-01 -3.85408610e-01 4.21425194e-01 2.45736584e-01
-2.16824085e-01 4.89684016e-01 -5.50838187e-02 1.05552711e-01
-5.70155144e-01 -7.20134258e-01 -4.73605096e-01 -9.51680541e-01
-2.00707048e-01 1.87917829e-01 -6.29789829e-02 5.51127791e-02
3.92124981e-01 6.51273966e-01 -1.56727958e+00 -3.35405946e-01
9.88658845e-01 9.73723292e-01 7.78184831e-02 2.54681796e-01
-1.27522618e-01 2.14007437e-01 -2.06715137e-01 4.58097905e-02
4.76836145e-01 -1.94619238e-01 -7.37884045e-01 6.90085053e-01
4.12923485e-01 -1.58282429e-01 -4.27957237e-01 1.29524505e+00
1.86876789e-01 3.65522265e-01 1.00924933e+00 -1.33117723e+00
-8.65922809e-01 1.78208146e-02 -4.76253957e-01 -9.31834951e-02
-1.44246118e-02 -6.14132524e-01 9.39209580e-01 -1.24028599e+00
4.58266199e-01 1.08336258e+00 5.11200786e-01 4.60691988e-01
-1.32151496e+00 -1.65517509e-01 -5.60543686e-02 2.43182763e-01
-1.69057059e+00 -1.84405178e-01 1.28917682e+00 -5.53956032e-01
1.67163625e-01 1.89586267e-01 1.88619450e-01 3.33272964e-01
-9.71232504e-02 6.86335206e-01 8.07345748e-01 -7.44438171e-01
4.82574791e-01 3.90345156e-01 3.05543602e-01 7.90011048e-01
2.57749707e-01 -1.13810174e-01 -1.51709333e-01 -2.93828547e-01
9.11769271e-01 3.49257678e-01 -3.92654091e-01 -1.00924432e+00
-1.09891570e+00 1.06614769e+00 4.56092596e-01 4.56667393e-01
-1.84462532e-01 -3.68584245e-01 1.66708738e-01 3.46357703e-01
6.31459415e-01 3.01181376e-01 -1.21893078e-01 3.60698014e-01
-7.25960374e-01 -1.27030894e-01 6.33053422e-01 1.01490831e+00
1.09257340e+00 -1.47412062e-01 4.26884443e-01 9.20375884e-01
1.60326660e-01 2.73023486e-01 5.62060356e-01 -8.02728117e-01
4.17353421e-01 9.34238493e-01 1.71191484e-01 -1.26595771e+00
-2.93847769e-01 -1.09923340e-01 -8.70871007e-01 1.93975329e-01
4.57585603e-01 9.90070403e-02 -4.41345692e-01 1.36456621e+00
6.54300392e-01 3.61638814e-01 1.60163820e-01 8.83673668e-01
3.30098122e-01 5.80052972e-01 -3.18729758e-01 -2.35408992e-01
7.74935126e-01 -6.27314270e-01 -7.91880906e-01 4.35829461e-01
8.82647514e-01 -5.76156557e-01 1.03862858e+00 3.91641259e-01
-7.07437634e-01 -7.45638430e-01 -1.27427328e+00 1.70364534e-03
-2.85099477e-01 4.89864647e-01 3.64426225e-01 5.23287654e-01
-7.70265460e-01 8.60422194e-01 -9.52832997e-01 -3.35979372e-01
2.74990737e-01 4.22300756e-01 -5.42977512e-01 -1.45324051e-01
-6.78384721e-01 5.66317797e-01 4.39650595e-01 2.52210587e-01
-4.53846723e-01 -4.74723399e-01 -9.87290084e-01 -2.51804948e-01
4.72777151e-02 -2.09118515e-01 5.64357936e-01 -7.37907469e-01
-1.27796233e+00 8.09440315e-01 -1.63715839e-01 -1.88684523e-01
3.93924296e-01 -2.01125160e-01 -4.81036335e-01 3.93053353e-01
-6.72277808e-02 3.07116359e-01 1.25149131e+00 -1.21293223e+00
-5.71130872e-01 -7.37880528e-01 -1.62041217e-01 1.52970240e-01
-1.13560343e+00 -4.83127803e-01 8.32851138e-03 -6.72919512e-01
6.45412207e-01 -8.59092474e-01 -2.28368253e-01 3.17158967e-01
-2.31317863e-01 -6.00277185e-01 1.39055729e+00 -5.20062566e-01
1.16250002e+00 -2.50933194e+00 6.13059521e-01 2.44094610e-01
1.94592044e-01 1.69697274e-02 -1.11555360e-01 4.53653783e-01
-1.60725266e-01 -2.00922176e-01 -4.81452286e-01 -4.36454982e-01
-2.00104982e-01 2.61680394e-01 -3.52826536e-01 1.18477237e+00
2.87275314e-01 5.10473490e-01 -8.51077974e-01 -5.34861207e-01
3.51198405e-01 5.68975508e-01 -4.10141289e-01 3.29411894e-01
8.91065001e-02 5.58826327e-01 -6.25444233e-01 4.09092993e-01
7.14618087e-01 -2.05039367e-01 -3.51539887e-02 -3.58539931e-02
5.10107540e-02 -6.43359274e-02 -1.59468246e+00 1.78377807e+00
-2.12117374e-01 4.42472249e-01 2.06443995e-01 -1.65617597e+00
1.41907156e+00 3.07063162e-01 8.88270378e-01 6.91255331e-02
-1.89377051e-02 2.89925545e-01 -4.14450794e-01 -6.84030116e-01
3.20666105e-01 -1.20479748e-01 1.91643283e-01 3.33450317e-01
-1.33627459e-01 7.53334314e-02 -1.21222824e-01 -1.81816995e-01
4.12951410e-01 -5.05682863e-02 1.54510319e-01 -5.99696755e-01
9.40943360e-01 -4.84203883e-02 4.78400707e-01 -3.11765522e-02
-5.00817113e-02 4.34856743e-01 3.79096791e-02 -5.47501206e-01
-1.07539046e+00 -1.02959561e+00 -7.19714701e-01 6.78683817e-01
5.10401011e-01 -1.55892938e-01 -1.02710593e+00 -8.30427110e-01
8.31710771e-02 3.60458136e-01 -6.27529383e-01 -2.09538609e-01
-5.86257160e-01 -4.93560672e-01 6.52290583e-02 5.00498593e-01
5.97610235e-01 -5.38904369e-01 -1.16257139e-01 9.27232876e-02
-1.66174378e-02 -9.52287078e-01 -5.39153218e-01 -2.96942800e-01
-1.47970033e+00 -1.23884416e+00 -6.57515764e-01 -1.43995798e+00
1.22677290e+00 5.04706979e-01 3.74874353e-01 1.96825236e-01
-2.42595628e-01 5.57861745e-01 -2.69667417e-01 -5.12728142e-03
-3.25688660e-01 9.32083651e-02 3.69063854e-01 5.70044875e-01
5.96134245e-01 -5.58754861e-01 -4.03923273e-01 6.58976495e-01
-1.08205891e+00 -5.06013036e-01 2.96036333e-01 9.36278164e-01
7.06803739e-01 5.60310781e-01 7.93398917e-01 -5.00866115e-01
2.58700758e-01 -4.62183148e-01 -8.21348548e-01 1.18372887e-01
-7.33509243e-01 3.20555598e-01 6.91608131e-01 -6.00688159e-01
-6.81470573e-01 2.37066388e-01 3.43485653e-01 -6.15662336e-01
-3.01719993e-01 2.33640626e-01 -4.95781541e-01 -3.09051514e-01
6.62559450e-01 3.42164427e-01 4.73448098e-01 -7.88579762e-01
3.52955192e-01 9.22506630e-01 4.95264590e-01 -5.74919164e-01
1.07763886e+00 9.16736186e-01 2.25745827e-01 -1.30076945e+00
-6.05601311e-01 -7.88152814e-01 -1.25731027e+00 -4.28773798e-02
5.94041288e-01 -5.70158601e-01 -2.12685183e-01 1.08935945e-01
-8.44295919e-01 1.40497372e-01 -4.55249637e-01 8.71371806e-01
-8.01221907e-01 7.36765265e-01 -3.44764233e-01 -7.78524756e-01
-1.01188108e-01 -9.47511852e-01 1.00536180e+00 -1.20240778e-01
3.35397273e-02 -1.59289968e+00 1.22057490e-01 1.05181105e-01
1.59618687e-02 3.97916019e-01 9.75659549e-01 -5.63686073e-01
-4.44516212e-01 -4.22258943e-01 4.85507324e-02 6.41203701e-01
6.30176604e-01 -2.95979440e-01 -8.47959936e-01 -5.68796098e-01
5.49804628e-01 2.17436682e-02 5.81420898e-01 3.15864533e-01
1.17598712e+00 -2.96327710e-01 -4.41506803e-01 5.87138474e-01
1.48705232e+00 -1.07814111e-01 4.95766580e-01 3.16279918e-01
6.60113275e-01 8.85461092e-01 7.17378318e-01 3.23702067e-01
1.31632060e-01 4.39160496e-01 4.57851738e-01 -1.84114780e-02
2.25840077e-01 -2.73080498e-01 1.54977635e-01 9.13640916e-01
9.91014391e-02 5.28074205e-01 -4.99094337e-01 7.55763054e-01
-1.72780621e+00 -6.90310538e-01 -1.46384791e-01 2.75555348e+00
5.76256216e-01 -2.28238121e-01 2.29057923e-01 7.83770800e-01
1.12445390e+00 -1.19314305e-01 -4.62810844e-01 -4.05817702e-02
-3.87582220e-02 -2.28197426e-01 2.81596124e-01 6.04101598e-01
-1.33041346e+00 5.20869017e-01 6.28464794e+00 7.22737253e-01
-1.08205187e+00 -2.23385587e-01 2.12278515e-01 3.30845028e-01
-1.37536108e-01 -1.80675369e-02 -8.62694740e-01 1.94039628e-01
4.74836051e-01 -1.04166962e-01 4.25213009e-01 1.01124442e+00
2.02888325e-01 4.09052372e-01 -1.28640890e+00 1.11803186e+00
1.41800091e-01 -1.06971288e+00 6.31425530e-02 3.68908733e-01
8.53422582e-01 -5.64587951e-01 2.22436801e-01 -7.42856786e-02
-3.75422776e-01 -8.41327608e-01 1.90766186e-01 4.05436575e-01
6.30331457e-01 -8.32935095e-01 4.79718864e-01 6.70712411e-01
-1.18317986e+00 -2.15701327e-01 -7.01779962e-01 -2.33503096e-02
-2.32984096e-01 6.02721930e-01 -8.62207770e-01 4.80689138e-01
4.13744181e-01 1.14354050e+00 -5.90945363e-01 1.04824209e+00
-4.40513641e-02 4.51554567e-01 -2.57956684e-01 7.56182075e-02
8.21275786e-02 -8.14461887e-01 7.41539478e-01 7.35289574e-01
2.67603278e-01 -8.06967840e-02 2.92675376e-01 8.28830779e-01
8.31896812e-02 5.96786499e-01 -7.63588667e-01 1.62021339e-01
3.85439217e-01 1.29282343e+00 -4.63625520e-01 -8.42797607e-02
-4.47156698e-01 9.27776396e-01 3.80783498e-01 3.73220861e-01
-4.14127260e-01 -5.12874842e-01 4.77106959e-01 2.36709803e-01
4.26597506e-01 -5.31601250e-01 -3.03271532e-01 -1.11794758e+00
5.11041939e-01 -2.76738316e-01 5.14124751e-01 -3.80151207e-03
-1.36830401e+00 3.42956185e-01 -1.56641938e-02 -1.74258721e+00
-1.87320143e-01 -7.72517920e-01 -5.93459904e-01 6.48121834e-01
-1.57842016e+00 -9.20064211e-01 -1.97486356e-01 7.75423229e-01
3.74695122e-01 -3.30712855e-01 8.85572493e-01 1.66051298e-01
-3.89176190e-01 4.50417548e-01 5.65289438e-01 4.14719522e-01
5.10339737e-01 -1.31309152e+00 -2.66836017e-01 5.67955494e-01
1.49203315e-01 6.48444176e-01 4.72035229e-01 -2.86503196e-01
-1.44441056e+00 -1.28989863e+00 6.60662889e-01 -3.08744073e-01
6.43968642e-01 -1.89241573e-01 -1.31018484e+00 6.28619969e-01
-2.33794600e-01 3.28743666e-01 7.53820002e-01 -1.05848275e-01
-3.79027009e-01 -2.14168176e-01 -1.25774324e+00 2.56865352e-01
6.55820251e-01 -5.17394722e-01 -7.35074580e-01 4.20286864e-01
4.42396611e-01 7.72359669e-02 -1.37394369e+00 3.00681382e-01
2.65725732e-01 -5.80574453e-01 8.94490898e-01 -7.86485553e-01
-4.57447506e-02 -4.89153028e-01 -3.51978779e-01 -1.13923049e+00
-1.89960241e-01 -6.31297231e-01 -3.26967806e-01 1.23481023e+00
-3.28964740e-02 -7.46734381e-01 9.99673903e-01 5.00308275e-01
1.59901716e-02 -6.14561260e-01 -1.00204873e+00 -1.21186471e+00
3.36521178e-01 1.36826500e-01 3.93642098e-01 1.18282104e+00
2.58461356e-01 2.26194099e-01 -2.06564158e-01 4.42555398e-01
1.16207314e+00 3.20669085e-01 8.16991925e-01 -1.61090851e+00
-7.33945593e-02 -1.57118857e-01 -9.90645051e-01 -1.14729762e+00
4.65829074e-01 -1.07914567e+00 -2.80908704e-01 -1.19192839e+00
-6.00870922e-02 -5.99623382e-01 -1.70620903e-01 -8.55925009e-02
-1.31828904e-01 -1.92894146e-01 -3.38219553e-01 6.59357548e-01
-2.33641356e-01 8.78444493e-01 1.27936721e+00 -9.60706845e-02
-4.92274106e-01 2.56249934e-01 -4.41269398e-01 6.69703364e-01
8.04331243e-01 -2.01886326e-01 -4.94425178e-01 -3.77009362e-01
-5.09046435e-01 -2.81735778e-01 2.28813976e-01 -8.64139378e-01
8.70307982e-02 -1.38944462e-01 1.83644518e-01 -5.12168407e-01
3.38230491e-01 -1.18857312e+00 -2.15423748e-01 2.61495352e-01
-3.88433427e-01 -2.91193992e-01 -1.92359805e-01 9.52521145e-01
-3.76134723e-01 -3.95834416e-01 9.58124220e-01 4.45968747e-01
-4.98619497e-01 5.72572172e-01 2.54069120e-01 -1.40451983e-01
1.23636162e+00 -2.91638911e-01 1.42087221e-01 -1.65062100e-01
-7.33492732e-01 5.77148534e-02 5.70481062e-01 4.62909102e-01
8.61996293e-01 -1.72772765e+00 -6.60515010e-01 4.43087548e-01
3.30539972e-01 1.72782943e-01 -2.44916171e-01 7.59624541e-01
-3.53964984e-01 4.19147581e-01 -5.97903058e-02 -8.16893220e-01
-8.99437904e-01 1.14701259e+00 3.57772082e-01 2.63800561e-01
-8.92689466e-01 1.09719083e-01 3.89715321e-02 -4.15274769e-01
3.79052818e-01 -2.79045343e-01 -3.44289988e-01 -9.96469930e-02
6.76395714e-01 7.43727982e-01 -2.74711754e-02 -8.30467522e-01
-4.32192907e-02 9.17934656e-01 1.55813143e-01 9.94761139e-02
1.36871552e+00 -3.03315312e-01 -2.42455706e-01 5.78716099e-01
1.87604535e+00 -6.38354942e-02 -1.24243605e+00 -7.03309834e-01
4.02141541e-01 -6.30855381e-01 4.30455320e-02 5.05780458e-01
-9.67703819e-01 9.71162200e-01 5.93049824e-01 4.87452418e-01
1.02781236e+00 4.47803773e-02 6.63845181e-01 4.60411966e-01
3.21901798e-01 -1.22537827e+00 1.97902247e-01 -8.19382519e-02
8.04939091e-01 -1.24623096e+00 -1.69567447e-02 -6.95520699e-01
-1.43988714e-01 1.41544282e+00 2.83393264e-01 -6.33093953e-01
1.03271377e+00 -3.35087687e-01 -1.55208305e-01 -1.00090168e-01
-9.12108943e-02 1.28186360e-01 4.48460490e-01 6.28763974e-01
2.16060877e-01 -3.10080615e-03 -3.92588437e-01 2.37426355e-01
-9.01841298e-02 -2.40002364e-01 2.48366743e-01 8.30907464e-01
-7.47935951e-01 -9.90257740e-01 -7.70098805e-01 6.05728999e-02
-1.89200506e-01 4.03279096e-01 -1.42728314e-01 7.86806166e-01
-1.69101134e-01 9.84118342e-01 1.01231851e-01 -1.55086786e-01
2.30493978e-01 5.31920791e-02 4.50631410e-01 -5.29954731e-01
3.47190946e-01 1.77236900e-01 -4.84391689e-01 -1.23933911e-01
-4.78750437e-01 -8.43745112e-01 -1.43299317e+00 9.21850726e-02
-6.07097447e-01 5.24571598e-01 7.93832004e-01 8.33508134e-01
2.16278240e-01 -2.29278445e-01 1.26708019e+00 -7.18779683e-01
-1.12069309e+00 -8.76958787e-01 -1.02629054e+00 6.05290353e-01
6.08938456e-01 -8.82358849e-01 -7.76159644e-01 3.99478108e-01]
|
[7.930380344390869, 4.131680488586426]
|
edc1f7d3-dda5-4f90-b82a-6826ab75b009
|
learnable-hollow-kernels-for-anatomical
|
2007.05103
| null |
https://arxiv.org/abs/2007.05103v2
|
https://arxiv.org/pdf/2007.05103v2.pdf
|
LORCK: Learnable Object-Resembling Convolution Kernels
|
Segmentation of certain hollow organs, such as the bladder, is especially hard to automate due to their complex geometry, vague intensity gradients in the soft tissues, and a tedious manual process of the data annotation routine. Yet, accurate localization of the walls and the cancer regions in the radiologic images of such organs is an essential step in oncology. To address this issue, we propose a new class of hollow kernels that learn to 'mimic' the contours of the segmented organ, effectively replicating its shape and structural complexity. We train a series of the U-Net-like neural networks using the proposed kernels and demonstrate the superiority of the idea in various spatio-temporal convolution scenarios. Specifically, the dilated hollow-kernel architecture outperforms state-of-the-art spatial segmentation models, whereas the addition of temporal blocks with, e.g., Bi-LSTM, establishes a new multi-class baseline for the bladder segmentation challenge. Our spatio-temporal model based on the hollow kernels reaches the mean dice scores of 0.936, 0.736, and 0.712 for the bladder's inner wall, the outer wall, and the tumor regions, respectively. The results pave the way towards other domain-specific deep learning applications where the shape of the segmented object could be used to form a proper convolution kernel for boosting the segmentation outcome.
|
['Dmitry V. Dylov', 'Oleg Rogov', 'Denis Larionov', 'Olga Shegai', 'Elizaveta Lazareva']
|
2020-07-09
| null | null | null | null |
['bladder-segmentation']
|
['medical']
|
[-9.30245128e-03 3.15932900e-01 -2.76924103e-01 -3.40279102e-01
-5.16125262e-01 -6.70921803e-01 3.84300053e-01 2.86908805e-01
-6.52933776e-01 4.42789406e-01 -2.01100521e-02 -5.50566435e-01
-1.22588217e-01 -6.01291656e-01 -6.13743305e-01 -9.29686844e-01
-2.05229148e-01 3.15679193e-01 4.36785966e-01 3.76278795e-02
1.12525702e-01 7.19290614e-01 -5.18657088e-01 2.17440560e-01
1.10150242e+00 1.11571145e+00 5.70277631e-01 6.10588849e-01
-2.80339807e-01 5.49421847e-01 -1.77275509e-01 -2.15256244e-01
-9.65331867e-02 -3.53253126e-01 -9.38493729e-01 -1.03372842e-01
2.52453566e-01 -1.25370443e-01 -4.13738966e-01 1.13962352e+00
2.47853413e-01 2.89412607e-02 6.38845921e-01 -4.75991309e-01
-6.76327050e-01 4.70743001e-01 -4.83042687e-01 2.78297782e-01
-3.31187487e-01 1.37149662e-01 4.61930126e-01 -4.62589979e-01
6.46264374e-01 6.48118854e-01 1.10103452e+00 5.76099753e-01
-9.61454511e-01 -3.33010674e-01 -1.12966616e-02 -2.19537929e-01
-1.12322116e+00 5.45202792e-02 3.38701159e-01 -4.68414098e-01
4.62309629e-01 4.37150687e-01 5.10932803e-01 8.36474359e-01
7.00564504e-01 1.01791477e+00 9.22126949e-01 -2.80748487e-01
1.61890492e-01 3.48417275e-02 1.40271127e-01 8.74392807e-01
1.91209972e-01 -5.25502563e-02 2.88959920e-01 -1.61440279e-02
1.18537259e+00 3.20020974e-01 -4.08330500e-01 -4.36571807e-01
-1.26521587e+00 6.24323189e-01 1.20294905e+00 7.35608280e-01
-2.65357167e-01 3.70962054e-01 5.62662363e-01 -4.60463375e-01
5.45335650e-01 4.81165141e-01 -1.49839789e-01 2.49549061e-01
-1.04574966e+00 -1.20233968e-01 5.90884864e-01 7.59745121e-01
2.16065332e-01 -4.38570291e-01 -6.72467053e-01 6.73507333e-01
2.61732757e-01 -4.37362120e-02 6.97183013e-01 -5.80802500e-01
9.95267183e-02 7.24390507e-01 9.84912440e-02 -4.52065170e-01
-1.00725174e+00 -8.18135738e-01 -1.16789734e+00 -2.27405634e-02
8.39468956e-01 -1.82373896e-01 -1.55557823e+00 1.43041337e+00
4.39526439e-01 3.37358475e-01 -8.06807131e-02 1.08416533e+00
9.70276773e-01 2.65713215e-01 1.86643586e-01 2.45898575e-01
1.53771460e+00 -1.32307541e+00 -6.45265639e-01 -2.85116196e-01
1.09218299e+00 -5.72956979e-01 8.96884918e-01 -2.44772702e-01
-1.08754015e+00 -3.78891438e-01 -8.56950164e-01 -2.41579749e-02
-2.49934614e-01 4.78667885e-01 9.74292755e-01 6.66293800e-01
-1.19727004e+00 9.31902587e-01 -1.24017906e+00 -2.97141224e-01
6.94557846e-01 3.04734260e-01 -2.95795918e-01 -1.06930628e-01
-9.92384791e-01 1.03673899e+00 4.60321873e-01 4.72122252e-01
-6.19233668e-01 -9.81136918e-01 -9.89522755e-01 2.55821925e-02
-6.67385906e-02 -6.56469703e-01 1.35045874e+00 -7.00700283e-01
-1.20354080e+00 1.04733944e+00 1.20817222e-01 -4.27534252e-01
9.25283372e-01 1.33990973e-01 4.15847301e-02 2.09056646e-01
1.73149243e-01 6.22941852e-01 2.84631103e-01 -1.12882960e+00
-3.72018188e-01 -5.80273449e-01 -2.97808796e-01 1.52214468e-01
-1.61354885e-01 -2.36206636e-01 -8.71371925e-01 -6.15343928e-01
3.02575350e-01 -1.03041804e+00 -8.23868215e-01 3.75563025e-01
-5.74488580e-01 5.39177395e-02 6.79256260e-01 -7.98882782e-01
1.09674203e+00 -2.27886987e+00 -1.96818024e-01 1.98972344e-01
8.80518630e-02 2.39663377e-01 2.13078991e-01 -2.07804054e-01
1.46460552e-02 3.32012139e-02 -4.73688900e-01 -2.44588137e-01
-3.40746552e-01 2.69776553e-01 3.32472920e-01 8.63705099e-01
-1.38137629e-02 1.30797720e+00 -1.14467120e+00 -6.26036942e-01
1.86971664e-01 4.89464492e-01 -2.23399356e-01 8.83244425e-02
1.25366300e-01 7.18964040e-01 -4.77911890e-01 4.93167490e-01
6.83591902e-01 -5.36798775e-01 6.24303035e-02 -2.37146750e-01
-1.80458009e-01 -9.27762762e-02 -6.81656599e-01 2.09700608e+00
-4.04855460e-01 4.37156826e-01 3.64774376e-01 -8.64139497e-01
6.15297735e-01 3.58091384e-01 8.07307243e-01 -8.79737854e-01
7.90731534e-02 5.11785746e-01 3.39577168e-01 -5.57175994e-01
2.27552950e-01 -3.18620920e-01 5.41317947e-02 1.29208967e-01
-3.20014209e-02 -1.95622668e-01 1.69805259e-01 -9.72459838e-02
1.01896715e+00 1.07063651e-01 3.81961688e-02 -6.67444229e-01
3.14307302e-01 1.33449873e-02 2.92031974e-01 7.69074559e-01
-4.68163610e-01 9.55649674e-01 4.16408926e-01 -5.93003511e-01
-9.40641642e-01 -1.06376576e+00 -7.53176272e-01 5.66684306e-01
4.22060490e-01 4.60709542e-01 -8.14323545e-01 -7.72832692e-01
1.00054137e-01 3.75828385e-01 -1.09187293e+00 -2.38469318e-01
-7.41189122e-01 -9.32365239e-01 7.19874322e-01 8.95097256e-01
5.17408013e-01 -1.01257634e+00 -7.24280596e-01 3.54649723e-01
-1.38165951e-01 -1.11092496e+00 -7.08391845e-01 5.24650156e-01
-1.06253219e+00 -9.40059721e-01 -1.43734264e+00 -9.62573528e-01
9.62550223e-01 -4.83717658e-02 8.77230346e-01 1.21018499e-01
-4.98506635e-01 -1.33834779e-01 -2.59859674e-02 -2.27566659e-01
-3.06899637e-01 1.00705706e-01 -4.74952668e-01 -1.97803259e-01
-2.27200657e-01 -9.50855166e-02 -9.81303394e-01 4.31049705e-01
-1.00386882e+00 2.27451667e-01 6.79987848e-01 1.03559494e+00
6.32478237e-01 -2.13202670e-01 1.95649356e-01 -9.95792568e-01
5.28332770e-01 -5.49319148e-01 -3.98660451e-01 4.08100843e-01
-2.58432060e-01 -7.69682899e-02 5.40812850e-01 -5.66216946e-01
-1.04257214e+00 2.94747114e-01 -1.94903299e-01 -2.91322410e-01
-2.38625631e-02 5.18669844e-01 5.40141523e-01 -3.62164021e-01
7.94039071e-01 2.64719099e-01 2.28406172e-02 -1.51478991e-01
1.65771097e-01 3.08372140e-01 7.43838489e-01 -5.21001041e-01
3.45202565e-01 7.59379447e-01 5.51350601e-02 -4.94746566e-01
-9.03866053e-01 -8.08040619e-01 -6.35078013e-01 -2.73796823e-02
1.08600104e+00 -4.96332258e-01 -5.66639364e-01 5.28461576e-01
-1.19828475e+00 -7.70791948e-01 -4.30612803e-01 5.13097107e-01
-5.35653651e-01 3.19118589e-01 -1.18595672e+00 -5.68933606e-01
-5.31970263e-01 -1.37438619e+00 1.14274836e+00 7.08612859e-01
-1.21758804e-01 -1.32123482e+00 -1.70485564e-02 1.60523459e-01
7.09370613e-01 3.73020500e-01 9.53898847e-01 -6.42413735e-01
-3.32765818e-01 -2.57009029e-01 -6.73012435e-01 9.97893810e-02
2.80386090e-01 -2.03739807e-01 -9.92106974e-01 -2.15530083e-01
-7.90778175e-02 -1.47073179e-01 1.00630248e+00 1.13351083e+00
1.51793575e+00 1.44017458e-01 -7.85203516e-01 9.99201894e-01
1.24345839e+00 3.54203045e-01 4.80786055e-01 1.36041641e-01
7.12665379e-01 3.50714624e-01 3.67844701e-01 1.94624007e-01
8.38967636e-02 3.72648627e-01 6.43414319e-01 -9.16423380e-01
-1.73135146e-01 1.31665900e-01 -3.76639634e-01 3.83281559e-01
-7.87986144e-02 2.47376747e-02 -1.09473670e+00 7.39408433e-01
-1.74458921e+00 -3.59963655e-01 -2.59946406e-01 2.03757572e+00
9.00147021e-01 1.02410965e-01 -2.37602815e-01 -3.88716966e-01
5.69824696e-01 -2.51630936e-02 -6.40232444e-01 -2.94632792e-01
1.88264862e-01 1.90483361e-01 8.55343699e-01 3.95323426e-01
-1.53094089e+00 6.29170716e-01 5.84298563e+00 7.69854486e-01
-1.30965209e+00 -1.94033273e-02 1.16414356e+00 3.75028551e-01
-6.41962662e-02 -2.94778973e-01 -6.13142550e-01 4.14425939e-01
6.38414979e-01 3.26862156e-01 -5.87341897e-02 6.02314889e-01
1.32187307e-01 -3.07777315e-01 -1.03839731e+00 6.30818605e-01
-3.06603432e-01 -1.70938051e+00 -4.87415850e-01 8.33620653e-02
9.37442064e-01 3.33563566e-01 8.93109664e-02 1.72892094e-01
2.07536995e-01 -1.35399902e+00 4.60631520e-01 5.84591806e-01
8.26278210e-01 -3.36953491e-01 1.01834166e+00 3.17715198e-01
-1.02857459e+00 1.18588194e-01 -2.34021977e-01 5.12647152e-01
-5.27485944e-02 5.23026884e-01 -1.04577732e+00 4.91815895e-01
7.11376727e-01 3.12427670e-01 -5.53755224e-01 1.51706469e+00
1.96281955e-01 4.14123505e-01 -3.84740770e-01 2.30691135e-02
8.26790571e-01 -2.02252910e-01 -7.14040250e-02 1.55957901e+00
2.91330159e-01 -1.12175353e-01 1.65579747e-02 1.16962683e+00
-1.81086197e-01 1.76361091e-02 -3.73519182e-01 7.66642615e-02
-6.77307397e-02 1.51327193e+00 -1.21348405e+00 -2.15221971e-01
-3.11350256e-01 9.94441807e-01 2.95730948e-01 5.26197076e-01
-8.84777904e-01 -2.02626422e-01 2.02594757e-01 -5.67917749e-02
4.81831729e-02 2.16944981e-03 -8.26081455e-01 -8.35072100e-01
7.85746351e-02 -1.31974950e-01 3.96279722e-01 -3.69360805e-01
-1.10283709e+00 5.25900483e-01 -3.12256128e-01 -9.02457416e-01
2.19502300e-01 -7.70678997e-01 -9.39865172e-01 9.75575149e-01
-1.64950132e+00 -1.03504741e+00 -5.65902770e-01 3.33390653e-01
1.75398052e-01 5.03378868e-01 8.68838847e-01 2.90741563e-01
-4.32999939e-01 5.21759689e-01 4.43841100e-01 5.53510427e-01
4.88312930e-01 -1.49747157e+00 3.29266101e-01 4.74533141e-01
-3.21891308e-01 5.47119677e-01 3.17256719e-01 -5.34310997e-01
-1.05257881e+00 -1.17158163e+00 4.87013549e-01 -1.67203143e-01
7.06991076e-01 -1.55009732e-01 -1.08998883e+00 6.25017285e-01
6.53836280e-02 7.72561848e-01 5.85310817e-01 -1.79722756e-01
2.41455346e-01 1.75419167e-01 -1.22520041e+00 6.72333419e-01
5.49558640e-01 -1.04872607e-01 -7.62390122e-02 6.40263140e-01
5.06286561e-01 -1.14403903e+00 -1.13902116e+00 6.54879928e-01
4.53112274e-01 -7.65506566e-01 1.12414539e+00 -4.98114973e-01
5.27670383e-01 6.48077577e-02 3.88472617e-01 -1.26098859e+00
-3.66418928e-01 -1.49911851e-01 1.51291505e-01 5.45310557e-01
5.41871309e-01 -4.24033076e-01 1.17633533e+00 7.47318506e-01
-7.63952434e-01 -1.29218554e+00 -1.03880024e+00 -5.28852463e-01
4.31878269e-01 -1.89953253e-01 7.90381208e-02 1.13268805e+00
-7.56921768e-02 -3.69219571e-01 2.43956700e-01 -1.64750274e-02
4.91958439e-01 1.36707425e-01 1.52840763e-01 -8.80522192e-01
-1.32190526e-01 -8.58241141e-01 -2.42215589e-01 -1.01049805e+00
-1.70329899e-01 -1.13966942e+00 4.83069420e-02 -1.82858646e+00
2.35507622e-01 -8.61756563e-01 -6.06259882e-01 4.43360507e-01
-3.35464835e-01 1.84403047e-01 -7.32087940e-02 1.35201931e-01
-2.40817979e-01 3.12866390e-01 2.21386552e+00 -3.18521708e-01
-2.72829860e-01 4.28553402e-01 -3.36947531e-01 7.72990644e-01
7.14811742e-01 -1.59625545e-01 -9.55113024e-02 -2.78270066e-01
-3.97244513e-01 3.11386287e-01 4.32018876e-01 -8.58433664e-01
5.32859802e-01 -1.44056737e-01 7.35327005e-01 -4.15822059e-01
2.43370026e-01 -8.36693764e-01 -1.32000640e-01 9.37670350e-01
-2.64097482e-01 -3.82201821e-01 5.11611581e-01 2.93810844e-01
-2.07211852e-01 -3.87472093e-01 1.13069284e+00 -5.41977525e-01
-3.66761029e-01 4.72111553e-01 -9.79486331e-02 3.58023159e-02
1.20151889e+00 -4.52355325e-01 -1.47999570e-01 1.68677047e-01
-9.20694649e-01 3.97662580e-01 2.54324406e-01 2.97047403e-02
5.10196567e-01 -1.22382367e+00 -5.30257344e-01 -2.84108706e-02
2.77447235e-03 6.82584465e-01 4.91093218e-01 1.51801872e+00
-1.00129485e+00 5.60611308e-01 -2.77566742e-02 -9.90727067e-01
-7.42298424e-01 2.48554870e-01 1.09030378e+00 -8.04810166e-01
-9.55997407e-01 1.05095351e+00 6.56038642e-01 -6.46243930e-01
3.14758420e-01 -9.46187019e-01 -3.56644802e-02 -2.55942762e-01
2.68840343e-01 -9.45388749e-02 2.25146428e-01 -3.25435966e-01
-4.07083362e-01 3.37078631e-01 -2.17968881e-01 4.57409263e-01
1.18920314e+00 2.42215067e-01 2.66333832e-03 1.58903018e-01
1.17523837e+00 -5.43231428e-01 -1.50874722e+00 -4.24675226e-01
1.94265366e-01 -9.60655585e-02 1.82485893e-01 -9.25036371e-01
-1.33810616e+00 8.31930995e-01 7.19398201e-01 -2.64506023e-02
8.95516634e-01 4.51247059e-02 8.33155572e-01 -1.23484299e-01
-9.22608674e-02 -8.35406482e-01 -8.34821537e-02 4.50202197e-01
6.17241442e-01 -1.39420164e+00 -2.96744615e-01 -3.70327741e-01
-5.43139398e-01 1.25660574e+00 6.67785525e-01 -4.18924838e-02
5.01833141e-01 5.24204433e-01 4.08440381e-01 -2.22817019e-01
-6.52684569e-02 6.44999295e-02 5.63760638e-01 2.95206994e-01
7.35849321e-01 2.14426219e-01 -1.93879694e-01 4.29744691e-01
1.99534386e-01 -4.50402871e-02 2.80716479e-01 8.27299178e-01
-2.86772966e-01 -6.51648581e-01 -2.94007242e-01 4.95170414e-01
-6.96781874e-01 -9.59287286e-02 -1.45694586e-02 9.17681992e-01
-9.57094356e-02 1.84415683e-01 1.21322898e-02 2.22249135e-01
2.86759734e-01 -1.89542413e-01 4.73930508e-01 -5.36144793e-01
-8.62525463e-01 2.29712129e-01 -3.06205183e-01 -5.16012490e-01
-7.22974911e-02 -2.69413471e-01 -1.73667848e+00 1.39322236e-01
-2.92831004e-01 1.77655593e-02 8.23129892e-01 9.78731036e-01
-6.14743158e-02 8.38652372e-01 2.36580029e-01 -9.89856720e-01
-6.34666443e-01 -9.44888353e-01 -6.18901134e-01 5.78588724e-01
3.84417951e-01 -4.28336084e-01 9.45304409e-02 -3.08784455e-01]
|
[14.651639938354492, -2.582350730895996]
|
7565fc6e-a122-49a3-8ffc-0c0959586277
|
dual-attention-model-for-aspect-level
|
2303.07689
| null |
https://arxiv.org/abs/2303.07689v1
|
https://arxiv.org/pdf/2303.07689v1.pdf
|
Dual-Attention Model for Aspect-Level Sentiment Classification
|
I propose a novel dual-attention model(DAM) for aspect-level sentiment classification. Many methods have been proposed, such as support vector machines for artificial design features, long short-term memory networks based on attention mechanisms, and graph neural networks based on dependency parsing. While these methods all have decent performance, I think they all miss one important piece of syntactic information: dependency labels. Based on this idea, this paper proposes a model using dependency labels for the attention mechanism to do this task. We evaluate the proposed approach on three datasets: laptop and restaurant are from SemEval 2014, and the last one is a twitter dataset. Experimental results show that the dual attention model has good performance on all three datasets.
|
['Mengfei Ye']
|
2023-03-14
| null | null | null | null |
['dependency-parsing']
|
['natural-language-processing']
|
[-3.46143275e-01 -2.52146013e-02 -4.97656018e-01 -5.49584746e-01
-2.61206597e-01 -1.85807794e-01 3.25825304e-01 3.91295969e-01
-2.17389539e-01 4.65677470e-01 4.75530475e-01 -6.06077015e-01
9.95808318e-02 -9.15523410e-01 -5.45280933e-01 -3.06619525e-01
-3.42742838e-02 2.55458534e-01 3.81634645e-02 -4.49662298e-01
8.25971544e-01 1.78650215e-01 -1.16550291e+00 3.26369941e-01
8.79243612e-01 9.57767785e-01 -1.43555645e-02 5.48811734e-01
-5.67032158e-01 1.23546374e+00 -5.89889348e-01 -6.69370234e-01
-1.14571735e-01 -1.97929099e-01 -9.89132524e-01 -2.57395148e-01
4.26661409e-03 1.91135108e-02 7.43619492e-03 9.64600861e-01
4.14271355e-01 1.65534452e-01 7.00898886e-01 -1.22434938e+00
-1.37948787e+00 9.08470750e-01 -8.59845936e-01 4.26153481e-01
3.21357608e-01 -2.64741689e-01 1.55642200e+00 -8.88275683e-01
1.41951665e-01 1.42954934e+00 8.93283963e-01 2.76935309e-01
-5.40002108e-01 -7.04767704e-01 7.31179655e-01 3.46860290e-01
-9.03601587e-01 -3.66088189e-03 9.71722960e-01 -3.22922230e-01
1.35831034e+00 -9.90623012e-02 5.31993747e-01 1.03882802e+00
6.82049572e-01 9.36693132e-01 1.01045418e+00 -5.63236594e-01
1.24450715e-03 2.52958953e-01 1.22235680e+00 7.59039700e-01
3.91770691e-01 -3.72508109e-01 -3.78732979e-01 -2.75573432e-01
3.32863718e-01 -7.41017982e-03 -9.50560197e-02 3.71547997e-01
-8.02464843e-01 1.19339693e+00 6.91106558e-01 6.18953884e-01
-1.57868415e-01 -5.84873781e-02 4.66641247e-01 3.61717016e-01
7.71509826e-01 2.94069380e-01 -8.79297733e-01 2.58504659e-01
-1.45825952e-01 -1.85200721e-01 1.09976792e+00 1.05996823e+00
4.87881184e-01 6.66076178e-03 -6.49814010e-02 7.30749369e-01
8.77007127e-01 2.09977776e-01 7.56777287e-01 2.96068359e-02
6.53326809e-01 9.83430922e-01 -2.42694855e-01 -1.40013921e+00
-7.74801016e-01 -3.96538675e-01 -7.92087078e-01 -1.27042770e-01
-1.33925125e-01 -2.88104832e-01 -1.12982094e+00 1.41807663e+00
6.77499594e-03 -1.71254762e-02 3.65550257e-02 7.61159658e-01
1.48682356e+00 8.26565146e-01 4.07339752e-01 6.74826503e-02
1.42226863e+00 -1.52542078e+00 -8.85936439e-01 -4.15876716e-01
9.56989408e-01 -7.03877926e-01 1.05232489e+00 4.09684986e-01
-7.48110771e-01 -6.48061156e-01 -1.27343023e+00 -3.11480612e-01
-9.03576195e-01 -7.54048079e-02 1.03339362e+00 1.03367543e+00
-9.45824265e-01 7.45939612e-01 -4.46286500e-01 -3.37695003e-01
2.42115840e-01 6.72434866e-01 -1.12877153e-01 7.63114318e-02
-1.34813654e+00 6.84255779e-01 1.38827726e-01 3.60435516e-01
-3.84800106e-01 -1.48918062e-01 -1.01408148e+00 2.08433226e-01
-1.69734731e-02 -8.02628756e-01 1.00679398e+00 -1.23686075e+00
-1.36082649e+00 6.35786951e-01 -6.50816655e-04 -1.81267634e-01
-3.01995307e-01 -3.95942152e-01 -5.34305811e-01 -2.94815153e-01
4.90810759e-02 1.46923140e-01 6.82952702e-01 -1.14904499e+00
-4.96927530e-01 -8.69088233e-01 4.08056885e-01 6.67641610e-02
-6.42543197e-01 2.52879739e-01 -2.89634258e-01 -5.29786468e-01
-1.08940922e-01 -8.81039798e-01 -3.99370968e-01 -8.43821108e-01
-6.12725675e-01 -6.75673008e-01 8.65827799e-01 -7.38253713e-01
1.64848661e+00 -1.85591543e+00 -1.03768064e-02 -5.89745045e-02
4.66945544e-02 2.41927668e-01 -1.84399173e-01 2.59765685e-01
-2.95436382e-01 7.48208642e-01 -9.98946428e-02 -5.45643747e-01
-1.93578124e-01 1.50061250e-01 -1.60218552e-02 3.33884805e-01
1.16681553e-01 1.07679725e+00 -6.29420638e-01 -4.34993327e-01
-9.26565081e-02 4.45543230e-01 -4.94663715e-01 3.07426453e-01
-1.21497750e-01 4.11036164e-01 -6.35825276e-01 7.40011036e-01
6.27652705e-01 -3.93334806e-01 1.66186571e-01 -2.18872532e-01
6.33187406e-03 4.49434698e-01 -8.22097957e-01 1.51829374e+00
-5.41715026e-01 2.08288893e-01 -2.65814438e-02 -9.34820116e-01
1.14312005e+00 1.62602887e-01 -9.69300494e-02 -6.71224415e-01
6.83822632e-01 -2.21617311e-01 1.27890594e-02 -6.05662346e-01
4.88211155e-01 -2.40253597e-01 -4.86975074e-01 4.10523802e-01
5.59766367e-02 1.91438079e-01 5.55293635e-02 1.38698712e-01
7.56087422e-01 -1.97980497e-02 1.37217447e-01 -5.04995525e-01
8.00613344e-01 -1.46448001e-01 7.36647904e-01 5.36444068e-01
-1.03425838e-01 4.98984456e-01 6.89024150e-01 -7.10966468e-01
-6.93227172e-01 -4.74302247e-02 2.30094105e-01 1.35821855e+00
1.33067340e-01 -6.67583942e-01 -5.05025923e-01 -1.22568202e+00
-2.28368446e-01 7.28356421e-01 -9.57454860e-01 -1.90171033e-01
-5.87644577e-01 -1.06672752e+00 -2.75209006e-02 8.08901489e-01
6.07830882e-01 -1.36701429e+00 7.78471977e-02 2.19052136e-02
2.81482972e-02 -1.01418734e+00 -3.51378918e-01 3.57050985e-01
-7.56711602e-01 -8.87701213e-01 -2.96680182e-01 -1.24741888e+00
4.70031083e-01 2.68921167e-01 1.48321784e+00 4.58998203e-01
3.94478887e-02 2.38037910e-02 -6.88998997e-01 -6.51078522e-01
-6.82348981e-02 4.25661922e-01 -2.22719654e-01 8.02699104e-02
6.51340961e-01 -6.09722435e-01 -4.18057024e-01 -1.23607065e-03
-5.78338027e-01 -2.59972095e-01 6.26947701e-01 9.03152347e-01
3.14495891e-01 -2.50759751e-01 7.12035000e-01 -1.40436482e+00
9.32636559e-01 -8.54080200e-01 -3.45441729e-01 1.59874842e-01
-8.21874440e-01 1.05551563e-01 8.47807884e-01 7.65053779e-02
-1.01105869e+00 -3.65705907e-01 -6.45066857e-01 -1.38274943e-02
-1.23059236e-01 9.83503938e-01 -2.93467909e-01 -3.65575403e-02
2.56937653e-01 -2.34115526e-01 -3.78484368e-01 -6.28956854e-01
1.20239466e-01 8.88485491e-01 -2.47566268e-01 -3.97515595e-01
1.58730790e-01 1.23129718e-01 -1.13310039e-01 -7.39003181e-01
-1.33896267e+00 -3.19940776e-01 -7.20739901e-01 1.90390587e-01
1.21396267e+00 -7.40581512e-01 -7.72002637e-01 4.86024588e-01
-1.21402752e+00 -8.83224085e-02 3.23631048e-01 3.80728483e-01
-8.87788087e-02 5.26181817e-01 -9.87202942e-01 -7.55866468e-01
-8.28524172e-01 -1.13899887e+00 8.83811116e-01 2.58385926e-01
-9.81424656e-03 -1.52158689e+00 7.83770606e-02 3.29337150e-01
4.02294129e-01 1.78990752e-01 1.42603302e+00 -7.96345890e-01
-9.88649130e-02 -3.06519717e-01 -1.89582750e-01 4.98090804e-01
-5.18781543e-02 7.64451921e-02 -8.59093547e-01 -8.14751089e-02
7.44742826e-02 -3.54799360e-01 9.71970201e-01 4.63864356e-01
1.19837701e+00 -1.26854613e-01 -4.38146800e-01 4.73033845e-01
1.55165207e+00 3.27140510e-01 5.74777484e-01 3.46647739e-01
1.12083411e+00 4.95055526e-01 4.91565228e-01 2.45910078e-01
8.22139621e-01 2.04968885e-01 7.07459033e-01 -1.15721934e-01
2.08605658e-02 -1.23470902e-01 3.49052310e-01 1.46574891e+00
-5.32269031e-02 -5.64304948e-01 -9.09447312e-01 5.80807090e-01
-2.01555085e+00 -4.44191277e-01 -6.86240196e-01 1.73584318e+00
1.45919204e-01 5.24185896e-01 -8.20583384e-03 1.02061972e-01
8.29692721e-01 2.17132881e-01 -2.28357181e-01 -1.07462883e+00
-1.13179468e-01 3.00983876e-01 3.50717217e-01 4.25234318e-01
-1.22956443e+00 1.01337397e+00 6.60216379e+00 4.96791899e-01
-9.30180311e-01 2.35552043e-01 1.02602088e+00 5.73485732e-01
-3.78956020e-01 -1.41202416e-02 -1.07048631e+00 3.82067025e-01
1.37640524e+00 4.80937511e-02 -3.05187732e-01 1.03475726e+00
-3.32813352e-01 6.65289909e-02 -7.42748380e-01 8.12269449e-01
4.89717424e-01 -1.06422448e+00 3.26749146e-01 -1.56094000e-01
5.56685448e-01 -8.27275664e-02 1.83676444e-02 8.75039399e-01
3.30710322e-01 -1.20125008e+00 2.33052492e-01 2.72132367e-01
1.29274786e-01 -1.18052888e+00 1.21846354e+00 3.23353171e-01
-1.22053623e+00 -3.29172403e-01 -3.80307317e-01 -5.54140270e-01
-6.53583333e-02 5.12853801e-01 -2.21275225e-01 7.91748226e-01
7.61797011e-01 1.11530733e+00 -8.16877007e-01 6.94213092e-01
-5.87108314e-01 8.63750398e-01 4.10605520e-02 -6.11850739e-01
4.77167070e-01 -2.41572127e-01 1.67425483e-01 1.18239236e+00
-4.56605628e-02 1.78125501e-02 1.31193697e-01 4.74205524e-01
-1.26408786e-01 4.85376626e-01 -8.08346391e-01 5.99075668e-03
-1.99821234e-01 1.54894757e+00 -8.77676070e-01 -2.67201722e-01
-9.09018874e-01 8.48132491e-01 4.34374124e-01 5.59727326e-02
-1.04287946e+00 -5.44944763e-01 2.84438133e-01 -1.08861715e-01
3.73831511e-01 -1.78391621e-01 -6.20004296e-01 -1.05450296e+00
-1.62974358e-01 -6.22885346e-01 6.66000903e-01 -7.38912284e-01
-1.57650816e+00 1.05356312e+00 -3.78842592e-01 -8.63618255e-01
1.53074011e-01 -8.46145928e-01 -8.39342475e-01 7.69281983e-01
-1.81422997e+00 -1.47456622e+00 -1.77379146e-01 5.64033210e-01
6.42118812e-01 -3.75291668e-02 1.03769445e+00 4.56375062e-01
-9.00850475e-01 7.37292826e-01 -2.31995255e-01 3.22043121e-01
5.07714689e-01 -1.28041029e+00 6.91379011e-01 4.15588289e-01
5.67765608e-02 6.91763103e-01 3.37545484e-01 -5.82502663e-01
-1.58487737e+00 -8.56493056e-01 1.15750122e+00 -5.49847543e-01
6.34984910e-01 -3.09467733e-01 -7.91817367e-01 1.01604462e+00
7.75049865e-01 -3.46503183e-02 9.52591360e-01 5.75361013e-01
-2.78345168e-01 2.36401886e-01 -8.36423576e-01 1.08985439e-01
8.54244113e-01 -1.09092019e-01 -8.39317322e-01 5.53356528e-01
1.03180754e+00 3.30806673e-02 -7.39462078e-01 5.44800937e-01
3.51553738e-01 -9.63812053e-01 4.94533062e-01 -8.44933212e-01
7.52051353e-01 -6.29998147e-02 1.58418249e-02 -1.41356516e+00
-7.53735542e-01 -1.66951492e-01 -9.08713564e-02 1.52556503e+00
8.20859969e-01 -6.97322071e-01 6.76721096e-01 5.29763162e-01
-3.74065965e-01 -1.03446329e+00 -4.50962484e-01 -5.05427122e-01
2.43911326e-01 -3.37492347e-01 5.71923316e-01 1.15305972e+00
-2.92827375e-02 1.49634194e+00 -4.04218376e-01 -1.51107758e-02
2.08400875e-01 5.07967412e-01 6.11132681e-01 -1.44971752e+00
-2.26220950e-01 -4.45985943e-01 -3.30495685e-01 -1.09836841e+00
5.21431088e-01 -8.75913858e-01 -4.51128423e-01 -1.87286341e+00
3.11265081e-01 -3.22306901e-01 -5.52779973e-01 4.15511787e-01
-3.38169396e-01 -8.40691011e-03 -1.12891831e-01 -1.17524609e-01
-7.31124043e-01 5.86456120e-01 1.14287734e+00 -2.51168519e-01
-1.30115509e-01 7.15215281e-02 -1.19236898e+00 8.89660716e-01
9.03613448e-01 -5.29656708e-01 -1.98256932e-02 -5.52421570e-01
5.60805798e-01 -5.66996373e-02 -3.44385594e-01 -7.21985877e-01
1.29949048e-01 7.83546716e-02 2.19208539e-01 -6.13136411e-01
9.75434408e-02 -8.42659533e-01 -5.17344832e-01 3.92175466e-01
-1.00235604e-01 4.83902067e-01 2.21596479e-01 5.97857237e-01
-4.12083715e-01 -4.18681115e-01 4.81450975e-01 -3.42603981e-01
-8.20909798e-01 4.04994398e-01 -3.73670846e-01 -1.63661629e-01
7.57612050e-01 1.12582222e-01 -4.15196329e-01 -2.74604112e-01
-8.82202744e-01 4.41996336e-01 -6.46786839e-02 7.94744670e-01
4.32655185e-01 -1.28098679e+00 -6.06070876e-01 1.46867633e-01
1.13355227e-01 -2.13657647e-01 7.34905675e-02 6.81255043e-01
-4.09318000e-01 4.93286908e-01 -1.98178709e-01 -1.44529670e-01
-1.21182168e+00 9.17812526e-01 1.93307362e-03 -7.31973112e-01
-1.73146993e-01 1.20947313e+00 2.30079025e-01 -5.49138665e-01
-1.30763287e-02 -4.90206033e-01 -9.58745718e-01 2.09118351e-01
4.95114416e-01 -1.40748862e-02 2.80696213e-01 -7.91892946e-01
-4.69658673e-01 7.84015119e-01 -3.87367904e-01 4.14217710e-01
1.49881721e+00 -8.62969607e-02 -4.76464748e-01 6.55721545e-01
1.34575224e+00 1.65418182e-02 -2.95951426e-01 7.92315751e-02
1.27451509e-01 -2.11577073e-01 1.78839192e-01 -6.50183320e-01
-1.19183016e+00 1.07168877e+00 2.87069798e-01 6.42511427e-01
1.00135040e+00 -2.04053462e-01 1.01277494e+00 3.02235037e-01
3.24738264e-01 -9.64543521e-01 2.73817986e-01 1.00171053e+00
7.45801330e-01 -1.33804476e+00 7.50644803e-02 -5.25766492e-01
-6.76194787e-01 1.08439267e+00 9.89409685e-01 -4.31353867e-01
1.18279910e+00 2.54989147e-01 -4.48054373e-02 -4.49938178e-01
-7.57648647e-01 -2.69173950e-01 1.16096601e-01 3.31724882e-01
9.20419753e-01 -1.65885642e-01 -6.50332153e-01 1.13886833e+00
8.35203938e-03 -3.62370104e-01 5.29556930e-01 1.06558979e+00
-3.71609360e-01 -1.12895131e+00 -1.09109282e-01 5.04380405e-01
-9.24494267e-01 -3.91277015e-01 -6.63706124e-01 7.77381361e-01
2.77229715e-02 1.12741590e+00 -1.48642749e-01 -7.64920652e-01
4.80043143e-01 1.66749835e-01 1.89294994e-01 -8.51822317e-01
-1.22930479e+00 -2.68554807e-01 2.52727658e-01 -1.51982456e-01
-5.00456095e-01 -3.59126832e-04 -1.22714353e+00 -3.10023278e-01
-7.11450398e-01 4.04700935e-01 7.17764974e-01 8.91789794e-01
5.20376146e-01 7.03664541e-01 8.12040508e-01 -6.30040705e-01
-5.89487627e-02 -1.26567733e+00 -5.64842045e-01 4.45832282e-01
7.49937743e-02 -6.10940516e-01 -4.77010578e-01 -3.92962754e-01]
|
[11.416118621826172, 6.719272136688232]
|
dec49dbc-ad65-4e49-abc5-79f3b0b88160
|
hawkes-processes-for-continuous-time-sequence
| null | null |
https://aclanthology.org/P16-2064
|
https://aclanthology.org/P16-2064.pdf
|
Hawkes Processes for Continuous Time Sequence Classification: an Application to Rumour Stance Classification in Twitter
| null |
['Arkaitz Zubiaga', 'Kalina Bontcheva', 'P. K. Srijith', 'Michal Lukasik', 'Trevor Cohn', 'Duy Vu']
|
2016-08-01
| null | null | null |
acl-2016-8
|
['rumour-detection']
|
['natural-language-processing']
|
[-8.63703638e-02 1.71006292e-01 -6.22772932e-01 -4.08054382e-01
-8.41685571e-03 -9.08429027e-01 6.55310392e-01 -6.53472245e-01
-2.85945535e-01 1.06888819e+00 -4.63127941e-02 -1.01159286e+00
-3.91567826e-01 -9.63214397e-01 -4.95059669e-01 -6.31337762e-01
-9.79754329e-01 7.25764990e-01 3.30370307e-01 -6.93831444e-01
7.03166842e-01 7.88774848e-01 -1.68942046e+00 7.18545914e-01
7.04417467e-01 8.52217197e-01 2.49141872e-01 1.14950800e+00
-1.95044339e-01 1.55633950e+00 -7.48382092e-01 -5.46825826e-01
3.13719302e-01 -1.23176083e-01 -7.22945035e-01 -1.01074085e-01
9.28529128e-02 -8.59008506e-02 -2.09758401e-01 9.22211111e-01
5.37373662e-01 4.49454933e-02 1.08379531e+00 -1.42548037e+00
-5.91619551e-01 6.10313773e-01 -4.01565880e-02 1.21627934e-01
1.03678203e+00 -5.39447069e-01 1.19919395e+00 -1.13026452e+00
7.20913768e-01 1.26888943e+00 8.66221786e-01 5.44149756e-01
-1.22286928e+00 -1.94712028e-01 -3.26822817e-01 -9.51717794e-02
-1.46558487e+00 -3.25250506e-01 4.25783843e-02 -2.08119690e-01
1.66093647e+00 1.26596653e+00 1.20609856e+00 1.01401424e+00
1.26658809e+00 8.34431887e-01 1.04267764e+00 -5.13792276e-01
3.35295945e-01 3.66983831e-01 1.54683650e-01 6.33519173e-01
8.40953708e-01 5.26628852e-01 -7.06372619e-01 -9.13127720e-01
9.33553874e-01 -2.94925272e-01 1.71355158e-01 -5.05680561e-01
-9.05919552e-01 6.91228509e-01 1.78732842e-01 3.83959889e-01
-1.39880210e-01 9.89067405e-02 1.26390755e-01 5.30987144e-01
-2.58292928e-02 6.47037446e-01 -9.11868811e-01 -1.33165747e-01
-8.71728659e-01 5.10332465e-01 1.25398111e+00 1.52653182e+00
1.24482810e-01 2.94908643e-01 -9.34252143e-02 3.17179203e-01
8.92314315e-01 1.01808000e+00 4.28362608e-01 -1.36146402e+00
-6.87414408e-02 1.72361732e-01 5.01781464e-01 -8.52631688e-01
-6.33224547e-01 -9.64177120e-03 -8.93263519e-01 4.49267089e-01
3.49161088e-01 4.57367361e-01 -8.02827001e-01 5.07305264e-01
4.33481112e-02 -2.34125629e-01 4.53833073e-01 5.55570945e-02
4.99930978e-01 3.76208365e-01 -1.34477139e-01 -5.73289394e-01
1.06082785e+00 -1.36716676e+00 -1.35299087e+00 2.33215362e-01
9.05734658e-01 -1.07320261e+00 4.35900748e-01 5.33875942e-01
-1.55548143e+00 -1.37560293e-01 -1.08699942e+00 1.78573877e-01
-7.27255583e-01 -3.14239264e-01 8.57801437e-01 1.43120694e+00
-1.60129595e+00 9.73287821e-01 -4.91727620e-01 6.59165755e-02
1.20568443e-02 8.24621081e-01 -2.64718989e-03 4.62812334e-01
-1.33193445e+00 1.08501506e+00 2.22979754e-01 -1.21242590e-01
-1.65216476e-01 -2.13068098e-01 -8.23704481e-01 -5.63443303e-01
-4.78693932e-01 -5.29636025e-01 1.44139910e+00 -2.59346128e-01
-1.65295815e+00 9.71794367e-01 -1.42069459e-01 -1.97814897e-01
6.14786744e-01 -1.28011424e-02 -8.31891418e-01 2.42498964e-01
-1.89849049e-01 5.76383233e-01 9.28263724e-01 -1.35132408e+00
-7.59897232e-01 -1.67359829e-01 -1.23336017e-01 2.66287565e-01
-1.25510961e-01 1.89734384e-01 2.11616129e-01 -1.12999000e-01
3.27147305e-01 -7.20919967e-01 -2.53068686e-01 -5.32041907e-01
-1.46512717e-01 -7.10518599e-01 7.70373225e-01 -4.51523662e-01
1.83705616e+00 -1.67618537e+00 -2.06720144e-01 3.98590982e-01
3.57815564e-01 -1.24705513e-03 2.40583986e-01 1.08380008e+00
-2.76906848e-01 7.32199550e-01 4.11965609e-01 -1.10722095e-01
2.24991128e-01 4.85861301e-01 -4.16602850e-01 3.05609167e-01
-1.29282743e-01 1.15307164e+00 -1.16605783e+00 -5.23096442e-01
4.11106765e-01 9.42391157e-02 -4.58732933e-01 4.79237735e-01
2.99364805e-01 1.47170946e-01 -3.56553018e-01 1.39399457e+00
1.15709066e+00 -1.31984919e-01 1.45911396e-01 5.30878425e-01
-3.79135728e-01 3.55090618e-01 -6.96863770e-01 1.05554795e+00
7.08333924e-02 5.00173986e-01 1.02364108e-01 -7.94621468e-01
3.33247900e-01 8.61540735e-01 4.37155962e-01 -9.67555881e-01
-2.26398129e-02 6.92409754e-01 1.12803578e-01 -6.07703328e-01
7.58228302e-01 3.94563079e-02 -3.64872098e-01 6.40070081e-01
-2.37588286e-01 -6.59476995e-01 9.64643434e-02 3.08100313e-01
6.22585893e-01 -5.18246442e-02 5.62923312e-01 -1.05613089e+00
7.86340594e-01 -1.82965681e-01 -1.81299388e-01 1.03415680e+00
-3.09923887e-01 3.19085121e-01 2.99841821e-01 -6.65102363e-01
-6.45341039e-01 -1.12307119e+00 -4.89381433e-01 1.30636716e+00
3.24267983e-01 -4.39044595e-01 -9.54439282e-01 -2.49762803e-01
1.77620783e-01 6.89606130e-01 -5.90509653e-01 3.84124845e-01
-5.03739953e-01 -8.32535863e-01 7.39044368e-01 3.45434904e-01
-5.07752821e-02 -1.33414865e+00 -6.58416986e-01 1.25490099e-01
-2.20292807e-01 -6.63697243e-01 -6.23428151e-02 4.48765576e-01
-1.35989368e+00 -5.18594682e-01 -6.66252747e-02 -8.20914626e-01
5.87345481e-01 2.46782884e-01 1.27047324e+00 5.39230824e-01
-2.31483161e-01 4.26904231e-01 -1.21292919e-01 -4.95818377e-01
-4.59671497e-01 -8.00336525e-02 5.28869390e-01 -5.87835789e-01
5.19427478e-01 -2.50617653e-01 -7.29350567e-01 5.37953973e-01
-6.88540697e-01 1.62748516e-01 1.79803044e-01 1.04410267e+00
1.35816500e-01 -9.34035778e-02 1.22507080e-01 -6.38007045e-01
8.72274399e-01 -1.69219792e-01 -3.78732830e-01 5.77745810e-02
-6.77108407e-01 -3.74140263e-01 3.21430594e-01 -3.25342178e-01
-1.01981449e+00 -4.87835288e-01 -9.82677937e-02 2.45538145e-01
1.11353043e-02 -1.46784872e-01 6.47139177e-02 -5.24923325e-01
8.02199244e-01 9.25758183e-02 1.99174434e-02 -6.80815242e-03
3.01039815e-01 7.09525108e-01 -6.82967342e-03 -6.68678164e-01
8.44880998e-01 4.91470337e-01 7.98524171e-02 -9.57177758e-01
-1.52186140e-01 -2.60129690e-01 -9.51962709e-01 -6.54426932e-01
6.56643391e-01 -6.78531289e-01 -9.10833478e-01 3.91110867e-01
-9.38691139e-01 -3.38627815e-01 -3.91645581e-01 4.25431967e-01
-1.01278400e+00 2.75717527e-02 -3.90154392e-01 -1.27895141e+00
-5.10977268e-01 -1.02017939e+00 9.43384409e-01 5.30070923e-02
-5.10597289e-01 -1.26927447e+00 5.87685481e-02 2.71537274e-01
1.81734428e-01 -1.73075795e-01 6.90226793e-01 -2.38256708e-01
-4.24233019e-01 -1.53791070e-01 2.34436691e-02 -1.39755070e-01
1.70832314e-02 4.95917559e-01 -9.81751978e-01 -5.31145096e-01
6.65065646e-02 -1.92070693e-01 -1.08835101e-01 6.52520418e-01
5.91872573e-01 -2.29931593e-01 -8.56000841e-01 5.40386558e-01
1.38545322e+00 3.85070026e-01 5.32770038e-01 7.28214979e-01
1.41836226e-01 5.53460240e-01 9.17806149e-01 4.63203549e-01
1.30579369e-02 3.28798652e-01 2.40537539e-01 1.49327129e-01
1.11720070e-01 -1.54819340e-01 3.77893507e-01 1.16112018e+00
-8.18235934e-01 -2.69281328e-01 -5.07867396e-01 4.42987174e-01
-1.72482407e+00 -1.40330648e+00 -4.32368398e-01 6.90478683e-01
6.25676990e-01 1.56016424e-01 -1.48347050e-01 3.35214496e-01
4.99015123e-01 -2.03574806e-01 -1.19133167e-01 -1.06291151e+00
-1.43546045e-01 3.15233678e-01 7.37729073e-01 1.00061214e+00
-7.20721722e-01 1.03317809e+00 1.29781246e+01 1.02230716e+00
2.21112028e-01 1.03134915e-01 5.16071796e-01 3.48020852e-01
-4.36954498e-01 -4.56139445e-02 -1.04416132e+00 2.72933897e-02
1.38140702e+00 -4.30666685e-01 6.85999811e-01 5.44219851e-01
3.44648361e-01 -4.23268199e-01 -1.26188684e+00 5.26221812e-01
9.73738134e-02 -1.40886843e+00 -2.83300440e-04 6.85225725e-01
7.73699820e-01 -5.08050561e-01 6.22419357e-01 3.24184299e-01
6.09259963e-01 -1.14389277e+00 8.60300779e-01 2.53660440e-01
1.03040910e+00 -6.05088234e-01 5.67372203e-01 1.68872893e-01
-1.14389896e+00 -2.20873043e-01 -8.77727985e-01 -1.00755692e+00
3.93533185e-02 -1.81779593e-01 -4.29956943e-01 3.48861217e-01
9.58353162e-01 2.99398601e-01 -3.93658698e-01 9.95779395e-01
-4.78694476e-02 1.04875881e-02 -2.71853864e-01 -4.48467314e-01
4.83122796e-01 -3.54241252e-01 4.66730654e-01 1.00164843e+00
2.48499006e-01 3.51035744e-01 -9.84472036e-02 4.01770771e-01
5.45058846e-01 3.29446048e-02 -1.19659424e+00 -1.78908288e-01
2.83276141e-01 9.16795909e-01 -4.83487815e-01 -4.22520459e-01
-2.00212970e-01 8.62069130e-01 -3.55488248e-02 5.01107454e-01
-6.11489356e-01 -4.35615242e-01 9.72222984e-01 -1.27327025e-01
-1.14700586e-01 -3.48497719e-01 -6.23769283e-01 -7.30352640e-01
-5.89872956e-01 -4.54965204e-01 5.93606755e-02 -5.53365827e-01
-1.39813089e+00 5.79277515e-01 -2.27688253e-02 -1.40553558e+00
-6.99901402e-01 -1.27676582e+00 -4.76714373e-01 4.92853165e-01
-1.11898029e+00 -1.10984349e+00 2.50124663e-01 4.52870727e-01
1.64141744e-01 -5.34416080e-01 1.39563632e+00 3.57715860e-02
1.00637585e-01 9.24474537e-01 6.69434488e-01 -7.37814724e-01
5.56605101e-01 -1.27867436e+00 5.68737745e-01 -1.37897313e-01
-4.31265175e-01 9.05828118e-01 6.28349900e-01 -5.39804697e-01
-1.41196322e+00 -3.66917729e-01 1.08350635e+00 -9.83769417e-01
6.55218959e-01 -3.86345625e-01 4.23767231e-02 7.88592756e-01
7.15902448e-01 -6.18741751e-01 8.21781039e-01 -1.83753878e-01
1.80774391e-01 5.75296998e-01 -1.39248300e+00 6.12354755e-01
1.66275799e+00 -4.63594139e-01 -6.25784039e-01 7.60327101e-01
8.13696027e-01 -6.94087505e-01 -1.30082703e+00 3.34633321e-01
8.65424156e-01 -8.75409484e-01 1.61978090e+00 -1.32660246e+00
-4.43697497e-02 2.81152606e-01 -2.61993498e-01 -9.32519078e-01
-5.96193194e-01 -1.23518765e+00 -5.33532679e-01 -5.83747849e-02
5.96577883e-01 -1.13057327e+00 3.42365682e-01 8.74560475e-01
-2.82833427e-01 -6.42737269e-01 -1.06996536e+00 -1.32016802e+00
-3.35779637e-02 -1.45572275e-01 4.90409225e-01 7.63798356e-01
6.83744550e-01 1.09839931e-01 -6.36873543e-02 -1.00294888e-01
5.33176839e-01 9.55312885e-03 4.41501856e-01 -1.34294486e+00
3.86843324e-01 -5.75816095e-01 -3.07655483e-01 -9.45992947e-01
-8.85957032e-02 -8.12076271e-01 -6.53862000e-01 -1.28511906e+00
-8.32044985e-03 -1.91056758e-01 -1.11109078e-01 -1.65725678e-01
3.67937148e-01 2.13746816e-01 1.20859891e-02 1.03788137e-01
-3.71160030e-01 6.18435517e-02 1.29639816e+00 6.91750320e-05
-1.62315920e-01 4.85058486e-01 -4.81304944e-01 7.84440815e-01
8.58408585e-02 -2.99253196e-01 -6.78878546e-01 6.11881316e-02
6.69384480e-01 4.61409837e-02 3.24159935e-02 -7.42885649e-01
5.37211418e-01 -3.75702560e-01 4.78586555e-01 -1.32223868e+00
1.30741090e-01 -9.61415648e-01 6.95283338e-02 9.40189242e-01
2.74610907e-01 1.20424610e-02 8.66204947e-02 5.39500564e-02
-1.44871444e-01 -5.70943117e-01 9.21121240e-01 -4.07591403e-01
-4.92852688e-01 -5.20386267e-03 -1.03226590e+00 8.97834301e-02
9.92593169e-01 -7.84614205e-01 -3.59281451e-01 -4.20183957e-01
-8.29068601e-01 -1.95836127e-02 6.50830388e-01 3.03609259e-02
7.30431557e-01 -1.51530886e+00 -2.30721906e-01 7.18729138e-01
-3.22939813e-01 -3.74400020e-01 -1.70157343e-01 6.58265352e-01
-1.32361674e+00 1.02442718e+00 -5.41665435e-01 -4.55340147e-01
-1.14228773e+00 4.78126436e-01 4.28307921e-01 -2.41845414e-01
-2.15481281e-01 1.11954463e+00 2.71224789e-02 -8.23025763e-01
1.85185194e-01 -9.21545625e-02 -7.54407048e-01 3.55081353e-03
6.88606799e-01 1.05194807e+00 -2.91290224e-01 -6.04341030e-01
-4.56784427e-01 6.65885091e-01 2.32151806e-01 -2.87484169e-01
9.17833567e-01 -2.19243199e-01 -9.89108324e-01 4.28274393e-01
8.38715494e-01 -1.36269778e-01 -3.69319022e-02 4.16855574e-01
1.25943512e-01 -8.02164078e-01 -4.32406247e-01 -3.55811834e-01
-1.85641110e-01 5.54822803e-01 5.34874737e-01 8.99602413e-01
8.57008278e-01 -3.02566767e-01 8.18335712e-01 9.66778398e-01
5.72402716e-01 -1.68019545e+00 -2.13140488e-01 6.89524531e-01
9.12339568e-01 -9.28350806e-01 5.44190466e-01 -7.27165341e-01
-4.14997995e-01 1.32979155e+00 4.68304873e-01 -1.55325383e-01
1.27306652e+00 5.74917436e-01 1.14069022e-02 -3.70670199e-01
-9.44949508e-01 1.12705544e-01 3.75366658e-01 1.11147714e+00
5.06513238e-01 5.07374525e-01 -9.66500878e-01 3.21953118e-01
-7.28706717e-01 -2.34555230e-01 4.90474731e-01 1.41972518e+00
-6.43810987e-01 -1.20391917e+00 -7.23931909e-01 4.61561680e-01
-5.49773455e-01 -1.16372630e-01 -5.28106689e-01 8.46754074e-01
-5.76629937e-02 1.49448860e+00 -1.97535474e-03 -5.03491640e-01
4.11356747e-01 1.41089618e-01 7.21762300e-01 -1.23501487e-01
-9.04846430e-01 4.15413082e-01 3.84890139e-01 -1.23056793e+00
-8.58632207e-01 -1.05834293e+00 -1.40667629e+00 -1.19437599e+00
-5.12782812e-01 1.89310342e-01 3.83317530e-01 3.90289724e-01
-2.06836104e-01 2.85260603e-02 9.80917513e-01 -1.07949340e+00
-3.90341938e-01 -9.39418614e-01 -1.00026262e+00 -7.84516707e-02
2.89751232e-01 -8.00943017e-01 -7.83523321e-01 2.83909619e-01]
|
[-7.236628532409668, 3.856640338897705]
|
08fef9ea-f782-4912-a651-13b18db34e2c
|
multi-scale-multi-modal-micro-expression
|
2301.02969
| null |
https://arxiv.org/abs/2301.02969v2
|
https://arxiv.org/pdf/2301.02969v2.pdf
|
Multi-scale multi-modal micro-expression recognition algorithm based on transformer
|
A micro-expression is a spontaneous unconscious facial muscle movement that can reveal the true emotions people attempt to hide. Although manual methods have made good progress and deep learning is gaining prominence. Due to the short duration of micro-expression and different scales of expressed in facial regions, existing algorithms cannot extract multi-modal multi-scale facial region features while taking into account contextual information to learn underlying features. Therefore, in order to solve the above problems, a multi-modal multi-scale algorithm based on transformer network is proposed in this paper, aiming to fully learn local multi-grained features of micro-expressions through two modal features of micro-expressions - motion features and texture features. To obtain local area features of the face at different scales, we learned patch features at different scales for both modalities, and then fused multi-layer multi-headed attention weights to obtain effective features by weighting the patch features, and combined cross-modal contrastive learning for model optimization. We conducted comprehensive experiments on three spontaneous datasets, and the results show the accuracy of the proposed algorithm in single measurement SMIC database is up to 78.73% and the F1 value on CASMEII of the combined database is up to 0.9071, which is at the leading level.
|
['Pan Wang', 'Lin Wang', 'Chun Qi', 'Jie Li', 'Fengping Wang']
|
2023-01-08
| null | null | null | null |
['micro-expression-recognition']
|
['computer-vision']
|
[-2.62098797e-02 -3.30552906e-01 -2.18798980e-01 -4.62501109e-01
-8.72653663e-01 7.47974589e-02 2.93681681e-01 -6.89008474e-01
-2.86578953e-01 4.52447653e-01 2.85451144e-01 6.92580879e-01
-4.51210439e-02 -4.45160329e-01 -4.66771394e-01 -1.19935489e+00
-1.09432101e-01 -1.02142349e-01 -1.42385930e-01 -3.58299226e-01
6.50161728e-02 6.12280846e-01 -1.71782756e+00 7.67548501e-01
4.37702566e-01 1.53926921e+00 -4.15430851e-02 2.29011104e-01
-1.75793603e-01 9.64706182e-01 -4.70429957e-01 -3.43764365e-01
-4.63272408e-02 -4.83280510e-01 -6.42494261e-01 1.43741846e-01
3.67167562e-01 -2.53307641e-01 -2.54530579e-01 1.35370004e+00
7.04464316e-01 -1.08554982e-01 5.11962295e-01 -1.26646197e+00
-5.69544137e-01 7.10069984e-02 -1.05320537e+00 2.97979027e-01
1.49323523e-01 2.70245038e-02 7.98223376e-01 -1.00233388e+00
4.66557175e-01 1.50474393e+00 4.61397767e-01 7.22415805e-01
-9.35685039e-01 -1.05715930e+00 -2.16013771e-02 4.98216540e-01
-1.66964602e+00 -5.61435878e-01 1.11618793e+00 -2.14552492e-01
8.20152044e-01 9.84918773e-02 6.92467272e-01 1.13232899e+00
4.53417659e-01 9.09625113e-01 1.64076698e+00 -1.49093509e-01
-3.39583814e-01 1.34252816e-01 -3.48990470e-01 9.40708876e-01
-5.56189895e-01 6.16637990e-02 -7.32907236e-01 1.83730915e-01
7.34431028e-01 1.26517296e-01 -2.71153841e-02 1.78175122e-01
-8.87240410e-01 7.20480740e-01 5.19171119e-01 6.85718358e-01
-5.51191986e-01 6.94621773e-03 6.39420152e-01 3.13135535e-01
6.12669706e-01 -1.82898611e-01 -5.30084610e-01 -3.30775231e-01
-7.64698863e-01 -1.84850320e-02 4.05268818e-02 5.01177669e-01
9.41019893e-01 2.22591043e-01 -1.34708971e-01 1.11909270e+00
1.45488366e-01 7.88867176e-01 7.92731404e-01 -7.63494849e-01
1.73016265e-01 6.36092901e-01 -4.09908593e-01 -1.43078780e+00
-6.44007087e-01 -7.14923665e-02 -1.03592575e+00 2.87083596e-01
8.55336338e-02 -1.98683321e-01 -6.94451451e-01 2.00791383e+00
4.31726843e-01 2.12721348e-01 -2.25147251e-02 1.18545425e+00
1.01323915e+00 7.00568140e-01 1.26957253e-01 -5.87083876e-01
1.49281955e+00 -7.37581551e-01 -9.90592241e-01 6.99356422e-02
3.51701677e-01 -7.01641023e-01 9.17478383e-01 3.74141812e-01
-9.06037569e-01 -7.45852232e-01 -8.17621946e-01 1.24647260e-01
-2.17554867e-01 2.98052609e-01 7.06484079e-01 2.82562882e-01
-7.83735275e-01 4.38958049e-01 -5.53543627e-01 -1.47779092e-01
6.98713720e-01 5.70474565e-01 -7.32683241e-01 1.33197665e-01
-1.24635482e+00 8.69048655e-01 1.11162968e-01 5.65836251e-01
-7.72084177e-01 -4.44425046e-01 -6.99238002e-01 -1.26696363e-01
1.72461957e-01 -5.25375605e-02 7.93820083e-01 -1.73792529e+00
-1.71784818e+00 1.13159573e+00 -3.09851676e-01 2.94027150e-01
-2.45267339e-02 -7.20293000e-02 -9.65620458e-01 5.29300094e-01
-1.08942121e-01 7.68172324e-01 1.24544656e+00 -8.87311637e-01
-5.34575045e-01 -8.22386384e-01 -9.88311321e-02 2.72231638e-01
-5.36660850e-01 6.11011446e-01 -3.08466703e-01 -5.24897695e-01
-1.73679758e-02 -8.13609004e-01 4.11012709e-01 -1.36790976e-01
-3.00510451e-02 -2.84518301e-01 1.04412961e+00 -8.01261485e-01
1.11609602e+00 -2.23292565e+00 4.78079617e-01 -3.14680859e-02
2.19502345e-01 1.40673891e-01 -2.50730097e-01 -2.31878623e-01
-2.43019700e-01 -2.74566144e-01 3.73571813e-02 -5.86337671e-02
-1.75524488e-01 4.64218408e-02 -6.80480003e-02 5.66697955e-01
4.08624738e-01 1.13515723e+00 -4.51262772e-01 -8.11834335e-01
9.00356919e-02 6.06615007e-01 -2.83817708e-01 3.24610054e-01
1.71877548e-01 4.64857250e-01 -5.84614158e-01 1.05834687e+00
8.89995992e-01 -5.43979853e-02 -1.52013019e-01 -6.53769195e-01
1.62175730e-01 -5.35638213e-01 -8.68309140e-01 1.72155178e+00
-6.67755485e-01 5.95642328e-01 4.58122224e-01 -1.22386670e+00
8.86505961e-01 5.24272621e-01 8.29980850e-01 -1.07064009e+00
6.01486981e-01 9.92760286e-02 -1.30627677e-01 -1.04101324e+00
-5.14281392e-02 -8.16439033e-01 -5.58260754e-02 1.37086794e-01
3.65040332e-01 2.35085592e-01 -4.48920757e-01 -2.70282835e-01
5.12078881e-01 1.34102315e-01 4.98475991e-02 -2.72106409e-01
8.07479322e-01 -5.10669887e-01 7.26424873e-01 -2.35315949e-01
-3.75070333e-01 3.99340957e-01 4.59793866e-01 -5.39582968e-01
-6.50473416e-01 -6.81783855e-01 -2.63895720e-01 1.25038266e+00
2.00003400e-01 -2.08974868e-01 -7.24128246e-01 -6.74011052e-01
-1.82198644e-01 -4.78152707e-02 -1.01722014e+00 -3.57495427e-01
-3.25363547e-01 -1.02813995e+00 5.15282989e-01 4.73008990e-01
7.71228075e-01 -1.52252007e+00 -3.56409580e-01 -5.82994446e-02
-3.76923412e-01 -1.19759929e+00 -3.77845079e-01 -1.79742977e-01
-5.51008642e-01 -8.68054211e-01 -8.12469542e-01 -8.15566599e-01
4.51553702e-01 -1.13991745e-01 6.47720277e-01 8.68083350e-03
-6.24990046e-01 9.55244750e-02 -1.45077914e-01 -1.56466082e-01
3.66405621e-02 -2.74840057e-01 2.48669013e-01 6.89072132e-01
6.68525755e-01 -7.14537859e-01 -3.41185629e-01 2.32476041e-01
-7.31897414e-01 -8.72775987e-02 9.13170576e-01 1.06132233e+00
5.25935829e-01 4.87702787e-02 6.60039365e-01 -4.08997476e-01
3.17247808e-01 -3.48766953e-01 -2.29084119e-01 1.36341274e-01
-6.92761317e-02 -2.17536867e-01 5.39934278e-01 -7.08089709e-01
-1.23429072e+00 -1.41016722e-01 -2.54855812e-01 -8.58934343e-01
-1.47537380e-01 3.60257208e-01 -5.23062885e-01 -2.57085502e-01
2.93175936e-01 4.59394157e-01 3.33263874e-01 -1.30853668e-01
1.70668706e-01 7.35869050e-01 4.22132939e-01 -5.81799269e-01
4.38459843e-01 5.87942660e-01 3.83683518e-02 -9.60278571e-01
-8.84316504e-01 -6.42128587e-02 -6.32532477e-01 -3.91181380e-01
1.11797559e+00 -1.06852257e+00 -1.01559150e+00 8.99652004e-01
-8.07911932e-01 -9.34268907e-02 1.35237440e-01 4.30167556e-01
-5.99073768e-01 2.76993394e-01 -7.64881670e-01 -7.02847362e-01
-3.41697425e-01 -1.24482441e+00 1.22010934e+00 4.22378808e-01
7.76895732e-02 -7.95105815e-01 -2.17318848e-01 4.03871030e-01
5.99936426e-01 4.35482085e-01 5.69313824e-01 -1.02942005e-01
-1.66359514e-01 -1.41012937e-01 -4.43918616e-01 4.03249115e-01
2.96774566e-01 -7.17946813e-02 -1.24782228e+00 -1.72116444e-01
1.76233098e-01 -8.45891714e-01 5.72520435e-01 4.48187530e-01
1.43824303e+00 -2.97593832e-01 -1.47087008e-01 8.18322897e-01
1.26931775e+00 1.49004564e-01 7.07625449e-01 7.86713064e-02
6.87811852e-01 7.05792546e-01 7.05238640e-01 5.67152917e-01
1.63921028e-01 8.76595616e-01 3.93526554e-01 -2.03736663e-01
9.80767235e-02 5.51195852e-02 5.63550889e-01 8.07933271e-01
-2.99422711e-01 6.71736121e-01 -2.93586880e-01 4.34637398e-01
-1.52269828e+00 -1.25823712e+00 4.43114460e-01 1.56656349e+00
9.24827516e-01 -8.92921165e-02 2.06815898e-01 -2.55524546e-01
5.80341637e-01 3.91314805e-01 -6.45047665e-01 -3.05490166e-01
-5.68742216e-01 1.79987535e-01 -2.93620564e-02 2.38089472e-01
-1.22894382e+00 1.02514803e+00 5.48824406e+00 1.30936396e+00
-1.84518826e+00 2.69137025e-01 6.88748479e-01 -3.83324176e-01
-1.65850320e-03 -4.20789182e-01 -7.30625749e-01 5.53289473e-01
8.33425105e-01 3.73842344e-02 5.50874114e-01 7.49107778e-01
4.25737686e-02 2.51672328e-01 -5.26479602e-01 1.53657711e+00
3.32003772e-01 -1.01594019e+00 -1.52277514e-01 3.53047997e-02
5.88344574e-01 -1.04291834e-01 2.20603526e-01 4.63454396e-01
-2.78317899e-01 -1.32943535e+00 4.75398988e-01 7.41405427e-01
1.12023687e+00 -9.24596429e-01 7.40700781e-01 1.41943410e-01
-1.32570517e+00 -2.09356964e-01 -4.54548180e-01 4.43306118e-02
-1.71176374e-01 1.92035481e-01 -4.21931259e-02 4.73183632e-01
1.03225172e+00 9.00272310e-01 -3.84546399e-01 2.35449225e-01
9.90236774e-02 4.24801588e-01 -3.20850760e-01 -7.07577541e-02
1.65115237e-01 -1.03575371e-01 1.79740727e-01 1.14723599e+00
2.99499542e-01 2.54737884e-01 -1.23777635e-01 7.44628310e-01
-6.15286417e-02 4.12637025e-01 -3.62086207e-01 -9.75214764e-02
-9.72563028e-03 1.85000932e+00 -2.39143431e-01 -2.34157637e-01
-5.51588714e-01 1.17035651e+00 3.10846329e-01 2.11708739e-01
-1.05016899e+00 -4.21475589e-01 7.79115438e-01 -1.72795221e-01
2.22007871e-01 1.92182362e-01 1.07444376e-01 -1.38063848e+00
3.91160846e-02 -1.09548616e+00 3.98639590e-01 -8.67731929e-01
-1.24892223e+00 8.83581400e-01 -1.99644685e-01 -1.21543062e+00
-8.91421735e-02 -8.16626072e-01 -5.56521773e-01 8.14118147e-01
-1.44591153e+00 -1.64288247e+00 -4.77917284e-01 1.10613394e+00
3.54356766e-01 -4.67118680e-01 8.83111477e-01 4.28223759e-01
-6.68160856e-01 9.62015212e-01 -1.03392765e-01 2.30838448e-01
7.61759520e-01 -6.70826852e-01 -6.53752387e-01 3.97086769e-01
-1.56038990e-02 5.16734004e-01 3.53136510e-01 -1.68430105e-01
-1.46190357e+00 -7.35215008e-01 3.77288371e-01 -4.64881510e-02
5.95923305e-01 -2.23107204e-01 -9.30837393e-01 5.11104941e-01
1.72360659e-01 5.37081599e-01 5.57702363e-01 5.76489605e-02
-2.99835622e-01 -5.29184103e-01 -1.17496908e+00 3.54474902e-01
7.93811023e-01 -7.98729837e-01 -3.47656220e-01 1.46021917e-01
3.46929580e-01 -3.08697253e-01 -1.27266002e+00 6.63321376e-01
8.05267155e-01 -9.85719919e-01 7.89149642e-01 -7.21906662e-01
6.55119002e-01 -7.47985840e-02 -4.71666574e-01 -1.15244246e+00
-2.37023264e-01 -3.99008244e-01 2.24880166e-02 1.21806955e+00
8.46412033e-02 -4.72053975e-01 6.26795113e-01 1.88279912e-01
1.66453913e-01 -1.12920809e+00 -1.22096992e+00 -2.43023098e-01
2.76531670e-02 -2.84639686e-01 6.65471077e-01 1.06078684e+00
2.35783532e-01 4.98544753e-01 -7.60976374e-01 6.05252478e-03
4.66273516e-01 5.31624615e-01 6.66052818e-01 -8.20116758e-01
3.89486924e-02 -5.43974757e-01 -6.41452670e-01 -7.99402773e-01
6.15308702e-01 -8.14781666e-01 -2.83098608e-01 -7.68638074e-01
5.33498943e-01 -4.77562658e-02 -5.99334836e-01 6.13305926e-01
-1.40788659e-01 5.87366164e-01 -9.09858346e-02 1.48533126e-02
-5.39426506e-01 1.02263153e+00 1.50372362e+00 -1.99687943e-01
2.69258440e-01 -3.58584166e-01 -6.96285725e-01 7.96127737e-01
4.94077563e-01 -6.99386597e-02 -7.33537897e-02 -9.71998498e-02
-1.55464143e-01 3.17762256e-01 2.56367534e-01 -9.10991251e-01
-4.40894179e-02 -3.28719914e-01 7.56803811e-01 -4.13732141e-01
8.09318900e-01 -8.60591352e-01 -1.53755724e-01 1.86728358e-01
-1.96143806e-01 -1.67118147e-01 3.75354916e-01 2.15533197e-01
-6.89429104e-01 4.09882426e-01 1.14691734e+00 4.94523831e-02
-1.20252073e+00 6.85233414e-01 -7.37296268e-02 -6.96176812e-02
1.05053532e+00 -1.41351804e-01 3.98665890e-02 -3.61698896e-01
-9.11597908e-01 -1.07517362e-01 1.55514672e-01 6.73562944e-01
6.84236288e-01 -1.79464912e+00 -7.28667676e-01 4.99693811e-01
1.61144629e-01 -4.92227256e-01 1.00654149e+00 1.27622557e+00
-1.41246825e-01 1.99486822e-01 -7.75833189e-01 -7.07503140e-01
-1.47023666e+00 4.96783048e-01 6.98500216e-01 6.61097690e-02
-4.74647433e-01 1.03264737e+00 3.67400885e-01 -3.73692781e-01
-2.06564248e-01 7.88479894e-02 -4.61372763e-01 3.05689991e-01
7.97054172e-01 -1.35160424e-03 -9.38399360e-02 -1.46377647e+00
-4.84005511e-01 1.21009696e+00 1.56424455e-02 1.61834419e-01
1.38920367e+00 -2.06237331e-01 -4.62369949e-01 4.35642332e-01
1.85632789e+00 3.98262106e-02 -1.18334150e+00 -4.65988815e-01
-4.49751735e-01 -5.44168293e-01 1.06692828e-01 -5.84901631e-01
-1.64169776e+00 1.20597506e+00 8.79629433e-01 -3.27060580e-01
1.68115258e+00 5.69678098e-02 8.69421005e-01 -9.48248617e-03
2.61271685e-01 -1.12503672e+00 3.61622781e-01 1.76537573e-01
1.01917434e+00 -1.37875712e+00 -1.91773415e-01 -9.74764302e-02
-1.04109049e+00 1.18825185e+00 8.78779471e-01 -1.14180140e-01
7.95685291e-01 2.97797948e-01 3.57776105e-01 -4.82304186e-01
-6.99500442e-01 4.42991555e-02 3.61607373e-01 2.30273828e-01
3.25746179e-01 1.94557197e-02 -5.85441943e-04 8.86962414e-01
-7.53281787e-02 1.72923461e-01 -1.75265059e-01 5.44613123e-01
-3.86566341e-01 -5.09414256e-01 -3.36776614e-01 3.65442365e-01
-9.50590134e-01 1.92616120e-01 -3.84766236e-02 6.87831938e-01
2.78330654e-01 6.71974242e-01 3.79292890e-02 -7.89409041e-01
1.66132927e-01 4.87364531e-02 7.41318464e-01 5.52261956e-02
-1.21653840e-01 2.86400378e-01 -2.57268190e-01 -9.08222675e-01
-7.83067882e-01 -5.65382838e-01 -1.33616149e+00 -2.22746551e-01
-2.21119478e-01 1.32239629e-02 3.17519456e-01 1.02849007e+00
2.63605386e-01 4.31767941e-01 1.00975478e+00 -1.04957020e+00
-4.48431700e-01 -1.26989079e+00 -9.22055662e-01 7.83753037e-01
2.89952010e-01 -9.85316992e-01 -3.24385107e-01 -1.02941312e-01]
|
[13.646313667297363, 1.6899478435516357]
|
8a48a1e6-0b26-4870-b81f-d7c2f9ad8d71
|
foreground-guidance-and-multi-layer-feature
|
2210.13053
| null |
https://arxiv.org/abs/2210.13053v1
|
https://arxiv.org/pdf/2210.13053v1.pdf
|
Foreground Guidance and Multi-Layer Feature Fusion for Unsupervised Object Discovery with Transformers
|
Unsupervised object discovery (UOD) has recently shown encouraging progress with the adoption of pre-trained Transformer features. However, current methods based on Transformers mainly focus on designing the localization head (e.g., seed selection-expansion and normalized cut) and overlook the importance of improving Transformer features. In this work, we handle UOD task from the perspective of feature enhancement and propose FOReground guidance and MUlti-LAyer feature fusion for unsupervised object discovery, dubbed FORMULA. Firstly, we present a foreground guidance strategy with an off-the-shelf UOD detector to highlight the foreground regions on the feature maps and then refine object locations in an iterative fashion. Moreover, to solve the scale variation issues in object detection, we design a multi-layer feature fusion module that aggregates features responding to objects at different scales. The experiments on VOC07, VOC12, and COCO 20k show that the proposed FORMULA achieves new state-of-the-art results on unsupervised object discovery. The code will be released at https://github.com/VDIGPKU/FORMULA.
|
['Yongtao Wang', 'Zengyu Yang', 'Zhiwei Lin']
|
2022-10-24
| null | null | null | null |
['object-discovery']
|
['computer-vision']
|
[ 1.67327777e-01 -2.06738457e-01 2.06698161e-02 -3.15177947e-01
-7.09071219e-01 -3.00636709e-01 3.27702075e-01 1.15846144e-02
-6.72023967e-02 2.17316121e-01 9.33326315e-03 1.76009834e-01
-1.35271624e-01 -6.34585559e-01 -4.74431723e-01 -9.41289008e-01
1.71634108e-01 1.72757372e-01 8.88373315e-01 1.71266615e-01
2.56791770e-01 4.99520570e-01 -1.76894855e+00 1.76498413e-01
7.44583070e-01 1.25177264e+00 5.88661551e-01 5.13156056e-01
-1.51646063e-01 6.02775872e-01 -3.67229789e-01 -4.60592620e-02
2.17566743e-01 -2.61199147e-01 -4.86377686e-01 5.75197816e-01
5.08882225e-01 -4.27096993e-01 -3.32998365e-01 1.04676986e+00
7.28680670e-01 -7.15554059e-02 5.40705800e-01 -1.20994377e+00
-6.26794755e-01 5.72520673e-01 -6.93802357e-01 7.31286168e-01
-3.75389867e-02 -4.92353402e-02 1.00503695e+00 -1.29076469e+00
4.57459539e-01 9.77907598e-01 6.11354232e-01 3.03300917e-01
-9.72668052e-01 -6.68197036e-01 3.00823838e-01 5.59658885e-01
-1.55903566e+00 -5.28584659e-01 8.96726847e-01 -4.56252903e-01
5.64574718e-01 3.17436129e-01 3.98894101e-01 4.93009210e-01
-4.45561036e-02 1.13506985e+00 8.88091922e-01 -4.33436662e-01
7.19521418e-02 2.79742926e-01 2.38426939e-01 8.41551781e-01
4.02988344e-01 -4.18215171e-02 -7.85682619e-01 9.67004895e-02
7.66583979e-01 2.47658744e-01 -1.72575593e-01 -6.63364589e-01
-1.13003755e+00 5.34160197e-01 5.96786559e-01 3.66855681e-01
-3.10248673e-01 1.20866045e-01 1.28701359e-01 -1.77488357e-01
4.70515460e-01 1.95343152e-01 -4.08040583e-01 1.88148126e-01
-8.03976178e-01 -3.70122269e-02 1.91140443e-01 1.07660806e+00
8.48623157e-01 -1.93686172e-01 -4.01933849e-01 7.11772501e-01
4.71259266e-01 4.05316383e-01 2.09632307e-01 -6.99109793e-01
1.10484026e-01 8.75639856e-01 5.41419387e-02 -8.21680129e-01
-3.23286176e-01 -5.62439680e-01 -2.83415437e-01 5.05089983e-02
1.71636701e-01 7.07882419e-02 -1.16321874e+00 1.08969164e+00
7.30176747e-01 3.69315654e-01 -1.63031995e-01 1.14237213e+00
1.13949633e+00 4.92650688e-01 -1.40471324e-01 -2.98436023e-02
1.36193216e+00 -1.07532465e+00 -8.00031364e-01 3.84085700e-02
5.48004329e-01 -8.99209023e-01 7.70041645e-01 3.06697100e-01
-7.90974259e-01 -6.63249135e-01 -9.72634554e-01 -7.48628303e-02
-4.99367207e-01 5.74006021e-01 7.56690383e-01 5.57992876e-01
-7.81431019e-01 1.32984996e-01 -8.17638695e-01 -6.06892109e-01
6.28636956e-01 3.78196806e-01 -2.31442582e-02 3.15684490e-02
-5.81828833e-01 5.05072832e-01 4.61363643e-01 3.78132671e-01
-1.06347227e+00 -5.43660223e-01 -5.14563084e-01 -8.23518785e-04
8.15605700e-01 -3.11651468e-01 1.25785148e+00 -6.32846773e-01
-1.36023855e+00 4.94191647e-01 -1.26795560e-01 -2.33181804e-01
1.74992159e-01 -3.95408601e-01 -2.04012260e-01 2.40452692e-01
1.52277604e-01 5.65005839e-01 7.26500928e-01 -1.28933620e+00
-1.24851501e+00 -2.97782749e-01 -5.18811345e-02 1.84582114e-01
-5.24036467e-01 3.06775808e-01 -7.87494242e-01 -4.87036049e-01
4.79093730e-01 -5.78526556e-01 -9.98732746e-02 1.37535542e-01
-4.24390376e-01 -4.56931621e-01 1.40145195e+00 -4.33562756e-01
1.24508142e+00 -2.35632992e+00 -5.05649522e-02 1.63920715e-01
4.81936663e-01 1.81769416e-01 1.12968348e-01 -1.57847106e-02
2.74019361e-01 -2.97713339e-01 -1.41553074e-01 -2.36497238e-01
-6.35090098e-02 -7.50165433e-02 1.49106801e-01 6.12807870e-01
3.37744296e-01 7.21166730e-01 -9.66594756e-01 -7.55225122e-01
4.89422888e-01 4.18066621e-01 -4.42362040e-01 7.68009052e-02
-9.75437611e-02 2.37289503e-01 -7.41159976e-01 1.26080477e+00
8.04679334e-01 -3.85203809e-01 -6.49058968e-02 -4.31997627e-01
-5.06102562e-01 5.36473393e-02 -1.51874030e+00 1.37521875e+00
1.01985604e-01 6.37066305e-01 5.68714663e-02 -7.65481114e-01
9.46770847e-01 1.17471427e-01 6.19271934e-01 -4.60855484e-01
4.25001055e-01 2.88117051e-01 7.81629514e-03 -4.70118403e-01
5.18154681e-01 4.42479163e-01 2.91890889e-01 1.07217610e-01
2.52038419e-01 2.42514446e-01 2.37690210e-01 2.55254269e-01
1.01097643e+00 1.59844682e-01 1.14482842e-01 -3.48751754e-01
5.09220779e-01 -5.04328683e-02 5.88902831e-01 7.70807803e-01
-3.38360399e-01 7.15328932e-01 1.36484116e-01 -1.73825547e-01
-6.68280005e-01 -1.06775951e+00 -3.76394004e-01 1.25062847e+00
5.55739224e-01 -4.55627441e-01 -7.39453077e-01 -7.55365789e-01
5.98036721e-02 2.37667367e-01 -6.17407501e-01 3.74505529e-03
-3.28862369e-01 -9.53900397e-01 2.06832901e-01 7.69783080e-01
6.00033879e-01 -9.07154083e-01 -7.46772885e-01 2.52548486e-01
-1.01004988e-02 -1.13037670e+00 -5.20064533e-01 5.17373145e-01
-7.22770751e-01 -1.03918433e+00 -6.14398479e-01 -9.95860398e-01
9.55086529e-01 6.31027818e-01 6.80597246e-01 4.33911197e-02
-5.37299335e-01 5.20704985e-01 -6.99702322e-01 -4.73187059e-01
3.62689346e-01 1.18964233e-01 4.50149849e-02 4.00716960e-01
4.60125953e-01 -2.05632001e-01 -6.24945760e-01 6.62565053e-01
-6.56118274e-01 9.94734168e-02 7.37209618e-01 5.29844105e-01
9.50142503e-01 1.13342047e-01 3.28991294e-01 -6.57178938e-01
6.17387518e-02 -2.72197545e-01 -6.60578310e-01 2.43597209e-01
-5.09555638e-01 -2.34313846e-01 5.81435822e-02 -4.67934668e-01
-1.19661701e+00 2.51714468e-01 3.06669891e-01 -5.54150939e-01
-6.38976544e-02 -9.85858776e-03 -5.40501237e-01 -3.22035283e-01
4.39736277e-01 2.76530594e-01 -4.67517555e-01 -8.18953216e-01
4.41219628e-01 8.99298489e-01 5.31933427e-01 -3.44426036e-01
9.00946617e-01 6.55251026e-01 -4.13746059e-01 -8.41467679e-01
-8.10419440e-01 -9.09525156e-01 -7.93299317e-01 -4.32189971e-01
7.28367805e-01 -8.12899232e-01 -2.50699967e-01 5.46443701e-01
-8.49863470e-01 -2.11747721e-01 -3.25749874e-01 4.76125836e-01
-2.17391402e-01 1.23160683e-01 -3.20122153e-01 -8.29439998e-01
-3.75864089e-01 -9.99206543e-01 1.17630732e+00 6.34435534e-01
3.84838343e-01 -3.24342072e-01 -2.54090220e-01 3.20810944e-01
1.99771896e-01 1.60875283e-02 1.98328421e-01 -4.32346612e-01
-1.09078252e+00 -8.67098793e-02 -6.40043318e-01 1.81381196e-01
3.71148437e-01 1.03237294e-01 -1.08411837e+00 -1.59195453e-01
-2.31871769e-01 1.05141290e-01 8.93153787e-01 5.48922479e-01
1.13911247e+00 1.56358406e-01 -6.29021406e-01 7.63727367e-01
1.24578416e+00 2.11493716e-01 3.70498449e-01 3.97306144e-01
8.38338971e-01 2.96787024e-01 9.36331749e-01 5.46019971e-01
3.14338386e-01 6.79198205e-01 3.74479949e-01 -2.24784717e-01
-4.31777894e-01 4.38260809e-02 2.11149395e-01 5.99157155e-01
-9.30518359e-02 -5.29861376e-02 -7.82458663e-01 7.42526948e-01
-1.90764797e+00 -4.47560430e-01 -2.67774075e-01 1.79680455e+00
4.22507018e-01 2.64146239e-01 1.01696663e-01 1.01855680e-01
8.64435673e-01 -1.32788837e-01 -4.35706645e-01 2.15995088e-01
-1.47278503e-01 -5.69368191e-02 5.71930408e-01 2.29159907e-01
-1.28446889e+00 1.07626772e+00 5.33245897e+00 9.04309452e-01
-1.13879597e+00 3.98758560e-01 3.34168524e-01 -1.27539515e-01
5.50887957e-02 1.63409933e-02 -1.29479456e+00 3.82041931e-01
1.84715241e-01 -2.06210669e-02 5.53695448e-02 9.79614377e-01
2.13836297e-01 -3.29691440e-01 -8.14480841e-01 1.02082789e+00
1.09526604e-01 -1.11092556e+00 -2.04615980e-01 -1.30791172e-01
7.05594420e-01 7.87005424e-02 -1.56919062e-01 -1.76687837e-02
-2.37186849e-01 -3.81003231e-01 8.92456889e-01 3.74513924e-01
5.18992364e-01 -5.01912832e-01 7.47219741e-01 -3.79102938e-02
-1.65968430e+00 -1.85058624e-01 -3.88332248e-01 2.54891038e-01
3.90954241e-02 7.52998471e-01 -8.87835324e-01 3.83012950e-01
1.11397564e+00 6.06686771e-01 -9.03472960e-01 1.52943707e+00
-2.27931663e-01 6.04906440e-01 -5.71055233e-01 5.23080258e-03
1.61863133e-01 1.86239615e-01 5.97517431e-01 1.25253940e+00
2.67157763e-01 1.03081405e-01 1.92848548e-01 6.39711499e-01
6.15661331e-02 2.29415968e-01 -1.42622098e-01 -1.10532962e-01
3.60144615e-01 1.54622531e+00 -1.21005929e+00 -3.11783314e-01
-5.64720988e-01 7.47701764e-01 1.84949830e-01 2.13648260e-01
-9.61952031e-01 -4.03191268e-01 2.47392714e-01 3.53377998e-01
9.06229556e-01 -1.96632043e-01 -2.47125134e-01 -9.66508269e-01
7.15822205e-02 -4.39059615e-01 3.74870300e-01 -6.89805388e-01
-8.59696746e-01 3.54913503e-01 4.94400859e-02 -1.13127625e+00
5.07296920e-01 -6.39651537e-01 -6.32467270e-01 2.71244556e-01
-1.54776120e+00 -1.04604506e+00 -5.95999241e-01 5.14711022e-01
6.19411826e-01 -2.03684438e-03 1.22913525e-01 6.80096269e-01
-8.04634452e-01 4.33940291e-01 1.72032341e-01 3.91379707e-02
5.50437212e-01 -1.04443252e+00 -1.70774497e-02 1.11298025e+00
2.88478106e-01 3.37326467e-01 5.12085080e-01 -6.40321493e-01
-1.34339261e+00 -1.05295241e+00 6.72475219e-01 -3.73860151e-01
5.73704839e-01 -5.43617129e-01 -8.05204630e-01 4.94192630e-01
-8.13529044e-02 4.06270355e-01 3.82699877e-01 -6.69828057e-02
1.30999386e-01 -3.72972071e-01 -1.01063287e+00 2.60418296e-01
1.25987995e+00 -1.74538791e-01 -2.71218151e-01 2.81162292e-01
5.14290154e-01 -4.49790359e-01 -6.27960205e-01 6.37536645e-01
4.39385563e-01 -7.49846935e-01 7.30554819e-01 -5.43802008e-02
-1.61871523e-01 -9.79621410e-01 -2.00131983e-01 -5.74447095e-01
-6.47942066e-01 -4.90022480e-01 -4.94094491e-01 1.63350630e+00
2.81665772e-01 -2.99680531e-01 8.28999639e-01 3.15660723e-02
-2.44716316e-01 -9.13216770e-01 -7.57298529e-01 -5.11378407e-01
-9.55995440e-01 -2.90398777e-01 4.75659400e-01 4.94504839e-01
-3.77474666e-01 1.61669120e-01 -1.66574225e-01 5.95057130e-01
8.54124129e-01 2.38919571e-01 8.02711427e-01 -1.04757488e+00
-4.56487089e-02 -2.10664958e-01 -5.67481875e-01 -1.08335650e+00
-5.93591332e-01 -6.22322619e-01 4.37431186e-01 -1.58134937e+00
4.96318430e-01 -5.19179165e-01 -6.70565248e-01 4.42872345e-01
-4.18127477e-01 4.80494916e-01 1.66970134e-01 2.67586410e-01
-1.04752386e+00 6.06269956e-01 1.09471357e+00 8.76729656e-03
-2.33155042e-01 -1.40119091e-01 -6.73553169e-01 6.84670508e-01
9.13006008e-01 -5.56068480e-01 -1.81760103e-01 -4.14518505e-01
-4.11137164e-01 -7.16916025e-01 4.79482949e-01 -1.20734608e+00
3.71965170e-01 1.14168497e-02 5.92773855e-01 -1.08930230e+00
8.19287375e-02 -9.13858056e-01 -1.38109520e-01 3.25379103e-01
1.71277598e-01 -2.22317100e-01 1.38157725e-01 5.27549863e-01
7.32991323e-02 -1.35980487e-01 7.84875393e-01 6.49024025e-02
-1.34149504e+00 3.12844783e-01 -2.49164164e-01 -3.39655817e-01
1.16964579e+00 -3.51275772e-01 -2.33985022e-01 2.04964027e-01
-5.95920920e-01 4.55561519e-01 2.70644903e-01 4.98599887e-01
7.26951361e-01 -1.21963513e+00 -6.11377120e-01 3.11144501e-01
4.16005403e-01 2.16705456e-01 1.60782278e-01 1.11321580e+00
-3.20845515e-01 2.57814437e-01 6.98904768e-02 -8.84714723e-01
-1.37323177e+00 3.27905655e-01 3.56170356e-01 -2.64808908e-02
-7.70997465e-01 1.25440931e+00 4.92045343e-01 -1.07732140e-01
3.87482047e-01 -4.01086003e-01 -1.98636293e-01 2.48322397e-01
5.24097025e-01 3.98920268e-01 -4.08535823e-02 -5.60391545e-01
-7.01601207e-01 6.08909965e-01 -2.70671785e-01 3.18817079e-01
1.18661463e+00 -2.87090719e-01 7.17652291e-02 1.67885303e-01
7.54013240e-01 -3.17192450e-02 -1.46175969e+00 -5.79579294e-01
8.40839371e-02 -5.48782885e-01 4.40615982e-01 -5.36543131e-01
-1.25343859e+00 4.59567279e-01 1.09764135e+00 3.89910638e-02
1.11659598e+00 3.93572569e-01 6.33187950e-01 2.49841839e-01
2.94638544e-01 -1.19554913e+00 1.75916791e-01 3.02078068e-01
5.09847879e-01 -1.13021195e+00 9.90633368e-02 -7.83833623e-01
-2.83886433e-01 9.40579116e-01 8.72132182e-01 7.99438357e-03
5.69290459e-01 4.69689339e-01 4.25136238e-02 -3.15147966e-01
-3.25248480e-01 -7.98585176e-01 3.75023544e-01 5.06041646e-01
2.29501262e-01 -7.02130329e-03 -2.15296030e-01 5.29218495e-01
2.65695840e-01 1.81270063e-01 1.97881863e-01 1.10659254e+00
-9.17372882e-01 -7.87615061e-01 -5.70453763e-01 5.79544425e-01
-2.63229817e-01 -5.28736897e-02 -2.99278259e-01 5.30115962e-01
6.33132458e-01 7.88288534e-01 -9.75144748e-03 -4.31837112e-01
2.93891042e-01 -2.40795478e-01 5.68369806e-01 -6.76670551e-01
-3.16208512e-01 4.52963620e-01 -9.23815221e-02 -4.78184432e-01
-4.56361592e-01 -7.61000276e-01 -1.25924504e+00 1.44454345e-01
-9.68601704e-01 1.36273831e-01 5.09830475e-01 7.31662929e-01
3.15114349e-01 7.69260466e-01 5.00376225e-01 -1.00014973e+00
-5.16250059e-02 -1.05229461e+00 -7.01444745e-01 6.36332715e-03
2.51457661e-01 -1.07834649e+00 -2.93683499e-01 2.04405695e-01]
|
[9.279473304748535, 0.9578142166137695]
|
3db59908-f3ba-4717-b0b9-1935a88c01f4
|
towards-measuring-ethicality-of-an
|
2303.03929
| null |
https://arxiv.org/abs/2303.03929v1
|
https://arxiv.org/pdf/2303.03929v1.pdf
|
Towards Measuring Ethicality of an Intelligent Assistive System
|
Artificial intelligence (AI) based assistive systems, so called intelligent assistive technology (IAT) are becoming increasingly ubiquitous by each day. IAT helps people in improving their quality of life by providing intelligent assistance based on the provided data. A few examples of such IATs include self-driving cars, robot assistants and smart-health management solutions. However, the presence of such autonomous entities poses ethical challenges concerning the stakeholders involved in using these systems. There is a lack of research when it comes to analysing how such IAT adheres to provided ethical regulations due to ethical, logistic and cost issues associated with such an analysis. In the light of the above-mentioned problem statement and issues, we present a method to measure the ethicality of an assistive system. To perform this task, we utilised our simulation tool that focuses on modelling navigation and assistance of Persons with Dementia (PwD) in indoor environments. By utilising this tool, we analyse how well different assistive strategies adhere to provided ethical regulations such as autonomy, justice and beneficence of the stakeholders.
|
['Thomas Kirste', 'Sebastian Bader', 'J. -C. Põder', 'M. Salman Shaukat']
|
2023-02-28
| null | null | null | null |
['self-driving-cars']
|
['computer-vision']
|
[-1.47320643e-01 7.69330740e-01 5.75582147e-01 -2.87206054e-01
4.56923366e-01 -1.99106589e-01 6.44691110e-01 4.01105694e-02
-1.09005892e+00 1.28195965e+00 3.50667715e-01 -4.42723453e-01
-3.49071890e-01 -5.73626161e-01 -2.04036742e-01 -2.57678419e-01
1.34386774e-02 4.95675832e-01 1.31440625e-01 -3.71461123e-01
1.37062177e-01 4.53844577e-01 -1.72815597e+00 -2.26904929e-01
1.15241408e+00 5.78928530e-01 1.92772403e-01 4.68373030e-01
1.63209468e-01 6.66099787e-01 -4.57800955e-01 -3.37182403e-01
6.58123493e-02 -6.88873380e-02 -9.37462091e-01 -4.47415352e-01
-7.80925214e-01 -9.23339009e-01 2.25018755e-01 8.24966669e-01
6.01206958e-01 -9.37685594e-02 9.25011337e-01 -1.66133356e+00
-3.09514403e-02 -6.03647865e-02 5.66641808e-01 -1.30472273e-01
6.11298025e-01 1.45887524e-01 2.84312755e-01 -1.88034311e-01
6.54397726e-01 1.08291900e+00 4.47576284e-01 6.68959379e-01
-7.96360254e-01 -2.38229334e-01 -2.94825613e-01 4.58075792e-01
-1.00674510e+00 -5.90362787e-01 4.38164592e-01 -6.12555623e-01
1.00569165e+00 2.34732479e-01 9.44605410e-01 9.99817133e-01
1.96019873e-01 5.35239160e-01 9.32656109e-01 -7.84301698e-01
9.79723036e-01 8.78771722e-01 2.93344259e-01 -6.46600649e-02
7.41047382e-01 -6.91606030e-02 1.38828442e-01 -3.35076600e-01
3.17445368e-01 -1.53930590e-01 -4.82638925e-02 -6.98230326e-01
-8.98486614e-01 7.28625298e-01 2.47036129e-01 9.02287304e-01
-8.87345135e-01 -1.28459662e-01 6.85936928e-01 2.39310861e-01
1.41442329e-01 3.12925398e-01 -4.23708647e-01 -5.56102812e-01
-1.09044202e-01 2.24153265e-01 1.14837515e+00 6.21546686e-01
2.30277225e-01 -3.74079764e-01 4.19785202e-01 2.76016831e-01
7.96808839e-01 2.37721384e-01 2.02511415e-01 -1.13096321e+00
6.66371733e-02 7.96866894e-01 7.48338342e-01 -8.71813774e-01
-4.98555899e-01 2.71841794e-01 -6.55124068e-01 6.77048504e-01
2.82425135e-01 -5.07948875e-01 -7.67467991e-02 1.38307500e+00
5.04744709e-01 -5.15820920e-01 4.35824364e-01 8.50511193e-01
3.72724444e-01 1.68646090e-02 5.27051330e-01 -3.23316276e-01
1.46802628e+00 -6.47038966e-02 -1.24172056e+00 -1.33089170e-01
9.38339233e-01 -1.34421840e-01 7.04906881e-01 3.05541962e-01
-7.27650285e-01 3.89653146e-02 -8.94980013e-01 2.76243865e-01
-6.95196390e-01 -2.19932109e-01 3.88405442e-01 1.20540702e+00
-1.28749669e+00 3.26223999e-01 -6.68168545e-01 -1.09648454e+00
1.10456094e-01 5.65198243e-01 -6.33747399e-01 1.29760072e-01
-1.55493987e+00 1.52775276e+00 2.95802116e-01 -5.04130162e-02
1.57857195e-01 -3.73494953e-01 -7.94176161e-01 -3.70622307e-01
-1.30198598e-01 -8.22020888e-01 1.14466953e+00 -8.28428447e-01
-1.36245215e+00 8.50784600e-01 2.88668960e-01 -7.49284685e-01
1.11561584e+00 -6.00124300e-01 -3.68339539e-01 8.96000043e-02
4.28111851e-01 4.72897708e-01 6.65555447e-02 -9.23952460e-01
-6.45651698e-01 -6.96751952e-01 1.99194923e-01 3.79270405e-01
-3.43759626e-01 4.64061677e-01 4.75844055e-01 2.09412083e-01
-6.33405924e-01 -7.91904211e-01 -3.79447967e-01 2.99248099e-01
-1.06454603e-01 -8.41884613e-02 9.99125540e-01 -7.49887049e-01
1.02277398e+00 -1.88821125e+00 -1.55993879e-01 3.08260117e-02
5.79419658e-02 6.16674840e-01 6.78689539e-01 4.59128737e-01
5.26327252e-01 -6.79757819e-02 -2.32891113e-01 -1.66607685e-02
5.35891235e-01 4.72611457e-01 3.82723927e-01 2.24277243e-01
-2.17175245e-01 2.83139110e-01 -7.63119161e-01 -5.54283202e-01
8.75073969e-01 8.18868816e-01 -3.36809069e-01 2.83580244e-01
3.32390130e-01 5.23006320e-01 -7.83761024e-01 1.95698708e-01
3.82256866e-01 6.45199060e-01 2.47559592e-01 1.73030868e-01
-2.88273573e-01 6.73818663e-02 -1.03517997e+00 1.01546860e+00
-4.24895942e-01 3.26782167e-01 4.85153913e-01 -1.01155448e+00
6.28271639e-01 8.02729785e-01 4.01860923e-01 -8.63505840e-01
4.20233339e-01 7.25978971e-01 3.97927649e-02 -1.04320931e+00
2.25658804e-01 1.07786199e-02 1.68969050e-01 4.86484706e-01
-2.82738119e-01 1.61680028e-01 -2.65827924e-01 -3.26346867e-02
1.34750414e+00 3.59731883e-01 9.12314057e-01 -5.74527204e-01
8.51111054e-01 -2.93252617e-01 1.37118265e-01 2.06995726e-01
-1.02292502e+00 -1.60387963e-01 2.90685415e-01 -7.42206752e-01
-1.21270454e+00 -2.81408072e-01 -1.87475443e-01 4.40643311e-01
-1.00178085e-01 9.01559833e-03 -1.33849061e+00 -4.80342686e-01
2.22962778e-02 1.31416905e+00 -4.19356316e-01 -1.08328581e-01
-1.71219319e-01 -2.55464852e-01 5.05567729e-01 1.30024374e-01
8.28343928e-01 -1.41343760e+00 -2.03322148e+00 4.86999094e-01
-1.40496358e-01 -1.15378046e+00 6.21530712e-01 6.61477298e-02
-7.93429554e-01 -1.08548653e+00 -7.74928272e-01 -3.02991837e-01
5.34581423e-01 -1.97005212e-01 8.13710213e-01 -2.71913521e-02
-2.66041636e-01 8.00176024e-01 -4.61859316e-01 -7.18698680e-01
-7.65006244e-01 -1.66543975e-01 5.42791128e-01 -3.69482428e-01
7.96700120e-01 -7.95418143e-01 -7.86637187e-01 4.04025853e-01
-6.64769471e-01 -9.36482027e-02 3.56168836e-01 -1.39353210e-02
-4.11448896e-01 -5.11574699e-03 9.24669623e-01 -4.85658616e-01
1.12280655e+00 -6.65410101e-01 3.10144331e-02 9.35603604e-02
-5.35188615e-01 3.36500630e-02 3.50823700e-01 9.64841098e-02
-9.57832575e-01 -3.30279693e-02 -2.33308911e-01 5.33175051e-01
-1.14339066e+00 1.65397320e-02 -7.11453021e-01 1.22522257e-01
7.52313793e-01 -3.00103486e-01 4.26644713e-01 -2.36638188e-01
1.35363370e-01 1.41247678e+00 1.71421394e-01 -2.44282678e-01
2.81845272e-01 2.93587178e-01 -5.92069626e-02 -1.01092231e+00
2.18959838e-01 -2.64471978e-01 -6.67529523e-01 -7.82715976e-01
9.36451733e-01 -3.34509194e-01 -1.02720475e+00 3.99116725e-01
-1.11173332e+00 -1.91754758e-01 -1.81126133e-01 5.93721449e-01
-6.94487810e-01 4.24469978e-01 1.21084794e-01 -1.90741909e+00
-6.56638324e-01 -8.97700489e-01 6.29144132e-01 4.87505645e-03
-9.43453372e-01 -9.55749094e-01 -1.17677309e-01 4.70178753e-01
7.27681816e-01 5.37891328e-01 8.11404645e-01 -5.49426377e-01
2.82363474e-01 -4.02260780e-01 6.20513707e-02 4.43515033e-01
3.33239436e-01 -4.70234513e-01 -6.79893613e-01 1.71617508e-01
4.10835981e-01 -1.08670913e-01 -3.39530051e-01 3.31129342e-01
-9.71134380e-02 -6.96148813e-01 -3.73455971e-01 -5.54516733e-01
1.09626675e+00 6.65122151e-01 1.10247576e+00 1.17592227e+00
6.03346154e-02 1.41934395e+00 1.16922486e+00 6.02028847e-01
5.64088881e-01 7.83617437e-01 5.64340293e-01 3.66692007e-01
5.95874429e-01 5.54322839e-01 3.28988075e-01 -7.47115090e-02
-5.32542527e-01 1.91967115e-01 -1.03026199e+00 5.59208632e-01
-2.26173043e+00 -6.58589721e-01 -3.71030301e-01 2.21821356e+00
3.57777894e-01 9.80314612e-02 4.22768503e-01 7.56481588e-01
6.45364165e-01 -6.02009416e-01 -4.39884007e-01 -7.34987140e-01
5.13988256e-01 -3.74042034e-01 3.32668126e-01 3.39986205e-01
-8.64138603e-01 2.81127989e-01 5.06448317e+00 6.73761219e-02
-3.05053174e-01 9.36525315e-03 3.77248704e-01 4.48615998e-01
-7.82079175e-02 -9.76166204e-02 -8.18356276e-02 6.58959508e-01
1.14083910e+00 -6.43446892e-02 3.31332862e-01 1.07774615e+00
7.50847518e-01 -6.84997559e-01 -9.16075826e-01 6.46534443e-01
-2.57818639e-01 -3.48676294e-01 -6.25783622e-01 3.46602678e-01
-1.73087865e-01 -4.42270011e-01 -5.28100252e-01 2.51650941e-02
9.20822099e-02 -6.93159103e-01 8.11307371e-01 8.16044331e-01
2.40563646e-01 -9.92549121e-01 1.37677062e+00 6.77173913e-01
-7.32520700e-01 -1.73311993e-01 -3.27182189e-02 -6.12367392e-01
4.61575031e-01 2.69847572e-01 -8.91840041e-01 4.03872907e-01
7.92589247e-01 -8.79950300e-02 -1.70512795e-01 7.80460417e-01
-3.12173665e-02 2.69260705e-02 -5.72951078e-01 -4.99938130e-01
7.11729825e-02 -4.50701207e-01 5.48160136e-01 7.55945802e-01
5.27462423e-01 1.89367384e-01 -8.57308388e-01 7.19478548e-01
9.47031677e-01 2.14729846e-01 -1.22544789e+00 2.56191909e-01
3.83324385e-01 8.90587330e-01 -4.26416993e-01 3.16356532e-02
-1.75933078e-01 1.15339613e+00 -3.83327633e-01 9.85169709e-02
-4.35632139e-01 -2.31840655e-01 8.08621705e-01 5.67764759e-01
-3.92436266e-01 -3.04739154e-03 -3.11709493e-01 -3.88167411e-01
2.01214656e-01 -6.43286705e-01 -9.14453194e-02 -9.59344745e-01
-6.20956719e-01 4.43095982e-01 2.34407753e-01 -1.11673784e+00
-6.72194660e-01 -7.76275635e-01 -3.13325673e-01 5.63604414e-01
-8.74321997e-01 -1.12728572e+00 -2.48194501e-01 4.24775600e-01
-7.41772726e-02 -2.23858595e-01 1.30682909e+00 2.72446811e-01
-1.00438975e-01 -1.70070976e-01 7.52308890e-02 -4.34394985e-01
4.37077463e-01 -1.15302241e+00 4.04874422e-02 5.32066286e-01
-1.12692857e+00 4.98444438e-01 1.03390896e+00 -7.15958476e-01
-9.71369207e-01 -4.51710969e-01 1.25478280e+00 -5.07748663e-01
3.93551439e-01 -1.62921742e-01 -4.55527037e-01 3.94003242e-01
3.22100282e-01 -4.26920980e-01 6.19451344e-01 -2.64230847e-01
4.97615099e-01 1.94252625e-01 -2.07240844e+00 7.62199700e-01
1.08313870e+00 -2.73876548e-01 -7.88828254e-01 3.69406082e-02
3.93016487e-01 3.68841618e-01 -7.86940813e-01 3.39623779e-01
7.00494349e-01 -1.48137236e+00 8.32540929e-01 -2.64232874e-01
-4.12553325e-02 -7.16715455e-02 -9.15870667e-02 -1.05346143e+00
2.02368870e-01 -3.66722494e-01 -3.16157602e-02 1.22091258e+00
-9.42284837e-02 -1.26193094e+00 4.46264893e-01 1.94577229e+00
1.27535924e-01 -3.06053519e-01 -1.39104140e+00 -5.12911081e-01
-1.68068007e-01 -5.01036942e-01 5.72111607e-01 7.68288255e-01
6.07540369e-01 -1.90625992e-02 -3.06346655e-01 -1.00826688e-01
6.27315938e-01 -1.20804632e+00 7.80019045e-01 -1.64310730e+00
3.31098616e-01 -1.09945543e-01 -1.14212120e+00 1.74854100e-01
-3.47888581e-02 8.06184784e-02 -1.87621996e-01 -2.12377524e+00
-2.26424649e-01 -3.43603343e-01 1.23758711e-01 3.13182652e-01
1.32438570e-01 -3.45322430e-01 2.39079949e-02 9.35863033e-02
-3.34337115e-01 3.55912596e-01 4.54419225e-01 2.60964841e-01
-2.39093706e-01 1.99243173e-01 -6.14087164e-01 9.91959453e-01
9.87472415e-01 -1.81222826e-01 -4.42275852e-01 2.45259807e-01
3.18530887e-01 -1.42252564e-01 5.41953266e-01 -1.36271584e+00
1.72614843e-01 1.02396652e-01 -1.10963605e-01 -1.25517875e-01
1.47594169e-01 -1.77661848e+00 6.94538891e-01 1.03280807e+00
7.41019994e-02 -4.51943874e-01 -6.61281273e-02 1.36114329e-01
1.91422433e-01 -5.97583652e-01 4.29437518e-01 5.14023677e-02
-5.44985354e-01 -6.99987888e-01 -1.10211277e+00 -6.29381418e-01
1.55871344e+00 -9.05810773e-01 -1.74220830e-01 -6.34117782e-01
-6.76108599e-01 3.37318718e-01 6.21530414e-01 1.96055442e-01
2.19093964e-01 -1.05522490e+00 -2.18165547e-01 -1.24412112e-01
4.77237105e-02 -2.53417462e-01 3.23490381e-01 7.43806779e-01
-8.13547134e-01 7.81850934e-01 -7.76480138e-01 2.50779595e-02
-1.40417194e+00 4.73755240e-01 4.25293654e-01 -4.49792407e-02
-6.70185685e-01 -8.18330720e-02 -1.25680029e-01 -7.33423591e-01
3.02624881e-01 -6.03602938e-02 -8.25880885e-01 1.10965565e-01
3.92260581e-01 8.34239542e-01 -9.28275213e-02 -8.58076215e-01
-7.65361607e-01 2.47836605e-01 3.60281765e-01 -6.37581825e-01
1.38684869e+00 -4.90438521e-01 -8.20092112e-02 2.28600010e-01
5.46345174e-01 -4.51374531e-01 -9.06533301e-01 5.29546678e-01
3.05940747e-01 -1.59168676e-01 -2.42558062e-01 -8.20622504e-01
-1.76048249e-01 5.31982422e-01 1.02932239e+00 7.13724792e-01
9.78591919e-01 -5.60271323e-01 4.04262125e-01 5.98045766e-01
1.10417855e+00 -1.60404754e+00 -5.71006060e-01 -2.65146811e-02
1.09869123e+00 -1.06079090e+00 -2.16810673e-01 -2.80536860e-01
-7.63861835e-01 9.47317183e-01 3.49976987e-01 2.58494467e-01
8.27798843e-01 1.32737443e-01 2.72349510e-02 2.06853561e-02
2.78174952e-02 -2.73472965e-01 -4.60375875e-01 1.35699403e+00
1.83404163e-01 3.74312222e-01 -9.68058527e-01 9.34657216e-01
-1.55372337e-01 6.63996696e-01 4.92563337e-01 1.35486805e+00
-5.53095877e-01 -1.06771350e+00 -6.72694147e-01 3.63356829e-01
-2.60881901e-01 4.61233616e-01 -7.35814750e-01 1.05451381e+00
3.00192446e-01 1.33461726e+00 -3.01625699e-01 -2.27507189e-01
6.48575068e-01 1.65933087e-01 -4.39603291e-02 -1.04102328e-01
-5.50190568e-01 -6.31604910e-01 7.39022672e-01 -3.42439622e-01
-7.00444281e-01 -7.32746303e-01 -1.24121034e+00 -2.67066270e-01
5.18526770e-02 2.59596527e-01 1.14334607e+00 1.19272280e+00
4.68760729e-01 3.25629264e-01 2.47365758e-01 -8.42486322e-01
-1.59268320e-01 -1.01104426e+00 -5.26558220e-01 2.41600499e-01
-2.07020249e-02 -6.28377318e-01 -5.38134217e-01 -3.05147052e-01]
|
[4.97033166885376, 0.9773358106613159]
|
18439422-10b3-4577-8396-fd6391cb01bb
|
less-is-more-data-efficient-complex-question
|
2010.15881
| null |
https://arxiv.org/abs/2010.15881v1
|
https://arxiv.org/pdf/2010.15881v1.pdf
|
Less is More: Data-Efficient Complex Question Answering over Knowledge Bases
|
Question answering is an effective method for obtaining information from knowledge bases (KB). In this paper, we propose the Neural-Symbolic Complex Question Answering (NS-CQA) model, a data-efficient reinforcement learning framework for complex question answering by using only a modest number of training samples. Our framework consists of a neural generator and a symbolic executor that, respectively, transforms a natural-language question into a sequence of primitive actions, and executes them over the knowledge base to compute the answer. We carefully formulate a set of primitive symbolic actions that allows us to not only simplify our neural network design but also accelerate model convergence. To reduce search space, we employ the copy and masking mechanisms in our encoder-decoder architecture to drastically reduce the decoder output vocabulary and improve model generalizability. We equip our model with a memory buffer that stores high-reward promising programs. Besides, we propose an adaptive reward function. By comparing the generated trial with the trials stored in the memory buffer, we derive the curriculum-guided reward bonus, i.e., the proximity and the novelty. To mitigate the sparse reward problem, we combine the adaptive reward and the reward bonus, reshaping the sparse reward into dense feedback. Also, we encourage the model to generate new trials to avoid imitating the spurious trials while making the model remember the past high-reward trials to improve data efficiency. Our NS-CQA model is evaluated on two datasets: CQA, a recent large-scale complex question answering dataset, and WebQuestionsSP, a multi-hop question answering dataset. On both datasets, our model outperforms the state-of-the-art models. Notably, on CQA, NS-CQA performs well on questions with higher complexity, while only using approximately 1% of the total training samples.
|
['Daiqing Qi', 'Jingyao Zhang', 'Wei Wu', 'Guilin Qi', 'Yuan-Fang Li', 'Yuncheng Hua']
|
2020-10-29
| null | null | null | null |
['multi-hop-question-answering']
|
['knowledge-base']
|
[ 1.72527619e-02 1.23653881e-01 -2.58588523e-01 -3.34724247e-01
-1.06511116e+00 -6.50652409e-01 3.49406719e-01 1.95417747e-01
-5.42175531e-01 5.17019212e-01 5.28822513e-03 -5.87350309e-01
-1.84742454e-02 -1.17168891e+00 -1.07097650e+00 -2.81267613e-01
2.54439026e-01 3.66827965e-01 5.45426965e-01 -3.72936726e-01
2.37020910e-01 -4.66860123e-02 -1.63104200e+00 3.85513604e-01
1.33893037e+00 1.09428990e+00 5.78836799e-01 6.86552644e-01
-3.27548206e-01 1.46044254e+00 -6.11119688e-01 -5.14772415e-01
-7.82204196e-02 -6.52678847e-01 -1.10394192e+00 -4.71072674e-01
2.12228730e-01 -7.05070794e-01 -3.08223665e-01 8.13224375e-01
3.43648523e-01 2.11876601e-01 1.41119495e-01 -1.05394769e+00
-7.76874781e-01 8.11069608e-01 3.48370187e-02 1.73765376e-01
5.25171518e-01 6.12253368e-01 1.24534178e+00 -6.37584507e-01
3.95975113e-01 1.10588396e+00 2.23015055e-01 6.79086328e-01
-1.12190914e+00 -4.90252018e-01 3.26238543e-01 4.62299824e-01
-1.13348472e+00 -5.03856599e-01 6.61245465e-01 -6.05772473e-02
1.28886151e+00 2.55629539e-01 5.51700771e-01 8.02380025e-01
-1.17165051e-01 1.07672250e+00 7.23605871e-01 -3.29053432e-01
6.21369183e-01 -1.51078314e-01 3.65094900e-01 1.01516664e+00
-2.91185051e-01 2.43616030e-01 -5.34897625e-01 -2.96806186e-01
5.48252881e-01 -9.57701504e-02 -2.22422138e-01 -1.07181892e-01
-9.89725053e-01 8.79843533e-01 5.56323349e-01 8.77890810e-02
-4.24164623e-01 4.67845351e-01 1.88630044e-01 6.03705883e-01
-2.16711462e-01 7.80435562e-01 -5.28492630e-01 -4.82405573e-01
-5.64632297e-01 4.30735052e-01 9.23021317e-01 9.74013329e-01
1.02126706e+00 -2.61837207e-02 -5.62165141e-01 7.96945035e-01
2.38517821e-01 5.33504903e-01 4.91424322e-01 -1.14970028e+00
6.26507461e-01 9.52207625e-01 -7.71612814e-03 -7.21944213e-01
-1.37539238e-01 -2.19514742e-01 -3.26988727e-01 -3.33459169e-01
4.46143925e-01 6.77036196e-02 -7.87636042e-01 2.06641865e+00
3.31837535e-01 2.05951050e-01 1.09772429e-01 9.36556637e-01
7.02116191e-01 8.28284025e-01 6.32154793e-02 2.85279602e-02
1.38654721e+00 -1.31454027e+00 -4.27137554e-01 -3.90787274e-01
9.85687733e-01 -1.49241224e-01 1.60118890e+00 3.46032053e-01
-1.30963123e+00 -4.61452961e-01 -8.96253586e-01 -2.81044573e-01
-5.08033969e-02 6.67358488e-02 5.41232526e-01 3.78858060e-01
-8.34473312e-01 5.58400512e-01 -7.58566976e-01 2.85729855e-01
2.46135369e-01 2.16259375e-01 3.02833289e-01 -3.18410307e-01
-1.28945291e+00 7.32837975e-01 3.21147263e-01 -1.06562011e-01
-1.24429798e+00 -7.33180046e-01 -8.74500036e-01 3.42761040e-01
8.67968082e-01 -8.20689976e-01 1.57528913e+00 -6.66381061e-01
-1.78961670e+00 2.90084928e-01 -3.14675808e-01 -4.66931641e-01
-1.81535278e-02 -2.24741265e-01 -2.28938147e-01 2.08567008e-01
-7.26505965e-02 7.09217668e-01 8.71898413e-01 -8.68988812e-01
-4.61942583e-01 -1.24225870e-01 5.68872511e-01 -4.68702242e-02
-3.13094139e-01 -3.39359105e-01 -7.14577377e-01 -4.34225410e-01
-1.54881909e-01 -8.45208406e-01 -2.65515417e-01 -2.30656952e-01
-7.10582286e-02 -4.98332202e-01 1.98863328e-01 -6.78027451e-01
1.48985147e+00 -2.07910156e+00 2.76828468e-01 1.54645279e-01
1.72478288e-01 1.55851856e-01 -6.55492246e-01 1.94755659e-01
2.91259974e-01 -2.23094802e-02 -3.53488445e-01 2.42688823e-02
1.14840657e-01 5.56176722e-01 -6.00797057e-01 -2.23851323e-01
6.02967739e-01 1.33675063e+00 -1.11805356e+00 -4.21821356e-01
-3.85475904e-01 -2.59561062e-01 -1.16271555e+00 5.69094718e-01
-1.19155812e+00 4.51080576e-02 -7.77349770e-01 5.71123540e-01
3.67664874e-01 -5.38128853e-01 1.81303158e-01 9.44366455e-02
2.88747549e-01 8.11642408e-01 -9.98381674e-01 1.90116251e+00
-6.34842217e-01 3.49913165e-02 -1.29064485e-01 -8.60072076e-01
9.75485921e-01 1.38869444e-02 -8.95943642e-02 -1.27587962e+00
-1.27659976e-01 3.07901412e-01 1.31041303e-01 -7.16916084e-01
6.65584087e-01 1.32085621e-01 -1.39282286e-01 6.17166042e-01
7.86583647e-02 -1.02983937e-01 2.72975862e-01 3.08169723e-01
1.38413620e+00 2.15285987e-01 -6.52031600e-02 9.44836214e-02
7.56504238e-01 1.30033925e-01 5.60687959e-01 8.93993616e-01
-9.88344997e-02 1.44920826e-01 7.24460065e-01 -1.74550384e-01
-8.50474179e-01 -9.90827560e-01 4.10621494e-01 1.49023759e+00
5.15141636e-02 -5.43828666e-01 -8.54018807e-01 -8.30605090e-01
2.94357929e-02 9.08852935e-01 -3.22777271e-01 -5.71962774e-01
-9.67701852e-01 -2.91662484e-01 8.33368361e-01 5.30139565e-01
6.10567451e-01 -1.33006632e+00 -6.14228725e-01 4.42606747e-01
-3.70731473e-01 -7.78089464e-01 -6.58436298e-01 8.10967833e-02
-7.95532286e-01 -1.02572238e+00 -4.21940684e-01 -8.27033579e-01
5.70689023e-01 5.05406223e-02 1.39232540e+00 4.74969149e-01
7.93365389e-02 2.93656766e-01 -4.16112691e-01 8.93898606e-02
-6.14140511e-01 3.17997545e-01 -3.66216779e-01 -3.00682396e-01
3.30666482e-01 -5.32446682e-01 -6.08840764e-01 3.38517964e-01
-1.08742106e+00 1.23771816e-01 5.99598765e-01 1.09570742e+00
6.53121889e-01 -4.36939806e-01 1.08142626e+00 -6.01369679e-01
8.92815411e-01 -6.24980807e-01 -8.97305846e-01 4.66495574e-01
-6.29800141e-01 6.46708548e-01 9.88983750e-01 -6.58683300e-01
-8.43369246e-01 -2.21337974e-01 -2.94029564e-01 -4.16360587e-01
1.84072912e-01 8.68618727e-01 1.44789647e-03 1.29164755e-01
8.41580331e-01 5.62027693e-01 -6.42221719e-02 -3.55989963e-01
6.90908790e-01 4.37290400e-01 5.49772501e-01 -1.11932325e+00
6.09801769e-01 -1.09659366e-01 -4.18707192e-01 -2.46145859e-01
-8.13318551e-01 -2.37902462e-01 1.15272954e-01 -2.64512212e-03
5.88732719e-01 -7.02406466e-01 -1.12345016e+00 2.57168710e-01
-1.18086672e+00 -7.55790651e-01 -4.34695601e-01 1.59534678e-01
-7.30806351e-01 2.06362039e-01 -7.59768486e-01 -7.86467612e-01
-4.75172579e-01 -1.25861466e+00 7.34492660e-01 3.02602887e-01
-1.08997233e-01 -4.68588382e-01 1.08012937e-01 5.57276309e-01
5.28438926e-01 -2.47616827e-01 1.51258624e+00 -5.87779403e-01
-1.19061911e+00 3.27689916e-01 -6.47708848e-02 4.07048523e-01
-3.22067350e-01 -4.76169229e-01 -8.83781731e-01 -1.83627054e-01
-1.03635341e-02 -9.18311536e-01 8.59830439e-01 -1.67911738e-01
1.47418225e+00 -4.73877609e-01 7.86467791e-02 3.75369817e-01
1.16712093e+00 2.76383936e-01 6.57103539e-01 2.72251248e-01
4.08708036e-01 3.39399248e-01 6.09020829e-01 2.52224118e-01
7.89048851e-01 5.08977830e-01 5.32669723e-01 4.90795642e-01
4.76331590e-03 -7.06457496e-01 6.51246130e-01 1.15145409e+00
3.28268528e-01 7.64709041e-02 -1.00669992e+00 7.71674573e-01
-1.94416225e+00 -7.73416281e-01 1.61187738e-01 2.08402038e+00
1.38243580e+00 1.72854498e-01 3.90369482e-02 -1.29012793e-01
1.28940105e-01 1.45079225e-01 -1.05557323e+00 -3.98035049e-01
2.11004302e-01 5.11206865e-01 7.17714801e-02 6.11264765e-01
-4.39137757e-01 1.05015683e+00 5.54553413e+00 9.72462595e-01
-8.91009450e-01 -6.79829866e-02 3.93540651e-01 -2.12358326e-01
-7.17114151e-01 1.06079489e-01 -7.17734635e-01 4.34199631e-01
1.37158585e+00 -2.18368441e-01 1.05144048e+00 1.05602241e+00
-1.99376985e-01 -2.01168180e-01 -1.40019071e+00 6.39574230e-01
-1.26531631e-01 -1.52596068e+00 2.05338970e-01 -3.79433244e-01
3.86720806e-01 2.72030234e-02 1.93812400e-02 1.06946194e+00
4.26683396e-01 -9.54599202e-01 7.79020131e-01 6.63807571e-01
5.24127483e-01 -7.13329911e-01 2.66826987e-01 8.33341599e-01
-1.00409472e+00 -3.95610482e-01 -3.50784838e-01 -7.14024454e-02
1.75321996e-02 3.36971492e-01 -6.63167179e-01 5.23163021e-01
4.53635365e-01 3.00137252e-01 -7.24636436e-01 7.26883411e-01
-4.16723579e-01 7.72060275e-01 -3.65727663e-01 -5.32492816e-01
2.80114859e-01 -1.21673040e-01 1.34647176e-01 6.43259108e-01
1.40883997e-01 3.70638698e-01 8.30657873e-03 1.43380392e+00
-3.07648987e-01 -8.97027329e-02 -1.28634021e-01 -1.43867329e-01
6.80455863e-01 8.81778777e-01 6.84851259e-02 -4.35351759e-01
-4.52159971e-01 7.71776676e-01 8.24668765e-01 5.09840786e-01
-8.87622058e-01 -3.88963938e-01 4.48985279e-01 -1.77182585e-01
4.77440238e-01 -6.87133446e-02 -5.50943874e-02 -1.30667460e+00
3.46872747e-01 -1.42894626e+00 3.86420429e-01 -8.58212471e-01
-9.50033605e-01 4.52160567e-01 -1.19106233e-01 -7.13388920e-01
-6.44815385e-01 -2.74384797e-01 -4.55214322e-01 1.07348371e+00
-1.84650004e+00 -6.27419770e-01 -2.06506789e-01 6.67576790e-01
2.87599355e-01 -1.42840063e-02 8.94091785e-01 2.73925692e-01
-6.39829516e-01 8.74931514e-01 -1.23120211e-01 7.23291561e-02
3.66919041e-01 -1.09538126e+00 5.33948898e-01 4.24516976e-01
-5.64118698e-02 9.03694212e-01 2.12776661e-01 -4.03031021e-01
-1.96408427e+00 -1.02795351e+00 7.55872846e-01 -4.33823019e-01
6.92825079e-01 -2.73680300e-01 -1.30739295e+00 5.23160338e-01
-1.57043204e-01 -1.67871147e-01 5.29660821e-01 3.93407419e-02
-6.84268415e-01 -1.23693280e-01 -9.70907629e-01 7.97748983e-01
8.31128120e-01 -8.26507807e-01 -9.42083001e-01 1.07665285e-02
1.24964070e+00 -4.87764388e-01 -9.33649838e-01 1.48199394e-01
3.25255096e-01 -6.49283886e-01 7.87497818e-01 -7.01500535e-01
7.60461152e-01 -3.16003174e-01 -1.16882555e-01 -1.12434804e+00
-2.02254280e-01 -5.73800623e-01 -7.34434187e-01 1.02013469e+00
6.33001029e-01 -4.97546732e-01 1.02622080e+00 6.90415621e-01
-1.34704888e-01 -1.30138087e+00 -1.01760495e+00 -6.83850586e-01
2.11417317e-01 -5.83032012e-01 9.83234465e-01 6.52883649e-01
9.01807100e-02 5.37367582e-01 -6.23828098e-02 7.15889931e-02
2.41474345e-01 4.39394265e-01 7.95989752e-01 -6.95820332e-01
-8.58785212e-01 -3.36604595e-01 4.01416838e-01 -1.80575073e+00
9.99387577e-02 -1.01160812e+00 4.24205996e-02 -1.16097534e+00
1.10256821e-01 -6.14141107e-01 -2.48712644e-01 6.87664807e-01
-4.48231816e-01 -6.73046470e-01 2.28809223e-01 1.71103805e-01
-7.33277142e-01 8.45053673e-01 1.45625365e+00 -1.70316860e-01
-2.65008301e-01 -2.60325551e-01 -7.60901868e-01 3.84612590e-01
7.40107954e-01 -5.52382052e-01 -6.08872235e-01 -6.09765768e-01
6.56955063e-01 5.08004129e-01 4.38227415e-01 -7.45063782e-01
4.44408059e-01 -3.22424144e-01 -1.51157379e-01 -5.38408697e-01
1.96127951e-01 -6.02555871e-01 -4.15579766e-01 6.00352585e-01
-7.59210944e-01 4.87591065e-02 2.17359081e-01 6.33973360e-01
-1.83055133e-01 -5.90654433e-01 5.17321169e-01 -3.62362787e-02
-5.89320838e-01 1.66320890e-01 -1.43930048e-01 3.84826958e-01
7.29647934e-01 2.30864137e-01 -7.07010865e-01 -3.75207186e-01
-3.53649080e-01 8.64978135e-01 2.26867720e-01 4.63713527e-01
7.46868312e-01 -1.33822691e+00 -3.35994780e-01 2.44512767e-01
2.96921611e-01 3.09578270e-01 2.63358980e-01 5.81865132e-01
-3.12906414e-01 5.02024174e-01 1.55266240e-01 -3.08624685e-01
-7.07763016e-01 7.41450965e-01 4.94191140e-01 -6.46143496e-01
-2.96487123e-01 6.62344396e-01 -2.87617706e-02 -7.28921115e-01
3.03309828e-01 -8.31218898e-01 -1.41528035e-02 -3.95110279e-01
5.79338133e-01 2.26544902e-01 -1.24393158e-01 1.87094241e-01
-2.48454094e-01 4.17772643e-02 -3.01116824e-01 -1.21621639e-01
1.14727926e+00 3.23898375e-01 -3.65607627e-02 1.08521022e-01
1.08563280e+00 -1.76497892e-01 -1.18164551e+00 -5.48352122e-01
8.15644041e-02 -5.31152226e-02 -5.91175556e-02 -1.05891430e+00
-7.37724245e-01 7.59957433e-01 1.16592325e-01 1.18474320e-01
1.24331093e+00 1.20081343e-01 9.82151330e-01 1.11525464e+00
4.19036150e-01 -1.01402843e+00 5.78674257e-01 9.15417969e-01
8.88137937e-01 -9.81909275e-01 -3.84114355e-01 -6.65471926e-02
-5.04409254e-01 8.32435727e-01 1.03297949e+00 -6.09087236e-02
1.06733076e-01 -2.28606872e-02 -3.23631585e-01 -1.06986903e-01
-1.31617463e+00 -1.28831998e-01 -4.39562760e-02 4.09652293e-01
7.49319792e-02 -2.05198571e-01 -2.44688436e-01 1.21320796e+00
-2.35130042e-01 3.54739755e-01 2.87576497e-01 1.06377316e+00
-6.45641863e-01 -1.10877573e+00 -1.94645852e-01 5.64317524e-01
-1.22755811e-01 -3.73591840e-01 -1.24700591e-01 4.27476555e-01
-2.39961326e-01 9.68120813e-01 -7.18087479e-02 -5.12344658e-01
6.77625239e-01 2.71485597e-01 2.98803091e-01 -5.11802852e-01
-8.87303412e-01 -4.27272677e-01 3.22042316e-01 -8.15505087e-01
6.07826710e-02 -2.00034410e-01 -1.59635949e+00 -3.59994709e-01
-4.65961576e-01 3.24585915e-01 3.57955337e-01 9.34130192e-01
6.70969963e-01 6.19702041e-01 6.68179631e-01 -2.03431956e-02
-1.26124012e+00 -7.99614072e-01 -1.15564175e-01 2.02885881e-01
4.05482382e-01 -2.23035038e-01 -7.41506591e-02 -7.52358288e-02]
|
[10.898393630981445, 7.85999870300293]
|
68ea3923-a996-4e4a-9d40-70f37b9e443a
|
adversarial-intrinsic-motivation-for
|
2105.13345
| null |
https://arxiv.org/abs/2105.13345v3
|
https://arxiv.org/pdf/2105.13345v3.pdf
|
Adversarial Intrinsic Motivation for Reinforcement Learning
|
Learning with an objective to minimize the mismatch with a reference distribution has been shown to be useful for generative modeling and imitation learning. In this paper, we investigate whether one such objective, the Wasserstein-1 distance between a policy's state visitation distribution and a target distribution, can be utilized effectively for reinforcement learning (RL) tasks. Specifically, this paper focuses on goal-conditioned reinforcement learning where the idealized (unachievable) target distribution has full measure at the goal. This paper introduces a quasimetric specific to Markov Decision Processes (MDPs) and uses this quasimetric to estimate the above Wasserstein-1 distance. It further shows that the policy that minimizes this Wasserstein-1 distance is the policy that reaches the goal in as few steps as possible. Our approach, termed Adversarial Intrinsic Motivation (AIM), estimates this Wasserstein-1 distance through its dual objective and uses it to compute a supplemental reward function. Our experiments show that this reward function changes smoothly with respect to transitions in the MDP and directs the agent's exploration to find the goal efficiently. Additionally, we combine AIM with Hindsight Experience Replay (HER) and show that the resulting algorithm accelerates learning significantly on several simulated robotics tasks when compared to other rewards that encourage exploration or accelerate learning.
|
['Peter Stone', 'Scott Niekum', 'Mauricio Tec', 'Ishan Durugkar']
|
2021-05-27
| null |
http://proceedings.neurips.cc/paper/2021/hash/486c0401c56bf7ec2daa9eba58907da9-Abstract.html
|
http://proceedings.neurips.cc/paper/2021/file/486c0401c56bf7ec2daa9eba58907da9-Paper.pdf
|
neurips-2021-12
|
['multi-goal-reinforcement-learning']
|
['methodology']
|
[ 6.07844489e-03 5.69045246e-01 -1.50466204e-01 -3.33826840e-02
-8.51373613e-01 -4.04839844e-01 7.78603494e-01 5.91141172e-02
-8.97620738e-01 1.09286654e+00 -9.09310952e-02 -2.14839339e-01
-3.03678423e-01 -7.21494496e-01 -1.00339067e+00 -1.16918528e+00
-3.46857876e-01 6.91363871e-01 -1.01094976e-01 -8.07036534e-02
3.60915393e-01 5.40674388e-01 -1.28233194e+00 -5.71752071e-01
7.85424292e-01 4.77406710e-01 4.94296819e-01 8.97975683e-01
1.85336247e-01 8.52878332e-01 -6.36846185e-01 4.35381792e-02
4.85153913e-01 -8.20156038e-01 -8.10471773e-01 -1.18595041e-01
-3.61392684e-02 -4.30360228e-01 -2.48500228e-01 1.27730525e+00
4.09844816e-01 5.29724061e-01 9.07260358e-01 -1.50367975e+00
-3.83494854e-01 6.62885189e-01 -4.38434273e-01 2.70160418e-02
1.74718291e-01 2.75518328e-01 7.37738431e-01 -3.30684185e-01
6.45297766e-01 1.32929075e+00 2.90708005e-01 8.64852190e-01
-1.45725083e+00 -3.34017664e-01 -3.81093659e-02 -1.68899700e-01
-9.51145113e-01 1.03575021e-01 5.66243887e-01 -3.61422509e-01
6.07090473e-01 -2.96812236e-01 7.08135486e-01 1.25232983e+00
6.45330071e-01 9.01198030e-01 1.46219254e+00 -4.66527909e-01
8.85490417e-01 2.58412883e-02 -4.24038768e-01 7.71318436e-01
1.18997111e-03 7.18955994e-01 -2.62729645e-01 1.08698361e-01
9.20736670e-01 -3.54507565e-02 -1.39773309e-01 -8.66682649e-01
-1.12044418e+00 9.72909570e-01 3.53254080e-01 2.39571169e-01
-5.58146536e-01 7.23969817e-01 -1.52524799e-01 5.51112473e-01
-7.82113299e-02 6.18268132e-01 -8.07993412e-02 -5.38549483e-01
-5.44099629e-01 6.01377428e-01 9.87493455e-01 8.88326466e-01
9.38638568e-01 2.78894216e-01 -1.85767025e-01 4.37443674e-01
3.68546456e-01 8.50172400e-01 2.92509943e-01 -1.44063556e+00
3.45936358e-01 -3.31361815e-02 5.00950813e-01 -5.45170844e-01
-2.38574743e-01 -4.44277734e-01 -3.70834768e-01 9.95479465e-01
7.10024297e-01 -4.13221359e-01 -7.08322763e-01 2.30255795e+00
3.84843051e-01 5.07125184e-02 4.16355789e-01 7.61230826e-01
-1.94139645e-01 5.77189326e-01 2.38461699e-02 -3.20471734e-01
4.69130874e-01 -6.83242500e-01 -3.79241228e-01 -1.97231188e-01
6.52396381e-01 -3.89067829e-01 1.21181011e+00 4.22472954e-01
-1.11881614e+00 -1.94893494e-01 -1.04607308e+00 6.64108217e-01
7.94988498e-02 -3.30258727e-01 1.68776736e-01 5.67745984e-01
-1.10254741e+00 1.24629319e+00 -1.16315246e+00 -3.63026232e-01
2.60502845e-01 2.29500830e-01 -4.93473113e-02 1.51360199e-01
-8.39773476e-01 1.24724901e+00 3.30096096e-01 -3.78533632e-01
-1.83441687e+00 -3.03047121e-01 -6.73658133e-01 -2.20509052e-01
4.35848534e-01 -4.64126974e-01 1.51884401e+00 -7.42550373e-01
-1.96512043e+00 4.44781452e-01 3.36695462e-01 -7.49241710e-01
7.74700582e-01 -1.08570158e-01 1.97546721e-01 1.28173396e-01
1.44504279e-01 9.16783214e-01 1.13704133e+00 -1.26690674e+00
-6.02322221e-01 -4.21694905e-01 1.64483890e-01 5.16443133e-01
9.50489044e-02 -6.09419346e-01 2.86721915e-01 -4.09700394e-01
-4.85881828e-02 -1.10585749e+00 -4.49322104e-01 -1.94516137e-01
-2.63981104e-01 -2.98947692e-01 4.90046293e-01 -1.33115098e-01
4.23361093e-01 -1.77376986e+00 4.90684420e-01 2.72637457e-01
-5.00411615e-02 -9.18962657e-02 -3.01279575e-01 5.26494622e-01
3.85295749e-01 -2.35349745e-01 -4.72285241e-01 -3.80197078e-01
3.11335444e-01 5.97116411e-01 -5.71919680e-01 7.70359755e-01
-4.65559550e-02 7.26241529e-01 -1.43459153e+00 -3.29189688e-01
1.64101377e-01 2.80244529e-01 -5.41355848e-01 3.37975830e-01
-2.94274390e-01 7.89458036e-01 -4.08382744e-01 1.11761495e-01
3.27478081e-01 4.16598290e-01 7.10491985e-02 6.58508062e-01
-7.79976472e-02 4.89496551e-02 -1.03645241e+00 1.64562857e+00
-5.08728921e-01 4.23077494e-01 -7.90051669e-02 -9.70840454e-01
1.24714756e+00 -8.33422467e-02 5.68543017e-01 -5.34321189e-01
2.15697601e-01 2.49005973e-01 -5.32746762e-02 -2.14550152e-01
3.84817094e-01 -5.30332267e-01 -3.25243063e-02 6.80567324e-01
2.77994514e-01 -5.03526688e-01 2.85288393e-02 2.00239494e-01
1.12444484e+00 7.72760689e-01 2.16281623e-01 -4.32971954e-01
1.37552068e-01 -4.53477427e-02 4.44871902e-01 9.83864427e-01
-4.32655603e-01 1.98582560e-01 6.99202359e-01 1.34514332e-01
-1.13804805e+00 -1.52596200e+00 1.15803987e-01 8.03051531e-01
3.38907063e-01 3.55063044e-02 -8.89236867e-01 -8.16309214e-01
7.82321766e-02 1.15634835e+00 -8.56136084e-01 -3.99809331e-01
-4.68002826e-01 -1.96423963e-01 4.60444152e-01 4.45858628e-01
4.42658275e-01 -1.30008292e+00 -1.05282128e+00 3.42264235e-01
1.11130439e-01 -4.19913203e-01 -4.35972780e-01 5.86638331e-01
-9.75802004e-01 -1.01566792e+00 -9.83832657e-01 -5.54861307e-01
5.65566957e-01 -5.88978603e-02 6.89659715e-01 -6.86406374e-01
1.86264485e-01 9.30666208e-01 -2.67114490e-01 -3.93655449e-01
-8.27446043e-01 -2.50758827e-01 2.85684973e-01 -2.63538122e-01
1.73840359e-01 -6.82969511e-01 -4.22964990e-01 2.58893132e-01
-8.09194505e-01 -2.35380724e-01 3.86246592e-01 7.32383609e-01
6.63479507e-01 -8.24272707e-02 5.59308767e-01 -3.45989317e-01
7.75931954e-01 -4.13423061e-01 -9.43509996e-01 -6.24410696e-02
-7.93246865e-01 7.44652152e-01 6.61972582e-01 -6.26668751e-01
-8.69884253e-01 -8.13744813e-02 8.32723305e-02 -6.12297714e-01
-5.72960153e-02 9.06508192e-02 5.16478978e-02 3.24065872e-02
6.93889678e-01 4.00853783e-01 3.53717595e-01 -2.26780280e-01
5.54182410e-01 1.86272472e-01 5.65490723e-01 -8.70477974e-01
8.46697211e-01 4.43013102e-01 4.97826070e-01 -5.01834512e-01
-6.84027612e-01 -3.80458198e-02 -4.37767476e-01 -4.22681302e-01
6.51179194e-01 -4.75244433e-01 -1.06491053e+00 2.45876968e-01
-8.07826579e-01 -9.51002777e-01 -9.84584570e-01 9.14435983e-01
-1.53554010e+00 2.21839562e-01 -3.51533383e-01 -1.32061684e+00
9.50651318e-02 -1.01124942e+00 6.56021595e-01 2.51969934e-01
-7.00027915e-03 -1.05136931e+00 6.10187888e-01 -3.80385041e-01
2.40714535e-01 5.42851269e-01 5.60908437e-01 -3.88691187e-01
-5.24550259e-01 2.27346614e-01 4.76842314e-01 4.44275886e-01
-5.30790761e-02 -3.21578801e-01 -6.05162382e-01 -5.63164175e-01
3.48907709e-01 -5.07573128e-01 8.66018653e-01 5.36215425e-01
5.97961128e-01 -4.36638832e-01 1.18704804e-04 4.36448336e-01
1.38778841e+00 4.66827869e-01 5.47245502e-01 5.94663322e-01
4.06193078e-01 6.02828622e-01 9.81537461e-01 5.66679835e-01
1.78435177e-01 3.19510132e-01 9.08333778e-01 5.53119898e-01
2.24603742e-01 -7.72444367e-01 8.58044863e-01 3.83786678e-01
-3.08676418e-02 1.15108810e-01 -6.39590561e-01 5.18625140e-01
-1.94174993e+00 -1.11851943e+00 3.25373173e-01 2.62433100e+00
7.79934824e-01 1.84947491e-01 3.35749418e-01 -1.16377451e-01
5.83672345e-01 -7.68944994e-02 -1.00709522e+00 -6.52172506e-01
9.63625386e-02 2.23372579e-01 7.71441400e-01 9.12286937e-01
-7.57698059e-01 7.51755416e-01 6.67554283e+00 7.18282759e-01
-7.42341340e-01 1.06046163e-01 2.52191246e-01 -2.95283720e-02
-3.24380189e-01 6.33215159e-02 -7.06379533e-01 3.41515213e-01
9.43902075e-01 -3.23706061e-01 8.65824997e-01 1.22896600e+00
2.49250382e-01 -2.75681436e-01 -1.33288360e+00 6.66889668e-01
-2.88571537e-01 -8.31432641e-01 -3.06799233e-01 4.61855590e-01
8.23469877e-01 -7.76494155e-03 3.31856728e-01 5.78397512e-01
1.00920975e+00 -9.16009605e-01 8.25142980e-01 5.78760207e-01
4.34394628e-01 -1.22653782e+00 3.95791560e-01 7.25383103e-01
-8.45512569e-01 -1.55903429e-01 -4.61043864e-01 3.37550193e-02
2.39254180e-02 1.73027769e-01 -1.16130877e+00 1.12891480e-01
2.18814448e-01 6.74846411e-01 1.02250651e-01 9.12565768e-01
-5.77298701e-01 3.76249880e-01 -2.76762068e-01 -3.57281715e-01
7.10867763e-01 -6.03166401e-01 9.95640993e-01 8.12623560e-01
5.10332465e-01 -4.47407305e-01 2.50239909e-01 1.12014639e+00
1.37561619e-01 -4.47887890e-02 -8.24130893e-01 -6.72358349e-02
2.02831462e-01 9.14562523e-01 -6.45387590e-01 -9.99764800e-02
4.04579073e-01 9.44854677e-01 5.02154350e-01 3.69358540e-01
-8.46131682e-01 -3.95248830e-01 7.52587140e-01 -2.54272819e-01
2.03513607e-01 -2.98745751e-01 1.87522650e-01 -7.72337735e-01
-3.02719295e-01 -5.09694099e-01 1.74615204e-01 -5.60523033e-01
-9.91623342e-01 3.53785455e-01 3.19042951e-02 -1.25646782e+00
-6.62212431e-01 -2.97849834e-01 -4.78694558e-01 8.48340690e-01
-1.35276711e+00 -5.38670599e-01 1.27189323e-01 6.80283606e-01
3.23929667e-01 -2.68879771e-01 6.17851436e-01 -4.04474765e-01
-1.31626144e-01 3.69935602e-01 4.56088871e-01 -2.28033289e-01
4.19109315e-01 -1.98680556e+00 8.52496997e-02 6.93281412e-01
1.38560191e-01 3.45404863e-01 1.06013465e+00 -6.74670935e-01
-1.43517780e+00 -1.06087101e+00 2.82287598e-01 -3.79631817e-01
7.01884806e-01 1.08719833e-01 -5.60182512e-01 8.10364544e-01
9.57148895e-02 -3.33665162e-01 5.52916378e-02 -2.98626751e-01
-1.16544820e-01 -1.58363655e-01 -1.34580159e+00 7.44283557e-01
8.36065352e-01 -2.25992054e-01 -6.76635742e-01 2.55001247e-01
6.14072025e-01 -3.09147298e-01 -7.22652972e-01 3.02708475e-03
3.98959190e-01 -9.12852228e-01 6.74944162e-01 -7.64052570e-01
2.40465537e-01 -2.27026299e-01 -2.22604156e-01 -1.85436571e+00
-7.91591927e-02 -1.11760068e+00 -3.30483884e-01 8.00106704e-01
-2.39005014e-02 -8.01487207e-01 8.25196147e-01 1.10996291e-01
-1.55621823e-02 -8.12756002e-01 -9.71754074e-01 -1.30429280e+00
5.44303060e-01 -3.37209314e-01 2.35283434e-01 3.98552299e-01
1.29737690e-01 5.86595107e-03 -4.22366291e-01 1.16247153e-02
1.16600502e+00 1.12211905e-01 7.92366743e-01 -8.44552994e-01
-5.03329098e-01 -4.62608933e-01 -2.80128926e-01 -1.25508749e+00
2.44011909e-01 -9.65385139e-01 5.29574454e-01 -1.44675577e+00
-8.62069577e-02 -5.85375428e-01 -4.21260357e-01 1.19058460e-01
2.13520944e-01 -4.25781012e-01 4.30070072e-01 2.43771598e-01
-5.05465209e-01 9.34727669e-01 1.56796551e+00 7.43161887e-02
-5.41640162e-01 2.23731101e-01 -3.96572948e-01 7.27828801e-01
1.05132568e+00 -9.08585489e-01 -5.47100246e-01 5.13916239e-02
8.62302035e-02 3.62813503e-01 3.45147371e-01 -1.04966450e+00
8.36531073e-02 -4.97379780e-01 4.99247499e-02 -4.07686263e-01
3.78121018e-01 -5.83665371e-01 -1.68373600e-01 9.97256219e-01
-8.67556155e-01 -1.45151466e-01 -2.93691814e-01 1.04312074e+00
2.95499951e-01 -6.70340776e-01 9.87472832e-01 -1.41800597e-01
-3.36264253e-01 1.86189488e-01 -6.16650045e-01 3.51059556e-01
1.22853279e+00 7.02118203e-02 2.77118180e-02 -7.01518834e-01
-8.37297201e-01 3.15716714e-01 6.60720170e-01 4.89649102e-02
6.50182307e-01 -1.26668882e+00 -5.67410648e-01 -1.87793151e-01
-1.89973369e-01 -2.47153416e-01 -2.20271721e-01 7.93739498e-01
-2.40110219e-01 1.57153532e-01 -4.78739470e-01 -6.47758007e-01
-7.54215837e-01 6.27553344e-01 5.82187176e-01 -3.88164669e-01
-6.30375147e-01 6.68592453e-01 -4.55852859e-02 -5.07023036e-01
3.93954575e-01 -3.45920861e-01 8.30096379e-02 -2.37977251e-01
4.20320988e-01 5.84152281e-01 -5.74332535e-01 -3.51127535e-01
1.55399600e-03 3.74780148e-01 1.58593208e-01 -8.07476401e-01
1.21113658e+00 -1.73363432e-01 2.92226493e-01 7.06649840e-01
1.14852524e+00 -3.07813883e-01 -2.20625329e+00 -1.73115596e-01
6.79658502e-02 -5.39394855e-01 -1.61039144e-01 -5.03631532e-01
-7.62245297e-01 9.27710950e-01 6.10656142e-01 3.46933275e-01
5.68081021e-01 -4.44732979e-02 4.23511684e-01 6.73343539e-01
7.40962148e-01 -1.35437727e+00 4.77022946e-01 5.15248597e-01
8.71733785e-01 -8.89336765e-01 -4.50886101e-01 4.34779346e-01
-9.39487040e-01 9.80894744e-01 4.03526038e-01 -5.85767925e-01
2.90242523e-01 1.49183705e-01 -3.05846304e-01 2.79448926e-01
-5.77061176e-01 -4.95834887e-01 -2.77436435e-01 9.96262252e-01
-2.97732800e-01 3.30066681e-01 -7.19881207e-02 -3.46955329e-01
-2.66037732e-01 -1.21764906e-01 8.00033987e-01 1.16866755e+00
-9.88329291e-01 -1.09722853e+00 -3.96807998e-01 1.69859529e-01
-1.96658656e-01 2.16809869e-01 8.21861550e-02 7.51985610e-01
-2.64670551e-01 9.05619502e-01 -7.23308846e-02 -1.37418672e-01
4.57115434e-02 -6.12610132e-02 1.05459917e+00 -4.53900278e-01
-1.56902760e-01 1.45205548e-02 -4.13451046e-01 -7.20437646e-01
-4.13864076e-01 -7.46959865e-01 -1.47482038e+00 -2.25196734e-01
9.55660194e-02 3.67749393e-01 9.00694311e-01 9.00389433e-01
-1.04165323e-01 2.85101831e-01 8.86840224e-01 -7.70790279e-01
-1.39158547e+00 -8.13329160e-01 -8.81496429e-01 9.25469324e-02
4.92159009e-01 -8.69285703e-01 -5.78162134e-01 -2.48134017e-01]
|
[4.099189758300781, 2.1504199504852295]
|
1913b7ab-b1b0-4089-b6b5-cb44ae0b37a3
|
adapting-sequence-to-sequence-models-for-text
|
1904.06100
| null |
http://arxiv.org/abs/1904.06100v1
|
http://arxiv.org/pdf/1904.06100v1.pdf
|
Adapting Sequence to Sequence models for Text Normalization in Social Media
|
Social media offer an abundant source of valuable raw data, however informal
writing can quickly become a bottleneck for many natural language processing
(NLP) tasks. Off-the-shelf tools are usually trained on formal text and cannot
explicitly handle noise found in short online posts. Moreover, the variety of
frequently occurring linguistic variations presents several challenges, even
for humans who might not be able to comprehend the meaning of such posts,
especially when they contain slang and abbreviations. Text Normalization aims
to transform online user-generated text to a canonical form. Current text
normalization systems rely on string or phonetic similarity and classification
models that work on a local fashion. We argue that processing contextual
information is crucial for this task and introduce a social media text
normalization hybrid word-character attention-based encoder-decoder model that
can serve as a pre-processing step for NLP applications to adapt to noisy text
in social media. Our character-based component is trained on synthetic
adversarial examples that are designed to capture errors commonly found in
online user-generated text. Experiments show that our model surpasses neural
architectures designed for text normalization and achieves comparable
performance with state-of-the-art related work.
|
['ChengXiang Zhai', 'Kabir Manghnani', 'Ismini Lourentzou']
|
2019-04-12
| null | null | null | null |
['lexical-normalization']
|
['natural-language-processing']
|
[ 6.47834361e-01 3.04202107e-03 5.66052534e-02 -4.07235503e-01
-8.24269295e-01 -6.54321909e-01 4.85190511e-01 8.35024416e-01
-9.12236094e-01 5.21054685e-01 5.71309090e-01 -3.85156155e-01
3.81450593e-01 -6.61037385e-01 -7.00611830e-01 -4.73076701e-02
6.74955606e-01 3.75800580e-01 -8.62802863e-02 -8.00111294e-01
1.99157745e-01 3.57562959e-01 -1.17922556e+00 5.30860722e-01
9.95576382e-01 3.56786519e-01 2.33773425e-01 9.52558875e-01
-6.75382435e-01 6.54086232e-01 -8.61298025e-01 -9.33826506e-01
2.37161443e-01 -3.63833696e-01 -6.16051495e-01 -3.56074452e-01
4.93128568e-01 -1.09578200e-01 -2.29667321e-01 1.35396540e+00
7.50843763e-01 1.71145782e-01 7.33119965e-01 -7.40885675e-01
-9.13683593e-01 9.50696409e-01 -1.82888955e-01 3.70211452e-01
3.98122668e-01 -1.25533948e-02 8.72000873e-01 -9.30971324e-01
6.72104001e-01 1.16288030e+00 9.18531060e-01 8.65490556e-01
-1.23615015e+00 -3.04885536e-01 1.54633284e-01 -9.15774927e-02
-9.26629543e-01 -6.68420732e-01 5.47275305e-01 -2.33671010e-01
1.27640033e+00 3.28806192e-01 1.46316752e-01 1.56854510e+00
2.90078074e-01 7.46537864e-01 3.29786271e-01 -8.46839130e-01
-1.00056097e-01 1.09790966e-01 3.92268859e-02 2.02106446e-01
2.44165823e-01 -7.72118628e-01 -4.98191714e-01 6.14631698e-02
1.55473188e-01 1.14462338e-01 -1.17582805e-01 2.47890741e-01
-1.27649438e+00 7.19880164e-01 1.33734182e-01 4.78471905e-01
-4.03644651e-01 -9.30563062e-02 8.32358122e-01 4.89628315e-01
6.14144027e-01 7.95781076e-01 -6.80501640e-01 -3.70816320e-01
-1.03865790e+00 2.78391182e-01 9.45825994e-01 1.19514883e+00
4.15070087e-01 -1.03432447e-01 -3.20322841e-01 1.06214464e+00
-1.39426263e-02 6.40784681e-01 1.02302802e+00 -3.15583706e-01
8.42836201e-01 7.00081348e-01 -8.02382603e-02 -1.06343091e+00
-3.95569474e-01 -3.00389946e-01 -9.87969160e-01 -3.02658975e-01
5.50755203e-01 -2.72315323e-01 -7.29705215e-01 1.54331350e+00
6.06779493e-02 -5.97141087e-01 5.82320578e-02 1.95549309e-01
7.66092777e-01 6.56312585e-01 1.56397551e-01 -1.41212627e-01
1.16688132e+00 -9.70641851e-01 -1.05412197e+00 -6.42591596e-01
8.54794025e-01 -1.36458266e+00 1.74184322e+00 1.55998766e-01
-1.03008521e+00 -3.94911736e-01 -1.00095165e+00 -5.96047938e-01
-9.12752211e-01 -2.07133554e-02 5.59813827e-02 8.36738825e-01
-6.05989993e-01 8.57663035e-01 -7.00955808e-01 -7.79303849e-01
3.93240511e-01 4.66093458e-02 -2.83516228e-01 5.94672002e-02
-1.21060598e+00 9.31475222e-01 2.06667721e-01 7.33018667e-02
7.95048196e-03 -8.37788880e-01 -9.96464849e-01 -7.17172623e-02
2.22205773e-01 -6.18014812e-01 1.61123061e+00 -1.28916669e+00
-1.61423504e+00 7.90065646e-01 -3.18816751e-01 -4.55704242e-01
7.93293953e-01 -5.83102703e-01 -6.01788223e-01 -2.91167498e-01
1.82802707e-01 2.01535627e-01 1.08905208e+00 -7.08900034e-01
-3.26211005e-01 -1.32371932e-01 -3.14657360e-01 1.64718673e-01
-8.91750813e-01 5.70247948e-01 -3.88176858e-01 -1.19705403e+00
-1.43433586e-01 -6.72252953e-01 -2.40511835e-01 -2.04438478e-01
-6.81386054e-01 2.28386614e-02 6.12131417e-01 -1.13887537e+00
1.61483192e+00 -1.84763312e+00 -5.45136631e-02 2.17901945e-01
-1.58568844e-01 5.71376324e-01 -2.44528711e-01 7.48550415e-01
-3.17272241e-03 5.70316553e-01 -3.21673363e-01 -8.37135851e-01
2.76375741e-01 6.68546483e-02 -4.14564162e-01 1.98452190e-01
2.64909685e-01 9.52623487e-01 -9.67324972e-01 -2.52113253e-01
-2.73850728e-02 5.19613922e-01 -6.90885186e-01 2.64791936e-01
-2.31328592e-01 2.30778620e-01 -4.72530201e-02 3.53609890e-01
3.77537668e-01 1.70577794e-01 -8.86685029e-02 4.82255109e-02
5.79793788e-02 6.29934907e-01 -9.60895598e-01 1.75542045e+00
-6.13714039e-01 7.90417194e-01 -3.84847969e-02 -5.29981673e-01
6.29721582e-01 1.25176802e-01 -7.69482478e-02 -5.52582860e-01
4.56883639e-01 1.27903476e-01 1.33539075e-02 -6.22285783e-01
1.05574083e+00 -3.29047590e-02 -3.70416164e-01 5.24740100e-01
6.35727718e-02 -3.49737018e-01 5.90953827e-01 3.20828021e-01
1.29473913e+00 -1.02614291e-01 4.73030984e-01 7.63596296e-02
6.57731831e-01 -2.59447128e-01 3.77091229e-01 9.79821861e-01
-1.26324356e-01 9.74294066e-01 5.52425683e-01 -2.08061859e-01
-1.43675983e+00 -6.46896362e-01 1.48794800e-01 1.55437279e+00
-5.42635143e-01 -7.48795867e-01 -1.12510347e+00 -6.20905399e-01
-2.05175564e-01 9.23751891e-01 -5.75942576e-01 -2.32173756e-01
-6.72902346e-01 -5.89867532e-01 6.63919628e-01 5.20171106e-01
-8.73436928e-02 -1.19474089e+00 -4.62619141e-02 5.48507452e-01
-2.95854539e-01 -1.25056660e+00 -8.91059399e-01 1.88932106e-01
-3.77109826e-01 -7.80984163e-01 -4.46434677e-01 -7.91524470e-01
7.75885344e-01 -1.05038406e-02 9.61085320e-01 1.32461801e-01
-1.08477704e-01 2.84357429e-01 -6.73954308e-01 -1.02065444e+00
-1.10786176e+00 7.14027166e-01 1.24890245e-01 1.39300659e-01
6.30432904e-01 -5.19681275e-01 -8.83033574e-02 -5.38545661e-02
-1.24956834e+00 1.77865829e-02 3.41081530e-01 7.92673528e-01
2.91799158e-01 -3.18452179e-01 5.82113206e-01 -1.27019334e+00
1.19671392e+00 -3.00345212e-01 -1.07415080e-01 1.67491645e-01
-1.50390655e-01 2.98490990e-02 1.33831084e+00 -6.22478664e-01
-9.77021456e-01 8.89445841e-02 -5.33083558e-01 9.07813087e-02
-2.52012491e-01 5.83228290e-01 -4.73394483e-01 2.96613723e-01
9.81670797e-01 3.26932892e-02 -1.31973237e-01 -7.22576141e-01
4.73152280e-01 1.10038936e+00 7.09065855e-01 -3.77412677e-01
1.04059947e+00 1.41253784e-01 -6.09223425e-01 -1.00075209e+00
-1.14747322e+00 -5.67225158e-01 -9.24352467e-01 1.22022927e-01
6.44110322e-01 -5.94449699e-01 -3.08449626e-01 4.10421133e-01
-1.54001641e+00 -3.24850947e-01 -4.40989405e-01 1.33604854e-01
-1.88462764e-01 5.41848242e-01 -6.20947599e-01 -5.12946427e-01
-7.49619305e-01 -6.84389651e-01 8.03213298e-01 1.03991225e-01
-7.92577207e-01 -9.03649211e-01 1.86056402e-02 3.70449901e-01
7.23679900e-01 9.93027315e-02 8.90668392e-01 -1.22706354e+00
2.00068682e-01 -7.38285780e-01 -6.41652420e-02 7.13584304e-01
1.33252844e-01 2.24875450e-01 -1.02114153e+00 7.82622490e-03
-1.75766528e-01 -1.92314386e-01 6.95416570e-01 -1.29793748e-01
1.14239264e+00 -6.09015465e-01 1.56845257e-01 4.92253900e-01
8.58340919e-01 -3.76378208e-01 6.79729760e-01 5.21593332e-01
1.01009333e+00 4.82662976e-01 1.56584382e-01 3.42234850e-01
3.27890754e-01 2.40918607e-01 1.71774596e-01 1.45657146e-02
-1.99180841e-02 -5.80964148e-01 5.02451181e-01 1.25946510e+00
4.13333297e-01 -6.11417472e-01 -9.95715678e-01 5.93306601e-01
-1.77517521e+00 -8.28413725e-01 -1.65580690e-01 2.06496620e+00
1.33447540e+00 3.75793993e-01 -2.68620193e-01 3.68833065e-01
7.55560040e-01 7.21763121e-03 -2.90946573e-01 -8.57890368e-01
-2.61273205e-01 3.34719032e-01 6.31010771e-01 4.47664469e-01
-1.31614506e+00 1.14307714e+00 5.27522087e+00 8.91862512e-01
-1.00224757e+00 1.74184963e-01 1.72666714e-01 -1.11110896e-01
-1.06236704e-01 -3.94405901e-01 -9.21448946e-01 5.14208078e-01
1.13398385e+00 -1.31873265e-01 5.34677565e-01 7.50853479e-01
3.52507174e-01 3.23812008e-01 -1.38738000e+00 1.01405513e+00
4.17518020e-01 -9.39761519e-01 2.49013215e-01 -4.59588349e-01
7.64601111e-01 1.25866458e-01 -8.70230719e-02 4.55763221e-01
1.75849169e-01 -1.12198782e+00 8.97389352e-01 4.93782371e-01
5.80256760e-01 -7.03047991e-01 9.31042314e-01 4.74716723e-01
-6.15339577e-01 2.18371227e-01 -5.20454586e-01 -2.52633601e-01
1.39347240e-01 5.84000468e-01 -8.98720026e-01 2.54560169e-02
5.00117242e-01 6.36714697e-01 -8.46681476e-01 6.58922613e-01
-5.24342418e-01 5.00756919e-01 -4.52503324e-01 -3.25136125e-01
4.27609719e-02 2.66047508e-01 5.90592682e-01 1.78760850e+00
2.03241959e-01 -3.66732389e-01 -1.54473916e-01 4.84054476e-01
-7.02459037e-01 9.01123703e-01 -4.50068027e-01 -5.72538316e-01
3.06990176e-01 1.18753111e+00 -7.43220568e-01 -2.98428804e-01
-4.64652300e-01 1.28017461e+00 5.39695263e-01 1.90363035e-01
-5.60411215e-01 -9.90987659e-01 6.65235221e-01 2.09226921e-01
1.07941970e-01 -2.12013632e-01 -5.70322573e-01 -1.55984163e+00
3.24567646e-01 -1.25028348e+00 1.19274288e-01 -6.16354585e-01
-1.61922312e+00 6.53098226e-01 -6.45626783e-01 -1.06552541e+00
-1.69921577e-01 -6.58868194e-01 -5.20714998e-01 8.42704117e-01
-1.32513690e+00 -1.20599687e+00 -1.50983557e-01 4.71024007e-01
6.79801166e-01 -2.08883643e-01 1.08543241e+00 4.66870010e-01
-7.58308649e-01 1.06950724e+00 4.92325932e-01 6.33284807e-01
1.25769067e+00 -1.25264931e+00 9.12922978e-01 1.39089215e+00
1.55376583e-01 8.72806489e-01 7.83100963e-01 -7.77650654e-01
-1.19968295e+00 -1.41875160e+00 1.58666503e+00 -7.21927702e-01
8.91645372e-01 -7.03890264e-01 -1.05079532e+00 7.53555000e-01
2.44078487e-01 -1.26155868e-01 9.75989282e-01 7.97984228e-02
-4.16837782e-01 3.73035446e-02 -9.72092390e-01 1.17120457e+00
9.34895933e-01 -8.33749533e-01 -9.09146667e-01 4.80429411e-01
7.61222124e-01 -5.84792852e-01 -4.82590109e-01 -1.24042109e-01
3.38753849e-01 -3.34880412e-01 6.59928203e-01 -1.00218856e+00
7.07317591e-01 -7.33886957e-02 -2.88528632e-02 -1.52114630e+00
8.59999657e-03 -1.22330523e+00 2.62655854e-01 1.60594881e+00
7.77479887e-01 -4.33710963e-01 4.31681365e-01 8.44101191e-01
-1.84061363e-01 -1.19846195e-01 -6.21030927e-01 -5.30755818e-01
2.58042395e-01 -9.87697303e-01 5.73641896e-01 1.11805642e+00
1.93693057e-01 5.74419081e-01 -3.00740808e-01 -1.27140388e-01
1.54603183e-01 -6.92794561e-01 9.74315584e-01 -1.02788186e+00
-3.07557150e-03 -5.68783402e-01 -3.30301076e-01 -7.88990319e-01
3.52282256e-01 -1.07289994e+00 1.37447044e-01 -1.44851446e+00
-1.68074429e-01 2.31020339e-03 7.29399770e-02 4.97236282e-01
-2.86851257e-01 4.61310953e-01 3.58580321e-01 -1.50430322e-01
-5.67276537e-01 4.88940746e-01 8.44253063e-01 -2.24797785e-01
-4.02659982e-01 -8.76651853e-02 -9.03201163e-01 1.06613433e+00
7.85400569e-01 -8.08566213e-01 1.52813688e-01 -6.05774164e-01
9.18858111e-01 -7.64767110e-01 -1.04587294e-01 -8.67418826e-01
3.96769881e-01 -1.90865621e-02 3.45177323e-01 -3.04169983e-01
-9.03145149e-02 -8.53237212e-01 -6.22369289e-01 2.37083714e-02
-6.69914901e-01 1.40508980e-01 2.99613625e-01 3.20112705e-01
9.93438181e-04 -5.34815431e-01 6.74267411e-01 -5.52415149e-03
-8.89337733e-02 -2.27802284e-02 -8.19005430e-01 4.81471151e-01
4.19096202e-01 -6.70078248e-02 -1.65131420e-01 -5.11014223e-01
-5.74257553e-01 -7.95071200e-02 6.00731492e-01 6.65410340e-01
2.41404876e-01 -1.03213418e+00 -8.49923730e-01 3.89934123e-01
2.22820133e-01 4.14328948e-02 -6.40516579e-02 4.54552680e-01
-8.53259325e-01 8.03429540e-03 -1.78016983e-02 -4.99125011e-02
-1.20217705e+00 4.78124648e-01 2.19064295e-01 -3.61036003e-01
-4.18742001e-01 8.76308143e-01 -3.82429034e-01 -8.05190444e-01
4.56808299e-01 -7.76000261e-01 -1.53042898e-01 3.32842231e-01
9.80833530e-01 1.98253632e-01 6.91175282e-01 -7.19583809e-01
-6.51287613e-03 1.99573815e-01 -3.86525363e-01 2.25420538e-02
1.46704853e+00 -3.27403188e-01 -2.21127272e-01 4.51113433e-01
1.27795100e+00 4.66740042e-01 -7.36313820e-01 -5.32996476e-01
2.43453860e-01 -1.66675493e-01 -1.99473947e-01 -6.58517957e-01
-4.16682541e-01 9.62000430e-01 1.68517351e-01 1.79320738e-01
8.91370356e-01 -4.72736627e-01 1.14877295e+00 9.93858218e-01
-3.51659030e-01 -1.72789705e+00 -2.92350203e-02 1.27155197e+00
1.10102677e+00 -1.35793161e+00 -9.23344791e-02 -2.96581507e-01
-4.53150064e-01 1.39962614e+00 2.63406456e-01 1.06916770e-01
5.99684477e-01 4.36011821e-01 2.50237048e-01 4.11827594e-01
-3.70428801e-01 -4.85829674e-02 2.79652953e-01 6.93705261e-01
7.17527092e-01 -2.16998756e-01 -2.94294357e-01 1.08188331e+00
-8.05272877e-01 -3.53047729e-01 8.74624610e-01 9.87447560e-01
-1.61354586e-01 -1.33461976e+00 -4.53947872e-01 5.38362265e-01
-1.04502690e+00 -6.07594430e-01 -5.11688232e-01 3.34779412e-01
4.04480565e-03 8.42069328e-01 8.27859193e-02 -1.24711186e-01
6.55734539e-01 5.33815384e-01 2.12660834e-01 -1.03704417e+00
-1.17787898e+00 -1.41774774e-01 1.02249064e-01 -3.12371135e-01
5.90690188e-02 -7.72862315e-01 -1.09717834e+00 -5.73095202e-01
1.63559180e-05 -2.89328247e-01 9.79438007e-01 1.04812920e+00
1.85096890e-01 5.78064799e-01 1.57319054e-01 -9.29238141e-01
-5.62167704e-01 -1.39094841e+00 -1.33081421e-01 8.31786811e-01
3.77861261e-01 1.58070311e-01 -2.82822937e-01 5.40882885e-01]
|
[10.826698303222656, 9.993329048156738]
|
8d1ba5ae-e2b5-44bd-a5de-6b1d54a42043
|
improving-knowledge-extraction-from-llms-for
|
2306.06770
| null |
https://arxiv.org/abs/2306.06770v2
|
https://arxiv.org/pdf/2306.06770v2.pdf
|
Improving Knowledge Extraction from LLMs for Robotic Task Learning through Agent Analysis
|
Large language models (LLMs) offer significant promise as a knowledge source for robotic task learning. Prompt engineering has been shown to be effective for eliciting knowledge from an LLM but alone is insufficient for acquiring relevant, situationally grounded knowledge for an embodied robotic agent learning novel tasks. We describe a cognitive-agent approach that extends and complements prompt engineering, mitigating its limitations, and thus enabling a robot to acquire new task knowledge matched to its native language capabilities, embodiment, environment, and user preferences. The approach is to increase the response space of LLMs and deploy general strategies, embedded within the autonomous robot, to evaluate, repair, and select among candidate responses produced by the LLM. We describe the approach and experiments that show how a robot, by retrieving and evaluating a breadth of responses from the LLM, can achieve >75% task completion in one-shot learning without user oversight. The approach achieves 100% task completion when human oversight (such as indication of preference) is provided, while greatly reducing how much human oversight is needed.
|
['Peter Lindes', 'Robert E. Wray', 'James R. Kirk']
|
2023-06-11
| null | null | null | null |
['one-shot-learning', 'prompt-engineering']
|
['methodology', 'natural-language-processing']
|
[ 1.51812568e-01 6.82487428e-01 -5.12178093e-02 -3.45931143e-01
-8.87516260e-01 -8.82117808e-01 5.87999046e-01 3.19908768e-01
-5.84153235e-01 6.23897374e-01 2.67653167e-01 -2.91488439e-01
-4.50113922e-01 -1.92643553e-01 -5.39905012e-01 -1.71366557e-01
1.13328114e-01 6.76835716e-01 2.28267044e-01 -3.49644929e-01
6.28061056e-01 7.82528996e-01 -1.69950318e+00 1.93899944e-01
7.49142051e-01 5.24095118e-01 1.05684030e+00 5.96474946e-01
-8.23709890e-02 1.37120163e+00 -4.49200004e-01 3.12726229e-01
7.64033273e-02 -1.06258631e-01 -1.46610367e+00 5.37150949e-02
-8.96441266e-02 -5.84445059e-01 -1.21430129e-01 7.14978456e-01
2.89065808e-01 4.93488342e-01 5.42644739e-01 -1.28475142e+00
-5.39478004e-01 5.35171926e-01 2.14281991e-01 -1.62004501e-01
8.41214657e-01 3.64247203e-01 4.93830830e-01 -8.28518033e-01
5.96459270e-01 1.50698435e+00 2.93366224e-01 7.73732245e-01
-9.11883950e-01 -2.49738112e-01 2.14658380e-01 1.28370121e-01
-1.09571564e+00 -7.60531962e-01 5.45170248e-01 -6.50127888e-01
1.33986413e+00 -1.37526125e-01 2.87432045e-01 1.02414978e+00
7.44088292e-02 9.40865040e-01 9.40664470e-01 -7.99221098e-01
5.31198382e-01 5.49730659e-01 8.71042982e-02 8.13960135e-01
-5.21118753e-02 3.58057648e-01 -9.37269568e-01 -2.17143640e-01
7.60734260e-01 -4.22134787e-01 -2.06577927e-01 -7.55235672e-01
-1.36876595e+00 3.86536717e-01 2.32182771e-01 3.75547111e-01
-9.34648514e-01 1.73491195e-01 1.89210474e-01 6.36631966e-01
-2.76453018e-01 1.34290397e+00 -6.38235569e-01 -3.55243355e-01
-1.21423014e-01 2.49423325e-01 9.61696804e-01 1.40229475e+00
8.96004379e-01 5.56644984e-02 -1.86196744e-01 7.58420944e-01
6.57768548e-02 4.65739727e-01 5.12726843e-01 -1.72346675e+00
8.29058886e-02 7.51671851e-01 8.03000331e-01 -7.30945766e-01
-6.63412690e-01 1.21187698e-02 4.25765783e-01 5.42952478e-01
3.59944105e-02 -3.00451845e-01 -5.33139169e-01 1.76553106e+00
2.92895228e-01 -7.87608147e-01 7.49007881e-01 8.42086375e-01
4.16419923e-01 5.57876587e-01 3.39813173e-01 -2.04987288e-01
1.21321797e+00 -9.13280845e-01 -3.95359904e-01 -8.84689927e-01
9.72496271e-01 -4.37583745e-01 1.34812605e+00 5.33107698e-01
-9.45857346e-01 -4.79984730e-01 -9.19835091e-01 8.85277241e-02
-1.52344495e-01 1.74613178e-01 7.52086818e-01 3.88083607e-02
-1.17335558e+00 3.06262642e-01 -4.00862366e-01 -9.31804597e-01
-1.40195727e-01 4.23500508e-01 -5.45355380e-01 -4.04523283e-01
-8.81258965e-01 1.43886971e+00 6.22022033e-01 -5.05433194e-02
-1.40223825e+00 -1.16737440e-01 -9.25799847e-01 -7.63100758e-02
6.60252333e-01 -4.61607754e-01 1.95086300e+00 -7.57684410e-01
-1.70189917e+00 6.00352645e-01 1.64324239e-01 -1.29851907e-01
1.76733956e-02 -1.49376854e-01 -5.07600121e-02 2.91241437e-01
3.45972270e-01 1.04523814e+00 6.25407517e-01 -1.38724768e+00
-8.48840177e-01 -1.94028407e-01 4.87747490e-01 6.38550878e-01
-1.87560841e-01 2.53115237e-01 -1.98673876e-03 -6.19548894e-02
1.06094874e-01 -1.17217124e+00 -2.80596882e-01 -1.96253881e-01
2.64050007e-01 -6.20692074e-01 3.78948897e-01 -3.25440973e-01
3.13680530e-01 -2.03712893e+00 2.37962037e-01 -4.65551484e-03
-1.07772931e-01 1.00811094e-01 -6.13332391e-01 6.79884255e-01
3.42247397e-01 -2.52866447e-01 2.97879040e-01 1.63347751e-01
2.08067566e-01 2.27513537e-01 -2.26498082e-01 -1.70656621e-01
-1.31499572e-02 8.36255848e-01 -1.23271203e+00 -1.82749435e-01
1.89974055e-01 -4.85330410e-02 -4.03477341e-01 6.70247912e-01
-4.64257181e-01 3.40298176e-01 -6.60265565e-01 5.17216504e-01
-1.21840172e-01 3.92385162e-02 3.75383586e-01 2.41249800e-01
-1.56305984e-01 1.87784612e-01 -8.06919098e-01 1.86884558e+00
-7.45812774e-01 5.31589091e-01 4.09730285e-01 -5.53573012e-01
1.17452276e+00 4.80712771e-01 6.09324425e-02 -5.77398539e-01
6.37720451e-02 2.85411149e-01 -7.43495226e-02 -1.04112160e+00
3.91676784e-01 -1.01569928e-02 -3.99132699e-01 8.29913199e-01
2.06020057e-01 -4.68095422e-01 -2.24488825e-01 2.98495114e-01
1.34913540e+00 2.64004409e-01 4.12912697e-01 -3.61345738e-01
1.83556825e-01 6.05484903e-01 5.46524748e-02 1.18228662e+00
-2.83023983e-01 -3.43185306e-01 4.07904461e-02 -3.98141325e-01
-5.82253098e-01 -7.14179754e-01 7.09296644e-01 1.65615463e+00
1.90015018e-01 -2.59415731e-02 -6.99436128e-01 -2.23019093e-01
-2.69553632e-01 1.41328895e+00 -3.33494604e-01 -6.17480099e-01
-2.38263145e-01 1.68992966e-01 3.75866085e-01 5.45341313e-01
3.19691211e-01 -1.82031572e+00 -1.48323035e+00 3.78203005e-01
-4.52855378e-01 -1.02958417e+00 -1.92158863e-01 4.78028536e-01
-6.82585657e-01 -8.40265632e-01 -2.68371969e-01 -1.16303337e+00
8.73570561e-01 5.07030308e-01 6.68943942e-01 -9.98965427e-02
-4.75238711e-02 1.21721244e+00 -4.92368490e-01 -4.90191072e-01
-8.12363446e-01 -2.64976084e-01 4.84907985e-01 -5.47316670e-01
4.11821753e-01 -3.45710009e-01 6.04905980e-03 2.69751161e-01
-5.22424698e-01 1.31032422e-01 1.04284310e+00 5.44254482e-01
-6.31908700e-03 8.63272548e-02 7.67379284e-01 -3.43393713e-01
1.26411676e+00 -3.05687070e-01 -3.42951864e-01 5.41183591e-01
-8.70897114e-01 3.36449772e-01 1.98137805e-01 -7.22212672e-01
-1.46379673e+00 3.91282767e-01 5.31543970e-01 -9.97801125e-02
-5.19422650e-01 7.78488100e-01 -1.66569240e-02 -2.66916096e-01
1.15095735e+00 2.19535008e-01 1.57913774e-01 -4.55021560e-01
4.22463745e-01 8.69599640e-01 8.21010709e-01 -1.21554494e+00
2.78119534e-01 -1.95428506e-01 -5.65660715e-01 -5.31000376e-01
-5.83280921e-01 -4.70846474e-01 -5.85139513e-01 -4.03096110e-01
4.55142915e-01 -8.77062023e-01 -9.83297706e-01 2.19326556e-01
-1.22845924e+00 -9.50553060e-01 -1.67999327e-01 6.02100372e-01
-1.19929910e+00 -1.44869924e-01 -3.98543328e-01 -1.18662488e+00
-8.08867663e-02 -1.17389464e+00 6.35483265e-01 1.72555968e-01
-7.90878952e-01 -3.78980160e-01 -3.33720505e-01 1.85823977e-01
6.51461661e-01 -4.03057963e-01 1.28029597e+00 -9.60639536e-01
-4.29740578e-01 -3.13165396e-01 -1.08444855e-01 -1.76837116e-01
5.98948784e-02 -7.86822021e-01 -7.28945971e-01 -2.81206876e-01
1.79488361e-01 -1.03597105e+00 -7.25526735e-02 -1.55814394e-01
4.37699139e-01 -4.82816756e-01 -4.02177066e-01 -3.36434186e-01
9.91955161e-01 6.95664644e-01 2.42527649e-01 5.55657685e-01
5.86183332e-02 1.21303821e+00 1.01125121e+00 3.06857854e-01
5.63385427e-01 5.59987009e-01 1.82362035e-01 5.51769674e-01
1.43515602e-01 -3.22937697e-01 6.68705702e-01 6.82424707e-03
3.14523131e-01 7.53908455e-02 -1.05421543e+00 6.76848352e-01
-2.05061960e+00 -6.34301901e-01 5.10106802e-01 1.90385401e+00
7.87970543e-01 -5.80188222e-02 -8.77690539e-02 -3.82060289e-01
3.60198975e-01 -4.66090083e-01 -9.78419363e-01 -6.49791479e-01
3.96414995e-01 -3.40909779e-01 1.93790808e-01 7.32958972e-01
-2.74140835e-01 1.34020686e+00 7.02479792e+00 1.74264640e-01
-6.40169978e-01 -3.74596938e-02 -2.11169705e-01 2.05901694e-02
-2.02761933e-01 -1.68681261e-03 -4.16628391e-01 -3.29432398e-01
7.39702523e-01 -2.21830800e-01 9.96259987e-01 1.18009520e+00
2.05519140e-01 -5.35333693e-01 -1.51158333e+00 8.66387665e-01
2.03388706e-01 -7.91262805e-01 -9.14931893e-02 -3.76119107e-01
1.45113036e-01 -5.67748398e-02 -2.30674759e-01 7.31760979e-01
8.58611465e-01 -8.49054694e-01 9.55060363e-01 6.03839517e-01
5.63120127e-01 -4.71422702e-01 4.11419034e-01 1.16543305e+00
-5.91452122e-01 -6.18379414e-01 -2.44635940e-01 -4.07926768e-01
-7.72212744e-02 -4.19334739e-01 -1.45841777e+00 3.12064420e-02
5.57453573e-01 -1.34580329e-01 -3.71775717e-01 5.49820840e-01
-4.32399482e-01 2.18215361e-02 -1.95852175e-01 -4.85217005e-01
1.33282378e-01 3.48405838e-01 5.78667223e-01 8.25070500e-01
2.77317643e-01 6.49478495e-01 5.74551225e-01 8.43713522e-01
4.27471966e-01 -3.64143178e-02 -8.17161024e-01 -2.25461915e-01
1.04153073e+00 8.20067108e-01 -3.36471736e-01 -2.98666269e-01
-5.30324988e-02 9.49167132e-01 4.75012362e-01 6.58145785e-01
-7.37608150e-02 -4.22541946e-01 3.58474523e-01 -3.29713345e-01
-2.30182946e-01 -2.72533029e-01 8.38799477e-02 -4.76519078e-01
-7.51241669e-02 -1.21671438e+00 1.48660153e-01 -1.61533964e+00
-6.56357527e-01 6.48735464e-01 3.52397531e-01 -8.20215702e-01
-7.87506878e-01 -6.96802497e-01 -1.17434584e-01 8.92627358e-01
-1.04854298e+00 -1.09212124e+00 -4.42143530e-01 4.36277628e-01
8.54195952e-01 -4.95985448e-01 1.14256668e+00 -4.59540009e-01
4.47128788e-02 7.93829039e-02 -6.09376371e-01 -4.50740606e-01
7.36828804e-01 -7.76544988e-01 -2.27243453e-02 3.66254181e-01
-3.73680979e-01 1.02653348e+00 9.72762644e-01 -6.99311554e-01
-1.70016801e+00 -6.51540935e-01 8.38136673e-01 -7.14465916e-01
3.32880050e-01 -1.37056932e-01 -7.56023049e-01 9.97535110e-01
8.06461424e-02 -7.10043848e-01 4.05419201e-01 -3.15256082e-02
-2.75175631e-01 1.21215455e-01 -1.34907711e+00 8.84920120e-01
1.07017124e+00 -6.60110533e-01 -1.17245305e+00 4.60342407e-01
1.02044094e+00 -1.81504577e-01 -4.32979345e-01 2.32400194e-01
4.79292631e-01 -4.60758328e-01 6.70299292e-01 -6.36841536e-01
3.28886462e-03 -5.24393857e-01 -3.12027544e-01 -1.47466516e+00
-5.91978550e-01 -9.14612651e-01 1.32878065e-01 9.66910779e-01
4.47729826e-01 -5.11622667e-01 3.10574055e-01 1.35760701e+00
-2.94582754e-01 -2.50914335e-01 -3.35502982e-01 -8.44613314e-01
-3.20317566e-01 -5.80906212e-01 5.43185949e-01 6.43043041e-01
8.34562778e-01 6.02184057e-01 -1.50687069e-01 2.18854666e-01
3.11591297e-01 -1.36010915e-01 9.62797821e-01 -1.24532962e+00
1.76460985e-02 -3.63105804e-01 1.33034915e-01 -1.26051247e+00
4.06676114e-01 -7.61952519e-01 9.23251271e-01 -1.75546730e+00
1.29831672e-01 -6.06504798e-01 -1.51681513e-01 8.53347063e-01
3.47875983e-01 -6.58858776e-01 2.94649571e-01 5.00823796e-01
-6.99351490e-01 1.81114733e-01 9.43063736e-01 -7.45243505e-02
-5.89348018e-01 -2.99001575e-01 -1.12210238e+00 8.43464136e-01
8.88674021e-01 -3.85637492e-01 -8.03066432e-01 -6.50755346e-01
3.44679356e-01 2.96069801e-01 3.04751933e-01 -1.14068973e+00
8.93453419e-01 -5.76785564e-01 1.43923387e-01 -1.72496274e-01
3.45991105e-01 -9.30191100e-01 9.06740576e-02 4.46277291e-01
-9.95464444e-01 -5.89774037e-03 4.40592170e-01 5.42762160e-01
2.70019859e-01 -6.79879606e-01 2.99287081e-01 -6.25152349e-01
-1.56162906e+00 -3.36670607e-01 -1.19098043e+00 -4.35154408e-01
1.02575111e+00 -2.62560725e-01 -2.37498775e-01 -5.60046375e-01
-7.08650649e-01 6.46834970e-01 6.21985137e-01 7.70801365e-01
7.87095606e-01 -1.06376863e+00 -2.61653513e-01 -4.64153895e-03
7.20952511e-01 -1.39457792e-01 -1.76245496e-01 2.98000962e-01
-4.98735532e-02 6.80602789e-01 -3.87160957e-01 -1.63161695e-01
-8.74733806e-01 8.25942755e-01 2.85642743e-01 4.78153735e-01
-4.95237619e-01 9.07141984e-01 2.28076845e-01 -8.35709512e-01
5.18790305e-01 -3.85940112e-02 -3.78818810e-01 -2.83966243e-01
6.19755685e-01 6.46289289e-02 -1.25239387e-01 -2.58007646e-01
-4.49995518e-01 2.24969268e-01 5.88964336e-02 -5.97717762e-01
1.05028296e+00 -4.02363777e-01 -4.32566591e-02 5.73420942e-01
5.24471879e-01 -2.38198757e-01 -1.37060297e+00 -4.50761437e-01
3.44116986e-01 -1.13872424e-01 -1.32608533e-01 -1.35768592e+00
-8.50195903e-03 6.24741614e-01 2.62557834e-01 -1.34484455e-01
7.56386757e-01 3.30054075e-01 3.82490247e-01 1.43753207e+00
1.27185643e+00 -1.36675215e+00 6.37920439e-01 7.19499826e-01
1.22826648e+00 -1.14595318e+00 -2.41560459e-01 3.43273580e-02
-1.02854502e+00 1.18061829e+00 1.04992199e+00 2.88910478e-01
8.08077231e-02 6.37193397e-02 2.95743316e-01 -4.05902475e-01
-1.00014484e+00 -9.53527018e-02 -1.77848056e-01 1.08601022e+00
-8.82285163e-02 -3.25882901e-03 2.16221824e-01 5.92783809e-01
-1.31781891e-01 1.17317095e-01 6.46768987e-01 1.28270614e+00
-1.08207822e+00 -6.11394107e-01 -3.48183721e-01 1.35810569e-01
2.61454493e-01 2.18898997e-01 -8.28155994e-01 4.79155123e-01
-3.48472774e-01 1.49670410e+00 -2.44415343e-01 -3.52452010e-01
6.66925073e-01 3.75956476e-01 6.15782857e-01 -1.04454899e+00
-1.94926456e-01 -3.69803578e-01 4.06879216e-01 -7.05407858e-01
-5.95590025e-02 -5.93514502e-01 -1.50918794e+00 3.21665108e-01
-1.43549815e-01 2.68447548e-01 7.30431974e-01 1.11394727e+00
4.32238907e-01 1.60888597e-01 3.97731125e-01 -8.98475349e-01
-8.37459266e-01 -1.00720394e+00 -8.57405365e-02 1.09714039e-01
3.41118604e-01 -8.17792296e-01 -1.25664532e-01 6.71530142e-02]
|
[4.376928329467773, 0.9198378920555115]
|
36f0ca61-a5f0-45ca-a93e-2d1a6e154ae3
|
sca-streaming-cross-attention-alignment-for
|
2211.00589
| null |
https://arxiv.org/abs/2211.00589v1
|
https://arxiv.org/pdf/2211.00589v1.pdf
|
SCA: Streaming Cross-attention Alignment for Echo Cancellation
|
End-to-End deep learning has shown promising results for speech enhancement tasks, such as noise suppression, dereverberation, and speech separation. However, most state-of-the-art methods for echo cancellation are either classical DSP-based or hybrid DSP-ML algorithms. Components such as the delay estimator and adaptive linear filter are based on traditional signal processing concepts, and deep learning algorithms typically only serve to replace the non-linear residual echo suppressor. This paper introduces an end-to-end echo cancellation network with a streaming cross-attention alignment (SCA). Our proposed method can handle unaligned inputs without requiring external alignment and generate high-quality speech without echoes. At the same time, the end-to-end algorithm simplifies the current echo cancellation pipeline for time-variant echo path cases. We test our proposed method on the ICASSP2022 and Interspeech2021 Microsoft deep echo cancellation challenge evaluation dataset, where our method outperforms some of the other hybrid and end-to-end methods.
|
['Xin Lei', 'Sriram Srinivasan', 'Kaustubh Kalgaonkar', 'Yun Li', 'Yangyang Shi', 'Yang Liu']
|
2022-11-01
| null | null | null | null |
['speech-separation']
|
['speech']
|
[ 7.16779232e-02 -2.89246529e-01 6.39680505e-01 -3.01398933e-01
-9.61672246e-01 -5.23351192e-01 3.99354815e-01 -3.02884430e-01
-5.14913559e-01 2.67298281e-01 5.91228187e-01 -4.34688807e-01
-7.28926212e-02 1.90366641e-01 -5.16185284e-01 -6.07579172e-01
-7.59717301e-02 -1.93178296e-01 1.53951868e-01 -3.89399260e-01
-1.13720752e-01 1.62497729e-01 -1.30546057e+00 1.17653206e-01
1.13505077e+00 9.92718697e-01 2.81932741e-01 1.22895885e+00
2.57426947e-01 5.93446195e-01 -8.28219593e-01 -3.87949616e-01
3.68747592e-01 -5.35774291e-01 2.19426796e-01 -4.22212780e-01
6.37076497e-01 -6.31800532e-01 -8.85192156e-01 1.05616474e+00
1.52635312e+00 7.39403516e-02 3.55402529e-01 -7.94680655e-01
-7.25316033e-02 6.52552545e-01 -3.19165766e-01 3.03319037e-01
2.59703189e-01 1.87110722e-01 5.22965014e-01 -1.27018917e+00
1.10426396e-01 1.03835225e+00 1.23752916e+00 2.42788732e-01
-1.00108027e+00 -1.11283112e+00 -1.78407624e-01 3.57101917e-01
-9.57961440e-01 -1.40578640e+00 8.86820495e-01 6.44548051e-03
1.15930724e+00 2.66222566e-01 5.51588893e-01 9.91322041e-01
5.73468991e-02 1.08964562e+00 9.73205507e-01 -3.97898108e-01
5.72434627e-02 -3.48265409e-01 -3.00762411e-02 1.24078996e-01
-3.54351312e-01 6.32733047e-01 -7.47269094e-01 -1.16093889e-01
5.61941743e-01 -4.03505355e-01 -6.93345726e-01 8.01487416e-02
-1.14151347e+00 2.03903541e-01 2.28206322e-01 2.55549937e-01
-5.47336817e-01 1.24969631e-01 5.75781047e-01 5.55253923e-01
5.31451583e-01 2.77901530e-01 -5.70972621e-01 -3.64385098e-01
-1.59557772e+00 3.23052794e-01 9.08644557e-01 7.86580384e-01
3.69783342e-02 9.27135706e-01 -2.88676709e-01 1.14855850e+00
1.64039642e-01 7.88761556e-01 5.51051557e-01 -8.02654505e-01
4.78205055e-01 -9.17370677e-01 1.51205987e-01 -8.76730859e-01
-5.30534148e-01 -1.20229602e+00 -1.10079908e+00 7.04752877e-02
1.41027465e-01 -7.04905391e-01 -1.00587833e+00 1.87986851e+00
9.73929018e-02 7.51018822e-01 2.51254529e-01 1.23101890e+00
7.37716317e-01 8.29703271e-01 -4.12851155e-01 -3.71783704e-01
8.66121113e-01 -1.35489583e+00 -1.49423754e+00 -4.92462158e-01
1.33527696e-01 -1.15801084e+00 4.34176713e-01 7.09795952e-01
-1.46284413e+00 -6.92754269e-01 -1.14519632e+00 4.75464240e-02
2.63136387e-01 1.11197263e-01 2.51930475e-01 8.01276267e-01
-1.34456480e+00 6.07700527e-01 -7.09003091e-01 4.15895313e-01
-3.09547365e-01 2.35748321e-01 -1.55291269e-02 9.41980928e-02
-1.43155766e+00 5.57015896e-01 -1.50522664e-01 5.24044931e-01
-8.53339076e-01 -1.13842630e+00 -7.12646484e-01 3.57680112e-01
2.67275035e-01 -5.25271118e-01 1.70382750e+00 -1.19659984e+00
-1.90808392e+00 2.73606628e-01 -2.72744179e-01 -7.75175810e-01
5.58386087e-01 -7.97446191e-01 -1.10904181e+00 -3.96902598e-02
-3.23777854e-01 2.52856106e-01 1.48414862e+00 -8.44021440e-01
-5.03941894e-01 6.91920593e-02 -5.99364519e-01 2.84580410e-01
-2.38453373e-01 8.95165130e-02 -4.56065923e-01 -1.17186391e+00
2.80002862e-01 -6.30275726e-01 -1.59706309e-01 -2.92797595e-01
-1.83074653e-01 4.94048148e-01 6.56002641e-01 -1.48801744e+00
1.37806809e+00 -2.45171762e+00 4.57396591e-03 -5.11676781e-02
2.14539155e-01 9.64658976e-01 -4.44946557e-01 4.76794511e-01
-3.07494968e-01 -5.05177855e-01 -1.99193195e-01 -5.42297482e-01
2.16867179e-01 -6.19152606e-01 -4.71580029e-01 5.41741014e-01
-3.24142911e-02 4.33415979e-01 -7.98159778e-01 3.56051922e-02
3.78675610e-01 8.32702279e-01 -7.16460645e-01 4.18215036e-01
3.83662075e-01 3.50149423e-01 4.11116391e-01 1.76580697e-01
1.06146359e+00 6.15966856e-01 8.97018090e-02 -4.26405787e-01
-3.26170474e-01 6.65528893e-01 -1.43554425e+00 1.83608365e+00
-7.82652080e-01 1.14766598e+00 1.15638626e+00 -6.04044378e-01
8.60872030e-01 7.82366693e-01 1.75224900e-01 -1.10808682e+00
-1.17729850e-01 7.88625121e-01 5.18338740e-01 -2.40891203e-01
4.22901660e-01 2.20290665e-02 4.82261866e-01 8.16473439e-02
1.91223547e-01 -1.66061804e-01 -2.08620325e-01 9.51453596e-02
1.26669490e+00 -6.34560511e-02 7.71624371e-02 -1.71436161e-01
4.37742531e-01 -7.41632104e-01 6.66794896e-01 7.69762099e-01
-3.43180090e-01 8.91732395e-01 -1.34304076e-01 2.00646028e-01
-9.82966661e-01 -1.28992784e+00 1.30050838e-01 1.18587530e+00
-2.19834462e-01 -4.26204085e-01 -8.76328468e-01 1.07186206e-01
-3.15907449e-01 7.07365513e-01 2.92528242e-01 -6.61373809e-02
-7.74034560e-01 -1.23321995e-01 8.88178945e-01 2.47571528e-01
4.81682152e-01 -7.04558074e-01 -1.56833783e-01 8.16320002e-01
-4.84562218e-01 -1.12547588e+00 -1.11172509e+00 3.67395729e-01
-6.12455845e-01 -3.08798909e-01 -1.22139549e+00 -7.94061542e-01
1.80287197e-01 5.35416067e-01 8.87542129e-01 -3.33274007e-01
8.85277167e-02 3.00728500e-01 -1.29036888e-01 -5.47522128e-01
-2.64662832e-01 -2.25631654e-01 3.99896026e-01 9.91903022e-02
-1.53575717e-02 -8.08605969e-01 -9.82386112e-01 3.07633251e-01
-6.12659335e-01 -6.07222617e-02 7.64899850e-01 9.58646476e-01
7.99858794e-02 5.93346171e-03 8.31227660e-01 -2.66932428e-01
1.06396151e+00 1.25267673e-02 -5.93216240e-01 -8.52917284e-02
-3.89412314e-01 -4.16700840e-02 8.82820368e-01 -7.71998584e-01
-1.35469949e+00 -1.10002883e-01 -7.51687109e-01 -6.90453649e-01
1.54015809e-01 5.20956874e-01 -2.64453948e-01 9.31837335e-02
4.73137200e-01 2.63784468e-01 -1.79880396e-01 -5.33890188e-01
2.30924428e-01 9.73541021e-01 1.04455113e+00 1.58508971e-01
6.99567854e-01 -2.17872076e-02 -3.20623010e-01 -1.11712480e+00
-2.99488157e-01 -5.93095422e-01 -7.38776773e-02 -1.69108555e-01
2.24415615e-01 -1.21737039e+00 -4.42878276e-01 9.99743462e-01
-1.37147307e+00 -4.38960195e-01 1.61774769e-01 9.58902895e-01
-3.05835128e-01 3.52015078e-01 -7.61909604e-01 -1.04018033e+00
-8.76264751e-01 -9.59204435e-01 7.27049112e-01 8.17000866e-02
-1.08718053e-01 -6.57737195e-01 2.88320243e-01 -2.70039272e-02
1.16468167e+00 -4.42573011e-01 1.31759837e-01 -5.02183139e-01
-9.64220241e-02 -2.38706082e-01 1.85933396e-01 5.80979645e-01
-2.99914688e-01 -4.49859887e-01 -1.46547878e+00 -5.44835269e-01
4.00359511e-01 9.92211550e-02 9.49071109e-01 8.90704989e-01
5.78963459e-01 -1.76666439e-01 5.13874963e-02 1.02125883e+00
9.34958398e-01 3.99956942e-01 8.74563932e-01 -2.33177185e-01
3.23014915e-01 4.02244836e-01 3.04478705e-01 3.21796924e-01
-8.97725299e-03 5.88537753e-01 -5.35804518e-02 -5.78044474e-01
-5.63082337e-01 -1.35035485e-01 7.18739569e-01 1.63551581e+00
5.09574115e-01 -5.14384449e-01 -6.43121719e-01 5.11042833e-01
-1.50024211e+00 -9.50163603e-01 -1.92823708e-01 2.19956923e+00
8.34295452e-01 6.64591938e-02 -1.07776262e-01 3.76646429e-01
6.66641176e-01 3.13013315e-01 -7.66786814e-01 -4.56297010e-01
-2.16982171e-01 5.50065875e-01 5.19608438e-01 8.04732919e-01
-9.37236726e-01 6.54896855e-01 5.70334816e+00 1.04388177e+00
-1.40497804e+00 2.33424217e-01 1.70049235e-01 -2.09120750e-01
-5.32173663e-02 -3.39445770e-01 -4.00182843e-01 -7.72856409e-03
1.23738885e+00 -6.78161830e-02 7.85705149e-01 3.30685914e-01
4.62291181e-01 2.36758620e-01 -1.03731561e+00 1.36410570e+00
9.20860171e-02 -7.10071623e-01 -8.97863209e-01 -3.43572915e-01
4.57268715e-01 2.07440346e-01 2.14076445e-01 4.96774435e-01
-2.53430605e-01 -8.05870116e-01 1.04856908e+00 4.90461946e-01
9.10780489e-01 -6.61072075e-01 4.33640540e-01 3.68397117e-01
-1.28236890e+00 -1.12196982e-01 1.85944997e-02 5.30532654e-03
4.75580692e-01 9.64354277e-01 -6.20573640e-01 4.30787235e-01
3.72538418e-01 4.37216789e-01 1.89871311e-01 1.56435704e+00
-3.19403440e-01 1.01535285e+00 -4.01864946e-01 2.73577064e-01
3.47341150e-02 -6.05051965e-02 1.03132379e+00 1.72141039e+00
4.84384447e-01 3.11795175e-01 -4.25773352e-01 3.94535989e-01
-5.73013462e-02 -2.67796844e-01 -1.79667562e-01 -2.59522852e-02
5.68432212e-01 1.01345932e+00 -1.43294305e-01 -9.72438306e-02
-3.97961199e-01 1.10561740e+00 -2.48936340e-01 7.54778206e-01
-8.97904515e-01 -1.16003144e+00 7.77743757e-01 -8.19601938e-02
4.30241138e-01 -3.96707058e-01 4.58623134e-02 -1.04431915e+00
1.49005622e-01 -1.33765483e+00 -4.52250719e-01 -8.97222757e-01
-9.50933635e-01 7.30822623e-01 -5.93111813e-01 -1.26107991e+00
-4.94918108e-01 -3.43412161e-01 -7.07121074e-01 1.16810298e+00
-1.47149587e+00 -4.90905732e-01 1.64812803e-01 4.81197953e-01
6.44027710e-01 -1.97566092e-01 6.58703446e-01 1.00766361e+00
-1.68868214e-01 8.69477332e-01 6.18339360e-01 6.98105767e-02
1.14778972e+00 -1.02582514e+00 8.30190718e-01 1.41608369e+00
-9.57738757e-02 6.47156954e-01 1.16151571e+00 -5.41283965e-01
-1.41556978e+00 -8.37943017e-01 9.03189480e-01 4.81908083e-01
3.62653196e-01 -7.62503743e-01 -8.52320969e-01 1.33394718e-01
6.98907614e-01 -9.87927020e-02 1.83550641e-01 -7.12385476e-02
-1.84498236e-01 -5.80417633e-01 -6.80024505e-01 7.92812228e-01
1.02979314e+00 -6.41235054e-01 -3.56176049e-01 -3.98869067e-02
7.65463769e-01 -8.14749539e-01 -2.29336694e-01 2.06721529e-01
8.15076590e-01 -1.17770648e+00 1.10300636e+00 -1.33831263e-01
1.98164761e-01 -1.52330011e-01 -1.98440775e-02 -1.92377031e+00
-3.39172661e-01 -1.57012463e+00 -2.59861082e-01 1.31408668e+00
3.65420461e-01 -6.22094691e-01 2.61170447e-01 2.82490283e-01
-7.95915842e-01 -6.42438605e-02 -7.49292254e-01 -9.39042330e-01
-2.63105452e-01 -6.98183358e-01 2.94477969e-01 5.48900723e-01
-2.24333525e-01 2.20518231e-01 -9.76263821e-01 3.80820960e-01
8.33173990e-01 -3.86795074e-01 6.63306773e-01 -6.23678744e-01
-7.73416817e-01 -4.91676450e-01 -9.33136232e-03 -1.67308092e+00
1.40122818e-02 -4.76097375e-01 5.85634172e-01 -1.34718633e+00
-5.97909749e-01 5.34257926e-02 -1.51765272e-01 -2.97049522e-01
-2.80470133e-01 -2.40585744e-01 3.54700118e-01 -3.67483616e-01
-2.95907974e-01 7.28661716e-01 8.97822082e-01 -1.35731921e-01
-3.91821086e-01 1.89148515e-01 -3.53216499e-01 6.78369284e-01
5.47910988e-01 -5.04673123e-01 -3.68716955e-01 -8.04186106e-01
-1.54277205e-01 7.43066132e-01 7.80155212e-02 -1.31906235e+00
8.54494810e-01 5.20748496e-01 1.16654553e-01 -7.84502506e-01
6.28981829e-01 -8.98042440e-01 2.81739563e-01 4.79693830e-01
-4.37528908e-01 -1.73194006e-01 4.41860408e-01 4.16575193e-01
-5.44236541e-01 1.61566228e-01 8.60413253e-01 4.25950229e-01
-2.28696242e-01 2.09006649e-02 -6.77664280e-01 3.04679424e-01
2.16547057e-01 4.90603000e-02 -1.47106186e-01 -1.09546173e+00
-5.64867020e-01 1.05287485e-01 -4.64449644e-01 2.45266154e-01
7.54235804e-01 -1.17687440e+00 -1.18152153e+00 3.42767566e-01
-5.54813504e-01 -5.17970920e-01 6.66764081e-01 1.11532855e+00
-2.75597572e-01 5.19889891e-01 1.06807929e-02 -4.35604393e-01
-1.26106799e+00 3.32658619e-01 8.49669158e-01 -1.37346625e-01
-5.69799542e-01 9.73648846e-01 2.20280588e-01 -2.87989229e-01
9.08403873e-01 -2.35340506e-01 2.46713385e-01 -2.61439860e-01
7.75114655e-01 6.35560989e-01 3.67152601e-01 -3.32313359e-01
-2.65571207e-01 4.54480261e-01 8.67479071e-02 -7.66661406e-01
1.21692681e+00 -3.38873178e-01 2.32524037e-01 2.11010158e-01
1.24353421e+00 5.35113811e-01 -1.15605640e+00 -4.29254740e-01
-4.15593773e-01 -4.35268998e-01 6.35892391e-01 -1.09248710e+00
-1.24052548e+00 1.30915678e+00 1.03225756e+00 -7.94158429e-02
1.73879957e+00 -9.18441117e-01 1.16563594e+00 2.19235390e-01
2.62107477e-02 -1.01637292e+00 -4.98582758e-02 8.29984963e-01
1.15836823e+00 -9.45474446e-01 -3.82377356e-01 -8.11815411e-02
-3.94981742e-01 1.14031923e+00 2.30776623e-01 6.30555227e-02
7.26174951e-01 9.21632528e-01 4.19080079e-01 4.88528043e-01
-6.82904422e-01 -1.44628044e-02 2.91687846e-01 6.81784570e-01
6.08179629e-01 -9.71367583e-02 -3.18388432e-01 8.49930763e-01
-2.41102159e-01 -7.74572715e-02 2.64673442e-01 6.02752268e-01
-2.68960476e-01 -8.67605865e-01 -6.52767062e-01 1.96372509e-01
-5.83322525e-01 -6.55057490e-01 -2.40297541e-01 1.34854466e-01
-3.29657853e-01 1.43745661e+00 4.05655801e-02 -4.55889732e-01
6.45050764e-01 -3.97101827e-02 1.82060480e-01 1.62559614e-01
-1.16189849e+00 1.01606882e+00 2.07572579e-01 -4.14736748e-01
9.03665461e-03 -6.78518772e-01 -1.14997935e+00 -3.76281589e-01
-5.41324019e-01 1.28707495e-02 9.04831946e-01 5.66666007e-01
5.30433893e-01 1.03452039e+00 6.55953467e-01 -1.11026597e+00
-6.86847091e-01 -1.11185360e+00 -5.83304882e-01 1.73448294e-01
1.09366679e+00 5.41658490e-04 -5.05586743e-01 -9.31815952e-02]
|
[15.01501178741455, 5.968545913696289]
|
6766042c-9da3-4022-84c4-c5eee2f2b529
|
unsupervised-learning-of-discourse-aware-text
| null | null |
https://aclanthology.org/P19-2053
|
https://aclanthology.org/P19-2053.pdf
|
Unsupervised Learning of Discourse-Aware Text Representation for Essay Scoring
|
Existing document embedding approaches mainly focus on capturing sequences of words in documents. However, some document classification and regression tasks such as essay scoring need to consider discourse structure of documents. Although some prior approaches consider this issue and utilize discourse structure of text for document classification, these approaches are dependent on computationally expensive parsers. In this paper, we propose an unsupervised approach to capture discourse structure in terms of coherence and cohesion for document embedding that does not require any expensive parser or annotation. Extrinsic evaluation results show that the document representation obtained from our approach improves the performance of essay Organization scoring and Argument Strength scoring.
|
['Paul Reisert', 'Naoya Inoue', 'Kentaro Inui', 'Hiroki Ouchi', 'Farjana Sultana Mim']
|
2019-07-01
| null | null | null |
acl-2019-7
|
['document-embedding']
|
['methodology']
|
[ 1.60511546e-02 3.85589540e-01 -5.98233879e-01 -4.99633461e-01
-5.47942102e-01 -6.33903980e-01 9.12293077e-01 8.14006627e-01
-4.04826730e-01 6.09667122e-01 8.90217066e-01 -4.89158064e-01
-2.21201386e-02 -8.92555296e-01 -2.20135916e-02 -3.48521382e-01
3.04269016e-01 2.16286957e-01 1.37754709e-01 -3.63242686e-01
7.67011523e-01 -1.71320140e-01 -1.05066741e+00 4.82416093e-01
8.98920655e-01 3.95040125e-01 -2.86766496e-02 9.92478609e-01
-6.82666600e-01 1.32644606e+00 -1.07171524e+00 -4.81245935e-01
-4.88559753e-01 -6.90096676e-01 -1.08007443e+00 -8.51446465e-02
1.81102306e-01 -3.67956877e-01 -3.22611839e-01 9.70863163e-01
2.90691525e-01 -4.23268713e-02 9.19088900e-01 -7.37330616e-01
-1.06227469e+00 1.05982649e+00 -4.07843620e-01 3.51998121e-01
5.24368823e-01 -4.49507892e-01 1.75231910e+00 -7.98441768e-01
6.43621802e-01 1.03488970e+00 5.76245487e-01 3.52552354e-01
-1.13297796e+00 -2.39160866e-01 2.18984306e-01 4.14600760e-01
-5.13416111e-01 -2.60363847e-01 1.27367091e+00 -6.05163693e-01
9.82936621e-01 3.38227212e-01 6.88895166e-01 1.09136271e+00
4.65964936e-02 9.39508736e-01 9.08919632e-01 -9.75456476e-01
-6.60580993e-02 1.51525676e-01 1.26483130e+00 7.43712366e-01
4.72145140e-01 -4.70439136e-01 -4.87541974e-01 -1.36170074e-01
2.02209339e-01 -2.86769629e-01 -1.54336631e-01 8.66899565e-02
-9.81428504e-01 1.45025778e+00 -8.64877701e-02 5.56094825e-01
-1.68290362e-01 1.55216992e-01 7.42849648e-01 2.56210417e-01
6.86643124e-01 5.81113994e-01 4.37303670e-02 -4.09263939e-01
-9.48489726e-01 1.56088427e-01 9.97240067e-01 4.55411971e-01
4.68154728e-01 -1.47242799e-01 -3.80731791e-01 8.11078131e-01
5.82439005e-01 -8.21812823e-03 7.92000651e-01 -7.01177418e-01
7.95917988e-01 1.16113126e+00 -2.05918282e-01 -1.39169621e+00
-4.19989347e-01 2.02739257e-02 -4.08261865e-01 -1.76863581e-01
7.81378374e-02 -1.50439829e-01 -3.59949201e-01 1.22524810e+00
1.83195710e-01 -2.66809195e-01 2.75314599e-01 6.68319166e-01
1.22759473e+00 7.60075986e-01 -1.05987772e-01 -9.23218429e-02
1.49371994e+00 -1.31989110e+00 -1.17133725e+00 -6.42783269e-02
1.09542727e+00 -7.81968296e-01 1.19765425e+00 -1.09657541e-01
-1.16903055e+00 -3.17853659e-01 -1.44314885e+00 -5.11826873e-01
-8.65915269e-02 3.21778446e-01 7.19915926e-01 7.98833370e-01
-5.16749918e-01 3.16785634e-01 -7.80584276e-01 -1.55286829e-03
3.24553639e-01 8.85320529e-02 -2.48985156e-01 3.77230167e-01
-1.18465877e+00 1.02390659e+00 4.23161417e-01 -8.17005560e-02
-1.10490091e-01 -3.16368461e-01 -1.03039336e+00 3.81426215e-01
2.84854546e-02 -3.65529299e-01 1.24996209e+00 -6.33962214e-01
-1.69012570e+00 5.87217927e-01 -3.14421773e-01 -4.06557411e-01
-8.96720216e-02 -3.14858824e-01 7.79841021e-02 2.19837025e-01
-6.30752668e-02 6.04025126e-02 3.94824445e-01 -7.70680547e-01
-3.19415569e-01 -2.35579476e-01 6.46252334e-01 2.41015211e-01
-1.24575508e+00 3.63400161e-01 2.47420985e-02 -6.58922613e-01
9.87839792e-03 -6.53377771e-01 7.35353753e-02 -3.01101506e-01
-3.55268359e-01 -7.95355380e-01 9.58377898e-01 -8.34443152e-01
1.66923797e+00 -1.85982406e+00 2.39766091e-01 -1.52653828e-01
3.04210693e-01 1.80788964e-01 -6.71367999e-03 7.52818525e-01
3.70567024e-01 4.40036803e-01 -6.93981424e-02 -3.13618571e-01
3.91533226e-01 2.38803029e-01 -4.41712439e-01 3.55307341e-01
1.64556086e-01 9.28659618e-01 -7.79783607e-01 -9.83913243e-01
5.01624197e-02 3.35791796e-01 -7.66484737e-01 2.55838841e-01
-1.76094115e-01 -9.18930620e-02 -6.10392988e-01 3.83012980e-01
1.29017010e-01 -3.98496240e-01 7.47004330e-01 8.04399997e-02
-1.54745549e-01 1.23184001e+00 -7.86061585e-01 1.51426971e+00
-3.24365765e-01 1.16293299e+00 -3.45089495e-01 -1.28383064e+00
9.84969556e-01 5.56890905e-01 1.65528268e-01 -4.68488872e-01
5.83098903e-02 -2.11003739e-02 4.09875542e-01 -7.40797520e-01
9.15889800e-01 9.00654346e-02 -1.72734588e-01 1.05355418e+00
-7.73150995e-02 1.56037465e-01 6.09330714e-01 3.34393114e-01
1.16997981e+00 -2.08685607e-01 5.46103776e-01 -1.86963648e-01
7.04681754e-01 1.09429851e-01 4.14750874e-01 2.83334076e-01
-1.22144669e-01 3.10953438e-01 9.75141823e-01 -2.09743574e-01
-1.11230826e+00 -5.49897134e-01 -1.55445904e-01 1.24396729e+00
-2.19998434e-01 -9.20626998e-01 -7.11075127e-01 -8.89830410e-01
-1.05142653e-01 5.11292040e-01 -7.02176034e-01 1.76344469e-01
-9.76153374e-01 -7.02236772e-01 7.60397732e-01 8.26500416e-01
2.15820536e-01 -1.01167798e+00 -5.95224082e-01 4.34717447e-01
-2.78682739e-01 -8.07527542e-01 -3.28789085e-01 9.57871303e-02
-8.92046392e-01 -1.41867137e+00 -3.44348133e-01 -1.00610232e+00
7.12128341e-01 1.41548574e-01 8.94673765e-01 3.49290818e-01
-1.46088880e-02 2.79302627e-01 -6.15535975e-01 -3.91858190e-01
-4.71623063e-01 5.28168499e-01 -2.93483466e-01 -6.04542851e-01
5.91767311e-01 -3.11433434e-01 -3.79929304e-01 -2.18298107e-01
-6.99536085e-01 6.46492019e-02 2.11352900e-01 1.17008162e+00
-4.87784147e-02 -3.65855038e-01 7.11376846e-01 -1.12828338e+00
1.45710814e+00 -2.74306983e-01 -2.16047436e-01 4.02772665e-01
-8.32178295e-01 2.76117563e-01 4.75631654e-01 -3.07190269e-01
-1.03725934e+00 -6.10873222e-01 -2.59055495e-02 3.89175713e-01
3.20475399e-01 1.05066550e+00 2.87001103e-01 3.05847824e-01
4.79161114e-01 9.56733003e-02 -1.02267759e-02 -3.37047666e-01
1.83800519e-01 9.27648067e-01 3.93855944e-02 -6.41034067e-01
4.71825421e-01 1.99462637e-01 -2.49350384e-01 -8.61514390e-01
-1.05281818e+00 -4.29677725e-01 -7.03421593e-01 -3.80892046e-02
9.25857425e-01 -5.54366291e-01 -6.20480835e-01 -3.56720656e-01
-1.58041286e+00 3.65059853e-01 -9.87071320e-02 7.37919748e-01
-1.39161840e-01 8.30093563e-01 -8.44373167e-01 -7.15050936e-01
-5.84579945e-01 -9.47162807e-01 6.43229663e-01 1.03282273e-01
-8.00706267e-01 -1.45235586e+00 7.11592734e-01 6.76903486e-01
1.88371330e-01 2.26324737e-01 1.33152449e+00 -7.53620863e-01
-1.94459438e-01 -3.30426276e-01 -1.06873021e-01 3.76000851e-01
2.57711381e-01 1.74208209e-01 -7.87989914e-01 8.51075798e-02
-5.14914878e-02 -4.21106428e-01 8.44766557e-01 7.85036981e-02
9.07316983e-01 -6.67965114e-01 1.27101121e-02 1.80959463e-01
1.04373622e+00 -2.88342144e-02 3.76333207e-01 7.64292479e-01
6.77834511e-01 8.06136549e-01 5.49614012e-01 3.69814396e-01
6.29438281e-01 4.64252084e-01 3.77385952e-02 3.19384485e-01
-2.29398794e-02 -1.11706741e-01 6.59356236e-01 1.71223211e+00
-2.03545496e-01 -3.08594525e-01 -1.01513636e+00 7.19510019e-01
-1.99430990e+00 -1.11452329e+00 -4.48449373e-01 1.26350498e+00
1.23014081e+00 1.92056388e-01 -2.78502584e-01 6.44664228e-01
4.46366608e-01 7.01632202e-01 -4.19283099e-03 -8.51367593e-01
-6.87963888e-02 2.90072858e-01 -9.89827141e-02 6.60198569e-01
-9.10980523e-01 7.48469949e-01 6.98432636e+00 1.72447711e-01
-7.03276157e-01 2.60823339e-01 2.66468786e-02 -1.36793137e-01
-6.02814078e-01 1.69567972e-01 -7.57620871e-01 5.09664893e-01
9.02244270e-01 -5.58286130e-01 -2.73527741e-01 7.44283199e-01
-7.10617006e-02 7.25893825e-02 -1.18131924e+00 4.65806633e-01
2.10586816e-01 -1.64003873e+00 -7.41253197e-02 5.11185974e-02
6.28914952e-01 -5.31644344e-01 -1.06657267e-01 4.12553638e-01
1.33404553e-01 -1.16396737e+00 4.58968192e-01 2.92983651e-01
1.43690526e-01 -6.08609319e-01 9.78760123e-01 3.33253235e-01
-9.10375535e-01 1.01004981e-01 -5.31326234e-01 -4.78237510e-01
8.53795782e-02 4.31370080e-01 -8.98433805e-01 2.87097782e-01
1.37090892e-01 6.54113591e-01 -6.43010080e-01 3.53848159e-01
-6.78363681e-01 1.21286595e+00 2.31451035e-01 -6.84684753e-01
3.65395695e-01 -1.71269238e-01 3.92471492e-01 1.55760944e+00
-5.35834283e-02 1.10700488e-01 1.55722156e-01 4.89945173e-01
-3.33215356e-01 3.60464901e-01 -5.01226783e-01 -6.36053145e-01
5.01166046e-01 1.12560713e+00 -5.41839063e-01 -3.35877687e-01
-5.82762361e-01 5.77729940e-01 4.99056190e-01 -7.84001574e-02
-8.55343878e-01 -5.20241857e-01 4.66881573e-01 1.14937192e-02
1.89118013e-01 -6.48216546e-01 -7.18148291e-01 -1.14715695e+00
2.88328737e-01 -7.62208104e-01 2.54520088e-01 -2.74689525e-01
-9.54958498e-01 4.77449298e-01 -2.16078684e-01 -9.06688035e-01
-4.44637656e-01 -5.81787467e-01 -9.88365769e-01 4.49431151e-01
-1.56071377e+00 -1.07248235e+00 -1.67512938e-01 4.60955612e-02
6.21783257e-01 -3.28131676e-01 1.11315513e+00 1.13675192e-01
-6.33037031e-01 5.58513582e-01 1.98113039e-01 5.27729690e-01
6.62035406e-01 -1.52686703e+00 1.06654882e-01 5.55353343e-01
4.46150690e-01 8.88416350e-01 5.78213871e-01 -3.46992522e-01
-1.08656406e+00 -4.70932662e-01 1.57232082e+00 -4.81585532e-01
9.39781368e-01 -3.02418321e-01 -1.10930705e+00 6.42255783e-01
8.48933697e-01 -5.03441930e-01 1.40089571e+00 7.18604267e-01
-3.43308836e-01 3.46948415e-01 -5.26233554e-01 5.62247753e-01
5.61695695e-01 -8.45513582e-01 -1.39424741e+00 3.99454832e-01
5.57170808e-01 -2.05086172e-01 -9.54920709e-01 -3.42576317e-02
5.31670928e-01 -5.75140536e-01 7.58164108e-01 -1.03483593e+00
1.25752091e+00 -5.58519922e-02 -7.57750273e-02 -9.28213418e-01
-2.94707537e-01 -1.92465305e-01 -7.46176064e-01 1.38145304e+00
4.10692453e-01 -3.35259140e-01 9.59766448e-01 4.79717791e-01
-1.76152498e-01 -7.26515770e-01 -7.14691818e-01 -5.42022884e-01
2.17736781e-01 4.76819687e-02 3.31594527e-01 1.35399318e+00
6.50346339e-01 8.31130326e-01 -1.61306903e-01 -1.84349224e-01
4.41686034e-01 4.09747660e-01 6.21579587e-01 -1.44980478e+00
-1.42890319e-01 -5.66603541e-01 -3.39212954e-01 -7.26979613e-01
5.70334256e-01 -1.20974576e+00 -2.63588279e-01 -2.03310609e+00
4.53338593e-01 1.34018864e-02 -1.98508158e-01 3.03108186e-01
-3.40745121e-01 -1.01351872e-01 5.79487421e-02 3.75794679e-01
-6.90471113e-01 8.51971030e-01 8.66759479e-01 -5.13243437e-01
-9.52061452e-03 -5.61530292e-01 -7.87416935e-01 9.30189013e-01
9.72673833e-01 -6.39649570e-01 -4.15466398e-01 -7.69957066e-01
4.57330853e-01 -1.92590967e-01 7.41812140e-02 -5.58181047e-01
3.78502160e-01 -3.10501277e-01 1.87117502e-01 -6.35098636e-01
1.22451439e-01 -2.39058927e-01 -7.27290630e-01 3.84209424e-01
-7.62907505e-01 3.21823567e-01 -2.31290720e-02 3.37137848e-01
-6.70433342e-01 -1.16876686e+00 3.96137148e-01 1.63444683e-01
-1.91902071e-01 -3.99966151e-01 -6.58113897e-01 8.91666487e-02
8.35558712e-01 -3.06963831e-01 -6.63750648e-01 -2.37600312e-01
-3.69992733e-01 1.26606897e-01 1.99997395e-01 4.97024119e-01
7.01403677e-01 -1.39060342e+00 -8.80482256e-01 -2.77313828e-01
-1.15403347e-01 -1.09190948e-01 -2.72936195e-01 5.49727440e-01
-7.24117756e-01 8.72626543e-01 -1.02597758e-01 -2.78042674e-01
-1.77975357e+00 -9.12073925e-02 -3.19653779e-01 -9.16766763e-01
-5.96963823e-01 6.56278133e-01 -8.48634094e-02 -3.43302369e-01
6.70624748e-02 -2.64622152e-01 -8.38103175e-01 5.35151124e-01
5.57878971e-01 4.02529150e-01 -2.17402175e-01 -3.49101692e-01
-2.07828358e-01 3.18363160e-01 -4.50136870e-01 -2.63070792e-01
1.48499608e+00 1.58541203e-01 -4.01209593e-01 6.44126654e-01
1.12457740e+00 3.46017629e-01 -5.39995611e-01 -8.89718682e-02
8.18005979e-01 -3.84588212e-01 2.03485861e-01 -3.31506997e-01
-4.12758380e-01 1.13636506e+00 2.68340744e-02 4.41115379e-01
4.47936952e-01 -2.28491291e-01 8.52411211e-01 8.57128382e-01
-2.37175778e-01 -1.49956870e+00 5.56500971e-01 1.04671323e+00
8.96906734e-01 -1.21440816e+00 4.13712293e-01 -1.13999218e-01
-5.14961302e-01 1.49471128e+00 5.97016811e-01 -2.11558595e-01
5.27742863e-01 1.90460280e-01 -3.40638608e-02 -3.65249127e-01
-9.54205632e-01 4.21853550e-02 3.85127872e-01 2.19918936e-01
1.31788683e+00 -2.31681466e-01 -1.17445266e+00 8.24284136e-01
-4.56798911e-01 -3.74433488e-01 7.88884223e-01 1.27407765e+00
-6.92881227e-01 -1.51971054e+00 -1.44831359e-01 3.58564228e-01
-6.29258633e-01 -2.60114849e-01 -8.84995401e-01 6.89327180e-01
-3.02500844e-01 7.54232883e-01 1.73007146e-01 -1.78691998e-01
4.36884947e-02 3.66654307e-01 5.79395175e-01 -1.10422611e+00
-8.08954298e-01 -2.91401774e-01 6.00489914e-01 1.18693955e-01
-7.96758354e-01 -6.55866086e-01 -1.44455111e+00 -4.33265090e-01
-6.90042078e-01 4.16368812e-01 8.13097596e-01 8.49457383e-01
9.49013159e-02 8.30102921e-01 3.84029120e-01 -2.19595313e-01
-6.45283878e-01 -1.39562452e+00 -2.44004905e-01 9.96610522e-02
1.28708079e-01 -4.80316162e-01 -1.68221489e-01 1.28750890e-01]
|
[11.015473365783691, 9.332802772521973]
|
d1e38421-2131-4393-8233-41d5a9d1a847
|
using-drug-descriptions-and-molecular
| null | null |
https://academic.oup.com/bioinformatics/advance-article/doi/10.1093/bioinformatics/btaa907/5938075#209442351
|
https://academic.oup.com/bioinformatics/advance-article-pdf/doi/10.1093/bioinformatics/btaa907/34012017/btaa907.pdf
|
Using Drug Descriptions and Molecular Structures for Drug-Drug Interaction Extraction from Literature
|
Motivation
Neural methods to extract drug-drug interactions (DDIs) from literature require a large number of annotations. In this study, we propose a novel method to effectively utilize external drug database information as well as information from large-scale plain text for DDI extraction. Specifically, we focus on drug description and molecular structure information as the drug database information.
Results
We evaluated our approach on the DDIExtraction 2013 shared task data set. We obtained the following results. First, large-scale raw text information can greatly improve the performance of extracting DDIs when combined with the existing model and it shows the state-of-the-art performance. Second, each of drug description and molecular structure information is helpful to further improve the DDI performance for some specific DDI types. Finally, the simultaneous use of the drug description and molecular structure information can significantly improve the performance on all the DDI types. We showed that the plain text, the drug description information, and molecular structure information are complementary and their effective combination are essential for the improvement.
|
['Yutaka Sasaki', 'Makoto Miwa', 'Masaki Asada']
|
2020-10-24
| null | null | null | null |
['drug-drug-interaction-extraction']
|
['natural-language-processing']
|
[ 5.66507764e-02 -4.38922763e-01 -7.42134511e-01 -2.13353381e-01
-8.60202968e-01 -5.51910102e-01 4.71151859e-01 5.16021132e-01
-2.96697050e-01 1.28496528e+00 2.71922857e-01 -3.92507255e-01
-2.81251073e-01 -7.00002134e-01 -7.49774396e-01 -8.22337985e-01
1.69526618e-02 4.74612117e-01 -1.00837816e-02 -7.53899589e-02
1.00004010e-01 7.54962087e-01 -1.11167848e+00 2.96035826e-01
1.43493068e+00 6.92004144e-01 3.18931699e-01 -8.76524672e-02
-3.44104856e-01 6.34696007e-01 -5.87742984e-01 -6.09025471e-02
-4.22353186e-02 -4.31935966e-01 -6.89939797e-01 -4.92890626e-01
6.32659858e-03 -4.51680541e-01 -4.04312521e-01 9.29895222e-01
1.00324929e+00 -2.52090454e-01 8.56071413e-01 -8.40866029e-01
-5.58111012e-01 3.88185412e-01 -4.63980436e-01 -2.22308505e-02
3.93360704e-01 2.78891444e-01 7.77669132e-01 -1.09564030e+00
9.81186032e-01 1.00845921e+00 5.25530696e-01 3.91167402e-01
-7.39602447e-01 -8.19075465e-01 1.62803754e-02 3.70077282e-01
-1.69733262e+00 -3.29526633e-01 6.11013114e-01 -5.27513921e-01
1.52035594e+00 5.46329841e-02 3.06418866e-01 1.00314188e+00
1.81072906e-01 8.53705764e-01 5.51987171e-01 -1.17207669e-01
1.35534719e-01 -2.31593065e-02 2.73951024e-01 6.25235319e-01
4.61639881e-01 1.56635299e-01 -2.45835140e-01 -4.95574564e-01
3.04285109e-01 3.22664052e-01 -5.33644855e-01 1.00496292e-01
-9.14421976e-01 8.70664239e-01 4.21749771e-01 3.67464334e-01
-6.25520229e-01 -3.42526048e-01 5.47742486e-01 -1.79495931e-01
5.02503455e-01 7.67326117e-01 -8.77562046e-01 1.04531407e-01
-5.67906976e-01 1.43070623e-01 8.74367833e-01 8.68214369e-01
5.25112987e-01 -3.49756181e-01 -1.44012421e-01 1.00697637e+00
2.70880163e-01 4.80558723e-01 3.61375958e-01 -1.11729421e-01
6.09485447e-01 7.82756925e-01 2.79514212e-02 -8.57390225e-01
-6.38056159e-01 -3.05123687e-01 -9.50486660e-01 -4.14527059e-01
3.30920517e-02 -3.51147473e-01 -1.22303879e+00 1.71184778e+00
3.41558695e-01 9.70299020e-02 3.06346536e-01 5.78387201e-01
1.48656797e+00 6.83664262e-01 4.39735740e-01 -4.99703079e-01
1.26488554e+00 -6.85753107e-01 -1.14611030e+00 2.11430371e-01
1.09592378e+00 -9.10941482e-01 3.18289220e-01 3.61716092e-01
-6.42624676e-01 -2.65664369e-01 -1.41252863e+00 1.01158069e-02
-7.47987866e-01 3.42106462e-01 8.13991845e-01 1.58008844e-01
-2.05291241e-01 8.30887198e-01 -7.39030600e-01 1.08278111e-01
9.18029189e-01 7.18529880e-01 -6.01527274e-01 -4.08764601e-01
-1.52150738e+00 1.06918454e+00 6.01288259e-01 -1.28140122e-01
-7.73387194e-01 -9.70177948e-01 -7.81813562e-01 -3.39086279e-02
4.48057920e-01 -6.39982879e-01 8.81627798e-01 -1.45696715e-01
-1.10688043e+00 3.02194625e-01 -2.28579864e-01 -2.72661299e-01
1.49189651e-01 -1.42590150e-01 -4.25290406e-01 2.92639434e-01
4.16695848e-02 4.24877375e-01 -5.31287640e-02 -7.74524689e-01
-4.45813358e-01 -6.82235360e-01 -1.98060706e-01 1.09696478e-01
-2.43752435e-01 -1.50623009e-01 -6.34467125e-01 -5.99254906e-01
-1.77475542e-01 -9.61613059e-01 -2.92137295e-01 -4.05755430e-01
-6.43848002e-01 -5.48327208e-01 5.77335477e-01 -6.77942336e-01
1.51438987e+00 -1.79405999e+00 3.26592505e-01 1.57847688e-01
6.81316674e-01 7.08610237e-01 -2.31626034e-01 7.13019729e-01
-4.44330186e-01 4.85965669e-01 -2.14530021e-01 2.64929831e-01
-3.29286128e-01 7.53046423e-02 1.20536543e-01 3.82832497e-01
5.21119356e-01 1.03280187e+00 -7.13287830e-01 -4.98956472e-01
-4.54111323e-02 6.88240349e-01 -4.75488842e-01 2.32553422e-01
-3.81327271e-01 3.96332920e-01 -8.65958869e-01 7.71836460e-01
9.94174063e-01 -4.38270271e-01 1.58487990e-01 -5.22812963e-01
1.32173732e-01 5.54521263e-01 -8.36022258e-01 1.61037517e+00
-7.53097609e-02 -5.79025270e-03 -5.29625356e-01 -7.78481483e-01
7.65680850e-01 5.73331237e-01 9.69144642e-01 -6.44752979e-01
-8.53577908e-03 4.95809555e-01 3.20783496e-01 -6.01837397e-01
-3.79585057e-01 -5.28812874e-03 3.51180226e-01 4.54767235e-02
1.20322928e-01 2.66331971e-01 3.97596389e-01 1.19108513e-01
1.11766386e+00 3.14296894e-02 6.49181366e-01 -1.05575547e-01
6.14957988e-01 7.36196572e-03 7.94367909e-01 2.16617316e-01
2.19952404e-01 1.40715048e-01 5.74571073e-01 -5.19255519e-01
-8.46067667e-01 -4.26558554e-01 -5.55058897e-01 3.68653804e-01
-4.64274250e-02 -8.61290276e-01 -6.11795425e-01 -1.03328955e+00
-1.81942072e-03 3.28525275e-01 -5.09813190e-01 -1.50001451e-01
-4.37517494e-01 -1.53140414e+00 5.86058259e-01 3.41786534e-01
5.24324775e-01 -7.25731850e-01 3.47042173e-01 4.51645702e-01
1.75895507e-03 -9.63108301e-01 -3.94939959e-01 4.28857148e-01
-7.51106441e-01 -1.41664648e+00 -8.34333062e-01 -5.31734705e-01
5.01917779e-01 1.22424504e-02 8.26450884e-01 2.01939747e-01
-4.25322652e-01 -6.08126104e-01 -2.21426502e-01 -7.15594709e-01
-2.68030226e-01 3.14440936e-01 4.56496775e-02 -4.97311562e-01
6.91482306e-01 -5.39560556e-01 -4.88683343e-01 1.75745726e-01
-8.46484303e-01 2.38845516e-02 6.47223175e-01 8.08218479e-01
7.58414745e-01 6.60364479e-02 8.26288581e-01 -1.16421282e+00
7.59239912e-01 -6.54909372e-01 -3.73536468e-01 1.48618430e-01
-7.79991567e-01 4.42800373e-01 6.34984791e-01 -4.64570552e-01
-8.50465000e-01 4.02161866e-01 -7.48701036e-01 -2.29354538e-02
-1.48843557e-01 1.05880010e+00 -6.96057081e-01 9.35162511e-03
6.33053184e-01 1.41134307e-01 -1.74917176e-01 -8.77345920e-01
-1.10159315e-01 8.58892500e-01 -8.73730704e-02 -4.57177252e-01
1.60028383e-01 -1.74769074e-01 2.30359450e-01 -4.80671912e-01
-5.79418480e-01 -3.36822301e-01 -5.31887710e-01 8.35059404e-01
8.43286395e-01 -1.04556060e+00 -1.06770170e+00 4.50596809e-01
-1.38181221e+00 3.03049773e-01 3.87160212e-01 8.54098916e-01
2.59110704e-03 5.40277064e-01 -7.88352489e-01 -2.60382891e-01
-6.76633000e-01 -1.68258309e+00 7.84487903e-01 7.69010708e-02
-3.17855692e-03 -8.30710173e-01 3.03817391e-01 -3.45316567e-02
2.84713916e-02 2.76975721e-01 1.32327724e+00 -1.35208988e+00
-4.24219728e-01 -2.55047023e-01 -3.20851803e-01 1.10745035e-01
6.57037973e-01 -7.30730593e-03 -7.33285785e-01 1.38653681e-01
-6.33412823e-02 -3.76104489e-02 9.45880175e-01 3.77074033e-01
1.08006680e+00 -3.91272783e-01 -7.61012375e-01 6.05340719e-01
1.41253114e+00 9.49041843e-01 6.49374962e-01 -1.77144408e-02
1.06467199e+00 2.24642172e-01 5.22622108e-01 5.03555238e-01
1.58696309e-01 9.60034549e-01 1.37592137e-01 -5.91040313e-01
-1.20631546e-01 -1.70814618e-01 -1.37432814e-01 6.34190261e-01
-6.88386858e-02 -3.03652406e-01 -1.15157270e+00 2.50227332e-01
-1.97719109e+00 -6.51514292e-01 -2.93768674e-01 1.94157374e+00
1.55574942e+00 -1.99911654e-01 -3.98974866e-02 -2.83578008e-01
5.35283744e-01 -4.08222675e-01 -9.88434732e-01 8.58309120e-02
-1.25063822e-01 3.48966926e-01 6.25768960e-01 3.79537910e-01
-1.13752639e+00 8.97526026e-01 6.28412437e+00 1.26092672e+00
-1.06243610e+00 -3.67927521e-01 5.94081402e-01 2.22578973e-01
-1.66896656e-01 -3.95603299e-01 -1.21266329e+00 7.43610978e-01
1.07851005e+00 -2.64488071e-01 1.39162481e-01 7.37878144e-01
1.59627199e-01 1.92273900e-01 -1.33433139e+00 1.12698364e+00
-2.87313521e-01 -1.83188057e+00 3.26735288e-01 4.72898662e-01
5.41029930e-01 8.00292194e-02 -2.63519794e-01 6.72311932e-02
2.01303080e-01 -1.28319085e+00 -4.14152592e-01 3.61803889e-01
7.81841099e-01 -9.63055074e-01 1.23757195e+00 2.27601126e-01
-1.03226531e+00 4.96160388e-01 -1.81029186e-01 3.16320062e-01
-1.66569829e-01 6.93457603e-01 -1.06867886e+00 1.05831039e+00
8.60448256e-02 1.39994991e+00 -6.06945395e-01 1.20267045e+00
-1.04643747e-01 2.90889502e-01 -3.77991676e-01 -1.95267890e-02
2.46285602e-01 -8.74375775e-02 -2.11230703e-02 1.18762302e+00
-3.19939442e-02 4.90248829e-01 4.16014761e-01 5.43475032e-01
-3.41540217e-01 4.90773827e-01 -7.75207818e-01 -4.45581198e-01
4.61263806e-01 8.42861235e-01 -1.53634995e-01 -6.32500410e-01
-4.71762627e-01 6.71190739e-01 -1.11512821e-02 3.81663531e-01
-7.99871266e-01 -7.16208458e-01 8.60219240e-01 -2.65954137e-01
7.68734366e-02 1.15318589e-01 -8.53167698e-02 -1.12550545e+00
-3.32540572e-01 -1.15296435e+00 3.91662300e-01 -3.71957839e-01
-1.45996308e+00 7.61047065e-01 4.57462184e-02 -1.43821836e+00
-1.31037682e-01 -7.14743793e-01 -7.31154159e-02 1.16914427e+00
-1.52113664e+00 -8.90200794e-01 5.78351691e-02 4.54369932e-01
4.49706644e-01 -2.98581898e-01 1.15393293e+00 9.51250553e-01
-1.00868952e+00 6.08771443e-01 3.33757728e-01 2.08047688e-01
9.42089379e-01 -9.45844054e-01 1.92891419e-01 1.26106516e-01
-1.90880865e-01 9.91678059e-01 4.44834352e-01 -1.01822960e+00
-1.42329907e+00 -1.04798520e+00 7.52452135e-01 -2.48981476e-01
3.26500773e-01 4.04321030e-02 -1.10880387e+00 1.98378474e-01
-3.52537185e-02 -2.65111476e-01 1.20772135e+00 7.06079975e-02
-3.02764684e-01 1.76540360e-01 -1.17601538e+00 4.32063848e-01
8.68505239e-01 -4.11273718e-01 -4.96641845e-01 8.04127991e-01
8.97529483e-01 -5.14829397e-01 -1.19148529e+00 7.56477892e-01
3.61951709e-01 -1.08727865e-01 1.20486462e+00 -1.01379132e+00
7.27286994e-01 -4.56688553e-01 -4.85306606e-02 -1.24286616e+00
-1.17809348e-01 -2.31477723e-01 -4.89529110e-02 1.00346828e+00
9.36671972e-01 -6.16823316e-01 4.61740285e-01 6.80328667e-01
-8.55754390e-02 -1.08036947e+00 -6.83206141e-01 -6.47502959e-01
1.22768156e-01 -5.93658276e-02 8.76923263e-01 1.03618252e+00
2.31042042e-01 1.00447512e+00 -5.04414439e-01 -5.84667809e-02
1.27676189e-01 -3.25740688e-02 4.83351767e-01 -1.37572491e+00
-2.75280625e-01 -2.20083565e-01 -2.97369033e-01 -9.03193474e-01
1.36180967e-02 -1.11314094e+00 -2.43843302e-01 -1.73231876e+00
5.57805896e-01 -1.69072151e-01 -4.25142765e-01 8.77937615e-01
-5.23193896e-01 -2.18335077e-01 -5.33420686e-03 3.20247889e-01
-3.31363410e-01 5.92637420e-01 1.43286836e+00 -3.90622735e-01
-5.32896042e-01 -2.12056816e-01 -8.45358551e-01 6.03091359e-01
5.28689802e-01 -6.96470261e-01 -3.26712310e-01 4.14317921e-02
-5.61882295e-02 1.56490445e-01 -2.52790809e-01 -3.92041534e-01
1.46404281e-01 -3.70003998e-01 6.88714266e-01 -7.61561751e-01
2.29408234e-01 -7.88618743e-01 2.66346216e-01 7.39060938e-01
-1.44832090e-01 -1.99603155e-01 6.19071364e-01 5.61450958e-01
-2.85991520e-01 -1.17462724e-01 6.81807339e-01 -2.05830947e-01
-2.68735260e-01 6.79658890e-01 -5.69988638e-02 -2.88347304e-01
7.46136427e-01 6.77570552e-02 -4.27359879e-01 4.92652506e-02
-6.04500353e-01 2.35689938e-01 -4.11930232e-04 3.21245134e-01
6.20412886e-01 -1.39405859e+00 -7.11812556e-01 1.10468954e-01
2.50297993e-01 -2.78104782e-01 8.94301012e-02 6.84792519e-01
-4.45245743e-01 8.22194874e-01 -1.18474722e-01 -3.56925339e-01
-1.51927352e+00 8.64102483e-01 2.57924527e-01 -5.77187121e-01
-3.06451946e-01 5.75685859e-01 5.86258054e-01 -1.65067807e-01
2.54902244e-01 -2.26188853e-01 -5.03831685e-01 2.89733261e-01
8.56737733e-01 -9.04298052e-02 3.65718961e-01 -2.81003177e-01
-7.64454901e-01 6.95625365e-01 -7.37974703e-01 5.15534222e-01
1.77682102e+00 4.35304821e-01 -2.28403613e-01 -1.44346863e-01
1.59236562e+00 -1.28735930e-01 -6.03299379e-01 -2.77518988e-01
6.21695891e-02 -3.30341399e-01 9.30165052e-02 -1.33232343e+00
-1.00501764e+00 8.45462441e-01 6.96076393e-01 -3.73642981e-01
9.61772442e-01 -3.24114673e-02 8.68446231e-01 5.96088409e-01
8.70324075e-02 -7.92940676e-01 -3.29803169e-01 4.16417778e-01
8.28624010e-01 -1.33262539e+00 4.52451527e-01 -6.78736746e-01
-6.40421867e-01 1.15033865e+00 4.23808455e-01 3.70770305e-01
7.68452942e-01 1.52482286e-01 -4.30942237e-01 -5.38354397e-01
-7.47725070e-01 -2.79420167e-01 7.03859448e-01 3.49371523e-01
7.81886518e-01 -1.67412922e-01 -8.24896216e-01 9.17851865e-01
4.73006070e-01 2.76779681e-01 1.44629151e-01 7.91345656e-01
-2.74244636e-01 -1.63823366e+00 1.17195368e-01 5.03734767e-01
-7.55639613e-01 -7.05748498e-01 -8.36610019e-01 7.11043179e-01
2.25506768e-01 8.82398367e-01 -6.83504701e-01 -4.47108328e-01
4.09804732e-01 1.47743002e-01 3.27329010e-01 -7.92063475e-01
-5.05423009e-01 3.58548641e-01 2.53254324e-01 -3.37787718e-01
-2.80295461e-01 -9.05864872e-03 -1.62658012e+00 -2.86011845e-01
-5.29832244e-01 2.04606950e-01 8.32646072e-01 1.16617382e+00
8.79108727e-01 5.89677572e-01 5.21856129e-01 -2.15381816e-01
-2.12828070e-01 -1.04336154e+00 -2.65199482e-01 2.25567684e-01
4.37840730e-01 -6.86319411e-01 2.45616175e-02 5.84701486e-02]
|
[8.301756858825684, 8.61764144897461]
|
65f80e28-1141-4cbe-ad95-ffa7023bdbb4
|
good-exploring-geometric-cues-for-detecting
|
2212.11720
| null |
https://arxiv.org/abs/2212.11720v3
|
https://arxiv.org/pdf/2212.11720v3.pdf
|
GOOD: Exploring Geometric Cues for Detecting Objects in an Open World
|
We address the task of open-world class-agnostic object detection, i.e., detecting every object in an image by learning from a limited number of base object classes. State-of-the-art RGB-based models suffer from overfitting the training classes and often fail at detecting novel-looking objects. This is because RGB-based models primarily rely on appearance similarity to detect novel objects and are also prone to overfitting short-cut cues such as textures and discriminative parts. To address these shortcomings of RGB-based object detectors, we propose incorporating geometric cues such as depth and normals, predicted by general-purpose monocular estimators. Specifically, we use the geometric cues to train an object proposal network for pseudo-labeling unannotated novel objects in the training set. Our resulting Geometry-guided Open-world Object Detector (GOOD) significantly improves detection recall for novel object categories and already performs well with only a few training classes. Using a single "person" class for training on the COCO dataset, GOOD surpasses SOTA methods by 5.0% AR@100, a relative improvement of 24%.
|
['Dan Zhang', 'Andreas Geiger', 'Haiwen Huang']
|
2022-12-22
| null | null | null | null |
['class-agnostic-object-detection', 'open-world-object-detection']
|
['computer-vision', 'computer-vision']
|
[ 1.32795021e-01 1.63616553e-01 6.16609640e-02 -4.58677351e-01
-9.09461617e-01 -6.91687644e-01 4.89794850e-01 1.93907797e-01
-6.09124660e-01 3.09012681e-01 -2.45573968e-01 2.14469969e-01
3.50087017e-01 -6.87942922e-01 -1.09523523e+00 -5.27419567e-01
1.82013556e-01 7.10528016e-01 9.56332862e-01 2.15967391e-02
1.57191992e-01 7.35381246e-01 -1.96573102e+00 3.38156104e-01
4.28094864e-01 1.09337699e+00 1.42938778e-01 7.66647160e-01
-1.38309794e-02 2.50964284e-01 -4.24249768e-01 -4.82072771e-01
4.83507752e-01 -5.88306487e-02 -5.23750842e-01 2.84758121e-01
1.14906108e+00 -4.65431273e-01 -4.40642774e-01 8.69934201e-01
3.74350309e-01 1.28168046e-01 6.25859022e-01 -1.18676436e+00
-6.80115283e-01 1.15974188e-01 -6.96479321e-01 4.10485208e-01
4.86054659e-01 4.62817252e-01 7.73298800e-01 -1.34133375e+00
6.09811783e-01 1.29828775e+00 6.76816285e-01 8.18656504e-01
-1.36145353e+00 -6.83215141e-01 5.06375074e-01 2.63203681e-01
-1.62731373e+00 -4.75842893e-01 7.20071375e-01 -3.32282692e-01
9.59411681e-01 3.97952199e-01 6.02711141e-01 1.04940796e+00
-2.36383118e-02 1.04023552e+00 1.12757468e+00 -3.16340715e-01
2.11663887e-01 5.27651966e-01 8.58104846e-04 7.51530051e-01
5.08144736e-01 3.98911685e-01 -4.65427756e-01 4.05412428e-02
5.96082211e-01 1.29310176e-01 2.15967894e-02 -8.95604491e-01
-1.09580564e+00 6.10156894e-01 1.22011483e+00 -4.28296067e-02
-5.39779253e-02 2.55046576e-01 2.11091004e-02 -1.65575996e-01
4.08209920e-01 4.22665358e-01 -5.72272897e-01 2.39837050e-01
-7.47818828e-01 2.90002078e-01 2.42421091e-01 1.19998765e+00
9.24902558e-01 -3.03465158e-01 -5.09697795e-02 5.83361328e-01
3.53856474e-01 6.69417560e-01 2.39221349e-01 -7.84062445e-01
2.39602461e-01 8.21229160e-01 1.87745512e-01 -7.28467286e-01
-6.35402858e-01 -7.04499483e-01 -1.59558013e-01 3.07671547e-01
6.65216208e-01 3.46234143e-01 -1.43818712e+00 1.43815160e+00
6.51894510e-01 -1.81686521e-01 -2.68664658e-01 1.04802120e+00
9.66778815e-01 3.86812150e-01 9.35595781e-02 2.39852026e-01
1.26071692e+00 -1.05581748e+00 8.96199197e-02 -7.12958157e-01
3.91853452e-01 -7.01898634e-01 1.17552865e+00 3.63931984e-01
-9.25204694e-01 -7.49010623e-01 -8.26238453e-01 3.97280417e-02
-6.77194595e-01 2.04781368e-01 7.81718671e-01 9.51537848e-01
-7.00200319e-01 3.01156044e-01 -7.98846185e-01 -7.09869027e-01
7.46214211e-01 1.81249201e-01 -3.92457753e-01 -4.59516943e-01
-2.98682541e-01 8.88817549e-01 6.67665184e-01 -1.73564136e-01
-9.99707222e-01 -6.93692684e-01 -7.76167512e-01 -1.52891234e-01
5.66977084e-01 -6.91284597e-01 1.15400565e+00 -9.28776741e-01
-9.82551098e-01 1.13922906e+00 -3.86527851e-02 -2.85265714e-01
7.15152860e-01 -2.45278791e-01 -3.42756778e-01 2.59056121e-01
3.24690610e-01 1.07389677e+00 1.03696942e+00 -1.53788102e+00
-9.23389256e-01 -6.68764114e-01 1.14206873e-01 6.50105253e-02
-1.97387576e-01 -9.07037780e-02 -5.76002896e-01 -4.74824518e-01
6.28505945e-01 -1.00506449e+00 -2.42008463e-01 6.14982307e-01
-5.53211153e-01 -1.80208981e-01 7.59231031e-01 -1.58488944e-01
3.88291597e-01 -1.94924843e+00 -3.25816423e-01 8.25297087e-02
9.03622881e-02 2.07785182e-02 -1.39894232e-01 -1.15650192e-01
1.41793177e-01 -1.82556644e-01 7.73641020e-02 -3.09529364e-01
-1.06124707e-01 8.38275403e-02 -2.60328263e-01 5.67867696e-01
4.88911688e-01 9.48490679e-01 -1.07249773e+00 -4.41583902e-01
5.20713449e-01 1.74052075e-01 -4.85226661e-01 9.71100181e-02
-4.25582349e-01 3.83428693e-01 -2.35079870e-01 1.17090404e+00
8.46745014e-01 -2.82662034e-01 -3.43150944e-01 -2.17050731e-01
1.43335074e-01 -3.76936197e-02 -1.34904373e+00 1.72939432e+00
-2.87823439e-01 6.23466015e-01 -4.17764813e-01 -4.60536897e-01
8.05928349e-01 -2.56916940e-01 1.85594589e-01 -5.67699850e-01
8.39006826e-02 3.74646038e-01 -3.90825681e-02 -2.61558533e-01
3.36003631e-01 -8.47638492e-03 9.38740522e-02 2.76412010e-01
2.38946557e-01 -1.27380282e-01 -2.20120642e-02 2.29555666e-01
1.16540658e+00 2.54089028e-01 1.20459057e-01 1.93722427e-01
1.20030031e-01 1.86058909e-01 4.00961190e-01 1.20110524e+00
-3.83495182e-01 1.08182061e+00 -1.68268830e-01 -7.52364516e-01
-1.07119000e+00 -1.46079302e+00 -4.54141915e-01 1.22883451e+00
4.59993303e-01 -7.38868043e-02 -3.97022188e-01 -9.76855636e-01
3.37138265e-01 5.86638033e-01 -7.95784652e-01 -2.23384231e-01
-3.56349736e-01 -6.79443777e-01 3.02017301e-01 8.69898021e-01
2.49267876e-01 -9.35124636e-01 -1.04701197e+00 8.01296160e-02
1.17870018e-01 -1.23926520e+00 -1.00534432e-01 2.82351106e-01
-7.92992651e-01 -1.36600137e+00 -8.28169525e-01 -5.56494951e-01
9.07270670e-01 7.05247104e-01 1.08147168e+00 -1.28518865e-01
-9.60533619e-01 6.80519938e-01 -2.61341959e-01 -5.10842800e-01
-1.60578713e-01 -8.95198286e-02 2.30756447e-01 1.07382022e-01
5.01279831e-01 -5.80265187e-02 -8.46007109e-01 4.40508544e-01
-4.79772270e-01 -1.93784982e-02 8.78250062e-01 5.72236180e-01
5.73108733e-01 -3.67701322e-01 2.50237077e-01 -5.87596059e-01
-3.62761915e-01 -1.01828158e-01 -4.96688902e-01 2.90795594e-01
-4.38116521e-01 -1.92695975e-01 1.13747187e-01 -7.15601265e-01
-1.10283387e+00 6.26064658e-01 2.35144541e-01 -6.18133485e-01
-5.25040805e-01 -3.02386999e-01 -5.63893430e-02 -4.50243801e-01
1.20395470e+00 1.61498070e-01 -4.42882150e-01 -5.11082530e-01
7.03100681e-01 4.57891881e-01 6.97824061e-01 -4.83064592e-01
1.02717400e+00 9.71052706e-01 -2.62746364e-01 -7.05023050e-01
-1.13208175e+00 -1.02711523e+00 -8.90197635e-01 -2.96580583e-01
6.72157586e-01 -1.14987445e+00 -5.21619499e-01 2.75549114e-01
-1.19902551e+00 -6.81902170e-02 -6.24005854e-01 4.03667599e-01
-5.43666244e-01 1.08174540e-01 -2.11993933e-01 -1.06316149e+00
-5.72064985e-03 -6.87892437e-01 1.47322297e+00 4.04619694e-01
1.15174793e-01 -2.50733405e-01 -2.54764318e-01 4.29158151e-01
8.97069275e-02 3.90289634e-01 3.58054936e-01 -4.78122950e-01
-1.02278352e+00 -6.68744624e-01 -5.75501978e-01 1.20464951e-01
-1.38301611e-01 -2.24865541e-01 -1.42412555e+00 -3.87672395e-01
-3.43218923e-01 -5.74061692e-01 1.20629966e+00 7.95310512e-02
1.14419198e+00 1.34338796e-01 -8.33053768e-01 5.98048329e-01
1.43928719e+00 -1.04575783e-01 3.26242089e-01 3.93891186e-01
8.28153014e-01 4.58037853e-01 7.73805678e-01 4.98579741e-01
2.29479477e-01 7.15163052e-01 5.16265631e-01 -1.16646662e-01
-3.94628972e-01 -3.39778036e-01 4.45596203e-02 -2.91597039e-01
7.90985227e-02 5.53358048e-02 -1.02855599e+00 6.90644622e-01
-1.70748627e+00 -5.65854847e-01 -2.07609504e-01 2.07380104e+00
6.38908923e-01 5.89075565e-01 3.60992372e-01 -1.11261241e-01
5.63441277e-01 -4.15192336e-01 -9.34701264e-01 2.69343168e-01
-2.12869540e-01 2.01076254e-01 7.86006987e-01 2.08371524e-02
-1.30720365e+00 1.05602324e+00 5.83741713e+00 5.26023149e-01
-8.29190195e-01 2.72595346e-01 3.66068602e-01 -3.13834190e-01
1.31979123e-01 -2.32508266e-03 -1.13520849e+00 1.36077315e-01
4.32196379e-01 3.85853916e-01 -2.58091930e-03 1.24468267e+00
-2.41784453e-01 -4.00621742e-01 -1.37753415e+00 1.24413371e+00
3.84130180e-01 -1.06273699e+00 4.72953729e-03 -2.57121652e-01
8.97866249e-01 2.85464466e-01 7.30857849e-02 3.33555728e-01
8.67277384e-03 -6.87524319e-01 1.09266436e+00 5.05681813e-01
5.70731819e-01 -5.33681393e-01 5.08406639e-01 3.62815768e-01
-1.06886494e+00 -3.77248734e-01 -7.09954977e-01 3.23056877e-02
-1.75518185e-01 2.17654988e-01 -9.70874131e-01 7.30872080e-02
1.01901960e+00 4.59357917e-01 -1.22681510e+00 1.64864159e+00
-2.89900482e-01 2.07737342e-01 -7.22517133e-01 1.36815449e-02
1.45665482e-01 5.26743352e-01 5.66621482e-01 1.10943747e+00
-6.36238903e-02 1.16238050e-01 3.52166653e-01 1.06651354e+00
1.16204217e-01 -2.13824794e-01 -3.81013185e-01 3.92727137e-01
3.09524626e-01 1.21133673e+00 -1.11329687e+00 -2.22755134e-01
-4.18494970e-01 1.09955621e+00 4.28633690e-01 2.79659122e-01
-7.61713088e-01 -3.12153488e-01 3.77002388e-01 4.41941857e-01
5.64984679e-01 -1.52785525e-01 -2.50174433e-01 -1.23295271e+00
3.27664524e-01 -4.33158308e-01 3.85030717e-01 -9.92125988e-01
-1.31487656e+00 4.50261742e-01 -1.10600214e-03 -1.19462514e+00
4.84539121e-02 -1.01305664e+00 -2.84397811e-01 4.96799082e-01
-1.47618699e+00 -1.70819068e+00 -7.95061707e-01 4.25498307e-01
5.51898360e-01 -1.73550528e-02 6.31821215e-01 1.65073380e-01
-2.10533664e-01 7.28838503e-01 -1.57270599e-02 1.69630304e-01
6.99690819e-01 -1.22877955e+00 5.84494293e-01 9.21015084e-01
6.49620950e-01 4.12882745e-01 6.88054085e-01 -6.02669835e-01
-1.23653769e+00 -1.25187755e+00 4.50917333e-01 -1.14918852e+00
3.32369715e-01 -7.57509708e-01 -6.54831707e-01 6.90417469e-01
-6.79764628e-01 7.16173947e-01 2.12206155e-01 1.45200625e-01
-8.65437031e-01 -1.33022934e-01 -1.13592160e+00 4.89658386e-01
1.26360345e+00 -4.90121543e-01 -4.98070866e-01 7.07100391e-01
6.98202431e-01 -5.52753031e-01 -3.62993121e-01 3.86139125e-01
5.82367897e-01 -1.04176593e+00 1.37616909e+00 -8.22135031e-01
-1.41916215e-01 -4.83502179e-01 -2.34326079e-01 -7.97732770e-01
-2.64984518e-01 -1.42654136e-01 -2.83589423e-01 9.04744208e-01
2.40989730e-01 -3.72032255e-01 1.14479506e+00 6.37906432e-01
-8.24755877e-02 -5.29162824e-01 -9.67739999e-01 -1.15866530e+00
-2.65461385e-01 -7.13141918e-01 1.79898620e-01 5.33477187e-01
-4.78427052e-01 1.12761855e-01 4.51666117e-02 3.50954801e-01
9.95801270e-01 2.68810034e-01 1.18186569e+00 -1.21471262e+00
-4.66931045e-01 -2.88776726e-01 -8.97659302e-01 -1.01993930e+00
-3.35500866e-01 -6.64814353e-01 2.96605438e-01 -1.31158471e+00
5.00127673e-01 -9.12105560e-01 -2.74388164e-01 4.87311631e-01
-2.20818460e-01 9.14774299e-01 3.85093510e-01 1.89350575e-01
-1.03765345e+00 4.94137973e-01 9.50478554e-01 -3.27749193e-01
-1.52103171e-01 1.49393275e-01 -5.16112149e-01 8.90205681e-01
4.11875129e-01 -6.29730999e-01 9.41126496e-02 -2.94982791e-01
7.93414861e-02 -4.47768539e-01 1.09617543e+00 -1.50581610e+00
1.88840970e-01 3.65469642e-02 9.62301075e-01 -9.60141778e-01
6.58411860e-01 -8.20045829e-01 -1.20947257e-01 5.58006465e-01
-7.03880116e-02 -5.23109019e-01 3.15635413e-01 1.04066968e+00
4.18642193e-01 -7.93887749e-02 1.00003505e+00 -3.64620119e-01
-1.19231963e+00 2.70527542e-01 8.11464414e-02 6.83763698e-02
1.25344646e+00 -6.90929532e-01 -3.18668902e-01 7.01261908e-02
-8.97685528e-01 -6.61047846e-02 5.18619180e-01 7.80683219e-01
8.73340368e-01 -1.06017041e+00 -5.12991905e-01 3.79835129e-01
8.00506949e-01 2.07542285e-01 1.00288451e-01 4.38138783e-01
-4.14171189e-01 3.49542260e-01 -1.16030000e-01 -1.05965018e+00
-1.07990801e+00 8.06618750e-01 4.53711450e-01 4.64307725e-01
-7.20593035e-01 1.38133657e+00 4.41186160e-01 -4.49435145e-01
4.61638093e-01 -4.08468932e-01 4.52591538e-01 -2.57092535e-01
4.21549946e-01 3.47172409e-01 1.04646556e-01 -5.96061528e-01
-7.17175782e-01 6.71335876e-01 -2.50893563e-01 1.43179774e-01
1.00203419e+00 -1.62326723e-01 3.81179750e-01 5.23631692e-01
1.09722733e+00 -3.35770398e-01 -1.44896197e+00 -6.39828682e-01
1.19521969e-03 -9.01583195e-01 -7.09293857e-02 -9.87145364e-01
-6.34928763e-01 8.17689419e-01 1.23080981e+00 -2.97594994e-01
6.11381292e-01 6.25211239e-01 5.38803279e-01 6.50344133e-01
7.96276510e-01 -1.08609486e+00 6.70629323e-01 2.21518248e-01
6.65284455e-01 -1.79276109e+00 1.61900938e-01 -5.12349725e-01
-2.04379454e-01 1.09233022e+00 1.01417768e+00 -2.61077523e-01
4.65215236e-01 -2.46026009e-01 3.21496539e-02 -2.83118963e-01
-2.54876524e-01 -6.30565405e-01 6.15835726e-01 9.21681941e-01
-1.53629884e-01 1.52556464e-01 5.06705463e-01 3.40743423e-01
-1.08127534e-01 -3.87491792e-01 2.93124765e-01 9.67555940e-01
-8.00437570e-01 -4.99161571e-01 -6.17347598e-01 5.11740208e-01
8.30161348e-02 1.34798750e-01 -4.75367099e-01 8.92562032e-01
4.40307379e-01 7.17486799e-01 3.05211544e-01 -2.05823302e-01
5.02380550e-01 -1.47795841e-01 8.78297210e-01 -9.82316136e-01
-3.05149227e-01 -6.81798756e-02 -3.68543804e-01 -6.95675075e-01
-3.74065310e-01 -8.25541556e-01 -1.00150776e+00 1.63510919e-01
-6.99602544e-01 -5.21577179e-01 6.37337029e-01 8.45567286e-01
1.33982241e-01 3.63592774e-01 3.45751405e-01 -1.13559377e+00
-5.64224303e-01 -7.27128029e-01 -4.21110809e-01 4.37059253e-01
3.47776234e-01 -9.03782964e-01 -1.42923981e-01 -1.05467126e-01]
|
[9.432851791381836, 1.3351812362670898]
|
75218828-c85a-444d-a5e8-194ea9386095
|
synthetic-yet-natural-properties-of-wordnet
| null | null |
https://aclanthology.org/2019.gwc-1.18
|
https://aclanthology.org/2019.gwc-1.18.pdf
|
Synthetic, yet natural: Properties of WordNet random walk corpora and the impact of rare words on embedding performance
|
Creating word embeddings that reflect semantic relationships encoded in lexical knowledge resources is an open challenge. One approach is to use a random walk over a knowledge graph to generate a pseudo-corpus and use this corpus to train embeddings. However, the effect of the shape of the knowledge graph on the generated pseudo-corpora, and on the resulting word embeddings, has not been studied. To explore this, we use English WordNet, constrained to the taxonomic (tree-like) portion of the graph, as a case study. We investigate the properties of the generated pseudo-corpora, and their impact on the resulting embeddings. We find that the distributions in the psuedo-corpora exhibit properties found in natural corpora, such as Zipf’s and Heaps’ law, and also observe that the proportion of rare words in a pseudo-corpus affects the performance of its embeddings on word similarity.
|
['John Kelleher', 'Abhijit Mahalunkar', 'Alfredo Maldonado', 'Filip Klubička']
| null | null | null | null |
gwc-2019-7
|
['word-similarity']
|
['natural-language-processing']
|
[-1.67459637e-01 1.65201247e-01 -2.40975201e-01 -1.51106909e-01
-5.66205122e-02 -8.53349984e-01 8.20611775e-01 8.24678838e-01
-8.69964361e-01 4.51863497e-01 7.88008630e-01 -3.56597781e-01
-2.70291328e-01 -1.27945924e+00 -5.10412872e-01 -4.81762737e-01
-4.77497913e-02 3.62497658e-01 3.99593949e-01 -3.42901617e-01
3.80156964e-01 1.85336173e-01 -1.40493929e+00 -1.31248683e-01
8.29504013e-01 4.51469451e-01 2.12739617e-01 4.86602604e-01
-5.10960281e-01 1.73159093e-01 -6.79014742e-01 -6.56848252e-01
2.07233727e-01 -3.51346463e-01 -9.51246083e-01 -1.37244567e-01
1.22463547e-01 2.80977398e-01 -4.25038457e-01 1.15285718e+00
3.77128333e-01 2.93053120e-01 1.09701431e+00 -8.92579556e-01
-8.29786777e-01 8.45488489e-01 -2.10251063e-01 4.50161636e-01
2.38413185e-01 -2.92562526e-02 1.49449003e+00 -5.64402699e-01
9.91811872e-01 1.19184017e+00 6.76435530e-01 1.37587279e-01
-1.18693650e+00 -3.51122320e-01 -2.81962156e-01 1.90536290e-01
-1.60573280e+00 3.88644747e-02 4.15639609e-01 -4.52316016e-01
8.92414808e-01 -2.88044930e-01 7.69079924e-01 9.85638201e-01
4.80362587e-02 9.36699137e-02 7.70610690e-01 -8.18097889e-01
4.84844595e-01 2.43001014e-01 2.88020343e-01 5.05768657e-01
8.57053936e-01 -1.84823647e-01 -3.19370955e-01 -4.41438824e-01
5.16393363e-01 -4.17765915e-01 -3.09890866e-01 -5.46013415e-01
-8.65792155e-01 1.24026513e+00 4.78996158e-01 6.61274433e-01
-4.04523790e-01 2.48048425e-01 5.31338215e-01 1.78891361e-01
5.63057959e-01 8.53135884e-01 -4.34030116e-01 -2.06595987e-01
-4.49428439e-01 4.70407546e-01 9.61177707e-01 5.30609727e-01
9.75436270e-01 -3.13682377e-01 4.50981446e-02 1.05403852e+00
1.67592034e-01 -6.99931756e-02 9.45750058e-01 -3.90394062e-01
2.34347761e-01 6.25445068e-01 9.07718390e-02 -1.10070097e+00
-3.10780108e-01 -2.37851158e-01 1.15414202e-01 -2.12810576e-01
7.70835638e-01 -2.76485622e-01 -7.50365973e-01 1.90176308e+00
3.16852301e-01 5.56105264e-02 1.89072624e-01 7.48684168e-01
5.57543397e-01 4.31817204e-01 3.35663438e-01 1.65266469e-01
1.53683054e+00 -4.10133630e-01 -3.25151414e-01 -3.34862113e-01
9.24896955e-01 -6.78394258e-01 1.37643182e+00 -2.89046973e-01
-4.73987222e-01 -2.22557887e-01 -1.08856273e+00 3.35022509e-02
-8.96995127e-01 -4.41864014e-01 4.35345203e-01 8.42185795e-01
-6.94624245e-01 8.42999399e-01 -6.08130276e-01 -8.94951463e-01
3.42402428e-01 -1.23366654e-01 -3.76533240e-01 -1.69390664e-01
-1.47544813e+00 1.34200144e+00 8.75226140e-01 -7.06369102e-01
-2.03348294e-01 -6.83048427e-01 -9.20612812e-01 1.74908236e-01
7.70341903e-02 -5.86173773e-01 7.24005222e-01 -8.38013530e-01
-8.42060328e-01 9.15570796e-01 1.47116452e-01 -2.34764487e-01
3.73830879e-03 2.38284513e-01 -4.05548364e-01 -1.08122483e-01
6.00893050e-02 5.39558828e-01 4.59830642e-01 -1.04856443e+00
-3.55525553e-01 -3.81455034e-01 3.55710238e-01 2.01675311e-01
-6.52496934e-01 -2.05078259e-01 -1.69674560e-01 -7.60511518e-01
-3.75014991e-01 -8.57473493e-01 -2.17389658e-01 -2.80212283e-01
-1.15886472e-01 -4.92905349e-01 1.32813379e-01 -3.46297830e-01
1.42328906e+00 -2.30120182e+00 -5.99391013e-02 5.08975267e-01
-2.67103799e-02 1.76610410e-01 -6.54542208e-01 8.65535617e-01
-3.24245580e-02 3.95706356e-01 -1.44832492e-01 2.19625294e-01
2.13958949e-01 4.96487558e-01 -6.61957487e-02 5.30560017e-01
7.55203515e-02 7.20025659e-01 -1.16057694e+00 -2.31650576e-01
-1.13086015e-01 3.22273433e-01 -6.62173510e-01 -1.04091968e-02
-3.15312147e-01 -4.07803386e-01 -3.24615568e-01 -1.14479132e-01
3.15456510e-01 4.05166559e-02 4.81666476e-01 -1.82542861e-01
-4.81448062e-02 7.94462144e-01 -9.13468003e-01 1.35950994e+00
-7.58957088e-01 8.21906805e-01 -5.50138235e-01 -7.67318547e-01
9.81768668e-01 3.06307431e-02 5.51585816e-02 -5.86114645e-01
1.28132969e-01 2.97935098e-01 6.02930486e-01 -3.78045410e-01
9.31411028e-01 -4.39345270e-01 4.07560309e-03 7.05609024e-01
2.06186652e-01 -1.44710243e-01 5.77155352e-01 2.70685494e-01
1.28189290e+00 -1.20789647e-01 3.96528184e-01 -6.86242104e-01
-3.90715189e-02 2.16510952e-01 2.52418429e-01 4.70029712e-01
1.57986119e-01 4.62487370e-01 7.98218787e-01 -1.38124451e-01
-1.18985736e+00 -1.17735374e+00 -4.81063157e-01 9.07824993e-01
1.66354090e-01 -8.95027220e-01 -7.18988240e-01 -6.34053349e-01
2.64217734e-01 1.20966709e+00 -8.32441330e-01 -5.41013539e-01
-2.59428769e-01 -8.36534142e-01 5.12124360e-01 4.32916105e-01
-3.96123350e-01 -9.93958056e-01 -6.82703912e-01 2.30032295e-01
9.52631906e-02 -1.10301495e+00 -4.60007578e-01 2.27416605e-01
-5.52720785e-01 -1.21471632e+00 -3.32501501e-01 -7.06737161e-01
5.58736980e-01 -7.65737221e-02 1.38319981e+00 -7.42939264e-02
-2.99199313e-01 3.49382043e-01 -9.20423985e-01 -3.73589069e-01
-5.36601424e-01 2.21011028e-01 8.64328593e-02 -2.13756174e-01
6.37807786e-01 -6.51441514e-01 -3.23672682e-01 5.25475144e-02
-1.35271394e+00 -7.25090444e-01 2.06724226e-01 7.85794199e-01
2.57128209e-01 3.14260840e-01 5.46503365e-01 -9.80593979e-01
1.31839859e+00 -7.41639555e-01 -3.67550492e-01 1.40949130e-01
-4.80354816e-01 4.50725526e-01 4.46446836e-01 -7.38222301e-01
-4.27974671e-01 -4.22688872e-01 -5.33082783e-02 -1.54534299e-02
1.87338755e-01 8.00976813e-01 7.80372545e-02 1.61352783e-01
7.80509591e-01 -4.75661345e-02 -2.04960592e-02 -4.26891625e-01
8.41914833e-01 7.92616427e-01 9.36852396e-02 -6.19932473e-01
7.27511644e-01 8.62840265e-02 -3.70138317e-01 -1.18029809e+00
-6.51390314e-01 -2.94592649e-01 -4.94594276e-01 3.86281937e-01
9.92734253e-01 -6.18568122e-01 -2.93802414e-02 -3.49697202e-01
-1.13574088e+00 -2.63396263e-01 -5.90712368e-01 5.92870235e-01
-2.27031708e-01 1.94729328e-01 -2.14118704e-01 -3.73768449e-01
-6.94859028e-03 -9.30068612e-01 4.07355696e-01 8.77708271e-02
-8.32272649e-01 -1.48364115e+00 5.43287635e-01 -2.61918008e-01
5.70666909e-01 1.77123532e-01 1.73563242e+00 -1.14071321e+00
-4.07946073e-02 -1.97184652e-01 -2.22043306e-01 3.25248718e-01
4.30357009e-01 6.69836858e-03 -6.31392360e-01 -4.77282181e-02
-3.88544589e-01 -1.56672865e-01 7.43198156e-01 8.26460719e-02
6.09607756e-01 -2.79828846e-01 -2.05507040e-01 2.14811683e-01
1.64472187e+00 -1.70525461e-01 7.22090244e-01 4.76337463e-01
4.24310148e-01 6.99833632e-01 2.37009332e-01 3.19655806e-01
3.96448374e-01 5.65741122e-01 1.27165005e-01 4.08181608e-01
-1.08024083e-01 -5.67738414e-01 3.70532334e-01 1.04978085e+00
2.58279443e-01 -3.25365037e-01 -1.06036425e+00 9.86995041e-01
-1.21487677e+00 -7.68781960e-01 5.77767305e-02 2.17356110e+00
1.03861284e+00 2.16373146e-01 2.48211682e-01 9.06895008e-03
5.70506990e-01 4.66740608e-01 3.56623977e-02 -6.92477763e-01
1.18183695e-01 7.24148512e-01 6.50232434e-01 5.26558638e-01
-5.48623085e-01 1.02812052e+00 6.20949745e+00 8.58112693e-01
-9.82160330e-01 2.41017252e-01 -5.22046722e-02 1.32748425e-01
-7.32743144e-01 1.97853297e-01 -4.78435546e-01 5.60029030e-01
1.12612128e+00 -6.84039533e-01 3.11277360e-01 4.76568401e-01
-1.26871124e-01 -8.94092619e-02 -1.11294031e+00 4.43121880e-01
-1.35421939e-02 -1.12086320e+00 6.48996159e-02 2.15990588e-01
5.80177009e-01 2.21775606e-01 -2.95086473e-01 2.48562187e-01
5.97734690e-01 -1.04107451e+00 5.69468617e-01 -2.19064299e-02
7.97198713e-01 -8.14246297e-01 7.09585369e-01 3.77927683e-02
-1.05107188e+00 1.78039536e-01 -6.76880121e-01 -6.97583780e-02
7.27154985e-02 7.84167588e-01 -9.58195746e-01 3.37734610e-01
2.92123765e-01 3.57464343e-01 -8.59654307e-01 9.02358174e-01
-3.33009154e-01 7.80727625e-01 -3.15412730e-01 -5.08049846e-01
3.58197331e-01 -3.43237191e-01 3.86312425e-01 1.26086545e+00
2.98321515e-01 -2.42186427e-01 -9.04598013e-02 9.03122306e-01
-2.07692236e-01 3.87978464e-01 -7.62246728e-01 -7.55194187e-01
8.18156362e-01 1.02489293e+00 -7.73254454e-01 -1.39134273e-01
-5.89313149e-01 5.62071979e-01 4.79045570e-01 2.80050427e-01
-5.14926851e-01 -7.17917681e-01 1.11246431e+00 5.12062132e-01
5.48399270e-01 -2.79977024e-01 -1.04024328e-01 -9.12849188e-01
-4.65759784e-02 -6.82103932e-01 2.64097720e-01 -5.69980919e-01
-1.65163136e+00 5.27564585e-01 1.50126725e-01 -6.78519487e-01
-8.35443065e-02 -5.52285969e-01 -6.79156303e-01 8.89569163e-01
-1.09263241e+00 -3.97446454e-01 -7.27108773e-03 1.47762731e-01
6.45160377e-02 3.43647413e-02 8.35869610e-01 6.92899600e-02
-3.67558300e-01 4.66990709e-01 2.14692891e-01 2.63514012e-01
3.93494636e-01 -1.32366621e+00 7.25619793e-01 4.69076276e-01
6.29810214e-01 9.27336633e-01 9.15752292e-01 -6.12305462e-01
-1.09625232e+00 -1.04027224e+00 8.74822199e-01 -6.64461434e-01
1.18636882e+00 -3.72273952e-01 -9.45112765e-01 4.79336262e-01
3.12264830e-01 -1.28048865e-04 7.79811919e-01 3.53676021e-01
-7.99785137e-01 2.03824162e-01 -9.38473105e-01 6.57429755e-01
1.07649231e+00 -6.38761938e-01 -9.02300477e-01 1.92889988e-01
9.05159354e-01 1.15228869e-01 -9.46074069e-01 6.44342527e-02
4.69164550e-01 -6.28787935e-01 7.44454205e-01 -8.29740882e-01
5.56919038e-01 4.79393974e-02 -3.26744467e-01 -2.01603365e+00
-3.86052936e-01 -7.93750808e-02 3.07061106e-01 1.28803635e+00
6.60766661e-01 -7.65496016e-01 6.99416757e-01 2.44520873e-01
2.10076988e-01 -7.91958749e-01 -8.72524023e-01 -9.38413382e-01
6.34113669e-01 -1.46938577e-01 6.14549398e-01 9.71029460e-01
2.30555892e-01 5.30347943e-01 5.06818354e-01 -1.71769992e-01
1.82690054e-01 -2.12595612e-01 5.43861628e-01 -1.22705543e+00
-7.82435909e-02 -4.51715648e-01 -7.31709421e-01 -5.29194117e-01
3.13202322e-01 -1.25365078e+00 -2.57750303e-01 -1.51353025e+00
-1.05390787e-01 -7.11431682e-01 -1.42172307e-01 9.35884565e-02
-2.20132336e-01 2.10789754e-03 5.73418476e-02 -2.66704887e-01
-1.59324661e-01 4.44680363e-01 7.18091786e-01 1.92082837e-01
-4.31649350e-02 -6.46583557e-01 -8.00921321e-01 6.63314760e-01
8.26080859e-01 -7.79963315e-01 -5.45871019e-01 -4.96425629e-01
5.18877089e-01 -5.56451023e-01 -7.15850368e-02 -6.00148380e-01
-2.11277738e-01 -2.27651149e-01 -1.29083782e-01 2.42535278e-01
1.00940883e-01 -6.76361799e-01 -5.35701327e-02 3.23134094e-01
-4.18787271e-01 5.81862211e-01 9.82046872e-02 5.88275135e-01
-2.24866629e-01 -7.33298481e-01 7.06055522e-01 -1.09977357e-01
-4.62826669e-01 -4.88797948e-02 -3.74661833e-01 8.31126630e-01
8.13584745e-01 -1.18534729e-01 -1.86716437e-01 -1.06474556e-01
-1.18617713e-01 -1.14976391e-01 8.86311650e-01 6.01884305e-01
1.96593851e-01 -1.34892058e+00 -5.13212800e-01 -1.39522273e-02
5.03030241e-01 -4.31795955e-01 -3.79049659e-01 3.34030330e-01
-6.53718770e-01 6.60845414e-02 -8.11031759e-02 6.93171918e-02
-7.41043270e-01 3.03066254e-01 2.30255678e-01 -1.35404631e-01
-4.56808627e-01 6.53496146e-01 -1.81718677e-01 -5.71637213e-01
-1.98682994e-02 -2.61177540e-01 -1.69735774e-01 4.48680997e-01
1.76540628e-01 2.09200725e-01 5.39945364e-02 -5.58919430e-01
-4.75263625e-01 4.69002426e-01 1.55001730e-01 -2.31101170e-01
1.42119551e+00 2.54741848e-01 -1.40988499e-01 4.87913311e-01
1.48872769e+00 2.91887015e-01 -4.23229903e-01 -2.44141698e-01
3.99646282e-01 -5.96812844e-01 -7.96673223e-02 -3.18394005e-01
-7.49781251e-01 5.58624029e-01 3.34983170e-01 5.10167181e-01
4.08316284e-01 1.18672207e-01 8.07656467e-01 1.21061862e-01
3.03244948e-01 -1.04495513e+00 4.78255786e-02 6.80896521e-01
5.47527015e-01 -6.67930543e-01 3.59997526e-02 -3.02147627e-01
-4.96296018e-01 9.39908147e-01 3.22858930e-01 -3.85220468e-01
7.43733704e-01 1.06673144e-01 1.48237556e-01 -2.87674665e-01
-7.26536810e-01 -5.91601253e-01 2.05021575e-02 6.67819083e-01
6.79778457e-01 1.01079278e-01 -7.72693574e-01 2.99108773e-01
-7.01561749e-01 -2.60748923e-01 6.63589537e-01 6.75214112e-01
-3.35046500e-01 -1.36571884e+00 -2.86591183e-02 7.18225479e-01
-5.12270153e-01 -3.76297772e-01 -5.72128832e-01 7.87284613e-01
2.00908408e-01 7.04068661e-01 4.55997169e-01 -3.71462494e-01
2.42409974e-01 2.83545494e-01 6.52124941e-01 -1.07047796e+00
-4.56520289e-01 -7.10854590e-01 5.29195666e-01 -6.45265505e-02
-7.27059916e-02 -3.65804702e-01 -1.12654150e+00 -3.34641963e-01
-5.34095645e-01 4.43371385e-01 7.40310848e-01 9.79386926e-01
2.66626418e-01 2.63792455e-01 1.28325433e-01 -4.17239964e-01
-7.86424935e-01 -1.20627880e+00 -6.88287735e-01 1.00781631e+00
-2.59305030e-01 -8.49791169e-01 -5.31459570e-01 -3.89744788e-01]
|
[10.37275218963623, 8.920268058776855]
|
d39b53ab-24e0-446a-8f05-33deb37ef2b4
|
semantic-scene-completion-using-local-deep
|
2011.09141
| null |
https://arxiv.org/abs/2011.09141v3
|
https://arxiv.org/pdf/2011.09141v3.pdf
|
Semantic Scene Completion using Local Deep Implicit Functions on LiDAR Data
|
Semantic scene completion is the task of jointly estimating 3D geometry and semantics of objects and surfaces within a given extent. This is a particularly challenging task on real-world data that is sparse and occluded. We propose a scene segmentation network based on local Deep Implicit Functions as a novel learning-based method for scene completion. Unlike previous work on scene completion, our method produces a continuous scene representation that is not based on voxelization. We encode raw point clouds into a latent space locally and at multiple spatial resolutions. A global scene completion function is subsequently assembled from the localized function patches. We show that this continuous representation is suitable to encode geometric and semantic properties of extensive outdoor scenes without the need for spatial discretization (thus avoiding the trade-off between level of scene detail and the scene extent that can be covered). We train and evaluate our method on semantically annotated LiDAR scans from the Semantic KITTI dataset. Our experiments verify that our method generates a powerful representation that can be decoded into a dense 3D description of a given scene. The performance of our method surpasses the state of the art on the Semantic KITTI Scene Completion Benchmark in terms of geometric completion intersection-over-union (IoU).
|
['Dariu M. Gavrila', 'Markus Enzweiler', 'David Emmerichs', 'Christoph B. Rist']
|
2020-11-18
| null | null | null | null |
['3d-semantic-scene-completion']
|
['computer-vision']
|
[ 4.92064714e-01 1.05962642e-01 3.50638956e-01 -7.36910880e-01
-9.91362929e-01 -6.73791289e-01 5.69174170e-01 3.82639647e-01
-2.49589384e-01 3.62137914e-01 -9.97057036e-02 -3.57895158e-02
-3.54185253e-02 -1.19371092e+00 -1.17769444e+00 -2.85919249e-01
-1.19606871e-02 8.74422431e-01 3.86547059e-01 1.05109960e-01
1.02573976e-01 8.22971821e-01 -1.66834617e+00 1.59854159e-01
1.08817458e+00 1.12217021e+00 5.45940101e-01 5.61102271e-01
-5.12398481e-01 3.49276483e-01 -2.00371951e-01 5.36280796e-02
4.95242864e-01 9.84553844e-02 -8.42189670e-01 4.50737953e-01
8.98547947e-01 -3.94733727e-01 -8.92969444e-02 8.68611991e-01
-1.57795995e-01 2.64291197e-01 7.39014685e-01 -8.91576707e-01
-6.37231991e-02 -1.06903270e-01 -5.30441940e-01 -4.27390605e-01
1.54537365e-01 -3.69817466e-02 8.34861159e-01 -1.20725703e+00
6.49964094e-01 1.43727100e+00 8.43371332e-01 -1.13086127e-01
-1.59711683e+00 -3.87483209e-01 3.70838642e-01 -3.23226601e-01
-1.59275854e+00 -1.81950212e-01 7.68669426e-01 -6.60757840e-01
9.41508234e-01 2.34571561e-01 5.07801175e-01 2.17582598e-01
-5.34498282e-02 4.55683738e-01 9.17954087e-01 -2.27597103e-01
5.16318202e-01 -2.46986702e-01 7.95963407e-02 6.23244166e-01
1.42288059e-01 -2.36960307e-01 -2.56743163e-01 -1.53661832e-01
1.21614134e+00 1.66881591e-01 7.71709345e-03 -8.03400815e-01
-1.11217320e+00 9.45246935e-01 7.44925022e-01 -1.88684255e-01
-3.34448546e-01 5.75553834e-01 1.25261202e-01 -1.52635857e-01
9.18031931e-01 2.26665333e-01 -3.73183370e-01 3.72917414e-01
-1.16691375e+00 4.38407987e-01 6.70015872e-01 1.10800850e+00
1.47044516e+00 -1.80287997e-03 1.72065392e-01 7.27171600e-01
2.80107588e-01 6.45659685e-01 -4.81074154e-01 -1.37765229e+00
3.28314334e-01 6.49374187e-01 1.72872454e-01 -7.38833010e-01
-1.48837015e-01 -3.37663978e-01 -5.21096706e-01 4.30497438e-01
2.23587394e-01 3.67876679e-01 -1.31345010e+00 1.39770782e+00
7.21105337e-01 3.89138669e-01 -2.79276520e-01 7.84917951e-01
8.56424928e-01 7.35251307e-01 2.30654046e-01 2.87655920e-01
1.09122205e+00 -8.21444035e-01 -1.80465043e-01 -4.87502873e-01
4.23331618e-01 -6.12814665e-01 1.06098115e+00 2.32942328e-01
-8.98587823e-01 -6.29827559e-01 -9.00507689e-01 -5.54099083e-01
-1.79915562e-01 -1.15836628e-01 1.08614302e+00 3.67762387e-01
-1.07846677e+00 5.04001856e-01 -1.00927532e+00 -3.18972349e-01
8.23725641e-01 2.37403452e-01 -4.91822869e-01 -4.71343309e-01
-4.65650707e-01 4.74163771e-01 5.02790987e-01 -9.35782418e-02
-1.25765133e+00 -1.14545977e+00 -1.28355241e+00 3.97547297e-02
3.17456931e-01 -8.58968198e-01 1.16749716e+00 -6.13184154e-01
-9.72997010e-01 9.25446570e-01 -3.11031491e-01 -4.34382796e-01
3.88635129e-01 -2.02989146e-01 3.65193397e-01 1.93235502e-01
5.77284932e-01 1.04374266e+00 6.39845014e-01 -1.55536830e+00
-4.99477506e-01 -4.30682927e-01 2.11131200e-01 3.51415604e-01
3.44567299e-01 -5.35834491e-01 -5.87740541e-01 -3.92400265e-01
5.95398068e-01 -5.88650286e-01 -6.01993442e-01 4.15063977e-01
-4.32928979e-01 -1.38144925e-01 8.99813414e-01 -5.58463216e-01
4.11374807e-01 -2.10546899e+00 -8.49808529e-02 2.71045864e-01
1.70149371e-01 -3.52578223e-01 -1.97677508e-01 3.04525942e-01
5.36455065e-02 1.86243206e-01 -9.68581140e-01 -8.76062036e-01
-6.59286082e-02 4.09864366e-01 -5.93769252e-01 7.09918320e-01
2.73203731e-01 8.16217661e-01 -8.34094524e-01 -3.09848875e-01
6.54363275e-01 7.14922190e-01 -8.74330282e-01 3.51411313e-01
-8.61585736e-01 7.84429491e-01 -5.05735397e-01 6.19006991e-01
1.15183818e+00 -1.91952646e-01 -1.25625521e-01 -1.11531340e-01
-2.69133687e-01 4.56205249e-01 -1.17754066e+00 2.49366093e+00
-6.75743222e-01 4.50789571e-01 1.94653288e-01 -9.02437508e-01
9.78314936e-01 -1.24707177e-01 5.78897715e-01 -4.01171535e-01
-1.62654564e-01 2.00751662e-01 -8.99137974e-01 3.63144130e-02
4.90747124e-01 -2.07102001e-01 -1.40753895e-01 1.79368153e-01
-1.04206512e-02 -1.09755349e+00 -1.49784684e-01 2.64284253e-01
9.46354866e-01 5.41074038e-01 4.16855142e-02 -4.76727456e-01
2.18124762e-01 5.12266040e-01 4.75690097e-01 5.66613555e-01
5.86234868e-01 8.89652967e-01 1.87802300e-01 -4.12970990e-01
-1.23664367e+00 -1.47961199e+00 -3.74010772e-01 6.51774645e-01
4.33453828e-01 -3.75876993e-01 -7.85399735e-01 -3.00707072e-01
1.95590630e-01 7.94709265e-01 -3.59584898e-01 4.48076308e-01
-4.79195505e-01 -1.65297613e-01 1.44141108e-01 4.70183074e-01
5.68060875e-01 -7.86327064e-01 -7.29230285e-01 1.62411824e-01
-2.12956116e-01 -1.66300297e+00 -2.55093366e-01 8.97963345e-03
-1.17143118e+00 -1.05703318e+00 -2.54909694e-01 -7.38173783e-01
8.53682339e-01 5.39981067e-01 1.28665817e+00 -1.27339557e-01
-4.69617665e-01 6.74639046e-01 -5.51127121e-02 -2.51700908e-01
-1.58813953e-01 -1.18221201e-01 -3.03397298e-01 -2.26941202e-02
-3.15184742e-02 -9.02526796e-01 -6.38926506e-01 -1.60143189e-02
-1.06955791e+00 5.49789310e-01 1.56737670e-01 3.96633983e-01
1.27246904e+00 1.72250018e-01 5.07195340e-03 -1.01841617e+00
-2.36249417e-01 -4.85043019e-01 -9.53361869e-01 -6.30501881e-02
-4.93103813e-04 -3.05453558e-02 2.08190382e-01 1.37161449e-01
-1.23220372e+00 6.38989627e-01 -1.80808172e-01 -5.88870287e-01
-6.01411879e-01 3.12129945e-01 -2.41814256e-01 -1.42290294e-01
4.56163973e-01 3.53115164e-02 -2.72382259e-01 -8.15745711e-01
7.74078012e-01 1.27653226e-01 7.87780643e-01 -7.21004188e-01
8.50052118e-01 1.17514563e+00 2.05396950e-01 -1.10982335e+00
-9.31892574e-01 -6.49132609e-01 -1.01407337e+00 2.21730862e-02
1.00539792e+00 -1.28942049e+00 -3.40806693e-01 1.77597299e-01
-1.41219008e+00 -6.75379038e-01 -6.47302806e-01 2.37962395e-01
-7.55809188e-01 5.42980373e-01 -3.02877933e-01 -6.76068902e-01
-1.01847924e-01 -9.93272126e-01 1.74119306e+00 -2.26304591e-01
-4.80745435e-02 -8.47705603e-01 7.90301263e-02 3.06376994e-01
2.23745182e-02 8.12959433e-01 8.97680342e-01 2.31996372e-01
-1.32284641e+00 1.22919410e-01 -4.89869118e-01 2.69599497e-01
4.95523959e-02 -2.30272770e-01 -1.26109481e+00 -9.37745273e-02
1.44908816e-01 -2.54959047e-01 1.07248998e+00 5.52208364e-01
1.36622691e+00 8.78100097e-03 -2.72991717e-01 1.03667247e+00
1.97019029e+00 -2.91885763e-01 7.26284385e-01 5.31801442e-03
1.12027609e+00 8.98050070e-01 6.11465693e-01 3.27335417e-01
6.01332188e-01 5.42203307e-01 8.99985194e-01 -3.72934133e-01
-3.07047665e-01 -5.94244003e-01 -6.06998280e-02 2.73033857e-01
1.87960178e-01 -5.13623245e-02 -1.07088578e+00 6.49544477e-01
-1.77086401e+00 -3.32121521e-01 -5.55480480e-01 2.35634303e+00
4.17779297e-01 -1.00229448e-03 -2.78791636e-01 -5.84318936e-02
4.13254201e-01 8.02968889e-02 -4.96181577e-01 -2.47102916e-01
-7.10620955e-02 6.92049265e-01 7.05913842e-01 9.29071605e-01
-1.17222655e+00 1.35253131e+00 5.86770821e+00 8.54696870e-01
-7.95607686e-01 3.65916312e-01 6.41070485e-01 2.65719444e-01
-7.03377426e-01 3.19894910e-01 -5.56452394e-01 -2.79804710e-02
5.60015321e-01 3.68205518e-01 2.53885478e-01 1.04630327e+00
1.24255612e-01 -3.89422596e-01 -1.18321216e+00 1.05178761e+00
-2.44256616e-01 -1.48809230e+00 1.08121969e-01 1.00475535e-01
9.02890146e-01 4.46521789e-01 -2.96057791e-01 8.33622068e-02
4.86526102e-01 -1.17772806e+00 1.12139308e+00 4.22220886e-01
1.32551813e+00 -6.01001918e-01 5.07604554e-02 4.47935134e-01
-1.59046543e+00 2.91868895e-01 -4.75781232e-01 -1.94235057e-01
3.57756019e-01 9.62937653e-01 -8.83486152e-01 5.63722372e-01
6.34323359e-01 6.35804713e-01 -4.38708723e-01 1.05969942e+00
-2.34960884e-01 4.91850346e-01 -7.85032928e-01 7.05783010e-01
1.75491959e-01 -4.00308698e-01 5.44949472e-01 1.10090280e+00
2.43914261e-01 7.26120993e-02 5.68642080e-01 1.49504185e+00
-4.05419469e-02 2.08687186e-02 -8.21559370e-01 2.28365302e-01
3.13121885e-01 1.17497671e+00 -1.09062910e+00 -3.61957669e-01
-1.71948746e-01 1.01596296e+00 4.05875832e-01 4.31236446e-01
-6.19469881e-01 5.22348545e-02 6.61660552e-01 3.75538647e-01
3.64863575e-01 -8.57137144e-01 -8.02265763e-01 -9.63924468e-01
8.31817463e-02 -4.65469621e-02 6.24900591e-03 -7.68234491e-01
-1.08231890e+00 3.92758399e-01 2.77442217e-01 -1.05913723e+00
-2.70634275e-02 -3.28735143e-01 -4.21834260e-01 9.18597877e-01
-1.61972094e+00 -1.49480057e+00 -7.18610287e-01 5.46160102e-01
7.97499895e-01 5.97020864e-01 7.79065371e-01 1.39211506e-01
1.10055327e-01 -2.70672768e-01 -2.33454421e-01 -1.44859403e-01
1.62681006e-02 -1.24154603e+00 7.94150651e-01 7.84235001e-01
1.34915054e-01 2.35833779e-01 4.35905695e-01 -8.66852939e-01
-1.18125951e+00 -1.55522406e+00 5.77899575e-01 -5.33159494e-01
7.19327405e-02 -8.33195210e-01 -9.09925580e-01 7.50318229e-01
-3.37781340e-01 1.29005343e-01 2.24108607e-01 -1.15307674e-01
-4.91621822e-01 -3.74227986e-02 -1.18486190e+00 2.37415105e-01
1.32909369e+00 -6.06887221e-01 -2.09331572e-01 5.18832266e-01
1.15065777e+00 -6.03100836e-01 -6.75501943e-01 7.66969979e-01
9.92328525e-02 -9.81801569e-01 1.24618816e+00 -8.06251168e-02
3.44540924e-01 -5.29398739e-01 -5.57221353e-01 -8.56101513e-01
-1.22250021e-01 -2.84493625e-01 2.06425011e-01 9.75948036e-01
-7.33330026e-02 -4.24730927e-01 8.36450756e-01 5.72888851e-01
-5.77549100e-01 -5.21023989e-01 -9.13135648e-01 -6.05323315e-01
1.24494739e-01 -1.03061235e+00 7.52371430e-01 7.22581089e-01
-8.95052314e-01 1.90884233e-01 2.30562046e-01 4.68569726e-01
1.04434407e+00 5.14783382e-01 1.06167662e+00 -1.46692228e+00
-5.23936041e-02 -6.82777762e-02 -3.91945958e-01 -1.38953543e+00
1.90717146e-01 -1.05968416e+00 2.82245725e-01 -1.99507856e+00
4.43741642e-02 -8.98680687e-01 3.66499156e-01 4.15837348e-01
2.68456012e-01 5.03142416e-01 -6.05304614e-02 2.14464679e-01
-4.78834122e-01 7.37748563e-01 1.15729809e+00 -1.32005468e-01
-3.20174664e-01 -3.10885787e-01 -3.53555560e-01 8.64734113e-01
5.67930222e-01 -4.45822984e-01 -5.49957395e-01 -7.23539174e-01
7.00842813e-02 -2.69104639e-04 9.17136550e-01 -9.80842292e-01
2.46668439e-02 -2.15094358e-01 2.19880760e-01 -1.08556402e+00
8.31416488e-01 -8.45121801e-01 4.78037298e-01 1.78236049e-02
9.16951597e-02 -6.04333401e-01 3.89745414e-01 7.04797804e-01
-2.08043233e-02 8.42201486e-02 6.94996715e-01 -3.01947027e-01
-9.81883526e-01 6.95792079e-01 2.01448016e-02 -1.28455400e-01
8.77348483e-01 -3.82645905e-01 2.40475431e-01 -2.08835721e-01
-6.19767249e-01 2.21175775e-01 7.87124276e-01 1.65792972e-01
8.41637015e-01 -1.17600656e+00 -6.97593629e-01 3.11983824e-01
1.31307304e-01 1.13507664e+00 4.34834391e-01 2.94743448e-01
-1.10745394e+00 4.90722060e-01 1.44197404e-01 -1.22434998e+00
-8.85437489e-01 1.49410263e-01 3.54569376e-01 -8.73039812e-02
-1.02980053e+00 8.77546906e-01 1.15004897e+00 -8.20020437e-01
1.72676314e-02 -7.12693512e-01 2.60552406e-01 -4.33585286e-01
-5.12454920e-02 1.17485330e-01 1.10778175e-01 -7.59018719e-01
-2.05007628e-01 8.31450760e-01 4.15756822e-01 -1.78172261e-01
1.44766307e+00 -1.35825604e-01 -4.05862719e-01 3.71893376e-01
1.18656409e+00 -1.33256093e-01 -1.63237500e+00 -3.11465889e-01
-2.48817995e-01 -8.12883317e-01 1.84193388e-01 -3.43662620e-01
-8.13404322e-01 1.02555013e+00 4.10807729e-01 -9.63329449e-02
8.37455034e-01 2.72301108e-01 7.66376853e-01 1.60457581e-01
7.62828231e-01 -7.58839369e-01 -4.78857756e-02 5.34438312e-01
1.06635785e+00 -1.19294477e+00 1.49828956e-01 -1.20236838e+00
-1.49400249e-01 9.63431954e-01 4.00459945e-01 -5.18045187e-01
7.99650908e-01 1.42980129e-01 -4.47330594e-01 -6.20289266e-01
-1.68667093e-01 -4.57401067e-01 4.31287825e-01 6.60798788e-01
1.65046286e-02 3.29130769e-01 2.57992923e-01 1.57199308e-01
-3.79757077e-01 -2.96168447e-01 1.48362219e-01 8.63519311e-01
-5.95969439e-01 -9.30794597e-01 -4.85100299e-01 2.94087410e-01
6.62387982e-02 -9.34072360e-02 -1.08629860e-01 6.28652751e-01
3.98257732e-01 8.09220672e-01 4.42052484e-01 -3.01791281e-02
2.71095246e-01 -2.63312429e-01 5.30151606e-01 -1.34593141e+00
1.94469482e-01 2.81391501e-01 -7.09743500e-02 -8.98249328e-01
-3.06714773e-01 -5.20176768e-01 -1.56773746e+00 -3.37983780e-02
-5.64905591e-02 -1.38167098e-01 9.93700504e-01 8.45003784e-01
2.23427117e-01 5.06633103e-01 5.03907442e-01 -1.37767947e+00
1.75206870e-01 -6.38611197e-01 -5.87304235e-01 3.80284250e-01
4.01128173e-01 -7.26534605e-01 -8.43842998e-02 1.98352531e-01]
|
[8.549749374389648, -2.9744179248809814]
|
a290c4d0-1a2a-407e-b65c-5ef5ba578043
|
findings-of-the-constraint-2022-shared-task
| null | null |
https://aclanthology.org/2022.constraint-1.1
|
https://aclanthology.org/2022.constraint-1.1.pdf
|
Findings of the CONSTRAINT 2022 Shared Task on Detecting the Hero, the Villain, and the Victim in Memes
|
We present the findings of the shared task at the CONSTRAINT 2022 Workshop: Hero, Villain, and Victim: Dissecting harmful memes for Semantic role labeling of entities. The task aims to delve deeper into the domain of meme comprehension by deciphering the connotations behind the entities present in a meme. In more nuanced terms, the shared task focuses on determining the victimizing, glorifying, and vilifying intentions embedded in meme entities to explicate their connotations. To this end, we curate HVVMemes, a novel meme dataset of about 7000 memes spanning the domains of COVID-19 and US Politics, each containing entities and their associated roles: hero, villain, victim, or none. The shared task attracted 105 participants, but eventually only 6 submissions were made. Most of the successful submissions relied on fine-tuning pre-trained language and multimodal models along with ensembles. The best submission achieved an F1-score of 58.67.
|
['Tanmoy Chakraborty', 'Md. Shad Akhtar', 'Preslav Nakov', 'Himanshi Mathur', 'Atharva Kulkarni', 'Tharun Suresh', 'Shivam Sharma']
| null | null | null | null |
constraint-acl-2022-5
|
['semantic-role-labeling']
|
['natural-language-processing']
|
[-1.29487380e-01 4.79368985e-01 -1.39321819e-01 -2.21038297e-01
-4.49640214e-01 -1.14665246e+00 1.08768964e+00 6.30804420e-01
-4.26394641e-01 7.75461018e-01 1.06317031e+00 -2.49711141e-01
2.43482932e-01 -4.44023401e-01 -3.44512254e-01 -1.82535335e-01
3.15928221e-01 5.30799270e-01 -6.40890673e-02 -6.86541021e-01
6.15513444e-01 -2.87233829e-01 -1.17022061e+00 7.32492089e-01
8.32153380e-01 3.97498399e-01 -3.16564798e-01 5.04038513e-01
-4.00614858e-01 1.49830210e+00 -8.83985698e-01 -9.22949791e-01
-2.69961655e-01 -5.03916204e-01 -1.23640025e+00 -2.57656872e-01
5.91907918e-01 3.14781547e-01 -3.75329882e-01 7.57517338e-01
2.96248406e-01 2.85294414e-01 1.00805342e+00 -1.11028600e+00
-1.03637171e+00 1.12320578e+00 -3.45852673e-01 5.45611024e-01
6.79062009e-01 -1.06456056e-01 1.50454783e+00 -1.12986577e+00
1.21575427e+00 1.47832465e+00 5.78375101e-01 9.23092961e-01
-1.26437032e+00 -3.91710132e-01 -1.15890376e-01 1.14874139e-01
-1.24545968e+00 -6.42072678e-01 3.44760507e-01 -9.42135274e-01
9.44160521e-01 4.89321381e-01 1.31174818e-01 1.49128246e+00
-2.32040063e-01 5.57087064e-01 1.04129028e+00 -1.17642209e-01
-8.46256837e-02 6.07399881e-01 3.75401855e-01 6.53525770e-01
5.34849502e-02 -5.59039652e-01 -8.09452951e-01 -4.32895184e-01
-1.39100000e-01 -5.26675105e-01 -2.51725286e-01 3.10748190e-01
-1.26258469e+00 9.32372034e-01 4.58333164e-01 4.66415018e-01
-4.64570895e-02 -9.90050193e-03 7.39242852e-01 2.55103648e-01
7.71000385e-01 1.07315123e+00 -2.00298816e-01 -1.73870817e-01
-5.76863408e-01 4.01630461e-01 1.21201932e+00 6.55969977e-01
3.80702645e-01 -6.10921979e-01 -4.34917420e-01 9.63076949e-01
1.40895993e-01 3.27067554e-01 -8.50954500e-04 -1.01635396e+00
6.58918262e-01 8.34902585e-01 5.08634925e-01 -1.51131821e+00
-6.31379545e-01 -1.66237235e-01 -4.56272423e-01 -5.94821036e-01
4.52879578e-01 -3.24623913e-01 -4.71789479e-01 1.75610673e+00
2.58249730e-01 -2.14661330e-01 6.78731501e-03 9.73390937e-01
1.37016225e+00 9.00938690e-01 5.97796023e-01 2.30344057e-01
1.51538444e+00 -7.74163663e-01 -6.90232575e-01 -3.71650159e-01
8.61232460e-01 -8.19561779e-01 1.11463845e+00 -2.92125016e-01
-9.29269969e-01 4.23258962e-03 -7.74983287e-01 -6.30970180e-01
-6.63394034e-01 6.12701699e-02 3.90228927e-01 3.43765914e-01
-7.81144798e-01 3.75259101e-01 -2.91641116e-01 -5.40229082e-01
1.84576854e-01 -2.44577512e-01 -4.59351391e-01 3.39224756e-01
-1.56964135e+00 1.34290457e+00 2.96230316e-01 -1.17011860e-01
-6.94340587e-01 -8.65837753e-01 -7.71389067e-01 -5.66596426e-02
4.39464688e-01 -6.88660979e-01 8.66740465e-01 -8.16148043e-01
-9.14918661e-01 1.84986937e+00 -3.07117790e-01 -1.65714353e-01
4.69600350e-01 -1.69707060e-01 -5.14469862e-01 -1.21137202e-01
5.69152594e-01 4.77150351e-01 5.83678901e-01 -1.22726238e+00
-5.17795086e-01 -1.97793558e-01 3.68177474e-01 2.10448995e-01
-5.08065939e-01 6.79168403e-01 2.63275236e-01 -4.36370850e-01
-1.21019669e-01 -8.15122128e-01 3.65418285e-01 -4.70652580e-01
-8.56385589e-01 -6.94000125e-01 4.78992105e-01 -8.59919369e-01
1.47913122e+00 -2.26869154e+00 5.54166079e-01 -1.17842101e-01
8.56387317e-01 -1.47754356e-01 -3.14173438e-02 8.72737050e-01
7.15517327e-02 7.14961410e-01 -1.57386005e-01 -5.36418438e-01
3.95512015e-01 -6.96010515e-02 -6.08223498e-01 4.53084230e-01
3.73911224e-02 9.14817274e-01 -1.15214646e+00 -4.10539180e-01
-3.48332465e-01 1.62427053e-01 -3.19651872e-01 7.84822181e-02
-3.74019414e-01 3.46898317e-01 -2.95053542e-01 5.55710673e-01
3.47494066e-01 -4.30161893e-01 4.20616955e-01 -1.38738737e-01
-2.48471692e-01 8.94419849e-01 -5.31423151e-01 1.09201288e+00
-3.28647017e-01 1.12168670e+00 5.22964954e-01 -3.06884468e-01
6.90970004e-01 1.99031323e-01 1.60844088e-01 -4.73231554e-01
1.40703708e-01 3.24296534e-01 -1.95962340e-01 -6.54729843e-01
1.02859545e+00 -4.23248023e-01 -7.26870418e-01 7.68020630e-01
-1.41823843e-01 9.05143544e-02 3.27229172e-01 7.77167499e-01
1.08331776e+00 -2.19483256e-01 2.80003399e-01 -6.37057364e-01
5.95950842e-01 2.77068406e-01 4.15361881e-01 5.58032870e-01
-3.42054814e-01 3.62473428e-01 1.05618095e+00 -4.34207231e-01
-1.11932170e+00 -9.83085990e-01 -8.22747964e-03 1.71581185e+00
3.40428382e-01 -7.94681489e-01 -4.91836727e-01 -7.58011222e-01
1.50244474e-01 1.05872858e+00 -8.69150341e-01 -6.61833864e-03
-5.89725435e-01 -6.59451008e-01 9.62653041e-01 1.15274899e-01
3.76037717e-01 -9.47975218e-01 -3.07766438e-01 1.16258167e-01
-1.00921273e+00 -1.14413536e+00 -3.67848873e-01 -3.57552588e-01
-1.07277714e-01 -1.09591186e+00 -2.23466858e-01 -4.73090261e-01
3.80593777e-01 -5.82002327e-02 1.62285769e+00 3.72202486e-01
5.36080487e-02 1.95318058e-01 -4.13589627e-01 -3.22362810e-01
-6.31700337e-01 3.28518301e-01 -1.22848287e-01 -9.68566313e-02
4.85464036e-01 -3.26537281e-01 -1.72162235e-01 2.35465482e-01
-6.18244886e-01 -8.83955881e-02 -2.67042130e-01 6.64668918e-01
-2.53159434e-01 -5.54641366e-01 4.17479455e-01 -1.36357713e+00
9.16490555e-01 -1.35479331e+00 1.16641834e-01 3.43918651e-01
-2.21215233e-01 -2.25208864e-01 3.77143651e-01 2.22550966e-02
-1.09893477e+00 -7.44730473e-01 1.60617068e-01 3.04044276e-01
3.53323370e-01 4.76176888e-01 2.04228610e-01 5.23355603e-01
1.08916974e+00 -2.57680893e-01 -4.09779102e-01 -4.00200844e-01
6.49326026e-01 6.70797706e-01 5.65555871e-01 -9.22287405e-01
6.57510400e-01 4.08897072e-01 -3.52701098e-01 -8.42835069e-01
-1.35755813e+00 -6.84879839e-01 -1.60253882e-01 -1.28288418e-01
9.65089262e-01 -9.67387617e-01 -8.92381012e-01 2.93191224e-01
-1.48924041e+00 -1.84645116e-01 9.31363776e-02 -2.11507365e-01
-7.49031454e-02 9.19451416e-02 -8.97208810e-01 -5.24355769e-01
-2.31065601e-01 -3.14614654e-01 7.28855968e-01 1.24602333e-01
-9.10646260e-01 -1.19226813e+00 1.64173231e-01 8.10056269e-01
5.15282214e-01 5.63674271e-01 1.27447796e+00 -9.93060589e-01
-6.65536746e-02 2.59191841e-01 -3.95589709e-01 -1.92535222e-01
-2.18434483e-01 1.05974928e-01 -7.48473525e-01 5.11995032e-02
-4.09882247e-01 -6.91496372e-01 9.35995638e-01 -3.95414084e-01
6.50612473e-01 -6.32621109e-01 -3.98828954e-01 6.32116050e-02
1.08638799e+00 -3.77901345e-01 5.40459871e-01 5.11232376e-01
7.78544426e-01 1.09258246e+00 2.74263501e-01 5.78778803e-01
7.99565911e-01 3.73330802e-01 4.26527262e-01 3.95672470e-01
-5.13083413e-02 -5.50592422e-01 4.88422513e-01 3.19845587e-01
8.32540840e-02 -6.92271411e-01 -1.23130178e+00 8.77237082e-01
-1.82303667e+00 -1.25266087e+00 -5.55471063e-01 1.44740486e+00
8.20239246e-01 -1.78014264e-01 7.69163072e-02 -4.78766441e-01
9.83437359e-01 7.68747509e-01 -1.56753212e-01 -5.95851779e-01
-4.00615007e-01 -2.62955785e-01 6.13493472e-03 7.40845144e-01
-1.08644927e+00 1.26865196e+00 6.17064142e+00 6.19932413e-01
-8.10970485e-01 5.09209871e-01 4.17181820e-01 -1.16458848e-01
-7.65333772e-01 6.31114021e-02 -7.44159281e-01 5.81626356e-01
8.77826512e-01 -3.35323006e-01 4.25417721e-01 6.31203353e-01
-6.24257326e-02 -1.91733003e-01 -1.05586588e+00 5.44738472e-01
4.04983580e-01 -1.53899217e+00 -2.50032663e-01 -2.14489236e-01
7.48409748e-01 1.77114546e-01 7.83132091e-02 5.14527738e-01
2.46149153e-01 -1.29200923e+00 1.13098156e+00 3.53857279e-01
5.32616735e-01 -2.26537123e-01 6.28972650e-01 3.87930810e-01
-4.93709028e-01 -1.46085367e-01 -2.31372833e-01 -4.03868020e-01
2.84860075e-01 3.26118767e-01 -7.98744977e-01 1.07583240e-01
3.62712026e-01 8.32054794e-01 -5.35527647e-01 2.80643940e-01
-5.91834128e-01 7.11402833e-01 -1.88548282e-01 -4.82208908e-01
2.75550067e-01 1.42248094e-01 1.07846951e+00 1.62890875e+00
-2.08931774e-01 5.10430574e-01 -3.25178914e-02 1.39440835e+00
-7.31861353e-01 3.40553485e-02 -3.38756949e-01 -7.77121544e-01
8.14756691e-01 1.53475165e+00 -6.13565326e-01 -2.06560984e-01
5.95286861e-02 9.77226794e-01 5.43431401e-01 1.88556522e-01
-8.83496821e-01 -2.58819878e-01 6.76309109e-01 2.96288222e-01
-2.64283895e-01 -1.89124495e-01 -4.89536971e-01 -1.28306818e+00
-1.60965040e-01 -6.48668468e-01 6.16996706e-01 -8.42257082e-01
-1.79248250e+00 4.76329178e-01 -1.58617258e-01 -3.88780653e-01
-1.24439083e-01 -5.44853628e-01 -7.79025793e-01 9.29845750e-01
-9.42136943e-01 -1.28323281e+00 -3.15013349e-01 1.26972854e-01
1.57554075e-01 -8.60862248e-03 7.57962883e-01 2.85609365e-01
-6.40666544e-01 4.15476710e-01 -1.76093861e-01 3.54006857e-01
9.42384064e-01 -1.21162617e+00 3.82750094e-01 5.39346814e-01
-4.84338664e-02 9.57308292e-01 1.12788498e+00 -7.20201015e-01
-7.49769330e-01 -7.04694927e-01 1.81797373e+00 -1.43515968e+00
1.14953566e+00 -5.29836774e-01 -7.78027952e-01 8.60110760e-01
5.70370555e-01 -4.14823920e-01 8.11990499e-01 7.17061996e-01
-8.73549223e-01 7.51912296e-01 -1.16959226e+00 5.32427609e-01
1.23229849e+00 -1.09116042e+00 -1.06651318e+00 5.19911587e-01
6.98339581e-01 -5.43795645e-01 -5.84796667e-01 -2.01071084e-01
2.03837410e-01 -8.30689490e-01 6.01109445e-01 -1.24672651e+00
1.19319975e+00 -1.64816678e-01 -2.61478454e-01 -1.25052536e+00
-2.21810460e-01 -6.25741065e-01 5.35208248e-02 1.53050303e+00
9.11636472e-01 -4.18825448e-01 3.17767203e-01 9.73003447e-01
-1.97689131e-01 -3.14499438e-01 -7.75808334e-01 -2.75239170e-01
2.13111371e-01 -5.26878722e-02 2.34148428e-01 1.56264246e+00
5.02443373e-01 9.11093891e-01 -2.91014999e-01 -1.45652458e-01
3.40780705e-01 7.95956776e-02 4.59498584e-01 -1.12842476e+00
4.34257053e-02 -5.33730268e-01 -7.82733187e-02 -4.99595284e-01
6.99845076e-01 -1.32103217e+00 -2.29121402e-01 -1.64626479e+00
7.28123605e-01 -4.17743295e-01 1.46432161e-01 5.67064524e-01
-2.94297457e-01 3.75166088e-01 3.66632879e-01 3.25622261e-01
-9.38978076e-01 2.37180397e-01 8.14480007e-01 -3.34449738e-01
-1.64905906e-01 -5.17753661e-01 -1.22498631e+00 6.77753150e-01
5.04259825e-01 -6.05634212e-01 3.95374484e-02 -5.80766141e-01
1.13471043e+00 -3.85902524e-01 7.74872601e-01 -9.97222811e-02
1.46675035e-01 -2.94927001e-01 -1.03546634e-01 -1.39784098e-01
3.69119197e-01 -8.03143978e-02 -3.93816642e-02 9.28722918e-02
-8.06523561e-01 3.01272683e-02 4.31992151e-02 3.55389118e-01
-1.61888674e-01 -1.81708112e-01 5.26085556e-01 -3.05380464e-01
-8.90654087e-01 -3.14184874e-01 -5.28053641e-01 9.30987120e-01
8.19144607e-01 1.64078012e-01 -1.03863502e+00 -5.13187885e-01
-9.92171168e-01 3.93508792e-01 4.73944783e-01 7.39762247e-01
2.80888200e-01 -1.06098378e+00 -1.20881462e+00 -6.80739164e-01
3.36255014e-01 -6.22833312e-01 3.79150689e-01 7.39858747e-01
-3.28360915e-01 2.15728208e-01 -4.71727783e-03 4.00028564e-02
-1.03113103e+00 2.11757019e-01 4.39172894e-01 -9.35432035e-03
-2.90316552e-01 1.19784236e+00 1.94732249e-02 -6.60741627e-01
-3.60068440e-01 4.31290179e-01 -4.80914950e-01 6.49423361e-01
5.76164424e-01 7.96590328e-01 -3.52572232e-01 -1.00346124e+00
-5.51332533e-01 1.59657493e-01 1.41908035e-01 -3.06422234e-01
1.15149856e+00 -4.28790569e-01 -9.72689986e-01 5.77354133e-01
1.24861002e+00 5.19870222e-01 -2.87793338e-01 -1.12158909e-01
4.75056738e-01 -3.77704293e-01 -2.12436438e-01 -1.30598259e+00
-2.96036124e-01 6.02951050e-01 -4.38920677e-01 4.89715368e-01
2.98774093e-01 4.79508758e-01 9.58808899e-01 1.59698099e-01
1.69553742e-01 -1.28205550e+00 7.32013360e-02 1.04172063e+00
1.27086830e+00 -1.14738131e+00 -2.40282074e-01 -5.70450723e-01
-9.45787251e-01 8.68499756e-01 8.88707697e-01 1.73663288e-01
1.41222805e-01 -2.95701623e-01 8.63193870e-02 -8.46199572e-01
-8.32318425e-01 2.29988117e-02 4.03646260e-01 7.96799287e-02
7.04934359e-01 3.40936601e-01 -8.01060200e-01 7.72484481e-01
-3.20348889e-01 -7.49716520e-01 9.76288438e-01 3.73622358e-01
-3.12100410e-01 -4.33536500e-01 -1.60433903e-01 7.94066861e-02
-5.75777709e-01 -3.04582566e-01 -1.28951621e+00 6.29407108e-01
2.73826778e-01 1.20402277e+00 1.07262991e-01 -4.78043437e-01
1.45125344e-01 2.20756471e-01 1.64336652e-01 -8.85823429e-01
-1.04654205e+00 -8.66051614e-01 1.02357423e+00 -4.40024883e-01
-3.30273479e-01 -5.12895405e-01 -1.45863616e+00 -9.71031249e-01
9.10995454e-02 3.03606927e-01 6.04097724e-01 9.93441164e-01
5.88002682e-01 4.24232408e-02 3.96876246e-01 -2.97680438e-01
-4.69959885e-01 -1.05573487e+00 -3.15398455e-01 7.07732439e-01
2.53457487e-01 -5.84469318e-01 -6.69784963e-01 -3.00706536e-01]
|
[8.561338424682617, 10.665306091308594]
|
059eb18d-780f-4591-a76c-8d3345e19582
|
droneattention-sparse-weighted-temporal
|
2212.03384
| null |
https://arxiv.org/abs/2212.03384v1
|
https://arxiv.org/pdf/2212.03384v1.pdf
|
DroneAttention: Sparse Weighted Temporal Attention for Drone-Camera Based Activity Recognition
|
Human activity recognition (HAR) using drone-mounted cameras has attracted considerable interest from the computer vision research community in recent years. A robust and efficient HAR system has a pivotal role in fields like video surveillance, crowd behavior analysis, sports analysis, and human-computer interaction. What makes it challenging are the complex poses, understanding different viewpoints, and the environmental scenarios where the action is taking place. To address such complexities, in this paper, we propose a novel Sparse Weighted Temporal Attention (SWTA) module to utilize sparsely sampled video frames for obtaining global weighted temporal attention. The proposed SWTA is comprised of two parts. First, temporal segment network that sparsely samples a given set of frames. Second, weighted temporal attention, which incorporates a fusion of attention maps derived from optical flow, with raw RGB images. This is followed by a basenet network, which comprises a convolutional neural network (CNN) module along with fully connected layers that provide us with activity recognition. The SWTA network can be used as a plug-in module to the existing deep CNN architectures, for optimizing them to learn temporal information by eliminating the need for a separate temporal stream. It has been evaluated on three publicly available benchmark datasets, namely Okutama, MOD20, and Drone-Action. The proposed model has received an accuracy of 72.76%, 92.56%, and 78.86% on the respective datasets thereby surpassing the previous state-of-the-art performances by a margin of 25.26%, 18.56%, and 2.94%, respectively.
|
['Peter Corcoran', 'Hari Mohan Pandey', 'Heena Rathore', 'Kamlesh Tiwari', 'Esha Pahwa', 'Achleshwar Luthra', 'Santosh Kumar Yadav']
|
2022-12-07
| null | null | null | null |
['human-activity-recognition', 'human-activity-recognition']
|
['computer-vision', 'time-series']
|
[ 2.13443846e-01 -4.82310683e-01 -2.57481076e-02 -6.06677197e-02
-2.46105865e-01 -1.90415651e-01 6.34375274e-01 -1.03671685e-01
-6.62551701e-01 6.47480607e-01 2.41737977e-01 1.95290864e-01
2.45524999e-02 -5.27521729e-01 -5.95921040e-01 -9.02037740e-01
-1.39187962e-01 -9.47190374e-02 4.83352095e-01 -2.15240389e-01
1.39846906e-01 4.57979620e-01 -1.74588037e+00 8.53541642e-02
6.78058445e-01 1.28409719e+00 -2.31336933e-02 7.01846421e-01
3.04449826e-01 1.24913943e+00 -5.16579151e-01 -2.90457547e-01
2.41582785e-02 -3.19540799e-01 -4.74694937e-01 2.40616173e-01
2.85679698e-01 -5.06898880e-01 -6.55303717e-01 6.93149388e-01
3.78044963e-01 5.98123431e-01 2.32062802e-01 -1.18661392e+00
-2.72712022e-01 7.73573592e-02 -6.76552355e-01 7.16262877e-01
3.15816522e-01 5.17956614e-01 6.32898629e-01 -7.34256804e-01
4.78961945e-01 9.71325636e-01 4.21336889e-01 3.97010505e-01
-6.01629436e-01 -5.47841787e-01 2.64004469e-01 6.42288744e-01
-1.27812457e+00 -3.97786111e-01 8.42422843e-01 -6.49285793e-01
9.69164312e-01 -8.27295799e-03 1.14456487e+00 1.15730739e+00
1.31410450e-01 8.91710877e-01 7.36806214e-01 7.44137168e-02
1.14234686e-01 -3.47684026e-01 1.14377484e-01 6.85301542e-01
2.83261746e-01 1.45185236e-02 -6.38489127e-01 2.54838139e-01
6.82476878e-01 3.00954312e-01 -5.38054228e-01 -2.57665187e-01
-1.31215858e+00 5.06187201e-01 5.84791780e-01 2.69352287e-01
-6.81275368e-01 2.42334902e-01 5.39111376e-01 -2.27106750e-01
3.56762558e-01 9.50974897e-02 -1.72467858e-01 -5.37943661e-01
-7.91369259e-01 2.21787214e-01 4.31321144e-01 6.01900399e-01
4.87780988e-01 4.34422374e-01 -3.14426869e-01 6.56535387e-01
1.99717373e-01 4.50691879e-01 4.85579073e-01 -9.10796642e-01
5.50849497e-01 5.86059570e-01 1.92186609e-01 -1.26288283e+00
-3.50989133e-01 -4.58526850e-01 -9.26483989e-01 -8.93234387e-02
3.82950306e-01 -2.19405159e-01 -6.83546722e-01 1.57464051e+00
3.85775983e-01 8.26075196e-01 3.32991295e-02 1.25690472e+00
6.96070373e-01 8.47775578e-01 1.30465388e-01 -2.54491180e-01
1.22863424e+00 -1.26550734e+00 -7.89451480e-01 -3.05762887e-01
2.46893451e-01 -4.54895109e-01 6.87038600e-01 5.64666212e-01
-9.17119205e-01 -7.75319159e-01 -1.22548246e+00 -2.47438941e-02
-2.33487085e-01 2.40699813e-01 4.54098046e-01 3.12467635e-01
-7.50360310e-01 3.80887240e-01 -1.14362299e+00 -4.12368536e-01
5.31876028e-01 2.82936275e-01 -4.83387172e-01 -1.96513578e-01
-9.91970181e-01 6.72891319e-01 3.27487081e-01 5.42043328e-01
-1.25543845e+00 -4.60064769e-01 -1.06588078e+00 1.28687829e-01
5.13159215e-01 -5.77902555e-01 1.02630949e+00 -1.04737711e+00
-1.54918313e+00 3.98121476e-01 -1.97692141e-01 -5.81387103e-01
5.67028463e-01 -5.79476058e-01 -4.78188872e-01 3.33584845e-01
-4.13350537e-02 4.32689905e-01 7.97676802e-01 -7.09908724e-01
-7.12342024e-01 -3.35897475e-01 1.65415213e-01 2.20735639e-01
-3.45221132e-01 4.84577799e-03 -8.06793511e-01 -7.43207157e-01
-4.15794402e-01 -9.75349903e-01 -9.61522087e-02 -8.69554803e-02
-1.04754522e-01 -1.38171703e-01 1.01057720e+00 -8.07234049e-01
1.33102071e+00 -2.06039858e+00 3.76177579e-01 -2.69103825e-01
9.39607024e-02 7.62328923e-01 -2.82496754e-02 1.89919353e-01
8.13170336e-03 -3.65768075e-01 -3.73997867e-01 -3.41044337e-01
-4.25659746e-01 4.74786274e-02 1.29810840e-01 6.31102443e-01
5.35252631e-01 7.86572337e-01 -9.98513877e-01 -2.90033966e-01
5.74865520e-01 7.40344107e-01 -5.79735994e-01 3.95586669e-01
5.62019423e-02 5.14480829e-01 -3.02634895e-01 7.03521848e-01
4.91510689e-01 -9.08787549e-02 -1.30678518e-02 -1.72706842e-01
-3.54744554e-01 -1.13093905e-01 -1.21638250e+00 1.84275627e+00
-2.26897553e-01 8.55157137e-01 -1.08337685e-01 -1.05584979e+00
7.53576756e-01 3.97101551e-01 8.42273355e-01 -6.58995867e-01
4.38224733e-01 -1.99946873e-02 4.53839600e-02 -9.13988233e-01
5.08970201e-01 1.82110980e-01 -5.82013186e-03 1.74805000e-01
2.47782767e-01 5.03435671e-01 4.67623502e-01 1.10106878e-01
1.12151492e+00 2.40189642e-01 3.04361969e-01 2.38060914e-02
1.06275725e+00 -8.62878561e-02 8.84250998e-01 2.23078355e-01
-6.38292372e-01 5.57149887e-01 2.87137568e-01 -8.46954226e-01
-7.87358284e-01 -6.61769152e-01 4.68843788e-01 7.18002379e-01
3.66210163e-01 -3.22563112e-01 -6.24635160e-01 -5.12571812e-01
-3.88480902e-01 1.77778438e-01 -6.90617442e-01 -3.20572287e-01
-8.55527699e-01 -6.24031305e-01 4.19036955e-01 7.84629345e-01
1.06680465e+00 -1.17734122e+00 -1.03347600e+00 2.96498090e-01
-4.87825304e-01 -1.48120558e+00 -4.25055474e-01 -3.23051304e-01
-6.67477608e-01 -1.37439919e+00 -8.41147542e-01 -3.49611551e-01
2.66293973e-01 5.63418865e-01 7.85211682e-01 2.26245783e-02
-2.75665969e-01 3.74519944e-01 -4.49578643e-01 -3.45630884e-01
4.24676925e-01 1.43757567e-01 1.55877648e-02 7.10096002e-01
4.13752675e-01 -3.78346115e-01 -7.68073380e-01 2.63006657e-01
-9.42269564e-01 -7.58037567e-02 5.17320573e-01 8.65943015e-01
3.73957574e-01 -1.67195186e-01 2.53553391e-01 -4.23167378e-01
2.16984734e-01 -7.02260971e-01 -6.02783203e-01 1.80069227e-02
-1.00949919e-02 -2.55021363e-01 5.89306891e-01 -4.18062687e-01
-1.20343876e+00 2.72647887e-01 -1.54766757e-02 -6.82345152e-01
-1.38557941e-01 4.12866861e-01 -5.18388115e-02 7.54766464e-02
3.42583299e-01 3.44894201e-01 8.25755298e-02 -2.11409763e-01
-1.25734538e-01 4.48300600e-01 7.05985487e-01 -2.33743355e-01
6.65973723e-01 7.04738617e-01 -7.43372813e-02 -9.45621490e-01
-8.87698174e-01 -6.76895380e-01 -5.77271044e-01 -5.68451583e-01
1.22841406e+00 -1.12673175e+00 -7.77692735e-01 9.29946959e-01
-1.11185956e+00 -1.89674094e-01 7.03743612e-03 6.81988418e-01
-3.33860129e-01 3.99079561e-01 -4.41523105e-01 -8.26578200e-01
-3.21731150e-01 -1.16240525e+00 9.10082936e-01 5.81259370e-01
1.02365658e-01 -7.59635568e-01 7.72617757e-02 7.87894487e-01
4.81680036e-01 5.83496690e-01 -4.14478732e-03 -2.75874853e-01
-8.11420977e-01 -1.78362757e-01 -7.38294423e-02 4.57953513e-01
6.11178316e-02 1.07418664e-01 -1.00767314e+00 -2.83056021e-01
-2.12613404e-01 -2.64098704e-01 8.16404521e-01 5.04337192e-01
1.06744635e+00 -5.57575710e-02 -4.36851867e-02 8.16607714e-01
1.24099874e+00 4.51736122e-01 7.24057436e-01 3.83353680e-01
8.63819897e-01 4.41383064e-01 7.81110227e-01 6.46761715e-01
5.08711934e-01 8.47683609e-01 6.49738431e-01 -6.70265639e-03
-5.76937981e-02 3.41394134e-02 5.80976367e-01 7.42932320e-01
-5.54932237e-01 -3.22844982e-01 -8.59915972e-01 7.68995643e-01
-2.14553165e+00 -1.22817481e+00 2.06012949e-02 1.92852855e+00
1.46650314e-01 3.03823315e-02 3.28201354e-01 3.87969702e-01
6.81204975e-01 6.00773394e-01 -5.21427631e-01 -2.21418485e-01
7.94871971e-02 1.16913356e-01 2.69073188e-01 2.47793660e-01
-1.45291054e+00 8.20163250e-01 4.51296377e+00 4.26120758e-01
-1.22847855e+00 6.88523501e-02 4.41075504e-01 -2.11630657e-01
4.43095982e-01 -3.20245743e-01 -5.43325245e-01 6.17219508e-01
9.26830053e-01 4.17751372e-02 4.22777563e-01 7.01497495e-01
5.36777914e-01 -1.56462133e-01 -7.05223680e-01 1.15723896e+00
3.80168945e-01 -1.23298073e+00 -1.17383912e-01 -7.25259259e-02
6.32042885e-01 -5.05580306e-02 -1.07410774e-01 3.29027563e-01
-1.20509438e-01 -8.89037967e-01 6.01679265e-01 8.18263710e-01
4.47318375e-01 -9.55253243e-01 1.14002454e+00 4.17089313e-01
-1.56666279e+00 -3.30754310e-01 2.00787187e-02 -3.36233974e-01
3.43725443e-01 3.32848340e-01 -4.07945395e-01 6.21676981e-01
9.73660290e-01 1.21488392e+00 -3.36132646e-01 1.15422404e+00
-2.55484641e-01 5.37488699e-01 -1.20060384e-01 1.95096340e-02
4.93349969e-01 -1.61262244e-01 6.13473177e-01 1.11137676e+00
2.89629728e-01 3.19797695e-01 1.00357555e-01 3.95512611e-01
2.37866938e-02 -1.57541916e-01 -5.00934184e-01 -6.15079068e-02
1.41335890e-01 1.28552210e+00 -4.70559508e-01 -4.13710564e-01
-5.34407794e-01 8.96996796e-01 1.98450401e-01 2.95325488e-01
-1.19287181e+00 -5.02181649e-01 9.27561522e-01 -1.32246350e-03
4.77040231e-01 -5.18545628e-01 2.97946334e-01 -1.45835924e+00
1.16828814e-01 -7.90766656e-01 4.99469519e-01 -7.86912203e-01
-7.23618865e-01 6.84636414e-01 3.99810672e-02 -1.32865775e+00
-1.78755686e-01 -4.42244768e-01 -6.29448354e-01 5.57334006e-01
-1.62495744e+00 -9.93559241e-01 -1.04552162e+00 8.04693937e-01
9.18507218e-01 -1.76639363e-01 2.47602493e-01 6.45044923e-01
-1.25100732e+00 3.11434865e-01 -5.23409843e-01 4.12236303e-01
4.35292363e-01 -8.62836719e-01 9.05761123e-02 1.47825527e+00
-1.61519691e-01 3.03092182e-01 3.24912488e-01 -2.55841970e-01
-1.41838491e+00 -1.33376300e+00 7.78622448e-01 -1.80729240e-01
4.90352213e-01 -1.09305792e-01 -7.25026608e-01 6.28633857e-01
2.84194380e-01 5.60084879e-01 4.73105639e-01 -5.20125806e-01
5.75853102e-02 -3.82769525e-01 -8.87973785e-01 4.41449136e-01
1.04780793e+00 -1.86653450e-01 -3.56222481e-01 -1.07203260e-01
4.58124220e-01 -4.10698622e-01 -7.79999852e-01 3.94975096e-01
5.90486825e-01 -1.16805768e+00 9.16015625e-01 -4.34363484e-01
5.04451513e-01 -6.31785452e-01 -2.26184148e-02 -1.12843180e+00
-3.44403982e-01 -6.07088029e-01 -5.18279195e-01 9.29374039e-01
-1.93783417e-01 -4.45571274e-01 7.72896647e-01 3.86687756e-01
-3.73873532e-01 -7.92200863e-01 -8.70447278e-01 -5.21316350e-01
-6.50873959e-01 -4.09364134e-01 4.02944505e-01 8.14771056e-01
-3.25617105e-01 2.78370053e-01 -7.62076318e-01 1.65098220e-01
5.45370579e-01 -2.53716260e-01 9.23450232e-01 -1.04250002e+00
-1.42271012e-01 -2.95706064e-01 -7.09575832e-01 -1.14400625e+00
-2.98299287e-02 -3.02042484e-01 -5.91173843e-02 -1.33490920e+00
6.49304762e-02 8.84998962e-02 -5.08421242e-01 3.21817607e-01
-2.59219944e-01 3.83132756e-01 3.40606868e-01 2.84863766e-02
-8.32779646e-01 9.42404091e-01 1.16717637e+00 -2.91023195e-01
-1.40348002e-01 -3.34352143e-02 -3.85308921e-01 7.01560795e-01
7.86589742e-01 -2.11821660e-01 -4.66400206e-01 -5.80981433e-01
-2.31140107e-01 1.07031494e-01 5.07273018e-01 -1.53262281e+00
3.65154058e-01 -1.31676629e-01 2.64343470e-01 -4.80253905e-01
6.79033637e-01 -8.75692844e-01 1.47362843e-01 6.83330834e-01
1.53306993e-02 3.48844945e-01 1.71512440e-01 7.33376682e-01
-7.26716101e-01 2.18809426e-01 7.93030739e-01 -5.92238232e-02
-1.18149626e+00 5.61003685e-01 -4.16204244e-01 1.87092703e-02
1.44361258e+00 -3.48779976e-01 -2.39982009e-01 -2.83621490e-01
-4.92421091e-01 3.25701207e-01 2.40092613e-02 6.09222531e-01
7.83460975e-01 -1.29975224e+00 -6.47191346e-01 2.73614675e-01
1.08292863e-01 6.66674376e-02 6.58169866e-01 1.14195859e+00
-7.28013277e-01 5.32457232e-01 -5.12956023e-01 -6.80722654e-01
-1.18835068e+00 3.61018538e-01 3.83971512e-01 -2.22918317e-01
-4.43020254e-01 6.55941665e-01 -1.78234741e-01 1.37886643e-01
2.12810725e-01 -4.91561562e-01 -6.26761794e-01 1.45768642e-01
6.46774828e-01 7.43415594e-01 1.90652329e-02 -1.01839340e+00
-4.95596021e-01 7.74312317e-01 2.57042915e-01 1.54500693e-01
1.41547310e+00 1.06804349e-01 2.77425379e-01 3.03801358e-01
9.99959886e-01 -3.16853851e-01 -1.67768872e+00 -2.47541979e-01
-2.39006594e-01 -6.05106533e-01 1.04714014e-01 -4.20044243e-01
-1.48837650e+00 1.13773620e+00 5.26329041e-01 -3.77337672e-02
1.35343051e+00 -4.78621453e-01 1.03571773e+00 1.25542179e-01
2.48612359e-01 -8.73037159e-01 3.46901447e-01 5.77078223e-01
7.78489530e-01 -1.24845541e+00 -3.82420197e-02 -1.38895392e-01
-8.16952348e-01 9.79792714e-01 8.37523460e-01 -2.93817163e-01
5.21771491e-01 -1.22023717e-01 -1.53581411e-01 -1.35854781e-01
-9.00072336e-01 -3.38970065e-01 3.47454667e-01 4.34617937e-01
2.11840674e-01 -1.90809295e-01 -1.30925655e-01 5.26560724e-01
2.92624116e-01 3.32771719e-01 4.48883057e-01 1.04940617e+00
-3.05051327e-01 -5.40727913e-01 -2.29133084e-01 7.11370632e-02
-4.66529727e-01 2.61423260e-01 -4.99364659e-02 8.24750364e-01
4.22325432e-01 9.94336307e-01 1.38465673e-01 -6.06940031e-01
4.69218940e-01 -2.70070285e-01 1.07646577e-01 -2.98564017e-01
-6.34393394e-01 -1.32625088e-01 6.69458956e-02 -8.90156209e-01
-9.92270947e-01 -7.76944578e-01 -9.06254947e-01 -2.64453351e-01
-4.99594398e-02 -7.17399195e-02 3.02269757e-01 9.66232181e-01
4.44908798e-01 7.38269269e-01 5.19713998e-01 -1.08316815e+00
5.55842035e-02 -9.41839576e-01 -1.90839380e-01 4.68650371e-01
4.91912276e-01 -9.81679022e-01 -1.90830901e-01 2.03661129e-01]
|
[8.084348678588867, 0.6136142015457153]
|
97294a3a-6db0-4d30-a256-039b013fa1ac
|
distributed-dual-quaternion-based
|
2203.06278
| null |
https://arxiv.org/abs/2203.06278v1
|
https://arxiv.org/pdf/2203.06278v1.pdf
|
Distributed Dual Quaternion Based Localization of Visual Sensor Networks
|
In this paper we consider the localization problem for a visual sensor network. Inspired by the alternate attitude and position distributed optimization framework discussed in [1], we propose an estimation scheme that exploits the unit dual quaternion algebra to describe the sensors pose. This representation is beneficial in the formulation of the optimization scheme allowing to solve the localization problem without designing two interlaced position and orientation estimators, thus improving the estimation error distribution over the two pose components and the overall localization performance. Furthermore, the numerical experimentation asserts the robustness of the proposed algorithm w.r.t. the initial conditions.
|
['Angelo Cenedese', 'Giulia Michieletto', 'Marco Fabris', 'Luca Varotto']
|
2022-03-11
| null | null | null | null |
['distributed-optimization']
|
['methodology']
|
[-7.72111043e-02 2.83582032e-01 -1.07513912e-01 5.00194095e-02
-2.37714816e-02 -6.50040209e-01 5.70480466e-01 5.53130865e-01
-8.55564415e-01 9.73909676e-01 -3.85831118e-01 3.18910391e-03
-3.28339040e-01 -5.88752270e-01 -6.49201989e-01 -7.99636602e-01
-3.94240506e-02 9.47267339e-02 -8.35828781e-02 -3.55702221e-01
2.14915499e-01 7.82244921e-01 -1.39201963e+00 -1.12971556e+00
7.24277258e-01 1.14600754e+00 1.51188731e-01 7.51143277e-01
5.83876014e-01 3.32451969e-01 -6.12368941e-01 -1.83680058e-02
2.51224011e-01 -9.48438793e-02 -2.16066703e-01 4.15277153e-01
3.00097823e-01 -1.25702918e-01 2.25598440e-01 1.33947825e+00
5.13652265e-01 1.77193403e-01 5.62389076e-01 -1.11430776e+00
1.44383386e-01 6.37373254e-02 -5.41871250e-01 -5.67005761e-02
5.90098917e-01 -5.03280520e-01 7.67246544e-01 -8.97539198e-01
7.81415522e-01 7.47746050e-01 7.09216058e-01 -1.94779739e-01
-1.17541087e+00 1.90394580e-01 -2.88887799e-01 1.66235238e-01
-1.91475976e+00 -2.83828974e-01 8.61574829e-01 -3.82219404e-01
1.63572177e-01 3.01155388e-01 7.95865595e-01 6.08782232e-01
3.30562830e-01 1.71383560e-01 8.95810246e-01 -5.76197147e-01
4.42110777e-01 4.31460328e-02 2.77155694e-02 6.78723335e-01
1.09945536e+00 -6.22724481e-02 -2.21606180e-01 -1.22335128e-01
7.86931038e-01 -1.41398579e-01 -3.02724808e-01 -1.11284935e+00
-1.20643735e+00 8.79425287e-01 5.99403560e-01 3.68902117e-01
-6.93325758e-01 1.09848201e-01 1.87854484e-01 -2.57499311e-02
2.72000909e-01 2.75450885e-01 1.42561868e-01 -6.91501016e-04
-5.91104448e-01 1.38373077e-01 1.01777804e+00 9.43691850e-01
8.55532408e-01 3.59947234e-01 5.56316257e-01 2.03599915e-01
5.53270102e-01 1.07646918e+00 6.91191703e-02 -6.01751745e-01
2.77426451e-01 4.50604409e-01 4.96519148e-01 -1.49502885e+00
-6.48711145e-01 -6.22980595e-01 -9.75339293e-01 1.11692846e-01
2.92340904e-01 -5.67510009e-01 -2.51579791e-01 1.50149977e+00
6.26529157e-01 -3.17663252e-01 2.89908409e-01 1.00709510e+00
-4.06007692e-02 4.03324991e-01 -3.39416772e-01 -4.48455095e-01
1.07834971e+00 -3.82415086e-01 -1.02108479e+00 -5.92541359e-02
1.72405943e-01 -8.26184452e-01 8.02255198e-02 4.12695467e-01
-9.00538802e-01 -5.00921547e-01 -1.49981904e+00 3.91603768e-01
-3.67668718e-01 6.41050696e-01 9.97829735e-02 5.39210200e-01
-1.02995574e+00 3.23200762e-01 -8.40326250e-01 -7.47034252e-01
-6.53584838e-01 5.33059537e-01 -3.58781308e-01 5.02725661e-01
-9.76155639e-01 1.30223501e+00 7.17464805e-01 5.85702956e-01
-1.28166035e-01 6.21368252e-02 -7.91892290e-01 -2.67227262e-01
2.43976176e-01 -5.34848571e-01 7.79970169e-01 -7.56666899e-01
-1.85522819e+00 3.90936136e-01 -1.53727785e-01 -5.12515128e-01
6.52444243e-01 -1.04565941e-01 -3.42090912e-02 3.42472643e-01
8.33974853e-02 1.98048756e-01 8.59148502e-01 -1.21096182e+00
-4.52460736e-01 -4.85862166e-01 8.43819976e-02 4.09970224e-01
-2.78623134e-01 -7.87970304e-01 -1.27696261e-01 -5.08927703e-01
5.22168100e-01 -1.13038754e+00 -3.92429084e-01 -6.59539625e-02
-2.88323104e-01 3.11415166e-01 5.89760363e-01 -4.48129416e-01
8.56464148e-01 -2.04387760e+00 8.36433709e-01 7.19659925e-01
-5.41695878e-02 -6.61805719e-02 2.31824249e-01 6.22624755e-01
1.88254211e-02 -5.45985818e-01 1.36291146e-01 -3.74787480e-01
-1.55681565e-01 4.02360857e-01 -1.68532625e-01 1.40102506e+00
-1.35929435e-01 1.61830723e-01 -7.16541946e-01 -3.00410420e-01
4.31436628e-01 5.98556757e-01 -1.54486582e-01 2.45696738e-01
2.69656479e-01 6.02385461e-01 -5.53093195e-01 3.57135475e-01
6.77375615e-01 2.02245235e-01 4.77646798e-01 -6.07499003e-01
-5.88852108e-01 -3.71773928e-01 -1.86503494e+00 1.48136044e+00
-4.09392953e-01 3.83435428e-01 6.29283428e-01 -1.11927760e+00
1.27206290e+00 3.65626186e-01 7.70722985e-01 -4.87521321e-01
6.89059615e-01 3.48167449e-01 -3.23431641e-01 -9.82533693e-02
7.57736564e-01 2.55371444e-03 -1.21897615e-01 -1.02253325e-01
7.43224993e-02 -2.19808429e-01 3.43450338e-01 -1.10304326e-01
3.35823834e-01 1.82321981e-01 1.10286677e+00 -6.04247034e-01
9.75903928e-01 -6.22064285e-02 4.08542365e-01 3.16646159e-01
2.79072411e-02 1.14986204e-01 2.32951149e-01 -1.54107139e-01
-1.02161562e+00 -9.36403394e-01 -2.77860433e-01 3.57755572e-01
5.88893652e-01 -4.04864401e-02 -3.91281873e-01 5.52236848e-02
1.41888797e-01 1.03977762e-01 -4.07287091e-01 3.89966890e-02
-5.28238595e-01 -3.67575437e-01 3.37454855e-01 4.34723049e-02
4.27239180e-01 1.25573780e-02 -1.12223804e+00 2.05147788e-01
-2.28105024e-01 -1.04427552e+00 1.18490204e-01 3.55838537e-01
-8.36109817e-01 -1.23956764e+00 -8.03372502e-01 -4.49158847e-01
9.80822563e-01 -2.29446143e-02 5.23230910e-01 -2.61764079e-01
2.09954157e-02 9.68423784e-01 -5.21191418e-01 -3.38295937e-01
-1.57099843e-01 8.47506672e-02 4.87791806e-01 2.52347708e-01
-2.52188623e-01 -3.26288164e-01 -3.09555173e-01 2.37407297e-01
-8.16482186e-01 -4.60177571e-01 4.09168482e-01 6.29525781e-01
4.49712366e-01 -2.28145361e-01 1.17333323e-01 -1.07336253e-01
5.24774253e-01 -3.79371494e-01 -1.32845759e+00 7.89524093e-02
-6.56878054e-01 2.38200635e-01 6.40373588e-01 -2.23634452e-01
-5.49154639e-01 4.75197554e-01 -3.86280171e-03 -2.16073152e-02
2.14421108e-01 4.62798178e-01 -1.25510925e-02 -9.18073714e-01
3.91962349e-01 2.65303493e-01 3.96144331e-01 -3.99920255e-01
4.65989947e-01 4.21363384e-01 4.57782567e-01 -3.15497309e-01
9.33511794e-01 4.99603987e-01 7.22800732e-01 -1.39628291e+00
-2.80782133e-01 -6.31590068e-01 -7.92572558e-01 -4.03557271e-01
6.81600511e-01 -7.83128381e-01 -1.14799333e+00 2.03961745e-01
-1.46225345e+00 6.58216476e-01 -2.89013356e-01 6.44681036e-01
-7.76040792e-01 7.26324916e-01 5.67908287e-02 -1.13748431e+00
-1.69077888e-01 -1.34827733e+00 8.55796635e-01 1.91348746e-01
-1.56850249e-01 -1.25284028e+00 3.92447978e-01 -5.29378295e-01
2.63394564e-01 6.15896046e-01 -1.45441741e-01 -1.47894368e-01
-4.81676012e-01 -5.13980746e-01 1.79382548e-01 2.35465020e-01
-1.52824149e-01 -1.09150521e-01 -4.28684592e-01 -7.48639405e-01
3.11563969e-01 -2.54155844e-02 2.55999476e-01 9.76421386e-02
2.03724988e-02 -2.13310853e-01 -2.60390937e-01 6.96185946e-01
1.88638949e+00 6.87007084e-02 7.47872982e-04 5.58113992e-01
4.57568645e-01 3.47075492e-01 8.64892781e-01 7.96372414e-01
3.66935015e-01 9.82920885e-01 9.63551879e-01 -1.59046315e-02
5.44708133e-01 8.28977581e-03 1.96675971e-01 8.53112698e-01
-1.79969430e-01 -7.56411850e-02 -6.52276754e-01 5.48211753e-01
-1.96667659e+00 -4.21504676e-01 -1.22167706e-01 2.29042578e+00
2.27671459e-01 -4.30824667e-01 -1.52231395e-01 2.60940790e-01
7.09012866e-01 3.73927891e-01 -1.01231687e-01 -4.34306651e-01
-2.61779428e-01 -4.82583530e-02 1.23025846e+00 9.57631588e-01
-1.07097530e+00 1.82174042e-01 6.30701637e+00 3.10880989e-01
-1.33677053e+00 -1.41390368e-01 -3.36039364e-01 5.69353521e-01
1.14985794e-01 1.43007770e-01 -8.61844003e-01 2.07156375e-01
5.72172940e-01 -4.30585653e-01 2.36905903e-01 8.92805457e-01
1.50563434e-01 -6.64492607e-01 -6.85179353e-01 1.04132414e+00
3.90586078e-01 -8.33935440e-01 -1.17904574e-01 1.16285019e-01
4.30878073e-01 -2.70506471e-01 -7.86445215e-02 -4.77958620e-01
-4.36129391e-01 -3.00341517e-01 8.43104780e-01 5.73466480e-01
3.18096429e-01 -7.92843163e-01 9.30306196e-01 4.25365925e-01
-1.22650409e+00 -1.26387756e-02 -3.96789551e-01 -4.19667721e-01
4.19684917e-01 5.51291764e-01 -1.01055133e+00 1.14770985e+00
2.83846110e-02 6.30273700e-01 -4.81126070e-01 1.19877458e+00
-3.51250231e-01 -5.55854812e-02 -8.51751328e-01 -4.65170026e-01
2.38626689e-01 -8.58819246e-01 1.11796784e+00 7.55782425e-01
6.14333153e-01 -2.66129524e-01 1.82812601e-01 4.35203612e-01
4.15410221e-01 3.08484584e-01 -7.57293701e-01 3.41400713e-01
2.48815075e-01 1.18392479e+00 -6.30490482e-01 -2.12418865e-02
-7.89064541e-02 8.13729525e-01 7.31025562e-02 4.02041852e-01
-6.37001812e-01 -6.22398376e-01 4.52835083e-01 -1.37221649e-01
4.74240720e-01 -6.91392779e-01 -3.27105299e-02 -1.24351513e+00
1.49594694e-01 -2.90061235e-01 -6.45033419e-02 -3.32544506e-01
-5.89385331e-01 4.63512391e-01 1.94219559e-01 -1.56701434e+00
-7.69424796e-01 -7.01208353e-01 -2.11619705e-01 7.84308195e-01
-1.30898631e+00 -8.37250650e-01 -1.16831437e-01 4.48051631e-01
-2.46225625e-01 1.67720124e-01 7.16551423e-01 2.17894644e-01
-3.86557698e-01 1.19296052e-01 4.08861637e-01 -1.32727653e-01
4.37884539e-01 -1.20863223e+00 -4.85556334e-01 1.18991292e+00
-2.41531029e-01 7.86512494e-01 1.24951398e+00 -3.36425841e-01
-1.71131909e+00 -5.94056189e-01 8.10319960e-01 1.04311332e-01
8.97140324e-01 -9.46346596e-02 -1.00696459e-01 6.72677100e-01
2.87251770e-01 2.14833370e-03 1.13944963e-01 -2.06747681e-01
2.68988013e-01 -4.66584593e-01 -1.07990241e+00 3.56996804e-01
1.25503138e-01 -3.40063155e-01 -5.40784717e-01 1.58529133e-01
3.74990255e-02 -5.91774642e-01 -1.09912062e+00 3.16851646e-01
5.18581152e-01 -7.95480669e-01 9.51772869e-01 2.62526363e-01
-6.16995335e-01 -5.81204355e-01 -3.51245344e-01 -1.31449234e+00
-3.39446515e-02 -5.45836270e-01 -5.97414598e-02 1.05095243e+00
-5.72342128e-02 -9.35780168e-01 6.86569452e-01 -1.13137953e-01
4.18121219e-01 -3.43162626e-01 -1.42684305e+00 -6.51002824e-01
-5.37541330e-01 3.69102620e-02 -1.12909272e-01 4.42078859e-01
5.16382642e-02 3.66566002e-01 -4.72823322e-01 7.08242118e-01
1.09480739e+00 1.07748164e-02 8.09285402e-01 -1.09031129e+00
-2.97145277e-01 -1.12877379e-03 -9.72455084e-01 -1.10321736e+00
4.06341851e-02 -2.63869584e-01 8.58956575e-02 -1.16393256e+00
-6.01921618e-01 -1.07918456e-01 -1.97583754e-02 -1.33145168e-01
1.59536377e-01 3.25430989e-01 3.42956215e-01 9.45720002e-02
-5.13437927e-01 5.71732700e-01 7.80562818e-01 8.94575119e-02
6.36942014e-02 1.75216928e-01 -1.72615051e-04 8.35421443e-01
6.58086956e-01 -1.02506980e-01 -3.86666268e-01 -3.66935253e-01
4.13670778e-01 2.31126636e-01 4.59510475e-01 -1.18408096e+00
4.36829507e-01 1.40545785e-01 1.62469447e-01 -6.36773109e-01
5.31036675e-01 -1.39276958e+00 3.12999815e-01 8.65371227e-01
2.74092704e-01 4.63061154e-01 -1.38751015e-01 6.39913142e-01
-5.62844753e-01 -4.41637367e-01 9.36992288e-01 2.20575154e-01
-8.03085864e-01 -1.95010751e-01 -3.24708223e-01 -4.14597064e-01
1.17354572e+00 -3.33370894e-01 6.64486317e-03 -4.55675602e-01
-9.31560755e-01 4.70006801e-02 7.44014442e-01 2.03588158e-02
3.72372091e-01 -1.21874046e+00 -9.96005759e-02 2.48068511e-01
1.09485701e-01 -1.49575740e-01 -1.36416689e-01 1.19732785e+00
-1.00924373e+00 4.84640092e-01 -3.96628737e-01 -7.91124821e-01
-1.06179142e+00 2.91260451e-01 4.36031371e-01 3.68454717e-02
1.22663938e-02 2.48653278e-01 -6.73989713e-01 -3.14303100e-01
2.49548584e-01 -2.11549670e-01 -3.70805383e-01 3.48285407e-01
2.49627046e-02 7.26009071e-01 2.65150573e-02 -1.02455914e+00
-5.94858110e-01 1.08414185e+00 7.51076937e-01 -3.22081596e-01
1.09692121e+00 -6.18108034e-01 -3.86705190e-01 5.86561382e-01
1.25374568e+00 3.96321982e-01 -9.22457039e-01 6.00522123e-02
1.43620834e-01 -1.73592001e-01 4.75708470e-02 -9.80950668e-02
-6.06581211e-01 4.01855201e-01 6.87239170e-01 2.59846300e-01
9.22572136e-01 -7.27308214e-01 -4.36551031e-03 6.43291056e-01
6.85170829e-01 -1.07298684e+00 -3.19652587e-01 5.53743005e-01
8.32645357e-01 -8.58908474e-01 5.13644636e-01 -4.62812185e-01
-1.94364846e-01 1.40831351e+00 1.10611558e-01 -6.27877116e-01
5.70256352e-01 2.94663161e-01 9.84814242e-02 3.67753357e-01
4.53121774e-02 -4.04927880e-01 7.32351840e-02 5.05300999e-01
4.27952409e-01 6.95276186e-02 -1.20346606e+00 -2.65824169e-01
1.50017813e-01 -1.88337207e-01 7.26528943e-01 1.05183339e+00
-5.82804918e-01 -1.14419353e+00 -9.31289673e-01 -4.41684663e-01
-7.77997896e-02 5.63068688e-01 -2.07137510e-01 1.09813190e+00
2.36041769e-01 7.88203895e-01 -1.12061203e-01 5.36615290e-02
7.02620447e-01 -2.42004871e-01 6.36461914e-01 -9.93730575e-02
-2.24657580e-01 2.85060108e-01 6.16300991e-03 -5.44118404e-01
-7.35128582e-01 -6.75735116e-01 -1.02703989e+00 5.20543493e-02
-2.63878524e-01 7.22702682e-01 1.26120830e+00 6.56213224e-01
1.48175701e-01 2.93420911e-01 7.74564326e-01 -9.14254725e-01
-1.13084137e+00 -8.41552258e-01 -8.99793983e-01 -3.14057246e-02
6.96467400e-01 -9.29157972e-01 -2.70946532e-01 -1.94574088e-01]
|
[7.790817737579346, -2.2239720821380615]
|
b75a0f71-5fe1-42ea-b771-20bd126ea21d
|
learn-an-effective-lip-reading-model-without
|
2011.07557
| null |
https://arxiv.org/abs/2011.07557v1
|
https://arxiv.org/pdf/2011.07557v1.pdf
|
Learn an Effective Lip Reading Model without Pains
|
Lip reading, also known as visual speech recognition, aims to recognize the speech content from videos by analyzing the lip dynamics. There have been several appealing progress in recent years, benefiting much from the rapidly developed deep learning techniques and the recent large-scale lip-reading datasets. Most existing methods obtained high performance by constructing a complex neural network, together with several customized training strategies which were always given in a very brief description or even shown only in the source code. We find that making proper use of these strategies could always bring exciting improvements without changing much of the model. Considering the non-negligible effects of these strategies and the existing tough status to train an effective lip reading model, we perform a comprehensive quantitative study and comparative analysis, for the first time, to show the effects of several different choices for lip reading. By only introducing some easy-to-get refinements to the baseline pipeline, we obtain an obvious improvement of the performance from 83.7% to 88.4% and from 38.2% to 55.7% on two largest public available lip reading datasets, LRW and LRW-1000, respectively. They are comparable and even surpass the existing state-of-the-art results.
|
['Xilin Chen', 'Shiguang Shan', 'Shuang Yang', 'Dalu Feng']
|
2020-11-15
| null | null | null | null |
['lipreading']
|
['computer-vision']
|
[ 8.25336352e-02 6.44045621e-02 -5.14200330e-01 -2.42605135e-01
-1.01741242e+00 -1.93124935e-01 6.18809283e-01 -3.76841962e-01
-3.71113569e-01 6.25709176e-01 4.60610360e-01 -2.32829556e-01
1.75361603e-01 -1.04360335e-01 -5.96202493e-01 -8.21728110e-01
1.66374803e-01 6.54703975e-02 4.09088880e-01 -2.10517691e-03
3.87628049e-01 3.36705148e-01 -2.14343357e+00 1.35630235e-01
8.46690118e-01 1.09613252e+00 2.47758001e-01 5.97518682e-01
-3.66238236e-01 4.91995662e-01 -4.01332766e-01 -5.16009986e-01
-2.75379587e-02 -2.03987032e-01 -6.87121868e-01 -7.69503340e-02
3.74809891e-01 -3.56561393e-01 -2.30481818e-01 8.36773217e-01
1.18313241e+00 -1.82808965e-01 5.04351020e-01 -1.24234354e+00
-5.50545216e-01 3.66806448e-01 -6.08247161e-01 -6.94793761e-02
4.36908990e-01 3.72488409e-01 7.62168705e-01 -9.33068871e-01
4.91545200e-01 1.08694923e+00 6.95608974e-01 8.52342069e-01
-8.33167732e-01 -5.75010478e-01 -5.72618656e-02 4.33462054e-01
-1.35728979e+00 -1.38601422e+00 8.89541566e-01 -2.88919210e-01
9.15667057e-01 1.13976538e-01 4.47022796e-01 1.21499360e+00
-3.33509833e-01 1.02453673e+00 1.23900771e+00 -5.40393889e-01
-8.66437480e-02 1.91614270e-01 -1.08743634e-03 5.48469365e-01
-9.44269225e-02 1.20241893e-02 -6.61070585e-01 2.18627304e-01
2.68866509e-01 -4.34135616e-01 -6.09842420e-01 -3.92013997e-01
-1.13392150e+00 4.13139939e-01 6.68080077e-02 1.34056449e-01
-1.09334804e-01 -1.62251502e-01 6.94465816e-01 -7.69580156e-02
4.28860575e-01 -1.86605811e-01 -5.43159068e-01 -4.54398990e-01
-9.79366958e-01 -2.10366137e-02 7.31843710e-01 7.28697300e-01
5.22391796e-01 -5.42617813e-02 -6.44029826e-02 1.22086442e+00
4.24494743e-01 5.31690001e-01 7.98170328e-01 -7.77284682e-01
4.48803157e-01 1.22369193e-01 -7.60210380e-02 -5.86996436e-01
-4.45762038e-01 -1.33918315e-01 -6.83151603e-01 3.39041263e-01
7.63353825e-01 -1.74183235e-01 -9.54510450e-01 1.79791594e+00
1.82215959e-01 8.38852823e-02 3.96455750e-02 5.92194855e-01
1.24703920e+00 3.10133576e-01 6.99299425e-02 -3.37401479e-01
1.28326368e+00 -1.06026614e+00 -9.77974355e-01 2.33391579e-02
2.71853507e-01 -9.84257996e-01 1.48554504e+00 2.64507979e-01
-1.24159968e+00 -5.49287081e-01 -8.37837636e-01 -2.08040074e-01
-4.90667701e-01 3.70096117e-01 5.58613122e-01 8.43360364e-01
-1.49480164e+00 3.34981620e-01 -4.30810362e-01 -3.80012929e-01
5.66225886e-01 3.62493128e-01 -4.73421216e-01 8.70018080e-02
-9.35298502e-01 7.17757881e-01 -2.19517469e-01 -2.29901896e-04
-5.60501218e-01 -6.44387722e-01 -8.03785563e-01 -6.56175986e-02
6.11504078e-01 -4.16216791e-01 1.35827768e+00 -6.86671555e-01
-2.10845017e+00 1.08831429e+00 -6.91508114e-01 -2.61109143e-01
8.18576455e-01 -1.16610564e-01 -3.03698123e-01 1.74794883e-01
-3.07514936e-01 1.02005148e+00 1.00923979e+00 -1.11204255e+00
-5.04428685e-01 -8.08058456e-02 -1.48881197e-01 7.16246665e-02
-1.59840837e-01 3.60814363e-01 -8.41318727e-01 -4.49424922e-01
-1.07075393e-01 -7.82239020e-01 4.91851002e-01 2.02439800e-01
-3.53061408e-01 -6.16398633e-01 7.48932242e-01 -8.82300019e-01
9.42684352e-01 -2.07388377e+00 -1.14936419e-01 -5.78596413e-01
2.14346647e-01 6.81514978e-01 -2.28947937e-01 2.09797367e-01
-9.15863886e-02 3.26853961e-01 -1.07810637e-02 -7.61381388e-01
1.92260280e-01 -2.87433654e-01 -3.70876752e-02 4.57495600e-01
1.22660331e-01 1.16408610e+00 -5.31881511e-01 -5.68745732e-01
3.46353173e-01 8.43565285e-01 -3.00872177e-01 2.02370688e-01
-3.98552753e-02 8.09725076e-02 2.34884426e-01 9.17108119e-01
8.29088926e-01 -3.10659949e-02 -2.52058655e-01 -2.53807902e-01
-1.48549527e-01 6.02658987e-01 -7.60702133e-01 1.59822416e+00
-4.04906541e-01 1.18297172e+00 2.88287610e-01 -9.98457730e-01
6.66902244e-01 5.25543690e-01 4.44521010e-01 -8.00368071e-01
1.04983076e-01 2.37403601e-01 5.68071902e-02 -9.65150118e-01
1.26626879e-01 -7.17964023e-02 4.24318969e-01 2.37975642e-01
4.11118492e-02 7.22930431e-02 -4.85622063e-02 -1.11551225e-01
4.57562655e-01 3.19632106e-02 2.16756538e-01 -2.91712850e-01
8.62398326e-01 -7.11370349e-01 2.42380187e-01 4.41589504e-01
-6.78193986e-01 7.55023599e-01 5.93819022e-01 1.37717894e-03
-7.35012293e-01 -8.52368057e-01 -3.57585758e-01 1.12803388e+00
-8.72539431e-02 -3.58109564e-01 -9.52266514e-01 -5.97158372e-01
-2.30326176e-01 2.04984993e-01 -4.89223003e-01 6.16570078e-02
-3.43669564e-01 -6.10695362e-01 7.52443492e-01 5.07414162e-01
7.87475705e-01 -1.32574034e+00 -3.19428772e-01 -2.56731689e-01
-3.07625175e-01 -1.27280891e+00 -4.05808121e-01 -1.28579482e-01
-2.58737504e-01 -8.32146168e-01 -1.29282737e+00 -9.77450490e-01
2.35118032e-01 3.04337710e-01 9.31895614e-01 1.80508092e-01
-9.56550539e-02 1.35605693e-01 -2.04173446e-01 -5.12341976e-01
-4.50703502e-01 2.75599688e-01 1.67204767e-01 -1.59196053e-02
2.70424634e-01 -3.38166088e-01 -5.09494662e-01 3.19802940e-01
-4.47042704e-01 1.21665955e-01 6.04676545e-01 7.56692171e-01
2.04705864e-01 -3.87127012e-01 9.32587028e-01 -1.64114892e-01
6.17785573e-01 -1.27962753e-01 -3.21000904e-01 2.16611847e-01
-7.12055683e-01 -2.18496248e-02 3.58369559e-01 -7.28270352e-01
-9.35585320e-01 -1.60209283e-01 -8.31558228e-01 -1.93717703e-01
-5.37742972e-01 6.55423254e-02 -5.27795076e-01 -1.34449422e-01
6.77323267e-02 3.73546839e-01 3.78883898e-01 -7.27547169e-01
2.43374735e-01 1.19533217e+00 5.88001311e-01 -2.29766533e-01
3.79982024e-01 2.96384901e-01 -3.67968440e-01 -9.97020841e-01
-5.05151272e-01 -3.90306115e-01 -5.44778049e-01 -2.69127101e-01
7.14239597e-01 -7.44343221e-01 -1.21915293e+00 1.32221246e+00
-9.64658320e-01 -6.20581090e-01 -1.61309242e-02 2.32541397e-01
-6.53487265e-01 5.98063648e-01 -4.49111402e-01 -8.71347368e-01
-2.70302385e-01 -1.48169744e+00 1.09933245e+00 2.65064895e-01
4.51586992e-02 -8.71572018e-01 -1.36862531e-01 5.51533282e-01
7.68457174e-01 -2.77795255e-01 7.47460723e-01 -5.03541946e-01
-2.62078881e-01 4.63465638e-02 -5.53305686e-01 4.95121032e-01
3.58144075e-01 1.23508751e-01 -1.55481875e+00 -1.73863858e-01
-3.24326158e-01 -4.57251549e-01 1.04118490e+00 5.07871807e-01
1.24793816e+00 -2.15494037e-01 -3.32572639e-01 6.22641623e-01
1.06464541e+00 5.52221164e-02 8.66104245e-01 2.39252597e-02
4.96454805e-01 7.66416728e-01 9.92985442e-02 1.06164522e-01
5.86370945e-01 9.93048191e-01 3.99892658e-01 -2.85195202e-01
-8.02644610e-01 -3.99020880e-01 4.96661693e-01 8.29728961e-01
-9.01010111e-02 -2.23681346e-01 -9.58164871e-01 5.67988634e-01
-1.43893611e+00 -9.32866812e-01 2.05946431e-01 2.30711603e+00
9.49829578e-01 8.52834955e-02 4.85797435e-01 4.22520399e-01
6.86104774e-01 4.02655303e-01 -5.85415065e-01 -1.48098618e-01
-1.69922173e-01 8.47470537e-02 1.71650842e-01 7.45219052e-01
-1.03115714e+00 1.12049866e+00 7.16688633e+00 1.03456700e+00
-1.64823854e+00 -4.24564779e-02 5.37735224e-01 -6.81314319e-02
6.77809864e-02 -5.16943157e-01 -9.59830582e-01 6.24109805e-01
9.38591540e-01 1.28368651e-02 4.94990289e-01 5.62549472e-01
4.15199429e-01 -2.18252584e-01 -8.27962875e-01 1.32873273e+00
3.19295347e-01 -1.27613664e+00 -2.45730385e-01 9.65832323e-02
2.52259612e-01 2.76045144e-01 2.55854934e-01 3.06889594e-01
-3.38345051e-01 -1.13763809e+00 6.95893347e-01 4.80202019e-01
1.20638120e+00 -5.33484280e-01 7.16054678e-01 3.44687611e-01
-1.22312868e+00 3.98091190e-02 -2.17111230e-01 1.59720570e-01
1.18637308e-01 2.31001407e-01 -6.94728076e-01 1.74841061e-01
7.38218009e-01 6.94265485e-01 -5.19265711e-01 1.18830943e+00
-2.56139398e-01 6.84351623e-01 -3.70146155e-01 -2.28751168e-01
-2.20758080e-01 3.52600187e-01 4.27251756e-01 1.23226881e+00
1.11905187e-02 -4.18225795e-01 -3.97388190e-01 6.20085061e-01
-2.27791592e-01 3.09026450e-01 -4.81474906e-01 6.61367476e-02
2.40655005e-01 1.19994211e+00 -2.81659633e-01 -3.28749746e-01
-5.63452840e-01 6.26209915e-01 2.27627322e-01 3.82035494e-01
-8.68594527e-01 -4.80237097e-01 9.36216235e-01 3.35418850e-01
2.85603762e-01 -7.86465555e-02 -1.25805855e-01 -1.10740077e+00
1.88880056e-01 -1.05507469e+00 -2.06451476e-01 -9.20031786e-01
-9.34822798e-01 5.62675476e-01 -1.22197971e-01 -8.39701355e-01
-4.16762799e-01 -6.32237256e-01 -6.23328865e-01 8.76434684e-01
-2.05884910e+00 -1.04604912e+00 -3.59002650e-01 5.53355455e-01
8.11347008e-01 -3.00264567e-01 7.31232703e-01 3.65894914e-01
-6.91403508e-01 1.15717721e+00 -1.10809416e-01 1.57185808e-01
9.61940110e-01 -9.22826648e-01 2.84033686e-01 6.53333724e-01
-1.28171053e-02 3.50114793e-01 5.28931081e-01 -1.16959333e-01
-1.11105704e+00 -3.46346647e-01 1.24774778e+00 -2.91483849e-01
5.97191691e-01 -6.20023251e-01 -9.39212263e-01 2.08408102e-01
5.63557088e-01 -2.08527483e-02 3.74006420e-01 6.38842732e-02
-3.03379774e-01 -1.78885326e-01 -1.13885939e+00 6.87287092e-01
1.19254780e+00 -6.22924566e-01 -5.25282800e-01 1.28574548e-02
7.17274129e-01 -2.63526797e-01 -4.86730129e-01 5.23399591e-01
9.38298047e-01 -1.17217958e+00 1.08149278e+00 -4.91589427e-01
1.97403789e-01 1.35536967e-02 4.68585938e-02 -9.74292278e-01
2.27630839e-01 -8.86088252e-01 -8.84822831e-02 1.60322702e+00
3.76391590e-01 -9.45633113e-01 7.11112142e-01 4.71129447e-01
-1.68084726e-02 -1.11691058e+00 -9.29185927e-01 -6.18924916e-01
2.37976685e-01 -4.35123980e-01 5.61889946e-01 4.86041933e-01
2.05762625e-01 2.34073371e-01 -4.81237084e-01 -1.49085119e-01
4.05077517e-01 -1.79922119e-01 8.39583397e-01 -1.14678907e+00
-1.91500504e-02 -7.50269592e-01 -2.94601202e-01 -1.49085343e+00
5.05693257e-01 -6.53542638e-01 2.58109599e-01 -1.56006289e+00
3.47964227e-01 -2.55050063e-01 -2.69746989e-01 7.50068724e-01
-3.03218752e-01 3.60958844e-01 2.38948613e-01 1.50195599e-01
-3.60711068e-01 4.87218589e-01 1.16091073e+00 -1.65518746e-01
-9.89124626e-02 3.31053376e-01 -7.58746684e-01 8.08024108e-01
9.28960681e-01 -8.77499394e-03 -3.91354024e-01 -3.40546489e-01
-1.68988109e-01 -5.40474243e-02 3.60857159e-01 -9.26859319e-01
2.52215832e-01 2.60228902e-01 2.99156718e-02 -6.50454462e-01
5.36999702e-01 -3.76314580e-01 -4.71912056e-01 1.26245841e-01
-3.36563975e-01 -2.44110912e-01 3.73771608e-01 1.77909851e-01
-1.75734773e-01 -2.13194713e-01 1.08012676e+00 3.31925422e-01
-7.64908969e-01 1.98929355e-01 -2.91483462e-01 1.16721459e-01
7.33248353e-01 -2.33738095e-01 -3.97309989e-01 -7.05987334e-01
-5.58163881e-01 -1.06578030e-01 4.77925003e-01 5.80580056e-01
4.09737110e-01 -9.99288738e-01 -5.53830683e-01 4.78478789e-01
-8.42512250e-02 -3.12788844e-01 1.34240657e-01 1.11882865e+00
-2.79359426e-02 6.88528061e-01 -1.81417808e-01 -6.69589579e-01
-1.52149904e+00 5.40701866e-01 5.07452667e-01 8.84928331e-02
-6.26307786e-01 8.49092185e-01 -6.01714887e-02 -2.24807009e-01
8.00065637e-01 -4.89144981e-01 -3.80379915e-01 2.03588590e-01
6.56329274e-01 4.20633763e-01 4.41044122e-02 -7.92483687e-01
-4.95999902e-01 9.61927593e-01 2.17187539e-01 -4.73432206e-02
1.11195505e+00 -3.31828058e-01 2.89263725e-01 2.81408429e-01
1.30134845e+00 5.67576647e-01 -1.47515130e+00 6.30791187e-02
-1.76575631e-01 -2.24157035e-01 -5.29792942e-02 -9.94008899e-01
-1.18293059e+00 1.37082875e+00 5.95900416e-01 1.87253430e-01
1.20198786e+00 2.61648595e-01 7.03535378e-01 3.24125984e-03
1.27157345e-01 -8.48954380e-01 -2.57151965e-02 3.18151653e-01
8.98819089e-01 -1.56926095e+00 -2.93998897e-01 -5.21940470e-01
-5.59073508e-01 9.08645391e-01 2.20979720e-01 4.20760512e-01
5.80512166e-01 4.99956280e-01 4.73728031e-01 3.99387330e-01
-6.50364280e-01 -5.46027482e-01 3.51918727e-01 8.85917127e-01
4.99971777e-01 -1.75516486e-01 -1.47501916e-01 4.36757296e-01
-1.32633179e-01 2.91047692e-01 9.59073454e-02 2.05253661e-01
-3.85971576e-01 -1.22725201e+00 -4.22397703e-02 1.65266052e-01
-5.84112406e-01 -1.64432183e-01 -2.70229220e-01 1.05671918e+00
-1.96438916e-02 9.84055400e-01 -6.75536841e-02 -2.86472738e-01
2.23502651e-01 4.58114058e-01 3.11214030e-01 -2.58607119e-02
-5.38467467e-02 1.79569304e-01 5.94878197e-02 -5.63819945e-01
-5.08270144e-01 -7.82391191e-01 -1.11955786e+00 -4.02479827e-01
-3.54842216e-01 -3.17394972e-01 9.04661953e-01 9.21902001e-01
4.90096480e-01 2.59964556e-01 3.27521831e-01 -1.15654373e+00
-6.40027165e-01 -1.09918880e+00 -2.83347189e-01 1.89538494e-01
7.11270630e-01 -7.88406909e-01 -7.26600945e-01 -2.40886044e-02]
|
[14.309409141540527, 4.987491607666016]
|
619c0352-7e88-4812-88e9-2fc09b44d334
|
spotr-spatio-temporal-pose-transformers-for
|
2303.06277
| null |
https://arxiv.org/abs/2303.06277v1
|
https://arxiv.org/pdf/2303.06277v1.pdf
|
SPOTR: Spatio-temporal Pose Transformers for Human Motion Prediction
|
3D human motion prediction is a research area of high significance and a challenge in computer vision. It is useful for the design of many applications including robotics and autonomous driving. Traditionally, autogregressive models have been used to predict human motion. However, these models have high computation needs and error accumulation that make it difficult to use them for realtime applications. In this paper, we present a non-autogressive model for human motion prediction. We focus on learning spatio-temporal representations non-autoregressively for generation of plausible future motions. We propose a novel architecture that leverages the recently proposed Transformers. Human motion involves complex spatio-temporal dynamics with joints affecting the position and rotation of each other even though they are not connected directly. The proposed model extracts these dynamics using both convolutions and the self-attention mechanism. Using specialized spatial and temporal self-attention to augment the features extracted through convolution allows our model to generate spatio-temporally coherent predictions in parallel independent of the activity. Our contributions are threefold: (i) we frame human motion prediction as a sequence-to-sequence problem and propose a non-autoregressive Transformer to forecast a sequence of poses in parallel; (ii) our method is activity agnostic; (iii) we show that despite its simplicity, our approach is able to make accurate predictions, achieving better or comparable results compared to the state-of-the-art on two public datasets, with far fewer parameters and much faster inference.
|
['Misha Sra', 'Avinash Ajit Nargund']
|
2023-03-11
| null | null | null | null |
['motion-prediction']
|
['computer-vision']
|
[ 9.96419564e-02 2.11513881e-02 -1.14056304e-01 -7.11235255e-02
-2.54617214e-01 -3.31069440e-01 1.02439630e+00 -4.61923450e-01
-4.49558914e-01 4.26254183e-01 5.91664910e-01 -1.35098353e-01
1.67701095e-01 -6.20393157e-01 -9.56808865e-01 -5.56332946e-01
-1.81359798e-01 3.84295493e-01 7.51369774e-01 -4.76393789e-01
7.87658393e-02 6.49004579e-01 -1.64326537e+00 3.59130979e-01
3.49180579e-01 7.42570221e-01 3.91229331e-01 1.09165001e+00
4.61666882e-01 1.34100676e+00 -4.75512892e-02 -6.76895976e-02
2.86380470e-01 -3.70587945e-01 -9.59781766e-01 -4.73221391e-02
2.90437400e-01 -3.76909524e-01 -8.88255775e-01 3.59342784e-01
3.79459530e-01 5.87793112e-01 6.58089221e-01 -1.13920343e+00
-2.42146254e-01 1.01186253e-01 -4.29284006e-01 3.20458710e-01
2.49493569e-01 5.61819971e-01 9.22929287e-01 -7.50866473e-01
7.48139381e-01 1.28966463e+00 7.68392622e-01 7.10418820e-01
-1.07429695e+00 -3.98455501e-01 2.62460023e-01 6.93091571e-01
-1.08040893e+00 -4.41441417e-01 6.82005167e-01 -6.06252730e-01
1.54431069e+00 1.30564481e-01 1.05268419e+00 1.42816496e+00
5.01174092e-01 1.11821520e+00 3.95952016e-01 -1.25827685e-01
1.26369238e-01 -6.19879723e-01 -3.54196280e-01 6.98791087e-01
-3.91973466e-01 3.85413915e-01 -6.92818105e-01 9.68044251e-02
1.00005150e+00 7.60599738e-03 -4.93700020e-02 -4.12824631e-01
-1.63687479e+00 6.66952193e-01 5.46764970e-01 1.71257675e-01
-7.26602077e-01 8.21281254e-01 2.55021423e-01 -2.66392440e-01
2.37127721e-01 2.10198656e-01 -4.14388806e-01 -4.98052031e-01
-9.56375718e-01 7.15556264e-01 4.18112457e-01 7.33695030e-01
4.98649389e-01 2.21112847e-01 -2.33883157e-01 3.46680403e-01
2.38723502e-01 4.04503495e-01 7.35785007e-01 -1.21350205e+00
3.22730899e-01 1.67621762e-01 3.73309433e-01 -1.25183034e+00
-7.58430898e-01 -1.72112435e-01 -8.77931356e-01 2.07829431e-01
3.77303928e-01 1.08093303e-02 -8.45373213e-01 1.74679816e+00
2.72933245e-01 8.09838355e-01 -8.86637941e-02 1.05711484e+00
3.35503280e-01 9.44384098e-01 2.80154616e-01 1.64597765e-01
1.21968234e+00 -1.47421038e+00 -4.99299228e-01 -5.78248799e-01
5.38308263e-01 -5.02540290e-01 8.10964763e-01 2.01972261e-01
-1.07053447e+00 -8.95396650e-01 -9.56446290e-01 -4.04583246e-01
-4.65936493e-03 1.75944641e-01 6.83881164e-01 4.08897139e-02
-1.06079519e+00 8.43868852e-01 -1.52199912e+00 -4.56699580e-01
1.63053423e-01 3.12513918e-01 -3.97048652e-01 2.16852114e-01
-9.89318430e-01 1.09965861e+00 2.14562356e-01 9.05166864e-02
-7.21438408e-01 -4.01117295e-01 -1.08759904e+00 -8.74198005e-02
1.73721775e-01 -1.20788097e+00 1.35275936e+00 -7.83955216e-01
-1.79632962e+00 3.50893229e-01 -5.56960642e-01 -9.34679687e-01
7.88505316e-01 -6.16957664e-01 -1.71714753e-01 1.10593870e-01
9.63328630e-02 1.19302082e+00 7.87707210e-01 -5.89803755e-01
-7.64759541e-01 -9.25499797e-02 -2.24968836e-01 3.31558734e-01
1.21341065e-01 -2.36689255e-01 -6.74841881e-01 -9.18352067e-01
4.79899235e-02 -1.33794105e+00 -6.63266063e-01 8.85058418e-02
-1.56804711e-01 -2.46264413e-01 8.22115004e-01 -6.82117581e-01
1.05641937e+00 -1.79540539e+00 4.19358999e-01 -1.79062903e-01
1.47147980e-02 3.54954958e-01 -1.32012904e-01 3.01529020e-01
4.17083986e-02 -2.34060958e-01 -2.15255931e-01 -5.07534206e-01
-1.11637853e-01 2.96368569e-01 -5.41523814e-01 4.54788148e-01
5.37757516e-01 1.29755962e+00 -1.03992856e+00 -2.60102183e-01
6.54770374e-01 6.86654389e-01 -6.79513037e-01 2.01062053e-01
-3.09739590e-01 6.92513525e-01 -4.06169116e-01 1.73715770e-01
-3.29927988e-02 -2.80889869e-01 2.97138747e-02 -2.23633274e-01
-1.82393357e-01 3.27785790e-01 -8.91638458e-01 1.67147517e+00
-2.94161052e-01 7.67586291e-01 -6.52691305e-01 -1.00844443e+00
6.45533264e-01 3.07562143e-01 6.14955902e-01 -5.12365162e-01
2.89928745e-02 1.08451797e-02 -6.46506548e-02 -5.87594748e-01
7.39239573e-01 4.28672619e-02 -2.52970010e-02 2.77924418e-01
-7.27906451e-02 -3.54948603e-02 -1.22086024e-02 -4.45333943e-02
1.29486585e+00 9.06614125e-01 4.15107369e-01 1.06960140e-01
5.03030956e-01 2.20266342e-01 6.29848063e-01 5.13132274e-01
-3.05889189e-01 7.43007362e-01 1.43944561e-01 -9.55816746e-01
-1.28292310e+00 -8.94513309e-01 6.76426351e-01 1.01485705e+00
1.51727647e-01 -2.62212574e-01 -3.96910638e-01 -3.62573624e-01
-2.26118863e-01 5.87827146e-01 -7.11699724e-01 -1.65774241e-01
-1.21097827e+00 -4.82386410e-01 5.55234730e-01 1.02016032e+00
4.01215434e-01 -1.25710690e+00 -1.17354703e+00 4.28051502e-01
-5.27855456e-01 -1.29992163e+00 -5.97533822e-01 -2.00762540e-01
-8.56904209e-01 -8.35018277e-01 -8.24644446e-01 -5.55052757e-01
2.64246494e-01 3.06387037e-01 9.57898736e-01 -2.04828337e-01
-2.22558722e-01 3.45070869e-01 -1.38962299e-01 -1.77760348e-01
-2.10713968e-01 5.10536060e-02 1.38497710e-01 9.30398330e-02
1.97958037e-01 -6.73541427e-01 -7.86082327e-01 4.23708051e-01
-5.62554061e-01 4.73063141e-01 6.31281972e-01 8.43203485e-01
6.05763435e-01 -3.91767412e-01 3.71382326e-01 -3.97591352e-01
8.91308561e-02 -5.47782481e-01 -4.26872075e-01 -2.07674369e-01
-1.58337310e-01 2.77745932e-01 5.78570604e-01 -5.91680348e-01
-1.05977750e+00 7.72090554e-01 -1.94499090e-01 -6.85640514e-01
-2.00614139e-01 1.33519024e-01 2.76783735e-01 1.84984565e-01
6.66900814e-01 3.35640728e-01 9.22097787e-02 -2.94154555e-01
4.54508752e-01 -9.10219103e-02 9.65723813e-01 -1.56364262e-01
7.09831417e-01 8.64554703e-01 3.34009975e-01 -8.86878133e-01
-4.19783264e-01 -5.71239650e-01 -8.45313132e-01 -3.11183333e-01
1.10806251e+00 -1.04649615e+00 -8.75487030e-01 5.78517318e-01
-1.37535417e+00 -6.99017167e-01 -1.44474268e-01 6.20969892e-01
-1.13699055e+00 5.27428269e-01 -7.13423550e-01 -8.23197663e-01
-1.79183513e-01 -1.05001986e+00 1.12166739e+00 -1.11022498e-02
-7.56824970e-01 -8.42790186e-01 1.69224486e-01 3.34353000e-01
2.29969725e-01 3.62584502e-01 4.24244642e-01 -2.53802389e-01
-8.35404873e-01 -8.25172216e-02 1.78339109e-01 -2.28686973e-01
-1.68444350e-01 -1.98697746e-02 -8.09673965e-01 5.89664578e-02
-4.17494535e-01 -1.38900429e-01 1.04942632e+00 5.80561578e-01
9.87060368e-01 -3.12757283e-01 -5.60148418e-01 5.45011878e-01
7.63716817e-01 -5.06371818e-02 9.34616923e-01 4.30960119e-01
8.59557629e-01 8.03703189e-01 6.82649434e-01 4.35074061e-01
6.65192664e-01 1.09091949e+00 3.56476784e-01 1.96504667e-01
-2.49735877e-01 -5.31934798e-01 6.38644099e-01 5.09604692e-01
-5.70082009e-01 -2.22922042e-01 -9.41172242e-01 9.08288896e-01
-2.49027824e+00 -1.42276883e+00 -3.20310533e-01 1.85253286e+00
3.37962598e-01 2.94508096e-02 4.29381549e-01 1.25205517e-03
3.23060691e-01 2.70159602e-01 -6.00396574e-01 -1.29462630e-01
-2.98193041e-02 2.25869134e-01 4.21178907e-01 5.01888514e-01
-1.35377085e+00 1.26154900e+00 6.21927929e+00 4.94702876e-01
-1.24469256e+00 -7.92381465e-02 4.55364853e-01 -2.77207077e-01
4.70981412e-02 -1.07984334e-01 -6.48023784e-01 2.60525167e-01
9.34415162e-01 1.02201901e-01 1.67303219e-01 7.09866524e-01
5.58308065e-01 -2.10779719e-02 -1.05276680e+00 8.42232287e-01
-6.22328930e-02 -1.63314366e+00 -6.82613999e-03 2.08377577e-02
6.61954522e-01 2.26154611e-01 -1.66872796e-02 1.50410905e-01
3.77168715e-01 -1.13172948e+00 1.10180235e+00 8.34711611e-01
1.36790723e-01 -6.91762984e-01 5.38425326e-01 6.06574833e-01
-1.40204430e+00 -2.15932325e-01 -8.64818096e-02 -4.82350320e-01
6.72859371e-01 7.82458037e-02 -8.66979837e-01 3.73167038e-01
7.59580374e-01 1.03113890e+00 -2.93705076e-01 8.14490676e-01
-2.90842354e-01 5.08499444e-01 -2.29061753e-01 -2.79697143e-02
4.18289483e-01 1.55879334e-01 6.01937473e-01 1.17549646e+00
5.33457160e-01 1.72460765e-01 1.14085414e-01 6.96958482e-01
4.32122797e-01 -2.63799012e-01 -6.12257600e-01 1.74703166e-01
1.87009841e-01 8.68865967e-01 -5.38500786e-01 -3.84170502e-01
-3.23237449e-01 1.23947334e+00 2.93656737e-01 3.08689177e-01
-9.67737794e-01 6.81697652e-02 8.74035597e-01 2.17764780e-01
6.58322394e-01 -7.01070130e-01 -2.30086654e-01 -1.00921416e+00
3.97415124e-02 -5.10359585e-01 2.36023515e-01 -7.83763885e-01
-8.73138249e-01 5.65895259e-01 -5.28207421e-02 -1.39746535e+00
-1.03059268e+00 -5.33035874e-01 -5.15137851e-01 7.41467476e-01
-1.22217679e+00 -1.44722426e+00 -2.28152946e-01 4.49369729e-01
7.47687995e-01 -2.18611024e-03 6.63839042e-01 -3.37602347e-02
-2.52249539e-01 2.35299915e-01 -4.82428670e-01 9.60841179e-02
5.02338111e-01 -8.85007501e-01 1.18505299e+00 9.93163466e-01
2.60840297e-01 2.97286212e-01 8.27257276e-01 -7.38792539e-01
-1.29847074e+00 -1.35365236e+00 1.21210825e+00 -6.88891828e-01
5.93685389e-01 -5.02820984e-02 -7.54134953e-01 9.13072348e-01
-3.29774767e-02 2.52179891e-01 2.02588663e-01 -4.39491481e-01
-1.50029853e-01 2.27921158e-01 -5.85517049e-01 1.00066519e+00
1.23652875e+00 -2.79460222e-01 -6.19621575e-01 5.26054166e-02
6.54960990e-01 -4.10406530e-01 -3.85391653e-01 4.96152371e-01
9.06048775e-01 -1.06490207e+00 1.19550204e+00 -6.44431353e-01
6.55694544e-01 -5.36373436e-01 1.41708143e-02 -9.46161866e-01
-7.31776357e-01 -7.18409657e-01 -4.62049007e-01 4.24228758e-01
2.19368309e-01 -3.22123915e-01 1.02358723e+00 5.07629454e-01
-2.58810550e-01 -8.40345979e-01 -9.26938057e-01 -6.20986760e-01
-2.09774226e-01 -6.53102577e-01 2.42596343e-01 6.12855315e-01
-7.06957653e-02 3.73788804e-01 -9.24097240e-01 1.75322250e-01
2.39262938e-01 -9.91091505e-02 1.10358548e+00 -9.09077525e-01
-5.41141629e-01 -4.75467592e-01 -6.80713058e-01 -1.65458131e+00
2.57905126e-01 -4.10706282e-01 3.90635937e-01 -1.37650919e+00
-1.26924664e-01 3.61758657e-02 1.49620056e-01 5.95848799e-01
-5.92380166e-02 3.11932027e-01 3.26342404e-01 4.12753850e-01
-4.09022331e-01 7.54671454e-01 1.17021024e+00 -6.97419867e-02
-2.48956636e-01 1.88077793e-01 -2.75107529e-02 9.73627746e-01
5.58409452e-01 -1.87700421e-01 -4.11125392e-01 -3.94603699e-01
6.60519451e-02 2.34904036e-01 7.95471072e-01 -1.36700499e+00
4.47585583e-01 -2.93200463e-01 5.68713248e-01 -8.38383853e-01
8.84520888e-01 -4.50783283e-01 3.50815862e-01 7.18757093e-01
-2.16210619e-01 3.45581561e-01 1.67050600e-01 9.30232286e-01
-5.71351685e-02 3.57464671e-01 5.53539693e-01 -2.64755934e-01
-1.23491621e+00 3.74854267e-01 -8.20930719e-01 -4.04942811e-01
1.04022276e+00 -3.17624122e-01 8.18813145e-02 -7.45271742e-01
-7.63455868e-01 8.79891142e-02 3.82362068e-01 7.29580224e-01
6.38267815e-01 -1.32861590e+00 -5.92374980e-01 -2.94045079e-02
6.01081513e-02 4.00842540e-02 4.82215464e-01 9.59685862e-01
-6.97851539e-01 6.24224544e-01 -2.89381742e-01 -8.64213765e-01
-1.04966700e+00 4.39052671e-01 2.90260017e-01 -3.52669746e-01
-9.06688094e-01 6.78277791e-01 3.06549728e-01 -2.68061936e-01
-7.01303314e-03 -3.60652894e-01 -2.78180122e-01 -2.83443630e-01
6.11569047e-01 5.22872686e-01 -2.38307849e-01 -1.18122494e+00
-3.29499096e-01 5.82576275e-01 3.19149762e-01 -3.60747278e-01
1.31131065e+00 -1.71037748e-01 2.99115002e-01 4.83560205e-01
8.58330905e-01 -3.58291060e-01 -1.72834194e+00 -2.89151706e-02
6.64928555e-02 -2.45836779e-01 -3.19566339e-01 -3.79048973e-01
-7.24975228e-01 1.09707642e+00 3.47173065e-01 -2.69006550e-01
8.83365750e-01 -6.20037802e-02 1.14428353e+00 4.19547319e-01
4.09266174e-01 -1.05170047e+00 3.79277557e-01 9.08855736e-01
8.75966966e-01 -1.00230825e+00 -1.66311145e-01 -3.28654200e-01
-9.08742547e-01 1.06123078e+00 6.85337484e-01 -3.50349635e-01
3.99841249e-01 -7.30487034e-02 -9.35181156e-02 8.75614285e-02
-1.08040082e+00 -2.43244529e-01 5.85999966e-01 6.54048622e-01
3.91219586e-01 -6.46200404e-02 1.97130684e-02 5.29019833e-01
-2.81339496e-01 1.38876230e-01 3.50626707e-01 8.05197358e-01
-3.63039255e-01 -8.50406349e-01 -2.18086958e-01 9.34625603e-03
-1.55870751e-01 1.88955024e-01 -1.63506255e-01 6.38572097e-01
2.36369222e-02 7.88155913e-01 2.51312375e-01 -6.26780212e-01
2.45201156e-01 2.74548922e-02 4.44359958e-01 -2.89108902e-01
-2.91333735e-01 3.33193243e-02 1.19187996e-01 -1.09764099e+00
-6.12124085e-01 -7.91942596e-01 -1.39544761e+00 -2.68487334e-01
2.78356910e-01 -4.48512316e-01 3.19044590e-01 1.08266366e+00
7.50362992e-01 5.87110281e-01 2.53344297e-01 -1.54947948e+00
-3.53585482e-01 -7.67915130e-01 3.13793570e-02 4.45816964e-01
3.87821406e-01 -8.09277296e-01 6.68307021e-02 4.48995858e-01]
|
[7.32735013961792, -0.1318986564874649]
|
114bf613-3f71-4cc1-a627-e2b1c31e5e5f
|
dual-embeddings-and-metrics-for-relational
| null | null |
https://aclanthology.org/W17-6924
|
https://aclanthology.org/W17-6924.pdf
|
Dual Embeddings and Metrics for Relational Similarity
| null |
['D. Li', 'Douglas Summers-Stay', 'an']
|
2017-01-01
| null | null | null |
ws-2017-1
|
['learning-word-embeddings']
|
['methodology']
|
[-8.63703638e-02 1.71006292e-01 -6.22772932e-01 -4.08054382e-01
-8.41685571e-03 -9.08429027e-01 6.55310392e-01 -6.53472245e-01
-2.85945535e-01 1.06888819e+00 -4.63127941e-02 -1.01159286e+00
-3.91567826e-01 -9.63214397e-01 -4.95059669e-01 -6.31337762e-01
-9.79754329e-01 7.25764990e-01 3.30370307e-01 -6.93831444e-01
7.03166842e-01 7.88774848e-01 -1.68942046e+00 7.18545914e-01
7.04417467e-01 8.52217197e-01 2.49141872e-01 1.14950800e+00
-1.95044339e-01 1.55633950e+00 -7.48382092e-01 -5.46825826e-01
3.13719302e-01 -1.23176083e-01 -7.22945035e-01 -1.01074085e-01
9.28529128e-02 -8.59008506e-02 -2.09758401e-01 9.22211111e-01
5.37373662e-01 4.49454933e-02 1.08379531e+00 -1.42548037e+00
-5.91619551e-01 6.10313773e-01 -4.01565880e-02 1.21627934e-01
1.03678203e+00 -5.39447069e-01 1.19919395e+00 -1.13026452e+00
7.20913768e-01 1.26888943e+00 8.66221786e-01 5.44149756e-01
-1.22286928e+00 -1.94712028e-01 -3.26822817e-01 -9.51717794e-02
-1.46558487e+00 -3.25250506e-01 4.25783843e-02 -2.08119690e-01
1.66093647e+00 1.26596653e+00 1.20609856e+00 1.01401424e+00
1.26658809e+00 8.34431887e-01 1.04267764e+00 -5.13792276e-01
3.35295945e-01 3.66983831e-01 1.54683650e-01 6.33519173e-01
8.40953708e-01 5.26628852e-01 -7.06372619e-01 -9.13127720e-01
9.33553874e-01 -2.94925272e-01 1.71355158e-01 -5.05680561e-01
-9.05919552e-01 6.91228509e-01 1.78732842e-01 3.83959889e-01
-1.39880210e-01 9.89067405e-02 1.26390755e-01 5.30987144e-01
-2.58292928e-02 6.47037446e-01 -9.11868811e-01 -1.33165747e-01
-8.71728659e-01 5.10332465e-01 1.25398111e+00 1.52653182e+00
1.24482810e-01 2.94908643e-01 -9.34252143e-02 3.17179203e-01
8.92314315e-01 1.01808000e+00 4.28362608e-01 -1.36146402e+00
-6.87414408e-02 1.72361732e-01 5.01781464e-01 -8.52631688e-01
-6.33224547e-01 -9.64177120e-03 -8.93263519e-01 4.49267089e-01
3.49161088e-01 4.57367361e-01 -8.02827001e-01 5.07305264e-01
4.33481112e-02 -2.34125629e-01 4.53833073e-01 5.55570945e-02
4.99930978e-01 3.76208365e-01 -1.34477139e-01 -5.73289394e-01
1.06082785e+00 -1.36716676e+00 -1.35299087e+00 2.33215362e-01
9.05734658e-01 -1.07320261e+00 4.35900748e-01 5.33875942e-01
-1.55548143e+00 -1.37560293e-01 -1.08699942e+00 1.78573877e-01
-7.27255583e-01 -3.14239264e-01 8.57801437e-01 1.43120694e+00
-1.60129595e+00 9.73287821e-01 -4.91727620e-01 6.59165755e-02
1.20568443e-02 8.24621081e-01 -2.64718989e-03 4.62812334e-01
-1.33193445e+00 1.08501506e+00 2.22979754e-01 -1.21242590e-01
-1.65216476e-01 -2.13068098e-01 -8.23704481e-01 -5.63443303e-01
-4.78693932e-01 -5.29636025e-01 1.44139910e+00 -2.59346128e-01
-1.65295815e+00 9.71794367e-01 -1.42069459e-01 -1.97814897e-01
6.14786744e-01 -1.28011424e-02 -8.31891418e-01 2.42498964e-01
-1.89849049e-01 5.76383233e-01 9.28263724e-01 -1.35132408e+00
-7.59897232e-01 -1.67359829e-01 -1.23336017e-01 2.66287565e-01
-1.25510961e-01 1.89734384e-01 2.11616129e-01 -1.12999000e-01
3.27147305e-01 -7.20919967e-01 -2.53068686e-01 -5.32041907e-01
-1.46512717e-01 -7.10518599e-01 7.70373225e-01 -4.51523662e-01
1.83705616e+00 -1.67618537e+00 -2.06720144e-01 3.98590982e-01
3.57815564e-01 -1.24705513e-03 2.40583986e-01 1.08380008e+00
-2.76906848e-01 7.32199550e-01 4.11965609e-01 -1.10722095e-01
2.24991128e-01 4.85861301e-01 -4.16602850e-01 3.05609167e-01
-1.29282743e-01 1.15307164e+00 -1.16605783e+00 -5.23096442e-01
4.11106765e-01 9.42391157e-02 -4.58732933e-01 4.79237735e-01
2.99364805e-01 1.47170946e-01 -3.56553018e-01 1.39399457e+00
1.15709066e+00 -1.31984919e-01 1.45911396e-01 5.30878425e-01
-3.79135728e-01 3.55090618e-01 -6.96863770e-01 1.05554795e+00
7.08333924e-02 5.00173986e-01 1.02364108e-01 -7.94621468e-01
3.33247900e-01 8.61540735e-01 4.37155962e-01 -9.67555881e-01
-2.26398129e-02 6.92409754e-01 1.12803578e-01 -6.07703328e-01
7.58228302e-01 3.94563079e-02 -3.64872098e-01 6.40070081e-01
-2.37588286e-01 -6.59476995e-01 9.64643434e-02 3.08100313e-01
6.22585893e-01 -5.18246442e-02 5.62923312e-01 -1.05613089e+00
7.86340594e-01 -1.82965681e-01 -1.81299388e-01 1.03415680e+00
-3.09923887e-01 3.19085121e-01 2.99841821e-01 -6.65102363e-01
-6.45341039e-01 -1.12307119e+00 -4.89381433e-01 1.30636716e+00
3.24267983e-01 -4.39044595e-01 -9.54439282e-01 -2.49762803e-01
1.77620783e-01 6.89606130e-01 -5.90509653e-01 3.84124845e-01
-5.03739953e-01 -8.32535863e-01 7.39044368e-01 3.45434904e-01
-5.07752821e-02 -1.33414865e+00 -6.58416986e-01 1.25490099e-01
-2.20292807e-01 -6.63697243e-01 -6.23428151e-02 4.48765576e-01
-1.35989368e+00 -5.18594682e-01 -6.66252747e-02 -8.20914626e-01
5.87345481e-01 2.46782884e-01 1.27047324e+00 5.39230824e-01
-2.31483161e-01 4.26904231e-01 -1.21292919e-01 -4.95818377e-01
-4.59671497e-01 -8.00336525e-02 5.28869390e-01 -5.87835789e-01
5.19427478e-01 -2.50617653e-01 -7.29350567e-01 5.37953973e-01
-6.88540697e-01 1.62748516e-01 1.79803044e-01 1.04410267e+00
1.35816500e-01 -9.34035778e-02 1.22507080e-01 -6.38007045e-01
8.72274399e-01 -1.69219792e-01 -3.78732830e-01 5.77745810e-02
-6.77108407e-01 -3.74140263e-01 3.21430594e-01 -3.25342178e-01
-1.01981449e+00 -4.87835288e-01 -9.82677937e-02 2.45538145e-01
1.11353043e-02 -1.46784872e-01 6.47139177e-02 -5.24923325e-01
8.02199244e-01 9.25758183e-02 1.99174434e-02 -6.80815242e-03
3.01039815e-01 7.09525108e-01 -6.82967342e-03 -6.68678164e-01
8.44880998e-01 4.91470337e-01 7.98524171e-02 -9.57177758e-01
-1.52186140e-01 -2.60129690e-01 -9.51962709e-01 -6.54426932e-01
6.56643391e-01 -6.78531289e-01 -9.10833478e-01 3.91110867e-01
-9.38691139e-01 -3.38627815e-01 -3.91645581e-01 4.25431967e-01
-1.01278400e+00 2.75717527e-02 -3.90154392e-01 -1.27895141e+00
-5.10977268e-01 -1.02017939e+00 9.43384409e-01 5.30070923e-02
-5.10597289e-01 -1.26927447e+00 5.87685481e-02 2.71537274e-01
1.81734428e-01 -1.73075795e-01 6.90226793e-01 -2.38256708e-01
-4.24233019e-01 -1.53791070e-01 2.34436691e-02 -1.39755070e-01
1.70832314e-02 4.95917559e-01 -9.81751978e-01 -5.31145096e-01
6.65065646e-02 -1.92070693e-01 -1.08835101e-01 6.52520418e-01
5.91872573e-01 -2.29931593e-01 -8.56000841e-01 5.40386558e-01
1.38545322e+00 3.85070026e-01 5.32770038e-01 7.28214979e-01
1.41836226e-01 5.53460240e-01 9.17806149e-01 4.63203549e-01
1.30579369e-02 3.28798652e-01 2.40537539e-01 1.49327129e-01
1.11720070e-01 -1.54819340e-01 3.77893507e-01 1.16112018e+00
-8.18235934e-01 -2.69281328e-01 -5.07867396e-01 4.42987174e-01
-1.72482407e+00 -1.40330648e+00 -4.32368398e-01 6.90478683e-01
6.25676990e-01 1.56016424e-01 -1.48347050e-01 3.35214496e-01
4.99015123e-01 -2.03574806e-01 -1.19133167e-01 -1.06291151e+00
-1.43546045e-01 3.15233678e-01 7.37729073e-01 1.00061214e+00
-7.20721722e-01 1.03317809e+00 1.29781246e+01 1.02230716e+00
2.21112028e-01 1.03134915e-01 5.16071796e-01 3.48020852e-01
-4.36954498e-01 -4.56139445e-02 -1.04416132e+00 2.72933897e-02
1.38140702e+00 -4.30666685e-01 6.85999811e-01 5.44219851e-01
3.44648361e-01 -4.23268199e-01 -1.26188684e+00 5.26221812e-01
9.73738134e-02 -1.40886843e+00 -2.83300440e-04 6.85225725e-01
7.73699820e-01 -5.08050561e-01 6.22419357e-01 3.24184299e-01
6.09259963e-01 -1.14389277e+00 8.60300779e-01 2.53660440e-01
1.03040910e+00 -6.05088234e-01 5.67372203e-01 1.68872893e-01
-1.14389896e+00 -2.20873043e-01 -8.77727985e-01 -1.00755692e+00
3.93533185e-02 -1.81779593e-01 -4.29956943e-01 3.48861217e-01
9.58353162e-01 2.99398601e-01 -3.93658698e-01 9.95779395e-01
-4.78694476e-02 1.04875881e-02 -2.71853864e-01 -4.48467314e-01
4.83122796e-01 -3.54241252e-01 4.66730654e-01 1.00164843e+00
2.48499006e-01 3.51035744e-01 -9.84472036e-02 4.01770771e-01
5.45058846e-01 3.29446048e-02 -1.19659424e+00 -1.78908288e-01
2.83276141e-01 9.16795909e-01 -4.83487815e-01 -4.22520459e-01
-2.00212970e-01 8.62069130e-01 -3.55488248e-02 5.01107454e-01
-6.11489356e-01 -4.35615242e-01 9.72222984e-01 -1.27327025e-01
-1.14700586e-01 -3.48497719e-01 -6.23769283e-01 -7.30352640e-01
-5.89872956e-01 -4.54965204e-01 5.93606755e-02 -5.53365827e-01
-1.39813089e+00 5.79277515e-01 -2.27688253e-02 -1.40553558e+00
-6.99901402e-01 -1.27676582e+00 -4.76714373e-01 4.92853165e-01
-1.11898029e+00 -1.10984349e+00 2.50124663e-01 4.52870727e-01
1.64141744e-01 -5.34416080e-01 1.39563632e+00 3.57715860e-02
1.00637585e-01 9.24474537e-01 6.69434488e-01 -7.37814724e-01
5.56605101e-01 -1.27867436e+00 5.68737745e-01 -1.37897313e-01
-4.31265175e-01 9.05828118e-01 6.28349900e-01 -5.39804697e-01
-1.41196322e+00 -3.66917729e-01 1.08350635e+00 -9.83769417e-01
6.55218959e-01 -3.86345625e-01 4.23767231e-02 7.88592756e-01
7.15902448e-01 -6.18741751e-01 8.21781039e-01 -1.83753878e-01
1.80774391e-01 5.75296998e-01 -1.39248300e+00 6.12354755e-01
1.66275799e+00 -4.63594139e-01 -6.25784039e-01 7.60327101e-01
8.13696027e-01 -6.94087505e-01 -1.30082703e+00 3.34633321e-01
8.65424156e-01 -8.75409484e-01 1.61978090e+00 -1.32660246e+00
-4.43697497e-02 2.81152606e-01 -2.61993498e-01 -9.32519078e-01
-5.96193194e-01 -1.23518765e+00 -5.33532679e-01 -5.83747849e-02
5.96577883e-01 -1.13057327e+00 3.42365682e-01 8.74560475e-01
-2.82833427e-01 -6.42737269e-01 -1.06996536e+00 -1.32016802e+00
-3.35779637e-02 -1.45572275e-01 4.90409225e-01 7.63798356e-01
6.83744550e-01 1.09839931e-01 -6.36873543e-02 -1.00294888e-01
5.33176839e-01 9.55312885e-03 4.41501856e-01 -1.34294486e+00
3.86843324e-01 -5.75816095e-01 -3.07655483e-01 -9.45992947e-01
-8.85957032e-02 -8.12076271e-01 -6.53862000e-01 -1.28511906e+00
-8.32044985e-03 -1.91056758e-01 -1.11109078e-01 -1.65725678e-01
3.67937148e-01 2.13746816e-01 1.20859891e-02 1.03788137e-01
-3.71160030e-01 6.18435517e-02 1.29639816e+00 6.91750320e-05
-1.62315920e-01 4.85058486e-01 -4.81304944e-01 7.84440815e-01
8.58408585e-02 -2.99253196e-01 -6.78878546e-01 6.11881316e-02
6.69384480e-01 4.61409837e-02 3.24159935e-02 -7.42885649e-01
5.37211418e-01 -3.75702560e-01 4.78586555e-01 -1.32223868e+00
1.30741090e-01 -9.61415648e-01 6.95283338e-02 9.40189242e-01
2.74610907e-01 1.20424610e-02 8.66204947e-02 5.39500564e-02
-1.44871444e-01 -5.70943117e-01 9.21121240e-01 -4.07591403e-01
-4.92852688e-01 -5.20386267e-03 -1.03226590e+00 8.97834301e-02
9.92593169e-01 -7.84614205e-01 -3.59281451e-01 -4.20183957e-01
-8.29068601e-01 -1.95836127e-02 6.50830388e-01 3.03609259e-02
7.30431557e-01 -1.51530886e+00 -2.30721906e-01 7.18729138e-01
-3.22939813e-01 -3.74400020e-01 -1.70157343e-01 6.58265352e-01
-1.32361674e+00 1.02442718e+00 -5.41665435e-01 -4.55340147e-01
-1.14228773e+00 4.78126436e-01 4.28307921e-01 -2.41845414e-01
-2.15481281e-01 1.11954463e+00 2.71224789e-02 -8.23025763e-01
1.85185194e-01 -9.21545625e-02 -7.54407048e-01 3.55081353e-03
6.88606799e-01 1.05194807e+00 -2.91290224e-01 -6.04341030e-01
-4.56784427e-01 6.65885091e-01 2.32151806e-01 -2.87484169e-01
9.17833567e-01 -2.19243199e-01 -9.89108324e-01 4.28274393e-01
8.38715494e-01 -1.36269778e-01 -3.69319022e-02 4.16855574e-01
1.25943512e-01 -8.02164078e-01 -4.32406247e-01 -3.55811834e-01
-1.85641110e-01 5.54822803e-01 5.34874737e-01 8.99602413e-01
8.57008278e-01 -3.02566767e-01 8.18335712e-01 9.66778398e-01
5.72402716e-01 -1.68019545e+00 -2.13140488e-01 6.89524531e-01
9.12339568e-01 -9.28350806e-01 5.44190466e-01 -7.27165341e-01
-4.14997995e-01 1.32979155e+00 4.68304873e-01 -1.55325383e-01
1.27306652e+00 5.74917436e-01 1.14069022e-02 -3.70670199e-01
-9.44949508e-01 1.12705544e-01 3.75366658e-01 1.11147714e+00
5.06513238e-01 5.07374525e-01 -9.66500878e-01 3.21953118e-01
-7.28706717e-01 -2.34555230e-01 4.90474731e-01 1.41972518e+00
-6.43810987e-01 -1.20391917e+00 -7.23931909e-01 4.61561680e-01
-5.49773455e-01 -1.16372630e-01 -5.28106689e-01 8.46754074e-01
-5.76629937e-02 1.49448860e+00 -1.97535474e-03 -5.03491640e-01
4.11356747e-01 1.41089618e-01 7.21762300e-01 -1.23501487e-01
-9.04846430e-01 4.15413082e-01 3.84890139e-01 -1.23056793e+00
-8.58632207e-01 -1.05834293e+00 -1.40667629e+00 -1.19437599e+00
-5.12782812e-01 1.89310342e-01 3.83317530e-01 3.90289724e-01
-2.06836104e-01 2.85260603e-02 9.80917513e-01 -1.07949340e+00
-3.90341938e-01 -9.39418614e-01 -1.00026262e+00 -7.84516707e-02
2.89751232e-01 -8.00943017e-01 -7.83523321e-01 2.83909619e-01]
|
[-7.264651298522949, 3.6799516677856445]
|
2a65aedd-f8c9-47ad-a516-80a4681ee7e6
|
boosting-the-performance-of-transformer
|
2306.00708
| null |
https://arxiv.org/abs/2306.00708v1
|
https://arxiv.org/pdf/2306.00708v1.pdf
|
Boosting the Performance of Transformer Architectures for Semantic Textual Similarity
|
Semantic textual similarity is the task of estimating the similarity between the meaning of two texts. In this paper, we fine-tune transformer architectures for semantic textual similarity on the Semantic Textual Similarity Benchmark by tuning the model partially and then end-to-end. We experiment with BERT, RoBERTa, and DeBERTaV3 cross-encoders by approaching the problem as a binary classification task or a regression task. We combine the outputs of the transformer models and use handmade features as inputs for boosting algorithms. Due to worse test set results coupled with improvements on the validation set, we experiment with different dataset splits to further investigate this occurrence. We also provide an error analysis, focused on the edges of the prediction range.
|
['Vladimir Čeperić', 'Ivan Rep']
|
2023-06-01
| null | null | null | null |
['semantic-textual-similarity']
|
['natural-language-processing']
|
[ 3.56486082e-01 6.86096102e-02 6.29852861e-02 -6.43259287e-01
-8.07567894e-01 -5.92183173e-01 9.28263009e-01 3.50455225e-01
-5.81568956e-01 3.53818476e-01 5.65512002e-01 -3.80716056e-01
-6.28167018e-02 -5.00758171e-01 -4.04516906e-01 -2.27193430e-01
2.88749546e-01 6.91875637e-01 2.65509814e-01 -3.83059919e-01
3.54483902e-01 -2.62056708e-01 -1.31502748e+00 8.07700396e-01
7.01844096e-01 1.13107562e+00 4.55415696e-02 6.71759367e-01
-1.17593497e-01 7.88405240e-01 -3.98577094e-01 -8.45166922e-01
1.50246963e-01 -5.05772591e-01 -1.16703153e+00 -2.41421729e-01
5.28510392e-01 3.37903537e-02 -1.13846965e-01 7.52535462e-01
3.46650064e-01 1.63604051e-01 7.18657732e-01 -1.21483910e+00
-4.86323148e-01 9.40766871e-01 -2.10867912e-01 2.22266197e-01
3.05373073e-01 -3.78148034e-02 1.55302620e+00 -8.57517958e-01
3.87784004e-01 1.36044359e+00 9.07029390e-01 1.99620545e-01
-1.08949876e+00 -3.82168800e-01 -7.69734457e-02 4.93661731e-01
-1.05942154e+00 -6.14642680e-01 5.45704961e-01 -2.98447251e-01
1.13990533e+00 1.45061731e-01 1.83180094e-01 1.25548732e+00
-4.27392125e-02 6.30872071e-01 1.02504027e+00 -2.74028450e-01
2.75471270e-01 2.39683092e-01 2.75685519e-01 4.31799769e-01
-2.03049496e-01 9.97337103e-02 -4.34200406e-01 -2.74997175e-01
-6.33263588e-03 -1.68871373e-01 -2.02993095e-01 -3.89930367e-01
-1.11479902e+00 9.42343950e-01 7.43679523e-01 5.51347435e-01
-1.81092814e-01 7.64311701e-02 7.17187285e-01 7.05283880e-01
5.97104311e-01 6.23440266e-01 -6.72949255e-01 -2.60872960e-01
-8.46867561e-01 1.93560511e-01 8.42506409e-01 5.58947265e-01
3.74841958e-01 -3.85718077e-01 -4.66095507e-01 9.91781473e-01
-2.28761211e-02 5.71572455e-03 9.37438309e-01 -7.57656097e-01
5.34884155e-01 5.51760972e-01 -1.37105942e-01 -7.05040932e-01
-3.61996174e-01 -4.47565496e-01 -4.30617064e-01 -3.06907117e-01
3.97948146e-01 2.88153365e-02 -6.95457816e-01 1.68244243e+00
6.39331639e-02 1.20823897e-01 1.32818401e-01 1.00573778e+00
6.87211633e-01 3.52086663e-01 2.78590322e-01 3.25040340e-01
1.40990317e+00 -1.09104300e+00 -2.06027880e-01 -3.84559333e-01
9.81982946e-01 -7.34813988e-01 1.37199807e+00 -1.40055018e-02
-9.71084893e-01 -4.81127590e-01 -8.53776217e-01 -3.19980890e-01
-3.52359742e-01 2.47273207e-01 1.61108673e-01 3.29488605e-01
-8.22472751e-01 9.79341924e-01 -6.72480762e-01 -5.24625599e-01
1.76712170e-01 -2.57056821e-02 -1.29860163e-01 6.27320185e-02
-1.26610935e+00 1.19750535e+00 2.55039215e-01 -4.59033430e-01
-2.49588430e-01 -8.17700744e-01 -7.55406499e-01 4.79338109e-01
2.91150082e-02 -8.60731363e-01 1.34590673e+00 -1.07698202e+00
-1.19085813e+00 1.10430741e+00 -1.47484709e-02 -6.39310718e-01
4.85214859e-01 -5.41704297e-02 -1.67938635e-01 -2.12065250e-01
3.13611150e-01 5.53850114e-01 5.29581368e-01 -7.40330458e-01
-5.23754358e-01 -4.67730194e-01 1.21565359e-02 1.58262178e-01
-3.34397376e-01 7.14755729e-02 -3.55836423e-03 -7.85629332e-01
-1.39172450e-01 -9.04344797e-01 -2.74594370e-02 -3.78562719e-01
-6.97404668e-02 -3.60606372e-01 5.51789761e-01 -6.27751470e-01
8.51243854e-01 -2.26239991e+00 9.52593237e-02 6.34973720e-02
-9.76033881e-02 -1.70003138e-02 -4.20482218e-01 4.45870399e-01
-3.15683991e-01 3.87569261e-03 -2.01377079e-01 -4.84577507e-01
1.41817808e-01 6.55869488e-03 -4.76709783e-01 2.04076007e-01
3.11920583e-01 7.39511728e-01 -9.14833784e-01 -2.02411324e-01
-1.33678839e-01 1.98309422e-01 -6.44779980e-01 2.08895817e-01
-3.05273205e-01 5.54824732e-02 -3.11433256e-01 1.20320968e-01
1.57731667e-01 -3.27462196e-01 2.16694251e-01 -3.33236724e-01
4.10932124e-01 9.24270213e-01 -7.80754387e-01 1.85635507e+00
-9.36248779e-01 6.80322051e-01 -3.05722326e-01 -1.09890676e+00
9.26730275e-01 2.41666481e-01 1.83539450e-01 -9.70845342e-01
3.71917516e-01 2.44084299e-01 5.73576540e-02 -2.33975887e-01
5.90573013e-01 -4.53639656e-01 -1.63997442e-01 6.96552753e-01
8.39567557e-02 -2.32259586e-01 1.24229990e-01 3.95379141e-02
1.23214149e+00 1.13505803e-01 1.19707666e-01 -3.74395698e-01
4.20029998e-01 2.14287173e-02 -5.23402728e-03 5.10787129e-01
5.18163927e-02 6.17536128e-01 5.53984761e-01 -3.46735358e-01
-9.65776861e-01 -8.47949028e-01 -1.20356850e-01 1.39908302e+00
9.97451842e-02 -6.96775198e-01 -5.76621830e-01 -1.10434818e+00
1.91753045e-01 1.12395263e+00 -7.24457443e-01 -5.35128474e-01
-3.68577510e-01 -4.34161246e-01 4.45853829e-01 6.99652433e-01
2.84879416e-01 -7.71732092e-01 -5.61183691e-01 -3.91411372e-02
-3.57207417e-01 -1.29092431e+00 -4.85035151e-01 4.60168630e-01
-5.99132061e-01 -1.00054693e+00 -3.31571758e-01 -6.44152761e-01
-1.48053728e-02 8.58634710e-02 1.52401543e+00 5.49880322e-03
-4.93048579e-02 1.13860659e-01 -4.67670232e-01 -3.09431329e-02
-6.38444662e-01 4.18895036e-01 -3.01877916e-01 -1.78782195e-01
3.26214463e-01 -5.83571315e-01 -3.51749003e-01 3.45943034e-01
-6.27633095e-01 -6.71417043e-02 2.57630199e-01 1.04705977e+00
1.76769704e-01 -2.70880401e-01 5.33943057e-01 -6.48280442e-01
7.96126485e-01 -5.65408468e-01 -3.20023060e-01 3.03727299e-01
-6.23227656e-01 6.94045484e-01 7.44316816e-01 -2.07607806e-01
-7.89448559e-01 -3.76359880e-01 -1.29403964e-01 -4.79134440e-01
2.50195321e-02 3.81950110e-01 6.29659221e-02 4.07399118e-01
8.07310939e-01 -1.91046208e-01 -4.20352183e-02 -5.67429006e-01
4.39872801e-01 1.05581367e+00 3.10373425e-01 -4.42038208e-01
3.57077569e-01 4.63204905e-02 -2.80843437e-01 -3.03557187e-01
-1.26923633e+00 -4.28239942e-01 -3.38961124e-01 2.80385435e-01
6.08003199e-01 -8.77134144e-01 -4.16904122e-01 -7.39292949e-02
-1.00739455e+00 -4.79499817e-01 -3.40537429e-01 3.23572159e-01
-5.70116758e-01 1.59537509e-01 -4.23027068e-01 -7.25256056e-02
-4.56098229e-01 -9.58385348e-01 1.21492410e+00 -3.16769600e-01
-6.79454029e-01 -1.07776260e+00 1.39651760e-01 5.80639482e-01
7.17354536e-01 -2.65812457e-01 1.06143653e+00 -1.31671906e+00
7.07615316e-02 -1.60484090e-01 -4.32865679e-01 2.19473749e-01
-3.18237878e-02 -3.33633065e-01 -1.18440378e+00 -2.71070838e-01
7.29596540e-02 -5.63149035e-01 1.10679674e+00 -1.67324647e-01
1.20991683e+00 -2.31116936e-01 -3.05480599e-01 4.88201201e-01
9.88043547e-01 -2.48890460e-01 3.02474141e-01 5.38422287e-01
3.88513505e-01 6.89177811e-01 5.93319178e-01 3.09156537e-01
5.54943264e-01 1.10059059e+00 1.41926780e-01 2.08300903e-01
-2.57664829e-01 -3.10702920e-01 2.64797270e-01 4.80519503e-01
5.83959341e-01 -3.16689909e-01 -9.54895437e-01 5.06587565e-01
-1.71685159e+00 -8.60682964e-01 1.44456208e-01 2.04661012e+00
8.94140005e-01 1.97058976e-01 1.55355290e-01 1.88342050e-01
5.83283186e-01 3.87233160e-02 -2.29061872e-01 -6.15691900e-01
1.82920426e-01 2.79616714e-01 3.33986193e-01 4.88876969e-01
-9.34159815e-01 1.00420785e+00 5.71515274e+00 7.80135095e-01
-1.33339500e+00 7.85952508e-02 6.74504459e-01 -3.32880139e-01
-4.08279061e-01 1.34948224e-01 -3.45101029e-01 6.10653222e-01
1.22619355e+00 -5.38726151e-03 3.92591506e-01 5.82069159e-01
-2.28934318e-01 3.69027108e-02 -1.50546372e+00 8.11155379e-01
4.33130525e-02 -1.03733635e+00 6.71167001e-02 -3.81350577e-01
3.15530777e-01 4.75822568e-01 -1.25396669e-01 4.93239939e-01
4.61354464e-01 -1.06384361e+00 7.84263849e-01 -3.48173245e-03
5.20801544e-01 -4.89572197e-01 8.89379621e-01 1.26049235e-01
-8.06149900e-01 -1.51728705e-01 -2.29105741e-01 -1.08138546e-01
-7.58630633e-02 4.26519901e-01 -1.04801440e+00 4.09452170e-01
7.48690605e-01 8.75905454e-01 -8.45080614e-01 7.28438675e-01
-6.52869418e-02 5.19968688e-01 -2.18617722e-01 -5.08726053e-02
3.47975612e-01 7.07309321e-02 3.08580667e-01 1.19647491e+00
1.73039377e-01 -3.18625450e-01 -1.51826767e-03 9.41698194e-01
-1.26001149e-01 -7.21032098e-02 -3.07599276e-01 -1.19743481e-01
4.05433655e-01 1.09776175e+00 -4.31438148e-01 -3.72778296e-01
-3.48855704e-01 1.07330406e+00 5.91368377e-01 3.07331756e-02
-9.41572607e-01 -3.91655385e-01 5.70600808e-01 1.37773648e-01
3.97526264e-01 3.56653422e-01 -5.05762398e-01 -1.01966202e+00
6.39965944e-03 -8.41517687e-01 4.91067469e-01 -8.65223646e-01
-1.41499877e+00 6.06058538e-01 -6.86529055e-02 -9.54074979e-01
-4.90722448e-01 -4.57530230e-01 -7.00103223e-01 8.23003531e-01
-1.31596792e+00 -9.90687490e-01 -3.25428337e-01 3.38371634e-01
4.57673699e-01 -1.22890145e-01 7.43848205e-01 1.22941442e-01
-2.23638773e-01 8.41551960e-01 1.57017410e-01 6.15401007e-02
9.66258287e-01 -1.13721204e+00 7.62339294e-01 2.79583573e-01
1.91027716e-01 1.67400807e-01 6.99027181e-01 -2.21110418e-01
-7.30720222e-01 -9.90234613e-01 1.33890665e+00 -5.24039805e-01
9.68929946e-01 -3.46307397e-01 -9.26496089e-01 5.43309867e-01
3.48516673e-01 -1.66789263e-01 5.52091837e-01 3.87665302e-01
-6.56538785e-01 -5.87372901e-03 -1.04402614e+00 4.32939857e-01
1.12015593e+00 -7.67110527e-01 -8.96625102e-01 2.26043224e-01
7.24670231e-01 -1.91135749e-01 -8.38187039e-01 4.92875248e-01
4.25145209e-01 -1.08969259e+00 7.41887331e-01 -8.23404729e-01
7.21959114e-01 2.11826980e-01 -2.71760464e-01 -1.67355585e+00
-1.89990729e-01 4.75106090e-02 5.68957984e-01 1.21401048e+00
4.84368920e-01 -3.10793251e-01 6.87437177e-01 2.25700438e-01
-1.13464303e-01 -7.05650210e-01 -8.32732975e-01 -6.85588121e-01
2.86972284e-01 -2.88513809e-01 6.83265269e-01 1.14753652e+00
3.57970834e-01 1.03290522e+00 2.32069552e-01 -3.55931163e-01
1.72599614e-01 4.26717430e-01 3.94359112e-01 -1.00862992e+00
-5.35895228e-01 -8.15314293e-01 -3.48925948e-01 -6.66720212e-01
5.31204760e-01 -1.18757379e+00 -7.72831321e-04 -9.86934900e-01
4.12093878e-01 -3.94070506e-01 -2.42758751e-01 5.82867384e-01
-4.48760241e-01 1.42429486e-01 9.19325426e-02 -2.87651084e-02
-5.50685644e-01 6.86459243e-01 7.87320852e-01 -3.05453837e-01
1.32516250e-01 -1.67902913e-02 -5.93054533e-01 3.39978099e-01
9.78409052e-01 -4.52893615e-01 -5.04813910e-01 -7.28327751e-01
2.17792571e-01 1.83806866e-01 6.09129369e-01 -8.36383343e-01
-8.86226371e-02 1.80159926e-01 5.19592911e-02 -1.52021587e-01
1.82771742e-01 -7.77238131e-01 -2.60538220e-01 3.57340515e-01
-9.37734842e-01 3.13459605e-01 -1.14072803e-02 2.37185776e-01
-3.33654940e-01 -2.35621989e-01 8.81333470e-01 1.03829838e-01
-2.47471750e-01 -1.82957992e-01 -7.04532564e-02 4.49989259e-01
6.05539083e-01 -4.74718995e-02 -2.17397749e-01 -3.89058709e-01
-6.74616933e-01 2.00058103e-01 6.76797569e-01 7.77130544e-01
1.75656036e-01 -1.15719140e+00 -8.88148427e-01 1.94336116e-01
4.06718373e-01 -4.86464292e-01 -2.08609611e-01 7.69680262e-01
9.63713527e-02 4.82431829e-01 -1.61380529e-01 -5.13910592e-01
-1.34110940e+00 5.56076586e-01 6.44743204e-01 -3.71527344e-01
-1.89091519e-01 6.76444173e-01 4.82950881e-02 -6.65035963e-01
1.82504714e-01 -2.85866320e-01 -9.99950990e-02 2.57707071e-02
1.21223398e-01 3.86806220e-01 5.07923543e-01 -3.45363945e-01
-4.49776858e-01 3.16305369e-01 -8.69860426e-02 -4.59044315e-02
1.05393648e+00 -1.17631227e-01 -8.23517144e-03 3.18952978e-01
1.54148388e+00 -2.70809382e-01 -6.54683173e-01 -5.43891847e-01
3.69990081e-01 -3.04191262e-01 2.16378540e-01 -9.72632289e-01
-9.87818599e-01 7.90724695e-01 5.19009709e-01 1.26509562e-01
8.11430871e-01 2.47804806e-01 7.65815198e-01 3.64857554e-01
-1.72933400e-01 -9.41787481e-01 1.39177725e-01 8.63353848e-01
8.24371576e-01 -1.33031487e+00 -2.24852428e-01 1.13397706e-02
-7.03387380e-01 1.03590882e+00 4.21966672e-01 3.24261636e-02
3.45482737e-01 1.93872318e-01 -3.79179530e-02 -8.63517746e-02
-1.29545999e+00 -1.86452508e-01 4.28335279e-01 1.20341271e-01
8.74418914e-01 -1.08784594e-01 -3.85677099e-01 7.12090611e-01
-6.58550799e-01 3.72664072e-02 5.80043942e-02 6.73243701e-01
-2.48161480e-01 -1.01186585e+00 -8.80608987e-03 4.77056235e-01
-3.36853862e-01 -4.44997191e-01 -7.47970879e-01 4.24538285e-01
-1.58949420e-01 9.40466344e-01 4.09889817e-01 -5.74094296e-01
3.42710525e-01 2.85302788e-01 4.35471743e-01 -4.00389075e-01
-9.72157776e-01 -5.00112653e-01 5.33716738e-01 -6.40543103e-01
-1.54021345e-02 -6.35231137e-01 -1.13604593e+00 -2.94028997e-01
-1.91304788e-01 3.54652643e-01 6.18969679e-01 1.14197624e+00
4.12683934e-01 4.18976516e-01 7.25033224e-01 -4.55363154e-01
-1.17071962e+00 -1.29645288e+00 -2.37169787e-02 9.32746232e-01
1.18830711e-01 -3.92297596e-01 -4.95950490e-01 -2.25891456e-01]
|
[11.142507553100586, 8.72928524017334]
|
aaf5e2dc-9a9f-46df-8791-25d5d2b5faee
|
cross-modal-consensus-network-for-weakly
|
2107.12589
| null |
https://arxiv.org/abs/2107.12589v1
|
https://arxiv.org/pdf/2107.12589v1.pdf
|
Cross-modal Consensus Network for Weakly Supervised Temporal Action Localization
|
Weakly supervised temporal action localization (WS-TAL) is a challenging task that aims to localize action instances in the given video with video-level categorical supervision. Both appearance and motion features are used in previous works, while they do not utilize them in a proper way but apply simple concatenation or score-level fusion. In this work, we argue that the features extracted from the pretrained extractor, e.g., I3D, are not the WS-TALtask-specific features, thus the feature re-calibration is needed for reducing the task-irrelevant information redundancy. Therefore, we propose a cross-modal consensus network (CO2-Net) to tackle this problem. In CO2-Net, we mainly introduce two identical proposed cross-modal consensus modules (CCM) that design a cross-modal attention mechanism to filter out the task-irrelevant information redundancy using the global information from the main modality and the cross-modal local information of the auxiliary modality. Moreover, we treat the attention weights derived from each CCMas the pseudo targets of the attention weights derived from another CCM to maintain the consistency between the predictions derived from two CCMs, forming a mutual learning manner. Finally, we conduct extensive experiments on two common used temporal action localization datasets, THUMOS14 and ActivityNet1.2, to verify our method and achieve the state-of-the-art results. The experimental results show that our proposed cross-modal consensus module can produce more representative features for temporal action localization.
|
['Wei-Shi Zheng', 'Ying Shan', 'Dan Xu', 'Jia-Chang Feng', 'Fa-Ting Hong']
|
2021-07-27
| null | null | null | null |
['weakly-supervised-action-localization', 'weakly-supervised-temporal-action']
|
['computer-vision', 'computer-vision']
|
[ 9.61031392e-02 -3.06748569e-01 -3.57109666e-01 -1.49206698e-01
-7.56777346e-01 -1.00992836e-01 5.73176324e-01 -3.24043512e-01
-4.91304606e-01 5.16006470e-01 5.78908980e-01 4.63277757e-01
-1.36786178e-01 -1.85215741e-01 -5.55735707e-01 -9.83547091e-01
7.71381184e-02 -1.12844683e-01 7.11732090e-01 -1.29991651e-01
1.17962882e-01 -7.42327198e-02 -1.59751582e+00 7.80344605e-01
8.58088195e-01 1.08920050e+00 3.62905174e-01 7.44589642e-02
5.98426722e-02 1.18535578e+00 -3.46223652e-01 2.60824692e-02
2.15994745e-01 -8.14301252e-01 -6.91823006e-01 1.82000369e-01
2.13471070e-01 -2.95401454e-01 -4.86271143e-01 9.71417964e-01
4.03511375e-01 2.27075949e-01 4.17844504e-01 -1.49847245e+00
-4.06200737e-01 5.03705680e-01 -7.14122713e-01 3.75162840e-01
3.17900866e-01 4.28860962e-01 9.15820897e-01 -8.07230890e-01
5.87084353e-01 1.32789338e+00 4.13646847e-01 4.16230589e-01
-9.45717216e-01 -6.92222536e-01 6.95960164e-01 7.50319600e-01
-1.38884866e+00 -4.88727808e-01 1.02093565e+00 -3.66783231e-01
7.18682468e-01 1.32295072e-01 6.01333141e-01 1.36016703e+00
3.03829432e-01 1.11022258e+00 1.13712859e+00 -3.98390628e-02
-2.02314835e-02 -1.05653755e-01 -1.08279653e-01 6.58194840e-01
-2.59118468e-01 -1.71538517e-01 -7.60432959e-01 4.92356047e-02
7.38417327e-01 2.79325306e-01 -5.24327397e-01 -3.78230274e-01
-1.64289486e+00 4.03780013e-01 4.80373263e-01 7.19088376e-01
-4.23091322e-01 1.91890091e-01 6.32989168e-01 2.59448975e-01
2.50602424e-01 -8.46198276e-02 -6.02694154e-01 -2.30985358e-01
-5.04134059e-01 -1.30699366e-01 1.48819864e-01 7.90655375e-01
7.92341352e-01 -1.96315393e-01 -6.13061309e-01 6.59360945e-01
3.83521020e-01 2.89332658e-01 1.02662301e+00 -9.07121241e-01
6.77956820e-01 8.79021049e-01 6.50536269e-02 -1.20212603e+00
-3.66021991e-01 -2.49120399e-01 -8.58019173e-01 4.18709554e-02
1.02447815e-01 -8.04594681e-02 -7.87337422e-01 2.16346192e+00
3.15281004e-01 5.65779448e-01 7.23828822e-02 1.15361595e+00
7.53783047e-01 4.30843443e-01 2.87271410e-01 -4.22437549e-01
1.10855758e+00 -1.34184551e+00 -8.13766003e-01 -2.23116308e-01
6.47523105e-01 -6.06002212e-01 8.38703454e-01 1.14436574e-01
-7.88185954e-01 -9.19153452e-01 -1.00069118e+00 6.54762760e-02
-1.32928938e-01 4.24654007e-01 4.99883801e-01 -3.06550741e-01
-5.56109667e-01 6.67564273e-01 -9.35565472e-01 -3.58701289e-01
3.39561224e-01 4.45447490e-02 -7.87178576e-01 1.17818482e-01
-1.28601837e+00 8.19574118e-01 6.66218162e-01 3.84682536e-01
-9.25531983e-01 -2.79194832e-01 -7.18351960e-01 -7.97926858e-02
5.89177847e-01 -5.12923896e-01 9.99547064e-01 -1.47295308e+00
-1.32907605e+00 4.92703736e-01 -5.38084283e-02 -3.01046409e-02
5.50301075e-01 -2.58538306e-01 -4.86997694e-01 1.37827203e-01
3.42929274e-01 7.19446778e-01 7.69519866e-01 -1.09099293e+00
-8.86845946e-01 -2.65399128e-01 1.66767493e-01 5.11456788e-01
-3.55485827e-01 -1.29310280e-01 -7.77283549e-01 -7.96980321e-01
2.92250663e-01 -8.69015932e-01 -5.84491640e-02 -1.36223566e-02
-2.19574258e-01 -4.76503074e-01 9.25310731e-01 -6.55356407e-01
1.31003273e+00 -2.28102708e+00 6.14684582e-01 -8.57779682e-02
2.25453153e-02 2.18199059e-01 -3.90792042e-01 3.78117114e-01
-2.96440452e-01 -1.44121185e-01 -1.65553287e-01 -3.39262247e-01
-2.05394715e-01 2.15118051e-01 1.28201544e-01 5.32791376e-01
3.82910281e-01 8.11253309e-01 -1.10282803e+00 -8.43949616e-01
4.09004033e-01 2.57133901e-01 -2.02606529e-01 2.45857850e-01
-7.51257762e-02 8.44851255e-01 -7.23963499e-01 6.62847877e-01
3.65504950e-01 -1.68784142e-01 1.31112933e-01 -7.82483220e-01
-1.80059299e-02 2.59477831e-02 -1.24717140e+00 2.29639554e+00
-2.40209118e-01 1.83827639e-01 -1.79973096e-01 -1.16071665e+00
5.30557394e-01 3.85077238e-01 1.01185286e+00 -8.02299380e-01
2.69019663e-01 5.52448519e-02 1.24691300e-01 -9.11713481e-01
-1.31640911e-01 7.93954730e-02 -3.46461795e-02 2.85132110e-01
4.38149959e-01 6.57165706e-01 2.39993364e-01 4.95765842e-02
1.21215916e+00 5.81102073e-01 2.77499586e-01 -1.26661792e-01
9.75997508e-01 -1.61123037e-01 1.15990758e+00 4.46434677e-01
-5.43935895e-01 6.26460135e-01 3.69281530e-01 -3.40597332e-01
-6.39432847e-01 -7.12119997e-01 1.73354149e-01 1.01748514e+00
5.05073726e-01 -5.44120669e-01 -4.33987498e-01 -1.18131804e+00
-2.66662687e-01 1.69094622e-01 -8.16803157e-01 -4.54046935e-01
-4.94414717e-01 -4.72803593e-01 3.00312132e-01 7.71271348e-01
8.44933391e-01 -1.35847807e+00 -4.75084722e-01 1.26722053e-01
-6.02998435e-01 -1.08669031e+00 -7.11198032e-01 3.16500328e-02
-5.89268267e-01 -1.24770892e+00 -7.21932471e-01 -6.41853809e-01
5.59612572e-01 4.89179105e-01 6.78247750e-01 3.00973710e-02
1.50993377e-01 3.32000226e-01 -6.52143717e-01 1.03218168e-01
5.40959686e-02 -1.28690913e-01 1.33941099e-01 6.43730164e-01
4.20532674e-01 -8.35376680e-01 -5.96346021e-01 5.82103312e-01
-8.48775268e-01 3.80075604e-01 8.68559599e-01 8.57521594e-01
5.56787729e-01 -8.33784118e-02 4.67973918e-01 -2.72587270e-01
3.66230875e-01 -5.01966298e-01 -9.97036994e-02 3.96549612e-01
-1.28141224e-01 1.31045789e-01 5.49977720e-01 -7.71648765e-01
-1.09761083e+00 3.15610886e-01 1.20002113e-01 -9.64855075e-01
-1.98202297e-01 6.61129236e-01 -5.49448252e-01 9.50510204e-02
2.81996489e-01 4.64042813e-01 -1.43904656e-01 -5.00327766e-01
2.05780640e-01 4.13635314e-01 4.45711643e-01 -3.98282886e-01
7.39022136e-01 4.87138957e-01 -9.89671350e-02 -2.32580751e-01
-9.84911859e-01 -5.75375915e-01 -8.40687513e-01 -4.22234654e-01
1.19365156e+00 -1.01474249e+00 -6.17162108e-01 6.12056077e-01
-1.25236177e+00 -4.08919267e-02 -1.88852683e-01 7.66305745e-01
-5.75459659e-01 5.43864071e-01 -3.54532391e-01 -6.18421316e-01
-7.39173368e-02 -1.24442422e+00 1.21959412e+00 2.63134360e-01
9.72525179e-02 -6.69790924e-01 2.83277541e-01 3.53823453e-01
-2.17846893e-02 3.14819455e-01 4.54947263e-01 -6.20405674e-01
-5.16291797e-01 4.26720753e-02 -2.40046069e-01 4.71854925e-01
4.03330922e-01 -2.14962214e-01 -8.70195508e-01 -1.60975650e-01
-5.23272641e-02 -4.24121678e-01 1.10486686e+00 1.01480119e-01
1.12173641e+00 -1.17188931e-01 -2.60477275e-01 4.21495408e-01
1.15018129e+00 1.24405034e-01 6.38756990e-01 1.84102461e-01
9.89324987e-01 4.45108473e-01 9.40832794e-01 3.46405596e-01
3.50497663e-01 8.63486767e-01 4.87506628e-01 -6.30348027e-02
-9.82040539e-02 -3.33131403e-01 7.96885550e-01 9.43996608e-01
-4.56050128e-01 6.54909983e-02 -4.41065550e-01 4.51409191e-01
-2.61914206e+00 -1.18620670e+00 -8.34488645e-02 2.11299586e+00
6.94661856e-01 5.71628213e-02 8.05877224e-02 -3.48722190e-02
7.44361520e-01 3.89373690e-01 -5.44461966e-01 3.50427449e-01
-1.92962602e-01 -3.02000821e-01 2.01081276e-01 8.66918359e-03
-1.44137132e+00 7.07868934e-01 4.80982971e+00 1.08363497e+00
-1.00289965e+00 3.47687513e-01 3.34826678e-01 -5.62178753e-02
6.65291473e-02 1.16913661e-01 -3.58621478e-01 7.74835050e-01
4.96968865e-01 5.09713627e-02 1.88518375e-01 7.11366236e-01
4.22427237e-01 -6.21978939e-02 -1.10257471e+00 1.10226524e+00
2.51714706e-01 -8.73885214e-01 4.72941808e-02 -2.02359110e-01
6.04674459e-01 -4.28064093e-02 -2.14992747e-01 6.01163089e-01
-5.27493432e-02 -5.78520894e-01 7.74766862e-01 8.84970784e-01
3.94238263e-01 -4.86758053e-01 9.21299815e-01 4.61301625e-01
-1.65010571e+00 -3.07263702e-01 -2.34111428e-01 1.29092738e-01
2.41823763e-01 2.16656163e-01 7.57898986e-02 1.02167678e+00
8.87437403e-01 1.38709259e+00 -6.49146140e-01 8.92251432e-01
-3.29825282e-01 3.43820661e-01 -1.18839718e-01 3.91039431e-01
4.11080241e-01 -1.29594743e-01 5.44819117e-01 9.07116413e-01
1.77711919e-01 1.08917855e-01 5.37398517e-01 5.78820407e-01
2.09536821e-01 2.16677085e-01 -3.57588112e-01 -4.90341103e-03
2.69995816e-02 1.28023398e+00 -3.95949781e-01 -3.28351766e-01
-6.87192380e-01 1.24552035e+00 2.70265341e-01 3.18648010e-01
-1.20253932e+00 -8.54519084e-02 6.24549091e-01 -2.98134267e-01
2.96281785e-01 -7.10777789e-02 3.14617902e-01 -1.63705432e+00
3.13070148e-01 -9.12111998e-01 5.32491744e-01 -9.11752164e-01
-1.44373202e+00 4.35796380e-01 1.40679955e-01 -1.87855721e+00
5.18332981e-02 -2.90973574e-01 -7.34652579e-01 6.82173491e-01
-1.32403278e+00 -1.55393553e+00 -5.68753183e-01 9.00727034e-01
7.50521541e-01 -9.41550508e-02 5.60479701e-01 4.85195339e-01
-8.39811921e-01 4.28030759e-01 -2.48784065e-01 2.29516700e-01
8.90452385e-01 -8.47208083e-01 -3.27751994e-01 8.51138055e-01
1.50248893e-02 4.89510328e-01 3.11337441e-01 -7.19482899e-01
-1.24635959e+00 -1.09321618e+00 4.72767919e-01 -2.62844652e-01
7.38245308e-01 -9.20887068e-02 -9.34938133e-01 6.53408051e-01
2.34896988e-01 2.57285774e-01 3.89476120e-01 -1.07557565e-01
-1.94945335e-01 -2.98733681e-01 -6.97267830e-01 5.00855684e-01
1.39693451e+00 -4.76893991e-01 -6.83417737e-01 2.76819319e-01
7.45602727e-01 -1.26728848e-01 -9.03891742e-01 7.11285830e-01
6.51278079e-01 -1.09172833e+00 7.46275127e-01 -7.24486530e-01
5.49668491e-01 -7.24442422e-01 -1.16505392e-01 -1.07828212e+00
-5.86525738e-01 -2.14284331e-01 -1.80962145e-01 1.37526441e+00
6.16909824e-02 -3.56018215e-01 4.22590554e-01 1.64405540e-01
-3.30926359e-01 -7.78810740e-01 -1.18560886e+00 -7.62673199e-01
-3.47749710e-01 -2.40636453e-01 3.81795377e-01 1.05548334e+00
1.53040633e-01 4.69296426e-01 -7.57715404e-01 5.59616787e-03
2.92519271e-01 7.17997104e-02 6.54730618e-01 -1.02020037e+00
-3.35503489e-01 -4.46022481e-01 -6.90012157e-01 -1.05541873e+00
3.70663166e-01 -6.99219882e-01 1.59984395e-01 -1.28703582e+00
5.68778872e-01 -1.23583384e-01 -9.27819073e-01 8.54250669e-01
-4.00239289e-01 7.94960931e-02 2.82146335e-01 4.14087951e-01
-1.09846997e+00 1.11054075e+00 1.37321866e+00 -3.20522904e-01
-1.34087056e-01 -2.83656836e-01 -4.81092244e-01 8.70155156e-01
5.77996075e-01 -4.24499720e-01 -5.94813943e-01 -3.85104746e-01
-3.02355647e-01 3.85231264e-02 4.95654672e-01 -1.30416644e+00
2.81741679e-01 -5.84628642e-01 3.82879049e-01 -6.56436980e-01
3.16744566e-01 -1.14230597e+00 7.33324513e-02 2.94710994e-01
-2.48212144e-01 -9.76326540e-02 -4.16192748e-02 7.90400386e-01
-4.37796235e-01 7.69078583e-02 5.37714601e-01 -2.13407397e-01
-1.04545629e+00 4.97304261e-01 -1.78227842e-01 -1.68103307e-01
1.16728497e+00 -8.14516917e-02 -3.62161458e-01 -2.73096204e-01
-7.63771534e-01 4.47002798e-01 3.49916518e-01 7.78633356e-01
4.72116143e-01 -1.79433537e+00 -6.78963482e-01 -2.55624913e-02
4.95008707e-01 -3.24553877e-01 6.69762433e-01 1.45504034e+00
1.38790324e-01 2.33887836e-01 -4.48592782e-01 -7.51477301e-01
-1.12418735e+00 7.16875315e-01 4.59845394e-01 -3.50282937e-01
-5.52166164e-01 5.63191652e-01 5.37499905e-01 -1.01911969e-01
9.46004689e-02 -3.37760717e-01 -5.34785151e-01 3.96545567e-02
5.06358683e-01 1.26265898e-01 -2.22223774e-01 -9.84501123e-01
-6.67938948e-01 6.77167416e-01 1.76423401e-01 -2.55168416e-02
1.24960077e+00 -2.08666682e-01 -7.47449398e-02 6.74830258e-01
1.20407999e+00 -3.60936791e-01 -1.36911976e+00 -4.26562279e-01
-2.33253121e-01 -4.53223318e-01 -8.97253156e-02 -5.09379983e-01
-1.41324437e+00 7.03950822e-01 8.39090765e-01 -2.18476027e-01
1.42258334e+00 1.21616749e-02 5.77994943e-01 2.42361665e-01
3.41689914e-01 -1.20703840e+00 3.77959281e-01 2.22933352e-01
1.08235347e+00 -1.33323240e+00 -6.99234530e-02 -1.61618516e-01
-8.38546336e-01 9.01897728e-01 1.16663122e+00 -6.35812655e-02
6.03370428e-01 -2.97667056e-01 -1.43343762e-01 -1.13447294e-01
-8.46243739e-01 -5.23591816e-01 5.01485884e-01 4.44555074e-01
2.05621555e-01 -2.61708379e-01 -6.80193484e-01 9.01838303e-01
6.99397504e-01 1.19551376e-01 2.18855757e-02 1.12172937e+00
-3.08686227e-01 -1.05093300e+00 -7.68877417e-02 2.35716060e-01
-4.50188257e-02 1.74059287e-01 -3.60792607e-01 6.73913658e-01
7.00497329e-01 9.27871227e-01 -1.59818068e-01 -9.43311870e-01
3.28961521e-01 2.51579867e-03 3.74874353e-01 -3.03922683e-01
-4.50474709e-01 2.22896397e-01 -5.23949489e-02 -1.03385794e+00
-1.07036746e+00 -6.83645070e-01 -1.24092793e+00 1.43222585e-01
-2.39736900e-01 -4.65938151e-02 4.07213941e-02 1.27431154e+00
4.83541012e-01 7.04709113e-01 6.73489988e-01 -9.30644274e-01
-3.39354992e-01 -1.23551250e+00 -3.89460593e-01 8.22900951e-01
6.68644682e-02 -1.07348263e+00 -3.38746428e-01 1.99087992e-01]
|
[8.570191383361816, 0.7265576124191284]
|
31571bca-6da8-4789-b034-6a7f9b0c5635
|
one-shot-face-reenactment-on-megapixels
|
2205.13368
| null |
https://arxiv.org/abs/2205.13368v1
|
https://arxiv.org/pdf/2205.13368v1.pdf
|
One-Shot Face Reenactment on Megapixels
|
The goal of face reenactment is to transfer a target expression and head pose to a source face while preserving the source identity. With the popularity of face-related applications, there has been much research on this topic. However, the results of existing methods are still limited to low-resolution and lack photorealism. In this work, we present a one-shot and high-resolution face reenactment method called MegaFR. To be precise, we leverage StyleGAN by using 3DMM-based rendering images and overcome the lack of high-quality video datasets by designing a loss function that works without high-quality videos. Also, we apply iterative refinement to deal with extreme poses and/or expressions. Since the proposed method controls source images through 3DMM parameters, we can explicitly manipulate source images. We apply MegaFR to various applications such as face frontalization, eye in-painting, and talking head generation. Experimental results show that our method successfully disentangles identity from expression and head pose, and outperforms conventional methods.
|
['Nam Ik Cho', 'Hyung Il Koo', 'Geonsu Lee', 'Wonjun Kang']
|
2022-05-26
| null | null | null | null |
['talking-head-generation', 'face-reenactment', 'facial-editing', 'talking-face-generation']
|
['computer-vision', 'computer-vision', 'computer-vision', 'computer-vision']
|
[ 2.69897610e-01 8.90315622e-02 1.23323947e-02 -5.31264067e-01
-3.54917139e-01 -2.73294985e-01 4.60593700e-01 -9.56728935e-01
7.36350566e-02 5.18578231e-01 2.86698699e-01 2.29178488e-01
1.53837889e-01 -5.03394604e-01 -4.96116251e-01 -7.24246323e-01
3.50179136e-01 4.28188927e-02 -6.69011474e-02 -3.75193000e-01
1.02097809e-01 7.29424119e-01 -1.77793229e+00 6.32397905e-02
6.21514320e-01 8.98863196e-01 -2.19322234e-01 3.80533248e-01
1.32264793e-01 5.58408618e-01 -5.99138379e-01 -6.82261586e-01
3.65293771e-01 -6.88858807e-01 -3.47111672e-01 5.18379331e-01
8.59215677e-01 -6.73997343e-01 -2.71767378e-01 9.97054458e-01
6.90393269e-01 -7.95269310e-02 2.73254901e-01 -1.59936047e+00
-5.91512680e-01 2.85567362e-02 -1.19602656e+00 -3.31574559e-01
7.92530119e-01 -7.80735835e-02 3.58709663e-01 -1.05167294e+00
6.46461725e-01 1.80248225e+00 4.51781660e-01 1.00363326e+00
-1.34975576e+00 -1.34923768e+00 1.57705829e-01 -1.74153317e-02
-1.59836435e+00 -1.01685047e+00 9.63737667e-01 -2.44459942e-01
-6.49278462e-02 3.53937387e-01 7.35834658e-01 1.09350431e+00
-1.93038911e-01 4.89160448e-01 1.29691470e+00 -2.19051331e-01
-1.07025899e-01 4.05595684e-03 -5.12630165e-01 8.10683191e-01
-6.21358901e-02 1.24819279e-01 -8.84731650e-01 -2.23245248e-01
1.38824666e+00 -4.36471682e-03 -6.54589713e-01 -4.44992423e-01
-9.07427430e-01 6.01467729e-01 1.48481205e-01 -7.89660215e-02
1.66551378e-02 -6.64952695e-02 -6.69782013e-02 1.49052039e-01
5.01403511e-01 2.26367041e-01 -7.74626732e-02 1.00265540e-01
-9.90141988e-01 3.29403669e-01 5.69224954e-01 1.07648349e+00
7.44732201e-01 1.46032706e-01 -1.11635305e-01 9.27569866e-01
3.41451883e-01 5.20308733e-01 3.33402693e-01 -1.42754900e+00
-5.71044721e-02 4.38265383e-01 2.57291365e-02 -1.14874864e+00
-2.05367148e-01 1.03991367e-01 -8.19740713e-01 5.21480680e-01
2.30401739e-01 -1.57028466e-01 -8.06221426e-01 1.91386831e+00
7.87006080e-01 4.37320024e-01 -2.15254337e-01 1.10877764e+00
8.60085666e-01 3.56936246e-01 -3.81085277e-01 -4.82642293e-01
1.41403818e+00 -7.25612998e-01 -9.91532326e-01 6.83423355e-02
1.46325808e-02 -9.49863613e-01 1.11740172e+00 3.91767681e-01
-1.19353878e+00 -4.26621974e-01 -8.49962234e-01 -3.04908782e-01
3.46090615e-01 4.97577377e-02 5.59815764e-01 8.07108462e-01
-1.09537053e+00 3.02747190e-01 -5.94513595e-01 -3.30231130e-01
4.24283743e-01 5.60706317e-01 -6.99472070e-01 8.56409129e-03
-8.75660062e-01 5.68468988e-01 -1.29733846e-01 1.18768513e-01
-6.09544635e-01 -7.87797689e-01 -9.09964621e-01 -2.41011098e-01
3.36968213e-01 -6.95991218e-01 1.14204526e+00 -1.20938063e+00
-2.19991684e+00 1.08794796e+00 -3.46651614e-01 2.12719992e-01
6.24630451e-01 -2.80177116e-01 -3.16935092e-01 1.66153133e-01
-1.61377773e-01 8.07321608e-01 1.22769713e+00 -1.32115781e+00
-3.43204767e-01 -6.96397245e-01 2.02678069e-01 2.96845376e-01
-5.60069323e-01 3.09543729e-01 -8.06318521e-01 -7.13577390e-01
1.82396278e-01 -8.55426133e-01 2.28756249e-01 6.76099300e-01
-3.52200717e-01 2.02675149e-01 1.24228203e+00 -5.31172335e-01
8.67360711e-01 -2.33899426e+00 4.18716013e-01 2.33428576e-03
4.00389433e-01 1.65858239e-01 -1.39387414e-01 -5.15964665e-02
-1.61733165e-01 -3.71185392e-02 -3.47921140e-02 -7.00713992e-01
-2.11168930e-01 8.74701291e-02 -2.89603591e-01 6.57797933e-01
-6.86858641e-03 5.59022188e-01 -5.75610995e-01 -7.36129284e-01
1.45807773e-01 1.06378710e+00 -8.73980999e-01 3.73551488e-01
4.91588190e-02 9.02429163e-01 -2.46588662e-01 6.49035394e-01
1.08805680e+00 1.11777328e-01 3.03572714e-01 -5.06166220e-01
-2.79815122e-02 -1.92409933e-01 -1.07864177e+00 1.69522190e+00
-4.11829561e-01 6.13424122e-01 5.43507636e-01 -2.10809588e-01
8.88954699e-01 2.61875451e-01 4.24296498e-01 -5.51256359e-01
3.22577715e-01 -4.34385203e-02 -2.35445276e-01 -4.35053647e-01
2.76106685e-01 -2.53318071e-01 4.29522276e-01 2.06478372e-01
-1.72097757e-01 -3.05510789e-01 -1.35302350e-01 1.58528537e-02
3.69188398e-01 3.92821342e-01 7.73491412e-02 -1.25570431e-01
5.78126907e-01 -6.96342766e-01 7.86284983e-01 -1.62693039e-01
8.05504397e-02 9.98269856e-01 5.60764432e-01 -1.51912704e-01
-8.08372259e-01 -9.12755549e-01 -1.51371256e-01 9.93221998e-01
3.26719105e-01 -4.95693415e-01 -1.08217692e+00 -2.54299939e-01
-1.67068556e-01 3.37143362e-01 -7.21055508e-01 -4.44558403e-03
-6.75776064e-01 -6.05138302e-01 5.54587841e-01 2.59952128e-01
7.46832788e-01 -7.19447315e-01 -4.65530515e-01 -2.53887981e-01
-1.87703326e-01 -1.19278646e+00 -9.87078786e-01 -9.74012971e-01
-6.02449238e-01 -1.17614472e+00 -8.62038314e-01 -7.45036900e-01
9.61425185e-01 3.65360588e-01 7.12697089e-01 1.86073199e-01
-2.65133113e-01 2.46728107e-01 -8.17945749e-02 -1.95689768e-01
-1.22852400e-01 -2.47984841e-01 2.89689064e-01 6.21382892e-01
-1.30931046e-02 -8.82116258e-01 -8.00789416e-01 6.45009100e-01
-8.40737224e-01 4.19433445e-01 2.32130617e-01 5.53727150e-01
3.11811060e-01 -2.05404893e-01 2.91257948e-01 -8.20320308e-01
4.51038390e-01 1.34103343e-01 -7.57991254e-01 2.02697162e-02
-3.24633092e-01 -3.05039316e-01 3.25606287e-01 -6.99634969e-01
-1.29843855e+00 2.32222839e-03 -1.35229856e-01 -7.30822861e-01
7.41885155e-02 -2.91411966e-01 -7.03532040e-01 -4.17324901e-01
3.38942766e-01 -7.05448464e-02 3.54104042e-01 -3.82012248e-01
3.56223345e-01 5.33728421e-01 6.09857142e-01 -6.01916790e-01
1.02993345e+00 7.84889817e-01 1.70667972e-02 -1.01817298e+00
-7.36074805e-01 1.37658522e-01 -5.63128531e-01 -2.92477429e-01
6.92982793e-01 -9.12784040e-01 -1.15195990e+00 7.12846756e-01
-1.00694299e+00 -9.06481072e-02 2.40547936e-02 2.76473284e-01
-4.48900253e-01 3.51076722e-01 -4.95195836e-01 -6.73945010e-01
-1.84059113e-01 -1.20598960e+00 1.39756405e+00 4.39279556e-01
-1.53669789e-01 -5.62600732e-01 -1.72900274e-01 4.27970678e-01
3.87067407e-01 3.87400597e-01 4.68520910e-01 3.24151695e-01
-5.75359404e-01 9.47151035e-02 -2.07463741e-01 8.46808329e-02
4.18079853e-01 2.73152769e-01 -1.17117155e+00 -4.71442074e-01
1.84457883e-01 -3.18754673e-01 3.42030764e-01 1.48139834e-01
1.18780804e+00 -4.98781443e-01 -2.58248121e-01 1.13689268e+00
7.32314587e-01 9.62495431e-02 8.38243783e-01 -1.41055528e-02
9.40938056e-01 9.43125546e-01 5.26563883e-01 4.54664171e-01
4.06258732e-01 1.07003164e+00 2.80340433e-01 -2.81769156e-01
-2.92502135e-01 -4.76599425e-01 3.91408920e-01 4.87738907e-01
-1.75551042e-01 1.10158183e-01 -3.39880437e-01 6.63240766e-03
-1.57468617e+00 -8.44715238e-01 2.44092196e-01 2.21752262e+00
9.82137322e-01 -3.15501302e-01 1.63083643e-01 6.92084879e-02
8.12718511e-01 1.06997840e-01 -6.30150735e-01 1.50441334e-01
-8.25481117e-02 5.80062680e-02 7.58755803e-02 6.19449258e-01
-7.10656762e-01 1.00318527e+00 5.95998859e+00 6.28227413e-01
-1.39466989e+00 -6.28635928e-04 6.29819930e-01 -4.20360357e-01
-3.90355706e-01 -1.31047472e-01 -8.23857963e-01 2.00516224e-01
1.85983062e-01 -2.30268881e-01 4.85544622e-01 5.77898264e-01
3.09973210e-01 8.35268199e-02 -1.07985950e+00 1.48001897e+00
5.12515843e-01 -9.57061768e-01 6.62689731e-02 1.88234270e-01
5.94660699e-01 -8.19694221e-01 3.35770011e-01 -1.34279847e-01
-7.70770833e-02 -1.14761651e+00 6.63744271e-01 3.00053984e-01
1.31656122e+00 -7.78939664e-01 1.03589892e-01 5.48595972e-02
-1.03584063e+00 1.59155384e-01 -1.95026517e-01 1.28804371e-01
1.48933783e-01 2.77644694e-01 -4.74628121e-01 1.64118409e-01
6.00263298e-01 6.14146650e-01 -3.31720114e-01 6.86108530e-01
-4.23580319e-01 5.15642427e-02 -2.99183637e-01 3.76264751e-01
-5.87398171e-01 -3.17203134e-01 4.71629888e-01 5.98945558e-01
3.78741056e-01 4.42591041e-01 3.94562073e-02 7.64127016e-01
-3.92420322e-01 1.14617668e-01 -6.25597179e-01 3.51434737e-01
4.56490964e-01 1.33957875e+00 -4.04699624e-01 7.96505287e-02
-2.25263894e-01 1.22201025e+00 -8.55786167e-03 3.70751202e-01
-8.99508893e-01 -6.59563392e-02 1.15041494e+00 5.11705399e-01
-3.98737043e-02 -1.37738749e-01 1.13461703e-01 -1.39956141e+00
-1.35772547e-03 -1.04548168e+00 -4.14356254e-02 -9.62162673e-01
-8.48792076e-01 7.50191391e-01 1.30068600e-01 -1.02925813e+00
-1.86924249e-01 -3.25739831e-01 -5.08638740e-01 6.48972332e-01
-1.42194569e+00 -1.40388715e+00 -5.50542712e-01 9.28204536e-01
5.14528573e-01 5.12489006e-02 8.06196928e-01 3.81821930e-01
-6.94478929e-01 9.77730036e-01 -4.18261200e-01 1.17107801e-01
9.95351851e-01 -5.93584061e-01 2.53174603e-01 5.45788109e-01
-9.33882296e-02 7.16367185e-01 6.41225517e-01 -3.86121154e-01
-1.53996372e+00 -8.74243915e-01 3.84754419e-01 -2.88352787e-01
1.43116415e-01 -6.31488264e-01 -7.86215186e-01 6.73187971e-01
1.55101791e-01 1.31828785e-01 5.71704686e-01 -1.30740970e-01
-5.14038742e-01 -3.80465150e-01 -1.43804169e+00 9.55556929e-01
1.32687008e+00 -4.69439596e-01 -1.40058070e-01 2.23045312e-02
5.28179228e-01 -7.19113350e-01 -7.07674325e-01 3.95616353e-01
9.60721016e-01 -1.14004934e+00 9.01826084e-01 -2.44939417e-01
2.09253237e-01 -5.00357985e-01 -1.18471803e-02 -1.13171852e+00
-1.74906999e-01 -1.06740069e+00 8.95614028e-02 1.50845826e+00
-1.28136426e-01 -6.35164440e-01 8.78138125e-01 7.94927895e-01
3.37831229e-01 -5.13909698e-01 -7.81693161e-01 -3.99290264e-01
-2.46808365e-01 7.97855482e-02 9.25360501e-01 1.01386261e+00
-1.93194509e-01 2.69497424e-01 -8.46698225e-01 2.75607467e-01
7.74116635e-01 3.00278544e-01 1.27969086e+00 -1.07231939e+00
-4.51929606e-02 -2.37795711e-01 -4.20369893e-01 -1.13490188e+00
2.61931330e-01 -3.27754050e-01 -2.41537347e-01 -8.56324315e-01
1.30732134e-01 -1.75474212e-01 3.71942520e-01 4.52063471e-01
6.70667440e-02 6.85142457e-01 2.70919144e-01 2.06057727e-01
-9.48374644e-02 7.02872813e-01 1.60579717e+00 2.01054245e-01
-1.73219025e-01 -1.18519731e-01 -8.80668938e-01 9.61759329e-01
6.52893066e-01 -1.59852773e-01 -6.76488459e-01 -5.16726851e-01
-9.11706612e-02 2.78143615e-01 4.43203032e-01 -6.84657276e-01
1.19578905e-01 -3.38498801e-01 5.69351435e-01 -1.52488172e-01
8.79931390e-01 -6.68013036e-01 4.51332450e-01 1.06301963e-01
-6.97419867e-02 5.74990213e-02 5.13233580e-02 2.08136514e-01
-4.02878597e-02 3.06346834e-01 1.15663409e+00 1.29443839e-01
-2.90736884e-01 6.90036893e-01 1.00518748e-01 3.41940671e-02
1.05484557e+00 -2.81217128e-01 -3.98794599e-02 -7.75732040e-01
-4.00349647e-01 4.56658751e-02 1.02577877e+00 6.06646478e-01
9.11135674e-01 -1.50759029e+00 -8.24067593e-01 7.84928739e-01
-4.81422581e-02 1.19976327e-01 1.16931617e-01 6.96689785e-01
-4.34475541e-01 -1.83160082e-01 -3.24884176e-01 -6.63830280e-01
-1.71632946e+00 2.82113403e-01 5.37614107e-01 5.53353131e-01
-6.78565681e-01 7.75061667e-01 7.63819039e-01 -2.67226756e-01
2.13568673e-01 2.76010126e-01 -1.94015175e-01 -4.14990857e-02
9.16277230e-01 2.39591345e-01 -2.43862346e-01 -1.01215112e+00
-2.38252953e-01 1.07223439e+00 -1.12051629e-01 -1.98421746e-01
1.07038796e+00 -2.90093333e-01 -2.12386429e-01 4.89914604e-02
1.19282675e+00 4.03929085e-01 -1.61331952e+00 -2.48544868e-02
-6.20007575e-01 -1.22534871e+00 -1.40706643e-01 -2.23175302e-01
-1.73616362e+00 8.32995832e-01 5.76470017e-01 -3.46634150e-01
1.39808953e+00 -1.85877487e-01 7.94074118e-01 -7.80258328e-02
5.29586554e-01 -6.61429524e-01 2.80266106e-01 1.08135454e-01
1.10241926e+00 -9.74856496e-01 2.65101399e-02 -8.67929876e-01
-4.38455582e-01 9.57951665e-01 1.01758540e+00 1.88915417e-01
6.41611457e-01 4.51876104e-01 2.71398067e-01 -8.37504342e-02
-3.70797753e-01 1.64321363e-01 2.35032260e-01 6.97228789e-01
5.39042354e-01 -2.01657429e-01 -1.15360785e-03 2.33241335e-01
-5.05478621e-01 7.13467300e-02 4.31580245e-01 5.80220282e-01
-1.38967469e-01 -1.22161674e+00 -7.06848621e-01 7.42999045e-03
-4.58749712e-01 1.64811835e-01 -3.95530343e-01 7.88224757e-01
1.01025842e-01 8.41760278e-01 5.27227335e-02 -3.22409451e-01
5.32683551e-01 -2.95650095e-01 9.47553813e-01 -5.05827785e-01
-1.24199718e-01 3.83954823e-01 -2.89998829e-01 -6.61609054e-01
-5.56984067e-01 -5.15561104e-01 -1.00725687e+00 -7.88592637e-01
-3.24500114e-01 -9.36808884e-02 3.73397976e-01 5.20724654e-01
4.77081597e-01 1.43485934e-01 8.95640910e-01 -1.22586155e+00
-6.43202364e-02 -7.88552701e-01 -6.71397328e-01 5.08094847e-01
4.14331347e-01 -9.04307783e-01 -2.31996566e-01 1.89568549e-01]
|
[12.841500282287598, -0.2555060386657715]
|
97d3deba-1c2c-49e4-a868-9035f0050ae0
|
column-type-annotation-using-chatgpt
|
2306.00745
| null |
https://arxiv.org/abs/2306.00745v1
|
https://arxiv.org/pdf/2306.00745v1.pdf
|
Column Type Annotation using ChatGPT
|
Column type annotation is the task of annotating the columns of a relational table with the semantic type of the values contained in each column. Column type annotation is a crucial pre-processing step for data search and integration in the context of data lakes. State-of-the-art column type annotation methods either rely on matching table columns to properties of a knowledge graph or fine-tune pre-trained language models such as BERT for the column type annotation task. In this work, we take a different approach and explore using ChatGPT for column type annotation. We evaluate different prompt designs in zero- and few-shot settings and experiment with providing task definitions and detailed instructions to the model. We further implement a two-step table annotation pipeline which first determines the class of the entities described in the table and depending on this class asks ChatGPT to annotate columns using only the relevant subset of the overall vocabulary. Using instructions as well as the two-step pipeline, ChatGPT reaches F1 scores of over 85% in zero- and one-shot setups. To reach a similar F1 score a RoBERTa model needs to be fine-tuned with 300 examples. This comparison shows that ChatGPT is able deliver competitive results for the column type annotation task given no or only a minimal amount of task-specific demonstrations.
|
['Christian Bizer', 'Keti Korini']
|
2023-06-01
| null | null | null | null |
['table-annotation', 'table-annotation', 'column-type-annotation']
|
['knowledge-base', 'natural-language-processing', 'natural-language-processing']
|
[-1.68585964e-02 5.01353085e-01 -2.49379218e-01 -3.59032452e-01
-1.07078099e+00 -9.30832565e-01 5.85124195e-01 9.07026112e-01
-6.83434784e-01 6.50585055e-01 1.43129468e-01 -1.98815495e-01
-7.88560659e-02 -7.60676086e-01 -7.40850270e-01 4.24760319e-02
1.38961837e-01 1.15026355e+00 6.60302401e-01 -4.94310468e-01
2.62684882e-01 5.05749546e-02 -1.57258832e+00 8.69942725e-01
5.92257500e-01 1.05335784e+00 1.75959751e-01 7.83804536e-01
-6.59719050e-01 9.96029973e-01 -8.27298760e-01 -6.03224754e-01
1.41861215e-01 -1.75927907e-01 -1.27507067e+00 -3.00140619e-01
4.86426026e-01 1.72071662e-02 8.39153081e-02 6.74638927e-01
3.57626736e-01 2.09064215e-01 4.66254294e-01 -1.08287573e+00
-2.99786963e-02 1.27781045e+00 6.30090013e-02 1.37225136e-01
7.84566402e-01 2.27359325e-01 1.41434097e+00 -7.18966782e-01
1.06299150e+00 1.14510548e+00 5.95469058e-01 4.02426362e-01
-1.31549275e+00 -3.42964858e-01 1.23281598e-01 1.94823653e-01
-1.26289976e+00 -3.53981137e-01 1.15504973e-01 -5.60979187e-01
1.61888778e+00 2.95008093e-01 4.27367091e-01 7.57007599e-01
-2.74924934e-01 4.96308506e-01 8.06042075e-01 -5.66241324e-01
9.42111090e-02 5.80293477e-01 3.85316342e-01 6.94290936e-01
1.54855847e-01 -7.10174918e-01 -8.00876915e-01 1.11942336e-01
3.12866151e-01 -5.89243948e-01 1.40772760e-01 -4.36944723e-01
-1.39457715e+00 5.75089574e-01 2.41184905e-01 3.33430171e-01
-1.80287138e-01 1.20606475e-01 9.47756529e-01 3.76844049e-01
8.28804299e-02 1.04170716e+00 -8.09799016e-01 -5.11934042e-01
-5.69924593e-01 3.24917406e-01 1.39479887e+00 1.56024420e+00
7.84797430e-01 -6.12541020e-01 -6.80077136e-01 8.01886559e-01
-1.59231171e-01 -6.41283095e-02 3.29642326e-01 -7.81445682e-01
1.05276895e+00 1.06811798e+00 3.27032208e-01 -4.01559651e-01
-5.92543185e-01 -1.49038792e-01 7.05756573e-03 -9.63233933e-02
9.41902101e-01 -8.65964741e-02 -6.82568431e-01 1.48410392e+00
4.75936502e-01 -5.29575169e-01 2.59172529e-01 4.95058298e-01
1.16424012e+00 3.47296208e-01 2.03918308e-01 2.16349289e-02
2.00742054e+00 -8.93588126e-01 -7.92035997e-01 -2.93088764e-01
1.26050103e+00 -6.46468222e-01 1.58425283e+00 2.47500911e-01
-9.56124485e-01 -4.79821593e-01 -1.01040590e+00 -6.04409099e-01
-9.13671553e-01 3.39165002e-01 3.08064640e-01 3.46869648e-01
-5.49137414e-01 4.81069744e-01 -7.42125452e-01 -7.86440909e-01
8.39095861e-02 3.10843945e-01 -5.26101291e-01 2.37900037e-02
-1.12790227e+00 1.00309575e+00 6.55234456e-01 -3.58221054e-01
-5.68668902e-01 -9.62926686e-01 -1.04639757e+00 2.87329525e-01
1.15964317e+00 -5.29733419e-01 1.47571445e+00 -3.78556140e-02
-1.12527156e+00 1.02019000e+00 -2.25929413e-02 -6.73563540e-01
4.81706142e-01 -3.70930433e-01 -3.09543330e-02 -6.84411451e-02
2.96708912e-01 5.52271724e-01 9.36330184e-02 -8.33667159e-01
-7.69384146e-01 -1.11429296e-01 3.93507153e-01 1.22316904e-01
-4.48720604e-02 5.61645329e-02 -8.69800687e-01 -1.71539485e-01
-2.73006201e-01 -8.59409273e-01 1.80555824e-02 -3.17508787e-01
-7.35681951e-01 -6.10518157e-01 5.70938408e-01 -5.31785786e-01
1.64303994e+00 -1.95278215e+00 -8.18240941e-02 7.65670016e-02
1.08852899e-02 -2.48213187e-02 6.74343109e-02 8.56958210e-01
1.28241792e-01 3.11739385e-01 -5.41101545e-02 -3.95646334e-01
4.16756570e-01 2.33683184e-01 -1.17134757e-01 -2.08809972e-01
2.88501740e-01 8.02719414e-01 -8.01281571e-01 -5.91953993e-01
2.63702441e-02 6.39925003e-02 -6.79352462e-01 3.69892418e-01
-7.61841416e-01 1.72644436e-01 -3.01333874e-01 4.25078750e-01
-9.99811757e-03 -2.67747313e-01 4.22239959e-01 -5.19152641e-01
-2.36424953e-01 9.08478081e-01 -1.40137196e+00 1.80578256e+00
-5.59490323e-01 5.16657591e-01 -1.51049882e-01 -6.75241828e-01
6.85023189e-01 2.36700684e-01 1.80942520e-01 -5.74758053e-01
4.80628461e-02 3.39105487e-01 9.84127894e-02 -9.12151933e-01
6.27832234e-01 2.77588218e-01 -6.16772532e-01 2.85514295e-01
4.58854675e-01 -3.59057337e-01 9.75361526e-01 5.22340775e-01
1.42506647e+00 3.21637064e-01 5.47458291e-01 -1.87480271e-01
5.56083381e-01 5.18173814e-01 1.93969697e-01 8.14370871e-01
2.87049651e-01 3.28416497e-01 1.04586947e+00 -3.38993877e-01
-9.97663558e-01 -3.77270579e-01 7.00937733e-02 1.36821520e+00
-2.01468877e-02 -1.16789508e+00 -9.13261235e-01 -9.07341599e-01
-2.92605609e-02 1.09268808e+00 -8.28203976e-01 1.14043385e-01
-6.28712595e-01 -2.12711394e-01 5.14165044e-01 5.59638321e-01
4.07033712e-01 -1.13954699e+00 -8.71870756e-01 1.01584136e-01
-2.69222111e-01 -1.50325239e+00 -2.83569634e-01 7.54776955e-01
-2.51263559e-01 -1.30275977e+00 8.05083141e-02 -6.33580983e-01
3.85565251e-01 -4.90041405e-01 1.51790118e+00 1.55811101e-01
-2.02553913e-01 2.56903529e-01 -4.21731532e-01 -4.79161710e-01
-5.27707040e-01 6.26554489e-01 -4.83184487e-01 -4.26900566e-01
5.06783128e-01 -1.10977829e-01 -2.32421935e-01 3.22220117e-01
-5.61662734e-01 1.77516684e-01 4.41273898e-01 4.58633751e-01
5.59494555e-01 -1.90638378e-01 1.41443014e-01 -1.44959319e+00
6.17418587e-01 -1.83773175e-01 -8.48090291e-01 5.39108694e-01
-6.06399655e-01 5.86164892e-01 6.69474065e-01 -3.35154355e-01
-7.52604663e-01 2.94295698e-01 -7.90206045e-02 -9.58009213e-02
-9.79620516e-02 7.24799156e-01 -3.69864970e-01 4.72325265e-01
8.14047575e-01 -2.26399347e-01 -2.22518235e-01 -6.92041457e-01
5.17673314e-01 3.49741191e-01 5.28927565e-01 -7.97125161e-01
6.04439020e-01 -1.04301661e-01 -5.39538339e-02 -3.82156670e-01
-1.00290406e+00 -7.38908291e-01 -9.30932522e-01 5.64494133e-02
1.05666423e+00 -7.49985218e-01 -1.05211508e+00 -4.33721989e-02
-9.12384450e-01 -8.31110477e-01 -5.88746011e-01 1.92836031e-01
-5.70906997e-01 -1.66634083e-01 -4.90779161e-01 -5.94606102e-01
-1.61963761e-01 -1.22916162e+00 1.20953429e+00 -2.13810196e-03
-4.91318762e-01 -7.13176668e-01 -9.98837650e-02 3.29868078e-01
1.95785075e-01 1.01158684e-02 1.17284155e+00 -1.40866113e+00
-5.20952702e-01 -3.59271467e-01 -8.93569142e-02 -1.82206288e-01
-2.65390217e-01 -4.24084999e-02 -8.04840386e-01 2.15980679e-01
-6.08996511e-01 -6.07650459e-01 6.26995444e-01 -3.29067796e-01
9.83107805e-01 -4.45532560e-01 -4.52300161e-01 2.95918167e-01
1.29990017e+00 1.09960467e-01 4.00471687e-01 6.85508609e-01
7.24426270e-01 8.26754332e-01 9.66538012e-01 3.44019711e-01
6.54616296e-01 1.10003114e+00 3.35022599e-01 3.88001919e-01
-1.99480057e-01 -4.96906668e-01 1.86918247e-02 2.49000236e-01
3.55096132e-01 -3.93671542e-01 -1.06000268e+00 6.03484511e-01
-1.86388874e+00 -9.05498922e-01 -1.96781427e-01 2.26973510e+00
1.26681912e+00 5.51290870e-01 3.52977693e-01 1.13320857e-01
3.13220620e-01 -1.64860487e-01 -3.33695203e-01 -4.92550194e-01
1.69205025e-01 2.85049877e-03 4.78146851e-01 5.46865284e-01
-1.14209855e+00 1.18359971e+00 5.58267593e+00 8.24783862e-01
-7.33579397e-01 -9.89765376e-02 2.13338837e-01 -1.32990584e-01
-2.91439220e-02 2.14930207e-01 -1.28948843e+00 2.17538282e-01
1.03001213e+00 -2.64254436e-02 4.34467852e-01 7.69988358e-01
-1.65143430e-01 -3.69055629e-01 -1.55789006e+00 7.16099143e-01
-9.62450355e-02 -1.34733582e+00 -2.13903233e-01 -1.18598379e-01
1.97380409e-01 -3.11476197e-02 -4.88039047e-01 8.93351972e-01
4.68529612e-01 -8.36771131e-01 9.65700030e-01 4.45354134e-01
8.07547390e-01 -3.67016673e-01 7.52175570e-01 3.56581032e-01
-1.39028347e+00 -3.35315168e-02 -1.13663085e-01 6.09568618e-02
2.65706964e-02 3.94933149e-02 -1.25104237e+00 4.39856648e-01
8.42828572e-01 4.26488310e-01 -8.58957171e-01 8.89063597e-01
-4.65133607e-01 4.68169779e-01 -5.83471239e-01 -2.66177386e-01
5.35814799e-02 1.40485600e-01 3.99808586e-01 1.41746461e+00
7.31435940e-02 1.53261488e-02 3.84470910e-01 7.52928197e-01
-3.71864945e-01 1.84106559e-01 -2.72834390e-01 -2.21628681e-01
6.82036161e-01 1.49884832e+00 -8.32200766e-01 -7.60111272e-01
-4.11545038e-01 5.85275054e-01 6.84067070e-01 4.92177047e-02
-6.42638683e-01 -7.67412603e-01 3.94226968e-01 3.65014136e-01
5.73467612e-01 1.37246819e-02 -3.53263259e-01 -8.29777658e-01
1.37080118e-01 -8.09845388e-01 7.88663983e-01 -1.00016999e+00
-8.01095247e-01 6.36667430e-01 4.50027794e-01 -1.00252533e+00
-4.61553872e-01 -6.23802066e-01 -4.27447796e-01 6.81629896e-01
-1.11355054e+00 -1.10948431e+00 -5.63178480e-01 3.24092209e-01
6.39263570e-01 1.29995510e-01 1.03851926e+00 1.14877023e-01
-6.18947566e-01 5.69988906e-01 -4.54091519e-01 5.13323724e-01
9.53857422e-01 -1.86358523e+00 4.96987224e-01 6.76331401e-01
2.01391801e-01 8.28463197e-01 1.02369452e+00 -6.92725658e-01
-1.35175169e+00 -9.41121519e-01 1.12612355e+00 -8.31267655e-01
8.45750451e-01 -8.61336052e-01 -1.02368939e+00 8.69644463e-01
2.98529893e-01 6.41411990e-02 5.09582698e-01 4.40449357e-01
-6.76805139e-01 -1.29548926e-02 -9.84305799e-01 3.29858035e-01
1.01190460e+00 -6.60939395e-01 -7.22360134e-01 4.84784037e-01
9.42585468e-01 -8.47537041e-01 -1.11920631e+00 9.06051472e-02
4.84430760e-01 -6.13881648e-01 6.08740389e-01 -6.81315720e-01
3.09745491e-01 -3.83992106e-01 -4.38624732e-02 -1.14941382e+00
8.85241404e-02 -5.59851408e-01 -7.35042393e-02 1.45835972e+00
1.00498307e+00 -1.40467554e-01 5.32346129e-01 7.58018672e-01
-3.34870458e-01 -7.00734317e-01 -5.83634019e-01 -6.47974908e-01
-3.05353791e-01 -4.02726978e-01 5.51458776e-01 5.94323814e-01
4.92806822e-01 8.80487859e-01 1.17239714e-01 -1.19631417e-01
4.60530333e-02 -4.14966084e-02 1.15612161e+00 -1.26647282e+00
-2.69698471e-01 -3.69758844e-01 4.02394868e-02 -6.86134756e-01
-4.19801325e-02 -9.24596429e-01 1.91109598e-01 -2.00315595e+00
2.08934397e-02 -5.51502347e-01 2.17135604e-02 8.10737193e-01
-1.15168549e-01 1.90326720e-02 1.95114151e-01 1.60130009e-01
-1.02100646e+00 -1.57638073e-01 7.49346793e-01 -1.23346604e-01
-3.44294697e-01 -1.42706975e-01 -5.73366582e-01 3.65147799e-01
5.27974427e-01 -5.26872218e-01 -3.62665802e-01 -2.86228687e-01
6.34739876e-01 -9.71200876e-03 1.14065027e-02 -1.15165854e+00
4.55353290e-01 6.42367918e-03 6.69569597e-02 -5.85759997e-01
4.52115595e-01 -8.10995221e-01 -6.89407215e-02 3.16687167e-01
-8.42353106e-01 1.80836871e-01 3.83618265e-01 2.80063838e-01
-4.38537002e-02 -4.30707753e-01 4.41871732e-01 -3.43236715e-01
-7.99324572e-01 -1.32997930e-01 -3.79540592e-01 4.65834677e-01
8.17329466e-01 -1.03419729e-01 -5.82803071e-01 -4.10002023e-02
-8.65502357e-01 4.48262036e-01 3.49676013e-01 4.39486712e-01
-5.16186953e-02 -7.87159264e-01 -3.08489889e-01 3.83657925e-02
6.96742415e-01 2.16888368e-01 -1.79504171e-01 9.58640158e-01
-5.05820215e-01 5.87050557e-01 -1.33913606e-01 -4.68931913e-01
-1.36792994e+00 7.22054124e-01 3.31455112e-01 -7.58985400e-01
-5.19907176e-01 9.05261874e-01 -1.70029402e-01 -5.19343495e-01
4.21785712e-01 -7.77127624e-01 -5.75566649e-01 5.18442392e-01
4.67131346e-01 4.75857779e-03 5.16689956e-01 -3.43815267e-01
-4.97690767e-01 3.94979626e-01 -1.13141619e-01 -1.85500666e-01
1.11090660e+00 1.43605143e-01 1.33565649e-01 8.64778936e-01
9.38468933e-01 2.83012897e-01 -7.93798625e-01 -1.39105156e-01
5.38336098e-01 -8.21764302e-03 -4.13467199e-01 -1.26758564e+00
-5.16851485e-01 7.33932793e-01 1.01172797e-01 4.45372045e-01
6.23298824e-01 2.91873783e-01 3.38986933e-01 7.48227060e-01
3.22372258e-01 -1.14920866e+00 1.43698737e-01 7.09487081e-01
8.40551436e-01 -1.25021946e+00 -2.32839826e-02 -6.79808915e-01
-7.75141180e-01 1.08151758e+00 9.38095689e-01 2.18217432e-01
2.68292427e-01 4.41957712e-01 -3.27595621e-02 -4.90108460e-01
-1.23643494e+00 -5.81450641e-01 4.60247070e-01 4.34412748e-01
7.77258217e-01 -2.26444945e-01 -4.01847690e-01 8.00894558e-01
-5.14727890e-01 -1.78787708e-01 3.83619457e-01 1.00179374e+00
-4.64848697e-01 -1.21537876e+00 1.34822484e-02 5.39930701e-01
-4.18576211e-01 -1.63705394e-01 -5.89394808e-01 1.25858188e+00
-1.13298178e-01 8.34295213e-01 2.03533500e-01 -4.35534269e-01
8.68143976e-01 3.95205021e-01 3.47235918e-01 -1.20728850e+00
-1.33731318e+00 8.38561729e-02 7.96879709e-01 -7.43416011e-01
-6.19181469e-02 -5.47872782e-01 -1.57029939e+00 1.63150728e-01
-3.06662858e-01 4.73199189e-01 6.38966322e-01 1.04588163e+00
3.27252299e-01 6.54323936e-01 -3.45949203e-01 -3.06078613e-01
-4.37553883e-01 -1.28022647e+00 9.36263334e-03 5.81174970e-01
-4.65066172e-02 -7.03667581e-01 3.50878462e-02 5.09763844e-02]
|
[9.650516510009766, 8.292823791503906]
|
ac428919-7d57-484c-9a6c-983925bfb997
|
plop-learning-without-forgetting-for
|
2011.11390
| null |
https://arxiv.org/abs/2011.11390v3
|
https://arxiv.org/pdf/2011.11390v3.pdf
|
PLOP: Learning without Forgetting for Continual Semantic Segmentation
|
Deep learning approaches are nowadays ubiquitously used to tackle computer vision tasks such as semantic segmentation, requiring large datasets and substantial computational power. Continual learning for semantic segmentation (CSS) is an emerging trend that consists in updating an old model by sequentially adding new classes. However, continual learning methods are usually prone to catastrophic forgetting. This issue is further aggravated in CSS where, at each step, old classes from previous iterations are collapsed into the background. In this paper, we propose Local POD, a multi-scale pooling distillation scheme that preserves long- and short-range spatial relationships at feature level. Furthermore, we design an entropy-based pseudo-labelling of the background w.r.t. classes predicted by the old model to deal with background shift and avoid catastrophic forgetting of the old classes. Our approach, called PLOP, significantly outperforms state-of-the-art methods in existing CSS scenarios, as well as in newly proposed challenging benchmarks.
|
['Matthieu Cord', 'Arnaud Dapogny', 'Yifu Chen', 'Arthur Douillard']
|
2020-11-23
| null |
http://openaccess.thecvf.com//content/CVPR2021/html/Douillard_PLOP_Learning_Without_Forgetting_for_Continual_Semantic_Segmentation_CVPR_2021_paper.html
|
http://openaccess.thecvf.com//content/CVPR2021/papers/Douillard_PLOP_Learning_Without_Forgetting_for_Continual_Semantic_Segmentation_CVPR_2021_paper.pdf
|
cvpr-2021-1
|
['overlapped-100-5', 'overlapped-10-1', 'disjoint-15-5', 'disjoint-10-1', 'disjoint-15-1', 'overlapped-15-5', 'overlapped-100-10', 'overlapped-15-1', 'overlapped-50-50', 'overlapped-100-50', 'continual-semantic-segmentation']
|
['computer-vision', 'computer-vision', 'computer-vision', 'computer-vision', 'computer-vision', 'computer-vision', 'computer-vision', 'computer-vision', 'computer-vision', 'computer-vision', 'computer-vision']
|
[ 5.67044199e-01 4.51396368e-02 1.14973485e-01 -3.75278622e-01
-5.35152972e-01 -4.71720129e-01 4.51146305e-01 5.38508952e-01
-8.29371095e-01 9.01234090e-01 -2.40897313e-01 2.58762062e-01
1.84870332e-01 -8.58377397e-01 -8.24389756e-01 -9.34611499e-01
3.09383541e-01 3.88392299e-01 1.06574047e+00 1.63986266e-01
1.66638762e-01 3.74557078e-01 -1.62690675e+00 3.18759769e-01
1.11872637e+00 1.06412530e+00 5.10292470e-01 3.53763878e-01
-4.14246023e-01 6.73233807e-01 -4.77447033e-01 -5.61681092e-01
1.41701683e-01 -3.85775030e-01 -8.17432523e-01 1.91937134e-01
6.90082192e-01 1.88953299e-02 -2.89485723e-01 1.26829278e+00
2.90743530e-01 3.98460478e-01 3.56762648e-01 -9.31711018e-01
-3.32984030e-01 3.69273067e-01 -8.33205700e-01 2.47093216e-01
-3.59655946e-01 1.42667606e-01 7.27006614e-01 -7.38587976e-01
8.19514215e-01 1.07337737e+00 7.92510211e-01 6.04669571e-01
-1.29003453e+00 -4.82471824e-01 5.80153108e-01 3.53180408e-01
-1.31533337e+00 -4.15252112e-02 7.07879841e-01 -3.48145485e-01
5.62131524e-01 2.04530880e-01 8.10533106e-01 9.11044180e-01
2.50810355e-01 1.13892531e+00 1.12207329e+00 -7.49138966e-02
5.63548803e-01 1.24744728e-01 2.68406719e-01 6.52520716e-01
4.21534300e-01 -4.07619178e-01 -5.35112560e-01 1.76018968e-01
4.37849283e-01 5.98880164e-02 -9.43434164e-02 -6.78054452e-01
-9.31441724e-01 5.12111843e-01 6.07269645e-01 2.86614984e-01
-3.00324917e-01 7.55415708e-02 3.44549268e-01 -2.19215438e-01
6.59486711e-01 2.93815523e-01 -7.34009385e-01 1.49811998e-01
-1.29453325e+00 2.82088667e-01 3.80753964e-01 6.17582858e-01
1.03866982e+00 -3.15421194e-01 -3.84869188e-01 8.31228018e-01
-2.23304685e-02 6.58115819e-02 6.56863034e-01 -8.03686082e-01
1.64513394e-01 7.29293287e-01 6.84389398e-02 -8.57443571e-01
-4.56077397e-01 -1.00963259e+00 -8.97607028e-01 1.60401508e-01
4.59217131e-01 1.73416063e-01 -1.41853416e+00 1.58656919e+00
5.30666471e-01 7.17938244e-01 1.02531455e-01 5.61966181e-01
5.35116255e-01 7.01343179e-01 3.65588337e-01 -2.45455295e-01
1.03220618e+00 -1.36959231e+00 -5.13107777e-01 -6.63701236e-01
1.98983133e-01 -4.47588772e-01 9.36649978e-01 4.82819885e-01
-9.27391887e-01 -7.41807342e-01 -8.37332785e-01 -1.25614345e-01
-5.23564517e-01 8.56425613e-02 5.22355914e-01 5.05737424e-01
-9.09118712e-01 1.02083170e+00 -1.15536332e+00 -2.79165745e-01
1.02216160e+00 2.31648713e-01 -1.93321541e-01 -1.34677216e-01
-9.28306460e-01 7.15731561e-01 8.73277605e-01 1.18023947e-01
-7.60872483e-01 -7.61576414e-01 -6.62026644e-01 1.04439579e-01
5.97711265e-01 -5.02653301e-01 1.03950489e+00 -9.61414754e-01
-1.32457507e+00 8.42639387e-01 -1.35173947e-01 -6.94028914e-01
8.06382358e-01 -5.15509427e-01 3.71477683e-03 -1.12725906e-02
2.84054969e-02 9.82690156e-01 1.10759819e+00 -1.44721532e+00
-8.88344765e-01 -4.99212593e-01 -2.57685810e-01 3.15949112e-01
-3.75977427e-01 -5.87767005e-01 -7.26353288e-01 -8.27598274e-01
4.95110184e-01 -8.87665570e-01 -5.74357152e-01 2.04271004e-01
-3.31115723e-01 -7.75047317e-02 9.89680052e-01 -8.06349993e-01
1.19198489e+00 -2.15941262e+00 4.15233791e-01 -1.47157311e-01
1.46679386e-01 7.63714314e-01 9.94258374e-03 -3.45567256e-01
2.71300137e-01 -5.67200407e-02 -8.58931899e-01 -6.99909627e-01
-1.64609358e-01 3.67318869e-01 -1.64028227e-01 2.92821676e-01
3.45013261e-01 7.65181661e-01 -9.73336220e-01 -5.46167731e-01
3.41527551e-01 4.25209224e-01 -4.83881295e-01 -4.20708060e-02
-4.71539646e-01 5.95672131e-01 -1.74404025e-01 4.15263027e-01
1.04079413e+00 -1.49195075e-01 -1.62180513e-01 -1.75839551e-02
-1.47194043e-01 -2.05440313e-01 -1.10676360e+00 2.04062653e+00
-1.41593710e-01 3.64475608e-01 -1.87530398e-01 -9.97995436e-01
7.68174171e-01 -2.33515799e-01 2.84463584e-01 -5.74419379e-01
-3.81089188e-02 3.96751136e-01 -3.36234689e-01 -1.83909640e-01
5.58575392e-01 -5.41506112e-02 6.73041567e-02 -1.61628097e-01
2.38360703e-01 -3.15085173e-01 2.11893812e-01 -6.50417246e-03
9.12369907e-01 3.65890205e-01 1.80762976e-01 -2.02256218e-01
6.82797492e-01 -1.08681567e-01 1.14417505e+00 8.66154373e-01
-4.20934081e-01 7.77455270e-01 3.98294121e-01 -5.55448830e-01
-8.27950776e-01 -1.10810709e+00 -1.04821004e-01 6.99869394e-01
4.76273894e-01 -4.66625206e-02 -9.58813190e-01 -9.39340949e-01
-1.62849799e-02 8.09500813e-01 -5.45973480e-01 -3.89676720e-01
-6.85479403e-01 -1.11625290e+00 2.10353985e-01 4.57083076e-01
9.60523129e-01 -1.13756454e+00 -8.13366830e-01 6.13908470e-01
-1.81864858e-01 -1.20407557e+00 -2.30475545e-01 2.82345474e-01
-1.11682940e+00 -8.98742318e-01 -1.06904459e+00 -7.15135276e-01
6.20940208e-01 2.00483724e-01 9.11781132e-01 -1.51740119e-01
-5.92345119e-01 -1.06886029e-01 -2.04491690e-01 -2.07171991e-01
-2.11940110e-01 4.15022761e-01 -3.14308137e-01 3.65609288e-01
2.00761780e-01 -4.19390231e-01 -7.59211481e-01 9.80760977e-02
-1.15774024e+00 3.33483845e-01 6.08845651e-01 7.79328883e-01
9.89966631e-01 3.69323492e-01 6.08710170e-01 -9.35040951e-01
-3.86324935e-02 -2.50228316e-01 -6.66300535e-01 3.94881517e-01
-3.39139611e-01 1.13909677e-01 3.86717409e-01 -3.94719779e-01
-1.43503273e+00 2.44576603e-01 -2.59821147e-01 -1.92585781e-01
-1.98228851e-01 1.51919618e-01 -2.50545084e-01 -1.89531878e-01
3.99849594e-01 3.09824884e-01 -4.58318174e-01 -7.22163260e-01
2.66715467e-01 3.14337343e-01 7.37179220e-01 -3.67747277e-01
5.57239473e-01 6.50216818e-01 -6.16768785e-02 -6.64301038e-01
-1.28149712e+00 -5.04278362e-01 -9.83411551e-01 -1.72315374e-01
1.05737782e+00 -8.01848590e-01 2.25700978e-02 1.23406422e+00
-1.12597728e+00 -4.55656558e-01 -5.88286638e-01 1.13639325e-01
-4.79216218e-01 4.20492411e-01 -4.15898353e-01 -6.06932521e-01
-1.62857369e-01 -9.59570706e-01 1.06059003e+00 7.55619645e-01
1.33199424e-01 -8.02941978e-01 2.29261983e-02 4.31684077e-01
3.35379511e-01 5.48108697e-01 7.52729177e-01 -4.12726551e-01
-6.96254373e-01 -1.57996356e-01 -3.69433165e-01 7.06931949e-01
1.09014414e-01 -2.88045198e-01 -1.05243790e+00 -4.27154005e-01
-2.02008158e-01 -1.84744701e-01 1.67198312e+00 2.68658996e-01
1.54933190e+00 -3.07524409e-02 -5.89698493e-01 5.27774751e-01
1.37910688e+00 1.67934850e-01 6.10777676e-01 3.22569251e-01
8.18793535e-01 5.73413968e-01 6.51941836e-01 3.34864199e-01
2.31734335e-01 4.87516284e-01 4.03892487e-01 -1.30427880e-02
-4.03358161e-01 -2.28007168e-01 1.16112027e-02 4.85629588e-01
2.40154058e-01 -2.53357977e-01 -8.91630709e-01 6.30128443e-01
-1.95725322e+00 -6.18039012e-01 6.27259314e-02 2.38689065e+00
9.28283393e-01 5.28438509e-01 -3.63384038e-01 5.00095859e-02
8.23940694e-01 2.58790761e-01 -9.33560312e-01 1.72101513e-01
-4.55072075e-01 4.74547684e-01 5.09805202e-01 4.22619343e-01
-1.49784482e+00 1.41921437e+00 4.89817286e+00 1.07947922e+00
-1.08570588e+00 3.79730135e-01 9.78340983e-01 1.07070785e-02
7.81172439e-02 -4.04378809e-02 -8.77765775e-01 5.53005397e-01
4.08199102e-01 1.97061330e-01 1.14958845e-01 7.57677674e-01
-1.53148606e-01 -5.86159945e-01 -6.53758705e-01 8.28227937e-01
4.45567258e-02 -1.35255766e+00 2.88740080e-02 -3.52038980e-01
1.13925731e+00 4.85471748e-02 3.52690741e-02 3.09101462e-01
7.10773692e-02 -5.24153590e-01 8.13939750e-01 5.21574318e-01
3.58809799e-01 -7.37632930e-01 7.82501280e-01 3.23670954e-01
-1.12589514e+00 -2.81992137e-01 -3.99280101e-01 3.27930808e-01
1.79860368e-01 9.36163247e-01 -3.09522063e-01 5.63123703e-01
9.40640211e-01 7.61292577e-01 -9.27352190e-01 1.44920743e+00
-2.87777781e-01 4.74374712e-01 -3.18742424e-01 3.46355289e-01
4.15874124e-01 -1.44644780e-02 5.65935254e-01 1.13392806e+00
3.75668585e-01 -1.28890276e-01 1.61577106e-01 6.78923965e-01
-1.21329486e-01 -5.36834486e-02 -2.95605808e-02 1.36523753e-01
1.07401907e-01 1.10593426e+00 -1.34785450e+00 -5.77470958e-01
-7.52873868e-02 1.48925507e+00 3.33551675e-01 2.33703226e-01
-7.39301205e-01 -2.83709288e-01 4.94624615e-01 -3.37378234e-02
5.71252525e-01 -1.81436718e-01 -2.65528232e-01 -1.27470839e+00
1.93005279e-01 -3.01839292e-01 4.74507213e-01 -5.68511188e-01
-1.01339245e+00 6.24757349e-01 -1.10488914e-01 -7.74399400e-01
4.83619958e-01 -4.43017721e-01 -4.70324725e-01 4.79454786e-01
-1.77181697e+00 -1.05145895e+00 -4.70079452e-01 3.12819868e-01
9.02425408e-01 1.85019404e-01 4.57126170e-01 4.09920782e-01
-6.91246212e-01 3.55089128e-01 2.48594180e-01 -3.05017799e-01
5.79690754e-01 -1.27155995e+00 4.91467059e-01 9.76706624e-01
9.37580541e-02 4.74023670e-02 6.37479842e-01 -8.55678320e-01
-4.43817258e-01 -1.57971239e+00 7.71840334e-01 -3.01759392e-02
3.99241656e-01 -4.15509284e-01 -1.36368644e+00 2.75249779e-01
-4.04279456e-02 4.27421033e-01 2.56235659e-01 -3.41242433e-01
-1.00664862e-01 -3.33361953e-01 -1.36604440e+00 4.90485102e-01
1.06935143e+00 -1.14464715e-01 -3.92260790e-01 1.83602810e-01
7.06013143e-01 -4.43314075e-01 -2.54158258e-01 6.37806535e-01
2.89158344e-01 -9.06145453e-01 8.88130784e-01 -3.36971223e-01
4.54097502e-02 -5.43037355e-01 8.52220207e-02 -1.37446272e+00
-1.74031168e-01 -6.66979194e-01 -2.52572656e-01 1.18509984e+00
8.86507928e-02 -5.20554900e-01 1.03970969e+00 3.17343593e-01
-2.82626748e-01 -5.99238515e-01 -1.30014932e+00 -8.47945154e-01
1.58429533e-01 -4.14290801e-02 4.17067081e-01 7.86953866e-01
-7.77574718e-01 -4.40752208e-02 -2.07282320e-01 2.36607715e-01
8.69822800e-01 -4.36239392e-02 4.40793306e-01 -1.41422737e+00
-2.09890649e-01 -4.83013630e-01 -6.00367069e-01 -8.54699075e-01
-1.32520767e-02 -6.38528764e-01 2.19110832e-01 -1.58145297e+00
2.84129173e-01 -5.94645917e-01 -6.98245466e-01 5.09395838e-01
-5.71574152e-01 5.03520608e-01 2.95694083e-01 1.68586206e-02
-8.11462343e-01 8.21232498e-01 1.21167064e+00 -2.68073142e-01
-4.25148427e-01 1.57876119e-01 -3.30237180e-01 9.39842284e-01
1.00540113e+00 -6.43862367e-01 -2.97057927e-01 -3.40219647e-01
3.43554392e-02 -4.49017525e-01 5.42534232e-01 -1.47930849e+00
1.93078384e-01 -4.79579950e-03 4.13823634e-01 -7.64275432e-01
2.92506695e-01 -5.89152873e-01 8.37009475e-02 5.78469872e-01
-1.45428285e-01 -3.97930235e-01 5.28599858e-01 8.49634647e-01
-2.52174109e-01 -3.66014689e-01 1.34167695e+00 -2.96800107e-01
-1.00668597e+00 4.31980103e-01 -2.56847858e-01 7.91652054e-02
1.25310564e+00 -1.74176872e-01 -8.12560320e-02 2.84350365e-01
-1.14005053e+00 2.57112056e-01 3.78866524e-01 5.63500702e-01
4.23301160e-01 -9.56203997e-01 -4.14966375e-01 2.01106910e-02
-5.49801365e-02 5.98284960e-01 6.63653076e-01 6.10303402e-01
-6.34379148e-01 9.09250304e-02 -1.65509760e-01 -5.97229838e-01
-1.16790199e+00 4.85567927e-01 3.74277800e-01 -4.94265050e-01
-7.90409923e-01 1.27239895e+00 4.19888496e-01 -2.78113008e-01
3.31887037e-01 -3.90737474e-01 -1.13729544e-01 2.23977149e-01
3.30789715e-01 2.75239170e-01 1.74975783e-01 -4.57264364e-01
-2.84105211e-01 5.86057484e-01 -3.41969252e-01 1.01470843e-01
1.41791558e+00 -2.82547653e-01 -2.00900465e-01 5.04878342e-01
8.95838916e-01 -4.32876199e-01 -1.91561115e+00 -4.27984565e-01
2.66750515e-01 -4.61747140e-01 1.50989175e-01 -9.10736561e-01
-1.27995372e+00 1.06092668e+00 9.20499086e-01 -2.40130261e-01
1.21923614e+00 -3.99941131e-02 1.08467174e+00 3.26577902e-01
3.44956964e-01 -1.44627202e+00 2.54844040e-01 3.70035768e-01
5.27930856e-01 -1.21311700e+00 6.10874668e-02 -4.68550086e-01
-6.69736981e-01 9.15171027e-01 8.12565386e-01 1.30166952e-02
5.77790558e-01 -4.65342738e-02 -1.93435207e-01 1.29415050e-01
-2.83177048e-01 -2.49367520e-01 1.06268547e-01 3.52014422e-01
-1.74231276e-01 -6.46034554e-02 -1.16064727e-01 3.74214143e-01
2.37641543e-01 4.24425490e-02 3.59627187e-01 1.09790504e+00
-6.63283467e-01 -1.02630055e+00 -1.64614007e-01 3.33152801e-01
-2.92216241e-01 3.10275666e-02 -2.30934188e-01 4.23323184e-01
8.23614836e-01 4.72486734e-01 1.36789307e-01 -9.02872905e-03
1.24718659e-01 2.22290352e-01 4.73008096e-01 -5.57551742e-01
-3.42958272e-01 4.71550040e-02 -4.16436493e-01 -4.75389898e-01
-4.60388809e-01 -1.01934123e+00 -1.29635453e+00 2.62341559e-01
-3.25551003e-01 -6.46832734e-02 5.39244294e-01 7.98408628e-01
3.65740031e-01 7.24451184e-01 2.34085560e-01 -8.87074828e-01
-2.44459420e-01 -7.01095223e-01 -4.85285759e-01 2.40244210e-01
4.30802077e-01 -8.15813959e-01 -2.04240486e-01 1.60754874e-01]
|
[9.40975570678711, 1.94631028175354]
|
b34db5bd-5486-47b0-8d15-2faec7cc9e31
|
massive-online-crowdsourced-study-of
|
1511.02919
| null |
http://arxiv.org/abs/1511.02919v1
|
http://arxiv.org/pdf/1511.02919v1.pdf
|
Massive Online Crowdsourced Study of Subjective and Objective Picture Quality
|
Most publicly available image quality databases have been created under
highly controlled conditions by introducing graded simulated distortions onto
high-quality photographs. However, images captured using typical real-world
mobile camera devices are usually afflicted by complex mixtures of multiple
distortions, which are not necessarily well-modeled by the synthetic
distortions found in existing databases. The originators of existing legacy
databases usually conducted human psychometric studies to obtain statistically
meaningful sets of human opinion scores on images in a stringently controlled
visual environment, resulting in small data collections relative to other kinds
of image analysis databases. Towards overcoming these limitations, we designed
and created a new database that we call the LIVE In the Wild Image Quality
Challenge Database, which contains widely diverse authentic image distortions
on a large number of images captured using a representative variety of modern
mobile devices. We also designed and implemented a new online crowdsourcing
system, which we have used to conduct a very large-scale, multi-month image
quality assessment subjective study. Our database consists of over 350000
opinion scores on 1162 images evaluated by over 7000 unique human observers.
Despite the lack of control over the experimental environments of the numerous
study participants, we demonstrate excellent internal consistency of the
subjective dataset. We also evaluate several top-performing blind Image Quality
Assessment algorithms on it and present insights on how mixtures of distortions
challenge both end users as well as automatic perceptual quality prediction
models.
|
['Alan C. Bovik', 'Deepti Ghadiyaram']
|
2015-11-09
| null | null | null | null |
['blind-image-quality-assessment']
|
['computer-vision']
|
[ 8.25409740e-02 -5.97658753e-01 1.88630804e-01 -3.96951914e-01
-1.14636183e+00 -9.06085312e-01 4.59853113e-01 -1.24235883e-01
-6.51504338e-01 4.63021576e-01 3.77104700e-01 -2.31086850e-01
9.40838456e-02 -3.26397330e-01 -6.75492048e-01 -2.90498435e-01
7.20120221e-02 1.00677721e-01 2.67936975e-01 -1.45951286e-01
2.81803966e-01 1.81602597e-01 -1.83067095e+00 2.64055610e-01
1.04693270e+00 1.00773704e+00 -1.37529187e-02 1.04777455e+00
6.52774751e-01 6.11287773e-01 -1.04977584e+00 -1.12287331e+00
8.07069838e-01 -2.49968320e-01 -2.53944695e-01 6.12311602e-01
1.00389254e+00 -8.25574934e-01 -3.67444068e-01 1.21055353e+00
1.00875378e+00 -2.43133157e-02 2.34756216e-01 -1.36382973e+00
-1.12979758e+00 -3.11765730e-01 -3.34050953e-01 4.87273723e-01
9.36553180e-01 1.05707026e+00 8.30965042e-01 -8.04546058e-01
5.73211432e-01 1.08740985e+00 7.13692486e-01 2.21813709e-01
-1.27449358e+00 -5.88652730e-01 -4.99021649e-01 3.80833060e-01
-1.34138072e+00 -1.04158163e+00 2.61757612e-01 -3.91092837e-01
7.34373391e-01 1.85642019e-01 7.19989896e-01 1.14516008e+00
1.42611057e-01 2.12046042e-01 1.64854181e+00 -2.69138008e-01
2.85478622e-01 5.19357026e-01 -4.31870401e-01 3.58162791e-01
1.84279889e-01 2.51191974e-01 -5.73656559e-01 -5.30604661e-01
7.08476722e-01 -3.65370303e-01 -6.24317765e-01 -3.89736563e-01
-1.22578585e+00 4.11763638e-01 3.02003950e-01 -1.53018653e-01
-3.37108642e-01 -2.81261832e-01 1.04125880e-01 4.99963373e-01
2.50406355e-01 3.33142191e-01 -1.92877069e-01 -2.68802434e-01
-8.49159598e-01 2.45347008e-01 5.51751196e-01 9.90136266e-01
5.02088845e-01 -2.28049010e-01 -1.81157961e-01 1.07036388e+00
1.86057717e-01 8.63096833e-01 5.74885368e-01 -1.43875003e+00
4.52447116e-01 2.29698956e-01 8.17004681e-01 -1.40398502e+00
3.01299356e-02 -2.71368865e-02 -4.29701746e-01 6.81203485e-01
7.54920542e-01 6.46452159e-02 -8.02537322e-01 1.47210073e+00
6.88225701e-02 -2.93795317e-01 -2.69896358e-01 1.23735559e+00
6.18292868e-01 2.25223318e-01 1.09750547e-01 -2.34918997e-01
1.48139966e+00 -6.01986587e-01 -5.92400014e-01 -4.14240547e-02
-1.31856706e-02 -1.17170644e+00 1.75832784e+00 7.73058951e-01
-1.59057307e+00 -7.74294794e-01 -1.04734993e+00 -1.77622251e-02
-1.22701399e-01 -1.27937227e-01 2.29437560e-01 1.21249127e+00
-1.57682014e+00 1.39081776e-01 -7.90759251e-02 -3.67618114e-01
5.56661606e-01 6.73162714e-02 -6.26966357e-01 -5.42053759e-01
-9.56671059e-01 9.43960428e-01 -4.72391456e-01 -2.34283924e-01
-9.94185686e-01 -5.11623859e-01 -6.64729178e-01 -4.44442004e-01
2.23770037e-01 -7.09159553e-01 1.41096151e+00 -1.36888599e+00
-1.25794041e+00 1.32355201e+00 -3.55922312e-01 -2.34863862e-01
7.00262904e-01 4.82252473e-03 -1.01888537e+00 4.57661301e-01
2.50421315e-01 5.02803385e-01 1.00975299e+00 -1.60610461e+00
-4.65915680e-01 -2.79779106e-01 2.57931441e-01 3.26254666e-01
-2.19468400e-01 4.76290017e-01 -8.72144222e-01 -7.10791111e-01
-3.27901125e-01 -7.20470309e-01 4.80831899e-02 2.92051345e-01
9.24988240e-02 5.19157469e-01 1.17449716e-01 -8.72537673e-01
1.08654547e+00 -2.10723853e+00 -3.55291128e-01 3.35725471e-02
4.53394353e-01 4.77007270e-01 -3.84585708e-01 1.83700293e-01
1.55044720e-01 2.35792965e-01 -1.65569596e-02 -3.53339911e-01
4.93764654e-02 -5.46899959e-02 -3.12232375e-02 7.62281358e-01
-1.11620136e-01 7.53714859e-01 -9.98264074e-01 -5.36640167e-01
2.05442235e-01 1.79304332e-01 -5.55713236e-01 4.01002526e-01
2.40335301e-01 1.58749044e-01 3.52486700e-01 9.37735260e-01
8.50476801e-01 -1.91404805e-01 -2.14299321e-01 -2.84191549e-01
2.30371341e-01 -2.20711455e-01 -1.06770229e+00 1.63258684e+00
-2.99376816e-01 7.08924651e-01 7.57935271e-03 2.36029029e-02
3.77921432e-01 4.41631734e-01 1.92989245e-01 -1.19513595e+00
3.90249081e-02 2.57740140e-01 1.64406165e-03 -8.00446212e-01
7.63355613e-01 -1.20653853e-01 2.35884875e-01 3.57954204e-01
-3.13288905e-02 -4.31873053e-01 2.29741856e-01 2.08727613e-01
1.12730908e+00 -5.21981120e-01 1.65465608e-01 1.85441956e-01
3.41825604e-01 1.90372299e-02 4.73794013e-01 7.71076560e-01
-1.14012766e+00 1.15542734e+00 1.19471848e-01 -2.70343184e-01
-1.49312365e+00 -1.49394500e+00 -3.49930942e-01 7.46273160e-01
3.39096457e-01 -4.16411012e-01 -7.76281655e-01 -2.71515876e-01
-2.46172488e-01 6.15552925e-02 -2.93245345e-01 8.38821009e-02
1.56044811e-01 -7.73106873e-01 7.56050467e-01 1.07774228e-01
7.10502803e-01 -8.50394368e-01 -3.18596154e-01 -1.38817862e-01
-2.78802663e-01 -1.41692710e+00 -8.14321637e-01 -8.09036970e-01
-2.66614556e-01 -1.49364972e+00 -1.01703203e+00 -5.59406102e-01
4.52114910e-01 7.18730390e-01 1.57910192e+00 6.76283240e-02
-3.46629918e-01 7.63875961e-01 -4.11447525e-01 -2.62871981e-01
-4.42562312e-01 -8.07188213e-01 4.28733736e-01 1.65993601e-01
4.03372884e-01 -3.17401260e-01 -1.05678737e+00 9.78201807e-01
-1.10459578e+00 -5.20546257e-01 4.66965675e-01 5.85619211e-01
4.12911654e-01 3.66203099e-01 4.18512017e-01 -1.41001135e-01
1.09977460e+00 -3.50301981e-01 -6.40301406e-01 2.30393469e-01
-5.49762487e-01 -6.66793883e-01 3.74859184e-01 -5.08578420e-01
-9.97691333e-01 -5.83140433e-01 1.89749315e-01 -2.80025691e-01
-3.28820974e-01 -1.13185495e-01 -3.51247728e-01 -4.70575511e-01
1.08944285e+00 -4.30939756e-02 3.44827920e-02 -1.34592965e-01
3.80891949e-01 9.67614949e-01 9.51629102e-01 -3.67455274e-01
9.43875730e-01 4.72604990e-01 -5.90551972e-01 -6.59365296e-01
-2.01972187e-01 -4.61942554e-01 -8.97665024e-02 -4.56776470e-01
7.26294935e-01 -1.39456105e+00 -7.56920218e-01 1.09175169e+00
-8.16329837e-01 -3.46513480e-01 -1.72883049e-01 3.32885385e-01
-3.79804403e-01 6.54663086e-01 -5.98108709e-01 -7.65016913e-01
6.05306029e-02 -1.58069599e+00 9.33520496e-01 1.99623495e-01
-1.63640991e-01 -6.63006008e-01 1.47643328e-01 9.05536592e-01
5.96319199e-01 -8.38565081e-02 4.13548410e-01 1.79171249e-01
-4.53085572e-01 -3.83477449e-01 -4.43821341e-01 7.36876249e-01
1.50779426e-01 6.82284236e-02 -1.15881348e+00 -5.21162093e-01
7.24525899e-02 -6.87591016e-01 1.64885968e-01 3.52038532e-01
9.05348241e-01 -9.91432462e-03 4.58602518e-01 5.92645884e-01
1.39013767e+00 1.58964068e-01 1.01942110e+00 4.65395302e-01
2.49973416e-01 4.73562002e-01 4.69141901e-01 3.94017041e-01
5.44625103e-01 8.21380913e-01 1.64133295e-01 -2.06393883e-01
-1.62108958e-01 -1.22602567e-01 4.21349615e-01 4.54234332e-01
-1.20906375e-01 -4.36669499e-01 -8.58531296e-01 6.25013530e-01
-1.12110186e+00 -9.15065408e-01 -4.54754308e-02 2.42889071e+00
9.30611312e-01 3.01370304e-02 5.03539145e-01 1.88187391e-01
8.22652817e-01 5.86785115e-02 -4.35249954e-01 -1.02903657e-01
-5.80406725e-01 3.57739888e-02 6.11435890e-01 3.44837159e-01
-8.48902822e-01 5.21109164e-01 7.66754627e+00 7.04640687e-01
-6.87935472e-01 1.53038308e-01 7.97658741e-01 -5.01318336e-01
-2.06713721e-01 -2.03000098e-01 -1.03609368e-01 6.95268750e-01
1.04381001e+00 -2.20953196e-01 7.72444069e-01 5.24542034e-01
6.44480228e-01 -6.11268520e-01 -7.44104743e-01 1.50411189e+00
3.48677307e-01 -8.59843433e-01 -8.07168186e-02 2.08343610e-01
9.43448126e-01 6.23214105e-03 5.30016065e-01 -2.38882452e-01
2.78232694e-01 -1.21519995e+00 8.98175240e-01 6.18268371e-01
1.21724522e+00 -4.24630493e-01 6.28770769e-01 -8.56188238e-02
-6.08954072e-01 -1.73289984e-01 -4.79477435e-01 -8.09960961e-02
1.67146876e-01 7.15044856e-01 -3.02459359e-01 1.14469513e-01
1.07222366e+00 3.41806978e-01 -1.27436996e+00 1.43910980e+00
-2.57527586e-02 3.94504189e-01 1.17788613e-01 4.96853858e-01
-2.94593215e-01 -5.77355884e-02 4.54745978e-01 7.37160742e-01
2.48605490e-01 4.24223617e-02 -3.59040767e-01 6.02613032e-01
-3.25437486e-01 2.24695858e-02 -6.27436936e-01 7.43101612e-02
5.28481662e-01 1.12069893e+00 -2.61838615e-01 -3.33775461e-01
-6.74486637e-01 1.05910802e+00 -1.34522095e-01 5.96708536e-01
-8.32104146e-01 -1.77737534e-01 9.63327527e-01 2.10213125e-01
-2.12328956e-01 -2.44476318e-01 -1.62985593e-01 -1.33801544e+00
5.95706463e-01 -1.90073109e+00 -1.90512910e-02 -1.51425743e+00
-1.66701901e+00 7.07546175e-01 -2.28732765e-01 -1.52970684e+00
1.12188704e-01 -5.94500661e-01 -3.77500415e-01 1.19444919e+00
-1.25034356e+00 -5.46099842e-01 -7.24380553e-01 1.15136003e+00
3.63293111e-01 -3.38004470e-01 6.75563753e-01 5.09356558e-01
-3.56131792e-01 8.16927075e-01 1.19297169e-02 -2.83225384e-02
1.44287026e+00 -1.09120798e+00 4.80156928e-01 1.01656091e+00
1.44457482e-02 5.95894754e-01 6.64665699e-01 -4.74345267e-01
-1.26213157e+00 -5.88041008e-01 5.04513621e-01 -9.17355716e-01
5.34196436e-01 -1.39981017e-01 -7.44648337e-01 1.80549204e-01
3.98682326e-01 1.37704641e-01 8.33511412e-01 -3.90758336e-01
-5.17142475e-01 -1.40709162e-01 -1.41491163e+00 5.83169639e-01
1.11000228e+00 -9.52908933e-01 -4.45680380e-01 1.46975040e-01
4.63760704e-01 -2.17170164e-01 -9.82004523e-01 1.46108702e-01
6.38135433e-01 -1.52399838e+00 1.02416527e+00 -3.42292875e-01
5.27530730e-01 -4.82891619e-01 -4.20913756e-01 -1.44548118e+00
-2.25303113e-01 -6.05314493e-01 2.34266028e-01 1.31572914e+00
2.80293077e-01 -5.42400479e-01 4.11101013e-01 1.09444952e+00
2.62930244e-01 -2.19869316e-01 -5.65080881e-01 -7.33473241e-01
-1.94770217e-01 -6.82051778e-01 6.86292648e-01 6.50578082e-01
-2.27574721e-01 -2.61583269e-01 -5.03921211e-01 2.38100857e-01
8.33467901e-01 -5.78452468e-01 1.15771592e+00 -6.02188826e-01
-5.00950038e-01 -3.05402011e-01 -8.28484595e-01 -6.75341070e-01
-3.76244694e-01 -2.09283754e-01 -1.50304465e-02 -1.17678773e+00
2.91523814e-01 -2.07986668e-01 1.03555992e-01 -2.46765494e-01
-4.75874066e-01 8.64631116e-01 7.71534145e-02 5.68722665e-01
-7.27893353e-01 2.87016392e-01 1.35637617e+00 -8.82007852e-02
7.30804354e-02 -2.67390192e-01 -9.31916714e-01 6.66575372e-01
4.24463600e-01 -5.59565052e-02 -6.12973213e-01 -7.68498778e-01
3.70492041e-01 -3.28276306e-02 7.82164276e-01 -1.29133761e+00
2.38703504e-01 -6.44577742e-02 6.78079903e-01 1.40134826e-01
2.62990713e-01 -7.78716028e-01 2.45650724e-01 -8.66907611e-02
-1.32293865e-01 3.79299819e-01 8.07481557e-02 6.02525651e-01
-3.18703800e-01 2.29969114e-01 9.70261216e-01 -2.52966315e-01
-8.67164612e-01 2.82828540e-01 -3.09813708e-01 3.62058938e-01
8.77972424e-01 -3.44948709e-01 -6.60138190e-01 -8.22857738e-01
-3.68090987e-01 -7.37819597e-02 1.32972705e+00 4.41193283e-01
6.88802898e-01 -1.34701943e+00 -5.86867392e-01 2.52985716e-01
4.90909249e-01 -5.64353585e-01 3.75151336e-01 6.32103503e-01
-6.63345158e-01 -9.97902676e-02 -4.84335274e-01 -5.12611985e-01
-1.23711598e+00 7.02673137e-01 4.63900656e-01 2.11945564e-01
-2.70769536e-01 5.54034770e-01 -1.82411894e-01 -8.13857187e-03
1.80890560e-01 -7.31197000e-03 1.00175470e-01 -1.57451719e-01
9.02173638e-01 4.70129341e-01 2.76120543e-01 -1.03899026e+00
-1.41903386e-01 4.54352468e-01 3.45453560e-01 -5.49867332e-01
8.07202816e-01 -6.15251660e-01 1.13513626e-01 1.64779112e-01
1.19773817e+00 2.77363569e-01 -1.27711320e+00 -1.86637908e-01
-4.47974473e-01 -1.26769197e+00 -4.53290828e-02 -1.02622688e+00
-9.91669714e-01 4.83867526e-01 1.15649951e+00 1.62658736e-01
1.60401857e+00 -3.89282286e-01 7.91315794e-01 4.87656184e-02
7.09431350e-01 -1.05941725e+00 2.76623428e-01 -1.03932805e-01
9.29184139e-01 -1.81282556e+00 -1.72957867e-01 -5.29417321e-02
-8.32507908e-01 4.92526203e-01 3.52382869e-01 1.01327514e-02
3.55693668e-01 -7.25015905e-03 6.08445406e-01 -6.43781424e-02
-4.79157150e-01 -1.77228063e-01 4.26326424e-01 1.11956155e+00
1.66033715e-01 -5.78181706e-02 4.59372951e-03 4.53534484e-01
-4.31114823e-01 1.72322303e-01 7.48389304e-01 6.45385325e-01
-1.83344528e-01 -9.22480106e-01 -9.08300042e-01 2.81746507e-01
-5.98708510e-01 -1.03553869e-01 -3.87651473e-01 5.22328019e-01
2.71156073e-01 1.77170861e+00 -2.22615451e-01 -5.74411154e-01
6.61771536e-01 -2.84206927e-01 3.81400406e-01 -2.56289124e-01
-5.24056554e-01 -2.41737142e-01 7.47313350e-02 -8.81339550e-01
-5.64541757e-01 -5.44166267e-01 -2.15934008e-01 -7.12576449e-01
1.37512878e-01 -2.94517785e-01 5.82453668e-01 6.17014885e-01
2.77922869e-01 1.19673848e-01 6.68306947e-01 -9.87192273e-01
-4.14808512e-01 -8.13619375e-01 -7.57293940e-01 1.23581207e+00
3.77144456e-01 -5.03374219e-01 -5.99115670e-01 4.10006076e-01]
|
[11.870537757873535, -1.7967212200164795]
|
2fc915c5-f84d-4980-a43f-ad4fc085c703
|
the-effect-of-points-dispersion-on-the-k-nn-1
|
2302.13160
| null |
https://arxiv.org/abs/2302.13160v1
|
https://arxiv.org/pdf/2302.13160v1.pdf
|
The Effect of Points Dispersion on the $k$-nn Search in Random Projection Forests
|
Partitioning trees are efficient data structures for $k$-nearest neighbor search. Machine learning libraries commonly use a special type of partitioning trees called $k$d-trees to perform $k$-nn search. Unfortunately, $k$d-trees can be ineffective in high dimensions because they need more tree levels to decrease the vector quantization (VQ) error. Random projection trees rpTrees solve this scalability problem by using random directions to split the data. A collection of rpTrees is called rpForest. $k$-nn search in an rpForest is influenced by two factors: 1) the dispersion of points along the random direction and 2) the number of rpTrees in the rpForest. In this study, we investigate how these two factors affect the $k$-nn search with varying $k$ values and different datasets. We found that with larger number of trees, the dispersion of points has a very limited effect on the $k$-nn search. One should use the original rpTree algorithm by picking a random direction regardless of the dispersion of points.
|
['Masahiro Takatsuka', 'Adel F. Ahmed', 'John Stavrakakis', 'Mashaan Alshammari']
|
2023-02-25
| null | null | null | null |
['instance-search', 'vector-quantization-k-means-problem']
|
['computer-vision', 'miscellaneous']
|
[-4.22857493e-01 -4.66210842e-01 -4.71397400e-01 -4.21343505e-01
-4.77170140e-01 -4.58354384e-01 1.00486703e-01 2.23504409e-01
-3.41005832e-01 6.34282351e-01 2.59910464e-01 -3.66157055e-01
-4.99681026e-01 -1.40708518e+00 -2.84110069e-01 -7.37697542e-01
-4.57293428e-02 5.78800976e-01 5.86512208e-01 2.14817245e-02
9.02289569e-01 6.71713650e-01 -1.64804566e+00 2.60642558e-01
5.60769796e-01 9.30251122e-01 7.11401179e-02 3.85424942e-01
-7.33537197e-01 1.99750811e-01 -5.54460287e-01 -2.35611815e-02
6.85558736e-01 -1.58219770e-01 -8.42752814e-01 -4.15018886e-01
1.17089599e-01 -1.01894051e-01 8.98618773e-02 7.66877472e-01
5.90137959e-01 2.76000708e-01 9.75669563e-01 -1.09057426e+00
-1.54699394e-02 7.01039255e-01 -9.30462658e-01 1.75962523e-01
1.36354148e-01 -1.10103078e-01 9.37248170e-01 -9.00052726e-01
6.15679085e-01 1.17416584e+00 1.03266430e+00 -1.93574522e-02
-1.44546068e+00 -9.09452319e-01 -3.39791030e-01 1.54484466e-01
-1.94454455e+00 -2.41697073e-01 9.03602719e-01 -4.25613761e-01
1.10432661e+00 3.20548564e-01 7.04965830e-01 1.16046816e-01
3.47129047e-01 1.93678498e-01 9.97067451e-01 -5.34075022e-01
6.57821774e-01 1.54435351e-01 5.72210252e-01 2.84716636e-01
3.07584167e-01 -5.58104888e-02 -4.69606698e-01 -7.05098093e-01
7.51452088e-01 6.52028248e-02 -8.91958177e-02 -8.57658505e-01
-8.04717422e-01 1.21859515e+00 6.30072713e-01 3.96562606e-01
-1.61979392e-01 1.95733398e-01 4.69334871e-01 7.13567138e-02
3.74368876e-02 6.20566905e-01 -5.95207214e-01 -1.63319573e-01
-9.12206352e-01 3.26155007e-01 7.25779951e-01 8.28320444e-01
1.02684128e+00 -5.15415370e-01 9.58815292e-02 1.26324749e+00
1.62397519e-01 1.64086461e-01 5.45172095e-01 -8.78764570e-01
7.26611912e-01 7.30811596e-01 4.24861833e-02 -1.35259843e+00
-3.24277580e-01 -1.10679626e-01 -7.34752476e-01 1.78789660e-01
3.81027430e-01 1.05540425e-01 -8.58293831e-01 1.21943223e+00
4.65104342e-01 -4.43437487e-01 -1.33530587e-01 6.50180459e-01
5.61408341e-01 7.53925204e-01 -4.60352823e-02 -2.74661720e-01
9.68602598e-01 -4.81857449e-01 -2.91285455e-01 9.80811343e-02
9.46315169e-01 -8.85539293e-01 1.25531113e+00 2.41345420e-01
-7.53253222e-01 -3.46339732e-01 -9.14125323e-01 1.26802057e-01
-4.95042086e-01 -1.83909267e-01 5.11758268e-01 7.03705370e-01
-1.02883470e+00 6.74270093e-01 -5.80753803e-01 -4.81260687e-01
3.20794493e-01 5.95509350e-01 -1.27179980e-01 -1.62313268e-01
-8.52731347e-01 6.49391770e-01 3.39556336e-01 -4.24746484e-01
-5.32385595e-02 -4.76233453e-01 -3.74556512e-01 9.73174945e-02
-5.39733935e-03 -3.55913103e-01 6.19226635e-01 8.79862607e-02
-1.06067204e+00 5.18596590e-01 -3.64120960e-01 -1.73397452e-01
1.36356100e-01 8.61180073e-04 5.88108227e-02 -1.26592293e-01
2.54865587e-01 7.31638312e-01 6.00647032e-01 -1.02277231e+00
-4.89399999e-01 -8.62986267e-01 -5.02486646e-01 5.68564236e-01
-1.89341769e-01 -3.98247808e-01 -2.86826164e-01 -3.91155630e-01
7.94493079e-01 -9.04964745e-01 -5.10983288e-01 -2.66487151e-01
-3.98138672e-01 -4.16540086e-01 1.02928936e+00 1.69036329e-01
1.65100431e+00 -2.32483792e+00 -3.37844849e-01 6.43400848e-01
1.86511874e-01 1.31697049e-02 1.93642318e-01 6.52475953e-01
-4.61288244e-02 3.25165570e-01 -2.95253117e-02 3.30389172e-01
-3.03035885e-01 2.79177070e-01 -2.28433013e-01 2.07557321e-01
-6.88563287e-01 1.94268823e-01 -6.01586342e-01 -6.72388136e-01
2.65206516e-01 2.40901545e-01 -6.07582629e-01 -2.53798008e-01
1.38525784e-01 -3.46598744e-01 -5.26787519e-01 3.03167611e-01
9.26753640e-01 -1.50488719e-01 -6.59012794e-02 -4.86088581e-02
-3.81174594e-01 4.00897175e-01 -1.56002200e+00 1.20903873e+00
-3.69858220e-02 3.16667050e-01 -2.71115392e-01 -9.74084198e-01
1.51928961e+00 -4.82104085e-02 6.11447811e-01 -3.78976762e-01
-2.41815969e-01 4.56680894e-01 -5.68693392e-02 7.51266554e-02
6.27135217e-01 -3.01754355e-01 5.22243790e-02 5.37934422e-01
-6.20523691e-01 -5.10255635e-01 5.48807085e-02 -4.94642332e-02
1.22945285e+00 -5.73290408e-01 4.79162693e-01 -6.85051978e-01
1.57312080e-01 4.53414202e-01 7.35224247e-01 7.50322640e-01
-1.64693028e-01 4.96446341e-01 2.76192069e-01 -8.04023325e-01
-9.39224958e-01 -1.02698100e+00 -6.64002359e-01 9.83199120e-01
1.82000309e-01 -6.71536922e-01 -4.51541871e-01 -5.81657827e-01
1.96445465e-01 8.57457161e-01 -2.63816595e-01 1.33565396e-01
-5.66877246e-01 -9.29492652e-01 2.31320664e-01 3.79622102e-01
6.19653165e-01 -8.80333781e-01 -8.21860552e-01 2.48614192e-01
-1.68649435e-01 -3.55137289e-01 -4.46470946e-01 7.76990294e-01
-1.44300425e+00 -9.81779873e-01 -5.32990992e-01 -6.70747161e-01
5.73280275e-01 6.63567007e-01 1.00743365e+00 -3.30069005e-01
-1.03137195e-01 -1.35035694e-01 -4.34014648e-01 -2.35504881e-01
1.06816292e-01 4.92481142e-01 -1.26659162e-02 -8.39132667e-01
8.90529692e-01 -5.94392180e-01 -8.44948232e-01 7.58149981e-01
-5.89410007e-01 -4.66573149e-01 3.19008231e-01 7.35329628e-01
1.01982486e+00 9.91464496e-01 1.11756437e-01 -8.70246530e-01
8.47718477e-01 -4.18829978e-01 -5.24756432e-01 -1.52695909e-01
-9.55588698e-01 3.05671841e-01 6.90074861e-01 -2.24966168e-01
-3.09294105e-01 1.49402812e-01 -7.79117793e-02 -3.77794862e-01
2.97375508e-02 3.49569231e-01 4.90446612e-02 1.05187066e-01
9.61737156e-01 1.46060601e-01 -3.83338720e-01 -4.44137931e-01
1.95873961e-01 9.06471789e-01 -2.32647702e-01 -3.03867489e-01
3.39024007e-01 3.92099977e-01 1.61384970e-01 -9.23694134e-01
-1.39786437e-01 -6.33910716e-01 -5.73741853e-01 2.76957124e-01
7.00430989e-01 -4.91316348e-01 -2.82848001e-01 1.53578520e-01
-6.85299277e-01 -1.86979845e-01 -5.04990578e-01 4.33892548e-01
-4.25419062e-01 1.87731624e-01 -2.34449431e-01 -5.30346692e-01
-4.12044108e-01 -1.27365255e+00 7.53980756e-01 2.17002332e-02
-6.43916905e-01 -4.48850691e-01 3.42943162e-01 -4.21864586e-03
2.29582131e-01 -1.86971650e-01 1.32797098e+00 -4.37217712e-01
-3.11115086e-01 -1.72774449e-01 -2.41635412e-01 -1.24405079e-01
2.49069959e-01 -5.29926084e-02 -4.68289047e-01 -1.91642299e-01
5.66681437e-02 1.21175230e-01 6.45020068e-01 8.21018994e-01
1.16777837e+00 -4.19165522e-01 -8.65709186e-01 5.75657845e-01
1.48010170e+00 6.28097653e-01 7.58951664e-01 4.47993398e-01
4.75361705e-01 3.47572029e-01 6.47784889e-01 3.88385504e-01
2.50934958e-01 7.30451703e-01 8.15847814e-02 2.18401298e-01
9.23138335e-02 -3.87602866e-01 -2.39421502e-01 6.97652459e-01
3.08463424e-01 3.31568439e-03 -1.40637803e+00 6.24694884e-01
-1.41430783e+00 -8.42650235e-01 -2.05732793e-01 2.48125863e+00
8.46109092e-01 9.27606151e-02 1.76918522e-01 6.67026222e-01
8.01177621e-01 1.45537183e-01 -5.59614778e-01 -8.23562205e-01
3.41165274e-01 8.52949172e-02 8.67296934e-01 4.17299747e-01
-7.71767914e-01 6.96829498e-01 7.17506790e+00 1.12692487e+00
-1.21988583e+00 -4.06708449e-01 6.72860444e-01 -1.05396494e-01
-2.73613662e-01 1.27273604e-01 -1.33706248e+00 8.22160780e-01
6.69821560e-01 1.27174571e-01 7.19101429e-02 1.15575886e+00
1.14732891e-01 -4.42863166e-01 -8.40011895e-01 1.26856148e+00
-4.74007040e-01 -1.36214423e+00 2.58817859e-02 3.88480991e-01
5.24208188e-01 1.82983324e-01 5.51468991e-02 -1.79231260e-02
6.54968262e-01 -1.00308979e+00 -7.17048869e-02 -6.47676066e-02
8.25112760e-01 -9.27010477e-01 5.70013225e-01 4.87465888e-01
-1.34080696e+00 -1.01286717e-01 -7.34916866e-01 -2.68909447e-02
-1.17347702e-01 1.15856814e+00 -1.09818256e+00 -1.52655691e-01
1.22579932e+00 1.49035543e-01 -3.09369326e-01 8.92553151e-01
3.66015822e-01 4.12953675e-01 -8.84089947e-01 -2.67426044e-01
2.15672255e-01 -4.13430870e-01 3.73833328e-01 8.72879326e-01
5.75797260e-01 3.17114472e-01 -1.31094471e-01 3.58362645e-01
4.42840964e-01 3.06020707e-01 -9.44148242e-01 1.61335334e-01
1.19312441e+00 4.99189824e-01 -1.07009995e+00 -9.37061310e-02
-2.28125006e-02 4.35470849e-01 1.94408730e-01 5.87886199e-02
-1.91621110e-01 -7.77108431e-01 5.69807947e-01 7.33547032e-01
7.46599793e-01 -3.42304021e-01 -8.03989530e-01 -5.33899069e-01
-2.12707847e-01 -5.91644585e-01 6.97113872e-01 -4.79071528e-01
-1.18534684e+00 4.97858703e-01 2.11017892e-01 -1.18230689e+00
-3.00291747e-01 -1.69970900e-01 -1.87769756e-01 8.85928631e-01
-6.81504846e-01 -1.88452572e-01 3.27987745e-02 5.57011485e-01
2.99594671e-01 5.32347858e-02 9.23738778e-01 -1.60641313e-01
-2.84145564e-01 5.33205628e-01 5.23487210e-01 -1.87823363e-02
4.90127414e-01 -1.09692645e+00 1.39609471e-01 3.29173386e-01
7.51053318e-02 9.72877860e-01 6.41002655e-01 -7.04184115e-01
-1.06123328e+00 -8.13010335e-01 9.86896455e-01 -3.18063438e-01
1.04130045e-01 -2.26208605e-02 -7.71934986e-01 3.21881622e-01
-4.14266825e-01 3.72101292e-02 1.11091554e+00 5.22534132e-01
-4.01052564e-01 -4.54085916e-01 -1.68283176e+00 4.91839945e-01
9.34017479e-01 -2.69419014e-01 -3.61584634e-01 -1.17903918e-01
3.62595916e-01 -1.34323061e-01 -1.08485126e+00 5.04514754e-01
6.91412508e-01 -1.29587293e+00 1.13762724e+00 1.74650535e-01
2.33649388e-01 -4.47608441e-01 -5.50473869e-01 -1.22229970e+00
-5.94787717e-01 6.29396364e-02 3.58848065e-01 7.97969460e-01
3.97096068e-01 -7.17241585e-01 1.21250296e+00 5.80456376e-01
4.55984622e-01 -1.00800443e+00 -1.17342222e+00 -6.15212262e-01
4.37312603e-01 -3.82405788e-01 8.95005405e-01 8.77020776e-01
2.09474564e-01 5.48861027e-01 1.86912715e-01 -1.42972246e-01
4.77158934e-01 2.77905226e-01 8.72835517e-01 -1.28473866e+00
-1.03288375e-01 -3.91545862e-01 -5.34263134e-01 -1.26157582e+00
-6.68850720e-01 -5.49586535e-01 -2.93290261e-02 -1.36966205e+00
1.57459602e-01 -1.37979650e+00 8.84124413e-02 2.55835623e-01
4.11380753e-02 4.87774499e-02 7.00328499e-02 7.00282872e-01
-2.02864453e-01 2.70646602e-01 9.97580647e-01 1.68427125e-01
-8.69299293e-01 5.80344386e-02 -4.69771564e-01 7.22853363e-01
8.13909769e-01 -7.50185132e-01 -7.00462699e-01 -4.01864529e-01
1.26848921e-01 2.23658457e-01 -3.85550827e-01 -1.03010738e+00
2.71557510e-01 -2.68838972e-01 5.99457264e-01 -1.31778014e+00
3.44694793e-01 -8.37934434e-01 3.88049692e-01 5.17775536e-01
-1.64521858e-01 3.19002062e-01 2.92630102e-02 3.84347469e-01
-1.05732486e-01 -4.37524885e-01 8.54834676e-01 -3.64395082e-01
-3.69441926e-01 1.12871937e-02 -4.22033399e-01 -3.01568955e-01
1.11795223e+00 -8.76971304e-01 -2.66585108e-02 -1.44090906e-01
-4.94891763e-01 1.30462542e-01 7.08042920e-01 -1.16195396e-01
4.98057753e-01 -1.16242158e+00 -1.01083234e-01 4.09922451e-01
-4.86484542e-02 3.53944004e-01 -1.87965274e-01 2.06110865e-01
-7.27358520e-01 6.58324659e-01 2.58019250e-02 -8.52820516e-01
-1.37002671e+00 4.35625046e-01 1.41319737e-01 -3.63163054e-01
-5.79884946e-01 1.05751765e+00 -1.63936958e-01 -7.13739336e-01
1.59610718e-01 -2.10262120e-01 -7.93261975e-02 1.74719140e-01
3.63771886e-01 7.11969852e-01 1.11502953e-01 -3.83449763e-01
-5.18916070e-01 1.01165617e+00 -3.43653440e-01 -2.50547826e-01
1.07223547e+00 -1.44910604e-01 -8.32449496e-02 5.07018030e-01
1.24260807e+00 -1.25029692e-02 -7.81334341e-01 7.50555918e-02
-9.94316265e-02 -9.10024464e-01 1.00791737e-01 -3.02659482e-01
-8.78763497e-01 6.61056638e-01 5.48280060e-01 3.28992218e-01
1.04429579e+00 -8.10343921e-02 7.21693814e-01 2.93646574e-01
8.67964387e-01 -1.19230855e+00 -2.01611936e-01 5.08934438e-01
4.39455181e-01 -9.44544315e-01 3.40006262e-01 -5.54397702e-01
-4.60977644e-01 1.03183675e+00 5.21406531e-01 -1.03248894e-01
1.18554020e+00 2.47490689e-01 1.36110857e-01 -2.90766567e-01
-4.86451298e-01 2.60204792e-01 -2.26102814e-01 5.86999178e-01
4.40413743e-01 1.74662948e-01 -7.60872722e-01 2.02675804e-01
-8.69069457e-01 2.27282643e-01 1.10723130e-01 9.41896260e-01
-7.31194019e-01 -1.32554448e+00 -6.98804259e-01 9.32118773e-01
-2.92581990e-02 -2.63520088e-02 -2.83911347e-01 6.25612497e-01
2.96866745e-01 9.72763956e-01 4.31141675e-01 -6.00331783e-01
1.57248929e-01 5.34635708e-02 1.32884979e-01 -4.84302759e-01
-2.45940819e-01 -7.84638375e-02 -2.79692888e-01 -6.00976348e-01
-4.86417487e-03 -6.94233239e-01 -1.33446431e+00 -8.09920132e-01
-5.13285458e-01 7.20871687e-01 6.97406828e-01 4.04584706e-01
6.11139059e-01 -2.90623367e-01 6.59797966e-01 -3.50170761e-01
-6.17924929e-01 -6.90131724e-01 -1.01159239e+00 1.73511561e-02
-2.11513713e-01 -5.90176225e-01 -4.63504195e-01 -4.51851457e-01]
|
[7.474233627319336, 4.689890384674072]
|
e1842e24-b68a-4fb9-9560-c8310773372a
|
nima-neural-image-assessment
|
1709.05424
| null |
http://arxiv.org/abs/1709.05424v2
|
http://arxiv.org/pdf/1709.05424v2.pdf
|
NIMA: Neural Image Assessment
|
Automatically learned quality assessment for images has recently become a hot
topic due to its usefulness in a wide variety of applications such as
evaluating image capture pipelines, storage techniques and sharing media.
Despite the subjective nature of this problem, most existing methods only
predict the mean opinion score provided by datasets such as AVA [1] and TID2013
[2]. Our approach differs from others in that we predict the distribution of
human opinion scores using a convolutional neural network. Our architecture
also has the advantage of being significantly simpler than other methods with
comparable performance. Our proposed approach relies on the success (and
retraining) of proven, state-of-the-art deep object recognition networks. Our
resulting network can be used to not only score images reliably and with high
correlation to human perception, but also to assist with adaptation and
optimization of photo editing/enhancement algorithms in a photographic
pipeline. All this is done without need for a "golden" reference image,
consequently allowing for single-image, semantic- and perceptually-aware,
no-reference quality assessment.
|
['Peyman Milanfar', 'Hossein Talebi']
|
2017-09-15
| null | null | null | null |
['aesthetics-quality-assessment']
|
['computer-vision']
|
[ 3.45041543e-01 -2.42658213e-01 2.17563435e-01 -7.05910683e-01
-6.67409241e-01 -5.77268124e-01 5.46373665e-01 4.04674977e-01
-6.78577960e-01 4.30556834e-01 1.26856431e-01 -1.17562756e-01
-9.55306590e-02 -6.66427851e-01 -5.80612183e-01 -5.09889543e-01
9.74110886e-02 2.61913568e-01 3.91209632e-01 -2.77869761e-01
4.19291615e-01 6.35247767e-01 -2.05132532e+00 2.67946601e-01
7.20140934e-01 1.60525286e+00 3.44535649e-01 8.80959332e-01
1.58439085e-01 6.60705090e-01 -7.89642036e-01 -8.22797596e-01
3.71035218e-01 -1.78530306e-01 -7.28427947e-01 3.18969637e-02
9.50163245e-01 -6.02864444e-01 -1.10333554e-01 1.12046218e+00
8.15422297e-01 7.06857443e-02 4.00281280e-01 -1.08825564e+00
-7.23455429e-01 1.59053743e-01 -2.89143562e-01 -2.60301377e-03
2.97805756e-01 4.88286197e-01 8.47945929e-01 -6.18544459e-01
7.08927155e-01 8.92244875e-01 5.89677989e-01 3.37996751e-01
-1.11597896e+00 -4.55441982e-01 -4.41872954e-01 7.81457484e-01
-1.01491153e+00 -6.72046483e-01 6.90981269e-01 -4.68897283e-01
9.61551845e-01 1.92516521e-01 6.86557233e-01 7.95415699e-01
-3.24987620e-02 5.42892933e-01 1.40368795e+00 -4.84774411e-01
4.35968399e-01 2.55042642e-01 -4.83699143e-01 5.22118628e-01
7.97543861e-03 -5.71112260e-02 -5.50080121e-01 3.22849423e-01
5.06734192e-01 -1.93593875e-01 -2.49602348e-01 -5.23541033e-01
-1.13733256e+00 2.38219708e-01 6.03873372e-01 2.77771622e-01
-6.55580759e-01 2.69356728e-01 3.97923857e-01 4.41626608e-01
4.53356713e-01 5.52452564e-01 -5.13498783e-01 -3.92281622e-01
-1.41089785e+00 6.16798326e-02 5.54959953e-01 4.02844220e-01
8.59283328e-01 -7.51366690e-02 -1.56837210e-01 8.76544476e-01
3.94183397e-02 6.86586678e-01 4.34521496e-01 -1.38569951e+00
1.81731209e-01 5.59713185e-01 4.76224005e-01 -1.20397294e+00
-2.76480675e-01 -4.54478890e-01 -8.48715901e-01 9.93527293e-01
4.16759819e-01 3.14798534e-01 -8.83015215e-01 1.51690435e+00
1.34254903e-01 -9.98911709e-02 -1.50251627e-01 1.18281209e+00
5.23577392e-01 3.22379291e-01 9.57493037e-02 2.44086366e-02
1.25910056e+00 -6.44643307e-01 -4.60976958e-01 -5.47631942e-02
2.04848960e-01 -1.06845224e+00 1.21537983e+00 9.12532628e-01
-1.34579992e+00 -9.28030252e-01 -1.22390699e+00 -1.82454780e-01
-5.12183726e-01 3.25235575e-01 5.37607253e-01 8.98948431e-01
-1.72575986e+00 1.01150072e+00 -4.57713574e-01 -4.54256296e-01
6.96719646e-01 4.14840877e-01 -5.97498357e-01 -1.27296209e-01
-8.79168153e-01 1.24771416e+00 1.61357194e-01 2.86335438e-01
-6.74806774e-01 -6.18135571e-01 -5.99096298e-01 1.72235593e-01
7.02354982e-02 -7.49204695e-01 1.22984076e+00 -1.49696183e+00
-1.65492225e+00 1.19772983e+00 -1.72243640e-02 -5.45181751e-01
5.87769032e-01 -2.35749513e-01 -4.89808679e-01 3.92944485e-01
-2.00436532e-01 8.85528147e-01 8.42634022e-01 -1.05960000e+00
-5.01367092e-01 -3.52699637e-01 2.25978374e-01 6.84678629e-02
-5.64478815e-01 1.19486995e-01 -5.69913149e-01 -2.70737171e-01
-4.36769426e-01 -5.94928980e-01 -9.34145227e-02 5.38192213e-01
6.05664402e-02 -2.27138177e-02 7.01640725e-01 -9.08918560e-01
9.72021639e-01 -1.85750449e+00 5.19863404e-02 8.84701032e-03
1.01446390e-01 8.71451497e-01 -3.06298703e-01 2.92262763e-01
8.82308483e-02 -6.13542162e-02 -6.45739883e-02 -5.22079170e-01
1.11646280e-01 -1.69010863e-01 -1.70869697e-02 5.34492612e-01
5.14732897e-01 8.59025538e-01 -9.39614654e-01 -5.03726780e-01
9.49206412e-01 5.94265461e-01 -3.56079102e-01 3.74156207e-01
-1.85043007e-01 3.46218407e-01 1.83034956e-01 5.49592376e-01
8.03406477e-01 -2.31397927e-01 -5.32558076e-02 -6.97746098e-01
-2.62816876e-01 5.73222190e-02 -1.22128224e+00 1.92067122e+00
-8.28786910e-01 1.10616267e+00 2.51589082e-02 -7.97998428e-01
9.20985758e-01 8.96971151e-02 4.43998128e-01 -1.15173090e+00
2.94628799e-01 2.61662006e-01 -8.57886747e-02 -6.04825020e-01
7.59539962e-01 3.02680075e-01 3.10825199e-01 2.97474295e-01
2.73411959e-01 -2.66689271e-01 3.21405321e-01 -3.87272015e-02
9.88698483e-01 1.93577617e-01 3.02443147e-01 -1.20197730e-02
7.09514976e-01 -2.39765510e-01 1.29453897e-01 5.40460885e-01
-5.31984150e-01 9.92679358e-01 1.77523732e-01 -5.56684434e-01
-1.46586978e+00 -9.62436020e-01 -2.32275948e-02 7.99617290e-01
2.87867069e-01 -1.14685223e-01 -8.04970324e-01 -3.14279050e-01
-3.00005257e-01 5.38440228e-01 -6.35157347e-01 -1.43263151e-03
-3.06413293e-01 -2.10299224e-01 1.87357008e-01 4.07696515e-01
6.72066331e-01 -1.29509640e+00 -9.40370739e-01 9.32323113e-02
-7.52004460e-02 -1.15312505e+00 -9.23616663e-02 -1.64020643e-01
-5.27384341e-01 -1.09187698e+00 -9.00011957e-01 -3.22518528e-01
3.41421843e-01 2.13190541e-01 1.58904886e+00 1.57868072e-01
-3.50501865e-01 5.46362817e-01 -3.15782785e-01 -2.81718731e-01
-4.21455383e-01 -1.07746117e-01 -2.01417729e-01 1.90788522e-01
3.22935283e-01 -5.55090606e-01 -1.28617752e+00 3.97406608e-01
-1.12728465e+00 -5.98736517e-02 8.53326499e-01 5.76218247e-01
5.34788251e-01 -5.80728836e-02 3.05357039e-01 -4.97107357e-01
3.88007373e-01 -7.12031790e-04 -7.41140306e-01 3.73130828e-01
-8.78751814e-01 -2.14711055e-01 5.92660606e-01 -1.36584252e-01
-9.62055326e-01 -1.17710546e-01 -3.25824857e-01 -4.27181304e-01
-4.84797508e-01 8.47559497e-02 -2.53372729e-01 -2.64546216e-01
6.73724949e-01 3.26094069e-02 -1.24926738e-01 -2.19683111e-01
5.18193185e-01 8.19291651e-01 8.84778321e-01 -1.56369600e-02
6.46207154e-01 5.90214610e-01 1.33025393e-01 -7.95052528e-01
-4.91710097e-01 -6.30040407e-01 -7.10833192e-01 -7.24136472e-01
8.30600441e-01 -7.87531078e-01 -9.53929603e-01 8.17880630e-01
-1.23231065e+00 -3.19400311e-01 -3.44396323e-01 2.47160658e-01
-6.17055714e-01 4.55458134e-01 -2.15282589e-01 -9.50619042e-01
-5.10335922e-01 -1.21844697e+00 1.22099030e+00 4.36768144e-01
-1.24361910e-01 -8.19877744e-01 -1.18076533e-01 5.94087720e-01
9.17887390e-01 1.51377678e-01 4.38992381e-01 -6.23640306e-02
-6.17839992e-01 -3.35415870e-01 -6.42320096e-01 9.19189572e-01
-1.71682864e-01 5.25054447e-02 -1.30777955e+00 -1.44929022e-01
-3.41563761e-01 -6.56530380e-01 6.47202492e-01 4.66506571e-01
1.29129767e+00 -3.03546842e-02 3.00095290e-01 6.12887681e-01
1.62265372e+00 -3.57310206e-01 1.09561992e+00 4.20048803e-01
3.43386024e-01 6.04072094e-01 5.84226310e-01 3.26183110e-01
3.55596989e-01 1.03737199e+00 7.70635545e-01 -3.80165488e-01
-5.30596197e-01 8.15155208e-02 2.25224704e-01 4.01588202e-01
-1.45713091e-01 -2.40083545e-01 -7.82289326e-01 7.09521532e-01
-1.62828362e+00 -1.03598857e+00 -1.38280004e-01 2.44079781e+00
7.39772320e-01 1.42178647e-02 9.90204662e-02 1.88560277e-01
5.34422338e-01 1.50533065e-01 -5.57902813e-01 -5.95527470e-01
-2.87952155e-01 4.29705709e-01 6.33871377e-01 1.12168118e-01
-1.00246680e+00 6.04229808e-01 5.79502487e+00 8.09412837e-01
-1.30880761e+00 1.52360499e-01 8.75390410e-01 -8.30806866e-02
5.20625710e-02 -2.38391250e-01 -3.14262569e-01 5.18585920e-01
1.04318523e+00 1.37461841e-01 5.35778046e-01 8.38643551e-01
3.93014610e-01 -3.79576087e-01 -1.00623953e+00 1.31173158e+00
4.80423391e-01 -1.41845691e+00 -2.33848870e-01 -1.54530808e-01
7.68564880e-01 2.02826560e-01 2.82817662e-01 -1.63414389e-01
2.49750745e-02 -1.04171777e+00 1.01679599e+00 7.85560131e-01
1.11470151e+00 -6.09198213e-01 9.67646718e-01 -1.30449682e-01
-8.58997762e-01 -2.76922602e-02 -4.53118771e-01 4.45693312e-03
2.40704760e-01 7.72342861e-01 -5.18200040e-01 4.58229721e-01
1.08298278e+00 5.04174590e-01 -9.75849628e-01 1.33387601e+00
-1.16867349e-01 3.41914713e-01 -1.81915924e-01 4.11379822e-02
-4.76571247e-02 9.22851730e-03 5.51735125e-02 1.18533909e+00
4.52287853e-01 -2.73076952e-01 -2.11406976e-01 7.03395426e-01
-1.80298135e-01 1.89349100e-01 -3.57100934e-01 1.00954667e-01
2.14921042e-01 1.52637589e+00 -7.70882070e-01 -2.52156496e-01
-3.55906755e-01 1.31308413e+00 1.66963995e-01 6.20548278e-02
-4.65350509e-01 -4.66601104e-01 6.60610378e-01 1.75255507e-01
4.84177023e-01 -2.64986396e-01 -2.84791738e-01 -9.05780077e-01
3.27803612e-01 -9.08216476e-01 -1.87213451e-01 -1.26179516e+00
-1.11725080e+00 9.20154929e-01 -3.67549151e-01 -1.32593191e+00
-3.10425013e-01 -1.00993657e+00 -4.99649376e-01 9.92772043e-01
-1.82153249e+00 -1.32218599e+00 -7.33961284e-01 5.27083158e-01
3.21723461e-01 -1.24621697e-01 7.70960391e-01 5.18440604e-01
-9.72919911e-02 4.78225350e-01 5.02289683e-02 -8.25864375e-02
1.06900406e+00 -1.41976380e+00 3.09657544e-01 9.07502353e-01
7.69488662e-02 2.61148155e-01 9.11688387e-01 -7.72228688e-02
-1.17691803e+00 -7.84152865e-01 7.43550062e-01 -5.08523166e-01
5.22803307e-01 -1.64011315e-01 -7.08560586e-01 -2.51476049e-01
5.24875998e-01 1.15144126e-01 4.93514061e-01 1.53773772e-02
-3.45459104e-01 -6.89582765e-01 -1.13800788e+00 4.77631360e-01
8.07854295e-01 -5.54938138e-01 -1.42195314e-01 1.62785903e-01
2.48599276e-01 -1.72834516e-01 -9.24043775e-01 3.30216736e-01
8.59042287e-01 -1.74980116e+00 1.13628161e+00 -1.37946665e-01
7.96936035e-01 -3.86558205e-01 -1.22711003e-01 -1.30997527e+00
-1.03933737e-01 -3.34209323e-01 -3.98164578e-02 1.24641740e+00
2.02000543e-01 -2.52650112e-01 8.38963926e-01 6.24134481e-01
-6.75111264e-02 -6.06570363e-01 -8.94176066e-01 -4.93312061e-01
-3.34018379e-01 -6.62918746e-01 5.83912611e-01 4.25208747e-01
-5.64594626e-01 -1.20767817e-01 -6.29631817e-01 6.60042837e-02
6.54374361e-01 1.22788981e-01 9.41655397e-01 -1.13414109e+00
-5.03534257e-01 -7.82818317e-01 -1.01928544e+00 -6.26960993e-01
-1.42433941e-01 -4.79125410e-01 -6.60036355e-02 -1.69544518e+00
9.67552513e-03 -3.27560455e-01 -5.47374070e-01 1.48955345e-01
1.38178281e-02 9.12007511e-01 1.72796920e-01 1.02945179e-01
-1.07355416e+00 3.87946039e-01 9.43290591e-01 -1.12224348e-01
2.30067506e-01 -3.37713182e-01 -4.64492470e-01 5.40347695e-01
6.18200719e-01 -3.76230776e-02 -2.63681501e-01 -6.12856448e-01
6.03680313e-01 -3.07137787e-01 7.20786393e-01 -1.50288272e+00
2.56826669e-01 1.83821559e-01 7.35691369e-01 -4.85328794e-01
3.48384619e-01 -9.50505555e-01 1.14406765e-01 2.10814700e-01
-3.29685658e-01 3.44387889e-02 -5.55904815e-03 2.90843368e-01
-4.42396224e-01 -2.63667047e-01 9.17735279e-01 1.96329150e-02
-1.13580203e+00 2.57446915e-01 1.64632767e-01 -3.04086894e-01
8.29696417e-01 -3.53455245e-01 -4.24308181e-01 -5.61538935e-01
-3.26816320e-01 -5.03253303e-02 9.89954352e-01 4.11686063e-01
5.78870893e-01 -1.08458221e+00 -6.86771810e-01 3.84752825e-02
3.76331389e-01 -4.00618970e-01 6.42767191e-01 6.51390314e-01
-8.31086338e-01 4.79624197e-02 -6.40943527e-01 -6.95533693e-01
-1.15500855e+00 4.71117616e-01 2.84120828e-01 -2.09580198e-01
-2.34321341e-01 5.78007102e-01 -2.62128234e-01 -6.25865757e-02
1.95761561e-01 1.99810639e-02 -1.23121805e-01 -1.23999543e-01
8.39262366e-01 2.89776623e-01 5.52427888e-01 -6.25463486e-01
-1.86302572e-01 5.32066047e-01 1.81466982e-01 -2.08405666e-02
1.47774434e+00 -3.40379804e-01 -9.87211894e-03 3.49704653e-01
1.07274032e+00 -2.08991319e-01 -1.38608944e+00 -5.60787059e-02
-1.40231833e-01 -8.10115159e-01 3.99736971e-01 -1.22343862e+00
-1.18757880e+00 1.19360304e+00 1.23738539e+00 1.53722167e-01
1.55328250e+00 -2.86072284e-01 7.35885859e-01 1.70071810e-01
4.83384132e-01 -1.26776612e+00 1.22831441e-01 -3.85868326e-02
9.09543872e-01 -1.69575167e+00 2.07370855e-02 1.42459944e-01
-5.57510018e-01 1.19608355e+00 3.82619292e-01 -1.58852376e-02
3.43802571e-01 7.12779239e-02 4.45497394e-01 1.02829617e-02
-6.17863536e-01 -3.78087610e-01 5.63980401e-01 9.56595004e-01
5.42450726e-01 -1.33029297e-01 -5.16370684e-02 1.39726058e-01
-1.29277483e-01 3.33065271e-01 4.09878284e-01 3.98680925e-01
-3.04595470e-01 -1.16223311e+00 -1.58280745e-01 3.69433790e-01
-4.36995953e-01 -1.20345503e-01 -1.43783033e-01 4.22802120e-01
4.34761465e-01 1.06419718e+00 4.79226699e-03 -4.58209038e-01
4.73214239e-01 -1.93641931e-01 5.41300893e-01 -2.35160500e-01
-6.96668446e-01 -3.82885188e-01 -3.85203101e-02 -1.03790259e+00
-7.05668330e-01 -5.43854058e-01 -3.40235263e-01 -3.74223888e-01
-4.60736193e-02 -3.11530501e-01 1.18334794e+00 7.53095090e-01
5.24291337e-01 4.29862857e-01 4.91073042e-01 -1.26524627e+00
-3.20371598e-01 -9.44705963e-01 -4.26830083e-01 9.27641451e-01
1.52579844e-01 -5.33584356e-01 -6.24805056e-02 2.96682775e-01]
|
[11.77366828918457, -1.8061224222183228]
|
588341f8-9423-48ad-84ba-5b72eeaabe68
|
feddef-robust-federated-learning-based
|
2210.04052
| null |
https://arxiv.org/abs/2210.04052v2
|
https://arxiv.org/pdf/2210.04052v2.pdf
|
FedDef: Defense Against Gradient Leakage in Federated Learning-based Network Intrusion Detection Systems
|
Deep learning (DL) methods have been widely applied to anomaly-based network intrusion detection system (NIDS) to detect malicious traffic. To expand the usage scenarios of DL-based methods, the federated learning (FL) framework allows multiple users to train a global model on the basis of respecting individual data privacy. However, it has not yet been systematically evaluated how robust FL-based NIDSs are against existing privacy attacks under existing defenses. To address this issue, we propose two privacy evaluation metrics designed for FL-based NIDSs, including (1) privacy score that evaluates the similarity between the original and recovered traffic features using reconstruction attacks, and (2) evasion rate against NIDSs using Generative Adversarial Network-based adversarial attack with the reconstructed benign traffic. We conduct experiments to show that existing defenses provide little protection that the corresponding adversarial traffic can even evade the SOTA NIDS Kitsune. To defend against such attacks and build a more robust FL-based NIDS, we further propose FedDef, a novel optimization-based input perturbation defense strategy with theoretical guarantee. It achieves both high utility by minimizing the gradient distance and strong privacy protection by maximizing the input distance. We experimentally evaluate four existing defenses on four datasets and show that our defense outperforms all the baselines in terms of privacy protection with up to 7 times higher privacy score, while maintaining model accuracy loss within 3% under optimal parameter combination.
|
['Xuewei Feng', 'Ke Xu', 'Qi Li', 'Yi Zhao', 'Jiahui Chen']
|
2022-10-08
| null | null | null | null |
['network-intrusion-detection']
|
['miscellaneous']
|
[ 6.97273090e-02 -2.24507213e-01 -2.03888297e-01 -3.78625542e-01
-6.50621176e-01 -9.77499485e-01 6.27033651e-01 -2.64192730e-01
-1.24889478e-01 5.82166135e-01 -3.44476372e-01 -8.75961125e-01
-1.65644914e-01 -9.39996123e-01 -6.08630776e-01 -6.36764288e-01
-1.47126734e-01 3.15144777e-01 2.95687139e-01 -4.44538966e-02
1.65703192e-01 1.09703469e+00 -9.71317768e-01 2.21921012e-01
5.86217940e-01 1.23173404e+00 -9.75778699e-01 4.34437931e-01
-1.15641706e-01 5.22793531e-01 -7.04539418e-01 -9.64392185e-01
1.16800165e+00 -3.11449349e-01 -5.69121480e-01 -4.55381840e-01
4.43559319e-01 -7.02650666e-01 -7.15159237e-01 1.26316762e+00
3.17045689e-01 1.35608641e-02 2.17594430e-01 -1.96830070e+00
-6.55317605e-01 2.97955066e-01 -5.53888261e-01 2.19663545e-01
7.67658502e-02 6.79737985e-01 7.22783804e-01 -3.20969671e-01
3.43981683e-01 1.48625445e+00 5.16927779e-01 1.01380742e+00
-1.23375523e+00 -1.11538494e+00 7.29304552e-02 9.82418209e-02
-1.22397482e+00 -4.96617496e-01 6.62438393e-01 9.88835543e-02
5.20226479e-01 7.44563639e-01 -8.65596086e-02 1.51767969e+00
2.95457635e-02 6.48295879e-01 9.84147251e-01 2.18078736e-02
4.01499391e-01 3.16032141e-01 1.33765057e-01 7.20320284e-01
6.04941308e-01 5.33496857e-01 -1.23670593e-01 -1.01733637e+00
3.97062600e-01 3.37934285e-01 -1.13035887e-01 -4.65806693e-01
-3.85530472e-01 1.01959777e+00 3.47092032e-01 -1.45120770e-01
-1.38526797e-01 -2.91282944e-02 6.97679281e-01 6.49020493e-01
2.45953262e-01 3.22011381e-01 -7.37897515e-01 2.01184213e-01
-4.48070139e-01 4.12129998e-01 1.23960590e+00 6.48085713e-01
5.95337927e-01 3.87105852e-01 -2.20969394e-01 3.53963643e-01
2.56535560e-01 5.06079674e-01 1.38014898e-01 -1.04458308e+00
3.56930852e-01 7.00530946e-01 -1.78015828e-01 -1.23811567e+00
2.92639256e-01 -2.07577825e-01 -7.63546467e-01 3.46938401e-01
4.19470519e-01 -4.89472538e-01 -4.80635852e-01 1.92283118e+00
4.79745477e-01 5.10955334e-01 1.41064197e-01 5.85649669e-01
2.87390482e-02 3.07131797e-01 3.74544449e-02 -2.02735528e-01
9.23644781e-01 -6.51925564e-01 -3.36640984e-01 6.74747080e-02
3.98520529e-01 -6.36385456e-02 9.34504449e-01 1.25837475e-01
-7.17293918e-01 6.16516322e-02 -9.92868125e-01 5.18849313e-01
-6.74901307e-01 -7.51718462e-01 6.39709651e-01 1.36602092e+00
-7.83885062e-01 5.24551034e-01 -8.00361037e-01 -2.28808179e-01
9.02415335e-01 5.91712356e-01 -4.45937902e-01 8.94923955e-02
-1.29875028e+00 3.06784809e-01 -1.09778326e-02 -4.71328110e-01
-1.31424022e+00 -1.04775023e+00 -5.24910569e-01 2.17841014e-01
7.19717205e-01 -6.39701366e-01 9.77706492e-01 -5.30483484e-01
-1.30968750e+00 8.07891548e-01 1.46872953e-01 -1.11128461e+00
7.62805462e-01 2.16674015e-01 -9.37723935e-01 3.37492600e-02
-8.59579593e-02 3.60634439e-02 9.96351480e-01 -1.35013425e+00
-5.17317295e-01 -7.12012231e-01 2.23453015e-01 -3.85089219e-01
-6.71530843e-01 3.00066978e-01 -9.05180126e-02 -6.36717081e-01
-3.71776760e-01 -7.25287259e-01 -4.51326728e-01 3.23988020e-01
-7.13441491e-01 1.32710829e-01 1.87808180e+00 -1.77559286e-01
1.09743881e+00 -2.13521647e+00 -5.47709405e-01 5.95042586e-01
4.46202666e-01 9.55411494e-01 -3.14055473e-01 1.42263189e-01
1.15919523e-01 6.03296340e-01 -3.27131599e-01 -1.98886976e-01
3.46375018e-01 3.00966024e-01 -8.50083530e-01 3.86429608e-01
-1.43805984e-02 8.82568598e-01 -5.76145649e-01 -1.13695674e-01
9.52778906e-02 3.37766200e-01 -7.30272174e-01 5.71591079e-01
-2.41420433e-01 4.38811362e-01 -5.96953690e-01 9.86970246e-01
9.91445065e-01 7.11887628e-02 1.50668144e-01 7.88663924e-02
3.97046387e-01 -9.34619606e-02 -1.05286682e+00 7.19459236e-01
-1.87502220e-01 1.70172751e-01 4.26850408e-01 -7.24590003e-01
9.81150150e-01 1.10027678e-01 6.44716918e-01 -5.28685451e-01
2.00544104e-01 7.70854801e-02 -7.85991699e-02 -1.84633672e-01
-2.02124625e-01 2.66478926e-01 -1.37596950e-01 7.74451554e-01
-2.40892977e-01 7.00317681e-01 -2.96039701e-01 2.36220524e-01
1.53674173e+00 -6.47485852e-01 7.90980607e-02 -2.89435182e-02
1.01252627e+00 -5.28571725e-01 8.22882473e-01 1.06546414e+00
-7.98612833e-01 -2.98085287e-02 6.95995688e-01 -7.13424385e-01
-7.24160135e-01 -1.30313420e+00 -1.69445630e-02 1.13058829e+00
1.31208450e-03 -3.91541541e-01 -8.41112435e-01 -1.64134002e+00
5.85222661e-01 6.90659642e-01 -5.37299454e-01 -4.74642307e-01
-5.61477959e-01 -7.08260238e-01 1.40413523e+00 1.62873805e-01
9.19250607e-01 -6.99795961e-01 4.37229834e-02 -2.13183641e-01
2.27074295e-01 -1.15709090e+00 -5.26584923e-01 -1.51981324e-01
-4.29773569e-01 -1.31120610e+00 2.47348640e-02 -1.23015828e-01
5.24149597e-01 2.08184361e-01 5.57789326e-01 1.90623105e-01
-3.73284012e-01 4.65881675e-01 1.01711027e-01 -4.74754125e-01
-7.01545477e-01 -5.55924848e-02 4.25015807e-01 5.19615531e-01
8.60533178e-01 -8.12082589e-01 -4.47499871e-01 4.72211152e-01
-1.06245148e+00 -9.69516635e-01 3.42426181e-01 5.11526108e-01
2.88527131e-01 1.09869204e-01 5.94838977e-01 -1.27028823e+00
1.00453413e+00 -7.91586876e-01 -6.73499286e-01 1.95436299e-01
-8.06178212e-01 -2.19479506e-03 1.17590272e+00 -4.83541548e-01
-7.97547936e-01 -2.97259837e-01 -1.36670738e-01 -8.98699403e-01
-3.69095355e-01 -3.65418404e-01 -6.92689538e-01 -8.30383301e-01
7.70481706e-01 3.02812725e-01 1.93956822e-01 -4.13087666e-01
2.90400147e-01 7.71729827e-01 4.56440896e-01 -8.29065740e-01
1.25741804e+00 5.93826175e-01 5.22198498e-01 -5.15585721e-01
-6.05212271e-01 -4.52904813e-02 -1.07962780e-01 1.91699922e-01
3.22349221e-01 -2.72717685e-01 -1.26433921e+00 3.96184176e-01
-9.35265899e-01 2.01581448e-01 -1.38996020e-01 -8.02060068e-02
-4.38725442e-01 7.44104624e-01 -6.75738871e-01 -8.57235730e-01
-7.83174872e-01 -1.20471716e+00 4.12995487e-01 -6.55667409e-02
1.46240622e-01 -8.81291032e-01 -8.84093111e-04 3.77459109e-01
9.39115644e-01 8.04958463e-01 1.01927829e+00 -1.47376299e+00
-6.66436672e-01 -5.87384582e-01 -2.44366512e-01 6.35926247e-01
1.42187759e-01 -9.40095708e-02 -8.68806720e-01 -4.86057043e-01
2.88684040e-01 -1.73326164e-01 4.90766019e-01 -2.52426058e-01
1.65948236e+00 -1.06708705e+00 -1.61093697e-01 1.14376032e+00
1.45864677e+00 4.09605265e-01 6.52197659e-01 3.45678598e-01
7.06233144e-01 4.18553859e-01 9.28842276e-02 5.81873178e-01
-1.55229241e-01 3.32195669e-01 8.12329710e-01 2.74810642e-01
3.59646648e-01 -4.43404734e-01 3.52908969e-01 -2.18232334e-01
3.69431764e-01 -4.06833678e-01 -5.67804515e-01 4.81428765e-02
-1.68501925e+00 -1.09100366e+00 1.64970249e-01 2.31611633e+00
4.49309677e-01 2.00938076e-01 4.10665512e-01 -7.55344033e-02
6.16183937e-01 3.32933575e-01 -1.13326943e+00 -8.90139580e-01
-1.20264463e-01 2.81972647e-01 8.09175372e-01 3.25242668e-01
-1.11153448e+00 9.81263459e-01 6.29535055e+00 7.27349579e-01
-9.14924860e-01 7.06618130e-02 7.10170805e-01 -2.00676724e-01
-2.16651876e-02 8.45604241e-02 -7.06885517e-01 6.33856595e-01
1.26949966e+00 -5.85783422e-01 8.64626110e-01 1.28320563e+00
-8.77513885e-02 9.42355573e-01 -9.73389506e-01 7.42367327e-01
-1.81032017e-01 -1.35703444e+00 5.12100160e-01 5.94110191e-01
4.64495629e-01 -1.15983054e-01 3.77760619e-01 4.79984164e-01
7.52520859e-01 -9.95909810e-01 -1.56475633e-01 2.90133893e-01
6.86398804e-01 -1.27292871e+00 3.78548592e-01 3.26947719e-01
-8.70080590e-01 -3.85584205e-01 -3.50667715e-01 4.64756936e-01
-1.41736984e-01 1.20154902e-01 -5.56983888e-01 5.42978823e-01
4.94294792e-01 4.17712219e-02 -5.15516639e-01 4.92060065e-01
5.99892661e-02 7.86262155e-01 -4.60685581e-01 1.90598547e-01
4.88916636e-01 -6.44463003e-02 1.08430552e+00 9.96211410e-01
-1.02950282e-01 -1.36940688e-01 3.85932595e-01 1.07984495e+00
-4.92689043e-01 -6.65435046e-02 -9.25922692e-01 -4.81322110e-02
8.88607025e-01 1.17141700e+00 7.68925995e-02 8.01295415e-02
-1.16432600e-01 9.21401739e-01 2.14882821e-01 2.82996356e-01
-8.70909989e-01 -5.03596127e-01 1.61896038e+00 2.28786647e-01
8.08676705e-03 7.22457767e-02 -1.16338797e-01 -1.16152990e+00
-2.99173947e-02 -1.30389869e+00 7.81424880e-01 1.16548181e-01
-1.81630647e+00 7.23943651e-01 -2.72988409e-01 -1.00685072e+00
-2.24113628e-01 -4.85503107e-01 -8.99136484e-01 7.61811256e-01
-1.12902093e+00 -1.37363791e+00 2.27087185e-01 1.18992901e+00
-1.12517916e-01 -9.19269562e-01 9.02771652e-01 3.10752749e-01
-1.02915192e+00 1.45008707e+00 2.24243943e-02 3.53252828e-01
5.08958459e-01 -9.33638811e-01 6.58381701e-01 1.16492546e+00
-7.46756569e-02 9.89784420e-01 3.87721688e-01 -6.26364529e-01
-1.50055707e+00 -1.52467084e+00 2.64912784e-01 -5.43572128e-01
7.22802222e-01 -4.26306188e-01 -1.07727158e+00 8.96586001e-01
-2.13986203e-01 5.45716643e-01 9.65195417e-01 -3.16189885e-01
-9.93980229e-01 -3.77054870e-01 -2.19381332e+00 8.08204889e-01
1.07314396e+00 -5.59834659e-01 9.88379046e-02 3.48628074e-01
1.10022736e+00 -2.48567807e-03 -5.98634064e-01 4.43770617e-01
3.31498086e-01 -1.15992332e+00 1.09678674e+00 -1.41517138e+00
-4.59234923e-01 -1.93051755e-01 -4.95937049e-01 -5.80270886e-01
-1.97196141e-01 -1.09029579e+00 -9.46704507e-01 1.37035668e+00
4.99128997e-02 -1.13853669e+00 1.19192564e+00 8.78179908e-01
5.52189887e-01 -9.06947255e-01 -9.99101639e-01 -1.15615308e+00
1.13745831e-01 -5.27360141e-01 1.33874691e+00 1.10021579e+00
-5.66328704e-01 -3.27866107e-01 -5.73881626e-01 5.48040986e-01
1.14575994e+00 -2.65205979e-01 1.09620273e+00 -1.13401449e+00
-2.02330589e-01 -3.69677842e-01 -6.29563153e-01 -3.83228302e-01
5.17216444e-01 -8.75782430e-01 -8.95033658e-01 -4.90440100e-01
-1.48301303e-01 -2.75706440e-01 -6.35652184e-01 7.65824914e-01
2.22179830e-01 7.54627287e-02 1.65850699e-01 1.53811961e-01
-5.35588741e-01 3.19706470e-01 6.00120664e-01 -2.03003827e-02
-6.03077859e-02 3.63066614e-01 -1.21449602e+00 5.73712528e-01
9.98558700e-01 -5.68147004e-01 -4.82445091e-01 3.46012190e-02
-5.61625957e-01 -2.16352165e-01 5.28829634e-01 -9.36925769e-01
2.38004804e-01 -5.92098475e-01 7.15931430e-02 -1.29252210e-01
-2.56242733e-02 -1.06538343e+00 1.24680094e-01 5.77122569e-01
-2.68222272e-01 6.87679499e-02 9.87900272e-02 8.78363907e-01
2.21353590e-01 3.12824279e-01 1.13990390e+00 3.59632145e-03
-3.46065521e-01 1.13763809e+00 -1.04409933e-01 1.70189857e-01
1.41097200e+00 -1.56092912e-01 -4.71894890e-01 -3.10405761e-01
-3.62485230e-01 2.27428675e-01 5.34503281e-01 3.80968153e-01
5.73604703e-01 -1.27834570e+00 -6.23673618e-01 7.83792496e-01
-1.89534184e-02 -7.20966280e-01 8.71578231e-02 2.10523278e-01
-2.88792193e-01 4.71519828e-01 -2.64575481e-01 -1.06068619e-01
-1.33903790e+00 9.63047206e-01 5.64613998e-01 -4.07868177e-01
-4.23507273e-01 5.88940322e-01 2.18422003e-02 -7.04567313e-01
5.47187150e-01 5.29319942e-01 4.28082287e-01 -8.45925510e-01
8.99093390e-01 7.98762083e-01 -1.69605669e-03 -5.87934494e-01
-6.29811585e-01 -1.12996124e-01 -3.73768300e-01 2.12479442e-01
9.95106637e-01 1.20164892e-02 -2.65432417e-01 -5.96118033e-01
1.65695786e+00 2.14446977e-01 -1.12468410e+00 -3.73993099e-01
6.29392564e-02 -1.11634362e+00 -1.06907360e-01 -9.03699756e-01
-1.52845478e+00 4.14252877e-01 6.57003462e-01 4.53783959e-01
1.18413305e+00 -5.71363151e-01 1.33664095e+00 4.86631095e-01
3.56760383e-01 -4.27890658e-01 -1.54122472e-01 3.80671114e-01
3.65514398e-01 -1.11558461e+00 -3.07257593e-01 -2.83932865e-01
-5.79576850e-01 9.32749450e-01 1.00801158e+00 -2.38871813e-01
7.48854399e-01 3.28351736e-01 -8.42167065e-02 -3.74030583e-02
-8.56080592e-01 3.12436342e-01 9.06203464e-02 8.75276625e-01
-4.77757573e-01 -2.81387493e-02 2.14569941e-02 8.36864412e-01
1.15513399e-01 -4.63481784e-01 1.36359558e-01 8.05169880e-01
-1.27249092e-01 -1.48011053e+00 -2.32710347e-01 4.82060492e-01
-8.36553454e-01 3.64472687e-01 -9.23770249e-01 5.15214801e-01
3.80434878e-02 9.38246846e-01 -3.31745237e-01 -8.03029239e-01
3.66968215e-01 -1.53851314e-02 -2.05651581e-01 -1.49043977e-01
-8.51637661e-01 -6.80126667e-01 -2.23866954e-01 -9.94394660e-01
3.44006985e-01 -3.14322203e-01 -8.04289639e-01 -1.20407176e+00
-1.06605396e-01 2.14226663e-01 4.38099563e-01 6.32878244e-01
8.44549179e-01 -4.92042154e-02 1.45920277e+00 -7.04476312e-02
-1.14808619e+00 -1.82554990e-01 -5.89337766e-01 5.40557742e-01
3.00479442e-01 -2.83580810e-01 -8.19031894e-01 -8.16721201e-01]
|
[5.728770732879639, 7.211320877075195]
|
1fef6c62-b606-4b2e-8314-c86634a7f2c6
|
mini-model-adaptation-efficiently-extending
|
2212.10503
| null |
https://arxiv.org/abs/2212.10503v2
|
https://arxiv.org/pdf/2212.10503v2.pdf
|
Mini-Model Adaptation: Efficiently Extending Pretrained Models to New Languages via Aligned Shallow Training
|
Prior work shows that it is possible to expand pretrained Masked Language Models (MLMs) to new languages by learning a new set of embeddings, while keeping the transformer body frozen. Despite learning a small subset of parameters, this approach is not compute-efficient, as training the new embeddings requires a full forward and backward pass over the entire model. We propose mini-model adaptation, a compute-efficient alternative that builds a shallow mini-model from a fraction of a large model's parameters. New language-specific embeddings can then be efficiently trained over the mini-model and plugged into the aligned large model for rapid cross-lingual transfer. We explore two approaches to learn mini-models: MiniJoint, which jointly pretrains the primary model and the mini-model using a single transformer with a secondary MLM head at a middle layer; and MiniPost, where we start from a regular pretrained model, build a mini-model by extracting and freezing a few layers, and learn a small number of parameters on top. Experiments on XNLI, MLQA and PAWS-X show that mini-model adaptation matches the performance of the standard approach using 2.3x less compute on average.
|
['Mikel Artetxe', 'Yihong Chen', 'Patrick Lewis', 'Kelly Marchisio']
|
2022-12-20
| null | null | null | null |
['cross-lingual-transfer']
|
['natural-language-processing']
|
[-7.59407580e-02 3.44160110e-01 -2.30873913e-01 -6.92504704e-01
-1.28279543e+00 -7.27482677e-01 5.94628632e-01 -5.20555563e-02
-8.20358157e-01 5.14487624e-01 2.02847913e-01 -5.55009604e-01
7.42043436e-01 -5.10827601e-01 -1.13932204e+00 -4.22033876e-01
-1.90073252e-02 8.85368705e-01 5.11338234e-01 -2.56525576e-01
-2.32185170e-01 3.29023033e-01 -9.65869784e-01 5.11391282e-01
7.84110069e-01 6.29069209e-01 3.33961904e-01 6.76480651e-01
-3.88277322e-01 3.85933697e-01 -2.51972526e-01 -5.07485569e-01
3.13867122e-01 -2.66255945e-01 -9.34596241e-01 6.70699105e-02
4.55394983e-01 -4.27421421e-01 -8.86714682e-02 5.35312057e-01
3.52493137e-01 -1.54833356e-02 2.52736211e-01 -9.32709336e-01
-9.25468743e-01 1.02010155e+00 -4.85604972e-01 1.30927280e-01
-9.05369222e-02 -1.00165650e-01 1.02187359e+00 -1.30317926e+00
5.74198604e-01 1.20480227e+00 7.70547688e-01 8.56126070e-01
-1.50167143e+00 -8.25042546e-01 3.83231521e-01 -2.34232754e-01
-1.38854361e+00 -8.24577630e-01 3.14080566e-01 -2.63559937e-01
1.52368546e+00 -1.74030915e-01 4.45569098e-01 6.71756089e-01
9.37135518e-02 6.09883308e-01 9.24173892e-01 -8.69303524e-01
2.04568226e-02 7.29174912e-01 6.23554364e-02 9.49991941e-01
-1.33356266e-02 -7.13672563e-02 -4.86301392e-01 -2.01894924e-01
3.38659942e-01 -1.27552047e-01 6.42247200e-02 -4.75664169e-01
-9.31905568e-01 8.79538774e-01 4.53611225e-01 4.15510207e-01
-8.84484053e-02 7.24105984e-02 3.27414870e-01 5.29281795e-01
5.74145555e-01 2.54775077e-01 -9.18871880e-01 1.79401666e-01
-1.28265810e+00 -5.43189384e-02 7.75784373e-01 1.01692140e+00
1.36903799e+00 -1.27441034e-01 3.45842838e-01 8.34822416e-01
2.80018598e-01 2.76968092e-01 8.36640358e-01 -6.17591023e-01
6.93017602e-01 3.95024538e-01 -1.51234522e-01 1.43715873e-01
-5.99039942e-02 -1.80182129e-01 -2.49560952e-01 -1.55603392e-02
4.50841784e-02 -3.54770899e-01 -1.19844997e+00 1.98550153e+00
3.52760643e-01 1.31590024e-01 7.46734515e-02 2.29212239e-01
1.64182946e-01 8.85940015e-01 1.13793977e-01 1.41993031e-01
1.15390754e+00 -1.36841142e+00 -9.20851603e-02 -8.18912387e-01
1.26912391e+00 -8.29636574e-01 1.22248971e+00 3.13710421e-02
-1.38328135e+00 -7.25701451e-01 -1.13832557e+00 -5.23792982e-01
-6.86480463e-01 9.45920646e-02 4.21595007e-01 4.49075133e-01
-1.60752618e+00 3.45309258e-01 -1.06595230e+00 -4.58091915e-01
1.90008819e-01 7.83706367e-01 -7.15495110e-01 -4.36729282e-01
-1.15656126e+00 1.22575831e+00 5.52394629e-01 -1.17213458e-01
-9.20085669e-01 -9.31591451e-01 -1.08086324e+00 -1.07007042e-01
-1.24419145e-01 -5.58760226e-01 1.36174238e+00 -8.22502494e-01
-1.69587505e+00 9.99169409e-01 -5.48469901e-01 -5.40605545e-01
2.81714145e-02 -3.63158882e-01 -3.26123953e-01 -2.06366196e-01
-8.55375156e-02 9.99112785e-01 7.61577427e-01 -1.23400199e+00
-5.35507262e-01 -3.27045053e-01 2.21111804e-01 1.73174292e-01
-5.44783890e-01 3.16441000e-01 -7.25596786e-01 -3.18837702e-01
8.10415298e-03 -1.05964375e+00 -2.58686334e-01 -2.98365533e-01
-5.37429862e-02 -1.91633731e-01 6.87462926e-01 -1.07343543e+00
1.38969374e+00 -2.04903173e+00 2.60360688e-01 7.18208924e-02
-5.02555398e-03 5.46360552e-01 -6.16202414e-01 6.08391285e-01
-1.07993320e-01 2.42624477e-01 -5.20840406e-01 -1.09331405e+00
7.27064610e-02 5.88622928e-01 -4.30101424e-01 1.25959888e-01
3.14116031e-01 1.02649224e+00 -5.39910614e-01 -4.61268216e-01
2.36269720e-02 5.59960008e-01 -1.11987817e+00 4.78048742e-01
-3.19917828e-01 8.44733939e-02 3.86696756e-02 2.70810753e-01
7.53734589e-01 -5.03729135e-02 4.02315766e-01 5.88684753e-02
-2.08743036e-01 8.73490691e-01 -8.59071434e-01 1.94747818e+00
-9.97062266e-01 3.45774561e-01 1.62767589e-01 -9.74161088e-01
6.75178230e-01 3.07683527e-01 9.37513784e-02 -4.52426970e-01
-1.94981322e-01 4.23198670e-01 -1.12300031e-01 -2.43488118e-01
4.60848302e-01 -5.05123854e-01 -2.64642030e-01 8.29008222e-01
6.02970302e-01 -2.64159203e-01 3.44102308e-02 3.84570181e-01
7.97322154e-01 2.93927610e-01 6.25249743e-02 -7.76408538e-02
5.42233169e-01 -2.67237008e-01 3.87341112e-01 3.82713705e-01
1.99076340e-01 3.57117802e-01 7.41340518e-02 -4.20375019e-01
-1.20375550e+00 -1.27364707e+00 1.75670981e-01 1.64678884e+00
-6.16838932e-01 -6.73853755e-01 -9.36207592e-01 -8.63407969e-01
2.08133813e-02 7.64051914e-01 -6.57577455e-01 -2.46535972e-01
-1.06344593e+00 -6.43165708e-01 6.28107071e-01 6.75030887e-01
2.38686353e-01 -7.96794832e-01 8.23199973e-02 3.63808960e-01
8.81136656e-02 -1.22173727e+00 -7.24121749e-01 5.62932730e-01
-9.54250336e-01 -3.45345199e-01 -4.91211534e-01 -1.32419908e+00
9.34952855e-01 -1.16162464e-01 1.19177055e+00 2.29277954e-01
1.49653301e-01 -1.15886532e-01 -1.17950309e-02 -1.09863155e-01
-7.21127033e-01 6.22770965e-01 1.84762686e-01 -6.12684451e-02
6.03662312e-01 -5.72541177e-01 1.81666464e-02 -2.24921837e-01
-9.30239737e-01 -6.82467641e-03 6.63903296e-01 7.26421595e-01
5.34763753e-01 -4.55925465e-01 2.44209692e-01 -1.16730356e+00
3.22889745e-01 -3.19158196e-01 -7.08073854e-01 4.69740182e-01
-5.13346970e-01 4.38234717e-01 8.14646423e-01 -4.35510010e-01
-9.38461363e-01 1.93935797e-01 -3.96385342e-01 -2.96660185e-01
2.25393161e-01 4.50252622e-01 -2.86706686e-01 -5.16501069e-02
5.22288024e-01 1.72594786e-01 -1.16200745e-01 -9.07949328e-01
8.65939617e-01 7.24779844e-01 3.77046973e-01 -6.79724216e-01
1.20518017e+00 1.89577773e-01 -7.73376584e-01 -6.38708472e-01
-8.20659518e-01 -4.15469170e-01 -1.22003102e+00 5.84810913e-01
7.38514245e-01 -1.12793040e+00 1.88951358e-01 3.30246776e-01
-1.26587951e+00 -9.19755638e-01 -4.36685234e-01 5.12818515e-01
-2.56501317e-01 1.35633603e-01 -7.93717742e-01 -1.87347472e-01
-3.29689771e-01 -9.13403332e-01 9.99272883e-01 -1.12935424e-01
-1.39666930e-01 -1.35290134e+00 4.76497501e-01 1.65152907e-01
6.08113468e-01 -6.18234396e-01 1.14837015e+00 -8.89042854e-01
-4.07449514e-01 -2.45168746e-01 1.72870681e-02 8.27275753e-01
6.42169416e-02 -2.47329175e-01 -1.09558630e+00 -6.10429645e-01
-1.98170945e-01 -6.18615448e-01 9.64016020e-01 -7.76327550e-02
7.53164589e-01 -4.17234719e-01 -5.06367803e-01 8.71181905e-01
1.48053896e+00 -1.93423554e-01 3.94869983e-01 2.23725021e-01
6.44419909e-01 9.72653180e-02 -3.11644450e-02 -1.64951280e-01
7.10363865e-01 4.50360119e-01 -1.46669000e-01 -2.87910432e-01
-1.20057054e-01 -5.01106262e-01 1.04461873e+00 1.69555116e+00
4.18557167e-01 1.11120574e-01 -8.89513791e-01 7.09933281e-01
-1.23023164e+00 -5.20704627e-01 8.37313950e-01 2.27109981e+00
1.40585923e+00 2.53436089e-01 1.43789947e-02 -2.34827518e-01
2.68302441e-01 3.14918347e-02 -3.52781773e-01 -8.90728891e-01
9.32656899e-02 8.52497101e-01 4.95066494e-01 1.28452194e+00
-8.24177504e-01 1.68035817e+00 7.00550938e+00 4.48552459e-01
-1.43475962e+00 5.76101482e-01 2.49682143e-01 -2.65017956e-01
-5.57053030e-01 4.95603293e-01 -1.42236662e+00 2.07931161e-01
1.60306251e+00 -1.31644890e-01 5.96136093e-01 7.90995181e-01
-2.59344637e-01 1.88108891e-01 -1.36099255e+00 3.41662288e-01
2.97308356e-01 -1.13083017e+00 2.91892260e-01 1.16395332e-01
7.45238483e-01 7.35946476e-01 -4.68098298e-02 8.40158522e-01
5.86053252e-01 -8.13761294e-01 5.53869128e-01 1.39510363e-01
9.13864434e-01 -6.52618766e-01 4.47375864e-01 4.43134010e-01
-1.23218417e+00 9.61730108e-02 -6.27763569e-01 3.15238573e-02
2.12796926e-01 2.00646132e-01 -1.02478504e+00 2.63209671e-01
5.08984208e-01 3.57148021e-01 -6.66704297e-01 5.76969206e-01
-3.90505224e-01 8.83514643e-01 -5.53842604e-01 4.45627630e-01
2.99059302e-01 -2.26848163e-02 -3.15014087e-02 1.39443946e+00
2.53716528e-01 -2.62486607e-01 2.73795515e-01 5.78312635e-01
-3.99788380e-01 9.99416187e-02 -3.61605346e-01 -2.59975582e-01
5.56595385e-01 1.18797898e+00 -9.83445123e-02 -7.33036935e-01
-8.34188521e-01 1.34583509e+00 9.17415559e-01 3.83886665e-01
-7.38127291e-01 -4.25244451e-01 6.70584381e-01 2.22839668e-01
6.46087766e-01 -5.08152246e-01 1.90631077e-01 -1.42395473e+00
6.98161274e-02 -9.30467665e-01 3.50755841e-01 -6.17177546e-01
-9.32288110e-01 9.85855043e-01 -8.41958746e-02 -5.95586598e-01
-6.30995631e-01 -6.48561537e-01 -5.13224900e-01 1.30526960e+00
-1.79764593e+00 -1.58979499e+00 2.96180516e-01 5.30171812e-01
4.47146475e-01 -7.51191452e-02 1.30221021e+00 3.47786605e-01
-5.39797902e-01 1.01907945e+00 1.14281744e-01 2.90307581e-01
8.73106360e-01 -1.07365596e+00 8.94400656e-01 1.05647922e+00
2.89099962e-01 1.02537799e+00 3.01588088e-01 -3.90249014e-01
-1.13508022e+00 -1.11280143e+00 1.71921480e+00 -6.05238616e-01
8.99805546e-01 -9.86761212e-01 -1.15615976e+00 1.53776979e+00
5.46179056e-01 2.53490239e-01 8.43154073e-01 2.82751143e-01
-6.99531913e-01 -1.73359767e-01 -8.12107265e-01 4.27465171e-01
6.38705254e-01 -1.02500165e+00 -9.62165534e-01 2.28801042e-01
1.06248629e+00 -3.07888389e-01 -8.94376040e-01 -6.72236234e-02
4.70120043e-01 -5.12728631e-01 9.23588336e-01 -9.19102252e-01
-8.83341208e-02 -5.32901287e-02 -2.66000271e-01 -1.44361818e+00
-2.45340243e-01 -7.65315235e-01 -2.06016585e-01 1.18420303e+00
9.07301188e-01 -8.30588877e-01 5.98952830e-01 5.90707481e-01
-2.56527066e-01 -7.85797775e-01 -8.55366230e-01 -7.52295256e-01
8.31176400e-01 -4.30662453e-01 7.30048478e-01 8.98851097e-01
2.20707636e-02 5.79835176e-01 -1.72626629e-01 3.36880296e-01
3.30198228e-01 -1.82357073e-01 9.29372907e-01 -8.13526273e-01
-4.60026920e-01 6.49385387e-03 1.24367185e-01 -1.29019845e+00
3.86154175e-01 -1.39120328e+00 2.37517487e-02 -1.57026565e+00
2.44620442e-02 -6.92080498e-01 -4.56017017e-01 8.62698138e-01
-1.42021850e-01 2.80344009e-01 1.73562095e-01 1.09921843e-01
-3.18298042e-01 2.89918482e-01 6.44678235e-01 8.00172624e-04
-2.68518120e-01 -2.71586359e-01 -6.85110807e-01 8.03691566e-01
6.74606323e-01 -7.15755522e-01 -3.56692731e-01 -1.03865087e+00
7.17591196e-02 -4.96497214e-01 -1.12422504e-01 -9.35057402e-01
2.79296398e-01 1.15029782e-01 3.36052716e-01 -2.70494998e-01
4.66438025e-01 -4.90283608e-01 -3.90327871e-01 2.72022516e-01
-3.51969481e-01 3.45053226e-01 6.16693377e-01 -7.20809400e-02
-1.02303073e-01 -2.97755092e-01 9.73867118e-01 -2.78119296e-01
-5.81921875e-01 3.85201037e-01 -1.72215506e-01 2.14796662e-01
5.10799587e-01 3.66357118e-02 1.72806740e-01 1.48866713e-01
-1.01781130e+00 1.96231917e-01 6.40014470e-01 5.38121343e-01
1.98485211e-01 -1.27867496e+00 -5.91829896e-01 6.79782033e-01
-2.60691382e-02 4.09063511e-03 -7.21062049e-02 5.63074112e-01
-2.97710329e-01 5.07768095e-01 9.62337554e-02 -3.48794371e-01
-1.01767433e+00 6.55098319e-01 5.04882455e-01 -6.72844529e-01
-3.45461160e-01 1.33264136e+00 3.41938883e-01 -1.14671004e+00
2.34266743e-02 -4.75884885e-01 3.54806960e-01 -3.92628498e-02
6.42475128e-01 -3.19829136e-01 2.53540784e-01 -7.22977042e-01
-4.83873010e-01 6.90212071e-01 -5.81185877e-01 -5.65175772e-01
1.44006836e+00 -1.76026419e-01 -3.20442617e-01 6.36063814e-01
1.59606111e+00 4.18521166e-01 -1.16243029e+00 -7.43389785e-01
9.25520286e-02 1.19709402e-01 -1.23585030e-01 -7.13897109e-01
-8.29891086e-01 1.25240600e+00 2.62201339e-01 -4.44973975e-01
9.54230309e-01 1.85240105e-01 1.15117264e+00 4.44256604e-01
3.11160147e-01 -1.19405496e+00 -6.57859668e-02 7.22792327e-01
6.47267878e-01 -9.07301366e-01 -1.83894113e-01 4.93169837e-02
-3.96840334e-01 9.49761868e-01 6.88685000e-01 -1.60236686e-01
9.03156996e-01 6.54868722e-01 3.70231181e-01 3.62699598e-01
-1.08484435e+00 -3.76081700e-03 2.36032382e-01 3.45201254e-01
6.01095378e-01 -2.66479384e-02 3.35679978e-01 6.33391857e-01
-4.63167727e-01 6.68543577e-02 1.04659952e-01 1.09754491e+00
-5.89353204e-01 -1.73536468e+00 8.19033850e-03 1.31407395e-01
-3.23572874e-01 -6.29318058e-01 -1.29591137e-01 9.05153692e-01
3.64088774e-01 4.36340898e-01 2.68023133e-01 -4.30361927e-01
2.04139739e-01 8.14625740e-01 5.56104124e-01 -1.18712294e+00
-6.06049001e-01 -3.37489485e-03 -6.64578527e-02 -4.99159485e-01
-8.85646269e-02 -5.93992233e-01 -1.43748593e+00 -2.63126493e-01
-3.93779606e-01 4.49693561e-01 7.33008444e-01 1.19325280e+00
5.25416493e-01 6.43813834e-02 5.07588625e-01 -8.26718569e-01
-7.17109084e-01 -1.10612297e+00 -1.71070486e-01 4.48558331e-02
4.90074068e-01 -1.03405967e-01 -4.53438073e-01 3.23712289e-01]
|
[10.909208297729492, 9.489717483520508]
|
f340fcd0-250a-4c51-b22a-6bb2e437ae7c
|
keystroke-patterns-as-prosody-in-digital
| null | null |
https://aclanthology.org/D14-1155
|
https://aclanthology.org/D14-1155.pdf
|
Keystroke Patterns as Prosody in Digital Writings: A Case Study with Deceptive Reviews and Essays
| null |
['Yejin Choi', 'Song Feng', 'Jun Seok Kang', 'Ritwik Banerjee']
|
2014-10-01
| null | null | null |
emnlp-2014-10
|
['deception-detection']
|
['miscellaneous']
|
[-8.63703638e-02 1.71006292e-01 -6.22772932e-01 -4.08054382e-01
-8.41685571e-03 -9.08429027e-01 6.55310392e-01 -6.53472245e-01
-2.85945535e-01 1.06888819e+00 -4.63127941e-02 -1.01159286e+00
-3.91567826e-01 -9.63214397e-01 -4.95059669e-01 -6.31337762e-01
-9.79754329e-01 7.25764990e-01 3.30370307e-01 -6.93831444e-01
7.03166842e-01 7.88774848e-01 -1.68942046e+00 7.18545914e-01
7.04417467e-01 8.52217197e-01 2.49141872e-01 1.14950800e+00
-1.95044339e-01 1.55633950e+00 -7.48382092e-01 -5.46825826e-01
3.13719302e-01 -1.23176083e-01 -7.22945035e-01 -1.01074085e-01
9.28529128e-02 -8.59008506e-02 -2.09758401e-01 9.22211111e-01
5.37373662e-01 4.49454933e-02 1.08379531e+00 -1.42548037e+00
-5.91619551e-01 6.10313773e-01 -4.01565880e-02 1.21627934e-01
1.03678203e+00 -5.39447069e-01 1.19919395e+00 -1.13026452e+00
7.20913768e-01 1.26888943e+00 8.66221786e-01 5.44149756e-01
-1.22286928e+00 -1.94712028e-01 -3.26822817e-01 -9.51717794e-02
-1.46558487e+00 -3.25250506e-01 4.25783843e-02 -2.08119690e-01
1.66093647e+00 1.26596653e+00 1.20609856e+00 1.01401424e+00
1.26658809e+00 8.34431887e-01 1.04267764e+00 -5.13792276e-01
3.35295945e-01 3.66983831e-01 1.54683650e-01 6.33519173e-01
8.40953708e-01 5.26628852e-01 -7.06372619e-01 -9.13127720e-01
9.33553874e-01 -2.94925272e-01 1.71355158e-01 -5.05680561e-01
-9.05919552e-01 6.91228509e-01 1.78732842e-01 3.83959889e-01
-1.39880210e-01 9.89067405e-02 1.26390755e-01 5.30987144e-01
-2.58292928e-02 6.47037446e-01 -9.11868811e-01 -1.33165747e-01
-8.71728659e-01 5.10332465e-01 1.25398111e+00 1.52653182e+00
1.24482810e-01 2.94908643e-01 -9.34252143e-02 3.17179203e-01
8.92314315e-01 1.01808000e+00 4.28362608e-01 -1.36146402e+00
-6.87414408e-02 1.72361732e-01 5.01781464e-01 -8.52631688e-01
-6.33224547e-01 -9.64177120e-03 -8.93263519e-01 4.49267089e-01
3.49161088e-01 4.57367361e-01 -8.02827001e-01 5.07305264e-01
4.33481112e-02 -2.34125629e-01 4.53833073e-01 5.55570945e-02
4.99930978e-01 3.76208365e-01 -1.34477139e-01 -5.73289394e-01
1.06082785e+00 -1.36716676e+00 -1.35299087e+00 2.33215362e-01
9.05734658e-01 -1.07320261e+00 4.35900748e-01 5.33875942e-01
-1.55548143e+00 -1.37560293e-01 -1.08699942e+00 1.78573877e-01
-7.27255583e-01 -3.14239264e-01 8.57801437e-01 1.43120694e+00
-1.60129595e+00 9.73287821e-01 -4.91727620e-01 6.59165755e-02
1.20568443e-02 8.24621081e-01 -2.64718989e-03 4.62812334e-01
-1.33193445e+00 1.08501506e+00 2.22979754e-01 -1.21242590e-01
-1.65216476e-01 -2.13068098e-01 -8.23704481e-01 -5.63443303e-01
-4.78693932e-01 -5.29636025e-01 1.44139910e+00 -2.59346128e-01
-1.65295815e+00 9.71794367e-01 -1.42069459e-01 -1.97814897e-01
6.14786744e-01 -1.28011424e-02 -8.31891418e-01 2.42498964e-01
-1.89849049e-01 5.76383233e-01 9.28263724e-01 -1.35132408e+00
-7.59897232e-01 -1.67359829e-01 -1.23336017e-01 2.66287565e-01
-1.25510961e-01 1.89734384e-01 2.11616129e-01 -1.12999000e-01
3.27147305e-01 -7.20919967e-01 -2.53068686e-01 -5.32041907e-01
-1.46512717e-01 -7.10518599e-01 7.70373225e-01 -4.51523662e-01
1.83705616e+00 -1.67618537e+00 -2.06720144e-01 3.98590982e-01
3.57815564e-01 -1.24705513e-03 2.40583986e-01 1.08380008e+00
-2.76906848e-01 7.32199550e-01 4.11965609e-01 -1.10722095e-01
2.24991128e-01 4.85861301e-01 -4.16602850e-01 3.05609167e-01
-1.29282743e-01 1.15307164e+00 -1.16605783e+00 -5.23096442e-01
4.11106765e-01 9.42391157e-02 -4.58732933e-01 4.79237735e-01
2.99364805e-01 1.47170946e-01 -3.56553018e-01 1.39399457e+00
1.15709066e+00 -1.31984919e-01 1.45911396e-01 5.30878425e-01
-3.79135728e-01 3.55090618e-01 -6.96863770e-01 1.05554795e+00
7.08333924e-02 5.00173986e-01 1.02364108e-01 -7.94621468e-01
3.33247900e-01 8.61540735e-01 4.37155962e-01 -9.67555881e-01
-2.26398129e-02 6.92409754e-01 1.12803578e-01 -6.07703328e-01
7.58228302e-01 3.94563079e-02 -3.64872098e-01 6.40070081e-01
-2.37588286e-01 -6.59476995e-01 9.64643434e-02 3.08100313e-01
6.22585893e-01 -5.18246442e-02 5.62923312e-01 -1.05613089e+00
7.86340594e-01 -1.82965681e-01 -1.81299388e-01 1.03415680e+00
-3.09923887e-01 3.19085121e-01 2.99841821e-01 -6.65102363e-01
-6.45341039e-01 -1.12307119e+00 -4.89381433e-01 1.30636716e+00
3.24267983e-01 -4.39044595e-01 -9.54439282e-01 -2.49762803e-01
1.77620783e-01 6.89606130e-01 -5.90509653e-01 3.84124845e-01
-5.03739953e-01 -8.32535863e-01 7.39044368e-01 3.45434904e-01
-5.07752821e-02 -1.33414865e+00 -6.58416986e-01 1.25490099e-01
-2.20292807e-01 -6.63697243e-01 -6.23428151e-02 4.48765576e-01
-1.35989368e+00 -5.18594682e-01 -6.66252747e-02 -8.20914626e-01
5.87345481e-01 2.46782884e-01 1.27047324e+00 5.39230824e-01
-2.31483161e-01 4.26904231e-01 -1.21292919e-01 -4.95818377e-01
-4.59671497e-01 -8.00336525e-02 5.28869390e-01 -5.87835789e-01
5.19427478e-01 -2.50617653e-01 -7.29350567e-01 5.37953973e-01
-6.88540697e-01 1.62748516e-01 1.79803044e-01 1.04410267e+00
1.35816500e-01 -9.34035778e-02 1.22507080e-01 -6.38007045e-01
8.72274399e-01 -1.69219792e-01 -3.78732830e-01 5.77745810e-02
-6.77108407e-01 -3.74140263e-01 3.21430594e-01 -3.25342178e-01
-1.01981449e+00 -4.87835288e-01 -9.82677937e-02 2.45538145e-01
1.11353043e-02 -1.46784872e-01 6.47139177e-02 -5.24923325e-01
8.02199244e-01 9.25758183e-02 1.99174434e-02 -6.80815242e-03
3.01039815e-01 7.09525108e-01 -6.82967342e-03 -6.68678164e-01
8.44880998e-01 4.91470337e-01 7.98524171e-02 -9.57177758e-01
-1.52186140e-01 -2.60129690e-01 -9.51962709e-01 -6.54426932e-01
6.56643391e-01 -6.78531289e-01 -9.10833478e-01 3.91110867e-01
-9.38691139e-01 -3.38627815e-01 -3.91645581e-01 4.25431967e-01
-1.01278400e+00 2.75717527e-02 -3.90154392e-01 -1.27895141e+00
-5.10977268e-01 -1.02017939e+00 9.43384409e-01 5.30070923e-02
-5.10597289e-01 -1.26927447e+00 5.87685481e-02 2.71537274e-01
1.81734428e-01 -1.73075795e-01 6.90226793e-01 -2.38256708e-01
-4.24233019e-01 -1.53791070e-01 2.34436691e-02 -1.39755070e-01
1.70832314e-02 4.95917559e-01 -9.81751978e-01 -5.31145096e-01
6.65065646e-02 -1.92070693e-01 -1.08835101e-01 6.52520418e-01
5.91872573e-01 -2.29931593e-01 -8.56000841e-01 5.40386558e-01
1.38545322e+00 3.85070026e-01 5.32770038e-01 7.28214979e-01
1.41836226e-01 5.53460240e-01 9.17806149e-01 4.63203549e-01
1.30579369e-02 3.28798652e-01 2.40537539e-01 1.49327129e-01
1.11720070e-01 -1.54819340e-01 3.77893507e-01 1.16112018e+00
-8.18235934e-01 -2.69281328e-01 -5.07867396e-01 4.42987174e-01
-1.72482407e+00 -1.40330648e+00 -4.32368398e-01 6.90478683e-01
6.25676990e-01 1.56016424e-01 -1.48347050e-01 3.35214496e-01
4.99015123e-01 -2.03574806e-01 -1.19133167e-01 -1.06291151e+00
-1.43546045e-01 3.15233678e-01 7.37729073e-01 1.00061214e+00
-7.20721722e-01 1.03317809e+00 1.29781246e+01 1.02230716e+00
2.21112028e-01 1.03134915e-01 5.16071796e-01 3.48020852e-01
-4.36954498e-01 -4.56139445e-02 -1.04416132e+00 2.72933897e-02
1.38140702e+00 -4.30666685e-01 6.85999811e-01 5.44219851e-01
3.44648361e-01 -4.23268199e-01 -1.26188684e+00 5.26221812e-01
9.73738134e-02 -1.40886843e+00 -2.83300440e-04 6.85225725e-01
7.73699820e-01 -5.08050561e-01 6.22419357e-01 3.24184299e-01
6.09259963e-01 -1.14389277e+00 8.60300779e-01 2.53660440e-01
1.03040910e+00 -6.05088234e-01 5.67372203e-01 1.68872893e-01
-1.14389896e+00 -2.20873043e-01 -8.77727985e-01 -1.00755692e+00
3.93533185e-02 -1.81779593e-01 -4.29956943e-01 3.48861217e-01
9.58353162e-01 2.99398601e-01 -3.93658698e-01 9.95779395e-01
-4.78694476e-02 1.04875881e-02 -2.71853864e-01 -4.48467314e-01
4.83122796e-01 -3.54241252e-01 4.66730654e-01 1.00164843e+00
2.48499006e-01 3.51035744e-01 -9.84472036e-02 4.01770771e-01
5.45058846e-01 3.29446048e-02 -1.19659424e+00 -1.78908288e-01
2.83276141e-01 9.16795909e-01 -4.83487815e-01 -4.22520459e-01
-2.00212970e-01 8.62069130e-01 -3.55488248e-02 5.01107454e-01
-6.11489356e-01 -4.35615242e-01 9.72222984e-01 -1.27327025e-01
-1.14700586e-01 -3.48497719e-01 -6.23769283e-01 -7.30352640e-01
-5.89872956e-01 -4.54965204e-01 5.93606755e-02 -5.53365827e-01
-1.39813089e+00 5.79277515e-01 -2.27688253e-02 -1.40553558e+00
-6.99901402e-01 -1.27676582e+00 -4.76714373e-01 4.92853165e-01
-1.11898029e+00 -1.10984349e+00 2.50124663e-01 4.52870727e-01
1.64141744e-01 -5.34416080e-01 1.39563632e+00 3.57715860e-02
1.00637585e-01 9.24474537e-01 6.69434488e-01 -7.37814724e-01
5.56605101e-01 -1.27867436e+00 5.68737745e-01 -1.37897313e-01
-4.31265175e-01 9.05828118e-01 6.28349900e-01 -5.39804697e-01
-1.41196322e+00 -3.66917729e-01 1.08350635e+00 -9.83769417e-01
6.55218959e-01 -3.86345625e-01 4.23767231e-02 7.88592756e-01
7.15902448e-01 -6.18741751e-01 8.21781039e-01 -1.83753878e-01
1.80774391e-01 5.75296998e-01 -1.39248300e+00 6.12354755e-01
1.66275799e+00 -4.63594139e-01 -6.25784039e-01 7.60327101e-01
8.13696027e-01 -6.94087505e-01 -1.30082703e+00 3.34633321e-01
8.65424156e-01 -8.75409484e-01 1.61978090e+00 -1.32660246e+00
-4.43697497e-02 2.81152606e-01 -2.61993498e-01 -9.32519078e-01
-5.96193194e-01 -1.23518765e+00 -5.33532679e-01 -5.83747849e-02
5.96577883e-01 -1.13057327e+00 3.42365682e-01 8.74560475e-01
-2.82833427e-01 -6.42737269e-01 -1.06996536e+00 -1.32016802e+00
-3.35779637e-02 -1.45572275e-01 4.90409225e-01 7.63798356e-01
6.83744550e-01 1.09839931e-01 -6.36873543e-02 -1.00294888e-01
5.33176839e-01 9.55312885e-03 4.41501856e-01 -1.34294486e+00
3.86843324e-01 -5.75816095e-01 -3.07655483e-01 -9.45992947e-01
-8.85957032e-02 -8.12076271e-01 -6.53862000e-01 -1.28511906e+00
-8.32044985e-03 -1.91056758e-01 -1.11109078e-01 -1.65725678e-01
3.67937148e-01 2.13746816e-01 1.20859891e-02 1.03788137e-01
-3.71160030e-01 6.18435517e-02 1.29639816e+00 6.91750320e-05
-1.62315920e-01 4.85058486e-01 -4.81304944e-01 7.84440815e-01
8.58408585e-02 -2.99253196e-01 -6.78878546e-01 6.11881316e-02
6.69384480e-01 4.61409837e-02 3.24159935e-02 -7.42885649e-01
5.37211418e-01 -3.75702560e-01 4.78586555e-01 -1.32223868e+00
1.30741090e-01 -9.61415648e-01 6.95283338e-02 9.40189242e-01
2.74610907e-01 1.20424610e-02 8.66204947e-02 5.39500564e-02
-1.44871444e-01 -5.70943117e-01 9.21121240e-01 -4.07591403e-01
-4.92852688e-01 -5.20386267e-03 -1.03226590e+00 8.97834301e-02
9.92593169e-01 -7.84614205e-01 -3.59281451e-01 -4.20183957e-01
-8.29068601e-01 -1.95836127e-02 6.50830388e-01 3.03609259e-02
7.30431557e-01 -1.51530886e+00 -2.30721906e-01 7.18729138e-01
-3.22939813e-01 -3.74400020e-01 -1.70157343e-01 6.58265352e-01
-1.32361674e+00 1.02442718e+00 -5.41665435e-01 -4.55340147e-01
-1.14228773e+00 4.78126436e-01 4.28307921e-01 -2.41845414e-01
-2.15481281e-01 1.11954463e+00 2.71224789e-02 -8.23025763e-01
1.85185194e-01 -9.21545625e-02 -7.54407048e-01 3.55081353e-03
6.88606799e-01 1.05194807e+00 -2.91290224e-01 -6.04341030e-01
-4.56784427e-01 6.65885091e-01 2.32151806e-01 -2.87484169e-01
9.17833567e-01 -2.19243199e-01 -9.89108324e-01 4.28274393e-01
8.38715494e-01 -1.36269778e-01 -3.69319022e-02 4.16855574e-01
1.25943512e-01 -8.02164078e-01 -4.32406247e-01 -3.55811834e-01
-1.85641110e-01 5.54822803e-01 5.34874737e-01 8.99602413e-01
8.57008278e-01 -3.02566767e-01 8.18335712e-01 9.66778398e-01
5.72402716e-01 -1.68019545e+00 -2.13140488e-01 6.89524531e-01
9.12339568e-01 -9.28350806e-01 5.44190466e-01 -7.27165341e-01
-4.14997995e-01 1.32979155e+00 4.68304873e-01 -1.55325383e-01
1.27306652e+00 5.74917436e-01 1.14069022e-02 -3.70670199e-01
-9.44949508e-01 1.12705544e-01 3.75366658e-01 1.11147714e+00
5.06513238e-01 5.07374525e-01 -9.66500878e-01 3.21953118e-01
-7.28706717e-01 -2.34555230e-01 4.90474731e-01 1.41972518e+00
-6.43810987e-01 -1.20391917e+00 -7.23931909e-01 4.61561680e-01
-5.49773455e-01 -1.16372630e-01 -5.28106689e-01 8.46754074e-01
-5.76629937e-02 1.49448860e+00 -1.97535474e-03 -5.03491640e-01
4.11356747e-01 1.41089618e-01 7.21762300e-01 -1.23501487e-01
-9.04846430e-01 4.15413082e-01 3.84890139e-01 -1.23056793e+00
-8.58632207e-01 -1.05834293e+00 -1.40667629e+00 -1.19437599e+00
-5.12782812e-01 1.89310342e-01 3.83317530e-01 3.90289724e-01
-2.06836104e-01 2.85260603e-02 9.80917513e-01 -1.07949340e+00
-3.90341938e-01 -9.39418614e-01 -1.00026262e+00 -7.84516707e-02
2.89751232e-01 -8.00943017e-01 -7.83523321e-01 2.83909619e-01]
|
[-7.314663410186768, 3.725912094116211]
|
e08de16d-dcdd-4b00-aed9-3cdb208540aa
|
distance-based-authorship-verification-across
| null | null |
https://aclanthology.org/W19-5611
|
https://aclanthology.org/W19-5611.pdf
|
Distance-Based Authorship Verification Across Modern Standard Arabic Genres
| null |
['Hossam Ahmed']
|
2019-07-01
| null | null | null |
ws-2019-7
|
['authorship-verification']
|
['natural-language-processing']
|
[-8.63703638e-02 1.71006292e-01 -6.22772932e-01 -4.08054382e-01
-8.41685571e-03 -9.08429027e-01 6.55310392e-01 -6.53472245e-01
-2.85945535e-01 1.06888819e+00 -4.63127941e-02 -1.01159286e+00
-3.91567826e-01 -9.63214397e-01 -4.95059669e-01 -6.31337762e-01
-9.79754329e-01 7.25764990e-01 3.30370307e-01 -6.93831444e-01
7.03166842e-01 7.88774848e-01 -1.68942046e+00 7.18545914e-01
7.04417467e-01 8.52217197e-01 2.49141872e-01 1.14950800e+00
-1.95044339e-01 1.55633950e+00 -7.48382092e-01 -5.46825826e-01
3.13719302e-01 -1.23176083e-01 -7.22945035e-01 -1.01074085e-01
9.28529128e-02 -8.59008506e-02 -2.09758401e-01 9.22211111e-01
5.37373662e-01 4.49454933e-02 1.08379531e+00 -1.42548037e+00
-5.91619551e-01 6.10313773e-01 -4.01565880e-02 1.21627934e-01
1.03678203e+00 -5.39447069e-01 1.19919395e+00 -1.13026452e+00
7.20913768e-01 1.26888943e+00 8.66221786e-01 5.44149756e-01
-1.22286928e+00 -1.94712028e-01 -3.26822817e-01 -9.51717794e-02
-1.46558487e+00 -3.25250506e-01 4.25783843e-02 -2.08119690e-01
1.66093647e+00 1.26596653e+00 1.20609856e+00 1.01401424e+00
1.26658809e+00 8.34431887e-01 1.04267764e+00 -5.13792276e-01
3.35295945e-01 3.66983831e-01 1.54683650e-01 6.33519173e-01
8.40953708e-01 5.26628852e-01 -7.06372619e-01 -9.13127720e-01
9.33553874e-01 -2.94925272e-01 1.71355158e-01 -5.05680561e-01
-9.05919552e-01 6.91228509e-01 1.78732842e-01 3.83959889e-01
-1.39880210e-01 9.89067405e-02 1.26390755e-01 5.30987144e-01
-2.58292928e-02 6.47037446e-01 -9.11868811e-01 -1.33165747e-01
-8.71728659e-01 5.10332465e-01 1.25398111e+00 1.52653182e+00
1.24482810e-01 2.94908643e-01 -9.34252143e-02 3.17179203e-01
8.92314315e-01 1.01808000e+00 4.28362608e-01 -1.36146402e+00
-6.87414408e-02 1.72361732e-01 5.01781464e-01 -8.52631688e-01
-6.33224547e-01 -9.64177120e-03 -8.93263519e-01 4.49267089e-01
3.49161088e-01 4.57367361e-01 -8.02827001e-01 5.07305264e-01
4.33481112e-02 -2.34125629e-01 4.53833073e-01 5.55570945e-02
4.99930978e-01 3.76208365e-01 -1.34477139e-01 -5.73289394e-01
1.06082785e+00 -1.36716676e+00 -1.35299087e+00 2.33215362e-01
9.05734658e-01 -1.07320261e+00 4.35900748e-01 5.33875942e-01
-1.55548143e+00 -1.37560293e-01 -1.08699942e+00 1.78573877e-01
-7.27255583e-01 -3.14239264e-01 8.57801437e-01 1.43120694e+00
-1.60129595e+00 9.73287821e-01 -4.91727620e-01 6.59165755e-02
1.20568443e-02 8.24621081e-01 -2.64718989e-03 4.62812334e-01
-1.33193445e+00 1.08501506e+00 2.22979754e-01 -1.21242590e-01
-1.65216476e-01 -2.13068098e-01 -8.23704481e-01 -5.63443303e-01
-4.78693932e-01 -5.29636025e-01 1.44139910e+00 -2.59346128e-01
-1.65295815e+00 9.71794367e-01 -1.42069459e-01 -1.97814897e-01
6.14786744e-01 -1.28011424e-02 -8.31891418e-01 2.42498964e-01
-1.89849049e-01 5.76383233e-01 9.28263724e-01 -1.35132408e+00
-7.59897232e-01 -1.67359829e-01 -1.23336017e-01 2.66287565e-01
-1.25510961e-01 1.89734384e-01 2.11616129e-01 -1.12999000e-01
3.27147305e-01 -7.20919967e-01 -2.53068686e-01 -5.32041907e-01
-1.46512717e-01 -7.10518599e-01 7.70373225e-01 -4.51523662e-01
1.83705616e+00 -1.67618537e+00 -2.06720144e-01 3.98590982e-01
3.57815564e-01 -1.24705513e-03 2.40583986e-01 1.08380008e+00
-2.76906848e-01 7.32199550e-01 4.11965609e-01 -1.10722095e-01
2.24991128e-01 4.85861301e-01 -4.16602850e-01 3.05609167e-01
-1.29282743e-01 1.15307164e+00 -1.16605783e+00 -5.23096442e-01
4.11106765e-01 9.42391157e-02 -4.58732933e-01 4.79237735e-01
2.99364805e-01 1.47170946e-01 -3.56553018e-01 1.39399457e+00
1.15709066e+00 -1.31984919e-01 1.45911396e-01 5.30878425e-01
-3.79135728e-01 3.55090618e-01 -6.96863770e-01 1.05554795e+00
7.08333924e-02 5.00173986e-01 1.02364108e-01 -7.94621468e-01
3.33247900e-01 8.61540735e-01 4.37155962e-01 -9.67555881e-01
-2.26398129e-02 6.92409754e-01 1.12803578e-01 -6.07703328e-01
7.58228302e-01 3.94563079e-02 -3.64872098e-01 6.40070081e-01
-2.37588286e-01 -6.59476995e-01 9.64643434e-02 3.08100313e-01
6.22585893e-01 -5.18246442e-02 5.62923312e-01 -1.05613089e+00
7.86340594e-01 -1.82965681e-01 -1.81299388e-01 1.03415680e+00
-3.09923887e-01 3.19085121e-01 2.99841821e-01 -6.65102363e-01
-6.45341039e-01 -1.12307119e+00 -4.89381433e-01 1.30636716e+00
3.24267983e-01 -4.39044595e-01 -9.54439282e-01 -2.49762803e-01
1.77620783e-01 6.89606130e-01 -5.90509653e-01 3.84124845e-01
-5.03739953e-01 -8.32535863e-01 7.39044368e-01 3.45434904e-01
-5.07752821e-02 -1.33414865e+00 -6.58416986e-01 1.25490099e-01
-2.20292807e-01 -6.63697243e-01 -6.23428151e-02 4.48765576e-01
-1.35989368e+00 -5.18594682e-01 -6.66252747e-02 -8.20914626e-01
5.87345481e-01 2.46782884e-01 1.27047324e+00 5.39230824e-01
-2.31483161e-01 4.26904231e-01 -1.21292919e-01 -4.95818377e-01
-4.59671497e-01 -8.00336525e-02 5.28869390e-01 -5.87835789e-01
5.19427478e-01 -2.50617653e-01 -7.29350567e-01 5.37953973e-01
-6.88540697e-01 1.62748516e-01 1.79803044e-01 1.04410267e+00
1.35816500e-01 -9.34035778e-02 1.22507080e-01 -6.38007045e-01
8.72274399e-01 -1.69219792e-01 -3.78732830e-01 5.77745810e-02
-6.77108407e-01 -3.74140263e-01 3.21430594e-01 -3.25342178e-01
-1.01981449e+00 -4.87835288e-01 -9.82677937e-02 2.45538145e-01
1.11353043e-02 -1.46784872e-01 6.47139177e-02 -5.24923325e-01
8.02199244e-01 9.25758183e-02 1.99174434e-02 -6.80815242e-03
3.01039815e-01 7.09525108e-01 -6.82967342e-03 -6.68678164e-01
8.44880998e-01 4.91470337e-01 7.98524171e-02 -9.57177758e-01
-1.52186140e-01 -2.60129690e-01 -9.51962709e-01 -6.54426932e-01
6.56643391e-01 -6.78531289e-01 -9.10833478e-01 3.91110867e-01
-9.38691139e-01 -3.38627815e-01 -3.91645581e-01 4.25431967e-01
-1.01278400e+00 2.75717527e-02 -3.90154392e-01 -1.27895141e+00
-5.10977268e-01 -1.02017939e+00 9.43384409e-01 5.30070923e-02
-5.10597289e-01 -1.26927447e+00 5.87685481e-02 2.71537274e-01
1.81734428e-01 -1.73075795e-01 6.90226793e-01 -2.38256708e-01
-4.24233019e-01 -1.53791070e-01 2.34436691e-02 -1.39755070e-01
1.70832314e-02 4.95917559e-01 -9.81751978e-01 -5.31145096e-01
6.65065646e-02 -1.92070693e-01 -1.08835101e-01 6.52520418e-01
5.91872573e-01 -2.29931593e-01 -8.56000841e-01 5.40386558e-01
1.38545322e+00 3.85070026e-01 5.32770038e-01 7.28214979e-01
1.41836226e-01 5.53460240e-01 9.17806149e-01 4.63203549e-01
1.30579369e-02 3.28798652e-01 2.40537539e-01 1.49327129e-01
1.11720070e-01 -1.54819340e-01 3.77893507e-01 1.16112018e+00
-8.18235934e-01 -2.69281328e-01 -5.07867396e-01 4.42987174e-01
-1.72482407e+00 -1.40330648e+00 -4.32368398e-01 6.90478683e-01
6.25676990e-01 1.56016424e-01 -1.48347050e-01 3.35214496e-01
4.99015123e-01 -2.03574806e-01 -1.19133167e-01 -1.06291151e+00
-1.43546045e-01 3.15233678e-01 7.37729073e-01 1.00061214e+00
-7.20721722e-01 1.03317809e+00 1.29781246e+01 1.02230716e+00
2.21112028e-01 1.03134915e-01 5.16071796e-01 3.48020852e-01
-4.36954498e-01 -4.56139445e-02 -1.04416132e+00 2.72933897e-02
1.38140702e+00 -4.30666685e-01 6.85999811e-01 5.44219851e-01
3.44648361e-01 -4.23268199e-01 -1.26188684e+00 5.26221812e-01
9.73738134e-02 -1.40886843e+00 -2.83300440e-04 6.85225725e-01
7.73699820e-01 -5.08050561e-01 6.22419357e-01 3.24184299e-01
6.09259963e-01 -1.14389277e+00 8.60300779e-01 2.53660440e-01
1.03040910e+00 -6.05088234e-01 5.67372203e-01 1.68872893e-01
-1.14389896e+00 -2.20873043e-01 -8.77727985e-01 -1.00755692e+00
3.93533185e-02 -1.81779593e-01 -4.29956943e-01 3.48861217e-01
9.58353162e-01 2.99398601e-01 -3.93658698e-01 9.95779395e-01
-4.78694476e-02 1.04875881e-02 -2.71853864e-01 -4.48467314e-01
4.83122796e-01 -3.54241252e-01 4.66730654e-01 1.00164843e+00
2.48499006e-01 3.51035744e-01 -9.84472036e-02 4.01770771e-01
5.45058846e-01 3.29446048e-02 -1.19659424e+00 -1.78908288e-01
2.83276141e-01 9.16795909e-01 -4.83487815e-01 -4.22520459e-01
-2.00212970e-01 8.62069130e-01 -3.55488248e-02 5.01107454e-01
-6.11489356e-01 -4.35615242e-01 9.72222984e-01 -1.27327025e-01
-1.14700586e-01 -3.48497719e-01 -6.23769283e-01 -7.30352640e-01
-5.89872956e-01 -4.54965204e-01 5.93606755e-02 -5.53365827e-01
-1.39813089e+00 5.79277515e-01 -2.27688253e-02 -1.40553558e+00
-6.99901402e-01 -1.27676582e+00 -4.76714373e-01 4.92853165e-01
-1.11898029e+00 -1.10984349e+00 2.50124663e-01 4.52870727e-01
1.64141744e-01 -5.34416080e-01 1.39563632e+00 3.57715860e-02
1.00637585e-01 9.24474537e-01 6.69434488e-01 -7.37814724e-01
5.56605101e-01 -1.27867436e+00 5.68737745e-01 -1.37897313e-01
-4.31265175e-01 9.05828118e-01 6.28349900e-01 -5.39804697e-01
-1.41196322e+00 -3.66917729e-01 1.08350635e+00 -9.83769417e-01
6.55218959e-01 -3.86345625e-01 4.23767231e-02 7.88592756e-01
7.15902448e-01 -6.18741751e-01 8.21781039e-01 -1.83753878e-01
1.80774391e-01 5.75296998e-01 -1.39248300e+00 6.12354755e-01
1.66275799e+00 -4.63594139e-01 -6.25784039e-01 7.60327101e-01
8.13696027e-01 -6.94087505e-01 -1.30082703e+00 3.34633321e-01
8.65424156e-01 -8.75409484e-01 1.61978090e+00 -1.32660246e+00
-4.43697497e-02 2.81152606e-01 -2.61993498e-01 -9.32519078e-01
-5.96193194e-01 -1.23518765e+00 -5.33532679e-01 -5.83747849e-02
5.96577883e-01 -1.13057327e+00 3.42365682e-01 8.74560475e-01
-2.82833427e-01 -6.42737269e-01 -1.06996536e+00 -1.32016802e+00
-3.35779637e-02 -1.45572275e-01 4.90409225e-01 7.63798356e-01
6.83744550e-01 1.09839931e-01 -6.36873543e-02 -1.00294888e-01
5.33176839e-01 9.55312885e-03 4.41501856e-01 -1.34294486e+00
3.86843324e-01 -5.75816095e-01 -3.07655483e-01 -9.45992947e-01
-8.85957032e-02 -8.12076271e-01 -6.53862000e-01 -1.28511906e+00
-8.32044985e-03 -1.91056758e-01 -1.11109078e-01 -1.65725678e-01
3.67937148e-01 2.13746816e-01 1.20859891e-02 1.03788137e-01
-3.71160030e-01 6.18435517e-02 1.29639816e+00 6.91750320e-05
-1.62315920e-01 4.85058486e-01 -4.81304944e-01 7.84440815e-01
8.58408585e-02 -2.99253196e-01 -6.78878546e-01 6.11881316e-02
6.69384480e-01 4.61409837e-02 3.24159935e-02 -7.42885649e-01
5.37211418e-01 -3.75702560e-01 4.78586555e-01 -1.32223868e+00
1.30741090e-01 -9.61415648e-01 6.95283338e-02 9.40189242e-01
2.74610907e-01 1.20424610e-02 8.66204947e-02 5.39500564e-02
-1.44871444e-01 -5.70943117e-01 9.21121240e-01 -4.07591403e-01
-4.92852688e-01 -5.20386267e-03 -1.03226590e+00 8.97834301e-02
9.92593169e-01 -7.84614205e-01 -3.59281451e-01 -4.20183957e-01
-8.29068601e-01 -1.95836127e-02 6.50830388e-01 3.03609259e-02
7.30431557e-01 -1.51530886e+00 -2.30721906e-01 7.18729138e-01
-3.22939813e-01 -3.74400020e-01 -1.70157343e-01 6.58265352e-01
-1.32361674e+00 1.02442718e+00 -5.41665435e-01 -4.55340147e-01
-1.14228773e+00 4.78126436e-01 4.28307921e-01 -2.41845414e-01
-2.15481281e-01 1.11954463e+00 2.71224789e-02 -8.23025763e-01
1.85185194e-01 -9.21545625e-02 -7.54407048e-01 3.55081353e-03
6.88606799e-01 1.05194807e+00 -2.91290224e-01 -6.04341030e-01
-4.56784427e-01 6.65885091e-01 2.32151806e-01 -2.87484169e-01
9.17833567e-01 -2.19243199e-01 -9.89108324e-01 4.28274393e-01
8.38715494e-01 -1.36269778e-01 -3.69319022e-02 4.16855574e-01
1.25943512e-01 -8.02164078e-01 -4.32406247e-01 -3.55811834e-01
-1.85641110e-01 5.54822803e-01 5.34874737e-01 8.99602413e-01
8.57008278e-01 -3.02566767e-01 8.18335712e-01 9.66778398e-01
5.72402716e-01 -1.68019545e+00 -2.13140488e-01 6.89524531e-01
9.12339568e-01 -9.28350806e-01 5.44190466e-01 -7.27165341e-01
-4.14997995e-01 1.32979155e+00 4.68304873e-01 -1.55325383e-01
1.27306652e+00 5.74917436e-01 1.14069022e-02 -3.70670199e-01
-9.44949508e-01 1.12705544e-01 3.75366658e-01 1.11147714e+00
5.06513238e-01 5.07374525e-01 -9.66500878e-01 3.21953118e-01
-7.28706717e-01 -2.34555230e-01 4.90474731e-01 1.41972518e+00
-6.43810987e-01 -1.20391917e+00 -7.23931909e-01 4.61561680e-01
-5.49773455e-01 -1.16372630e-01 -5.28106689e-01 8.46754074e-01
-5.76629937e-02 1.49448860e+00 -1.97535474e-03 -5.03491640e-01
4.11356747e-01 1.41089618e-01 7.21762300e-01 -1.23501487e-01
-9.04846430e-01 4.15413082e-01 3.84890139e-01 -1.23056793e+00
-8.58632207e-01 -1.05834293e+00 -1.40667629e+00 -1.19437599e+00
-5.12782812e-01 1.89310342e-01 3.83317530e-01 3.90289724e-01
-2.06836104e-01 2.85260603e-02 9.80917513e-01 -1.07949340e+00
-3.90341938e-01 -9.39418614e-01 -1.00026262e+00 -7.84516707e-02
2.89751232e-01 -8.00943017e-01 -7.83523321e-01 2.83909619e-01]
|
[-7.201930046081543, 3.826484203338623]
|
b0404fab-7333-409d-aec3-c60d94261c05
|
amortized-inference-for-gaussian-process
|
2306.09819
| null |
https://arxiv.org/abs/2306.09819v1
|
https://arxiv.org/pdf/2306.09819v1.pdf
|
Amortized Inference for Gaussian Process Hyperparameters of Structured Kernels
|
Learning the kernel parameters for Gaussian processes is often the computational bottleneck in applications such as online learning, Bayesian optimization, or active learning. Amortizing parameter inference over different datasets is a promising approach to dramatically speed up training time. However, existing methods restrict the amortized inference procedure to a fixed kernel structure. The amortization network must be redesigned manually and trained again in case a different kernel is employed, which leads to a large overhead in design time and training time. We propose amortizing kernel parameter inference over a complete kernel-structure-family rather than a fixed kernel structure. We do that via defining an amortization network over pairs of datasets and kernel structures. This enables fast kernel inference for each element in the kernel family without retraining the amortization network. As a by-product, our amortization network is able to do fast ensembling over kernel structures. In our experiments, we show drastically reduced inference time combined with competitive test performance for a large set of kernels and datasets.
|
['Christoph Zimmer', 'Mona Meister', 'Matthias Bitzer']
|
2023-06-16
| null | null | null | null |
['active-learning', 'gaussian-processes', 'bayesian-optimization', 'active-learning']
|
['methodology', 'methodology', 'methodology', 'natural-language-processing']
|
[-2.14950189e-01 -2.71732062e-01 -5.06330058e-02 -5.98194003e-01
-8.30528855e-01 -8.35677266e-01 3.17135483e-01 3.92913401e-01
-8.54409456e-01 5.34029663e-01 -4.02226120e-01 -5.63767910e-01
-3.59692514e-01 -8.60497415e-01 -7.99798369e-01 -7.17889428e-01
-2.48652115e-01 7.61556447e-01 4.29976523e-01 5.31367362e-01
1.82946652e-01 8.20289850e-01 -1.27398181e+00 -1.44475684e-01
9.22525406e-01 7.62276649e-01 -1.17094621e-01 8.90921533e-01
-2.15999633e-01 4.05672580e-01 -4.22240466e-01 -6.93180740e-01
2.74629503e-01 -5.80245852e-02 -8.32472801e-01 -1.69435088e-02
5.23481846e-01 -1.85755298e-01 1.60884783e-01 1.05450797e+00
4.19695884e-01 3.19501996e-01 6.55564368e-01 -1.23993289e+00
1.33545220e-01 9.98453557e-01 -4.92105812e-01 -8.96351412e-03
-2.76162565e-01 5.93554527e-02 8.73700917e-01 -7.22798526e-01
2.68351454e-02 9.98715878e-01 7.66921580e-01 9.29320827e-02
-1.87985289e+00 -5.29646099e-01 3.39583397e-01 -7.56705329e-02
-1.55550027e+00 -3.00036341e-01 3.08043540e-01 -3.20640564e-01
8.46682668e-01 3.26932728e-01 5.58370888e-01 5.40550351e-01
-1.75444156e-01 6.77699924e-01 8.08557153e-01 -2.92135358e-01
8.06814969e-01 3.26789796e-01 4.80645269e-01 8.62089634e-01
4.77760792e-01 -4.41765487e-01 -1.40774280e-01 -7.14704990e-01
8.04866970e-01 2.27145925e-01 -2.36263275e-01 -6.80010378e-01
-1.11334515e+00 7.86502719e-01 4.43216087e-03 -2.75920123e-01
-2.43292972e-01 6.28178596e-01 6.22675121e-01 5.17134666e-01
4.98538017e-01 4.73722547e-01 -7.41559565e-01 -3.19269717e-01
-8.87636125e-01 3.52770984e-01 1.35484123e+00 5.40189266e-01
1.24247015e+00 -1.22771479e-01 -2.08246261e-02 1.02708149e+00
4.87995185e-02 5.29468954e-01 -1.79459061e-02 -8.37958753e-01
2.03136444e-01 4.36891049e-01 2.24273503e-01 -4.61275637e-01
-2.37703338e-01 -4.24433574e-02 -7.64325321e-01 3.72463971e-01
7.41250277e-01 -3.81181806e-01 -8.47093225e-01 1.68851578e+00
5.97656250e-01 2.53748655e-01 -2.55351901e-01 4.66620028e-01
-1.93230599e-01 5.51786721e-01 5.79136983e-02 -4.34659235e-02
1.39294410e+00 -6.85652971e-01 -1.70611143e-01 8.83151516e-02
8.69351447e-01 -7.06985474e-01 1.24189365e+00 6.11030340e-01
-1.00824022e+00 -2.61528164e-01 -8.71522427e-01 2.24621028e-01
-3.60533148e-01 2.64727056e-01 1.16501725e+00 7.77287066e-01
-8.37458193e-01 7.44373381e-01 -1.11805952e+00 -2.10002866e-02
4.08427268e-01 6.81431949e-01 -3.95586401e-01 4.83473800e-02
-7.54357278e-01 5.69089353e-01 6.66846871e-01 3.76016193e-04
-6.68026388e-01 -1.20589411e+00 -6.92396581e-01 3.15720409e-01
6.93909347e-01 -6.55868649e-01 1.25100720e+00 -5.86160481e-01
-1.80712044e+00 4.50922757e-01 1.03472821e-01 -6.65117860e-01
5.91159046e-01 -3.97011667e-01 1.10334806e-01 -1.44743383e-01
-4.01979029e-01 2.38257691e-01 1.25133586e+00 -5.77701509e-01
-3.26654673e-01 -2.23123625e-01 4.40674461e-02 -4.75253984e-02
-2.91305244e-01 -2.57054679e-02 -5.14355659e-01 -4.38327163e-01
-1.14885315e-01 -1.05935550e+00 -3.90257239e-01 5.59103899e-02
-3.86374623e-01 -3.71176451e-01 4.56701338e-01 -1.89557433e-01
1.09674275e+00 -2.13837028e+00 5.21369949e-02 4.75656539e-01
2.29408488e-01 2.02958748e-01 2.79938966e-01 1.38061970e-01
-5.28267324e-02 -7.01555610e-02 -4.50488567e-01 -4.45049107e-01
2.52797335e-01 2.47079611e-01 -4.27816451e-01 5.69452524e-01
1.55079812e-01 5.76599658e-01 -5.49310386e-01 -2.97441989e-01
1.23049729e-01 2.34743103e-01 -6.91755176e-01 3.82899046e-01
-3.70228589e-01 -2.29609057e-01 -2.20682144e-01 1.42076358e-01
8.53885412e-01 -4.39899117e-01 3.77104357e-02 -1.74888030e-01
-1.90711785e-02 1.88980967e-01 -1.62766171e+00 1.48043907e+00
-7.88326144e-01 2.21019343e-01 -6.75220713e-02 -1.15094316e+00
5.78019857e-01 1.20888792e-01 2.56479621e-01 1.37347698e-01
-6.37815148e-02 2.08858937e-01 -1.89276665e-01 -3.23908962e-02
1.96716085e-01 -4.76212949e-02 -2.38056213e-01 9.08512533e-01
1.56552330e-01 -2.58765351e-02 2.57630527e-01 1.38812810e-01
1.04171455e+00 -7.09165409e-02 1.55739918e-01 -4.01188999e-01
4.43347663e-01 -2.21563816e-01 4.13554579e-01 8.96191776e-01
2.74568677e-01 1.31024912e-01 8.25848162e-01 -5.89252114e-01
-1.14250791e+00 -1.34182441e+00 -1.95554569e-01 1.41568232e+00
-4.43234503e-01 -4.30201739e-01 -7.53641248e-01 -8.99968684e-01
2.36632898e-01 5.83501458e-01 -5.61460197e-01 -1.93530887e-01
-5.21046937e-01 -1.07479787e+00 7.36555517e-01 5.27734101e-01
5.33805728e-01 -4.96817201e-01 -5.37423134e-01 3.56839523e-02
5.87665319e-01 -8.82157564e-01 -5.65069675e-01 7.38826931e-01
-9.80701208e-01 -8.69753957e-01 -5.54377735e-01 -3.85633975e-01
8.86151195e-01 -3.66393998e-02 9.08064604e-01 -3.80833387e-01
-4.01139826e-01 2.49205038e-01 1.95119455e-01 -3.59539360e-01
-3.20207328e-01 2.72613883e-01 3.97886410e-02 1.75993592e-02
2.16159225e-01 -6.40548408e-01 -4.04387176e-01 2.01970086e-01
-9.70038533e-01 -8.54153112e-02 4.29358006e-01 7.82081842e-01
4.97238934e-01 2.40694717e-01 2.76177406e-01 -1.15008187e+00
5.07167101e-01 -3.83964419e-01 -1.45747602e+00 4.15830672e-01
-6.24717116e-01 7.33066976e-01 7.75096238e-01 -8.88787866e-01
-1.02795291e+00 3.42087388e-01 1.55787498e-01 -5.46733737e-01
-8.16295967e-02 3.21850121e-01 -1.21964410e-01 4.77401577e-02
7.28777230e-01 -7.63015747e-02 1.73831195e-01 -5.11667252e-01
4.09842521e-01 3.21806073e-01 2.08364129e-01 -8.54921460e-01
6.99195027e-01 3.96588624e-01 1.68235153e-01 -6.39008462e-01
-7.46106327e-01 -4.80459780e-01 -4.54493374e-01 2.58205295e-01
5.93742490e-01 -7.25669146e-01 -1.03303552e+00 5.98094702e-01
-1.07821143e+00 -6.37711585e-01 -4.52241480e-01 7.29928672e-01
-5.08981824e-01 1.64509043e-01 -4.65912521e-01 -8.55270743e-01
-3.54346573e-01 -1.02390099e+00 8.20217490e-01 1.05300307e-01
-1.17661059e-01 -1.21667862e+00 2.24935174e-01 -2.17099383e-01
4.20991689e-01 -3.45320880e-01 1.05664170e+00 -9.20197546e-01
-5.66184580e-01 -4.46216017e-01 -2.74066597e-01 4.81745332e-01
9.32907090e-02 -2.50570476e-02 -8.73795748e-01 -2.61169225e-01
-3.04045111e-01 -3.17224979e-01 1.03116691e+00 2.75366992e-01
1.48996258e+00 -2.35587716e-01 -1.00465469e-01 7.69001245e-01
1.30051839e+00 -3.90610784e-01 3.38514686e-01 -1.73228234e-01
6.94218457e-01 3.80258143e-01 2.49080122e-01 4.46864814e-01
-9.35352072e-02 4.04690266e-01 -9.16477069e-02 -8.02266598e-03
4.87352163e-01 4.64081988e-02 4.08955842e-01 3.53303105e-01
5.49657233e-02 3.85831632e-02 -8.57335746e-01 3.47473145e-01
-1.90933645e+00 -7.06225634e-01 1.89709872e-01 2.83679366e+00
1.20801413e+00 2.23014534e-01 3.31785709e-01 -6.62287623e-02
6.37169123e-01 -2.93861657e-01 -6.45517170e-01 -4.46274251e-01
3.77227783e-01 4.31882828e-01 8.16240609e-01 6.43541336e-01
-1.06939721e+00 8.79019558e-01 6.40473175e+00 1.08246577e+00
-1.13779032e+00 -3.57578844e-02 3.85982633e-01 -1.64048553e-01
8.72807391e-03 3.36780220e-01 -1.12149668e+00 4.02492017e-01
1.12238789e+00 -1.30698532e-01 5.14821351e-01 1.18261397e+00
-2.19139874e-01 -4.59581196e-01 -1.25624943e+00 1.00109994e+00
-2.63194591e-01 -1.20394266e+00 -9.32303965e-02 2.07146734e-01
2.49654710e-01 -7.92355984e-02 -2.01796025e-01 3.71173561e-01
7.27072120e-01 -8.39196682e-01 3.14333916e-01 4.27374572e-01
5.49920201e-01 -1.04464853e+00 6.00684166e-01 2.30340943e-01
-9.56878364e-01 1.91942081e-01 -5.85457265e-01 2.83383787e-01
5.19078858e-02 1.12144887e+00 -9.36947525e-01 9.45125967e-02
5.45742452e-01 8.00073296e-02 -4.35422480e-01 1.23296261e+00
4.01445590e-02 9.28524852e-01 -7.54196882e-01 -5.50084747e-02
-2.04300001e-01 -3.70754957e-01 2.17537001e-01 1.36300921e+00
5.14303781e-02 -3.40628773e-01 2.75895149e-01 8.97096157e-01
5.84212095e-02 -6.46066200e-03 -4.36910987e-01 -2.49162734e-01
5.24943888e-01 1.38236821e+00 -9.76080418e-01 -5.25525093e-01
-3.13765287e-01 1.20332885e+00 6.93367660e-01 3.58024091e-01
-1.03891313e+00 -7.24438846e-01 9.13976014e-01 -6.24997206e-02
6.55370295e-01 -4.19486344e-01 3.40314656e-02 -1.21279967e+00
-1.64309189e-01 -3.89901906e-01 5.07693708e-01 -2.82922238e-01
-1.52616882e+00 5.23843691e-02 5.54227054e-01 -6.71310842e-01
-2.21575022e-01 -6.74392760e-01 -5.36601663e-01 1.05546677e+00
-9.85970080e-01 -8.10494900e-01 -8.50076005e-02 5.70476711e-01
9.98490453e-02 1.31045327e-01 8.54016185e-01 2.29670152e-01
-6.73349500e-01 7.59252429e-01 -3.40136178e-02 3.85149242e-03
8.19698155e-01 -1.53893173e+00 5.28571963e-01 5.76317072e-01
1.01722144e-02 9.82751727e-01 5.81392586e-01 -4.99451697e-01
-1.49995661e+00 -9.88264978e-01 2.71369487e-01 -1.99388251e-01
1.00697303e+00 -5.68207383e-01 -1.10282600e+00 5.92154086e-01
-3.08898300e-01 2.18233496e-01 8.94170523e-01 7.18344390e-01
-7.94416547e-01 -3.49625677e-01 -8.99189413e-01 5.42542875e-01
6.21644557e-01 -6.68428421e-01 -7.00637922e-02 1.47909522e-01
6.59072995e-01 -1.34949654e-01 -1.02020395e+00 1.86223134e-01
5.87971210e-01 -5.82070112e-01 9.57541823e-01 -6.89140201e-01
-2.72668898e-01 -4.46730196e-01 1.49038807e-01 -1.18651807e+00
-1.39401346e-01 -7.87246346e-01 -4.97307241e-01 1.10438466e+00
5.22609770e-01 -9.03264582e-01 7.60953605e-01 6.71318114e-01
2.42112726e-01 -7.86953092e-01 -5.89653373e-01 -8.58452141e-01
-4.82191294e-02 -7.44353056e-01 6.93773389e-01 7.45843232e-01
-3.22400957e-01 4.89571065e-01 -1.91034988e-01 2.59439796e-01
8.08736563e-01 1.83724001e-01 1.08561993e+00 -1.43465698e+00
-8.55884731e-01 -3.46917838e-01 -2.65720010e-01 -9.49612677e-01
1.16080195e-01 -8.74462247e-01 2.36796867e-02 -8.53052557e-01
3.00097764e-01 -7.91103005e-01 -1.02904268e-01 6.05299354e-01
-2.92512327e-01 -6.02239370e-02 -2.83253372e-01 3.72359878e-03
-5.86574793e-01 2.44112030e-01 6.54537439e-01 2.44489953e-01
-3.14006627e-01 3.20807219e-01 -3.16367000e-01 8.20577741e-01
9.54379201e-01 -4.53693718e-01 -6.82377934e-01 -8.81300047e-02
5.38820982e-01 -1.69819340e-01 5.55826724e-01 -8.27646077e-01
3.57829332e-01 -2.28009105e-01 4.22416478e-01 -5.21100998e-01
2.09780276e-01 -6.85591936e-01 2.28737921e-01 1.90976888e-01
-3.26400310e-01 3.08692195e-02 4.04274583e-01 6.81739151e-01
2.75776595e-01 -3.94531250e-01 1.05053699e+00 5.04788011e-03
-1.88886330e-01 3.93213540e-01 -3.63303035e-01 2.66988724e-02
8.73528361e-01 -1.29439175e-01 -2.98613440e-02 5.07586785e-02
-8.00806582e-01 1.26891494e-01 5.33405066e-01 -1.53282881e-01
2.16542616e-01 -9.37333107e-01 -3.13858598e-01 3.40509355e-01
1.08417803e-02 3.88045669e-01 1.10930823e-01 9.70147908e-01
-5.28812468e-01 1.09513337e-02 1.23043045e-01 -6.74151063e-01
-1.21899223e+00 4.64673012e-01 3.50479484e-01 -4.74362791e-01
-5.14758468e-01 8.73464465e-01 1.07318930e-01 -4.63208169e-01
4.30184007e-01 -5.43503165e-01 3.85136217e-01 1.56769693e-01
6.23511910e-01 5.87345779e-01 1.33181319e-01 3.64124954e-01
-5.41148968e-02 2.96372414e-01 -5.09264767e-01 -3.93075764e-01
1.13178444e+00 2.25565016e-01 -3.13402236e-01 7.32760668e-01
1.24429524e+00 6.60971552e-02 -1.38094497e+00 -2.93078184e-01
1.24807656e-01 -2.66553998e-01 8.36843997e-02 -3.64968389e-01
-6.83955193e-01 5.82816958e-01 3.92857939e-01 1.69272900e-01
9.17021155e-01 1.04863003e-01 4.84213084e-01 9.43415940e-01
3.64291400e-01 -1.19346118e+00 -1.17979713e-01 4.64854002e-01
2.73595899e-01 -1.02156937e+00 2.48884112e-01 -3.84150416e-01
-2.46037602e-01 1.06025708e+00 3.82910103e-01 -2.62968898e-01
1.00708115e+00 6.23625457e-01 -4.10705358e-01 -1.16527773e-01
-7.96686292e-01 -2.01042928e-02 1.58365950e-01 1.92582846e-01
2.11475819e-01 8.98027718e-02 2.06230227e-02 3.26683939e-01
-1.35213643e-01 -3.45501527e-02 2.60355264e-01 9.33106065e-01
-5.10300577e-01 -1.28721571e+00 -2.01255560e-01 6.30645394e-01
-4.02772874e-01 -1.63829952e-01 -3.73682007e-03 5.94153166e-01
-1.61579713e-01 2.74498016e-01 3.86102706e-01 5.32840088e-04
6.98697716e-02 4.49981332e-01 6.86496317e-01 -6.53056622e-01
-4.38027024e-01 -4.00954112e-02 -2.92818714e-02 -3.98543090e-01
3.26450765e-02 -5.09454131e-01 -1.26012337e+00 -4.99679774e-01
-6.11206472e-01 3.91780347e-01 7.23989666e-01 7.64185905e-01
3.90708864e-01 1.14806294e-01 4.35335904e-01 -4.29560453e-01
-9.93134081e-01 -8.04133415e-01 -7.30770350e-01 9.25177485e-02
1.11190461e-01 -5.43063462e-01 -4.94774491e-01 6.30882978e-02]
|
[7.398060321807861, 4.1565093994140625]
|
cde73df9-12a5-4a41-9438-bfcc2c07892c
|
emotion-recognition-in-conversation-research
|
1905.02947
| null |
https://arxiv.org/abs/1905.02947v1
|
https://arxiv.org/pdf/1905.02947v1.pdf
|
Emotion Recognition in Conversation: Research Challenges, Datasets, and Recent Advances
|
Emotion is intrinsic to humans and consequently emotion understanding is a key part of human-like artificial intelligence (AI). Emotion recognition in conversation (ERC) is becoming increasingly popular as a new research frontier in natural language processing (NLP) due to its ability to mine opinions from the plethora of publicly available conversational data in platforms such as Facebook, Youtube, Reddit, Twitter, and others. Moreover, it has potential applications in health-care systems (as a tool for psychological analysis), education (understanding student frustration) and more. Additionally, ERC is also extremely important for generating emotion-aware dialogues that require an understanding of the user's emotions. Catering to these needs calls for effective and scalable conversational emotion-recognition algorithms. However, it is a strenuous problem to solve because of several research challenges. In this paper, we discuss these challenges and shed light on the recent research in this field. We also describe the drawbacks of these approaches and discuss the reasons why they fail to successfully overcome the research challenges in ERC.
|
['Eduard Hovy', 'Soujanya Poria', 'Navonil Majumder', 'Rada Mihalcea']
|
2019-05-08
| null | null | null | null |
['emotion-recognition-in-conversation']
|
['natural-language-processing']
|
[ 8.14253837e-02 2.37589628e-01 -1.30351156e-01 -5.16788363e-01
-2.58083761e-01 -3.59693766e-01 4.37046319e-01 6.13869727e-01
-4.06267852e-01 9.37083066e-01 4.82196957e-01 -1.41320571e-01
2.44305357e-02 -5.32203734e-01 3.90721597e-02 -4.52838182e-01
1.48900807e-01 2.01993376e-01 -3.77268583e-01 -5.57391763e-01
4.45467889e-01 5.26768744e-01 -1.97836411e+00 3.18454266e-01
1.18109059e+00 8.33457768e-01 -2.10813031e-01 7.51661003e-01
-5.94095290e-01 1.01568389e+00 -7.50573695e-01 -6.28828883e-01
-6.41760826e-01 -6.56934917e-01 -1.12537527e+00 3.29551958e-02
-3.07333499e-01 3.46892685e-01 4.23717260e-01 9.30628359e-01
4.99124944e-01 3.70873988e-01 4.37215030e-01 -1.51768529e+00
-1.22959428e-02 1.31890222e-01 -2.01650560e-01 -4.37368788e-02
8.97069395e-01 -4.58328098e-01 6.14612460e-01 -4.88980860e-01
6.37172818e-01 1.05865729e+00 2.56375849e-01 9.18584406e-01
-5.32588661e-01 -4.40602452e-01 5.42005040e-02 4.80943382e-01
-7.78328717e-01 -3.72071773e-01 9.49678600e-01 -2.50379473e-01
8.09808731e-01 5.93230546e-01 9.02644634e-01 1.48706245e+00
7.01128095e-02 1.03344810e+00 1.33281696e+00 -7.12957501e-01
5.58642447e-01 7.83137202e-01 3.20967585e-01 3.42428803e-01
-3.31431389e-01 -5.80634534e-01 -7.34560072e-01 -2.33750433e-01
2.75908351e-01 -2.54351974e-01 -3.18648070e-01 1.24171078e-01
-8.64117742e-01 9.49827969e-01 -1.28464088e-01 5.63611984e-01
-6.09368026e-01 -4.50608283e-01 6.78294003e-01 4.95104730e-01
5.95574081e-01 6.69923842e-01 -4.43753451e-01 -1.29549193e+00
-2.57657021e-01 1.00822002e-01 1.39616299e+00 4.88042116e-01
3.36917132e-01 -5.34687713e-02 3.67814362e-01 1.12022769e+00
1.16436891e-01 -9.64894071e-02 8.25766623e-01 -9.18681860e-01
-3.22128199e-02 1.05072081e+00 4.91037406e-02 -1.21986592e+00
-5.11236787e-01 -1.17498107e-01 -9.47200835e-01 -2.68854320e-01
2.60627478e-01 -3.63070846e-01 -6.54185414e-02 1.57636702e+00
6.88175857e-01 1.12705283e-01 4.66364324e-01 7.46738374e-01
1.05051267e+00 6.92138553e-01 3.49635690e-01 -5.80396831e-01
1.75841570e+00 -5.88413477e-01 -1.23500133e+00 -2.56642252e-01
8.00533056e-01 -8.71228576e-01 9.28109765e-01 7.26028383e-01
-7.69216835e-01 -3.26794647e-02 -5.02117574e-01 -1.92530118e-02
-8.26752543e-01 -3.44844274e-02 1.02721405e+00 9.08583939e-01
-7.74096549e-01 1.38340101e-01 -4.38534290e-01 -8.12171280e-01
1.03063837e-01 1.51708975e-01 -3.12556654e-01 7.45039806e-02
-1.38498890e+00 1.05029678e+00 1.61709003e-02 7.42301121e-02
2.00676277e-01 -3.37280750e-01 -9.49929833e-01 2.80943215e-02
3.83736938e-01 -1.49524242e-01 1.22013628e+00 -1.19385207e+00
-2.02460003e+00 9.93233025e-01 -2.55812645e-01 -2.16724679e-01
1.39701664e-01 -1.28447086e-01 -6.12321913e-01 1.20666087e-01
-3.14051688e-01 4.80423748e-01 3.81702513e-01 -8.64292502e-01
-4.10390496e-01 -4.81371045e-01 -8.33978727e-02 4.51893359e-01
-4.91103023e-01 4.99714702e-01 -1.19523726e-01 -2.56197989e-01
-1.11283049e-01 -8.16045403e-01 -3.15761417e-01 -5.12229979e-01
-1.40848756e-01 -6.81898713e-01 9.46846247e-01 -3.71478230e-01
1.22451055e+00 -1.92186856e+00 6.07772544e-02 1.29661202e-01
-2.11333577e-02 5.09327769e-01 2.70030797e-01 8.18271577e-01
-2.00258475e-02 1.74441054e-01 6.02081753e-02 -5.20839505e-02
4.74285707e-02 1.63651913e-01 -1.23092286e-01 3.54942456e-02
2.53265232e-01 7.33665705e-01 -9.43592131e-01 -5.19234300e-01
3.42122018e-01 6.84513271e-01 -3.59650344e-01 3.98873031e-01
-6.83651567e-02 5.51524758e-01 -8.37201416e-01 5.35876155e-01
6.89258873e-02 -2.77793612e-02 2.58680373e-01 3.40462565e-01
-3.96598488e-01 1.77830786e-01 -9.42500174e-01 1.13976121e+00
-5.62252283e-01 7.49229908e-01 2.88656265e-01 -1.37485898e+00
1.16603982e+00 7.59445488e-01 5.84802866e-01 -5.40250659e-01
3.95048946e-01 7.54489377e-02 -1.00389734e-01 -9.91694212e-01
5.65261662e-01 -2.64637709e-01 -3.61517400e-01 6.59892857e-01
-1.87610358e-01 -3.11177999e-01 1.71438441e-01 2.10396618e-01
8.48015070e-01 -2.85857081e-01 6.63710594e-01 -2.14020666e-02
9.28990602e-01 -1.32858036e-02 6.72754526e-01 2.16787368e-01
-7.74222016e-01 -8.83483365e-02 7.89653480e-01 -5.63153386e-01
-2.46794358e-01 -5.24904653e-02 1.51909679e-01 1.00741506e+00
-2.66099960e-01 -5.21216452e-01 -8.25002849e-01 -3.88356119e-01
-5.91436088e-01 7.01720417e-01 -3.85271251e-01 -1.47270143e-01
-7.05509558e-02 -6.42276287e-01 2.92681545e-01 1.07158646e-01
5.18358827e-01 -1.64694810e+00 -8.50985229e-01 3.55782568e-01
-6.40768647e-01 -1.31390369e+00 3.04722726e-01 3.61716859e-02
-5.27602613e-01 -9.69677866e-01 -3.72647285e-01 -7.52840579e-01
4.00904894e-01 1.28053948e-01 1.12641621e+00 1.52758285e-01
-4.16921824e-01 7.99965501e-01 -7.35616744e-01 -9.85978246e-01
-4.79822725e-01 4.61479872e-02 9.23064277e-02 5.29700518e-02
9.16144967e-01 -5.15457153e-01 -2.20461145e-01 1.09243810e-01
-8.27609062e-01 1.13304712e-01 1.63423523e-01 4.99585778e-01
3.33460569e-02 2.43351534e-01 1.14802241e+00 -1.08278084e+00
1.36178458e+00 -7.15990603e-01 -9.92479026e-02 3.25759351e-01
-2.68120855e-01 -3.84930015e-01 5.34031987e-01 -2.95014679e-01
-1.27694595e+00 -1.47437707e-01 -3.36209029e-01 1.99575096e-01
-7.62865484e-01 6.33017182e-01 -4.01556641e-02 9.60758924e-02
3.10867518e-01 1.05804941e-02 1.50176167e-01 -1.73567191e-01
3.52615938e-02 1.14725149e+00 1.44301817e-01 -7.02266932e-01
5.88478111e-02 -1.92615353e-02 -3.86794895e-01 -1.36568558e+00
-9.49769199e-01 -7.28433669e-01 -2.84423947e-01 -8.63038480e-01
1.00235391e+00 -6.22394383e-01 -1.46996093e+00 2.31515691e-01
-1.26264071e+00 4.99425493e-02 -5.53654172e-02 5.56969762e-01
-3.57730508e-01 2.29170635e-01 -6.48163259e-01 -1.43288517e+00
-4.79833841e-01 -1.07547486e+00 6.12141907e-01 9.28123713e-01
-7.75811672e-01 -1.12225318e+00 -3.35964039e-02 9.06770289e-01
5.29251397e-01 3.23195636e-01 8.99734378e-01 -7.69650877e-01
1.95987061e-01 -3.81828427e-01 8.92682150e-02 2.42046535e-01
-1.30568013e-01 1.73589215e-01 -1.01231849e+00 3.48766863e-01
3.14506978e-01 -8.04662168e-01 3.72439437e-03 -6.07165061e-02
1.16626942e+00 -3.12353730e-01 7.50567392e-02 -2.84131229e-01
9.42537963e-01 4.96893227e-01 5.18938899e-01 1.99913204e-01
1.57350376e-01 1.37776899e+00 8.06422889e-01 7.11937428e-01
6.32053852e-01 4.03056502e-01 2.59766608e-01 8.02533776e-02
7.07503021e-01 1.79991767e-01 4.14339870e-01 1.22202837e+00
-1.82784528e-01 -1.78904966e-01 -9.61541414e-01 3.14276814e-01
-1.97725594e+00 -1.09462833e+00 -3.53415042e-01 1.89077437e+00
9.91590798e-01 -2.95798421e-01 -7.78611302e-02 3.44742596e-01
4.91707444e-01 -9.13753435e-02 -3.17623317e-01 -1.45887208e+00
1.27634034e-01 2.41884634e-01 -4.87094760e-01 3.07130933e-01
-7.94033885e-01 9.92613614e-01 5.48436165e+00 5.06588161e-01
-1.07805336e+00 -3.09596002e-01 9.06306326e-01 3.69083762e-01
-2.34227851e-02 -4.03410435e-01 -4.93868440e-01 2.16621518e-01
1.14805889e+00 -2.20285654e-01 3.81855935e-01 9.76503432e-01
4.84975457e-01 -5.79866052e-01 -7.44403064e-01 1.10042238e+00
2.30342045e-01 -8.72278094e-01 -5.05882204e-01 -1.87633000e-02
5.70016801e-01 -5.27260900e-01 -2.42232800e-01 6.49984300e-01
-7.10121170e-02 -9.78220761e-01 -1.64544106e-01 3.27787578e-01
1.23656899e-01 -1.09459233e+00 9.52488303e-01 6.66611910e-01
-5.90819001e-01 1.40473068e-01 -8.85495767e-02 -5.99190772e-01
8.96208920e-03 7.21072793e-01 -7.88784206e-01 2.80900061e-01
6.53064191e-01 5.46232343e-01 -2.00065091e-01 7.54233122e-01
-2.02800170e-01 5.88601112e-01 -1.63392290e-01 -6.20392263e-01
2.32021451e-01 -5.16038239e-01 3.04426908e-01 1.13819182e+00
1.34537429e-01 6.01479769e-01 1.28718987e-01 3.49411130e-01
-6.18530251e-02 6.76107287e-01 -6.43132627e-01 -5.67167103e-01
3.91939759e-01 1.67421710e+00 -8.59425426e-01 -2.27341712e-01
-4.13021982e-01 9.61806476e-01 2.49992162e-01 1.47635654e-01
-3.57754916e-01 -3.53902459e-01 8.22908401e-01 -4.28947568e-01
-3.89828503e-01 9.17033851e-03 -2.88074650e-02 -1.11239123e+00
-3.37120414e-01 -1.27999580e+00 3.63607824e-01 -7.21214414e-01
-1.19496858e+00 5.10210574e-01 -2.38949776e-01 -8.87011886e-01
-5.52961349e-01 -5.50192416e-01 -8.45623255e-01 4.19463068e-01
-1.46465003e+00 -6.33374751e-01 -3.83450121e-01 5.25458515e-01
5.46625972e-01 1.13745987e-01 1.41458702e+00 1.50317386e-01
-9.22709286e-01 6.22481853e-02 -3.45208704e-01 -3.55895124e-02
7.72118926e-01 -1.16959918e+00 -4.07518774e-01 2.14398697e-01
-1.26044482e-01 3.76453638e-01 8.89924884e-01 -2.62537539e-01
-1.57643473e+00 -4.94752496e-01 1.39168656e+00 -1.42594680e-01
5.52306771e-01 -2.21064597e-01 -1.00717437e+00 2.71511048e-01
5.62628508e-01 -4.85322773e-01 1.47876501e+00 4.26324844e-01
1.23132847e-01 2.81813920e-01 -1.22104788e+00 6.82438612e-01
2.89573520e-01 -3.91125411e-01 -5.07905960e-01 4.03890818e-01
3.53824675e-01 -2.93562382e-01 -1.00573552e+00 1.33458242e-01
4.70278382e-01 -1.19963837e+00 5.41252494e-01 -7.45119274e-01
5.76191604e-01 3.96388412e-01 3.57577384e-01 -1.32827187e+00
5.46933651e-01 -8.22698295e-01 2.27478035e-02 1.47734475e+00
8.78240094e-02 -7.32107103e-01 9.05323446e-01 1.26845264e+00
1.00718059e-01 -1.11115026e+00 -6.85887575e-01 -9.39983875e-02
-2.53956199e-01 -7.91388512e-01 3.32918286e-01 1.40837407e+00
8.42497647e-01 5.97607434e-01 -3.72487307e-01 -3.65240753e-01
9.21375006e-02 6.75314292e-02 8.02999258e-01 -1.52082777e+00
2.00327933e-01 -4.79893655e-01 -3.92954171e-01 -4.50631082e-01
4.56187040e-01 -3.40815365e-01 -1.09407388e-01 -1.45167756e+00
-1.00042142e-01 -6.62693456e-02 6.56346828e-02 4.08823133e-01
-1.24978624e-01 -2.95270281e-03 -7.41738186e-04 -3.78476888e-01
-6.77293122e-01 5.82026601e-01 1.16885567e+00 3.55602384e-01
-3.80644709e-01 1.12552509e-01 -8.68871152e-01 9.81938183e-01
1.12302649e+00 -1.32538274e-01 -2.90615499e-01 3.46842110e-01
6.28121734e-01 3.46426547e-01 -1.05033353e-01 -6.11970544e-01
2.88344204e-01 -4.91339892e-01 6.89750314e-02 -3.30642313e-01
5.54521620e-01 -8.48343253e-01 -3.17345321e-01 1.16514929e-01
-4.37748432e-01 6.71079978e-02 1.58011079e-01 3.03971946e-01
-6.51809216e-01 -3.80511314e-01 5.35328567e-01 -1.61560535e-01
-7.44851768e-01 -1.75158098e-01 -1.01414800e+00 8.10886919e-02
1.29008591e+00 -9.82660204e-02 -2.48914957e-01 -9.82564628e-01
-4.70874310e-01 4.45124924e-01 7.20729455e-02 6.43979967e-01
6.15858316e-01 -8.49793851e-01 -3.61073017e-01 1.18866734e-01
1.41426638e-01 -1.52685046e-01 4.12928820e-01 8.90563428e-01
-1.68405876e-01 4.03328121e-01 -1.87453493e-01 -1.80097803e-01
-1.74299061e+00 7.44579285e-02 -1.02088377e-01 -3.22908312e-01
-2.48837784e-01 7.49231219e-01 -3.73598099e-01 -6.15234733e-01
4.77504492e-01 1.00345939e-01 -8.22256863e-01 3.81672084e-01
7.44806290e-01 2.58371621e-01 -3.65676805e-02 -7.32819319e-01
-2.89396107e-01 1.02041744e-01 1.06665500e-01 -1.65903922e-02
1.45751512e+00 -4.97856736e-02 -3.59285384e-01 6.83917761e-01
9.76110101e-01 -7.51129836e-02 -2.32287765e-01 9.65084285e-02
4.12401885e-01 -7.31234625e-02 -2.90417001e-02 -8.63145947e-01
-6.24477506e-01 1.00167882e+00 6.58532232e-02 6.05576038e-01
1.16608143e+00 -2.18392432e-01 7.18713343e-01 5.14602959e-01
1.55082881e-01 -1.53759181e+00 1.46985203e-01 6.97115004e-01
7.56591499e-01 -1.48202050e+00 -1.61970198e-01 -6.00588739e-01
-1.15401876e+00 1.22346365e+00 6.07634962e-01 4.86995608e-01
6.81233168e-01 3.09840553e-02 4.44456607e-01 -3.97468179e-01
-1.01091659e+00 -1.54792294e-01 1.11052074e-01 4.15497035e-01
1.07929981e+00 1.11974113e-01 -7.50263631e-01 6.55355811e-01
-3.56175393e-01 6.39198571e-02 7.41766393e-01 1.08646011e+00
-4.28610742e-01 -1.39690435e+00 -3.58361781e-01 3.95473540e-01
-6.50050461e-01 2.83692211e-01 -9.50613081e-01 5.45630217e-01
-1.59544662e-01 1.44454324e+00 -2.98964679e-01 -1.17769718e-01
2.03505531e-01 6.19065940e-01 -1.13605283e-01 -6.13317370e-01
-9.49645877e-01 -2.51165330e-01 4.55122322e-01 -5.52244365e-01
-8.07490170e-01 -5.03279448e-01 -1.34911442e+00 -2.43976265e-01
-1.36039481e-01 7.98356712e-01 1.15641427e+00 1.13159275e+00
5.22569299e-01 2.56617844e-01 6.25785232e-01 -5.59040666e-01
6.98297918e-02 -8.46951425e-01 -3.77904862e-01 4.35139209e-01
-2.61851221e-01 -3.79824281e-01 -2.69081026e-01 -1.62242770e-01]
|
[12.953849792480469, 6.271485805511475]
|
54a5ef81-0548-4e42-9b41-fca52060e502
|
differentiable-patch-selection-for-image
|
2104.03059
| null |
https://arxiv.org/abs/2104.03059v1
|
https://arxiv.org/pdf/2104.03059v1.pdf
|
Differentiable Patch Selection for Image Recognition
|
Neural Networks require large amounts of memory and compute to process high resolution images, even when only a small part of the image is actually informative for the task at hand. We propose a method based on a differentiable Top-K operator to select the most relevant parts of the input to efficiently process high resolution images. Our method may be interfaced with any downstream neural network, is able to aggregate information from different patches in a flexible way, and allows the whole model to be trained end-to-end using backpropagation. We show results for traffic sign recognition, inter-patch relationship reasoning, and fine-grained recognition without using object/part bounding box annotations during training.
|
['Thomas Unterthiner', 'Jakob Uszkoreit', 'Dirk Weissenborn', 'Alexey Dosovitskiy', 'Aravindh Mahendran', 'Jean-Baptiste Cordonnier']
|
2021-04-07
| null |
http://openaccess.thecvf.com//content/CVPR2021/html/Cordonnier_Differentiable_Patch_Selection_for_Image_Recognition_CVPR_2021_paper.html
|
http://openaccess.thecvf.com//content/CVPR2021/papers/Cordonnier_Differentiable_Patch_Selection_for_Image_Recognition_CVPR_2021_paper.pdf
|
cvpr-2021-1
|
['traffic-sign-recognition']
|
['computer-vision']
|
[ 4.08953279e-01 2.62088507e-01 1.69355527e-01 -5.62351823e-01
-8.10197771e-01 -2.56988764e-01 4.88676637e-01 -1.30480289e-01
-5.73973835e-01 5.31160712e-01 3.03053148e-02 -9.60755348e-02
-3.35535020e-01 -1.04603243e+00 -9.82433796e-01 -5.71016550e-01
1.10150471e-01 8.56803775e-01 7.90237784e-01 -7.11055845e-02
2.08841383e-01 1.08976328e+00 -1.84991205e+00 9.87907648e-01
4.87687290e-01 1.27354407e+00 3.99376631e-01 8.38267386e-01
-1.54208571e-01 9.80188608e-01 -5.10448635e-01 -3.66066754e-01
4.10763144e-01 8.08778629e-02 -8.70514095e-01 3.06934476e-01
1.18853939e+00 -2.76412159e-01 -1.46500319e-01 8.47699761e-01
4.02006894e-01 2.12401658e-01 4.54695880e-01 -6.43631876e-01
-4.27258551e-01 3.73833895e-01 -2.61292428e-01 3.77150655e-01
-4.39707860e-02 3.56601954e-01 7.84291744e-01 -1.06670058e+00
8.38440299e-01 1.32209396e+00 5.79080164e-01 3.06926966e-01
-1.32485938e+00 -2.58399129e-01 4.46099907e-01 2.84662545e-01
-1.09712589e+00 -4.00235325e-01 5.43924749e-01 -3.19633812e-01
1.24315524e+00 1.71930984e-01 7.53009737e-01 5.07386267e-01
-1.12040564e-01 7.31562972e-01 1.04832363e+00 -1.54205978e-01
1.35526538e-01 -3.95531356e-02 4.27194268e-01 9.71893311e-01
7.65800700e-02 -1.22435078e-01 -5.01354694e-01 1.18143484e-02
9.97201025e-01 4.34682295e-02 -1.31205752e-01 -3.49410325e-01
-1.01878428e+00 3.79567206e-01 7.19237030e-01 7.78591931e-02
-7.45344222e-01 2.38703221e-01 4.32487540e-02 1.66104913e-01
1.66483819e-01 2.30213031e-01 -5.82302451e-01 1.78040534e-01
-9.59937811e-01 2.35832334e-01 7.77162671e-01 6.32430673e-01
9.12538648e-01 -1.63473904e-01 -2.34363511e-01 9.34323132e-01
4.05072293e-04 2.30418295e-01 6.44295737e-02 -1.23530865e+00
5.83035946e-01 6.81396306e-01 -1.00670662e-03 -5.87974429e-01
-4.27956194e-01 -3.64625245e-01 -7.40823507e-01 9.17121351e-01
6.95617199e-01 1.05505466e-01 -1.41251504e+00 1.08243763e+00
2.31866404e-01 -1.77058503e-01 2.37655286e-02 1.15427220e+00
8.03640604e-01 5.38279474e-01 1.28269032e-01 4.45337832e-01
1.48536968e+00 -1.11388314e+00 2.29610205e-02 -4.70716715e-01
1.82956737e-02 -4.48599517e-01 1.05810654e+00 3.69096309e-01
-1.50972795e+00 -6.67806268e-01 -7.68782020e-01 -4.47696775e-01
-7.76252806e-01 1.99484617e-01 3.63815904e-01 1.23550445e-01
-1.19997966e+00 7.52017856e-01 -6.69082522e-01 -1.55713141e-01
9.41361129e-01 5.65565467e-01 -4.47116047e-01 -2.12478727e-01
-6.79144442e-01 1.03660786e+00 7.24461198e-01 3.28341752e-01
-5.50517678e-01 -7.13199139e-01 -4.87698376e-01 3.52148890e-01
3.23442161e-01 -9.98323441e-01 1.01835060e+00 -1.05274487e+00
-1.34101248e+00 8.75954270e-01 -1.27203509e-01 -4.22318071e-01
6.88830256e-01 -1.06940329e-01 -8.17687716e-03 4.91744876e-01
-2.00236499e-01 1.04389477e+00 1.06687987e+00 -9.44493890e-01
-1.01391244e+00 -4.29634750e-01 1.31881639e-01 2.20370159e-01
2.84204423e-01 1.96606994e-01 -6.86977506e-01 -3.58991086e-01
2.00929359e-01 -6.56729698e-01 -2.96958238e-01 3.93990874e-01
-8.80721062e-02 -1.12971641e-01 9.90469337e-01 -7.55796552e-01
3.22446764e-01 -1.95954311e+00 -2.92770825e-02 5.84431589e-01
8.83577242e-02 3.98087174e-01 -1.63657069e-01 -8.21015462e-02
-1.28727376e-01 -8.76825228e-02 -4.30087715e-01 -1.01520181e-01
-6.73705265e-02 3.77979100e-01 -3.50176573e-01 1.70672759e-01
6.51678979e-01 1.02797902e+00 -4.91991043e-01 -4.84705240e-01
3.27385664e-01 6.88810170e-01 -3.96179348e-01 3.48215774e-02
-2.99212128e-01 1.32109508e-01 -3.42216551e-01 6.30515099e-01
5.21275997e-01 -5.68289340e-01 -3.01977634e-01 -3.95870984e-01
-1.25313457e-02 1.10643923e-01 -1.34573877e+00 1.20385098e+00
-4.04977918e-01 7.00885355e-01 2.00844556e-01 -8.23486328e-01
7.99382925e-01 -1.04497954e-01 7.42862076e-02 -8.08196664e-01
-1.80957422e-01 5.58263436e-02 -3.29232186e-01 -3.83142531e-01
4.01987642e-01 -5.68989143e-02 1.19702399e-01 3.91783029e-01
-4.45063114e-02 2.00904816e-01 3.87191266e-01 -1.61667511e-01
1.06765401e+00 2.05647275e-01 1.60803840e-01 1.53003022e-01
4.14247274e-01 2.50897706e-01 2.39837036e-01 8.60744357e-01
2.10727900e-02 7.21677661e-01 4.59387958e-01 -8.81584644e-01
-1.15718305e+00 -1.07179546e+00 -2.08408296e-01 1.43401277e+00
7.94702917e-02 2.06751376e-01 -5.06137729e-01 -6.98781908e-01
-8.21103528e-03 4.01420832e-01 -5.97678602e-01 3.44083071e-01
-1.10367823e+00 -5.00803053e-01 4.99019563e-01 9.59597170e-01
5.81764996e-01 -1.41001225e+00 -9.33991134e-01 7.89870396e-02
9.33163166e-02 -1.09705031e+00 -1.25471562e-01 4.26177740e-01
-1.06345046e+00 -1.07305074e+00 -6.59734607e-01 -7.33869851e-01
9.31307018e-01 -1.65063888e-02 1.10893786e+00 -9.29637030e-02
-2.73611695e-01 9.40760598e-02 7.03678578e-02 2.30963230e-02
-1.93415284e-01 -9.06207860e-02 -5.82722843e-01 -3.26089896e-02
1.72913909e-01 -5.47252178e-01 -4.22435135e-01 3.73383641e-01
-7.37748861e-01 1.84072658e-01 1.02082205e+00 7.84231126e-01
8.76868248e-01 8.81038886e-03 1.69987708e-01 -8.98878694e-01
4.22722876e-01 1.34319901e-01 -6.92035794e-01 3.96510124e-01
-1.80357814e-01 4.01513845e-01 7.42191255e-01 -2.77458817e-01
-1.13315403e+00 3.40213835e-01 -8.51770267e-02 -4.90553349e-01
-6.89152360e-01 8.98725167e-02 -2.27563933e-01 -9.24579278e-02
6.54604912e-01 1.88034981e-01 -1.74473166e-01 -6.97445095e-01
6.56821370e-01 3.16836536e-01 8.25882316e-01 -3.69413644e-01
5.35524487e-01 6.49198294e-01 1.62541475e-02 -7.09405005e-01
-7.58595347e-01 -2.01399803e-01 -1.01104760e+00 -1.12769246e-01
8.90718281e-01 -6.81121469e-01 -6.28106236e-01 3.65929306e-01
-1.15362060e+00 -5.08964121e-01 -5.69472253e-01 1.07716829e-01
-5.80846190e-01 -8.36387277e-02 -6.45139396e-01 -3.67181063e-01
-3.87383342e-01 -1.00919652e+00 1.27232492e+00 2.54273772e-01
1.66428447e-01 -5.33363760e-01 -3.95497561e-01 4.74695593e-01
3.68993163e-01 4.14538011e-02 9.55886364e-01 -5.31981766e-01
-1.35691559e+00 -4.25859213e-01 -8.35027039e-01 2.36384735e-01
-3.32758754e-01 -7.29313418e-02 -1.02535987e+00 1.77799389e-01
-5.35946250e-01 -3.73685181e-01 1.40303147e+00 4.50415254e-01
1.35236812e+00 -4.65687335e-01 -2.17631951e-01 6.80076718e-01
1.25685346e+00 -2.58269578e-01 7.12002456e-01 3.23861688e-01
8.08478832e-01 7.09726453e-01 3.40817720e-01 9.01260972e-02
2.99717903e-01 5.47740459e-01 4.08156574e-01 -1.27599761e-01
-3.27427566e-01 -3.69659849e-02 6.48229346e-02 -1.02210708e-01
-3.10455620e-01 2.55469888e-01 -8.32937002e-01 3.33285540e-01
-1.88169253e+00 -1.28778636e+00 4.24389988e-02 1.91884255e+00
5.91418564e-01 2.80357808e-01 2.62433290e-01 -9.78367180e-02
6.54264271e-01 1.38063252e-01 -5.40013254e-01 -3.47346485e-01
-1.89552009e-01 3.48445326e-01 5.99445403e-01 5.53955019e-01
-1.07996511e+00 1.06961012e+00 7.04347897e+00 6.77010238e-01
-1.21466112e+00 -2.22758904e-01 6.06261730e-01 -2.09053919e-01
1.03900805e-01 -1.39163405e-01 -9.97884989e-01 4.80923504e-02
6.68125391e-01 5.87592423e-01 2.98452377e-01 9.52184141e-01
3.85992825e-02 -2.17672467e-01 -1.11487842e+00 7.89510608e-01
-9.20676067e-02 -1.62907028e+00 1.56742543e-01 -7.67469928e-02
4.48322237e-01 2.45129451e-01 -1.29751176e-01 1.03412539e-01
3.14016521e-01 -1.20203316e+00 7.00407922e-01 9.34595346e-01
4.92257357e-01 -6.15562499e-01 3.86679798e-01 5.45744598e-01
-1.07811427e+00 -3.04039389e-01 -5.74565172e-01 1.29379272e-01
2.21549422e-01 4.84861821e-01 -8.28007519e-01 2.11632606e-02
8.22147191e-01 2.66068965e-01 -7.57371902e-01 1.04996777e+00
-1.72433496e-01 2.28443757e-01 -6.35331273e-01 1.52110741e-01
3.03435236e-01 -2.67411321e-02 4.79760855e-01 1.35260296e+00
-1.11993654e-02 1.58089325e-02 1.69039354e-01 9.26522315e-01
-1.59392394e-02 -2.00303942e-01 -3.26095551e-01 4.04420674e-01
1.33676156e-01 1.27424312e+00 -8.39259386e-01 -6.58356607e-01
-3.37852001e-01 1.03965044e+00 7.68956006e-01 5.82387865e-01
-5.60585856e-01 -3.99321258e-01 4.75793332e-01 4.01855230e-01
1.03251624e+00 -9.25595164e-02 -5.17868340e-01 -8.77809525e-01
4.72202361e-01 -5.94277799e-01 6.05415046e-01 -1.21578979e+00
-1.27981186e+00 6.13244236e-01 -1.34174645e-01 -9.67683434e-01
-3.70000839e-01 -9.59637582e-01 -5.19484043e-01 1.12068379e+00
-1.68476331e+00 -1.25832021e+00 -4.49730605e-01 7.00539112e-01
3.06283206e-01 -3.23245227e-02 6.62737370e-01 1.46365583e-01
-2.70930618e-01 1.18351571e-01 -3.06595474e-01 3.92691612e-01
3.69621992e-01 -1.28746819e+00 4.01345164e-01 8.38541508e-01
8.95749554e-02 3.59710604e-01 3.80221307e-01 -4.81489956e-01
-9.77455378e-01 -1.20880532e+00 7.83722758e-01 -3.71515483e-01
3.98174852e-01 -4.73484099e-02 -1.10595500e+00 5.71758628e-01
-4.08652425e-02 6.08521223e-01 2.50135869e-01 1.05815694e-01
-6.58446491e-01 -4.47719514e-01 -1.11726260e+00 5.35653710e-01
1.08785462e+00 -5.41953743e-01 -8.09893608e-01 1.68078974e-01
2.72845328e-01 -5.73248029e-01 -8.00340712e-01 3.48334074e-01
5.74465215e-01 -9.46950912e-01 1.28550005e+00 -7.79404461e-01
3.17275226e-01 -5.17634451e-01 -7.91606754e-02 -9.01838660e-01
-4.86142546e-01 -7.89921358e-02 -3.34449321e-01 6.58244431e-01
6.75648987e-01 -4.77756381e-01 9.44279730e-01 9.69168484e-01
-1.15086146e-01 -6.98414624e-01 -7.74476230e-01 -3.58896524e-01
-2.30436489e-01 -5.07892668e-01 4.14384395e-01 3.39871913e-01
-4.42145109e-01 2.11485371e-01 -1.37446344e-01 5.27234077e-01
7.59476840e-01 6.07994318e-01 7.11366415e-01 -1.27951169e+00
-5.92141628e-01 -8.68359566e-01 -4.84770954e-01 -1.10599434e+00
3.95061783e-02 -8.43856514e-01 9.68615040e-02 -1.88932979e+00
-4.17262018e-02 -3.60053420e-01 -1.61743194e-01 8.07454705e-01
-5.71230948e-02 5.15461445e-01 3.12452912e-01 3.21313411e-01
-6.03336692e-01 -2.23997459e-02 1.33843565e+00 -1.19145915e-01
-2.46647924e-01 1.37035832e-01 -5.25759280e-01 9.68932450e-01
6.53620183e-01 -2.28671029e-01 -2.30195709e-02 -3.40174526e-01
-2.92670578e-02 5.52895144e-02 9.08865452e-01 -1.11437440e+00
4.53350097e-01 -8.80528316e-02 1.09696305e+00 -7.61992872e-01
5.28192878e-01 -9.63768899e-01 1.96409337e-02 2.25839987e-02
-7.10143507e-01 -2.08750233e-01 1.88693807e-01 5.01870036e-01
-2.14127123e-01 -1.66672990e-01 1.03377092e+00 -3.70547354e-01
-1.12688553e+00 4.37930822e-01 -3.35527301e-01 -3.94485481e-02
7.23529994e-01 -5.69498360e-01 -5.11584103e-01 -3.51828746e-02
-1.15579271e+00 7.25886375e-02 3.94507855e-01 2.31085062e-01
6.73497975e-01 -1.08366847e+00 -6.16829813e-01 3.02697808e-01
-9.65575222e-03 2.21047521e-01 3.42431843e-01 6.53057098e-01
-5.68873942e-01 2.58054018e-01 -4.78346556e-01 -6.90004885e-01
-1.42111838e+00 3.55390728e-01 6.66951358e-01 -3.83123636e-01
-1.24567032e+00 6.64140284e-01 8.70219320e-02 -1.52507648e-01
3.70795637e-01 -5.59144735e-01 -1.51280180e-01 -2.43202299e-02
7.79316843e-01 2.13348791e-01 1.64541781e-01 -5.96715212e-01
-3.17783177e-01 8.31054151e-01 -2.97248475e-02 -1.13093682e-01
1.39703774e+00 1.86495617e-01 -1.19988710e-01 1.39464691e-01
1.00403881e+00 -2.55957961e-01 -1.69077432e+00 -3.28149915e-01
-3.27956714e-02 -4.75744933e-01 1.05453975e-01 -1.09079361e+00
-1.05853546e+00 8.53835523e-01 4.70830679e-01 -3.59984189e-02
1.16755164e+00 1.78614900e-01 5.18839896e-01 1.04139149e+00
1.15555823e-01 -1.33981311e+00 -5.89938499e-02 5.40133357e-01
9.31711853e-01 -1.02660882e+00 -6.61495700e-02 -1.76234946e-01
-7.50354290e-01 1.34140444e+00 5.95724761e-01 -4.11522806e-01
6.71507657e-01 5.80098629e-01 1.13121830e-01 -2.13258773e-01
-8.80321145e-01 -6.15584195e-01 5.62450469e-01 6.28041863e-01
2.30951514e-02 -1.48508340e-01 3.07286352e-01 3.47654819e-02
5.41702509e-02 7.29196146e-02 1.55221149e-01 8.67190182e-01
-9.53151464e-01 -9.01488006e-01 -4.33150530e-01 7.34140396e-01
-2.29498476e-01 -1.34297386e-02 -3.45357239e-01 6.44825220e-01
2.64882058e-01 1.85797691e-01 3.72942299e-01 1.45363197e-01
5.79216003e-01 4.22522217e-01 5.74114859e-01 -4.08035070e-01
-6.39714062e-01 -4.46555205e-03 3.45006496e-01 -9.11513567e-01
-3.56301755e-01 -6.06536984e-01 -1.49144077e+00 -9.75312740e-02
1.50776982e-01 -4.28823233e-01 5.23038030e-01 1.10385001e+00
5.24398148e-01 5.24848163e-01 -5.90039715e-02 -1.12197208e+00
-1.86055571e-01 -7.61745930e-01 -2.39203036e-01 2.48408318e-01
4.40511942e-01 -3.16845775e-01 1.29440501e-01 2.20888332e-01]
|
[9.498353958129883, 0.7429116368293762]
|
e372a9c6-a72f-4b92-9932-dd6c5e7e0355
|
autorl-hyperparameter-landscapes
|
2304.02396
| null |
https://arxiv.org/abs/2304.02396v4
|
https://arxiv.org/pdf/2304.02396v4.pdf
|
AutoRL Hyperparameter Landscapes
|
Although Reinforcement Learning (RL) has shown to be capable of producing impressive results, its use is limited by the impact of its hyperparameters on performance. This often makes it difficult to achieve good results in practice. Automated RL (AutoRL) addresses this difficulty, yet little is known about the dynamics of the hyperparameter landscapes that hyperparameter optimization (HPO) methods traverse in search of optimal configurations. In view of existing AutoRL approaches dynamically adjusting hyperparameter configurations, we propose an approach to build and analyze these hyperparameter landscapes not just for one point in time but at multiple points in time throughout training. Addressing an important open question on the legitimacy of such dynamic AutoRL approaches, we provide thorough empirical evidence that the hyperparameter landscapes strongly vary over time across representative algorithms from RL literature (DQN, PPO, and SAC) in different kinds of environments (Cartpole, Bipedal Walker, and Hopper) This supports the theory that hyperparameters should be dynamically adjusted during training and shows the potential for more insights on AutoRL problems that can be gained through landscape analyses. Our code can be found at https://github.com/automl/AutoRL-Landscape
|
['Marius Lindauer', 'Alexander Dockhorn', 'Konrad Wienecke', 'Carolin Benjamins', 'Aditya Mohan']
|
2023-04-05
| null | null | null | null |
['automl', 'hyperparameter-optimization', 'open-question']
|
['methodology', 'methodology', 'natural-language-processing']
|
[-9.85385403e-02 -1.65061355e-01 -3.00307363e-01 4.67934785e-03
-6.99669003e-01 -9.70779538e-01 4.70469564e-01 2.52366245e-01
-4.79013622e-01 9.67298210e-01 -7.20274262e-03 -3.47091526e-01
-7.04941094e-01 -7.13194966e-01 -5.92213273e-01 -8.97969186e-01
-2.39344358e-01 4.76704240e-01 1.05287768e-02 -5.73622823e-01
5.26761472e-01 5.65500677e-01 -1.59708166e+00 -2.19785213e-01
6.35128558e-01 2.39740357e-01 2.32856721e-02 7.51400411e-01
1.29449204e-01 2.55517662e-01 -7.80546188e-01 -3.38072069e-02
3.21916193e-01 -5.80212593e-01 -6.45174086e-01 -1.38516650e-01
-7.07345977e-02 -5.38156778e-02 9.38037857e-02 8.58543456e-01
7.91425586e-01 2.79010028e-01 5.51583767e-01 -1.25245619e+00
-2.91870713e-01 6.75675154e-01 -5.48501492e-01 2.59610057e-01
3.28195155e-01 5.97170591e-01 9.50322151e-01 -1.90633729e-01
6.23284101e-01 1.22206283e+00 6.79158568e-01 2.56627053e-01
-1.56166756e+00 -5.03310502e-01 1.04559258e-01 9.40452889e-02
-1.41981995e+00 -2.05011576e-01 5.98660231e-01 -3.13703358e-01
7.41899967e-01 1.13672828e-02 1.00143814e+00 1.14881694e+00
3.18845034e-01 5.43455243e-01 1.18607521e+00 -6.43435776e-01
5.34613013e-01 -4.34302129e-02 -2.73318700e-02 5.76399386e-01
5.79932034e-01 3.62241805e-01 -4.90110785e-01 -1.62149340e-01
8.13933074e-01 -6.03051543e-01 -1.37067840e-01 -7.70152330e-01
-1.05936551e+00 9.52691317e-01 1.85667649e-01 4.44014430e-01
-1.19864076e-01 4.86824661e-01 3.76356751e-01 4.84041542e-01
2.45257351e-03 1.05764127e+00 -5.18407404e-01 -6.43612683e-01
-3.60519201e-01 6.53591692e-01 7.31429815e-01 4.71456140e-01
7.39864945e-01 -2.48418804e-02 1.58266678e-01 9.98516142e-01
9.10012498e-02 3.60129952e-01 4.81937587e-01 -1.10524333e+00
4.46772128e-01 5.78711629e-01 4.07190979e-01 -8.74043226e-01
-6.68364465e-01 -3.75226855e-01 -1.79124221e-01 5.01414657e-01
6.90745115e-01 -3.29334885e-01 -3.66735518e-01 1.72231042e+00
5.04299879e-01 -3.67987841e-01 8.28450471e-02 7.81691551e-01
1.24748632e-01 5.07278681e-01 1.16497204e-01 -9.91973728e-02
1.02479970e+00 -5.82022071e-01 -4.37940925e-01 -1.08699784e-01
7.31222510e-01 -4.87851471e-01 1.54827809e+00 4.34717983e-01
-9.80574787e-01 -2.07515880e-01 -1.21353745e+00 4.34312910e-01
-4.05968815e-01 -2.22931027e-01 8.03772986e-01 9.05553341e-01
-8.73403430e-01 9.15140808e-01 -1.00505626e+00 -4.79796350e-01
1.14328817e-01 5.10679483e-01 -2.44680084e-02 1.31177396e-01
-1.01937342e+00 1.02482331e+00 5.13106227e-01 1.78000569e-01
-6.38482213e-01 -5.86543560e-01 -5.22135615e-01 -5.88766113e-02
5.66316903e-01 -4.06968176e-01 1.43334174e+00 -9.35920894e-01
-1.84991121e+00 2.20633537e-01 3.39378268e-01 -1.41329706e-01
5.68312943e-01 -2.64664263e-01 -9.05117914e-02 -9.89740416e-02
-2.17682853e-01 4.94715661e-01 6.05510473e-01 -1.31517172e+00
-3.18753272e-01 -3.58547300e-01 1.26233429e-01 5.89145422e-01
-6.39247447e-02 -3.74847442e-01 -7.48298969e-03 -4.04606760e-01
-2.05188543e-01 -1.16984165e+00 -2.10749418e-01 -3.93105298e-01
-1.22968458e-01 -2.04009250e-01 4.20544028e-01 -2.07638927e-03
1.36504519e+00 -1.89923680e+00 2.41411790e-01 3.14329654e-01
-2.89625853e-01 1.17205016e-01 -2.21209303e-01 9.07516837e-01
2.33433962e-01 2.93884099e-01 -1.62731290e-01 4.38433021e-01
2.08316326e-01 3.18000466e-01 -1.57694146e-02 4.93330359e-01
-1.33268863e-01 7.92258143e-01 -1.12801647e+00 -3.36898148e-01
2.82939404e-01 4.10756141e-01 -6.30213678e-01 -7.52285346e-02
-3.90914083e-01 3.23109716e-01 -5.89700758e-01 6.79278314e-01
1.48038954e-01 -2.25486383e-01 4.92653430e-01 1.20814465e-01
-3.36444944e-01 -9.26785246e-02 -1.24205184e+00 1.18999434e+00
-4.00591612e-01 6.47488773e-01 -2.72125810e-01 -8.07909012e-01
7.04737425e-01 8.16055313e-02 6.80977702e-01 -8.65177989e-01
2.34961703e-01 1.52980372e-01 1.91361919e-01 -5.13208985e-01
4.79083180e-01 -7.53943473e-02 -7.19580427e-02 7.33731091e-01
-1.07121378e-01 -3.34370196e-01 5.06668448e-01 -4.55634624e-01
1.15880597e+00 3.83470625e-01 3.94333422e-01 -2.92798817e-01
2.86787748e-01 4.16907758e-01 2.72811830e-01 7.65676260e-01
-2.08878204e-01 1.91082567e-01 6.74066722e-01 -3.85962933e-01
-1.15137577e+00 -1.02655804e+00 -3.96885037e-01 1.21751547e+00
2.37024859e-01 -3.19388062e-01 -8.36620629e-01 -2.62363851e-01
2.24206477e-01 8.46496344e-01 -8.14931333e-01 -3.07569593e-01
-7.56287098e-01 -9.77086365e-01 5.56929410e-01 2.55240798e-01
4.14648712e-01 -1.23014688e+00 -1.21394229e+00 2.95476615e-01
-2.11740639e-02 -6.58023953e-01 -3.62334996e-02 2.53791928e-01
-9.25117433e-01 -1.24056268e+00 -4.71859187e-01 -2.26238683e-01
5.07576048e-01 -3.49729508e-02 1.12639594e+00 1.46645695e-01
-3.99610966e-01 7.33004987e-01 -4.80307400e-01 -3.80080730e-01
-5.34875929e-01 3.60232383e-01 -3.14088096e-03 -5.46624422e-01
4.19583201e-04 -5.31419992e-01 -7.39342868e-01 7.46004462e-01
-8.35238218e-01 -3.45935404e-01 5.43632507e-01 6.87426627e-01
6.31757259e-01 2.02327281e-01 7.40155160e-01 -7.02068508e-01
1.09529567e+00 -5.78323722e-01 -9.24997151e-01 3.60914528e-01
-9.71909702e-01 4.33018953e-01 4.25990909e-01 -5.90240121e-01
-7.52976239e-01 -1.22011080e-01 2.63634045e-02 -1.02578357e-01
-1.01469569e-01 2.31385216e-01 3.11635196e-01 -2.14289144e-01
9.23362672e-01 -1.03165440e-01 1.47608221e-01 -2.67895907e-01
2.53651023e-01 1.63042337e-01 6.48768917e-02 -8.64435792e-01
7.39593327e-01 8.06543827e-02 6.31996840e-02 -8.29669535e-01
-6.69216216e-01 -7.66437501e-03 -4.07824308e-01 -5.40488720e-01
6.06714725e-01 -3.57167482e-01 -9.70066905e-01 8.20461139e-02
-2.89974123e-01 -1.04946923e+00 -4.75384355e-01 2.19790682e-01
-9.55552280e-01 1.78960140e-03 -7.21725598e-02 -8.32338512e-01
-7.51571432e-02 -1.08690929e+00 7.55733252e-01 4.76905197e-01
-5.38138032e-01 -1.19320655e+00 4.85570103e-01 2.22764820e-01
5.09000242e-01 4.42071587e-01 1.15909922e+00 -3.45047146e-01
-2.92101800e-01 6.51961342e-02 4.42946434e-01 -1.69227272e-01
3.09623946e-02 2.77210385e-01 -4.18295175e-01 -5.05946875e-01
-3.70185941e-01 -5.60272098e-01 2.35625371e-01 5.14561176e-01
9.35001314e-01 -2.29478404e-01 -1.09528974e-01 3.87549758e-01
1.71651161e+00 2.33429849e-01 4.88594055e-01 1.01122808e+00
1.21192746e-01 3.79938394e-01 7.70605147e-01 4.85740602e-01
1.62042692e-01 7.54925191e-01 3.63664418e-01 2.97867328e-01
1.84381664e-01 -3.98213536e-01 3.50834250e-01 2.17038244e-01
-1.03715517e-01 -3.41601044e-01 -1.09517539e+00 2.91402340e-01
-1.82743847e+00 -8.39336812e-01 2.98429817e-01 2.24213767e+00
7.21903801e-01 7.37479627e-02 5.50721645e-01 2.50614733e-01
5.04327118e-01 7.38207623e-02 -8.50710809e-01 -7.41495192e-01
4.95137088e-02 -2.27652001e-03 5.97616553e-01 4.42332178e-01
-6.09074593e-01 8.83423984e-01 7.30155706e+00 5.70873320e-01
-1.05814755e+00 -3.36746424e-01 4.86657530e-01 -3.08686525e-01
-1.90658763e-01 6.84985667e-02 -6.05144382e-01 3.49467576e-01
9.02775466e-01 -2.98959851e-01 9.43398833e-01 7.71821380e-01
4.70142245e-01 -5.56664348e-01 -8.88709724e-01 6.65460706e-01
-3.73495758e-01 -1.09934783e+00 -2.71038979e-01 1.06524974e-01
7.11379409e-01 -1.57027721e-01 1.25843629e-01 3.83722901e-01
6.62031293e-01 -1.05622292e+00 4.85857487e-01 3.16302776e-01
4.64424372e-01 -1.03836095e+00 5.39474249e-01 2.26487175e-01
-9.33286905e-01 -5.72781920e-01 -3.34889889e-01 1.49616212e-01
-1.15227066e-01 1.20689206e-01 -8.93729150e-01 1.86162129e-01
6.18258715e-01 1.50957137e-01 -7.05374599e-01 1.16645753e+00
-1.87968135e-01 7.67850459e-01 -3.50308955e-01 -4.38326031e-01
2.72902817e-01 -3.62959802e-01 5.33161998e-01 7.64868975e-01
2.20799148e-01 2.08949372e-02 -2.10519031e-01 6.90068722e-01
4.20400828e-01 2.55736202e-01 -6.25632405e-01 -2.83310324e-01
6.21492982e-01 9.98619497e-01 -9.61553097e-01 3.23522449e-01
1.38212353e-01 3.84758741e-01 4.25759524e-01 5.12355447e-01
-7.91481256e-01 -3.13049793e-01 5.70319831e-01 1.62931517e-01
2.00884759e-01 -5.27509451e-01 -2.37320542e-01 -5.74357092e-01
-2.06895500e-01 -9.92117524e-01 4.93494511e-01 -6.07622385e-01
-8.77513766e-01 1.43675014e-01 4.73718941e-01 -1.00551856e+00
-3.63244206e-01 -3.53280753e-01 -5.68476498e-01 1.74618766e-01
-9.81159925e-01 -4.55663711e-01 -2.65612841e-01 3.35562199e-01
5.65745056e-01 -3.66951376e-02 5.56494892e-01 -2.11110458e-01
-7.38812506e-01 5.22967398e-01 4.61048126e-01 -4.21479762e-01
5.18127203e-01 -1.25968051e+00 1.16568081e-01 1.56243846e-01
-4.43043709e-02 4.55661744e-01 1.01081288e+00 -3.94313425e-01
-1.72809327e+00 -4.91145521e-01 -4.35843840e-02 -4.66970086e-01
7.60472298e-01 1.85251962e-02 -7.13946521e-01 2.99059570e-01
2.38588862e-02 -4.29233104e-01 5.82555830e-01 3.28484774e-01
1.64387852e-01 -1.45257590e-02 -1.04977000e+00 9.20329332e-01
9.53751862e-01 1.63998187e-01 -9.52239856e-02 2.00705335e-01
3.05320919e-01 -3.87482703e-01 -9.93383944e-01 3.32794875e-01
6.87078595e-01 -1.02135885e+00 9.18531060e-01 -4.75956112e-01
1.77113011e-01 -1.87808916e-01 1.01580638e-02 -1.56722629e+00
-1.01944268e-01 -8.00825119e-01 1.00151869e-02 9.34056222e-01
4.17404830e-01 -8.39937568e-01 6.81778073e-01 4.68084246e-01
1.71876162e-01 -1.21813929e+00 -8.05777907e-01 -8.76018643e-01
3.97383451e-01 -2.19060093e-01 7.02246904e-01 6.41122341e-01
1.00922495e-01 2.05632165e-01 1.79486558e-01 -6.40338287e-02
5.74490488e-01 9.65230465e-02 9.15263891e-01 -1.00393987e+00
-3.21290374e-01 -7.54538417e-01 -1.26568004e-01 -4.29348141e-01
1.01602934e-01 -5.80511749e-01 -4.72130865e-04 -1.39967763e+00
-9.92720276e-02 -8.25999260e-01 -4.75531071e-02 4.02580291e-01
-2.33840421e-02 -1.81529209e-01 1.29069179e-01 2.25249052e-01
-5.61540008e-01 5.58789432e-01 1.38918293e+00 3.85517329e-01
-7.02677548e-01 -3.19122225e-02 -5.66288352e-01 6.50286794e-01
1.25342476e+00 -4.93370295e-01 -6.96862042e-01 -2.55377203e-01
6.47227347e-01 1.52990580e-01 1.26618758e-01 -1.06340373e+00
-1.13978207e-01 -6.42862260e-01 3.66586924e-01 -1.35337412e-01
6.31644158e-03 -6.28538728e-01 1.50092736e-01 4.04149532e-01
-4.47009385e-01 6.19700432e-01 3.76744300e-01 5.87152064e-01
3.24050009e-01 -3.66705835e-01 8.87942374e-01 -6.76136240e-02
-4.17753577e-01 -8.38561207e-02 -6.63755536e-01 3.49605858e-01
1.10728848e+00 -4.46378082e-01 -3.54979038e-01 -3.14624995e-01
-5.56209326e-01 4.35883701e-01 7.46663392e-01 3.39267880e-01
1.08898103e-01 -1.01078379e+00 -1.97286487e-01 -4.59036715e-02
-1.24733016e-01 -1.75569430e-01 6.80885911e-02 6.82611704e-01
-7.73331761e-01 1.74271122e-01 -3.37177813e-01 -5.33684075e-01
-9.61985290e-01 2.91405797e-01 6.76285684e-01 -2.84805477e-01
-7.75438905e-01 3.35748196e-01 -3.77094865e-01 -3.58816773e-01
2.62507468e-01 -1.13721304e-02 -1.19944960e-01 1.20949447e-02
-5.46384305e-02 6.05603576e-01 -1.74031496e-01 -1.43286601e-01
-8.58242214e-02 8.22344542e-01 1.77373588e-01 -3.83278072e-01
1.42053556e+00 -5.95454238e-02 2.15663269e-01 5.17236233e-01
7.14595973e-01 -7.63706043e-02 -1.46329570e+00 3.81052196e-01
1.33384749e-01 -3.75609726e-01 -1.54046997e-01 -8.45305264e-01
-6.73358977e-01 4.33519959e-01 7.20513940e-01 2.24575490e-01
1.00951850e+00 -2.55475163e-01 3.38977724e-01 7.00859666e-01
4.30651546e-01 -1.55283320e+00 4.51082021e-01 3.46312076e-01
9.37079608e-01 -9.72460926e-01 3.07558894e-01 1.97908789e-01
-7.97873497e-01 1.15727723e+00 6.37214959e-01 -1.19689532e-01
4.02621269e-01 3.06566924e-01 7.38201365e-02 -1.84649184e-01
-8.66393387e-01 -9.79807451e-02 -2.85015017e-01 4.93826866e-01
3.36816937e-01 3.23221721e-02 -3.23844939e-01 -4.20086905e-02
-6.00648999e-01 -2.43295863e-01 5.64833701e-01 1.10373712e+00
-6.24140203e-01 -1.28554606e+00 -4.33563203e-01 2.60638714e-01
-2.13413864e-01 4.51899171e-01 -3.59267771e-01 1.30600178e+00
-2.89106965e-01 8.01348746e-01 -1.03510268e-01 -1.60177246e-01
4.21067923e-01 1.72276169e-01 7.43134797e-01 -2.36020833e-01
-7.62096584e-01 -1.54042184e-01 1.55307010e-01 -5.72349548e-01
-3.19487564e-02 -8.60113800e-01 -1.43181157e+00 -2.92805761e-01
-2.74227083e-01 2.65577257e-01 8.04316759e-01 7.86861837e-01
1.98831707e-01 4.31107640e-01 5.19491732e-01 -7.97734797e-01
-7.16834307e-01 -5.57097375e-01 -4.02149051e-01 1.57883778e-01
1.98735699e-01 -1.09139049e+00 -5.77526689e-01 -4.54233825e-01]
|
[4.349639415740967, 1.9754964113235474]
|
9d47d90d-134b-41c3-b573-604be6c0dd0d
|
confidence-guided-adaptive-gate-and-dual
|
2105.06714
| null |
https://arxiv.org/abs/2105.06714v1
|
https://arxiv.org/pdf/2105.06714v1.pdf
|
Confidence-guided Adaptive Gate and Dual Differential Enhancement for Video Salient Object Detection
|
Video salient object detection (VSOD) aims to locate and segment the most attractive object by exploiting both spatial cues and temporal cues hidden in video sequences. However, spatial and temporal cues are often unreliable in real-world scenarios, such as low-contrast foreground, fast motion, and multiple moving objects. To address these problems, we propose a new framework to adaptively capture available information from spatial and temporal cues, which contains Confidence-guided Adaptive Gate (CAG) modules and Dual Differential Enhancement (DDE) modules. For both RGB features and optical flow features, CAG estimates confidence scores supervised by the IoU between predictions and the ground truths to re-calibrate the information with a gate mechanism. DDE captures the differential feature representation to enrich the spatial and temporal information and generate the fused features. Experimental results on four widely used datasets demonstrate the effectiveness of the proposed method against thirteen state-of-the-art methods.
|
['Huajun Zhou', 'Guangcong Wang', 'JianHuang Lai', 'Peijia Chen']
|
2021-05-14
| null | null | null | null |
['video-salient-object-detection']
|
['computer-vision']
|
[ 1.73028946e-01 -4.61107969e-01 -2.55214721e-01 -2.36119345e-01
-4.61310506e-01 -3.01167428e-01 4.05285090e-01 1.03020087e-01
-4.97508138e-01 6.55964494e-01 8.88439938e-02 2.26808205e-01
-1.75817627e-02 -5.53062797e-01 -5.36460578e-01 -8.47504914e-01
-2.10197821e-01 -2.63430625e-01 1.05295038e+00 7.83682540e-02
4.71549690e-01 3.79685283e-01 -1.66254175e+00 2.11544424e-01
9.35654700e-01 1.42238200e+00 5.76550066e-01 4.67319459e-01
7.68694431e-02 1.01950645e+00 1.96413770e-02 2.46455558e-02
4.26527947e-01 -3.33439559e-01 -2.82542765e-01 2.59677052e-01
1.93263844e-01 -5.77839971e-01 -3.84553820e-01 1.08902311e+00
1.89886898e-01 5.57554722e-01 2.23192811e-01 -1.39135611e+00
-3.21298540e-01 2.16496661e-01 -7.98723936e-01 8.53196025e-01
3.89158279e-01 4.92439806e-01 7.17807055e-01 -1.11768615e+00
6.36050045e-01 1.18934035e+00 2.43462205e-01 2.21669927e-01
-8.43792975e-01 -5.62256455e-01 6.77737832e-01 5.85982323e-01
-1.14399576e+00 -4.01730597e-01 1.15498567e+00 -5.36141872e-01
2.28994712e-01 4.77054641e-02 8.15654635e-01 6.40395045e-01
1.30407766e-01 1.06957662e+00 9.20132518e-01 -7.84143284e-02
1.30236253e-01 1.96641251e-01 -2.40965560e-02 8.94590259e-01
1.70436800e-01 5.18720984e-01 -7.80567288e-01 -2.86500584e-02
9.00218129e-01 2.13934362e-01 -6.53005123e-01 -5.55523276e-01
-1.45485532e+00 4.94290292e-01 6.25279844e-01 1.58636034e-01
-5.98105192e-01 -9.47579890e-02 1.32321537e-01 -2.56983638e-01
1.19896427e-01 9.52727422e-02 -2.13548616e-01 1.49885379e-02
-1.01470649e+00 1.28136545e-01 1.54416904e-01 9.29224670e-01
9.52361345e-01 2.59471595e-01 -4.97730374e-01 3.63218248e-01
4.56227571e-01 3.06434840e-01 4.84866887e-01 -1.05376780e+00
4.27493602e-01 5.93851089e-01 4.40809309e-01 -1.28079617e+00
-1.74908921e-01 -4.41150337e-01 -6.08422458e-01 2.28829131e-01
3.92302245e-01 1.17072098e-01 -7.86778092e-01 1.51648426e+00
7.41456151e-01 8.34416032e-01 -2.71145284e-01 1.50715041e+00
7.33546615e-01 7.47095525e-01 2.48640075e-01 -5.89350045e-01
8.72559071e-01 -1.00569701e+00 -6.24551833e-01 -3.21305901e-01
1.90989628e-01 -6.91689253e-01 7.08600283e-01 1.65224046e-01
-1.05213583e+00 -9.11987543e-01 -1.02544630e+00 2.64601409e-01
1.25553995e-01 1.89886853e-01 5.14616251e-01 2.15005919e-01
-5.62399566e-01 4.56308216e-01 -8.42168033e-01 6.18551299e-02
3.75473320e-01 1.06961772e-01 -2.32938498e-01 -2.12586552e-01
-1.06712735e+00 4.16303664e-01 6.26029909e-01 4.31184739e-01
-1.07096231e+00 -5.77143431e-01 -8.69684100e-01 -2.18804237e-02
6.06060505e-01 -3.69914323e-01 7.66550481e-01 -1.26783419e+00
-1.33829141e+00 2.29176477e-01 -2.40509033e-01 -3.12956810e-01
5.60690045e-01 -1.38551429e-01 -2.61900961e-01 4.93976474e-01
1.83304504e-01 6.85721934e-01 1.03132319e+00 -1.24073255e+00
-1.16627538e+00 -1.97308332e-01 -8.59806612e-02 3.62626761e-01
-1.08138524e-01 -3.17288972e-02 -5.45607626e-01 -6.81905985e-01
3.16823900e-01 -6.85363114e-01 -2.26472303e-01 5.39118052e-01
-6.80352226e-02 9.35481265e-02 1.09050584e+00 -7.09982455e-01
1.11246920e+00 -2.23045588e+00 9.90022868e-02 6.30541444e-02
6.72847927e-02 3.45435649e-01 1.24893396e-03 -3.04331303e-01
2.06048042e-01 -4.08432543e-01 -1.51903823e-01 2.96686571e-02
-4.40813482e-01 7.35850558e-02 -1.95736095e-01 5.44276059e-01
4.20029551e-01 6.39378607e-01 -1.30667841e+00 -9.42754626e-01
7.53748655e-01 4.10893828e-01 -3.63582045e-01 3.63370508e-01
-1.57587007e-01 7.87005842e-01 -8.53892028e-01 8.79447997e-01
7.69354880e-01 -1.66868776e-01 -1.97270676e-01 -2.62400627e-01
-3.83478075e-01 -1.21584371e-01 -1.70220387e+00 1.41309977e+00
1.63118988e-01 4.84777838e-01 1.85912605e-02 -8.07017744e-01
7.91414976e-01 -1.11111943e-02 5.54023564e-01 -7.26411402e-01
2.91404635e-01 1.35382682e-01 -1.93927273e-01 -6.24777615e-01
5.56977749e-01 1.98873907e-01 4.54944819e-01 -1.25432253e-01
5.27855800e-03 4.74805593e-01 1.17133625e-01 1.21309407e-01
8.31519008e-01 4.08615619e-01 2.50029773e-01 -2.51442306e-02
9.55087781e-01 -1.48510352e-01 1.21165621e+00 6.28349364e-01
-8.16754401e-01 6.65015936e-01 1.19234107e-01 -4.34736013e-01
-6.38232589e-01 -1.00782943e+00 5.18840775e-02 9.33986068e-01
9.78835404e-01 -7.41875321e-02 -3.82700771e-01 -6.08610749e-01
-8.90683234e-02 4.25611079e-01 -7.01914787e-01 -1.14608973e-01
-5.00459671e-01 -4.81464833e-01 -1.74938947e-01 6.92255437e-01
7.76237547e-01 -9.68524098e-01 -9.55327094e-01 4.68934089e-01
-4.64574665e-01 -1.29343951e+00 -7.01485395e-01 -2.46894017e-01
-8.10200512e-01 -1.05217803e+00 -5.95909834e-01 -5.53686917e-01
4.86946285e-01 7.75920510e-01 6.62566066e-01 1.58477202e-01
-3.21410447e-01 1.02254868e-01 -4.74801093e-01 -1.40641481e-02
-4.23715711e-02 -4.78550911e-01 -9.10137966e-02 5.96006155e-01
-7.24716019e-03 -2.90025026e-01 -9.11413312e-01 5.19375861e-01
-8.21211636e-01 2.66699255e-01 6.29232585e-01 7.72278428e-01
6.88131869e-01 -2.06854582e-01 2.50248849e-01 -2.54615575e-01
-6.07449226e-02 -4.70306575e-01 -7.95807064e-01 1.81178227e-01
-4.19665545e-01 -2.00957041e-02 2.41174355e-01 -6.33177638e-01
-1.26366997e+00 2.91536152e-01 3.63244772e-01 -7.74831891e-01
-1.06488507e-04 2.43897244e-01 -1.82804376e-01 -1.38556197e-01
2.65661299e-01 4.44534123e-01 -3.17622393e-01 -7.77339414e-02
1.16853692e-01 3.99000913e-01 7.96151102e-01 -4.40494180e-01
7.76581287e-01 6.55831218e-01 -3.58491056e-02 -5.38132370e-01
-7.45327175e-01 -7.46542573e-01 -6.78617001e-01 -6.69227481e-01
9.08289850e-01 -1.07731497e+00 -4.83137101e-01 4.57473993e-01
-1.06065464e+00 1.69376265e-02 -1.13867030e-01 8.78830612e-01
-3.79285574e-01 5.33331692e-01 -3.81029278e-01 -9.18610454e-01
-2.12655261e-01 -1.21645582e+00 1.06147039e+00 7.48021841e-01
3.55619937e-01 -6.52964592e-01 -3.19698572e-01 2.31374890e-01
1.63036063e-01 4.34585214e-01 3.81878585e-01 -4.44118939e-02
-1.28321719e+00 1.01977520e-01 -3.86767238e-01 2.36845002e-01
2.29445174e-02 2.33562723e-01 -8.81752253e-01 -1.08653858e-01
-1.25457510e-01 5.46977334e-02 8.36980700e-01 5.78418791e-01
1.02266598e+00 3.50528844e-02 -3.81171614e-01 6.23377442e-01
1.40463340e+00 3.71584952e-01 4.12588626e-01 3.52288842e-01
6.16944194e-01 4.59766179e-01 1.32245445e+00 7.82957971e-01
2.48602748e-01 6.86279893e-01 5.94557106e-01 1.19814962e-01
-1.19527616e-02 -1.96169868e-01 3.82421523e-01 3.75112027e-01
-1.96079507e-01 -1.09402873e-01 -4.79694724e-01 7.56068051e-01
-2.15967178e+00 -1.17031336e+00 -8.65773633e-02 2.18003440e+00
6.13780618e-01 3.09444249e-01 -7.03649744e-02 2.22014770e-01
1.02232420e+00 2.35220343e-01 -6.23055220e-01 3.36677462e-01
-3.11945230e-01 -4.25469935e-01 3.62265587e-01 3.10321689e-01
-1.16329610e+00 6.88040793e-01 5.08645535e+00 5.78944564e-01
-1.23564506e+00 3.30254398e-02 6.50616407e-01 -1.01236608e-02
-1.35482743e-01 6.39554262e-02 -6.92264497e-01 8.00060749e-01
1.87064245e-01 -6.29981831e-02 3.46748710e-01 8.65911424e-01
3.93994361e-01 -4.41743165e-01 -7.96254337e-01 9.84466910e-01
2.18499433e-02 -1.29871416e+00 -1.04667455e-01 -3.85729134e-01
7.21296072e-01 -1.13526233e-01 1.01718552e-01 5.00580017e-03
-6.71082810e-02 -3.32379282e-01 1.05580986e+00 6.70806825e-01
3.33604217e-01 -4.81041461e-01 6.35761619e-01 2.87443250e-01
-1.57514083e+00 -3.75307173e-01 -3.23051184e-01 1.21526398e-01
1.93069547e-01 4.88653451e-01 -5.13195097e-01 6.29014730e-01
8.36851239e-01 1.03540432e+00 -5.77706635e-01 1.36891186e+00
-1.01011366e-01 4.06647176e-01 -3.00222278e-01 1.58061221e-01
4.10382926e-01 -2.42016166e-01 8.39560330e-01 1.01454139e+00
1.81455880e-01 3.23441833e-01 3.45770150e-01 8.45221817e-01
4.72036302e-01 -1.19874686e-01 -2.81570032e-02 2.31517300e-01
5.11327386e-01 1.50143397e+00 -8.41067314e-01 -3.95058900e-01
-4.79441375e-01 9.26643610e-01 1.20396376e-01 5.14260471e-01
-1.04468524e+00 -9.36901644e-02 4.20889169e-01 2.01253295e-01
5.45567155e-01 -1.97297543e-01 1.04109876e-01 -1.31859362e+00
2.95233041e-01 -3.65281165e-01 5.25321245e-01 -1.22929740e+00
-8.25579762e-01 4.55380440e-01 5.96316345e-02 -1.66305411e+00
-1.10889845e-01 -3.60818028e-01 -7.37245977e-01 6.88659847e-01
-1.89311266e+00 -9.88461912e-01 -8.88933718e-01 6.93101466e-01
7.02742040e-01 2.55798567e-02 6.01752102e-03 2.83752203e-01
-5.63401937e-01 1.12438194e-01 -3.30819450e-02 1.00643873e-01
7.19661593e-01 -8.13044727e-01 -3.26242626e-01 1.31118226e+00
-1.31159797e-01 2.22166806e-01 6.70972168e-01 -5.88215351e-01
-1.30398071e+00 -1.23595011e+00 2.44156480e-01 1.00793175e-01
4.46170062e-01 9.60221216e-02 -1.00185931e+00 4.18300062e-01
-2.52827197e-01 7.19636559e-01 5.05391136e-02 -6.50843561e-01
1.00445366e-02 -2.94044793e-01 -1.05634332e+00 2.04130545e-01
9.19246316e-01 -1.41307503e-01 -3.16750318e-01 -1.82903945e-01
6.87885761e-01 -5.58493674e-01 -4.72194225e-01 5.29664338e-01
5.87000132e-01 -1.15725458e+00 8.70850742e-01 -2.34243840e-01
2.40721405e-01 -8.92950892e-01 -2.59767830e-01 -9.43291605e-01
-3.93471837e-01 -5.10884821e-01 -2.95738518e-01 1.23712492e+00
-2.53539711e-01 -2.04973713e-01 6.41363263e-01 5.65066814e-01
-1.96302414e-01 -6.80157304e-01 -9.42705870e-01 -4.92841095e-01
-7.97224581e-01 -4.46046978e-01 4.94563043e-01 8.61174822e-01
-2.60267079e-01 -2.69489765e-01 -3.37160468e-01 4.86141741e-01
7.88617313e-01 3.62988710e-01 4.60441589e-01 -1.09861851e+00
-1.98565736e-01 -2.44738430e-01 -8.41770470e-01 -1.15534329e+00
3.47595587e-02 -3.11448097e-01 2.63736963e-01 -1.15591049e+00
2.60934979e-01 -4.71616238e-01 -7.36981928e-01 1.57080099e-01
-6.77526176e-01 1.56485811e-01 3.30771238e-01 1.65313050e-01
-8.64402175e-01 8.49149823e-01 1.28556442e+00 -1.95016354e-01
-3.23556632e-01 -1.37193426e-01 -4.10530001e-01 6.95555151e-01
4.59467620e-01 -2.97167212e-01 -3.79091293e-01 -3.19915265e-01
-3.12072366e-01 4.10181552e-01 7.37825632e-01 -1.25816095e+00
2.29435578e-01 -6.84069872e-01 7.50175476e-01 -7.25523353e-01
2.56025672e-01 -8.94670129e-01 -1.09537020e-01 3.33366603e-01
-1.62439212e-01 -1.40542537e-01 9.19189528e-02 1.02116406e+00
-4.20544624e-01 3.46120298e-02 1.06047404e+00 1.82107333e-02
-1.40996099e+00 4.88585114e-01 -1.60757706e-01 -3.47258858e-02
1.31836605e+00 -4.71305430e-01 -1.58333331e-01 -2.75046736e-01
-5.58120012e-01 4.55766916e-01 2.89663404e-01 5.94239295e-01
1.09757507e+00 -1.27775645e+00 -6.22734427e-01 5.06576955e-01
2.20567584e-01 -6.57956153e-02 5.18022656e-01 9.99281526e-01
-3.22417915e-01 9.04207528e-02 -4.63923126e-01 -9.24152017e-01
-1.11361337e+00 7.64411867e-01 4.52380359e-01 1.76323161e-01
-4.13827270e-01 7.16342092e-01 3.92408550e-01 3.94544780e-01
2.33291373e-01 -4.92243320e-01 -3.43095750e-01 -1.32259652e-01
7.70975590e-01 4.58508223e-01 -2.73485929e-01 -8.11190724e-01
-4.74064648e-01 5.84255695e-01 9.64772850e-02 -2.75117308e-02
1.03592408e+00 -5.74435174e-01 2.62579709e-01 3.25534999e-01
9.29346621e-01 -2.73262143e-01 -1.91575015e+00 -5.53855300e-01
-2.23035991e-01 -1.08820939e+00 2.82630414e-01 -4.31956381e-01
-1.20375919e+00 7.87251472e-01 8.50082636e-01 -2.08061635e-01
1.37025523e+00 -2.54288942e-01 7.73332477e-01 -3.81564349e-02
3.52570653e-01 -1.02162695e+00 2.59977162e-01 6.52252659e-02
5.92996001e-01 -1.45634544e+00 2.25505885e-02 -6.17529988e-01
-8.78348231e-01 1.00691688e+00 9.56576288e-01 -1.05437718e-01
6.37098491e-01 9.56505723e-03 -1.87482417e-01 3.35239023e-01
-7.12061048e-01 -3.60184699e-01 4.50709015e-01 5.31462491e-01
-8.52608532e-02 -3.09694797e-01 -1.63998976e-01 5.18145680e-01
6.82934821e-01 2.18407050e-01 4.31499124e-01 1.00745773e+00
-6.79756761e-01 -4.82224494e-01 -4.26525533e-01 2.44841397e-01
-2.08188534e-01 1.18733004e-01 2.22235769e-01 5.83460271e-01
4.17184919e-01 9.75892007e-01 2.00413689e-01 -4.52329695e-01
-5.88359721e-02 -4.80347425e-01 2.69199342e-01 -1.12061284e-01
-2.09970489e-01 3.01899791e-01 -3.44685882e-01 -7.82310724e-01
-8.33131552e-01 -7.59546518e-01 -1.46035147e+00 4.46535945e-02
-5.11149585e-01 1.25271097e-01 2.91046798e-01 7.87842155e-01
3.22890580e-01 4.97087717e-01 8.19638133e-01 -1.08597398e+00
-1.11068346e-01 -6.63856626e-01 -5.08333921e-01 4.35543597e-01
6.33286536e-01 -1.07869411e+00 -3.06874335e-01 3.18571478e-01]
|
[9.4002103805542, -0.4027370810508728]
|
a3a4c45f-aa1b-43b6-9cf2-9e3b9e043798
|
methods-for-sparse-and-low-rank-recovery
|
1605.00507
| null |
http://arxiv.org/abs/1605.00507v1
|
http://arxiv.org/pdf/1605.00507v1.pdf
|
Methods for Sparse and Low-Rank Recovery under Simplex Constraints
|
The de-facto standard approach of promoting sparsity by means of
$\ell_1$-regularization becomes ineffective in the presence of simplex
constraints, i.e.,~the target is known to have non-negative entries summing up
to a given constant. The situation is analogous for the use of nuclear norm
regularization for low-rank recovery of Hermitian positive semidefinite
matrices with given trace. In the present paper, we discuss several strategies
to deal with this situation, from simple to more complex. As a starting point,
we consider empirical risk minimization (ERM). It follows from existing theory
that ERM enjoys better theoretical properties w.r.t.~prediction and
$\ell_2$-estimation error than $\ell_1$-regularization. In light of this, we
argue that ERM combined with a subsequent sparsification step like thresholding
is superior to the heuristic of using $\ell_1$-regularization after dropping
the sum constraint and subsequent normalization.
At the next level, we show that any sparsity-promoting regularizer under
simplex constraints cannot be convex. A novel sparsity-promoting regularization
scheme based on the inverse or negative of the squared $\ell_2$-norm is
proposed, which avoids shortcomings of various alternative methods from the
literature. Our approach naturally extends to Hermitian positive semidefinite
matrices with given trace. Numerical studies concerning compressed sensing,
sparse mixture density estimation, portfolio optimization and quantum state
tomography are used to illustrate the key points of the paper.
|
['Syama Sundar Rangapuram', 'Martin Slawski', 'Ping Li']
|
2016-05-02
| null | null | null | null |
['quantum-state-tomography']
|
['medical']
|
[ 5.19333661e-01 4.08176124e-01 -2.06744090e-01 -1.68245658e-01
-8.02309275e-01 -2.67483920e-01 8.83262232e-02 1.93251017e-02
-6.36076152e-01 1.03079271e+00 -9.13116243e-03 -3.72569889e-01
-4.91847992e-01 -6.41234636e-01 -6.01712525e-01 -1.16001821e+00
-9.93438438e-02 4.52721566e-01 -2.82154799e-01 -2.94697434e-01
2.59968489e-01 4.37203497e-01 -1.03019166e+00 -4.94232982e-01
9.97173309e-01 9.25997198e-01 -3.35003100e-02 2.28744745e-02
1.36178881e-01 4.23635334e-01 -2.24817932e-01 -4.19251174e-01
4.36308354e-01 -5.43024540e-01 -3.98174763e-01 3.91561627e-01
1.38720393e-01 1.97331831e-01 -1.97857425e-01 1.60257876e+00
3.60974252e-01 1.26269653e-01 7.32449532e-01 -9.54591036e-01
-3.36680323e-01 6.70335650e-01 -9.77841675e-01 2.04084277e-01
2.55043507e-01 -2.12540329e-01 1.04270744e+00 -1.05192554e+00
5.84425807e-01 8.32797587e-01 6.13826454e-01 2.35852003e-01
-1.51557493e+00 -7.02854335e-01 1.44247357e-02 -2.01280281e-01
-1.74981141e+00 -5.26531577e-01 8.00428092e-01 -4.40480590e-01
6.53525591e-01 2.68075734e-01 2.45463118e-01 5.60447514e-01
-7.92448968e-02 4.01333690e-01 1.27344584e+00 -6.11356795e-01
1.50331497e-01 3.29462349e-01 2.36733854e-01 9.31499720e-01
9.90352988e-01 3.83351445e-02 -4.61007953e-01 -2.43111268e-01
6.32563770e-01 -4.98325944e-01 -5.31013846e-01 -7.23665178e-01
-1.19630551e+00 9.88261998e-01 5.52942930e-03 5.97220600e-01
-3.08806956e-01 -2.44101696e-02 1.74485654e-01 2.70283550e-01
4.96840626e-01 3.25197786e-01 1.58708364e-01 1.57186374e-01
-1.20129716e+00 3.97070218e-03 8.01608562e-01 9.02934730e-01
7.66332090e-01 5.10128617e-01 2.12406084e-01 6.12484932e-01
5.37969470e-01 8.31392229e-01 -2.56857485e-01 -8.21271956e-01
6.32401705e-01 5.15178069e-02 1.51913121e-01 -1.02668476e+00
-3.08503687e-01 -8.76629293e-01 -1.20837283e+00 1.74570814e-01
6.11473203e-01 -2.40891591e-01 -5.62126994e-01 1.98693204e+00
4.24100943e-02 1.84369877e-01 1.29472479e-01 9.18059886e-01
2.45086968e-01 5.61149120e-01 -2.84163654e-01 -9.65040147e-01
8.39698493e-01 -2.51112968e-01 -1.06559861e+00 -3.19765285e-02
6.41350091e-01 -7.21762598e-01 5.62600672e-01 5.92647135e-01
-1.28750551e+00 1.54082239e-01 -1.15747523e+00 3.41624886e-01
1.01780511e-01 2.69579560e-01 6.59759820e-01 1.16598868e+00
-9.36487138e-01 4.99778658e-01 -7.11611331e-01 -1.61021873e-01
3.99945416e-02 6.36368752e-01 -4.16388571e-01 -7.25492090e-02
-8.66906464e-01 8.32638025e-01 3.46991569e-01 2.87377387e-01
-3.43939543e-01 -5.15645146e-01 -9.02203560e-01 -3.09022330e-02
4.03829157e-01 -4.97651368e-01 7.08737135e-01 -7.40269780e-01
-1.35792923e+00 8.87365103e-01 -3.51998806e-01 -3.85032564e-01
2.79400438e-01 1.24454223e-01 -3.03315014e-01 2.54181206e-01
1.69171378e-01 -2.38477260e-01 9.32852209e-01 -1.26319158e+00
-8.05869848e-02 -3.49643409e-01 -2.14921787e-01 1.16784550e-01
-1.26837075e-01 -2.52836011e-02 -1.20460533e-01 -6.13241315e-01
6.97905540e-01 -9.26852286e-01 -6.06695712e-01 -5.21826327e-01
-6.69449806e-01 2.47959435e-01 3.99094939e-01 -4.46312189e-01
1.47832763e+00 -2.20852280e+00 6.47722006e-01 1.10859537e+00
1.95905104e-01 1.38376832e-01 1.42570764e-01 4.20898348e-01
-5.38237214e-01 5.28506786e-02 -7.61781812e-01 -3.86938214e-01
3.12807858e-02 7.14724362e-02 -1.92451879e-01 1.11183035e+00
9.05197114e-02 1.91278979e-01 -7.85625875e-01 -2.52271116e-01
8.05670917e-02 4.05716240e-01 -7.06206977e-01 -4.29952621e-01
1.04945287e-01 5.97093642e-01 -4.38582927e-01 6.32181644e-01
8.16397786e-01 -3.04608434e-01 4.46198106e-01 -3.21270823e-01
-3.06538075e-01 -3.91731374e-02 -1.85319746e+00 1.45040953e+00
-1.17471777e-02 2.72829622e-01 8.06338012e-01 -1.53412938e+00
6.70519292e-01 3.63091946e-01 7.24444509e-01 -3.84731770e-01
3.34538519e-01 6.70076668e-01 1.03064567e-01 -5.14721386e-02
4.56525832e-01 -9.22508180e-01 -9.47225317e-02 4.70498621e-01
8.09949562e-02 -7.34772533e-02 4.01142150e-01 4.80775148e-01
9.56453323e-01 -4.69805330e-01 4.85116690e-01 -6.77275479e-01
7.84319818e-01 -2.73000509e-01 6.04545295e-01 6.80430889e-01
2.95392536e-02 3.37969691e-01 6.06648684e-01 4.24072295e-01
-9.03320909e-01 -1.02198112e+00 -4.87356842e-01 5.64291894e-01
8.31530169e-02 -3.25332552e-01 -5.55261672e-01 -2.04759687e-01
-2.35684648e-01 7.06603169e-01 -4.01273161e-01 7.90829584e-02
-5.03835678e-01 -1.41246998e+00 3.46661478e-01 -6.72245771e-02
5.23995221e-01 -3.67639750e-01 7.88888261e-02 1.76514208e-01
-1.30261391e-01 -8.81739795e-01 -2.69764781e-01 5.55828512e-01
-9.15926099e-01 -8.79099905e-01 -7.07992733e-01 -5.50095260e-01
1.03283978e+00 2.92810977e-01 8.40296566e-01 -1.25241131e-01
-1.40942512e-02 6.72829986e-01 -1.45503491e-01 -5.86212762e-02
-1.52358875e-01 -2.35930428e-01 3.90981555e-01 3.05304646e-01
-2.03455724e-02 -7.77222514e-01 -2.97600716e-01 8.93989801e-02
-1.03457439e+00 -4.16665345e-01 5.37675083e-01 9.12664294e-01
8.22512746e-01 1.55595213e-01 3.90384495e-01 -1.02418447e+00
6.11402631e-01 -5.71262896e-01 -6.64437473e-01 9.08516049e-02
-5.64623356e-01 2.02462971e-01 4.35837001e-01 -9.76334736e-02
-8.65262151e-01 7.81033188e-02 -5.28881736e-02 -2.74943948e-01
4.59883869e-01 8.31932068e-01 -2.05312580e-01 -4.67822403e-01
6.79035246e-01 3.31447542e-01 -2.18065642e-02 -3.22689593e-01
4.14959282e-01 2.70139784e-01 2.68684119e-01 -6.30165875e-01
1.18239331e+00 7.38153756e-01 5.11346936e-01 -1.21708238e+00
-9.96910632e-01 -6.19082212e-01 -3.31662863e-01 3.87988724e-02
6.67635679e-01 -8.73127878e-01 -5.80493510e-01 9.44755822e-02
-7.34686792e-01 1.21207990e-01 -5.23881257e-01 9.63019609e-01
-7.73246109e-01 8.62186551e-01 -4.44887638e-01 -1.15905094e+00
-9.73855481e-02 -1.07234466e+00 5.99899054e-01 -2.01931253e-01
2.04528019e-01 -9.53434825e-01 1.04194961e-01 3.61786991e-01
3.62252057e-01 1.09129846e-01 8.54144216e-01 -4.16770279e-01
-5.20651579e-01 -2.44791225e-01 -1.92350000e-01 6.14825964e-01
-2.88691044e-01 -3.53729218e-01 -4.54004824e-01 -5.51756203e-01
6.31715477e-01 -1.51573390e-01 8.84605050e-01 5.25710225e-01
7.29404867e-01 -2.47376248e-01 -1.26013398e-01 6.40771925e-01
1.66371500e+00 -9.41403732e-02 7.19353378e-01 1.14637658e-01
4.79196280e-01 4.77621198e-01 5.49849093e-01 7.50455141e-01
-1.29137680e-01 5.41760445e-01 4.58704293e-01 3.29895765e-02
3.98881435e-01 2.60912478e-01 4.06740189e-01 1.15512347e+00
-3.41063768e-01 -5.13004772e-02 -5.25775969e-01 2.14363486e-01
-1.52486479e+00 -1.24752438e+00 -4.63084996e-01 2.66371465e+00
6.38640761e-01 3.14549394e-02 4.64391373e-02 4.10291970e-01
8.46428275e-01 2.19204411e-01 -9.31902826e-02 -1.40469283e-01
-3.53006721e-01 5.27338028e-01 8.90942454e-01 8.76622021e-01
-9.34490204e-01 4.63792175e-01 5.65965176e+00 1.15298700e+00
-9.51357603e-01 2.61632055e-01 2.56119967e-01 -9.56949890e-02
-5.61809957e-01 1.44120246e-01 -7.42669165e-01 3.29122633e-01
6.51281834e-01 -1.63017958e-01 5.13436854e-01 4.45835769e-01
2.96480685e-01 -4.58286524e-01 -7.89515018e-01 1.20789087e+00
2.34335616e-01 -1.19114530e+00 -2.99485058e-01 3.95339251e-01
6.73559844e-01 -6.32412583e-02 2.90843695e-01 6.29720315e-02
-5.20640984e-02 -9.15885746e-01 5.19500494e-01 3.21883291e-01
8.55833948e-01 -8.43162000e-01 5.47989547e-01 4.04147923e-01
-1.05448890e+00 1.70823202e-01 -2.89795518e-01 -2.71029044e-02
5.44644654e-01 1.11992598e+00 -3.74022663e-01 7.62631118e-01
-1.31369671e-02 6.05670035e-01 1.63308933e-01 1.06015587e+00
-6.00161357e-03 6.72956169e-01 -7.06424952e-01 2.65168875e-01
2.57759720e-01 -1.05845010e+00 1.02168930e+00 1.21733844e+00
6.93319619e-01 4.52392489e-01 1.17101856e-01 7.71836281e-01
4.00583819e-02 3.61587346e-01 -7.99830794e-01 -7.47830719e-02
8.66860002e-02 9.01710510e-01 -6.95516825e-01 -1.92715287e-01
-5.27814507e-01 6.91777587e-01 6.57312796e-02 4.70979303e-01
-6.48561597e-01 -1.68304995e-01 3.14955384e-01 3.14100444e-01
2.51254648e-01 -5.20146191e-01 -4.17836994e-01 -1.57156706e+00
1.14106111e-01 -8.37328494e-01 2.53163487e-01 -1.13761954e-01
-1.19706452e+00 2.62886643e-01 1.33057432e-02 -1.23648751e+00
-1.84301995e-02 -5.34534931e-01 -2.78895646e-01 9.54836190e-01
-1.51701033e+00 -5.93964279e-01 4.18658346e-01 7.92630494e-01
-1.91922203e-01 1.31088654e-02 7.03003526e-01 4.99209344e-01
-8.75284195e-01 3.64862472e-01 5.55560827e-01 -1.67811289e-01
2.34745011e-01 -1.16810000e+00 -7.67126441e-01 1.14580250e+00
1.13513984e-01 9.39116538e-01 1.02343035e+00 -6.18123353e-01
-1.50921750e+00 -6.23454571e-01 9.23326373e-01 1.13520168e-01
1.05094230e+00 -6.52535781e-02 -6.40339971e-01 6.43317163e-01
9.26099867e-02 -3.91538702e-02 1.00975084e+00 2.21071407e-01
-2.22129852e-01 6.25377148e-02 -1.26050830e+00 3.11778635e-01
8.42936099e-01 -5.94624758e-01 -3.76695454e-01 7.08399594e-01
-4.94731404e-02 -1.54079586e-01 -9.54985261e-01 6.56563163e-01
1.74559280e-01 -9.60484922e-01 1.03830528e+00 -2.44088873e-01
-1.57674313e-01 -4.38668489e-01 -5.63831806e-01 -9.08807576e-01
-2.36016661e-01 -1.02138674e+00 1.58095479e-01 6.65262222e-01
5.26004553e-01 -6.65778458e-01 8.47250283e-01 2.93247342e-01
-2.17236921e-01 -5.80749691e-01 -1.37238300e+00 -9.09469962e-01
2.41749063e-02 -4.54985350e-01 -4.55987722e-01 1.14895499e+00
4.69173133e-01 4.48128164e-01 -6.96704149e-01 2.88723320e-01
1.04143608e+00 5.03805885e-03 3.69201481e-01 -1.03980339e+00
-7.13285506e-01 -3.44803303e-01 -2.79102415e-01 -1.14274633e+00
1.10052571e-01 -1.03763890e+00 -7.60005563e-02 -1.07298589e+00
2.35214010e-01 -7.24459648e-01 -2.16489464e-01 -9.38592758e-03
1.86317295e-01 4.24697280e-01 2.71608144e-01 1.87744141e-01
-5.09468079e-01 5.62684953e-01 1.20758951e+00 -1.32681429e-02
-1.89298272e-01 1.64414123e-01 -6.34956717e-01 8.50915551e-01
6.68301046e-01 -4.22259450e-01 -3.31474632e-01 2.93028308e-04
6.99770391e-01 5.10874271e-01 8.46832693e-02 -9.11670864e-01
1.05066612e-01 4.10730578e-02 -2.62439311e-01 -3.68648589e-01
6.50511503e-01 -8.48904431e-01 3.45072985e-01 3.66285950e-01
-4.83422503e-02 -2.63384283e-01 -1.72160730e-01 7.50897884e-01
-2.37563223e-01 -8.40413153e-01 9.74015772e-01 -1.04893386e-01
-1.54866338e-01 4.22286727e-02 -5.40294707e-01 1.48244217e-01
8.96500707e-01 -2.64645845e-01 2.33120620e-02 -5.22202373e-01
-1.32732034e+00 -1.63206115e-01 3.07393312e-01 -6.49934769e-01
5.24941325e-01 -1.11161363e+00 -6.64339960e-01 6.38001487e-02
-1.57988861e-01 -3.32378060e-01 2.10689157e-01 1.67498744e+00
-3.32501471e-01 6.11783504e-01 2.65527874e-01 -4.29447562e-01
-1.04178786e+00 3.53199691e-01 2.69498020e-01 -4.68879879e-01
-2.01226428e-01 9.17111218e-01 5.94024397e-02 -6.83208257e-02
1.95989788e-01 -1.25924125e-01 2.48638634e-02 5.31138852e-02
1.31939366e-01 4.06456023e-01 9.21991765e-02 -1.09467745e+00
-2.32249826e-01 6.59175694e-01 1.89468488e-01 -4.91734236e-01
1.25849557e+00 -1.90647423e-01 -5.86127222e-01 3.25200677e-01
1.03038430e+00 6.77101731e-01 -6.91399813e-01 -2.87551582e-01
-9.04500205e-03 -3.91688675e-01 5.82644641e-02 -3.06206524e-01
-1.19197786e+00 6.67509139e-01 2.16067642e-01 1.53342232e-01
1.04411411e+00 -6.84102178e-02 2.37820804e-01 4.94053930e-01
6.61898971e-01 -9.66981590e-01 -8.74256790e-02 4.28404540e-01
6.32123888e-01 -1.02115262e+00 2.55041689e-01 -9.73522484e-01
-3.66750509e-01 9.32240248e-01 6.37417194e-04 -3.76437306e-01
8.06242168e-01 1.02164239e-01 -5.30470252e-01 -3.87262225e-01
-8.61800760e-02 -4.34172779e-01 1.21359535e-01 2.87399799e-01
4.83350337e-01 5.49193211e-02 -8.85908663e-01 2.22105592e-01
5.84505610e-02 -3.07186991e-01 7.23928690e-01 8.55523169e-01
-5.68534672e-01 -1.39211416e+00 -6.77335024e-01 6.25287175e-01
-6.05142951e-01 -3.11934769e-01 5.26131094e-02 7.40274251e-01
1.04973733e-01 1.01628673e+00 -4.60308343e-01 -1.36953872e-02
1.39529154e-01 -3.47264633e-02 7.74623096e-01 -8.30281079e-01
-2.31507316e-01 4.51892227e-01 3.13956618e-01 -4.01737005e-01
-8.30900967e-01 -1.02682960e+00 -1.06312156e+00 -3.37340564e-01
-5.83857715e-01 5.24735689e-01 4.30538446e-01 8.90738249e-01
-2.32984290e-01 5.39179184e-02 4.91887510e-01 -6.68455839e-01
-9.21041369e-01 -6.77831888e-01 -1.22677076e+00 1.18185036e-01
1.58854336e-01 -7.01011658e-01 -8.31485868e-01 -2.65975982e-01]
|
[6.833868980407715, 4.6317596435546875]
|
1f079489-52a2-4d48-b57b-067fd44f2fba
|
life-learning-individual-features-for
|
2109.14844
| null |
https://arxiv.org/abs/2109.14844v2
|
https://arxiv.org/pdf/2109.14844v2.pdf
|
LIFE: Learning Individual Features for Multivariate Time Series Prediction with Missing Values
|
Multivariate time series (MTS) prediction is ubiquitous in real-world fields, but MTS data often contains missing values. In recent years, there has been an increasing interest in using end-to-end models to handle MTS with missing values. To generate features for prediction, existing methods either merge all input dimensions of MTS or tackle each input dimension independently. However, both approaches are hard to perform well because the former usually produce many unreliable features and the latter lacks correlated information. In this paper, we propose a Learning Individual Features (LIFE) framework, which provides a new paradigm for MTS prediction with missing values. LIFE generates reliable features for prediction by using the correlated dimensions as auxiliary information and suppressing the interference from uncorrelated dimensions with missing values. Experiments on three real-world data sets verify the superiority of LIFE to existing state-of-the-art models.
|
['Zhi-Hua Zhou', 'Yuan Jiang', 'Shao-Qun Zhang', 'Zhao-Yu Zhang']
|
2021-09-30
| null | null | null | null |
['time-series-prediction']
|
['time-series']
|
[ 1.08397059e-01 -4.70298231e-01 -1.92759529e-01 -4.91515726e-01
-8.30542624e-01 -1.16201349e-01 3.87682259e-01 1.38524905e-01
-1.63681600e-02 9.35649157e-01 2.10845679e-01 1.53988637e-02
-3.43213737e-01 -5.85876226e-01 -3.54284495e-01 -9.09949541e-01
-1.98567152e-01 2.19361886e-01 2.38366857e-01 -4.12270933e-01
2.23478094e-01 9.83441249e-02 -1.79568195e+00 3.47666323e-01
1.21658981e+00 1.37085509e+00 6.49977103e-02 1.20858833e-01
-2.80368268e-01 7.81586945e-01 -8.20013165e-01 -1.88990444e-01
2.34845877e-01 -2.12933481e-01 -2.89304882e-01 1.70835096e-03
-1.77841246e-01 7.92088285e-02 -3.69346321e-01 4.71799314e-01
5.74111879e-01 -2.46444426e-04 4.92099285e-01 -1.63037431e+00
-3.03213149e-01 5.10940731e-01 -6.46765828e-01 2.86564589e-01
1.92303985e-01 -8.92080739e-03 9.42724705e-01 -8.50930750e-01
1.48912892e-01 9.59051073e-01 6.63694739e-01 1.85299009e-01
-1.26508915e+00 -7.98540711e-01 2.26239100e-01 5.42651772e-01
-1.16569674e+00 -2.92644769e-01 1.06149602e+00 -3.54511559e-01
7.19554722e-01 3.35895330e-01 5.15002251e-01 1.33063388e+00
5.70835590e-01 1.26795495e+00 1.16596866e+00 -1.25454739e-01
1.64247006e-01 -2.25092262e-01 1.19443230e-01 -1.84387445e-01
-4.24643308e-02 2.04587877e-01 -7.62997270e-01 -3.14097285e-01
2.69203126e-01 5.87926447e-01 -3.00218668e-02 -2.18921512e-01
-1.58109689e+00 5.71372211e-01 -8.65045786e-02 4.21589077e-01
-6.04587317e-01 -4.03830767e-01 5.27584314e-01 6.08524919e-01
6.89101040e-01 2.95834485e-02 -8.66827488e-01 -5.40475607e-01
-7.86254168e-01 3.75732392e-01 4.74478304e-01 1.03657830e+00
4.99452710e-01 1.03623338e-01 -3.34479272e-01 8.88748884e-01
-9.99647528e-02 5.59608459e-01 8.56990218e-01 -4.82048601e-01
9.21416163e-01 6.45190358e-01 1.81462735e-01 -9.76406038e-01
-6.09494746e-01 -8.41243148e-01 -1.19457829e+00 -2.11493313e-01
2.94369102e-01 -2.56318480e-01 -5.24996579e-01 1.67316210e+00
2.05450058e-01 4.75934863e-01 1.53632119e-01 8.07704866e-01
3.96736175e-01 7.13866591e-01 -1.64554223e-01 -6.08242869e-01
6.21078908e-01 -7.41642177e-01 -8.67138088e-01 -8.17305595e-03
5.60383976e-01 -8.03876996e-01 9.75299418e-01 8.43023956e-01
-6.65184081e-01 -7.65807748e-01 -8.80061686e-01 2.34133527e-01
-5.38394181e-03 1.80417463e-01 5.96368015e-01 3.27315807e-01
-5.98451912e-01 8.54021847e-01 -1.01142955e+00 -6.63611293e-02
1.76572114e-01 5.77326655e-01 -3.41042578e-01 8.64048302e-02
-1.20089841e+00 5.56575716e-01 2.77754396e-01 1.92977503e-01
-3.78252864e-01 -3.91298115e-01 -5.57758272e-01 -6.33997843e-02
3.42223823e-01 -2.92290449e-01 1.24155819e+00 -9.18560803e-01
-1.32652700e+00 -1.93497673e-01 -3.23806107e-01 -3.39879364e-01
5.01410723e-01 -4.10088748e-01 -1.02802324e+00 -5.23229718e-01
2.16460824e-01 -4.02681306e-02 1.12040460e+00 -8.04394186e-01
-1.12955046e+00 -5.25082886e-01 -5.24775922e-01 -9.82698426e-03
-7.61348307e-01 -2.91699111e-01 -2.94390246e-02 -9.21281219e-01
3.63368899e-01 -8.27752292e-01 -3.90736878e-01 -5.55438221e-01
-4.67886329e-01 -4.33331430e-01 1.25578427e+00 -5.59698939e-01
1.53707314e+00 -2.22176409e+00 5.26283383e-02 2.54546791e-01
1.86691642e-01 3.03687379e-02 -3.23356360e-01 7.87198246e-01
-2.32463926e-01 -3.35970253e-01 -8.05242360e-02 -5.36844850e-01
-7.52938092e-02 2.55335838e-01 -5.91593087e-01 3.33440483e-01
2.65938818e-01 5.32037139e-01 -8.95661831e-01 -2.56805688e-01
2.39248514e-01 2.59842813e-01 -1.07579827e-01 2.23029420e-01
7.99521655e-02 8.93566370e-01 -6.33159459e-01 6.97670817e-01
6.00003719e-01 -4.40171883e-02 -5.72853312e-02 -7.17499182e-02
-1.81439683e-01 1.11929365e-01 -1.28913248e+00 1.51699150e+00
-1.70232728e-01 3.01588893e-01 -7.09114790e-01 -1.02878559e+00
1.25797474e+00 2.43804842e-01 1.22782266e+00 -7.58070648e-01
-7.39353150e-02 4.53574121e-01 -4.47445288e-02 -5.32600462e-01
5.28251052e-01 -1.18977234e-01 -2.84138620e-01 1.95725843e-01
-5.63643128e-02 2.05648556e-01 1.99899957e-01 -2.51712918e-01
1.43972552e+00 5.71242943e-02 1.81209520e-02 2.96277791e-01
5.82610905e-01 -3.39130938e-01 1.25072336e+00 2.80620039e-01
-1.18985660e-01 7.00438917e-01 4.18003917e-01 -6.97539151e-01
-8.94816220e-01 -1.03248167e+00 5.94376400e-02 8.63510311e-01
-3.26481014e-02 -7.34605134e-01 -4.81669009e-02 -8.37484837e-01
-6.01258362e-04 5.84973454e-01 -3.95482808e-01 -2.93781489e-01
-4.33428019e-01 -8.43030453e-01 2.03472629e-01 6.90217674e-01
1.14893802e-01 -8.85986686e-01 -3.29007715e-01 7.83254027e-01
-5.97172022e-01 -1.00099075e+00 -2.44945884e-01 4.11086053e-01
-1.22948027e+00 -7.60081649e-01 -5.52506864e-01 -3.97395879e-01
2.82004654e-01 3.36101651e-01 9.49138224e-01 -3.07011575e-01
5.61413765e-02 -2.30921417e-01 -7.05946565e-01 -6.38277769e-01
3.23865674e-02 3.25721473e-01 3.50426286e-01 3.65691870e-01
4.18536872e-01 -8.95577252e-01 -4.18512344e-01 4.41296250e-01
-8.91609728e-01 -3.34264860e-02 6.51077449e-01 9.55894530e-01
7.58128524e-01 4.06254888e-01 1.00698388e+00 -5.59217811e-01
6.84151709e-01 -7.87549555e-01 -3.04427654e-01 -4.89807837e-02
-7.58064806e-01 8.47184807e-02 1.21610034e+00 -5.03949165e-01
-7.86169708e-01 2.20530972e-01 -2.05921941e-02 -5.72265327e-01
-1.28171548e-01 8.74593794e-01 -1.32248685e-01 4.36891347e-01
4.01770294e-01 5.08561671e-01 -5.69637045e-02 -7.63459146e-01
-2.30008498e-01 8.64175320e-01 3.59588891e-01 -4.38058496e-01
8.63247991e-01 2.81916887e-01 3.04016829e-01 -6.22971296e-01
-7.41689205e-01 -5.68402827e-01 -7.06125557e-01 2.28437800e-02
3.72336134e-02 -8.97891641e-01 -2.82571226e-01 6.24343157e-01
-7.68074572e-01 1.59574732e-01 -4.48571771e-01 6.85509861e-01
-7.27708757e-01 5.24020076e-01 -3.77562940e-01 -8.91696990e-01
-3.91774058e-01 -9.17251945e-01 1.25703382e+00 6.06645532e-02
-1.49310574e-01 -4.33934510e-01 6.79439530e-02 1.41008124e-01
3.41624588e-01 4.54031736e-01 7.96733439e-01 -8.21996629e-01
-1.12843350e-01 -5.16492665e-01 2.34468102e-01 3.15712631e-01
3.49991411e-01 5.92737517e-04 -8.78275156e-01 -3.09352845e-01
3.20836790e-02 -1.84435248e-01 5.39913952e-01 1.75849557e-01
1.37707007e+00 -1.78254687e-03 -2.01015338e-01 3.02374184e-01
1.09417546e+00 2.14841083e-01 5.72815001e-01 2.88860977e-01
4.78349298e-01 5.21126509e-01 1.32640064e+00 1.16341722e+00
5.99411130e-01 5.97707033e-01 4.04817492e-01 6.99250549e-02
5.62932849e-01 -9.62482244e-02 4.79968011e-01 1.28928602e+00
-2.43260533e-01 -1.25143245e-01 -8.99953067e-01 3.76902819e-01
-2.38342166e+00 -1.06336927e+00 -5.95961392e-01 2.32356787e+00
6.07268870e-01 3.48393112e-01 4.69050467e-01 9.04029489e-01
5.40835857e-01 4.14859764e-02 -6.21653497e-01 2.00704113e-01
-3.35072011e-01 1.28200427e-01 1.54350266e-01 -4.01607305e-01
-1.23496783e+00 5.02361059e-01 5.61029816e+00 9.32550669e-01
-1.25339675e+00 -1.24490140e-02 4.83041883e-01 -9.66556594e-02
-1.32322401e-01 -5.31908982e-02 -6.25850439e-01 8.68228674e-01
1.03321087e+00 -2.27700591e-01 8.99431854e-02 8.24157119e-01
5.74684203e-01 1.39995620e-01 -9.76159215e-01 1.23052990e+00
-2.42342621e-01 -8.26677740e-01 -2.15368822e-01 6.88437894e-02
7.46441841e-01 8.23638961e-02 1.43835708e-01 6.82807148e-01
-1.55136973e-01 -6.96937203e-01 6.40265048e-01 8.75954449e-01
5.09270906e-01 -1.01813734e+00 1.09025085e+00 7.63779581e-01
-1.43100965e+00 -5.26603162e-01 -2.79171199e-01 -2.42189631e-01
1.09454565e-01 9.94247437e-01 -4.98413533e-01 1.10015368e+00
6.39199913e-01 1.18709540e+00 -7.25360096e-01 1.41258204e+00
2.11137787e-01 7.29162097e-01 -3.91971558e-01 4.48312499e-02
4.84250709e-02 -1.14956900e-01 4.27391052e-01 7.07456946e-01
7.88408935e-01 -2.44578689e-01 4.58322316e-01 3.23126227e-01
3.71098310e-01 2.72934139e-01 -4.13043797e-01 4.21325676e-02
4.20914352e-01 1.07459688e+00 -3.33375454e-01 -1.50094628e-01
-7.86528111e-01 8.38831007e-01 -5.90831526e-02 2.21490592e-01
-8.96098912e-01 -3.55845869e-01 4.92778867e-01 -5.24667688e-02
1.39230087e-01 -3.62320215e-01 -4.60053235e-01 -1.38472867e+00
4.15860921e-01 -9.54776168e-01 3.92869830e-01 -5.20623267e-01
-1.83060467e+00 4.62129027e-01 -3.47826421e-01 -2.40736175e+00
-4.86984074e-01 -4.01527770e-02 -4.79806632e-01 8.71455014e-01
-1.26735318e+00 -1.11283541e+00 -3.05587649e-01 9.02809501e-01
8.80589664e-01 -2.87611216e-01 7.37779856e-01 4.22810078e-01
-4.66380626e-01 4.81428266e-01 6.37323439e-01 -1.91003039e-01
8.14175427e-01 -1.09369779e+00 3.78754467e-01 6.38515055e-01
1.24509996e-02 2.84511298e-01 7.19900072e-01 -6.64040565e-01
-1.56305289e+00 -1.21088815e+00 1.09845364e+00 -2.70892560e-01
6.88138425e-01 -1.62772715e-01 -8.77955258e-01 3.54988575e-01
-2.51760364e-01 1.69493139e-01 9.79921222e-01 2.59618968e-01
4.33053486e-02 -6.51358485e-01 -8.50728393e-01 4.26712334e-01
8.15951586e-01 -1.90054387e-01 -4.23001230e-01 5.89948408e-02
4.37298805e-01 -1.34112984e-01 -1.12047613e+00 7.19528019e-01
5.50850809e-01 -8.72057438e-01 6.05718136e-01 -5.14382780e-01
1.79129258e-01 -4.59459513e-01 -3.41439217e-01 -1.56185293e+00
-5.35545290e-01 -5.98874211e-01 -4.40082848e-01 1.29718316e+00
3.61056745e-01 -5.56086421e-01 8.12663794e-01 4.62176561e-01
-3.25148433e-01 -8.59017551e-01 -1.06619489e+00 -1.08980441e+00
-1.54228315e-01 -8.29667032e-01 1.05659986e+00 9.73806143e-01
2.44844884e-01 3.81373376e-01 -6.70347393e-01 -1.01731144e-01
5.24055779e-01 3.49198610e-01 9.50650036e-01 -1.79730439e+00
-1.89069882e-01 -1.35514408e-01 -7.45487154e-01 -6.45315886e-01
-6.45677419e-03 -4.69525903e-01 -1.21531554e-01 -1.27069545e+00
-1.37371689e-01 -5.59771299e-01 -7.92133689e-01 4.80220526e-01
-2.46557146e-01 -1.73548535e-01 1.34229466e-01 4.97307569e-01
-6.41891718e-01 1.11517489e+00 9.70405936e-01 -8.02994892e-02
-3.72487724e-01 6.09025359e-01 -4.69588518e-01 4.88004655e-01
1.02022874e+00 -4.51795191e-01 -6.97524309e-01 -2.02426150e-01
-3.55303399e-02 3.62408549e-01 -9.14904103e-02 -1.50941205e+00
1.33943871e-01 -3.72617751e-01 7.61480749e-01 -1.08752429e+00
3.78619432e-01 -1.20822334e+00 2.56875187e-01 1.64417282e-01
-7.66622499e-02 2.07146451e-01 -2.06582218e-01 7.55256295e-01
-5.98768830e-01 3.74712467e-01 2.88328052e-01 4.98488933e-01
-6.06720746e-01 6.52450800e-01 -2.84892738e-01 -2.89478034e-01
9.90742147e-01 -2.21041963e-01 -2.60392241e-02 -4.77872878e-01
-8.58738005e-01 3.83724034e-01 9.46942046e-02 7.78815448e-01
7.69462526e-01 -1.60979283e+00 -6.32430971e-01 4.16307092e-01
3.09933841e-01 -1.48741856e-01 2.97927260e-01 1.15326154e+00
3.69578928e-01 1.04633972e-01 -2.20806271e-01 -7.09501445e-01
-1.04243135e+00 6.81592286e-01 -1.67652190e-01 -4.33309078e-01
-4.42035139e-01 1.50363788e-01 -3.13735247e-01 -3.34735513e-01
2.23865554e-01 -3.93278658e-01 -3.47975105e-01 2.27171764e-01
5.32828450e-01 6.15260541e-01 3.58871192e-01 -7.47027218e-01
-1.67397842e-01 3.74406010e-01 -2.00351123e-02 -1.74726006e-02
1.76195335e+00 -3.48764926e-01 2.12099478e-01 1.05462039e+00
9.28861916e-01 -2.42169827e-01 -1.14488661e+00 -4.85569715e-01
3.93513620e-01 -6.48310900e-01 -2.05753550e-01 -6.43496573e-01
-1.05337417e+00 8.57727408e-01 7.14603364e-01 4.45253491e-01
1.53148258e+00 -5.00329554e-01 1.13551426e+00 3.28777701e-01
9.19114232e-01 -1.36228693e+00 -4.08023000e-02 6.84070885e-01
7.66074061e-01 -1.18859971e+00 -3.64410520e-01 -3.21829975e-01
-8.17825377e-01 1.23278332e+00 7.21960068e-01 -6.99089691e-02
6.72350407e-01 2.52522618e-01 -1.39859021e-01 4.31506217e-01
-1.18826365e+00 -2.14480132e-01 1.48935631e-01 6.61261201e-01
4.67849195e-01 1.01125009e-01 -5.70916712e-01 1.19424570e+00
-1.16240136e-01 3.98589730e-01 2.70362318e-01 1.29587567e+00
-2.07110122e-01 -1.57140052e+00 -5.62826455e-01 1.02192068e+00
-6.09531224e-01 1.70986623e-01 -1.80253521e-01 4.44038838e-01
2.58546948e-01 1.34141362e+00 -2.33118191e-01 -1.08613229e+00
6.56006575e-01 2.57073194e-02 7.09732547e-02 -2.60770172e-01
-5.53508818e-01 5.56828856e-01 -2.96756011e-02 -5.80155671e-01
-2.18576565e-01 -1.03938448e+00 -1.26830280e+00 -2.06638351e-01
-2.93989211e-01 1.17244124e-01 3.86687696e-01 9.44820940e-01
5.90078413e-01 4.51205552e-01 1.35146511e+00 -8.27027261e-01
-7.83431411e-01 -1.12264931e+00 -9.65621710e-01 4.32354987e-01
4.19702977e-01 -7.00824559e-01 -8.50776434e-02 -1.41149955e-02]
|
[7.174736022949219, 2.85905385017395]
|
18e698ea-ed1a-47ba-b47b-b7e1eb21ec1e
|
semantic-sensor-network-ontology-based
|
2204.03059
| null |
https://arxiv.org/abs/2204.03059v2
|
https://arxiv.org/pdf/2204.03059v2.pdf
|
Semantic Sensor Network Ontology based Decision Support System for Forest Fire Management
|
The forests are significant assets for every country. When it gets destroyed, it may negatively impact the environment, and forest fire is one of the primary causes. Fire weather indices are widely used to measure fire danger and are used to issue bushfire warnings. It can also be used to predict the demand for emergency management resources. Sensor networks have grown in popularity in data collection and processing capabilities for a variety of applications in industries such as medical, environmental monitoring, home automation etc. Semantic sensor networks can collect various climatic circumstances like wind speed, temperature, and relative humidity. However, estimating fire weather indices is challenging due to the various issues involved in processing the data streams generated by the sensors. Hence, the importance of forest fire detection has increased day by day. The underlying Semantic Sensor Network (SSN) ontologies are built to allow developers to create rules for calculating fire weather indices and also the convert dataset into Resource Description Framework (RDF). This research describes the various steps involved in developing rules for calculating fire weather indices. Besides, this work presents a Web-based mapping interface to help users visualize the changes in fire weather indices over time. With the help of the inference rule, it designed a decision support system using the SSN ontology and query on it through SPARQL. The proposed fire management system acts according to the situation, supports reasoning and the general semantics of the open-world followed by all the ontologies
|
['Sonali Agarwal', 'Kumar Abhishek', 'Navjot Singh', 'Ritesh Chandra']
|
2022-04-03
| null | null | null | null |
['fire-detection']
|
['time-series']
|
[ 1.18583433e-01 -2.95702457e-01 -2.60916889e-01 -4.23530996e-01
6.97237492e-01 -5.01376927e-01 6.18897021e-01 5.21981478e-01
-3.06301028e-01 9.40509260e-01 3.05573702e-01 -2.86711663e-01
-6.97716594e-01 -1.90020251e+00 1.67379931e-01 -4.17213768e-01
-2.84400098e-02 2.28042066e-01 5.92229486e-01 -5.49835086e-01
1.49342462e-01 6.90098882e-01 -2.21653652e+00 -1.42718712e-02
6.73282444e-01 1.27838933e+00 5.56616783e-01 3.66099179e-01
-6.37377501e-01 4.79660153e-01 -5.88493347e-01 3.46343517e-01
-3.52291577e-02 7.42297024e-02 -6.26585603e-01 -4.81990039e-01
-4.44200009e-01 -4.38164741e-01 3.13433141e-01 1.07726920e+00
2.54816532e-01 2.76045382e-01 2.72776991e-01 -1.57651734e+00
-2.01508701e-01 3.59216958e-01 9.21284482e-02 1.13116354e-01
4.10151780e-01 -2.20165104e-01 3.69101644e-01 -1.07550561e-01
4.99816954e-01 9.83051181e-01 3.04110110e-01 1.58520862e-01
-3.20145398e-01 -7.30948746e-01 -1.12280212e-01 6.36104167e-01
-1.13525534e+00 -2.86466032e-01 1.81306541e-01 -2.02524185e-01
1.18770182e+00 8.47665310e-01 9.45503771e-01 3.44147801e-01
1.22057088e-01 -3.52966011e-01 1.27335668e+00 -4.87840265e-01
4.96920735e-01 -1.19183317e-01 8.91378671e-02 2.45086938e-01
5.69289863e-01 1.98536143e-01 -4.52884465e-01 -3.46860588e-02
3.79352570e-01 3.58025044e-01 -1.43092662e-01 2.02619866e-01
-9.25884783e-01 4.14754570e-01 7.13061213e-01 5.29090285e-01
-6.10534132e-01 -2.89136648e-01 1.73159778e-01 3.90670970e-02
2.39871487e-01 -4.03180346e-02 -4.33026105e-01 -1.32469684e-01
-6.26372695e-01 2.83542067e-01 9.65164363e-01 6.10109985e-01
7.10896432e-01 -1.10124785e-03 5.22156596e-01 7.08052158e-01
6.26376390e-01 1.22956634e+00 2.00096384e-01 -8.65472674e-01
6.45118430e-02 8.80887389e-01 3.23726088e-01 -1.11343253e+00
-6.09241903e-01 2.19387963e-01 -8.55302930e-01 4.67353463e-01
7.22439587e-02 4.10908610e-02 -9.57889318e-01 1.33057487e+00
6.41738534e-01 -6.73396140e-02 2.48782560e-01 8.39204848e-01
9.64313388e-01 8.23376298e-01 7.42877007e-01 -2.59376228e-01
1.54971135e+00 3.43430117e-02 -9.43707049e-01 -7.05278516e-02
2.05885395e-02 -5.74883759e-01 7.68061876e-01 6.53045950e-03
-3.42762560e-01 -1.27899311e-02 -9.50276613e-01 4.61076230e-01
-1.39865708e+00 -5.67805648e-01 5.94958305e-01 4.74693149e-01
-7.30879247e-01 3.84698391e-01 -4.95370805e-01 -1.25593901e+00
8.66820514e-02 -1.75972525e-02 -1.36253759e-01 -6.08617580e-03
-1.72730565e+00 1.26602530e+00 9.07601833e-01 1.17547616e-01
-2.75428325e-01 -5.59532106e-01 -6.57331407e-01 1.78629413e-01
1.58180162e-01 -5.27107656e-01 6.89320982e-01 -1.67945355e-01
-9.74552751e-01 5.33213019e-01 -3.02092787e-02 -2.66949177e-01
-3.64382565e-02 1.42639026e-01 -1.46287322e+00 -8.91197920e-02
4.51994598e-01 1.34226590e-01 -8.69827121e-02 -6.91924930e-01
-1.19793105e+00 -7.58944690e-01 2.73218095e-01 2.26175100e-01
-2.45282605e-01 3.27709466e-01 2.77143776e-01 -1.78303212e-01
-7.87818432e-02 -3.94118458e-01 -3.25087368e-01 -4.93378937e-02
1.68722391e-01 1.11496933e-01 1.25261867e+00 -5.70994139e-01
1.50788307e+00 -1.82610500e+00 -7.11987257e-01 7.07642257e-01
-3.29684973e-01 4.41236645e-01 3.71451288e-01 5.30724406e-01
3.03951561e-01 4.71372366e-01 -3.27586561e-01 1.01975441e+00
-8.42849910e-02 7.08649218e-01 -2.47952953e-01 -2.75304526e-01
-3.90764743e-01 -2.72941552e-02 -8.29134047e-01 -3.87209445e-01
5.66095114e-01 6.54702187e-01 2.21522465e-01 1.03956148e-01
-4.25898671e-01 8.76861736e-02 -6.35754347e-01 6.72362089e-01
7.02153385e-01 2.12516725e-01 3.99931610e-01 -2.03649595e-01
-6.67482972e-01 2.50690252e-01 -1.52896965e+00 1.12534201e+00
-5.95229268e-01 3.74912284e-02 2.04622060e-01 -6.99138641e-01
1.17334878e+00 6.53739691e-01 5.51658630e-01 -9.92910624e-01
-1.69665232e-01 4.07104641e-01 -4.74967182e-01 -1.01575398e+00
2.76839554e-01 -2.08154291e-01 4.77398932e-02 4.80189443e-01
-5.23540556e-01 -6.12146780e-02 6.23877347e-01 -2.45761350e-01
8.03084493e-01 3.44907731e-01 6.74812257e-01 -2.00225592e-01
6.53527200e-01 2.35570312e-01 7.30383158e-01 4.10989815e-05
-6.70437962e-02 -3.76485318e-01 -2.79459506e-01 -9.85740721e-01
-7.94392526e-01 -1.00873089e+00 -3.14599574e-01 9.30793464e-01
2.22246036e-01 -3.64210904e-01 -3.84048074e-01 -5.87616721e-03
6.47401735e-02 8.68963659e-01 -8.05310979e-02 2.40685999e-01
-8.62606391e-02 -7.75693536e-01 5.29976189e-01 3.79930168e-01
1.28878856e+00 -1.44296944e+00 -1.35412991e+00 3.94809693e-01
-5.34001231e-01 -9.07175779e-01 5.95991671e-01 -1.64527267e-01
-7.74445355e-01 -1.25476527e+00 1.25321671e-01 -2.95906246e-01
1.96725473e-01 3.00665110e-01 8.58711064e-01 3.54392588e-01
-4.47224408e-01 1.54887602e-01 -4.57317978e-01 -1.05172074e+00
-1.37795836e-01 1.48246614e-02 -2.45947111e-02 -5.15188277e-01
6.88939929e-01 -8.32614541e-01 -6.03741348e-01 3.56620610e-01
-1.30831409e+00 -9.80755761e-02 -4.16356837e-03 -4.05361891e-01
4.42848176e-01 7.41704404e-01 6.89990461e-01 -7.75234342e-01
6.79714382e-01 -8.11691701e-01 -6.69843793e-01 5.15702546e-01
-8.69018257e-01 -4.10742834e-02 3.17322016e-01 4.94450390e-01
-1.26362574e+00 -2.58175462e-01 6.65967464e-02 5.73167264e-01
-7.17608988e-01 7.77374387e-01 -5.21130860e-01 2.97286272e-01
3.62113118e-01 -2.80431688e-01 -1.16163276e-01 -3.98826182e-01
-6.17778744e-04 1.03461790e+00 5.45378447e-01 -2.26045415e-01
6.69890046e-01 5.99648058e-01 3.34886044e-01 -1.07121837e+00
-5.49144328e-01 -5.17421484e-01 -2.27208227e-01 -7.90955424e-01
1.03814399e+00 -3.90055597e-01 -8.54675233e-01 3.61564189e-01
-9.42055106e-01 1.07580855e-01 -3.07660121e-02 3.79913121e-01
-1.60190627e-01 -8.80617276e-02 1.20046988e-01 -1.00199902e+00
-6.37739003e-01 -3.55492055e-01 2.77101338e-01 6.17252111e-01
-2.27987498e-01 -8.61915052e-01 1.20955959e-01 1.71232596e-01
1.05683863e+00 6.49327636e-01 1.00116813e+00 8.89302988e-04
-3.04192036e-01 6.75011054e-02 -2.87807405e-01 3.89160849e-02
6.81841552e-01 2.68202931e-01 -7.39275992e-01 4.55134630e-01
-2.63467103e-01 4.34413552e-01 2.60741502e-01 1.22120772e-02
8.28373253e-01 -4.15545732e-01 -3.81084740e-01 4.10454035e-01
2.01792049e+00 6.29319787e-01 9.72877622e-01 9.07894850e-01
2.04520330e-01 7.53575981e-01 9.13541675e-01 6.71369493e-01
4.55876589e-01 3.48589420e-01 9.57106113e-01 1.26337349e-01
8.29798821e-03 3.04084755e-02 1.13916568e-01 2.89127082e-01
-7.68486023e-01 -1.52813345e-01 -1.00534761e+00 3.76272857e-01
-1.75010026e+00 -1.64929569e+00 -3.19721490e-01 2.34534717e+00
4.99288052e-01 -3.29062819e-01 7.20227882e-02 5.18726170e-01
7.79317498e-01 3.75303119e-01 -3.06045324e-01 -6.50453091e-01
-9.31724012e-02 6.17496967e-01 7.30447471e-01 4.11410749e-01
-8.30587745e-01 6.41713440e-01 5.41007328e+00 -1.51802003e-01
-1.26180887e+00 6.84479624e-02 -3.04949313e-01 2.55434096e-01
-4.37793970e-01 1.79767862e-01 -6.38289571e-01 4.80996400e-01
1.14364231e+00 -3.58235478e-01 7.21705973e-01 4.55320328e-01
8.56063843e-01 -5.49423397e-01 -3.09519414e-02 4.70677078e-01
-5.35218120e-01 -1.19708896e+00 -7.55041046e-03 -3.52735966e-02
6.69361725e-02 8.14253166e-02 -8.38038564e-01 -3.72677207e-01
5.13713062e-01 -6.20543599e-01 2.90924102e-01 8.94951940e-01
7.88016737e-01 -7.17149019e-01 6.20723844e-01 2.02239230e-01
-1.66452813e+00 -2.68368036e-01 -3.46700579e-01 -2.77441561e-01
4.67541039e-01 9.54527080e-01 -6.39960110e-01 8.30778420e-01
1.31628895e+00 3.31063986e-01 -4.79943678e-02 1.09589362e+00
-4.14763540e-01 4.14322764e-01 -9.49472189e-01 5.41080758e-02
-2.34695062e-01 -5.41933417e-01 3.09962243e-01 7.50620186e-01
6.53112173e-01 3.85784060e-01 -1.05241604e-01 3.63791764e-01
4.99161154e-01 3.21462363e-01 -7.14084029e-01 9.97270495e-02
8.41475964e-01 1.09195650e+00 -6.21596456e-01 -2.92283952e-01
-2.04077289e-01 2.94974655e-01 -3.87614220e-01 3.35500807e-01
-5.49520195e-01 -7.36297309e-01 1.01735055e+00 7.08519518e-01
-9.30164307e-02 -1.87090382e-01 9.09056664e-02 -7.30967402e-01
-5.00908587e-03 -4.06525522e-01 6.04261160e-01 -1.11111093e+00
-8.10495734e-01 2.93885171e-01 1.81940198e-01 -1.01513004e+00
-2.15001836e-01 -3.92320871e-01 -5.85704446e-01 9.31528151e-01
-1.78834081e+00 -1.01181328e+00 -9.46253181e-01 8.42921734e-01
-7.80886784e-02 3.15506458e-01 1.55342054e+00 4.64321941e-01
-2.48555839e-01 -7.25494087e-01 -1.98169723e-01 -2.21532688e-01
4.88178700e-01 -8.50205600e-01 -2.33765185e-01 8.96326423e-01
-4.00297284e-01 1.78397566e-01 5.23997664e-01 -9.38547611e-01
-7.47245371e-01 -1.03671420e+00 1.37016416e+00 5.07383227e-01
4.25999314e-01 4.15701240e-01 -5.85435331e-01 2.99413472e-01
1.47284359e-01 1.51322298e-02 8.04091930e-01 -2.07511738e-01
-2.47881621e-01 -8.53661954e-01 -1.70612025e+00 3.53828996e-01
8.70655417e-01 -2.31579617e-01 -3.13732415e-01 2.08566770e-01
2.75490135e-01 2.63994962e-01 -1.19711423e+00 5.10925412e-01
8.19181383e-01 -1.13262439e+00 7.48075366e-01 -5.37038982e-01
-1.47870404e-03 -8.94816935e-01 -4.30176318e-01 -1.16596878e+00
-3.75672936e-01 1.56059965e-01 2.75380552e-01 1.40243781e+00
2.42691144e-01 -9.45747077e-01 2.25380629e-01 8.01181495e-01
1.64444134e-01 1.37073770e-01 -7.75783598e-01 -3.71543467e-01
-8.57232809e-01 -4.73010719e-01 1.55495048e+00 8.93075347e-01
-1.10039522e-03 -1.22394666e-01 1.55134082e-01 5.56992888e-01
6.49028242e-01 -7.06894398e-02 2.97987431e-01 -1.96525633e+00
5.12018561e-01 -1.46666422e-01 -6.71651125e-01 2.28916317e-01
-4.49499965e-01 -5.68096101e-01 -4.01494771e-01 -2.38322353e+00
-4.15428191e-01 -7.52364099e-01 -3.94201875e-01 8.33503544e-01
2.34876603e-01 3.74437198e-02 2.84961879e-01 2.98985511e-01
7.40415454e-02 -2.46405657e-02 7.74745882e-01 5.90449609e-02
-2.03779731e-02 1.27839074e-01 -3.40328693e-01 6.82598114e-01
1.23314846e+00 -4.83474255e-01 -4.92492825e-01 -3.59828234e-01
9.63071585e-01 -9.51211229e-02 2.79567152e-01 -1.36281085e+00
1.05933718e-01 -9.16170061e-01 1.79133594e-01 -4.34804976e-01
-1.55937314e-01 -1.69162941e+00 9.98722792e-01 5.65795720e-01
2.85947859e-01 1.12545811e-01 -1.27282009e-01 1.51399791e-01
-1.91837087e-01 -2.80578882e-01 3.93781602e-01 -1.52886137e-01
-1.19889832e+00 2.18802169e-01 -6.30827725e-01 -3.83626401e-01
1.10382235e+00 -1.84708446e-01 -6.95623696e-01 -1.02937900e-01
-6.98092520e-01 3.58025104e-01 3.93383384e-01 4.83486205e-01
2.35756814e-01 -1.17214715e+00 -2.97969013e-01 3.16454619e-01
1.46598920e-01 -2.23773960e-02 8.70123580e-02 2.44760364e-01
-1.03565371e+00 3.12330216e-01 -7.00478733e-01 -1.25512434e-02
-1.20925105e+00 3.17290932e-01 2.33757123e-01 -1.47698179e-01
-2.62793362e-01 -1.09272644e-01 -6.93344474e-01 -5.24952292e-01
-5.84209077e-02 -4.29663926e-01 -7.51210034e-01 2.88224936e-01
8.22069287e-01 8.70855987e-01 2.52333969e-01 -5.77264667e-01
-7.65255094e-01 6.12656415e-01 8.32269073e-01 -1.73164591e-01
1.55439091e+00 -2.61964172e-01 -5.28007567e-01 3.30061316e-01
2.43369415e-01 -2.09701493e-01 -3.67308080e-01 3.91583797e-03
1.18066542e-01 -4.17228222e-01 1.25340357e-01 -1.22623146e+00
-1.26321793e+00 3.85333329e-01 8.02331448e-01 7.69531786e-01
1.54389513e+00 -3.39596808e-01 7.22652555e-01 3.78691792e-01
7.14275837e-01 -1.40740347e+00 -1.12862635e+00 2.48555213e-01
6.20792389e-01 -8.02160263e-01 2.45788366e-01 -6.52409017e-01
-1.91081345e-01 1.30829597e+00 1.55585647e-01 4.68068153e-01
1.27607000e+00 6.55283988e-01 3.50246906e-01 -3.67872179e-01
-3.47450167e-01 -6.06003404e-01 -5.72051108e-01 1.14666963e+00
2.25595236e-01 6.13859773e-01 -6.43507242e-01 1.31556749e-01
-2.46544421e-01 6.63128853e-01 1.62661716e-01 1.13238716e+00
-1.13963604e+00 -1.34800518e+00 -6.79909408e-01 6.36732042e-01
-2.03043908e-01 2.22064629e-01 -1.76444165e-02 5.83778381e-01
5.29659629e-01 1.34926796e+00 1.17414035e-01 -6.49376214e-02
6.28843248e-01 1.23852991e-01 -2.98929308e-02 -2.12716311e-01
-4.40690786e-01 -7.20164835e-01 4.43506837e-01 -5.53668678e-01
-1.07795334e+00 -3.36346209e-01 -1.60846615e+00 -7.25185573e-01
9.21473205e-02 2.23858863e-01 1.35898912e+00 9.11763847e-01
3.19701701e-01 3.41958016e-01 4.47589338e-01 -1.65356159e-01
-3.05290390e-02 -7.95016408e-01 -8.09413850e-01 4.63185549e-01
-2.62642413e-01 -7.68416226e-01 -2.39639014e-01 3.52472551e-02]
|
[9.162602424621582, 7.683679580688477]
|
e4c5d9b2-98b0-4d39-8ccf-db3f320e96b4
|
from-synthetic-to-real-unsupervised-domain
|
2103.14843
| null |
https://arxiv.org/abs/2103.14843v1
|
https://arxiv.org/pdf/2103.14843v1.pdf
|
From Synthetic to Real: Unsupervised Domain Adaptation for Animal Pose Estimation
|
Animal pose estimation is an important field that has received increasing attention in the recent years. The main challenge for this task is the lack of labeled data. Existing works circumvent this problem with pseudo labels generated from data of other easily accessible domains such as synthetic data. However, these pseudo labels are noisy even with consistency check or confidence-based filtering due to the domain shift in the data. To solve this problem, we design a multi-scale domain adaptation module (MDAM) to reduce the domain gap between the synthetic and real data. We further introduce an online coarse-to-fine pseudo label updating strategy. Specifically, we propose a self-distillation module in an inner coarse-update loop and a mean-teacher in an outer fine-update loop to generate new pseudo labels that gradually replace the old ones. Consequently, our model is able to learn from the old pseudo labels at the early stage, and gradually switch to the new pseudo labels to prevent overfitting in the later stage. We evaluate our approach on the TigDog and VisDA 2019 datasets, where we outperform existing approaches by a large margin. We also demonstrate the generalization ability of our model by testing extensively on both unseen domains and unseen animal categories. Our code is available at the project website.
|
['Gim Hee Lee', 'Chen Li']
|
2021-03-27
| null |
http://openaccess.thecvf.com//content/CVPR2021/html/Li_From_Synthetic_to_Real_Unsupervised_Domain_Adaptation_for_Animal_Pose_CVPR_2021_paper.html
|
http://openaccess.thecvf.com//content/CVPR2021/papers/Li_From_Synthetic_to_Real_Unsupervised_Domain_Adaptation_for_Animal_Pose_CVPR_2021_paper.pdf
|
cvpr-2021-1
|
['animal-pose-estimation']
|
['computer-vision']
|
[ 1.92283809e-01 -6.29274398e-02 7.18712881e-02 -6.22212112e-01
-5.84661663e-01 -6.94433272e-01 4.94375408e-01 1.76618487e-01
-7.80393124e-01 1.00879002e+00 -2.10591376e-01 2.39210322e-01
8.32313001e-02 -6.43383265e-01 -1.00574374e+00 -6.13204122e-01
1.50782809e-01 8.36564541e-01 6.81081355e-01 -8.33879560e-02
-6.01076223e-02 1.39586344e-01 -1.45277417e+00 5.04692970e-03
1.08492732e+00 9.22517419e-01 3.26998323e-01 1.27822280e-01
6.59855753e-02 3.93480867e-01 -4.44609106e-01 -3.26052994e-01
3.61318856e-01 -3.42034340e-01 -6.88296437e-01 2.33252272e-01
3.03245068e-01 -4.28724468e-01 -7.00548515e-02 1.15290344e+00
4.70656753e-01 1.48697853e-01 6.46521032e-01 -1.34783971e+00
-4.30153549e-01 4.10117745e-01 -7.08925009e-01 -8.78141597e-02
4.68850769e-02 2.50680685e-01 6.41604781e-01 -7.28747845e-01
7.10165739e-01 1.33736515e+00 6.61693215e-01 6.96141064e-01
-1.62711978e+00 -8.93385172e-01 3.60083282e-01 2.43411168e-01
-1.29301202e+00 -1.22834094e-01 7.83154905e-01 -5.92954993e-01
1.40402541e-01 -2.12578684e-01 4.87515956e-01 1.47947228e+00
-2.04119563e-01 6.84803069e-01 1.27298164e+00 -1.43384442e-01
4.96273309e-01 1.48783937e-01 1.39739942e-02 4.92451459e-01
3.39362085e-01 2.75562912e-01 -2.52322912e-01 -2.62711346e-01
7.29206979e-01 -2.62739677e-02 -5.60033992e-02 -9.07276154e-01
-1.16700184e+00 8.42423141e-01 6.18694127e-01 -1.06183529e-01
-3.00541669e-01 -6.77067637e-02 2.66097993e-01 2.78069019e-01
5.50053179e-01 4.79949474e-01 -7.99030304e-01 1.66612908e-01
-6.34269714e-01 3.98539662e-01 6.53677523e-01 9.22913909e-01
8.00577462e-01 -4.03909177e-01 -6.94271475e-02 1.21486521e+00
4.13565904e-01 4.46842700e-01 5.51411569e-01 -8.15782666e-01
3.34955096e-01 5.70716083e-01 3.18173796e-01 -7.78532088e-01
-4.19303209e-01 -7.02791333e-01 -7.11862206e-01 1.95731193e-01
6.36881411e-01 -2.39678115e-01 -1.05261111e+00 2.18929076e+00
8.88665497e-01 3.40037525e-01 -1.45374328e-01 9.48895991e-01
7.20932484e-01 5.09522915e-01 3.49030614e-01 6.75992742e-02
1.30999029e+00 -1.10465074e+00 -3.42182398e-01 -7.35009372e-01
5.66735685e-01 -5.06603956e-01 9.46934044e-01 2.16339111e-01
-4.86987561e-01 -6.82470143e-01 -1.00750613e+00 1.72407683e-02
-2.45113641e-01 3.23773265e-01 2.74108171e-01 4.93804105e-02
-4.75985259e-01 6.26768887e-01 -8.46590400e-01 -4.65675235e-01
5.04228830e-01 1.69927821e-01 -5.90649486e-01 -8.83929357e-02
-1.19036293e+00 7.78283536e-01 7.40242422e-01 8.28080028e-02
-9.40099895e-01 -5.59959650e-01 -8.77065897e-01 -3.96494985e-01
6.34276271e-01 -4.90530998e-01 1.36888206e+00 -9.31306303e-01
-1.31794345e+00 9.90085781e-01 2.75873244e-01 -4.21164095e-01
8.29834163e-01 -2.40276054e-01 -2.07998320e-01 -1.67795673e-01
1.88560724e-01 1.10502231e+00 9.49239016e-01 -1.39739680e+00
-7.51878440e-01 -4.30755436e-01 9.39525291e-02 2.47348785e-01
-1.12496883e-01 -4.36101019e-01 -5.16937077e-01 -9.38741446e-01
2.57428050e-01 -1.35662055e+00 -2.92405725e-01 3.28444004e-01
-5.37046418e-02 -1.21240027e-01 6.52767301e-01 -5.40955186e-01
9.22167659e-01 -2.17726636e+00 2.71068841e-01 -1.56898141e-01
1.93670988e-01 3.31840038e-01 -4.35147882e-01 2.88817603e-02
-1.57865271e-01 -4.62920219e-01 -5.83779454e-01 -3.27779412e-01
-7.73087442e-02 4.00968969e-01 -5.98500781e-02 4.57281023e-01
3.28738242e-01 6.42914236e-01 -1.05578566e+00 -4.93029326e-01
-8.39654636e-03 1.64557829e-01 -7.15495646e-01 3.33476514e-01
-4.59816873e-01 9.27964330e-01 -4.64050114e-01 3.95236343e-01
9.90823448e-01 -3.32788795e-01 1.08892553e-01 -9.81450081e-02
2.24877357e-01 7.48490691e-02 -1.34998524e+00 1.79025674e+00
-3.50054622e-01 1.90738559e-01 5.55994846e-02 -9.28235888e-01
8.88981640e-01 -2.93620955e-02 3.92564029e-01 -5.84847391e-01
1.78512543e-01 3.37654293e-01 8.84863175e-03 -2.93007314e-01
3.28581721e-01 -1.60137668e-01 -1.76208541e-01 2.77557313e-01
1.21714883e-01 -1.48688704e-01 2.17044681e-01 -1.27135023e-01
8.81469965e-01 5.91866553e-01 1.95311323e-01 -1.03282921e-01
5.86329341e-01 2.40534380e-01 9.38408971e-01 3.88401330e-01
-3.62194270e-01 6.24977231e-01 4.13515598e-01 -5.10287464e-01
-9.98208463e-01 -9.14373398e-01 -3.63756120e-01 1.14930189e+00
3.49659234e-01 -7.28416964e-02 -8.73184025e-01 -1.14933515e+00
2.71375179e-01 5.13308764e-01 -7.45605648e-01 -3.66887689e-01
-4.62272018e-01 -7.53351331e-01 1.82381645e-01 5.46515882e-01
6.76791906e-01 -1.17189014e+00 -4.82749820e-01 3.46074849e-01
-3.54654759e-01 -1.17868876e+00 -4.20959413e-01 3.86249185e-01
-7.97397494e-01 -9.81300533e-01 -8.06545377e-01 -9.07139480e-01
9.12060857e-01 -7.21307173e-02 8.48546267e-01 -3.23979169e-01
-1.40062556e-01 -2.89654821e-01 -3.86141241e-01 -1.75095722e-01
-5.37141621e-01 2.47839615e-01 6.23476431e-02 -5.00157662e-02
4.20886308e-01 -4.29823875e-01 -5.27441323e-01 7.50575185e-01
-9.44613218e-01 2.21215039e-01 5.62824905e-01 1.06261826e+00
8.63140643e-01 -8.36111754e-02 7.09282458e-01 -1.02256119e+00
2.29442880e-01 -6.05855882e-01 -1.00162184e+00 7.08339289e-02
-2.81899542e-01 4.79687423e-01 6.50780737e-01 -8.48726511e-01
-9.12936509e-01 4.22177106e-01 -2.67308056e-01 -2.66895503e-01
-3.36030424e-01 2.28977233e-01 -2.12518245e-01 -1.65090745e-03
8.06758225e-01 -1.33906171e-01 -1.04489274e-01 -8.64293337e-01
1.16390228e-01 5.49732268e-01 5.69167435e-01 -7.24587977e-01
1.03422701e+00 2.43408352e-01 -1.48921907e-01 -3.35589677e-01
-1.05680382e+00 -2.27214858e-01 -6.63749874e-01 1.00231757e-02
6.76013231e-01 -1.16901445e+00 -3.64021122e-01 7.99036324e-01
-9.66385841e-01 -5.61004996e-01 -3.45033973e-01 5.52653551e-01
-4.66814399e-01 3.19665253e-01 -4.74993408e-01 -1.96275041e-01
-4.72601131e-02 -1.17131388e+00 1.18433368e+00 2.14516252e-01
-1.90677077e-01 -6.63258135e-01 1.90827891e-01 2.19250724e-01
1.66952178e-01 4.34931278e-01 6.99986696e-01 -8.33115220e-01
-2.67564952e-01 -1.78202376e-01 -2.72767365e-01 5.56331158e-01
1.23591572e-01 -5.22117198e-01 -8.03187847e-01 -5.34733593e-01
-7.80753568e-02 -6.61217749e-01 7.46339560e-01 1.22868516e-01
1.08270478e+00 2.04185899e-02 -4.42233622e-01 5.79188347e-01
1.09350467e+00 -3.67830880e-03 9.57554355e-02 4.97119755e-01
6.13460720e-01 6.69211566e-01 1.03747177e+00 3.85831863e-01
4.77733165e-01 8.01385939e-01 3.91884446e-01 6.40591374e-03
-1.11804947e-01 -6.21107936e-01 4.23094146e-02 4.55202729e-01
3.69491518e-01 -1.24095589e-01 -8.66355956e-01 4.38806146e-01
-1.93336904e+00 -4.66588974e-01 1.02798425e-01 2.36172032e+00
1.10593009e+00 3.16467762e-01 1.79343179e-01 -1.28640145e-01
8.96140039e-01 -5.41845448e-02 -8.78902376e-01 2.22603351e-01
1.68474376e-01 2.17534951e-03 4.97388095e-01 2.72792310e-01
-1.49905992e+00 1.01434302e+00 5.54947615e+00 8.95264328e-01
-9.45824921e-01 4.72491272e-02 4.86521661e-01 2.25434050e-01
9.69401672e-02 -3.14343441e-03 -8.12136292e-01 7.47085869e-01
4.55926031e-01 9.40138921e-02 3.35407287e-01 1.19970846e+00
-1.08198039e-01 -2.81331927e-01 -1.35474503e+00 8.63757849e-01
-1.23029545e-01 -6.22981727e-01 -1.84904516e-01 -7.67653137e-02
7.05899119e-01 5.49461246e-02 -2.20792294e-01 4.89731461e-01
4.00392562e-01 -4.84428525e-01 8.02220166e-01 2.16544151e-01
8.10184479e-01 -5.68194926e-01 6.61447167e-01 7.25541115e-01
-1.12032580e+00 -6.97288364e-02 -4.79062408e-01 1.29616231e-01
-3.42481025e-02 4.48932141e-01 -7.29585648e-01 3.72511387e-01
6.56504571e-01 6.18248641e-01 -8.57324660e-01 1.19872224e+00
-4.75788116e-01 2.95084596e-01 -5.78757524e-01 1.66791022e-01
1.82105806e-02 -2.11675659e-01 4.13843930e-01 7.68184900e-01
1.81543812e-01 -7.61359408e-02 4.13842887e-01 7.60962486e-01
-1.58244550e-01 -1.24865972e-01 -2.71130294e-01 2.20474690e-01
5.57412326e-01 1.16779339e+00 -6.75980806e-01 -2.21075952e-01
-1.86324984e-01 1.02588439e+00 4.73270625e-01 1.61027536e-01
-1.08702087e+00 -2.92026311e-01 4.88351375e-01 1.30648941e-01
2.83637643e-01 -7.10225925e-02 5.98393157e-02 -1.25127327e+00
1.34574726e-01 -9.87667501e-01 4.04691607e-01 -5.75159132e-01
-1.40219223e+00 6.67091668e-01 2.53560156e-01 -1.56947136e+00
-2.68444240e-01 -4.15366322e-01 5.44917099e-02 7.19722569e-01
-1.52515149e+00 -9.54671621e-01 -5.63861728e-01 1.82617918e-01
5.86706281e-01 1.14428185e-01 6.46022797e-01 5.37360251e-01
-4.90176409e-01 5.91687620e-01 2.53797501e-01 8.96650255e-02
1.08834159e+00 -1.07262826e+00 5.17991781e-01 6.08742774e-01
-1.72998533e-01 2.41812438e-01 8.26249242e-01 -7.10353434e-01
-6.56583190e-01 -1.28484309e+00 5.76178432e-01 -3.56394231e-01
4.84653413e-01 -6.54594362e-01 -1.05888999e+00 6.52059734e-01
-3.89319897e-01 3.68014455e-01 2.31900916e-01 -1.57165244e-01
-3.23717058e-01 -1.92920491e-01 -1.37305379e+00 3.20283532e-01
1.10185087e+00 -9.07022059e-02 -7.37219989e-01 2.54919559e-01
6.48189306e-01 -7.20246196e-01 -5.63725412e-01 7.27337956e-01
6.02415025e-01 -5.47641098e-01 7.72834539e-01 -3.31357747e-01
5.02132773e-01 -6.69983983e-01 5.60722947e-02 -1.63380659e+00
-3.27564180e-01 -1.24318659e-01 1.22274384e-01 1.28317201e+00
3.32511723e-01 -7.45424271e-01 7.73124754e-01 3.36159587e-01
1.26737714e-01 -5.44797778e-01 -8.55852962e-01 -9.76383448e-01
3.58914137e-02 -1.07704505e-01 6.03622496e-01 8.16768050e-01
-5.36981881e-01 5.10692179e-01 -3.97515953e-01 1.46715313e-01
8.31418037e-01 1.77184612e-01 1.03629839e+00 -1.54753625e+00
-4.63748425e-01 7.59714097e-02 -3.19810361e-01 -1.26509428e+00
1.36208281e-01 -7.09784627e-01 4.17676985e-01 -1.28230500e+00
2.02316761e-01 -7.13448882e-01 -6.74864128e-02 6.67685628e-01
-3.30930889e-01 2.87695944e-01 -3.39556858e-03 9.05417949e-02
-6.16436303e-01 8.21796119e-01 1.40410399e+00 -3.16356570e-02
-1.15748666e-01 4.08100076e-02 -4.17329997e-01 9.10311878e-01
5.50837338e-01 -9.20937657e-01 -4.20067817e-01 -4.71705019e-01
-1.57606289e-01 -1.63748428e-01 4.20935571e-01 -1.21289611e+00
-1.36564776e-01 -1.57356262e-01 4.17565465e-01 -5.26194870e-01
2.53510475e-01 -1.10393572e+00 2.59394675e-01 5.15230536e-01
-3.25566739e-01 -1.85991213e-01 2.48913229e-01 7.49762237e-01
-1.26427621e-01 -1.84945181e-01 1.21501720e+00 8.32976475e-02
-7.01671600e-01 4.27147269e-01 1.13294512e-01 3.15566629e-01
1.25878203e+00 4.50446717e-02 -1.81892723e-01 4.70690466e-02
-8.40517759e-01 5.98846793e-01 7.38949180e-01 6.82343543e-01
1.12619445e-01 -1.40763664e+00 -6.24958038e-01 4.40989733e-01
4.89767849e-01 4.36904490e-01 2.22567424e-01 4.91993874e-01
-5.33411324e-01 -6.63667098e-02 -3.53349417e-01 -7.34922171e-01
-1.05555141e+00 6.49274051e-01 2.52976686e-01 -3.16028386e-01
-3.86599302e-01 8.21884513e-01 4.92904156e-01 -7.51209438e-01
3.93682390e-01 -3.40155363e-01 -2.00548157e-01 7.26242214e-02
3.02429557e-01 1.87716484e-01 -7.20392987e-02 -6.72831357e-01
-4.69958514e-01 6.94080889e-01 -1.59869403e-01 7.56212175e-02
1.45647061e+00 -1.20700903e-01 2.53903955e-01 3.55892509e-01
1.05652344e+00 -2.73488849e-01 -1.67578149e+00 -5.44387221e-01
1.20258808e-01 -5.14818549e-01 -3.54069144e-01 -8.07849288e-01
-8.72409225e-01 6.63450658e-01 8.91207814e-01 -8.94591585e-02
1.08023930e+00 7.61309192e-02 8.08269083e-01 3.29055548e-01
4.70463306e-01 -1.30102932e+00 1.52487084e-01 5.17470956e-01
8.46144617e-01 -1.53584957e+00 -8.26768428e-02 -3.93651247e-01
-7.00698316e-01 6.20484114e-01 9.99250293e-01 -1.58574715e-01
4.66362983e-01 5.70656769e-02 -2.20078342e-02 1.57630861e-01
-5.83976030e-01 -2.32579872e-01 1.26079708e-01 5.88165641e-01
1.50879055e-01 -3.46052926e-03 -4.50830370e-01 6.37713075e-01
-1.74976699e-02 2.18352765e-01 7.56539330e-02 9.42919314e-01
-3.44278336e-01 -1.40434515e+00 -4.23268199e-01 2.32159376e-01
-1.56977415e-01 2.85068333e-01 -1.40852854e-01 6.66011572e-01
4.62273657e-01 6.51108980e-01 -9.90460366e-02 -3.40974331e-01
5.90121567e-01 5.20467311e-02 3.46782833e-01 -7.88965881e-01
-2.25153103e-01 8.22048262e-02 -7.56530315e-02 -2.92854607e-01
-3.09704006e-01 -6.09108269e-01 -1.10608137e+00 -1.81454867e-02
-4.29925591e-01 1.73107386e-01 4.65357482e-01 8.30527186e-01
3.65110338e-01 5.06002486e-01 6.15888715e-01 -8.05523932e-01
-8.42074692e-01 -1.17263341e+00 -4.73165363e-01 7.34385073e-01
2.96058893e-01 -1.08220565e+00 -3.25948715e-01 8.17069039e-02]
|
[9.388365745544434, 1.30251145362854]
|
868ef9e4-3399-4ebe-9bfc-aaf9f6ea47ce
|
cohs-cqg-context-and-history-selection-for
|
2209.06652
| null |
https://arxiv.org/abs/2209.06652v2
|
https://arxiv.org/pdf/2209.06652v2.pdf
|
CoHS-CQG: Context and History Selection for Conversational Question Generation
|
Conversational question generation (CQG) serves as a vital task for machines to assist humans, such as interactive reading comprehension, through conversations. Compared to traditional single-turn question generation (SQG), CQG is more challenging in the sense that the generated question is required not only to be meaningful, but also to align with the occurred conversation history. While previous studies mainly focus on how to model the flow and alignment of the conversation, there has been no thorough study to date on which parts of the context and history are necessary for the model. We argue that shortening the context and history is crucial as it can help the model to optimise more on the conversational alignment property. To this end, we propose CoHS-CQG, a two-stage CQG framework, which adopts a CoHS module to shorten the context and history of the input. In particular, CoHS selects contiguous sentences and history turns according to their relevance scores by a top-p strategy. Our model achieves state-of-the-art performances on CoQA in both the answer-aware and answer-unaware settings.
|
['Ai Ti Aw', 'Shafiq Joty', 'Nancy F. Chen', 'Liangming Pan', 'Bowei Zou', 'Xuan Long Do']
|
2022-09-14
| null |
https://aclanthology.org/2022.coling-1.48
|
https://aclanthology.org/2022.coling-1.48.pdf
|
coling-2022-10
|
['question-generation']
|
['natural-language-processing']
|
[ 3.33990365e-01 4.82245833e-01 2.25853890e-01 -4.32614744e-01
-8.67244840e-01 -7.15535164e-01 7.12683141e-01 2.88447648e-01
-1.30752474e-01 7.31669068e-01 7.76261628e-01 -7.28853643e-01
1.22060284e-01 -7.80214727e-01 -3.32624674e-01 -2.52770036e-01
4.12675053e-01 6.07687950e-01 4.46003407e-01 -7.37330139e-01
5.16489327e-01 -2.13063344e-01 -1.49741089e+00 5.67367196e-01
1.38615882e+00 6.35766804e-01 4.62472171e-01 9.69736636e-01
-6.09495282e-01 1.02279544e+00 -7.91286886e-01 -5.88931620e-01
-2.65945703e-01 -1.22414184e+00 -1.50990212e+00 -1.05046630e-01
1.09575272e-01 -2.28463486e-01 7.02769235e-02 6.47619307e-01
5.50178587e-01 3.39708239e-01 5.34178138e-01 -9.90327179e-01
-3.48953754e-01 8.70553315e-01 9.64767262e-02 2.92263776e-01
8.67016375e-01 2.51232088e-01 1.37503242e+00 -5.25405586e-01
5.68677366e-01 1.34653831e+00 1.57376215e-01 7.40881741e-01
-7.40746379e-01 -2.26647854e-01 3.78388911e-01 5.55877507e-01
-6.19528770e-01 -5.05016327e-01 7.88295388e-01 -1.73185512e-01
8.48951876e-01 5.99719346e-01 7.28018284e-01 8.41018856e-01
1.71815529e-01 8.53464663e-01 8.20126951e-01 -6.59631610e-01
1.91501305e-01 -1.61767170e-01 1.24151863e-01 2.68515885e-01
-4.78579462e-01 -6.30600929e-01 -6.97024643e-01 -1.03787847e-01
2.42075086e-01 -5.89225233e-01 -4.88384664e-01 7.70339817e-02
-1.03261971e+00 9.34544265e-01 1.28451541e-01 4.61596817e-01
-4.71326500e-01 -1.97876200e-01 2.80775219e-01 4.40858036e-01
3.37978840e-01 9.66949284e-01 -2.55206496e-01 -6.68529689e-01
-5.96778870e-01 6.40515864e-01 1.26065958e+00 9.15319443e-01
5.22226989e-01 -5.78356802e-01 -7.92123854e-01 8.03325951e-01
1.59949034e-01 1.86952427e-01 2.83863455e-01 -1.32846618e+00
7.27033436e-01 8.94366920e-01 4.05171365e-01 -9.12258685e-01
-1.82668582e-01 -1.46856651e-01 -5.71336329e-01 -6.30796731e-01
6.52224183e-01 -1.34201363e-01 -2.62242764e-01 1.84282625e+00
5.02728403e-01 -1.09382287e-01 9.76623669e-02 8.98148715e-01
1.01634884e+00 8.58683407e-01 -9.42637548e-02 -3.17601621e-01
1.67050636e+00 -1.36179149e+00 -1.00892758e+00 -4.01421964e-01
6.65764034e-01 -1.04536748e+00 1.37056029e+00 1.32145837e-01
-1.32706559e+00 -4.37489331e-01 -6.60016060e-01 -4.65201855e-01
1.91756740e-01 -2.43431121e-01 3.22342873e-01 3.18412274e-01
-9.46681082e-01 2.19167069e-01 -2.98947394e-01 -3.17226976e-01
-1.43036276e-01 -2.48812318e-01 1.99409619e-01 -1.46443203e-01
-1.67607999e+00 8.90611947e-01 1.12753578e-01 3.03580701e-01
-4.20661300e-01 -4.24810380e-01 -7.46503234e-01 1.77474931e-01
5.49748361e-01 -9.86169696e-01 1.98423445e+00 -7.81263471e-01
-1.84165788e+00 5.44991553e-01 -6.73878253e-01 -4.33199674e-01
5.25768876e-01 -2.24917844e-01 5.40727116e-02 3.90996635e-01
-9.16007347e-03 5.93941987e-01 6.06555104e-01 -9.63055789e-01
-6.63817525e-01 -1.16760172e-01 6.19564652e-01 7.82833099e-01
1.02351017e-01 -7.33986422e-02 -5.69302619e-01 -2.93579519e-01
1.50992960e-01 -8.13259661e-01 -2.49679804e-01 -5.13200521e-01
-4.68256205e-01 -7.36403883e-01 4.46155101e-01 -9.29664671e-01
1.71877313e+00 -1.66187489e+00 2.97520548e-01 -1.38922557e-01
7.09138587e-02 2.32330218e-01 -3.07854384e-01 1.04288721e+00
4.01272207e-01 9.80710145e-03 -2.16334507e-01 -4.53856558e-01
1.01594634e-01 1.32765681e-01 -4.25697058e-01 -2.24216625e-01
2.13676080e-01 1.10715318e+00 -1.23598838e+00 -5.56980669e-01
-1.72528103e-01 -8.42226669e-03 -6.32093847e-01 9.07778263e-01
-7.78369069e-01 7.85328448e-01 -4.89714593e-01 2.38619283e-01
3.53949100e-01 -3.66660655e-01 3.44253272e-01 1.72350779e-01
-4.33898065e-03 1.09898758e+00 -6.66317105e-01 1.52474213e+00
-5.80604434e-01 4.11539555e-01 -3.31562869e-02 -5.99422157e-01
8.08493018e-01 4.49426979e-01 -7.05354661e-03 -8.27503502e-01
-5.01555726e-02 1.28698006e-01 2.58219540e-01 -7.88660228e-01
9.37255859e-01 6.37904461e-03 -1.82078362e-01 8.38025212e-01
-3.76365721e-01 -5.61841369e-01 4.52728510e-01 6.47749007e-01
9.72128749e-01 -2.19590906e-02 3.94708663e-01 -1.87967718e-02
9.24388051e-01 1.08086746e-02 3.28119755e-01 7.97628582e-01
-1.84087202e-01 5.87569594e-01 9.51176226e-01 1.24831228e-02
-7.97298312e-01 -5.58702290e-01 4.56469446e-01 1.21920979e+00
2.00789720e-01 -6.48904681e-01 -1.19913137e+00 -7.04643726e-01
-5.17052591e-01 1.12622702e+00 -3.73068869e-01 -6.80862889e-02
-9.33091819e-01 -9.75716487e-02 3.57235372e-01 1.92386389e-01
5.11507392e-01 -1.51278102e+00 -6.36239886e-01 5.28318882e-01
-1.13322616e+00 -1.02491701e+00 -8.85323226e-01 -3.93649161e-01
-5.94197929e-01 -1.21729767e+00 -4.08119738e-01 -7.12696552e-01
3.89189392e-01 4.88274783e-01 1.48954809e+00 4.97472346e-01
5.22318423e-01 3.16443175e-01 -9.98344243e-01 -2.30852813e-01
-8.12568724e-01 6.11293197e-01 -7.10198224e-01 -1.25660732e-01
3.45219940e-01 -5.26875794e-01 -8.98316920e-01 4.82484639e-01
-8.18864346e-01 5.29092968e-01 3.95978272e-01 8.01311493e-01
1.49610966e-01 -5.09240985e-01 9.63341415e-01 -1.03854072e+00
1.26657021e+00 -5.11018634e-01 -1.25553861e-01 5.16368210e-01
-3.81729394e-01 -5.51665127e-02 7.51849771e-01 -1.07638463e-01
-1.15884328e+00 -5.13796270e-01 -5.72098911e-01 4.53255147e-01
1.09455012e-01 5.94820142e-01 -2.34366044e-01 5.73324502e-01
4.03047025e-01 2.62799859e-01 7.32395872e-02 -2.70706922e-01
5.91633797e-01 8.62040401e-01 2.84330994e-01 -6.52580440e-01
3.52877885e-01 -7.90449604e-02 -4.40337598e-01 -6.71884894e-01
-1.03865600e+00 -7.19780385e-01 -4.82592493e-01 -4.43784654e-01
8.16526711e-01 -4.55097586e-01 -8.53320479e-01 4.43340093e-01
-1.62761712e+00 -4.39257860e-01 -3.56681854e-03 -3.72526720e-02
-5.62209249e-01 6.01726890e-01 -6.02349877e-01 -1.07416320e+00
-6.61354125e-01 -1.11860156e+00 8.77221107e-01 4.24023360e-01
-5.93479991e-01 -8.78243744e-01 1.71004329e-02 8.33731234e-01
5.32381177e-01 -1.74221784e-01 1.09832966e+00 -6.34534955e-01
-6.77839279e-01 1.48088232e-01 -8.29318613e-02 2.22048163e-01
1.23884387e-01 -2.30077818e-01 -8.06321561e-01 3.95674333e-02
1.74609289e-01 -3.91075999e-01 5.47625899e-01 -4.11640815e-02
9.06131923e-01 -6.80355489e-01 9.95941684e-02 -9.27153453e-02
6.82449222e-01 3.54976624e-01 8.98854077e-01 9.90746841e-02
4.20879632e-01 1.11066651e+00 7.49887764e-01 2.62622356e-01
1.09614480e+00 7.30442166e-01 3.78407419e-01 3.13640594e-01
-5.14285602e-02 -4.81581390e-01 1.99041680e-01 1.39061582e+00
1.75935835e-01 -5.55969059e-01 -8.01572263e-01 6.38103366e-01
-1.99671650e+00 -1.00926411e+00 -4.38346714e-01 1.82389617e+00
1.14358509e+00 5.35001382e-02 4.71351035e-02 1.08302370e-01
5.84438980e-01 4.63869631e-01 -3.60084265e-01 -7.26400912e-01
2.69533694e-02 7.56134987e-02 -5.00631392e-01 8.62031877e-01
-4.68973011e-01 9.89397168e-01 5.79394484e+00 5.81346571e-01
-7.25142598e-01 -6.50519179e-03 6.11525178e-01 2.79967010e-01
-7.96185493e-01 2.64431983e-01 -5.93741119e-01 4.63655770e-01
8.63126636e-01 -1.97112963e-01 4.92807597e-01 4.90831286e-01
4.95110184e-01 -4.74300086e-01 -1.10434854e+00 4.85433042e-01
1.07626095e-01 -1.19576919e+00 1.86539069e-01 -3.05844128e-01
5.02829492e-01 -6.63537025e-01 -4.43149894e-01 5.07677555e-01
-3.23782973e-02 -9.13101494e-01 7.22047985e-01 5.27096629e-01
2.42052048e-01 -6.28374159e-01 9.32004392e-01 7.78511643e-01
-1.17598474e+00 1.11872125e-02 2.14712452e-02 -3.03755075e-01
6.06665254e-01 3.78013372e-01 -1.15282917e+00 7.26270378e-01
2.63302147e-01 1.38919458e-01 -4.46846902e-01 7.44061291e-01
-7.43081629e-01 9.77231324e-01 1.09489188e-01 -5.20755053e-01
3.29438359e-01 -3.95339072e-01 5.90759456e-01 9.99248326e-01
1.54362097e-01 4.65750009e-01 1.39405683e-01 6.12309098e-01
-1.59558937e-01 3.14036161e-01 -1.39947191e-01 -6.38731867e-02
8.40085983e-01 9.71271038e-01 -2.65669018e-01 -2.46051162e-01
-1.96412593e-01 9.54291344e-01 3.51333469e-01 3.17263216e-01
-5.20358503e-01 -4.24826592e-01 4.17157263e-01 2.47509554e-01
2.12765917e-01 -1.73229054e-01 -1.73622623e-01 -9.84391809e-01
3.09813231e-01 -1.32709420e+00 3.80811304e-01 -7.47105062e-01
-1.07784200e+00 7.73740172e-01 -2.08757609e-01 -9.64023709e-01
-8.36528778e-01 1.07062519e-01 -9.56030846e-01 1.11763489e+00
-1.70336080e+00 -1.01853597e+00 -3.90421152e-01 1.81211233e-01
8.43109667e-01 4.73205626e-01 7.69539237e-01 -7.96530619e-02
-3.42673808e-01 4.07436877e-01 -4.72973526e-01 -1.45963263e-02
7.35413432e-01 -1.22552431e+00 6.45288885e-01 8.71541679e-01
-1.76818982e-01 7.91674554e-01 9.92862761e-01 -4.88235742e-01
-1.24290597e+00 -8.21225584e-01 1.65739524e+00 -4.75289673e-01
4.63998258e-01 -2.52469093e-01 -1.21666718e+00 2.33100340e-01
7.01605380e-01 -7.36082911e-01 6.96690857e-01 1.92015484e-01
-7.07594976e-02 -9.11051873e-03 -7.79658735e-01 6.95410669e-01
9.96383607e-01 -6.62533402e-01 -1.05739510e+00 9.63439792e-02
1.10908890e+00 -6.05423272e-01 -5.58007836e-01 5.68988398e-02
3.89019817e-01 -1.13381231e+00 4.52616841e-01 -5.31079888e-01
7.78147995e-01 -3.89685363e-01 6.91453069e-02 -1.43717122e+00
-2.18188912e-02 -1.06817222e+00 -3.30814779e-01 1.34489930e+00
5.41904569e-01 -5.12495160e-01 5.35986960e-01 5.35179198e-01
-4.68290865e-01 -1.05953658e+00 -9.13858414e-01 -2.38050863e-01
6.81017525e-04 -3.36516947e-01 9.29003835e-01 5.55835307e-01
2.71788746e-01 8.75724375e-01 -2.27555111e-01 -1.74885899e-01
-4.32741120e-02 4.29304123e-01 1.01579309e+00 -9.61623251e-01
-2.93695897e-01 -5.17665446e-01 4.75315243e-01 -1.74530828e+00
3.43890265e-02 -4.52556401e-01 4.68338758e-01 -2.09815717e+00
1.12058967e-02 -3.51229876e-01 3.11399996e-01 -1.79064609e-02
-8.45145762e-01 -2.40313604e-01 3.92233163e-01 2.04216525e-01
-8.87150764e-01 7.73857772e-01 1.73629379e+00 7.40800649e-02
-3.44591886e-01 2.68937290e-01 -1.04134297e+00 2.67402858e-01
7.89903224e-01 -2.71833576e-02 -6.73825204e-01 -4.08489019e-01
4.86630589e-01 5.89642525e-01 -8.58651549e-02 -4.44643021e-01
4.07304019e-01 -4.01247084e-01 -4.22378391e-01 -7.95266926e-01
2.29026273e-01 -8.69106576e-02 -1.88417166e-01 1.67504743e-01
-6.40260160e-01 3.19835752e-01 -2.28064150e-01 3.42855930e-01
-4.33848530e-01 -4.49766487e-01 3.62562656e-01 -3.08749646e-01
-3.90559584e-01 -5.69959357e-02 -5.65875828e-01 5.56784451e-01
6.61490560e-01 -3.51817794e-02 -4.97545719e-01 -1.06316555e+00
-2.16106862e-01 7.60817647e-01 2.94960648e-01 5.47709644e-01
4.50886816e-01 -9.96984363e-01 -7.68651664e-01 -2.21480235e-01
2.78069228e-01 4.72907156e-01 3.93873751e-01 7.06285894e-01
-3.79397392e-01 6.74881876e-01 2.59150594e-01 -4.22885358e-01
-1.15392578e+00 9.44892988e-02 3.12675625e-01 -7.70310998e-01
-2.54255712e-01 8.26011360e-01 -3.78537178e-03 -4.66768563e-01
2.51718551e-01 -4.51055020e-01 -6.56552434e-01 2.70232886e-01
8.73330235e-01 1.85721904e-01 9.00230259e-02 -3.55037957e-01
1.08505255e-02 1.26988426e-01 -5.90181760e-02 -3.19347441e-01
8.64670157e-01 -6.47610545e-01 -4.18131888e-01 4.47741777e-01
6.73681140e-01 5.62572554e-02 -1.06740570e+00 -2.22060204e-01
2.45253488e-01 -2.56151229e-01 -5.47793627e-01 -8.73145878e-01
-2.71132469e-01 7.89418161e-01 -3.99264485e-01 6.97245419e-01
1.00644326e+00 7.02808872e-02 1.32828641e+00 3.75296563e-01
2.77411550e-01 -9.68768656e-01 4.63674754e-01 1.05175054e+00
1.17788720e+00 -9.35839295e-01 -3.89093459e-01 -5.56225240e-01
-8.94913375e-01 9.56164479e-01 8.21395814e-01 4.04087603e-01
1.25949576e-01 -4.04821187e-01 2.99128294e-01 6.78335549e-03
-1.14524591e+00 -2.30642453e-01 3.11386734e-01 3.48414153e-01
6.64434612e-01 -1.34689035e-02 -7.11710632e-01 5.87487638e-01
-6.50164068e-01 -1.99240610e-01 5.24292648e-01 9.67271805e-01
-6.22212708e-01 -1.37547791e+00 -2.02587634e-01 2.67669529e-01
-1.52798936e-01 -1.87708870e-01 -7.31183350e-01 3.05230349e-01
-1.78546146e-01 1.56044412e+00 -4.55779359e-02 -2.32071981e-01
4.79227275e-01 3.05888653e-01 2.99480438e-01 -8.09296787e-01
-1.01859689e+00 -3.00379097e-01 6.87266529e-01 -4.18286294e-01
-3.87279779e-01 -5.67404687e-01 -1.20078790e+00 -3.38718355e-01
-4.19814169e-01 6.68624043e-01 5.03562868e-01 1.32887626e+00
4.25809205e-01 5.11777043e-01 8.23872626e-01 -3.64765465e-01
-6.32101536e-01 -1.38492489e+00 9.52670202e-02 4.32486534e-01
2.51653045e-01 -1.32180288e-01 -2.83767521e-01 -1.53127924e-01]
|
[11.873327255249023, 8.056998252868652]
|
4c293cbe-2e9a-45e9-b4c6-2ac01b5d8464
|
homophone-reveals-the-truth-a-reality-check
|
2209.10791
| null |
https://arxiv.org/abs/2209.10791v2
|
https://arxiv.org/pdf/2209.10791v2.pdf
|
Homophone Reveals the Truth: A Reality Check for Speech2Vec
|
Generating spoken word embeddings that possess semantic information is a fascinating topic. Compared with text-based embeddings, they cover both phonetic and semantic characteristics, which can provide richer information and are potentially helpful for improving ASR and speech translation systems. In this paper, we review and examine the authenticity of a seminal work in this field: Speech2Vec. First, a homophone-based inspection method is proposed to check the speech embeddings released by the author of Speech2Vec. There is no indication that these embeddings are generated by the Speech2Vec model. Moreover, through further analysis of the vocabulary composition, we suspect that a text-based model fabricates these embeddings. Finally, we reproduce the Speech2Vec model, referring to the official code and optimal settings in the original paper. Experiments showed that this model failed to learn effective semantic embeddings. In word similarity benchmarks, it gets a correlation score of 0.08 in MEN and 0.15 in WS-353-SIM tests, which is over 0.5 lower than those described in the original paper. Our data and code are available.
|
['Guangyu Chen']
|
2022-09-22
| null | null | null | null |
['word-similarity']
|
['natural-language-processing']
|
[-6.43098876e-02 2.89742172e-01 -1.06678963e-01 -3.06893826e-01
-7.76424646e-01 -7.34103382e-01 8.46780241e-01 3.52484226e-01
-5.85422754e-01 4.14726049e-01 7.34738052e-01 -6.31357908e-01
1.05716966e-01 -4.80896413e-01 -4.52312380e-01 -5.97733080e-01
7.24538490e-02 4.56178516e-01 1.39180139e-01 -6.73320115e-01
2.08797142e-01 -1.49502046e-02 -1.54154325e+00 -1.74169205e-02
7.34214962e-01 7.78062582e-01 2.37544924e-01 6.60588622e-01
-4.21156287e-01 1.93566188e-01 -7.92680442e-01 -8.01382184e-01
-1.15247192e-02 -3.68777096e-01 -8.04529607e-01 -3.01857889e-01
3.57498348e-01 1.30669530e-02 -6.42652154e-01 1.12062883e+00
7.55679905e-01 3.15362811e-01 6.47372246e-01 -1.05354524e+00
-1.28381181e+00 8.75082433e-01 1.79231182e-01 2.89193332e-01
3.70721221e-01 -9.22707617e-02 1.46369326e+00 -1.23529553e+00
5.28597474e-01 1.19055736e+00 5.49751878e-01 5.92614114e-01
-7.85319149e-01 -4.15473044e-01 -7.03019649e-02 3.34161729e-01
-1.54176056e+00 -6.89432025e-01 4.99975324e-01 -2.65337855e-01
1.10342920e+00 5.10777950e-01 4.79573369e-01 1.52008963e+00
-9.79048908e-02 6.52906954e-01 7.54601300e-01 -5.57216883e-01
2.85457104e-01 5.45481920e-01 9.37243998e-02 2.99481988e-01
2.67856896e-01 1.56385154e-01 -5.32559276e-01 -1.34767845e-01
5.38650155e-01 -2.94290274e-01 -2.26451844e-01 -2.44704202e-01
-1.31263804e+00 1.17632103e+00 5.68659678e-02 6.84525847e-01
-8.43340158e-02 9.38508436e-02 6.82393551e-01 5.29977977e-01
6.05696619e-01 7.16223776e-01 -4.60586965e-01 -5.60956001e-01
-7.09018350e-01 9.09013972e-02 6.94029808e-01 9.77201521e-01
3.50436300e-01 2.75192440e-01 -1.40782192e-01 1.08882809e+00
2.01271474e-01 7.46552348e-01 9.09591615e-01 -6.42613113e-01
2.41565838e-01 1.25112191e-01 -1.12334982e-01 -9.19325531e-01
-6.17630128e-03 -2.07817793e-01 -3.81034613e-01 -4.74296838e-01
1.74335018e-01 -2.06873696e-02 -6.79579616e-01 1.53118098e+00
-9.96624120e-03 1.58325046e-01 2.15494826e-01 9.26495254e-01
1.01423192e+00 6.49879277e-01 -1.01307981e-01 1.35863334e-01
1.61036313e+00 -1.04202974e+00 -1.00969839e+00 -2.06424981e-01
7.67646730e-01 -9.74574208e-01 1.44118357e+00 1.08160257e-01
-8.67358863e-01 -6.17202222e-01 -1.09890020e+00 5.46788238e-02
-7.34267652e-01 -5.58214933e-02 5.53356290e-01 9.91359055e-01
-1.10702872e+00 4.52008992e-01 -5.92496037e-01 -8.09485137e-01
-1.47478431e-01 -6.09827302e-02 -3.31268847e-01 1.09194383e-01
-1.67896795e+00 1.10236406e+00 1.86077192e-01 -4.12806273e-01
-4.32549685e-01 -5.72856843e-01 -1.04632294e+00 3.37779559e-02
6.05064221e-02 -2.52690256e-01 1.28403234e+00 -5.17452300e-01
-1.34813166e+00 7.26028085e-01 -3.37430358e-01 -3.96970928e-01
1.82172075e-01 -1.10905312e-01 -8.74834001e-01 1.68578789e-01
-3.66091356e-02 4.82581884e-01 7.11615324e-01 -9.56575990e-01
-3.52388263e-01 -1.93901882e-01 -9.97586325e-02 1.29805133e-01
-6.78461313e-01 1.18196562e-01 -4.80344892e-01 -9.46453512e-01
-1.77086219e-01 -8.25457752e-01 -3.84462625e-02 -3.03089529e-01
-1.90365106e-01 -6.03677273e-01 3.53249907e-01 -6.87528849e-01
1.51896989e+00 -2.25126576e+00 3.87371110e-04 1.71782032e-01
4.21041204e-03 5.12494326e-01 -3.59460086e-01 7.88675070e-01
-1.21950075e-01 4.73232448e-01 -1.29184380e-01 -4.48932052e-01
3.64164412e-01 3.71932983e-01 -5.40468276e-01 4.70510662e-01
1.89493701e-01 9.28384542e-01 -9.27541137e-01 -1.90834671e-01
2.85740584e-01 7.40992367e-01 -4.80703473e-01 1.02445826e-01
2.52264112e-01 -1.36813030e-01 -1.12671636e-01 3.92170250e-01
4.32493597e-01 2.86114812e-01 2.16897815e-01 -1.35290623e-01
-1.63870435e-02 7.78019965e-01 -9.45042610e-01 1.68162775e+00
-6.26036227e-01 9.38568175e-01 -3.27515513e-01 -1.07117856e+00
1.10191953e+00 5.02612829e-01 2.94782758e-01 -7.95073748e-01
2.26730049e-01 3.12813699e-01 5.99851785e-03 -4.03982311e-01
1.02879667e+00 -1.35473877e-01 -1.81478396e-01 3.36638331e-01
3.89260590e-01 -3.50543588e-01 -5.53581230e-02 1.68316334e-01
1.02877641e+00 -4.37043607e-01 4.82236557e-02 -3.86750787e-01
4.26142812e-01 -2.01919764e-01 1.21582806e-01 6.50160372e-01
-3.16996783e-01 8.66686702e-01 1.71982944e-01 -5.55957258e-02
-1.24287128e+00 -1.19966209e+00 -3.73129189e-01 1.09917092e+00
4.07027304e-02 -9.65458095e-01 -7.19389200e-01 -5.06646156e-01
1.38194542e-02 1.00149715e+00 -6.83525085e-01 -4.29314017e-01
-1.96441278e-01 -3.00914466e-01 9.29966986e-01 6.59646571e-01
-4.27589864e-01 -8.26058626e-01 -7.69347996e-02 7.67058283e-02
-2.90603638e-01 -1.18162179e+00 -7.84932733e-01 3.17325220e-02
-4.16419625e-01 -7.41457939e-01 -8.47805977e-01 -8.17712724e-01
3.79149973e-01 3.78075957e-01 1.05325961e+00 1.09549411e-01
-7.09261224e-02 5.12185335e-01 -1.06244612e+00 -4.47769791e-01
-5.89313447e-01 2.49631077e-01 5.17295957e-01 -2.86082894e-01
8.27988625e-01 -3.12557608e-01 -2.61712432e-01 2.49089286e-01
-8.73844564e-01 -6.98398054e-01 2.65074849e-01 1.00633144e+00
1.40323997e-01 -2.42453486e-01 5.95368266e-01 -4.59644228e-01
8.96749020e-01 -4.97836798e-01 -1.23895541e-01 1.72575876e-01
-8.24147582e-01 1.17891952e-01 5.62664628e-01 -4.82715994e-01
-4.53181833e-01 -4.59941357e-01 -6.01199329e-01 -4.21576053e-01
-8.29538554e-02 3.40872645e-01 1.43014550e-01 3.50509703e-01
4.81798917e-01 4.23680514e-01 1.17931746e-01 -5.95892966e-01
6.44769609e-01 1.05198324e+00 2.71362275e-01 -3.55012625e-01
8.16372216e-01 1.07061908e-01 -7.97126651e-01 -1.32647634e+00
-4.59354609e-01 -6.77239656e-01 -4.56402689e-01 2.72577316e-01
9.21262145e-01 -7.15381086e-01 -4.90042299e-01 1.08872838e-02
-1.37498212e+00 5.45368791e-02 -3.96271110e-01 7.98702300e-01
-3.91031414e-01 4.93233681e-01 -4.42880332e-01 -7.93331921e-01
-1.06616616e-01 -1.13940716e+00 1.11652970e+00 -1.63867399e-01
-6.62086904e-01 -1.35221767e+00 1.59345821e-01 3.04212838e-01
7.70885050e-01 -5.48461199e-01 7.13455021e-01 -1.30999017e+00
2.56706595e-01 -1.27907664e-01 3.83720212e-02 5.92918456e-01
3.65198553e-01 -2.22606538e-03 -1.21631634e+00 -2.95225471e-01
-2.15789005e-01 -7.30304196e-02 7.05609560e-01 1.30595684e-01
9.25224364e-01 -3.18657756e-01 2.86787301e-02 3.95212978e-01
1.05881667e+00 1.80433303e-01 6.10964239e-01 2.67444819e-01
5.56746721e-01 7.58053362e-01 4.19652551e-01 4.49758649e-01
3.78602803e-01 8.05707097e-01 5.11427522e-02 -7.03409547e-03
-1.46724969e-01 -4.91786927e-01 6.61973298e-01 1.57541311e+00
3.78645301e-01 -1.83971584e-01 -8.40660214e-01 7.61332750e-01
-1.45156205e+00 -8.20723116e-01 -1.41076490e-01 2.15590549e+00
7.67181814e-01 2.22707093e-02 1.17402397e-01 1.45565018e-01
5.53032935e-01 4.09341633e-01 -8.92589092e-02 -9.26803708e-01
-3.36531639e-01 6.13843501e-01 6.07790589e-01 7.26781607e-01
-7.60231495e-01 1.30580139e+00 7.05548000e+00 8.88529897e-01
-9.31581259e-01 3.94016176e-01 4.29541282e-02 1.54640555e-01
-9.66402292e-01 -1.61949739e-01 -6.69967592e-01 5.64460218e-01
1.38282633e+00 -5.68679750e-01 3.44947308e-01 7.38546848e-01
9.20902118e-02 3.11341614e-01 -1.00245380e+00 1.11204886e+00
4.74460393e-01 -1.19357824e+00 1.95764139e-01 6.59248084e-02
2.88060069e-01 1.05006270e-01 2.54401743e-01 5.83449841e-01
9.29622650e-02 -1.22523761e+00 6.76913023e-01 -2.41594911e-02
7.46899009e-01 -8.83998275e-01 1.03642368e+00 -1.02556951e-01
-1.06563616e+00 1.79053947e-01 -6.36911333e-01 6.79518282e-02
2.68495828e-01 4.41591531e-01 -9.78770912e-01 5.60539067e-01
6.04458511e-01 5.99612951e-01 -5.80989063e-01 6.19210601e-01
-1.21406466e-01 9.01719809e-01 2.34171818e-03 -3.61870527e-01
3.53438079e-01 -2.16740623e-01 6.35236859e-01 1.47210371e+00
5.15878618e-01 -3.10138315e-01 -2.15006575e-01 7.14568973e-01
1.05987377e-01 4.21803325e-01 -8.78971517e-01 -5.65064788e-01
7.87771940e-01 8.99089396e-01 -3.54242593e-01 -3.82625401e-01
-4.47268695e-01 1.21660948e+00 -3.51879071e-03 1.69465393e-01
-8.96445453e-01 -7.69131958e-01 1.39622927e+00 -3.13876085e-02
4.26542461e-01 -5.43494165e-01 -1.88454300e-01 -1.12715733e+00
-7.00652041e-03 -8.38556468e-01 -6.79338574e-02 -5.45897305e-01
-1.48730469e+00 9.22608852e-01 -1.63732931e-01 -1.05642414e+00
-2.98474073e-01 -8.30125690e-01 -4.62955028e-01 8.43678355e-01
-1.21924675e+00 -5.87284386e-01 1.24508217e-01 2.73141295e-01
7.69886792e-01 -4.34474885e-01 1.11347187e+00 5.14276206e-01
-2.69770235e-01 1.14224398e+00 2.42471576e-01 3.10570449e-01
9.40723300e-01 -1.20772171e+00 9.92327869e-01 6.83456004e-01
5.63740373e-01 8.99213195e-01 9.44772661e-01 -2.87922710e-01
-1.49577951e+00 -9.12187457e-01 1.41952980e+00 -7.97810197e-01
1.09907889e+00 -5.31405091e-01 -1.09094095e+00 4.00114119e-01
5.44725835e-01 -3.47898364e-01 9.82431769e-01 3.03613394e-01
-5.14416397e-01 1.91645548e-01 -7.47690022e-01 7.65418947e-01
1.16295195e+00 -8.47450197e-01 -1.03352988e+00 2.31401101e-01
1.36833119e+00 -1.00025535e-01 -8.57887864e-01 1.46683650e-02
5.22140443e-01 -6.58133030e-01 9.71160173e-01 -7.94411719e-01
2.42610380e-01 -2.52880082e-02 -7.04821825e-01 -1.75718927e+00
-1.44402340e-01 -4.34469759e-01 9.12607834e-02 1.32095635e+00
5.85570037e-01 -8.53176057e-01 3.02752346e-01 7.33444616e-02
-4.55744088e-01 -5.89736521e-01 -1.03649294e+00 -1.36780834e+00
4.77998614e-01 -8.30998719e-01 7.09071040e-01 1.22526729e+00
2.52673417e-01 3.45062464e-01 -1.30973279e-01 -1.14468791e-01
2.91849494e-01 -6.37540936e-01 5.46138525e-01 -1.02838778e+00
-9.53516141e-02 -5.77966452e-01 -4.93617505e-01 -1.13120067e+00
4.47711021e-01 -1.16761613e+00 -4.69549969e-02 -1.49958968e+00
-3.06355625e-01 -4.05860513e-01 -2.94497818e-01 3.23578238e-01
-1.83247104e-01 2.34801948e-01 9.02750194e-02 -9.37828347e-02
-1.52921796e-01 9.52218115e-01 7.68794715e-01 -1.82770327e-01
3.49800736e-02 -3.52256358e-01 -9.20531690e-01 2.62828350e-01
1.00023365e+00 -3.35799307e-01 -4.38663781e-01 -5.72687566e-01
2.11181659e-02 -4.63337749e-01 6.43511862e-02 -6.52921259e-01
1.71497576e-02 4.31596152e-02 -2.16141224e-01 -2.66094714e-01
5.61557531e-01 -6.50441766e-01 -3.67258906e-01 2.36372754e-01
-4.74376857e-01 3.72727752e-01 1.24904118e-01 4.16214824e-01
-4.93277133e-01 -2.88641453e-01 3.82440925e-01 1.54084727e-01
-7.19161868e-01 7.15142339e-02 -6.82305813e-01 2.76940286e-01
5.68595529e-01 -2.10040703e-01 -1.34897307e-01 -6.68137014e-01
-4.04418528e-01 -7.34684393e-02 3.14416826e-01 1.15009987e+00
7.37195909e-01 -1.67188597e+00 -8.29208374e-01 3.43937427e-01
6.19405091e-01 -7.64387131e-01 -2.49708854e-02 7.96449959e-01
-1.38806477e-01 5.75643778e-01 2.02350959e-01 -3.83454919e-01
-1.12370372e+00 4.90114838e-01 4.18820931e-03 3.83964986e-01
-4.68249559e-01 7.77910888e-01 -1.43193930e-01 -8.22068930e-01
3.66216362e-01 -4.66482013e-01 -1.85401127e-01 2.50027627e-01
7.05776989e-01 3.94730210e-01 3.82059246e-01 -8.85511875e-01
-6.72791362e-01 3.85641754e-01 1.85733870e-01 -4.69791114e-01
1.26917112e+00 -2.96226114e-01 1.12826623e-01 7.46648073e-01
1.65952218e+00 2.82478750e-01 -3.44983488e-01 -2.75352269e-01
8.31808150e-02 -4.47405607e-01 -3.24849933e-02 -3.96577358e-01
-8.04079354e-01 1.06124675e+00 4.21807021e-01 4.26531166e-01
5.27359724e-01 1.62948683e-01 1.16430318e+00 2.53029019e-01
1.05581798e-01 -1.37329304e+00 -3.37415151e-02 1.04852057e+00
7.97228456e-01 -1.19230306e+00 -5.03412366e-01 -2.23675579e-01
-9.93570149e-01 8.92560065e-01 3.18757832e-01 6.46451116e-02
7.06115305e-01 6.25143126e-02 2.25195780e-01 -1.25905843e-02
-7.66204655e-01 -4.40408170e-01 3.66393596e-01 8.33835661e-01
7.13326514e-01 2.43214637e-01 -4.33181912e-01 8.08123350e-01
-7.74896204e-01 -7.05969572e-01 4.98294353e-01 5.28533280e-01
-5.61073661e-01 -1.32832599e+00 -2.56514847e-01 1.06834814e-01
-2.93107510e-01 -5.51352918e-01 -5.81641316e-01 5.19584596e-01
-2.72897691e-01 1.30806363e+00 3.74182820e-01 -7.80237854e-01
4.75458652e-01 3.96283686e-01 2.25275263e-01 -7.05275118e-01
-3.74630153e-01 -2.70266593e-01 2.54889607e-01 -4.78500456e-01
-6.19310737e-02 -3.83039266e-01 -1.07301223e+00 -4.23402846e-01
-2.09591225e-01 5.58731019e-01 9.47738230e-01 7.05653965e-01
3.03510070e-01 4.55854475e-01 8.06572855e-01 -4.95345682e-01
-8.00712764e-01 -1.09079266e+00 -6.34019554e-01 5.22533953e-01
2.00446904e-01 -6.32563412e-01 -7.38389790e-01 -1.90029725e-01]
|
[10.794832229614258, 8.683478355407715]
|
3c2782e4-3a89-4ba5-b9d2-9d0f43344d65
|
counting-dense-objects-in-remote-sensing
|
2002.05928
| null |
https://arxiv.org/abs/2002.05928v1
|
https://arxiv.org/pdf/2002.05928v1.pdf
|
Counting dense objects in remote sensing images
|
Estimating accurate number of interested objects from a given image is a challenging yet important task. Significant efforts have been made to address this problem and achieve great progress, yet counting number of ground objects from remote sensing images is barely studied. In this paper, we are interested in counting dense objects from remote sensing images. Compared with object counting in natural scene, this task is challenging in following factors: large scale variation, complex cluttered background and orientation arbitrariness. More importantly, the scarcity of data severely limits the development of research in this field. To address these issues, we first construct a large-scale object counting dataset based on remote sensing images, which contains four kinds of objects: buildings, crowded ships in harbor, large-vehicles and small-vehicles in parking lot. We then benchmark the dataset by designing a novel neural network which can generate density map of an input image. The proposed network consists of three parts namely convolution block attention module (CBAM), scale pyramid module (SPM) and deformable convolution module (DCM). Experiments on the proposed dataset and comparisons with state of the art methods demonstrate the challenging of the proposed dataset, and superiority and effectiveness of our method.
|
['Qingjie Liu', 'Yunhong Wang', 'Guangshuai Gao']
|
2020-02-14
| null | null | null | null |
['object-counting']
|
['computer-vision']
|
[ 2.31088638e-01 -5.30009508e-01 3.68401617e-01 -3.63743007e-01
-2.67869532e-01 -2.05511838e-01 6.12130761e-01 -1.06794640e-01
-7.68729925e-01 8.91040802e-01 1.16589241e-01 -2.15606704e-01
-6.80600330e-02 -1.31271863e+00 -5.98456144e-01 -6.06709898e-01
4.02632803e-02 3.21490943e-01 5.94017208e-01 2.31045112e-01
3.38860631e-01 5.27716935e-01 -1.50690484e+00 -1.21950120e-01
1.01944065e+00 8.05573404e-01 7.63368964e-01 5.13059676e-01
-3.00617754e-01 1.06689537e+00 -5.78655243e-01 -7.32771233e-02
1.19291335e-01 -2.15162978e-01 -5.67798495e-01 1.31810874e-01
3.48364681e-01 -7.68427789e-01 -2.92796314e-01 1.39620841e+00
5.74964285e-01 2.61548579e-01 8.07078302e-01 -8.81616473e-01
-1.08664238e+00 4.79710996e-01 -1.04061770e+00 9.43947613e-01
-3.75584006e-01 -2.42152736e-02 4.69716281e-01 -8.65195274e-01
-1.47727802e-01 1.38791120e+00 5.85677207e-01 2.39549890e-01
-4.83654946e-01 -8.67032707e-01 1.65531546e-01 2.11617813e-01
-1.74297559e+00 -2.21961737e-02 3.67011040e-01 -4.72520173e-01
6.00438058e-01 6.98855668e-02 4.96983796e-01 3.87170643e-01
-9.41784307e-02 5.76517701e-01 1.14517140e+00 -1.41158581e-01
2.67696232e-01 1.26215532e-01 1.09057203e-01 5.18313766e-01
8.14066231e-01 -3.08690071e-01 3.02765280e-01 1.80374116e-01
1.13952565e+00 4.53030407e-01 -1.40400603e-01 2.52319157e-01
-1.19473541e+00 8.97227108e-01 7.69016504e-01 4.23982680e-01
-3.91277492e-01 1.71827838e-01 9.26521432e-04 -3.82238984e-01
5.91249585e-01 -3.90028626e-01 -1.24733575e-01 2.82454938e-01
-9.61573303e-01 3.03248972e-01 5.18444657e-01 8.83375049e-01
6.63458705e-01 9.13393348e-02 -2.43710488e-01 7.52450287e-01
2.88353086e-01 8.73689353e-01 3.49564523e-01 -4.74992901e-01
8.07975531e-01 6.14751160e-01 3.40353429e-01 -1.37523651e+00
-4.74729002e-01 -2.68458247e-01 -1.31903064e+00 5.41614853e-02
8.08242932e-02 -1.27763242e-01 -7.59727478e-01 1.18295491e+00
3.73632550e-01 2.56486893e-01 -2.28588328e-01 1.12434494e+00
1.05236602e+00 9.77682889e-01 4.49194282e-01 -2.66551208e-02
1.43427157e+00 -7.34670937e-01 -5.40437996e-01 -3.87736201e-01
4.98716086e-02 -3.77974331e-01 9.03173089e-01 -2.62390468e-02
-7.94801950e-01 -8.01827013e-01 -9.91639137e-01 -6.60855323e-02
-4.27625269e-01 5.95142245e-01 7.11293280e-01 5.27890265e-01
-5.63940704e-01 3.17876548e-01 -8.51236761e-01 -4.68949586e-01
7.82725751e-01 2.41823703e-01 -5.94074242e-02 -1.28306553e-01
-1.00075936e+00 7.55902886e-01 6.22645438e-01 5.21290958e-01
-7.22440660e-01 -3.83172244e-01 -7.89675474e-01 2.13372514e-01
1.09091930e-01 -4.59936827e-01 9.07383204e-01 -6.32932544e-01
-7.10097253e-01 7.89539397e-01 1.34002700e-01 -2.02199325e-01
5.96001565e-01 -1.25006929e-01 -3.93606961e-01 -1.34499341e-01
4.23966020e-01 5.44122696e-01 4.82195199e-01 -1.02282262e+00
-1.00032949e+00 -5.82352221e-01 5.49200699e-02 1.77267984e-01
-2.56271690e-01 1.69444948e-01 -1.17048934e-01 -4.68112588e-01
-6.24936726e-03 -4.64836568e-01 -1.61466897e-01 3.96122318e-03
-2.03973040e-01 -2.63481081e-01 9.68181551e-01 -7.19859958e-01
1.18621564e+00 -1.91734695e+00 -4.19161499e-01 -2.49383166e-01
1.17174767e-01 4.10336584e-01 1.11448623e-01 2.26291902e-02
3.23574066e-01 1.34274647e-01 -4.07575399e-01 6.20662086e-02
-1.88374072e-01 2.80134708e-01 -1.95901081e-01 5.47501385e-01
4.98884976e-01 7.67438352e-01 -9.51441586e-01 -9.63510215e-01
6.09278798e-01 6.02384865e-01 -3.02201927e-01 8.95795152e-02
-2.66720373e-02 6.84291959e-01 -8.68874490e-01 9.29640710e-01
1.34188080e+00 -4.67272490e-01 -2.59059131e-01 -1.95043623e-01
-3.84437203e-01 -2.88801402e-01 -1.42684245e+00 9.85325992e-01
-3.49508464e-01 2.45493129e-01 -3.25002037e-02 -9.22244251e-01
1.10786498e+00 -2.76135415e-01 4.69130836e-02 -6.74971700e-01
3.75764310e-01 1.37128964e-01 -1.87186226e-01 -7.17756033e-01
6.33534133e-01 -2.29785189e-01 9.83657539e-02 8.43939930e-02
-3.14705223e-01 3.54588288e-03 3.67989242e-01 -7.64954090e-02
9.48269248e-01 -1.55156419e-01 5.47152758e-01 -2.66158015e-01
6.65825486e-01 4.12293412e-02 4.24771935e-01 6.00071788e-01
-3.59774202e-01 4.27109152e-01 -1.11108739e-02 -6.96785629e-01
-9.75053132e-01 -9.83873010e-01 -4.09781724e-01 9.28931534e-01
4.54477310e-01 4.53008533e-01 -6.49330914e-01 -2.83827603e-01
1.81744862e-02 3.43406677e-01 -6.70010626e-01 4.29939389e-01
-7.85964906e-01 -1.15398479e+00 2.88287222e-01 8.39903831e-01
1.48284674e+00 -1.31804085e+00 -6.58189893e-01 2.52659589e-01
-3.64537239e-01 -1.35168684e+00 -3.66868854e-01 -2.82940298e-01
-8.64878416e-01 -1.07564175e+00 -8.76592040e-01 -8.76837969e-01
6.61400616e-01 8.17398965e-01 1.17104793e+00 2.47742400e-01
-4.62231129e-01 -1.96383387e-01 -2.08296776e-01 -8.24234784e-01
-2.34255232e-02 -8.22856836e-03 -1.48316681e-01 -1.59648266e-02
8.35420489e-01 -5.31144679e-01 -8.94534945e-01 3.96860123e-01
-1.17074203e+00 -2.25700766e-01 9.29188430e-01 5.25159061e-01
5.31629682e-01 5.70477426e-01 4.82721031e-01 -6.37654722e-01
5.42981982e-01 -7.47904181e-01 -9.58725631e-01 6.23168871e-02
-4.89353463e-02 -3.26089829e-01 5.47730267e-01 -2.71364510e-01
-9.84423637e-01 -9.65406094e-03 -1.37328459e-02 -1.34991139e-01
-3.19173992e-01 2.65412152e-01 -1.83864817e-01 6.95648417e-02
4.72101092e-01 4.37885761e-01 -3.10034424e-01 -4.22784418e-01
-3.08178905e-02 1.04430878e+00 5.06702125e-01 -4.72295254e-01
7.35592902e-01 7.61806905e-01 -1.33816078e-01 -9.13577616e-01
-1.02134061e+00 -6.19793713e-01 -6.08592331e-01 -8.91345590e-02
1.10794902e+00 -1.21931779e+00 -7.34253168e-01 4.42177922e-01
-1.19876683e+00 5.57542965e-02 -1.90390870e-02 6.18734896e-01
-4.43064384e-02 3.78452122e-01 -4.80556279e-01 -1.10855854e+00
-5.49090207e-01 -7.47767329e-01 1.20491135e+00 5.66745520e-01
6.08388007e-01 -7.14580297e-01 3.33737992e-02 3.48994166e-01
6.15924478e-01 3.56862068e-01 5.67099869e-01 -1.91663042e-01
-8.47538531e-01 -1.57875076e-01 -1.02735460e+00 5.92454910e-01
1.14132911e-01 -1.48696482e-01 -7.90111244e-01 -5.82505055e-02
-2.62006838e-02 -2.55761892e-01 1.03869605e+00 6.58050001e-01
1.33476913e+00 -2.72169650e-01 -3.66893172e-01 3.99704099e-01
1.73895586e+00 4.18664627e-02 8.55012238e-01 2.79055327e-01
8.36698771e-01 3.30610812e-01 4.76012856e-01 6.61604047e-01
6.04754627e-01 1.02916211e-01 6.82189167e-01 -1.90628752e-01
1.64269298e-01 -9.25628990e-02 -2.26799145e-01 7.35484064e-01
-6.70541406e-01 -5.85327782e-02 -8.37088823e-01 7.12100863e-01
-1.57579601e+00 -1.39517546e+00 -3.74491364e-01 1.74777973e+00
4.54778939e-01 -1.69517502e-01 8.37809369e-02 1.86814785e-01
1.06953013e+00 1.42260328e-01 -4.51804250e-01 2.05506891e-01
-4.28706817e-02 4.54053059e-02 7.47260273e-01 1.05459653e-01
-1.47681379e+00 6.68591619e-01 5.23003674e+00 9.41736341e-01
-8.43317688e-01 1.73207253e-01 8.24332356e-01 3.38867635e-01
8.71146247e-02 -4.36271936e-01 -1.20119858e+00 8.05730104e-01
3.57129216e-01 1.92054108e-01 2.82588840e-01 9.44309473e-01
2.46201411e-01 -2.43186221e-01 -4.83246535e-01 1.06882203e+00
2.18183566e-02 -1.05506384e+00 2.17478082e-01 -4.99341860e-02
6.89655423e-01 1.96424171e-01 -1.06374770e-01 5.49880981e-01
3.87284935e-01 -1.20325446e+00 6.01817727e-01 4.64843929e-01
7.86997855e-01 -6.74513638e-01 1.04457009e+00 7.32405841e-01
-1.65799260e+00 -1.76632836e-01 -1.09384787e+00 -4.61289853e-01
1.80921908e-02 8.71671975e-01 -3.03407550e-01 1.68480009e-01
9.16255653e-01 5.03466725e-01 -5.12889266e-01 1.07354283e+00
-6.47112448e-03 3.53073686e-01 -3.69954228e-01 -3.96697551e-01
4.47578937e-01 -3.63756597e-01 -7.20255300e-02 1.34182692e+00
4.19439405e-01 5.71117282e-01 3.04815859e-01 1.21887350e+00
-8.90215561e-02 4.20037098e-02 -5.92928171e-01 3.93279642e-02
4.68619406e-01 1.69413102e+00 -9.08841312e-01 -3.45958441e-01
-4.94923651e-01 6.04458511e-01 3.00454825e-01 -1.08778765e-02
-1.23827744e+00 -3.60513270e-01 1.47355869e-01 2.78688997e-01
4.54613507e-01 -2.31479973e-01 -6.02124026e-03 -1.16992569e+00
1.45363897e-01 -4.33668464e-01 3.63261759e-01 -7.24939466e-01
-1.36283278e+00 3.52167010e-01 5.04329614e-02 -1.16863251e+00
5.82898080e-01 -5.96790135e-01 -7.61157990e-01 1.00697744e+00
-2.02040553e+00 -1.21742857e+00 -1.12407482e+00 5.98193884e-01
6.53840482e-01 1.30593821e-01 4.02452856e-01 7.30010569e-01
-3.55574876e-01 5.95044978e-02 1.28251284e-01 6.08939171e-01
5.30637875e-02 -9.31081772e-01 4.03590322e-01 6.90321863e-01
-2.88781673e-01 3.28478366e-01 1.35249585e-01 -5.64726889e-01
-1.07364321e+00 -1.66407347e+00 6.54027164e-01 -3.22615355e-01
5.07272482e-01 -1.12950183e-01 -8.13948095e-01 6.85089707e-01
-3.30087006e-01 5.24461210e-01 1.97140381e-01 -3.78674060e-01
4.55628633e-02 -5.50594851e-02 -1.36726177e+00 1.41084328e-01
9.49398935e-01 5.22099398e-02 -4.14047301e-01 4.60181773e-01
4.46540713e-01 -2.56157964e-01 -4.44938600e-01 4.73286510e-01
3.05859298e-01 -9.66255605e-01 1.15482056e+00 -7.65977949e-02
7.18983471e-01 -4.92461383e-01 -3.32337171e-01 -8.05538893e-01
-6.54458940e-01 3.66146147e-01 1.92020729e-01 1.19791842e+00
-2.40073994e-01 -4.04914677e-01 8.47131491e-01 2.38241449e-01
-5.74597195e-02 -3.90007555e-01 -7.46621370e-01 -6.52944267e-01
2.62287498e-01 1.16773732e-01 8.60728860e-01 9.01852727e-01
-6.07326269e-01 3.37858707e-01 -3.31037581e-01 4.96854782e-01
7.63229072e-01 -2.05276646e-02 7.49543488e-01 -1.29558611e+00
-4.87358235e-02 -1.91011012e-01 -5.96655130e-01 -1.09842670e+00
-3.05547029e-01 -5.92128813e-01 5.02674580e-02 -1.87508976e+00
7.36055911e-01 -6.56808734e-01 -2.18934822e-03 1.67128086e-01
-5.22465408e-01 4.11193967e-01 -2.02590339e-02 3.61224055e-01
-6.97589815e-01 5.37041783e-01 1.46290290e+00 -5.45012772e-01
1.17502436e-01 -1.41602650e-01 -5.46429098e-01 9.53751683e-01
9.71330225e-01 -3.45236123e-01 -1.39236793e-01 -8.38974595e-01
-2.16447306e-03 -3.37574147e-02 7.49478400e-01 -1.40749121e+00
1.28228799e-01 -2.73569494e-01 8.06919992e-01 -9.91469979e-01
1.01853400e-01 -8.12199950e-01 -8.20116326e-02 5.76555490e-01
2.91053802e-01 -4.62743379e-02 1.90593660e-01 6.98893130e-01
-1.78278044e-01 -2.53270835e-01 9.26884294e-01 -6.69614434e-01
-9.11640048e-01 5.31648695e-01 -3.03005558e-02 1.56913027e-01
9.84778225e-01 -3.35470080e-01 -5.17203987e-01 1.52056068e-01
-3.02547961e-01 1.43280894e-01 -6.39748797e-02 1.32621989e-01
5.20706236e-01 -1.39219618e+00 -1.17216051e+00 8.32493231e-02
-2.32521556e-02 4.52912122e-01 5.09803772e-01 6.62818313e-01
-1.04492581e+00 3.69935393e-01 -2.42290661e-01 -5.99175870e-01
-8.67989302e-01 4.79772747e-01 2.61235654e-01 -3.16705972e-01
-5.50391138e-01 7.99336731e-01 3.86845917e-01 -5.51652372e-01
-1.04910754e-01 -6.51195407e-01 -5.68669498e-01 -2.10664943e-01
9.04987514e-01 6.08188510e-01 -8.02290961e-02 -7.57928967e-01
-4.08832252e-01 8.18092108e-01 5.26058339e-02 4.33680326e-01
1.30534852e+00 -1.11416690e-01 -1.59543052e-01 2.82176584e-01
9.33223605e-01 -2.92195439e-01 -1.04872727e+00 -3.16003293e-01
-4.83282954e-01 -6.89687014e-01 3.20406705e-02 -4.22253937e-01
-1.06166434e+00 1.15291774e+00 7.39065111e-01 3.12314987e-01
8.86936545e-01 -5.75022511e-02 7.21403480e-01 4.19642419e-01
5.66339254e-01 -9.21356857e-01 -6.20484240e-02 5.54834962e-01
6.75262094e-01 -1.69144428e+00 2.34083310e-01 -2.78079718e-01
-3.47877085e-01 8.40567172e-01 6.70576453e-01 -3.94183666e-01
7.04681337e-01 2.55451411e-01 -1.82387993e-01 -3.00451726e-01
4.64732572e-02 -4.87686098e-01 -3.79092693e-02 5.90440512e-01
3.13989818e-01 1.96528852e-01 -3.17788273e-01 5.89877188e-01
-1.24428710e-02 3.07015121e-01 4.63463515e-01 9.91748631e-01
-1.02710605e+00 -1.80505574e-01 -7.61856139e-01 7.86653757e-01
-5.97711205e-01 -3.23641002e-01 1.93084672e-01 9.01647031e-01
4.67646778e-01 8.94025445e-01 4.83560055e-01 -2.81798523e-02
2.38400266e-01 -6.76970780e-01 3.57494265e-01 -4.38571841e-01
-2.37914279e-01 -2.85603583e-01 -3.35944444e-01 3.74964960e-02
-7.84219861e-01 -4.24858540e-01 -1.14328599e+00 -7.24214673e-01
-6.15589380e-01 -5.16251884e-02 6.36255503e-01 8.07285845e-01
-1.52729318e-01 6.02506340e-01 4.32152748e-01 -1.04188561e+00
-6.90068185e-01 -1.27623236e+00 -1.08432651e+00 3.07158083e-01
-6.04552031e-02 -7.63515234e-01 -4.08505082e-01 1.82588771e-01]
|
[8.531953811645508, -0.29251232743263245]
|
d8da4b0e-d8fb-49cf-87e9-a9dea44a3001
|
text-based-inference-of-moral-sentiment-1
|
2001.07209
| null |
https://arxiv.org/abs/2001.07209v1
|
https://arxiv.org/pdf/2001.07209v1.pdf
|
Text-based inference of moral sentiment change
|
We present a text-based framework for investigating moral sentiment change of the public via longitudinal corpora. Our framework is based on the premise that language use can inform people's moral perception toward right or wrong, and we build our methodology by exploring moral biases learned from diachronic word embeddings. We demonstrate how a parameter-free model supports inference of historical shifts in moral sentiment toward concepts such as slavery and democracy over centuries at three incremental levels: moral relevance, moral polarity, and fine-grained moral dimensions. We apply this methodology to visualizing moral time courses of individual concepts and analyzing the relations between psycholinguistic variables and rates of moral sentiment change at scale. Our work offers opportunities for applying natural language processing toward characterizing moral sentiment change in society.
|
['Renato Ferreira Pinto Jr.', 'Yang Xu', 'Jing Yi Xie', 'Graeme Hirst']
|
2020-01-20
|
text-based-inference-of-moral-sentiment
|
https://aclanthology.org/D19-1472
|
https://aclanthology.org/D19-1472.pdf
|
ijcnlp-2019-11
|
['diachronic-word-embeddings']
|
['natural-language-processing']
|
[-1.82824537e-01 2.57783383e-01 -4.37569022e-01 -7.17419803e-01
2.25897282e-01 -4.94269222e-01 1.12526822e+00 5.42383671e-01
-1.04279101e+00 5.98625362e-01 1.20679057e+00 -4.73289788e-01
-8.25267360e-02 -8.76242399e-01 -1.30386025e-01 -5.05281448e-01
1.54916659e-01 2.44054511e-01 -6.64065182e-01 -9.06506598e-01
7.12590694e-01 2.90161997e-01 -8.72272313e-01 -3.23044807e-02
1.02669752e+00 1.19666956e-01 -7.19569385e-01 4.62149262e-01
3.29134464e-02 1.11164618e+00 -2.97238857e-01 -1.14642477e+00
-2.03732535e-01 -4.07966465e-01 -8.50353956e-01 -6.17480814e-01
5.22911847e-01 -4.20460105e-01 -1.29480511e-01 1.10747886e+00
3.19913179e-01 -2.03893427e-02 1.28990281e+00 -1.02354550e+00
-1.58930850e+00 5.90504408e-01 -2.89494306e-01 3.70836794e-01
3.11773539e-01 1.11260951e-01 1.40492094e+00 -4.28530246e-01
9.81402874e-01 1.73989868e+00 8.63499701e-01 9.49459076e-01
-1.33219182e+00 -4.44916815e-01 2.67786920e-01 1.15131252e-01
-5.14182448e-01 -3.90792251e-01 9.68715668e-01 -1.29434144e+00
5.57599187e-01 -1.81243151e-01 8.92821491e-01 1.55253446e+00
6.86637223e-01 2.79977888e-01 1.27386534e+00 -3.18042189e-01
1.46146238e-01 2.74667382e-01 8.03961277e-01 6.93406403e-01
6.43687189e-01 1.10817306e-01 -6.40103400e-01 -6.75717235e-01
3.60031605e-01 -2.32060209e-01 2.36463323e-01 3.91665883e-02
-1.23400414e+00 1.32760203e+00 2.34586611e-01 2.69225776e-01
-3.80906820e-01 5.15282869e-01 7.90612340e-01 4.59365547e-01
8.95714402e-01 8.65178645e-01 -4.63433981e-01 -4.09570128e-01
-2.76535600e-01 4.60204363e-01 5.54719925e-01 5.30265570e-02
3.87524992e-01 -1.78797945e-01 1.17098931e-02 8.03049862e-01
3.98693949e-01 6.76491678e-01 5.04638433e-01 -1.19536972e+00
-2.70418562e-02 4.65995133e-01 1.76523969e-01 -1.36945927e+00
-6.44004464e-01 3.12292635e-01 -2.71783620e-01 2.69576579e-01
4.92865622e-01 -3.11744332e-01 -1.27406195e-01 2.21118999e+00
8.99990350e-02 -9.58267093e-01 1.84496403e-01 5.78743696e-01
4.30090755e-01 3.39104444e-01 6.70143723e-01 -4.79613245e-02
1.57720089e+00 -1.07147153e-02 -7.15119004e-01 -2.50266463e-01
1.15381503e+00 -3.25301796e-01 1.31313956e+00 7.27327690e-02
-8.80724370e-01 -1.01223320e-01 -4.81445223e-01 -5.36201119e-01
-6.59936726e-01 -3.44760954e-01 8.89622927e-01 8.84802997e-01
-9.37073350e-01 7.39416301e-01 -4.63744313e-01 -9.38763857e-01
6.10187113e-01 -2.83715606e-01 -1.29879668e-01 4.81130898e-01
-1.39272380e+00 1.23377836e+00 -3.99556011e-01 -4.40344810e-02
-1.30130634e-01 -5.28233767e-01 -8.90587568e-01 -4.64545250e-01
-3.54733258e-01 -6.16170764e-01 9.19399858e-01 -1.23666739e+00
-1.06417143e+00 1.49174535e+00 -4.64781933e-03 -1.87266558e-01
4.87633914e-01 -4.64366108e-01 -5.55701315e-01 -2.32437894e-01
3.68614495e-01 6.00407124e-01 4.56788003e-01 -8.48015308e-01
-2.21464157e-01 -6.35649502e-01 3.04243624e-01 2.39663899e-01
-8.87694120e-01 4.35910285e-01 9.00799811e-01 -6.70538247e-01
-4.69777554e-01 -7.99726665e-01 2.53818594e-02 1.43103331e-01
-1.53161347e-01 -7.37768352e-01 1.65293440e-01 -6.73754513e-01
1.04874337e+00 -2.17302775e+00 -1.36120975e-01 -4.84017320e-02
3.51170212e-01 -4.62246835e-01 8.57484564e-02 6.69915974e-01
-6.95190625e-03 5.66276312e-01 -9.12177786e-02 1.11455135e-02
6.07490838e-01 1.06307231e-01 -3.97631794e-01 9.68747199e-01
2.05013245e-01 8.37496340e-01 -1.18714559e+00 -5.47065258e-01
-5.84909692e-03 3.39957386e-01 -1.18213904e+00 -2.59397954e-01
1.27757192e-01 1.66859820e-01 -3.35915923e-01 4.91313040e-01
3.16318184e-01 2.80274481e-01 4.68278289e-01 8.47654641e-02
-4.07178223e-01 5.52538693e-01 5.67699373e-02 9.73978162e-01
-2.61919022e-01 1.11810505e+00 -1.35178193e-01 -6.17184281e-01
9.45943534e-01 -8.14406797e-02 -1.12937026e-01 -9.56398785e-01
3.47070694e-01 -8.17485228e-02 2.58793145e-01 -6.36854887e-01
8.80975246e-01 -8.61391664e-01 -7.48918891e-01 8.45500529e-01
-4.22518969e-01 -2.04926759e-01 4.17588241e-02 2.14971766e-01
5.25513053e-01 4.35426861e-01 3.90290290e-01 -1.09984553e+00
3.08669627e-01 6.57429323e-02 8.14377964e-01 3.12173188e-01
-7.71557331e-01 -3.63410085e-01 1.25771618e+00 -1.01223409e+00
-1.04720604e+00 -1.14799666e+00 -5.98483145e-01 1.67279053e+00
8.42525810e-03 -1.88340750e-02 -4.69019651e-01 -5.04576504e-01
3.74344915e-01 1.39547968e+00 -1.21727109e+00 -2.55141705e-01
-5.92211664e-01 -1.08254981e+00 5.30667901e-01 3.18983316e-01
-5.67272715e-02 -1.30007994e+00 -7.86763132e-01 -2.89801925e-01
1.30288005e-01 -6.47489429e-01 -4.64291275e-02 -2.22437501e-01
-7.87588596e-01 -1.03161144e+00 -1.43733308e-01 -4.68361408e-01
8.20030630e-01 -4.21020418e-01 1.26850057e+00 -8.33237395e-02
-1.88490510e-01 4.82623965e-01 -1.41642526e-01 -5.60741544e-01
-8.20020437e-01 -2.84902573e-01 6.19896054e-01 -3.50339651e-01
8.81239951e-01 -4.18640494e-01 -5.73163807e-01 -3.36817175e-01
-7.21520424e-01 -3.09128910e-01 -2.24159837e-01 4.00989115e-01
-4.44010168e-01 -6.15284383e-01 7.12326765e-01 -1.26217544e+00
1.37984610e+00 -9.55571413e-01 -2.06090346e-01 -6.54666647e-02
-9.63475049e-01 2.42815670e-02 7.67575085e-01 -4.68403511e-02
-1.13004911e+00 -9.95885849e-01 1.87994286e-01 7.88950503e-01
-1.16100600e-02 1.30803987e-01 5.74272633e-01 5.84852815e-01
7.37969160e-01 -4.00361180e-01 -3.40524465e-02 -1.76924184e-01
8.45222771e-01 6.14046276e-01 4.48963255e-01 -1.03461611e+00
6.17992759e-01 7.96025455e-01 -3.78570139e-01 -9.99949992e-01
-8.21752429e-01 1.36358663e-01 -7.63974667e-01 -3.86820495e-01
1.39601803e+00 -9.30091560e-01 -1.14697587e+00 3.23078543e-01
-1.14195538e+00 -3.69915098e-01 -2.57687658e-01 5.29759169e-01
-4.91267145e-01 3.67839068e-01 -7.12318122e-01 -7.94750154e-01
-3.01089793e-01 -4.07264411e-01 6.37943745e-01 1.68433059e-02
-1.13524365e+00 -1.75977445e+00 8.78267407e-01 4.03355002e-01
-3.28622200e-03 5.52598774e-01 1.51658022e+00 -4.09322470e-01
5.06596923e-01 2.33763337e-01 1.92738529e-02 2.95524687e-01
1.79484069e-01 3.28369796e-01 -7.82315195e-01 2.12452728e-02
-1.51042283e-01 -7.93179333e-01 6.98557377e-01 2.41157413e-01
5.15155911e-01 -4.26669568e-01 3.17345224e-02 3.89574587e-01
1.20588481e+00 -6.15517609e-02 4.59322929e-01 8.27302754e-01
5.86568058e-01 1.16507053e+00 2.66003668e-01 4.85325724e-01
8.47843766e-01 4.75219972e-02 -1.71180710e-01 1.01432741e-01
5.56114674e-01 -1.62744373e-01 8.69789600e-01 5.72418272e-01
-1.16373010e-01 3.56773958e-02 -1.27486002e+00 9.01748240e-01
-1.42065406e+00 -1.31579733e+00 -2.41812110e-01 1.60252416e+00
9.48144913e-01 -5.87994158e-02 1.90978155e-01 -3.55850667e-01
5.02815664e-01 4.62770373e-01 -4.55271065e-01 -1.69236171e+00
-2.35593971e-02 -2.35140100e-01 4.31911558e-01 8.37725937e-01
-7.11149096e-01 1.14659500e+00 7.46550369e+00 -1.69652492e-01
-6.35059476e-01 -1.67106032e-01 7.41364300e-01 -1.37097463e-01
-1.01569188e+00 -1.41613439e-01 -3.34157854e-01 2.22283944e-01
8.03929269e-01 -3.19143534e-01 1.24599962e-02 6.29091620e-01
4.66373891e-01 -7.04685226e-02 -1.32905138e+00 4.49450135e-01
1.33440465e-01 -8.53943765e-01 -1.38308913e-01 1.88288823e-01
7.79577792e-01 -8.91772360e-02 3.92939061e-01 2.35163838e-01
1.06716824e+00 -8.77439082e-01 8.50376129e-01 6.38728738e-01
6.96682572e-01 -6.46601558e-01 5.89558125e-01 -1.82895973e-01
-3.68924104e-02 -1.79375142e-01 -3.19627941e-01 -7.99336612e-01
6.21236674e-02 4.42686975e-01 -5.13301849e-01 -3.35366637e-01
5.87882102e-01 9.77439404e-01 -3.12559724e-01 -5.47353983e-01
-3.52025181e-01 3.77571821e-01 2.84701258e-01 -2.05484688e-01
1.98278546e-01 -7.09381700e-01 2.25873441e-01 1.39473975e+00
-2.99849898e-01 -3.76034752e-02 -6.09241307e-01 1.15252340e+00
-2.48891786e-02 1.58259034e-01 -1.01994908e+00 -4.41196352e-01
2.03121766e-01 1.11441052e+00 -4.56755579e-01 -1.68066293e-01
-4.78744417e-01 9.25682664e-01 3.83914709e-01 2.87225306e-01
-6.69159114e-01 -1.95737362e-01 1.55643439e+00 -1.44445777e-01
-3.21460724e-01 -2.20544159e-01 -5.89945555e-01 -1.08840656e+00
-3.64467531e-01 -6.44206703e-01 2.20084295e-01 -5.57599008e-01
-1.75484812e+00 -6.95415810e-02 -2.06365228e-01 -2.20346257e-01
-7.46013597e-02 -8.95420671e-01 -7.41421938e-01 5.95139563e-01
-1.28526866e+00 -8.30737710e-01 2.22867459e-01 2.15810224e-01
-2.00403348e-01 3.01888492e-02 8.87821615e-01 -3.19351047e-01
-6.48571849e-01 3.69950384e-01 3.87479097e-01 3.23037058e-01
9.84056056e-01 -1.48865485e+00 6.72231853e-01 3.00808847e-01
-5.61401665e-01 8.74138653e-01 7.35692680e-01 -6.99681938e-01
-9.59644735e-01 -4.53930706e-01 1.44885683e+00 -1.08544481e+00
1.37149394e+00 -6.07597172e-01 -1.89166203e-01 8.38316321e-01
4.07218993e-01 -6.48166478e-01 1.30217922e+00 9.47434962e-01
-7.38748848e-01 5.06732538e-02 -1.18593669e+00 1.27485430e+00
1.22141266e+00 -1.07953131e+00 -1.15242743e+00 2.18076855e-01
6.90126121e-01 7.10948825e-01 -9.85004008e-01 -2.18705699e-01
8.21658313e-01 -1.08090329e+00 6.59660280e-01 -9.99584615e-01
1.07315195e+00 3.76582891e-01 -1.79553062e-01 -1.33305645e+00
-6.82219028e-01 -4.74893510e-01 7.48406589e-01 1.11944962e+00
1.63179696e-01 -9.56761122e-01 4.29706663e-01 7.43739367e-01
1.66344121e-01 -4.65511262e-01 -7.73572743e-01 -1.93322062e-01
1.15524757e+00 -2.50330865e-01 4.17365372e-01 1.51909828e+00
5.51835060e-01 4.47878271e-01 -1.45087689e-01 -2.62465030e-01
5.76245368e-01 9.92763042e-02 5.91382325e-01 -1.64715624e+00
4.12709951e-01 -8.19385767e-01 -4.27484006e-01 -2.12329865e-01
6.59595311e-01 -9.58049595e-01 -4.36492354e-01 -1.53817427e+00
3.90910089e-01 -1.48441866e-02 -8.45103487e-02 1.28678992e-01
-5.87566011e-02 -2.05635484e-02 1.13265134e-01 4.97308234e-03
-4.18188214e-01 7.81784892e-01 1.06411839e+00 2.34313861e-01
1.62119921e-02 -1.08184302e+00 -1.26674211e+00 1.30444169e+00
9.15620327e-01 -4.27070409e-01 -1.53200045e-01 -4.26906496e-01
1.14173031e+00 -4.66116250e-01 5.70331812e-01 -2.42130682e-01
-5.34760118e-01 -7.45429218e-01 3.08790028e-01 1.03956014e-01
4.74742018e-02 -4.60159600e-01 -7.38797963e-01 7.42247343e-01
-7.92118788e-01 4.52750415e-01 4.72103767e-02 3.77056420e-01
2.55092472e-01 2.64504433e-01 7.37342477e-01 2.49937963e-04
-5.31445265e-01 -2.34069124e-01 -8.14976931e-01 6.31218135e-01
8.49173725e-01 -1.06226832e-01 -7.85400748e-01 -1.58954263e-01
-4.67376769e-01 3.14980000e-01 1.01196289e+00 6.10178232e-01
2.59710878e-01 -1.37324429e+00 -1.01862049e+00 -2.12362766e-01
2.58264303e-01 -1.22983646e+00 9.23159868e-02 6.76207185e-01
-7.18655646e-01 1.13778174e-01 -3.51155072e-01 1.34842381e-01
-1.07738113e+00 4.00657028e-01 3.44295830e-01 3.43042165e-01
-4.55282122e-01 9.12183642e-01 5.38675010e-01 -7.72009790e-01
-4.01160389e-01 -7.40136951e-02 -5.14735162e-01 7.57180750e-01
1.77835509e-01 5.57925642e-01 -9.61843371e-01 -9.04921353e-01
-6.31434977e-01 8.65013123e-01 -1.53753916e-02 -2.32754037e-01
1.52058351e+00 -3.38532537e-01 -6.23586178e-01 1.15355444e+00
1.30844986e+00 3.72404844e-01 -8.10312510e-01 2.79751033e-01
9.09478292e-02 -3.52245271e-01 -2.10461453e-01 -7.12458074e-01
-2.79853523e-01 7.26575077e-01 3.41109782e-01 7.24884570e-02
4.48359102e-01 6.21115081e-02 4.89031404e-01 5.43772161e-01
3.85184251e-02 -1.85056078e+00 -6.21111691e-02 4.79087919e-01
6.65991306e-01 -1.23558426e+00 2.47237980e-01 4.32429254e-01
-9.75079596e-01 1.03287315e+00 5.34792721e-01 -4.26692814e-01
5.18482864e-01 -2.10982203e-01 5.24143100e-01 -6.60619557e-01
-8.47627163e-01 2.58189112e-01 -8.14469457e-02 4.14076298e-01
1.08562124e+00 5.91835618e-01 -9.76002753e-01 1.47721842e-01
-7.87549376e-01 -3.30761999e-01 1.02569687e+00 5.48788249e-01
-3.42048287e-01 -8.68659854e-01 -3.18771303e-01 1.75551757e-01
-7.39888728e-01 -4.09771591e-01 -9.53975677e-01 9.38414812e-01
3.44989210e-01 7.66800106e-01 5.42576134e-01 -1.59276769e-01
1.04946665e-01 2.87177801e-01 3.38708282e-01 -4.27915722e-01
-6.17163599e-01 -6.42124653e-01 5.06973982e-01 -5.93068540e-01
-5.95448792e-01 -1.07663488e+00 -1.13122976e+00 -1.02964914e+00
4.06756252e-01 6.45378008e-02 2.87534922e-01 7.33672678e-01
-1.11947572e-02 1.77296385e-01 3.81494075e-01 -7.49550462e-02
-4.31832761e-01 -7.71844506e-01 -6.18843973e-01 8.09984803e-01
5.16923606e-01 -3.47755343e-01 -6.15417898e-01 -1.68923642e-02]
|
[9.329949378967285, 10.164267539978027]
|
1e070dcd-70ec-48a5-8088-3cd5db736aab
|
deep-neural-network-for-musical-instrument
|
2105.00933
| null |
https://arxiv.org/abs/2105.00933v2
|
https://arxiv.org/pdf/2105.00933v2.pdf
|
Deep Neural Network for Musical Instrument Recognition using MFCCs
|
The task of efficient automatic music classification is of vital importance and forms the basis for various advanced applications of AI in the musical domain. Musical instrument recognition is the task of instrument identification by virtue of its audio. This audio, also termed as the sound vibrations are leveraged by the model to match with the instrument classes. In this paper, we use an artificial neural network (ANN) model that was trained to perform classification on twenty different classes of musical instruments. Here we use use only the mel-frequency cepstral coefficients (MFCCs) of the audio data. Our proposed model trains on the full London philharmonic orchestra dataset which contains twenty classes of instruments belonging to the four families viz. woodwinds, brass, percussion, and strings. Based on experimental results our model achieves state-of-the-art accuracy on the same.
|
['Partha Pakray', 'Abdullah Faiz Ur Rahman Khilji', 'Saranga Kingkor Mahanta']
|
2021-05-03
| null | null | null | null |
['instrument-recognition', 'music-classification']
|
['audio', 'music']
|
[ 3.24401915e-01 -4.42464739e-01 1.46108285e-01 1.62645280e-01
-6.66189075e-01 -9.38012719e-01 2.44741023e-01 2.14378517e-02
-3.25522095e-01 4.15974081e-01 9.70871449e-02 -9.95545983e-02
-4.37869549e-01 -3.10003310e-01 -1.81316495e-01 -4.55492526e-01
-3.81772012e-01 1.60432592e-01 -2.03076273e-01 -3.68473053e-01
6.18138731e-01 4.18131322e-01 -1.77452922e+00 4.97653097e-01
3.36030781e-01 1.31356680e+00 -1.15965903e-01 1.14340603e+00
9.41192433e-02 6.43241286e-01 -7.70266235e-01 -3.69316548e-01
2.75612265e-01 -4.98768002e-01 -6.98406756e-01 -4.08246845e-01
2.11376250e-01 2.02514082e-02 -2.38293651e-02 9.32824314e-01
5.38256466e-01 2.85220176e-01 7.49203503e-01 -8.60274136e-01
3.50823365e-02 1.01530969e+00 -2.19391689e-01 2.20223099e-01
3.80518228e-01 -4.32117671e-01 1.37526929e+00 -8.71301055e-01
1.81884125e-01 7.90166497e-01 1.19450891e+00 2.30741575e-01
-9.99230981e-01 -9.29052055e-01 -5.18769383e-01 3.08657587e-01
-1.42006838e+00 -3.67269278e-01 1.09089279e+00 -5.96959770e-01
6.85354471e-01 3.44048202e-01 1.00052953e+00 7.25875080e-01
6.18580058e-02 6.59875929e-01 8.08453321e-01 -7.97426999e-01
2.66535491e-01 -2.98137009e-01 1.64458621e-03 4.08720315e-01
-3.88230890e-01 2.82069862e-01 -7.96033919e-01 -2.72747248e-01
9.27820027e-01 -4.66148496e-01 -1.91149682e-01 9.12919268e-02
-1.23627818e+00 7.87931204e-01 1.23765178e-01 5.90753078e-01
-5.24787962e-01 1.78643972e-01 7.20103383e-01 5.00441730e-01
-6.46980107e-02 1.01605344e+00 -6.05259955e-01 -6.25746071e-01
-1.06082428e+00 4.14228380e-01 1.00478733e+00 1.90856308e-01
1.11975864e-01 4.72345263e-01 2.23066449e-01 1.15311146e+00
1.42616779e-02 -4.09273766e-02 9.19707894e-01 -1.11453080e+00
2.75357574e-01 3.43901932e-01 -9.29469392e-02 -9.47665036e-01
-4.02112186e-01 -8.73088479e-01 -9.33169246e-01 2.18358472e-01
5.25352180e-01 1.03175595e-01 -4.18125719e-01 1.40115762e+00
-1.19050138e-03 5.53666115e-01 2.02073134e-03 5.85742235e-01
6.63728654e-01 5.08292198e-01 -2.22320512e-01 -1.72637343e-01
1.21930945e+00 -6.80231988e-01 -5.44148207e-01 3.78364533e-01
1.24490261e-01 -1.06426740e+00 1.03577662e+00 1.09750414e+00
-1.01910245e+00 -1.10447741e+00 -1.31971371e+00 2.68373311e-01
-1.78491294e-01 4.04500633e-01 5.50404727e-01 6.80064201e-01
-3.76524866e-01 1.11152232e+00 -5.06003797e-01 2.37430856e-01
3.33041809e-02 5.23989201e-01 -1.71784878e-01 9.56355572e-01
-1.04796064e+00 3.59059602e-01 8.07567418e-01 -1.38154209e-01
-3.75614852e-01 -7.75566578e-01 -3.66558611e-01 1.85298845e-02
-2.14891940e-01 -2.54760236e-01 1.67469990e+00 -1.22454751e+00
-1.77891743e+00 7.63959229e-01 3.42650443e-01 -6.22665763e-01
1.54466927e-01 -2.39759684e-01 -6.82234228e-01 7.61031806e-02
-2.68864959e-01 2.46138841e-01 1.13282514e+00 -8.48016679e-01
-9.87399995e-01 -1.13133848e-01 -3.67421776e-01 -1.18863270e-01
-4.49506342e-01 9.73795280e-02 8.66185874e-02 -1.13988853e+00
3.88192058e-01 -1.25834072e+00 2.55978286e-01 -7.22633302e-01
-5.65853536e-01 -3.70969743e-01 4.30841267e-01 -8.14312220e-01
1.64831650e+00 -2.40237904e+00 1.55903906e-01 5.12278736e-01
-2.62998521e-01 1.72615200e-01 7.92826489e-02 3.92539769e-01
-3.80401820e-01 -3.02213639e-01 -7.02160150e-02 -3.85460369e-02
1.78378686e-01 -9.67402980e-02 -7.36807466e-01 7.86080658e-02
-4.85464418e-03 4.44156408e-01 -6.01240337e-01 -1.56576395e-01
5.75971277e-03 3.27973127e-01 -5.03910184e-01 2.14669704e-01
4.78882529e-02 4.54374790e-01 -2.18059167e-01 6.65799022e-01
7.02714846e-02 4.16987091e-01 1.29162539e-02 -4.88316268e-01
-1.65611237e-01 4.57747072e-01 -1.36827457e+00 1.79336393e+00
-2.79058009e-01 5.85591674e-01 -1.63403809e-01 -8.40089679e-01
1.10706639e+00 7.92037547e-01 6.04841173e-01 -7.89578333e-02
2.35144317e-01 6.85421884e-01 5.65179229e-01 -3.48459244e-01
6.38132334e-01 -4.15064126e-01 -2.15958238e-01 3.29099596e-01
2.41389498e-01 -3.26751053e-01 1.83124349e-01 -5.03715217e-01
9.00567591e-01 2.37592056e-01 4.58153695e-01 -8.99559557e-02
5.60657501e-01 -2.51349926e-01 2.16783851e-01 6.17919087e-01
-5.45666218e-02 5.98952651e-01 8.61340202e-03 -6.14653528e-01
-1.12454629e+00 -9.11331117e-01 -2.00810254e-01 1.39993000e+00
-6.81578815e-01 -5.03435254e-01 -6.84866011e-01 -2.88244095e-02
5.66801988e-02 3.59395355e-01 -4.26875114e-01 -1.25257298e-01
-7.70294011e-01 -2.90137857e-01 1.13524354e+00 5.86596608e-01
1.39648229e-01 -1.58385658e+00 -7.32334316e-01 5.69907904e-01
-5.60084432e-02 -7.65967667e-01 -3.13503414e-01 4.81967330e-01
-7.87971556e-01 -9.59139705e-01 -5.78623831e-01 -9.48477387e-01
-4.33257759e-01 -4.27247971e-01 1.14385581e+00 -3.59576672e-01
-3.45417500e-01 2.29918852e-01 -3.20649058e-01 -9.89038229e-01
-5.99794865e-01 3.52339953e-01 4.05278534e-01 2.37876445e-01
1.01846747e-01 -1.06605256e+00 -3.76663029e-01 1.09777197e-01
-7.74484217e-01 -3.78251016e-01 2.91391432e-01 8.09754491e-01
6.98983073e-01 3.53593051e-01 8.29774797e-01 -4.09670144e-01
8.48201454e-01 -1.08291000e-01 -1.57517746e-01 -1.65248483e-01
8.44195019e-03 -2.03229651e-01 8.06178868e-01 -8.52190137e-01
-3.98842067e-01 2.98451155e-01 -2.72825152e-01 -4.95912164e-01
-1.50156438e-01 6.82363868e-01 1.91292122e-01 -1.07467718e-01
6.46951497e-01 7.77553692e-02 -3.64268780e-01 -8.01930010e-01
1.83292434e-01 9.93615746e-01 1.23445237e+00 -5.43856144e-01
5.51898658e-01 9.17838216e-02 1.38938218e-01 -9.31198359e-01
-7.96137214e-01 -6.70120060e-01 -8.37845922e-01 -4.51514542e-01
4.72537339e-01 -4.51578230e-01 -1.04945600e+00 4.61860865e-01
-9.07209158e-01 2.20684826e-01 -5.09136558e-01 7.06514299e-01
-9.18525457e-01 6.46498725e-02 -7.26146519e-01 -1.29302049e+00
-6.93731487e-01 -5.74782789e-01 8.05486739e-01 1.53197959e-01
-8.20570588e-01 -6.40234053e-01 4.19014901e-01 -5.87051176e-02
7.93662816e-02 4.27177697e-01 1.06824350e+00 -8.06860864e-01
1.24222070e-01 -4.60421234e-01 5.55153608e-01 5.65616369e-01
2.41717651e-01 8.40654299e-02 -1.27742374e+00 -1.39401779e-01
1.21516354e-01 -4.07684684e-01 8.62606645e-01 2.90244043e-01
1.20320177e+00 -2.69619953e-02 2.81871080e-01 5.89030445e-01
1.08070052e+00 5.20244241e-01 3.18586975e-01 3.64459395e-01
5.63714266e-01 5.20660400e-01 4.62908030e-01 6.01024091e-01
-3.34086418e-01 7.67968416e-01 2.66947538e-01 3.59629929e-01
1.65767536e-01 -3.72747838e-01 2.67529309e-01 1.46740460e+00
-5.61367989e-01 2.24721894e-01 -8.68085206e-01 4.81518030e-01
-1.57054877e+00 -1.16156411e+00 -1.26454420e-03 2.22258282e+00
8.40240896e-01 2.71289974e-01 5.44214964e-01 1.50287008e+00
5.01638651e-01 -2.20695451e-01 -3.28562766e-01 -6.13137424e-01
9.50483978e-03 1.04315197e+00 -1.01588927e-01 -3.08921523e-02
-1.57942784e+00 3.93787682e-01 6.23387241e+00 1.07034945e+00
-1.12330711e+00 -3.80445838e-01 1.02045350e-01 -8.46915394e-02
4.72517222e-01 -4.21582997e-01 -4.32143241e-01 3.51908505e-01
1.10367930e+00 1.85335413e-01 7.40472853e-01 6.08566701e-01
-2.22017001e-02 4.62234229e-01 -1.17619014e+00 1.22774839e+00
-1.01538762e-01 -1.14804494e+00 3.06914784e-02 -9.96437147e-02
6.67282581e-01 -2.36952081e-01 3.15410197e-01 3.89109373e-01
-2.02609122e-01 -1.19553626e+00 1.04452491e+00 6.25034690e-01
7.81540811e-01 -1.16692579e+00 6.53716564e-01 2.93694466e-01
-1.42153537e+00 -4.29197162e-01 4.77652028e-02 -2.64358968e-01
-1.58773735e-01 1.11141816e-01 -8.31639707e-01 4.67939228e-01
6.45590425e-01 7.69586205e-01 -1.12328112e-01 1.27812779e+00
2.04363912e-01 1.11158907e+00 -2.95709610e-01 9.75023881e-02
7.74143115e-02 -1.48613080e-01 5.37618518e-01 1.16984963e+00
6.99174166e-01 -4.02853377e-02 1.90147042e-01 3.33705127e-01
-3.37452488e-03 3.42717230e-01 -5.73879257e-02 -3.43802810e-01
4.41741586e-01 1.07952762e+00 -5.11727154e-01 -1.51099294e-01
1.17099047e-01 5.16272664e-01 -3.11813354e-02 -1.07896319e-02
-3.50310206e-01 -5.79482496e-01 4.97238427e-01 -1.47066042e-01
4.02222872e-01 -1.38208568e-01 -3.13522995e-01 -6.08830571e-01
-1.94787726e-01 -1.16166699e+00 5.31901419e-01 -6.05575800e-01
-1.48658895e+00 5.98123372e-01 -6.11029625e-01 -1.70164728e+00
-7.10759044e-01 -7.93723106e-01 -6.98197007e-01 8.92642558e-01
-1.08322942e+00 -9.65681911e-01 -4.32602689e-02 5.46757162e-01
4.97276604e-01 -6.55740738e-01 1.43368757e+00 2.50465602e-01
-1.76005606e-02 4.83147413e-01 3.09402049e-01 5.47092974e-01
5.31511784e-01 -1.39768469e+00 2.28926182e-01 -3.24694552e-02
1.02030408e+00 3.80375117e-01 7.04232633e-01 -9.09822658e-02
-9.98619556e-01 -8.15328479e-01 7.91062176e-01 -2.24309325e-01
9.39814866e-01 1.99954748e-01 -7.93519855e-01 2.50132203e-01
-1.73254926e-02 -3.77395600e-01 1.16012883e+00 7.40587413e-02
-4.16017681e-01 -1.85062200e-01 -6.08796775e-01 1.49732679e-01
5.70873260e-01 -9.49537873e-01 -8.19578826e-01 -1.20847054e-01
1.74935147e-01 -2.56964207e-01 -1.01213562e+00 3.39769632e-01
1.27020705e+00 -1.00589919e+00 1.15082514e+00 -8.20280313e-01
4.71556246e-01 -2.73589820e-01 -3.44411820e-01 -1.22087300e+00
-3.31221461e-01 -6.93921983e-01 -4.10924107e-01 1.06752086e+00
1.92474872e-01 -1.22619964e-01 6.45050406e-01 -3.13573927e-01
-1.05469622e-01 -4.53777730e-01 -1.14604282e+00 -7.47045040e-01
-1.29988417e-01 -8.37880075e-01 4.85010296e-01 1.03773487e+00
2.29783222e-01 4.26774502e-01 -3.79288256e-01 -2.52301186e-01
4.34150338e-01 3.37884516e-01 5.56319594e-01 -1.88973391e+00
-8.31410348e-01 -7.68054426e-01 -7.21841335e-01 -4.78010118e-01
7.73546398e-02 -9.42025423e-01 -1.71714798e-01 -5.54668844e-01
-2.64159262e-01 -2.69318342e-01 -1.08408129e+00 1.63792118e-01
2.09338635e-01 8.78113925e-01 4.19853091e-01 4.88232076e-01
-6.07082061e-02 1.35146365e-01 7.43248403e-01 -1.80281997e-01
-5.63982308e-01 7.67320931e-01 -1.52427390e-01 1.09280050e+00
9.63980496e-01 -2.84293026e-01 1.43569726e-02 1.79454327e-01
3.14382136e-01 1.60845459e-01 2.20298678e-01 -1.58801806e+00
2.56873313e-02 3.31033170e-01 4.32484776e-01 -6.87120855e-01
7.38051176e-01 -6.05692983e-01 3.28582555e-01 5.51142097e-01
-6.82258606e-01 6.26860857e-02 1.80547819e-01 3.22930664e-01
-5.66529989e-01 -4.26686615e-01 5.96306443e-01 -4.37810179e-03
-3.18424433e-01 -1.82834104e-01 -2.83351958e-01 -1.78041309e-01
2.92720556e-01 -2.23134577e-01 4.70957696e-01 -3.64708275e-01
-1.02538621e+00 -7.21793115e-01 -1.68742269e-01 3.99785101e-01
3.33264530e-01 -1.55011141e+00 -8.47785115e-01 2.50845075e-01
1.47785485e-01 -7.01461494e-01 4.52268729e-03 5.65838456e-01
-3.60108584e-01 4.74744260e-01 -3.04558456e-01 -4.74134296e-01
-1.55337095e+00 2.05544740e-01 3.33876729e-01 -2.20467463e-01
-3.79392385e-01 9.43412840e-01 -2.51891971e-01 -3.23044747e-01
5.07357597e-01 -5.18134952e-01 -7.49041557e-01 2.80362457e-01
5.26481032e-01 4.01843011e-01 1.04119107e-01 -8.16178501e-01
-1.00557730e-01 6.94155931e-01 4.89523292e-01 -3.54294568e-01
1.25412679e+00 5.45918584e-01 -1.16610557e-01 1.17911065e+00
9.78520155e-01 2.47852594e-01 -5.41932821e-01 -9.18524414e-02
3.17115873e-01 -9.59415361e-02 4.15551616e-03 -8.70645642e-01
-5.46008706e-01 1.00276339e+00 7.15253234e-01 6.90769851e-01
1.26896715e+00 -5.02808571e-01 7.67964482e-01 4.77394491e-01
3.87999773e-01 -1.17322075e+00 1.32655641e-02 6.59623861e-01
8.44543993e-01 -5.25029600e-01 -3.21666598e-01 5.16369902e-02
-3.46405327e-01 1.53544712e+00 -1.55570477e-01 -4.89594579e-01
7.33514786e-01 1.05592236e-01 1.17925525e-01 2.49464735e-01
-4.82319117e-01 -6.50364906e-02 7.90871978e-01 2.68884450e-01
7.36105621e-01 2.44372606e-01 -1.08620502e-01 1.23844147e+00
-1.04401350e+00 2.06549410e-02 1.84147343e-01 6.16902232e-01
-5.46478093e-01 -1.06142986e+00 -7.49517083e-01 5.21402001e-01
-8.50684881e-01 -1.41637310e-01 -5.91217816e-01 4.00019914e-01
4.63117808e-01 9.59974051e-01 -6.32911734e-03 -7.28934467e-01
4.66300756e-01 4.10711914e-01 6.13264740e-01 -1.49239942e-01
-1.23496747e+00 2.54400462e-01 -4.28887531e-02 -3.97350639e-02
-5.00301361e-01 -5.17490685e-01 -1.14879000e+00 1.93888173e-01
-1.90807343e-01 5.12737513e-01 9.03251529e-01 7.96209574e-01
-9.85849053e-02 7.40250885e-01 7.79853225e-01 -1.14221525e+00
-8.20600033e-01 -1.27671576e+00 -1.00886059e+00 4.67734277e-01
4.84148175e-01 -3.71163517e-01 -1.87820658e-01 3.77911031e-01]
|
[15.840409278869629, 5.262752056121826]
|
a2c167ef-55eb-42ee-94a2-f4da5de30ca3
|
upgpt-universal-diffusion-model-for-person
|
2304.08870
| null |
https://arxiv.org/abs/2304.08870v1
|
https://arxiv.org/pdf/2304.08870v1.pdf
|
UPGPT: Universal Diffusion Model for Person Image Generation, Editing and Pose Transfer
|
Existing person image generative models can do either image generation or pose transfer but not both. We propose a unified diffusion model, UPGPT to provide a universal solution to perform all the person image tasks - generative, pose transfer, and editing. With fine-grained multimodality and disentanglement capabilities, our approach offers fine-grained control over the generation and the editing process of images using a combination of pose, text, and image, all without needing a semantic segmentation mask which can be challenging to obtain or edit. We also pioneer the parameterized body SMPL model in pose-guided person image generation to demonstrate new capability - simultaneous pose and camera view interpolation while maintaining a person's appearance. Results on the benchmark DeepFashion dataset show that UPGPT is the new state-of-the-art while simultaneously pioneering new capabilities of edit and pose transfer in human image generation.
|
['Andrew Gilbert', 'Armin Mustafa', 'Soon Yau Cheong']
|
2023-04-18
| null | null | null | null |
['pose-transfer']
|
['computer-vision']
|
[ 3.23501170e-01 4.03736383e-01 3.65613729e-01 -4.20456588e-01
-5.53966284e-01 -5.92370689e-01 9.47520494e-01 -5.70703030e-01
-2.52155930e-01 6.87108815e-01 -6.34607598e-02 2.65838325e-01
1.72159851e-01 -8.11062455e-01 -9.39580798e-01 -5.13402998e-01
4.33319002e-01 1.19552732e+00 7.06379637e-02 -3.49508345e-01
-2.07522824e-01 2.89784938e-01 -1.41141510e+00 6.27853423e-02
1.05830777e+00 6.20538890e-01 9.84403640e-02 9.44137752e-01
8.82504135e-02 3.17934632e-01 -6.38356745e-01 -9.11351562e-01
4.19209361e-01 -5.91020346e-01 -6.73280716e-01 5.71774364e-01
9.80898559e-01 -6.51636302e-01 -2.17822313e-01 8.24836373e-01
7.68309534e-01 2.00327635e-01 8.22748363e-01 -1.43094778e+00
-1.14385617e+00 4.53619093e-01 -7.53769577e-01 -4.34571087e-01
6.36817694e-01 3.93797338e-01 4.60076004e-01 -7.10842788e-01
8.27358902e-01 1.47550428e+00 6.04972661e-01 1.02404070e+00
-1.46943378e+00 -3.31783861e-01 2.43965685e-01 -3.03282797e-01
-1.25332022e+00 -2.11672917e-01 5.03483653e-01 -6.20331466e-01
4.72133040e-01 4.72826123e-01 1.14879167e+00 1.41474903e+00
4.13737074e-02 7.26003766e-01 1.26545155e+00 -3.43118697e-01
-2.20144495e-01 8.01064894e-02 -3.37314934e-01 1.05846560e+00
3.19567136e-02 2.38434106e-01 -5.97215056e-01 1.83898449e-01
1.49232721e+00 -1.60092771e-01 -2.43278086e-01 -5.60956538e-01
-1.40736842e+00 5.44363499e-01 2.79080063e-01 -2.83951104e-01
-2.23506719e-01 5.96963108e-01 -9.97494906e-02 1.19382516e-01
2.71615654e-01 3.97832096e-01 -1.30635977e-01 3.65206860e-02
-1.16912699e+00 7.60214627e-01 6.64625883e-01 1.44560099e+00
5.64010978e-01 1.10299639e-01 -7.80151784e-01 6.19014859e-01
1.48510024e-01 8.11207354e-01 4.50068489e-02 -1.07686520e+00
3.82027239e-01 4.67760473e-01 3.87917459e-01 -8.01294506e-01
-2.59930044e-01 -4.14463073e-01 -7.84676313e-01 3.07896256e-01
5.64964473e-01 -4.05861527e-01 -1.63676643e+00 1.91362464e+00
5.98943174e-01 -1.02719754e-01 -5.83370209e-01 1.03662193e+00
8.31462383e-01 3.81378919e-01 2.24112675e-01 2.38107398e-01
1.58510315e+00 -1.17878544e+00 -5.23870409e-01 -3.42725426e-01
-1.93496764e-01 -6.67019129e-01 1.15178418e+00 2.44356558e-01
-1.63640535e+00 -7.66376793e-01 -7.98461556e-01 -4.54822689e-01
-1.50327832e-01 2.49751121e-01 8.12328756e-01 7.60881245e-01
-1.23606849e+00 5.72191894e-01 -8.49490106e-01 -3.84145707e-01
3.66315037e-01 4.21410561e-01 -4.76194113e-01 9.05819014e-02
-8.85632336e-01 8.34355950e-01 1.48292243e-01 8.70112255e-02
-1.04435587e+00 -9.19668734e-01 -8.59604239e-01 -1.57463595e-01
3.03337187e-01 -1.90264392e+00 1.05521441e+00 -8.00909758e-01
-1.61940885e+00 1.16229296e+00 8.91584456e-02 -2.86564708e-01
1.37130666e+00 -4.43353266e-01 2.20615491e-02 1.60893768e-01
1.64922923e-01 1.40685606e+00 1.24268150e+00 -1.51421225e+00
-2.60456979e-01 -4.32197869e-01 2.08167378e-02 5.93453884e-01
-4.81842086e-02 -6.89257234e-02 -1.02058959e+00 -8.65724862e-01
-1.87568620e-01 -1.13052142e+00 -3.54517996e-01 3.17662001e-01
-7.72404134e-01 1.43815890e-01 6.19117916e-01 -8.81590664e-01
7.45080233e-01 -1.72927642e+00 7.82130659e-01 8.72248858e-02
3.45052719e-01 -5.66394767e-03 -2.38312423e-01 2.82880962e-01
9.09015760e-02 1.32420644e-01 -2.19048694e-01 -1.04035020e+00
4.52042252e-01 2.15253383e-01 8.36733845e-04 2.59556621e-01
3.71790193e-02 1.36324048e+00 -7.07727194e-01 -6.68665648e-01
4.03226465e-01 8.55903566e-01 -6.93434119e-01 3.04293811e-01
-2.91492820e-01 1.02269697e+00 -1.82026654e-01 4.25183922e-01
5.81249356e-01 -1.84745505e-01 -1.19583644e-01 -3.76870036e-01
2.14080065e-01 -4.56325561e-01 -1.11623812e+00 2.07419753e+00
-2.05230460e-01 1.72610864e-01 2.00550795e-01 -1.36314809e-01
5.25446236e-01 2.24104449e-01 3.71538848e-01 -3.23572725e-01
2.36436889e-01 -2.26205379e-01 -2.09351376e-01 -2.86694139e-01
6.52451813e-01 -1.96078494e-01 -1.40731871e-01 4.69912142e-01
3.57245028e-01 -5.39646387e-01 3.06048125e-01 4.04647321e-01
3.68207783e-01 7.29686737e-01 -8.45366493e-02 -2.01419741e-01
2.26424504e-02 -2.07655877e-01 1.56288832e-01 8.23144376e-01
7.78036788e-02 1.19825351e+00 1.78264678e-01 -4.68109474e-02
-1.20669067e+00 -1.46610701e+00 1.13466665e-01 1.20983636e+00
1.35267094e-01 -2.67474562e-01 -1.31393647e+00 -6.40733242e-01
-1.56590354e-03 6.38809919e-01 -8.73169601e-01 1.40955355e-02
-5.15964389e-01 -7.27661371e-01 5.05820513e-01 6.26032710e-01
6.21924937e-01 -9.34696198e-01 -5.19186795e-01 -1.57709774e-02
-2.95624405e-01 -1.09624553e+00 -1.20666027e+00 -5.03666461e-01
-5.04335821e-01 -7.61155069e-01 -1.24651206e+00 -6.70328438e-01
1.03983259e+00 -2.71879405e-01 1.29580188e+00 -5.14647365e-02
-5.42401373e-01 6.85682714e-01 -4.08848673e-02 -2.54295439e-01
-3.34627688e-01 3.09101678e-02 -1.62436515e-01 2.23726276e-02
-4.78847355e-01 -8.14987421e-01 -1.01140010e+00 3.65875006e-01
-7.09704280e-01 5.14523387e-01 3.86775255e-01 6.87818944e-01
5.75312972e-01 -3.19568098e-01 1.63617805e-01 -8.76115441e-01
5.85329890e-01 9.14587304e-02 -3.60936552e-01 3.06697875e-01
-4.19227362e-01 -2.28556931e-01 5.07325977e-02 -5.74346840e-01
-1.50422430e+00 2.09953889e-01 -2.02867538e-01 -3.13416034e-01
-2.53049254e-01 -2.11154222e-01 -3.29287559e-01 -2.54965961e-01
5.87333322e-01 3.00449401e-01 9.51066390e-02 -3.00478309e-01
1.09436905e+00 4.37423252e-02 1.00240052e+00 -8.93442988e-01
1.13407838e+00 6.07286632e-01 -1.52079016e-01 -5.10063350e-01
-4.54720646e-01 7.54325688e-02 -1.04845786e+00 -3.15670997e-01
1.36730134e+00 -9.51898694e-01 -7.41347551e-01 7.23432064e-01
-1.13963175e+00 -3.83332074e-01 -5.22232413e-01 -7.89907500e-02
-7.63205707e-01 4.10315484e-01 -9.54169869e-01 -4.11801189e-01
-5.27472615e-01 -1.16480768e+00 1.59151280e+00 2.56445974e-01
-5.18761933e-01 -9.92957890e-01 -1.61250427e-01 7.71831214e-01
3.41099501e-01 5.96764982e-01 5.97825050e-01 8.11363831e-02
-8.83784294e-01 -1.20500118e-01 -1.65641770e-01 -2.81827562e-02
-5.57728931e-02 -1.13120742e-01 -7.21919358e-01 -3.88951987e-01
-4.77114171e-01 -1.57339409e-01 7.84247637e-01 3.36821139e-01
9.23519075e-01 -2.65867889e-01 -3.29454482e-01 1.04437196e+00
1.02184510e+00 -2.70994697e-02 7.13575721e-01 -9.71191376e-02
1.25373638e+00 4.70054269e-01 2.79195398e-01 4.15754110e-01
7.56298006e-01 9.63658750e-01 1.66119516e-01 -4.83813405e-01
-5.91608465e-01 -8.04532409e-01 5.68156280e-02 3.22414398e-01
-7.20852911e-01 -2.75943398e-01 -5.35014629e-01 3.60238820e-01
-1.75462544e+00 -8.23797286e-01 -1.09406449e-01 2.06183124e+00
1.01127934e+00 -2.72873878e-01 3.86423022e-01 -4.09303010e-01
5.63875377e-01 -1.34046778e-01 -3.41296822e-01 6.37158938e-03
8.87139216e-02 2.56229401e-01 3.43292475e-01 5.06888330e-01
-1.11655080e+00 1.23445141e+00 6.78772020e+00 7.36231387e-01
-7.82989442e-01 2.02840716e-01 4.46503222e-01 -2.22829700e-01
-4.61641639e-01 -2.98410147e-01 -8.34391415e-01 3.35545599e-01
3.03267449e-01 1.33142561e-01 4.71254945e-01 5.19935191e-01
6.18414916e-02 -1.39970118e-02 -1.26117337e+00 1.11705005e+00
2.98883498e-01 -1.11204517e+00 5.39910257e-01 1.21326290e-01
9.65486944e-01 -6.61265254e-01 3.58344406e-01 1.65210798e-01
5.59187114e-01 -1.12245333e+00 1.26799548e+00 7.27032721e-01
1.07729805e+00 -5.81173956e-01 1.01567686e-01 2.46841982e-01
-9.83860910e-01 2.81488419e-01 6.06058575e-02 3.54928553e-01
8.18792164e-01 2.83701181e-01 -6.01166189e-01 5.91124356e-01
5.72365165e-01 2.67394692e-01 -5.95869899e-01 7.25045681e-01
-4.70791668e-01 -6.51485473e-02 -1.82434127e-01 5.09446204e-01
-2.33730540e-01 -2.95076042e-01 6.94359362e-01 1.08241892e+00
3.91966939e-01 -1.65795479e-02 4.95813489e-01 1.32942140e+00
-4.57927398e-02 -1.95353940e-01 -1.29480943e-01 6.99727014e-02
2.60058522e-01 1.23382962e+00 -7.20469058e-01 -4.06664342e-01
2.16932699e-01 1.88270664e+00 2.15388611e-01 4.83286589e-01
-1.19589460e+00 8.45348090e-02 4.76859391e-01 4.50530976e-01
2.63949186e-01 -4.05745298e-01 -2.53386438e-01 -1.19405770e+00
-1.07201688e-01 -8.09476972e-01 4.30416688e-03 -1.05643940e+00
-1.15673566e+00 6.24404013e-01 2.57004887e-01 -6.86464548e-01
-3.36696446e-01 -4.20783222e-01 -4.19212967e-01 9.53719258e-01
-9.26719844e-01 -2.08888197e+00 -4.91241783e-01 8.63104761e-01
4.95396912e-01 1.89637631e-01 6.36348188e-01 3.42123538e-01
-4.94719207e-01 8.23611200e-01 -7.35662758e-01 9.73724574e-02
8.32034290e-01 -1.54345405e+00 6.38732553e-01 6.93067551e-01
9.39826742e-02 8.18526387e-01 8.20599794e-01 -8.50782514e-01
-1.16871822e+00 -1.12726545e+00 2.98451245e-01 -9.27644789e-01
1.05958596e-01 -5.18354118e-01 -2.44419381e-01 9.88403320e-01
5.03891945e-01 -3.12068373e-01 2.37125576e-01 -9.31533501e-02
-1.93213135e-01 1.34421974e-01 -1.14963245e+00 9.02303755e-01
1.47817743e+00 -3.37994635e-01 -3.52975309e-01 2.90522933e-01
7.77056456e-01 -1.02516103e+00 -9.05203819e-01 1.68736726e-01
6.10863626e-01 -9.74856377e-01 1.39258873e+00 -4.08266544e-01
4.31437552e-01 -3.58980328e-01 2.83740819e-01 -1.38982761e+00
-4.92373973e-01 -9.56585169e-01 1.85580086e-02 1.50373816e+00
4.02838796e-01 -4.77308035e-01 8.35147679e-01 9.33219016e-01
-1.56639457e-01 -4.48857248e-01 -5.49944997e-01 -5.67941666e-01
-1.26492400e-02 -2.21570671e-01 7.46408343e-01 6.70855880e-01
-4.71002102e-01 2.82223552e-01 -9.17867601e-01 4.43836786e-02
1.02026033e+00 2.89175455e-02 1.10648084e+00 -9.82197225e-01
-7.44737923e-01 -2.72394210e-01 -2.41683483e-01 -1.20526159e+00
-1.87246010e-01 -7.02933669e-01 -4.98087965e-02 -1.76091659e+00
3.74543816e-01 -2.25601360e-01 4.89352822e-01 4.82582629e-01
-3.48952919e-01 5.63474655e-01 5.80026686e-01 5.79104498e-02
-3.52493525e-01 5.73422134e-01 1.99972975e+00 -3.70677235e-03
-1.30005777e-01 -7.43406340e-02 -7.47972965e-01 7.08523571e-01
2.71991223e-01 -3.94938104e-02 -8.26118290e-01 -7.10887611e-01
-3.75428200e-02 1.49450287e-01 9.42398071e-01 -9.50630009e-01
1.61022499e-01 -5.90226613e-02 8.29300761e-01 -4.44117159e-01
7.69560635e-01 -6.10462964e-01 6.45665824e-01 1.74669236e-01
-1.00212708e-01 -3.44781484e-03 6.40511587e-02 5.86520553e-01
2.86833227e-01 3.74308735e-01 8.46148252e-01 -4.55763876e-01
-5.16573846e-01 7.26436734e-01 1.23403229e-01 1.46410435e-01
1.11023617e+00 -4.03214782e-01 -2.64818311e-01 -4.97454822e-01
-1.34747708e+00 2.06665754e-01 6.97060406e-01 4.58515644e-01
5.02266228e-01 -1.31160581e+00 -7.46124268e-01 2.94144750e-01
-2.49798298e-01 6.66425824e-02 6.18688464e-01 7.43762732e-01
-5.03558278e-01 -1.10364653e-01 -2.83824414e-01 -6.16398156e-01
-1.33697093e+00 4.13148254e-01 6.35274172e-01 -2.71792322e-01
-7.16658473e-01 1.17327678e+00 5.79581082e-01 -6.23386025e-01
-9.50761512e-02 -1.37546398e-02 1.98838800e-01 -6.38340265e-02
3.06995988e-01 2.88060278e-01 -3.52052629e-01 -6.37006879e-01
2.99512576e-02 8.16676617e-01 6.95463084e-03 -5.02512693e-01
9.15337980e-01 -4.33222294e-01 -5.83595112e-02 -5.79166003e-02
4.51578081e-01 -4.18174081e-02 -1.72758198e+00 3.61253858e-01
-8.33138645e-01 -5.12563407e-01 -3.34535211e-01 -1.24705565e+00
-1.10498619e+00 7.93498755e-01 4.75083888e-01 -2.21022755e-01
8.42363417e-01 5.85966147e-02 9.37895715e-01 -2.30271518e-01
7.40978658e-01 -1.04878724e+00 2.56437689e-01 1.33313715e-01
1.30911136e+00 -8.59381855e-01 -1.89267874e-01 -9.12117898e-01
-8.72275114e-01 5.43665886e-01 9.28305328e-01 -3.21224406e-02
3.62064600e-01 4.03755307e-01 2.14893874e-02 -2.20668346e-01
-1.92809105e-01 -1.95608526e-01 8.00706923e-01 1.04984367e+00
3.60475361e-01 2.61703789e-01 -4.21984270e-02 5.62772453e-01
-6.87142015e-01 -1.28255010e-01 1.96317267e-02 4.48387444e-01
-3.29477526e-02 -1.31439900e+00 -5.06158113e-01 1.05019182e-01
-1.12831056e-01 -1.10615902e-01 -4.54762071e-01 8.91550779e-01
4.37059879e-01 5.91467023e-01 -1.27096027e-01 -2.01560110e-01
4.87219393e-01 -2.68876944e-02 1.11542189e+00 -6.49270356e-01
-5.75106561e-01 2.90101707e-01 1.27930924e-01 -6.09038830e-01
-2.65314966e-01 -6.09820902e-01 -1.02594984e+00 -4.05692428e-01
1.90059766e-02 -3.60900164e-01 3.76744539e-01 8.84921372e-01
3.70307207e-01 8.15395594e-01 -7.04351664e-02 -1.38986492e+00
-3.22226584e-01 -8.59804153e-01 -6.09827280e-01 7.23939657e-01
-6.74579069e-02 -6.68652177e-01 1.23065375e-01 5.61151445e-01]
|
[11.941361427307129, -0.8137364387512207]
|
3fd971a1-6b97-416f-9c71-120cb26d2552
|
forward-modeling-for-partial-observation
|
1812.00054
| null |
http://arxiv.org/abs/1812.00054v1
|
http://arxiv.org/pdf/1812.00054v1.pdf
|
Forward Modeling for Partial Observation Strategy Games - A StarCraft Defogger
|
We formulate the problem of defogging as state estimation and future state
prediction from previous, partial observations in the context of real-time
strategy games. We propose to employ encoder-decoder neural networks for this
task, and introduce proxy tasks and baselines for evaluation to assess their
ability of capturing basic game rules and high-level dynamics. By combining
convolutional neural networks and recurrent networks, we exploit spatial and
sequential correlations and train well-performing models on a large dataset of
human games of StarCraft: Brood War. Finally, we demonstrate the relevance of
our models to downstream tasks by applying them for enemy unit prediction in a
state-of-the-art, rule-based StarCraft bot. We observe improvements in win
rates against several strong community bots.
|
['Jonas Gehring', 'Nicolas Usunier', 'Gabriel Synnaeve', 'Zeming Lin', 'Vegard Mella', 'Vasil Khalidov', 'Nicolas Carion', 'Dan Gant']
|
2018-11-30
|
forward-modeling-for-partial-observation-2
|
https://openreview.net/forum?id=B1nxTzbRZ
|
https://openreview.net/pdf?id=B1nxTzbRZ
|
iclr-2018-1
|
['real-time-strategy-games']
|
['playing-games']
|
[-5.59211858e-02 3.60233244e-03 -4.30683941e-01 3.21012974e-01
-2.88924962e-01 -7.92348623e-01 9.84231889e-01 -4.08204645e-01
-6.20444417e-01 7.47561336e-01 5.44137061e-01 -6.01292491e-01
-5.54912947e-02 -6.95681155e-01 -4.28172559e-01 -3.79885994e-02
-5.46722770e-01 7.02548563e-01 7.35497832e-01 -1.12147045e+00
2.49450609e-01 2.32628882e-01 -1.01249540e+00 5.20125329e-01
2.62875080e-01 4.62720513e-01 1.80871710e-01 1.15360034e+00
7.78848588e-01 1.95624018e+00 -8.96708906e-01 -6.01614594e-01
5.77300429e-01 -4.92146730e-01 -1.02968240e+00 -2.67187655e-01
-1.88397884e-01 -6.87299967e-01 -1.04485869e+00 7.38995731e-01
4.85543430e-01 2.23737806e-01 4.23173636e-01 -1.08049941e+00
-1.64296553e-01 8.62702668e-01 -2.35728458e-01 7.07278669e-01
7.01958537e-02 9.03769493e-01 1.30399442e+00 1.90262988e-01
9.80436265e-01 1.01798511e+00 8.62611949e-01 8.47668052e-01
-1.36188352e+00 -6.58467114e-01 2.78174520e-01 9.22697783e-02
-5.97205281e-01 -6.24927223e-01 4.39281493e-01 -5.50748348e-01
1.76982892e+00 -3.13060105e-01 7.01825619e-01 1.92148757e+00
5.78503907e-01 7.07239389e-01 7.10996389e-01 3.20926726e-01
-7.22389072e-02 -8.12608600e-01 -1.56556338e-01 9.13574994e-01
-2.30267961e-02 8.69469225e-01 -4.70236689e-01 -3.23016405e-01
1.01051903e+00 -1.27025619e-01 2.19941214e-01 -1.26190677e-01
-1.13083637e+00 1.11923826e+00 4.19604748e-01 6.35994673e-02
-4.85758722e-01 7.47390926e-01 8.20118546e-01 5.01402617e-01
5.37792206e-01 1.07490158e+00 -5.43201327e-01 -8.43668282e-01
-7.41262615e-01 7.32210457e-01 1.06176126e+00 4.22537863e-01
7.90592656e-02 4.06405807e-01 -3.45331311e-01 2.80796915e-01
-1.50384471e-01 1.82689771e-01 4.40571636e-01 -1.27442253e+00
4.43728626e-01 2.21081570e-01 1.58319503e-01 -7.62064874e-01
-5.92297673e-01 -6.47725642e-01 -1.45446405e-01 4.61249173e-01
5.04832625e-01 -7.26071656e-01 -8.20439994e-01 2.01215100e+00
-3.78513217e-01 7.41141617e-01 1.55169249e-01 5.94588816e-01
3.70767504e-01 5.60665131e-01 -1.87580600e-01 2.85375286e-02
8.27953339e-01 -1.15539849e+00 -6.06461428e-02 -8.00971448e-01
6.79421008e-01 -1.17843896e-01 7.15975285e-01 5.17726600e-01
-1.07520139e+00 -1.71610281e-01 -6.88397110e-01 3.48318130e-01
3.89264859e-02 -6.61370158e-02 8.28260362e-01 4.75666039e-02
-1.13672292e+00 1.14348280e+00 -1.48904061e+00 -3.45114827e-01
2.53031224e-01 4.19184834e-01 -9.41177830e-02 7.30080426e-01
-1.30158830e+00 1.25864410e+00 5.67062497e-01 -3.01029474e-01
-1.99846721e+00 -3.62825900e-01 -5.91539562e-01 1.63127884e-01
6.12197757e-01 -4.16130990e-01 1.76727509e+00 -3.65566194e-01
-1.77497029e+00 6.91057980e-01 4.20555174e-01 -9.41407263e-01
1.93552688e-01 -5.19735813e-02 -2.78269231e-01 -1.02087870e-01
2.69872546e-01 2.54981011e-01 7.70047903e-01 -6.67659760e-01
-8.38562906e-01 5.99424355e-03 7.47033238e-01 1.92069218e-01
3.09562124e-02 3.16894442e-01 1.51794747e-01 -4.67735887e-01
-6.25672579e-01 -1.10104549e+00 -6.30122602e-01 -9.39409435e-01
-2.59961665e-01 -7.46783167e-02 7.35788405e-01 -8.81256521e-01
1.18223262e+00 -1.78475320e+00 5.79514503e-01 -1.61702707e-01
4.98730958e-01 4.70223397e-01 -5.04697442e-01 6.57887816e-01
6.75885305e-02 -1.98146001e-01 3.08025569e-01 -2.71417528e-01
2.88559813e-02 2.12667063e-01 -6.62683129e-01 1.12450793e-01
2.75122792e-01 1.32445323e+00 -1.24579227e+00 4.21758071e-02
5.92732290e-03 -1.32079363e-01 -1.03240693e+00 2.39991859e-01
-8.06350589e-01 6.30252838e-01 -3.78113687e-01 2.49631181e-01
-5.13599992e-01 -2.74300665e-01 5.90013146e-01 4.79557157e-01
1.26640936e-02 9.90121841e-01 -2.77337551e-01 1.74963343e+00
-4.32600588e-01 7.31851280e-01 3.26421186e-02 -9.20128405e-01
5.07203698e-01 2.03635350e-01 4.49912280e-01 -4.06238914e-01
5.15811443e-01 -9.91166979e-02 6.58270478e-01 -1.23885222e-01
6.86692178e-01 -3.34575363e-02 -5.98656535e-01 7.07293928e-01
5.44415653e-01 -2.04993457e-01 4.66368437e-01 4.22096312e-01
1.87317050e+00 3.79825264e-01 2.80413628e-01 2.31029633e-02
-2.32411981e-01 4.74278271e-01 5.61519563e-01 1.08111882e+00
-4.34583306e-01 -8.05653334e-02 1.11544907e+00 -4.99760628e-01
-1.08034647e+00 -8.66302311e-01 7.37737358e-01 1.62691975e+00
-1.95519656e-01 -1.00933933e+00 -3.88462722e-01 -7.86297858e-01
-2.75000602e-01 5.69402575e-01 -7.31700897e-01 -4.87013698e-01
-8.99375379e-01 -5.65713346e-01 1.05210161e+00 4.50676054e-01
4.18846965e-01 -1.42965639e+00 -8.70989382e-01 7.29764462e-01
-1.73212767e-01 -1.10098159e+00 -2.13298723e-01 3.04999352e-01
-5.89805245e-01 -1.35929823e+00 -5.19046001e-03 -4.06727910e-01
-4.35949415e-01 5.15181273e-02 1.43812513e+00 1.02645561e-01
-3.21121849e-02 5.57501949e-02 -3.17511767e-01 -8.36333260e-02
-7.00366318e-01 6.39783025e-01 3.14363033e-01 -8.02848279e-01
3.27266566e-03 -1.04231846e+00 -2.47382730e-01 1.91453829e-01
-3.81954849e-01 8.54555666e-02 3.69920015e-01 1.03147972e+00
-3.57931465e-01 -6.31300285e-02 2.09677771e-01 -6.78979039e-01
1.17593479e+00 -6.45404756e-01 -9.87026513e-01 -9.06006023e-02
-1.09799430e-01 -9.43428488e-04 7.99644828e-01 -5.28773785e-01
-8.29953492e-01 -2.89455056e-01 -2.29726173e-02 -4.85980511e-01
1.32002592e-01 5.40789425e-01 5.43197334e-01 1.80898048e-02
1.18721032e+00 2.68452764e-01 1.84694126e-01 -1.75790370e-01
5.86422503e-01 2.79647619e-01 7.75984168e-01 -7.03861237e-01
7.26961434e-01 1.87570959e-01 -5.69107309e-02 -2.92218208e-01
-8.58324826e-01 -1.07108384e-01 -2.97254741e-01 -6.10611513e-02
7.78785467e-01 -1.06973553e+00 -1.35340142e+00 6.54233694e-01
-1.38420594e+00 -1.27705467e+00 -3.22691530e-01 1.33203670e-01
-9.68279719e-01 1.15092590e-01 -1.32962525e+00 -5.87138891e-01
-1.68028653e-01 -1.16948724e+00 7.88116395e-01 -8.03385600e-02
-2.63825059e-01 -1.09035408e+00 8.10766518e-01 1.51242211e-01
4.82659429e-01 2.04596266e-01 6.02296710e-01 -9.44894314e-01
-3.50177437e-01 -1.04217641e-01 1.86390311e-01 6.92233145e-02
-1.02331944e-01 -2.78453887e-01 -4.20754820e-01 -3.73533040e-01
-2.73323983e-01 -7.79781044e-01 1.22993779e+00 2.73110360e-01
3.77816141e-01 -5.13672113e-01 -4.03894275e-01 4.96153891e-01
8.16075444e-01 1.36319458e-01 4.18887138e-01 2.35081032e-01
4.74610269e-01 2.64874488e-01 3.27697247e-01 6.06087983e-01
2.69724190e-01 7.93245256e-01 7.51159370e-01 4.58590358e-01
-6.43937141e-02 -7.05485046e-01 9.18331206e-01 4.69965965e-01
-5.62177777e-01 -3.24752271e-01 -9.46950018e-01 5.73948562e-01
-2.10605359e+00 -1.67803586e+00 3.42925310e-01 1.49429297e+00
7.77268052e-01 6.38944447e-01 9.07988489e-01 -3.93785477e-01
4.58626837e-01 5.52872002e-01 -5.21658540e-01 -4.07880127e-01
2.04646289e-01 5.40504694e-01 6.97360277e-01 7.40797937e-01
-1.30068934e+00 1.96023643e+00 7.63214302e+00 1.09855342e+00
-1.01758575e+00 3.51880193e-01 2.53802627e-01 -6.00293815e-01
2.50452489e-01 2.08448648e-01 -7.21155405e-01 3.28235149e-01
1.12314141e+00 -4.66685463e-03 1.12831342e+00 7.52724230e-01
1.95182279e-01 2.10687101e-01 -7.39131153e-01 2.19017774e-01
-2.28745386e-01 -1.68841517e+00 -3.96494716e-01 4.15551305e-01
7.84867585e-01 8.50539148e-01 9.84669775e-02 8.84015083e-01
1.97655261e+00 -1.23677397e+00 6.15949988e-01 1.26173392e-01
5.81401289e-01 -4.91151124e-01 4.00103897e-01 6.64814293e-01
-1.03005326e+00 -6.67197764e-01 -3.03917199e-01 -9.86069977e-01
2.76013911e-01 -3.20435286e-01 -1.05899835e+00 2.86204219e-01
1.84287146e-01 1.25770068e+00 -4.08118963e-01 3.63047361e-01
-6.18761241e-01 9.37920153e-01 -2.52408028e-01 -6.77619874e-02
7.57391930e-01 5.18962741e-02 7.85127640e-01 7.86773264e-01
-2.20880121e-01 1.96553856e-01 3.71937007e-01 9.31144834e-01
1.64997697e-01 -7.69054115e-01 -9.34115529e-01 -4.93457735e-01
2.22167209e-01 1.05087817e+00 -4.87069219e-01 -1.16575547e-01
1.42428549e-02 9.47438598e-01 7.28284776e-01 1.24153085e-01
-9.76059854e-01 6.48204163e-02 1.23607063e+00 -7.60398135e-02
4.09040838e-01 -5.68407357e-01 1.53709263e-01 -1.47060955e+00
-7.71471500e-01 -1.23034942e+00 3.90750378e-01 -5.18309236e-01
-1.04209924e+00 7.37805128e-01 -3.19727361e-02 -8.55723321e-01
-1.10223365e+00 -6.99738801e-01 -1.16821039e+00 4.19560581e-01
-9.24840510e-01 -1.26186049e+00 4.29484129e-01 5.90441883e-01
5.34704268e-01 -8.49379003e-01 6.30948365e-01 -3.43240261e-01
-5.57789743e-01 1.69057310e-01 -1.06174737e-01 6.06986642e-01
1.03631616e-01 -9.12497044e-01 1.27991819e+00 1.10837400e+00
3.51756662e-01 2.64578104e-01 9.09588158e-01 -9.14498031e-01
-8.74986589e-01 -9.50477660e-01 3.96146506e-01 -9.77757037e-01
1.36471140e+00 -6.16626799e-01 -1.99204355e-01 1.28822911e+00
1.26542419e-01 -3.76445323e-01 1.14405870e-01 4.31409776e-01
-3.83923292e-01 6.33493066e-01 -5.39796054e-01 1.05887699e+00
1.63562214e+00 -7.09312499e-01 -7.15653241e-01 3.37060153e-01
8.57455909e-01 -5.10798216e-01 -3.98993105e-01 1.67315975e-01
4.44794297e-01 -9.79469955e-01 1.01022732e+00 -1.65432167e+00
9.83334541e-01 2.16911837e-01 5.64019158e-02 -1.70533168e+00
-7.67874420e-01 -1.07647657e+00 -2.87266970e-01 3.75332206e-01
3.75861585e-01 -4.03870612e-01 1.15313148e+00 -2.82303710e-02
-9.69990715e-02 -4.62393284e-01 -8.38617921e-01 -9.10242081e-01
3.12452853e-01 -2.47192994e-01 1.85958609e-01 7.27937281e-01
7.09907413e-01 6.53402984e-01 -1.02160382e+00 -1.72500938e-01
2.56103724e-01 -1.03250846e-01 1.03085148e+00 -1.02919996e+00
-1.05756998e+00 -5.92492819e-01 -5.41762114e-01 -1.23035240e+00
7.45255828e-01 -8.76430809e-01 -1.00610100e-01 -1.18448973e+00
7.24186823e-02 7.97206163e-02 -8.14342871e-02 7.00594366e-01
6.29467666e-02 3.25763226e-01 4.63545859e-01 2.44308054e-01
-9.77714300e-01 6.05542064e-01 9.69886065e-01 -7.68863857e-02
-1.07197449e-01 1.74221456e-01 -6.92641437e-01 6.79343700e-01
9.40286338e-01 -5.88496566e-01 -3.82969499e-01 -4.41472024e-01
4.38701570e-01 4.34947014e-01 4.54486966e-01 -9.97621179e-01
3.33788276e-01 -6.22932434e-01 -2.69319654e-01 1.01927809e-01
7.01401353e-01 -1.35016084e-01 -1.32320866e-01 9.56033647e-01
-3.42356443e-01 2.31748521e-01 1.97214261e-01 5.83531559e-01
2.39007637e-01 4.79547009e-02 4.66315597e-01 -4.98798549e-01
-9.01325643e-01 3.43199909e-01 -1.00618291e+00 4.34062392e-01
8.98848593e-01 6.00144416e-02 -6.67628229e-01 -9.74579275e-01
-7.64090002e-01 2.42639512e-01 3.72773826e-01 3.46725643e-01
2.36921445e-01 -8.28534722e-01 -9.36962783e-01 -1.20544776e-01
-2.76509881e-01 -8.05287898e-01 -1.51233405e-01 6.69857979e-01
-7.33363688e-01 3.47394526e-01 -6.59944594e-01 -1.94491521e-02
-9.76119876e-01 3.93198013e-01 7.70941079e-01 -1.22499645e+00
-5.17426312e-01 8.85976076e-01 5.72826266e-02 -7.60618567e-01
-1.40202314e-01 -1.14779599e-01 -2.15375841e-01 -5.56669950e-01
2.07844198e-01 2.57837564e-01 -3.23249876e-01 -3.84467781e-01
-2.40628347e-01 -3.25136304e-01 -8.43332633e-02 -4.60416496e-01
1.46491015e+00 4.56476957e-01 5.93615137e-02 7.53130242e-02
4.81428295e-01 -1.66363776e-01 -1.69391418e+00 -2.26600453e-01
2.95271911e-02 -2.14608293e-02 -1.00933991e-01 -8.86246800e-01
-8.66274238e-01 6.24901235e-01 -1.19141780e-01 1.71161190e-01
5.90624392e-01 -6.83832020e-02 8.35062325e-01 8.82811606e-01
5.96551836e-01 -8.19838881e-01 4.56943840e-01 1.54899526e+00
3.30525398e-01 -8.32535505e-01 -3.59838337e-01 2.25049809e-01
-8.96113634e-01 7.10060596e-01 6.87055945e-01 -8.70462835e-01
4.39540654e-01 5.96245170e-01 -1.32164732e-01 -5.25105476e-01
-1.61983490e+00 -5.27219117e-01 -2.66728938e-01 6.79031551e-01
-1.60244238e-02 -2.79281344e-02 5.67866676e-02 8.03885579e-01
-6.41461551e-01 6.14953563e-02 7.09174514e-01 8.29848766e-01
-4.78337049e-01 -1.00583649e+00 1.80692703e-01 5.01050055e-01
-6.36068761e-01 -2.93007135e-01 -8.85376394e-01 7.30843484e-01
-1.49785891e-01 1.09904468e+00 4.26258780e-02 -9.99646008e-01
1.98034182e-01 -2.69090861e-01 6.74447119e-01 -9.10596669e-01
-1.26476848e+00 -3.81526768e-01 6.42861962e-01 -8.98854911e-01
8.20590034e-02 -3.77731681e-01 -6.62083566e-01 -8.78206134e-01
-1.11102074e-01 2.11590871e-01 -5.01325689e-02 9.77025986e-01
4.20579731e-01 6.13808990e-01 4.01031584e-01 -1.06884277e+00
-9.65129137e-01 -1.17635691e+00 -4.36147183e-01 3.58045399e-01
1.03565067e-01 -6.40086174e-01 -1.18935429e-01 -3.84585887e-01]
|
[3.667001485824585, 1.4796781539916992]
|
93f7a7c5-ae67-431c-bd65-357754d463e8
|
a-multiplicative-value-function-for-safe-and
|
2303.04118
| null |
https://arxiv.org/abs/2303.04118v1
|
https://arxiv.org/pdf/2303.04118v1.pdf
|
A Multiplicative Value Function for Safe and Efficient Reinforcement Learning
|
An emerging field of sequential decision problems is safe Reinforcement Learning (RL), where the objective is to maximize the reward while obeying safety constraints. Being able to handle constraints is essential for deploying RL agents in real-world environments, where constraint violations can harm the agent and the environment. To this end, we propose a safe model-free RL algorithm with a novel multiplicative value function consisting of a safety critic and a reward critic. The safety critic predicts the probability of constraint violation and discounts the reward critic that only estimates constraint-free returns. By splitting responsibilities, we facilitate the learning task leading to increased sample efficiency. We integrate our approach into two popular RL algorithms, Proximal Policy Optimization and Soft Actor-Critic, and evaluate our method in four safety-focused environments, including classical RL benchmarks augmented with safety constraints and robot navigation tasks with images and raw Lidar scans as observations. Finally, we make the zero-shot sim-to-real transfer where a differential drive robot has to navigate through a cluttered room. Our code can be found at https://github.com/nikeke19/Safe-Mult-RL.
|
['Luc van Gool', 'Fisher Yu', 'Alexander Liniger', 'Zhejun Zhang', 'Nick Bührer']
|
2023-03-07
| null | null | null | null |
['robot-navigation']
|
['robots']
|
[ 8.90023783e-02 4.16711152e-01 -1.93099916e-01 -4.27714318e-01
-9.40456986e-01 -4.07631904e-01 4.85571951e-01 1.27062146e-02
-1.03279102e+00 1.13111567e+00 -1.32079929e-01 -3.78171176e-01
-1.90644443e-01 -6.34765208e-01 -8.17757666e-01 -7.23410428e-01
-4.56438094e-01 5.64616203e-01 1.04286119e-01 -3.55840772e-01
1.47382528e-01 4.60000217e-01 -1.46952629e+00 -4.71036553e-01
7.97545373e-01 9.32183146e-01 3.08652669e-01 6.97194517e-01
6.14472389e-01 1.08202553e+00 -2.90255278e-01 1.66321352e-01
4.98583227e-01 -1.24527030e-01 -7.24766791e-01 -4.09702569e-01
-2.62427300e-01 -5.57148755e-01 -1.61443189e-01 7.97858179e-01
5.30116618e-01 5.79198718e-01 5.46858847e-01 -1.77484238e+00
-1.35088321e-02 4.82082874e-01 -5.64234197e-01 -3.23190719e-01
3.54201555e-01 7.03579783e-01 8.29703629e-01 -2.61779994e-01
3.94671381e-01 1.50455666e+00 1.31423086e-01 7.39016652e-01
-1.35952950e+00 -5.87592483e-01 5.50288796e-01 8.06773975e-02
-1.07418907e+00 -4.09964681e-01 3.12761486e-01 -2.82639951e-01
1.06836462e+00 -3.65831591e-02 3.89329106e-01 1.57424188e+00
4.38265145e-01 9.41190958e-01 9.25949693e-01 -5.56479543e-02
7.26803422e-01 -2.26724058e-01 -3.94723356e-01 6.87538445e-01
-5.13696694e-04 8.03785086e-01 -2.10852697e-01 -1.18008047e-01
3.56091619e-01 -4.77855355e-02 3.88201289e-02 -7.18648434e-01
-1.14248860e+00 9.86426711e-01 4.92232889e-01 -4.34444636e-01
-4.93569672e-01 6.93340063e-01 3.96665663e-01 3.55531812e-01
-1.66774113e-02 4.64040101e-01 -4.57679600e-01 -2.92805552e-01
-2.21090898e-01 7.72710264e-01 7.66968668e-01 1.04454374e+00
5.20139337e-01 1.88376665e-01 -3.64857018e-01 5.05242705e-01
4.08512980e-01 7.93387473e-01 -1.06652021e-01 -1.59108174e+00
4.10223782e-01 3.43416911e-03 7.83816993e-01 -2.58885205e-01
-7.05657840e-01 1.49756847e-02 -3.38366568e-01 1.13427210e+00
3.27023149e-01 -4.87308055e-01 -6.70610130e-01 1.95772183e+00
5.33828795e-01 9.21647251e-02 2.25435331e-01 9.73610163e-01
-4.85238843e-02 4.64139581e-01 1.32511005e-01 -2.57753551e-01
8.76214981e-01 -9.55185890e-01 -6.74489021e-01 -4.80383366e-01
6.48452818e-01 -4.43884104e-01 1.32154727e+00 8.04640353e-01
-1.03290665e+00 -9.77792442e-02 -1.03237987e+00 1.33483469e-01
-1.59007162e-02 -2.44513363e-01 3.68386894e-01 7.75889084e-02
-7.56030679e-01 6.53934419e-01 -1.03257978e+00 -4.66330983e-02
4.81465548e-01 4.73915130e-01 -1.67860091e-01 -1.36387721e-01
-9.78930235e-01 1.21783864e+00 3.47500920e-01 -1.19390756e-01
-1.61880541e+00 -3.59531671e-01 -9.89831686e-01 -1.60655007e-01
1.09364665e+00 -4.54420388e-01 1.79158235e+00 -2.94615656e-01
-1.88329971e+00 3.96694630e-01 3.54087085e-01 -5.76438725e-01
9.66507852e-01 -6.56451166e-01 1.51526868e-01 -3.11306983e-01
2.88931966e-01 6.89238846e-01 1.00624168e+00 -1.36295295e+00
-8.47176552e-01 -8.60282108e-02 2.27601409e-01 4.61155623e-01
4.41597790e-01 -3.16795647e-01 1.79644916e-02 -1.53912753e-01
-6.16076410e-01 -1.31969643e+00 -6.97293282e-01 8.89461413e-02
-3.29802364e-01 -2.45733470e-01 5.09898305e-01 -1.06507219e-01
6.31327331e-01 -2.21601248e+00 3.25071841e-01 5.83976991e-02
-1.57973260e-01 -2.98921108e-01 -3.75487775e-01 3.37025553e-01
3.90422255e-01 -3.47113401e-01 -4.81008202e-01 -5.04294157e-01
3.42078686e-01 5.56749940e-01 -5.23983896e-01 7.41025150e-01
1.07487328e-01 6.89173281e-01 -1.31591606e+00 -3.34897697e-01
2.90070742e-01 1.65798292e-01 -7.02340245e-01 5.59135139e-01
-5.88495255e-01 6.95244193e-01 -6.41349614e-01 4.80012566e-01
3.93886626e-01 4.49940532e-01 1.62703231e-01 6.53840125e-01
-3.38896483e-01 2.53894001e-01 -1.22473383e+00 1.81105340e+00
-6.28734291e-01 3.08457036e-02 5.11513889e-01 -6.49779499e-01
6.42561615e-01 -1.09062195e-01 4.61148530e-01 -9.85382617e-01
2.46445552e-01 1.28255561e-01 -2.59490222e-01 -4.45861101e-01
6.63517118e-01 -5.94812743e-02 -5.37269294e-01 5.47767162e-01
-2.59007514e-01 -5.83870530e-01 -7.31345490e-02 4.06569205e-02
1.41482604e+00 7.59863257e-01 3.61676216e-01 -1.65184006e-01
1.95494369e-01 -2.53410544e-02 7.68059194e-01 9.99586105e-01
-4.22343731e-01 -3.78321074e-02 6.59812272e-01 -1.98143587e-01
-6.80989265e-01 -1.20155787e+00 2.66386807e-01 1.44036078e+00
3.14794391e-01 -1.29959285e-02 -3.94462496e-01 -9.28219259e-01
4.15021539e-01 1.39121389e+00 -5.85146129e-01 -3.31171185e-01
-7.25395381e-01 -2.32104704e-01 4.73734736e-01 2.87314951e-01
2.05642749e-02 -1.35302353e+00 -1.52151716e+00 7.58339167e-02
-9.42702591e-02 -8.29677761e-01 -3.80334854e-01 8.23651373e-01
-4.12176639e-01 -1.05750191e+00 -2.10413501e-01 -2.61736810e-01
5.16470253e-01 1.85123906e-02 8.82083118e-01 -1.12962522e-01
-4.57578242e-01 5.97556174e-01 -3.08596939e-01 -6.20962441e-01
-2.74119347e-01 -2.25939482e-01 4.29374337e-01 -4.64373201e-01
-1.67256027e-01 -3.18156868e-01 -4.16711867e-01 2.85659403e-01
-5.58207989e-01 -1.94206655e-01 3.90878290e-01 7.33078957e-01
5.97298443e-01 -3.11739534e-01 4.38058645e-01 -5.72120726e-01
6.94609463e-01 -6.91687882e-01 -1.19855201e+00 -2.33797859e-02
-6.45125866e-01 5.10879815e-01 6.91444516e-01 -4.62108582e-01
-9.93041992e-01 3.86884779e-01 6.08965696e-04 -3.19420427e-01
-1.49243340e-01 1.55480821e-02 -5.37440926e-02 1.34052217e-01
7.42738008e-01 -6.14335835e-02 8.60906243e-02 -2.77796481e-02
5.40234864e-01 1.52945742e-01 4.48043436e-01 -1.09059501e+00
7.85760105e-01 3.42896193e-01 2.43313298e-01 -3.40825826e-01
-6.63669586e-01 -3.90894786e-02 -2.27279887e-01 -3.80171746e-01
7.03928113e-01 -8.74529004e-01 -1.50101566e+00 4.57526743e-02
-9.52528417e-01 -1.36784458e+00 -7.18529701e-01 4.44929153e-01
-1.34255970e+00 3.17056403e-02 -2.07899734e-01 -1.32134283e+00
1.23312948e-02 -1.24891853e+00 1.02733159e+00 1.68890938e-01
-1.20487176e-01 -4.69916821e-01 2.68111795e-01 -7.33201802e-02
3.69643778e-01 4.58372653e-01 5.60069442e-01 -2.28952914e-01
-5.84631085e-01 2.92179674e-01 2.59196728e-01 7.37595884e-03
-2.36425877e-01 -3.76833916e-01 -8.40598404e-01 -6.49596512e-01
-9.03576240e-02 -1.05230367e+00 6.94619417e-01 4.41918641e-01
9.25532043e-01 -4.88245040e-01 -1.90095618e-01 5.78976572e-01
1.27139544e+00 4.41565126e-01 3.49158674e-01 6.29010379e-01
2.46316016e-01 8.03855479e-01 1.39090776e+00 9.70649362e-01
5.97948313e-01 6.92242682e-01 1.21730077e+00 3.20108563e-01
5.59972644e-01 -3.09540004e-01 8.77467394e-01 -2.21821144e-01
6.50145561e-02 -3.03860396e-01 -7.77728915e-01 3.15254092e-01
-2.45419478e+00 -7.80649364e-01 2.76273817e-01 2.37238932e+00
6.81785285e-01 2.98952729e-01 2.33395085e-01 -2.88606167e-01
7.86738694e-02 5.68695627e-02 -1.21358585e+00 -5.01295388e-01
4.93795007e-01 -1.40473619e-01 8.31946552e-01 9.81817007e-01
-1.02927017e+00 1.15248537e+00 5.62729692e+00 6.20446980e-01
-8.27115774e-01 7.34820291e-02 2.77354956e-01 -8.03377986e-01
2.26987563e-02 -9.49458033e-02 -7.34257936e-01 2.52211750e-01
1.00719285e+00 -4.39502560e-02 1.04684579e+00 1.21507597e+00
5.28020024e-01 -7.29041994e-01 -1.41679478e+00 7.06346989e-01
-4.81959611e-01 -6.45222664e-01 -8.13289583e-01 9.35590118e-02
2.86337078e-01 3.92645717e-01 2.46559694e-01 7.98793852e-01
1.33507621e+00 -1.25004780e+00 1.15111268e+00 5.17396688e-01
6.80151224e-01 -1.08839738e+00 3.56657773e-01 7.35681117e-01
-9.44122970e-01 -5.46035528e-01 -1.78646952e-01 -2.27372229e-01
3.77154678e-01 8.02636445e-02 -8.69543076e-01 3.25175822e-01
7.89812803e-01 5.12480915e-01 7.17910826e-02 6.44955397e-01
-7.79468656e-01 4.88054194e-02 -4.26098436e-01 -3.43030803e-02
4.63260740e-01 -2.65507907e-01 6.24335110e-01 8.80893052e-01
1.98695436e-01 1.52700081e-01 6.18912220e-01 9.81149077e-01
2.87156612e-01 -2.87989527e-01 -8.99088323e-01 4.13056165e-01
5.08933008e-01 1.23403811e+00 -5.33115089e-01 1.44107029e-01
1.33178551e-02 8.72783899e-01 6.13301098e-01 3.75465006e-01
-1.06847167e+00 -1.94906488e-01 1.04902160e+00 -2.81893283e-01
9.48190913e-02 -4.30771530e-01 -1.67642578e-01 -7.52428234e-01
-8.06982517e-02 -8.18339586e-01 3.00226837e-01 -6.09554887e-01
-9.20200586e-01 3.40804785e-01 1.62506402e-01 -1.06921721e+00
-4.80097592e-01 -4.74550754e-01 -4.44436908e-01 5.99450469e-01
-1.83381116e+00 -8.46256077e-01 -1.85681701e-01 7.40134835e-01
6.12609386e-01 -6.19968027e-02 7.69962430e-01 -4.01676707e-02
-6.55846655e-01 2.93602794e-01 -1.43195093e-01 -4.78370458e-01
9.42557335e-01 -1.28512406e+00 2.75584847e-01 5.51193833e-01
-7.05422044e-01 2.19635606e-01 9.81185734e-01 -7.39045858e-01
-1.45613885e+00 -1.07660902e+00 1.42158702e-01 -5.14764428e-01
5.40217400e-01 -5.47936380e-01 -4.77270514e-01 8.03461313e-01
1.44650154e-02 2.00943112e-01 1.68791622e-01 -9.90152210e-02
-9.83101875e-02 1.18755244e-01 -1.37349451e+00 8.46588314e-01
1.07102466e+00 -1.16409725e-02 -3.47853631e-01 2.11863935e-01
9.45120752e-01 -5.86650372e-01 -4.22185570e-01 2.09121376e-01
5.29566944e-01 -8.43836606e-01 8.89135361e-01 -5.37345767e-01
2.82665551e-01 -3.37692857e-01 -3.05166692e-01 -1.59104693e+00
-1.94289759e-01 -8.44426453e-01 -2.49677882e-01 6.34508491e-01
2.89920896e-01 -6.66218221e-01 5.94316483e-01 7.08751678e-01
-2.17940271e-01 -7.06833720e-01 -1.16446531e+00 -1.06717324e+00
6.60835356e-02 -5.50553679e-01 4.68331248e-01 4.93606359e-01
-7.27849733e-03 1.34895176e-01 -7.37233281e-01 2.18653411e-01
9.58546877e-01 -3.45838040e-01 8.13485265e-01 -9.07888591e-01
-2.99602270e-01 -3.04039299e-01 4.50792015e-01 -7.76476622e-01
6.97371304e-01 -6.23474360e-01 8.48662555e-01 -1.35385096e+00
-1.82163611e-01 -7.81224132e-01 -2.12987125e-01 8.44723403e-01
1.16674587e-01 -5.18550098e-01 4.21685606e-01 -1.46452207e-02
-9.71726835e-01 1.08629107e+00 1.13145745e+00 -1.16090804e-01
-5.04954875e-01 1.35185152e-01 -4.42320108e-01 6.96658373e-01
1.05003858e+00 -7.68898964e-01 -5.86249292e-01 -3.00441563e-01
4.40968692e-01 3.66490453e-01 2.45791793e-01 -8.03627908e-01
2.09714264e-01 -9.30152953e-01 -1.12069383e-01 -3.88817340e-01
5.45391440e-01 -1.07621455e+00 -2.68565208e-01 9.00951385e-01
-5.56135356e-01 9.94483903e-02 2.38100186e-01 7.04802454e-01
3.71862590e-01 -1.59755588e-01 1.02666509e+00 -1.19306214e-01
-7.01813936e-01 2.43198022e-01 -6.01367772e-01 2.83113450e-01
1.46051800e+00 3.36178213e-01 -3.28818440e-01 -4.24817771e-01
-4.52746332e-01 1.11714184e+00 5.55321515e-01 4.77373958e-01
7.86049426e-01 -1.05110908e+00 -6.47142887e-01 1.29791543e-01
1.24692880e-01 3.44140500e-01 5.72846308e-02 8.29309225e-01
-1.87611192e-01 1.45146087e-01 -4.52714324e-01 -3.25654626e-01
-9.66779709e-01 6.82965815e-01 3.32983881e-01 -3.52468669e-01
-6.68323636e-01 8.67641866e-01 1.11798435e-01 -8.85122716e-01
7.01469660e-01 -3.49442542e-01 -9.30154994e-02 -9.11989585e-02
3.38513106e-01 6.21425271e-01 -4.97224331e-01 -9.15538818e-02
-4.79112923e-01 1.14508800e-01 1.11763433e-01 -5.39759099e-01
1.44801176e+00 -1.49155751e-01 3.07114065e-01 4.88891721e-01
5.88611484e-01 -2.84887791e-01 -2.09041548e+00 -2.23347880e-02
2.45638102e-01 -3.57196957e-01 -7.27160880e-03 -9.34127808e-01
-4.90862399e-01 5.50884724e-01 6.16283000e-01 -2.37998053e-01
7.66991258e-01 -2.98326731e-01 5.11550248e-01 7.39360392e-01
9.57508504e-01 -1.41298664e+00 4.20294017e-01 9.10297215e-01
1.09339118e+00 -1.52701378e+00 5.14227822e-02 3.22947502e-01
-1.14640939e+00 6.19520843e-01 8.52980971e-01 -3.16888660e-01
2.13808358e-01 6.16705298e-01 -7.62229115e-02 1.84072167e-01
-1.05577266e+00 -2.68112093e-01 -4.12187994e-01 7.51203954e-01
-2.52109975e-01 3.25366437e-01 1.44577339e-01 3.30150515e-01
-5.70183769e-02 -3.83893140e-02 6.73321247e-01 1.31129909e+00
-6.24991536e-01 -1.03932297e+00 -4.53713328e-01 1.11570619e-01
-1.45303324e-01 2.18996927e-01 -2.76463293e-02 6.96437776e-01
-9.40181687e-02 9.35525060e-01 -1.73174843e-01 -1.56009465e-01
4.45993006e-01 -1.46923751e-01 4.26852167e-01 -6.83891773e-01
-4.31619197e-01 -6.37993068e-02 1.83808148e-01 -1.35099614e+00
4.91753705e-02 -7.18333900e-01 -1.71168923e+00 -1.96863219e-01
1.37247130e-01 -8.06217343e-02 6.82322264e-01 7.57436514e-01
8.43441412e-02 7.84622371e-01 7.24711180e-01 -1.26224351e+00
-1.19318724e+00 -4.91513819e-01 -5.75226545e-01 -1.57179177e-01
8.19691837e-01 -9.13337946e-01 -3.12143594e-01 -4.68572289e-01]
|
[4.468740940093994, 1.9916068315505981]
|
c6c89078-9e4e-4a27-bf39-75a517c14227
|
neural-arabic-text-diacritization-state-of-1
|
1911.03531
| null |
https://arxiv.org/abs/1911.03531v1
|
https://arxiv.org/pdf/1911.03531v1.pdf
|
Neural Arabic Text Diacritization: State of the Art Results and a Novel Approach for Machine Translation
|
In this work, we present several deep learning models for the automatic diacritization of Arabic text. Our models are built using two main approaches, viz. Feed-Forward Neural Network (FFNN) and Recurrent Neural Network (RNN), with several enhancements such as 100-hot encoding, embeddings, Conditional Random Field (CRF) and Block-Normalized Gradient (BNG). The models are tested on the only freely available benchmark dataset and the results show that our models are either better or on par with other models, which require language-dependent post-processing steps, unlike ours. Moreover, we show that diacritics in Arabic can be used to enhance the models of NLP tasks such as Machine Translation (MT) by proposing the Translation over Diacritization (ToD) approach.
|
['Mahmoud Al-Ayyoub', "Bara' Al-Jawarneh", 'Ibraheem Tuffaha', 'Ali Fadel']
|
2019-11-08
|
neural-arabic-text-diacritization-state-of
|
https://aclanthology.org/D19-5229
|
https://aclanthology.org/D19-5229.pdf
|
ws-2019-11
|
['arabic-text-diacritization']
|
['natural-language-processing']
|
[ 2.00263426e-01 -2.34470926e-02 -6.16486706e-02 -4.66266990e-01
-6.83190703e-01 -5.05131483e-01 1.02691424e+00 8.35206360e-02
-6.23037875e-01 7.29799211e-01 5.12042880e-01 -8.05764914e-01
1.69903785e-01 -7.95385480e-01 -7.63030052e-01 -6.46008134e-01
1.27796745e-02 7.69273102e-01 -6.43778667e-02 -8.41140747e-01
1.92210227e-01 6.19677007e-01 -7.95278430e-01 5.77534258e-01
1.05944693e+00 5.95257819e-01 -1.73730642e-01 5.24935722e-01
-7.46202171e-02 1.00798750e+00 -5.01013100e-01 -8.33012283e-01
-2.70746518e-02 -3.18066925e-01 -1.19506860e+00 -4.12880212e-01
-1.15980960e-01 -3.68927866e-01 -2.77158976e-01 7.19595671e-01
4.85275924e-01 1.63082302e-01 8.94156218e-01 -5.97245097e-01
-1.28023732e+00 1.06373358e+00 -4.65011686e-01 1.44416183e-01
6.68176115e-02 -3.58846515e-01 1.00266945e+00 -1.34224129e+00
7.51568854e-01 1.39837623e+00 6.56365812e-01 3.47376019e-01
-9.57679927e-01 -1.08260937e-01 2.17117686e-02 6.28370523e-01
-1.05522680e+00 -3.95724922e-01 5.00324786e-01 1.17291473e-02
1.56990385e+00 6.95168823e-02 9.95080099e-02 1.24021590e+00
4.19171542e-01 1.23077130e+00 1.16891718e+00 -1.16328633e+00
-1.79189727e-01 -1.04084156e-01 2.74409235e-01 6.30431831e-01
-1.56239375e-01 -5.85797951e-02 -3.88229549e-01 1.82082579e-01
4.41914946e-01 -1.87046811e-01 9.87994522e-02 2.77392656e-01
-1.22792196e+00 1.05095696e+00 5.57626128e-01 7.12028444e-01
-5.18087149e-01 -1.10971324e-01 4.96684015e-01 5.06132245e-01
6.59433663e-01 3.32085073e-01 -6.88637733e-01 -1.97793379e-01
-8.91558409e-01 3.77709642e-02 7.71755874e-01 7.38870323e-01
7.46650159e-01 2.73004115e-01 -1.81959316e-01 1.08610106e+00
5.01967669e-01 5.85655868e-01 9.59456682e-01 -2.54910111e-01
8.64679456e-01 3.84234220e-01 -2.65285652e-02 -8.91458750e-01
-4.28352445e-01 -3.98829207e-02 -8.70094836e-01 -2.21644372e-01
3.67319018e-01 -6.43077970e-01 -1.10873175e+00 1.39929879e+00
-1.36949793e-01 -4.43131834e-01 2.93616354e-01 6.87992811e-01
6.82539701e-01 1.39420652e+00 -3.26254964e-02 -1.12047613e-01
1.18862915e+00 -1.50295949e+00 -1.14627481e+00 -3.83424044e-01
8.72197807e-01 -1.11107862e+00 9.16485548e-01 3.93127263e-01
-1.33493400e+00 -4.19524729e-01 -1.07638919e+00 -4.70384836e-01
-8.85371864e-01 4.27446127e-01 6.09433174e-01 6.48588777e-01
-1.16588914e+00 8.31185043e-01 -1.15063858e+00 -4.50892389e-01
-7.07204938e-02 5.01605809e-01 -3.31801772e-01 2.79698838e-02
-1.57405782e+00 1.55102980e+00 4.98735279e-01 7.84990251e-01
-5.74838936e-01 1.48887737e-02 -9.56638217e-01 -6.21228814e-02
-2.00468272e-01 -2.14704543e-01 1.08173311e+00 -1.35683346e+00
-2.20827007e+00 7.62727976e-01 -2.70979822e-01 -9.12850797e-01
4.40351404e-02 -7.27742255e-01 -4.43368852e-01 -5.80797531e-02
-5.63821673e-01 7.53937483e-01 6.79177642e-01 -9.74913597e-01
-3.38851869e-01 -2.76599526e-01 4.42858897e-02 3.02070379e-01
-3.77851427e-01 7.79663384e-01 -3.15299369e-02 -7.95013428e-01
6.65643439e-02 -1.09445918e+00 -2.59284914e-01 -7.64175057e-01
-3.60401750e-01 -3.70789409e-01 7.01699018e-01 -1.56609368e+00
1.45988595e+00 -1.76052523e+00 5.83712876e-01 2.33704504e-02
-4.12272245e-01 6.77741885e-01 -4.05318975e-01 7.87410617e-01
-1.56494722e-01 3.64176631e-02 -4.90559548e-01 -2.50598431e-01
1.86035708e-01 4.96756196e-01 -3.02298993e-01 4.17848349e-01
8.35345030e-01 1.30936038e+00 -5.47878206e-01 -1.44994110e-01
-1.80816036e-02 6.37217343e-01 -1.46462440e-01 6.92452788e-02
-3.01115870e-01 1.01503678e-01 1.56921431e-01 6.04072273e-01
7.71445811e-01 1.70081481e-01 4.83582616e-01 7.18856454e-02
-2.20856339e-01 1.01217973e+00 -6.48892164e-01 1.63360238e+00
-6.37156069e-01 7.20336914e-01 -3.18728387e-01 -1.09231639e+00
1.10159683e+00 4.12833691e-01 -1.16327576e-01 -7.62256444e-01
2.98615247e-01 4.29639727e-01 2.78675824e-01 -2.02348173e-01
7.54196048e-01 7.41951317e-02 1.60890236e-01 7.74859428e-01
3.90844494e-01 2.50335008e-01 3.03948402e-01 1.79437082e-02
6.83518648e-01 4.55366075e-01 7.84942433e-02 -2.69299120e-01
6.01767838e-01 -1.78019926e-01 3.18049908e-01 2.95515299e-01
-1.69376303e-02 6.47220731e-01 5.06602705e-01 -5.18116474e-01
-1.04946637e+00 -8.87021244e-01 -3.77066880e-02 1.49493682e+00
-4.32886928e-01 -4.23074663e-01 -8.84211123e-01 -8.84092927e-01
-5.97055554e-01 8.57834518e-01 -6.69996560e-01 -8.87899250e-02
-1.41657650e+00 -1.46955001e+00 8.70347917e-01 7.59654999e-01
4.17440563e-01 -1.50393450e+00 3.16569060e-02 5.02871692e-01
-5.20625889e-01 -8.96010220e-01 3.43830585e-02 8.12168837e-01
-9.83022451e-01 -3.39261681e-01 -9.38960016e-01 -1.28989470e+00
4.38625693e-01 -2.94406563e-01 1.16491795e+00 -2.17509732e-01
4.80979353e-01 -4.84111965e-01 -7.98308015e-01 -3.92712206e-01
-6.92090392e-01 4.66732293e-01 -9.92225558e-02 -2.87502594e-02
7.48486340e-01 -3.66672784e-01 -3.89818288e-02 -4.38838862e-02
-9.72304165e-01 -1.29150048e-01 8.16007316e-01 1.14251971e+00
2.49167606e-01 -6.16471052e-01 4.77838099e-01 -1.05269766e+00
7.37420917e-01 -3.81705970e-01 -3.73192906e-01 4.47603285e-01
-5.26914656e-01 2.93335855e-01 6.79235101e-01 -3.04838836e-01
-1.23229420e+00 -1.29163027e-01 -7.40589142e-01 2.91460037e-01
3.59746665e-02 1.01397312e+00 -2.21395455e-02 2.13948876e-01
7.33886600e-01 3.49260271e-01 -1.88120529e-01 -6.61606371e-01
6.17821276e-01 9.12829399e-01 9.34169441e-02 -3.66047412e-01
5.39145589e-01 7.04943240e-02 -3.75589520e-01 -5.85351408e-01
-6.20378375e-01 8.28384683e-02 -1.23935556e+00 3.68340462e-01
8.32108736e-01 -7.26522207e-01 -3.56695019e-02 7.63135970e-01
-1.46231890e+00 -4.86388803e-01 1.67868704e-01 5.67018032e-01
-3.36164385e-01 5.16744733e-01 -1.49894512e+00 -4.48724598e-01
-6.23463511e-01 -1.09910703e+00 7.71325529e-01 1.76233634e-01
8.42271373e-03 -1.33762574e+00 2.69034117e-01 9.16135758e-02
7.01500118e-01 -9.58342701e-02 1.35483527e+00 -9.12530482e-01
-1.06331281e-01 4.94062249e-03 -3.63589704e-01 6.46249652e-01
7.03116581e-02 3.84340018e-01 -9.56692696e-01 -2.02663705e-01
-2.85045624e-01 -5.21751404e-01 8.70127976e-01 3.63596708e-01
5.11987090e-01 -4.90532339e-01 1.25803351e-01 3.74437809e-01
1.17737007e+00 1.85673371e-01 1.20163500e+00 6.67327702e-01
5.61072588e-01 5.77408254e-01 4.68015820e-01 9.36259180e-02
4.81230378e-01 5.14691353e-01 3.23655963e-01 -1.72263607e-01
-8.30861777e-02 8.04837644e-02 9.94040728e-01 1.65372598e+00
-2.77351022e-01 -4.45833713e-01 -1.11173391e+00 6.77428663e-01
-1.83296275e+00 -5.43236017e-01 -3.42367411e-01 1.74868584e+00
1.12730074e+00 -9.02603865e-02 -9.65014566e-03 1.50455132e-01
5.92776060e-01 1.39101416e-01 1.49791420e-01 -1.52881479e+00
-5.45839310e-01 7.99696922e-01 3.47938389e-01 7.30184436e-01
-1.25735581e+00 1.69329965e+00 6.72362709e+00 8.30541372e-01
-1.20297146e+00 2.91506082e-01 6.64862454e-01 6.14724874e-01
-1.69894621e-01 -4.10802029e-02 -8.74284685e-01 1.86854392e-01
1.63341618e+00 3.36963236e-01 4.50053900e-01 3.71356726e-01
-3.12134903e-02 1.58113241e-01 -8.12980294e-01 3.14979464e-01
2.61223108e-01 -1.18765891e+00 2.00081289e-01 -2.32366100e-01
7.39923179e-01 4.41187501e-01 3.68283726e-02 6.76053107e-01
7.12988973e-01 -1.15167046e+00 3.97126079e-01 2.23452285e-01
5.23482978e-01 -9.94446456e-01 1.14560521e+00 -2.27517611e-03
-5.88875830e-01 5.56174777e-02 -6.04813576e-01 -1.09947003e-01
8.34881235e-03 2.74070740e-01 -8.99346590e-01 7.66580403e-01
4.58764225e-01 8.32108498e-01 -7.93384731e-01 5.20330131e-01
-6.57899141e-01 1.11731124e+00 -2.68233150e-01 -3.18972528e-01
9.56530869e-01 -4.24696863e-01 1.90822899e-01 1.59568763e+00
5.18255122e-02 -5.55581748e-01 -4.07443076e-01 4.41766322e-01
-1.12173595e-01 4.89707112e-01 -3.04908186e-01 -3.25350583e-01
-1.90310366e-03 1.24802458e+00 -6.65186644e-01 -2.59312928e-01
-3.32296520e-01 1.27358484e+00 6.09308958e-01 3.60510677e-01
-8.86170089e-01 -8.05574477e-01 2.72877216e-01 -5.05552828e-01
5.22340357e-01 -5.45738637e-01 -2.49609694e-01 -1.07404268e+00
-3.94316241e-02 -1.09966934e+00 2.82325059e-01 -6.61569476e-01
-1.33193278e+00 1.14608622e+00 -4.23116118e-01 -9.61524487e-01
-6.29353940e-01 -1.16781747e+00 -2.73945779e-01 1.01550019e+00
-1.74557757e+00 -1.52186549e+00 6.04329407e-01 4.53657568e-01
4.66475546e-01 -2.84518600e-01 1.25849116e+00 5.16310930e-01
-5.69306970e-01 7.06579983e-01 7.17984974e-01 6.41203344e-01
9.52878416e-01 -1.30774951e+00 7.48045623e-01 1.02957904e+00
5.60342669e-01 7.13000953e-01 1.17018633e-01 -3.19982737e-01
-1.18367708e+00 -1.16529119e+00 1.80349803e+00 -4.46313262e-01
8.31359565e-01 -5.08660913e-01 -8.98189723e-01 1.06379831e+00
9.27563727e-01 -3.68804306e-01 5.90572178e-01 4.40924674e-01
-2.06769705e-01 1.84173375e-01 -7.31241524e-01 7.09303498e-01
4.35472399e-01 -3.50386232e-01 -7.65091062e-01 5.51181674e-01
7.24967062e-01 -3.07928145e-01 -7.09535301e-01 4.12273228e-01
3.86305809e-01 -5.76014161e-01 7.36387253e-01 -9.93668616e-01
9.00158942e-01 -3.03829871e-02 -2.17730552e-01 -1.63533413e+00
-5.04778445e-01 -7.28575528e-01 -1.69342414e-01 1.17812121e+00
9.53680992e-01 -4.07362133e-01 1.95929185e-01 3.78027707e-02
-2.59570718e-01 -6.77633166e-01 -7.47207940e-01 -6.19427919e-01
7.93829083e-01 -2.32740149e-01 4.57550317e-01 1.07404602e+00
1.18121356e-01 6.69111550e-01 -5.23630679e-01 -9.61851552e-02
-4.31693971e-01 -2.36495640e-02 1.81576699e-01 -7.41623044e-01
-1.02980159e-01 -2.86382258e-01 -1.37534395e-01 -1.05130148e+00
3.52971882e-01 -1.05646288e+00 2.95856055e-02 -1.72119737e+00
-2.94493079e-01 -3.59310120e-01 -3.69403154e-01 7.99783945e-01
-3.66295390e-02 4.50344235e-01 1.65490329e-01 3.10366452e-02
-4.34815943e-01 6.66167736e-01 8.99477780e-01 -1.65051982e-01
-1.89169049e-01 -1.65329441e-01 -3.26833218e-01 7.14938581e-01
9.86455381e-01 -4.75829065e-01 3.39739919e-01 -1.14693463e+00
5.75512707e-01 -2.54571587e-01 -2.85286844e-01 -6.30674481e-01
-7.80113116e-02 2.00209454e-01 4.10311729e-01 -7.01006293e-01
2.15516582e-01 -4.28824723e-01 -7.21095324e-01 4.98321950e-01
-3.00377041e-01 8.03005040e-01 3.61014575e-01 -1.17317967e-01
-4.48334903e-01 -4.91819948e-01 6.67447925e-01 3.71281914e-02
-3.95282835e-01 -8.88391733e-02 -1.11036444e+00 -2.93626636e-01
3.53558123e-01 2.62590945e-01 -2.72432983e-01 -2.88556308e-01
-6.98498130e-01 -2.40484983e-01 -1.88557714e-01 5.02042532e-01
4.16240185e-01 -1.29179883e+00 -1.14046335e+00 1.95691556e-01
-2.17775851e-01 -3.69687617e-01 -3.72418553e-01 1.05357850e+00
-1.07035172e+00 6.87427402e-01 -4.27889287e-01 -2.43375540e-01
-8.76991451e-01 3.22597414e-01 2.07266465e-01 -8.66608799e-01
-2.25764886e-01 9.65549588e-01 -3.92315805e-01 -1.07666361e+00
1.64243862e-01 -3.00779492e-01 -5.09491444e-01 8.86073261e-02
4.57045764e-01 3.12857240e-01 7.11174548e-01 -8.19201946e-01
-3.71848494e-01 3.29318941e-02 -6.60349488e-01 -3.64123166e-01
1.63509274e+00 -2.32866891e-02 -6.48818672e-01 4.22984898e-01
9.86848414e-01 -1.24108209e-03 -6.82791591e-01 -3.99787307e-01
4.80913490e-01 1.63435683e-01 -8.51368308e-02 -1.32917130e+00
-7.88434744e-01 1.39763582e+00 4.55645353e-01 -4.78099063e-02
9.64011312e-01 -5.60877442e-01 1.05110526e+00 8.61708581e-01
-9.81948078e-02 -1.37550080e+00 -2.22971886e-01 1.45717943e+00
7.88362801e-01 -1.04019952e+00 -3.26590955e-01 3.07356846e-02
-7.46059477e-01 1.51277685e+00 4.80233401e-01 -4.67673779e-01
6.85761988e-01 4.53579187e-01 4.26650167e-01 2.89737284e-01
-7.51075625e-01 -2.60943800e-01 1.77570075e-01 5.81542432e-01
1.13763726e+00 5.78388758e-02 -8.32151711e-01 6.19771719e-01
-3.38425666e-01 -2.98193216e-01 5.23598015e-01 1.01077986e+00
-1.30403638e-01 -1.61714876e+00 -2.47215316e-01 -1.55107421e-03
-7.32348144e-01 -7.05670655e-01 -7.19576597e-01 6.58374906e-01
-7.03282803e-02 8.82455587e-01 4.09876704e-02 -3.23058128e-01
-1.36062518e-01 3.82512033e-01 6.42470598e-01 -4.26715255e-01
-1.04910445e+00 7.75636807e-02 2.61106580e-01 -1.38726667e-01
-5.36700010e-01 -5.33701062e-01 -1.12021911e+00 -3.27092230e-01
-6.09994352e-01 1.54942954e-02 8.55460405e-01 1.18074012e+00
3.20946813e-01 3.06224018e-01 6.56153023e-01 -8.02638650e-01
-4.50751334e-01 -1.64232635e+00 -2.60873646e-01 3.67992669e-02
3.93512193e-03 -1.02194138e-01 7.27665052e-02 2.99421221e-01]
|
[10.915863990783691, 10.31955623626709]
|
ee2ea8fc-54b9-4a0a-8edd-76927ac25e2e
|
deepvo-towards-end-to-end-visual-odometry
|
1709.08429
| null |
http://arxiv.org/abs/1709.08429v1
|
http://arxiv.org/pdf/1709.08429v1.pdf
|
DeepVO: Towards End-to-End Visual Odometry with Deep Recurrent Convolutional Neural Networks
|
This paper studies monocular visual odometry (VO) problem. Most of existing
VO algorithms are developed under a standard pipeline including feature
extraction, feature matching, motion estimation, local optimisation, etc.
Although some of them have demonstrated superior performance, they usually need
to be carefully designed and specifically fine-tuned to work well in different
environments. Some prior knowledge is also required to recover an absolute
scale for monocular VO. This paper presents a novel end-to-end framework for
monocular VO by using deep Recurrent Convolutional Neural Networks (RCNNs).
Since it is trained and deployed in an end-to-end manner, it infers poses
directly from a sequence of raw RGB images (videos) without adopting any module
in the conventional VO pipeline. Based on the RCNNs, it not only automatically
learns effective feature representation for the VO problem through
Convolutional Neural Networks, but also implicitly models sequential dynamics
and relations using deep Recurrent Neural Networks. Extensive experiments on
the KITTI VO dataset show competitive performance to state-of-the-art methods,
verifying that the end-to-end Deep Learning technique can be a viable
complement to the traditional VO systems.
|
['Sen Wang', 'Ronald Clark', 'Hongkai Wen', 'Niki Trigoni']
|
2017-09-25
| null | null | null | null |
['monocular-visual-odometry']
|
['robots']
|
[-3.33055228e-01 -4.02092129e-01 -3.51549357e-01 -2.34727845e-01
-2.19598755e-01 -2.63979048e-01 5.48746943e-01 -7.39426970e-01
-3.96164984e-01 4.31047827e-01 7.45386854e-02 -7.55564049e-02
1.12693541e-01 -4.24668312e-01 -7.67745495e-01 -6.42435849e-01
-1.16328783e-01 4.69629407e-01 2.14849815e-01 -3.51320505e-01
1.98023856e-01 5.75404108e-01 -1.65634573e+00 -3.37255895e-01
3.42321724e-01 1.02427554e+00 4.32270944e-01 1.10771239e+00
7.94906914e-02 1.19310856e+00 -3.39963794e-01 3.14262807e-01
4.71311748e-01 -4.76287007e-01 -7.18098462e-01 5.12404926e-02
4.46780384e-01 -2.79757112e-01 -8.76485765e-01 8.39257240e-01
8.70579660e-01 6.27820641e-02 3.98847729e-01 -1.12784028e+00
-4.33057815e-01 7.80630531e-03 -2.90427327e-01 2.09851250e-01
3.96224052e-01 3.68557185e-01 8.59278202e-01 -9.63953257e-01
1.02312624e+00 1.13155425e+00 7.01352239e-01 3.79296452e-01
-8.41816247e-01 -3.55064988e-01 1.12978734e-01 4.34850156e-01
-1.36084199e+00 -4.59172040e-01 8.29786599e-01 -3.92978519e-01
1.39599824e+00 -1.85751811e-01 1.08221936e+00 8.52670670e-01
4.38992262e-01 9.83074665e-01 5.55152595e-01 -1.77380815e-01
-3.40834409e-02 -2.19494849e-01 -3.43832225e-01 8.91754031e-01
-1.86593980e-02 4.15970474e-01 -6.89527094e-01 2.79953480e-01
1.07721972e+00 1.90739892e-02 -4.49351907e-01 -1.17528605e+00
-1.45001960e+00 6.80187762e-01 9.02046204e-01 1.76842779e-01
-4.40466464e-01 7.40002215e-01 5.34847677e-01 4.86688107e-01
2.50538379e-01 7.76875317e-02 -6.11488044e-01 -2.31397256e-01
-9.00459111e-01 2.75366843e-01 5.45325994e-01 1.00319505e+00
9.31832492e-01 3.21011871e-01 1.53823331e-01 4.77380604e-01
5.05123556e-01 3.22267652e-01 9.79952216e-01 -8.58330369e-01
3.20729643e-01 5.02724946e-01 1.41775087e-01 -1.03038085e+00
-7.78309107e-01 -5.38568258e-01 -7.60323822e-01 3.76355886e-01
4.91215214e-02 -1.71707153e-01 -1.13077056e+00 1.59993720e+00
4.80119735e-01 2.86652893e-01 4.07463424e-02 1.39622200e+00
9.53720748e-01 4.29424405e-01 -4.18578178e-01 8.68061110e-02
8.52975965e-01 -1.11559951e+00 -7.17673659e-01 -5.02544343e-01
7.11799324e-01 -8.19306612e-01 6.53400481e-01 1.86451480e-01
-8.35895360e-01 -7.12817311e-01 -1.35590196e+00 -4.02183980e-01
-3.07990253e-01 2.23840788e-01 6.97473526e-01 2.65604436e-01
-1.31248331e+00 6.85756087e-01 -1.04090512e+00 -5.42623103e-01
-9.36467797e-02 6.29510880e-01 -3.94349068e-01 1.32787660e-01
-1.01478863e+00 9.80353713e-01 1.47987381e-01 6.77318990e-01
-1.03860319e+00 -8.82561728e-02 -1.38554978e+00 -3.73375893e-01
2.36472219e-01 -1.24698651e+00 1.28649890e+00 -9.25055563e-01
-2.03582287e+00 7.26950645e-01 -4.34146166e-01 -5.60768008e-01
6.94054723e-01 -5.39448440e-01 5.88821173e-02 4.17181291e-02
1.73216034e-02 7.74239838e-01 9.88479376e-01 -8.65304649e-01
-5.63467324e-01 -2.02554077e-01 -1.92266144e-02 4.14406508e-01
3.91185075e-01 -1.85447037e-01 -6.07778490e-01 -3.30267549e-01
7.16987908e-01 -1.23518515e+00 -4.21043783e-01 5.47053851e-02
-2.39282444e-01 -4.25952189e-02 9.22778964e-01 -4.68143433e-01
8.72523963e-01 -1.61502612e+00 6.15471601e-01 -1.16109535e-01
1.26064405e-01 2.74649769e-01 2.09589422e-01 2.95852393e-01
7.54138902e-02 -4.89655793e-01 1.02362968e-01 -5.11120796e-01
4.51897755e-02 2.85905033e-01 -2.50498466e-02 1.09181798e+00
1.56215265e-01 1.20685053e+00 -9.10769522e-01 -2.07633942e-01
6.69427931e-01 6.50721610e-01 -4.91893768e-01 3.92212659e-01
-2.65010595e-01 6.85075521e-01 -1.35896459e-01 7.91050136e-01
3.19321215e-01 -3.56451958e-01 1.52024329e-01 -1.77462623e-01
-2.12921932e-01 4.68115300e-01 -1.35574329e+00 2.22833228e+00
-6.13496065e-01 9.76993382e-01 -2.66577929e-01 -9.85895038e-01
9.68276203e-01 2.33012840e-01 4.63029116e-01 -6.90221369e-01
3.81853670e-01 4.32663232e-01 -2.32516497e-01 -6.47388041e-01
7.61258483e-01 1.00565385e-02 -1.73486009e-01 5.51945344e-02
2.78771430e-01 -2.23866135e-01 -1.80505246e-01 -1.82152241e-01
7.45542347e-01 1.08765697e+00 5.70712507e-01 1.16293594e-01
7.17105329e-01 2.32417047e-01 8.32157016e-01 4.14260745e-01
-4.73510355e-01 6.96889877e-01 3.76396589e-02 -7.26749122e-01
-1.10431063e+00 -7.04070151e-01 2.22206339e-01 5.78794956e-01
6.04486823e-01 -2.34994441e-01 -2.24842787e-01 -3.24821800e-01
2.30002981e-02 -1.45912796e-01 -5.57955086e-01 -1.77836463e-01
-7.40725815e-01 -4.34195757e-01 4.09551114e-01 6.25476539e-01
7.79529035e-01 -1.00711977e+00 -1.09764278e+00 5.30325174e-01
-1.27247870e-01 -1.31307697e+00 -3.88679445e-01 1.80913866e-01
-8.55333805e-01 -1.06280386e+00 -7.83328176e-01 -8.36264968e-01
1.60389140e-01 6.56951249e-01 9.65194523e-01 -9.81933549e-02
-2.69510597e-01 2.12020576e-01 -1.82066828e-01 -5.15944175e-02
1.26646176e-01 2.69795239e-01 3.01858455e-01 -1.85728937e-01
3.30534875e-01 -5.44689357e-01 -7.27294266e-01 3.36772799e-01
-4.61966366e-01 -5.11236042e-02 6.30424261e-01 1.08651483e+00
4.81323361e-01 -5.11619508e-01 1.12574354e-01 -2.84998447e-01
-1.13561619e-02 -3.06235462e-01 -8.56128037e-01 -1.97174817e-01
-4.49847281e-01 2.76293188e-01 3.21639091e-01 -3.54213774e-01
-7.15701878e-01 6.03570342e-01 -7.75142238e-02 -9.09906387e-01
-4.30918410e-02 6.51489019e-01 1.26313102e-02 -5.10339677e-01
4.79535431e-01 2.64156878e-01 8.19396302e-02 -3.89727890e-01
4.83833373e-01 5.74166834e-01 7.20858037e-01 1.31917536e-01
9.88223732e-01 7.90744662e-01 1.60053089e-01 -8.54056060e-01
-6.74394190e-01 -8.42907906e-01 -1.23055720e+00 -1.99378461e-01
8.82148445e-01 -1.41850615e+00 -9.20879602e-01 6.13290668e-01
-1.09035087e+00 -5.21847367e-01 1.36193395e-01 7.30531156e-01
-8.73479724e-01 3.81549209e-01 -6.59085631e-01 -7.22509801e-01
-3.54192019e-01 -1.14745080e+00 1.22440398e+00 2.33840138e-01
-5.22822440e-02 -1.09606540e+00 4.06778663e-01 3.03559024e-02
3.14156592e-01 4.98907298e-01 6.74639940e-02 1.36579588e-01
-1.00080419e+00 -1.83827430e-01 -2.76329387e-02 1.85712539e-02
5.83820529e-02 -8.69270880e-03 -9.46197510e-01 -4.38034266e-01
-2.06105784e-01 -3.27962309e-01 9.56569076e-01 5.75360835e-01
4.05836552e-01 2.31716987e-02 -2.54858702e-01 1.22377622e+00
1.52452874e+00 3.52670550e-02 5.88573217e-01 8.01425397e-01
1.20895100e+00 3.14623266e-01 7.00134575e-01 3.02589089e-01
9.04494047e-01 1.07576239e+00 7.59914815e-01 7.75871705e-03
2.65194904e-02 -1.98622718e-01 7.47592747e-01 9.99091208e-01
3.17683369e-02 3.08157235e-01 -7.76982427e-01 7.94572115e-01
-2.27612257e+00 -8.37192059e-01 5.15985303e-03 1.92645705e+00
3.66540611e-01 -7.57024856e-03 2.44018197e-01 2.24983871e-01
2.93854207e-01 4.74996388e-01 -7.07021058e-01 -3.84757787e-01
-1.77020848e-01 -1.13326162e-01 6.80655718e-01 5.03314197e-01
-1.26384509e+00 1.42328632e+00 5.99910879e+00 7.23123178e-02
-1.75866151e+00 -4.22469452e-02 -1.22647099e-01 -1.13748260e-01
3.36744457e-01 1.89131945e-01 -8.25449705e-01 -8.01053196e-02
7.86025107e-01 1.47607625e-01 4.18305337e-01 1.02232170e+00
2.56374985e-01 -6.53808936e-02 -8.51781845e-01 1.27144003e+00
1.37522638e-01 -1.28073561e+00 -3.01504463e-01 1.52882487e-02
6.98999882e-01 6.15931928e-01 -1.69901490e-01 2.81269699e-01
3.42717499e-01 -1.01615107e+00 1.04965377e+00 5.65011919e-01
5.88381112e-01 -7.46960878e-01 9.83408868e-01 4.28860337e-01
-1.73544812e+00 3.05092917e-03 -7.63451993e-01 -5.93476534e-01
3.11305761e-01 3.00238639e-01 -9.65416610e-01 9.01878655e-01
6.99674428e-01 1.27671170e+00 -5.05818069e-01 1.16965330e+00
-4.60908026e-01 -8.66703615e-02 -2.52647936e-01 -1.00803092e-01
5.24968266e-01 -2.68151215e-03 5.15453458e-01 9.71892655e-01
2.62799501e-01 -3.74137282e-01 1.78109884e-01 5.42708457e-01
1.81730404e-01 -1.65624753e-01 -1.01127994e+00 1.38745204e-01
6.19866624e-02 1.10884035e+00 -3.25067580e-01 -2.73337811e-01
-3.81034493e-01 1.22814906e+00 4.38840628e-01 2.59715110e-01
-6.56478345e-01 -4.91792411e-01 1.08084631e+00 -4.05262560e-01
7.42507219e-01 -7.04215586e-01 -1.33449316e-01 -1.46226358e+00
-4.76744398e-02 -5.87849975e-01 1.44203842e-01 -9.49097097e-01
-8.36975813e-01 6.90886378e-01 -3.79211962e-01 -1.62281346e+00
-8.92091513e-01 -6.75162971e-01 -4.39226598e-01 8.32416058e-01
-1.90878880e+00 -1.14202225e+00 -7.25143850e-01 6.92775607e-01
4.38594460e-01 9.09704808e-03 4.27128226e-01 1.91141143e-01
-4.84479129e-01 2.29701012e-01 -1.17193617e-01 9.89627019e-02
5.29345632e-01 -1.15352857e+00 7.97437906e-01 9.99020457e-01
3.07558239e-01 3.91101956e-01 8.44838679e-01 -5.05150557e-01
-1.87786078e+00 -1.07656753e+00 9.37586248e-01 -5.25434554e-01
4.33466434e-01 -2.35756740e-01 -5.57532370e-01 9.48634565e-01
8.66502896e-02 3.23770434e-01 -1.03326604e-01 -2.59511799e-01
-4.51907441e-02 5.17638624e-02 -6.93071425e-01 6.84645832e-01
1.43535423e+00 -7.17275918e-01 -4.62845862e-01 1.86190218e-01
6.50830328e-01 -1.10110474e+00 -6.57704234e-01 5.39666653e-01
6.20009422e-01 -1.04199529e+00 1.18015349e+00 -5.13760388e-01
1.38493791e-01 -7.55746841e-01 -1.23208232e-01 -1.24650443e+00
-2.54050404e-01 -8.61209571e-01 -3.66299629e-01 6.12217009e-01
1.23906545e-02 -4.77879047e-01 8.81836593e-01 -3.96470614e-02
-7.93052167e-02 -6.16772234e-01 -1.01949859e+00 -7.83423424e-01
-3.92496526e-01 -4.79733318e-01 4.98048395e-01 7.79627323e-01
-2.04005048e-01 4.24318105e-01 -9.52358484e-01 4.28500235e-01
5.16998947e-01 3.44820201e-01 1.49556553e+00 -9.86898303e-01
-2.79022783e-01 -2.39985839e-01 -1.06883788e+00 -1.50536633e+00
1.07861064e-01 -6.79268599e-01 2.66625911e-01 -1.69681478e+00
-3.47254187e-01 -3.38148661e-02 1.09164258e-02 2.17010796e-01
-3.92836556e-02 4.85223740e-01 2.43936136e-01 4.16043848e-01
-5.00223815e-01 8.50727320e-01 1.23275006e+00 1.41717315e-01
-4.94173199e-01 -1.18554376e-01 -7.95409158e-02 7.68996894e-01
5.55906355e-01 -3.11835408e-01 -2.29369819e-01 -3.90654266e-01
2.10953608e-01 1.30901232e-01 5.94661891e-01 -1.01396930e+00
6.04750276e-01 1.65551789e-02 5.23135304e-01 -9.21202540e-01
5.64707041e-01 -6.43741727e-01 1.91618279e-01 6.88688159e-01
3.02305698e-01 3.73001009e-01 7.97817186e-02 5.01203239e-01
-3.70313108e-01 1.68558389e-01 6.35764778e-01 -2.73412168e-01
-1.28057265e+00 4.48146522e-01 -2.48155907e-01 -1.80702507e-01
8.68195295e-01 -3.73703837e-01 -4.28537056e-02 -6.26805604e-01
-6.61129057e-01 2.34832302e-01 6.14040494e-01 4.80136395e-01
6.57116592e-01 -1.39829290e+00 -2.09331647e-01 2.64256537e-01
2.01158240e-01 2.74580538e-01 2.04243921e-02 8.80758941e-01
-9.54320014e-01 7.12636292e-01 -2.77810723e-01 -1.18227470e+00
-1.02332520e+00 4.97881502e-01 7.49661982e-01 -1.51034653e-01
-9.38586533e-01 7.31625855e-01 -6.69225305e-02 -7.60307312e-01
2.51642376e-01 -5.16394675e-01 -1.83109328e-01 -7.36944601e-02
1.34400219e-01 1.71586841e-01 -6.20610341e-02 -9.88883197e-01
-5.10081291e-01 1.08227479e+00 3.37925673e-01 -1.75952211e-01
1.36017835e+00 -5.49063027e-01 2.67493784e-01 6.64950311e-01
1.25898707e+00 -4.29102033e-01 -1.53012502e+00 -4.31884646e-01
1.87877029e-01 -4.48806047e-01 1.85774595e-01 -1.85052752e-01
-1.16657376e+00 1.00183630e+00 5.19025385e-01 -4.49201256e-01
9.34541643e-01 -3.16627562e-01 7.87436664e-01 5.21692693e-01
6.29996121e-01 -9.59467888e-01 1.31288946e-01 9.49841142e-01
7.46476412e-01 -1.30304909e+00 4.07977290e-02 -1.41551286e-01
-6.47573054e-01 1.22383654e+00 4.21184748e-01 -5.61942637e-01
6.63185120e-01 -5.41275889e-02 6.12096786e-01 -2.32662663e-01
-8.63890767e-01 -6.09382331e-01 3.77797365e-01 3.92853320e-01
3.46712828e-01 -2.78981984e-01 8.93280208e-02 -7.74895102e-02
-3.55320185e-01 2.18531623e-01 4.81050938e-01 1.15467608e+00
-3.86657119e-01 -8.76812398e-01 -1.78651169e-01 -2.35116988e-01
1.27851442e-02 2.02333078e-01 -1.88169688e-01 1.08546734e+00
-8.42473581e-02 5.44683099e-01 -1.18559539e-01 -6.27432466e-01
4.16626364e-01 -9.66332108e-02 5.76458812e-01 -3.45434606e-01
-6.79373980e-01 1.74467728e-01 5.06438650e-02 -9.89780545e-01
-6.85861468e-01 -6.76776350e-01 -1.19771600e+00 -3.99609804e-01
-1.93688437e-01 -4.16366845e-01 7.97107041e-01 1.20031881e+00
3.85056794e-01 6.72888935e-01 5.86769938e-01 -1.64307165e+00
-3.96300077e-01 -9.17848170e-01 -2.84815937e-01 7.54585266e-02
7.35963106e-01 -9.89327013e-01 -1.67789429e-01 -1.49767563e-01]
|
[8.196529388427734, -2.1178882122039795]
|
ee6c1615-4b90-4141-b32c-14b53ababff5
|
decentralized-multi-agent-reinforcement-5
|
2306.12926
| null |
https://arxiv.org/abs/2306.12926v1
|
https://arxiv.org/pdf/2306.12926v1.pdf
|
Decentralized Multi-Agent Reinforcement Learning with Global State Prediction
|
Deep reinforcement learning (DRL) has seen remarkable success in the control of single robots. However, applying DRL to robot swarms presents significant challenges. A critical challenge is non-stationarity, which occurs when two or more robots update individual or shared policies concurrently, thereby engaging in an interdependent training process with no guarantees of convergence. Circumventing non-stationarity typically involves training the robots with global information about other agents' states and/or actions. In contrast, in this paper we explore how to remove the need for global information. We pose our problem as a Partially Observable Markov Decision Process, due to the absence of global knowledge on other agents. Using collective transport as a testbed scenario, we study two approaches to multi-agent training. In the first, the robots exchange no messages, and are trained to rely on implicit communication through push-and-pull on the object to transport. In the second approach, we introduce Global State Prediction (GSP), a network trained to forma a belief over the swarm as a whole and predict its future states. We provide a comprehensive study over four well-known deep reinforcement learning algorithms in environments with obstacles, measuring performance as the successful transport of the object to the goal within a desired time-frame. Through an ablation study, we show that including GSP boosts performance and increases robustness when compared with methods that use global knowledge.
|
['Carlo Pinciroli', 'Apratim Mukherjee', 'Pranjal Paliwal', 'Joshua Bloom']
|
2023-06-22
| null | null | null | null |
['multi-agent-reinforcement-learning']
|
['methodology']
|
[-5.22910990e-02 9.98339131e-02 -5.38406298e-02 7.36397505e-02
-3.16525638e-01 -5.38163245e-01 6.72169626e-01 3.38866025e-01
-8.04880619e-01 1.16906548e+00 -3.75737041e-01 -1.95533231e-01
-4.37927842e-01 -7.52069771e-01 -1.04377353e+00 -1.29610503e+00
-7.59735465e-01 8.12647402e-01 4.34728980e-01 -3.98843497e-01
2.29252484e-02 4.53666538e-01 -1.19470680e+00 -3.52453142e-01
4.90404934e-01 8.44179869e-01 4.45000350e-01 9.91682887e-01
4.06644166e-01 1.09784126e+00 -9.80491936e-01 2.77013391e-01
2.16288626e-01 -3.91001940e-01 -9.63552177e-01 2.28316948e-01
-3.00971836e-01 -4.33465332e-01 -2.46286169e-01 7.16267526e-01
3.83046895e-01 2.23273247e-01 3.70007157e-01 -1.59596753e+00
-1.31610096e-01 5.19139767e-01 -3.01724851e-01 -7.64207467e-02
1.01853758e-02 5.42601109e-01 5.12099981e-01 1.84515566e-01
5.14081240e-01 1.17458189e+00 7.50642717e-01 5.71748614e-01
-1.25597751e+00 -3.70075613e-01 3.14656317e-01 -1.26231655e-01
-9.09989238e-01 -3.55665058e-01 2.48940542e-01 -2.29699165e-01
1.05438077e+00 -2.40991324e-01 7.46142268e-01 1.05246997e+00
6.84874833e-01 5.60092211e-01 1.12137651e+00 -2.45431796e-01
7.59581327e-01 -2.57398158e-01 -4.39946264e-01 7.95176029e-01
3.25135916e-01 3.79130840e-01 -3.02907944e-01 -2.05006436e-01
8.67029130e-01 -2.67913789e-01 7.55045637e-02 -4.32399303e-01
-1.36708939e+00 7.69046724e-01 4.62309361e-01 5.36078028e-02
-7.80965090e-01 7.59318829e-01 3.12538475e-01 8.53540003e-01
2.34142952e-02 7.00296104e-01 -7.10867226e-01 -2.31502995e-01
-1.56381175e-01 2.47465596e-01 1.25365734e+00 6.49610758e-01
1.01087224e+00 1.57457441e-01 3.94891083e-01 2.79869825e-01
3.00099730e-01 5.80492318e-01 1.66090012e-01 -1.53569162e+00
5.62906004e-02 1.84081405e-01 6.52762711e-01 -8.12621117e-01
-6.90575242e-01 -2.75326759e-01 -6.66816592e-01 7.04693198e-01
2.98692763e-01 -9.68732238e-01 -6.67148769e-01 2.01295567e+00
5.81139803e-01 2.08095878e-01 6.88318551e-01 6.01947188e-01
-1.06496945e-01 7.62250900e-01 -2.63558012e-02 -3.49260598e-01
7.47577190e-01 -9.78820801e-01 -4.47841734e-01 -4.83256698e-01
7.59854674e-01 -2.70572752e-01 1.94322765e-01 4.80090320e-01
-8.99721444e-01 -2.25996703e-01 -1.00839889e+00 7.04107583e-01
-3.32640886e-01 -3.30515504e-01 6.17130578e-01 1.57922775e-01
-1.45897329e+00 8.45983446e-01 -1.53856194e+00 -6.97483182e-01
1.49069190e-01 9.66468394e-01 -2.74249911e-01 2.61113495e-01
-9.08622384e-01 1.20367038e+00 4.14747447e-01 2.15802528e-03
-1.61280107e+00 -3.01754232e-02 -6.94957852e-01 -2.42911905e-01
7.21398056e-01 -6.20396852e-01 1.72541356e+00 -1.06584144e+00
-2.05360746e+00 3.17623466e-02 2.96979755e-01 -6.77973986e-01
3.53281289e-01 -8.04923773e-02 1.54647171e-01 7.62472153e-02
2.54534245e-01 7.86246121e-01 6.94718659e-01 -1.64989293e+00
-1.05380440e+00 -1.99694224e-02 4.35748249e-01 4.27503735e-01
-1.50373057e-01 -3.67512941e-01 -1.85128018e-01 -8.77233744e-02
-1.80696115e-01 -1.34447634e+00 -7.32540965e-01 -2.18763545e-01
-2.54270062e-02 -3.25797588e-01 1.05325413e+00 -1.48740709e-01
3.93042713e-01 -1.81503284e+00 4.76142675e-01 1.21022142e-01
2.52884477e-02 7.42687732e-02 -4.37292635e-01 1.02268374e+00
6.07626021e-01 -1.20771274e-01 -1.33751139e-01 -6.77437246e-01
4.99020144e-03 1.07316780e+00 -8.74191895e-03 5.65450311e-01
1.82518199e-01 5.68836749e-01 -1.31554329e+00 -3.53082180e-01
4.81526507e-03 2.81358689e-01 -4.84730095e-01 1.62265792e-01
-4.56329525e-01 1.00330675e+00 -7.18883157e-01 1.17566220e-01
2.48625159e-01 -3.57016474e-01 6.65903151e-01 4.74905819e-01
-5.18532217e-01 2.48135015e-01 -1.14482856e+00 1.50365019e+00
-6.24701798e-01 4.29070324e-01 7.58842587e-01 -1.23562360e+00
5.77985168e-01 4.83078599e-01 9.68315125e-01 -4.68284935e-01
2.76418239e-01 2.61170864e-01 1.20433770e-01 -4.70343351e-01
3.54443699e-01 -9.43392515e-02 -2.19480976e-01 6.00157738e-01
1.84391171e-01 -2.55784005e-01 3.61887246e-01 7.72707835e-02
1.63621831e+00 2.80416440e-02 4.14496996e-02 -1.53014719e-01
4.97393794e-02 3.73758614e-01 6.43667936e-01 1.27371514e+00
-1.92692816e-01 -3.15094352e-01 5.89727700e-01 -3.73864800e-01
-8.22463214e-01 -8.76148343e-01 5.16515911e-01 1.05938244e+00
6.26324236e-01 -1.45897761e-01 -4.34552073e-01 -7.00068474e-01
1.07273251e-01 4.78637546e-01 -7.73420334e-01 -1.80120751e-01
-7.92206049e-01 -7.17005968e-01 3.90034199e-01 1.44638225e-01
5.84636986e-01 -1.33128881e+00 -1.29687452e+00 5.94256580e-01
2.01254636e-01 -1.04274905e+00 5.43788783e-02 7.12131739e-01
-6.22874916e-01 -1.02190650e+00 -3.73623222e-01 -9.28309858e-01
6.66242898e-01 1.98248878e-01 7.82176256e-01 2.18359590e-01
1.23433739e-01 9.68210518e-01 -4.42686409e-01 -2.58851320e-01
-8.56015086e-01 1.49768338e-01 3.12253475e-01 -2.57304698e-01
-4.38176721e-01 -5.42607546e-01 -4.37743336e-01 4.46734250e-01
-6.34040415e-01 -1.42038345e-01 7.76076317e-01 8.24692070e-01
2.26002485e-01 4.15813476e-01 5.98338783e-01 -1.81135103e-01
6.99711442e-01 -6.32715225e-01 -8.16328764e-01 -1.79371648e-02
-1.82029322e-01 1.36212870e-01 6.92767143e-01 -7.14800179e-01
-7.72987008e-01 1.27494186e-01 2.15180293e-01 -1.59128264e-01
-2.04101011e-01 5.09461641e-01 4.56051916e-01 -3.67317528e-01
2.80981034e-01 3.21015835e-01 5.79122603e-01 -1.01041965e-01
-7.59325996e-02 3.45699281e-01 2.06898972e-01 -8.02398920e-01
5.90907514e-01 5.48198998e-01 3.30520511e-01 -8.86094213e-01
-5.36140203e-01 -1.53012514e-01 -4.74387676e-01 -3.55387002e-01
7.50689685e-01 -6.36139035e-01 -1.47320807e+00 6.23026729e-01
-1.08913350e+00 -1.30715728e+00 -4.59700763e-01 4.85300869e-01
-8.84895265e-01 1.12556256e-01 -6.57616317e-01 -8.69919121e-01
1.05430685e-01 -1.23703218e+00 8.63709569e-01 2.70448476e-01
2.08435774e-01 -1.25977898e+00 4.61581051e-01 -3.28463525e-01
7.04930961e-01 2.74641544e-01 3.53494525e-01 -6.16064191e-01
-4.87248898e-01 4.32967357e-02 3.34864199e-01 3.36504847e-01
3.14941138e-01 -2.82991290e-01 -2.35288516e-01 -8.68669629e-01
-4.48462367e-02 -7.25371599e-01 3.05066615e-01 2.30528966e-01
3.57607305e-01 -4.97383386e-01 -5.71303964e-01 -2.10603438e-02
1.29180455e+00 4.24155563e-01 1.48304597e-01 7.13157535e-01
2.33927533e-01 5.96291780e-01 6.04797363e-01 7.33924806e-01
6.89720392e-01 5.07865787e-01 9.46605563e-01 1.47763416e-01
1.95438787e-01 8.84293690e-02 7.62523949e-01 6.11461341e-01
-7.61473626e-02 -5.40449321e-01 -8.76930356e-01 4.66764987e-01
-2.20374322e+00 -7.55662262e-01 2.09374607e-01 2.08036947e+00
5.57175219e-01 4.50387262e-02 9.55121443e-02 -3.33537787e-01
5.51215470e-01 -1.01409979e-01 -7.86432624e-01 -2.67955393e-01
9.14434716e-02 -2.63509095e-01 8.89422953e-01 7.22948551e-01
-1.14260864e+00 1.02688134e+00 6.03580999e+00 2.67947108e-01
-1.28859985e+00 2.25296915e-01 1.22605763e-01 2.15232037e-02
5.75749874e-01 2.02932701e-01 -5.09259284e-01 2.72614390e-01
9.60093677e-01 2.63050348e-01 8.81524205e-01 5.31916916e-01
3.51596802e-01 -5.84028721e-01 -9.64445829e-01 3.12513024e-01
-4.24606830e-01 -1.09128892e+00 -3.88876498e-01 3.76696438e-01
9.50706482e-01 5.77510595e-01 -1.45069182e-01 5.06479204e-01
1.29717839e+00 -6.28794551e-01 8.19903135e-01 4.04392391e-01
7.10685775e-02 -6.26792610e-01 7.18405724e-01 8.36556435e-01
-1.01314116e+00 -5.75231731e-01 -1.96793884e-01 -4.30259734e-01
2.28628948e-01 -2.23188266e-01 -9.81522143e-01 5.63299119e-01
7.11018801e-01 6.07241154e-01 -9.92298350e-02 9.84944046e-01
-6.20817803e-02 5.54852426e-01 -8.41075003e-01 -3.72519076e-01
8.09637785e-01 -1.57487050e-01 6.71631753e-01 7.59971976e-01
2.31000602e-01 -6.09487332e-02 7.60833621e-01 3.04017842e-01
1.84878677e-01 -4.66921568e-01 -5.42435884e-01 1.12260051e-01
3.40417743e-01 9.91291285e-01 -7.69375920e-01 -2.81627983e-01
-2.66627669e-01 8.73573601e-01 4.97811347e-01 3.70895386e-01
-6.68598831e-01 -1.29761338e-01 7.85370052e-01 -5.16273975e-01
4.86370862e-01 -7.79425263e-01 3.94559383e-01 -7.39169121e-01
-3.97438407e-01 -7.43805647e-01 1.09686323e-01 -3.75441730e-01
-1.10485709e+00 3.31873059e-01 -1.01816930e-01 -9.91808236e-01
-3.79475236e-01 -3.79040629e-01 -5.60687065e-01 1.53234050e-01
-1.51576567e+00 -1.06584275e+00 -7.99900666e-02 5.44772446e-01
4.78258938e-01 -1.29138723e-01 7.31453836e-01 -1.66956037e-01
-5.33621788e-01 -6.49869591e-02 4.58677143e-01 -7.43196607e-02
4.29187924e-01 -1.16774213e+00 -9.36376974e-02 4.57599193e-01
-3.53079140e-01 3.52291852e-01 1.00219727e+00 -8.31836283e-01
-1.85160887e+00 -1.04258299e+00 4.84543405e-02 -2.37164423e-01
9.42239702e-01 -1.46999016e-01 -5.58533788e-01 9.80879843e-01
4.71977621e-01 -2.01192405e-03 1.00217842e-01 -1.53369561e-01
4.55216467e-01 3.79112782e-04 -8.61950696e-01 4.75453615e-01
5.93740821e-01 1.97576657e-01 -1.64346382e-01 4.12708133e-01
6.28554940e-01 -2.08868295e-01 -9.36354280e-01 3.56699556e-01
2.76614070e-01 -6.25044227e-01 5.43003619e-01 -5.12878060e-01
-2.07511619e-01 -3.31102103e-01 -1.10460103e-01 -1.80287409e+00
-1.03660852e-01 -1.03334236e+00 -7.51994178e-03 7.65142739e-01
2.54341722e-01 -9.89490926e-01 7.88029253e-01 1.76213998e-02
-2.93520838e-01 -5.38966477e-01 -1.24368751e+00 -1.03434670e+00
1.99627891e-01 -4.31341352e-03 2.54183531e-01 6.48250759e-01
3.99203151e-02 2.24399537e-01 -4.50352430e-01 7.83483744e-01
5.64640820e-01 -1.32909864e-01 1.12562621e+00 -9.39726233e-01
-4.59387273e-01 -2.12556139e-01 -1.22529060e-01 -1.29600871e+00
2.91973501e-01 -3.24077278e-01 6.90660954e-01 -1.73251629e+00
-2.10667744e-01 -9.09689426e-01 -3.84640843e-02 6.83398783e-01
4.98291999e-01 -1.26911923e-01 2.24527538e-01 3.50649029e-01
-1.28884411e+00 8.57278228e-01 1.34222066e+00 -5.95312454e-02
-4.05288398e-01 1.77777007e-01 -1.57338813e-01 6.47080898e-01
1.25575566e+00 -7.54732847e-01 -3.24352562e-01 -7.36432076e-01
2.98548728e-01 3.88738424e-01 5.29820085e-01 -1.34151852e+00
7.05551803e-01 -4.75515097e-01 -2.55181164e-01 4.42787074e-02
7.51181662e-01 -8.61421943e-01 5.63051887e-02 1.06127810e+00
-2.39684790e-01 2.91569024e-01 1.71843901e-01 9.21437383e-01
8.40282813e-02 -3.02018315e-01 6.15633130e-01 -2.55610555e-01
-6.46845281e-01 9.62986574e-02 -1.09616268e+00 -2.15300158e-01
1.41601753e+00 2.48085454e-01 -3.63908768e-01 -6.16107106e-01
-8.30496728e-01 7.92595029e-01 6.32787347e-01 1.44316480e-01
3.68776977e-01 -5.77323556e-01 -5.06730556e-01 -8.35615471e-02
-4.13423061e-01 5.55329509e-02 -6.33263066e-02 1.08289659e+00
-5.03686070e-01 1.25052333e-01 -2.12879479e-01 -6.17824793e-01
-8.27575922e-01 4.54451710e-01 7.71304309e-01 -3.80511582e-01
-3.20868343e-01 4.31498498e-01 -1.46876231e-01 -6.14848018e-01
1.41406372e-01 -1.10720478e-01 -4.49108966e-02 -2.71161139e-01
-7.11490214e-02 3.13217521e-01 -2.90479541e-01 -4.32541519e-01
-8.92807245e-02 3.63107800e-01 7.18857870e-02 -4.06112522e-01
1.57132399e+00 -4.00154859e-01 -3.18723023e-01 4.63747472e-01
7.14330375e-01 -3.57722431e-01 -1.92209840e+00 1.67290233e-02
2.79016308e-02 1.15015708e-01 3.12438086e-02 -7.01451659e-01
-8.84161174e-01 3.89437526e-01 3.56948704e-01 8.07842970e-01
6.49852633e-01 1.59775779e-01 5.68533123e-01 1.00875676e+00
8.93500984e-01 -1.14446497e+00 4.26353723e-01 1.11326146e+00
6.10552132e-01 -1.32706130e+00 -2.18341932e-01 2.20882431e-01
-6.08988762e-01 1.05762768e+00 7.23559976e-01 -3.40177804e-01
5.42632401e-01 4.74162608e-01 2.38479464e-04 -1.17561780e-01
-1.20238447e+00 -3.58506292e-01 -8.35992932e-01 7.46755779e-01
-5.57210803e-01 -5.89317605e-02 3.29773545e-01 -1.77078426e-01
5.49720973e-02 -1.26062319e-01 7.32695997e-01 1.68369591e+00
-8.36575508e-01 -1.08228850e+00 -3.36064428e-01 1.87672555e-01
-1.69869363e-01 5.33821344e-01 -1.34049907e-01 9.92534935e-01
6.61697164e-02 1.32025337e+00 2.18806028e-01 -2.00075537e-01
2.38180961e-02 -5.02334893e-01 4.37252492e-01 -4.11019266e-01
-5.08756578e-01 -1.26805410e-01 1.81409910e-01 -5.51879585e-01
-6.03197455e-01 -8.36100757e-01 -1.38310516e+00 -2.95469433e-01
-1.94336712e-01 5.03719866e-01 9.16375279e-01 1.09300625e+00
4.40609574e-01 6.23044610e-01 7.07621932e-01 -1.39810407e+00
-7.92119503e-01 -7.85859883e-01 -2.60167807e-01 -9.87114161e-02
9.51223969e-01 -8.65666866e-01 -4.98252094e-01 -1.35274768e-01]
|
[4.014683246612549, 1.973343014717102]
|
311fa6ee-d60d-4be4-a68d-e292bda210c2
|
bayesian-inference-and-neural-estimation-of
|
2305.17749
| null |
https://arxiv.org/abs/2305.17749v1
|
https://arxiv.org/pdf/2305.17749v1.pdf
|
Bayesian inference and neural estimation of acoustic wave propagation
|
In this work, we introduce a novel framework which combines physics and machine learning methods to analyse acoustic signals. Three methods are developed for this task: a Bayesian inference approach for inferring the spectral acoustics characteristics, a neural-physical model which equips a neural network with forward and backward physical losses, and the non-linear least squares approach which serves as benchmark. The inferred propagation coefficient leads to the room impulse response (RIR) quantity which can be used for relocalisation with uncertainty. The simplicity and efficiency of this framework is empirically validated on simulated data.
|
['Hong Ge', 'Yuhang He', 'Yongchao Huang']
|
2023-05-28
| null | null | null | null |
['room-impulse-response', 'bayesian-inference']
|
['audio', 'methodology']
|
[ 3.26710671e-01 1.46958396e-01 6.70492530e-01 -5.54627895e-01
-7.02796161e-01 1.68168113e-01 4.75321054e-01 -1.14268521e-02
-4.88212645e-01 8.92286539e-01 6.06636629e-02 -4.57087755e-01
-7.32577801e-01 -8.06176901e-01 -3.77566814e-01 -1.02434063e+00
-2.21861795e-01 6.26446381e-02 3.22813869e-01 -1.47187188e-01
2.20657676e-01 6.23491764e-01 -1.58205616e+00 -1.62182286e-01
5.69712520e-01 1.29475534e+00 1.67382821e-01 8.36565673e-01
8.98091495e-02 6.61598265e-01 -5.20471275e-01 1.55455217e-01
-1.42609045e-01 -2.29189590e-01 -5.87891579e-01 -7.16134429e-01
-1.39486015e-01 -2.62615025e-01 -2.99849033e-01 8.59230161e-01
6.46019220e-01 5.90882301e-01 1.03086865e+00 -7.67808437e-01
-5.77995889e-02 6.44432783e-01 1.86820000e-01 1.18481822e-01
5.50145745e-01 -2.14532286e-01 6.95383430e-01 -5.56278884e-01
-2.45211422e-02 1.23509514e+00 8.78755450e-01 4.20399547e-01
-1.22527826e+00 -2.64643282e-01 -5.65739334e-01 3.88280928e-01
-1.45068336e+00 -5.36040127e-01 1.04068828e+00 -4.16620463e-01
8.43460321e-01 4.76320803e-01 2.98366368e-01 1.00839543e+00
1.15309462e-01 3.32189620e-01 1.43168902e+00 -1.03859699e+00
6.20122433e-01 5.02595484e-01 3.32066327e-01 6.08120978e-01
-2.02161625e-01 7.68035412e-01 -6.49146616e-01 -3.56831461e-01
3.94728780e-01 -7.01055765e-01 -3.95841986e-01 7.82911479e-02
-6.09869301e-01 5.86921632e-01 3.31875175e-01 5.35323501e-01
-3.98566186e-01 3.43303651e-01 8.86421055e-02 1.41399711e-01
4.17127043e-01 2.08861634e-01 -1.37504026e-01 3.83855738e-02
-9.35493946e-01 1.17452137e-01 1.12357211e+00 3.84447992e-01
6.24535382e-01 2.35600725e-01 9.83745307e-02 1.08542609e+00
1.18383682e+00 8.94109607e-01 1.92694608e-02 -8.88704538e-01
-1.19444482e-01 -3.44777972e-01 5.66597395e-02 -9.53447282e-01
-5.94797373e-01 -6.59573525e-02 -6.22154117e-01 4.28273380e-01
3.31002563e-01 -1.70860842e-01 -6.73538268e-01 1.60043538e+00
2.85584509e-01 4.36346412e-01 1.69759467e-01 5.72573066e-01
1.00099230e+00 9.33289766e-01 -9.45400670e-02 -2.87795246e-01
9.93860602e-01 -5.49665391e-01 -9.53786373e-01 3.11871078e-02
3.01661342e-02 -7.36815691e-01 4.35008258e-01 7.83002734e-01
-1.00836265e+00 -6.36966527e-01 -1.11931574e+00 4.63956475e-01
-4.40959752e-01 -9.45182666e-02 1.48442984e-01 1.05967951e+00
-1.07388937e+00 9.24577534e-01 -7.30366707e-01 4.92203087e-02
-3.25293899e-01 2.91062772e-01 1.25821277e-01 3.71208906e-01
-1.40244496e+00 1.02588773e+00 5.63433051e-01 6.74904764e-01
-6.51532114e-01 -7.89814770e-01 -6.84331000e-01 -6.30599707e-02
-5.20995371e-02 -4.21716928e-01 1.25641918e+00 -2.56675392e-01
-2.31487131e+00 2.32816592e-01 2.65253596e-02 -2.84790844e-01
3.07217389e-01 -1.79111913e-01 -8.55300725e-01 1.15374148e-01
-5.88759840e-01 -5.71469031e-02 1.10510945e+00 -1.75690556e+00
-1.95514262e-01 6.20571338e-03 -3.63793701e-01 -2.42975861e-01
4.04168069e-02 -7.79280737e-02 9.51749757e-02 -2.39537179e-01
4.96166408e-01 -6.86541259e-01 -6.35388866e-02 -3.48133296e-01
-7.44135201e-01 -1.63835287e-01 3.37784618e-01 -9.79167581e-01
1.10960162e+00 -2.12874532e+00 7.74365887e-02 9.57487524e-01
-1.04624450e-01 2.03985274e-01 1.16228461e-01 7.33746171e-01
6.60671592e-02 -3.13230276e-01 -6.34592295e-01 -4.19359773e-01
1.41487643e-01 2.79082526e-02 -3.14486682e-01 3.95257622e-01
-1.15761943e-01 3.00594002e-01 -7.19000041e-01 -3.59990567e-01
5.97817004e-01 8.67360055e-01 -4.84072790e-02 4.65328813e-01
1.55448139e-01 4.70584810e-01 -4.57624197e-01 3.23879123e-01
8.82366896e-01 5.35511971e-01 -3.21417600e-01 -4.62914050e-01
-3.40574622e-01 4.47774887e-01 -1.42685223e+00 1.39284337e+00
-6.82080746e-01 5.17593384e-01 5.42229235e-01 -9.02178645e-01
1.34924436e+00 3.10586214e-01 1.13536105e-01 -4.96163487e-01
1.38867572e-01 3.90071392e-01 -2.19568387e-01 -8.22116911e-01
3.59992385e-01 -4.96426910e-01 2.67347842e-01 3.48643541e-01
2.39441723e-01 -5.33578277e-01 -3.95629615e-01 -1.84194803e-01
1.06747866e+00 3.90845120e-01 -1.08850740e-01 -4.44421262e-01
9.82486010e-01 -5.51782429e-01 6.16968647e-02 7.88016319e-01
-3.80646549e-02 2.43373469e-01 -3.37157287e-02 -1.45062581e-01
-8.46992552e-01 -1.38082778e+00 -4.39274639e-01 7.55543113e-01
7.50547042e-03 9.74269062e-02 -7.42801070e-01 1.74804524e-01
-8.93925354e-02 1.21446812e+00 -2.91223645e-01 -2.75910407e-01
-3.96740317e-01 -8.00772429e-01 6.66743457e-01 3.80036905e-02
-5.85253127e-02 -1.08603966e+00 -4.27691549e-01 3.18535715e-01
-1.51060656e-01 -9.91439342e-01 6.12090826e-01 5.63092709e-01
-6.63682699e-01 -4.84118611e-01 -1.97633103e-01 -1.32348448e-01
2.57292718e-01 -3.17229658e-01 9.19179976e-01 -8.48041102e-02
-4.60054070e-01 9.55632567e-01 -2.98359841e-01 -5.97539723e-01
-1.15316796e+00 -5.19546509e-01 2.55304933e-01 5.09243496e-02
-5.02519347e-02 -1.00615144e+00 -4.69065040e-01 -3.72825153e-02
-8.38240087e-01 -5.22457778e-01 6.07780635e-01 3.25136006e-01
1.25515372e-01 2.59904414e-01 3.24299783e-01 -5.08605123e-01
8.35963845e-01 -3.08593124e-01 -6.14317536e-01 1.77560017e-01
-5.97516000e-01 3.75682920e-01 4.27544802e-01 -7.71987950e-03
-1.48464084e+00 -1.65492579e-01 -8.12581420e-01 1.59923017e-01
-6.37297988e-01 3.09038371e-01 -1.66182905e-01 -4.38514978e-01
7.21191823e-01 2.40024328e-01 -1.41047567e-01 -6.67654455e-01
3.30657899e-01 7.28411734e-01 6.83809221e-01 -8.04999173e-01
1.03262222e+00 2.59454519e-01 3.67787391e-01 -1.42450058e+00
-4.68575597e-01 -5.83772063e-01 -6.78789556e-01 -7.48598576e-01
7.15347052e-01 -3.53114605e-01 -8.63272488e-01 6.27370596e-01
-1.28662336e+00 -2.84531832e-01 -1.98905185e-01 1.05452251e+00
-6.36833429e-01 4.63213205e-01 -5.87511420e-01 -1.70674777e+00
-1.80035442e-01 -7.54663169e-01 8.24787557e-01 9.78544354e-02
-6.93183392e-02 -1.31878531e+00 5.42290032e-01 1.06970012e-01
5.08307278e-01 1.69483081e-01 9.39753056e-01 -2.04929054e-01
-4.46519405e-01 -1.17792539e-01 -8.51898789e-02 6.72168016e-01
-3.72952312e-01 2.10892528e-01 -1.63418949e+00 3.60390171e-02
4.13906306e-01 -3.60979646e-01 9.91000652e-01 6.05940819e-01
8.63345861e-01 2.60476023e-01 -2.67529607e-01 4.15790230e-01
1.83894932e+00 2.05436021e-01 7.72976935e-01 1.33303449e-01
2.79809117e-01 7.98135519e-01 1.21597558e-01 4.77856487e-01
-1.17958963e-01 4.48123872e-01 5.99510670e-01 2.02056825e-01
2.61650588e-02 8.97461455e-03 2.28169337e-01 1.31379235e+00
-4.80064988e-01 -2.52439231e-01 -7.74877846e-01 6.57660421e-04
-1.46349430e+00 -8.90788436e-01 -5.45265257e-01 2.36888528e+00
5.49567938e-01 2.19060987e-01 -5.22521555e-01 5.16954601e-01
4.18216914e-01 1.44569993e-01 7.59981722e-02 -9.12167609e-01
3.02848309e-01 4.40691501e-01 5.00049829e-01 9.38150465e-01
-8.41839552e-01 1.18853234e-01 7.72686958e+00 8.46551180e-01
-6.13636553e-01 1.21856347e-01 7.88677633e-02 5.95561385e-01
-3.89435709e-01 -2.79706925e-01 -7.06606746e-01 2.25623667e-01
1.54691708e+00 6.37142062e-01 7.38259673e-01 3.03788006e-01
3.91857386e-01 -5.30033350e-01 -8.26475203e-01 6.33279860e-01
-9.80978087e-03 -7.27905273e-01 -5.48258066e-01 -1.06542386e-01
1.22869007e-01 -1.04395770e-01 -1.13491416e-01 1.64405927e-01
7.35179111e-02 -8.24481606e-01 7.45063901e-01 1.44893014e+00
3.34780246e-01 -4.69868302e-01 6.26387477e-01 4.66181159e-01
-9.95943189e-01 -4.82696407e-02 -4.15832818e-01 -8.34389850e-02
5.72409451e-01 1.08749366e+00 -7.90192544e-01 8.33730578e-01
8.30012858e-01 -6.79722987e-04 -1.01859927e-01 1.19640398e+00
-3.71708989e-01 1.00645471e+00 -8.14990819e-01 -3.05511147e-01
1.36360466e-01 -3.99608880e-01 8.93041730e-01 1.38542986e+00
4.81510788e-01 -2.79303715e-02 -3.48470628e-01 1.37426329e+00
5.80291748e-01 -2.07940373e-03 -4.37676221e-01 4.96997327e-01
4.39965725e-01 1.34961390e+00 -5.41434765e-01 2.00576752e-01
-1.64694041e-02 5.93142390e-01 -1.70045629e-01 5.67153692e-01
-7.37152696e-01 -4.21245009e-01 9.12584737e-02 -2.66110927e-01
1.31994218e-01 -2.17521638e-01 1.23533994e-01 -2.57727027e-01
-2.04768121e-01 -5.79474196e-02 -2.81708360e-01 -8.82589936e-01
-1.42058218e+00 5.25561750e-01 6.16716743e-01 -6.92543685e-01
-3.50509763e-01 -1.02327824e+00 -8.98501754e-01 1.14272964e+00
-1.58076024e+00 -7.80858815e-01 -1.89957842e-01 2.70552725e-01
-1.60138488e-01 -4.62333374e-02 1.07104075e+00 3.68708223e-01
-4.06102955e-01 1.16045929e-01 5.78746855e-01 -2.47366682e-01
3.55062455e-01 -1.45970869e+00 -1.70472320e-02 5.08900166e-01
-1.99790537e-01 4.33008671e-01 1.32151341e+00 -1.49545297e-01
-1.28524518e+00 -6.89026952e-01 7.75129616e-01 -2.23321617e-01
8.15885842e-01 -2.50089020e-01 -1.00157475e+00 1.47330677e-02
7.43198693e-02 -1.14631556e-01 7.26486266e-01 1.25115722e-01
-2.82337636e-01 -2.60581464e-01 -1.19773281e+00 -1.38798254e-02
3.61276180e-01 -7.59692550e-01 -7.47808754e-01 1.29751325e-01
5.45123041e-01 -5.85093983e-02 -1.15846682e+00 4.34865475e-01
6.38847470e-01 -1.21331143e+00 1.35311651e+00 1.95516557e-01
8.65550786e-02 -2.39332631e-01 -3.40041459e-01 -1.32741749e+00
-1.84013143e-01 -6.33300364e-01 -1.17114082e-01 1.42788184e+00
5.34366727e-01 -6.83026075e-01 2.70806372e-01 5.13153136e-01
-1.83688998e-01 -4.71303493e-01 -1.27784145e+00 -8.26653600e-01
-5.89184873e-02 -1.14043450e+00 3.20828617e-01 3.36242884e-01
-3.23253632e-01 2.32919469e-01 -3.33121151e-01 5.85515618e-01
9.75748301e-01 -4.91510421e-01 8.69498998e-02 -1.51032794e+00
-6.05299950e-01 -4.30095404e-01 -3.99937004e-01 -8.53108585e-01
2.48904347e-01 -3.94821078e-01 8.19849133e-01 -1.29897010e+00
-2.68388718e-01 -7.07800508e-01 -6.50927484e-01 -2.94120401e-01
2.74107456e-01 7.33050182e-02 -1.87295198e-01 -1.80927083e-01
-4.70032766e-02 5.57263851e-01 5.96076488e-01 -4.97929053e-03
-2.40742683e-01 4.31300521e-01 1.93804681e-01 8.53836834e-01
7.57077277e-01 -5.28045237e-01 -3.56959909e-01 -9.33335163e-03
3.36792260e-01 1.37126535e-01 7.83915877e-01 -1.46726429e+00
1.88216567e-01 2.09598631e-01 5.37168235e-02 -7.34971285e-01
7.40253568e-01 -9.29093838e-01 2.69224077e-01 4.83270854e-01
-3.70803177e-01 -6.97097898e-01 1.57306805e-01 8.53706181e-01
-1.54343709e-01 -1.03989649e+00 7.70897329e-01 9.71439406e-02
-4.83456671e-01 -3.44871491e-01 -5.82721233e-01 -5.57487607e-01
4.29104477e-01 -8.00804645e-02 8.03335607e-02 -3.07632536e-01
-8.33484709e-01 -5.37373483e-01 -4.01597142e-01 -9.97644216e-02
6.66621804e-01 -1.08735788e+00 -6.90108299e-01 6.78737611e-02
-1.97347075e-01 -4.56416368e-01 4.28974956e-01 9.25787151e-01
-6.10795736e-01 5.03955245e-01 2.01650351e-01 -6.81691468e-01
-1.05529106e+00 2.91085958e-01 7.06468940e-01 -1.66581217e-02
-3.17034274e-01 9.82723355e-01 -3.52294713e-01 -8.22045624e-01
3.36270362e-01 -3.62495899e-01 -4.85701233e-01 -3.60306650e-01
5.56601346e-01 8.98791373e-01 7.68982470e-02 -8.41119587e-01
-2.98726201e-01 6.22388065e-01 8.87617707e-01 -5.75291574e-01
1.15133584e+00 -5.10221839e-01 -5.83630383e-01 1.24968970e+00
1.17525578e+00 5.77022210e-02 -1.00888133e+00 -3.23551059e-01
-4.50424589e-02 -3.07843834e-02 4.99461114e-01 -9.67579365e-01
-3.44130009e-01 1.04068255e+00 9.50953245e-01 7.81088054e-01
1.09132338e+00 -1.46231428e-01 5.16736627e-01 6.94364309e-01
3.15809637e-01 -1.35562825e+00 -4.10958469e-01 4.25314307e-01
8.60949039e-01 -9.60995793e-01 6.06404766e-02 -5.58886409e-01
1.67340130e-01 1.51763034e+00 -1.00822069e-01 2.40728352e-02
1.35079181e+00 4.17277694e-01 -9.75058526e-02 4.13234122e-02
-3.34129006e-01 -2.10462231e-02 4.36316520e-01 8.21209133e-01
3.79462987e-01 3.64293724e-01 -7.25574717e-02 3.66383314e-01
-1.44243836e-01 -2.18986005e-01 1.67326882e-01 5.47666192e-01
-9.46209311e-01 -9.23919559e-01 -9.35373604e-01 1.10527370e-02
-1.77482903e-01 -3.73955742e-02 1.77205607e-01 4.27390784e-01
2.24712938e-01 1.37895179e+00 -3.29754740e-01 -3.83468896e-01
4.19149458e-01 2.72478193e-01 5.23819983e-01 -3.07967644e-02
-2.90144950e-01 1.47479683e-01 2.57597536e-01 -4.86822158e-01
-8.29386234e-01 -4.84576911e-01 -1.15112305e+00 -7.85648301e-02
-5.50681710e-01 3.38784814e-01 1.41599131e+00 1.22349334e+00
-4.22222257e-01 9.57058966e-01 7.68089771e-01 -9.84431088e-01
-6.83411419e-01 -1.00150609e+00 -9.80593622e-01 -7.82315433e-02
4.74441528e-01 -6.04520440e-01 -7.18394518e-01 -7.34482482e-02]
|
[15.209113121032715, 5.603954315185547]
|
b7ea7a70-14b0-4213-9110-4b3bbbf4813c
|
visa-an-ambiguous-subtitles-dataset-for
|
2201.08054
| null |
https://arxiv.org/abs/2201.08054v3
|
https://arxiv.org/pdf/2201.08054v3.pdf
|
VISA: An Ambiguous Subtitles Dataset for Visual Scene-Aware Machine Translation
|
Existing multimodal machine translation (MMT) datasets consist of images and video captions or general subtitles, which rarely contain linguistic ambiguity, making visual information not so effective to generate appropriate translations. We introduce VISA, a new dataset that consists of 40k Japanese-English parallel sentence pairs and corresponding video clips with the following key features: (1) the parallel sentences are subtitles from movies and TV episodes; (2) the source subtitles are ambiguous, which means they have multiple possible translations with different meanings; (3) we divide the dataset into Polysemy and Omission according to the cause of ambiguity. We show that VISA is challenging for the latest MMT system, and we hope that the dataset can facilitate MMT research. The VISA dataset is available at: https://github.com/ku-nlp/VISA.
|
['Sadao Kurohashi', 'Chenhui Chu', 'Weiqi Gu', 'Shuichiro Shimizu', 'Yihang Li']
|
2022-01-20
| null |
https://aclanthology.org/2022.lrec-1.725
|
https://aclanthology.org/2022.lrec-1.725.pdf
|
lrec-2022-6
|
['multimodal-machine-translation']
|
['natural-language-processing']
|
[ 2.48498961e-01 -2.63339996e-01 -4.64393497e-01 -3.32804203e-01
-9.96111393e-01 -9.98470902e-01 6.36102855e-01 -2.18529388e-01
-1.39626727e-01 8.52473557e-01 4.71777081e-01 -5.11020005e-01
4.07653123e-01 -2.72162288e-01 -8.34864378e-01 -3.58291358e-01
3.44900638e-01 5.39431393e-01 -9.56233442e-02 -4.81218010e-01
1.54902622e-01 -1.18388899e-01 -1.07931209e+00 8.56218040e-01
8.58662248e-01 8.51898551e-01 4.05718058e-01 4.25566822e-01
-3.12546581e-01 2.30537519e-01 -5.63892186e-01 -7.71167219e-01
4.04572934e-01 -8.16587508e-01 -8.66351128e-01 1.12851091e-01
7.97716618e-01 -2.25003660e-01 -4.02400970e-01 1.12294841e+00
5.50764799e-01 -3.18181455e-01 5.33234477e-01 -1.59645069e+00
-1.05314183e+00 5.11421740e-01 -6.13511205e-01 1.66293547e-01
7.31453955e-01 2.15863273e-01 1.06502903e+00 -1.15570664e+00
9.81368661e-01 1.34375191e+00 2.53700435e-01 6.48838580e-01
-1.07992911e+00 -6.66714072e-01 5.62773086e-02 5.22949517e-01
-1.48548377e+00 -5.70407867e-01 6.18132710e-01 -2.66325682e-01
7.87756503e-01 5.21271586e-01 8.21514189e-01 1.73445094e+00
1.23620063e-01 1.06287050e+00 1.31941962e+00 -4.03802991e-01
-1.14431277e-01 5.19329272e-02 -2.76257575e-01 4.88405049e-01
-8.57535079e-02 -3.99237663e-01 -6.10178709e-01 8.16868171e-02
5.92886984e-01 -2.41893411e-01 -3.57006907e-01 -9.99727920e-02
-1.98748767e+00 5.45268059e-01 1.51649058e-01 3.25956136e-01
-1.42185101e-02 -3.20536196e-01 3.09308022e-01 6.87084854e-01
-5.08892387e-02 3.41588706e-01 -2.81300485e-01 -3.79145086e-01
-5.84365785e-01 6.85172006e-02 3.62136871e-01 1.46473145e+00
6.77122355e-01 -3.36915076e-01 -1.33690029e-01 1.06249833e+00
4.62659299e-02 8.79140556e-01 6.12910688e-01 -1.11192334e+00
1.22076988e+00 4.85150993e-01 2.42729276e-01 -1.01734245e+00
-1.70134202e-01 1.48785099e-01 -4.89862740e-01 -4.79035020e-01
4.69667554e-01 -1.49858907e-01 -7.43295133e-01 1.78160501e+00
4.88387560e-03 -3.28683048e-01 1.78538412e-01 1.11996782e+00
1.25252867e+00 8.48639429e-01 -2.53882915e-01 -2.52648979e-01
1.59330118e+00 -1.09846330e+00 -1.02497876e+00 -4.27320451e-01
5.52317202e-01 -1.34077609e+00 1.50587428e+00 5.38133942e-02
-1.10355389e+00 -2.43425623e-01 -9.48280156e-01 -2.92343467e-01
-3.84257555e-01 2.70479172e-01 2.10457146e-01 1.95132390e-01
-8.71454358e-01 -1.18696019e-01 -4.11818534e-01 -5.00134349e-01
2.67459363e-01 -3.30014303e-02 -6.01938546e-01 -3.05339932e-01
-1.30746293e+00 8.23593974e-01 2.79463142e-01 3.09910595e-01
-6.03471220e-01 -8.62597227e-02 -9.95470107e-01 -3.99747550e-01
4.69710022e-01 -7.90360034e-01 1.30001521e+00 -1.42869437e+00
-1.17563117e+00 1.18989813e+00 -6.13733947e-01 1.44127011e-01
7.31530786e-01 -3.31144338e-03 -5.60366750e-01 4.67025876e-01
3.53673279e-01 1.07165575e+00 8.01616907e-01 -1.12492573e+00
-4.01249617e-01 -3.16276431e-01 1.11585662e-01 5.43008268e-01
-2.17452183e-01 1.47209480e-01 -9.07353282e-01 -9.11889911e-01
1.31331041e-01 -1.09825504e+00 2.69959092e-01 -4.70811315e-02
-8.56935978e-01 -7.48157576e-02 6.88358128e-01 -9.12736952e-01
1.09769225e+00 -2.23518062e+00 4.64660257e-01 -2.75701851e-01
7.93852061e-02 4.34375815e-02 -6.43101990e-01 7.04521120e-01
-1.03078336e-01 3.19831252e-01 -1.84968203e-01 -3.05756599e-01
1.10807940e-01 2.61522710e-01 -3.27006757e-01 1.35447308e-01
1.13110512e-01 1.31053090e+00 -8.85030091e-01 -8.10251892e-01
6.81167170e-02 1.64928120e-02 8.73590857e-02 -3.18099596e-02
-2.49018312e-01 6.77269638e-01 -3.10625911e-01 1.01722193e+00
5.33694804e-01 -1.85063064e-01 2.11858109e-01 -4.26712543e-01
-3.31771038e-02 4.47426945e-01 -6.66264772e-01 1.80160010e+00
-3.52647267e-02 1.16059446e+00 -1.30367488e-01 -4.61138248e-01
5.44420004e-01 4.03029263e-01 2.61198163e-01 -1.00599539e+00
1.68884203e-01 3.28686237e-01 -8.92098323e-02 -8.50089610e-01
5.43491542e-01 4.75618057e-02 -3.17655832e-01 2.86862940e-01
-1.75619438e-01 -1.30112410e-01 6.33478701e-01 3.83449972e-01
6.68343186e-01 2.57420144e-03 2.28608660e-02 1.89505443e-01
7.53139034e-02 4.30166423e-01 7.58231282e-01 5.36760688e-01
-3.84275228e-01 9.64177787e-01 6.31310344e-01 -3.37157369e-01
-1.17457271e+00 -1.26129758e+00 3.31336074e-02 6.93973184e-01
4.32524204e-01 -6.30945504e-01 -6.11863315e-01 -4.69977617e-01
-5.45088649e-01 5.46937883e-01 -4.28480119e-01 2.41763726e-01
-6.09960616e-01 -4.83590662e-01 4.39500809e-01 2.93755531e-01
5.64456284e-01 -9.13125277e-01 -1.15997992e-01 -2.25056946e-01
-1.19864285e+00 -1.49453843e+00 -1.09059417e+00 -4.52243328e-01
-7.08159387e-01 -1.05817556e+00 -1.04990220e+00 -1.10051811e+00
8.98551583e-01 5.80129683e-01 1.24638093e+00 -5.06907105e-02
5.52132539e-02 3.35061520e-01 -4.73704308e-01 -1.41723335e-01
-4.94012535e-01 -1.80594087e-01 8.48328471e-02 -1.20539494e-01
4.42781419e-01 -4.11398858e-01 -4.98495996e-01 6.86134577e-01
-7.32948661e-01 7.88923979e-01 5.88584781e-01 8.08619201e-01
7.10553408e-01 -5.24330556e-01 3.72060984e-01 -4.11828518e-01
6.58164203e-01 -4.08086151e-01 -4.18295801e-01 5.59487641e-01
-2.14790404e-02 -4.76978332e-01 3.97641510e-01 -6.39633417e-01
-6.68503344e-01 -1.45304680e-01 1.51842430e-01 -5.40249705e-01
-1.30975872e-01 5.19250035e-01 -2.53244877e-01 3.20601642e-01
3.34710032e-01 4.96918738e-01 -7.38881081e-02 -3.52285802e-01
1.79230422e-01 8.34243238e-01 6.02419138e-01 -5.52608490e-01
6.23145342e-01 2.01797977e-01 -2.19716176e-01 -7.66216636e-01
-4.30505931e-01 -5.59991188e-02 -6.10539556e-01 -4.02212858e-01
8.44835222e-01 -9.93963957e-01 -3.18596780e-01 4.06521171e-01
-1.21429253e+00 -2.25328043e-01 2.42104217e-01 4.71395880e-01
-5.12733757e-01 7.30152130e-01 -9.43954110e-01 -2.44174615e-01
-1.30446747e-01 -1.49149156e+00 1.02606511e+00 3.24297845e-01
-4.78561521e-01 -6.18040562e-01 -2.40877658e-01 9.38568652e-01
6.73598126e-02 -5.55090792e-03 1.08766723e+00 -2.99980700e-01
-5.60042024e-01 -2.64095049e-02 -1.57278284e-01 2.33603027e-02
2.89881080e-01 1.77528024e-01 -4.11275983e-01 -2.15533316e-01
-3.27041060e-01 -5.54451525e-01 4.95914251e-01 1.71960473e-01
7.09117234e-01 -4.48379785e-01 -1.49942234e-01 3.08779985e-01
8.58023405e-01 1.87272266e-01 6.88262403e-01 3.80591124e-01
6.69732213e-01 6.25386000e-01 7.14390755e-01 -7.00718760e-02
6.82120085e-01 1.01504040e+00 3.32181692e-01 -9.39360354e-03
-2.62761474e-01 -5.20852089e-01 6.88681185e-01 1.34037495e+00
5.91452494e-02 -3.12202781e-01 -1.04222178e+00 7.47718453e-01
-2.00810814e+00 -8.82914901e-01 -2.54716605e-01 1.96991575e+00
9.46005404e-01 -1.82865098e-01 1.49567872e-01 -4.36167568e-01
9.89326239e-01 1.06839813e-01 -3.22223365e-01 -3.61537427e-01
-6.54061913e-01 -5.60185254e-01 1.54840589e-01 3.89973283e-01
-8.41953337e-01 9.27489698e-01 6.24607086e+00 8.93082678e-01
-1.06579018e+00 1.28227219e-01 4.55778152e-01 -2.80552655e-01
-5.92658043e-01 -2.95456536e-02 -4.52460945e-01 7.94379652e-01
5.60997248e-01 -1.30178481e-01 5.17583787e-01 3.25080365e-01
2.52235621e-01 -9.50628519e-02 -1.04163861e+00 1.23274910e+00
2.63996154e-01 -1.24561822e+00 3.24413419e-01 -1.21808030e-01
6.53109431e-01 1.65607274e-01 2.79237926e-01 1.68165058e-01
-2.27121487e-01 -7.84931600e-01 9.36322331e-01 2.42187008e-01
9.04432178e-01 -3.37642789e-01 5.33435643e-01 2.70304531e-01
-9.56696749e-01 2.43668452e-01 -2.82006502e-01 1.09876432e-01
3.81415427e-01 1.27236128e-01 -5.91141403e-01 6.51406348e-01
7.81088531e-01 9.92949069e-01 -7.84063697e-01 9.33129370e-01
-3.88793439e-01 4.20972854e-01 -3.12370837e-01 -1.32124111e-01
2.02978343e-01 -5.41445434e-01 9.52297986e-01 1.03423893e+00
5.47784269e-01 -1.51506409e-01 7.37528736e-03 6.75282300e-01
-3.50800544e-01 2.20164403e-01 -7.15970337e-01 -4.35159564e-01
5.97997904e-01 9.97985244e-01 -7.14986444e-01 -3.34666044e-01
-8.03368449e-01 1.29768574e+00 7.48692602e-02 6.20823503e-01
-8.13979268e-01 -5.55604510e-02 5.57401121e-01 -2.90759027e-01
1.59051806e-01 -2.49785289e-01 -1.16488181e-01 -1.83968508e+00
5.51379025e-01 -1.22860444e+00 4.37382340e-01 -1.42776132e+00
-1.43347979e+00 7.69354403e-01 -3.45577225e-02 -1.69201016e+00
-2.28466973e-01 -5.34143925e-01 -2.32801601e-01 6.47359192e-01
-1.08525312e+00 -1.25871503e+00 -1.65445700e-01 5.57775259e-01
8.70960653e-01 -1.12506136e-01 7.02353895e-01 4.53004658e-01
-8.60481799e-01 6.15846097e-01 -1.87697075e-02 3.38672847e-01
1.14549685e+00 -8.52423251e-01 3.75973880e-01 6.78280950e-01
3.72546524e-01 5.75121760e-01 8.08713794e-01 -6.39905572e-01
-1.49565363e+00 -8.22739065e-01 1.21271300e+00 -8.64679456e-01
6.39762163e-01 -5.28194249e-01 -6.33212507e-01 9.65102971e-01
7.76913941e-01 -3.20102334e-01 6.47152126e-01 -1.57693043e-01
-4.83019173e-01 1.99934691e-01 -7.29649186e-01 1.22806239e+00
1.04772031e+00 -5.43647230e-01 -8.00225377e-01 5.19760072e-01
5.87452471e-01 -7.52474666e-01 -5.55146635e-01 3.27240676e-01
6.28600717e-01 -7.19775498e-01 8.15755725e-01 -6.13222122e-01
9.22779441e-01 -4.48480725e-01 -3.92377585e-01 -1.17949533e+00
1.46043617e-02 -5.52240729e-01 1.11804880e-01 1.17041671e+00
8.59352112e-01 -4.37747955e-01 3.53951484e-01 5.19141436e-01
-2.91546196e-01 -7.35808134e-01 -1.25591826e+00 -8.32300901e-01
1.00071855e-01 -3.14934850e-01 5.85270882e-01 1.36011720e+00
1.67952701e-01 6.43553734e-01 -7.06932962e-01 -1.76086739e-01
1.70519635e-01 5.61513841e-01 6.46774113e-01 -4.40749496e-01
-3.69148850e-02 -6.35122538e-01 -1.97399691e-01 -1.26579356e+00
-3.06728017e-03 -9.20263529e-01 -1.47924826e-01 -1.51903987e+00
5.54643333e-01 1.09696515e-01 5.94626367e-02 8.79766762e-01
-2.74124034e-02 5.93409657e-01 3.68446618e-01 6.47112668e-01
-8.41501355e-01 6.09771550e-01 1.64003909e+00 -4.47375059e-01
-2.97273193e-02 -4.29240286e-01 -5.27027547e-01 4.27197874e-01
8.88268888e-01 -2.95450836e-01 -2.35172108e-01 -8.23100626e-01
4.94239926e-01 3.15136880e-01 3.72137189e-01 -2.83121139e-01
-1.22911498e-01 -4.84217584e-01 4.29556161e-01 -6.33704364e-01
5.17608881e-01 -7.14712620e-01 3.28720063e-01 1.50744498e-01
-3.66671026e-01 6.32548869e-01 3.24099600e-01 1.88140914e-01
-3.52834821e-01 -8.28069374e-02 2.91469783e-01 -8.99994075e-02
-6.89897776e-01 7.38503560e-02 -5.81962049e-01 2.35255003e-01
8.04877877e-01 -1.77819014e-01 -7.32942462e-01 -7.95840442e-01
-5.11213541e-01 5.58225870e-01 8.69106352e-01 9.09322441e-01
6.13565445e-01 -1.78429699e+00 -8.36046875e-01 -1.85417473e-01
4.58197951e-01 -3.37203652e-01 2.97468543e-01 1.14971447e+00
-3.91503960e-01 3.17415833e-01 -4.42276537e-01 -6.26265049e-01
-1.48428822e+00 4.54547048e-01 1.50230601e-01 4.84230816e-01
-6.26114130e-01 5.11839211e-01 2.42122993e-01 -3.81401777e-01
-1.11183889e-01 -1.75545271e-02 -8.01815540e-02 -8.82412959e-03
4.93608266e-01 -6.05569147e-02 -2.93965816e-01 -1.16742790e+00
-3.56097877e-01 5.34345746e-01 -1.40939057e-01 -3.12628806e-01
8.97816122e-01 -6.88100994e-01 -3.20294648e-01 6.76589072e-01
1.14155364e+00 8.39041173e-02 -8.85577738e-01 -3.86980832e-01
-2.71346569e-01 -7.03936696e-01 -5.93828499e-01 -9.94624734e-01
-7.97734499e-01 8.38588893e-01 3.21073145e-01 -1.33093446e-01
1.11412358e+00 3.73939484e-01 1.09281647e+00 3.73340279e-01
3.10194910e-01 -1.09502089e+00 5.45294881e-02 5.41478634e-01
1.15154374e+00 -1.41638255e+00 -4.03395236e-01 -5.17356932e-01
-8.29937577e-01 1.00827157e+00 5.66602588e-01 7.09905684e-01
-2.50563741e-01 -2.23161533e-01 5.81209838e-01 1.39241852e-02
-8.17704260e-01 -2.19857749e-02 4.26836550e-01 5.31756997e-01
3.05868745e-01 1.35641336e-01 -5.89291215e-01 7.39469767e-01
-4.17556286e-01 -3.46558303e-01 5.35240948e-01 7.95416117e-01
1.21333845e-01 -1.19687247e+00 -4.24876034e-01 1.98149323e-01
-1.84632108e-01 -1.90133750e-01 -8.73836875e-01 6.97960973e-01
-1.18316971e-01 1.06067061e+00 -1.89725488e-01 -4.50494289e-01
2.02319637e-01 1.08099155e-01 5.77627063e-01 -3.07602137e-01
-6.36744276e-02 1.45933256e-01 4.31225330e-01 -5.39315522e-01
-3.48493397e-01 -7.57086933e-01 -1.00377989e+00 -3.09225649e-01
1.69516623e-01 4.98837940e-02 5.27590930e-01 9.34704483e-01
5.22851765e-01 1.62175689e-02 3.30572814e-01 -6.46420896e-01
9.26163122e-02 -9.43606496e-01 -9.51103643e-02 6.44200563e-01
2.91675806e-01 -4.00767446e-01 -2.13907227e-01 2.73572147e-01]
|
[11.357680320739746, 1.5629624128341675]
|
b0da5a26-2697-4f7d-886a-76a0d202c32e
|
linguistic-more-taking-a-further-step-toward
|
2305.05140
| null |
https://arxiv.org/abs/2305.05140v2
|
https://arxiv.org/pdf/2305.05140v2.pdf
|
Linguistic More: Taking a Further Step toward Efficient and Accurate Scene Text Recognition
|
Vision model have gained increasing attention due to their simplicity and efficiency in Scene Text Recognition (STR) task. However, due to lacking the perception of linguistic knowledge and information, recent vision models suffer from two problems: (1) the pure vision-based query results in attention drift, which usually causes poor recognition and is summarized as linguistic insensitive drift (LID) problem in this paper. (2) the visual feature is suboptimal for the recognition in some vision-missing cases (e.g. occlusion, etc.). To address these issues, we propose a $\textbf{L}$inguistic $\textbf{P}$erception $\textbf{V}$ision model (LPV), which explores the linguistic capability of vision model for accurate text recognition. To alleviate the LID problem, we introduce a Cascade Position Attention (CPA) mechanism that obtains high-quality and accurate attention maps through step-wise optimization and linguistic information mining. Furthermore, a Global Linguistic Reconstruction Module (GLRM) is proposed to improve the representation of visual features by perceiving the linguistic information in the visual space, which gradually converts visual features into semantically rich ones during the cascade process. Different from previous methods, our method obtains SOTA results while keeping low complexity (92.4% accuracy with only 8.11M parameters). Code is available at https://github.com/CyrilSterling/LPV.
|
['Yongdong Zhang', 'Jianjun Xu', 'Yuxin Wang', 'Hongtao Xie', 'Boqiang Zhang']
|
2023-05-09
| null | null | null | null |
['scene-text-recognition']
|
['computer-vision']
|
[-4.64244634e-02 -5.11770010e-01 -1.25002369e-01 -2.39153832e-01
-5.58406293e-01 -2.21640795e-01 5.60581803e-01 -1.63898379e-01
-5.35448909e-01 5.65238416e-01 1.41886428e-01 -1.45091951e-01
1.01840518e-01 -5.18843472e-01 -5.97826064e-01 -6.97940946e-01
8.02795708e-01 3.58952172e-02 1.28312707e-01 -2.12846115e-01
3.83125186e-01 4.44907635e-01 -1.81044960e+00 3.78209233e-01
1.26088321e+00 1.16483879e+00 6.37933791e-01 4.47418839e-01
-6.20266855e-01 8.33663166e-01 -6.04654133e-01 -3.76787454e-01
-5.59927970e-02 -5.03029704e-01 -6.40941083e-01 1.51483044e-01
4.17200506e-01 -2.62781799e-01 -4.26059812e-01 1.32823575e+00
5.56169689e-01 1.65595621e-01 5.36258817e-01 -1.04816294e+00
-1.21755016e+00 3.35550487e-01 -9.32389021e-01 5.32451630e-01
1.55717418e-01 3.12858284e-01 8.32427144e-01 -1.49349296e+00
2.87561417e-01 1.39377952e+00 3.14558148e-01 6.67762101e-01
-8.47049594e-01 -5.46081722e-01 4.39718455e-01 5.51033080e-01
-1.61412442e+00 -6.62649512e-01 8.18910360e-01 -3.62666845e-01
1.07042003e+00 3.97695631e-01 5.55051744e-01 9.35828209e-01
1.55800223e-01 1.03906274e+00 8.04558516e-01 -5.20359635e-01
-2.09547564e-01 2.20917568e-01 3.52095634e-01 7.53593624e-01
2.53452510e-01 -1.18392430e-01 -7.89873004e-01 5.38751960e-01
5.92327535e-01 1.65016711e-01 -4.30067241e-01 1.84967965e-01
-1.00253642e+00 5.37901998e-01 7.22928524e-01 5.32176018e-01
-2.74761856e-01 5.61916456e-02 1.94042549e-01 8.55183601e-02
3.38106871e-01 1.06186327e-02 -1.46881893e-01 8.82024914e-02
-8.22719574e-01 -2.04297394e-01 1.22261792e-01 8.68818343e-01
7.64769495e-01 4.81564432e-01 -4.16684687e-01 1.10329044e+00
5.54505050e-01 7.67772257e-01 7.19237030e-01 -5.77915013e-01
6.48163021e-01 8.34876835e-01 5.07743936e-03 -1.19432294e+00
-3.42384309e-01 -4.31085110e-01 -9.26842868e-01 4.31951024e-02
1.77906305e-01 1.56546772e-01 -1.07851696e+00 1.43257952e+00
9.58910864e-03 -3.41369033e-01 -1.55482104e-03 1.20570111e+00
1.21796048e+00 7.47278869e-01 8.95381197e-02 -2.67171443e-01
1.40244639e+00 -1.09275019e+00 -8.64031494e-01 -5.22573531e-01
3.19472849e-01 -8.20087492e-01 1.41928506e+00 2.79269636e-01
-1.02742934e+00 -8.55552495e-01 -9.59616303e-01 -4.62941557e-01
-3.86770070e-01 4.96923089e-01 3.24838638e-01 4.96342480e-01
-1.25874627e+00 -4.05007228e-02 -5.21303833e-01 -5.56816101e-01
6.38942301e-01 3.11380029e-01 3.23894247e-02 -3.69127363e-01
-1.09464061e+00 8.42736840e-01 3.15705210e-01 4.83128518e-01
-7.83840358e-01 -1.57885507e-01 -7.47696519e-01 -9.29982364e-02
2.88930178e-01 -6.54533088e-01 8.54870677e-01 -1.33601058e+00
-1.31739843e+00 8.35736811e-01 -5.86405993e-01 -2.16214180e-01
4.95846003e-01 -2.69183040e-01 -4.28789318e-01 7.63294548e-02
9.12984610e-02 7.95107007e-01 1.11822891e+00 -1.23243678e+00
-5.13291717e-01 -6.03215814e-01 -1.32806107e-01 5.84678829e-01
-5.94512761e-01 1.41393691e-01 -8.25792551e-01 -6.91278636e-01
2.86640733e-01 -2.86069930e-01 2.50548661e-01 -2.62554102e-02
-3.34129125e-01 -2.97089249e-01 8.80631804e-01 -8.05005431e-01
1.42004919e+00 -2.15138698e+00 1.00147955e-01 -3.22436601e-01
2.56432533e-01 5.54921746e-01 -1.17585182e-01 7.80313015e-02
1.33093297e-01 2.53673702e-01 -1.39758825e-01 -4.87029195e-01
-3.63882422e-01 6.35348260e-02 -3.68077248e-01 2.78771400e-01
2.21036628e-01 1.24662638e+00 -6.76455736e-01 -7.08142340e-01
4.57835585e-01 4.33734328e-01 -2.79211402e-01 -8.45554695e-02
-2.66052455e-01 3.38105291e-01 -5.62768877e-01 1.01823890e+00
6.56297445e-01 -2.77489126e-01 -4.44660693e-01 -3.36464942e-01
-2.98631877e-01 -1.46945179e-01 -8.07307005e-01 1.72200358e+00
-2.36568168e-01 6.41341209e-01 4.48692404e-02 -9.35481310e-01
1.16006970e+00 -7.57249743e-02 1.17515884e-01 -9.43768859e-01
4.17682022e-01 3.93920913e-02 -2.28724867e-01 -7.37781048e-01
4.90633786e-01 -3.71202603e-02 2.31795982e-01 1.99581042e-01
-1.04031399e-01 1.42672569e-01 4.48801555e-03 2.27824971e-02
6.32446349e-01 2.05270767e-01 3.24045233e-02 -1.25818968e-01
9.06301022e-01 1.35179475e-01 5.05741835e-01 7.36761808e-01
-4.63204771e-01 5.68378627e-01 2.83664297e-02 -3.89593244e-01
-7.35801578e-01 -7.17996836e-01 -1.17667921e-01 1.05600059e+00
4.44730967e-01 -1.69966429e-01 -5.61398089e-01 -4.79282290e-01
-2.10846603e-01 8.92560363e-01 -3.72380257e-01 -4.39087540e-01
-3.12079042e-01 -9.52482164e-01 4.23501760e-01 4.78876591e-01
1.05761516e+00 -1.26863658e+00 -5.67749023e-01 -7.64793307e-02
-3.57481569e-01 -8.18462968e-01 -6.19480789e-01 -2.12319434e-01
-7.49764025e-01 -7.78884768e-01 -6.88528001e-01 -8.73664439e-01
7.62264669e-01 7.20366716e-01 7.41109371e-01 2.57382780e-01
-4.04046744e-01 3.16799104e-01 -3.60844910e-01 -3.84330273e-01
-1.70457326e-02 -1.89848930e-01 -2.23769732e-02 2.50178784e-01
8.04258823e-01 -4.20693755e-02 -6.76578104e-01 1.47592753e-01
-8.09268475e-01 2.60905087e-01 5.65633595e-01 1.02932072e+00
7.23318398e-01 6.81967363e-02 5.62647045e-01 -2.91224509e-01
6.78590715e-01 -1.91511363e-01 -5.76824486e-01 3.33754510e-01
-7.14922965e-01 -1.60401940e-01 7.21380293e-01 -2.38839358e-01
-1.19068491e+00 -8.39091912e-02 -1.24642299e-02 -8.05054247e-01
-2.00166553e-01 4.96000350e-01 -3.04722250e-01 -4.39162254e-02
5.90989411e-01 7.86191761e-01 -3.06510702e-02 -4.49428618e-01
2.28990033e-01 8.96103919e-01 4.06464607e-01 -2.35513270e-01
4.61841583e-01 5.05330384e-01 -4.27117378e-01 -9.75397050e-01
-8.69719863e-01 -2.31661558e-01 -4.99799430e-01 -2.83374935e-01
8.75989556e-01 -9.28797007e-01 -8.44154716e-01 7.01062202e-01
-1.19720483e+00 -1.36784427e-02 -1.14858836e-01 4.78118062e-01
-2.43319273e-01 6.52418911e-01 -4.25733149e-01 -1.07929420e+00
-5.60714245e-01 -1.07819474e+00 9.58005130e-01 5.44022202e-01
2.38324031e-01 -6.74670756e-01 -5.65966308e-01 6.67256892e-01
2.71611780e-01 -3.36485952e-01 7.84125388e-01 -2.33216211e-01
-7.02075362e-01 -5.26464581e-02 -7.41744041e-01 4.64654773e-01
6.27083629e-02 -5.39355204e-02 -1.11174440e+00 -2.73153871e-01
2.51356572e-01 -3.06461602e-01 1.30742192e+00 5.41556060e-01
1.15627408e+00 -2.20197156e-01 -1.69872910e-01 6.67349756e-01
1.40815294e+00 4.64549392e-01 6.12837493e-01 2.51459122e-01
1.09178221e+00 4.48949337e-01 5.67840993e-01 3.14619571e-01
4.91443545e-01 4.69496042e-01 4.27398443e-01 -1.53154895e-01
-5.40703237e-01 -2.99957812e-01 4.92315143e-01 9.28643942e-01
5.22014238e-02 -4.76796120e-01 -1.07184303e+00 5.15761018e-01
-1.90984452e+00 -6.64013147e-01 -8.24750215e-02 1.98785937e+00
6.30745173e-01 3.62144075e-02 -2.22683266e-01 -7.80029073e-02
9.74769592e-01 1.93694010e-01 -9.66633558e-01 -2.27470711e-01
-5.48583686e-01 -2.43538618e-01 2.34422535e-01 4.73732263e-01
-8.64127398e-01 1.31235540e+00 5.25460529e+00 8.89147997e-01
-1.19872916e+00 1.21080592e-01 6.37913585e-01 -1.33589223e-01
-2.93866694e-01 -2.29802907e-01 -9.82112765e-01 5.38702905e-01
3.18883419e-01 -1.23353958e-01 6.16366863e-01 5.35261452e-01
3.19246083e-01 -1.36948317e-01 -6.28963649e-01 1.49904037e+00
6.25858963e-01 -1.08557510e+00 5.63940763e-01 -3.10109645e-01
4.26346421e-01 2.42697783e-02 2.71058559e-01 1.86304688e-01
8.52670446e-02 -1.10190809e+00 7.31204033e-01 8.25218081e-01
1.09350574e+00 -6.70283973e-01 5.64694047e-01 4.95421082e-01
-1.08457339e+00 -2.92256474e-01 -5.44706583e-01 8.07116255e-02
-1.82883620e-01 5.03452241e-01 -4.50486839e-01 6.18309200e-01
9.72332954e-01 9.82272148e-01 -7.56615639e-01 7.90071309e-01
-2.56688774e-01 3.86071444e-01 -1.91518068e-01 -1.52972698e-01
1.97339013e-01 -2.12850600e-01 6.84976459e-01 9.39163804e-01
3.33994210e-01 1.92226395e-02 2.54912861e-02 1.05587506e+00
-1.59332946e-01 2.49187261e-01 -4.62897688e-01 2.97348499e-02
3.79941314e-01 9.86821055e-01 -5.84333718e-01 -1.54064432e-01
-4.44878757e-01 1.13384926e+00 3.68634313e-01 6.12981617e-01
-6.93713248e-01 -4.29343820e-01 2.97828823e-01 -1.52067512e-01
3.46785426e-01 -9.02799442e-02 -4.97613817e-01 -1.47506118e+00
2.92692125e-01 -6.73138618e-01 4.16321844e-01 -1.14758337e+00
-1.12955761e+00 7.20520198e-01 -4.37990963e-01 -9.64284837e-01
2.19203696e-01 -4.99968201e-01 -2.72417992e-01 1.12654793e+00
-1.73325098e+00 -1.23360634e+00 -5.80525875e-01 8.49456966e-01
1.04084063e+00 -3.41914415e-01 4.32526827e-01 2.38666400e-01
-9.37256992e-01 7.26571858e-01 1.71546973e-02 7.45037049e-02
5.63889623e-01 -8.45344901e-01 1.29043832e-01 1.08772719e+00
1.57218575e-01 4.44568366e-01 4.04935151e-01 -7.09147930e-01
-1.56179082e+00 -1.11648214e+00 9.73184109e-01 -5.46942532e-01
3.81741494e-01 -7.94237107e-02 -1.01298666e+00 3.91487032e-01
7.10800812e-02 -1.98840663e-01 1.98574409e-01 -3.27568889e-01
-3.35456580e-01 -3.69668484e-01 -9.98194098e-01 7.83304930e-01
1.15637934e+00 -4.75469798e-01 -6.08823001e-01 1.53639302e-01
6.73665583e-01 -1.63180500e-01 -2.80989528e-01 2.04914302e-01
2.67691076e-01 -1.00058150e+00 8.50488484e-01 -2.28280634e-01
1.55910656e-01 -6.57288134e-01 -1.57341301e-01 -7.64239311e-01
-3.52149725e-01 -2.51386672e-01 -5.26423119e-02 1.27539122e+00
1.68057337e-01 -7.06304133e-01 3.48189265e-01 3.64795595e-01
-1.52601883e-01 -5.57519257e-01 -1.06607139e+00 -5.50952137e-01
-5.40269241e-02 -4.10747111e-01 4.09229100e-01 8.26414347e-01
-2.57176101e-01 6.37945712e-01 -4.10267621e-01 -5.07367849e-02
4.87370193e-01 2.48085275e-01 2.40676671e-01 -1.04596651e+00
4.50666286e-02 -6.76605225e-01 -2.50965893e-01 -1.29168677e+00
8.22038278e-02 -1.00028205e+00 1.92480367e-02 -1.66714549e+00
2.65691757e-01 -2.24435449e-01 -4.54372108e-01 6.01471901e-01
-2.40001425e-01 1.88017994e-01 3.29920799e-01 6.16761923e-01
-7.52216816e-01 8.65705848e-01 1.41471922e+00 -3.19943756e-01
-1.06843032e-01 -2.40708217e-01 -8.58513892e-01 6.12914503e-01
8.05865705e-01 -1.01553604e-01 -3.54933560e-01 -9.26180840e-01
1.04931906e-01 -7.81166181e-02 4.95233476e-01 -7.34822214e-01
4.70051020e-01 -9.02482867e-02 7.05470085e-01 -7.85552382e-01
4.01503623e-01 -6.23375833e-01 -3.13831240e-01 4.90953386e-01
-1.32748872e-01 3.58874910e-02 2.68651903e-01 4.79705244e-01
-3.31306100e-01 -2.10140690e-01 7.82747090e-01 -2.65818417e-01
-9.86104608e-01 2.43053198e-01 -2.57953167e-01 -3.70206572e-02
7.01181769e-01 -4.29075688e-01 -5.73871255e-01 -2.66022980e-01
-3.90300959e-01 4.81996238e-01 4.93116021e-01 6.04124129e-01
9.49229360e-01 -1.25737321e+00 -7.83631504e-01 3.12082112e-01
2.08187521e-01 4.50330414e-02 5.13941646e-01 8.31678569e-01
-3.95663291e-01 6.20308280e-01 -1.37638431e-02 -7.10988760e-01
-1.02654064e+00 7.74677455e-01 4.79267955e-01 1.55031830e-01
-5.58020234e-01 9.55705404e-01 4.02055651e-01 7.52185360e-02
2.31908694e-01 -1.07628271e-01 -4.49845195e-01 1.62315607e-01
6.05066597e-01 3.10129464e-01 2.53341552e-02 -9.23978627e-01
-5.71033359e-01 8.40989113e-01 -3.06705683e-01 7.94474706e-02
9.07815337e-01 -5.95721245e-01 -1.85954809e-01 5.47306657e-01
8.49081635e-01 -2.78135389e-01 -1.25500560e+00 -3.33919674e-01
-1.68398902e-01 -5.59313595e-01 3.20689023e-01 -9.04945254e-01
-1.08409679e+00 1.08837605e+00 7.02322602e-01 -8.28210339e-02
1.39247417e+00 -5.32495603e-02 6.13630831e-01 3.21055710e-01
1.60843402e-01 -1.21637237e+00 1.46348223e-01 5.29000580e-01
1.11868131e+00 -1.33218420e+00 -1.00781515e-01 -8.19355994e-02
-8.95316124e-01 9.38346565e-01 7.95616210e-01 1.46301314e-01
4.41078693e-01 -1.26863912e-01 2.25583255e-01 -1.46222115e-01
-6.39611125e-01 -4.92697239e-01 3.04573148e-01 5.19268215e-01
2.10209042e-01 -2.02249676e-01 -3.05159688e-01 6.85731411e-01
8.79168361e-02 -2.62612142e-02 1.54412568e-01 6.66489184e-01
-6.20308995e-01 -6.62937999e-01 -4.68871921e-01 4.30925220e-01
-2.59160280e-01 -3.74069422e-01 -4.75241065e-01 4.14902240e-01
2.04107136e-01 1.19782424e+00 4.67748269e-02 -3.29993010e-01
2.83719242e-01 1.56980187e-01 2.59967446e-01 -4.28751260e-01
-4.66067612e-01 3.34407657e-01 -3.41918379e-01 -3.47352982e-01
-3.47064584e-01 -5.54062068e-01 -1.26505256e+00 -1.31442234e-01
-3.99740517e-01 -1.93563342e-01 4.79711801e-01 7.93849826e-01
5.26165187e-01 5.56313515e-01 5.36942542e-01 -4.05564070e-01
-2.60570049e-01 -8.49584520e-01 -4.61221218e-01 3.35061789e-01
3.37005913e-01 -6.54889405e-01 -3.47744286e-01 1.18442237e-01]
|
[11.7047119140625, 2.0527307987213135]
|
349e3b6b-1523-426d-ab78-3eee9ae6d5c5
|
portfolio-optimization-with-idiosyncratic-and
|
2111.11286
| null |
https://arxiv.org/abs/2111.11286v1
|
https://arxiv.org/pdf/2111.11286v1.pdf
|
Portfolio optimization with idiosyncratic and systemic risks for financial networks
|
In this study, we propose a new multi-objective portfolio optimization with idiosyncratic and systemic risks for financial networks. The two risks are measured by the idiosyncratic variance and the network clustering coefficient derived from the asset correlation networks, respectively. We construct three types of financial networks in which nodes indicate assets and edges are based on three correlation measures. Starting from the multi-objective model, we formulate and solve the asset allocation problem. We find that the optimal portfolios obtained through the multi-objective with networked approach have a significant over-performance in terms of return measures in an out-of-sample framework. This is further supported by the less drawdown during the periods of the stock market fluctuating downward. According to analyzing different datasets, we also show that improvements made to portfolio strategies are robust.
|
['Jihui Han', 'Chao Wang', 'Lin Chen', 'Longfeng Zhao', 'Yajie Yang']
|
2021-11-22
| null | null | null | null |
['portfolio-optimization']
|
['time-series']
|
[-4.17349756e-01 3.85234095e-02 -2.14361817e-01 -1.70395654e-02
4.68227863e-02 -7.85860896e-01 4.41448718e-01 -1.49085268e-01
-8.70325416e-03 8.21701527e-01 3.06122690e-01 -9.13309529e-02
-1.41104043e+00 -1.37999582e+00 -1.20874099e-01 -5.70791185e-01
-5.49874783e-01 3.28124911e-01 -2.33910326e-02 -1.93511754e-01
5.18019319e-01 4.36042637e-01 -8.50876272e-01 -2.13278219e-01
7.21150994e-01 1.34625161e+00 -4.47064698e-01 1.40044019e-01
-1.78587258e-01 7.89290667e-01 -5.24804771e-01 -8.96412671e-01
7.32584178e-01 -2.43054256e-01 -3.30279559e-01 7.31431227e-03
-4.64787662e-01 -1.69509590e-01 -1.29194289e-01 1.22951722e+00
4.41327661e-01 1.58825263e-01 7.55669355e-01 -1.25503135e+00
-6.19080961e-01 9.92800176e-01 -4.89220738e-01 4.22414899e-01
2.70980112e-02 -8.27800483e-02 1.61577559e+00 -5.37208378e-01
4.58199143e-01 1.21473360e+00 7.60031819e-01 -1.81788355e-01
-1.00245321e+00 -3.21266472e-01 2.12806702e-01 -1.84660599e-01
-8.29321206e-01 1.31510794e-01 8.70171785e-01 -5.36662579e-01
7.17557669e-01 3.86255741e-01 6.19772077e-01 5.48916638e-01
3.90020370e-01 -2.77370936e-03 9.88429785e-01 -1.11557588e-01
6.07904196e-02 2.51149744e-01 2.17904031e-01 1.47636622e-01
1.05726469e+00 3.48575383e-01 -3.73675942e-01 -3.19889098e-01
6.86262608e-01 8.84177685e-02 -1.44232392e-01 1.23564459e-01
-8.33147764e-01 9.39746916e-01 1.64727643e-01 4.84609812e-01
-6.97835565e-01 -6.22552149e-02 8.92035365e-02 6.70984566e-01
7.87443757e-01 6.22464418e-01 -2.47161791e-01 1.81717142e-01
-8.52529049e-01 -2.39965711e-02 1.25996900e+00 5.18838823e-01
1.48042813e-01 3.47642243e-01 -1.01901799e-01 4.55969930e-01
6.25188231e-01 4.69441950e-01 2.76805937e-01 -1.07146072e+00
7.88172126e-01 6.43276453e-01 1.59094676e-01 -1.64328027e+00
-5.12733102e-01 -1.09583497e+00 -9.36361492e-01 4.32387143e-01
4.44677562e-01 -6.25537276e-01 1.93732560e-01 1.46214366e+00
-1.27408788e-01 4.36695404e-02 1.60835341e-01 4.34743941e-01
1.13491118e-01 3.96128178e-01 -3.88125837e-01 -6.43915117e-01
1.11353147e+00 -1.13272989e+00 -9.49491560e-01 3.15979838e-01
8.33981261e-02 -3.43224972e-01 1.08518355e-01 5.13414443e-01
-1.24198186e+00 -1.38321340e-01 -9.30267453e-01 1.10105073e+00
-2.48501822e-01 -4.24603879e-01 4.26917493e-01 1.03185129e+00
-1.18873358e+00 1.13761616e+00 -1.82100877e-01 1.07897483e-01
2.26927444e-01 1.77758113e-01 1.96447372e-01 5.36335230e-01
-1.28453040e+00 8.66477251e-01 2.71777898e-01 7.81678483e-02
-5.44311345e-01 -6.37996435e-01 -9.44643375e-03 4.16969091e-01
6.25668705e-01 -8.67090881e-01 3.92915845e-01 -8.87249410e-01
-1.47331274e+00 -1.39461691e-02 7.84757376e-01 -6.07789755e-01
9.83081400e-01 -1.51416123e-01 -3.88450325e-01 3.27546775e-01
-1.42446607e-01 -6.52947187e-01 4.89080936e-01 -9.02586341e-01
-3.18737894e-01 -1.45620331e-01 7.60304779e-02 -2.34981440e-02
-8.74268353e-01 4.43027169e-01 1.95107430e-01 -1.21952474e+00
-1.90542936e-01 -5.60404480e-01 -4.36164558e-01 -4.75161403e-01
-6.79037213e-01 3.11344475e-01 2.72272021e-01 -8.13561916e-01
1.64232051e+00 -1.68553078e+00 1.97860360e-01 1.10985601e+00
2.02645436e-01 -2.71322966e-01 6.13107197e-02 6.31274760e-01
-2.91015416e-01 7.55073786e-01 -9.74544266e-04 -9.45877060e-02
3.64716560e-01 -2.79982299e-01 -9.01035294e-02 3.17723662e-01
4.16669920e-02 6.75578237e-01 -6.29965007e-01 -1.77378327e-01
-2.38033175e-01 -1.80272922e-01 -1.68210208e-01 1.83720917e-01
6.31620660e-02 5.66563830e-02 -7.99556494e-01 6.80629969e-01
5.03864050e-01 -4.73616093e-01 4.40575480e-01 6.40955418e-02
1.38002932e-02 -4.51481715e-02 -1.39700067e+00 7.02289820e-01
1.53127154e-02 1.71137914e-01 -1.12218350e-01 -9.26338732e-01
1.16145873e+00 4.89116997e-01 6.73117161e-01 -3.90002102e-01
6.05185740e-02 3.46645862e-01 3.25642765e-01 -2.12273076e-01
2.25059167e-01 -3.97917449e-01 1.45330459e-01 1.07363808e+00
-7.27757514e-02 3.35157186e-01 3.50187510e-01 2.61797197e-02
1.11919808e+00 -4.24411207e-01 2.22112715e-01 -6.52749240e-01
4.58082944e-01 -5.00604153e-01 4.25326496e-01 6.88712656e-01
1.06007107e-01 1.01553254e-01 1.33480215e+00 7.77385384e-02
-8.19524944e-01 -1.03435886e+00 -1.31803413e-03 3.57699752e-01
-4.13690358e-01 1.39846459e-01 -5.99660337e-01 -6.98570788e-01
5.65941513e-01 4.74466443e-01 -4.69357491e-01 2.23201096e-01
-6.89242780e-02 -1.24625504e+00 3.62684011e-01 3.08464497e-01
4.91722167e-01 -8.40420187e-01 -3.89916986e-01 3.69129509e-01
4.43980902e-01 -7.04983771e-01 -2.82879859e-01 -9.43550095e-02
-1.00292325e+00 -1.18340111e+00 -9.58119869e-01 7.22855479e-02
4.68244761e-01 -2.23917559e-01 1.12415099e+00 1.79711744e-01
1.06138840e-01 5.69414079e-01 -3.53814304e-01 -2.89063215e-01
-4.08002548e-03 -4.07270081e-02 2.38420129e-01 4.45436597e-01
-3.40983897e-01 -9.49515760e-01 -6.93202853e-01 3.32636267e-01
-6.65811062e-01 -7.69378901e-01 3.88003320e-01 5.43159246e-01
1.22567460e-01 7.15784431e-01 1.32796776e+00 -7.23101437e-01
1.19368136e+00 -1.06302214e+00 -1.01147985e+00 4.26175743e-01
-1.18603766e+00 -7.40807205e-02 2.78590411e-01 -2.00940207e-01
-1.12550747e+00 -8.20293188e-01 6.27626479e-01 -1.32012025e-01
6.51371121e-01 1.09586203e+00 2.84752273e-03 -3.38465929e-01
-1.25231907e-01 -3.60943049e-01 1.66337177e-01 -5.54792881e-01
9.66948792e-02 1.68444365e-01 -1.69588000e-01 -5.20607412e-01
1.04758275e+00 2.71219969e-01 5.68257689e-01 -2.33244583e-01
-6.33486390e-01 2.00907588e-01 -3.96884590e-01 -6.75802946e-01
6.42257214e-01 -4.23024654e-01 -7.92298615e-01 6.75113976e-01
-8.26846957e-01 1.23087645e-01 -2.90759742e-01 7.59378254e-01
-3.39491248e-01 4.13635939e-01 -6.64370954e-01 -1.40747833e+00
-4.07944649e-01 -6.22333229e-01 -2.68426687e-01 1.24129266e-01
-8.58344585e-02 -1.76197839e+00 2.69796342e-01 2.59538442e-01
8.15349102e-01 8.93775105e-01 8.73569965e-01 -1.17937577e+00
-8.86459053e-01 -1.96873143e-01 -2.39109010e-01 5.23942053e-01
1.79904625e-01 1.88940495e-01 -4.21252161e-01 -1.76190436e-01
3.90759766e-01 2.48782307e-01 5.99763572e-01 3.64658684e-01
7.68706501e-01 -6.33832335e-01 1.14756788e-03 4.59738553e-01
1.84667206e+00 5.20688772e-01 3.45781714e-01 8.19395125e-01
1.09628037e-01 1.24017155e+00 6.26791775e-01 9.38702524e-01
1.67834722e-02 3.65768492e-01 6.06758952e-01 3.23362857e-01
5.41804135e-01 3.22533399e-02 3.84311944e-01 1.03882074e+00
-5.39977849e-01 -5.45541167e-01 -9.48294282e-01 4.35846269e-01
-1.95751083e+00 -1.29832888e+00 -1.62281916e-01 2.08005524e+00
3.64538014e-01 4.57960874e-01 5.82236052e-01 -1.63546633e-02
6.93671286e-01 3.21419209e-01 -3.38536024e-01 -8.52698162e-02
-5.97965002e-01 1.43471673e-01 9.04616416e-01 4.81877506e-01
-7.14837015e-01 2.17259657e-02 7.17082357e+00 5.69439530e-01
-5.00738442e-01 -7.05388635e-02 1.06707668e+00 -5.33004999e-01
-7.42614508e-01 -1.02740198e-01 -6.31141126e-01 8.10736597e-01
1.10555530e+00 -7.95602500e-01 3.05499673e-01 4.16287750e-01
2.96042234e-01 1.37646347e-01 -5.94760120e-01 2.49456838e-01
-2.10084096e-01 -1.28028512e+00 -2.14649364e-01 5.53763509e-01
8.87131989e-01 -3.99570018e-01 2.94481397e-01 -4.33135539e-01
5.49502552e-01 -1.07267809e+00 7.50200629e-01 1.43827593e+00
6.17817864e-02 -1.09441876e+00 1.04992211e+00 2.90054120e-02
-1.14494836e+00 -4.79775935e-01 -1.74161419e-01 -3.01440842e-02
5.75274885e-01 1.02714467e+00 5.50667457e-02 1.11194193e+00
6.36874259e-01 6.71983004e-01 -3.07044834e-01 1.19560885e+00
-1.19968817e-01 4.50623423e-01 -1.06701724e-01 -1.96418107e-01
1.61378264e-01 -1.24314988e+00 7.44284689e-01 5.47767937e-01
9.18940246e-01 1.17100306e-01 -3.93243402e-01 1.19315064e+00
-8.73826537e-03 2.02487335e-02 -7.30434895e-01 -2.99224794e-01
4.29458529e-01 1.23700869e+00 -7.24672139e-01 -9.24873129e-02
-5.07270277e-01 2.24010780e-01 -1.79570183e-01 4.74319130e-01
-6.08543932e-01 -6.79903328e-01 4.50611562e-01 -8.04833695e-02
2.60274112e-01 -1.15474820e-01 -6.27041757e-01 -8.92164171e-01
2.59575337e-01 -5.60625613e-01 3.74146283e-01 -4.07160997e-01
-1.76319289e+00 4.26976383e-01 5.13567962e-02 -1.22128785e+00
-2.60435224e-01 -7.11069524e-01 -1.06384921e+00 1.04824460e+00
-1.61732149e+00 -3.16218495e-01 2.07180440e-01 2.98521310e-01
-3.76038760e-01 -7.61328995e-01 3.31734508e-01 3.83249342e-01
-8.86039019e-01 4.42774981e-01 6.54058099e-01 4.10792157e-02
1.67473271e-01 -1.26521122e+00 9.64934900e-02 7.94882119e-01
-1.91336870e-01 4.19844896e-01 3.80310506e-01 -9.92101610e-01
-8.64064217e-01 -7.57611036e-01 7.42878258e-01 -2.92797804e-01
1.48518693e+00 2.66974002e-01 -5.96668899e-01 5.22599876e-01
6.43841743e-01 -3.73557836e-01 8.65055263e-01 -1.44827411e-01
2.64298748e-02 -2.19438300e-01 -1.47760046e+00 2.50167817e-01
9.60644364e-01 -1.17590353e-01 -4.41565931e-01 3.80215496e-01
7.76501238e-01 5.54861367e-01 -1.66831732e+00 1.63291737e-01
4.25377429e-01 -1.22979820e+00 1.07091916e+00 -6.53165460e-01
3.23574960e-01 1.83969200e-01 -1.23279653e-01 -1.33063757e+00
-5.67533195e-01 -8.07210982e-01 -7.39995092e-02 1.65914309e+00
7.84589708e-01 -1.52703619e+00 7.36744523e-01 4.55845147e-01
5.49669981e-01 -7.41330743e-01 -1.01293135e+00 -1.30932164e+00
1.57981157e-01 2.18249977e-01 9.64412212e-01 9.29561913e-01
1.20959920e-03 -1.86483815e-01 -4.10435766e-01 -1.33664340e-01
1.09205830e+00 1.16432153e-01 1.20419972e-01 -1.48305988e+00
-6.40072405e-01 -7.90318787e-01 -9.81177464e-02 -1.26291052e-01
1.89321309e-01 -6.82789385e-01 -8.28517377e-01 -1.05747700e+00
1.89708576e-01 -3.40100050e-01 -8.31405342e-01 -9.14966036e-03
1.87882617e-01 -1.79213077e-01 6.55326188e-01 3.06732416e-01
-2.60983586e-01 5.11145473e-01 9.19648767e-01 7.32387826e-02
1.11870952e-01 5.06763995e-01 -8.01896811e-01 7.32449710e-01
1.07182288e+00 -5.14854014e-01 -3.20713222e-01 3.62174101e-02
8.37482631e-01 6.35227025e-01 1.09858043e-01 -7.34046519e-01
-1.46183996e-02 -4.68560874e-01 3.18952277e-02 -4.44179952e-01
8.22589621e-02 -9.86817181e-01 6.02259934e-01 7.13212788e-01
-1.39444366e-01 6.87856495e-01 -2.19460025e-01 7.07284927e-01
-3.29756945e-01 -7.07640827e-01 4.27001566e-01 -1.59599125e-01
2.75140077e-01 2.19286248e-01 -1.34737924e-01 -2.10869815e-02
1.19472647e+00 -8.17185938e-02 -6.93778217e-01 -6.47378862e-01
-8.15899968e-01 3.72380793e-01 8.95062312e-02 -5.71054965e-02
4.19384420e-01 -1.53372371e+00 -8.38925302e-01 -4.26320016e-01
-6.93275392e-01 -6.90196216e-01 -1.82522796e-02 1.03646851e+00
-4.13811177e-01 5.90275884e-01 -3.94910008e-01 2.50031710e-01
-9.35809076e-01 2.34816149e-01 7.61398435e-01 -8.73463988e-01
6.04909770e-02 6.05905771e-01 -1.79502130e-01 -1.55411102e-02
1.00215934e-01 5.80160916e-02 -5.91524780e-01 8.56792271e-01
1.57477215e-01 1.21295083e+00 -3.32343340e-01 -2.98729628e-01
-2.33205423e-01 7.96185076e-01 4.43395972e-01 -4.54088748e-01
1.77775812e+00 -3.12192649e-01 -6.77584708e-01 3.99563760e-01
8.11932981e-01 2.94729590e-01 -1.01145828e+00 -5.27429692e-02
7.09953249e-01 -5.44997752e-01 -2.66744763e-01 -7.65180647e-01
-1.79202914e+00 3.00376594e-01 1.25288725e-01 1.03796828e+00
1.06534719e+00 -5.01702249e-01 3.08186561e-01 2.07424000e-01
3.53985697e-01 -1.05166709e+00 3.92375886e-01 4.14096326e-01
9.99516785e-01 -5.30334651e-01 2.56843656e-01 -4.09918904e-01
-4.67113048e-01 1.22766149e+00 1.88687101e-01 -4.69253540e-01
1.27631569e+00 1.14266805e-01 -2.92472154e-01 -3.73398244e-01
-7.70383298e-01 1.04541853e-01 3.82375836e-01 3.86829436e-01
7.81209916e-02 2.94837244e-02 -5.83326995e-01 1.16087437e+00
-2.01643929e-01 -5.02862394e-01 7.09322929e-01 4.71911520e-01
-4.04864848e-01 -1.18094182e+00 -4.62969869e-01 6.50052667e-01
-9.72366452e-01 -1.51857540e-01 -6.61865830e-01 8.47868443e-01
-4.58085865e-01 8.91403615e-01 8.98973048e-02 -3.24485719e-01
4.60433632e-01 -6.07573576e-02 -1.38716310e-01 -3.31540942e-01
-1.10429764e+00 1.93528518e-01 3.96127343e-01 -2.81944215e-01
-5.28400123e-01 -8.23593795e-01 -5.86836398e-01 -5.31096995e-01
-5.43784797e-01 1.65515870e-01 2.86687374e-01 6.58437490e-01
2.45727807e-01 9.14416075e-01 1.26706433e+00 -4.35131103e-01
-1.19840372e+00 -8.72705162e-01 -1.36673784e+00 -3.98528390e-02
-4.49356772e-02 -5.63247979e-01 -1.00263190e+00 -7.05972970e-01]
|
[4.999718189239502, 4.041356086730957]
|
7e573cc0-6b76-4f2a-ac47-971ffdebe93d
|
introduction-to-latent-variable-energy-based
|
2306.02572
| null |
https://arxiv.org/abs/2306.02572v1
|
https://arxiv.org/pdf/2306.02572v1.pdf
|
Introduction to Latent Variable Energy-Based Models: A Path Towards Autonomous Machine Intelligence
|
Current automated systems have crucial limitations that need to be addressed before artificial intelligence can reach human-like levels and bring new technological revolutions. Among others, our societies still lack Level 5 self-driving cars, domestic robots, and virtual assistants that learn reliable world models, reason, and plan complex action sequences. In these notes, we summarize the main ideas behind the architecture of autonomous intelligence of the future proposed by Yann LeCun. In particular, we introduce energy-based and latent variable models and combine their advantages in the building block of LeCun's proposal, that is, in the hierarchical joint embedding predictive architecture (H-JEPA).
|
['Yann Lecun', 'Anna Dawid']
|
2023-06-05
| null | null | null | null |
['self-driving-cars']
|
['computer-vision']
|
[-3.54655176e-01 5.30509591e-01 -4.06413704e-01 -1.30021527e-01
6.23835362e-02 -2.25152835e-01 1.02391469e+00 -4.27768677e-01
-1.36058524e-01 9.23387647e-01 1.55637234e-01 -8.03636312e-02
-3.74750763e-01 -1.03375340e+00 -3.36402237e-01 -4.83780593e-01
-6.75562546e-02 8.03462267e-01 1.10239647e-01 -6.11536920e-01
1.20623656e-01 3.83351684e-01 -1.46426141e+00 -4.25375938e-01
1.05253673e+00 5.27213812e-01 4.70246404e-01 8.60362530e-01
1.33409768e-01 1.38542056e+00 -2.90019453e-01 -3.82913053e-01
2.88984329e-02 -2.67753124e-01 -6.80561543e-01 -3.71531427e-01
-5.14167547e-01 -3.04384381e-01 -9.36672449e-01 7.93300509e-01
-4.86075059e-02 3.06653470e-01 1.12420130e+00 -1.63077104e+00
-1.02908123e+00 4.64082032e-01 1.23179607e-01 -3.05204749e-01
1.48771703e-01 3.82591933e-01 8.18596840e-01 -5.53147614e-01
8.51748407e-01 1.30345201e+00 5.41384697e-01 8.56718957e-01
-1.06027496e+00 -1.70040876e-01 1.26862690e-01 8.35569501e-01
-1.14256930e+00 -5.73526740e-01 9.59552050e-01 -4.85154241e-01
1.46958911e+00 -1.65546417e-01 9.06676531e-01 1.31374407e+00
8.70242774e-01 8.07412863e-01 6.43175781e-01 -3.97635013e-01
3.98229688e-01 1.44295588e-01 1.95193410e-01 9.58441556e-01
6.95773482e-01 2.53815740e-01 -6.08057499e-01 4.68163528e-02
4.69922364e-01 -9.11724493e-02 2.69636482e-01 -7.49799609e-01
-1.33037031e+00 9.71764565e-01 4.17321265e-01 4.41795290e-01
-8.14340413e-01 5.86162448e-01 -9.32841189e-03 1.33631807e-02
1.30013958e-01 6.38508439e-01 -4.62274313e-01 -4.65332180e-01
-4.61220413e-01 6.70788214e-02 7.74500787e-01 9.36156392e-01
6.81687117e-01 5.22369444e-01 4.38642025e-01 2.75246382e-01
4.53809887e-01 3.83463800e-01 6.26174450e-01 -1.41827905e+00
-7.52685741e-02 6.17605269e-01 3.94917041e-01 -1.12536609e+00
-6.51732028e-01 -1.78266577e-02 -7.18973577e-01 1.08972177e-01
-3.18518192e-01 -3.30258787e-01 -6.10590935e-01 1.40925407e+00
-8.17674175e-02 4.15599011e-02 6.80089772e-01 4.82826352e-01
4.49154168e-01 8.56167197e-01 2.84289658e-01 2.43906975e-02
1.10447490e+00 -1.20360684e+00 -1.07094479e+00 -6.01910889e-01
4.06223863e-01 1.07272349e-01 4.93048638e-01 3.28789622e-01
-1.06719255e+00 -7.67828822e-01 -1.56486678e+00 -3.42087448e-01
-7.90934205e-01 -5.41252270e-03 1.21118903e+00 3.94752681e-01
-1.26915777e+00 4.78841275e-01 -1.28561175e+00 -4.37941283e-01
9.06926692e-02 3.32734227e-01 -3.01915437e-01 2.69546151e-01
-1.38809574e+00 1.58585072e+00 6.77848458e-01 -2.91685224e-01
-9.21806931e-01 -1.31962582e-01 -1.00411534e+00 -7.07514808e-02
1.84593663e-01 -1.06437218e+00 1.11537123e+00 -4.77819473e-01
-1.85256886e+00 4.24350798e-01 -3.03695410e-01 -8.68113101e-01
4.99815568e-02 -3.36277634e-01 -5.71302712e-01 1.21143125e-01
-1.13890074e-01 9.38463509e-01 5.20451546e-01 -1.15261149e+00
-7.33492494e-01 -3.01056296e-01 7.30946437e-02 1.15564197e-01
-4.33702767e-01 -4.70530570e-01 -2.65732229e-01 -4.98282909e-02
-5.56749590e-02 -1.11332977e+00 -6.58211231e-01 -2.23656267e-01
4.49482761e-02 -5.29834270e-01 7.04633832e-01 -5.36957860e-01
8.73443067e-01 -2.08432150e+00 7.14297473e-01 -2.35663727e-01
5.95888317e-01 -2.95353681e-02 -1.34261101e-01 5.75302958e-01
3.82095605e-01 -2.67569125e-02 3.88600528e-02 -1.41243875e-01
4.23221260e-01 5.41384518e-01 -1.86818585e-01 2.92899132e-01
1.77948341e-01 1.33692634e+00 -1.10113502e+00 -2.97288865e-01
7.08998919e-01 7.09966958e-01 -3.59265387e-01 -1.16053961e-01
-3.10561597e-01 4.09393907e-02 -8.36488426e-01 1.82324961e-01
-8.10737014e-02 -2.76874542e-01 3.73933911e-01 -2.62910426e-02
-2.69247264e-01 2.03461736e-01 -6.87745333e-01 1.57221019e+00
-4.43182290e-01 1.01872098e+00 -7.79993609e-02 -1.04879618e+00
8.70412350e-01 2.57660031e-01 7.94143558e-01 -7.00095594e-01
4.01586071e-02 1.02573022e-01 -1.88009202e-01 -5.07015526e-01
8.07942569e-01 7.91997612e-02 -2.74910867e-01 1.38972014e-01
1.57240883e-01 -3.92697781e-01 2.07386166e-01 2.23891243e-01
1.13133657e+00 4.66506422e-01 5.91515899e-01 -1.43338889e-01
4.24829543e-01 2.57049859e-01 5.40532887e-01 5.70537388e-01
-5.56103766e-01 -2.55073905e-01 3.39437962e-01 -7.77669132e-01
-1.16637087e+00 -1.04953992e+00 2.67836779e-01 8.60499620e-01
2.15531111e-01 -3.91943663e-01 -5.09987235e-01 -3.71872634e-01
-1.65198848e-01 1.38335812e+00 -4.15480167e-01 -4.60972071e-01
-3.24449003e-01 -5.07613063e-01 2.58703381e-01 5.66623569e-01
5.58068335e-01 -1.05157137e+00 -1.19227517e+00 3.85082453e-01
-2.33659193e-01 -1.14188516e+00 5.17852545e-01 3.07901800e-01
-7.28682756e-01 -9.06553447e-01 -1.18459880e-01 -6.10362768e-01
1.42257228e-01 9.80133712e-02 1.08251262e+00 -3.35059375e-01
-2.79971153e-01 6.19799733e-01 -1.86526790e-01 -5.85681558e-01
-3.78003120e-01 3.27448994e-02 6.09108984e-01 -6.47094607e-01
5.28375864e-01 -6.76992238e-01 -3.16133976e-01 -6.58678859e-02
-4.11909848e-01 3.25870752e-01 8.11798573e-01 7.80059934e-01
1.15882486e-01 3.10568184e-01 6.13681734e-01 -1.28876597e-01
4.45560396e-01 -6.49012923e-01 -5.05536079e-01 2.83241034e-01
-8.36741805e-01 3.31284195e-01 5.39172888e-01 -1.32204890e-01
-1.02678478e+00 7.99533650e-02 1.02763511e-01 -5.59489466e-02
-6.67509213e-02 3.58118176e-01 -2.56667405e-01 1.84324011e-01
4.51471627e-01 4.42247719e-01 3.72072607e-02 -7.30542243e-02
8.53878915e-01 7.86702216e-01 5.42129278e-01 -1.94126770e-01
8.47984791e-01 5.21708429e-01 9.62940678e-02 -9.71494734e-01
-2.48232529e-01 1.59162730e-01 -7.82009244e-01 -1.88934520e-01
1.22190881e+00 -9.15298939e-01 -7.77778387e-01 1.99847668e-01
-1.18155265e+00 -4.01760668e-01 -4.19808358e-01 6.16554320e-01
-1.18632281e+00 3.99688892e-02 -6.58522546e-01 -1.07888961e+00
-1.14225522e-01 -9.00363684e-01 6.15998268e-01 3.35855603e-01
-3.10837835e-01 -1.09437239e+00 2.65657067e-01 3.35229993e-01
5.75482666e-01 1.90290824e-01 9.18086171e-01 -4.44353372e-01
-5.39731622e-01 -1.70701548e-01 1.61454380e-01 5.15243709e-02
5.99909313e-02 1.53405905e-01 -7.98712730e-01 6.50764480e-02
-5.09288944e-02 -1.96248233e-01 6.22428060e-01 3.68004918e-01
4.79743004e-01 -3.36820900e-01 -8.00086021e-01 3.84995699e-01
1.19111252e+00 6.54339969e-01 5.66528916e-01 3.94284934e-01
2.85249799e-01 6.20633543e-01 3.54902744e-01 2.39873812e-01
8.42479289e-01 5.07300198e-01 2.62802452e-01 3.16293210e-01
1.10155523e-01 -3.52951020e-01 6.42913103e-01 1.23706496e+00
-3.54139537e-01 -1.26723975e-01 -9.32331920e-01 5.43320537e-01
-2.13017273e+00 -1.03542864e+00 1.69010162e-02 1.42676055e+00
2.27624923e-01 1.78997487e-01 -2.31230617e-01 -1.16593026e-01
2.48320788e-01 1.11500300e-01 -7.88456321e-01 -4.53239948e-01
-1.14302225e-01 -2.97475398e-01 4.33582515e-01 6.13764226e-01
-1.08056366e+00 1.31615937e+00 7.85046005e+00 3.74785304e-01
-5.43839872e-01 2.45716929e-01 2.88771778e-01 1.96898282e-01
-3.96225154e-01 -1.79842245e-02 -4.26315486e-01 3.17079455e-01
1.60769498e+00 -3.03436220e-01 7.03634143e-01 1.14354181e+00
7.95448497e-02 1.00803390e-01 -8.84324968e-01 8.60754550e-01
2.29212176e-02 -1.25244319e+00 -1.41550273e-01 1.61605895e-01
5.99263251e-01 2.23305419e-01 1.00160584e-01 8.09836745e-01
8.82425964e-01 -9.39855278e-01 5.98927379e-01 8.92439783e-01
2.54135758e-01 -7.47213840e-01 5.71916819e-01 4.60325629e-01
-9.65683341e-01 -5.10681033e-01 -5.34666002e-01 -4.71311063e-01
2.41519660e-01 2.88796961e-01 -5.01703441e-01 5.75356901e-01
2.74204940e-01 7.80729711e-01 -3.18894207e-01 3.98396730e-01
-3.69500458e-01 2.80490220e-01 -8.02920237e-02 -4.58459258e-01
3.62101942e-01 -3.75643939e-01 6.33988023e-01 6.95932388e-01
2.85913944e-01 1.63932979e-01 -1.55851603e-01 8.00109029e-01
2.51289070e-01 -4.61205751e-01 -9.53427434e-01 -4.01226670e-01
5.78467786e-01 9.92298245e-01 -5.39137602e-01 -3.84024292e-01
-4.77476269e-01 1.10805488e+00 2.52183527e-01 3.14824253e-01
-9.27284002e-01 -4.79654938e-01 8.64515364e-01 -3.98927778e-01
-6.26448616e-02 -9.47663188e-01 -4.07213837e-01 -1.17444611e+00
-3.32750112e-01 -5.85505486e-01 -1.04409046e-01 -9.12215829e-01
-8.48125577e-01 3.61602217e-01 -1.82714779e-02 -5.82126498e-01
-7.18437195e-01 -8.28051448e-01 -3.47570568e-01 3.38903934e-01
-1.31523705e+00 -1.38274622e+00 -7.04322234e-02 3.13355982e-01
7.52058566e-01 -4.81513172e-01 1.15341258e+00 -2.37993643e-01
-5.08627892e-01 3.23639922e-02 4.05370831e-01 -8.40269327e-02
-1.47068165e-02 -1.16405296e+00 7.34534442e-01 7.18386352e-01
2.71091729e-01 3.85567188e-01 9.56875801e-01 -6.84870422e-01
-1.98903847e+00 -8.04198503e-01 8.85723889e-01 -7.56953478e-01
9.49692249e-01 -2.31364042e-01 -2.55539626e-01 1.05611801e+00
3.80905569e-01 -5.38453698e-01 3.78627598e-01 2.42975235e-01
7.81294182e-02 -7.63237700e-02 -1.14704180e+00 8.78971577e-01
9.68837857e-01 -4.17621434e-01 -8.73436332e-01 2.53887653e-01
1.02363348e+00 3.93764675e-01 -6.77144885e-01 3.27378541e-01
7.84125507e-01 -6.89048409e-01 1.03473389e+00 -7.31108904e-01
1.67854831e-01 -1.54716447e-01 -2.63383001e-01 -1.18059814e+00
-8.49522471e-01 -8.82932663e-01 -7.03880668e-01 6.42287076e-01
2.78801382e-01 -8.40848863e-01 6.42879784e-01 8.22925985e-01
-1.83356479e-01 -4.42687124e-01 -9.92727816e-01 -8.43909204e-01
1.35555506e-01 -5.36218703e-01 5.86382508e-01 6.71779037e-01
3.99161190e-01 3.79566193e-01 -4.46549386e-01 2.14490704e-02
6.63989365e-01 -4.12266940e-01 6.81418777e-01 -1.45265126e+00
3.20213214e-02 -4.83840793e-01 -8.47676039e-01 -7.65965641e-01
5.30648947e-01 -4.31968123e-01 1.61954444e-02 -2.15872240e+00
1.28566518e-01 -1.33903116e-01 -3.70355546e-01 4.04283553e-01
5.70765547e-02 -1.82211235e-01 3.29294056e-01 2.83787519e-01
-8.47743750e-01 9.82246637e-01 7.04795241e-01 -2.30399773e-01
-2.73541898e-01 -4.27244991e-01 -7.43837118e-01 1.02036405e+00
9.58807111e-01 -2.30410900e-02 -4.02015388e-01 -3.20670217e-01
3.66235733e-01 3.22790444e-02 3.12802911e-01 -1.42637122e+00
5.60475349e-01 -4.80015635e-01 4.42887634e-01 -2.70498902e-01
8.79524291e-01 -8.88673246e-01 2.20940679e-01 9.21508849e-01
2.96666054e-03 1.08217083e-01 -2.52336171e-03 5.79337716e-01
1.91266924e-01 4.04408462e-02 5.29100657e-01 -4.64275852e-02
-1.12389183e+00 -1.24976650e-01 -1.08809960e+00 -6.30985379e-01
1.51725745e+00 -4.71903235e-02 -3.23540837e-01 -4.39710915e-01
-5.33859611e-01 4.07080501e-01 4.44567949e-01 7.66712189e-01
6.55525804e-01 -1.03857780e+00 -5.45357704e-01 5.42035922e-02
-1.65470183e-01 -5.48026145e-01 1.32145420e-01 3.20622832e-01
-4.62467432e-01 1.08858037e+00 -4.91798520e-01 -1.54629657e-02
-4.93669868e-01 9.61932778e-01 1.46657541e-01 -2.26619571e-01
-6.93267822e-01 3.52611959e-01 -2.00895548e-01 -4.80185032e-01
-9.49866995e-02 -6.16668500e-02 -4.38503385e-01 -2.01101258e-01
2.75586993e-01 6.92017555e-01 -6.21211886e-01 -5.70002079e-01
-4.06951666e-01 2.87376881e-01 4.50203478e-01 -3.62700462e-01
1.41649711e+00 -3.60007882e-01 -7.49323815e-02 8.76730025e-01
6.71643257e-01 -4.01202828e-01 -9.64278221e-01 3.23564529e-01
1.79454535e-01 1.58582970e-01 2.60842174e-01 -7.25787401e-01
-4.44783688e-01 7.59516656e-01 6.49003327e-01 1.62152901e-01
7.90680647e-01 2.76848637e-02 8.48359764e-01 9.97426689e-01
8.64823341e-01 -1.34418654e+00 -2.51835346e-01 6.29743993e-01
5.49128056e-01 -1.04314101e+00 1.66238248e-02 1.89791799e-01
-7.00107098e-01 1.03375149e+00 5.71313798e-01 -1.09174989e-01
7.55799532e-01 1.42397851e-01 -4.13953155e-01 -2.91781306e-01
-1.24052012e+00 -2.67226905e-01 -4.40508537e-02 1.03171539e+00
1.13377526e-01 5.30573785e-01 -3.70109737e-01 6.84715807e-01
-2.82522231e-01 7.11473078e-02 4.25822794e-01 8.27507973e-01
-9.48535740e-01 -8.15420687e-01 2.19707154e-02 2.00575739e-01
1.35578990e-01 2.44326070e-01 -3.56484979e-01 8.18906486e-01
1.16334021e-01 1.16935503e+00 3.02485954e-02 -7.55696476e-01
8.55945721e-02 4.20064926e-01 2.61775970e-01 -4.88374755e-02
2.31320992e-01 -6.07396781e-01 -8.00658390e-03 -7.96758533e-01
-3.10048103e-01 -5.56402743e-01 -1.32974660e+00 -2.29763746e-01
8.46598297e-02 2.74306744e-01 1.06502140e+00 8.09760690e-01
7.33798265e-01 5.63488781e-01 4.68463212e-01 -9.97071147e-01
-2.40239933e-01 -1.01469839e+00 -2.51222402e-01 -3.64145339e-01
1.92184940e-01 -8.55875790e-01 -2.49271333e-01 1.01247065e-01]
|
[4.528379917144775, 1.167421579360962]
|
fa4176c1-1c0e-4bd0-9855-e634b59b73e0
|
considerations-for-meaningful-sign-language
|
2211.15464
| null |
https://arxiv.org/abs/2211.15464v1
|
https://arxiv.org/pdf/2211.15464v1.pdf
|
Considerations for meaningful sign language machine translation based on glosses
|
Automatic sign language processing is gaining popularity in Natural Language Processing (NLP) research (Yin et al., 2021). In machine translation (MT) in particular, sign language translation based on glosses is a prominent approach. In this paper, we review recent works on neural gloss translation. We find that limitations of glosses in general and limitations of specific datasets are not discussed in a transparent manner and that there is no common standard for evaluation. To address these issues, we put forward concrete recommendations for future research on gloss translation. Our suggestions advocate awareness of the inherent limitations of gloss-based approaches, realistic datasets, stronger baselines and convincing evaluation.
|
['Sarah Ebling', 'Annette Rios', 'Amit Moryossef', 'Zifan Jiang', 'Mathias Müller']
|
2022-11-28
| null | null | null | null |
['sign-language-translation']
|
['computer-vision']
|
[ 3.83391351e-01 -1.32179230e-01 -6.84592187e-01 -5.51244795e-01
-1.04617941e+00 -4.78794843e-01 6.85039520e-01 -4.30657506e-01
-6.97462618e-01 9.16486144e-01 9.30232167e-01 -3.08720887e-01
3.27309482e-02 -2.90432036e-01 -4.61946964e-01 -3.59775007e-01
5.16836941e-01 4.47664738e-01 -2.48248074e-02 -2.35726640e-01
2.46695846e-01 3.02430928e-01 -1.50367582e+00 3.06462318e-01
9.22727764e-01 6.13134682e-01 -1.21753164e-01 3.45368683e-01
-3.05323869e-01 5.03059030e-01 -7.28056848e-01 -8.13980758e-01
3.26976418e-01 -6.52139246e-01 -7.74269700e-01 -1.69645831e-01
1.10679996e+00 -5.34916639e-01 -1.52986929e-01 8.71347666e-01
8.26455414e-01 -1.73638970e-01 7.49163628e-01 -1.19151676e+00
-1.20918238e+00 4.22080487e-01 -1.62498772e-01 2.19537728e-02
1.74999714e-01 5.73483765e-01 1.14136851e+00 -8.60221803e-01
9.77573693e-01 1.32429624e+00 4.17971194e-01 1.07873452e+00
-6.77595913e-01 -6.35318458e-01 3.32267374e-01 7.14385882e-02
-9.29737926e-01 -4.39296722e-01 3.87187660e-01 -1.31009787e-01
9.24281895e-01 3.35555375e-01 1.04758310e+00 1.55068147e+00
2.10104644e-01 1.14585757e+00 1.53059292e+00 -7.18933582e-01
-1.34299353e-01 -1.93687528e-01 1.76524878e-01 6.57657266e-01
7.00705886e-01 2.79546708e-01 -1.06243241e+00 1.28927574e-01
7.67146409e-01 -6.40700758e-01 3.27693187e-02 4.13736962e-02
-1.43838573e+00 6.91945791e-01 5.48224598e-02 3.66827756e-01
-3.68115515e-01 5.76962709e-01 5.62463701e-01 5.32644808e-01
2.02226311e-01 3.74870688e-01 -2.07389608e-01 -5.20518422e-01
-9.73539650e-01 4.38572884e-01 6.70390844e-01 9.93635058e-01
6.61218837e-02 2.16433585e-01 -2.82702535e-01 8.44612896e-01
6.24267638e-01 9.85901892e-01 4.67335820e-01 -1.10113370e+00
5.63466907e-01 1.90219477e-01 -6.62015676e-02 -4.53257382e-01
-1.28967449e-01 -1.46943349e-02 -1.37172580e-01 3.68782967e-01
7.96089828e-01 -1.36588886e-01 -1.31181586e+00 1.84598446e+00
-2.35478356e-01 -3.66905898e-01 -5.62939495e-02 1.24058735e+00
9.25672472e-01 1.21038295e-01 4.17368710e-01 -5.22659719e-02
1.41511405e+00 -1.05872536e+00 -8.44366670e-01 -3.68300706e-01
3.94059926e-01 -1.17223394e+00 1.47515595e+00 4.05878842e-01
-1.12379301e+00 1.25451595e-01 -6.89214528e-01 -4.99575615e-01
-3.22256058e-01 3.74268174e-01 9.66878295e-01 7.58658707e-01
-1.03505611e+00 9.14025530e-02 -8.40024233e-01 -1.13577020e+00
1.57430530e-01 2.56556034e-01 -1.98880717e-01 -5.10798208e-02
-1.05986965e+00 1.48043692e+00 -1.40154347e-01 3.35028291e-01
-3.90562594e-01 -5.31301014e-02 -5.34437060e-01 -7.96763718e-01
1.15020633e-01 -1.25152028e+00 1.46212983e+00 -1.00962830e+00
-1.93523991e+00 1.14604926e+00 -4.10276830e-01 -2.71721840e-01
5.36549211e-01 -5.35363495e-01 -5.28485537e-01 9.92717892e-02
-9.42755491e-03 9.35049951e-01 6.91824257e-01 -1.06751943e+00
-6.70878708e-01 -9.16396454e-02 -1.72525588e-02 5.77561378e-01
5.40618598e-02 8.68341327e-01 -2.07084715e-01 -7.85778165e-01
2.39636917e-02 -1.05154192e+00 1.34727314e-01 4.86631244e-01
8.37427005e-02 -3.20907086e-01 3.30430001e-01 -8.44904661e-01
1.16978085e+00 -1.59551620e+00 1.65213458e-02 3.51439230e-02
4.20142636e-02 3.61320883e-01 -2.94898480e-01 6.44591630e-01
5.74205399e-01 2.00252816e-01 -2.85297394e-01 -3.85947853e-01
3.17530274e-01 7.72254348e-01 -4.94200528e-01 2.61781216e-01
8.17412362e-02 1.59292698e+00 -8.92576754e-01 -7.03683555e-01
1.17762327e-01 5.10129929e-01 -1.65160790e-01 -5.20574212e-01
-2.27103099e-01 3.43850493e-01 -2.81042278e-01 1.04376113e+00
1.06165573e-01 9.78523120e-02 4.96924333e-02 -4.82848510e-02
-5.00024855e-01 8.73974502e-01 -6.63113654e-01 1.82463348e+00
-3.18966180e-01 1.05113673e+00 -1.95942894e-01 -4.05606180e-01
6.99059665e-01 4.34870243e-01 1.89321086e-01 -5.86214960e-01
1.54599428e-01 8.56902659e-01 2.13050857e-01 -6.30891740e-01
4.38242733e-01 -3.88242722e-01 1.39153555e-01 7.88632989e-01
-4.93222550e-02 -4.29321617e-01 4.49360818e-01 -2.13369906e-01
5.80337167e-01 5.48975706e-01 2.24788517e-01 -2.71645933e-02
-4.84358110e-02 2.87590355e-01 2.71571815e-01 5.42899072e-01
-5.97877145e-01 6.85612738e-01 -1.31595552e-01 -2.06563920e-01
-8.57524693e-01 -1.16633940e+00 8.48288015e-02 9.01491165e-01
-2.66802937e-01 -3.85553986e-01 -4.88836616e-01 -5.39199054e-01
-2.29366615e-01 7.31852174e-01 -5.23741543e-01 1.76614940e-01
-1.00148511e+00 -5.43663502e-01 1.06807101e+00 5.82405090e-01
4.55421329e-01 -1.44129539e+00 -9.73011315e-01 -1.10506229e-01
-5.01809001e-01 -1.05793059e+00 -6.29378676e-01 -4.54547495e-01
-1.16129160e+00 -8.40464711e-01 -1.12537670e+00 -1.06006336e+00
5.09997666e-01 3.08072090e-01 9.28861380e-01 3.38530466e-02
5.82840964e-02 8.35155547e-01 -3.49005491e-01 -8.62208128e-01
-5.10038435e-01 -2.25572050e-01 1.81029424e-01 -4.76338387e-01
8.28083992e-01 -2.99760848e-01 -5.16252518e-01 5.72944097e-02
-7.68459320e-01 7.73907974e-02 1.13087285e+00 7.94981003e-01
3.79382432e-01 -9.57684994e-01 3.94062102e-01 -1.04898885e-01
1.24606705e+00 3.22376877e-01 -1.71488598e-01 5.93018770e-01
-1.01119459e+00 -8.24653581e-02 -1.62112594e-01 -5.16911566e-01
-9.76836681e-01 -5.30030191e-01 5.19637354e-02 1.02509752e-01
-2.37463474e-01 6.09004557e-01 3.04957092e-01 -2.53312021e-01
6.90527499e-01 3.81409466e-01 2.54021496e-01 -3.87665957e-01
5.46358347e-01 5.80483198e-01 2.90087968e-01 -8.76416087e-01
8.03988814e-01 4.04100776e-01 -1.56954899e-01 -7.91294396e-01
-4.99012291e-01 -1.33447766e-01 -6.93841636e-01 -4.59375799e-01
6.87961936e-01 -6.13629878e-01 -2.11635754e-01 2.99059868e-01
-1.30423129e+00 -4.38087285e-01 -4.41326320e-01 1.19326127e+00
-8.06209028e-01 5.03787637e-01 -7.71303952e-01 -6.67681694e-01
-6.72662437e-01 -1.16468453e+00 1.17264640e+00 2.15076387e-01
-8.47762227e-01 -7.92594254e-01 1.77285239e-01 8.04006100e-01
6.72953248e-01 7.77229741e-02 6.23630285e-01 5.21680787e-02
-6.08530879e-01 5.08959591e-03 -1.36097983e-01 4.39549536e-01
4.95901197e-01 2.79537439e-01 -6.65740013e-01 2.02955648e-01
-3.35040480e-01 -5.75410843e-01 7.65053511e-01 5.66280186e-01
4.20406349e-02 -3.74323010e-01 1.23826675e-02 1.54308334e-01
1.07982457e+00 -3.47627401e-02 5.88907599e-01 3.31033707e-01
3.20830226e-01 7.96808243e-01 6.76889122e-01 -3.53352100e-01
6.45514369e-01 6.54858768e-01 -2.63473541e-01 7.55860731e-02
-9.33344066e-01 -3.23058039e-01 6.72304749e-01 1.11341417e+00
-8.48143876e-01 -1.68494537e-01 -8.10203195e-01 5.68154275e-01
-1.73021233e+00 -9.77203250e-01 -1.25755474e-01 2.04312873e+00
8.95500660e-01 -7.09071979e-02 2.88211316e-01 -2.09429950e-01
3.52688462e-01 9.23351720e-02 -1.47343472e-01 -6.65386736e-01
-5.06954014e-01 3.58182609e-01 5.08295000e-01 5.47644079e-01
-6.44820511e-01 1.47966492e+00 7.49701118e+00 2.67196476e-01
-1.32869935e+00 1.02613524e-01 -2.57794172e-01 -2.14019597e-01
-3.89036387e-01 3.67897227e-02 -7.33962476e-01 1.33186489e-01
5.88062704e-01 -2.53576070e-01 5.14669120e-01 2.85728186e-01
4.52510148e-01 -9.82578960e-04 -9.58504856e-01 9.39617395e-01
5.18102169e-01 -1.09561300e+00 4.25553828e-01 1.58772051e-01
6.48603439e-01 4.57215726e-01 3.51446301e-01 7.25667700e-02
3.29866737e-01 -8.15689325e-01 8.23201537e-01 4.90174472e-01
9.82556224e-01 4.14750315e-02 4.80733663e-01 -2.14931488e-01
-7.81182826e-01 3.32236439e-01 -1.32240012e-01 -1.80653349e-01
6.48146987e-01 -5.83742224e-02 -6.30280912e-01 3.00490677e-01
3.90102148e-01 4.40305173e-01 -2.94936925e-01 1.07046711e+00
-9.33942974e-01 1.08082247e+00 -5.71556747e-01 -4.74275678e-01
3.59354347e-01 -2.57804811e-01 5.59010029e-01 1.26690519e+00
3.22194010e-01 -4.67082225e-02 -2.46303722e-01 5.47940016e-01
1.28788993e-01 4.68972027e-01 -7.17695236e-01 -4.20201033e-01
2.00529732e-02 6.34877205e-01 -6.08532131e-01 -1.71294421e-01
-8.02912116e-01 1.02185392e+00 -6.42208159e-02 7.16694355e-01
-5.32490075e-01 1.61980867e-01 8.75443816e-01 -1.08450875e-01
-2.02566236e-01 -5.56103826e-01 -6.72747791e-01 -1.25823128e+00
4.14042175e-01 -1.12401402e+00 2.60736197e-01 -8.14855456e-01
-1.31266844e+00 1.64986908e-01 2.23347649e-01 -1.44267869e+00
-1.99287638e-01 -9.10270452e-01 -2.55331755e-01 7.22001314e-01
-1.59916735e+00 -1.78106642e+00 -3.79160978e-02 3.34778011e-01
5.79874218e-01 1.01179451e-01 9.42801416e-01 3.08269680e-01
-2.26092301e-02 8.20717871e-01 -3.50420535e-01 2.34733656e-01
1.11124694e+00 -9.60052371e-01 5.80030560e-01 8.90316963e-01
5.38857698e-01 1.07806706e+00 6.46155953e-01 -8.26377511e-01
-1.25024712e+00 -4.82468396e-01 1.64888597e+00 -7.22710073e-01
7.59773374e-01 1.35122359e-01 -3.02012146e-01 9.46622908e-01
4.91224021e-01 -6.35143220e-01 7.78164744e-01 -1.26959413e-01
-3.12314093e-01 1.78159133e-01 -9.17455077e-01 1.28462899e+00
1.54814196e+00 -5.41813135e-01 -9.53745008e-01 4.25066084e-01
4.63730365e-01 -4.65776026e-01 -6.85661376e-01 3.32988799e-01
1.26075602e+00 -3.78947258e-01 5.76243579e-01 -6.84068501e-01
2.96399593e-01 -2.50936598e-01 -2.12714940e-01 -1.09663713e+00
8.06750953e-02 -7.33462632e-01 6.08055554e-02 8.31218004e-01
4.49388981e-01 -9.11727190e-01 6.48564279e-01 8.49075556e-01
-3.45232964e-01 -7.26700127e-01 -9.97029901e-01 -8.53587568e-01
4.71450537e-01 -6.74601078e-01 2.93237209e-01 7.22536027e-01
3.55322249e-02 3.52038026e-01 -5.13816595e-01 -2.86532432e-01
5.06351054e-01 5.77595904e-02 9.11014020e-01 -1.03669107e+00
2.18304489e-02 -8.87107611e-01 -2.55576223e-01 -1.18308449e+00
-7.81733990e-02 -8.58974218e-01 -1.02892205e-01 -2.32969189e+00
-6.23631775e-02 7.65346512e-02 -1.22514643e-01 5.85488975e-01
1.30596787e-01 4.74398941e-01 4.73901153e-01 3.75674963e-01
-1.34540871e-01 4.49794888e-01 1.72239840e+00 -1.82870045e-01
-1.73053909e-02 -1.77878454e-01 -6.87322259e-01 6.03537917e-01
1.17687011e+00 -2.15349451e-01 -3.30594689e-01 -1.01813483e+00
1.40848622e-01 -6.31945908e-01 2.12941423e-01 -6.08623922e-01
1.14815354e-01 -5.64519167e-01 -2.04297423e-01 -3.81745487e-01
5.08856416e-01 -4.93596762e-01 -2.89881021e-01 4.70660090e-01
-3.95454466e-01 2.16164798e-01 1.24668464e-01 5.94445057e-02
-1.59544870e-01 -1.07395016e-01 3.82044137e-01 -2.70836353e-01
-7.21023679e-01 9.82674360e-02 -6.78656042e-01 1.09014809e-01
4.66749638e-01 -5.78927577e-01 -5.60250461e-01 -6.34623408e-01
-3.63634080e-01 1.52213231e-01 4.59745288e-01 7.65016496e-01
7.43035495e-01 -1.31108212e+00 -9.43009436e-01 -2.75954187e-01
2.94268578e-01 -5.20439804e-01 -3.51814628e-01 1.05593777e+00
-7.04641521e-01 8.34748685e-01 -4.05113280e-01 -5.76562807e-02
-1.38248634e+00 -8.29760358e-02 2.05968171e-01 4.50851955e-02
-6.91276550e-01 6.66468501e-01 -3.08840692e-01 -4.15763289e-01
2.99771219e-01 -8.80522072e-01 2.21071020e-02 -3.06480974e-01
3.99121165e-01 2.43367031e-01 -3.15585762e-01 -8.29050481e-01
-2.70375043e-01 8.98291051e-01 2.14298755e-01 -8.83305252e-01
9.73254621e-01 -8.28722715e-02 -1.73411444e-01 6.21129334e-01
4.71453875e-01 3.54172856e-01 -5.96074462e-01 -1.78802684e-01
1.94058865e-01 -3.22196871e-01 -4.25971627e-01 -1.30186832e+00
-6.92000866e-01 8.69303823e-01 4.90432501e-01 -7.39032924e-01
1.01756608e+00 -5.91649767e-03 1.01343322e+00 6.50050461e-01
5.95593154e-01 -1.50101626e+00 -1.38317525e-01 5.81910968e-01
1.39258444e+00 -1.14152706e+00 3.92565280e-02 -3.06120366e-01
-8.21076810e-01 1.20430267e+00 4.86903638e-01 1.14929497e-01
2.02086627e-01 6.63779229e-02 9.19017255e-01 -2.12628096e-01
-5.34865320e-01 -5.45460463e-01 4.60871696e-01 7.54047334e-01
7.36441195e-01 2.20588744e-01 -1.54469943e+00 -4.44476306e-02
-5.53189635e-01 3.47309798e-01 1.72762856e-01 1.02695596e+00
-3.43134522e-01 -1.63439763e+00 -4.23928499e-01 5.29024959e-01
-2.29221374e-01 -4.33566809e-01 -9.61045682e-01 1.01542950e+00
6.62273541e-02 9.47071552e-01 -4.77958441e-01 -1.89782068e-01
3.56668562e-01 5.88612556e-01 1.14873827e+00 -5.03681481e-01
-5.02607346e-01 -2.19138861e-02 5.61052918e-01 -3.47532898e-01
-9.39513087e-01 -1.01738977e+00 -1.12286580e+00 -2.35883147e-03
-1.14715703e-01 -3.35746586e-01 8.21111321e-01 1.09859908e+00
1.04904741e-01 -9.77952927e-02 -5.36386132e-01 -2.06623390e-01
-6.33173883e-01 -1.13585949e+00 -1.40098616e-01 1.96009338e-01
6.70939358e-03 -6.11806035e-01 -1.98009491e-01 2.19938159e-01]
|
[9.166788101196289, -6.493274211883545]
|
b0f28c59-5816-49ce-9108-34e6e4063d5d
|
probabilistic-relations-for-modelling
|
2303.09692
| null |
https://arxiv.org/abs/2303.09692v2
|
https://arxiv.org/pdf/2303.09692v2.pdf
|
Probabilistic relations for modelling epistemic and aleatoric uncertainty: semantics and automated reasoning with theorem proving
|
Probabilistic programming combines general computer programming, statistical inference, and formal semantics to help systems make decisions when facing uncertainty. Probabilistic programs are ubiquitous, including having a significant impact on machine intelligence. While many probabilistic algorithms have been used in practice in different domains, their automated verification based on formal semantics is still a relatively new research area. In the last two decades, it has attracted much interest. Many challenges, however, remain. The work presented in this paper, probabilistic relations, takes a step towards our vision to tackle these challenges. Our work is based on Hehner's predicative probabilistic programming, but there are several obstacles to the broader adoption of his work. Our contributions here include (1) the formalisation of its syntax and semantics by introducing an Iverson bracket notation to separate relations from arithmetic; (2) the formalisation of relations using Unifying Theories of Programming (UTP) and probabilities outside the brackets using summation over the topological space of the real numbers; (3) the constructive semantics for probabilistic loops using Kleene's fixed-point theorem; (4) the enrichment of its semantics from distributions to subdistributions and superdistributions to deal with the constructive semantics; (5) the unique fixed-point theorem to simplify the reasoning about probabilistic loops; and (6) the mechanisation of our theory in Isabelle/UTP, an implementation of UTP in Isabelle/HOL, for automated reasoning using theorem proving. We demonstrate our work with six examples, including problems in robot localisation, classification in machine learning, and the termination of probabilistic loops.
|
['Simon Foster', 'Jim Woodcock', 'Kangfeng Ye']
|
2023-03-16
| null | null | null | null |
['probabilistic-programming', 'automated-theorem-proving', 'automated-theorem-proving']
|
['methodology', 'miscellaneous', 'reasoning']
|
[-3.97726241e-03 3.67831439e-01 7.99019635e-02 -4.74562943e-01
-4.41562772e-01 -6.63165271e-01 1.00967908e+00 3.88011456e-01
-3.29137623e-01 6.53899670e-01 -1.57026753e-01 -8.87669623e-01
-6.67750180e-01 -1.02035224e+00 -7.57882714e-01 -6.88899577e-01
-7.09766984e-01 7.53190339e-01 7.40364254e-01 -2.73506671e-01
2.99981773e-01 4.43645507e-01 -1.76117945e+00 -1.23530880e-01
5.91860116e-01 8.59651148e-01 -1.44477174e-01 7.20750868e-01
-3.33537877e-01 1.09644461e+00 -3.52520227e-01 -3.01644474e-01
-7.24142864e-02 -9.81361791e-02 -1.12651074e+00 -8.16633344e-01
-3.46031398e-01 5.41563109e-02 1.37750655e-01 1.26052022e+00
-1.09301424e-02 -5.18551692e-02 6.85382247e-01 -2.11944008e+00
-2.32441783e-01 1.11158192e+00 -3.47022951e-01 -4.39386964e-01
5.40724874e-01 -1.77742720e-01 9.17943716e-01 -3.51145029e-01
3.54202688e-01 1.73635805e+00 7.05204189e-01 2.93984830e-01
-1.33986557e+00 -3.08896214e-01 -9.35855657e-02 5.95712736e-02
-1.64115608e+00 1.74482882e-01 3.20611864e-01 -6.23574495e-01
7.96809316e-01 4.48916525e-01 3.31801802e-01 5.02466261e-01
3.47731322e-01 7.26305187e-01 1.30666447e+00 -9.08634722e-01
5.69340765e-01 1.94873720e-01 4.34150189e-01 5.28213620e-01
2.35857189e-01 4.56447423e-01 -8.18980411e-02 -6.70256078e-01
4.87801790e-01 -2.90248692e-01 9.96354595e-02 -8.11885118e-01
-1.35895681e+00 1.19995880e+00 -7.24368915e-02 3.32181513e-01
3.57845128e-02 6.38964653e-01 5.96024036e-01 -7.45811313e-02
-7.98960328e-02 -1.97197441e-02 -5.95959127e-01 -3.32545072e-01
-4.17122036e-01 8.45237195e-01 1.46137428e+00 8.87817979e-01
6.05415940e-01 -3.92529875e-01 3.82888645e-01 4.41576570e-01
8.86045456e-01 8.05090129e-01 -1.63999647e-02 -1.05477262e+00
1.27731953e-02 2.16183305e-01 3.46378744e-01 -9.87569273e-01
-3.12097669e-01 2.95501411e-01 -3.36466521e-01 5.66714644e-01
4.82166260e-01 1.56682376e-02 -5.94695508e-01 1.76242924e+00
2.88605213e-01 -1.90835044e-01 2.59491712e-01 4.45544034e-01
2.13760048e-01 7.37942576e-01 2.25728571e-01 1.14519618e-01
1.28454590e+00 -1.77560776e-01 -4.54364330e-01 2.26575390e-01
7.29416490e-01 -4.35314685e-01 5.91707408e-01 5.52177250e-01
-6.99276209e-01 1.97574884e-01 -9.89487946e-01 7.06249997e-02
-5.93196511e-01 -4.97965604e-01 1.00125766e+00 1.14605427e+00
-1.14442146e+00 2.87517369e-01 -1.04866946e+00 -4.18684959e-01
-7.30272606e-02 5.90585709e-01 -2.42727950e-01 6.41607344e-02
-1.43456292e+00 1.19595456e+00 8.00516188e-01 -2.33386885e-02
-4.34844881e-01 -3.22427303e-01 -1.23744261e+00 -1.11500494e-01
5.65232396e-01 -4.33756292e-01 1.37603188e+00 -5.52909493e-01
-1.71803141e+00 6.46467149e-01 8.59881490e-02 -8.77165794e-01
3.94555539e-01 -8.18539858e-02 -1.64032280e-01 -2.75229305e-01
1.64162323e-01 4.09673572e-01 6.31376132e-02 -1.03303313e+00
-7.73721397e-01 -4.45891678e-01 4.73103017e-01 -1.72833264e-01
6.11370862e-01 3.93688351e-01 1.59002855e-01 -2.22346902e-01
2.61497349e-01 -1.04354930e+00 -3.97779584e-01 -4.60552275e-01
-3.47436786e-01 -4.93074894e-01 4.95080024e-01 5.19211814e-02
8.53476405e-01 -2.09358048e+00 1.34667113e-01 6.65468574e-01
-3.47803347e-02 -1.39026567e-01 4.39198315e-01 5.34540772e-01
-1.28818542e-01 3.46555173e-01 -5.22197604e-01 3.48937333e-01
7.97999144e-01 5.28683841e-01 -6.90936387e-01 6.62575781e-01
1.32046968e-01 4.37340200e-01 -1.08531773e+00 -5.88696122e-01
5.28418243e-01 2.38884017e-01 -3.64858568e-01 -2.59267092e-01
-5.17503321e-01 -3.55255067e-01 -3.93459916e-01 4.23605114e-01
7.03405738e-01 5.10736883e-01 4.69360828e-01 4.22250211e-01
-4.66204286e-01 4.60635930e-01 -1.80767560e+00 1.23098636e+00
-2.70801455e-01 1.29039794e-01 7.05064833e-02 -8.74185205e-01
7.42115974e-01 1.19458690e-01 2.72205174e-01 -9.22290534e-02
2.28824764e-01 2.76093364e-01 -3.30503136e-02 -9.90328416e-02
5.40176570e-01 -4.27654773e-01 -7.79403806e-01 5.93920767e-01
-1.24829784e-01 -8.36056888e-01 3.47632259e-01 2.63925791e-01
1.00170195e+00 8.10789287e-01 5.21260619e-01 -6.13997459e-01
5.73960483e-01 2.98446447e-01 5.57153046e-01 7.36068726e-01
-2.51839310e-01 1.52459607e-01 1.10705364e+00 -5.33073306e-01
-6.37579858e-01 -1.40746760e+00 -3.36356372e-01 1.03478050e+00
5.14047295e-02 -5.88831186e-01 -6.42434239e-01 -4.91228849e-01
3.53585370e-02 1.20249033e+00 -4.92140770e-01 9.77797210e-02
-2.24234536e-01 -6.91005349e-01 1.04204106e+00 3.73815268e-01
2.04190657e-01 -8.00097346e-01 -9.37655509e-01 1.10657565e-01
1.79570075e-02 -6.57698572e-01 2.82419294e-01 5.55581868e-01
-7.05235958e-01 -1.24022079e+00 -1.48881515e-02 -3.52474034e-01
2.35661566e-01 -2.31513724e-01 9.39967752e-01 -3.42397571e-01
1.00575499e-01 5.90028584e-01 -2.17781663e-01 -9.38802779e-01
-7.08274126e-01 -4.04467374e-01 3.29070725e-02 -7.94450819e-01
4.74588186e-01 -5.35752654e-01 3.03652555e-01 2.48869032e-01
-1.03226995e+00 -2.59042889e-01 2.01475814e-01 8.50275099e-01
2.11485490e-01 6.45088434e-01 -2.21250746e-02 -1.03215563e+00
4.63423640e-01 -4.64964271e-01 -1.19086182e+00 1.23875283e-01
-3.47386569e-01 4.74281251e-01 1.07730418e-01 9.95069817e-02
-8.72883856e-01 1.33553311e-01 1.95498653e-02 2.71772623e-01
-1.70743570e-01 7.43120074e-01 -4.05043691e-01 1.39429361e-01
7.29566753e-01 3.35292122e-03 1.39488459e-01 2.58001000e-01
7.82600462e-01 4.38787043e-01 5.88956177e-01 -1.40168047e+00
6.47673965e-01 5.21581531e-01 5.04616618e-01 -5.69012165e-01
-2.28606209e-01 -2.18868002e-01 -5.74590266e-01 9.67862383e-02
3.93698484e-01 -5.02968013e-01 -1.10922909e+00 4.72005248e-01
-1.20065677e+00 -3.40911865e-01 -2.86911488e-01 4.20023859e-01
-8.44244540e-01 4.87628192e-01 -4.24417913e-01 -1.47379172e+00
1.59837574e-01 -1.40926158e+00 9.27649617e-01 -1.01029247e-01
-5.64966261e-01 -1.02484632e+00 3.70509118e-01 -2.58670628e-01
5.13511226e-02 4.26271915e-01 1.12266541e+00 -7.51721144e-01
-3.78449738e-01 -2.17811540e-01 -1.68407276e-01 3.08867663e-01
-4.41277415e-01 6.45476341e-01 -7.54404366e-01 5.73669374e-01
3.73892263e-02 1.03219189e-01 1.91138476e-01 7.72070810e-02
5.10560334e-01 -2.47756347e-01 -4.82915461e-01 -1.07797563e-01
1.40572631e+00 1.88356370e-01 9.21267569e-01 7.01795518e-01
2.86928654e-01 1.01347148e+00 7.51619577e-01 2.84894139e-01
8.72208595e-01 5.47366619e-01 5.59247911e-01 6.92805767e-01
4.81685221e-01 -7.93475583e-02 5.15842676e-01 2.74249256e-01
-4.41242307e-02 4.94179636e-01 -1.44530177e+00 5.86761773e-01
-2.13757825e+00 -1.05060387e+00 -6.59572363e-01 2.50693178e+00
9.40937281e-01 9.86537412e-02 2.13785395e-01 3.86908203e-01
8.00997913e-01 -2.65796244e-01 3.36721659e-01 -7.92217731e-01
2.00585186e-01 3.55949640e-01 6.33011162e-01 8.40111017e-01
-1.13525844e+00 6.83769405e-01 6.13098764e+00 6.17463171e-01
-7.91027486e-01 1.57049581e-01 4.80823182e-02 6.10597432e-01
-4.94627535e-01 6.29098356e-01 -4.87415820e-01 4.89193469e-01
1.11847758e+00 -1.53236508e-01 3.99143517e-01 1.18555880e+00
-3.32828254e-01 -7.10145056e-01 -1.18365443e+00 5.75377584e-01
-2.27341413e-01 -8.45209360e-01 -4.94669139e-01 -3.54214497e-02
3.87760282e-01 6.62338659e-02 -1.73897967e-01 4.73804742e-01
1.28613055e+00 -1.06165230e+00 1.30725908e+00 3.62792104e-01
2.03271076e-01 -1.03151762e+00 1.11826050e+00 3.83802295e-01
-7.23920107e-01 5.26455231e-02 -6.27489388e-02 -4.07881320e-01
-4.05215137e-02 1.11014497e+00 -8.76952112e-01 7.97499657e-01
6.03376150e-01 -4.39411737e-02 4.42745984e-02 9.03763652e-01
-6.75547540e-01 4.24503952e-01 -8.35588336e-01 -3.66559327e-01
2.36726273e-02 -1.95336223e-01 5.20093203e-01 1.26506066e+00
9.34368595e-02 -1.62458852e-01 -5.38675906e-03 1.11745417e+00
7.13736773e-01 -3.42194051e-01 -5.78571200e-01 1.57129422e-01
3.63516480e-01 7.21987426e-01 -9.30934191e-01 -1.42688259e-01
-1.76053181e-01 1.03472419e-01 -1.83803707e-01 5.84897958e-03
-9.41944003e-01 -8.33735287e-01 4.32818115e-01 -1.50747716e-01
-3.94528061e-02 -5.38135529e-01 -6.35496259e-01 -6.04081988e-01
-5.04072905e-02 -6.23394132e-01 4.65514094e-01 -6.98321104e-01
-9.35858488e-01 2.77628750e-01 7.78369486e-01 -5.44240415e-01
-6.28571391e-01 -8.34729373e-01 -3.03809673e-01 1.17941737e+00
-9.80172276e-01 -1.01716757e+00 4.00526047e-01 1.26049265e-01
-4.74881202e-01 4.95681047e-01 1.20740998e+00 -1.51675090e-01
-5.04436009e-02 -4.32910956e-02 -1.42960399e-01 -8.28115791e-02
4.70964611e-01 -1.70263183e+00 2.51664191e-01 9.65889096e-01
-1.97697625e-01 1.07418525e+00 1.15480733e+00 -5.14741480e-01
-1.62916577e+00 -7.01375782e-01 1.12024152e+00 -6.72957718e-01
1.13967121e+00 -3.23328942e-01 -5.01176238e-01 9.61857975e-01
-1.95960417e-01 -6.04424551e-02 3.36300462e-01 4.74144787e-01
-8.02684426e-01 -7.01043829e-02 -1.49082613e+00 7.01094925e-01
3.68345499e-01 -5.70923686e-01 -1.05074978e+00 1.39885232e-01
5.38623333e-01 -4.81560737e-01 -8.91305983e-01 4.95419234e-01
8.12400281e-01 -1.15305281e+00 6.15519106e-01 1.25166431e-01
4.22712639e-02 -9.69839752e-01 -4.53756630e-01 -8.24216962e-01
1.88564345e-01 -6.26761556e-01 3.01546961e-01 1.42840791e+00
1.74005613e-01 -1.24801862e+00 3.16501081e-01 9.57183540e-01
1.06790103e-01 -4.75045204e-01 -1.01621938e+00 -5.88066041e-01
3.33943993e-01 -1.27246451e+00 6.85760498e-01 6.65933549e-01
6.31034553e-01 -8.81449878e-02 4.00779963e-01 3.03307116e-01
5.56432784e-01 -9.12636593e-02 8.51689398e-01 -1.38985312e+00
-2.03918979e-01 -5.97011685e-01 -1.01534104e+00 -3.57178867e-01
2.66080230e-01 -7.68524289e-01 6.60806715e-01 -1.40154839e+00
3.77044082e-02 -8.86676848e-01 2.22631037e-01 6.25992715e-01
4.13902342e-01 -2.73742706e-01 8.04613531e-02 -2.09414288e-02
-6.19800329e-01 3.03819656e-01 1.94721192e-01 3.14582605e-03
1.36561304e-01 1.87756583e-01 -6.28333151e-01 1.13587773e+00
6.41256988e-01 -7.11894810e-01 -2.92858005e-01 3.51713076e-02
8.41544747e-01 -5.82616813e-02 6.71182096e-01 -9.59235430e-01
3.00667882e-01 -3.77656251e-01 -3.55420411e-01 -4.90515560e-01
-3.28622945e-02 -9.14203525e-01 3.51454467e-01 5.85700035e-01
-7.59884715e-02 -6.64675608e-02 3.86193067e-01 3.95540863e-01
-1.28769532e-01 -7.48507202e-01 4.46024835e-01 7.28100073e-03
-7.87154198e-01 -4.59536403e-01 -6.50289774e-01 -8.20043311e-02
1.23091817e+00 1.41233400e-01 -1.71970800e-01 -6.73036128e-02
-6.44146323e-01 2.90578514e-01 7.64752388e-01 1.36764180e-02
1.47658780e-01 -8.90225530e-01 -2.94592142e-01 -2.64243275e-01
-4.35453691e-02 2.79696792e-01 -1.09026857e-01 8.24030161e-01
-7.59674847e-01 5.90102732e-01 4.88769189e-02 -6.80831790e-01
-9.53469694e-01 4.51342851e-01 5.63681349e-02 -1.50680512e-01
-1.78338110e-01 6.78814709e-01 4.55882326e-02 -1.02071595e+00
2.47524172e-01 -7.48684347e-01 2.30764046e-01 -4.58452463e-01
6.55288041e-01 4.52637583e-01 -1.24022178e-01 -5.31814396e-01
-7.29032040e-01 2.57593811e-01 3.35154831e-01 -7.48637676e-01
1.11884773e+00 -6.18836656e-02 -1.08345938e+00 1.02560186e+00
4.18363214e-01 2.15672836e-01 -6.01069212e-01 1.53815702e-01
5.27556360e-01 1.56585835e-02 -3.77299845e-01 -9.00526166e-01
9.07362613e-04 6.79438472e-01 2.89977193e-01 5.03120661e-01
6.12651348e-01 1.88678190e-01 1.15628809e-01 4.00691569e-01
1.16731465e+00 -7.22531497e-01 -9.53925371e-01 9.62806821e-01
7.04948366e-01 -6.92249715e-01 1.79870710e-01 -7.51782179e-01
-4.30793852e-01 1.30472243e+00 -1.07753552e-01 -1.11722343e-01
6.46858335e-01 7.09726572e-01 -2.35420704e-01 -6.32628947e-02
-4.53183532e-01 -2.06640631e-01 -4.05592263e-01 9.64407027e-01
2.80032188e-01 5.75433433e-01 -3.40590954e-01 8.70135307e-01
-4.94719356e-01 2.57821798e-01 7.67691195e-01 1.35199010e+00
-4.00650918e-01 -1.43186855e+00 -9.52628016e-01 -7.39927739e-02
-6.49511039e-01 -1.25444159e-01 -8.34641140e-03 1.22117865e+00
2.49927044e-01 9.75251555e-01 -3.01380217e-01 -2.12969691e-01
7.92564526e-02 1.97563291e-01 7.36795962e-01 -9.13231313e-01
-2.14363053e-01 -5.51376760e-01 1.89779177e-01 -4.29234296e-01
-5.02686024e-01 -9.36634123e-01 -1.62372434e+00 -4.80062366e-01
-2.38862887e-01 6.81478739e-01 1.22084284e+00 1.09736061e+00
-4.16722149e-02 3.47729743e-01 -8.99646711e-03 -7.13459194e-01
-8.95637631e-01 -6.68488204e-01 -9.37140584e-01 -5.50107479e-01
-1.19481340e-01 -7.85198987e-01 -4.82927114e-01 -1.00490347e-01]
|
[8.544563293457031, 6.615253448486328]
|
3ce99edb-205f-40ea-be98-c1658e8b1fe7
|
symmetry-detection-of-occluded-point-cloud
|
2003.06520
| null |
https://arxiv.org/abs/2003.06520v1
|
https://arxiv.org/pdf/2003.06520v1.pdf
|
Symmetry Detection of Occluded Point Cloud Using Deep Learning
|
Symmetry detection has been a classical problem in computer graphics, many of which using traditional geometric methods. In recent years, however, we have witnessed the arising deep learning changed the landscape of computer graphics. In this paper, we aim to solve the symmetry detection of the occluded point cloud in a deep-learning fashion. To the best of our knowledge, we are the first to utilize deep learning to tackle such a problem. In such a deep learning framework, double supervisions: points on the symmetry plane and normal vectors are employed to help us pinpoint the symmetry plane. We conducted experiments on the YCB- video dataset and demonstrate the efficacy of our method.
|
['Hongyan Jiang', 'Zhelun Wu', 'Siyun He']
|
2020-03-14
| null | null | null | null |
['symmetry-detection', 'occluded-3d-object-symmetry-detection']
|
['computer-vision', 'computer-vision']
|
[ 8.18227082e-02 -2.72021145e-02 1.77805975e-01 -2.89324701e-01
-3.03523451e-01 -1.39417350e-01 4.67838228e-01 -2.58086622e-01
-2.01060995e-01 1.96347445e-01 -1.67025864e-01 -4.42589611e-01
-7.13470206e-03 -6.98558629e-01 -9.51383948e-01 -3.91949594e-01
6.44865707e-02 3.02491933e-01 1.81938440e-01 -2.16734767e-01
6.31837249e-01 6.57905281e-01 -1.40966737e+00 1.18704788e-01
6.95911109e-01 1.09942949e+00 6.56638807e-03 1.41866133e-01
1.07330076e-01 3.03426653e-01 -2.48700902e-01 -2.51695573e-01
4.91825074e-01 3.83544937e-02 -7.01079488e-01 2.21024960e-01
7.50961781e-01 -5.30811965e-01 -3.21047544e-01 1.10872972e+00
2.81107098e-01 -2.09168568e-02 5.81172526e-01 -1.26407933e+00
-4.00658160e-01 2.53929403e-02 -1.23699081e+00 1.22746028e-01
3.35156024e-01 -1.65961653e-01 9.43589449e-01 -1.03808534e+00
4.86449957e-01 1.10441422e+00 7.55462646e-01 3.16679478e-01
-6.64576352e-01 -8.23350668e-01 2.39034742e-01 4.03307736e-01
-1.38425446e+00 -2.21581817e-01 1.22024059e+00 -4.51721728e-01
7.11475372e-01 -1.82617247e-01 6.17963552e-01 8.64558518e-01
-1.15349114e-01 9.05923665e-01 7.45168567e-01 -5.10105312e-01
-3.45574617e-02 -3.47613007e-01 -2.27446392e-01 9.96617496e-01
1.19050682e-01 -1.98049452e-02 -2.11757690e-01 7.36508220e-02
1.32889354e+00 4.33933407e-01 -4.77459967e-01 -6.98853910e-01
-9.72033143e-01 9.61999059e-01 7.71353722e-01 7.03448355e-02
-2.51723945e-01 1.83782041e-01 2.07687184e-01 -3.78077738e-02
4.95831192e-01 3.13381165e-01 -3.59561712e-01 3.56843951e-03
-7.23746955e-01 2.20138177e-01 5.87089002e-01 9.31356192e-01
6.66393578e-01 -9.60257370e-03 4.00873065e-01 5.60808361e-01
2.66798705e-01 1.75382778e-01 3.00738335e-01 -9.41034794e-01
5.49611211e-01 6.74222171e-01 -1.14928512e-02 -1.41583407e+00
-5.07089317e-01 -3.21598381e-01 -9.83755589e-01 4.91358012e-01
4.24141228e-01 -2.29761213e-01 -7.69735813e-01 1.34434867e+00
3.13183039e-01 5.72978377e-01 -2.49888048e-01 1.18857086e+00
5.43186963e-01 5.62274814e-01 -6.09481096e-01 2.81687140e-01
1.06960666e+00 -1.03728998e+00 -3.74648929e-01 1.41013041e-01
5.09646833e-01 -8.76736581e-01 9.94430780e-01 6.41171396e-01
-8.52660358e-01 -4.75620657e-01 -1.18976653e+00 -3.09211046e-01
-9.46749896e-02 1.25261277e-01 8.99732530e-01 2.94811517e-01
-8.08381498e-01 6.24728084e-01 -1.01003933e+00 -3.13620478e-01
6.63637519e-01 4.49274391e-01 -2.71574825e-01 -2.48052850e-02
-8.12262356e-01 5.04632652e-01 5.13426319e-04 3.50627869e-01
-3.93004566e-01 -4.88064885e-01 -8.38841438e-01 2.82493204e-01
6.44042850e-01 -9.19728696e-01 1.36490858e+00 -9.54477727e-01
-1.51645887e+00 9.73681092e-01 -1.49069577e-01 -3.69639695e-01
7.82173216e-01 -6.18300557e-01 3.23865153e-02 9.38644558e-02
8.34491849e-02 5.40117323e-01 8.71934175e-01 -1.30596519e+00
-7.28158236e-01 -6.74197793e-01 1.71008751e-01 7.88071826e-02
-4.49915305e-02 -7.30550885e-02 -5.71056366e-01 -4.69154954e-01
5.81243157e-01 -9.06881332e-01 -2.45024890e-01 3.76694918e-01
-5.33702433e-01 -4.93294954e-01 1.13159776e+00 -3.76617879e-01
7.70028234e-01 -2.08643532e+00 -2.73522455e-02 1.03194140e-01
3.93704325e-01 3.34322900e-01 2.15078592e-01 1.40481874e-01
-2.36533731e-01 1.48501918e-01 -1.48115322e-01 -3.72067660e-01
-6.39495701e-02 -9.88190621e-02 -5.32678246e-01 8.04702699e-01
2.71065891e-01 8.61848235e-01 -6.27241790e-01 -4.08747405e-01
3.13880742e-01 6.61841393e-01 -7.10192323e-01 2.05621019e-01
-9.52033550e-02 3.40903312e-01 -5.01812994e-01 6.40982985e-01
9.92142081e-01 -3.54589015e-01 -1.99179530e-01 -1.61134586e-01
-3.33622634e-01 -4.08248417e-02 -1.18654573e+00 1.77721250e+00
-3.77013952e-01 7.17169881e-01 -1.22965477e-01 -1.12518120e+00
8.10095608e-01 6.96282983e-02 6.11667156e-01 -6.78974748e-01
2.45400533e-01 1.36493027e-01 1.62661523e-01 -2.96949297e-01
3.12147141e-01 1.77812402e-03 3.03094864e-01 2.38836735e-01
-3.75753313e-01 -2.40874901e-01 -2.97286183e-01 -1.52635440e-01
6.54922247e-01 2.32225999e-01 2.78348535e-01 1.31316885e-01
5.04381061e-01 -1.44706964e-01 5.22719085e-01 5.35216570e-01
-1.74812198e-01 1.04019141e+00 6.05408609e-01 -9.20381188e-01
-1.11891663e+00 -6.82632208e-01 -2.38323852e-01 5.55818975e-01
2.53358185e-01 -3.71006548e-01 -6.47994220e-01 -4.27767128e-01
-3.81485512e-03 1.78578570e-01 -3.25762570e-01 1.74292982e-01
-1.05649602e+00 -2.20126152e-01 1.27401084e-01 6.05833292e-01
7.78820395e-01 -8.38170230e-01 -7.01863825e-01 2.16806744e-04
1.67527393e-01 -1.40381777e+00 -3.06848526e-01 -8.89326185e-02
-9.88545239e-01 -1.34536982e+00 -8.22606802e-01 -1.04653311e+00
6.79705918e-01 5.77959478e-01 7.74955571e-01 2.60332018e-01
-3.25896531e-01 6.75686868e-03 -3.54421586e-02 -4.28264171e-01
2.90690988e-01 1.12193964e-01 -1.97558299e-01 6.49610385e-02
2.30692357e-01 -5.83148956e-01 -9.38580453e-01 3.68191928e-01
-7.09002674e-01 3.21652621e-01 5.67454338e-01 5.45451641e-01
6.01915121e-01 2.79975921e-01 1.06518999e-01 -8.58584523e-01
1.74809590e-01 -3.15992266e-01 -9.34886992e-01 -1.02309465e-01
-2.59205580e-01 -2.11639777e-01 7.26011038e-01 -1.85770635e-02
-7.42925346e-01 2.59609789e-01 -2.83659577e-01 -7.95593798e-01
-2.96274871e-01 4.80627865e-01 -1.48186654e-01 -3.50611746e-01
2.15397179e-01 1.34749621e-01 -1.80951491e-01 -5.16723275e-01
1.50095552e-01 4.00724769e-01 6.14408135e-01 -4.97501105e-01
8.37357759e-01 9.77939010e-01 4.32697266e-01 -8.61211360e-01
-8.04881930e-01 -4.54508722e-01 -8.37725163e-01 -7.33151212e-02
8.11155319e-01 -8.13715219e-01 -9.92877245e-01 5.66709816e-01
-1.54065633e+00 -4.42411145e-03 2.67414510e-01 2.05696672e-01
-6.52274251e-01 7.21129596e-01 -3.11557233e-01 -5.85987091e-01
-3.29574972e-01 -1.16495466e+00 1.48529077e+00 2.51351714e-01
2.01274812e-01 -9.42948401e-01 1.54704815e-02 2.60635555e-01
1.04580574e-01 4.62736487e-01 6.85392559e-01 -2.76260316e-01
-8.81076813e-01 -2.53052622e-01 -5.83224416e-01 -4.30622660e-02
1.57333121e-01 2.10378602e-01 -9.77815330e-01 -1.61682501e-01
2.31525823e-01 -1.16579570e-01 8.80392551e-01 2.94097036e-01
1.74591827e+00 1.81469209e-02 -3.30926776e-01 1.21397662e+00
1.41163492e+00 6.85009360e-02 4.97737914e-01 6.96015358e-01
1.11298263e+00 3.13246071e-01 4.77571934e-01 4.46282983e-01
3.68546277e-01 8.39285374e-01 6.80468559e-01 -4.86692727e-01
1.54048324e-01 -3.14578831e-01 -3.15952718e-01 4.89622980e-01
-2.69974083e-01 -8.21437538e-02 -1.10083258e+00 -1.62628926e-02
-1.87029874e+00 -7.12197244e-01 -2.56886780e-01 1.89424586e+00
3.34425271e-01 2.22582459e-01 -1.81558162e-01 2.09072888e-01
7.95386255e-01 2.31613621e-01 -6.12082124e-01 -1.74213186e-01
3.55905481e-02 5.83182275e-02 2.29570553e-01 3.69174749e-01
-1.54387951e+00 1.08173645e+00 5.88306236e+00 4.18442637e-01
-1.53271687e+00 -2.89264709e-01 5.35153329e-01 2.61824310e-01
5.89476302e-02 7.76565671e-02 -6.96535230e-01 3.37556243e-01
4.25405875e-02 -1.93109326e-02 1.97640836e-01 1.10985303e+00
1.56098992e-01 1.58320785e-01 -1.18430805e+00 1.41098344e+00
2.92491704e-01 -1.25251245e+00 6.76897839e-02 2.75208235e-01
8.25566590e-01 -8.33276543e-04 2.89805412e-01 1.43540040e-01
-1.92247480e-01 -9.34662282e-01 4.32176441e-01 2.26625130e-01
6.04001522e-01 -7.51098931e-01 6.56942964e-01 4.06941086e-01
-1.09286845e+00 5.23404777e-02 -5.12039781e-01 -2.79005200e-01
2.79620409e-01 5.13849318e-01 -7.63222694e-01 5.61807275e-01
6.28508568e-01 9.82712686e-01 -3.67466867e-01 1.30462396e+00
-3.54216665e-01 1.04368120e-01 -3.85383606e-01 3.33084375e-01
4.44507033e-01 -5.04588604e-01 3.10460031e-01 6.76017225e-01
3.07382107e-01 1.51337638e-01 2.48472050e-01 8.99011493e-01
-2.29745477e-01 4.73574027e-02 -8.58516335e-01 4.07851487e-01
-8.44401494e-02 1.13809752e+00 -9.01860595e-01 -1.37941033e-01
-6.55893445e-01 1.01653600e+00 5.14654040e-01 2.51422316e-01
-8.49715114e-01 -2.93592572e-01 5.67845523e-01 3.02246660e-01
5.67782998e-01 -5.49807787e-01 -4.56782997e-01 -1.45398867e+00
1.73074633e-01 -5.85867286e-01 1.84922352e-01 -8.94911468e-01
-1.17326391e+00 4.32037920e-01 -2.82158703e-01 -1.37106621e+00
1.05774580e-02 -9.11683977e-01 -9.62520301e-01 6.38910949e-01
-1.75966072e+00 -9.23341751e-01 -4.50683504e-01 5.93167663e-01
5.80130756e-01 -7.56825432e-02 5.53680241e-01 2.39874616e-01
-6.13884926e-01 3.53298724e-01 2.02445164e-01 3.99215460e-01
4.26040679e-01 -1.12443709e+00 6.46284223e-01 7.74723709e-01
2.54548162e-01 7.61807561e-01 4.26471502e-01 -4.82894778e-01
-1.52114415e+00 -8.22833419e-01 5.16122818e-01 -1.36597574e-01
6.43644989e-01 -4.22782451e-01 -9.78040755e-01 7.93433309e-01
6.21927008e-02 1.81552261e-01 2.98529446e-01 1.24499947e-01
-3.63985151e-01 3.60621847e-02 -7.20172167e-01 7.39030659e-01
1.10985231e+00 -4.65044796e-01 -5.23761749e-01 4.77791548e-01
6.32343411e-01 -8.58847558e-01 -4.79242414e-01 6.64006233e-01
5.33082664e-01 -1.15457821e+00 1.03854334e+00 -5.51104844e-01
7.11193264e-01 -2.91895688e-01 4.75402325e-02 -1.03478134e+00
-1.39802247e-01 -5.86389661e-01 1.02323163e-02 7.95090079e-01
-7.44028464e-02 -5.29262543e-01 1.27975738e+00 2.87936836e-01
-3.35726202e-01 -1.13127398e+00 -9.90704298e-01 -4.61531788e-01
1.06832534e-01 -3.58760446e-01 6.68898284e-01 9.19550538e-01
-3.10849428e-01 1.62028208e-01 -2.80501932e-01 4.37842816e-01
5.69515944e-01 6.95899665e-01 1.08569825e+00 -1.23704433e+00
-6.72547594e-02 -7.93200791e-01 -6.23891830e-01 -1.55166817e+00
2.79773444e-01 -5.62250912e-01 -2.33171448e-01 -1.36263371e+00
6.52482212e-02 -3.14779729e-01 -5.40755056e-02 3.00022960e-01
-1.39553934e-01 1.85894683e-01 1.11263424e-01 2.22781241e-01
-6.69438064e-01 7.74071574e-01 1.32631373e+00 -3.71002518e-02
-4.25471440e-02 2.97364831e-01 -6.98187590e-01 1.15951157e+00
9.40623939e-01 -2.76938766e-01 -2.76098788e-01 -8.97678792e-01
2.62242943e-01 -1.47808313e-01 4.05459464e-01 -9.62528646e-01
5.11997223e-01 1.26981391e-02 6.26329601e-01 -9.80650485e-01
3.33318710e-01 -9.68967319e-01 -3.95331264e-01 3.81992131e-01
1.18098982e-01 1.97241053e-01 8.84113014e-02 4.82875913e-01
-3.70708555e-01 -1.03003338e-01 5.98490536e-01 -1.03458047e-01
-7.14990854e-01 7.64120698e-01 1.11615762e-01 -4.06245664e-02
9.62260842e-01 -3.82030338e-01 -2.33841717e-01 -2.61588812e-01
-2.80187905e-01 2.42429063e-01 7.40281880e-01 3.49187016e-01
9.92108583e-01 -1.21059585e+00 -4.38585997e-01 4.21541423e-01
2.40221694e-02 6.82091713e-01 1.71415970e-01 6.23151302e-01
-9.73179996e-01 5.74602246e-01 -2.26946607e-01 -8.96333098e-01
-1.00524509e+00 7.49748826e-01 4.60188836e-01 1.86151147e-01
-1.04633939e+00 5.91632843e-01 5.80840349e-01 -2.56886482e-01
3.95736456e-01 -3.77287775e-01 -3.75154644e-01 -3.18470091e-01
3.53482783e-01 2.81572342e-01 1.37042776e-01 -6.00827336e-01
-2.51137644e-01 1.17508626e+00 -2.22059458e-01 1.63743228e-01
1.51710474e+00 1.17223851e-01 -4.86866310e-02 2.23949894e-01
1.44335854e+00 -1.08155310e-01 -1.38645017e+00 -2.43615165e-01
4.20132056e-02 -8.07033539e-01 -3.92498523e-02 -2.36983504e-02
-1.37292755e+00 1.24564862e+00 5.19022524e-01 1.56105503e-01
9.06849742e-01 -2.92713165e-01 1.01163328e+00 4.40065593e-01
4.06683475e-01 -7.81178653e-01 1.65095627e-01 5.99929035e-01
8.49473894e-01 -1.65040147e+00 8.32206756e-02 -6.72439933e-01
-3.19399208e-01 1.45121562e+00 9.24560308e-01 -4.74373609e-01
8.77956629e-01 3.10291033e-02 9.99759287e-02 -4.12679732e-01
-4.39530909e-01 -1.67331547e-01 1.41523302e-01 4.42424834e-01
3.74928206e-01 -2.15966120e-01 4.51617595e-03 2.94378489e-01
-3.94693643e-01 -2.13344004e-02 5.57915986e-01 9.86937225e-01
-4.21022356e-01 -9.66449499e-01 -3.87267351e-01 5.17168529e-02
-4.52666104e-01 8.26822519e-02 -2.44890139e-01 9.23101306e-01
-1.94203050e-03 5.12192547e-01 2.97378778e-01 -2.26451159e-02
5.39405346e-01 -3.68133843e-01 4.98928666e-01 -6.26791418e-01
-1.90871730e-01 1.21835753e-01 -4.66475844e-01 -5.85123956e-01
-2.45738551e-01 -5.55925846e-01 -1.43670475e+00 -2.76455820e-01
-2.26228386e-01 -1.05060302e-01 8.02932143e-01 7.49151409e-01
3.40276539e-01 2.23783687e-01 9.16325510e-01 -1.20717394e+00
-5.14681458e-01 -6.36453629e-01 -5.43691516e-01 4.29548264e-01
3.82583380e-01 -8.38088870e-01 -4.04161036e-01 -2.47428522e-01]
|
[8.255020141601562, -2.4914276599884033]
|
ca547966-37d5-4d85-aadb-49dd146ee073
|
self-supervision-can-be-a-good-few-shot
|
2207.09176
| null |
https://arxiv.org/abs/2207.09176v1
|
https://arxiv.org/pdf/2207.09176v1.pdf
|
Self-Supervision Can Be a Good Few-Shot Learner
|
Existing few-shot learning (FSL) methods rely on training with a large labeled dataset, which prevents them from leveraging abundant unlabeled data. From an information-theoretic perspective, we propose an effective unsupervised FSL method, learning representations with self-supervision. Following the InfoMax principle, our method learns comprehensive representations by capturing the intrinsic structure of the data. Specifically, we maximize the mutual information (MI) of instances and their representations with a low-bias MI estimator to perform self-supervised pre-training. Rather than supervised pre-training focusing on the discriminable features of the seen classes, our self-supervised model has less bias toward the seen classes, resulting in better generalization for unseen classes. We explain that supervised pre-training and self-supervised pre-training are actually maximizing different MI objectives. Extensive experiments are further conducted to analyze their FSL performance with various training settings. Surprisingly, the results show that self-supervised pre-training can outperform supervised pre-training under the appropriate conditions. Compared with state-of-the-art FSL methods, our approach achieves comparable performance on widely used FSL benchmarks without any labels of the base classes.
|
['Xinmei Tian', 'Yajing Liu', 'Jianzhuang Liu', 'Liangjian Wen', 'Yuning Lu']
|
2022-07-19
| null | null | null | null |
['cross-domain-few-shot-learning', 'few-shot-image-classification', 'unsupervised-few-shot-learning', 'unsupervised-few-shot-image-classification']
|
['computer-vision', 'computer-vision', 'computer-vision', 'computer-vision']
|
[ 4.44294184e-01 3.67954522e-01 -7.35569119e-01 -7.93761671e-01
-6.89561963e-01 -2.13997573e-01 6.26802504e-01 1.92616671e-01
-3.57882291e-01 8.19432735e-01 2.27840289e-01 1.56791881e-01
-8.37915391e-02 -9.29210722e-01 -6.22512043e-01 -6.71215236e-01
1.01277977e-02 4.77309138e-01 1.13102607e-01 -2.60506757e-03
2.10744655e-03 7.59840831e-02 -1.78705502e+00 2.97973037e-01
9.14597094e-01 1.09959543e+00 1.36917248e-01 1.63524657e-01
-2.51893699e-01 1.04353750e+00 -3.98374170e-01 -1.44169003e-01
1.81552857e-01 -6.59051061e-01 -7.66606748e-01 4.21134800e-01
1.97454229e-01 -1.41007572e-01 -4.69355106e-01 1.07617915e+00
1.82662830e-01 5.06544054e-01 9.85619068e-01 -1.13409770e+00
-7.00370848e-01 6.39420152e-01 -2.98135310e-01 8.41575339e-02
1.17795587e-01 7.90092871e-02 1.36926436e+00 -9.75642443e-01
4.93197858e-01 1.07693386e+00 4.64394718e-01 6.29400849e-01
-1.24886954e+00 -5.79871833e-01 1.39400840e-01 3.60268414e-01
-1.35623693e+00 -5.22360563e-01 8.60016763e-01 -3.03474873e-01
6.48507953e-01 -6.10380918e-02 2.86484092e-01 1.04003286e+00
-3.14588904e-01 1.11607182e+00 1.03816652e+00 -6.73233986e-01
6.11667156e-01 4.35731858e-01 5.44601440e-01 6.40711963e-01
3.74252081e-01 3.33536386e-01 -5.99002302e-01 -6.42112941e-02
3.79686624e-01 4.50298071e-01 -7.05408975e-02 -6.61136091e-01
-9.39628839e-01 9.89980519e-01 4.75043595e-01 3.76508564e-01
-2.01102048e-01 -1.86361611e-01 3.62881362e-01 3.27932090e-01
5.66145658e-01 3.45573723e-01 -3.44409913e-01 2.83733726e-01
-9.51478481e-01 -3.84733081e-01 6.85974419e-01 1.10501516e+00
1.36234963e+00 2.76991129e-02 -3.09170604e-01 9.32724237e-01
2.39633814e-01 3.53789836e-01 8.19595754e-01 -6.53926790e-01
3.30767691e-01 7.84458220e-01 -2.18422830e-01 -5.06143630e-01
-1.54648572e-01 -5.74107945e-01 -7.51237392e-01 -1.35339051e-01
2.05682173e-01 -2.79079914e-01 -9.95588243e-01 1.79021120e+00
1.82605051e-02 3.00618798e-01 4.05955672e-01 5.41456521e-01
7.06948340e-01 6.09459043e-01 1.16413444e-01 -5.19050419e-01
8.16354930e-01 -1.00666714e+00 -7.51316547e-01 -4.26147461e-01
7.82126725e-01 -2.30620533e-01 9.50068831e-01 5.97036704e-02
-5.44070244e-01 -6.64491534e-01 -1.26731145e+00 3.13721746e-01
-4.38551307e-01 1.44012034e-01 7.72273242e-01 5.74301004e-01
-5.05389929e-01 8.32551479e-01 -7.15870500e-01 -3.21373880e-01
7.96070099e-01 1.27626687e-01 -3.64486754e-01 -3.68290544e-01
-1.20389080e+00 7.04810679e-01 7.30860353e-01 -3.65361094e-01
-1.15840280e+00 -5.74648857e-01 -1.15919900e+00 3.00928235e-01
7.49353707e-01 -6.47049993e-02 1.09049523e+00 -1.07692230e+00
-1.31180727e+00 8.30753624e-01 -2.22056925e-01 -5.98148406e-01
1.26615494e-01 -2.03212500e-01 -4.50679660e-01 9.26186144e-02
8.07433501e-02 5.68560541e-01 8.55502784e-01 -1.31204367e+00
-4.24648881e-01 -2.51403660e-01 -5.66860624e-02 1.32351011e-01
-7.20612347e-01 -4.32779640e-01 -2.53376991e-01 -4.95940268e-01
5.54464795e-02 -6.89947069e-01 -3.45543236e-01 -1.38912842e-01
-4.68400896e-01 -3.17913443e-01 8.28676581e-01 2.01171771e-01
1.18483412e+00 -2.38984132e+00 -3.53559703e-01 1.78888187e-01
2.05266953e-01 4.81278121e-01 -1.88872129e-01 4.60495770e-01
-3.47159915e-02 -1.66612968e-01 -3.62369210e-01 -2.81023264e-01
-4.22351211e-02 5.09765863e-01 -2.16819867e-01 4.60015148e-01
4.47370559e-01 7.73633659e-01 -1.33759499e+00 -6.50264084e-01
3.37175995e-01 1.34675190e-01 -3.78512055e-01 4.88506705e-01
-2.38425791e-01 2.18497351e-01 -4.18083072e-01 5.76102793e-01
3.84771287e-01 -6.06620789e-01 2.55434334e-01 5.45077911e-03
3.66154820e-01 1.88046053e-01 -1.10842764e+00 1.77210712e+00
-4.44291294e-01 6.01536632e-01 -5.62244713e-01 -1.50304604e+00
1.17535508e+00 3.00085604e-01 5.02764106e-01 -4.37461197e-01
1.68846324e-01 5.45503795e-02 -1.10158868e-01 -2.96514988e-01
-1.29153997e-01 -3.34997654e-01 -2.27942895e-02 5.47208250e-01
7.95310855e-01 1.78435370e-01 3.71288031e-01 3.45847666e-01
9.50851381e-01 -1.19467089e-02 6.80591166e-01 -2.02349767e-01
3.41925442e-01 -2.37557918e-01 7.99506605e-01 8.82181466e-01
-3.07110608e-01 5.10335863e-01 1.96012050e-01 -4.12174255e-01
-5.80281854e-01 -1.17099452e+00 -2.17629001e-01 1.19069421e+00
2.33482301e-01 -5.88566065e-01 -5.41930079e-01 -1.20703924e+00
-3.39987651e-02 9.82861817e-01 -7.53668010e-01 -4.78790492e-01
1.44817740e-01 -7.35780954e-01 1.95305511e-01 5.62134385e-01
4.73057568e-01 -1.11256015e+00 -5.09411752e-01 1.37459487e-01
1.86980754e-01 -9.55685854e-01 -2.25037128e-01 6.95987284e-01
-9.41285193e-01 -1.15988266e+00 -5.23378193e-01 -7.86939740e-01
1.00723505e+00 4.26921010e-01 1.04722643e+00 -9.35601443e-02
-3.98988128e-01 2.49590546e-01 -4.69816446e-01 -3.88024807e-01
-2.20117435e-01 -2.70073619e-02 1.18036203e-01 2.55335867e-01
9.09047663e-01 -5.77770054e-01 -3.71165067e-01 3.22448254e-01
-9.00087535e-01 -1.17549106e-01 6.64903879e-01 1.11609674e+00
6.51930988e-01 2.45209098e-01 8.24079990e-01 -1.47941780e+00
1.92611560e-01 -8.04463267e-01 -2.47570172e-01 5.43561518e-01
-9.31053400e-01 3.65676492e-01 7.12468147e-01 -5.22243202e-01
-1.19866502e+00 2.55007446e-01 2.35518306e-01 -6.04001820e-01
-4.22738105e-01 3.59431475e-01 -3.39046091e-01 2.55538970e-01
9.13915753e-01 3.38134795e-01 -1.05010971e-01 -4.69065398e-01
3.93089324e-01 7.45169044e-01 3.65265995e-01 -4.90050048e-01
7.74307013e-01 5.46227813e-01 -1.63650170e-01 -8.01097810e-01
-1.65990102e+00 -6.33148909e-01 -8.91447783e-01 -1.05926320e-01
4.26391512e-01 -8.45603049e-01 -1.73754483e-01 9.02562514e-02
-4.88915592e-01 -1.22227594e-01 -8.24646771e-01 6.53391421e-01
-6.12890840e-01 2.56285906e-01 -3.40111792e-01 -9.62431550e-01
-1.01425588e-01 -7.83996761e-01 7.48312712e-01 4.24457133e-01
-8.74094740e-02 -1.14774942e+00 2.63370484e-01 1.46931157e-01
2.94828743e-01 -4.65007462e-02 6.66925788e-01 -1.41126060e+00
-1.43449128e-01 -4.39601034e-01 -8.34435821e-02 6.81714058e-01
6.08824253e-01 -3.99570763e-01 -1.33764780e+00 -3.43553513e-01
-3.11501995e-02 -9.20573771e-01 1.16118634e+00 1.11252248e-01
1.19799924e+00 -2.05123678e-01 -3.10710967e-01 4.90188688e-01
1.57427776e+00 1.18326917e-01 3.40111494e-01 -5.94686456e-02
5.25816321e-01 6.15431309e-01 8.02640855e-01 6.16247237e-01
6.93232194e-02 1.28182963e-01 3.18578362e-01 -4.48947102e-02
7.34375939e-02 -6.36159539e-01 2.85753608e-01 6.58531249e-01
1.98940918e-01 -7.12762848e-02 -6.25193477e-01 4.50372785e-01
-1.92626965e+00 -1.00276268e+00 4.64973867e-01 2.47959781e+00
8.96102786e-01 3.01906854e-01 1.02502818e-03 2.63637900e-01
8.39097619e-01 2.73809940e-01 -9.21667993e-01 7.23410174e-02
-1.08012445e-01 1.99553803e-01 4.99986559e-01 2.11991236e-01
-1.34558058e+00 9.55464780e-01 6.68103981e+00 1.09967721e+00
-8.72371197e-01 9.40252468e-02 6.59808397e-01 -1.25023454e-01
-2.61992216e-01 7.26805404e-02 -9.19397235e-01 4.27539140e-01
9.18667555e-01 -3.61274451e-01 2.57980943e-01 1.10597205e+00
-2.55992413e-01 2.97167953e-02 -1.37060177e+00 9.27487373e-01
2.14571968e-01 -1.26835799e+00 1.57347471e-02 -4.45219837e-02
9.33576524e-01 1.20385513e-01 -2.43376717e-01 7.39652932e-01
5.50413668e-01 -8.74308288e-01 2.08372504e-01 4.89496082e-01
7.08868682e-01 -7.21978307e-01 7.55137086e-01 6.91816807e-01
-1.06681299e+00 -2.24839225e-01 -7.03355968e-01 -1.24159150e-01
-1.77969813e-01 8.23000908e-01 -6.75172329e-01 4.70235229e-01
2.81112343e-01 1.25262737e+00 -5.22361577e-01 7.89510429e-01
-5.56614697e-01 8.77612591e-01 -1.16044529e-01 -1.23543583e-01
2.57909209e-01 -1.68081179e-01 1.25002295e-01 1.16218150e+00
-8.32014382e-02 2.51149386e-02 5.48972785e-01 8.05890441e-01
-2.28343382e-01 1.51907355e-01 -8.60557675e-01 -3.65221709e-01
5.31621337e-01 1.00171947e+00 -5.17163038e-01 -7.69832790e-01
-7.14489460e-01 7.57585764e-01 5.88166654e-01 3.19798112e-01
-3.26352656e-01 -4.72756535e-01 3.40787023e-01 -7.91879743e-02
2.68157154e-01 2.03197122e-01 -7.45480582e-02 -1.28675508e+00
-2.56824225e-01 -5.17437637e-01 7.04983234e-01 -3.51108432e-01
-1.73878562e+00 5.04504681e-01 5.43032102e-02 -1.51205242e+00
-4.57079440e-01 -5.15592456e-01 -7.29979157e-01 3.94727290e-01
-1.65796626e+00 -8.39641869e-01 -1.27625465e-01 5.92361629e-01
7.04492152e-01 -4.81787324e-01 1.01362896e+00 -2.58295257e-02
-6.66314662e-01 6.80684209e-01 4.47600335e-01 3.17885101e-01
7.33626127e-01 -1.26679754e+00 -9.21688974e-02 6.49740458e-01
5.74342966e-01 6.45297289e-01 4.36196864e-01 -4.85935956e-01
-1.11432147e+00 -1.08148932e+00 6.61026120e-01 -3.32052028e-03
6.12551630e-01 -2.58800685e-01 -1.03039074e+00 6.67336941e-01
1.60011388e-02 4.33532447e-01 1.24649751e+00 2.76558638e-01
-5.20601332e-01 -1.71080664e-01 -1.11499202e+00 1.60473421e-01
1.04286444e+00 -6.97337568e-01 -8.61689746e-01 4.98711765e-01
5.59218347e-01 2.45858029e-01 -6.49830163e-01 4.14490253e-01
1.64946631e-01 -8.82834077e-01 8.17940295e-01 -9.39611673e-01
4.72651571e-01 -2.64900941e-02 -2.17101589e-01 -1.51253569e+00
-3.24437588e-01 -2.70682037e-01 -3.24498385e-01 1.23593903e+00
4.64998394e-01 -4.61741954e-01 8.76055479e-01 4.64977890e-01
7.40792528e-02 -7.45538116e-01 -5.15838027e-01 -1.11262953e+00
-1.44031852e-01 -3.36290449e-01 3.44454825e-01 1.07654071e+00
4.26783681e-01 5.34717619e-01 -5.85537553e-01 -9.24886912e-02
9.56279159e-01 3.74764323e-01 5.35092413e-01 -1.42516303e+00
-3.52650315e-01 4.83282767e-02 -5.12301505e-01 -8.12681615e-01
5.24937809e-01 -1.13857996e+00 3.09736937e-01 -1.24842787e+00
5.66478550e-01 -2.78583318e-01 -8.38365197e-01 8.21077406e-01
-4.44991589e-01 1.90112591e-02 1.37205496e-01 3.47158581e-01
-1.09913993e+00 7.42521405e-01 1.07021749e+00 -2.92412490e-01
-1.40013680e-01 9.16911960e-02 -6.41332090e-01 7.73320138e-01
7.29314446e-01 -6.13072097e-01 -8.09472680e-01 -1.49416300e-02
-3.03953230e-01 -2.08111733e-01 -1.79921284e-01 -1.05995476e+00
1.82370976e-01 -2.68883139e-01 4.64181304e-01 -3.79455626e-01
2.80659050e-01 -7.12289870e-01 -3.90894443e-01 4.06817168e-01
-8.65063787e-01 -9.24241126e-01 -2.72301793e-01 7.53685236e-01
-4.01029557e-01 -7.10867345e-01 9.21608686e-01 -2.85747290e-01
-8.54245186e-01 4.46182191e-01 -3.08632106e-01 3.57752591e-01
1.12908185e+00 -6.31494969e-02 -1.54874563e-01 -4.68902886e-01
-8.53499353e-01 3.19789499e-01 4.17128652e-01 2.46906087e-01
6.56643152e-01 -1.37381935e+00 -5.00496924e-01 4.75636244e-01
6.44338250e-01 -7.86711723e-02 4.55937944e-02 2.22934589e-01
2.35125646e-01 3.59310061e-01 -2.08925605e-01 -4.67108011e-01
-8.08637738e-01 8.52515280e-01 -4.24583256e-02 -3.15491170e-01
-4.19045031e-01 6.14457011e-01 3.15757155e-01 -3.90177071e-01
4.64822143e-01 1.82226479e-01 -2.20920593e-01 1.38794109e-01
7.87969530e-01 3.12653333e-01 -9.24544707e-02 -3.51521701e-01
-1.90105006e-01 9.97047424e-02 -3.05845380e-01 1.76270023e-01
1.41841602e+00 4.16532122e-02 5.22659361e-01 8.34627330e-01
1.44352674e+00 -3.77113014e-01 -1.30444312e+00 -9.45523679e-01
2.19387233e-01 -5.26279092e-01 2.06703827e-01 -4.60957408e-01
-1.18251348e+00 1.05400503e+00 4.57351178e-01 4.24136035e-03
9.39738572e-01 1.49274766e-01 4.26418900e-01 7.76834190e-01
5.66240609e-01 -1.28384662e+00 2.96354026e-01 4.24326360e-01
3.14231217e-01 -1.65185761e+00 -6.92209555e-03 -4.31094080e-01
-8.52383375e-01 9.86075461e-01 7.12870181e-01 -2.91243285e-01
8.27114880e-01 -6.25888770e-03 -1.26978263e-01 -1.10205628e-01
-9.05496359e-01 -6.45370245e-01 4.13566351e-01 5.55762053e-01
3.51066977e-01 8.72238129e-02 5.24879061e-02 6.58667445e-01
2.42399991e-01 1.12370685e-01 1.68383732e-01 1.14051223e+00
-7.46053934e-01 -9.47427034e-01 7.20813721e-02 8.45884144e-01
-1.44345745e-01 -4.14902950e-03 -4.32820916e-01 5.01056552e-01
-8.32127482e-02 1.06558108e+00 3.42756361e-02 -4.44036335e-01
3.39509617e-03 3.20806146e-01 2.01333746e-01 -1.17848873e+00
-1.21092163e-01 -7.88188577e-02 -1.75183058e-01 -4.40619320e-01
-4.62515563e-01 -3.86549622e-01 -1.25079656e+00 1.74111664e-01
-6.77030861e-01 2.74873853e-01 2.02481613e-01 1.29618812e+00
1.05375126e-01 3.46857339e-01 1.12140417e+00 -7.14911103e-01
-8.80879283e-01 -1.05008888e+00 -8.88363123e-01 7.36888289e-01
1.38783306e-01 -9.06695485e-01 -7.19200194e-01 -1.58716515e-02]
|
[9.956490516662598, 3.021212577819824]
|
ed675361-283f-4d9b-b55f-44f77a960153
|
asking-clarifying-questions-in-open-domain
|
1907.06554
| null |
https://arxiv.org/abs/1907.06554v1
|
https://arxiv.org/pdf/1907.06554v1.pdf
|
Asking Clarifying Questions in Open-Domain Information-Seeking Conversations
|
Users often fail to formulate their complex information needs in a single query. As a consequence, they may need to scan multiple result pages or reformulate their queries, which may be a frustrating experience. Alternatively, systems can improve user satisfaction by proactively asking questions of the users to clarify their information needs. Asking clarifying questions is especially important in conversational systems since they can only return a limited number of (often only one) result(s). In this paper, we formulate the task of asking clarifying questions in open-domain information-seeking conversational systems. To this end, we propose an offline evaluation methodology for the task and collect a dataset, called Qulac, through crowdsourcing. Our dataset is built on top of the TREC Web Track 2009-2012 data and consists of over 10K question-answer pairs for 198 TREC topics with 762 facets. Our experiments on an oracle model demonstrate that asking only one good question leads to over 170% retrieval performance improvement in terms of P@1, which clearly demonstrates the potential impact of the task. We further propose a retrieval framework consisting of three components: question retrieval, question selection, and document retrieval. In particular, our question selection model takes into account the original query and previous question-answer interactions while selecting the next question. Our model significantly outperforms competitive baselines. To foster research in this area, we have made Qulac publicly available.
|
['Fabio Crestani', 'Mohammad Aliannejadi', 'W. Bruce Croft', 'Hamed Zamani']
|
2019-07-15
| null | null | null | null |
['question-selection']
|
['natural-language-processing']
|
[-4.06910181e-02 1.38864145e-01 -3.83754708e-02 -4.69510645e-01
-1.56424785e+00 -1.12110293e+00 7.07625985e-01 2.94725627e-01
-6.77072525e-01 6.43847227e-01 4.71567720e-01 -4.42261279e-01
-1.57526005e-02 -3.03565472e-01 -4.23946202e-01 -1.33713828e-02
4.10097510e-01 7.49739885e-01 6.04327142e-01 -6.99892581e-01
5.34478962e-01 -5.53109162e-02 -1.29261363e+00 7.29375303e-01
1.19652414e+00 7.87094295e-01 3.49201798e-01 7.91492224e-01
-3.86326998e-01 6.57839894e-01 -7.41938055e-01 -7.88926244e-01
-9.44947451e-02 -3.03495556e-01 -1.54389381e+00 3.16980481e-02
4.18012083e-01 -5.82770348e-01 3.90686691e-02 6.61187887e-01
4.98427331e-01 3.44583750e-01 2.97450781e-01 -1.02615619e+00
-7.46508121e-01 1.45666033e-01 -3.89592745e-03 2.15252981e-01
1.08978224e+00 1.35114282e-01 1.45675993e+00 -1.14145935e+00
3.77006471e-01 1.00815952e+00 1.07113533e-01 5.98969281e-01
-9.75857496e-01 -1.97829276e-01 1.56976119e-01 2.17274457e-01
-1.15031040e+00 -5.15816748e-01 3.39015722e-01 -1.71216875e-01
9.08671618e-01 7.46958315e-01 2.36623913e-01 7.14323521e-01
-3.38765234e-01 9.21486676e-01 9.48284388e-01 -3.56371075e-01
2.23320678e-01 3.52147877e-01 6.25535190e-01 2.46713802e-01
2.03181747e-02 -5.03176033e-01 -5.42926908e-01 -6.07513547e-01
2.06508245e-02 -2.11249918e-01 -4.08020496e-01 1.04916513e-01
-6.91499710e-01 7.31276572e-01 3.43830496e-01 1.52687013e-01
-4.38624114e-01 -2.21610546e-01 8.98997709e-02 5.94611526e-01
3.29986989e-01 9.16827381e-01 -4.24183547e-01 -3.82175446e-01
-4.02208328e-01 7.39296317e-01 1.53722858e+00 1.07172251e+00
8.36442530e-01 -1.15833402e+00 -7.14937389e-01 1.22950959e+00
1.68018386e-01 5.56816101e-01 1.64514020e-01 -1.25212586e+00
5.76799035e-01 7.76314259e-01 8.50832522e-01 -8.43066037e-01
-1.17352702e-01 -1.04664616e-01 -5.22899851e-02 -6.84865296e-01
4.27831024e-01 -2.53254801e-01 -3.29456717e-01 1.32012367e+00
3.99312198e-01 -6.98090851e-01 -2.96221338e-02 1.00393987e+00
1.11888170e+00 5.60989320e-01 -9.50225368e-02 -6.27084225e-02
2.00926185e+00 -1.13088453e+00 -8.50409985e-01 -3.31190467e-01
6.08795702e-01 -1.18588996e+00 1.43961775e+00 3.64131667e-02
-1.17074633e+00 -1.42991766e-01 -3.79790783e-01 -4.61665779e-01
-1.86199427e-01 4.23666164e-02 1.93826377e-01 5.23350954e-01
-1.07617724e+00 -1.43453673e-01 -1.78553328e-01 -4.67821449e-01
-3.08768094e-01 8.01591724e-02 3.66249010e-02 -3.35061729e-01
-1.47701144e+00 7.01602340e-01 -4.74837244e-01 3.77569459e-02
-5.50281048e-01 -5.24749100e-01 -3.82871181e-01 2.61257440e-01
7.74567187e-01 -5.13332963e-01 2.01224232e+00 -3.11142921e-01
-1.25457871e+00 7.67536044e-01 -6.88515663e-01 -5.26976027e-02
2.78319746e-01 -5.54725170e-01 -1.54216886e-01 3.28804195e-01
4.86173630e-01 5.06970465e-01 4.30842608e-01 -1.23451102e+00
-8.21331263e-01 -2.28104353e-01 8.35171223e-01 5.80196679e-01
-1.33492202e-01 2.64140964e-01 -1.18888998e+00 -1.28787570e-02
2.16549374e-02 -9.83180463e-01 -7.64051899e-02 -3.44839722e-01
-2.98425287e-01 -7.39797056e-01 4.20550048e-01 -6.97836459e-01
1.44297850e+00 -1.75246155e+00 -3.22994977e-01 -3.25477198e-02
2.73878664e-01 2.80424207e-01 -3.18509400e-01 1.02489185e+00
5.49938262e-01 4.75015521e-01 4.12927531e-02 -3.38753253e-01
9.94445756e-02 -4.12093364e-02 -3.48605424e-01 -1.92821443e-01
1.32038295e-01 1.15119946e+00 -8.92459691e-01 -5.12225866e-01
-5.28603375e-01 1.86071135e-02 -5.31999648e-01 5.49352348e-01
-6.65171564e-01 2.31169730e-01 -7.66665697e-01 6.59517348e-01
5.50492764e-01 -6.09409571e-01 -1.78727601e-02 9.71268266e-02
1.85210228e-01 6.87888861e-01 -6.58688426e-01 1.42018652e+00
-5.27045488e-01 4.46078956e-01 3.50956529e-01 -2.82878965e-01
5.90116680e-01 4.49443430e-01 3.51196796e-01 -1.10974073e+00
-2.80664206e-01 2.11468771e-01 -3.20421159e-01 -8.82474363e-01
9.13115382e-01 3.30041319e-01 -2.12287709e-01 9.22775984e-01
-4.30661231e-01 -3.52143228e-01 3.34465861e-01 6.11434221e-01
1.39746273e+00 -5.58878183e-01 -9.76159349e-02 -1.14988163e-01
7.82930613e-01 1.57864586e-01 9.47676972e-03 1.24883246e+00
-3.70537668e-01 4.61747169e-01 7.47242033e-01 -4.29403819e-02
-5.80655158e-01 -7.06253588e-01 9.37298760e-02 1.53774166e+00
3.06064725e-01 -5.34828186e-01 -8.33974719e-01 -8.08244288e-01
-1.66526753e-02 6.29594326e-01 -3.78292978e-01 7.03549013e-02
-5.97019851e-01 -2.32394785e-01 3.25483620e-01 1.01770321e-02
5.90243459e-01 -9.52278137e-01 -1.97665259e-01 1.29281998e-01
-1.00315666e+00 -1.26662600e+00 -1.14587879e+00 -4.44836050e-01
-4.48320806e-01 -1.22340298e+00 -9.75193799e-01 -8.52644920e-01
3.37542027e-01 9.77207303e-01 1.53468442e+00 5.27300119e-01
3.23964089e-01 9.54099834e-01 -7.45333076e-01 -1.25886694e-01
-7.29047582e-02 4.72373515e-01 -5.79091787e-01 -1.05348393e-01
6.88321292e-01 -1.15711145e-01 -1.12161458e+00 9.44377601e-01
-9.43351626e-01 -2.94372499e-01 2.46110097e-01 6.20238602e-01
2.26902775e-02 -5.85592449e-01 9.98116851e-01 -9.35115576e-01
1.66762292e+00 -6.40981555e-01 -4.69955295e-01 7.03733027e-01
-7.04101562e-01 7.72811333e-03 2.27767363e-01 -9.04492736e-02
-1.18128824e+00 -2.34137803e-01 -2.91212589e-01 4.47554648e-01
2.03554690e-01 6.41080439e-01 9.33902189e-02 3.35963592e-02
8.79837096e-01 1.22348547e-01 3.27749215e-02 -5.73328018e-01
3.99757773e-01 1.05953026e+00 2.14658573e-01 -7.84271121e-01
6.05431974e-01 1.33673087e-01 -9.17601466e-01 -6.42746031e-01
-1.12535477e+00 -1.27114475e+00 -4.24653180e-02 -4.01916116e-01
6.97478831e-01 -8.08065951e-01 -1.31083465e+00 1.25166729e-01
-1.40232980e+00 -1.40064701e-01 1.67951450e-01 -8.25853348e-02
-1.40369743e-01 4.79598105e-01 -7.05169857e-01 -1.02701068e+00
-4.80802268e-01 -1.22654033e+00 1.19477272e+00 4.10800546e-01
-3.54201674e-01 -7.49354541e-01 1.52026206e-01 1.16906774e+00
7.03593194e-01 -6.57240391e-01 8.53864372e-01 -9.46711659e-01
-7.67685652e-01 -4.56590593e-01 -3.25915605e-01 2.57460549e-02
1.27660990e-01 -5.51083684e-01 -8.56313944e-01 -1.68158621e-01
7.91841894e-02 -6.33268952e-01 5.92607796e-01 -2.91199744e-01
9.41208899e-01 -4.07405853e-01 -5.34250997e-02 -4.19442236e-01
9.43763852e-01 1.58948466e-01 5.05246937e-01 1.60243422e-01
1.61383808e-01 1.05536425e+00 8.11801136e-01 3.10129166e-01
9.70084488e-01 8.24303687e-01 1.64101765e-01 2.66308933e-01
2.18042493e-01 -1.76121086e-01 6.30091652e-02 6.14336610e-01
3.68078798e-01 -5.59117377e-01 -9.59786534e-01 6.52234256e-01
-1.76773286e+00 -6.15173757e-01 -1.89653173e-01 2.17765617e+00
1.17807806e+00 -1.73062608e-01 8.81908387e-02 -4.03364927e-01
6.37651622e-01 7.14904815e-02 -5.88714600e-01 -1.89105198e-01
2.12016433e-01 1.88083723e-01 9.83831822e-04 9.41708207e-01
-5.64560533e-01 8.69663417e-01 6.04584503e+00 5.11451423e-01
-5.92575431e-01 8.89198631e-02 6.29526079e-01 2.88801696e-02
-7.86330163e-01 1.85547605e-01 -8.83793116e-01 4.45510894e-01
8.26152623e-01 -4.55545306e-01 5.79972565e-01 5.94234586e-01
2.07215711e-01 -5.48600018e-01 -1.05169022e+00 7.44631588e-01
8.59092921e-02 -1.07619977e+00 4.92167622e-02 -4.35620733e-02
4.30792183e-01 -1.73514277e-01 -1.41138494e-01 7.89965689e-01
3.14118594e-01 -8.46981704e-01 1.01319693e-01 4.79461789e-01
3.70388299e-01 -2.81269699e-01 7.68808484e-01 6.07216001e-01
-8.11044633e-01 1.23548768e-01 -9.87309292e-02 -3.04521178e-04
3.16934764e-01 2.44551525e-01 -1.04413545e+00 2.41477787e-01
6.86087847e-01 -8.28374326e-02 -7.90692389e-01 9.99731064e-01
-6.12630472e-02 6.13764107e-01 -2.04020053e-01 -7.34918773e-01
2.31715426e-01 -1.89141799e-02 3.38181764e-01 8.44045162e-01
-1.24511719e-01 6.62203729e-01 1.01887442e-01 4.94484663e-01
-5.48218310e-01 4.41266745e-01 -1.99271113e-01 -3.39974184e-03
7.54920959e-01 1.32640219e+00 -1.90780327e-01 -3.38265270e-01
-5.70238411e-01 1.01912081e+00 3.75865489e-01 6.89447045e-01
-4.91329283e-01 -4.60546255e-01 3.46261770e-01 7.00332597e-02
-7.11783469e-02 -7.20271841e-02 3.94628495e-02 -1.11370063e+00
8.01157713e-01 -1.32957065e+00 3.98978561e-01 -7.56799817e-01
-1.25344181e+00 5.70914805e-01 -2.51228034e-01 -7.64886081e-01
-4.35765266e-01 -1.50647536e-01 -5.10002136e-01 1.18042719e+00
-1.63984728e+00 -5.06620407e-01 -6.18661582e-01 4.25931871e-01
7.32656956e-01 3.32987875e-01 6.21586084e-01 4.65133518e-01
-2.30345458e-01 5.60796857e-01 -8.45485330e-02 -5.00414744e-02
1.14935815e+00 -1.18430579e+00 4.44413602e-01 3.66269410e-01
-1.75161615e-01 9.41520751e-01 5.96865773e-01 -6.17020726e-01
-1.55669916e+00 -5.68873286e-01 1.42585814e+00 -8.81198585e-01
4.89417315e-01 -5.08384049e-01 -1.31186068e+00 2.43849859e-01
4.28636849e-01 -4.37674195e-01 8.60489249e-01 4.21326786e-01
-1.34811237e-01 -5.66657931e-02 -9.17096734e-01 7.91371286e-01
6.36676848e-01 -1.00223875e+00 -6.12477481e-01 8.46316874e-01
1.09726071e+00 -3.08149308e-01 -6.51746154e-01 1.81448430e-01
5.13149798e-01 -7.70340681e-01 7.74084508e-01 -6.79232836e-01
1.98423058e-01 1.27749005e-02 -1.44889772e-01 -8.03727627e-01
2.44979165e-03 -8.06852221e-01 8.30110461e-02 1.21849334e+00
9.10941422e-01 -6.68780804e-01 7.99453557e-01 1.65700758e+00
1.14310198e-01 -7.71342933e-01 -7.16560602e-01 -3.28379661e-01
1.28433406e-01 -2.06142113e-01 6.54195726e-01 4.83640015e-01
1.54946983e-01 7.93386340e-01 -1.33646861e-01 1.06667154e-01
1.14197232e-01 1.51337042e-01 9.48296845e-01 -9.49916065e-01
-1.12906903e-01 -2.22788692e-01 7.17888296e-01 -1.80180502e+00
-2.52183247e-02 -5.06854117e-01 3.83223861e-01 -1.72979558e+00
3.06689888e-01 -4.54270005e-01 1.84663251e-01 1.19445987e-01
-7.46652007e-01 -2.09086850e-01 2.02017594e-02 5.64484537e-01
-1.19967473e+00 5.65411270e-01 1.36080611e+00 -9.40928329e-03
-3.52422476e-01 4.42072153e-01 -1.21801865e+00 1.93607062e-01
6.62005246e-01 -2.37436861e-01 -4.85293984e-01 -5.65887153e-01
6.20601833e-01 5.75226128e-01 2.72231638e-01 -4.02249753e-01
7.60703981e-01 -1.88300848e-01 -3.28047544e-01 -5.86967528e-01
5.03702641e-01 -5.93144834e-01 -5.87806940e-01 3.16822715e-02
-9.27592218e-01 2.76386023e-01 -1.42952710e-01 6.10144973e-01
-3.64142179e-01 -6.53971255e-01 8.39111283e-02 -2.06901476e-01
-2.96294481e-01 1.63719118e-01 -6.52575314e-01 5.95476210e-01
4.32379276e-01 3.57144654e-01 -6.55134976e-01 -1.16840827e+00
-4.24330801e-01 1.09841776e+00 2.90841818e-01 5.34543037e-01
3.95545661e-01 -1.07343769e+00 -7.81503558e-01 -4.15375084e-01
5.46027601e-01 -1.40592113e-01 2.57115602e-01 5.16290963e-01
-1.99304327e-01 9.70636129e-01 6.15479350e-01 -4.69778240e-01
-1.17470241e+00 2.24787161e-01 1.82067797e-01 -4.44412589e-01
6.35846928e-02 7.03470767e-01 1.56245694e-01 -7.22521245e-01
2.87402213e-01 -1.10744998e-01 -4.92386878e-01 1.27331614e-01
8.15809906e-01 2.20776543e-01 3.64428401e-01 -1.70342326e-01
-2.11391762e-01 9.82496962e-02 -4.46770340e-01 -6.30263269e-01
6.41372323e-01 -6.32233381e-01 -2.10472420e-01 6.42333329e-02
1.25158679e+00 3.12981635e-01 -6.33036673e-01 -7.46288896e-01
4.13161367e-01 -4.15331692e-01 -4.35175359e-01 -1.14029777e+00
-3.33422661e-01 4.97671366e-01 1.86651707e-01 7.04060137e-01
1.08474696e+00 2.78210968e-01 1.10382748e+00 1.12139499e+00
1.56399578e-01 -1.12028110e+00 5.26009381e-01 8.62046540e-01
1.19150615e+00 -1.58363760e+00 -4.00506347e-01 -4.85188127e-01
-6.70216978e-01 6.80403948e-01 7.09614277e-01 4.34568346e-01
3.83969635e-01 -4.73496377e-01 3.33276242e-01 -5.01706421e-01
-1.24562562e+00 -4.28060174e-01 5.33388376e-01 7.70015568e-02
6.52948081e-01 -2.43000835e-01 -6.51242971e-01 7.06838071e-01
-1.82219028e-01 -2.05406263e-01 3.96220058e-01 1.03643727e+00
-6.53569341e-01 -1.16967022e+00 -2.49656707e-01 5.52069068e-01
-5.97332537e-01 -3.19142342e-01 -8.29476535e-01 3.16794723e-01
-8.81199360e-01 1.76663196e+00 -3.68902773e-01 -1.86642021e-01
6.55116260e-01 1.59653187e-01 -2.14823276e-01 -7.06691682e-01
-1.03160489e+00 -2.56237984e-01 6.16338134e-01 -5.57804883e-01
-1.35423586e-01 -4.98068660e-01 -8.28381658e-01 -1.45715624e-01
-4.68200684e-01 8.57482195e-01 6.59458518e-01 8.95915091e-01
7.42538691e-01 -1.60601825e-01 8.27799201e-01 1.48431370e-02
-8.40832531e-01 -9.78964329e-01 -5.80192469e-02 5.10819495e-01
3.68139774e-01 -2.96542376e-01 -3.92639577e-01 -2.34013468e-01]
|
[12.065069198608398, 7.849086284637451]
|
d48bc21d-0af3-4ca2-85d8-4e055b91a903
|
interactive-language-acquisition-with-one
|
1805.00462
| null |
http://arxiv.org/abs/1805.00462v1
|
http://arxiv.org/pdf/1805.00462v1.pdf
|
Interactive Language Acquisition with One-shot Visual Concept Learning through a Conversational Game
|
Building intelligent agents that can communicate with and learn from humans
in natural language is of great value. Supervised language learning is limited
by the ability of capturing mainly the statistics of training data, and is
hardly adaptive to new scenarios or flexible for acquiring new knowledge
without inefficient retraining or catastrophic forgetting. We highlight the
perspective that conversational interaction serves as a natural interface both
for language learning and for novel knowledge acquisition and propose a joint
imitation and reinforcement approach for grounded language learning through an
interactive conversational game. The agent trained with this approach is able
to actively acquire information by asking questions about novel objects and use
the just-learned knowledge in subsequent conversations in a one-shot fashion.
Results compared with other methods verified the effectiveness of the proposed
approach.
|
['Wei Xu', 'Haonan Yu', 'Haichao Zhang']
|
2018-04-26
|
interactive-language-acquisition-with-one-1
|
https://aclanthology.org/P18-1243
|
https://aclanthology.org/P18-1243.pdf
|
acl-2018-7
|
['grounded-language-learning']
|
['natural-language-processing']
|
[ 1.41508549e-01 6.63924932e-01 1.11887991e-01 4.53590080e-02
-2.59931028e-01 -5.94368160e-01 1.04865122e+00 4.79461625e-02
-7.25130439e-01 1.27594972e+00 -1.18544929e-01 6.42873496e-02
-2.18495056e-01 -8.79964113e-01 -7.43241489e-01 -7.06089914e-01
-2.64790773e-01 9.85229492e-01 3.96229237e-01 -6.78720057e-01
2.02259138e-01 6.27898335e-01 -1.75383830e+00 2.47965418e-02
9.06359673e-01 2.29569301e-01 9.43986952e-01 7.92256832e-01
-4.27698582e-01 1.26491880e+00 -5.75350761e-01 -1.71505332e-01
-1.11783035e-02 -6.33122385e-01 -1.08715415e+00 3.23872298e-01
-5.51524937e-01 -6.84474051e-01 -3.38590115e-01 5.77609360e-01
2.91445464e-01 6.01351261e-01 3.52683395e-01 -1.06493139e+00
-1.62044331e-01 6.33584976e-01 2.56691247e-01 6.53763860e-02
9.41315114e-01 6.19265258e-01 4.72392201e-01 -6.63294673e-01
8.60269964e-01 1.31149328e+00 4.07345265e-01 1.08857489e+00
-8.12719226e-01 -2.26536229e-01 2.05679893e-01 2.74397522e-01
-8.01557124e-01 -3.85607392e-01 6.31073058e-01 -1.75960943e-01
1.02642739e+00 -1.50099486e-01 1.17054415e+00 1.38753009e+00
9.43324715e-02 1.07602811e+00 9.02893364e-01 -4.97908682e-01
5.15705884e-01 7.04947591e-01 -2.16158599e-01 8.88470829e-01
6.91112578e-02 4.02541220e-01 -8.74874532e-01 -5.71037382e-02
7.71917045e-01 1.55395031e-01 -8.42695534e-02 -4.83375490e-01
-1.30925536e+00 6.75171077e-01 1.72786236e-01 5.88280678e-01
-7.83090770e-01 9.46747810e-02 3.64216834e-01 9.03858066e-01
1.32314101e-01 7.42284715e-01 -5.49856603e-01 -8.08960021e-01
-3.39712292e-01 2.04819813e-01 1.23211217e+00 9.47606206e-01
8.79833579e-01 1.43454760e-01 9.80582759e-02 7.15352595e-01
2.74231955e-02 6.70534611e-01 8.67996097e-01 -1.06769538e+00
-1.12719886e-01 7.64742076e-01 2.27051347e-01 -4.71290857e-01
-4.07121539e-01 -1.08056724e-01 -3.33191335e-01 4.28809673e-01
2.69257396e-01 -4.18534309e-01 -3.15780878e-01 1.63550711e+00
5.20065904e-01 -5.91736473e-02 5.44823885e-01 1.73271611e-01
7.02369034e-01 6.40961587e-01 -9.13706049e-02 -6.10796809e-01
9.19730186e-01 -1.08182669e+00 -6.63977504e-01 -2.15560377e-01
4.99911666e-01 -6.54068664e-02 1.20221913e+00 5.07214308e-01
-1.10964882e+00 -5.48987329e-01 -6.12236917e-01 3.22881967e-01
-3.72152716e-01 -6.04702771e-01 6.91816807e-01 3.20308805e-01
-1.06286371e+00 5.35946548e-01 -7.50646770e-01 -7.94674158e-01
9.19564664e-02 4.58546132e-01 -3.85731161e-01 1.53956547e-01
-1.03319514e+00 1.02198386e+00 6.60127640e-01 -1.87445745e-01
-1.36964738e+00 -3.01919103e-01 -6.71461165e-01 -1.19383663e-01
7.41277695e-01 -7.55136967e-01 1.60697973e+00 -1.14410436e+00
-2.52447963e+00 6.23899996e-01 8.75712708e-02 -6.39046788e-01
7.08041847e-01 -2.96640337e-01 2.61790436e-02 3.45282763e-01
-1.14055149e-01 6.69995785e-01 7.69481778e-01 -1.44552290e+00
-6.59354091e-01 -1.32755473e-01 5.08659124e-01 6.05353773e-01
-1.31132111e-01 -5.01316071e-01 5.12850545e-02 -2.03062296e-01
-2.24847674e-01 -9.38478589e-01 -2.03773826e-01 -1.47315010e-01
3.87922317e-01 -4.13705766e-01 7.80029595e-01 -2.76606679e-01
4.63014603e-01 -1.79163623e+00 3.55420172e-01 -9.15103301e-04
1.38986632e-01 4.73852485e-01 -2.00737491e-01 1.02430391e+00
6.96464896e-01 -3.66151303e-01 -8.74013826e-03 -2.27922559e-01
-6.00368641e-02 4.74427402e-01 -5.19950204e-02 -1.02880314e-01
-1.42586619e-01 1.08317280e+00 -1.41380239e+00 -2.47670814e-01
3.76763105e-01 2.76859015e-01 -5.70812464e-01 9.15738106e-01
-5.51017702e-01 9.69341278e-01 -6.09996736e-01 2.09799200e-01
-5.44419773e-02 -2.17170507e-01 3.31928819e-01 6.29056990e-01
-6.86899424e-02 1.34448469e-01 -8.33279490e-01 1.68210351e+00
-9.23324287e-01 4.87602323e-01 2.25583762e-01 -9.55325723e-01
1.10827529e+00 5.45241475e-01 8.62335265e-02 -6.21596575e-01
2.54304647e-01 6.58290759e-02 1.80218238e-02 -9.51659977e-01
1.73925176e-01 -3.34162265e-01 -5.98549028e-04 8.11217785e-01
4.42660838e-01 -6.18322790e-01 3.80789675e-02 2.06891477e-01
1.13175678e+00 1.65178344e-01 7.36796856e-01 1.52761312e-02
7.83593714e-01 8.78680050e-02 5.83960898e-02 1.25638425e+00
-1.45067051e-01 -1.92331225e-01 -2.77070738e-02 -6.49093747e-01
-6.51398301e-01 -1.17247164e+00 6.50826454e-01 1.43860435e+00
1.52559519e-01 2.12930202e-01 -6.69474423e-01 -6.01566792e-01
-2.90170521e-01 1.08502233e+00 -4.35123533e-01 -4.84915704e-01
-5.34593761e-01 2.25752503e-01 1.66468441e-01 -1.58138908e-02
8.49093676e-01 -2.14410019e+00 -1.04415953e+00 5.86855292e-01
-1.07780844e-01 -9.17366982e-01 -1.21427149e-01 2.14818716e-01
-1.02219832e+00 -9.57694054e-01 -4.46892768e-01 -1.04631793e+00
5.17596722e-01 2.77824283e-01 1.00908399e+00 4.42474991e-01
-8.91099405e-03 1.17275453e+00 -7.03141809e-01 -5.16685307e-01
-1.11889529e+00 1.42574951e-01 3.44626307e-01 -3.41486521e-02
2.26732939e-01 -9.61512327e-01 -2.73614705e-01 7.10636973e-02
-7.38260388e-01 2.21823141e-01 7.52930641e-01 1.12683284e+00
-1.84295788e-01 -9.70858261e-02 9.46535349e-01 -8.40837955e-01
1.18215632e+00 -4.37394410e-01 -3.23376626e-01 1.98640808e-01
-3.78915638e-01 2.03309715e-01 8.10026526e-01 -7.97916770e-01
-1.49878180e+00 3.62137822e-03 -6.90476522e-02 9.90945548e-02
-4.25293744e-01 2.06342235e-01 1.44979581e-02 -1.45511061e-01
5.28961957e-01 8.48277330e-01 5.10676920e-01 -2.14233741e-01
4.52587873e-01 6.84578240e-01 4.15083051e-01 -7.64579058e-01
6.72532439e-01 3.23259920e-01 -5.57358742e-01 -1.23266661e+00
-3.03081959e-01 -2.19541997e-01 -9.15514529e-01 -6.54208422e-01
3.64427298e-01 -6.53404176e-01 -1.34259057e+00 6.65613830e-01
-1.07924211e+00 -7.40855098e-01 -8.45814228e-01 5.72605908e-01
-9.99507844e-01 2.41240740e-01 -4.39499557e-01 -1.21427155e+00
-3.00228775e-01 -8.38173509e-01 5.77652514e-01 4.28923905e-01
-2.83353508e-01 -1.09962904e+00 2.77362794e-01 3.55898261e-01
6.79218888e-01 -1.00565843e-01 5.82402587e-01 -9.65768099e-01
-7.59066403e-01 -2.34088466e-01 4.62680221e-01 9.45734754e-02
2.59220302e-01 -5.94926953e-01 -6.64438725e-01 -4.83241647e-01
2.69744813e-01 -9.67790842e-01 2.72945911e-01 -2.70146608e-01
3.86304289e-01 -5.97294569e-01 -2.74039239e-01 -1.96313530e-01
8.66613507e-01 6.25291765e-01 3.65193665e-01 3.06128919e-01
4.20603231e-02 6.75629914e-01 4.60221559e-01 6.37731135e-01
3.93887103e-01 4.82910782e-01 2.22277135e-01 4.58639979e-01
1.09043993e-01 -4.41843390e-01 5.67082882e-01 8.83051753e-01
-1.64428711e-01 -1.77292489e-02 -8.07846606e-01 6.67447984e-01
-1.94804025e+00 -1.11396289e+00 7.41800606e-01 2.01793313e+00
1.02877188e+00 3.33367527e-01 1.09478407e-01 -2.35035613e-01
4.32084054e-01 -3.61033767e-01 -7.68186808e-01 -4.74929899e-01
1.45198375e-01 1.97788090e-01 -1.27970606e-01 9.34164047e-01
-1.92127079e-01 1.23413837e+00 6.28638744e+00 4.96090204e-01
-8.19292545e-01 1.71762303e-01 7.56422244e-03 1.42529577e-01
-6.94023073e-02 -1.91141888e-01 -2.92224944e-01 5.57988584e-02
7.74959445e-01 -4.35721695e-01 9.05159533e-01 8.75423789e-01
3.46481770e-01 -5.21270037e-01 -1.14074385e+00 8.07313323e-01
5.26725389e-02 -1.19995773e+00 3.36838484e-01 -2.56655365e-01
4.47206944e-01 -2.11758018e-01 -2.50510544e-01 6.30737007e-01
5.68994105e-01 -8.31814229e-01 3.36272568e-01 1.00369000e+00
2.24405423e-01 -5.96325278e-01 6.92440391e-01 1.22973549e+00
-7.90608883e-01 -2.83245832e-01 1.09037086e-02 -7.37113774e-01
1.29271895e-01 -2.76109874e-01 -1.53472853e+00 1.03835933e-01
3.27292383e-01 3.08767587e-01 -1.85992330e-01 6.67811632e-01
-2.37256780e-01 4.44852382e-01 -3.02848369e-01 -9.65862811e-01
2.21376836e-01 -2.87616134e-01 7.99344838e-01 8.11764121e-01
7.73942545e-02 4.27412570e-01 2.74591774e-01 6.65033340e-01
1.87045738e-01 9.87496972e-02 -1.16183925e+00 -1.39973238e-01
4.52891558e-01 7.95635641e-01 -6.60835922e-01 -4.77842152e-01
-2.15910181e-01 1.17667365e+00 4.98841465e-01 2.22697139e-01
-2.45181084e-01 -2.03761548e-01 1.97357625e-01 -4.28591482e-02
2.84904540e-01 -3.60725909e-01 4.56344515e-01 -9.42809463e-01
-2.37747595e-01 -1.13737810e+00 -7.66194239e-02 -8.06423426e-01
-1.04607654e+00 7.94922769e-01 7.46456534e-02 -9.24551129e-01
-8.67120326e-01 -2.47916788e-01 -5.50160050e-01 1.67623997e-01
-1.06047714e+00 -8.97047818e-01 -4.63737577e-01 8.93172920e-01
1.02143824e+00 -8.26851606e-01 1.02677441e+00 -4.56967652e-01
1.49995491e-01 2.94669092e-01 4.15449440e-02 -2.51164883e-01
7.70348236e-02 -9.96315062e-01 -7.88992941e-02 1.90681890e-01
1.60355181e-01 4.54040021e-01 9.80727613e-01 -5.19904554e-01
-1.52950501e+00 -5.45346141e-01 6.02352142e-01 -2.92651027e-01
3.98821712e-01 -5.49389601e-01 -1.00580156e+00 5.85084677e-01
5.25719583e-01 -6.73041940e-01 3.28318149e-01 -7.99639300e-02
2.63089120e-01 -1.20017380e-01 -1.35060501e+00 9.00410712e-01
1.16247940e+00 -5.20042598e-01 -1.13408196e+00 3.60329598e-01
9.72318947e-01 -5.54164089e-02 -2.73168296e-01 7.76633713e-03
4.69501108e-01 -1.11090660e+00 6.62853062e-01 -5.40358543e-01
-1.69165075e-01 7.34495521e-02 3.16625744e-01 -1.47122848e+00
-2.44008340e-02 -1.17556131e+00 -6.19398896e-03 8.52468133e-01
3.63644600e-01 -9.00545359e-01 8.60771716e-01 2.55973756e-01
2.63231367e-01 -3.44739676e-01 -9.66004848e-01 -7.78091848e-01
-1.31411254e-01 -2.01358154e-01 4.08367932e-01 5.58023214e-01
5.98532379e-01 6.12711847e-01 -5.28159738e-01 -1.35586128e-01
3.71511459e-01 -1.27525836e-01 1.26155269e+00 -1.19706714e+00
-4.60777938e-01 -1.21178083e-01 -1.41671434e-01 -1.32533741e+00
3.16548616e-01 -5.43147683e-01 3.80700320e-01 -1.30759633e+00
1.01007849e-01 -3.97020094e-02 1.79653674e-01 1.47499859e-01
8.97916853e-02 -5.76105714e-01 2.54792750e-01 2.28211671e-01
-9.54109371e-01 9.49422002e-01 1.46831417e+00 -4.35032547e-02
-7.30209351e-01 3.38923216e-01 -2.57899612e-01 7.79631019e-01
7.74548769e-01 -3.62959951e-01 -7.42241442e-01 1.40113875e-01
1.45426109e-01 4.45715219e-01 1.73322484e-01 -1.11133718e+00
7.20614552e-01 -2.80739695e-01 -1.68942004e-01 -7.32387230e-02
3.82872343e-01 -1.07841957e+00 2.48501776e-03 9.17112648e-01
-4.79291886e-01 -1.47209167e-02 1.27455875e-01 8.25247049e-01
-8.65355134e-02 -4.31394041e-01 5.15972555e-01 -7.35751987e-01
-8.40037048e-01 -1.35735333e-01 -1.18060219e+00 2.85755266e-02
1.24101257e+00 -4.74547833e-01 2.02024654e-01 -1.09197497e+00
-9.66461301e-01 3.61276627e-01 3.03061306e-01 6.22050047e-01
7.29791522e-01 -8.80881071e-01 -5.37974656e-01 3.16807002e-01
4.04747762e-02 1.52462376e-02 -2.76649743e-02 2.51315743e-01
-5.10620475e-01 1.90720573e-01 -4.77246851e-01 -4.07227933e-01
-1.11201119e+00 6.34182632e-01 5.65284014e-01 -4.03879315e-01
-6.37422442e-01 6.24613225e-01 1.11085318e-01 -8.80153894e-01
5.41473210e-01 -1.17200010e-01 -3.70567918e-01 -2.07928464e-01
5.52653670e-01 2.22600341e-01 -3.61569136e-01 -1.21078245e-01
-2.69572791e-02 1.07175760e-01 -1.45666927e-01 -3.66107553e-01
1.37338519e+00 -4.88027841e-01 -1.84681881e-02 1.06732774e+00
7.18386531e-01 -2.02100605e-01 -1.27755618e+00 -5.76340616e-01
4.14983137e-03 -1.45215824e-01 -6.16736233e-01 -9.39858437e-01
-3.29000503e-01 5.06138623e-01 4.22588527e-01 3.85716915e-01
7.25994587e-01 1.88614115e-01 5.76220095e-01 1.34417081e+00
9.42903221e-01 -1.21720159e+00 9.54840720e-01 7.66993761e-01
1.14139712e+00 -1.35416269e+00 -3.20414811e-01 1.23484489e-02
-7.67231345e-01 1.26176119e+00 7.03126609e-01 -3.20422202e-02
6.64794683e-01 9.02925879e-02 1.05519734e-01 -2.28184432e-01
-1.07255602e+00 -2.83656627e-01 -2.13995486e-01 9.84120190e-01
-1.38753414e-01 -2.34738722e-01 -1.53597638e-01 2.27004901e-01
-3.61486405e-01 2.59280711e-01 6.81561232e-01 1.18166792e+00
-9.38594103e-01 -9.07008648e-01 2.21644640e-02 1.44591138e-01
2.71344066e-01 2.82114387e-01 -5.12818098e-01 8.95866454e-01
-2.80969385e-02 1.09195507e+00 -1.81259081e-01 -1.88551098e-01
4.07525182e-01 1.97549865e-01 5.37194550e-01 -8.30319226e-01
-6.53021574e-01 -4.87551838e-01 2.29568575e-02 -4.28241342e-01
-4.82495308e-01 -7.69893050e-01 -1.31542277e+00 -1.00785039e-01
-1.69278100e-01 5.20408928e-01 3.59832138e-01 1.30198634e+00
1.67720661e-01 3.24186146e-01 8.35919142e-01 -9.14113283e-01
-5.39328992e-01 -1.04105330e+00 -3.17241937e-01 3.76247495e-01
5.96734881e-01 -6.34473264e-01 -4.70189542e-01 -2.63487957e-02]
|
[4.092974662780762, 1.2989144325256348]
|
5a403bec-f52d-4790-aac4-577b310dbd01
|
adaptive-linear-span-network-for-object
|
2011.03972
| null |
https://arxiv.org/abs/2011.03972v1
|
https://arxiv.org/pdf/2011.03972v1.pdf
|
Adaptive Linear Span Network for Object Skeleton Detection
|
Conventional networks for object skeleton detection are usually hand-crafted. Although effective, they require intensive priori knowledge to configure representative features for objects in different scale granularity.In this paper, we propose adaptive linear span network (AdaLSN), driven by neural architecture search (NAS), to automatically configure and integrate scale-aware features for object skeleton detection. AdaLSN is formulated with the theory of linear span, which provides one of the earliest explanations for multi-scale deep feature fusion. AdaLSN is materialized by defining a mixed unit-pyramid search space, which goes beyond many existing search spaces using unit-level or pyramid-level features.Within the mixed space, we apply genetic architecture search to jointly optimize unit-level operations and pyramid-level connections for adaptive feature space expansion. AdaLSN substantiates its versatility by achieving significantly higher accuracy and latency trade-off compared with state-of-the-arts. It also demonstrates general applicability to image-to-mask tasks such as edge detection and road extraction. Code is available at \href{https://github.com/sunsmarterjie/SDL-Skeleton}{\color{magenta}github.com/sunsmarterjie/SDL-Skeleton}.
|
['Qixiang Ye', 'Jianbin Jiao', 'Yunjie Tian', 'Chang Liu']
|
2020-11-08
| null | null | null | null |
['object-skeleton-detection']
|
['computer-vision']
|
[ 2.47044042e-01 -9.68775749e-02 -2.45236903e-01 -2.09358141e-01
-7.51610756e-01 -2.34298483e-01 3.42933267e-01 -1.48648053e-01
-2.57887244e-01 4.19336706e-01 5.00360206e-02 -2.39144921e-01
-6.25652790e-01 -8.81385326e-01 -6.17222250e-01 -4.63264018e-01
1.86111089e-02 9.30514932e-02 4.95218724e-01 -2.37962261e-01
2.63632864e-01 9.97171760e-01 -1.77682269e+00 2.42652982e-01
6.44708157e-01 1.32228506e+00 3.23591560e-01 6.68536544e-01
4.69852751e-03 9.08828601e-02 -2.20386088e-01 -2.82282084e-01
6.16544545e-01 -2.79366314e-01 -5.14529705e-01 -1.15900651e-01
7.29490876e-01 -3.65434103e-02 -4.72865313e-01 9.89349365e-01
7.29540229e-01 2.59721987e-02 5.56466699e-01 -1.25403285e+00
-5.59237361e-01 4.64219987e-01 -7.75696993e-01 4.62961018e-01
-1.42327815e-01 3.63602966e-01 1.07321191e+00 -1.20775497e+00
3.73020142e-01 1.25130308e+00 6.47997439e-01 4.23076063e-01
-1.09159040e+00 -6.45663023e-01 1.76832005e-01 3.54099840e-01
-1.67290115e+00 -5.03676474e-01 7.68820167e-01 -2.14923799e-01
8.50158751e-01 4.42073613e-01 7.48829186e-01 7.72357702e-01
1.36383668e-01 8.69321764e-01 7.40151405e-01 -4.27085757e-01
1.24175437e-01 -4.77173090e-01 9.11610387e-03 1.02560472e+00
5.63511670e-01 2.35829130e-01 -8.51900399e-01 1.31971002e-01
1.28109050e+00 -5.82763925e-02 -1.61791384e-01 -3.68668348e-01
-1.22651029e+00 7.11362481e-01 6.95208669e-01 1.85119569e-01
-5.38557529e-01 5.60006142e-01 2.67270476e-01 1.06926180e-01
7.27274567e-02 4.56631929e-01 -5.31591654e-01 1.15564868e-01
-1.03616726e+00 2.41842970e-01 2.67177999e-01 7.04072952e-01
8.72145832e-01 3.37871283e-01 -3.09455484e-01 8.16887915e-01
2.18890488e-01 3.59727323e-01 4.44224656e-01 -1.25546324e+00
1.52745917e-01 7.75769055e-01 -3.12256336e-01 -8.42640102e-01
-6.98094189e-01 -6.56956255e-01 -8.48735750e-01 6.76729262e-01
2.97282636e-01 -1.12223908e-01 -1.08891416e+00 1.55547750e+00
5.47293842e-01 1.82139173e-01 -1.46306753e-01 8.69555235e-01
9.23484147e-01 3.62222373e-01 -1.83609560e-01 1.63560390e-01
1.64666557e+00 -1.04606462e+00 -3.20422873e-02 -3.58791232e-01
3.38358790e-01 -5.92630386e-01 1.16123664e+00 3.42392415e-01
-1.14596081e+00 -7.08430111e-01 -1.20483923e+00 -1.36212736e-01
-3.50015461e-01 2.50472337e-01 7.73285508e-01 6.17971659e-01
-1.31502068e+00 6.73363149e-01 -9.08941388e-01 -3.14353645e-01
7.87887096e-01 6.67055905e-01 -2.24446893e-01 1.63258806e-01
-9.12744641e-01 5.63669622e-01 4.10273254e-01 1.93343490e-01
-6.27384186e-01 -4.55884129e-01 -7.73083985e-01 1.30384520e-01
6.01452827e-01 -9.87852633e-01 1.16690183e+00 -6.59467340e-01
-1.32284415e+00 5.51555276e-01 -6.41322043e-03 -5.03203332e-01
3.39297354e-01 -2.40662426e-01 -2.82670051e-01 2.59187549e-01
6.98217824e-02 1.03951550e+00 9.08615947e-01 -1.04074407e+00
-7.89289415e-01 -4.39833283e-01 -6.92424700e-02 1.96547568e-01
-4.72311020e-01 -3.89092006e-02 -7.38341630e-01 -8.94806266e-01
2.16876119e-01 -6.78637385e-01 -4.13319677e-01 2.90028453e-01
-4.54588860e-01 -1.96498722e-01 8.44510078e-01 -3.74787778e-01
1.57463932e+00 -1.90718782e+00 -4.77095209e-02 3.58177155e-01
3.63643140e-01 3.21533263e-01 -3.78003389e-01 2.36711383e-01
-7.98002705e-02 3.43528762e-02 -2.47479826e-01 -1.47673756e-01
5.57658030e-03 -1.09397210e-01 2.26185352e-01 3.94908011e-01
3.96313608e-01 1.14204872e+00 -5.69968522e-01 -6.07652247e-01
2.87496686e-01 5.76556683e-01 -3.97944629e-01 -3.86083543e-01
-1.27597377e-01 1.13248311e-01 -6.41060650e-01 1.06525397e+00
3.82944524e-01 -4.41200525e-01 -2.99997807e-01 -5.44027746e-01
-2.15393066e-01 -1.17834449e-01 -1.45110857e+00 1.73543024e+00
-2.78592348e-01 6.37688637e-01 1.12162538e-01 -7.41936862e-01
9.41397667e-01 -7.47009814e-02 5.29018283e-01 -5.68172932e-01
2.22833797e-01 3.04969400e-01 5.15665486e-02 -2.23665774e-01
3.61359149e-01 4.14884925e-01 3.27043608e-02 1.65125027e-01
-7.98606128e-02 -6.29918426e-02 4.19965923e-01 -1.19683169e-01
1.28209186e+00 2.00629100e-01 4.98167515e-01 -2.59062201e-01
5.62441051e-01 -2.13729516e-01 5.37676692e-01 7.33166218e-01
-9.03149620e-02 5.79918981e-01 1.56948075e-01 -5.46032071e-01
-8.73973966e-01 -1.11667669e+00 -3.26213419e-01 1.25439715e+00
1.72287419e-01 -4.37334687e-01 -7.15888917e-01 -5.23130953e-01
1.10177258e-02 4.77175981e-01 -5.84276259e-01 -4.25473265e-02
-8.00202131e-01 -5.69253922e-01 5.44041574e-01 6.98146760e-01
6.48546815e-01 -1.11980534e+00 -1.23371327e+00 9.00173709e-02
1.97758645e-01 -7.13837624e-01 -5.43108881e-01 2.72660345e-01
-8.71394992e-01 -8.99286509e-01 -6.43597960e-01 -5.91073275e-01
5.90633035e-01 1.48437008e-01 1.04470026e+00 2.60333866e-01
-9.07310665e-01 3.38286489e-01 -2.71988481e-01 -3.46525252e-01
-3.54433134e-02 2.39160523e-01 -8.20255373e-03 -1.01023324e-01
1.83014035e-01 -9.30901408e-01 -1.18716156e+00 4.31033611e-01
-9.17872846e-01 1.87556967e-01 1.01308417e+00 7.30501354e-01
7.34686017e-01 1.99043807e-02 5.90822518e-01 -2.63633668e-01
4.76610750e-01 1.10883359e-02 -8.10998678e-01 1.58953384e-01
-7.14621067e-01 1.98172808e-01 2.84428418e-01 -2.25642145e-01
-6.09345376e-01 3.27779144e-01 -1.89414322e-01 -5.02996564e-01
-3.12389702e-01 3.24923754e-01 -2.07909092e-01 -2.68874675e-01
8.23975742e-01 1.26295328e-01 2.13152766e-02 -4.93089706e-01
5.34588695e-01 4.79058623e-01 7.27407455e-01 -5.60039461e-01
7.99557328e-01 6.07176721e-01 2.11655721e-01 -7.25657642e-01
-6.43279910e-01 -3.01246256e-01 -8.01530302e-01 -3.54211569e-01
8.01847935e-01 -7.34547079e-01 -5.97479343e-01 4.04871881e-01
-9.13714886e-01 -4.62272674e-01 -6.29740119e-01 2.18707144e-01
-6.99996531e-01 2.19413981e-01 -5.05754769e-01 -5.36450624e-01
-6.03009820e-01 -1.13770807e+00 1.22574925e+00 4.89194095e-01
-2.00953871e-01 -5.95957458e-01 -3.24521393e-01 2.15176627e-01
4.21276748e-01 3.35730165e-01 7.82160163e-01 -4.19717789e-01
-7.60552108e-01 -2.45702073e-01 -4.13194478e-01 2.18315661e-01
1.15073122e-01 9.56139714e-02 -7.69824862e-01 -1.86154932e-01
-4.92845476e-01 -9.87137705e-02 9.85956848e-01 7.07081735e-01
1.44689965e+00 -2.32761979e-01 -3.83616418e-01 7.56684721e-01
1.38661945e+00 8.50465074e-02 5.94713926e-01 5.27500451e-01
6.96188927e-01 3.56346190e-01 4.62470502e-01 7.56204486e-01
2.04989612e-01 8.37635696e-01 6.67393267e-01 -3.49646211e-01
-6.53566718e-01 1.10362247e-01 1.53907999e-01 2.50469625e-01
-2.32410714e-01 -1.11431554e-01 -9.37974095e-01 5.25042593e-01
-1.86344743e+00 -8.41317296e-01 6.02933615e-02 2.10203147e+00
7.05866456e-01 3.21576715e-01 3.52052689e-01 9.80603602e-03
7.95020521e-01 2.12490395e-01 -8.63468170e-01 -1.10524192e-01
-2.39478409e-01 4.01857376e-01 7.64329135e-01 3.32013190e-01
-1.19038057e+00 1.06844819e+00 4.68555069e+00 1.30706275e+00
-1.09672666e+00 1.01102754e-01 5.83667755e-01 -3.54732871e-01
-7.50808641e-02 -1.42972499e-01 -1.09881318e+00 8.36679116e-02
5.58906674e-01 -1.84173435e-01 3.71910483e-01 7.26310313e-01
2.77595222e-01 -1.79414563e-02 -8.91968310e-01 9.47423041e-01
-1.21040702e-01 -1.55565786e+00 -2.07877550e-02 9.14457664e-02
6.57067537e-01 2.97737747e-01 1.50527909e-01 5.56010604e-02
2.36815602e-01 -9.81011629e-01 8.65583718e-01 2.93957859e-01
9.55218434e-01 -5.99659264e-01 2.46873528e-01 1.20622208e-02
-1.62055504e+00 -4.39186752e-01 -2.07216650e-01 2.68868059e-01
3.34377199e-01 4.94087040e-01 -5.04552484e-01 5.28192699e-01
8.67327273e-01 4.63908166e-01 -7.90392399e-01 1.32687771e+00
-1.34452298e-01 3.43670487e-01 -5.45514643e-01 2.67512463e-02
2.83133894e-01 4.79925685e-02 5.51525116e-01 1.34717977e+00
5.80034256e-01 -1.39105588e-01 8.54835361e-02 6.34074152e-01
-3.81095819e-02 6.41305000e-02 -3.93459558e-01 1.64792687e-01
7.26887345e-01 1.35658514e+00 -1.30684829e+00 -2.25580186e-01
-2.51543730e-01 8.50260079e-01 1.47085741e-01 2.28968471e-01
-9.29165661e-01 -5.09169757e-01 6.62196517e-01 3.70088816e-01
6.66978478e-01 -1.52383968e-01 -5.28788745e-01 -6.45475626e-01
7.30792657e-02 -7.07857430e-01 4.12829936e-01 -6.50128007e-01
-1.06672597e+00 6.99757159e-01 1.54910877e-01 -1.29146373e+00
6.08283654e-02 -7.35055029e-01 -5.26660085e-01 5.35709441e-01
-1.15945292e+00 -1.34483695e+00 -4.05231059e-01 6.18915975e-01
9.46585715e-01 -2.43777737e-01 4.98152316e-01 3.45559150e-01
-5.68035126e-01 7.10019946e-01 -1.32076189e-01 6.50946349e-02
3.81912410e-01 -9.38180327e-01 7.71710992e-01 9.59108829e-01
2.52995282e-01 4.17736322e-01 4.81020421e-01 -7.15226173e-01
-1.24353421e+00 -1.09046161e+00 2.92204291e-01 -2.04953611e-01
6.46076798e-01 -1.26614109e-01 -6.60842896e-01 2.44469866e-01
2.60561751e-03 2.13002831e-01 3.32751125e-01 -2.29409099e-01
-3.09988260e-01 -3.33405256e-01 -9.07628655e-01 8.40375185e-01
1.39238429e+00 -1.33566126e-01 -9.35768485e-02 3.16701800e-01
7.23078012e-01 -4.11397159e-01 -8.46517026e-01 8.12027872e-01
7.33099103e-01 -1.09107864e+00 1.23244345e+00 -2.18714312e-01
3.15769672e-01 -5.93003094e-01 -3.22571427e-01 -6.91650867e-01
-5.82971931e-01 -5.80747485e-01 -2.63361126e-01 8.97779286e-01
6.04704797e-01 -5.56504428e-01 1.01762950e+00 1.97246745e-01
-4.82084185e-01 -1.36527872e+00 -1.23658121e+00 -1.00813282e+00
-2.30150923e-01 -5.27414083e-01 6.09242976e-01 4.88389611e-01
-5.37081182e-01 2.56395012e-01 -5.93890697e-02 2.64431268e-01
6.26442492e-01 9.11499262e-02 6.53816164e-01 -1.23971462e+00
-3.63148630e-01 -1.16083872e+00 -6.16127014e-01 -1.02991045e+00
-2.31085673e-01 -7.48219609e-01 -1.03020214e-01 -1.79934740e+00
-1.62032440e-01 -4.17389482e-01 -3.26010644e-01 8.23001146e-01
8.63002539e-02 5.86862326e-01 1.87896043e-01 1.28297195e-01
-5.72607160e-01 4.93992269e-01 1.12307620e+00 1.37566403e-01
-2.99093306e-01 2.38668904e-01 -8.42397153e-01 8.11267436e-01
9.91631985e-01 -3.00812274e-01 -3.22602868e-01 -3.28839362e-01
1.04510628e-01 -2.04027325e-01 6.92978501e-01 -1.37243021e+00
3.29433560e-01 -2.26792797e-01 4.36472893e-01 -7.65032172e-01
5.60752034e-01 -6.31137490e-01 2.07584381e-01 4.98129696e-01
-1.54385686e-01 3.20614457e-01 2.89566338e-01 4.76497829e-01
4.74168472e-02 -2.10399032e-01 7.59278595e-01 -1.97815001e-01
-9.99545217e-01 4.47621137e-01 -1.85297236e-01 -2.29651079e-01
9.82086062e-01 -8.45032871e-01 -3.07829499e-01 -2.02148948e-02
-6.66497827e-01 1.70580164e-01 3.95329833e-01 4.63293821e-01
8.55263233e-01 -1.27311337e+00 -7.31646359e-01 1.49798110e-01
3.87511142e-02 2.00042799e-01 4.96402353e-01 8.75491261e-01
-5.49895883e-01 3.06841731e-01 -2.26673096e-01 -6.21642053e-01
-1.32470012e+00 3.36923510e-01 3.23110402e-01 -1.71591893e-01
-6.63795769e-01 1.20806313e+00 2.74171919e-01 -1.29418641e-01
1.62529618e-01 -1.87526673e-01 4.92274351e-02 -6.57045767e-02
3.07378173e-01 6.07090414e-01 1.45374492e-01 -6.16597414e-01
-3.81947339e-01 8.12533081e-01 2.25395169e-02 3.70703749e-02
1.32670796e+00 -1.14609279e-01 -2.98138242e-02 3.46202999e-02
8.31608355e-01 -1.27504840e-01 -1.49511147e+00 -2.56743461e-01
2.90434565e-02 -2.93495834e-01 1.67339832e-01 -7.09391415e-01
-1.22205436e+00 6.55642211e-01 9.98514414e-01 -7.05462024e-02
1.40327263e+00 2.58544385e-01 7.45084703e-01 4.49348837e-01
3.05881977e-01 -1.10282922e+00 3.24816763e-01 3.90414804e-01
1.08627570e+00 -1.04319382e+00 2.90274650e-01 -3.03992718e-01
-3.53853166e-01 1.15857005e+00 8.02789092e-01 -8.11928138e-02
6.93438828e-01 4.17335898e-01 -1.27986789e-01 -5.46414256e-01
-5.33935070e-01 -5.25824726e-01 4.60955411e-01 4.08928543e-01
1.74585611e-01 9.44548938e-03 -1.81606621e-01 3.44194025e-01
-2.50610054e-01 -5.38369939e-02 1.11626826e-01 9.22821760e-01
-6.58611774e-01 -1.12932336e+00 -4.42981899e-01 7.25735426e-01
-2.22805738e-01 -2.29686275e-01 -1.80171803e-01 7.99509466e-01
2.94051915e-01 6.09153450e-01 -4.60560024e-02 -3.97754192e-01
3.90053749e-01 -2.51004547e-01 3.62673968e-01 -5.07903337e-01
-5.28421462e-01 3.17666948e-01 7.95563683e-02 -7.09186971e-01
-2.85373837e-01 -6.88616574e-01 -1.22541809e+00 -4.52367775e-02
-3.93214852e-01 -2.15571806e-01 5.83086014e-01 6.43984795e-01
6.55902863e-01 6.41418576e-01 2.22174644e-01 -1.27292919e+00
-3.86868447e-01 -6.37816489e-01 -3.63109022e-01 -2.24524081e-01
1.87636212e-01 -7.66468167e-01 -8.46157670e-02 -1.57444961e-02]
|
[9.241721153259277, -0.3051709234714508]
|
5b96c3cb-d76f-418e-b130-2cfd7ddf8efe
|
representing-additive-gaussian-processes-by
|
2305.00324
| null |
https://arxiv.org/abs/2305.00324v1
|
https://arxiv.org/pdf/2305.00324v1.pdf
|
Representing Additive Gaussian Processes by Sparse Matrices
|
Among generalized additive models, additive Mat\'ern Gaussian Processes (GPs) are one of the most popular for scalable high-dimensional problems. Thanks to their additive structure and stochastic differential equation representation, back-fitting-based algorithms can reduce the time complexity of computing the posterior mean from $O(n^3)$ to $O(n\log n)$ time where $n$ is the data size. However, generalizing these algorithms to efficiently compute the posterior variance and maximum log-likelihood remains an open problem. In this study, we demonstrate that for Additive Mat\'ern GPs, not only the posterior mean, but also the posterior variance, log-likelihood, and gradient of these three functions can be represented by formulas involving only sparse matrices and sparse vectors. We show how to use these sparse formulas to generalize back-fitting-based algorithms to efficiently compute the posterior mean, posterior variance, log-likelihood, and gradient of these three functions for additive GPs, all in $O(n \log n)$ time. We apply our algorithms to Bayesian optimization and propose efficient algorithms for posterior updates, hyperparameters learning, and computations of the acquisition function and its gradient in Bayesian optimization. Given the posterior, our algorithms significantly reduce the time complexity of computing the acquisition function and its gradient from $O(n^2)$ to $O(\log n)$ for general learning rate, and even to $O(1)$ for small learning rate.
|
['Liang Ding', 'HaoYuan Chen', 'Lu Zou']
|
2023-04-29
| null | null | null | null |
['additive-models']
|
['methodology']
|
[ 4.08164822e-02 -2.17388034e-01 4.04620767e-01 -1.90656275e-01
-1.19015086e+00 -4.71759498e-01 6.48523495e-02 2.53791898e-01
-7.74300337e-01 7.01329231e-01 -2.92715281e-01 -3.26660097e-01
-5.48694253e-01 -8.59732628e-01 -9.11798716e-01 -9.85291958e-01
-5.22028923e-01 7.13044107e-01 2.30568945e-01 1.46528214e-01
3.22863996e-01 3.72973263e-01 -1.36237776e+00 -7.18011141e-01
9.56569076e-01 1.37895036e+00 6.94292188e-02 9.49899435e-01
-2.86477897e-02 2.65900314e-01 -1.90669701e-01 -3.89845610e-01
3.51185173e-01 -2.14630023e-01 -2.17837185e-01 1.79499108e-02
3.96972567e-01 -2.82362908e-01 -2.47516394e-01 1.43601167e+00
5.52832305e-01 6.18100464e-01 9.20015693e-01 -9.81037676e-01
-2.28683829e-01 5.82585156e-01 -8.24059010e-01 9.92411375e-02
1.87600572e-02 -8.86290893e-02 6.61479414e-01 -8.36937428e-01
1.43888742e-01 1.44230151e+00 8.49041164e-01 5.77966012e-02
-1.43995059e+00 -8.66583884e-01 4.22055095e-01 -1.32601678e-01
-1.72294939e+00 -3.60048771e-01 4.27166998e-01 -5.71804106e-01
5.43173850e-01 1.06886961e-01 4.95603025e-01 3.82979214e-01
1.40717566e-01 5.65791607e-01 8.21905196e-01 -3.05862457e-01
5.36856294e-01 -6.63817525e-02 3.39770734e-01 9.91388917e-01
5.01389444e-01 -9.89200920e-02 -5.74583709e-01 -4.54172611e-01
8.73613417e-01 -5.18357046e-02 -2.34215111e-01 -2.53376395e-01
-5.83113849e-01 9.39944863e-01 -2.12591961e-01 -1.68912530e-01
-2.76542038e-01 8.18134248e-01 -6.13532551e-02 4.33952436e-02
5.86154163e-01 -9.14615300e-03 -5.15413284e-01 -6.29226744e-01
-6.24072373e-01 4.67114717e-01 1.01601636e+00 9.88980711e-01
1.02696800e+00 1.71971232e-01 7.90419653e-02 1.00386178e+00
6.24506652e-01 1.27301514e+00 6.70405477e-03 -1.32668042e+00
6.45729363e-01 1.49343848e-01 5.02725005e-01 -9.75769639e-01
-2.28740066e-01 -3.67151320e-01 -1.17582893e+00 -5.37373200e-02
6.57881439e-01 -6.65251791e-01 -6.91086531e-01 2.03190088e+00
4.64802206e-01 9.21373814e-02 -2.38133028e-01 4.26422179e-01
-3.80169526e-02 7.40032196e-01 -3.47101480e-01 -6.29891574e-01
1.23660433e+00 -3.25703382e-01 -5.96354008e-01 -4.82070476e-01
5.73618650e-01 -6.95323706e-01 9.52883124e-01 3.04756373e-01
-1.41183829e+00 -7.87667558e-02 -7.48566568e-01 1.75393760e-01
1.54883012e-01 8.78695771e-03 6.07283771e-01 7.64122725e-01
-8.69040012e-01 7.47501194e-01 -1.13482714e+00 2.01447323e-01
2.46689886e-01 6.76138878e-01 3.16813961e-03 -2.26174295e-01
-7.51898289e-01 4.03681278e-01 -9.79578421e-02 3.56868386e-01
-7.73286760e-01 -8.97534907e-01 -8.42482030e-01 2.68395156e-01
5.39357483e-01 -6.10728741e-01 1.12349081e+00 2.56138854e-02
-1.71079803e+00 1.10800425e-02 -5.87526143e-01 -3.63756835e-01
4.17617172e-01 -3.28258812e-01 3.01698267e-01 2.93986630e-02
-5.34232967e-02 7.12045133e-02 9.95863914e-01 -7.69846559e-01
-5.09918749e-01 -6.17144525e-01 -3.05219978e-01 1.49879485e-01
-2.00920880e-01 -2.08794266e-01 -7.07381427e-01 -3.78240705e-01
6.75024092e-01 -1.12546444e+00 -6.43665016e-01 1.56604707e-01
1.45113124e-02 1.35467425e-02 3.68100077e-01 -4.63574469e-01
1.15171039e+00 -2.29804039e+00 6.48860857e-02 4.16767806e-01
7.00403675e-02 -3.14611048e-01 1.66643411e-01 -2.02198662e-02
5.01441896e-01 1.32788373e-02 -5.76218426e-01 -6.12547934e-01
2.75375843e-01 2.13362113e-01 -1.46518677e-01 6.08539581e-01
-2.18492314e-01 2.66635448e-01 -7.25451648e-01 -3.78983110e-01
-3.29181030e-02 5.28788447e-01 -8.92433822e-01 -3.27893756e-02
-3.66766192e-02 -5.59557881e-03 -3.03748190e-01 2.10405409e-01
9.00498390e-01 -3.86610091e-01 2.31797591e-01 4.95151132e-02
4.16054279e-02 -9.35037993e-03 -2.02470112e+00 1.11644125e+00
-6.04029357e-01 4.24282670e-01 6.05036795e-01 -9.94442940e-01
7.59352744e-01 2.66437475e-02 6.58566773e-01 -9.92053673e-02
1.14339590e-01 3.87196481e-01 -1.91783458e-01 -2.20711268e-02
2.74459064e-01 -4.87804949e-01 -2.70041853e-01 6.89930499e-01
-6.30133674e-02 -3.95001620e-01 5.19537807e-01 1.33118004e-01
1.01860118e+00 -2.23523885e-01 -7.48146996e-02 -1.53060436e-01
3.48664761e-01 -4.93006259e-01 6.65659428e-01 1.07288039e+00
1.17605269e-01 2.23570526e-01 8.17333937e-01 -4.89930175e-02
-5.98936558e-01 -1.30461991e+00 -1.41943872e-01 1.14453590e+00
-2.76161125e-03 -3.94969076e-01 -6.00100398e-01 -1.42032206e-01
6.92666546e-02 5.04963756e-01 -4.27323550e-01 2.17345785e-02
-6.46372259e-01 -1.30745649e+00 1.78099543e-01 4.73182648e-01
5.05305588e-01 -1.89409316e-01 -2.33758315e-01 3.07795078e-01
1.27805024e-01 -9.30009544e-01 -7.42353559e-01 4.20357287e-01
-1.09991419e+00 -8.28456044e-01 -6.19829953e-01 -3.73690307e-01
7.87270308e-01 5.58253527e-02 4.46933210e-01 -4.38442975e-01
-3.84995863e-02 6.14431024e-01 2.75120854e-01 -5.73919415e-01
-9.62485373e-03 -2.97993571e-01 3.38419378e-01 -1.05813434e-02
-1.03313453e-01 -8.16589057e-01 -3.34735692e-01 2.75610715e-01
-6.81112826e-01 -4.47167844e-01 5.21230698e-01 6.87323272e-01
1.01974154e+00 1.85891420e-01 -5.46247624e-02 -7.31543303e-01
6.11760437e-01 -1.32544681e-01 -1.34711015e+00 -3.89594845e-02
-6.30907118e-01 3.95076185e-01 4.46009040e-01 -7.37281442e-01
-1.04676509e+00 1.85435936e-01 -1.50992724e-04 -4.71471995e-01
5.06307840e-01 7.52788246e-01 2.28357818e-02 -1.28082588e-01
4.50240016e-01 3.04191351e-01 1.43865775e-02 -6.53405249e-01
4.46461260e-01 3.20759892e-01 4.07077342e-01 -9.03445244e-01
7.76077330e-01 5.77937424e-01 5.70413470e-01 -8.08917046e-01
-1.02603257e+00 -2.73029894e-01 -2.50921696e-01 2.69262463e-01
4.93193567e-01 -8.03841829e-01 -1.21789372e+00 4.11462337e-01
-8.27189803e-01 -3.07823062e-01 -3.38102579e-01 8.77234817e-01
-7.47915030e-01 5.35053611e-01 -6.38600945e-01 -1.38867474e+00
-3.31339300e-01 -1.05843687e+00 8.35987270e-01 -1.11595713e-01
8.92676041e-02 -8.96885872e-01 2.52072942e-02 1.36318266e-01
4.48429525e-01 -3.36781979e-01 9.55474555e-01 -2.54149497e-01
-6.69782460e-01 -4.75983113e-01 -3.36253256e-01 3.34096253e-01
-6.75558895e-02 -1.80091992e-01 -2.69066662e-01 -3.12931240e-01
4.77896482e-01 2.76970446e-01 6.44140303e-01 8.13724220e-01
1.04755652e+00 -6.22339487e-01 -2.96073079e-01 7.96766937e-01
1.29676151e+00 1.71702638e-01 2.50398576e-01 -2.50500411e-01
4.37847376e-01 1.14371002e-01 4.68750268e-01 9.38212812e-01
9.33354422e-02 4.50839132e-01 8.65051895e-02 5.24281800e-01
4.11920965e-01 8.74240771e-02 5.94159901e-01 1.03910315e+00
-1.80286571e-01 4.06974629e-02 -7.98512995e-01 2.27795750e-01
-1.87423575e+00 -9.09773529e-01 -1.31175384e-01 2.52905917e+00
9.42254424e-01 2.19071195e-01 -1.22146390e-01 -7.65673369e-02
4.92164522e-01 -1.48062557e-01 -7.16313362e-01 -1.75269037e-01
1.28853902e-01 5.05129457e-01 8.23307097e-01 9.61258590e-01
-9.92388189e-01 5.72055161e-01 6.75239420e+00 1.12123275e+00
-7.02834070e-01 1.48636401e-01 3.33666116e-01 -5.00943661e-01
-1.11228332e-01 1.06058098e-01 -1.26669252e+00 6.69896901e-01
1.07695472e+00 -3.58215243e-01 6.00874960e-01 1.05994701e+00
1.36688009e-01 -4.38299477e-01 -9.29468989e-01 1.40789247e+00
-5.08672260e-02 -1.10122263e+00 -2.91846544e-01 4.42983657e-01
7.16888964e-01 -3.32544111e-02 7.06506670e-02 3.29517186e-01
6.78019345e-01 -7.49982059e-01 3.75340015e-01 4.25547302e-01
3.91351998e-01 -8.38034451e-01 7.20656335e-01 4.05054390e-01
-1.30871117e+00 -1.44672468e-01 -6.72839522e-01 5.82372304e-03
5.55251539e-01 1.07619083e+00 -4.50716108e-01 -3.91720347e-02
7.86734402e-01 2.59265900e-01 7.16266502e-03 9.29196358e-01
1.79314017e-02 4.38766420e-01 -1.24710655e+00 -1.84716046e-01
9.87532362e-02 -8.61246467e-01 5.87759376e-01 9.61611927e-01
9.44238901e-01 2.97873080e-01 1.50744036e-01 5.87108612e-01
-2.78240424e-02 1.29043266e-01 -3.70765403e-02 -1.76527217e-01
4.87614751e-01 8.01320255e-01 -6.18027389e-01 -3.90596479e-01
-1.65204301e-01 6.28071308e-01 2.27275625e-01 5.22387445e-01
-7.80298531e-01 -5.09721935e-01 9.68272150e-01 7.62246400e-02
7.17804849e-01 -7.63012350e-01 -2.75618702e-01 -9.80771780e-01
1.52561203e-01 -5.78091621e-01 3.86760533e-01 -2.96074241e-01
-1.14533114e+00 3.58353294e-02 2.15580225e-01 -8.05213153e-01
-3.93381804e-01 -6.44950509e-01 -4.50766683e-02 8.59532356e-01
-8.24073315e-01 -3.97697538e-01 5.82069233e-02 5.36863863e-01
1.97882816e-01 1.81754678e-02 5.60056210e-01 2.40665495e-01
-7.42183805e-01 3.96938354e-01 7.64026761e-01 -2.02817738e-01
4.03423697e-01 -1.12335324e+00 -1.01566024e-01 6.29844666e-01
2.23963931e-02 7.38440394e-01 1.01380205e+00 -4.91244912e-01
-1.62980831e+00 -5.79928398e-01 4.72694218e-01 -1.56735018e-01
9.01969910e-01 -2.90127784e-01 -7.72790492e-01 6.41435921e-01
-4.87141490e-01 -4.82922345e-02 8.12257826e-01 3.90547961e-01
-1.83300346e-01 -4.05690759e-01 -9.32379782e-01 6.96363270e-01
8.71268094e-01 -3.63168150e-01 7.01857954e-02 4.81796324e-01
5.59097886e-01 -6.23706937e-01 -1.03354061e+00 2.04081938e-01
5.49143016e-01 -6.03123426e-01 9.96982515e-01 -1.43678188e-01
-1.21183336e-01 -2.76961386e-01 -5.71325839e-01 -9.51798379e-01
-2.19862144e-02 -9.79610085e-01 -5.11855423e-01 7.64367998e-01
5.26544929e-01 -8.31071198e-01 7.21217752e-01 8.79113436e-01
1.38703622e-02 -9.04343128e-01 -1.19184840e+00 -7.83688426e-01
9.30418298e-02 -1.05591249e+00 2.63588965e-01 2.70640880e-01
-3.55978817e-01 2.15924114e-01 -3.30744922e-01 3.75271231e-01
9.56853867e-01 7.44256079e-02 7.55038261e-01 -1.29021847e+00
-8.11999559e-01 -4.00459737e-01 -2.11773857e-01 -1.64778984e+00
-1.17032297e-01 -5.48034430e-01 3.15481901e-01 -1.28135633e+00
2.81842709e-01 -6.31354570e-01 -1.61138270e-02 5.89983612e-02
-2.35481843e-01 -9.82153639e-02 1.81783676e-01 5.52102923e-02
-4.48201776e-01 5.96335232e-01 9.63004112e-01 -2.98632192e-03
-3.56985480e-01 2.38186255e-01 -5.61295092e-01 1.12817347e+00
4.14933473e-01 -8.07017684e-01 -3.34211588e-01 -4.82317626e-01
6.22326076e-01 1.83882460e-01 -3.84028628e-02 -9.75619435e-01
3.47876161e-01 -2.66902357e-01 2.84061998e-01 -6.47969961e-01
8.90865743e-01 -3.82936597e-01 6.98500127e-02 3.92664790e-01
-3.40056531e-02 -8.64697173e-02 2.25837320e-01 9.69568551e-01
2.71793772e-02 -6.36585891e-01 7.53068388e-01 -6.87765926e-02
-6.96166372e-03 6.30684257e-01 -7.08619237e-01 7.90314004e-02
7.43396699e-01 3.72503214e-02 4.81160022e-02 -7.77877033e-01
-9.48610365e-01 2.10195988e-01 1.38599016e-02 -4.80208993e-01
5.76787353e-01 -9.96910393e-01 -4.29547220e-01 1.77126646e-01
-8.05983961e-01 4.58736420e-01 5.26557565e-01 1.09849036e+00
-4.51666594e-01 1.45196021e-01 5.48455179e-01 -6.91499472e-01
-1.13968360e+00 3.20088625e-01 2.24718258e-01 -4.59655672e-01
-1.07524740e-02 1.24454308e+00 6.15430325e-02 -1.79539695e-01
3.51375937e-01 -3.23542625e-01 5.27725160e-01 3.74169946e-02
6.02274418e-01 7.67037988e-01 -3.41680974e-01 -1.54534891e-01
-2.46831194e-01 9.28568125e-01 5.84794991e-02 -6.88644052e-01
1.33981705e+00 -2.69294262e-01 -3.86654258e-01 4.73758668e-01
1.25981832e+00 2.08265662e-01 -1.38517153e+00 -2.64563560e-01
-1.40839294e-01 -3.44654024e-01 2.94780992e-02 -2.94373706e-02
-9.11736250e-01 1.01510572e+00 6.10165715e-01 -6.60240576e-02
9.67883766e-01 7.65571520e-02 4.50154036e-01 8.94232810e-01
5.09276927e-01 -1.14410853e+00 7.29271919e-02 8.67803574e-01
6.36218727e-01 -8.78144562e-01 3.75087738e-01 -7.40892112e-01
-2.26899192e-01 8.90511096e-01 2.94830412e-01 -1.60071179e-01
1.29206955e+00 4.20506060e-01 -4.07333463e-01 9.93075445e-02
-6.96528614e-01 -4.14845571e-02 9.88871530e-02 3.83943856e-01
2.33021200e-01 -5.06014861e-02 -1.95004806e-01 6.07541502e-01
-2.92568445e-01 -1.67142019e-01 3.93730968e-01 8.55824769e-01
-7.48744369e-01 -1.25045562e+00 -3.94913226e-01 6.80881739e-01
-5.61441898e-01 -8.84107873e-02 3.57983440e-01 3.22251141e-01
-2.60564893e-01 7.62564480e-01 2.21929446e-01 1.50767863e-01
1.61108166e-01 2.06810921e-01 6.51240349e-01 -5.76707184e-01
9.70890746e-02 3.94629776e-01 -1.32155478e-01 -4.27203059e-01
2.56387121e-03 -7.77974844e-01 -1.15372121e+00 -5.64591050e-01
-3.28910023e-01 2.45655194e-01 8.69523346e-01 1.01616192e+00
3.09914231e-01 -3.33424471e-02 3.83520424e-01 -8.01719427e-01
-1.00341308e+00 -9.45555747e-01 -8.13488603e-01 -5.42526320e-02
-4.24242802e-02 -6.81330562e-01 -6.77842975e-01 -1.21873654e-01]
|
[6.709073543548584, 4.24590539932251]
|
1ad072ec-2332-4f3a-ad9d-05132f584015
|
learning-roles-with-emergent-social-value
|
2301.13812
| null |
https://arxiv.org/abs/2301.13812v1
|
https://arxiv.org/pdf/2301.13812v1.pdf
|
Learning Roles with Emergent Social Value Orientations
|
Social dilemmas can be considered situations where individual rationality leads to collective irrationality. The multi-agent reinforcement learning community has leveraged ideas from social science, such as social value orientations (SVO), to solve social dilemmas in complex cooperative tasks. In this paper, by first introducing the typical "division of labor or roles" mechanism in human society, we provide a promising solution for intertemporal social dilemmas (ISD) with SVOs. A novel learning framework, called Learning Roles with Emergent SVOs (RESVO), is proposed to transform the learning of roles into the social value orientation emergence, which is symmetrically solved by endowing agents with altruism to share rewards with other agents. An SVO-based role embedding space is then constructed by individual conditioning policies on roles with a novel rank regularizer and mutual information maximizer. Experiments show that RESVO achieves a stable division of labor and cooperation in ISDs with different complexity.
|
['Hongyuan Zha', 'Jingyi Lu', 'Bo Jin', 'Xiangfeng Wang', 'Wenhao Li']
|
2023-01-31
| null | null | null | null |
['role-embedding']
|
['graphs']
|
[ 1.28213153e-03 8.16022754e-01 -1.60394654e-01 -9.14548784e-02
2.86546856e-01 -2.67630041e-01 5.16014874e-01 1.33230514e-03
-7.01080263e-01 1.01786482e+00 5.23419261e-01 4.74152006e-02
-7.85172462e-01 -6.74997866e-01 1.15396798e-01 -1.16908967e+00
-4.20124799e-01 3.64204496e-01 -4.35095161e-01 -9.66712177e-01
1.33391485e-01 -2.08753452e-01 -1.34089303e+00 -2.95189619e-01
1.03519213e+00 3.81669343e-01 1.07120380e-01 3.89352709e-01
2.79372960e-01 1.44929743e+00 -7.04829335e-01 -2.34346136e-01
6.26421094e-01 -6.13401115e-01 -9.08771932e-01 2.73596257e-01
-1.12926304e+00 -3.08183700e-01 -3.39631617e-01 7.77690113e-01
6.22586191e-01 5.31785727e-01 1.06920195e+00 -2.07654500e+00
-1.18399489e+00 9.98775303e-01 -7.73298919e-01 -1.89500585e-01
4.88585949e-01 2.10674219e-02 1.62112236e+00 -4.11718309e-01
6.19380355e-01 1.74445033e+00 2.85457402e-01 1.16584468e+00
-1.34541917e+00 -4.33751911e-01 3.60670984e-02 -1.08042896e-01
-7.45886266e-01 9.85264406e-02 7.99125969e-01 -4.58985686e-01
6.61199033e-01 -3.31611298e-02 9.25779760e-01 9.52475786e-01
2.39774868e-01 8.66577446e-01 1.26721478e+00 -4.30513531e-01
4.07837242e-01 -2.66374081e-01 -1.21329343e-02 4.17076945e-01
4.42898631e-01 2.59595484e-01 -5.66962183e-01 -2.67332882e-01
8.61346245e-01 -6.24080095e-03 1.56625450e-01 -6.86349273e-01
-1.24966073e+00 1.14428508e+00 4.73748356e-01 3.87053579e-01
-6.52309358e-01 3.03991973e-01 1.05162181e-01 1.16715097e+00
4.17967916e-01 1.13308108e+00 -2.66548932e-01 6.61809072e-02
2.80137211e-01 3.05896819e-01 6.39593124e-01 4.85218078e-01
9.90164399e-01 6.39451072e-02 -2.32072726e-01 8.31718385e-01
6.63638949e-01 1.85918078e-01 4.14143801e-01 -1.73220539e+00
2.96787210e-02 9.94733810e-01 4.44805533e-01 -8.63092482e-01
-8.34421694e-01 -1.60753399e-01 -8.35417449e-01 5.00997722e-01
3.29183906e-01 -5.82650959e-01 3.00668534e-02 2.17744827e+00
5.32620430e-01 -3.34426552e-01 7.98529267e-01 7.78036118e-01
4.90174681e-01 5.35832882e-01 -6.66071028e-02 -5.21224201e-01
1.11773610e+00 -7.59491861e-01 -7.70058990e-01 2.88613979e-02
9.79587197e-01 -1.30841101e-03 9.63407695e-01 -4.66486858e-03
-7.98325062e-01 4.04670648e-02 -8.74108613e-01 2.68680900e-01
5.34803830e-02 -7.12931037e-01 8.23896110e-01 5.07396400e-01
-1.48140407e+00 7.06643581e-01 -3.71681839e-01 -3.13992947e-01
5.86063147e-01 6.16307259e-01 -2.32815146e-01 5.36014259e-01
-1.48643410e+00 6.62017345e-01 4.26740833e-02 -2.75725007e-01
-1.12697339e+00 -3.87203425e-01 -9.48101878e-01 -3.34051512e-02
9.85517442e-01 -9.94859040e-01 1.09305608e+00 -1.26293242e+00
-1.90236259e+00 9.08867300e-01 5.96995234e-01 -3.05567861e-01
4.14641827e-01 2.19144866e-01 1.61178902e-01 3.69222872e-02
6.56845272e-01 5.43094456e-01 9.57861841e-01 -1.42993104e+00
-4.13257867e-01 -3.33088428e-01 7.10665286e-01 9.55687225e-01
-5.33635497e-01 2.18512744e-01 9.68461514e-01 -6.58210754e-01
-1.66577443e-01 -1.11060834e+00 -6.58921361e-01 -2.27640316e-01
-1.22558504e-01 -1.00853646e+00 1.46637827e-01 1.60025239e-01
8.60899568e-01 -2.07591677e+00 7.82974899e-01 4.74823937e-02
1.07571518e+00 -4.77319568e-01 -5.45472920e-01 4.45051730e-01
1.02354601e-01 2.08947390e-01 -1.80250585e-01 -1.73227072e-01
5.58709145e-01 4.50985163e-01 1.83437482e-01 5.13690114e-01
4.62156869e-02 1.12093568e+00 -1.34628534e+00 -2.70531654e-01
-4.23643589e-01 5.99591993e-04 -8.92906725e-01 4.88017678e-01
6.43977374e-02 5.68680584e-01 -7.65445828e-01 3.37617069e-01
2.82771498e-01 -4.09265190e-01 7.66727805e-01 7.89891839e-01
6.35778680e-02 1.45250916e-01 -1.18298924e+00 1.06601250e+00
-8.99033695e-02 1.74875394e-01 3.72575283e-01 -1.07102168e+00
8.72663379e-01 3.78122240e-01 8.74107897e-01 -5.20209968e-01
4.68722790e-01 9.59787890e-02 5.02148449e-01 -2.59078503e-01
5.72578907e-01 -3.87630343e-01 -3.87467295e-01 1.42966425e+00
1.12735435e-01 -2.06765145e-01 2.17074007e-01 5.15421689e-01
1.09281886e+00 4.25012633e-02 5.82980812e-01 -7.26646662e-01
4.65986043e-01 -3.09033036e-01 9.06954885e-01 7.79171646e-01
-6.38501406e-01 7.98280258e-03 1.24475467e+00 -2.68407106e-01
-9.01153862e-01 -8.90789151e-01 5.30908704e-01 1.51704419e+00
5.55155814e-01 -1.55744910e-01 -5.94422340e-01 -6.15087211e-01
3.62742871e-01 -2.92140176e-03 -8.29135358e-01 -4.07267630e-01
-5.04042566e-01 -9.67610538e-01 3.39724392e-01 1.32246092e-01
4.77536768e-01 -1.56435585e+00 -7.78277159e-01 1.44440338e-01
-6.12134766e-03 -5.38931787e-01 -4.84213769e-01 2.53036171e-01
-5.34246206e-01 -1.16751087e+00 -1.04824293e+00 -7.59361804e-01
7.68517613e-01 4.18790102e-01 8.59810591e-01 2.62329519e-01
1.52332857e-02 9.92548466e-01 -6.74216747e-01 -2.93386042e-01
-6.54867649e-01 -1.53251290e-01 9.63825524e-01 4.25885350e-01
6.33785650e-02 -8.33274126e-01 -6.87522948e-01 4.43225652e-01
-8.99217665e-01 -1.61604479e-01 3.94508868e-01 1.04831779e+00
-2.93237001e-01 -6.40363693e-02 1.26083565e+00 -6.04909062e-01
1.58642292e+00 -7.51273572e-01 -3.42450291e-01 2.01248556e-01
-9.18362617e-01 4.04028535e-01 4.77174789e-01 -3.96787167e-01
-1.14208889e+00 -6.78416431e-01 7.62180150e-01 3.43574911e-01
4.41858083e-01 4.96151328e-01 4.89271581e-02 1.37060463e-01
8.31627905e-01 -4.75642718e-02 7.18970358e-01 -8.72619003e-02
2.42754951e-01 8.27123106e-01 -4.03058171e-01 -7.88225889e-01
8.44553947e-01 2.92634159e-01 1.32406875e-01 -8.80576193e-01
-3.79690021e-01 -7.47149736e-02 -2.08158165e-01 -4.36615944e-01
9.66671109e-01 -8.31868768e-01 -1.56251788e+00 9.43431258e-01
-1.01452124e+00 -6.21797204e-01 -6.53398871e-01 5.14456272e-01
-9.56430674e-01 4.61779535e-01 -6.13761902e-01 -1.18527424e+00
2.39832371e-01 -6.03206277e-01 3.55767250e-01 4.33666885e-01
-1.64852932e-03 -9.65524077e-01 1.92842707e-01 4.92432237e-01
4.41525370e-01 5.91644011e-02 7.90704489e-01 -4.76457894e-01
-3.50659639e-01 6.02244556e-01 7.58855864e-02 3.43096852e-01
3.98478627e-01 -5.92936277e-01 -4.00929987e-01 -4.03110832e-01
2.47392401e-01 -8.65420461e-01 4.63550776e-01 4.57175016e-01
1.48403868e-01 -4.42261219e-01 7.65652731e-02 2.72066537e-02
8.16917360e-01 4.46919113e-01 3.33463490e-01 2.98931897e-01
3.04845273e-01 1.30004013e+00 6.16635799e-01 1.09364378e+00
9.63195622e-01 2.82819808e-01 4.77974415e-01 7.15911686e-02
5.61024249e-01 -9.94375572e-02 6.34267449e-01 6.06414258e-01
-6.51951253e-01 2.61022113e-02 -6.78913414e-01 1.41066894e-01
-2.23081827e+00 -9.53246534e-01 5.03248572e-01 1.79638040e+00
8.82644832e-01 -3.52659076e-01 5.73021591e-01 4.76431809e-02
8.39716733e-01 4.69732076e-01 -7.48627186e-01 -2.43795201e-01
-4.09849554e-01 -3.57751191e-01 9.81415510e-02 7.78313696e-01
-7.57890344e-01 1.00509191e+00 6.04328346e+00 3.17542821e-01
-2.28515193e-01 2.61965871e-01 7.03921139e-01 4.92190346e-02
-6.05183244e-01 3.84849422e-02 -2.31386408e-01 1.13036491e-01
1.71622679e-01 -7.05602288e-01 8.70856702e-01 5.37383497e-01
4.44590420e-01 -1.38263881e-01 -8.81797850e-01 9.29514289e-01
-1.63477823e-01 -8.55716765e-01 -3.99566889e-01 3.34107399e-01
9.60221052e-01 -6.05212092e-01 5.43964915e-02 3.84939909e-01
1.41020596e+00 -1.19494867e+00 6.20123029e-01 3.70691940e-02
5.57365716e-01 -5.81022263e-01 4.29316819e-01 4.62065011e-01
-9.17250633e-01 -8.00203323e-01 -4.73410845e-01 -9.51691449e-01
-1.83491021e-01 -9.39896405e-02 -6.19532108e-01 3.84247720e-01
5.14784038e-01 1.12475634e+00 -1.80902302e-01 2.48906687e-02
-3.99645776e-01 -1.51567504e-01 1.46857589e-01 -5.87827981e-01
3.11062276e-01 -6.23765945e-01 6.70708716e-01 -8.00646842e-02
-5.30166179e-02 3.59367043e-01 1.97031900e-01 9.50832069e-01
-1.64454401e-01 1.65868714e-01 -8.91631603e-01 -1.98306486e-01
3.35565090e-01 1.16740227e+00 -6.98560894e-01 1.57531753e-01
-7.88207576e-02 7.57693708e-01 4.16202337e-01 4.85356271e-01
-5.02698064e-01 -2.97594875e-01 1.09167850e+00 -4.78529111e-02
-2.36362785e-01 -9.82575268e-02 2.60064125e-01 -1.44434130e+00
-2.66310573e-01 -9.99074578e-01 5.02549231e-01 -5.16980350e-01
-1.39486980e+00 3.45914125e-01 4.12896201e-02 -1.15231133e+00
-1.67924628e-01 -4.82274234e-01 -4.03757304e-01 2.19651192e-01
-1.28639114e+00 -8.16395760e-01 3.03284585e-01 7.27780938e-01
1.14691645e-01 -6.72132969e-01 4.15725619e-01 -2.48306870e-01
-3.77665341e-01 3.53952348e-01 8.83065611e-02 -1.14478506e-01
3.43869120e-01 -1.46840143e+00 -2.06220254e-01 1.01990022e-01
-4.45567966e-01 4.31741327e-01 5.39697647e-01 -4.55274999e-01
-1.21017849e+00 -3.83846641e-01 7.88111567e-01 -3.79101187e-01
9.06258404e-01 -4.64952141e-01 -3.36327553e-01 3.64868522e-01
4.34211493e-01 -3.31891716e-01 7.71516681e-01 1.97016180e-01
-1.22639254e-01 1.22153964e-02 -1.21631074e+00 1.02659523e+00
1.77259123e+00 -3.51866305e-01 -7.58606613e-01 2.46997908e-01
1.22303510e+00 4.73683238e-01 -6.01389289e-01 3.03035714e-02
3.48438919e-01 -8.94515932e-01 8.38721454e-01 -9.92153347e-01
7.05080569e-01 1.81126874e-02 -8.11485052e-02 -1.77218616e+00
-4.61263984e-01 -1.43657064e+00 4.05615568e-01 7.52155125e-01
1.40067726e-01 -1.22513211e+00 7.51015902e-01 4.23236370e-01
2.19600514e-01 -5.56973875e-01 -1.23324716e+00 -8.66007566e-01
3.81027758e-01 3.85853320e-01 6.49928868e-01 1.28390348e+00
5.72716892e-01 7.10450709e-01 -7.66212225e-01 -5.28098702e-01
7.91882932e-01 -3.44812691e-01 7.03321338e-01 -1.61233556e+00
-5.53323269e-01 -5.02670467e-01 -1.93819851e-01 -6.87152803e-01
5.47823787e-01 -9.99939680e-01 -1.58662885e-01 -1.36858678e+00
1.74392968e-01 -5.49776375e-01 -3.76901925e-01 3.24255705e-01
-9.82510671e-02 -1.59838364e-01 5.36230445e-01 3.45809221e-01
-1.12257671e+00 1.13660395e+00 1.65800607e+00 -1.22529708e-01
-7.75073647e-01 -3.71493399e-02 -1.40420985e+00 7.89689600e-01
8.43670547e-01 -7.05896676e-01 -6.44501269e-01 3.29826586e-02
9.13559914e-01 4.93491799e-01 8.64923000e-02 -3.07788521e-01
2.45955065e-02 -7.18550205e-01 -5.49143970e-01 6.48215532e-01
1.71824563e-02 -5.16527355e-01 -2.50463396e-01 9.25836921e-01
-8.65787864e-01 7.22789615e-02 -8.37354362e-01 6.80560410e-01
1.12112425e-01 -2.68631577e-01 6.36301517e-01 -3.30980599e-01
-2.39199117e-01 -2.86851823e-02 -1.01303506e+00 4.23428893e-01
1.31577682e+00 -1.07951000e-01 -3.48739535e-01 -9.26896870e-01
-7.60741234e-01 8.92867088e-01 3.33963245e-01 3.59651238e-01
6.85978711e-01 -1.36246860e+00 -1.05179703e+00 -4.66700085e-02
8.18562731e-02 -3.17514449e-01 3.62995416e-01 5.79211295e-01
-2.35755041e-01 -1.78420752e-01 -6.23864412e-01 -7.73295620e-03
-1.12939847e+00 2.97937095e-01 4.48947549e-01 -6.57584667e-01
-8.21308419e-02 7.79354155e-01 5.38606286e-01 -1.10291624e+00
1.60888047e-03 1.37069464e-01 -7.89508462e-01 5.80035329e-01
1.21895030e-01 6.38452768e-01 -9.45891380e-01 -6.76663518e-01
-2.07617775e-01 3.23560685e-01 1.88424379e-01 -4.67696488e-01
1.65383279e+00 -3.48045141e-01 -6.37019753e-01 1.82178453e-01
6.74315691e-01 -4.00613189e-01 -1.35555148e+00 -3.67969960e-01
1.58692777e-01 -3.74883473e-01 -6.24002814e-01 -2.83317447e-01
-6.65697873e-01 3.15914184e-01 2.10532233e-01 8.06569993e-01
7.13708758e-01 2.34779105e-01 2.26189997e-02 8.61030340e-01
7.21474409e-01 -1.61305058e+00 1.20438492e+00 7.02068388e-01
9.75825727e-01 -1.42585993e+00 -1.30378529e-01 -1.90239981e-01
-1.35934901e+00 4.86121416e-01 7.06131637e-01 -3.87350291e-01
9.85650122e-01 -1.80910230e-01 7.52736628e-02 -2.45119572e-01
-9.78983760e-01 -3.40926409e-01 -4.06834513e-01 9.49179292e-01
3.48557457e-02 4.15333807e-01 -8.50912035e-01 8.65593374e-01
2.69467998e-02 -3.14186424e-01 1.20513117e+00 1.15490961e+00
-5.06654263e-01 -1.48765171e+00 -2.54402310e-01 4.40682083e-01
-3.07791322e-01 9.73164365e-02 -7.18139827e-01 4.94040996e-01
-1.04853109e-01 1.25619960e+00 5.82514592e-02 -4.99823421e-01
3.89173217e-02 -3.21440428e-01 3.60786349e-01 -8.30477834e-01
-7.70227551e-01 -1.58488050e-01 -1.57258883e-01 -3.79183888e-01
-8.98789465e-01 -6.50467038e-01 -1.24940848e+00 -3.44412148e-01
1.15504071e-01 3.46313685e-01 -2.13702843e-01 6.99921310e-01
4.41350877e-01 3.17126691e-01 1.33177769e+00 -6.50048971e-01
-1.23381972e+00 -7.69706309e-01 -1.31065238e+00 6.20489597e-01
3.86644095e-01 -1.03134143e+00 -7.41817594e-01 -5.71182907e-01]
|
[3.7948689460754395, 2.2220165729522705]
|
2c86dda6-d5b0-4ed0-a5f2-b896a8b758a5
|
multiplication-fusion-of-sparse-and
|
2001.07090
| null |
https://arxiv.org/abs/2001.07090v1
|
https://arxiv.org/pdf/2001.07090v1.pdf
|
Multiplication fusion of sparse and collaborative-competitive representation for image classification
|
Representation based classification methods have become a hot research topic during the past few years, and the two most prominent approaches are sparse representation based classification (SRC) and collaborative representation based classification (CRC). CRC reveals that it is the collaborative representation rather than the sparsity that makes SRC successful. Nevertheless, the dense representation of CRC may not be discriminative which will degrade its performance for classification tasks. To alleviate this problem to some extent, we propose a new method called sparse and collaborative-competitive representation based classification (SCCRC) for image classification. Firstly, the coefficients of the test sample are obtained by SRC and CCRC, respectively. Then the fused coefficient is derived by multiplying the coefficients of SRC and CCRC. Finally, the test sample is designated to the class that has the minimum residual. Experimental results on several benchmark databases demonstrate the efficacy of our proposed SCCRC. The source code of SCCRC is accessible at https://github.com/li-zi-qi/SCCRC.
|
['He-Feng Yin', 'Xiao-Jun Wu', 'Zi-Qi Li', 'Jun Sun']
|
2020-01-20
| null | null | null | null |
['sparse-representation-based-classification']
|
['computer-vision']
|
[ 2.57627100e-01 -4.25903738e-01 -3.80057484e-01 -2.32464880e-01
-7.37163007e-01 1.05552776e-02 2.75951087e-01 -3.47232670e-02
6.42806618e-03 4.61018562e-01 4.92648482e-01 5.14929891e-02
-3.61203969e-01 -5.68348467e-01 -6.46549389e-02 -1.07533193e+00
4.67279166e-01 -3.02157402e-01 -1.90077368e-02 -1.47775292e-01
4.41180795e-01 2.10361853e-01 -1.59328711e+00 4.39032763e-01
1.20243931e+00 1.26938915e+00 2.93924898e-01 -2.58663408e-02
-3.41860158e-03 8.75034630e-01 -5.34479856e-01 6.42291978e-02
1.64233282e-01 -6.38559818e-01 -2.23600134e-01 -4.77587841e-02
-2.36571297e-01 1.44842222e-01 -4.40780938e-01 1.13629055e+00
3.13305497e-01 3.32156420e-01 6.76816881e-01 -1.04641259e+00
-7.57978082e-01 4.39143062e-01 -8.20317388e-01 3.14623863e-01
3.02597970e-01 -3.50260258e-01 8.38079572e-01 -1.22445226e+00
3.04556787e-01 1.07630169e+00 3.88096869e-01 2.38519579e-01
-7.93210685e-01 -9.17272627e-01 2.85313666e-01 3.25337112e-01
-1.85512531e+00 -4.03546691e-01 1.01451528e+00 -3.61922771e-01
2.60047346e-01 4.26752716e-01 4.88449901e-01 7.30937898e-01
-2.50999629e-02 9.94211972e-01 1.19611931e+00 -3.95275176e-01
1.42737299e-01 9.36232954e-02 2.22852975e-01 5.48123658e-01
3.26800376e-01 -1.07347935e-01 -3.90943766e-01 -3.22273582e-01
6.98230863e-01 7.33124435e-01 -7.17701375e-01 -1.90946743e-01
-1.08703518e+00 1.04183030e+00 7.28162408e-01 6.00928962e-01
-4.45186555e-01 -2.58218288e-01 3.63484234e-01 1.29766658e-01
4.65008736e-01 2.83175502e-02 8.52069817e-03 1.00001119e-01
-8.26417983e-01 -2.07834825e-01 4.20623749e-01 7.12959349e-01
5.89928329e-01 1.23616330e-01 -2.35440340e-02 1.27287769e+00
3.31296384e-01 3.26088876e-01 9.04158890e-01 -6.23313129e-01
5.46019435e-01 7.47133613e-01 -3.38885337e-01 -1.36383355e+00
1.56621516e-01 -8.56747508e-01 -1.20551658e+00 -4.23648596e-01
-1.79315209e-01 5.24516255e-02 -5.02886057e-01 1.05824769e+00
2.07000539e-01 5.94094813e-01 2.79768407e-01 1.10990727e+00
9.90785718e-01 7.82894433e-01 3.16608734e-02 -2.34026328e-01
1.11540008e+00 -7.83553421e-01 -4.95498568e-01 -2.02322956e-02
5.42827368e-01 -7.77365386e-01 4.58354592e-01 3.16564947e-01
-5.30984879e-01 -6.26060665e-01 -9.59234536e-01 3.78276199e-01
-1.74789354e-02 6.42028987e-01 7.59442687e-01 3.04691494e-01
-4.99047667e-01 2.05410823e-01 -5.90694904e-01 1.30956899e-02
6.06738985e-01 1.19286358e-01 -3.26421261e-01 -6.53240621e-01
-1.00348330e+00 3.45840603e-01 2.26808310e-01 3.17453325e-01
-5.47442019e-01 -5.34500360e-01 -7.65540481e-01 3.28397937e-02
3.19568664e-01 -8.96124765e-02 8.14510643e-01 -1.15072703e+00
-1.10333955e+00 3.34360987e-01 -3.80942047e-01 -1.63697809e-01
5.60971461e-02 -9.70577896e-02 -5.08511305e-01 2.22275913e-01
2.60365129e-01 1.36080369e-01 8.69791985e-01 -1.24668896e+00
-6.89795494e-01 -2.91585803e-01 -2.79307246e-01 2.78482020e-01
-4.53012615e-01 1.86482184e-02 -4.69961137e-01 -8.25796962e-01
6.94448054e-01 -7.11761296e-01 -3.52968991e-01 -2.30532289e-01
-2.17474073e-01 -3.20295006e-01 9.85609353e-01 -6.23152614e-01
1.11541355e+00 -2.67984819e+00 2.63011664e-01 4.02180940e-01
1.25007600e-01 3.88663232e-01 -1.15906969e-01 3.75870496e-01
-3.75812203e-01 -5.47180548e-02 -2.22102955e-01 -3.75089794e-02
-7.07201600e-01 -1.01814471e-01 -3.41803104e-01 5.61960220e-01
1.28324986e-01 5.88663280e-01 -8.16299736e-01 -4.56878752e-01
1.86612099e-01 6.69979572e-01 -2.78186738e-01 9.92953777e-02
3.32662433e-01 5.70298553e-01 -9.10595953e-01 9.15895998e-01
7.22798526e-01 -3.68582606e-01 2.38338485e-01 -4.13385719e-01
-7.11187422e-02 3.44463959e-02 -1.23833847e+00 1.36562240e+00
-2.75951594e-01 2.46345043e-01 -7.35788792e-02 -1.39598608e+00
1.42927063e+00 2.68069506e-01 5.89220166e-01 -6.45533860e-01
1.31243870e-01 3.25925529e-01 -2.38095801e-02 -2.14392915e-01
3.31816643e-01 8.94348174e-02 2.97073908e-02 1.68342397e-01
-3.83512765e-01 3.57276738e-01 5.30487113e-02 3.97933632e-01
6.78717136e-01 -1.26835570e-01 4.96171415e-01 7.15908827e-03
9.12612319e-01 -4.89768088e-02 1.07876182e+00 1.75528303e-01
-9.43134874e-02 7.65237689e-01 2.29098186e-01 -1.76749155e-01
-4.28617120e-01 -7.81849027e-01 -7.08045810e-02 5.93082726e-01
3.91890973e-01 -4.19794232e-01 -1.50905773e-01 -5.34334302e-01
-1.64633431e-03 3.18268687e-01 -4.12101328e-01 -4.42190975e-01
-3.93790692e-01 -5.33149242e-01 -3.36429961e-02 4.69200969e-01
7.27606952e-01 -7.16891766e-01 -3.45493434e-04 1.78354144e-01
-3.28254282e-01 -6.86689079e-01 -4.33533400e-01 -1.68049082e-01
-9.37847614e-01 -1.12583268e+00 -9.87931609e-01 -9.19431329e-01
1.06667745e+00 1.13473380e+00 4.11955148e-01 5.60728788e-01
-1.45377249e-01 3.25602805e-03 -1.01036870e+00 -1.92727074e-02
-4.47690450e-02 -1.20831192e-01 -8.55585337e-02 4.72225487e-01
1.51370451e-01 -4.57420170e-01 -8.06627095e-01 3.97442728e-01
-5.88217437e-01 7.62867630e-02 8.05933714e-01 7.94626892e-01
8.22455168e-01 3.30579966e-01 7.55214572e-01 -8.75934064e-01
5.34267724e-01 -8.32806766e-01 -3.33880901e-01 -5.18802926e-02
-4.41428334e-01 -5.57377994e-01 8.36021483e-01 -4.68648583e-01
-9.39364016e-01 2.18808725e-01 6.62608594e-02 -7.27136135e-01
8.68049711e-02 9.51022029e-01 -3.91729968e-03 -1.20606255e-02
3.08708459e-01 7.91774094e-01 6.02972396e-02 -5.81759691e-01
-1.00176513e-01 9.19734061e-01 1.76379859e-01 -2.25508735e-01
7.40896940e-01 2.11231068e-01 -2.35746339e-01 -8.00646126e-01
-8.70161772e-01 -8.25298429e-01 -1.90202326e-01 -1.60666659e-01
4.83274013e-01 -1.40061772e+00 -1.68638170e-01 1.38014093e-01
-6.77048564e-01 3.21620136e-01 -9.92468968e-02 8.81425858e-01
1.74351949e-02 4.88400519e-01 -2.94456869e-01 -8.94816220e-01
-4.24131870e-01 -1.11244798e+00 7.73026705e-01 6.66049838e-01
8.28564614e-02 -5.55819333e-01 -2.26866841e-01 6.03483796e-01
3.50896806e-01 2.00787023e-01 7.82659292e-01 -5.38634002e-01
-5.39737701e-01 -4.87509519e-01 -4.02552068e-01 4.02558893e-01
3.41753751e-01 -5.74720502e-02 -6.74912393e-01 -4.66229379e-01
-3.97303980e-03 -6.78498745e-02 9.71128702e-01 2.82922506e-01
1.48710787e+00 -1.60676152e-01 -4.88594472e-01 6.05340600e-01
1.68516505e+00 4.77849931e-01 8.00253034e-01 8.24173912e-02
7.42179751e-01 2.21429735e-01 9.83740449e-01 4.69190419e-01
2.66713768e-01 6.29437864e-01 1.20165952e-01 -1.43790826e-01
-1.10163718e-01 -2.07114890e-01 3.87683928e-01 1.26912069e+00
-2.61714160e-01 1.33798793e-01 -7.44600415e-01 3.55169982e-01
-1.92131078e+00 -9.78707373e-01 -1.67043507e-01 2.09566569e+00
4.93471146e-01 -2.44489029e-01 -1.80997312e-01 5.39865553e-01
7.82657862e-01 1.58693790e-01 -3.72892350e-01 -9.48695559e-03
-5.11899628e-02 1.77859932e-01 1.11049518e-01 -7.88989589e-02
-9.66021657e-01 5.95025539e-01 4.65515661e+00 1.05250478e+00
-1.12758386e+00 1.80854678e-01 6.95743978e-01 1.70749322e-01
-8.63179266e-02 1.10641986e-01 -6.96985126e-01 7.76661158e-01
5.17208338e-01 -2.14226604e-01 2.25846305e-01 1.05579603e+00
2.12191701e-01 -8.52480531e-02 -4.50981319e-01 1.18410635e+00
2.60113508e-01 -1.25409925e+00 1.88162118e-01 -1.40851989e-01
8.44670713e-01 -1.08537182e-01 6.13840334e-02 3.71858597e-01
7.80027285e-02 -8.76495421e-01 2.79492229e-01 6.09937131e-01
7.77238011e-01 -7.90354311e-01 9.65090573e-01 3.53087395e-01
-1.62034404e+00 -3.99321228e-01 -6.60973370e-01 2.39567235e-01
-4.80489939e-01 8.08399260e-01 -3.32429618e-01 9.31482434e-01
5.48191071e-01 1.30782735e+00 -6.55539513e-01 1.25289047e+00
-2.09833622e-01 7.26035058e-01 1.28228694e-01 -4.27788775e-03
-6.60443008e-02 -4.89197791e-01 3.96989137e-01 8.78958464e-01
4.00493085e-01 3.51258665e-01 4.64720547e-01 4.25186604e-01
-7.22266585e-02 4.39170837e-01 -4.38577920e-01 -6.67160898e-02
7.86636591e-01 1.27445900e+00 -7.20345438e-01 -1.76863104e-01
-5.26187658e-01 7.52839625e-01 2.75831580e-01 3.06257993e-01
-5.76837897e-01 -4.51998293e-01 3.15675735e-01 -2.11239532e-01
5.91618776e-01 -1.03400603e-01 -9.36315134e-02 -1.38850725e+00
-3.04964948e-02 -1.07496440e+00 4.47936505e-01 -5.54625630e-01
-1.08244956e+00 4.42934930e-01 -2.88998961e-01 -1.86913574e+00
2.28898630e-01 3.28123122e-02 -7.20836878e-01 8.60881448e-01
-1.50399685e+00 -8.90381634e-01 -6.27327204e-01 6.59141719e-01
6.50247931e-01 -4.49173898e-01 7.71474659e-01 3.74552101e-01
-8.29332769e-01 4.91891116e-01 4.53117937e-01 2.54564106e-01
2.74247468e-01 -5.88732243e-01 -4.22160238e-01 7.81476140e-01
7.90972859e-02 7.01099277e-01 1.45613194e-01 -4.85689610e-01
-1.50177908e+00 -1.21812904e+00 4.87514794e-01 2.47578174e-01
2.33191743e-01 3.14399078e-02 -1.03805268e+00 2.17764780e-01
-3.78196418e-01 2.18296379e-01 8.53699088e-01 -1.43400773e-01
-5.33237219e-01 -4.31574821e-01 -9.08725262e-01 4.14102972e-01
6.80035055e-01 -5.15604496e-01 -2.55355954e-01 4.61610645e-01
5.37305713e-01 -1.29311875e-01 -9.24010754e-01 2.60559112e-01
4.04393792e-01 -8.19388926e-01 9.50802147e-01 1.39782652e-01
5.32733262e-01 -6.23589456e-01 -3.80251408e-01 -1.19207275e+00
-4.51215804e-01 1.87132120e-01 8.46301764e-02 1.25912333e+00
1.89364254e-01 -9.23623860e-01 6.44724369e-01 -2.31772512e-02
-2.29232945e-03 -1.10132515e+00 -7.76474297e-01 -6.74529791e-01
-2.23575264e-01 -5.66843189e-02 4.59686875e-01 9.90146756e-01
-1.57627255e-01 1.34642020e-01 -1.65630072e-01 1.06779180e-01
5.25329471e-01 6.59813702e-01 5.28644860e-01 -1.26412296e+00
-5.41550070e-02 -1.52561307e-01 -5.36093175e-01 -8.45111310e-01
3.20277028e-02 -1.16685736e+00 -2.37137556e-01 -1.50080717e+00
3.84966791e-01 -8.25242698e-01 -5.17942667e-01 4.60068285e-01
-2.43908510e-01 1.94609344e-01 2.43293762e-01 1.05362177e+00
-5.41320145e-01 8.01030278e-01 1.15460968e+00 -1.23394564e-01
-2.27407143e-01 1.08594343e-01 -1.19998789e+00 3.97843152e-01
1.11742830e+00 -5.09579778e-01 -4.75267470e-01 -7.20645040e-02
-3.97904873e-01 1.69850320e-01 1.57803118e-01 -1.02226126e+00
3.49422008e-01 -1.96039766e-01 7.18904257e-01 -4.81206417e-01
3.60712320e-01 -7.37605929e-01 3.66041303e-01 7.19503999e-01
-3.10625017e-01 -3.71559709e-01 -2.49355718e-01 8.13902915e-01
-6.24753535e-01 -1.66968793e-01 8.23174834e-01 5.42133041e-02
-6.96172774e-01 3.48078936e-01 -1.45614922e-01 -3.52172554e-01
1.00696564e+00 -1.96040824e-01 -2.27822736e-01 -3.77024323e-01
-3.06181073e-01 1.41057298e-01 1.14096344e-01 3.82919997e-01
1.16498435e+00 -1.35311151e+00 -8.39275122e-01 2.89733768e-01
2.56788343e-01 -9.02888924e-02 2.74219126e-01 9.74655509e-01
-1.77314460e-01 4.77395356e-01 1.84871048e-01 -4.27835226e-01
-1.42921436e+00 3.22123736e-01 8.27092305e-02 -3.03443596e-02
-6.80618823e-01 6.31622672e-01 1.38466433e-01 -1.07484959e-01
-8.89550894e-02 1.44399345e-01 -7.12000132e-01 -9.79782790e-02
6.53519213e-01 2.23238096e-01 -3.47976759e-02 -8.84316504e-01
-5.22498310e-01 6.43332422e-01 -4.48426813e-01 4.42819476e-01
1.42080915e+00 1.51206106e-02 -8.21305662e-02 7.52344579e-02
1.36567473e+00 -3.35986097e-03 -8.09488177e-01 -4.53443617e-01
-3.37670207e-01 -8.39161575e-01 3.54703069e-01 -5.07455587e-01
-1.57750237e+00 4.23454702e-01 6.18251741e-01 -1.14960454e-01
1.35216463e+00 -1.59862161e-01 6.83027864e-01 1.74675897e-01
4.72249210e-01 -6.04107916e-01 1.44470837e-02 2.64669329e-01
1.06826043e+00 -1.12051725e+00 3.68465960e-01 -8.51809382e-01
-7.94291496e-01 8.74161601e-01 4.18674797e-01 -4.72621113e-01
8.19016933e-01 -1.98608130e-01 -1.32370204e-01 -4.70277779e-02
-6.74270570e-01 -1.14609085e-01 2.26294830e-01 5.19511282e-01
4.45495903e-01 1.41050756e-01 -5.44328690e-01 6.28350377e-01
1.28541097e-01 2.02521048e-02 2.60558784e-01 1.18091500e+00
-3.15385133e-01 -1.03577614e+00 -4.24666792e-01 7.91358531e-01
-2.62848645e-01 -2.72653326e-02 -2.59923667e-01 3.04194450e-01
-9.67871547e-02 1.19432163e+00 -1.88285142e-01 -7.37309873e-01
6.87348843e-02 -3.12380195e-01 -2.34057195e-02 -8.27056706e-01
-3.91647756e-01 1.03653371e-01 -2.66516000e-01 -4.38119084e-01
-3.94102365e-01 -7.13914990e-01 -1.27672887e+00 -7.95454085e-02
-7.19311774e-01 6.04836583e-01 6.14484549e-01 5.93952894e-01
4.80394572e-01 2.59701967e-01 1.30727553e+00 -5.59814572e-01
-3.46590370e-01 -6.90785825e-01 -6.71915054e-01 1.78547859e-01
1.65117122e-02 -7.36645579e-01 -6.17880464e-01 -1.54813096e-01]
|
[12.461137771606445, 0.41552504897117615]
|
1b8a5eda-e47f-4b46-8247-b54f9df15500
|
modality-influence-in-multimodal-machine
|
2306.06476
| null |
https://arxiv.org/abs/2306.06476v1
|
https://arxiv.org/pdf/2306.06476v1.pdf
|
Modality Influence in Multimodal Machine Learning
|
Multimodal Machine Learning has emerged as a prominent research direction across various applications such as Sentiment Analysis, Emotion Recognition, Machine Translation, Hate Speech Recognition, and Movie Genre Classification. This approach has shown promising results by utilizing modern deep learning architectures. Despite the achievements made, challenges remain in data representation, alignment techniques, reasoning, generation, and quantification within multimodal learning. Additionally, assumptions about the dominant role of textual modality in decision-making have been made. However, limited investigations have been conducted on the influence of different modalities in Multimodal Machine Learning systems. This paper aims to address this gap by studying the impact of each modality on multimodal learning tasks. The research focuses on verifying presumptions and gaining insights into the usage of different modalities. The main contribution of this work is the proposal of a methodology to determine the effect of each modality on several Multimodal Machine Learning models and datasets from various tasks. Specifically, the study examines Multimodal Sentiment Analysis, Multimodal Emotion Recognition, Multimodal Hate Speech Recognition, and Multimodal Disease Detection. The study objectives include training SOTA MultiModal Machine Learning models with masked modalities to evaluate their impact on performance. Furthermore, the research aims to identify the most influential modality or set of modalities for each task and draw conclusions for diverse multimodal classification tasks. By undertaking these investigations, this research contributes to a better understanding of the role of individual modalities in multi-modal learning and provides valuable insights for future advancements in this field.
|
['Hadda Cherroun', 'Attia Nehar', 'Slimane Bellaouar', 'Abdelhamid Haouhat']
|
2023-06-10
| null | null | null | null |
['multimodal-sentiment-analysis', 'multimodal-emotion-recognition', 'genre-classification', 'sentiment-analysis', 'multimodal-sentiment-analysis', 'multimodal-emotion-recognition']
|
['computer-vision', 'computer-vision', 'computer-vision', 'natural-language-processing', 'natural-language-processing', 'speech']
|
[ 2.35078886e-01 -3.10241699e-01 -3.99011731e-01 -2.73824245e-01
-6.45163417e-01 -4.94726151e-01 9.32251513e-01 2.75848657e-01
-4.60219741e-01 4.57331955e-01 2.77067959e-01 -1.05806828e-01
-1.20702930e-01 -2.34226331e-01 -3.17736626e-01 -8.39531779e-01
1.62323698e-01 -6.09496869e-02 -4.96345729e-01 -2.86119491e-01
4.99073356e-01 4.58603978e-01 -1.74424076e+00 7.85919487e-01
3.93261641e-01 9.83728528e-01 -3.32951546e-01 6.65430963e-01
-1.37631029e-01 1.16823816e+00 -7.83671916e-01 -8.92545581e-01
-4.19306844e-01 -5.34008861e-01 -6.12954915e-01 2.91502148e-01
3.41640413e-01 -2.43512645e-01 -1.13157339e-01 7.17552841e-01
8.62381041e-01 3.89837623e-02 8.69610369e-01 -1.49861634e+00
-6.04707658e-01 6.06993675e-01 -4.78195727e-01 3.01978504e-03
6.29456341e-01 -3.70224565e-02 6.17769778e-01 -8.17303121e-01
4.58859295e-01 1.12615430e+00 4.92791623e-01 6.21192276e-01
-8.03838730e-01 -5.43950677e-01 -2.75022775e-01 4.13309664e-01
-1.26356959e+00 -5.74523509e-01 9.58404779e-01 -5.15021741e-01
9.54981208e-01 3.31403285e-01 2.38433495e-01 1.50568914e+00
1.37536809e-01 9.16272998e-01 1.41656494e+00 -7.60970235e-01
-2.36808509e-02 8.02312434e-01 2.27858528e-01 6.54468417e-01
-2.08647653e-01 -2.76526928e-01 -9.77274418e-01 -2.01271936e-01
4.56317142e-03 -2.05604762e-01 8.66134241e-02 1.42498210e-01
-1.00669050e+00 1.07253945e+00 -1.62155747e-01 8.03914547e-01
-2.53741264e-01 -3.23580056e-01 8.00258577e-01 2.78410465e-01
5.25912583e-01 2.73958713e-01 -2.94771433e-01 -3.58203381e-01
-8.04529130e-01 -1.83190316e-01 7.87147641e-01 2.95359284e-01
3.50852281e-01 1.49767110e-02 -8.90802033e-03 1.16472900e+00
4.27465856e-01 5.97704828e-01 6.31765902e-01 -7.42977977e-01
4.55166876e-01 6.82815373e-01 -2.31109858e-01 -1.28450310e+00
-7.30411589e-01 1.95080280e-01 -5.80798447e-01 -7.34934211e-02
3.88809949e-01 -4.74084675e-01 -5.74307382e-01 1.48115253e+00
7.17042089e-02 -3.82479608e-01 2.38723114e-01 8.94980073e-01
1.13534606e+00 5.54510534e-01 3.28824759e-01 -6.61386997e-02
1.54818356e+00 -7.07467556e-01 -8.78287852e-01 -4.84284833e-02
9.29576814e-01 -1.16216779e+00 8.39663088e-01 5.05804598e-01
-7.68404007e-01 -3.40370834e-01 -7.67069221e-01 -1.33057594e-01
-8.92315626e-01 4.19340491e-01 6.95987105e-01 1.26701915e+00
-7.16243625e-01 -5.32656536e-02 -5.28863907e-01 -6.63894713e-01
2.54455179e-01 2.37020597e-01 -4.95191664e-01 -3.35540771e-02
-1.16981435e+00 1.30341887e+00 2.45923430e-01 2.29980096e-01
-5.81764579e-01 -3.18197429e-01 -9.93072808e-01 -4.18361098e-01
-2.45721899e-02 -4.15360123e-01 9.51645732e-01 -1.52266443e+00
-1.37248826e+00 1.09196627e+00 -4.28762853e-01 -1.70798078e-01
4.97282073e-02 4.25753295e-02 -6.98764086e-01 9.42375734e-02
-1.99554265e-01 5.89318693e-01 9.16210532e-01 -1.27823818e+00
-6.17723465e-01 -6.25505567e-01 -1.69107839e-01 3.22568834e-01
-7.86771774e-01 5.71151793e-01 -6.48368597e-02 -3.54289114e-01
-3.75798196e-01 -9.42607701e-01 3.72010678e-01 -6.65030718e-01
-3.56474817e-01 -2.44149223e-01 8.17854464e-01 -6.91890061e-01
1.29868519e+00 -2.24353552e+00 4.18652058e-01 1.50965884e-01
-4.93105575e-02 -4.19211201e-03 -9.04915109e-02 9.48153019e-01
1.01416133e-01 3.20068032e-01 -6.93015531e-02 -6.32053196e-01
1.68529078e-01 1.85348421e-01 -2.02321574e-01 4.39860255e-01
3.40481043e-01 6.73532426e-01 -3.07235658e-01 -5.65798283e-01
5.50366819e-01 8.65352750e-01 3.57372798e-02 7.68631920e-02
3.29414904e-01 2.70372778e-01 -1.18614815e-01 1.07421386e+00
3.44556451e-01 1.38576835e-01 2.06002623e-01 -3.94622058e-01
-2.69925356e-01 -1.52616024e-01 -6.82966888e-01 1.21801925e+00
-4.46940154e-01 1.03717601e+00 -7.61644868e-03 -9.68170285e-01
7.39932537e-01 5.32687426e-01 4.11724448e-01 -8.21174562e-01
6.66828513e-01 1.59363464e-01 3.60263623e-02 -1.09529245e+00
8.10089767e-01 -3.06607813e-01 -1.83632627e-01 2.61730880e-01
1.65219437e-02 1.59242898e-01 2.73760021e-01 -1.98404826e-02
5.18119395e-01 -2.79199690e-01 6.41353503e-02 4.18702245e-01
7.09143281e-01 9.63589177e-02 -2.46401578e-01 5.25065780e-01
-3.06142777e-01 4.21803743e-01 6.91000938e-01 -1.40616804e-01
-7.09026396e-01 -5.36137760e-01 -2.63311595e-01 1.64310741e+00
-1.14008166e-01 -1.24952987e-01 -6.28573477e-01 -5.45312166e-01
-1.96042508e-01 7.22107887e-01 -7.62643218e-01 -1.89087585e-01
-1.51186958e-01 -1.29339874e+00 1.05792797e+00 2.87861437e-01
2.76594404e-02 -1.08845460e+00 -6.78665996e-01 -4.24933791e-01
-3.10363889e-01 -1.10876822e+00 1.61076590e-01 1.10145666e-01
-8.14404190e-01 -1.16967916e+00 -6.98847175e-01 -6.80584013e-01
5.36250770e-01 1.28344864e-01 7.44867563e-01 -1.17422290e-01
-6.33919835e-02 8.67238402e-01 -6.26084268e-01 -6.06100380e-01
-5.48869908e-01 2.96857029e-01 -7.95823485e-02 2.37224638e-01
8.11804473e-01 2.44068205e-01 -9.89274830e-02 2.35432103e-01
-1.25681663e+00 -3.34156215e-01 6.33110881e-01 7.87395000e-01
-7.99923986e-02 1.47472322e-01 5.06079257e-01 -6.73183382e-01
1.10665834e+00 -8.68943989e-01 1.25848770e-01 3.44886631e-01
-3.57846379e-01 -3.14421415e-01 3.05693716e-01 -4.91437823e-01
-1.17664123e+00 -2.18504295e-01 -1.30078271e-01 -2.56887078e-01
-5.75466633e-01 9.53845084e-01 8.91836658e-02 -1.99759066e-01
4.59360361e-01 1.65856287e-01 3.12626600e-01 -1.62657350e-01
2.50560075e-01 1.01665080e+00 -5.31690456e-02 -3.84929508e-01
1.73319697e-01 3.94169450e-01 -2.00855881e-01 -1.38763773e+00
-4.85114843e-01 -4.06922758e-01 -5.19569695e-01 -7.11603403e-01
1.03080451e+00 -8.19749534e-01 -7.38785028e-01 7.92317152e-01
-9.81679142e-01 1.29407719e-01 5.32108009e-01 5.94974041e-01
-2.35643126e-02 4.73180741e-01 -6.85440481e-01 -1.17233145e+00
-2.30771288e-01 -1.26722717e+00 9.23784494e-01 3.76506090e-01
-6.04488313e-01 -1.40338957e+00 1.24086812e-01 1.24165905e+00
3.29981893e-01 1.20817453e-01 1.00367093e+00 -8.95341337e-01
2.31709704e-01 -4.01268542e-01 -7.12792799e-02 4.52055275e-01
-1.98943336e-02 4.00323242e-01 -1.24475694e+00 3.21874814e-03
-2.10436627e-01 -8.54096889e-01 6.26048326e-01 2.28038192e-01
6.47214592e-01 -4.53485437e-02 8.17705914e-02 4.27898467e-02
1.01876187e+00 3.13178241e-01 5.54202735e-01 5.81388235e-01
6.34277046e-01 1.22118974e+00 5.76313019e-01 4.07277763e-01
4.53241199e-01 4.06969219e-01 4.60448354e-01 2.39932723e-02
1.69908151e-01 2.23474532e-01 7.33043969e-01 8.42850685e-01
-4.62673232e-02 -2.43195489e-01 -1.04639590e+00 2.44855344e-01
-1.54763234e+00 -1.03869581e+00 -1.46786720e-01 1.85024226e+00
4.27784383e-01 -4.25677985e-01 2.24239320e-01 2.60409087e-01
4.60982949e-01 2.08848089e-01 -5.20076826e-02 -9.39686835e-01
-6.41463161e-01 -2.02050105e-01 1.25533685e-01 1.76791295e-01
-1.31681418e+00 6.60172284e-01 6.00138950e+00 6.53666854e-01
-1.57533395e+00 -5.85792847e-02 6.68474376e-01 -1.65484518e-01
-1.12188056e-01 -5.88750184e-01 -5.24550200e-01 3.91861618e-01
1.03756011e+00 3.81292015e-01 3.70528758e-01 4.89701003e-01
3.68450463e-01 -4.81444567e-01 -9.05559599e-01 1.05371749e+00
8.35653603e-01 -7.81782568e-01 1.02554455e-01 4.00441214e-02
5.91600895e-01 -4.27623279e-02 5.55688739e-01 4.30504501e-01
-6.13530576e-01 -1.16543317e+00 6.12703681e-01 5.82608640e-01
2.52510726e-01 -8.53423119e-01 1.21665907e+00 4.18660305e-02
-3.91676158e-01 -2.50341326e-01 6.51693717e-02 -1.35966450e-01
7.39602325e-03 1.21019175e-02 -9.03223515e-01 4.48903978e-01
5.24403632e-01 3.79786223e-01 -7.17821300e-01 5.38842082e-01
1.73366472e-01 6.88409150e-01 -2.55889911e-03 -3.17748785e-01
2.70368993e-01 -1.32409617e-01 2.34344617e-01 1.63290679e+00
2.63051569e-01 -2.59909302e-01 -2.35803500e-01 2.74186254e-01
5.85719235e-02 5.73015034e-01 -7.75505602e-01 -6.96727574e-01
2.58344561e-01 1.36049986e+00 -6.72167718e-01 1.13850683e-02
-7.87959337e-01 7.65801489e-01 8.44660848e-02 3.26521575e-01
-9.46951449e-01 -1.46501347e-01 3.67440939e-01 -3.09592754e-01
-1.20515242e-01 2.53993110e-03 -7.22674906e-01 -8.85251403e-01
-2.70057946e-01 -1.27670276e+00 6.37442112e-01 -8.20485532e-01
-1.29167414e+00 3.55914682e-01 1.82838049e-02 -9.09431756e-01
-2.63597459e-01 -8.62983942e-01 -3.98904860e-01 8.08799565e-01
-1.24630356e+00 -1.50561798e+00 4.25587110e-02 6.39520407e-01
3.95651370e-01 -5.76485217e-01 9.70465958e-01 3.80427390e-01
-9.86881733e-01 6.37249589e-01 1.31196622e-02 1.50286749e-01
1.00537276e+00 -7.24846005e-01 -9.80514944e-01 5.33163846e-01
-7.03210309e-02 6.18709087e-01 5.53397357e-01 -4.46237415e-01
-1.55682313e+00 -4.01152700e-01 9.99604583e-01 -5.61890304e-01
8.57900918e-01 2.94401497e-01 -6.52310133e-01 2.69047439e-01
8.49722445e-01 -8.63040030e-01 1.54641080e+00 3.48214298e-01
-4.18810397e-01 6.87572211e-02 -9.44869757e-01 4.86997783e-01
-7.32757002e-02 -9.15494680e-01 -4.56077844e-01 -1.35441229e-01
-1.39762536e-01 -1.89903289e-01 -1.10533988e+00 2.94535309e-01
8.84578645e-01 -1.04200327e+00 7.33723581e-01 -8.13079178e-01
9.08471346e-01 9.22354311e-02 -2.99023420e-01 -1.10721171e+00
2.66056210e-01 8.43913555e-02 -4.14626226e-02 1.25421917e+00
7.51034379e-01 -4.11119848e-01 5.16297758e-01 1.01177609e+00
-7.50940219e-02 -7.86140680e-01 -9.25087512e-01 5.24523780e-02
6.32190183e-02 -6.22253358e-01 -2.33365577e-02 1.47737920e+00
2.55404800e-01 4.84739333e-01 -7.44770348e-01 6.50795847e-02
2.83020347e-01 -2.17186108e-01 7.87710130e-01 -6.19684279e-01
2.59544641e-01 -7.14819133e-01 -3.86949718e-01 -5.44039160e-02
3.52720827e-01 -6.88490748e-01 -5.79801381e-01 -1.23250866e+00
3.25827628e-01 1.97643489e-01 -2.63093174e-01 5.39759517e-01
-6.70958757e-02 4.51252371e-01 4.90167916e-01 2.05365848e-02
-5.27457893e-01 2.20960930e-01 9.06751096e-01 -2.39526853e-01
-1.34382620e-01 -1.19609937e-01 -7.47169316e-01 7.39553630e-01
1.01750302e+00 -1.59872428e-01 -3.34798902e-01 -4.58470613e-01
3.59349847e-01 -9.70187336e-02 3.13396066e-01 -4.51209962e-01
3.23766917e-01 -3.70967090e-02 5.37347794e-01 -5.01306653e-01
7.47060537e-01 -8.48157048e-01 -7.01800883e-02 9.17205364e-02
-5.15196383e-01 1.73199028e-01 4.74446833e-01 4.88463379e-02
-4.91928101e-01 -4.85303313e-01 3.91159087e-01 1.96846560e-01
-8.87737572e-01 -5.67732036e-01 -8.24690461e-01 -1.52943969e-01
1.11858940e+00 -2.95640647e-01 -4.44800347e-01 -4.24893409e-01
-7.15644062e-01 6.05237298e-02 2.43265718e-01 8.54375601e-01
6.65878654e-01 -1.04229498e+00 -5.84419191e-01 -1.83136597e-01
2.92057306e-01 -9.95383263e-01 5.14122546e-01 1.31326210e+00
-3.41955870e-01 5.03531039e-01 -2.32354641e-01 -4.33589250e-01
-1.87415409e+00 3.20068628e-01 2.63027489e-01 1.10420816e-01
4.33659881e-01 5.02079189e-01 -3.88678402e-01 -3.87804955e-01
4.77597296e-01 4.67415810e-01 -6.92500532e-01 8.60658944e-01
5.04123390e-01 7.20652103e-01 7.85489473e-03 -1.22355711e+00
-3.74706388e-01 4.25377637e-01 -1.15771163e-02 -2.19905198e-01
8.32905412e-01 -2.92828232e-01 -5.17829239e-01 9.03097332e-01
1.34518373e+00 8.95420313e-02 -2.21252963e-01 2.99722642e-01
7.50960261e-02 -1.49399951e-01 5.42408563e-02 -1.13324749e+00
-8.23792994e-01 1.01383173e+00 7.75730431e-01 4.93760169e-01
1.20840371e+00 -3.30880843e-02 3.27664375e-01 2.31582209e-01
-1.63499966e-01 -1.30294180e+00 1.65486075e-02 5.88679433e-01
6.53363705e-01 -1.73983371e+00 -1.77404031e-01 1.79587808e-02
-1.24439943e+00 1.35042262e+00 3.66844952e-01 7.64872134e-01
4.93656844e-01 7.75704160e-02 5.71440697e-01 -3.41732293e-01
-4.63987321e-01 -2.11132556e-01 6.16765976e-01 4.63107407e-01
1.06111968e+00 1.69500947e-01 -5.31813681e-01 5.25472820e-01
-2.52027623e-02 -1.46546841e-01 4.11099225e-01 1.00701547e+00
-1.43753111e-01 -1.10194290e+00 -8.26676726e-01 3.44638973e-01
-8.57829928e-01 -1.17426537e-01 -1.11068034e+00 8.59464645e-01
1.66436478e-01 1.38413560e+00 -1.77027181e-01 -7.32800364e-01
2.12059945e-01 5.16991496e-01 3.64258319e-01 -1.52509287e-01
-8.25810194e-01 6.40840232e-02 4.66127425e-01 -4.60995361e-02
-9.58291292e-01 -8.49616230e-01 -7.05903828e-01 -4.41109627e-01
-1.72172621e-01 4.50216606e-02 1.09484267e+00 1.05715942e+00
2.74624765e-01 2.79118329e-01 4.84365761e-01 -7.60251641e-01
-1.81232914e-01 -1.01701379e+00 -2.45808646e-01 5.39580405e-01
7.28826895e-02 -4.70386565e-01 -4.40790355e-01 1.39275014e-01]
|
[12.987695693969727, 5.306820392608643]
|
e01f83ef-7c2f-4c09-a599-089556f94fec
|
x-pool-cross-modal-language-video-attention
|
2203.15086
| null |
https://arxiv.org/abs/2203.15086v1
|
https://arxiv.org/pdf/2203.15086v1.pdf
|
X-Pool: Cross-Modal Language-Video Attention for Text-Video Retrieval
|
In text-video retrieval, the objective is to learn a cross-modal similarity function between a text and a video that ranks relevant text-video pairs higher than irrelevant pairs. However, videos inherently express a much wider gamut of information than texts. Instead, texts often capture sub-regions of entire videos and are most semantically similar to certain frames within videos. Therefore, for a given text, a retrieval model should focus on the text's most semantically similar video sub-regions to make a more relevant comparison. Yet, most existing works aggregate entire videos without directly considering text. Common text-agnostic aggregations schemes include mean-pooling or self-attention over the frames, but these are likely to encode misleading visual information not described in the given text. To address this, we propose a cross-modal attention model called X-Pool that reasons between a text and the frames of a video. Our core mechanism is a scaled dot product attention for a text to attend to its most semantically similar frames. We then generate an aggregated video representation conditioned on the text's attention weights over the frames. We evaluate our method on three benchmark datasets of MSR-VTT, MSVD and LSMDC, achieving new state-of-the-art results by up to 12% in relative improvement in Recall@1. Our findings thereby highlight the importance of joint text-video reasoning to extract important visual cues according to text. Full code and demo can be found at: https://layer6ai-labs.github.io/xpool/
|
['Guangwei Yu', 'Animesh Garg', 'Maksims Volkovs', 'Keyvan Golestan', 'Junwei Ma', 'Noel Vouitsis', 'Satya Krishna Gorti']
|
2022-03-28
| null |
http://openaccess.thecvf.com//content/CVPR2022/html/Gorti_X-Pool_Cross-Modal_Language-Video_Attention_for_Text-Video_Retrieval_CVPR_2022_paper.html
|
http://openaccess.thecvf.com//content/CVPR2022/papers/Gorti_X-Pool_Cross-Modal_Language-Video_Attention_for_Text-Video_Retrieval_CVPR_2022_paper.pdf
|
cvpr-2022-1
|
['video-text-retrieval']
|
['computer-vision']
|
[ 2.04359755e-01 -5.77418745e-01 -3.16563994e-01 -3.30010533e-01
-1.17068076e+00 -4.91496712e-01 6.85437799e-01 4.94074114e-02
-4.25327808e-01 3.17464918e-01 6.88845634e-01 1.63009554e-01
-3.04909591e-02 -4.10194844e-01 -8.54252279e-01 -6.30454600e-01
1.39038742e-01 4.29935679e-02 2.18033552e-01 -7.27316290e-02
3.52528721e-01 8.62627700e-02 -1.70157468e+00 9.23156917e-01
5.56611657e-01 1.21911645e+00 4.66980547e-01 6.64801300e-01
-1.26202792e-01 9.50341046e-01 -6.14384592e-01 -4.03983086e-01
2.25271165e-01 -4.82775182e-01 -7.25824714e-01 1.91061720e-01
9.97249782e-01 -7.37760901e-01 -1.01412618e+00 1.15302765e+00
4.61123109e-01 5.95273554e-01 7.24572778e-01 -1.18424439e+00
-9.72182035e-01 5.52916884e-01 -7.95896471e-01 5.91783464e-01
7.08284795e-01 2.19503000e-01 1.25408363e+00 -1.26697648e+00
7.33868182e-01 1.42778695e+00 -8.09076615e-03 3.63688856e-01
-8.63602221e-01 -7.03615546e-01 3.27521712e-01 4.91977245e-01
-1.55044031e+00 -4.78797704e-01 6.61387086e-01 -3.19857448e-01
8.73030722e-01 4.40277725e-01 5.50377905e-01 1.28727591e+00
3.27480525e-01 1.24057174e+00 4.26945299e-01 -9.18451510e-03
-1.14128880e-01 -1.74982592e-01 -7.45749027e-02 4.77471918e-01
1.65903997e-02 -3.58929276e-01 -9.13723052e-01 4.43422459e-02
4.88257855e-01 5.88331521e-01 -6.05581880e-01 -1.75527200e-01
-1.43551803e+00 7.08041728e-01 4.68950868e-01 3.52616608e-01
-4.60130423e-01 3.23462754e-01 5.33214986e-01 2.73491412e-01
5.06432354e-01 2.69797593e-01 -9.41716582e-02 -3.32503319e-02
-1.19705880e+00 2.58068442e-01 2.05985561e-01 9.64568615e-01
7.05647588e-01 -1.64642960e-01 -7.04444766e-01 7.69234478e-01
3.97557676e-01 5.93663633e-01 4.78084803e-01 -9.40128148e-01
7.62967110e-01 4.24933493e-01 1.11653265e-02 -1.28898549e+00
2.38137305e-01 -1.58239380e-01 -5.38944721e-01 -1.83488846e-01
1.48811176e-01 1.67879269e-01 -9.36541140e-01 1.53296375e+00
8.31359625e-02 1.99974522e-01 -4.96608391e-02 1.33659565e+00
1.06382060e+00 9.19840991e-01 1.15537651e-01 -1.67937309e-01
1.47584033e+00 -1.09382570e+00 -7.82462716e-01 -2.35371634e-01
3.20581615e-01 -1.02211201e+00 1.16735959e+00 1.53921425e-01
-1.36089301e+00 -5.65257847e-01 -8.78985941e-01 -4.90902185e-01
-3.88461530e-01 1.37225226e-01 5.03042005e-02 -5.05884513e-02
-1.00861979e+00 4.81169701e-01 -5.31532049e-01 -5.19431293e-01
4.73635405e-01 7.83535466e-02 -4.05689508e-01 -4.34545279e-01
-1.28210199e+00 5.40721714e-01 2.45289728e-01 -7.24925473e-02
-1.08934259e+00 -6.26814485e-01 -8.03692460e-01 1.65737957e-01
6.59617722e-01 -6.68875396e-01 9.77217615e-01 -1.31130445e+00
-8.73799920e-01 7.74021149e-01 -3.14003378e-01 -3.15766335e-01
4.30414230e-01 -5.94888508e-01 -3.94879371e-01 8.41024995e-01
2.35969201e-01 9.44792569e-01 1.30070114e+00 -1.01909995e+00
-6.98366880e-01 -2.70217657e-01 2.53228039e-01 5.17413139e-01
-5.46253979e-01 2.12613970e-01 -1.20714247e+00 -1.18426013e+00
-9.93625671e-02 -6.53801084e-01 3.22458744e-01 1.53928071e-01
-2.07246840e-01 -1.33106425e-01 1.14454401e+00 -8.37225616e-01
1.50529444e+00 -2.34592414e+00 3.67651761e-01 -1.83711573e-01
3.04668307e-01 1.51261806e-01 -4.67818201e-01 4.21625167e-01
-8.53996724e-02 9.28376466e-02 1.52612120e-01 -3.10567826e-01
3.61963026e-02 -2.27278069e-01 -5.19123197e-01 5.12901723e-01
2.48045772e-01 1.07258546e+00 -9.63707745e-01 -6.96605325e-01
4.03850704e-01 7.36379325e-01 -5.32831907e-01 1.98808715e-01
-2.97841579e-01 1.02811642e-01 -5.55790186e-01 7.19052970e-01
4.56125170e-01 -3.38691115e-01 -1.98183432e-01 -6.57715738e-01
3.04106593e-01 3.83671708e-02 -8.75333846e-01 1.92291129e+00
-8.33868012e-02 9.86320972e-01 -1.39431611e-01 -8.38108957e-01
4.44097131e-01 2.32511386e-01 6.78689599e-01 -7.23271549e-01
2.55273610e-01 -3.63262221e-02 -3.53585035e-01 -7.51683831e-01
7.74175942e-01 2.90312409e-01 -6.02802858e-02 3.69309157e-01
1.40916988e-01 2.19428942e-01 3.79578263e-01 6.98376477e-01
9.70632076e-01 1.78841382e-01 7.10989535e-02 -6.01246543e-02
5.52869618e-01 -2.86513209e-01 2.02699870e-01 7.83679307e-01
-2.09754914e-01 9.63273764e-01 3.81083995e-01 -2.92764634e-01
-9.02463675e-01 -9.24626589e-01 1.75044984e-01 1.40771735e+00
4.20008510e-01 -6.56638145e-01 -5.21770239e-01 -6.55483305e-01
-1.31924208e-02 4.24642295e-01 -8.00544858e-01 -3.01431388e-01
-3.98856163e-01 -3.08685482e-01 3.76219720e-01 4.26383227e-01
4.41374958e-01 -1.01836824e+00 -5.03121555e-01 1.72886308e-02
-7.00176716e-01 -1.19534016e+00 -1.22889936e+00 -3.07925254e-01
-5.81771731e-01 -1.02963829e+00 -1.12909639e+00 -6.95619524e-01
6.32832050e-01 8.02178442e-01 1.08716452e+00 2.36719951e-01
-2.81672686e-01 6.29317224e-01 -6.16070509e-01 2.50297755e-01
9.09097344e-02 -2.60243505e-01 -1.00731745e-01 3.85886580e-01
6.06547236e-01 -4.46377508e-02 -9.21145439e-01 3.01187545e-01
-1.24851370e+00 3.93159203e-02 5.32947421e-01 7.26818264e-01
6.01839483e-01 -1.67724833e-01 1.05078205e-01 -2.36479849e-01
3.66327971e-01 -6.77090347e-01 -1.00023165e-01 3.65121484e-01
-3.57011333e-02 -6.62607402e-02 4.60809171e-01 -4.92115885e-01
-8.04560900e-01 -1.56967193e-01 2.89747566e-01 -1.23458099e+00
7.73531245e-03 3.85669589e-01 -1.00547202e-01 3.65766495e-01
2.58182287e-01 4.03342754e-01 -2.60770142e-01 -1.51181489e-01
2.14310989e-01 5.51737189e-01 4.14371908e-01 -6.40381217e-01
5.10993481e-01 4.99600470e-01 -2.95188159e-01 -8.22433889e-01
-7.00563729e-01 -7.58360267e-01 -2.59306967e-01 -5.15808403e-01
1.08338368e+00 -1.21550548e+00 -4.22751337e-01 1.27741039e-01
-8.86220574e-01 -1.08888894e-01 5.56264399e-03 6.20705247e-01
-4.74059522e-01 7.34132707e-01 -5.96705198e-01 -4.49627608e-01
-4.23829764e-01 -1.41473043e+00 1.49202001e+00 5.08657880e-02
-1.78625539e-01 -6.39122903e-01 -3.86635125e-01 5.09974182e-01
1.61207542e-01 -6.05944581e-02 5.26132643e-01 -6.07201040e-01
-7.79967785e-01 -1.28566325e-01 -4.81457621e-01 1.62898690e-01
9.17912051e-02 2.90521145e-01 -8.76091003e-01 -7.00800300e-01
-2.76070207e-01 -3.61401677e-01 1.26929867e+00 5.33949971e-01
1.27162576e+00 -3.48844737e-01 -3.71766210e-01 4.79048043e-01
1.37060034e+00 1.29713312e-01 6.80267215e-01 1.71675831e-01
8.13890398e-01 5.50950944e-01 9.39768791e-01 3.70912611e-01
9.62787345e-02 7.90107489e-01 4.95041996e-01 1.60072491e-01
-1.96133032e-01 -1.45290956e-01 8.20848703e-01 6.21147692e-01
9.26969722e-02 -6.83745265e-01 -6.67029560e-01 6.65978253e-01
-1.90912771e+00 -1.31252062e+00 7.97154978e-02 2.29137802e+00
5.65802455e-01 -1.11663483e-01 3.69979739e-02 -2.07181811e-01
9.10218894e-01 4.78591204e-01 -4.73343998e-01 1.68300182e-01
-2.04026595e-01 -2.43337110e-01 1.42735630e-01 3.40306133e-01
-1.20312870e+00 8.41353536e-01 5.01701450e+00 1.22633064e+00
-1.12670505e+00 9.39397812e-02 7.52535701e-01 -8.65777612e-01
-2.67809570e-01 -2.41495892e-01 -6.32292449e-01 7.53200293e-01
6.80194438e-01 -1.85923681e-01 3.76249790e-01 4.98750716e-01
3.11170250e-01 -1.68634862e-01 -1.12731624e+00 1.16986358e+00
5.09013891e-01 -1.34514666e+00 6.49379849e-01 -7.54730850e-02
8.73810112e-01 7.42949024e-02 3.65892261e-01 3.34114522e-01
-1.72298774e-01 -8.23753834e-01 9.10540819e-01 5.79925239e-01
8.27049077e-01 -7.61650860e-01 5.87995529e-01 -2.91142642e-01
-1.46670055e+00 -4.14258763e-02 -4.36859608e-01 4.19406682e-01
4.32922132e-02 3.45698863e-01 -1.62213013e-01 6.08317018e-01
1.16890776e+00 1.12429130e+00 -7.57115722e-01 9.01993155e-01
1.89754426e-01 1.29081905e-01 -4.16505300e-02 1.13530576e-01
5.35723329e-01 -4.19876538e-03 6.09622896e-01 1.36594355e+00
5.10598004e-01 1.25274569e-01 1.21623896e-01 6.98906243e-01
-3.12499851e-01 2.54371762e-01 -5.58140099e-01 -1.81715339e-01
3.41353446e-01 1.15455329e+00 -6.77567840e-01 -6.90032840e-01
-7.99664497e-01 1.16035187e+00 6.22742660e-02 7.07433581e-01
-1.09329951e+00 -3.65348995e-01 7.92182446e-01 5.78169189e-02
5.52602887e-01 1.63963273e-01 2.68025339e-01 -1.41796470e+00
1.86265767e-01 -1.01574206e+00 6.22714460e-01 -1.23286307e+00
-1.21919072e+00 6.01833403e-01 -7.72717316e-03 -1.61365247e+00
-4.63361666e-02 -4.21412706e-01 -3.13902706e-01 7.01372087e-01
-1.26739383e+00 -9.91248488e-01 -5.55395246e-01 9.39189255e-01
1.09315538e+00 -1.58871368e-01 1.52486920e-01 4.80715692e-01
-2.86036879e-01 5.54788947e-01 2.09191188e-01 2.58022815e-01
1.14815116e+00 -8.78649414e-01 1.10970110e-01 9.96605337e-01
1.76668674e-01 5.45299172e-01 5.83158791e-01 -8.05286288e-01
-1.66615152e+00 -1.27255583e+00 5.70468664e-01 -4.02933568e-01
8.02274287e-01 -1.48973927e-01 -1.02830827e+00 7.04836488e-01
5.07710457e-01 1.63892657e-01 4.35822099e-01 -5.20139754e-01
-4.16656822e-01 -4.07963544e-02 -7.52342463e-01 7.57775605e-01
9.24039721e-01 -8.40242863e-01 -6.65014446e-01 1.68684468e-01
7.43753910e-01 -3.13192844e-01 -8.58142138e-01 1.73438713e-01
6.18402302e-01 -9.81481910e-01 1.19803441e+00 -4.81295854e-01
8.28549564e-01 -5.05222738e-01 -6.08000576e-01 -9.72015679e-01
-1.88901842e-01 -5.48062027e-01 -2.94950843e-01 1.10378349e+00
-1.82665661e-02 -5.50049134e-02 4.87108201e-01 3.32466781e-01
-9.57214534e-02 -5.72826564e-01 -7.48687387e-01 -4.33417469e-01
-1.12668879e-01 -4.35651034e-01 3.76364529e-01 9.71608758e-01
-7.77630229e-03 2.89087325e-01 -5.53948700e-01 1.10647619e-01
5.23855209e-01 1.18896678e-01 5.54519534e-01 -7.39580810e-01
-1.55210957e-01 -7.17764795e-01 -3.69137377e-01 -1.15842903e+00
1.60775006e-01 -7.95384824e-01 3.06338910e-02 -1.40396988e+00
6.59462631e-01 3.71686667e-01 -4.42007840e-01 2.99065322e-01
-4.97659206e-01 6.59145057e-01 4.71015096e-01 3.49029064e-01
-1.15094054e+00 5.27352989e-01 1.39143121e+00 -5.13014615e-01
1.13827087e-01 -5.70399344e-01 -4.74603921e-01 4.32144880e-01
4.66896564e-01 -3.31538737e-01 -4.65240180e-01 -6.87377512e-01
1.57056570e-01 2.40902975e-01 5.16718328e-01 -8.97887826e-01
3.41973811e-01 -4.23669256e-02 6.72312617e-01 -8.55873466e-01
4.72810000e-01 -8.67097557e-01 1.15562238e-01 1.77672744e-01
-5.76655149e-01 1.84271991e-01 1.42322138e-01 7.35339940e-01
-5.74581087e-01 -3.03073358e-02 5.44253349e-01 -4.60080095e-02
-9.56652045e-01 5.59553266e-01 -4.01871681e-01 2.13756457e-01
9.89906669e-01 -2.04950392e-01 -5.20622015e-01 -6.66103899e-01
-6.49351776e-01 3.40835571e-01 6.05359375e-01 8.10873210e-01
9.69025791e-01 -1.44145167e+00 -7.63596892e-01 5.17062843e-02
3.92799556e-01 -2.48248026e-01 6.33665204e-01 9.14759517e-01
-2.49971241e-01 5.97267449e-01 -2.38397479e-01 -6.72333419e-01
-1.42930007e+00 8.63672078e-01 2.03645587e-01 9.15479660e-02
-6.14735186e-01 6.54452085e-01 8.11527491e-01 3.89022529e-01
3.61725301e-01 -3.66655022e-01 -2.00679362e-01 4.17207539e-01
8.52843285e-01 1.46150142e-01 -1.57126114e-01 -1.04427421e+00
-4.03611034e-01 7.10885584e-01 -2.93623835e-01 1.05073273e-01
1.00458038e+00 -4.37989205e-01 1.74811080e-01 2.60427207e-01
1.54616630e+00 -1.85704947e-01 -1.47761643e+00 -4.41399515e-01
-3.73203754e-01 -9.29947317e-01 2.32350275e-01 -5.27739048e-01
-1.41116703e+00 8.36578667e-01 4.35667068e-01 -5.06210811e-02
1.35902858e+00 1.69572920e-01 7.16534615e-01 2.72051245e-01
3.05200908e-02 -9.84278083e-01 6.13861084e-01 1.75538898e-01
1.12639129e+00 -1.40219903e+00 1.72600999e-01 6.88426429e-03
-8.44600260e-01 1.04264224e+00 6.67219520e-01 -1.30642384e-01
3.32116246e-01 -7.24983960e-02 -1.77185953e-01 -1.70958892e-01
-9.99866188e-01 -3.06424767e-01 7.79072940e-01 2.54927486e-01
4.76544917e-01 -3.31793189e-01 -3.53254378e-02 3.22764099e-01
4.98436093e-01 -2.12293297e-01 3.66196424e-01 7.64666915e-01
-4.15587127e-01 -5.42712390e-01 -5.43853164e-01 5.75789988e-01
-8.14115703e-01 -3.16536963e-01 -2.82694727e-01 6.66998804e-01
-3.05999756e-01 8.54828775e-01 3.79675835e-01 -3.62119496e-01
1.27900392e-01 -4.59088886e-04 3.28006119e-01 -3.39692712e-01
-4.65245783e-01 6.14199221e-01 -3.02473038e-01 -8.78633320e-01
-6.54491067e-01 -7.56670535e-01 -1.09277189e+00 -5.37522137e-01
-1.96667872e-02 1.00652995e-02 1.24248274e-01 7.03747511e-01
4.24559385e-01 6.01744175e-01 4.38465208e-01 -1.15557301e+00
-3.18553559e-02 -7.69699514e-01 -4.05786902e-01 8.03687572e-01
4.75872040e-01 -7.12054491e-01 -4.96467382e-01 3.22380662e-01]
|
[10.229463577270508, 0.8641865253448486]
|
dc09c2d1-48d4-46e6-84ad-7e2b35424c95
|
adventures-in-mathematical-reasoning
|
2008.09067
| null |
https://arxiv.org/abs/2008.09067v1
|
https://arxiv.org/pdf/2008.09067v1.pdf
|
Adventures in Mathematical Reasoning
|
"Mathematics is not a careful march down a well-cleared highway, but a journey into a strange wilderness, where the explorers often get lost. Rigour should be a signal to the historian that the maps have been made, and the real explorers have gone elsewhere." W.S. Anglin, the Mathematical Intelligencer, 4 (4), 1982.
|
['Toby Walsh']
|
2020-08-20
| null | null | null | null |
['mathematical-reasoning']
|
['natural-language-processing']
|
[-3.43614876e-01 -2.97021475e-02 -2.26951048e-01 -3.02986681e-01
-3.12807709e-01 -5.45941949e-01 6.65661037e-01 -1.71368215e-02
-2.89317697e-01 1.04723048e+00 2.39425510e-01 -8.95665765e-01
-4.56913412e-01 -7.98526824e-01 -5.33119500e-01 -2.88447887e-01
-3.48033518e-01 5.97817004e-01 3.96847799e-02 -8.68429840e-01
1.03725481e+00 6.45208776e-01 -7.35104442e-01 -5.89707077e-01
7.41582870e-01 3.37853730e-01 -2.37423480e-02 8.78546059e-01
-2.42339894e-01 1.20490253e+00 -6.09981358e-01 -6.61191761e-01
1.40956312e-01 -6.10131800e-01 -1.18937898e+00 -3.09814513e-01
-2.06815861e-02 -1.94174290e-01 -5.50744355e-01 1.26875246e+00
-4.95356858e-01 -1.90278757e-02 4.27610785e-01 -1.21177697e+00
-1.01931918e+00 7.10105121e-01 -6.23469472e-01 5.65964878e-01
4.51241195e-01 1.73828170e-01 5.22903085e-01 -6.22362792e-01
8.63978386e-01 1.27336419e+00 8.55064511e-01 -3.31927180e-01
-9.69423950e-01 -7.84985721e-01 -1.00690320e-01 1.61631465e-01
-1.64615762e+00 -2.01716542e-01 6.33795142e-01 -6.65986836e-01
4.51710165e-01 5.02262056e-01 9.79746819e-01 3.75909150e-01
7.83801556e-01 2.31912971e-01 1.29564691e+00 -3.28471810e-01
1.60178959e-01 1.43115520e-01 -8.30051862e-03 4.25887018e-01
4.34065402e-01 -5.26637509e-02 -3.84843022e-01 2.59080715e-02
1.06261826e+00 9.43764448e-02 -3.91590685e-01 8.03405419e-02
-1.19857526e+00 7.39203155e-01 5.46699643e-01 8.05336595e-01
-2.47071713e-01 1.75372168e-01 2.76447058e-01 6.22834504e-01
2.21052319e-01 1.03313386e+00 -2.18040511e-01 -8.64529967e-01
-1.04031169e+00 2.79417157e-01 1.00580096e+00 7.36477137e-01
4.33310986e-01 -1.21754512e-01 8.84706855e-01 7.47173280e-02
2.84496456e-01 1.93286806e-01 2.38555953e-01 -1.02186990e+00
2.16380522e-01 4.93887722e-01 3.48528981e-01 -1.79409146e+00
-3.57348114e-01 -5.02808213e-01 -5.92419147e-01 7.07264125e-01
6.36447310e-01 -1.18576018e-02 -5.33945918e-01 9.90672708e-01
-2.36232445e-01 -2.29960948e-01 -1.06042244e-01 1.37606359e+00
2.90260732e-01 7.31383801e-01 7.20549654e-03 1.44687071e-01
9.01926041e-01 -5.65875709e-01 -8.12384009e-01 -7.08780363e-02
4.95284259e-01 -8.89385700e-01 1.09228873e+00 8.23618293e-01
-1.34622025e+00 -1.13106109e-01 -1.43557930e+00 -2.92306896e-02
-5.45733452e-01 -3.97932380e-01 8.66220117e-01 6.32083595e-01
-8.84923160e-01 1.08083236e+00 -7.43912101e-01 -7.19598889e-01
6.27253890e-01 -1.57255918e-01 -3.25166106e-01 7.93213546e-02
-8.94440055e-01 1.36522663e+00 4.72167641e-01 2.79747218e-01
-5.60095131e-01 -4.52750593e-01 -4.44481760e-01 -2.36141443e-01
1.39559150e-01 -1.17454685e-01 8.26296449e-01 -8.11582923e-01
-8.24715495e-01 9.63171184e-01 -1.32791236e-01 -3.57967466e-01
7.79397249e-01 -2.15320423e-01 -9.35445011e-01 2.29033269e-02
2.72202164e-01 -3.28379601e-01 1.17233917e-01 -1.40420794e+00
-6.83179080e-01 -5.12209773e-01 -2.84099989e-02 2.60694288e-02
2.88966686e-01 3.99116993e-01 4.45799917e-01 -8.78767133e-01
4.82515752e-01 -1.69151753e-01 -3.39160293e-01 -1.71582028e-01
-2.43811756e-01 -1.53773636e-01 5.00649393e-01 -6.55949712e-01
1.42235410e+00 -1.93450582e+00 -2.43523955e-01 7.16730118e-01
6.61779284e-01 -2.96990484e-01 5.65950990e-01 5.22240341e-01
3.47251520e-02 6.48231328e-01 8.32340214e-03 8.57605159e-01
1.69752553e-01 6.58064708e-02 -3.24904472e-01 8.54819655e-01
-1.37877136e-01 7.03671575e-01 -1.14906406e+00 -2.00224444e-01
-2.39327759e-01 1.63090616e-01 2.54276961e-01 -2.63611168e-01
5.03508568e-01 1.83402568e-01 -7.05944240e-01 7.48017192e-01
6.32264435e-01 -6.47117794e-01 1.01618513e-01 5.72339535e-01
-6.49673104e-01 4.48847115e-01 -9.07084465e-01 1.77934229e+00
8.74639601e-02 1.20322776e+00 1.28853276e-01 -5.80385327e-01
1.26210213e+00 2.35273361e-01 1.36096582e-01 -6.96788430e-01
2.81136602e-01 4.36921060e-01 2.02083111e-01 -5.11713147e-01
9.51223671e-01 -5.91398358e-01 -2.30248779e-01 8.33884597e-01
-7.38057792e-01 -5.72725654e-01 1.98128492e-01 3.54112208e-01
1.19217837e+00 4.48575109e-01 2.22083688e-01 -7.25119472e-01
-1.17081692e-02 1.00431609e+00 5.50521493e-01 6.83159292e-01
-1.39509842e-01 2.88788885e-01 4.55687821e-01 -1.06738663e+00
-1.44057190e+00 -1.32870507e+00 -5.09902775e-01 7.46542752e-01
3.66993427e-01 -4.07466024e-01 -5.08965373e-01 -2.87397467e-02
-1.20599292e-01 8.93325806e-01 -5.27794778e-01 8.91220346e-02
-4.32248592e-01 -2.92773277e-01 3.68438900e-01 8.96471590e-02
5.63155472e-01 -6.46861672e-01 -9.25219357e-01 2.45904729e-01
3.93881023e-01 -6.28671288e-01 4.95991111e-02 1.54143453e-01
-9.30294991e-01 -8.00197721e-01 -5.73536336e-01 -5.73035181e-01
4.65529501e-01 2.98480000e-02 1.22908187e+00 3.07597697e-01
-7.03024343e-02 -1.33464679e-01 -1.83675379e-01 -3.28099310e-01
-4.85786051e-01 -5.96981645e-02 -1.72559544e-01 -7.40581632e-01
7.94874310e-01 -9.55780625e-01 -3.68138641e-01 1.17667831e-01
-2.71953523e-01 4.74593677e-02 4.43737328e-01 4.10159200e-01
-8.00328776e-02 4.55219090e-01 7.24948704e-01 -8.85496140e-01
3.15594494e-01 -5.35447001e-01 -3.43148321e-01 2.70067930e-01
-4.78157222e-01 -2.78081268e-01 6.50794864e-01 1.36076555e-01
-7.43840098e-01 -9.42752719e-01 3.14369977e-01 2.23254725e-01
-2.47950226e-01 3.28865886e-01 -2.33901553e-02 -2.93317527e-01
1.02037907e+00 -7.44960550e-03 -1.04258895e-01 -6.65904999e-01
1.40799835e-01 6.68464482e-01 6.86054170e-01 -5.09616137e-01
1.14403856e+00 4.00514066e-01 -7.67570874e-03 -1.00539005e+00
-1.91615626e-01 -8.74579418e-03 -7.49596953e-01 -3.65369469e-01
5.36531448e-01 -6.25498176e-01 -5.27603507e-01 -2.23340079e-01
-1.28475261e+00 -2.17652410e-01 -4.17415470e-01 8.54508758e-01
-3.52247983e-01 -7.51665384e-02 -4.62507963e-01 -7.18660712e-01
4.47068155e-01 -5.74466467e-01 -1.33166134e-01 5.49908757e-01
-8.37012291e-01 -1.00525188e+00 8.20027441e-02 -5.11996448e-02
2.72730052e-01 6.46863759e-01 9.16469753e-01 -6.76180482e-01
-7.17648208e-01 -2.84170389e-01 -4.16757494e-01 -2.10994959e-01
2.80092776e-01 7.33167976e-02 -5.44441640e-01 2.32490152e-01
7.88085759e-02 1.87173896e-02 3.20486585e-03 -2.18761727e-01
6.80009007e-01 -8.08103740e-01 -3.49474162e-01 5.55793166e-01
1.80967140e+00 8.66586506e-01 7.34576404e-01 1.08850050e+00
3.69990915e-01 3.88311327e-01 3.93975496e-01 2.56121367e-01
5.83950818e-01 -1.60472363e-01 -1.37459949e-01 -9.19314250e-02
3.25497806e-01 -5.08196831e-01 -2.72550911e-01 9.00305569e-01
-5.71783185e-01 3.16276878e-01 -1.71336055e+00 6.83634162e-01
-1.59957850e+00 -8.27304482e-01 -7.11128831e-01 1.87956929e+00
6.86991692e-01 6.61361873e-01 -8.42120498e-02 1.96875677e-01
4.54395056e-01 7.73828626e-02 -1.40431136e-01 -8.18062007e-01
6.54436927e-03 1.77884847e-01 5.84529698e-01 5.72957098e-01
-6.81284845e-01 7.14885116e-01 8.41340828e+00 5.69834530e-01
-8.97489727e-01 -1.03677392e-01 5.05171001e-01 1.85225770e-01
-5.46520948e-01 4.72782314e-01 -1.55101314e-01 5.49440026e-01
6.50399446e-01 -6.77477598e-01 4.51974481e-01 6.84927881e-01
3.34442794e-01 -5.31637788e-01 -7.21588790e-01 5.79676747e-01
-5.81255667e-02 -1.42539477e+00 -1.14674851e-01 -1.12095103e-02
6.38999820e-01 -2.45890543e-01 7.29106972e-03 -3.79620939e-01
7.74531722e-01 -1.75481069e+00 6.90054417e-01 9.16042507e-01
3.81740719e-01 -9.48585331e-01 4.25471961e-01 2.46499464e-01
-5.91061473e-01 2.10784882e-01 -3.02664101e-01 -7.89390683e-01
2.40506455e-01 2.25680590e-01 -8.68449569e-01 2.96218932e-01
7.41092086e-01 4.80823606e-01 -6.77086174e-01 1.05432725e+00
-6.96653873e-02 5.08682668e-01 -2.04590559e-01 -4.81832683e-01
6.72196746e-01 -5.83680511e-01 9.36139941e-01 1.24013972e+00
7.81996846e-02 4.43482310e-01 -5.17722726e-01 1.05815864e+00
4.99974072e-01 -3.31787840e-02 -9.24159050e-01 -6.08047605e-01
4.33393627e-01 9.19386506e-01 -1.02697885e+00 -3.47026795e-01
-9.31487530e-02 3.04926813e-01 -3.09464425e-01 5.61599731e-01
-4.45699871e-01 -9.05716896e-01 5.33565283e-01 8.62694561e-01
-2.90424675e-01 -4.98937070e-01 -1.13325870e+00 -5.93723536e-01
-2.03543454e-01 -8.55801880e-01 -1.84655935e-01 -9.65932488e-01
-9.76572692e-01 3.44587386e-01 -2.55356163e-01 -8.29297423e-01
-6.31253943e-02 -6.06439948e-01 -9.95802045e-01 9.50473905e-01
-8.03604484e-01 -7.55450308e-01 5.62396869e-02 -1.15406953e-01
1.86200112e-01 -2.74504364e-01 4.54420328e-01 -1.24539837e-01
-3.06920409e-02 3.03303182e-01 4.05522197e-01 1.08229324e-01
1.01111876e-02 -1.15365469e+00 4.86550897e-01 7.88571715e-01
-8.21619481e-02 7.59734452e-01 1.18036139e+00 -6.53490067e-01
-1.25832617e+00 -2.03489751e-01 1.22437119e+00 -5.88738263e-01
1.30968261e+00 -6.14679083e-02 -1.33147073e+00 9.67978179e-01
3.64712805e-01 -9.36748147e-01 3.69355410e-01 1.90659687e-01
-4.84196059e-02 1.21086195e-01 -1.03986049e+00 7.57810712e-01
6.63404226e-01 -3.56841862e-01 -8.99603069e-01 2.77111501e-01
6.81598723e-01 -2.27324758e-02 -9.81413662e-01 -2.81864733e-01
7.81511903e-01 -9.51718688e-01 7.33962357e-01 -8.18082690e-01
5.41680157e-01 -3.34646851e-01 2.99865939e-02 -1.15942848e+00
-4.35058743e-01 -1.00791144e+00 1.04815102e+00 1.08906603e+00
6.04656219e-01 -8.26062202e-01 7.78774381e-01 8.63735378e-01
-1.54678021e-02 -5.22329152e-01 -7.96495259e-01 -7.05817461e-01
5.03624797e-01 -4.38153207e-01 8.99652123e-01 1.63904917e+00
9.33675170e-01 5.49985394e-02 5.23074493e-02 -2.53149658e-01
8.27080905e-01 -5.10060899e-02 6.59900367e-01 -1.05760467e+00
-3.38944495e-02 -1.02547419e+00 -8.71448517e-01 -6.29122257e-01
-6.69150710e-01 -7.57630467e-01 -1.09357916e-01 -1.62063408e+00
-1.38127536e-01 -7.50867426e-01 -3.24815512e-03 -1.00544415e-01
3.40538204e-01 9.56886336e-02 1.59944102e-01 3.49513203e-01
-2.52124220e-01 -3.36291082e-02 1.52619290e+00 1.40427083e-01
-1.42816871e-01 -3.25119406e-01 -1.04729390e+00 1.40734351e+00
1.17257655e+00 -2.79748291e-01 -1.73591405e-01 -2.63160646e-01
7.56128788e-01 7.80227333e-02 7.47650981e-01 -1.27790999e+00
5.08872628e-01 -6.80755973e-01 8.31301689e-01 -4.37325478e-01
2.06210464e-01 -8.98102939e-01 7.61611462e-01 5.45684576e-01
-1.07326224e-01 4.21675205e-01 7.61461928e-02 9.77854431e-03
-6.26752600e-02 -2.44344994e-01 6.04242682e-01 -3.53274196e-01
-6.98722243e-01 -3.01266283e-01 -2.08529085e-01 9.98661220e-02
9.95528042e-01 -7.14361250e-01 -5.14495850e-01 -3.28736633e-01
-4.74734724e-01 2.49546200e-01 1.01921999e+00 -1.55528888e-01
7.83111453e-01 -1.30752814e+00 -5.93870521e-01 -9.27780047e-02
-4.89575386e-01 8.36257860e-02 -8.86504501e-02 8.20608914e-01
-1.26383996e+00 2.82516301e-01 -5.83336949e-01 -2.57319599e-01
-4.89374071e-01 3.77176106e-01 7.39247724e-02 3.55162561e-01
-1.29236865e+00 8.71646583e-01 -6.20093718e-02 6.67981878e-02
-7.24759698e-03 -4.16866168e-02 -3.48125175e-02 -1.94807172e-01
7.87753284e-01 9.62452590e-01 -6.19548738e-01 -5.78568757e-01
-3.26530069e-01 6.55717134e-01 1.50630176e-02 -3.30704480e-01
1.54670167e+00 -2.94265211e-01 -4.62383658e-01 1.22818875e+00
8.91533017e-01 -1.93989858e-01 -6.59075379e-01 1.42743796e-01
3.25102955e-01 -1.20146728e+00 -8.17478523e-02 -9.99814034e-01
-2.24834993e-01 7.13898063e-01 1.54854149e-01 6.01227403e-01
7.94216275e-01 -5.60128205e-02 6.74182892e-01 8.58822986e-02
6.17163360e-01 -1.26441693e+00 -4.79963481e-01 3.07095945e-01
9.40572560e-01 -8.25905979e-01 3.94684494e-01 1.88856572e-01
-7.38379002e-01 9.76335704e-01 1.53984666e-01 -5.32791197e-01
7.44081497e-01 3.48586589e-01 -7.27269128e-02 -4.79154021e-01
-5.16095877e-01 4.54495817e-01 -2.90653050e-01 5.98936081e-01
4.21210289e-01 3.28463256e-01 -5.35595715e-01 4.65678751e-01
-9.01265502e-01 3.44648451e-01 7.66625524e-01 1.03192365e+00
-1.16718388e+00 -4.17230695e-01 -1.00223458e+00 1.23469949e-01
-4.29376602e-01 2.19516277e-01 -4.58387464e-01 1.42871308e+00
3.02653164e-01 6.61510348e-01 3.56574684e-01 -4.19877917e-01
2.46160552e-01 8.69601816e-02 3.48199785e-01 1.70551866e-01
-5.52101016e-01 -7.09084496e-02 6.08197674e-02 -3.69454145e-01
1.21575452e-01 -4.04009998e-01 -1.45564675e+00 -1.57795393e+00
2.66042978e-01 8.03769469e-01 8.16648126e-01 6.07094586e-01
-1.87752202e-01 1.61466390e-01 5.90960324e-01 -3.01305830e-01
-4.00431380e-02 -9.11089301e-01 -1.30454159e+00 -3.09384882e-01
4.40356642e-01 -3.78238738e-01 -2.98519492e-01 -1.04554869e-01]
|
[8.906855583190918, 6.264492034912109]
|
bb23a7e3-c11b-47f5-b1f2-40176f7946f5
|
rethinking-alignment-in-video-super
|
2207.08494
| null |
https://arxiv.org/abs/2207.08494v2
|
https://arxiv.org/pdf/2207.08494v2.pdf
|
Rethinking Alignment in Video Super-Resolution Transformers
|
The alignment of adjacent frames is considered an essential operation in video super-resolution (VSR). Advanced VSR models, including the latest VSR Transformers, are generally equipped with well-designed alignment modules. However, the progress of the self-attention mechanism may violate this common sense. In this paper, we rethink the role of alignment in VSR Transformers and make several counter-intuitive observations. Our experiments show that: (i) VSR Transformers can directly utilize multi-frame information from unaligned videos, and (ii) existing alignment methods are sometimes harmful to VSR Transformers. These observations indicate that we can further improve the performance of VSR Transformers simply by removing the alignment module and adopting a larger attention window. Nevertheless, such designs will dramatically increase the computational burden, and cannot deal with large motions. Therefore, we propose a new and efficient alignment method called patch alignment, which aligns image patches instead of pixels. VSR Transformers equipped with patch alignment could demonstrate state-of-the-art performance on multiple benchmarks. Our work provides valuable insights on how multi-frame information is used in VSR and how to select alignment methods for different networks/datasets. Codes and models will be released at https://github.com/XPixelGroup/RethinkVSRAlignment.
|
['Chao Dong', 'Yujiu Yang', 'Xintao Wang', 'Liangbin Xie', 'Jinjin Gu', 'Shuwei Shi']
|
2022-07-18
| null | null | null | null |
['video-super-resolution']
|
['computer-vision']
|
[ 1.92786977e-01 -2.15899020e-01 -3.22981447e-01 -1.65358275e-01
-4.52016681e-01 -3.72254074e-01 2.59215802e-01 -6.03750944e-01
-7.77713731e-02 5.18386066e-01 3.60672861e-01 -1.75609514e-01
-3.76811973e-03 -7.51250565e-01 -7.40621448e-01 -7.47876942e-01
2.32451454e-01 2.01397222e-02 5.29330552e-01 -6.10539079e-01
2.14846656e-01 4.89866793e-01 -1.49256325e+00 3.38258713e-01
8.76638591e-01 8.14262271e-01 6.83325350e-01 4.50988442e-01
-1.41825169e-01 7.85339713e-01 -4.84557152e-01 -2.64551669e-01
4.17722404e-01 -3.65684927e-01 -7.86421418e-01 1.79404784e-02
6.46457434e-01 -4.16844994e-01 -7.25797117e-01 1.17461002e+00
5.80805123e-01 -1.49369333e-02 -4.14886512e-03 -1.08655918e+00
-1.02652752e+00 1.00665307e+00 -8.70566428e-01 6.00979686e-01
3.74626517e-01 1.10378802e-01 1.17817485e+00 -8.01259398e-01
5.44045508e-01 1.09166873e+00 5.86187243e-01 6.66160405e-01
-9.69796360e-01 -9.45983052e-01 4.81212020e-01 6.36872947e-01
-1.50878453e+00 -6.99936569e-01 7.94802725e-01 -2.19491661e-01
6.97346330e-01 6.17883265e-01 6.84580743e-01 1.22594237e+00
-1.23634420e-01 6.25940502e-01 9.40872312e-01 -1.29585294e-02
-2.51288027e-01 -2.34568775e-01 -2.81298980e-02 5.31935096e-01
8.73747841e-02 -1.16204135e-01 -7.21804678e-01 2.84079611e-01
1.37205327e+00 1.03309164e-02 -6.95194602e-01 -1.14445329e-01
-1.41395700e+00 5.30334771e-01 4.31808412e-01 5.93199790e-01
-9.41976085e-02 1.09891355e-01 2.82209218e-01 2.65138537e-01
4.14322376e-01 3.68282378e-01 -2.71175057e-01 -2.68359538e-02
-8.24075997e-01 -1.44093364e-01 1.28215253e-01 1.07685351e+00
6.38724506e-01 9.13845524e-02 -1.46877438e-01 9.38964128e-01
-6.84817135e-02 3.77353251e-01 6.80024266e-01 -1.27253246e+00
5.37751198e-01 2.74680108e-01 2.19558161e-02 -1.26143169e+00
-3.05563867e-01 -5.08372903e-01 -1.08496761e+00 -4.03108336e-02
1.27528653e-01 1.13323361e-01 -6.54440463e-01 1.71367490e+00
1.52555645e-01 3.98376733e-01 -1.33380756e-01 1.24418533e+00
1.00318313e+00 5.74242711e-01 -3.96048784e-01 -1.97897553e-01
1.17341948e+00 -1.31635344e+00 -8.37073684e-01 -1.08012103e-01
2.73261130e-01 -8.17391455e-01 1.20076263e+00 2.44213149e-01
-1.31642258e+00 -7.49613225e-01 -1.00593781e+00 -2.30059430e-01
2.05080330e-01 1.18063636e-01 5.06906927e-01 2.81736702e-01
-1.24953306e+00 7.81323195e-01 -8.30123484e-01 -3.39320213e-01
4.52259958e-01 3.05947572e-01 -4.19423610e-01 -3.90541777e-02
-1.19035947e+00 6.57931387e-01 -6.58234358e-02 2.98789144e-01
-6.59139872e-01 -5.31170726e-01 -5.28183043e-01 7.57657737e-02
5.57418406e-01 -6.58625841e-01 9.81677413e-01 -1.23355293e+00
-1.49118626e+00 6.74434066e-01 -3.24435264e-01 -3.41931522e-01
5.55690944e-01 -4.26657826e-01 -4.82661486e-01 1.40397549e-01
-6.15730286e-02 4.13086325e-01 8.28766584e-01 -1.16024590e+00
-6.10097647e-01 -1.99885741e-01 3.19241285e-01 3.38410944e-01
-3.73762578e-01 1.23539776e-01 -7.87958920e-01 -8.49927366e-01
3.16295028e-01 -8.59316289e-01 -2.24009141e-01 -1.30187199e-01
-4.65123862e-01 1.17304824e-01 7.24528849e-01 -7.12752640e-01
1.49130547e+00 -2.16301584e+00 3.70997161e-01 -1.75866559e-01
5.75228691e-01 3.62912834e-01 -3.47802907e-01 7.36015216e-02
-3.09659272e-01 2.19378725e-01 1.49047062e-01 -1.13329932e-01
-4.38442707e-01 6.68109953e-02 -3.29335868e-01 3.28123480e-01
-1.77125350e-01 7.64918923e-01 -7.44177997e-01 -4.12789166e-01
1.13778122e-01 4.97546107e-01 -7.35645831e-01 1.80352256e-01
7.30926767e-02 6.91223860e-01 -4.78043050e-01 6.06225014e-01
6.01341069e-01 -5.55087507e-01 2.30053082e-01 -8.27646434e-01
-3.73515040e-01 3.65029693e-01 -1.09408116e+00 1.79661202e+00
-2.77810037e-01 7.97859967e-01 1.02893770e-01 -1.05746090e+00
9.50903356e-01 1.97003216e-01 7.11680233e-01 -7.51799226e-01
3.86732109e-02 2.76102894e-03 1.27215106e-02 -5.53799629e-01
6.53685927e-01 3.54715645e-01 4.06358331e-01 5.60595058e-02
-2.30549499e-01 5.70155084e-01 6.47128373e-02 1.55047715e-01
1.13345850e+00 1.29160643e-01 3.03870440e-01 -2.08586648e-01
5.30894816e-01 -2.14973301e-01 9.24658537e-01 6.07631505e-01
-1.70558468e-01 8.19729686e-01 2.90839493e-01 -5.08739233e-01
-1.12873852e+00 -9.91704345e-01 4.78491299e-02 1.03760290e+00
5.14906168e-01 -7.14775860e-01 -6.16181374e-01 -4.70364094e-01
-4.06931758e-01 1.77052498e-01 -4.47092086e-01 6.04973249e-02
-9.83941734e-01 -7.42749631e-01 2.97264934e-01 7.42406368e-01
7.08579481e-01 -7.32306480e-01 -4.64491457e-01 8.48803744e-02
-6.88487351e-01 -1.30916214e+00 -8.15307021e-01 -3.70567918e-01
-8.32466602e-01 -1.00535536e+00 -7.04523325e-01 -6.07359350e-01
6.70056880e-01 8.88768494e-01 1.08410013e+00 3.10138911e-01
6.81272298e-02 1.50163844e-01 -6.09265566e-01 2.59284586e-01
-1.46686837e-01 1.54828414e-01 7.38075301e-02 -1.09702749e-02
6.62965700e-02 -1.00517809e+00 -8.70761096e-01 8.19283962e-01
-6.11895919e-01 5.56253910e-01 6.37471378e-01 6.45623386e-01
7.28614450e-01 -4.50298609e-03 4.21882391e-01 -7.86831558e-01
1.49845719e-01 -2.01541007e-01 -5.50451577e-01 2.10401341e-01
-2.77653307e-01 8.40213336e-03 6.93039358e-01 -5.48912346e-01
-8.44077170e-01 -2.20587611e-01 -2.92715430e-01 -8.67359459e-01
1.34526670e-01 2.98533998e-02 -4.20917511e-01 -1.10604867e-01
3.15563381e-01 2.95825601e-01 -1.28544509e-01 -6.93996429e-01
2.95090407e-01 5.79634547e-01 5.67751408e-01 -4.73379165e-01
8.12547386e-01 6.35525703e-01 -2.44778603e-01 -7.71204293e-01
-7.26426482e-01 -3.07263911e-01 -3.23518127e-01 -2.50103891e-01
8.62496853e-01 -1.01086378e+00 -5.83547175e-01 3.96409512e-01
-8.97452950e-01 -4.72503215e-01 -1.95449460e-02 2.67340273e-01
-4.55363363e-01 5.54105043e-01 -7.64501274e-01 -1.98795110e-01
-4.07166302e-01 -1.51440561e+00 8.01133931e-01 2.99447566e-01
1.71591099e-02 -5.17377913e-01 -1.42535612e-01 6.74682260e-01
6.95141673e-01 -2.46241659e-01 4.88409668e-01 -4.04728115e-01
-9.83418405e-01 4.69792008e-01 -4.36529607e-01 2.63821900e-01
3.30823600e-01 1.29221097e-01 -7.75786221e-01 -4.03635144e-01
-1.04233608e-01 2.56638825e-01 8.25106859e-01 4.04521883e-01
1.60879636e+00 -4.18024868e-01 -2.62602150e-01 9.74260867e-01
1.33056784e+00 2.82984078e-01 9.25555050e-01 4.69916403e-01
1.09749150e+00 2.47740254e-01 6.66478217e-01 3.98475289e-01
4.80600297e-01 1.12011290e+00 5.55850685e-01 -1.28778666e-01
-2.84746468e-01 -7.34835714e-02 5.65603495e-01 1.28577888e+00
-6.55890465e-01 -2.04700083e-01 -5.97481072e-01 3.38550448e-01
-1.92006409e+00 -1.14064610e+00 -2.55328510e-02 2.13784933e+00
7.89480150e-01 3.91748622e-02 2.04620678e-02 -6.06848858e-04
1.02021062e+00 5.18785536e-01 -4.71388906e-01 3.16165835e-02
-3.10710341e-01 -3.74384634e-02 6.37398899e-01 3.95464480e-01
-9.71599340e-01 9.80373561e-01 5.90018940e+00 9.65364277e-01
-1.24555850e+00 1.56290084e-01 4.66824740e-01 -2.35963300e-01
-3.54941249e-01 -5.44222668e-02 -9.26587284e-01 6.46623671e-01
5.51538587e-01 -1.52616963e-01 7.35132575e-01 7.01418757e-01
3.60991091e-01 3.69692147e-01 -8.93741071e-01 1.16241467e+00
4.83845547e-02 -1.57839072e+00 8.96589980e-02 -5.92196845e-02
5.24607062e-01 1.50408357e-01 8.43304843e-02 -4.16583046e-02
1.42768875e-01 -9.32071269e-01 6.80821002e-01 3.61996114e-01
8.69540393e-01 -4.48607892e-01 5.71068585e-01 -2.05234900e-01
-1.55490232e+00 -5.45987636e-02 -5.42562962e-01 1.78088218e-01
3.40574205e-01 5.76578677e-01 -1.18215062e-01 8.19139540e-01
1.12017918e+00 1.07912624e+00 -6.49248600e-01 8.44169199e-01
-1.59532219e-01 2.59702384e-01 -1.33071989e-01 4.98472691e-01
-1.82029679e-01 -3.37506205e-01 7.86906540e-01 8.73037577e-01
6.54199719e-01 2.06772700e-01 -4.97563332e-02 6.15851820e-01
-2.21539717e-02 -1.67763811e-02 -3.15718919e-01 1.63548291e-01
7.55712569e-01 1.19773674e+00 -5.52986145e-01 -3.42751950e-01
-7.51516044e-01 1.01936805e+00 2.29258880e-01 2.71893233e-01
-1.27596617e+00 5.95588759e-02 1.10597694e+00 4.10955817e-01
5.57357311e-01 -1.51876375e-01 -4.46883701e-02 -1.51013029e+00
3.85723971e-02 -1.17794824e+00 2.53662854e-01 -1.01046097e+00
-9.94425178e-01 8.62629294e-01 -2.77767241e-01 -1.48889911e+00
4.29998368e-01 -2.39854753e-01 -4.47762430e-01 4.08819228e-01
-1.53054929e+00 -8.14187765e-01 -6.35897636e-01 6.92165554e-01
7.63757408e-01 -2.94489842e-02 3.68965685e-01 7.20326185e-01
-1.05684197e+00 6.78498864e-01 -6.27965331e-02 2.76385278e-01
8.24344814e-01 -9.09834325e-01 5.03505409e-01 9.03371692e-01
1.49707302e-01 5.72529137e-01 7.54561961e-01 -3.61308426e-01
-1.38993561e+00 -1.03780413e+00 2.82115340e-01 -1.74269840e-01
6.51200831e-01 -1.04859732e-01 -1.08101571e+00 9.95748639e-01
2.53850758e-01 1.51139244e-01 4.26082075e-01 3.93755175e-03
-3.66528988e-01 -3.98677289e-01 -7.28131771e-01 8.83665323e-01
1.63769782e+00 -3.86544615e-01 -3.86874288e-01 1.07382618e-01
1.01153338e+00 -5.51891446e-01 -9.87807691e-01 5.66048205e-01
4.70472813e-01 -1.32460475e+00 1.25848937e+00 -2.70868033e-01
5.90028346e-01 -5.35864055e-01 -3.51943135e-01 -1.10684144e+00
-6.19564116e-01 -7.37301111e-01 -1.71491206e-01 1.36618769e+00
1.67029575e-01 -6.68837130e-01 6.99756086e-01 3.26127589e-01
-2.58531332e-01 -8.26546907e-01 -8.52366149e-01 -8.26594710e-01
-2.84138888e-01 -1.09535739e-01 7.53069818e-01 1.15171301e+00
-2.52857238e-01 3.06183904e-01 -6.70542002e-01 4.36283767e-01
5.70040762e-01 2.56127864e-01 7.00771451e-01 -8.45684946e-01
-4.69918787e-01 -5.20815134e-01 -4.09025103e-01 -1.44739199e+00
-1.01958685e-01 -5.12317538e-01 -3.70185941e-01 -1.42538393e+00
2.29568005e-01 -6.73464537e-01 -3.76491636e-01 4.59520072e-01
-1.92062050e-01 3.30733240e-01 3.80173415e-01 4.06595141e-01
-6.58602476e-01 5.11370242e-01 1.48885512e+00 8.13108236e-02
-1.00422427e-01 -1.86665505e-01 -8.26853096e-01 7.79874802e-01
9.88379359e-01 -8.84952769e-02 -3.32337171e-01 -7.85177410e-01
1.45788789e-01 1.26412123e-01 1.99192405e-01 -9.38042104e-01
1.46417215e-01 -2.43341327e-01 1.39883280e-01 -4.61147606e-01
2.69593805e-01 -7.66317129e-01 5.12834966e-01 1.53855011e-01
-1.72582239e-01 5.01124322e-01 -4.24037836e-02 2.92430580e-01
-1.26531690e-01 1.80855617e-01 8.92429650e-01 -1.45330459e-01
-9.14178133e-01 4.69436646e-01 -9.15071964e-02 -1.54289398e-02
7.77723610e-01 -2.26275310e-01 -6.28650188e-01 -3.24277133e-01
-6.95203722e-01 1.41067594e-01 7.98650444e-01 5.80383718e-01
5.24306059e-01 -1.34202671e+00 -6.48386598e-01 6.80535659e-02
-3.40789616e-01 7.85563886e-02 7.03701496e-01 1.04139709e+00
-4.61442500e-01 1.91289604e-01 -3.63461733e-01 -5.92153251e-01
-1.64094377e+00 7.84151733e-01 3.59871298e-01 -1.50470257e-01
-1.06353009e+00 8.34081411e-01 5.63157320e-01 1.91457495e-01
1.06908850e-01 -1.72781259e-01 -3.47182900e-01 -3.01733911e-02
6.78410888e-01 4.39934701e-01 -1.16778933e-01 -7.62975216e-01
-3.13561916e-01 9.41241264e-01 -2.15842471e-01 3.15199614e-01
1.31340969e+00 -6.20497346e-01 -1.00489050e-01 4.95040007e-02
9.23799992e-01 2.75366873e-01 -1.43345654e+00 -3.06349635e-01
-4.84104753e-01 -7.67268836e-01 1.81206316e-02 -3.10208738e-01
-1.83247578e+00 6.50959671e-01 4.41747695e-01 9.85490307e-02
1.56882393e+00 3.75347398e-02 8.84673834e-01 -4.30384092e-02
4.95499223e-01 -7.67622173e-01 9.43660885e-02 1.26460508e-01
9.55437779e-01 -9.80332017e-01 1.23764098e-01 -6.31382287e-01
-6.48305476e-01 9.12700891e-01 9.08914149e-01 -1.00881957e-01
3.15683514e-01 3.66185784e-01 9.17323772e-03 -6.89393055e-05
-8.94288898e-01 -1.45125523e-01 8.10840353e-02 4.41875100e-01
4.55210507e-01 -1.23495944e-01 -2.95458496e-01 6.12069190e-01
-2.55070478e-01 -1.54164031e-01 7.42904484e-01 5.68118513e-01
-3.47008228e-01 -1.11497962e+00 -3.16190571e-01 2.04290956e-01
-6.13298535e-01 -1.83862165e-01 -8.15218538e-02 6.82248056e-01
-1.65285412e-02 7.49189854e-01 2.20777914e-02 -6.96381748e-01
3.72090757e-01 -4.94531274e-01 6.01381540e-01 -2.31409505e-01
-2.56318331e-01 1.50208801e-01 8.83602053e-02 -1.04793596e+00
-6.86321855e-01 -4.15950745e-01 -9.21781659e-01 -6.24742270e-01
-3.71051431e-01 -3.59925553e-02 7.17308894e-02 6.91954434e-01
6.28231823e-01 8.56765687e-01 5.96423626e-01 -9.66325939e-01
-2.39651844e-01 -7.06894338e-01 -3.10195714e-01 2.45937675e-01
3.03209603e-01 -6.58868968e-01 -3.21086317e-01 6.95177540e-02]
|
[11.064596176147461, -1.8610283136367798]
|
5af0c09b-08e3-4ea3-a936-cede99c57c3f
|
the-bussgang-decomposition-of-non-linear
|
2005.01597
| null |
https://arxiv.org/abs/2005.01597v1
|
https://arxiv.org/pdf/2005.01597v1.pdf
|
The Bussgang Decomposition of Non-Linear Systems: Basic Theory and MIMO Extensions
|
Many of the systems that appear in various signal processing applications are non-linear, for example, due to hardware impairments such as non-linear amplifiers and finite-resolution quantization. The Bussgang decomposition is a popular tool for analyzing the performance of systems that involve such non-linear components. In a nutshell, the decomposition provides an exact probabilistic relationship between the output and the input of a non-linearity: the output is equal to a scaled version of the input plus uncorrelated distortion. The decomposition can either be used to compute exact performance results or lower bounds where the uncorrelated distortion is treated as independent noise. This lecture note explains the basic theory, provides key examples, extends the theory to complex-valued vector signals, and clarifies some potential misconceptions.
|
['Emil Björnson', 'Özlem Tuğfe Demir']
|
2020-05-04
| null | null | null | null |
['misconceptions']
|
['miscellaneous']
|
[ 2.79079080e-01 -3.44442934e-01 -2.10467577e-01 -1.75745517e-01
-5.50527334e-01 -8.96338105e-01 2.61090755e-01 -1.70173950e-03
-2.39577796e-02 5.53163886e-01 1.28494278e-01 -3.83400887e-01
-2.63278604e-01 -3.30794364e-01 -4.86929178e-01 -1.04626656e+00
-4.32869107e-01 -4.19549085e-02 1.58025231e-02 -2.97683120e-01
6.84088916e-02 4.63845223e-01 -1.08216667e+00 1.43109471e-01
3.58405501e-01 1.05265951e+00 -3.46740693e-01 1.05115294e+00
4.61598992e-01 4.63559538e-01 -1.04030836e+00 -1.86839744e-01
4.76970196e-01 -6.58673406e-01 -1.96204647e-01 -1.07464381e-01
3.37419480e-01 -1.91065222e-01 -6.97984040e-01 1.55988729e+00
4.21098292e-01 -7.82695487e-02 8.06132436e-01 -1.18516243e+00
-3.69270235e-01 8.82235825e-01 -3.42461824e-01 4.01561469e-01
2.37290710e-01 -5.37721030e-02 7.75738895e-01 -6.52265072e-01
-2.51320660e-01 1.24603069e+00 8.36209595e-01 1.73409909e-01
-1.58734465e+00 -6.13232195e-01 -3.42296928e-01 2.29479000e-01
-1.72789383e+00 -5.98086894e-01 3.33748877e-01 -3.79470468e-01
7.32838631e-01 5.97081125e-01 2.32094556e-01 6.27821207e-01
6.47904694e-01 5.14527977e-01 7.98089087e-01 -5.57378829e-01
3.00493926e-01 1.80469692e-01 2.79302180e-01 3.77094626e-01
4.01692510e-01 2.72659034e-01 -3.16141635e-01 -4.02450562e-01
8.30937266e-01 -3.12137991e-01 -6.90516949e-01 -1.37997419e-01
-9.33550239e-01 4.32998568e-01 3.16931695e-01 5.85740507e-01
-1.46714538e-01 4.62148249e-01 1.97745025e-01 4.67211574e-01
-1.18733756e-01 5.07441044e-01 -2.48558879e-01 -1.41657472e-01
-1.16254938e+00 -3.37834992e-02 8.41600060e-01 8.41743588e-01
3.78439218e-01 5.84566116e-01 7.74169788e-02 2.43995368e-01
2.77960509e-01 5.92106760e-01 5.12838662e-01 -7.99646676e-01
3.95377666e-01 -9.51769203e-02 -4.85903770e-02 -9.29833174e-01
-4.86459345e-01 -8.11865509e-01 -1.27016008e+00 2.84248590e-01
6.57835782e-01 -2.55609274e-01 -6.26524508e-01 1.70947289e+00
-4.90165859e-01 6.53125122e-02 2.32755795e-01 7.31640995e-01
2.21305698e-01 1.03727949e+00 -4.07976836e-01 -5.47770381e-01
1.07733011e+00 -3.24750394e-01 -1.24716389e+00 -1.38188958e-01
1.50715962e-01 -1.01352811e+00 6.49340212e-01 7.00791776e-01
-1.10779762e+00 -4.10745054e-01 -1.54110515e+00 1.34292915e-01
8.01507533e-02 1.72819942e-01 3.84058058e-02 9.11433041e-01
-1.07880425e+00 6.22805119e-01 -7.47038126e-01 3.11366767e-01
-8.25734809e-02 4.56251383e-01 -7.91975297e-03 2.96620671e-02
-1.15127969e+00 9.36243832e-01 1.67656764e-01 1.92527220e-01
-4.01880562e-01 -6.49533987e-01 -4.79037732e-01 4.22671556e-01
1.81944445e-01 -4.40564007e-01 1.23834014e+00 -8.93981934e-01
-1.46851182e+00 2.66965441e-02 -2.04928175e-01 -4.83366728e-01
3.60743791e-01 -8.57102498e-02 -6.24053478e-01 -6.58440888e-02
-3.48823607e-01 -5.15242040e-01 8.58469784e-01 -8.14447105e-01
-2.05296725e-01 -3.10703635e-01 -2.50010967e-01 6.93055317e-02
9.60773453e-02 -1.91812769e-01 -1.48034349e-01 -8.01812768e-01
5.74082732e-01 -8.30884576e-01 -3.78913045e-01 -2.23818108e-01
-7.28399992e-01 3.04899246e-01 4.98542994e-01 -3.08675408e-01
1.42260134e+00 -2.57646513e+00 -1.63515043e-02 5.81184566e-01
-4.17865776e-02 2.46319205e-01 2.46529996e-01 6.38980806e-01
-3.15783411e-01 8.64540935e-02 -8.94253999e-02 1.13628596e-01
8.16934779e-02 1.57223269e-02 -4.88447338e-01 9.73028958e-01
-3.17964926e-02 4.83847141e-01 -6.76741779e-01 4.82467897e-02
2.87036955e-01 5.12049139e-01 -1.45163253e-01 -3.61500159e-02
4.77154553e-01 -5.51240779e-02 7.31432214e-02 1.68106988e-01
7.19359577e-01 4.37987782e-02 2.25924879e-01 -4.98970181e-01
-4.39526215e-02 4.01931316e-01 -1.57267570e+00 1.06851017e+00
-2.58417189e-01 1.28051889e+00 2.41974443e-01 -1.02315056e+00
6.16679847e-01 6.49458170e-01 8.94960687e-02 -3.24567586e-01
3.23804557e-01 1.60935894e-01 4.20737714e-01 -2.97876954e-01
3.23108464e-01 -4.74493116e-01 -9.58995298e-02 7.31655583e-02
2.28038818e-01 -3.78764868e-01 2.20053233e-02 2.05363154e-01
1.03619063e+00 -7.43666708e-01 8.68916750e-01 -5.33367157e-01
3.95436674e-01 -4.99712408e-01 5.74012816e-01 7.11474895e-01
-2.25566939e-01 5.97932637e-01 7.32440114e-01 2.43624553e-01
-9.39311683e-01 -1.42594111e+00 -3.26106399e-01 3.88914734e-01
2.76637048e-01 -3.90071005e-01 -4.70310718e-01 2.03482792e-01
-1.60863847e-01 7.16521680e-01 -1.14086628e-01 -3.92991871e-01
-2.06316948e-01 -7.19558537e-01 9.45846438e-01 5.25307119e-01
2.68608540e-01 1.19014442e-01 -3.29476476e-01 1.68043390e-01
6.72751665e-02 -1.12810421e+00 -5.77954292e-01 5.83335221e-01
-8.43238473e-01 -8.32176268e-01 -4.27592039e-01 -3.86939853e-01
6.14321232e-01 1.31404400e-01 7.37393200e-01 -2.21542969e-01
-8.55183229e-02 1.59203663e-01 3.12139094e-01 -1.90234780e-01
-4.28312868e-01 -7.07857847e-01 6.06907129e-01 -1.86752975e-02
-6.45190403e-02 -6.01408362e-01 -3.91068637e-01 2.76856095e-01
-8.35596204e-01 -4.50917482e-01 5.43631971e-01 7.55000293e-01
4.78459597e-01 7.03949332e-01 3.92743170e-01 -3.83100778e-01
7.60211647e-01 -3.41433764e-01 -7.93158472e-01 1.08465627e-01
-3.74187917e-01 3.93097550e-01 1.16941869e+00 -6.47676826e-01
-5.45363128e-01 2.23960564e-01 -2.58727074e-01 -2.56583720e-01
8.67164284e-02 3.17193657e-01 -5.09328783e-01 -1.81591973e-01
9.07963514e-01 1.97373748e-01 -3.55469286e-01 -1.50143236e-01
3.75578910e-01 7.57671833e-01 9.18789864e-01 -1.52666926e-01
1.06079066e+00 1.45202667e-01 4.05385673e-01 -1.15291035e+00
-3.50226611e-01 -2.62161881e-01 -2.93568701e-01 -5.43607911e-03
1.31514832e-01 -7.77687430e-01 -6.38102770e-01 9.22884867e-02
-1.10844219e+00 1.13331430e-01 -4.85452384e-01 8.76935899e-01
-5.51377237e-01 1.20532103e-01 -6.16412103e-01 -1.15248668e+00
2.89058201e-02 -1.18075109e+00 3.81824374e-01 3.74878138e-01
-2.00844720e-01 -8.23751926e-01 -4.84811105e-02 -3.85773927e-01
4.07246649e-01 1.23997986e-01 1.02425265e+00 -5.29214799e-01
-4.48940307e-01 -6.36556566e-01 1.93383083e-01 7.00574279e-01
-2.11061407e-02 2.30253473e-01 -8.46372962e-01 -2.41018727e-01
4.19232905e-01 3.75452489e-01 3.65704596e-01 5.69462299e-01
7.15584099e-01 -5.52355647e-01 -1.41618341e-01 5.20053029e-01
1.68540430e+00 2.85008818e-01 5.67429841e-01 -3.41974407e-01
9.84849930e-02 1.25199452e-01 -5.03999973e-03 3.07869047e-01
-2.82199144e-01 5.54609358e-01 2.23299831e-01 1.22434311e-01
1.73343375e-01 2.94362474e-02 4.34822977e-01 1.03845870e+00
1.51046038e-01 -3.64706755e-01 -4.50527608e-01 1.58662662e-01
-1.37748039e+00 -1.11067879e+00 -5.83846450e-01 2.48037434e+00
7.01559722e-01 2.59281337e-01 -9.96377543e-02 8.39751601e-01
6.12681806e-01 -2.20471561e-01 -4.28953588e-01 -6.12208486e-01
-3.15806836e-01 1.68709010e-01 9.56778944e-01 6.22697234e-01
-7.32337415e-01 6.41593486e-02 8.22507763e+00 7.79286444e-01
-1.15841424e+00 -1.13191813e-01 3.51991713e-01 -7.31319264e-02
4.54875827e-03 -2.26944223e-01 -4.64560032e-01 4.73407954e-01
1.27382576e+00 -7.94169664e-01 3.59600186e-01 5.40606678e-01
2.08571896e-01 -2.66411036e-01 -1.38167393e+00 1.05362558e+00
-9.16985422e-02 -9.45333838e-01 -3.58733714e-01 2.41005253e-02
6.18994057e-01 -3.93604636e-01 5.26981235e-01 -9.25300922e-03
3.21842320e-02 -1.01280320e+00 6.29039466e-01 4.72086161e-01
6.62625909e-01 -7.44049489e-01 8.24481130e-01 5.08789957e-01
-1.05144334e+00 -1.53149024e-01 -3.13536435e-01 -4.46647018e-01
1.91718489e-01 9.50747192e-01 -5.32772243e-01 3.27892810e-01
1.51669696e-01 -4.21183743e-02 -5.53612262e-02 1.38808286e+00
-3.31098258e-01 7.46581137e-01 -6.14858389e-01 -7.08494261e-02
-5.17804101e-02 -1.52205750e-01 7.83410668e-01 1.26301455e+00
4.24261808e-01 5.08861005e-01 -2.81902194e-01 5.92268348e-01
9.81646702e-02 -1.99826598e-01 -4.25765783e-01 1.39590442e-01
5.97992659e-01 8.78963649e-01 -4.66043770e-01 -3.63576353e-01
-3.85828167e-01 7.98991621e-01 -4.36235100e-01 5.48995376e-01
-6.36974871e-01 -8.27858388e-01 9.44472373e-01 8.86437595e-02
-1.61102608e-01 -5.04024863e-01 -5.50346911e-01 -9.77217376e-01
-9.08997506e-02 -7.86574006e-01 -5.25178835e-02 -5.81533313e-01
-9.28979874e-01 4.88885254e-01 -6.26365542e-02 -1.49665439e+00
-6.05817199e-01 -4.65543687e-01 -4.20987725e-01 1.05985653e+00
-8.46888781e-01 9.57836732e-02 9.07744095e-02 3.78353715e-01
7.84266517e-02 -5.36587127e-02 9.64857757e-01 3.80056381e-01
-5.46081781e-01 7.98570395e-01 6.58773422e-01 4.38239425e-02
3.94751906e-01 -1.29565275e+00 9.07034576e-02 1.29234123e+00
2.50213087e-01 7.36662209e-01 1.41215086e+00 7.47754648e-02
-1.56614375e+00 -6.46055102e-01 6.40892565e-01 -7.94016793e-02
7.13913560e-01 -2.93211132e-01 -8.29987586e-01 5.62351882e-01
2.04330191e-01 1.15722992e-01 8.65769029e-01 -3.75177890e-01
-2.55267322e-01 -2.51682967e-01 -1.10369205e+00 5.75506508e-01
2.84152865e-01 -6.14864349e-01 -6.67407930e-01 1.14053553e-02
2.67081469e-01 -7.49448478e-01 -6.68636501e-01 1.18613012e-01
6.77071869e-01 -1.03596151e+00 1.01619422e+00 -2.64867276e-01
1.56902611e-01 -5.31996012e-01 -6.28072798e-01 -1.54862142e+00
-3.65061015e-01 -8.54355752e-01 -1.26719967e-01 8.47519219e-01
2.93466091e-01 -5.25938690e-01 3.00712675e-01 5.42843223e-01
-4.79437076e-02 -4.97475445e-01 -1.23849392e+00 -1.17095745e+00
-5.25244065e-02 -4.96491760e-01 2.35479891e-01 5.85564911e-01
5.84546685e-01 4.20149863e-01 -3.03231567e-01 6.65317833e-01
6.74310446e-01 -3.16097885e-01 2.35790670e-01 -9.50217783e-01
-4.41727668e-01 -6.35698199e-01 -1.01150811e+00 -1.52406299e+00
-3.51830363e-01 -5.04805624e-01 2.40109205e-01 -1.02907121e+00
-4.27803665e-01 -1.98934734e-01 -3.13213527e-01 -6.96733668e-02
1.35897949e-01 3.62648845e-01 2.18440399e-01 -1.72186382e-02
-1.36595629e-02 1.81977227e-01 6.53097034e-01 -4.47210483e-02
-2.35740811e-01 6.61141157e-01 -6.80820584e-01 8.13137591e-01
6.64433837e-01 -4.36260939e-01 -3.59149992e-01 -1.61579520e-01
2.65128970e-01 3.80818546e-01 9.66486409e-02 -1.43833280e+00
5.68421423e-01 9.02253762e-02 3.41280043e-01 -4.47078645e-01
4.40739304e-01 -1.20611429e+00 3.00853133e-01 5.95501900e-01
-3.94962192e-01 1.54947315e-03 2.45039985e-01 5.12066722e-01
-4.33566093e-01 -4.79239285e-01 9.83889163e-01 3.31029415e-01
-1.03140725e-02 -2.24397629e-01 -8.36360395e-01 -4.17639278e-02
8.10419083e-01 -3.67182419e-02 -2.91685641e-01 -8.09066355e-01
-5.87831736e-01 -2.72161979e-02 -8.19344074e-02 -1.12456255e-01
3.91853422e-01 -1.39515483e+00 -5.11347890e-01 3.22186589e-01
-4.79190022e-01 -4.12090182e-01 1.06007241e-01 8.50822568e-01
-3.32032174e-01 7.03569949e-01 -4.44381125e-02 -5.03797293e-01
-1.43293929e+00 2.63709247e-01 6.36501014e-01 9.40814987e-02
-1.63634002e-01 7.92261541e-01 5.85167482e-02 4.73569423e-01
3.48170251e-01 -6.85293257e-01 2.62164712e-01 -2.10153744e-01
9.11767602e-01 4.45799887e-01 1.10341519e-01 -5.31432152e-01
-4.40288603e-01 5.09432495e-01 4.19027954e-01 -4.66402650e-01
7.47869790e-01 -2.13301182e-01 -1.73512269e-02 9.71390665e-01
1.40719199e+00 1.06961429e-01 -7.32930839e-01 -4.86601114e-01
-2.50309497e-01 -4.50998127e-01 1.52455524e-01 -5.45809269e-01
-8.93457413e-01 1.14311624e+00 8.19315195e-01 6.55805945e-01
1.51593137e+00 -3.19201112e-01 2.61227906e-01 4.48323697e-01
3.23298901e-01 -8.05606008e-01 -1.45923302e-01 4.68304694e-01
8.42699587e-01 -4.84912843e-01 2.28379354e-01 -3.35628390e-01
-2.13479668e-01 1.41703808e+00 4.47055250e-02 -3.67700249e-01
8.51639569e-01 9.90828991e-01 -3.97111140e-02 5.51969647e-01
-6.54724300e-01 1.67283088e-01 4.41308737e-01 5.43790400e-01
6.09376490e-01 2.15651259e-01 -3.56308818e-01 8.76058161e-01
-4.71124947e-01 -2.95436502e-01 1.06782019e+00 4.27533299e-01
-3.49386692e-01 -7.74120271e-01 -8.14490914e-01 3.98626506e-01
-6.84181452e-01 -2.38613307e-01 -1.62038729e-02 4.76393431e-01
-9.90520194e-02 9.85465348e-01 2.18730882e-01 -6.96856618e-01
5.74827850e-01 -6.96489878e-04 4.47274178e-01 -2.94867247e-01
-4.26009357e-01 3.35859030e-01 -3.19707870e-01 -4.31734651e-01
4.53955568e-02 -6.01727903e-01 -1.19775558e+00 -4.95967239e-01
-6.12304807e-01 1.82483226e-01 7.19460368e-01 9.56575871e-01
-1.99051406e-02 9.22201216e-01 6.22852087e-01 -5.71658432e-01
-1.01185811e+00 -7.56680787e-01 -8.89175713e-01 -1.93200603e-01
9.63948905e-01 -1.55053928e-01 -6.98913157e-01 3.53275314e-02]
|
[15.396178245544434, 5.471036434173584]
|
c019530b-4f76-43c4-bea1-6babec2bdb72
|
document-level-relation-extraction-with-cross
|
2303.03912
| null |
https://arxiv.org/abs/2303.03912v1
|
https://arxiv.org/pdf/2303.03912v1.pdf
|
Document-level Relation Extraction with Cross-sentence Reasoning Graph
|
Relation extraction (RE) has recently moved from the sentence-level to document-level, which requires aggregating document information and using entities and mentions for reasoning. Existing works put entity nodes and mention nodes with similar representations in a document-level graph, whose complex edges may incur redundant information. Furthermore, existing studies only focus on entity-level reasoning paths without considering global interactions among entities cross-sentence. To these ends, we propose a novel document-level RE model with a GRaph information Aggregation and Cross-sentence Reasoning network (GRACR). Specifically, a simplified document-level graph is constructed to model the semantic information of all mentions and sentences in a document, and an entity-level graph is designed to explore relations of long-distance cross-sentence entity pairs. Experimental results show that GRACR achieves excellent performance on two public datasets of document-level RE. It is especially effective in extracting potential relations of cross-sentence entity pairs. Our code is available at https://github.com/UESTC-LHF/GRACR.
|
['Fujun Hua', 'Ling Tian', 'Lizong Zhang', 'Zhao Kang', 'Hongfei Liu']
|
2023-03-07
| null | null | null | null |
['document-level-relation-extraction']
|
['natural-language-processing']
|
[-1.59088597e-01 4.11001503e-01 -3.36171657e-01 -3.39562595e-01
-5.96459329e-01 -6.41961753e-01 4.54712570e-01 7.59634852e-01
-1.40762106e-01 6.52026832e-01 5.07213116e-01 -3.11448932e-01
-4.78509992e-01 -1.26709473e+00 -3.67724150e-01 2.94921417e-02
-5.77895567e-02 2.67516941e-01 4.90075380e-01 -4.78758425e-01
7.84051046e-02 1.35296956e-01 -9.02280629e-01 5.60335696e-01
1.24524939e+00 4.60389942e-01 -8.29495564e-02 3.83993626e-01
-6.25356495e-01 1.04757643e+00 -7.35509753e-01 -1.14621043e+00
-2.24575594e-01 -4.47107166e-01 -1.23851109e+00 -1.40942439e-01
3.17006335e-02 1.41743183e-01 -5.22507191e-01 1.34722590e+00
1.66674003e-01 -9.45617929e-02 3.69749814e-01 -1.28088999e+00
-9.10021365e-01 1.32353008e+00 -6.66069150e-01 2.40027726e-01
6.76858246e-01 -6.46427691e-01 1.53803968e+00 -9.15084124e-01
1.01249719e+00 1.23160386e+00 6.91393435e-01 1.20551325e-01
-6.61471605e-01 -4.98912752e-01 3.67076814e-01 3.54262799e-01
-1.61202705e+00 1.33981099e-02 7.84550548e-01 -1.99754804e-01
1.38942730e+00 4.86042112e-01 4.41131353e-01 5.62652886e-01
6.71620294e-02 7.80851603e-01 5.95130682e-01 -3.43580663e-01
-2.35333875e-01 4.88004833e-02 7.70348489e-01 7.00784862e-01
8.25080037e-01 -8.71166766e-01 -3.72847766e-01 -1.17353305e-01
3.09472919e-01 -3.97885442e-02 -2.82651514e-01 5.54091018e-03
-9.03929591e-01 5.66548109e-01 7.67961502e-01 7.94709325e-01
-1.91693008e-01 -4.80078906e-01 5.78407228e-01 1.59476832e-01
6.49404347e-01 4.71997231e-01 -5.32585680e-01 2.52999842e-01
-6.19410515e-01 -6.56849369e-02 9.37520027e-01 1.49589264e+00
4.42824602e-01 -7.56262362e-01 -3.58250141e-01 9.68753636e-01
3.45963657e-01 1.89121261e-01 2.38692492e-01 -4.55396950e-01
1.17539823e+00 1.29574847e+00 -2.98653871e-01 -1.70853078e+00
-4.61302221e-01 -6.39496982e-01 -9.75226462e-01 -8.73682380e-01
-2.45640457e-01 -2.80308962e-01 -2.83693522e-01 1.35799372e+00
5.19083679e-01 2.89209932e-02 3.54217529e-01 5.66170454e-01
1.55841315e+00 4.40663278e-01 3.04508326e-03 -2.40389109e-01
1.68107438e+00 -1.18588567e+00 -9.71510768e-01 -1.40920088e-01
1.14897180e+00 -5.88371992e-01 6.56116068e-01 -2.08443776e-01
-9.33643520e-01 -2.11724401e-01 -9.27778423e-01 -4.71016884e-01
-7.91159213e-01 1.95628956e-01 4.59701955e-01 1.82829037e-01
-7.81464636e-01 4.02951777e-01 -5.36541104e-01 -3.46119344e-01
2.74727494e-01 4.03912738e-03 -4.18688983e-01 -1.61075592e-01
-1.80677259e+00 9.48787272e-01 6.96531653e-01 4.29574251e-01
2.54079849e-01 -6.20824158e-01 -1.10593414e+00 3.49137396e-01
7.97579169e-01 -9.00983632e-01 8.55693042e-01 -9.24440995e-02
-7.37616003e-01 6.04042292e-01 -3.88185561e-01 -1.91464037e-01
1.02509297e-01 -8.77785981e-02 -1.06393301e+00 1.49748370e-01
3.43497962e-01 -1.14304367e-02 -8.98486450e-02 -1.19905901e+00
-6.54605687e-01 -4.72720027e-01 3.64020348e-01 3.48979980e-01
-4.34656948e-01 2.85322875e-01 -8.11590254e-01 -5.87225020e-01
3.30390483e-01 -5.42524338e-01 3.07631362e-02 -7.94282258e-01
-1.11015999e+00 -5.47556162e-01 5.96977055e-01 -8.49026024e-01
2.05676460e+00 -1.64348328e+00 9.50794667e-02 3.03329110e-01
5.89085817e-01 2.07668141e-01 -5.98661453e-02 1.11591709e+00
-1.36871621e-01 5.29136777e-01 -2.52319604e-01 -2.59110302e-01
-3.80566418e-02 2.07228325e-02 -1.37040153e-01 -1.34537056e-01
9.39795598e-02 1.22715926e+00 -1.16329694e+00 -9.28493857e-01
-3.41220498e-01 3.52364421e-01 -2.90713817e-01 -1.79582134e-01
7.09283650e-02 -5.36775067e-02 -8.06757569e-01 5.74321032e-01
7.32758105e-01 -6.17243290e-01 7.06598938e-01 -5.21904707e-01
2.32667953e-01 7.25056887e-01 -9.47499335e-01 1.40652871e+00
-4.07380402e-01 3.21597755e-01 -2.90630162e-01 -6.52277231e-01
9.73129034e-01 5.36374748e-02 3.51839066e-01 -5.45555055e-01
-5.01492321e-02 1.31296605e-01 -9.99041498e-02 -6.92702830e-01
6.66735768e-01 3.53264123e-01 -2.40194887e-01 1.08458266e-01
-6.67441934e-02 1.90864336e-02 7.57360756e-01 9.62777913e-01
1.45146823e+00 -5.52201197e-02 3.98089856e-01 -6.19161762e-02
8.18163633e-01 1.19614571e-01 6.91385627e-01 2.71450460e-01
3.84541005e-01 2.84755558e-01 9.17781889e-01 1.89358249e-01
-5.74264228e-01 -7.86067545e-01 -6.18643723e-02 4.26266372e-01
4.75530893e-01 -1.37749743e+00 -6.28838658e-01 -1.15509677e+00
6.06556460e-02 8.70457292e-01 -3.69346231e-01 -9.53299701e-02
-6.45628452e-01 -6.42212808e-01 5.50192177e-01 4.73429531e-01
6.96978033e-01 -6.67761028e-01 3.70383292e-01 8.13260078e-02
-7.06819534e-01 -1.48635554e+00 -4.13436860e-01 -3.69145274e-01
-7.02252567e-01 -1.40829206e+00 -2.21753165e-01 -8.55056763e-01
1.06233835e+00 1.72768265e-01 1.33181393e+00 3.43829602e-01
1.23928990e-02 5.26356846e-02 -7.17373967e-01 -8.47276598e-02
-1.06084101e-01 3.74844313e-01 -2.95144349e-01 -4.06297386e-01
6.67231560e-01 -4.39732790e-01 -4.89446551e-01 1.90197840e-01
-6.80800378e-01 8.65213573e-02 4.76189762e-01 6.17005885e-01
4.51882005e-01 5.25229752e-01 6.81921482e-01 -1.48292243e+00
1.10454595e+00 -7.97038853e-01 -2.04560772e-01 9.62322295e-01
-8.16489577e-01 -9.71203819e-02 6.19768083e-01 2.42269188e-01
-1.28143573e+00 -5.33849180e-01 -1.95526049e-01 2.47060820e-01
1.49993017e-01 1.16649592e+00 -2.85769969e-01 5.52138746e-01
2.40883768e-01 1.56748686e-02 -6.41539752e-01 -4.65447664e-01
4.41813856e-01 8.79815578e-01 2.98567116e-01 -4.79218036e-01
8.11257780e-01 1.48712024e-01 6.17533475e-02 -3.87918085e-01
-1.32420802e+00 -6.65439010e-01 -9.30598795e-01 -8.07673146e-04
7.18110323e-01 -9.06763732e-01 -5.46048045e-01 2.11772010e-01
-1.35855961e+00 2.10370928e-01 -1.20532751e-01 3.26440632e-01
2.96108693e-01 5.85658669e-01 -9.32732522e-01 -4.37703252e-01
-4.06141222e-01 -7.23762751e-01 1.03187311e+00 3.50714982e-01
-2.66899109e-01 -1.26680803e+00 1.14231974e-01 6.81846082e-01
-1.58721402e-01 2.06555307e-01 9.56606090e-01 -8.15478206e-01
-4.64622051e-01 -3.87153387e-01 -5.47026753e-01 2.86360998e-02
5.87657571e-01 1.84718758e-01 -2.86468476e-01 4.95450050e-02
-4.30864245e-01 2.19819173e-01 7.86351085e-01 -1.74863786e-01
1.00529158e+00 -5.30039191e-01 -7.46280074e-01 2.48337001e-01
1.31164062e+00 -3.82016823e-02 5.16786397e-01 3.31808269e-01
9.93074477e-01 6.81600451e-01 7.37340569e-01 6.78676590e-02
1.18263936e+00 5.83558261e-01 -2.07675640e-02 -5.32687679e-02
-1.74632728e-01 -5.50758958e-01 4.82249968e-02 1.54794705e+00
-1.01899765e-01 -6.75927699e-01 -9.32227850e-01 5.22618175e-01
-2.01505637e+00 -1.00507665e+00 -8.17128062e-01 1.63184988e+00
9.91596997e-01 1.65727362e-01 -1.15558468e-01 2.84605119e-02
8.88963699e-01 1.15832547e-02 -2.25032508e-01 -1.52480647e-01
-2.77336091e-01 -1.65548459e-01 2.52947778e-01 6.16161883e-01
-8.48813534e-01 9.79856730e-01 4.96385431e+00 8.75772774e-01
-2.51316130e-01 1.15363568e-01 1.94159165e-01 3.17863077e-01
-7.77747393e-01 2.38156125e-01 -1.12727308e+00 4.86293823e-01
5.67140579e-01 -5.46386421e-01 -1.44746155e-01 4.89947617e-01
-1.61078587e-01 -7.99351335e-02 -7.91875303e-01 6.23379529e-01
1.43506274e-01 -1.37133014e+00 1.30422682e-01 -1.29670665e-01
6.85136974e-01 -2.18164861e-01 -5.41838586e-01 4.00277495e-01
5.16119659e-01 -6.61084414e-01 1.31034091e-01 5.78630626e-01
5.51354229e-01 -8.30650568e-01 1.18546009e+00 2.78118372e-01
-1.81404233e+00 1.03819832e-01 -2.31806770e-01 3.07019711e-01
3.11836421e-01 9.80369627e-01 -6.08909488e-01 1.57030749e+00
6.18873537e-01 1.13727295e+00 -8.16867173e-01 6.14755929e-01
-7.15439200e-01 2.53463566e-01 -8.07539970e-02 -2.86746114e-01
1.00171514e-01 -4.61183786e-01 4.56346482e-01 1.58880281e+00
2.46631697e-01 3.78371358e-01 -6.08991012e-02 6.32770717e-01
-4.53222811e-01 3.35389823e-01 -5.53234100e-01 -4.15789187e-02
8.68093193e-01 1.61037886e+00 -8.91578972e-01 -3.86918485e-01
-6.80977404e-01 9.48777556e-01 8.04573894e-01 2.53083140e-01
-7.38898754e-01 -1.06418252e+00 4.23361778e-01 -1.40822083e-01
1.41257912e-01 -1.69537991e-01 -9.38871279e-02 -1.55510437e+00
4.58459944e-01 -4.29225475e-01 8.62031579e-01 -8.50669384e-01
-1.52364814e+00 7.44850993e-01 7.16711059e-02 -1.12502313e+00
1.04375988e-01 -2.29506835e-01 -5.37364125e-01 7.05100358e-01
-1.40597296e+00 -1.11271000e+00 -2.33469024e-01 4.89660949e-01
2.50742614e-01 1.30852610e-01 5.86636543e-01 5.12594819e-01
-9.38397884e-01 7.71989346e-01 -1.01702116e-01 7.77776361e-01
2.76412666e-01 -1.29075563e+00 5.36491752e-01 1.06578052e+00
3.35769385e-01 1.17576194e+00 2.92772710e-01 -1.20398510e+00
-1.15010023e+00 -1.21854663e+00 1.68444932e+00 -7.05499887e-01
9.87577081e-01 -4.44654047e-01 -9.91442025e-01 9.87763584e-01
2.86388367e-01 -8.30057636e-02 7.54474282e-01 5.20114303e-01
-3.83646250e-01 7.99372885e-03 -1.01616967e+00 8.02578986e-01
1.59168947e+00 -7.34612823e-01 -8.63662422e-01 5.66690624e-01
1.07957208e+00 -5.61826110e-01 -1.36619437e+00 5.89203954e-01
8.54967684e-02 -5.48678935e-01 9.51885045e-01 -5.96614838e-01
7.18702018e-01 -5.43864965e-01 2.69166738e-01 -1.31519520e+00
-2.68104404e-01 -2.19525665e-01 -5.83827436e-01 1.89324749e+00
9.91494179e-01 -7.75757551e-01 3.90568852e-01 4.61285204e-01
-1.38025790e-01 -1.03125298e+00 -3.81615132e-01 -9.83405173e-01
-1.97520539e-01 -1.20900236e-01 9.09635901e-01 1.24505782e+00
6.77227199e-01 8.32393825e-01 9.19807181e-02 4.83438015e-01
3.44839573e-01 5.02397180e-01 3.35421264e-01 -1.10034561e+00
-6.50674552e-02 -3.42270881e-01 -2.18338922e-01 -1.00992382e+00
2.93942332e-01 -1.35889828e+00 -3.58070225e-01 -2.43140173e+00
4.96644408e-01 -4.44794118e-01 -7.06074238e-02 4.35077816e-01
-4.68801320e-01 -2.94431567e-01 -1.99972130e-02 2.51615822e-01
-8.79907548e-01 5.43066621e-01 1.47401333e+00 -9.66032147e-02
-2.00596694e-02 -2.12901682e-01 -9.36386764e-01 6.83158934e-01
7.58701742e-01 -5.73129654e-01 -6.68827295e-01 -5.39339423e-01
7.07268357e-01 -1.19588105e-02 -9.38565843e-03 -4.77800339e-01
6.80915773e-01 -1.04010522e-01 2.27010157e-02 -7.49942183e-01
-2.88524907e-02 -7.07982838e-01 1.19495407e-01 1.59060493e-01
-5.11750877e-01 1.60413489e-01 -1.75267264e-01 4.93224084e-01
-6.53380215e-01 -3.51586998e-01 -3.23528834e-02 -1.43939601e-02
-4.11468267e-01 3.03184748e-01 2.53656924e-01 3.86178493e-01
9.19984102e-01 1.46485537e-01 -9.52265322e-01 -1.54453084e-01
-6.75566971e-01 7.25270748e-01 5.64930663e-02 5.95682323e-01
5.49724340e-01 -1.37438154e+00 -7.86215782e-01 -4.10928905e-01
2.74142742e-01 3.55406493e-01 4.25725132e-01 8.66896808e-01
-2.94138163e-01 5.87486923e-01 4.94743019e-01 -4.15558442e-02
-1.68662143e+00 5.27147770e-01 7.43574053e-02 -8.06886435e-01
-6.78939342e-01 9.08138156e-01 -2.87257046e-01 -6.91915154e-01
-2.33334646e-01 -3.21137637e-01 -6.81941152e-01 2.12989181e-01
3.68648976e-01 3.44793886e-01 1.85591564e-01 -7.09055245e-01
-6.20429814e-01 6.45375490e-01 -2.37904251e-01 3.38821203e-01
1.19494998e+00 -4.27663147e-01 -7.36883342e-01 2.08003730e-01
1.10912800e+00 5.95127583e-01 -2.79922307e-01 -4.60511744e-01
5.19110024e-01 -1.99936971e-01 -2.22037271e-01 -7.12504745e-01
-1.08530545e+00 4.57724303e-01 -6.21323466e-01 5.65105736e-01
1.08308315e+00 3.70410234e-01 1.02312243e+00 5.30106723e-01
3.78746212e-01 -1.00621331e+00 -3.32480162e-01 7.20677257e-01
9.40427959e-01 -1.01903212e+00 3.49733055e-01 -1.31723571e+00
-7.40736008e-01 1.06034184e+00 7.74353266e-01 2.63186276e-01
8.38975191e-01 3.45953792e-01 -3.32830131e-01 -6.30622983e-01
-5.52135825e-01 -3.80009890e-01 5.41511238e-01 2.43137091e-01
6.33386910e-01 1.05629779e-01 -7.36369669e-01 9.92109299e-01
-3.36956054e-01 -2.50384271e-01 3.54274482e-01 8.72797966e-01
-5.92888147e-02 -1.33356023e+00 4.27698046e-01 4.83110934e-01
-3.46305221e-01 -5.24798214e-01 -8.23487520e-01 7.41990149e-01
6.42908216e-02 1.23651183e+00 -1.05317391e-01 -5.63395321e-01
4.79085684e-01 -1.08519383e-01 1.45149797e-01 -8.49886417e-01
-6.42396152e-01 -5.25616586e-01 8.09710920e-01 -2.49894783e-01
-4.86258209e-01 -6.13387823e-01 -1.73980486e+00 -3.73487890e-01
-5.87468863e-01 4.85231012e-01 2.05211222e-01 9.75894928e-01
8.17761779e-01 8.63804400e-01 6.55656099e-01 3.35033089e-01
7.67257512e-02 -9.84963119e-01 -5.99882364e-01 3.40785295e-01
-1.01280160e-01 -3.70724559e-01 -2.36631334e-01 -2.79534101e-01]
|
[9.226067543029785, 8.61160945892334]
|
cfb5fe49-a5c6-4c57-9808-b8b057f0e937
|
reduction-of-overfitting-in-diabetes
|
1707.08386
| null |
http://arxiv.org/abs/1707.08386v1
|
http://arxiv.org/pdf/1707.08386v1.pdf
|
Reduction of Overfitting in Diabetes Prediction Using Deep Learning Neural Network
|
Augmented accuracy in prediction of diabetes will open up new frontiers in
health prognostics. Data overfitting is a performance-degrading issue in
diabetes prognosis. In this study, a prediction system for the disease of
diabetes is pre-sented where the issue of overfitting is minimized by using the
dropout method. Deep learning neural network is used where both fully connected
layers are fol-lowed by dropout layers. The output performance of the proposed
neural network is shown to have outperformed other state-of-art methods and it
is recorded as by far the best performance for the Pima Indians Diabetes Data
Set.
|
['Jong-Myon Kim', 'Md. Rashedul Islam', 'Akm Ashiquzzaman', 'Abdul Kawsar Tushar']
|
2017-07-26
| null | null | null | null |
['diabetes-prediction']
|
['medical']
|
[ 6.22537546e-02 3.22788298e-01 -4.27904189e-01 -8.01080406e-01
-4.60148811e-01 5.43003321e-01 3.92562337e-02 5.88805914e-01
-5.29927433e-01 1.17819786e+00 2.36096367e-01 -1.57993540e-01
-2.43814200e-01 -6.01175070e-01 -5.16791582e-01 -6.60814047e-01
-2.58475095e-01 6.66750669e-01 -2.57925779e-01 -4.30202082e-04
-1.31412176e-02 4.61337537e-01 -1.08482325e+00 6.42524838e-01
9.74071622e-01 8.91937613e-01 -4.82437193e-01 5.82543314e-01
2.76373088e-01 1.06962025e+00 -4.72182989e-01 -3.08744580e-01
-8.68989453e-02 -1.96591422e-01 -5.73434651e-01 -1.99022204e-01
5.64589083e-01 -6.16791129e-01 -5.14964163e-01 5.13953328e-01
1.01255083e+00 -4.09800947e-01 4.36448127e-01 -1.07288992e+00
-6.44439220e-01 3.95212442e-01 -1.52495682e-01 4.32427227e-01
-2.29468241e-01 -1.16425440e-01 2.76763618e-01 -6.96764469e-01
5.24179578e-01 1.13814092e+00 1.18299699e+00 6.97085559e-01
-1.43179536e+00 -7.48261631e-01 -2.61297673e-01 2.03374159e-02
-1.00888216e+00 -3.83983076e-01 3.84662092e-01 -3.67218375e-01
1.16748559e+00 8.74216557e-02 6.70708537e-01 1.17351389e+00
6.71701074e-01 6.00199163e-01 9.94191229e-01 -2.51305789e-01
1.14333406e-01 2.88347125e-01 6.36654973e-01 6.28256202e-01
5.08862197e-01 3.99679929e-01 -2.99529910e-01 -3.39116842e-01
7.33177006e-01 1.87104568e-01 9.03234109e-02 7.18998462e-02
-8.61434102e-01 8.33657861e-01 4.73670304e-01 1.62377402e-01
-6.44124329e-01 1.62196904e-01 7.69553959e-01 6.67645931e-01
5.84390879e-01 3.77878875e-01 -8.41232300e-01 1.75220326e-01
-9.53968883e-01 1.98426992e-01 7.03236997e-01 4.90230739e-01
-2.35949650e-01 4.26999956e-01 -3.01454753e-01 9.22828853e-01
1.03795230e-01 -1.24044105e-01 5.93043685e-01 -6.52068377e-01
2.39293069e-01 1.07194328e+00 4.37590070e-02 -5.33443213e-01
-8.80392492e-01 -8.27719450e-01 -1.36664224e+00 6.00554943e-01
4.73460615e-01 -5.31586707e-01 -1.46301472e+00 1.27554071e+00
7.68859535e-02 1.79338932e-01 4.29203898e-01 6.47913039e-01
1.07722962e+00 2.70737141e-01 4.61617559e-01 -1.16947263e-01
1.18537545e+00 -8.07866931e-01 -8.19510221e-01 2.04564914e-01
7.49388099e-01 -3.93032968e-01 5.24021447e-01 6.94191813e-01
-1.06265306e+00 -4.83147562e-01 -9.80782747e-01 -3.04181308e-01
-3.42042387e-01 2.60775685e-01 8.61420751e-01 6.11358225e-01
-7.83114135e-01 9.25101221e-01 -9.24919307e-01 -6.38944685e-01
1.00356328e+00 7.72834063e-01 -6.03723586e-01 -2.55450785e-01
-1.24570143e+00 1.16869640e+00 4.71147835e-01 1.56910092e-01
-8.50163817e-01 -1.16323411e+00 -3.34236950e-01 -2.69586593e-01
-3.95643651e-01 -1.11832464e+00 9.63313282e-01 -1.17900026e+00
-1.27834237e+00 9.83025312e-01 1.35814726e-01 -8.23029280e-01
8.62629890e-01 -3.36335659e-01 -6.84819102e-01 -5.34286797e-02
-5.24149418e-01 5.95434427e-01 1.05633117e-01 -7.83769906e-01
-6.45706534e-01 -5.32025695e-01 -5.32697976e-01 8.31709430e-02
-7.54872411e-02 7.62126297e-02 3.20131004e-01 -6.50412679e-01
4.21112590e-02 -5.87679505e-01 -4.49793607e-01 2.04492509e-01
-3.86374503e-01 -1.15262613e-01 6.68455958e-01 -1.00582087e+00
1.19299555e+00 -1.82972908e+00 -3.62059057e-01 -1.04508825e-01
2.43124500e-01 4.78468359e-01 1.77403316e-01 2.83073276e-01
-4.00973618e-01 1.11734398e-01 2.42464215e-01 -8.77057537e-02
-5.69325626e-01 5.53294122e-01 4.34609205e-01 6.54467404e-01
4.53050315e-01 6.61166191e-01 -5.06077528e-01 -2.40908176e-01
7.64305145e-02 9.83189642e-01 -6.82508647e-01 2.96285987e-01
1.47815034e-01 5.53122640e-01 -2.12860107e-01 7.58220851e-01
6.60172880e-01 -1.22598626e-01 -4.91770729e-02 -1.87770635e-01
-1.51371151e-01 1.88566968e-01 -7.26459801e-01 1.33138192e+00
1.26409754e-01 2.69239217e-01 -2.75366068e-01 -1.25947404e+00
1.19368768e+00 6.52026474e-01 4.87012029e-01 -5.90204895e-01
1.86120883e-01 3.81117851e-01 2.75124907e-01 -1.00482476e+00
-2.50696480e-01 -4.77280319e-01 4.65345204e-01 -4.02790844e-01
1.87468186e-01 8.43071401e-01 -1.41390294e-01 -2.23828211e-01
9.52498138e-01 -2.47165617e-02 3.21790129e-01 -3.91407996e-01
1.79720253e-01 2.47863114e-01 1.00059569e+00 7.64483154e-01
-5.22870064e-01 5.98386288e-01 8.06416571e-01 -1.05796146e+00
-1.31037366e+00 -6.97193921e-01 -7.25796580e-01 6.87866747e-01
-7.03557670e-01 2.32807681e-01 -4.90523607e-01 -5.46672285e-01
4.71021026e-01 4.12048787e-01 -9.00629461e-01 -3.77268970e-01
-5.35279691e-01 -1.44785225e+00 8.90021145e-01 7.38513350e-01
4.86988336e-01 -8.58161926e-01 -4.42782700e-01 5.08048773e-01
3.97964180e-01 -6.42538607e-01 3.04351032e-01 6.64036214e-01
-1.72411621e+00 -1.09562373e+00 -1.08258390e+00 -1.13114691e+00
7.16394603e-01 -7.58723617e-01 1.22550178e+00 1.65418759e-01
-5.52539110e-01 -6.00513041e-01 -1.49077922e-02 -8.16238225e-01
-5.03556490e-01 1.38124540e-01 6.08997121e-02 -3.61617506e-01
8.90464306e-01 -5.17112374e-01 -1.00156593e+00 -3.73598605e-01
-4.06116754e-01 1.77991539e-01 9.15269256e-01 1.07205117e+00
5.28811693e-01 -7.17108250e-02 1.10247028e+00 -1.30192661e+00
4.45881188e-01 -6.51834905e-01 -1.69533610e-01 -3.96723710e-02
-1.06177330e+00 -6.61326200e-02 4.21999544e-01 -5.21940887e-01
-5.35889804e-01 1.02363989e-01 -2.12074250e-01 -1.06303382e-03
-4.48556870e-01 5.12914717e-01 1.10627346e-01 -4.40017246e-02
5.23345292e-01 -1.52171165e-01 2.82945275e-01 -1.06298816e+00
-4.98775691e-01 7.54963756e-01 2.48331517e-01 7.83164650e-02
-2.95387864e-01 4.20168159e-04 3.22749317e-01 -6.23722136e-01
-8.32617283e-01 -5.55993058e-02 -5.57253540e-01 7.77009353e-02
7.03607142e-01 -1.01747775e+00 -6.02247715e-01 8.36055338e-01
-7.38540888e-01 -1.28632307e-01 -5.75868599e-02 7.16774106e-01
-1.15980461e-01 -3.25623661e-01 -1.08109856e+00 -6.22345448e-01
-7.28303313e-01 -1.01687968e+00 2.79031157e-01 1.62362665e-01
-2.66179949e-01 -9.81602311e-01 8.53624716e-02 2.60264695e-01
5.34855127e-01 1.04850805e+00 1.33592105e+00 -1.11835063e+00
2.65204847e-01 -6.53116286e-01 -3.80248755e-01 6.68628097e-01
4.45020944e-02 -8.22029859e-02 -8.59898031e-01 -4.08325464e-01
-1.94440141e-01 -2.34034896e-01 8.35447669e-01 8.13519120e-01
9.34521377e-01 -4.45093185e-01 -2.40075842e-01 7.11000621e-01
1.75554132e+00 2.67064095e-01 7.80580342e-01 5.76489449e-01
4.57774550e-01 3.44592482e-01 -5.27592003e-02 3.18372279e-01
3.11842024e-01 9.23409909e-02 3.68690223e-01 -8.43573868e-01
-2.15265512e-01 8.32844526e-02 -6.82634860e-02 4.81308907e-01
-1.51808888e-01 9.65804532e-02 -1.04106653e+00 3.63992482e-01
-1.67510152e+00 -5.51111221e-01 -8.49837124e-01 2.25517178e+00
8.97712648e-01 3.18143815e-01 2.04512313e-01 3.44694525e-01
5.35928726e-01 -6.66128278e-01 -5.79546273e-01 -8.19283962e-01
-1.10501312e-01 3.84682685e-01 5.86618066e-01 3.35073262e-01
-1.21304631e+00 3.30730706e-01 7.55929661e+00 -2.14990109e-01
-1.03393304e+00 1.05488941e-01 1.08255696e+00 -2.36544356e-01
6.71243787e-01 -3.03210497e-01 -7.63221323e-01 5.32030225e-01
1.38021731e+00 3.51443827e-01 -2.48291358e-01 6.90127134e-01
6.87568426e-01 -1.03492968e-01 -1.15246105e+00 6.00426614e-01
-2.26651758e-01 -1.00963247e+00 1.70153767e-01 -1.17512532e-01
6.12391531e-01 1.43678248e-01 -7.20458329e-02 2.88411885e-01
8.00118968e-02 -1.50995755e+00 -1.75428033e-01 9.89132285e-01
8.03407073e-01 -1.07706654e+00 1.49513638e+00 -3.95522378e-02
-9.35167149e-02 -2.17953846e-01 -4.91648465e-01 -3.65585625e-01
-1.37988269e-01 9.07716870e-01 -7.58258164e-01 2.34468952e-01
7.60601580e-01 6.36180997e-01 -4.04266924e-01 1.50364769e+00
4.66452837e-01 9.11703229e-01 -1.26781642e-01 2.44925335e-01
1.69123054e-01 1.10013314e-01 1.32643357e-01 1.51364160e+00
2.05786884e-01 -6.05075993e-02 1.48491859e-02 3.69329512e-01
-1.19429082e-01 1.33044854e-01 -3.76731694e-01 2.42214855e-02
-1.00769117e-01 6.32754982e-01 -4.81495261e-02 -5.56899130e-01
-3.77250701e-01 6.97923958e-01 1.57273486e-01 1.95006087e-01
-4.41291898e-01 -2.97943920e-01 4.36552107e-01 4.37112868e-01
1.77426990e-02 4.67201293e-01 -6.80110157e-01 -7.66599715e-01
-4.57148641e-01 -9.39203024e-01 7.76498556e-01 -5.21872163e-01
-1.34708285e+00 4.60935205e-01 -5.72233737e-01 -1.00885844e+00
3.69586259e-01 -6.79147243e-01 -3.04574043e-01 1.13683844e+00
-1.66207838e+00 -1.12025702e+00 -2.71009326e-01 2.39554435e-01
2.58577317e-01 -3.81859124e-01 1.31950104e+00 8.46732438e-01
-9.49245572e-01 7.43262351e-01 4.41275358e-01 2.28096291e-01
7.51983464e-01 -1.21186423e+00 -2.68042505e-01 2.30264544e-01
-9.58178937e-01 4.25659865e-01 8.39909732e-01 -7.85776258e-01
-9.76844609e-01 -1.36809170e+00 1.39286804e+00 -3.67935821e-02
1.01421803e-01 8.53582099e-02 -1.16891050e+00 6.30005538e-01
1.01446532e-01 3.12579572e-01 1.14893997e+00 1.85366943e-01
1.15132436e-01 -4.48311418e-01 -1.58033764e+00 -1.91407595e-02
2.06134677e-01 2.03566074e-01 -5.86813092e-01 2.28924453e-01
1.95337161e-01 -5.31563878e-01 -1.34929001e+00 7.43233800e-01
8.18776369e-01 -7.60711432e-01 8.84442627e-01 -1.42914510e+00
5.97494960e-01 3.41347724e-01 1.23069078e-01 -9.89269733e-01
-5.30216277e-01 -2.48785466e-01 -3.96080881e-01 1.08787048e+00
5.20017982e-01 -4.91097301e-01 1.07475746e+00 6.21859431e-01
1.12209380e-01 -1.29895222e+00 -9.19784784e-01 -4.45811629e-01
4.61199224e-01 2.03461632e-01 8.25255513e-02 1.15108287e+00
-2.70320356e-01 3.35994989e-01 -5.97695589e-01 8.97153839e-02
6.79895043e-01 -6.27958477e-01 2.06895217e-01 -1.72888982e+00
2.75359184e-01 -8.92933756e-02 -7.37202883e-01 -1.43997297e-02
-6.08394325e-01 -8.54717076e-01 -6.03945255e-01 -1.70755112e+00
2.82492220e-01 -3.29237580e-01 -9.91027296e-01 5.11320591e-01
-8.25673714e-02 2.63087392e-01 -3.04700583e-01 1.64861958e-02
5.86346947e-02 6.91998303e-02 1.10857570e+00 1.08831879e-02
-3.39884639e-01 2.27415055e-01 -7.27005124e-01 4.89598602e-01
1.20125699e+00 -8.17167044e-01 2.71066409e-02 -3.25086415e-01
-1.43709719e-01 5.09831728e-03 4.11357701e-01 -1.22929132e+00
-1.64496601e-01 1.98044106e-02 1.13134503e+00 -5.26856184e-01
-4.83317003e-02 -7.91377425e-01 2.87507206e-01 1.17387116e+00
-6.71770155e-01 1.00392163e-01 1.78267568e-01 3.13780427e-01
-5.43823540e-02 -3.21988063e-03 1.14301383e+00 -2.05811724e-01
-2.19146907e-01 2.62236208e-01 -4.03201997e-01 -1.71511292e-01
8.87789428e-01 -3.81874710e-01 2.29462888e-02 3.47064883e-01
-1.13363421e+00 2.77032733e-01 -5.50078191e-02 2.00872198e-01
5.59476018e-01 -1.33833063e+00 -1.17259407e+00 2.54596144e-01
-8.51549730e-02 -1.64972812e-01 2.06974298e-01 1.10862637e+00
-7.90675223e-01 5.08712232e-01 -5.55764556e-01 -3.00099880e-01
-1.48980224e+00 3.55459005e-01 7.36704111e-01 -3.28395098e-01
-1.21193802e+00 6.76497519e-01 -4.35871333e-01 -2.87862509e-01
8.40767384e-01 -6.45304143e-01 -4.84075576e-01 -1.50021508e-01
6.54009342e-01 7.19643950e-01 2.32948452e-01 -1.16255611e-01
-3.04407358e-01 -1.57135129e-01 -4.45588499e-01 7.24044323e-01
1.84701419e+00 2.48424247e-01 -2.65415534e-02 6.89464509e-01
1.28364205e+00 -9.43342209e-01 -1.05784202e+00 3.60921770e-02
1.80053279e-01 -6.33944795e-02 5.71503997e-01 -1.68769896e+00
-1.12876081e+00 8.69000673e-01 1.51630807e+00 -1.91740647e-01
1.00160122e+00 -6.25960410e-01 8.11989367e-01 2.23802418e-01
-4.47898686e-01 -1.19975424e+00 -3.07869613e-01 2.83964515e-01
8.03553164e-01 -1.65327919e+00 1.92024857e-01 2.19686985e-01
-6.14905655e-01 1.53064597e+00 6.86325014e-01 -7.09782660e-01
8.00629556e-01 5.38254857e-01 4.08869952e-01 -9.37704295e-02
-9.38104093e-01 3.96482766e-01 1.96224049e-01 7.39956260e-01
9.45409834e-01 1.06790915e-01 -7.82619953e-01 8.92077267e-01
2.83754706e-01 1.02417040e+00 3.04604024e-01 8.96268785e-01
-4.06271636e-01 -8.50720525e-01 -8.51717666e-02 1.13336802e+00
-1.24831629e+00 -9.96909440e-02 -2.18330041e-01 1.03965998e+00
3.71154606e-01 4.74655211e-01 -4.78512188e-03 5.34964912e-02
5.50527155e-01 4.57093835e-01 2.34429032e-01 -1.75897494e-01
-1.03388035e+00 -4.96945158e-02 3.18592995e-01 -3.14624935e-01
-3.34298909e-01 -2.26543769e-01 -1.07831848e+00 -4.03243959e-01
-1.24920473e-01 -1.89388022e-01 5.29806256e-01 5.87142050e-01
5.05272329e-01 1.03470659e+00 1.79542914e-01 -1.52255103e-01
-6.30411685e-01 -1.25474346e+00 -5.26444733e-01 2.40695715e-01
7.84950435e-01 -3.98365676e-01 -1.12915099e-01 7.37333521e-02]
|
[8.02283763885498, 5.866252422332764]
|
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.