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 |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
6e474b2a-e422-4395-8d95-7fb581f5eca1 | low-cost-gnss-simulators-with-wireless-clock | 2306.00633 | null | https://arxiv.org/abs/2306.00633v1 | https://arxiv.org/pdf/2306.00633v1.pdf | Low-Cost GNSS Simulators with Wireless Clock Synchronization for Indoor Positioning | In regions where global navigation satellite systems (GNSS) signals are unavailable, such as underground areas and tunnels, GNSS simulators can be deployed for transmitting simulated GNSS signals. Then, a GNSS receiver in the simulator coverage outputs the position based on the received GNSS signals (e.g., Global Positioning System (GPS) L1 signals in this study) transmitted by the corresponding simulator. This approach provides periodic position updates to GNSS users while deploying a small number of simulators without modifying the hardware and software of user receivers. However, the simulator clock should be synchronized to the GNSS satellite clock to generate almost identical signals to the live-sky GNSS signals, which is necessary for seamless indoor and outdoor positioning handover. The conventional clock synchronization method based on the wired connection between each simulator and an outdoor GNSS antenna causes practical difficulty and increases the cost of deploying the simulators. This study proposes a wireless clock synchronization method based on a private time server and time delay calibration. Additionally, we derived the constraints for determining the optimal simulator coverage and separation between adjacent simulators. The positioning performance of the proposed GPS simulator-based indoor positioning system was demonstrated in the underground testbed for a driving vehicle with a GPS receiver and a pedestrian with a smartphone. The average position errors were 3.7 m for the vehicle and 9.6 m for the pedestrian during the field tests with successful indoor and outdoor positioning handovers. Since those errors are within the coverage of each deployed simulator, it is confirmed that the proposed system with wireless clock synchronization can effectively provide periodic position updates to users where live-sky GNSS signals are unavailable. | ['Jiwon Seo', 'Woohyun Kim'] | 2023-06-01 | null | null | null | null | ['outdoor-positioning'] | ['miscellaneous'] | [-9.11018103e-02 1.08141275e-02 1.32587567e-01 -3.80421570e-03
-4.87835795e-01 -6.21609032e-01 -6.95326552e-02 -1.44923449e-01
-3.74141991e-01 9.41496968e-01 -7.79111803e-01 -8.01910520e-01
1.26691721e-02 -1.00768805e+00 -8.81746531e-01 -9.56551373e-01
-3.23670030e-01 2.38366872e-01 5.55851161e-01 -3.06818396e-01
4.44515906e-02 3.71198416e-01 -1.27874565e+00 -8.22887361e-01
1.14895570e+00 8.91402423e-01 5.69434524e-01 7.12028801e-01
1.78429767e-01 -1.49483338e-01 -1.13545966e+00 2.20466033e-01
7.40650222e-02 -4.05040681e-01 2.25174278e-01 -3.97997648e-01
-4.35839355e-01 -2.51977831e-01 -4.83686149e-01 5.02104700e-01
7.80436277e-01 1.03567772e-01 -8.88868123e-02 -1.55951583e+00
2.08871663e-01 3.37786347e-01 -5.63887119e-01 1.08619861e-01
7.72594929e-01 -2.70784140e-01 -2.91897710e-02 5.72029017e-02
3.22093934e-01 3.72813761e-01 1.19997895e+00 1.65696591e-02
-6.12440288e-01 -8.17889750e-01 -2.11240560e-01 1.40588172e-02
-1.83531201e+00 -3.50860655e-01 2.97833890e-01 7.28857666e-02
5.54799676e-01 6.49569392e-01 8.14389169e-01 8.32228422e-01
7.04272509e-01 -2.44900435e-01 5.10988414e-01 -3.43376219e-01
5.77045321e-01 1.42819077e-01 -2.13383242e-01 -2.71586911e-03
1.02410734e+00 -5.52178584e-02 -1.28899723e-01 -3.77063751e-01
6.20822966e-01 -7.04367757e-02 -6.60238445e-01 -6.49677962e-02
-1.18492627e+00 1.97625026e-01 6.14596963e-01 6.01211965e-01
-4.86630142e-01 5.02251983e-01 -1.67332619e-01 3.43189240e-01
4.29440551e-02 -8.81399661e-02 -3.59908491e-01 -3.35966766e-01
-1.02295160e+00 -2.46225998e-01 8.73319566e-01 1.65161037e+00
4.92498308e-01 2.03500748e-01 4.34749722e-01 1.87445492e-01
6.40403807e-01 1.24800038e+00 4.21058506e-01 -6.80694044e-01
4.33759511e-01 -1.79583624e-01 6.19307160e-01 -1.23960292e+00
-6.28675103e-01 -1.14065349e+00 -5.79861939e-01 -5.15091300e-01
3.18278044e-01 -8.79449248e-01 -5.97513020e-01 1.34997952e+00
5.58763266e-01 9.21698213e-01 1.78546444e-01 5.34779489e-01
3.86543363e-01 5.88734984e-01 -3.29443038e-01 -3.12993705e-01
1.03599477e+00 -2.61291504e-01 -8.31668139e-01 -5.67392051e-01
9.79871571e-01 -7.97131538e-01 2.67325222e-01 -1.16736248e-01
-6.66633010e-01 -4.49084371e-01 -1.48266363e+00 8.82770896e-01
3.57799865e-02 8.61511603e-02 -3.31666730e-02 1.34928203e+00
-1.38497174e+00 9.90084931e-02 -1.02793801e+00 -7.27644503e-01
-5.34655869e-01 6.21979952e-01 1.51232362e-01 -8.13976526e-02
-1.59629524e+00 5.49935937e-01 -1.21496536e-01 4.35279161e-01
-8.95932317e-02 -2.57211715e-01 -7.41091013e-01 -1.38931602e-01
-1.22546181e-01 -8.47318769e-01 1.13746274e+00 -6.19342566e-01
-1.39544451e+00 -6.67801723e-02 -2.80784249e-01 -5.93265772e-01
6.11348033e-01 3.66161466e-01 -9.97038960e-01 8.05491433e-02
6.53766692e-01 -2.23239720e-01 1.48338780e-01 -1.27185333e+00
-9.08817112e-01 -4.47240882e-02 -1.59674825e-03 2.37587675e-01
4.80975330e-01 -4.95786339e-01 -4.84237403e-01 -1.03548899e-01
9.54858541e-01 -1.39345598e+00 -3.84521335e-01 -7.02849507e-01
-1.64405420e-01 7.55836129e-01 5.38218856e-01 -3.94849300e-01
1.07899082e+00 -2.25339651e+00 -8.65423083e-01 8.14119756e-01
-4.41192061e-01 -4.26669151e-01 6.02776051e-01 5.23322761e-01
1.85355678e-01 -1.12247221e-01 2.13548526e-01 -3.10208976e-01
-1.26208201e-01 3.35016936e-01 -1.05599619e-01 8.67366314e-01
-1.02284122e+00 2.78642774e-01 -9.07132566e-01 -2.66066374e-04
-6.16961950e-03 4.07986522e-01 -2.16498092e-01 -2.57939667e-01
7.82428265e-01 8.65204632e-01 -5.72792053e-01 4.19223338e-01
1.12730253e+00 5.71609400e-02 3.79177600e-01 3.12823325e-01
-5.43539524e-01 2.11516127e-01 -1.47613502e+00 1.06092405e+00
-5.96112490e-01 4.45476472e-01 2.32960671e-01 -5.82401216e-01
9.20053661e-01 5.43687463e-01 5.18256307e-01 -7.98817039e-01
2.98789322e-01 6.96678162e-01 -2.74673104e-01 -2.61354446e-01
6.28753722e-01 1.31740868e-01 -5.49042225e-01 1.56174108e-01
-3.85231674e-01 1.16068467e-01 -2.12620452e-01 1.25394523e-01
1.24705708e+00 -6.95479987e-03 6.37492761e-02 -4.70533036e-02
5.00507891e-01 -7.50221685e-02 6.40412450e-01 9.71762955e-01
1.72566399e-02 1.55875057e-01 -1.36513829e-01 3.03125650e-01
-5.17750025e-01 -7.65414059e-01 -1.46707371e-01 3.17295223e-01
1.07268631e+00 -4.83700544e-01 -5.81957042e-01 8.49317163e-02
1.68390870e-01 7.33651698e-01 4.93995883e-02 -2.67617088e-02
-4.31130677e-01 -3.86914611e-01 7.72190928e-01 1.09726667e-01
7.35417664e-01 1.50703236e-01 -3.50944251e-01 5.51098406e-01
-5.89430511e-01 -1.16049480e+00 -1.49367169e-01 -2.46842555e-03
-6.39210999e-01 -8.21818113e-01 -5.85761786e-01 -6.11428797e-01
8.31393480e-01 1.21185434e+00 2.57233649e-01 3.72662276e-01
5.75724542e-01 6.43680036e-01 -4.09546494e-01 -3.09064776e-01
5.66957109e-02 4.85008769e-02 3.95209819e-01 7.26979375e-02
-1.03268303e-01 -7.08039701e-01 -8.64733756e-01 1.21762419e+00
-1.03866458e-01 -2.81703502e-01 8.41331184e-02 1.56845048e-01
3.17343891e-01 4.62412566e-01 6.54302180e-01 -2.16303274e-01
-1.68055240e-02 -1.03594911e+00 -8.11474741e-01 -3.12563032e-01
-7.83662796e-02 -8.12364280e-01 3.77554208e-01 1.10733524e-01
-5.95617175e-01 -1.22246407e-01 -1.54115677e-01 6.04600191e-01
9.38306078e-02 5.17689586e-01 -3.84610891e-01 -7.16696262e-01
5.83047569e-01 2.37989813e-01 -4.28334594e-01 1.85011365e-02
-8.86271968e-02 1.06198788e+00 3.55107814e-01 1.07603113e-03
1.07081532e+00 6.51694179e-01 -1.11018658e-01 -1.16340828e+00
3.52694213e-01 -5.37325501e-01 7.49415830e-02 -4.36290920e-01
2.56375074e-01 -1.44348741e+00 -9.58212316e-01 3.99668097e-01
-8.90409887e-01 -2.10473463e-02 3.69151443e-01 8.58553350e-01
-3.38238180e-01 6.31457448e-01 -1.01882644e-01 -1.06447756e+00
1.47593781e-01 -9.14551258e-01 6.18366480e-01 7.58177876e-01
-2.99724519e-01 -7.72771239e-01 -2.54232466e-01 1.83178872e-01
7.61646032e-01 4.64698344e-01 -1.93363219e-01 -1.93395659e-01
-1.04611528e+00 -7.10577071e-01 4.65202183e-01 -6.74397349e-01
1.40723467e-01 -3.87872964e-01 -6.54088259e-01 -6.03519380e-01
3.45898867e-01 8.36189508e-01 -1.83831647e-01 8.72875154e-01
3.39714766e-01 -8.09595883e-02 -1.14883244e+00 8.89295280e-01
1.36935055e+00 7.88422465e-01 7.87142813e-01 9.07548368e-01
2.17079416e-01 -1.42036811e-01 9.78486300e-01 2.97035784e-01
5.87792456e-01 6.04654670e-01 3.86268049e-01 -2.65156105e-02
6.18526816e-01 5.44472151e-02 3.94174993e-01 5.28933704e-01
-6.37761876e-02 -7.87954748e-01 -6.98997200e-01 3.52154315e-01
-1.64146900e+00 -6.43394470e-01 -8.38118732e-01 2.53112197e+00
1.21079125e-01 4.34052438e-01 -3.12793672e-01 3.15377504e-01
1.18569684e+00 -1.78172544e-01 -2.25122392e-01 -1.06428199e-01
8.17844868e-02 -2.26543084e-01 1.65632892e+00 6.59265101e-01
-5.93727946e-01 1.79750636e-01 4.90023088e+00 3.32165599e-01
-9.98109043e-01 2.09893003e-01 -1.10693812e-01 3.50619286e-01
-5.30882180e-01 2.36218169e-01 -8.45168173e-01 1.11252642e+00
1.37333882e+00 -4.00611937e-01 -1.66056067e-01 8.46636355e-01
7.68232286e-01 -7.53142416e-01 -4.04861510e-01 8.55013251e-01
-2.97519356e-01 -1.09567630e+00 -9.32721138e-01 2.98552960e-01
6.95303500e-01 2.06303254e-01 -2.48387977e-01 -1.37725240e-02
-3.74743249e-03 -1.00257315e-01 7.56600022e-01 2.91623354e-01
5.75176537e-01 -9.45336878e-01 1.09709501e+00 4.56155568e-01
-1.67624521e+00 7.49604329e-02 -6.84447661e-02 -1.15281858e-01
7.38302946e-01 6.21017814e-01 -8.93229425e-01 1.20694757e+00
5.25840700e-01 -3.73762399e-02 1.38872191e-02 1.75249314e+00
-3.13918799e-01 5.69593370e-01 -9.87129033e-01 -2.24383473e-02
1.56514466e-01 -3.00385177e-01 7.22025454e-01 5.78732908e-01
1.42314100e+00 7.57999020e-04 -1.24684632e-01 1.03463084e-01
3.34073842e-01 -1.98523745e-01 -7.36343682e-01 8.59098196e-01
1.18413639e+00 7.78746665e-01 -7.61966884e-01 -8.40222090e-02
-2.53604621e-01 6.13536954e-01 -1.11122775e+00 8.73249233e-01
-1.33421767e+00 -1.09808886e+00 4.83781070e-01 6.72711253e-01
2.85501897e-01 -9.68061984e-01 -3.66150618e-01 -5.44621527e-01
1.02035291e-02 -8.17863271e-02 -1.95774555e-01 -7.26373672e-01
-2.83103794e-01 4.64702815e-01 -3.60833555e-01 -1.85006654e+00
-1.77291960e-01 3.63099068e-01 -6.75571680e-01 1.02010393e+00
-1.30864894e+00 -5.84154904e-01 -8.11118782e-01 4.55605239e-01
-2.91154087e-01 1.70647502e-01 4.13095087e-01 7.14301765e-01
-2.94997990e-01 7.05302358e-01 9.36669409e-01 -3.54238391e-01
5.88388681e-01 -7.13490367e-01 7.38871455e-01 1.04874992e+00
-6.69908226e-01 9.07844007e-01 1.01522338e+00 -1.05722952e+00
-1.45527482e+00 -9.43790734e-01 1.05527532e+00 2.19240010e-01
4.17141616e-01 -3.66083860e-01 -2.79246688e-01 8.41087759e-01
-6.95197061e-02 -2.58415192e-01 7.59356856e-01 -4.01204526e-01
9.37183261e-01 -2.57717997e-01 -1.42644179e+00 7.46169448e-01
8.81182909e-01 -2.88487859e-02 1.17321238e-01 2.46797472e-01
5.34838021e-01 -9.89859760e-01 -4.44858670e-01 9.02423263e-02
5.91969728e-01 -5.10817528e-01 7.99787700e-01 7.40809917e-01
-1.02018142e+00 -1.00026333e+00 -1.94830820e-01 -1.30975270e+00
-1.27245232e-01 -1.10587227e+00 3.97755295e-01 1.18178260e+00
2.91652501e-01 -1.47848356e+00 5.92051685e-01 4.82518971e-01
-1.61442369e-01 5.28620109e-02 -1.50890386e+00 -8.39257181e-01
-1.02880824e+00 -5.36489904e-01 9.68226075e-01 6.83476567e-01
-1.47992522e-01 -3.66886891e-02 -1.86150998e-01 1.26437318e+00
6.05624497e-01 -5.20527780e-01 1.24713886e+00 -9.28586245e-01
-1.73932388e-01 4.64432269e-01 -7.37370789e-01 -1.58561778e+00
-4.55551744e-01 -4.31851089e-01 1.85568616e-01 -1.77705073e+00
-1.06253147e+00 -9.49268043e-01 1.78319067e-01 -2.52306998e-01
4.07760620e-01 1.41687229e-01 -4.44866538e-01 2.24694032e-02
-2.49637797e-01 3.13588798e-01 6.14313364e-01 2.69873381e-01
-4.90759611e-01 9.91118252e-01 -5.10274649e-01 6.41852796e-01
8.19513142e-01 -5.80429733e-01 -6.10036612e-01 -3.74702513e-01
4.40846860e-01 5.04277587e-01 2.32882231e-01 -1.54483223e+00
7.81782031e-01 2.28587016e-01 4.05054122e-01 -7.28010237e-01
2.33115166e-01 -1.24027359e+00 9.65364516e-01 7.87504435e-01
9.26588118e-01 -1.07840523e-01 3.57989997e-01 9.48000550e-01
1.93182051e-01 1.21046081e-01 4.15219456e-01 4.28404093e-01
-2.98263609e-01 -1.31873786e-01 -9.50326324e-01 -7.76871443e-01
1.44014990e+00 -9.77814496e-01 -3.74237418e-01 -9.38256621e-01
-5.79982042e-01 6.02096915e-01 6.57135129e-01 -1.62857831e-01
3.20602596e-01 -1.36420655e+00 1.35424763e-01 3.35339844e-01
-1.59536883e-01 -3.19403470e-01 7.32265174e-01 1.25140822e+00
-8.76865566e-01 7.08875716e-01 6.52185902e-02 -8.78236890e-01
-1.14708364e+00 -3.10611963e-01 6.63290620e-01 6.71871841e-01
-3.02522838e-01 6.73691869e-01 -3.31358314e-01 -1.57586783e-01
6.92548752e-02 -6.28337324e-01 1.04850940e-01 -2.95368195e-01
2.78931677e-01 4.53250766e-01 2.47598320e-01 -7.82076240e-01
-8.01720858e-01 8.66177619e-01 6.72896385e-01 -3.26710999e-01
8.30841899e-01 -1.16108859e+00 2.54820883e-01 2.58957773e-01
6.61949217e-01 5.99078774e-01 -8.08232188e-01 9.31308195e-02
-1.48346886e-01 -6.93533182e-01 -2.05719948e-01 -1.34965152e-01
-5.31275868e-01 -1.93600342e-01 7.33901262e-01 4.18736875e-01
9.25478518e-01 -3.29979211e-01 1.01168907e+00 2.80597985e-01
1.43814492e+00 -9.77506876e-01 -8.44789207e-01 3.16786468e-01
8.11817870e-02 -4.38967407e-01 -1.75863862e-01 -3.89407068e-01
-9.47093815e-02 8.85905385e-01 3.30668122e-01 -1.52566750e-02
6.22324169e-01 2.41979659e-01 -6.02346547e-02 1.43488824e-01
1.94556251e-01 -4.81043719e-02 -5.64226866e-01 7.82829106e-01
3.85658853e-02 1.26664251e-01 -5.85564196e-01 7.94708431e-01
-6.00414693e-01 -1.26242191e-01 9.98395920e-01 1.43643081e+00
-7.06940234e-01 -9.92318690e-01 -8.68044019e-01 -1.95548330e-02
-2.59084612e-01 4.14546669e-01 4.56383109e-01 7.91059136e-01
2.40672395e-01 1.44598567e+00 -2.61608846e-02 -4.96119142e-01
6.42424762e-01 -3.84608895e-01 6.37250617e-02 -2.02184886e-01
-7.70319402e-02 2.30131373e-01 4.54925448e-01 -4.40212369e-01
-5.42156436e-02 -5.41222692e-01 -1.67463601e+00 -7.51596987e-01
-5.87121606e-01 7.75564492e-01 1.05867386e+00 8.17393422e-01
4.71843988e-01 7.68027663e-01 1.11734021e+00 -7.57167161e-01
2.19353288e-01 -4.64429855e-01 -8.97880733e-01 -7.49399424e-01
2.45527729e-01 -6.06833935e-01 -5.53028166e-01 -5.63661754e-01] | [6.310115814208984, 1.0392972230911255] |
7a77660f-2a5e-4a32-967f-94de41cc1717 | region-aware-adaptive-instance-normalization | 2106.02853 | null | https://arxiv.org/abs/2106.02853v1 | https://arxiv.org/pdf/2106.02853v1.pdf | Region-aware Adaptive Instance Normalization for Image Harmonization | Image composition plays a common but important role in photo editing. To acquire photo-realistic composite images, one must adjust the appearance and visual style of the foreground to be compatible with the background. Existing deep learning methods for harmonizing composite images directly learn an image mapping network from the composite to the real one, without explicit exploration on visual style consistency between the background and the foreground images. To ensure the visual style consistency between the foreground and the background, in this paper, we treat image harmonization as a style transfer problem. In particular, we propose a simple yet effective Region-aware Adaptive Instance Normalization (RAIN) module, which explicitly formulates the visual style from the background and adaptively applies them to the foreground. With our settings, our RAIN module can be used as a drop-in module for existing image harmonization networks and is able to bring significant improvements. Extensive experiments on the existing image harmonization benchmark datasets show the superior capability of the proposed method. Code is available at {https://github.com/junleen/RainNet}. | ['Xiao Gu', 'Rong Xie', 'Li Song', 'Han Xue', 'Jun Ling'] | 2021-06-05 | null | http://openaccess.thecvf.com//content/CVPR2021/html/Ling_Region-Aware_Adaptive_Instance_Normalization_for_Image_Harmonization_CVPR_2021_paper.html | http://openaccess.thecvf.com//content/CVPR2021/papers/Ling_Region-Aware_Adaptive_Instance_Normalization_for_Image_Harmonization_CVPR_2021_paper.pdf | cvpr-2021-1 | ['image-harmonization'] | ['computer-vision'] | [ 4.27598685e-01 -4.98713329e-02 2.41122693e-02 -3.84286493e-01
-2.98020154e-01 -4.99800682e-01 6.20384216e-01 -3.18831027e-01
-3.04788381e-01 5.18852651e-01 -1.41334802e-01 -1.56060353e-01
3.17697853e-01 -8.78097415e-01 -8.11650991e-01 -8.99435520e-01
6.98825955e-01 1.93228886e-01 3.44242245e-01 -3.58919919e-01
3.43522057e-02 5.12668133e-01 -1.35645545e+00 2.32745737e-01
1.15737665e+00 7.74431527e-01 4.27504808e-01 5.68498492e-01
-2.87131250e-01 5.24945259e-01 -6.07818305e-01 -4.18412387e-01
6.25316620e-01 -7.15444088e-01 -5.73159277e-01 3.76274884e-01
7.61086285e-01 -2.30609357e-01 -2.78933525e-01 1.30622578e+00
5.09077132e-01 1.63877845e-01 3.69916707e-01 -1.48401034e+00
-8.42318416e-01 4.47092265e-01 -9.49774444e-01 -2.77657062e-02
-9.58513394e-02 2.64053375e-01 6.54678643e-01 -6.72333896e-01
5.55121303e-01 1.12628508e+00 2.86594421e-01 5.97500265e-01
-1.27720582e+00 -8.02770376e-01 4.57296759e-01 1.87697023e-01
-1.25674951e+00 -4.54664737e-01 1.02141666e+00 -1.78092688e-01
1.54512718e-01 4.54459876e-01 7.45950162e-01 9.25449491e-01
1.63898006e-01 7.15275109e-01 1.15952611e+00 -4.11429435e-01
1.10289395e-01 1.70019612e-01 -3.85278136e-01 4.22797054e-01
1.34644479e-01 -1.99700356e-01 -2.45131969e-01 1.19904310e-01
9.73300099e-01 -5.20415790e-02 -5.52431345e-01 -5.59180796e-01
-1.09628451e+00 4.13878173e-01 6.65122986e-01 2.11695835e-01
-1.51697949e-01 2.09860146e-01 1.72125846e-01 1.51482582e-01
5.22405088e-01 2.09952816e-01 -1.27937719e-01 3.32854211e-01
-8.67797136e-01 1.78826302e-01 4.21384096e-01 9.37642276e-01
8.43889296e-01 1.60891280e-01 -4.39646542e-01 1.01082706e+00
2.90129334e-02 4.38523978e-01 3.01806480e-01 -1.07723022e+00
3.28750044e-01 6.76632464e-01 1.88597396e-01 -9.92791414e-01
4.10382487e-02 -4.93685186e-01 -1.01863647e+00 5.36481023e-01
3.45992237e-01 -5.04627563e-02 -1.04719210e+00 1.79091430e+00
7.05719411e-01 4.22740102e-01 -2.13857651e-01 1.00202358e+00
8.77183557e-01 6.71125472e-01 1.24016695e-01 -6.66453689e-02
1.22566247e+00 -1.21040487e+00 -7.37691104e-01 -3.55133295e-01
-4.85586412e-02 -1.09602523e+00 1.25486135e+00 1.35983065e-01
-1.27549970e+00 -7.61740208e-01 -1.11932385e+00 -3.00605714e-01
-1.47184744e-01 2.54503280e-01 4.53479826e-01 3.86268854e-01
-1.21957004e+00 3.22579265e-01 -6.00459039e-01 -2.47004583e-01
3.49170059e-01 1.37845770e-01 -3.69446576e-01 8.93465579e-02
-9.37661886e-01 6.84612334e-01 5.13051867e-01 3.28065246e-01
-6.45587504e-01 -6.40118062e-01 -7.52727866e-01 -2.21028253e-02
3.48518610e-01 -9.36312199e-01 1.13987827e+00 -1.84105217e+00
-1.67804921e+00 1.04501033e+00 -6.47227392e-02 8.24212879e-02
8.97393048e-01 -8.46724138e-02 -2.15809733e-01 2.14611441e-02
-1.59025546e-02 7.36884534e-01 1.07860768e+00 -1.78886080e+00
-7.24691510e-01 -1.52489841e-01 1.87392160e-01 4.27240729e-01
-2.63264775e-01 -1.62296861e-01 -1.11566174e+00 -9.50643599e-01
-6.40928671e-02 -9.44347620e-01 -8.94665718e-02 4.06068802e-01
-4.56141591e-01 1.31922841e-01 8.00118387e-01 -7.86859453e-01
1.04763782e+00 -2.13412476e+00 2.89676070e-01 1.71214774e-01
2.59850770e-01 4.07899231e-01 -3.96861404e-01 -7.65362196e-03
-2.69140333e-01 -1.77192867e-01 -4.59269345e-01 -4.15602922e-01
-1.75863802e-01 2.26851270e-01 -1.29956037e-01 5.10094941e-01
1.95273578e-01 9.06787872e-01 -7.80249000e-01 -5.06744146e-01
2.25235119e-01 6.68077409e-01 -4.21590656e-01 5.25516212e-01
-2.45783135e-01 6.94477618e-01 -2.41032913e-01 3.53992879e-01
1.09387100e+00 -1.64898336e-01 2.79976428e-01 -5.82852423e-01
-4.75737117e-02 -1.69300109e-01 -1.22979939e+00 1.69364977e+00
-4.68940347e-01 5.11727512e-01 3.89515370e-01 -7.39307821e-01
8.86955678e-01 4.06087153e-02 4.04515892e-01 -9.69236255e-01
2.64302224e-01 -1.04025029e-01 -1.29872262e-01 -1.49506494e-01
4.32462871e-01 3.24601270e-02 3.05293798e-01 3.66532654e-01
-2.15382218e-01 -2.15363398e-01 6.45483062e-02 5.18212318e-02
4.24962938e-01 3.32689673e-01 2.01590925e-01 -2.69932002e-01
7.22403049e-01 -2.97261000e-01 8.52407038e-01 5.56226492e-01
-3.63456793e-02 9.49996054e-01 3.52333009e-01 -4.38374877e-01
-1.24171472e+00 -1.17052126e+00 1.75923422e-01 1.05742109e+00
5.63586116e-01 -4.06871326e-02 -1.08321309e+00 -3.79155278e-01
-2.59901971e-01 5.55193782e-01 -6.56309187e-01 -1.93504587e-01
-7.31946886e-01 -6.41454041e-01 2.81906664e-01 4.45544630e-01
9.45030570e-01 -1.01101744e+00 -2.97897577e-01 -4.18983400e-02
-3.67687732e-01 -1.00154018e+00 -9.16012108e-01 -1.56312540e-01
-4.57324058e-01 -9.02111411e-01 -8.70894194e-01 -9.97864604e-01
8.04949582e-01 4.41468060e-01 1.06075644e+00 3.53052706e-01
-2.48260722e-01 1.83466494e-01 -2.00158656e-01 -3.08853000e-01
-5.75392723e-01 4.97441292e-02 -3.04804623e-01 4.93909866e-01
-2.49989927e-01 -6.87157929e-01 -8.63178968e-01 3.98443937e-01
-1.35328877e+00 6.56652391e-01 3.62397313e-01 6.63648844e-01
7.04209983e-01 -1.34425452e-02 1.15985878e-01 -9.70107913e-01
4.06229794e-01 -4.56116274e-02 -7.72827089e-01 5.36353648e-01
-3.77147824e-01 -9.12548900e-02 5.01883924e-01 -4.38621461e-01
-1.46163118e+00 1.84259191e-01 -3.39871310e-02 -4.85977739e-01
-4.63937595e-02 -4.94076572e-02 -6.92397356e-01 -2.11292580e-01
3.29160273e-01 2.65428811e-01 -2.76799202e-02 -4.67085004e-01
6.90446079e-01 3.70161563e-01 7.94680357e-01 -6.58469856e-01
1.16311252e+00 7.23550737e-01 -1.95386380e-01 -5.41780174e-01
-6.63449109e-01 -1.04499683e-01 -7.38644779e-01 -2.58269489e-01
9.22152936e-01 -9.43298101e-01 -4.37448263e-01 7.44360209e-01
-1.21344781e+00 -8.62036467e-01 -2.13815749e-01 -5.70892803e-02
-4.06450778e-01 4.37495977e-01 -4.74677116e-01 -3.34874898e-01
-3.93102139e-01 -1.10604870e+00 1.09468329e+00 5.79422116e-01
1.26492053e-01 -9.38114524e-01 5.43421693e-02 2.58628279e-01
5.55994511e-01 3.69347721e-01 8.13061893e-01 1.56323805e-01
-6.26558959e-01 1.28245965e-01 -4.67983842e-01 4.86933112e-01
4.81243730e-01 4.01937455e-01 -1.04135168e+00 -3.26583028e-01
-1.50493115e-01 1.83936402e-01 8.95679474e-01 3.07331502e-01
1.32245636e+00 -2.71231502e-01 -1.72763392e-02 9.86378849e-01
1.47813213e+00 2.94064254e-01 9.46484625e-01 3.99674863e-01
1.08847952e+00 5.15611649e-01 4.07155275e-01 2.90239185e-01
3.29997033e-01 9.53459799e-01 3.48600894e-01 -8.25433493e-01
-4.86824870e-01 -1.62152871e-01 2.06841484e-01 5.73089123e-01
-5.85320070e-02 -3.64029735e-01 -5.98389924e-01 2.97434419e-01
-1.93091047e+00 -8.76156569e-01 5.12104891e-02 2.12704778e+00
1.09308493e+00 -5.11860512e-02 4.56232689e-02 -2.11039543e-01
9.92558241e-01 2.11441308e-01 -5.90567231e-01 -2.34502509e-01
-3.03849548e-01 1.30768195e-01 3.48439217e-01 6.42032623e-01
-1.23055828e+00 1.01639342e+00 5.34578896e+00 9.57605898e-01
-1.52942598e+00 1.99086398e-01 1.02825189e+00 -1.52760997e-01
-5.50421178e-01 -1.34931847e-01 -4.27023351e-01 6.10537052e-01
1.91884682e-01 -1.97112471e-01 6.52692854e-01 5.75801849e-01
3.74714613e-01 -7.78853893e-04 -9.14794087e-01 1.09522569e+00
2.37746146e-02 -1.26330400e+00 3.43961656e-01 -2.96987325e-01
8.81592214e-01 -3.49495083e-01 2.03579456e-01 -1.12728275e-01
2.16001168e-01 -8.40311170e-01 9.77571070e-01 5.79363585e-01
8.31178308e-01 -6.91859782e-01 4.15261388e-01 -9.46859494e-02
-1.38758361e+00 2.49735653e-01 -3.56114835e-01 1.58378422e-01
2.03583494e-01 5.24730444e-01 -2.64818102e-01 5.29355824e-01
7.58642316e-01 6.48658514e-01 -8.05663288e-01 9.63193834e-01
-4.06386137e-01 1.69686496e-01 -6.77799359e-02 5.88769913e-01
-1.34054348e-01 -6.46326482e-01 4.37566757e-01 1.07146204e+00
1.19414173e-01 7.33506028e-03 1.80249244e-01 1.16437554e+00
-2.44004071e-01 1.45189390e-01 -3.80479395e-01 2.43964970e-01
3.40593964e-01 1.55343342e+00 -8.47255230e-01 -3.25873524e-01
-2.81040817e-01 1.39209127e+00 3.09572995e-01 6.56412601e-01
-1.12429094e+00 -2.90153682e-01 8.65585625e-01 3.56471799e-02
2.77477711e-01 3.08770891e-02 -4.69922394e-01 -1.24773765e+00
2.24104881e-01 -1.07100654e+00 1.35721594e-01 -1.03389263e+00
-1.20514131e+00 7.28042841e-01 -1.14275724e-01 -1.23540854e+00
3.71376604e-01 -3.39195967e-01 -9.50373888e-01 9.29128289e-01
-1.35589504e+00 -1.37498832e+00 -8.36196482e-01 7.51454234e-01
3.80357027e-01 9.26548466e-02 4.11662698e-01 4.90471035e-01
-7.44275510e-01 6.77310050e-01 -5.54144830e-02 1.56917363e-01
9.49799359e-01 -1.15962791e+00 4.79445070e-01 1.02275705e+00
-5.98672479e-02 4.59012449e-01 7.35967696e-01 -4.92988378e-01
-1.07485402e+00 -1.34455192e+00 3.20775002e-01 -2.32757255e-01
3.18728507e-01 -4.28976983e-01 -1.12947941e+00 5.45096517e-01
6.75456583e-01 -1.07686520e-01 2.71976709e-01 -5.15145540e-01
-2.60866165e-01 -5.35197496e-01 -9.21968400e-01 9.43732202e-01
9.65894043e-01 -4.06127781e-01 -1.27202511e-01 1.79777458e-01
8.00046980e-01 -6.40449643e-01 -6.85327828e-01 4.35567468e-01
4.34919715e-01 -9.79466617e-01 1.06048787e+00 -2.92680055e-01
4.29944932e-01 -7.65819430e-01 1.38978779e-01 -1.34375715e+00
-3.94371837e-01 -5.10364294e-01 3.89101535e-01 1.50223756e+00
1.95815444e-01 -5.23030221e-01 5.75802982e-01 6.47091031e-01
2.20573833e-03 -4.55709964e-01 -6.66157663e-01 -3.55977863e-01
5.82453907e-02 2.94510573e-02 8.00965786e-01 1.03685856e+00
-6.77192926e-01 1.15286291e-01 -4.47893500e-01 1.82018161e-01
6.04270220e-01 2.70279467e-01 1.06492281e+00 -7.75722384e-01
-4.77263778e-01 -6.15958095e-01 -7.98509717e-02 -8.26372623e-01
1.97751075e-01 -7.02151120e-01 1.34097159e-01 -1.49379277e+00
3.58986408e-01 -4.58673328e-01 -2.65872449e-01 5.57373464e-01
-5.51936448e-01 5.88957787e-01 3.87163758e-01 2.03615557e-02
-3.65096539e-01 7.52801001e-01 1.60471117e+00 -3.85420978e-01
-1.89210296e-01 -1.05706640e-01 -8.61252367e-01 6.29648447e-01
1.14459312e+00 -2.11071014e-01 -5.64125299e-01 -7.01423943e-01
2.30995030e-03 -3.02960277e-01 3.96576107e-01 -7.48044610e-01
9.45732668e-02 -3.77529353e-01 4.81837839e-01 -1.92494705e-01
2.20597908e-01 -8.30460012e-01 5.95631719e-01 2.79562533e-01
-2.67640561e-01 2.40796939e-01 3.90799642e-01 4.02492374e-01
-2.01897174e-01 2.81213839e-02 1.22155762e+00 -1.58527404e-01
-7.63184428e-01 4.81797189e-01 6.46351501e-02 5.15794232e-02
1.05753160e+00 -1.59272388e-01 -3.92557770e-01 -4.31766659e-01
-4.15299088e-01 7.94662982e-02 9.98590171e-01 5.63153803e-01
5.11335790e-01 -1.47672701e+00 -7.55018175e-01 2.70705789e-01
7.34062120e-02 6.45696521e-02 4.01793420e-01 7.07675397e-01
-8.24042857e-01 -2.09839255e-01 -4.85295326e-01 -4.63865787e-01
-1.38113964e+00 5.20297587e-01 7.62603998e-01 1.01180553e-01
-5.83048165e-01 7.62932658e-01 8.19756269e-01 -3.50178033e-01
2.61887759e-01 -2.24117085e-01 -8.29626024e-02 -2.74409473e-01
5.63897073e-01 1.70598865e-01 -1.05921097e-01 -6.80537343e-01
-1.42070010e-01 6.53464377e-01 -1.23082474e-02 4.05279659e-02
1.17491007e+00 -3.71679932e-01 -5.13398290e-01 2.70325154e-01
1.02816284e+00 2.80001257e-02 -1.52278149e+00 -2.71166652e-01
-4.14477319e-01 -6.16320848e-01 5.24747074e-02 -5.30146480e-01
-1.70895708e+00 6.75997019e-01 8.58537316e-01 4.70183976e-03
1.50103021e+00 -2.24829316e-01 6.17261827e-01 -1.59873486e-01
6.33556321e-02 -1.06188691e+00 1.14537306e-01 1.05784945e-01
1.11453736e+00 -1.13379955e+00 -2.10354812e-02 -5.07411718e-01
-7.25276232e-01 9.52858508e-01 1.00446308e+00 -2.12636024e-01
4.69853848e-01 2.12960407e-01 4.47465748e-01 5.53368628e-02
-3.79480809e-01 -1.58190262e-02 5.11473000e-01 5.17791450e-01
4.91177499e-01 6.90136328e-02 -2.51781434e-01 1.07946619e-01
8.37701783e-02 -2.44326904e-01 3.65370572e-01 5.63269734e-01
-1.76506728e-01 -1.39402866e+00 -4.19925272e-01 1.79098640e-02
-1.64689541e-01 -1.20571546e-01 -5.08689940e-01 6.65651560e-01
3.67296368e-01 7.24738419e-01 1.75886154e-01 -1.68199897e-01
3.98632646e-01 -3.20706099e-01 5.16063809e-01 -3.97308856e-01
-5.35087824e-01 2.78851628e-01 -2.46178597e-01 -8.25437903e-01
-6.33136332e-01 -3.03630292e-01 -1.07177913e+00 -4.85524446e-01
5.15506044e-02 -2.16966242e-01 3.78474057e-01 6.46412015e-01
2.84141928e-01 6.89357042e-01 5.70426226e-01 -1.11515367e+00
-2.81582214e-02 -6.59393251e-01 -6.42335773e-01 6.84788167e-01
1.60543993e-01 -4.64588314e-01 -1.26726583e-01 3.78721297e-01] | [11.28406047821045, -1.2622649669647217] |
abeab3b7-bc5b-4127-a30a-68b857e78fbf | using-active-speaker-faces-for-diarization-in | 2203.15961 | null | https://arxiv.org/abs/2203.15961v1 | https://arxiv.org/pdf/2203.15961v1.pdf | Using Active Speaker Faces for Diarization in TV shows | Speaker diarization is one of the critical components of computational media intelligence as it enables a character-level analysis of story portrayals and media content understanding. Automated audio-based speaker diarization of entertainment media poses challenges due to the diverse acoustic conditions present in media content, be it background music, overlapping speakers, or sound effects. At the same time, speaking faces in the visual modality provide complementary information and not prone to the errors seen in the audio modality. In this paper, we address the problem of speaker diarization in TV shows using the active speaker faces. We perform face clustering on the active speaker faces and show superior speaker diarization performance compared to the state-of-the-art audio-based diarization methods. We additionally report a systematic analysis of the impact of active speaker face detection quality on the diarization performance. We also observe that a moderately well-performing active speaker system could outperform the audio-based diarization systems. | ['Shrikanth Narayanan', 'Rahul Sharma'] | 2022-03-30 | null | null | null | null | ['face-clustering'] | ['computer-vision'] | [ 4.88168925e-01 -1.04649074e-01 2.87506551e-01 -3.54908198e-01
-1.37617874e+00 -6.79601192e-01 6.30956173e-01 4.92160320e-02
3.34797129e-02 1.49146244e-01 5.28594196e-01 4.26814735e-01
-2.02429444e-02 -4.29885209e-01 -5.65799177e-01 -9.52133536e-01
-2.32258793e-02 5.21882057e-01 8.47864673e-02 -6.82188794e-02
2.35000432e-01 4.44525868e-01 -2.30888557e+00 4.45232064e-01
5.35347283e-01 1.09357631e+00 1.81481298e-02 9.87026930e-01
-2.94391006e-01 4.73397642e-01 -1.04648244e+00 -2.26632133e-01
-1.20269328e-01 -4.05806243e-01 -2.93763310e-01 1.20020106e-01
8.98779988e-01 -2.55540133e-01 1.80176608e-02 8.98459315e-01
7.93979228e-01 4.69007762e-03 7.20240772e-01 -1.36688018e+00
-4.21808325e-02 8.10995638e-01 -6.10384464e-01 3.28553021e-01
8.60753715e-01 -3.09999913e-01 7.42511094e-01 -8.09954166e-01
5.93090713e-01 1.58495510e+00 4.15159702e-01 3.99292260e-01
-1.48232508e+00 -1.10360754e+00 -2.84427851e-02 3.00342202e-01
-1.91896021e+00 -1.22143126e+00 1.03167188e+00 -6.34987712e-01
5.39575100e-01 5.58823168e-01 2.82162338e-01 9.33781028e-01
-1.93114340e-01 6.79566562e-01 6.09815598e-01 -5.85688114e-01
2.52171636e-01 2.89227545e-01 3.26662213e-02 1.73334688e-01
-3.54144096e-01 -2.19194949e-01 -1.02664578e+00 -3.21106464e-01
2.73050129e-01 -3.30339611e-01 -1.68373838e-01 1.16513096e-01
-8.98884416e-01 6.09093070e-01 -2.10319579e-01 4.64895517e-01
-2.87324905e-01 5.04598469e-02 3.20759773e-01 1.70608386e-01
6.09640777e-01 3.24691683e-01 2.88354099e-01 -3.32796365e-01
-1.52413750e+00 2.80825704e-01 6.84268415e-01 7.76035845e-01
4.30983633e-01 3.34788173e-01 -2.22350493e-01 8.25255513e-01
3.88405263e-01 5.23778021e-01 2.04365149e-01 -9.26680505e-01
1.33609354e-01 8.50315168e-02 -7.33267143e-02 -1.15044546e+00
-1.24521971e-01 -5.56920320e-02 -3.51042151e-01 2.79061556e-01
3.31896752e-01 1.28722582e-02 -5.24112165e-01 1.54463017e+00
4.19474959e-01 3.88459489e-02 -3.09164822e-01 6.69624031e-01
1.19830441e+00 7.26864874e-01 -1.64594933e-01 -5.05830586e-01
1.34764850e+00 -3.68165582e-01 -1.19364560e+00 1.72598228e-01
-1.45453122e-02 -1.20067501e+00 8.02584589e-01 6.52618408e-01
-1.12687564e+00 -5.83958864e-01 -1.12343550e+00 3.42402995e-01
4.47535142e-02 -8.77680164e-03 2.60333776e-01 1.17294335e+00
-1.06065929e+00 -3.95195857e-02 -6.72912896e-01 -2.09224448e-01
2.04323545e-01 4.18488860e-01 -1.98014617e-01 2.13341579e-01
-8.09501529e-01 4.29829091e-01 -1.80908218e-01 -4.28469896e-01
-9.76803780e-01 -8.35554004e-01 -6.48029447e-01 1.55623913e-01
3.14029545e-01 1.66434079e-01 1.38653994e+00 -1.49768734e+00
-1.60906982e+00 9.49716389e-01 -6.13746166e-01 -3.03485960e-01
3.70342433e-01 -1.04818381e-01 -7.16238856e-01 5.46776414e-01
-1.82497919e-01 4.88299847e-01 1.28694832e+00 -1.39100361e+00
-5.45582116e-01 -5.04399657e-01 -3.88293892e-01 3.13040614e-01
-3.33162487e-01 5.84560990e-01 -4.83518034e-01 -7.05856800e-01
1.56761631e-01 -7.08147526e-01 6.34632468e-01 -1.49386287e-01
-3.31526309e-01 -4.68662288e-03 1.16117072e+00 -9.21934664e-01
1.32955527e+00 -2.74474168e+00 2.85933074e-02 3.52175564e-01
3.68224710e-01 -1.42988890e-01 1.95724964e-01 2.19212890e-01
-9.45476815e-02 -1.76376104e-02 3.50404680e-01 -5.91845155e-01
1.18689783e-01 -4.36808586e-01 -1.24755181e-01 4.15500998e-01
-7.83463717e-02 1.76372349e-01 -3.00785303e-01 -7.59567976e-01
7.15310648e-02 8.07201743e-01 -5.40523291e-01 2.98385229e-02
4.30968292e-02 3.95718038e-01 1.28830105e-01 8.02504778e-01
6.79718077e-01 4.42958385e-01 6.50120974e-02 -2.69625008e-01
-1.44172564e-01 2.84762949e-01 -1.44627666e+00 1.44653642e+00
-2.28117719e-01 1.12907779e+00 7.59949446e-01 -5.17328620e-01
9.57883239e-01 7.44873405e-01 6.03338003e-01 -6.20716929e-01
3.60561430e-01 5.60019445e-03 9.56692174e-02 -3.32319260e-01
6.79075122e-01 -2.21734598e-01 1.76503539e-01 3.30567986e-01
1.77918836e-01 -1.97792903e-01 1.47485420e-01 2.55986303e-01
6.66899860e-01 -5.09638131e-01 -8.04103240e-02 -2.71608949e-01
4.19692904e-01 -4.15289879e-01 2.63938338e-01 4.59013343e-01
-3.48293543e-01 9.49876487e-01 2.05477908e-01 4.47656393e-01
-6.44886255e-01 -1.29249918e+00 -2.74903089e-01 1.46825862e+00
-1.31765217e-01 -6.10372543e-01 -1.16869771e+00 -1.12645321e-01
-2.29011849e-01 8.10897052e-01 -4.50684130e-01 2.79164147e-02
-2.42273405e-01 -4.89834785e-01 7.51176775e-01 2.40109220e-01
1.00410149e-01 -5.71169198e-01 -1.06216781e-01 6.67167753e-02
-4.62544948e-01 -8.59298885e-01 -6.38288736e-01 -1.56435028e-01
-1.86408177e-01 -6.60848081e-01 -7.09344566e-01 -7.77565062e-01
1.57167345e-01 3.72502446e-01 1.03901088e+00 -3.32410097e-01
-3.27451974e-01 8.96774232e-01 -3.98419291e-01 -7.78586864e-01
-8.14134955e-01 -1.84659466e-01 3.55201542e-01 4.88916159e-01
5.41880786e-01 -7.11076200e-01 -3.64509493e-01 5.66624463e-01
-7.59763479e-01 -2.69487053e-01 -1.00512840e-01 2.75874048e-01
4.79106337e-01 4.57911462e-01 6.88085258e-01 -4.70176995e-01
4.63110030e-01 -1.79546729e-01 -3.80629063e-01 -2.68391799e-03
-1.67972550e-01 -4.24710572e-01 -7.55514354e-02 -7.06679821e-01
-1.30115640e+00 1.27558619e-01 -1.43025950e-01 -3.42197776e-01
-3.27065915e-01 6.44406229e-02 -7.41587758e-01 -1.32133558e-01
1.01057529e+00 -1.34192267e-03 1.28347248e-01 -3.89546901e-01
1.16864452e-02 1.06550777e+00 5.23530900e-01 -2.18959555e-01
4.85545129e-01 5.77872515e-01 -4.93507028e-01 -1.53699780e+00
-1.95370942e-01 -5.96918702e-01 -3.02586645e-01 -8.42683494e-01
9.86848950e-01 -9.68336225e-01 -6.62541747e-01 5.58102489e-01
-9.87454653e-01 1.28476068e-01 -1.57543540e-01 4.37029660e-01
-4.87863988e-01 2.08702222e-01 -3.54018688e-01 -1.20229793e+00
-1.68782070e-01 -1.33543837e+00 1.18783772e+00 1.69090480e-01
-6.25592470e-01 -5.34651101e-01 8.00682828e-02 7.10510254e-01
3.10210079e-01 2.18493074e-01 5.49341917e-01 -6.82687998e-01
-2.60674626e-01 -3.32754731e-01 2.80371279e-01 2.08187439e-02
3.39694649e-01 4.45996433e-01 -1.69023824e+00 -1.00396425e-01
1.48605667e-02 -5.58390394e-02 5.34410834e-01 7.20671237e-01
7.57734716e-01 -1.86149403e-01 -1.77355543e-01 2.38543823e-01
8.97515476e-01 4.30607378e-01 6.99093819e-01 -6.61041737e-02
5.55194497e-01 9.06777561e-01 4.71215695e-01 5.82239151e-01
-2.05573533e-02 9.14173543e-01 2.30213374e-01 -7.92195648e-02
-2.70832539e-01 -5.64346462e-02 5.42437494e-01 6.09508455e-01
1.10353790e-01 -2.85749108e-01 -8.68967891e-01 3.49337965e-01
-1.27782106e+00 -1.34421909e+00 -6.55623898e-02 2.27975988e+00
7.77356923e-01 1.56997174e-01 6.40946627e-01 8.24436069e-01
1.20537984e+00 -1.49309948e-01 -2.04698339e-01 -4.06623721e-01
-3.76454711e-01 -3.00663617e-02 2.21791178e-01 4.77096319e-01
-1.12215340e+00 5.43169618e-01 6.63824081e+00 1.14257824e+00
-1.17954063e+00 9.52566341e-02 4.36119825e-01 -7.49241769e-01
-1.79452091e-01 -5.83850622e-01 -7.39029646e-01 4.33473706e-01
8.36334407e-01 -1.76229045e-01 3.52205455e-01 7.19016075e-01
3.71808797e-01 -3.81099254e-01 -1.38549078e+00 1.51302898e+00
7.42311120e-01 -1.08591688e+00 -1.03765398e-01 1.68105185e-01
4.88888085e-01 -3.83368790e-01 1.95176840e-01 -1.70693532e-01
-2.30368063e-01 -9.22438622e-01 1.17733097e+00 3.59691381e-01
8.21383893e-01 -1.02995729e+00 2.52440929e-01 5.77198081e-02
-1.25628567e+00 3.21805403e-02 2.72546560e-01 2.29859889e-01
2.26617113e-01 3.66696358e-01 -9.87206459e-01 1.51797622e-01
8.69054019e-01 2.04815090e-01 -3.53006482e-01 1.07312524e+00
4.17855710e-01 8.05271387e-01 -4.55953181e-01 2.79790699e-01
-5.24457216e-01 1.47928998e-01 9.43076670e-01 1.24968553e+00
2.52670884e-01 -1.13769069e-01 -1.00119058e-02 4.75829393e-01
5.65889925e-02 1.99621290e-01 -4.18356657e-01 -2.02457622e-01
6.60011232e-01 1.12704110e+00 -8.85537863e-01 -2.42860749e-01
-1.74712017e-01 7.51149416e-01 -5.49604058e-01 1.93814605e-01
-5.52598536e-01 -2.34844059e-01 7.57421196e-01 5.09301603e-01
3.40109050e-01 -6.73796386e-02 -1.06839292e-01 -7.41819680e-01
-1.67667642e-01 -1.23214447e+00 2.67369837e-01 -6.53872252e-01
-1.04047287e+00 4.00919408e-01 -6.36578947e-02 -1.17575800e+00
-3.83274555e-01 -7.69156963e-02 -6.77848876e-01 6.19109869e-01
-7.16954589e-01 -9.85643864e-01 -2.73004472e-01 6.35921299e-01
7.45225310e-01 -2.17280090e-01 8.09779048e-01 6.15068555e-01
-3.48940790e-01 1.05534172e+00 3.42859030e-01 -4.88655642e-02
1.06388462e+00 -8.22836459e-01 4.52225208e-02 6.73897862e-01
1.42741203e-01 2.56914705e-01 1.20724738e+00 -4.12450850e-01
-1.30682564e+00 -5.95678806e-01 5.98436058e-01 -4.30240184e-01
3.99812877e-01 -7.89697468e-01 -7.18092263e-01 3.77821326e-01
1.12092398e-01 -7.50951648e-01 1.18865359e+00 2.83086389e-01
-2.78711975e-01 -2.67002285e-01 -1.26230752e+00 2.70679682e-01
6.12072527e-01 -7.49599814e-01 -4.64433014e-01 -9.17247832e-02
4.13376868e-01 -1.32600859e-01 -4.96132076e-01 -6.99299648e-02
8.50889087e-01 -9.62189198e-01 1.07264316e+00 8.11841339e-02
1.40553281e-01 -2.55152255e-01 -3.86163205e-01 -8.79455566e-01
-2.82466143e-01 -6.08694911e-01 9.09553394e-02 2.14099550e+00
3.70866328e-01 -2.25382119e-01 6.50530100e-01 5.99623561e-01
1.94817901e-01 2.35075161e-01 -1.07965887e+00 -6.70795739e-01
-4.07606006e-01 -5.15801072e-01 4.75199968e-01 9.16591406e-01
1.70702830e-01 3.32855374e-01 -3.17189485e-01 3.05333622e-02
6.17775679e-01 -1.45172969e-01 8.03417563e-01 -1.38284683e+00
-2.96654314e-01 -4.69505429e-01 -7.35439539e-01 -4.23648417e-01
4.68489081e-02 -5.24042964e-01 2.30855137e-01 -1.12491786e+00
6.39766008e-02 -7.83088878e-02 1.05446696e-01 -3.92886363e-02
9.89438146e-02 5.46501994e-01 3.31854969e-01 1.67450607e-01
-5.56261837e-01 3.68717015e-01 6.87621891e-01 -3.58165532e-01
-5.15046656e-01 -1.84339922e-04 -5.57407260e-01 7.91998148e-01
5.57435930e-01 -5.79256415e-01 -3.65959674e-01 -1.08801506e-01
9.30149332e-02 1.12868689e-01 -1.55565571e-02 -1.09010851e+00
2.17562616e-01 2.07047150e-01 2.61035919e-01 -7.90411949e-01
7.99461484e-01 -6.69807732e-01 2.93353736e-01 -5.50597683e-02
-4.06549454e-01 -2.37037510e-01 2.49642909e-01 4.53245550e-01
-3.31612825e-01 -2.16552690e-02 7.61132717e-01 8.86125788e-02
-1.68104798e-01 -1.36800662e-01 -1.13737774e+00 -5.30938394e-02
7.99902439e-01 -5.55245042e-01 6.33963570e-02 -8.39500785e-01
-7.02529073e-01 -2.31740221e-01 4.15428162e-01 4.24847990e-01
3.51529896e-01 -1.16284740e+00 -7.65335023e-01 1.96125939e-01
7.52387270e-02 -3.95366102e-01 4.27965492e-01 8.83840501e-01
-2.61364907e-01 -6.59843758e-02 -1.54997379e-01 -8.39373112e-01
-2.19930172e+00 4.05248821e-01 1.13303550e-01 5.51821291e-01
-1.00355513e-01 9.79702234e-01 6.27732515e-01 3.21505547e-01
7.52808392e-01 1.85227260e-01 -4.43856359e-01 7.84745634e-01
1.02978766e+00 7.20489979e-01 4.28354323e-01 -1.02257144e+00
-5.34242451e-01 2.68080771e-01 -1.79102574e-03 -6.67586923e-01
1.03910398e+00 -4.37381417e-01 2.53489375e-01 8.84604037e-01
1.03356993e+00 3.89727533e-01 -9.30650294e-01 -1.03266582e-01
-3.21188807e-01 -4.85853255e-01 3.20029110e-01 -4.90476966e-01
-8.86365116e-01 8.50801468e-01 9.40799594e-01 5.20789027e-01
1.23943162e+00 1.25604346e-01 3.66161913e-01 6.01842143e-02
2.39056814e-02 -1.35413003e+00 1.21031046e-01 4.55987640e-04
9.45547521e-01 -1.09071124e+00 -1.03766002e-01 -5.70004821e-01
-6.56381190e-01 8.74074876e-01 2.90214509e-01 3.06712389e-01
6.53339982e-01 6.22310698e-01 3.75298738e-01 -1.18812971e-01
-3.55850995e-01 -4.71123867e-02 5.41460693e-01 7.52530336e-01
5.86751044e-01 -5.95107116e-02 9.03259814e-02 7.57941008e-01
-6.45318449e-01 -5.71532607e-01 3.87405515e-01 7.65904129e-01
-4.63390142e-01 -7.05274403e-01 -1.17089832e+00 2.12191895e-01
-7.10032761e-01 2.09053494e-02 -9.84116733e-01 2.44280711e-01
-9.52291265e-02 1.50171924e+00 4.11339939e-01 -4.51008141e-01
6.39596358e-02 3.87794524e-01 5.15346229e-01 -3.72588962e-01
-8.02534878e-01 7.87038028e-01 2.87390143e-01 -2.14585826e-01
-5.77378154e-01 -1.06254041e+00 -1.10868406e+00 -5.22694826e-01
-6.56187594e-01 1.13288805e-01 1.01463783e+00 7.40948856e-01
3.24754536e-01 5.71819484e-01 6.49448872e-01 -1.22063029e+00
2.12976336e-01 -9.88906145e-01 -7.42026210e-01 3.90857011e-01
3.55347574e-01 -6.96814656e-01 -6.36563897e-01 4.46901858e-01] | [14.470536231994629, 5.223229885101318] |
303b35ab-1227-49a3-8d68-29c6a636fbc3 | query-guided-regression-network-with-context | 1708.01676 | null | http://arxiv.org/abs/1708.01676v1 | http://arxiv.org/pdf/1708.01676v1.pdf | Query-guided Regression Network with Context Policy for Phrase Grounding | Given a textual description of an image, phrase grounding localizes objects
in the image referred by query phrases in the description. State-of-the-art
methods address the problem by ranking a set of proposals based on the
relevance to each query, which are limited by the performance of independent
proposal generation systems and ignore useful cues from context in the
description. In this paper, we adopt a spatial regression method to break the
performance limit, and introduce reinforcement learning techniques to further
leverage semantic context information. We propose a novel Query-guided
Regression network with Context policy (QRC Net) which jointly learns a
Proposal Generation Network (PGN), a Query-guided Regression Network (QRN) and
a Context Policy Network (CPN). Experiments show QRC Net provides a significant
improvement in accuracy on two popular datasets: Flickr30K Entities and Referit
Game, with 14.25% and 17.14% increase over the state-of-the-arts respectively. | ['Ram Nevatia', 'Rama Kovvuri', 'Kan Chen'] | 2017-08-04 | query-guided-regression-network-with-context-1 | http://openaccess.thecvf.com/content_iccv_2017/html/Chen_Query-Guided_Regression_Network_ICCV_2017_paper.html | http://openaccess.thecvf.com/content_ICCV_2017/papers/Chen_Query-Guided_Regression_Network_ICCV_2017_paper.pdf | iccv-2017-10 | ['phrase-grounding'] | ['natural-language-processing'] | [ 1.43931821e-01 1.52984217e-01 -3.49650323e-01 -3.98889184e-01
-1.36391068e+00 -5.29739916e-01 7.99253345e-01 1.45804465e-01
-6.11146390e-01 6.40925944e-01 4.58461791e-01 1.76610122e-03
-1.17638826e-01 -6.55050576e-01 -8.84411335e-01 -5.12812972e-01
-6.32671174e-03 3.89708847e-01 6.65535212e-01 -2.08065435e-01
4.93097335e-01 1.20708428e-01 -1.20277596e+00 3.94553453e-01
4.55445617e-01 8.64785373e-01 5.74950218e-01 5.43430090e-01
-1.74363807e-01 7.62265921e-01 -5.13986409e-01 -2.89524764e-01
9.75479633e-02 -1.73379153e-01 -8.84744823e-01 -1.00683004e-01
5.45777142e-01 -3.07366192e-01 -6.01535916e-01 8.76694500e-01
4.45275664e-01 4.04675186e-01 4.78197604e-01 -1.41588795e+00
-9.32416618e-01 7.89301574e-01 -7.90113568e-01 1.69408724e-01
3.90041322e-01 -9.86421928e-02 1.39830756e+00 -1.06441295e+00
9.42736089e-01 1.31275678e+00 3.24422807e-01 7.41200864e-01
-1.15104187e+00 -7.74593711e-01 6.82445943e-01 2.15655446e-01
-1.75009513e+00 -1.73513368e-01 7.86433756e-01 -1.53562978e-01
9.31782067e-01 -2.86822673e-02 3.68529737e-01 1.08729076e+00
-1.17467836e-01 9.03897524e-01 7.74520040e-01 -3.00061554e-01
3.19747388e-01 1.09930567e-01 -1.80109352e-01 7.74139106e-01
-1.34573150e-02 7.73142725e-02 -7.37830520e-01 -1.60176963e-01
7.96299040e-01 -5.01883589e-02 -1.04437374e-01 -5.20614922e-01
-1.19629323e+00 9.85762775e-01 8.61248314e-01 7.50996126e-03
-4.85944271e-01 6.37551188e-01 2.78676569e-01 -2.43752658e-01
3.62921447e-01 4.83211875e-01 -3.73140961e-01 2.05712363e-01
-9.92003381e-01 7.02005804e-01 4.15556848e-01 1.39799964e+00
8.61316860e-01 -2.97942460e-01 -5.71816623e-01 7.68879235e-01
6.24455690e-01 5.81789792e-01 3.49319816e-01 -9.96805847e-01
6.14967883e-01 4.70920205e-01 2.77879983e-01 -1.13266158e+00
-1.21662490e-01 -4.47186947e-01 -3.52448344e-01 -5.37014842e-01
-9.35160890e-02 3.40976864e-02 -8.90793979e-01 1.81211627e+00
3.28095555e-01 4.24815625e-01 7.11836591e-02 1.14676476e+00
1.06458426e+00 9.09448624e-01 4.09513712e-01 1.21203721e-01
1.32368994e+00 -1.26736379e+00 -4.66497183e-01 -2.77506292e-01
4.22972620e-01 -5.79642177e-01 9.70083535e-01 5.36935078e-03
-7.54413009e-01 -5.24414182e-01 -7.99718559e-01 -9.98430327e-02
-4.34459776e-01 1.84182569e-01 3.57298970e-01 -1.01340882e-01
-1.18481398e+00 1.72147691e-01 -5.47861874e-01 -7.34681129e-01
3.75338346e-01 4.62538481e-01 -3.27681184e-01 -2.94937491e-01
-1.09952366e+00 5.18457174e-01 6.33209944e-01 -1.27880588e-01
-1.21439028e+00 -6.05165303e-01 -7.35804677e-01 1.18882827e-01
9.61391270e-01 -7.05417275e-01 1.30883515e+00 -7.32600331e-01
-1.08553255e+00 7.43162751e-01 -1.20008998e-01 -5.80418646e-01
2.30214015e-01 -3.80565882e-01 -4.16238338e-01 4.75266427e-01
5.77733517e-01 1.57063663e+00 7.77855754e-01 -1.43592393e+00
-1.20278811e+00 -6.87913820e-02 3.99879307e-01 3.50612819e-01
2.39202946e-01 4.29903939e-02 -1.20048976e+00 -6.94913924e-01
1.59303010e-01 -9.90265369e-01 -3.65458012e-01 1.49936061e-02
-5.63732028e-01 -6.28446460e-01 8.59146595e-01 -4.67713237e-01
1.36326921e+00 -2.02286577e+00 1.91894546e-02 -2.48844689e-03
2.21121497e-02 3.93477306e-02 -4.16228622e-01 5.21000922e-01
2.83941537e-01 3.49692762e-01 1.85544621e-02 -3.44402671e-01
1.09521560e-01 3.32559437e-01 -5.14937878e-01 1.46490097e-01
3.21098506e-01 1.10053790e+00 -1.12682724e+00 -6.81973994e-01
-1.57083616e-01 3.39955449e-01 -7.38093197e-01 1.71778515e-01
-6.95251644e-01 1.95140377e-01 -9.45372939e-01 6.57065868e-01
2.46518552e-01 -4.67414171e-01 2.25857757e-02 -3.24647307e-01
-2.56250310e-03 1.32009447e-01 -1.04787540e+00 2.11035800e+00
-2.66773194e-01 4.11683857e-01 -2.00594291e-01 -7.39942849e-01
8.89326751e-01 3.00756305e-01 2.62788922e-01 -7.70598412e-01
-3.53038669e-01 1.13385364e-01 -4.48247612e-01 -4.15523171e-01
1.05946517e+00 9.19140577e-02 -3.60218555e-01 3.54855061e-01
1.82914659e-01 2.02187121e-01 4.11584340e-02 5.80462456e-01
1.08729506e+00 5.21514058e-01 2.26709768e-01 -1.33285075e-01
4.88786221e-01 1.20069213e-01 6.12809122e-01 1.18937898e+00
-3.70949835e-01 6.59125030e-01 4.63675439e-01 -2.55752027e-01
-8.92823815e-01 -8.76882017e-01 4.62763429e-01 1.50295126e+00
5.52004158e-01 -6.04739964e-01 -5.22445917e-01 -1.13101625e+00
-2.37263204e-03 6.53205574e-01 -6.92788482e-01 -2.71271840e-02
-6.20689154e-01 -1.32053033e-01 2.68859446e-01 7.49374926e-01
6.72807634e-01 -1.12392867e+00 -4.25775677e-01 1.72851250e-01
-3.25086564e-01 -1.42181885e+00 -7.50540376e-01 -2.57542133e-01
-6.23015642e-01 -8.42612386e-01 -5.95706522e-01 -6.85542345e-01
7.16934502e-01 4.74654317e-01 1.11398530e+00 -6.39918745e-02
-9.37553216e-03 6.36578500e-01 -4.51936662e-01 -1.22351587e-01
1.43309310e-02 3.72308046e-01 -2.00203657e-01 -5.56845628e-02
2.61435717e-01 -8.73581097e-02 -8.70758653e-01 4.57586050e-01
-8.91054392e-01 1.30854890e-01 7.79987454e-01 8.08726013e-01
9.82576013e-01 -7.70074427e-02 6.82968140e-01 -8.75617921e-01
3.73680264e-01 -5.01663685e-01 -6.86178505e-01 3.90947938e-01
-5.82965136e-01 3.07000399e-01 8.31963643e-02 -2.24950448e-01
-1.10344350e+00 1.79137185e-01 1.91937476e-01 -5.83040237e-01
-3.09213381e-02 5.50000906e-01 5.05630746e-02 2.31776208e-01
5.86210012e-01 2.21255526e-01 -6.38660371e-01 -2.22088188e-01
6.30877316e-01 3.35119843e-01 6.30289733e-01 -7.15896130e-01
8.65394533e-01 4.75389451e-01 -2.00909466e-01 -3.42737138e-01
-1.21806383e+00 -9.97621834e-01 -3.98193985e-01 -1.03762858e-01
1.02011967e+00 -1.22946429e+00 -5.41400373e-01 -3.90602410e-01
-1.14473164e+00 -1.65677354e-01 -1.60272658e-01 4.20899600e-01
-5.29994905e-01 -1.40622020e-01 -2.04708219e-01 -6.66051924e-01
-4.05356526e-01 -9.43618774e-01 1.44723868e+00 4.01003838e-01
9.53908861e-02 -7.31174648e-01 -8.55318382e-02 2.56532937e-01
4.40899163e-01 6.79409951e-02 7.09243774e-01 -7.97096789e-01
-1.14811540e+00 -9.77831557e-02 -6.95708752e-01 -2.41674751e-01
-1.44444898e-01 -1.67013109e-01 -8.50031793e-01 -2.52191633e-01
-6.84847176e-01 -3.46231401e-01 1.04332983e+00 3.84907514e-01
1.05659807e+00 -4.70806152e-01 -6.33364260e-01 3.24363023e-01
1.81380737e+00 3.49883810e-02 3.77759218e-01 3.58710170e-01
7.19813704e-01 5.38788676e-01 1.03783131e+00 3.36503237e-01
6.63966000e-01 9.54871356e-01 5.99079311e-01 3.57941948e-02
-1.27902940e-01 -8.79766464e-01 4.86798584e-03 1.40000716e-01
1.92298293e-01 -4.71374899e-01 -8.08181763e-01 8.57838154e-01
-2.31066251e+00 -8.62885594e-01 3.35901886e-01 1.87946761e+00
4.68978375e-01 2.50925362e-01 9.77182165e-02 -5.83800495e-01
8.89603257e-01 2.67203212e-01 -5.57411551e-01 5.75978532e-02
9.60877612e-02 -1.97626784e-01 6.63027227e-01 3.96267205e-01
-1.32631922e+00 1.41827202e+00 5.67196751e+00 8.92222226e-01
-9.87600446e-01 9.59417373e-02 4.69721556e-01 1.18487485e-01
-3.60808402e-01 1.09807238e-01 -1.11219573e+00 3.63234989e-02
7.23379731e-01 -8.85081068e-02 1.33719131e-01 9.95029151e-01
3.77788723e-01 -1.51923761e-01 -1.12267077e+00 9.10797119e-01
2.72265166e-01 -1.34090292e+00 2.91549563e-01 3.10234055e-02
8.04072976e-01 2.45140150e-01 1.44331127e-01 6.90633595e-01
4.17205036e-01 -7.73408055e-01 6.79179013e-01 6.73205674e-01
6.30106509e-01 -7.08396912e-01 7.64563739e-01 1.12476289e-01
-1.23321152e+00 -1.59621686e-01 -4.02066231e-01 4.68614221e-01
4.84243445e-02 2.66999863e-02 -1.11094630e+00 4.94865268e-01
7.29485095e-01 7.12642252e-01 -6.69605434e-01 1.03558254e+00
-4.55021918e-01 6.26978397e-01 -1.41212672e-01 -2.39855796e-01
7.86918700e-01 2.98382163e-01 5.34682810e-01 1.20299363e+00
1.60138533e-01 2.20535681e-01 6.51385248e-01 9.74277556e-01
-5.04550278e-01 1.90542594e-01 -3.03086579e-01 -3.31629403e-02
7.48070836e-01 1.39376903e+00 -8.63103688e-01 -2.82041252e-01
-4.04682994e-01 9.47986662e-01 3.02245229e-01 6.11513495e-01
-9.74259377e-01 -2.37900853e-01 1.34094626e-01 5.95308281e-02
7.43339241e-01 2.58761924e-02 3.37650329e-01 -8.97400737e-01
1.49474172e-02 -4.37713206e-01 3.64340633e-01 -1.15272832e+00
-1.23208427e+00 4.42671508e-01 2.93780237e-01 -1.09784031e+00
-1.24818571e-01 -2.70072162e-01 -3.37000996e-01 8.04997504e-01
-1.92571425e+00 -1.47632515e+00 -2.47205943e-01 3.69447142e-01
7.86897957e-01 -1.85901582e-01 6.11131251e-01 5.73632158e-02
-3.17566872e-01 3.71366411e-01 -1.69012517e-01 1.67449489e-01
9.15616870e-01 -1.34733534e+00 3.11291426e-01 5.44110477e-01
3.36076796e-01 6.49518728e-01 5.09886622e-01 -5.57586133e-01
-1.23616707e+00 -1.52625930e+00 9.92930114e-01 -2.83067495e-01
6.19427025e-01 -3.58984292e-01 -7.23604918e-01 6.52046025e-01
1.49896711e-01 4.64879155e-01 5.66127062e-01 -1.16422698e-02
-4.94271398e-01 -3.04765403e-01 -1.01003802e+00 6.56751394e-01
8.83327127e-01 -6.63680255e-01 -5.40321171e-01 2.77186036e-01
1.15912759e+00 -3.86762142e-01 -5.25404930e-01 1.71942338e-01
6.09043181e-01 -6.01375043e-01 8.94322574e-01 -4.61915284e-01
4.08051610e-01 -6.21360719e-01 -5.98680198e-01 -8.87541473e-01
-3.26351672e-01 -4.15898651e-01 -6.74953535e-02 1.37643242e+00
6.19867742e-01 4.35489453e-02 1.05521822e+00 5.20815194e-01
-3.95161621e-02 -7.94434369e-01 -7.39997029e-01 -4.98882174e-01
-3.81052375e-01 -3.88109058e-01 5.98317683e-01 6.54473484e-01
-3.09378892e-01 5.75522661e-01 -2.22217575e-01 3.78153265e-01
3.84540379e-01 3.14404741e-02 6.30653679e-01 -8.39901984e-01
-2.03127846e-01 -1.07580483e-01 -3.20364207e-01 -1.30359316e+00
2.07443744e-01 -6.73649192e-01 3.59038681e-01 -1.78876531e+00
5.18602848e-01 -4.87988323e-01 -5.56365430e-01 5.73103309e-01
-2.57530183e-01 2.41809860e-01 3.68317127e-01 4.44018126e-01
-1.46567106e+00 5.33345938e-01 9.76288438e-01 -2.36608371e-01
-2.11820588e-01 -8.62222761e-02 -8.91231596e-01 5.06533861e-01
4.81105417e-01 -6.47374809e-01 -6.97726011e-01 -5.75755894e-01
3.59168172e-01 2.04350635e-01 5.33525407e-01 -7.39724219e-01
4.90086257e-01 -6.92255348e-02 2.46355668e-01 -9.78242218e-01
2.73594141e-01 -8.07640016e-01 -3.15639764e-01 3.14775780e-02
-9.43395436e-01 1.93584755e-01 1.45732492e-01 1.17979014e+00
-2.13622719e-01 -1.87274158e-01 2.98308372e-01 -2.52439052e-01
-1.12296903e+00 4.36449736e-01 7.45878788e-03 1.19341560e-01
9.50144768e-01 8.72090235e-02 -3.78581345e-01 -5.35466075e-01
-6.73544765e-01 5.72550058e-01 8.96788836e-02 5.69199562e-01
8.43960404e-01 -1.45555341e+00 -7.33966172e-01 -2.94889539e-01
5.87943017e-01 1.30522877e-01 1.31443739e-01 4.12562668e-01
-2.37542182e-01 7.96822608e-01 4.00245577e-01 -6.91881716e-01
-1.05499411e+00 5.97336292e-01 1.66311681e-01 -5.06527007e-01
-3.88276607e-01 9.54228282e-01 5.19019365e-01 -4.04151440e-01
3.56204093e-01 -3.14918607e-01 -3.19824487e-01 -2.22127438e-01
3.55732709e-01 3.05574830e-03 -2.00787097e-01 -7.71269441e-01
-2.87082404e-01 5.56086719e-01 -5.23905396e-01 -3.51481736e-01
1.10674512e+00 -3.25781703e-01 3.12530905e-01 1.83332354e-01
1.15376580e+00 -2.78866887e-01 -1.36407244e+00 -7.07454562e-01
2.58223802e-01 -2.76239574e-01 1.94434345e-01 -1.11069798e+00
-9.18788075e-01 4.68056411e-01 6.37063980e-01 -5.57759590e-02
1.03296220e+00 4.12639499e-01 5.20943880e-01 6.82312608e-01
2.90001988e-01 -1.06367481e+00 4.64327306e-01 2.86858082e-01
1.00872803e+00 -1.37687576e+00 6.84733838e-02 -3.86304796e-01
-7.92393029e-01 7.79975176e-01 8.39656413e-01 -2.48070002e-01
5.09776473e-01 -4.10624623e-01 -2.32610181e-01 -3.87602836e-01
-1.00229168e+00 -3.31717819e-01 4.80865180e-01 5.40949881e-01
2.45618939e-01 -1.45779327e-01 -1.60865739e-01 5.42194247e-01
1.52792707e-01 -1.09298348e-01 3.12319785e-01 8.47237468e-01
-5.82076967e-01 -9.14399564e-01 -2.14238137e-01 1.52236208e-01
-5.16022921e-01 -2.58927345e-01 -2.43571579e-01 9.91378188e-01
1.06099300e-01 9.70691562e-01 1.62818238e-01 -1.14689760e-01
4.09059763e-01 -7.03013614e-02 1.00622550e-01 -8.57404470e-01
-4.04792249e-01 4.78506297e-01 2.56053731e-02 -7.10428238e-01
-7.74931014e-01 -4.93436515e-01 -1.36604881e+00 3.06273758e-01
-5.56167781e-01 3.56111288e-01 7.55458534e-01 9.00711358e-01
6.15959764e-01 5.07696271e-01 4.85741198e-01 -4.97171879e-01
-2.32377023e-01 -9.07863259e-01 -4.60084856e-01 2.71556169e-01
2.26001352e-01 -5.52621305e-01 1.01099089e-01 5.06654419e-02] | [10.441947937011719, 1.4467012882232666] |
d3ea9c36-e570-451c-b013-4f01b00b4e24 | deep-submodular-networks-for-extractive-data | 2010.08593 | null | https://arxiv.org/abs/2010.08593v1 | https://arxiv.org/pdf/2010.08593v1.pdf | Deep Submodular Networks for Extractive Data Summarization | Deep Models are increasingly becoming prevalent in summarization problems (e.g. document, video and images) due to their ability to learn complex feature interactions and representations. However, they do not model characteristics such as diversity, representation, and coverage, which are also very important for summarization tasks. On the other hand, submodular functions naturally model these characteristics because of their diminishing returns property. Most approaches for modelling and learning submodular functions rely on very simple models, such as weighted mixtures of submodular functions. Unfortunately, these models only learn the relative importance of the different submodular functions (such as diversity, representation or importance), but cannot learn more complex feature representations, which are often required for state-of-the-art performance. We propose Deep Submodular Networks (DSN), an end-to-end learning framework that facilitates the learning of more complex features and richer functions, crafted for better modelling of all aspects of summarization. The DSN framework can be used to learn features appropriate for summarization from scratch. We demonstrate the utility of DSNs on both generic and query focused image-collection summarization, and show significant improvement over the state-of-the-art. In particular, we show that DSNs outperform simple mixture models using off the shelf features. Secondly, we also show that just using four submodular functions in a DSN with end-to-end learning performs comparably to the state-of-the-art mixture model with a hand-crafted set of 594 components and outperforms other methods for image collection summarization. | ['Rishabh Iyer', 'Chandrashekhar Lavania', 'Jiten Girdhar', 'Suraj Kothawade'] | 2020-10-16 | null | null | null | null | ['data-summarization'] | ['miscellaneous'] | [-1.18486350e-02 -5.09617142e-02 -4.47445273e-01 -1.59950837e-01
-9.59902525e-01 -5.13084531e-01 6.86650872e-01 3.65614861e-01
-1.73448354e-01 6.30096853e-01 7.67734587e-01 2.53433406e-01
-2.80709118e-01 -4.58347797e-01 -9.55418825e-01 -6.99042737e-01
-2.14439675e-01 7.83223689e-01 -4.14854437e-02 -2.57013798e-01
1.53607726e-01 3.63458693e-01 -1.67592490e+00 2.87769198e-01
1.01892328e+00 8.84856224e-01 3.28638613e-01 8.36680889e-01
-1.70291662e-01 6.57985926e-01 -7.79579282e-01 -1.21829972e-01
1.44277781e-01 -4.54396039e-01 -6.37269676e-01 3.95283401e-01
9.55540538e-01 -5.02685487e-01 -6.93005562e-01 7.64707983e-01
7.18999863e-01 3.40215743e-01 1.16265678e+00 -1.08339429e+00
-7.78030932e-01 6.06287420e-01 -5.26574135e-01 2.37308908e-02
2.65398264e-01 1.84499055e-01 1.50944448e+00 -7.81802118e-01
4.78427798e-01 1.45555985e+00 4.47279990e-01 3.98122847e-01
-1.23663449e+00 -2.75707901e-01 3.14901382e-01 2.77778029e-01
-1.01989841e+00 -3.90536904e-01 5.53319871e-01 -1.74756512e-01
1.03565085e+00 4.86414313e-01 6.59889221e-01 8.50504041e-01
2.45291203e-01 1.43003643e+00 4.26376522e-01 -8.98110941e-02
2.67805420e-02 9.17600691e-02 2.86735535e-01 6.79292977e-01
5.54840446e-01 -5.59537113e-01 -4.34964985e-01 -3.33518177e-01
4.93502676e-01 2.70543993e-01 -5.38931251e-01 -6.51520073e-01
-1.12489820e+00 1.01025021e+00 3.76833439e-01 9.66108218e-02
-4.75638092e-01 5.15205681e-01 4.85109508e-01 3.28981996e-01
5.10269523e-01 6.51126027e-01 -3.08600754e-01 -6.50777444e-02
-1.19858408e+00 4.71000195e-01 9.48935747e-01 1.06203187e+00
8.66311967e-01 9.45488960e-02 -7.13740647e-01 1.17313743e+00
-6.62022270e-03 4.49923903e-01 3.65968317e-01 -1.10585916e+00
3.62144828e-01 5.77129900e-01 -1.06363691e-01 -7.45166957e-01
-5.22595644e-01 -7.42233574e-01 -1.20846903e+00 -2.67549753e-01
-6.46744221e-02 -5.17018698e-02 -1.02961409e+00 1.76296532e+00
-9.16217268e-02 -1.34385645e-01 9.31617469e-02 5.69886386e-01
1.15874481e+00 1.01284099e+00 -4.23037380e-01 -2.39611745e-01
1.19257760e+00 -1.32720804e+00 -5.72204113e-01 -1.67717159e-01
6.15185142e-01 -4.75417346e-01 8.69914055e-01 4.76536483e-01
-1.49670553e+00 -4.31601465e-01 -8.85573626e-01 -3.03073436e-01
3.03918496e-02 5.51801026e-02 6.81973338e-01 2.72392720e-01
-1.37261021e+00 6.42945468e-01 -6.23082995e-01 -5.08586347e-01
7.00113177e-01 5.49558580e-01 -2.69663125e-01 -3.12354714e-01
-7.32591569e-01 8.71799946e-01 4.05896336e-01 -3.49223733e-01
-1.09377635e+00 -8.03643107e-01 -1.09032488e+00 6.46926343e-01
4.75779176e-01 -1.38084602e+00 1.31423604e+00 -6.85657024e-01
-1.37576103e+00 5.85807860e-01 -1.07699022e-01 -5.86857557e-01
3.56172234e-01 -4.34193492e-01 1.03443496e-01 3.62748295e-01
-2.85969358e-02 8.90878260e-01 9.72191811e-01 -1.29816496e+00
-4.71565813e-01 -1.39768079e-01 1.90319091e-01 4.00800437e-01
-6.54779553e-01 -1.93627492e-01 -4.62666720e-01 -5.57377100e-01
-2.65771955e-01 -6.49582088e-01 -3.36182177e-01 -1.41930252e-01
-6.14915967e-01 -3.44036162e-01 6.72178745e-01 -6.76158786e-01
1.36526060e+00 -1.86025512e+00 7.18706369e-01 -7.46792406e-02
3.59119743e-01 3.17217469e-01 -5.84028602e-01 6.91452861e-01
1.55720919e-01 1.05539322e-01 -3.45320016e-01 -7.42526770e-01
3.03927124e-01 8.68937150e-02 -7.97177926e-02 5.03815949e-01
2.80522406e-01 1.13433099e+00 -7.88233995e-01 -3.21012914e-01
7.00138956e-02 3.34305555e-01 -8.40516508e-01 9.09344405e-02
-6.46740675e-01 1.62548702e-02 -2.96109438e-01 4.60736752e-01
6.15666270e-01 -2.87353247e-01 -9.71518159e-02 -5.30371740e-02
3.85851264e-01 6.38198704e-02 -1.06279373e+00 1.77028728e+00
-5.69355726e-01 6.38153493e-01 1.27813876e-01 -1.36190081e+00
7.63326526e-01 1.85699299e-01 5.71257412e-01 -1.75653785e-01
1.43769085e-01 2.16073960e-01 -1.71280324e-01 -2.84180582e-01
9.60088432e-01 7.31498450e-02 -2.20007271e-01 5.90934873e-01
5.44451296e-01 -5.23905218e-01 4.56331939e-01 5.27183056e-01
9.73179698e-01 -4.12135810e-01 4.90373522e-01 -2.85957187e-01
5.42945623e-01 -1.90403119e-01 2.19249904e-01 9.82826352e-01
4.67498124e-01 9.57138062e-01 8.64978492e-01 -1.44529074e-01
-1.07388413e+00 -8.64761889e-01 7.58155668e-03 1.06954694e+00
-4.37789075e-02 -5.89925647e-01 -6.33731127e-01 -4.98460680e-01
1.42035842e-01 7.79972255e-01 -3.71625423e-01 -3.96580160e-01
-3.72749656e-01 -7.72700846e-01 2.57104397e-01 4.50168788e-01
4.50124234e-01 -8.60099077e-01 -2.26890802e-01 1.66673392e-01
-1.11027651e-01 -9.32277203e-01 -9.08304930e-01 4.69241571e-03
-1.00224864e+00 -7.74791777e-01 -1.07720780e+00 -7.37712026e-01
4.49371606e-01 6.56559169e-01 1.32255065e+00 2.41502821e-02
-1.83623865e-01 9.77369010e-01 -1.64616182e-01 -5.82543314e-01
-4.82143879e-01 4.10026908e-01 -9.84548256e-02 -3.89363542e-02
-4.70619798e-02 -7.20451295e-01 -6.53285027e-01 -3.37929390e-02
-1.34532261e+00 -2.35972002e-01 8.03307772e-01 9.75782812e-01
4.87510949e-01 -5.44650480e-02 7.40227103e-01 -8.41081023e-01
9.69226956e-01 -6.32394254e-01 -1.18324302e-01 4.03614551e-01
-1.85118228e-01 2.75721282e-01 6.43046379e-01 -4.74470973e-01
-5.36425233e-01 -4.01609927e-01 6.76295385e-02 -5.42630851e-01
1.88869074e-01 7.40716815e-01 -1.64038837e-01 3.07838291e-01
6.52360916e-01 5.46224654e-01 3.80189061e-01 -3.72770399e-01
6.18123055e-01 5.89235604e-01 5.19086182e-01 -4.44596231e-01
6.57028675e-01 4.65636402e-01 5.75829558e-02 -1.17510223e+00
-9.64025557e-01 -7.89460897e-01 -3.01237911e-01 2.68681049e-01
5.52483320e-01 -1.12600565e+00 -3.21513444e-01 4.71345425e-01
-1.23956954e+00 -1.37137756e-01 -7.72317171e-01 2.77995646e-01
-9.54275012e-01 7.44266391e-01 -3.48023415e-01 -5.74908257e-01
-7.61695325e-01 -9.63766873e-01 1.38526750e+00 1.96297765e-01
-2.80541420e-01 -1.02977264e+00 4.00831588e-02 3.37406725e-01
5.16295552e-01 9.59955230e-02 1.20334589e+00 -7.86039233e-01
-7.48849571e-01 -3.19860667e-01 -7.99947158e-02 6.75858200e-01
1.42441206e-02 -7.94842839e-03 -7.29544103e-01 -5.68459511e-01
-1.27438694e-01 -4.20680404e-01 1.69673085e+00 9.43509758e-01
1.22927296e+00 -6.28015816e-01 -9.70553458e-02 6.79320097e-01
1.13677561e+00 -4.20043200e-01 6.67782366e-01 3.67781073e-02
6.78216159e-01 6.05359554e-01 2.44210958e-01 6.29624426e-01
5.93897104e-01 5.15130699e-01 5.07381558e-01 -4.35433462e-02
-1.71569571e-01 -5.90180196e-02 6.33206487e-01 6.40760958e-01
6.02481291e-02 -6.54828489e-01 -4.60806906e-01 6.65353000e-01
-2.22046351e+00 -9.72477913e-01 1.09380305e-01 2.10076451e+00
7.08570361e-01 -1.60972059e-01 5.31467736e-01 -2.44925946e-01
4.47460026e-01 6.39161944e-01 -6.32508934e-01 -4.99590605e-01
-4.73096997e-01 8.72327387e-02 3.32420856e-01 3.66280079e-01
-1.07856786e+00 8.70607913e-01 6.24769115e+00 1.12293959e+00
-8.34622085e-01 -8.35301727e-02 6.27316475e-01 -4.25537765e-01
-7.83834219e-01 -2.65961170e-01 -7.74905443e-01 1.81517035e-01
6.25029504e-01 -3.42240721e-01 4.61158663e-01 7.24573374e-01
9.43776220e-02 -1.38099626e-01 -1.36670899e+00 1.12636161e+00
5.85295796e-01 -1.68558013e+00 5.83027005e-01 4.22353931e-02
1.12932110e+00 7.90473521e-02 1.43512636e-01 5.59489012e-01
2.71331042e-01 -1.27167869e+00 5.38591564e-01 4.80706573e-01
6.11727357e-01 -7.98890829e-01 7.71628380e-01 5.61240852e-01
-6.72286928e-01 -2.72438198e-01 -7.60421455e-01 2.44377226e-01
6.84683621e-02 6.24684691e-01 -5.89901805e-01 6.77370131e-01
2.57020593e-01 8.99802446e-01 -4.81656760e-01 1.18228853e+00
-1.86721212e-03 3.54545534e-01 -3.86023402e-01 -1.76184058e-01
4.47821259e-01 -2.30277687e-01 8.88739228e-01 1.34436536e+00
4.14075702e-01 -3.13017249e-01 2.10446537e-01 7.65196741e-01
-4.44399774e-01 1.08123638e-01 -7.07306743e-01 -1.28554672e-01
8.23464245e-02 1.17446280e+00 -2.16283932e-01 -3.89130622e-01
-3.34011912e-01 7.77015686e-01 3.31281096e-01 4.90564942e-01
-6.42439008e-01 -4.73052561e-02 6.67910576e-01 2.06822351e-01
7.54963696e-01 -1.23286232e-01 -6.98559582e-02 -1.37409508e+00
2.69527789e-02 -1.11806798e+00 3.19017768e-01 -2.82770455e-01
-1.28610361e+00 3.09526980e-01 4.07321453e-01 -1.04695928e+00
-3.63730878e-01 -2.79866815e-01 -9.51541543e-01 5.94028771e-01
-1.58917379e+00 -1.24846292e+00 -2.71901757e-01 4.55868185e-01
7.39131570e-01 -3.32671255e-01 6.12950206e-01 -1.74922675e-01
-4.11967158e-01 5.11123240e-01 4.86959875e-01 -3.86613250e-01
6.35556459e-01 -1.32990694e+00 1.92391425e-01 5.37325025e-01
2.02780560e-01 3.32650989e-01 8.12364876e-01 -1.69563189e-01
-1.46813476e+00 -1.08984995e+00 6.79503679e-01 -3.33036900e-01
3.40460837e-01 -3.40588689e-01 -8.20111513e-01 6.23341978e-01
4.02732164e-01 -2.55952269e-01 7.40151942e-01 2.44189307e-01
-2.02896729e-01 -8.03336725e-02 -8.34154069e-01 6.11445665e-01
1.04838884e+00 -1.80150941e-01 -5.19744754e-01 6.93057656e-01
8.11561942e-01 -1.26544669e-01 -5.67850053e-01 3.39124858e-01
4.55165595e-01 -9.04397070e-01 1.08521140e+00 -8.11082244e-01
8.09137106e-01 9.21282470e-02 -3.33149016e-01 -1.69659722e+00
-5.33207655e-01 -9.25586402e-01 -6.58753157e-01 1.13084137e+00
3.08514327e-01 -4.61207420e-01 7.11812556e-01 1.83554038e-01
-2.87507802e-01 -1.11278391e+00 -6.88464940e-01 -9.34786916e-01
2.03260869e-01 -6.97038472e-02 5.48175931e-01 4.13111895e-01
-3.10765296e-01 7.95243800e-01 -5.55542946e-01 -2.00786039e-01
4.48451251e-01 3.29921544e-01 1.25066435e+00 -1.21624315e+00
-4.24324393e-01 -8.46843421e-01 -5.47518790e-01 -1.57282329e+00
3.92905504e-01 -1.03330171e+00 -2.67827451e-01 -2.13317275e+00
6.75271571e-01 -2.48744749e-02 -9.39319506e-02 9.37254280e-02
-4.14500177e-01 -1.83456138e-01 4.07834768e-01 1.46083072e-01
-1.00285065e+00 1.20617723e+00 1.38685930e+00 -5.36195040e-01
-3.51627678e-01 2.02915609e-01 -1.26278329e+00 6.26185656e-01
6.33569539e-01 -2.54747480e-01 -5.82620323e-01 -5.89321434e-01
2.07010563e-02 1.02813825e-01 2.19453394e-01 -8.39203596e-01
3.36252242e-01 1.10224582e-01 7.80264139e-02 -8.90936673e-01
5.46635211e-01 -5.17284513e-01 -1.11850090e-01 2.66630590e-01
-3.15366685e-01 -2.10207865e-01 1.22079663e-01 8.93290997e-01
-4.24540997e-01 -4.04235750e-01 4.94787276e-01 -1.60238564e-01
-2.88574398e-01 5.62421858e-01 -1.94619060e-01 1.92662314e-01
9.22945738e-01 -3.21274161e-01 -2.70361423e-01 -1.19703877e+00
-3.74284655e-01 7.57891834e-01 4.38813567e-01 2.43899092e-01
6.67955220e-01 -1.12754190e+00 -1.34745634e+00 -5.97366989e-02
7.56876394e-02 5.48799276e-01 4.31734324e-01 9.32217956e-01
-5.32483757e-01 4.34669584e-01 -7.17181489e-02 -8.67116511e-01
-1.12011659e+00 5.21889508e-01 1.34753600e-01 -8.06780756e-01
-5.45157731e-01 7.09685802e-01 5.69747031e-01 -4.12933290e-01
3.92793000e-01 -3.68089408e-01 -2.78845012e-01 4.65066254e-01
3.63117069e-01 4.93104696e-01 -8.26827809e-02 -3.05252492e-01
-6.42065257e-02 5.14739811e-01 -4.17696804e-01 1.87027112e-01
1.75552940e+00 4.76596840e-02 -2.01431513e-01 1.83687091e-01
1.30903184e+00 -1.66484058e-01 -1.19439864e+00 -4.86168087e-01
-3.56193691e-01 -2.11962298e-01 -1.78813115e-02 -2.88532615e-01
-9.64060247e-01 8.47749591e-01 -2.56512016e-01 1.73554122e-01
1.07896292e+00 2.28461444e-01 9.38424170e-01 6.82214797e-01
-8.19734484e-03 -8.61779928e-01 4.25035805e-01 5.68705320e-01
1.23074234e+00 -1.08277917e+00 3.61988157e-01 -2.94392735e-01
-7.94323504e-01 1.11409736e+00 1.52291954e-01 -4.75360036e-01
4.91563052e-01 -1.11694589e-01 -6.10895097e-01 -2.79319137e-01
-1.07733119e+00 -4.84157465e-02 5.40408015e-01 5.86434245e-01
1.68664157e-01 -1.74762704e-03 -1.53055698e-01 5.22026420e-01
-1.69271544e-01 -3.86756897e-01 6.74261928e-01 5.38775563e-01
-5.98085523e-01 -8.20282280e-01 -2.07889482e-01 9.82432961e-01
-3.09304774e-01 -2.05055401e-01 -3.81858200e-01 6.92914248e-01
-4.68905926e-01 7.25896299e-01 9.04220492e-02 -1.30805895e-01
4.13999707e-01 -2.92943597e-01 6.57143831e-01 -1.01132083e+00
-6.42990589e-01 5.58409281e-02 1.45867318e-01 -2.92049140e-01
-3.71972442e-01 -7.44089842e-01 -6.85377836e-01 -4.97983366e-01
-3.47918481e-01 7.79232755e-02 3.41879040e-01 6.73020244e-01
4.65364695e-01 6.29617274e-01 6.66777372e-01 -1.17170691e+00
-1.19622433e+00 -1.15776265e+00 -7.40401804e-01 3.25724453e-01
5.85024297e-01 -3.27152729e-01 -2.21850470e-01 -2.63435036e-01] | [12.340677261352539, 9.351949691772461] |
c20562c2-9c2f-4875-a943-162821e91b90 | parents-and-children-distinguishing | 2304.00500 | null | https://arxiv.org/abs/2304.00500v1 | https://arxiv.org/pdf/2304.00500v1.pdf | Parents and Children: Distinguishing Multimodal DeepFakes from Natural Images | Recent advancements in diffusion models have enabled the generation of realistic deepfakes by writing textual prompts in natural language. While these models have numerous benefits across various sectors, they have also raised concerns about the potential misuse of fake images and cast new pressures on fake image detection. In this work, we pioneer a systematic study of the authenticity of fake images generated by state-of-the-art diffusion models. Firstly, we conduct a comprehensive study on the performance of contrastive and classification-based visual features. Our analysis demonstrates that fake images share common low-level cues, which render them easily recognizable. Further, we devise a multimodal setting wherein fake images are synthesized by different textual captions, which are used as seeds for a generator. Under this setting, we quantify the performance of fake detection strategies and introduce a contrastive-based disentangling strategy which let us analyze the role of the semantics of textual descriptions and low-level perceptual cues. Finally, we release a new dataset, called COCOFake, containing about 600k images generated from original COCO images. | ['Rita Cucchiara', 'Alberto del Bimbo', 'Lorenzo Baraldi', 'Marcella Cornia', 'Davide Morelli', 'Roberto Amoroso'] | 2023-04-02 | null | null | null | null | ['fake-image-detection'] | ['computer-vision'] | [ 2.62648225e-01 4.61613424e-02 -1.36788487e-01 -7.84005448e-02
-5.48348963e-01 -8.86064470e-01 1.36726999e+00 1.14303790e-01
-2.14319512e-01 3.57961088e-01 2.58553684e-01 -3.41184735e-01
3.00620556e-01 -5.80866098e-01 -7.53449559e-01 -4.67952132e-01
2.15183958e-01 7.14482293e-02 2.46208720e-02 -5.67434072e-01
5.02037823e-01 3.76249492e-01 -1.22518957e+00 7.75943816e-01
7.07073390e-01 6.58730507e-01 -2.33273819e-01 4.92091000e-01
2.82897890e-01 1.06178558e+00 -1.12549138e+00 -1.15275621e+00
1.60581589e-01 -6.01290226e-01 -5.38213432e-01 3.59842390e-01
7.41209626e-01 -5.90029061e-01 -9.48011279e-01 1.32825506e+00
3.39521110e-01 -4.56114054e-01 7.91415811e-01 -1.84242034e+00
-1.41681409e+00 4.21928346e-01 -4.88708824e-01 3.22958261e-01
4.20294374e-01 6.57996118e-01 8.79955947e-01 -9.18509483e-01
1.01669705e+00 1.47713614e+00 3.67016613e-01 6.96088016e-01
-1.21920478e+00 -7.01016724e-01 -2.33958066e-01 3.21158886e-01
-1.00167811e+00 -6.46562934e-01 8.69361877e-01 -6.61835253e-01
4.36822355e-01 2.89889693e-01 6.02317035e-01 1.94844031e+00
2.43483260e-01 9.66094077e-01 1.47890377e+00 -4.03825253e-01
-9.66322348e-02 5.65706730e-01 -3.59271914e-01 8.05651546e-01
4.81333405e-01 3.84103894e-01 -8.18916976e-01 -2.56584674e-01
7.39522576e-01 -4.08419222e-01 -4.28306043e-01 -3.02224934e-01
-1.28025365e+00 1.13075364e+00 3.39862704e-01 4.02669281e-01
-2.41474599e-01 1.87916532e-01 3.10464084e-01 5.03379881e-01
5.08272648e-01 6.75124168e-01 3.92552614e-01 1.65786877e-01
-8.24462831e-01 3.51247847e-01 5.57477593e-01 6.79811299e-01
3.36116344e-01 2.09762510e-02 -4.09360975e-01 5.96508443e-01
2.38944784e-01 6.88305795e-01 5.29189885e-01 -7.83063769e-01
6.89151645e-01 2.48362035e-01 3.42196286e-01 -1.89699554e+00
3.24077010e-01 -9.42038521e-02 -6.23592257e-01 4.21024971e-02
3.85867834e-01 1.32364213e-01 -6.83370113e-01 1.47751534e+00
-2.09132358e-01 -9.46095139e-02 2.22036261e-02 9.98789191e-01
4.51531798e-01 3.82299930e-01 -9.25933272e-02 8.12554881e-02
1.19582272e+00 -1.02421927e+00 -8.27484190e-01 -2.90121436e-01
6.12151027e-01 -9.10731912e-01 1.11210418e+00 3.25405359e-01
-9.66679335e-01 -1.62012368e-01 -1.08516359e+00 5.58987893e-02
-4.80548292e-01 5.32133467e-02 4.27420378e-01 1.12811494e+00
-9.77141678e-01 3.75020891e-01 -1.77628234e-01 -1.12060048e-01
8.09487581e-01 -2.41463304e-01 -4.86408710e-01 -2.72841513e-01
-1.20433307e+00 1.14689898e+00 4.39397208e-02 4.84428443e-02
-1.27180028e+00 -4.14959081e-02 -7.39469290e-01 -3.27930510e-01
3.04582238e-01 -3.69693249e-01 9.46086407e-01 -1.44875264e+00
-1.14247000e+00 1.21076703e+00 1.64617494e-01 -4.94988948e-01
9.33521390e-01 4.56031337e-02 -6.52145445e-01 4.46224123e-01
3.13119173e-01 8.33607852e-01 1.40027487e+00 -1.81825340e+00
-2.10079029e-01 -1.04273766e-01 1.31125867e-01 -4.04581688e-02
-4.72972244e-01 1.97031036e-01 -1.35087678e-02 -1.01092458e+00
-2.91768819e-01 -8.49956274e-01 9.58538055e-02 3.17078941e-02
-7.33014405e-01 2.38300577e-01 9.34001327e-01 -6.73208058e-01
9.60516930e-01 -2.38503289e+00 3.62945013e-02 1.49448633e-01
6.70879960e-01 2.77186364e-01 -4.96794194e-01 7.28400946e-01
8.93007889e-02 5.94219387e-01 -1.37654498e-01 -4.44052011e-01
9.13581997e-02 -6.20041564e-02 -6.75971210e-01 7.53781915e-01
5.53802669e-01 1.27256310e+00 -1.15371382e+00 -4.42877710e-01
7.84641206e-02 3.33124876e-01 -3.55826199e-01 4.58038971e-02
-2.23613098e-01 4.78457510e-01 -1.56613335e-01 7.18546093e-01
6.19575739e-01 -2.11361229e-01 -1.46674542e-02 -5.47306985e-02
9.25941765e-02 6.90926239e-02 -3.10555041e-01 1.19595706e+00
-8.00249279e-02 1.27538157e+00 -1.90215841e-01 -7.28666246e-01
5.57997823e-01 1.37618840e-01 -2.75930446e-02 -8.98820579e-01
2.77929902e-01 3.93363953e-01 8.21525138e-03 -6.02278769e-01
8.55464518e-01 -1.92101538e-01 -2.37037793e-01 5.57159603e-01
-8.92076641e-02 -4.90601599e-01 3.53005901e-02 5.44527888e-01
9.51461792e-01 -2.02112794e-01 -1.52279034e-01 1.22423284e-01
2.22962096e-01 2.96192408e-01 -2.10550874e-01 1.19887912e+00
-4.89331305e-01 6.81071460e-01 8.10201824e-01 -2.97959447e-01
-1.19498324e+00 -9.03792799e-01 1.74071744e-01 5.92296243e-01
4.45308089e-01 -3.75694335e-01 -8.74698043e-01 -8.86639655e-01
2.81089097e-02 6.79288924e-01 -9.04370487e-01 -4.15608644e-01
-3.26735139e-01 -7.96133995e-01 1.25997388e+00 -2.14715242e-01
6.37195647e-01 -9.71668363e-01 -6.36877656e-01 -1.56661943e-01
-4.12558168e-01 -1.34676933e+00 -3.67128998e-01 -5.52103519e-01
-4.33649123e-01 -1.08656132e+00 -8.25574398e-01 -5.07419646e-01
6.21134460e-01 5.78076303e-01 1.01608574e+00 3.86799395e-01
-1.84737459e-01 4.67139721e-01 -3.76080036e-01 -1.77820116e-01
-1.03259706e+00 -2.92241067e-01 -8.64244327e-02 3.30131918e-01
1.04853347e-01 -6.14222474e-02 -4.71289188e-01 3.50614101e-01
-1.29356897e+00 9.85800251e-02 5.41593492e-01 8.43726158e-01
-1.45635977e-01 -1.20985106e-01 3.72983962e-01 -7.86189973e-01
1.24114966e+00 -6.11231863e-01 -3.64926755e-01 2.69897789e-01
-4.96491939e-01 -7.67019913e-02 3.71378392e-01 -6.46605492e-01
-8.40051055e-01 -4.61196572e-01 4.71180171e-01 -5.56988716e-01
8.18408653e-02 2.73101240e-01 2.00054407e-01 -3.29548091e-01
9.63189900e-01 2.94065118e-01 -1.97145902e-02 -2.04995889e-02
8.56551051e-01 7.98325360e-01 4.79170978e-01 -4.54440504e-01
1.16621768e+00 8.06143820e-01 -3.80979419e-01 -9.88909125e-01
-5.02260506e-01 1.10288635e-01 -3.22357237e-01 -4.09187973e-01
6.64922595e-01 -7.66409218e-01 -3.94536048e-01 9.90146697e-01
-1.52444983e+00 -2.11053908e-01 3.32491696e-02 -1.40967622e-01
-4.02776480e-01 7.56397367e-01 -7.13374615e-01 -6.23542547e-01
2.47763544e-01 -1.26814890e+00 1.02692878e+00 -3.19588304e-01
-6.04036637e-02 -9.17847276e-01 -1.20137267e-01 7.34412491e-01
3.86043429e-01 3.94248694e-01 9.89365757e-01 -4.96314079e-01
-7.19081104e-01 -1.08225971e-01 -6.40310705e-01 3.54456872e-01
1.93841290e-02 -1.12413727e-01 -9.78227437e-01 -2.79020429e-01
1.21108470e-02 -5.88369787e-01 8.72411072e-01 -2.80983716e-01
7.43826389e-01 -6.93979919e-01 -3.00283432e-01 2.99419075e-01
1.32409620e+00 -9.25598368e-02 7.77307034e-01 3.95543247e-01
8.08283210e-01 9.08991158e-01 3.10662061e-01 1.41963825e-01
3.19376081e-01 5.51313281e-01 5.90743065e-01 -4.87457328e-02
-2.77487397e-01 -4.92690712e-01 4.77920800e-01 6.10015869e-01
2.15892985e-01 -7.41794169e-01 -9.10933137e-01 6.23959661e-01
-1.64489424e+00 -1.00544655e+00 -2.87568957e-01 1.86154687e+00
6.82546973e-01 2.78161056e-02 2.83228494e-02 7.89037719e-03
8.51717949e-01 5.52627802e-01 -1.65164173e-01 -3.67001355e-01
-6.88192666e-01 -2.52595544e-01 5.50918698e-01 4.24103618e-01
-9.86926019e-01 1.31928754e+00 6.87612629e+00 8.76015961e-01
-1.32117152e+00 3.14483434e-01 8.51552129e-01 1.39723584e-01
-6.35940492e-01 -9.95346010e-02 -5.65790869e-02 7.09281266e-01
7.26489425e-01 -5.49054630e-02 5.09110153e-01 4.14321989e-01
2.42014423e-01 -2.21924350e-01 -1.00996387e+00 1.00717878e+00
5.92717230e-01 -1.57933414e+00 4.22386706e-01 3.41890812e-01
8.59803736e-01 -8.61919001e-02 6.61996245e-01 -1.82825550e-01
1.99844167e-01 -1.24636471e+00 1.23401380e+00 2.04604790e-01
7.93035924e-01 -3.99536431e-01 5.35675287e-01 2.40558594e-01
-1.08820282e-01 -1.02151707e-01 -1.18655652e-01 -6.13353541e-03
8.42559859e-02 3.30980033e-01 -8.45084965e-01 2.68113792e-01
2.75344253e-01 5.54512262e-01 -1.00747013e+00 5.28053403e-01
-3.67794961e-01 4.03124571e-01 3.01739097e-01 2.56059468e-02
3.98179978e-01 -4.64588329e-02 6.68744028e-01 1.21822488e+00
9.82357264e-02 -3.62632722e-01 -3.06237340e-01 1.18992198e+00
-2.66461045e-01 -1.22290142e-01 -1.07759166e+00 -6.06658936e-01
3.38244796e-01 8.67339969e-01 -6.76236331e-01 -3.72727334e-01
-2.98657209e-01 1.51071501e+00 2.46778741e-01 4.01956260e-01
-1.03508949e+00 -7.95491412e-03 3.77596408e-01 1.78144425e-01
-1.62112173e-02 -2.35769212e-01 -1.57824263e-01 -1.48679447e+00
3.49840559e-02 -1.49433553e+00 -2.76445419e-01 -1.07368886e+00
-1.45197308e+00 5.50586402e-01 -2.55110096e-02 -8.97545516e-01
-5.75868338e-02 -6.19706929e-01 -2.71651179e-01 5.77827334e-01
-1.52016687e+00 -1.21475375e+00 -2.56129652e-01 4.62837785e-01
3.74969870e-01 -2.11053073e-01 5.13201892e-01 1.57912120e-01
-3.44823539e-01 4.44030374e-01 1.23739980e-01 4.11694348e-01
8.30851734e-01 -8.02581549e-01 6.33737981e-01 9.39701557e-01
4.61261511e-01 5.50726652e-01 8.12623262e-01 -7.80000925e-01
-1.15747976e+00 -7.13352382e-01 9.08551276e-01 -8.98565710e-01
9.37288702e-01 -4.45141822e-01 -6.13957345e-01 4.77848381e-01
4.32394683e-01 -1.24982342e-01 2.25029096e-01 -6.82840049e-01
-8.85214686e-01 5.74355483e-01 -1.22114933e+00 9.06221211e-01
9.08994019e-01 -9.91354704e-01 -4.18466747e-01 4.03301179e-01
5.61774254e-01 -5.31060062e-02 -2.29611963e-01 -5.93825318e-02
6.11819208e-01 -1.43275404e+00 9.06598032e-01 -7.06338227e-01
1.06997478e+00 7.12449178e-02 -9.63818729e-02 -1.42446661e+00
-1.14094336e-02 -7.34481931e-01 1.28805429e-01 8.76279056e-01
3.59481454e-01 -6.56136513e-01 5.88640392e-01 2.93116480e-01
2.90800601e-01 -2.89993346e-01 -8.55763137e-01 -7.84701943e-01
1.22279361e-01 -2.99964249e-01 3.11921716e-01 1.44834387e+00
-7.23509192e-02 2.80679673e-01 -7.58882165e-01 -5.70398755e-02
6.07458055e-01 -3.17884147e-01 6.84135616e-01 -7.72879243e-01
-2.02465624e-01 -5.51685393e-01 -5.00189245e-01 -9.39102232e-01
1.92330644e-01 -5.94229698e-01 -1.55762181e-01 -9.08140838e-01
2.02804700e-01 -2.45200798e-01 1.23741008e-01 1.66847557e-01
5.25312796e-02 6.83585167e-01 4.36520278e-01 6.82376802e-01
-3.31234515e-01 3.88007373e-01 1.50996101e+00 -3.91486287e-01
3.03546309e-01 -5.63092709e-01 -6.65282130e-01 5.43145120e-01
6.81879938e-01 -7.38081276e-01 -2.79368520e-01 -7.79391527e-01
5.92251420e-01 7.79088065e-02 9.82630074e-01 -6.21815503e-01
-9.20911580e-02 -2.06691578e-01 5.88529147e-02 -8.97083338e-03
4.67158198e-01 -5.16889691e-01 -2.38917023e-01 6.03171527e-01
-4.91558164e-01 3.56249094e-01 -6.52926937e-02 8.65428030e-01
-2.14303046e-01 -2.28394508e-01 6.35453045e-01 -2.63627917e-01
-3.82273495e-01 -1.18386596e-01 -7.28217006e-01 1.61207572e-01
8.96048844e-01 -2.62339324e-01 -9.65676665e-01 -8.22675765e-01
-2.84171551e-01 -3.57715279e-01 9.20882523e-01 7.84518659e-01
7.47067094e-01 -1.24798667e+00 -8.02022576e-01 1.02969995e-02
4.21918601e-01 -8.33844721e-01 -6.35865033e-02 7.34698355e-01
-6.10232174e-01 3.01528186e-01 -3.78741473e-01 -3.93333316e-01
-1.02161002e+00 6.88840508e-01 2.34012201e-01 1.45917758e-01
-2.28495479e-01 7.56634235e-01 2.34885469e-01 4.89870608e-02
-2.52889931e-01 -4.38940711e-03 1.32750094e-01 5.90923056e-02
2.65590310e-01 2.22798288e-01 -2.15340033e-01 -1.20708358e+00
-2.69241184e-01 -8.69296864e-02 5.71011901e-02 -6.85000241e-01
8.63867700e-01 -1.44360855e-01 -2.41467744e-01 1.63024247e-01
1.06909990e+00 3.66373569e-01 -1.05058348e+00 4.45165932e-02
5.55515215e-02 -1.02592421e+00 -7.08013996e-02 -9.84908521e-01
-8.80523503e-01 8.43622148e-01 3.51975888e-01 5.80502450e-01
6.00256085e-01 2.85506211e-02 7.26657152e-01 1.54617473e-01
2.77523130e-01 -7.72341788e-01 8.42466235e-01 2.60010421e-01
1.03209126e+00 -1.41478801e+00 -2.67913014e-01 -5.27791202e-01
-7.95177698e-01 7.62979269e-01 2.31667534e-01 -9.92101505e-02
-7.28633478e-02 -2.04934165e-01 2.56912053e-01 -4.50423062e-01
-4.98077601e-01 7.24786073e-02 1.57299548e-01 8.97673488e-01
9.94951203e-02 1.20135151e-01 -2.16116518e-01 1.63385794e-01
-2.28337839e-01 -3.42746168e-01 1.09751034e+00 7.33332574e-01
-1.12551913e-01 -1.07594931e+00 -5.65082788e-01 -1.19076818e-02
-4.96735930e-01 -4.09138590e-01 -1.09436929e+00 8.28751385e-01
1.51144803e-01 1.21818984e+00 -2.20744714e-01 -5.36642134e-01
-1.38823107e-01 -2.09024400e-01 7.51649022e-01 -4.14256930e-01
-5.76760650e-01 -4.41113323e-01 3.09417516e-01 -3.21293354e-01
-3.58448803e-01 -4.40718621e-01 -2.38733590e-01 -4.15455252e-01
-3.91039789e-01 -8.47287104e-02 8.22854161e-01 9.46157515e-01
4.07708853e-01 -1.03517503e-01 6.31788671e-01 -6.30722284e-01
-6.55547917e-01 -7.18250990e-01 -2.01667100e-01 7.93992698e-01
5.00553250e-01 -5.21724224e-01 -5.95226347e-01 1.92762747e-01] | [12.342619895935059, 1.1392738819122314] |
51681c95-d1dd-4c63-8a08-ea5750c17482 | bringing-the-state-of-the-art-to-customers-a | 2302.03222 | null | https://arxiv.org/abs/2302.03222v1 | https://arxiv.org/pdf/2302.03222v1.pdf | Bringing the State-of-the-Art to Customers: A Neural Agent Assistant Framework for Customer Service Support | Building Agent Assistants that can help improve customer service support requires inputs from industry users and their customers, as well as knowledge about state-of-the-art Natural Language Processing (NLP) technology. We combine expertise from academia and industry to bridge the gap and build task/domain-specific Neural Agent Assistants (NAA) with three high-level components for: (1) Intent Identification, (2) Context Retrieval, and (3) Response Generation. In this paper, we outline the pipeline of the NAA's core system and also present three case studies in which three industry partners successfully adapt the framework to find solutions to their unique challenges. Our findings suggest that a collaborative process is instrumental in spurring the development of emerging NLP models for Conversational AI tasks in industry. The full reference implementation code and results are available at \url{https://github.com/VectorInstitute/NAA} | ['Elham Dolatabadi', 'Frank Rudzicz', 'Xiaodan Zhu', 'Deval Pandya', 'Jaswinder Narain', 'Kanishk Patel', 'Kwesi P. Apponsah', 'Sean X. Zhang', 'Yu-Jia Shiah', 'Griffin Tanner', 'Michael Wang', 'Bukola Ishola', 'Erin Li', 'Waqar Muhammad', 'Iris Ren', 'Bencheng Wei', 'Karthik Raja K. Bhaskar', 'Alif Munim', 'Winnie Au', 'Jingcheng Niu', 'Sean Robertson', 'Elaine Lau', 'Tania Sidhom', 'Shirley Wang', 'Faiza Khan Khattak', 'Stephen Obadinma'] | 2023-02-07 | null | null | null | null | ['response-generation'] | ['natural-language-processing'] | [ 3.93132240e-01 3.93219531e-01 -2.59437636e-02 -6.22336268e-01
-8.58304322e-01 -7.63406634e-01 7.40586460e-01 1.29459485e-01
-1.99609548e-01 3.69211763e-01 7.34836817e-01 -5.66017985e-01
-1.26924187e-01 -4.47228611e-01 -1.97126195e-01 -2.73303270e-01
2.58564889e-01 1.13264823e+00 -4.36728835e-01 -5.02203345e-01
3.41995001e-01 4.01441216e-01 -1.21083331e+00 7.38496304e-01
6.88911319e-01 6.60164058e-01 4.34543669e-01 6.80295765e-01
-4.35322165e-01 1.28799939e+00 -5.57046771e-01 -1.70764983e-01
1.30654752e-01 -9.50289816e-02 -1.07497191e+00 -1.25213042e-01
6.83354959e-02 -3.93110812e-01 -8.49857405e-02 7.27980018e-01
4.61096436e-01 2.01451585e-01 3.51828039e-01 -1.57128716e+00
-9.50073242e-01 9.07069445e-01 -1.11274630e-01 -2.18511056e-02
5.80904424e-01 7.18689680e-01 1.10691500e+00 -7.20598996e-01
7.11535275e-01 1.19005084e+00 6.34814858e-01 7.78721869e-01
-9.72557783e-01 -6.06129587e-01 3.16751093e-01 1.79466888e-01
-8.28329444e-01 -6.89023912e-01 9.55119193e-01 -5.36072314e-01
1.64061725e+00 -1.84827726e-02 3.33782732e-01 1.66923666e+00
-2.12572485e-01 1.11529708e+00 7.63619781e-01 -5.08636653e-01
2.99489021e-01 5.61707854e-01 8.30278754e-01 4.35718447e-01
-1.36640906e-01 -2.26262793e-01 -6.42584383e-01 -5.50822854e-01
5.20314217e-01 -2.95037299e-01 -2.27052756e-02 -8.16179216e-02
-1.26733410e+00 1.03762579e+00 5.74393161e-02 4.59249020e-01
-1.06753206e+00 -1.01882026e-01 4.97692943e-01 2.90529281e-01
2.55943775e-01 9.14959311e-01 -5.49282610e-01 -7.59366632e-01
8.17096978e-02 5.60426831e-01 1.48183966e+00 1.10694242e+00
3.20185781e-01 1.60424665e-01 1.30920816e-04 1.26220310e+00
4.31564420e-01 9.05380845e-02 3.46474618e-01 -1.39407361e+00
4.55812007e-01 6.11590087e-01 4.28845286e-01 -7.85392582e-01
-5.88465154e-01 -1.43635154e-01 -2.66467333e-01 -2.12135464e-02
6.06707394e-01 -4.40344363e-01 -5.66353142e-01 1.50443339e+00
1.60764173e-01 -1.31345496e-01 3.25289488e-01 9.27319348e-01
6.66239202e-01 8.60785127e-01 3.11396331e-01 -1.29859120e-01
1.64802933e+00 -1.15250325e+00 -8.51974487e-01 -8.96099508e-01
7.98485637e-01 -7.70476460e-01 1.17761111e+00 4.34037268e-01
-1.19769001e+00 -3.80947024e-01 -9.25260007e-01 -1.46568626e-01
-4.35596764e-01 -5.25294319e-02 1.03855538e+00 3.71997803e-01
-1.03302073e+00 1.22797461e-02 -6.66088343e-01 -6.67242527e-01
1.77550137e-01 3.89738947e-01 -4.99314703e-02 -2.18690097e-01
-1.20386076e+00 8.86204779e-01 -1.04626209e-01 1.12316683e-01
-6.15450919e-01 -6.39581740e-01 -5.96182108e-01 -2.46749416e-01
3.05806726e-01 -9.97050464e-01 2.05233693e+00 -8.60638499e-01
-1.58937430e+00 5.73938906e-01 -2.50938058e-01 -3.76799971e-01
-1.05954148e-02 -3.96278650e-01 -2.98037440e-01 -3.10018539e-01
1.10145308e-01 5.73266327e-01 3.23629767e-01 -1.14775109e+00
-9.52379823e-01 -6.15972459e-01 2.21957099e-02 2.75984734e-01
-2.07042187e-01 4.42099005e-01 -1.37935534e-01 -2.66724288e-01
-4.21501279e-01 -9.13481295e-01 -2.54753977e-01 -6.65428519e-01
1.61749474e-03 -8.92565787e-01 6.97162271e-01 -8.02474916e-01
8.99049640e-01 -2.06215954e+00 6.64008334e-02 9.96672288e-02
1.05298087e-01 2.10333690e-01 -6.11871421e-01 7.72966325e-01
9.47227478e-02 -1.57172322e-01 3.72266173e-02 -3.29277962e-01
5.33275783e-01 5.89445718e-02 -5.30309618e-01 -2.32649773e-01
2.84783244e-01 7.83455193e-01 -1.01191950e+00 1.03618912e-02
1.48208350e-01 6.68859720e-01 -3.58554482e-01 2.79393524e-01
-5.38683832e-01 1.39500082e-01 -7.44094193e-01 7.55905867e-01
1.75771043e-01 -9.01560485e-02 4.07050312e-01 -3.09565980e-02
-1.80909127e-01 8.15132558e-01 -9.02744889e-01 1.59680581e+00
-5.64936697e-01 5.72041631e-01 8.88816714e-01 -6.63340926e-01
9.68553662e-01 4.58815038e-01 4.87806529e-01 -5.00949860e-01
2.90141970e-01 1.85290992e-01 2.73176074e-01 -5.77907324e-01
5.61493337e-01 1.79568172e-01 -7.95525387e-02 9.11908031e-01
-9.58667323e-02 -1.48012921e-01 1.61243051e-01 2.02456936e-01
1.47252047e+00 -1.17779024e-01 2.56372601e-01 3.91743518e-02
2.56711721e-01 4.78979349e-01 4.98502493e-01 9.78655994e-01
-6.47851944e-01 9.98713747e-02 2.01554149e-01 -6.28784001e-01
-1.05186808e+00 -6.43519819e-01 2.44248793e-01 1.47912693e+00
-5.08849382e-01 -3.76037598e-01 -9.05548930e-01 -5.40680349e-01
1.15042880e-01 1.43614256e+00 -1.07005276e-01 1.15830012e-01
-7.86823571e-01 -3.65307242e-01 5.58374703e-01 8.42700779e-01
1.91807002e-01 -1.75984538e+00 -3.38052303e-01 4.30767030e-01
-4.19991702e-01 -1.38144195e+00 -3.19079578e-01 1.65985212e-01
-3.24468583e-01 -5.56965053e-01 -2.14342773e-01 -1.01129353e+00
1.67887881e-01 1.91655695e-01 1.07721460e+00 -2.81767875e-01
-2.20464185e-01 7.56179571e-01 -1.94718257e-01 -9.84547615e-01
-8.77219200e-01 1.83604956e-01 4.31170374e-01 -4.66790825e-01
1.20921910e+00 -6.68739080e-01 -2.59176552e-01 3.74726236e-01
-4.43910271e-01 -4.06699628e-02 8.39102745e-01 4.26538289e-01
-1.57078281e-01 -1.84630856e-01 9.23298359e-01 -7.62308419e-01
1.59719336e+00 -4.57341522e-01 -3.45285535e-01 4.32867557e-02
-5.95992565e-01 -3.65238011e-01 3.22320461e-01 -4.92570668e-01
-1.23886573e+00 3.20137829e-01 -2.60386646e-01 6.64386451e-02
-7.90846705e-01 6.00150168e-01 -8.84481706e-03 4.79747623e-01
9.57228184e-01 -1.17594283e-03 3.64856929e-01 -2.91449904e-01
5.50650001e-01 1.25550520e+00 3.54992658e-01 -6.51939929e-01
2.70704269e-01 -7.16747940e-02 -7.21747398e-01 -9.03910339e-01
-3.69446814e-01 -6.58055842e-01 -2.71934479e-01 -1.33870244e-01
7.80126989e-01 -9.01644766e-01 -1.11904192e+00 3.04049283e-01
-1.64690053e+00 -8.22378993e-01 -3.15226287e-01 5.76319516e-01
-7.68641531e-01 -9.55688357e-02 -1.05314195e+00 -1.00883341e+00
-7.68468797e-01 -1.20874596e+00 8.66566300e-01 2.39236787e-01
-1.21446240e+00 -7.46795654e-01 8.04912224e-02 1.11717749e+00
7.16400325e-01 -6.29849434e-01 1.11928129e+00 -1.28734934e+00
-3.99163961e-01 -2.64154732e-01 -1.53867766e-01 2.97771037e-01
1.11579791e-01 -4.70733494e-01 -1.08402967e+00 1.59770072e-01
2.71497846e-01 -4.16215450e-01 6.05224781e-02 3.13653976e-01
3.28348070e-01 -4.36588556e-01 -4.16838378e-01 -2.32402503e-01
4.89365429e-01 8.18826258e-01 2.96509802e-01 4.85368609e-01
5.64135671e-01 9.82677937e-01 6.63802564e-01 3.30786437e-01
6.18249774e-01 6.42370105e-01 -5.66374697e-02 4.84231710e-01
5.49776591e-02 -6.96590589e-03 6.90022469e-01 8.89416397e-01
-6.97844476e-02 -2.40387723e-01 -1.30555439e+00 6.82538629e-01
-2.25522614e+00 -8.09254646e-01 4.83532362e-02 1.44673038e+00
8.98173988e-01 -2.45908555e-02 3.32028121e-01 -3.57906073e-01
3.65558356e-01 1.46717895e-02 -6.33284092e-01 -8.47143650e-01
3.61907065e-01 -3.09702784e-01 1.15804244e-02 7.34267950e-01
-8.98455083e-01 1.06952369e+00 6.22642088e+00 2.69498318e-01
-5.74718952e-01 -5.64867519e-02 3.48649681e-01 -1.42455161e-01
-2.02066690e-01 -2.58598983e-01 -8.08103502e-01 4.34984677e-02
1.10462439e+00 -1.43114477e-01 1.02876520e+00 1.41752505e+00
2.22738907e-01 4.40712541e-01 -1.40322983e+00 8.04126620e-01
1.33777320e-01 -1.27532721e+00 -2.13760659e-01 1.38694122e-01
1.44318327e-01 4.81485128e-01 -9.41287354e-02 6.63955986e-01
1.00750840e+00 -7.06347346e-01 1.90959886e-01 2.13403627e-01
-2.89483160e-01 -4.48218524e-01 8.07640254e-01 3.06496263e-01
-5.80957294e-01 -5.58541000e-01 2.65474338e-02 -2.51466870e-01
4.74810958e-01 2.56254077e-01 -1.66263926e+00 1.22191779e-01
5.40945292e-01 3.67428184e-01 1.44726545e-01 2.13418588e-01
-6.50209412e-02 6.53943419e-01 -3.13927740e-01 -3.41190159e-01
1.72588050e-01 -3.30189407e-01 8.76075864e-01 1.34557486e+00
-7.16482624e-02 3.48707825e-01 3.00130010e-01 9.81395245e-01
9.69768092e-02 -1.38721289e-02 -8.91563773e-01 -6.31757557e-01
7.66369522e-01 1.30610061e+00 -2.88084656e-01 -5.81229590e-02
-3.93650353e-01 8.85552526e-01 2.29377106e-01 5.24864078e-01
-4.22502637e-01 -3.30648243e-01 1.27314997e+00 -7.97090828e-02
-9.63616837e-03 -3.67421299e-01 -5.31465590e-01 -3.44212919e-01
6.43609986e-02 -1.60743451e+00 3.38439941e-01 -1.05403614e+00
-1.52221036e+00 6.41820252e-01 -2.89044768e-01 -4.91235107e-01
-9.50989604e-01 -8.34939957e-01 -5.65605462e-01 8.04969609e-01
-7.91956127e-01 -1.46253181e+00 -2.53170282e-01 2.62898773e-01
1.03185797e+00 -5.98888218e-01 1.19872057e+00 2.22113714e-01
-6.15265191e-01 1.92381799e-01 -4.97260123e-01 2.09730685e-01
6.12129271e-01 -9.61984932e-01 7.43368149e-01 3.03561091e-01
5.74974297e-03 1.13107216e+00 8.42475832e-01 -6.19204462e-01
-2.04358339e+00 -6.75427973e-01 1.10682476e+00 -9.52167511e-01
1.01702774e+00 -3.94308418e-01 -5.92011154e-01 1.21005166e+00
6.87505305e-01 -8.46667171e-01 6.91803277e-01 6.54545069e-01
-2.78903961e-01 1.11942757e-02 -9.77827609e-01 7.84197152e-01
1.06261349e+00 -8.13831687e-01 -6.93624258e-01 6.67633712e-01
9.93251324e-01 -9.15990025e-02 -6.77881658e-01 2.00794905e-01
6.15861833e-01 -4.34426039e-01 9.28016782e-01 -7.21801460e-01
2.78533608e-01 -1.93233728e-01 -2.29454279e-01 -1.33867788e+00
-6.57731891e-01 -8.75322938e-01 -5.80403954e-02 1.40833807e+00
6.65802002e-01 -8.65257382e-01 4.51674998e-01 1.27859640e+00
-4.81546700e-01 -4.91257370e-01 -3.79883289e-01 -3.28679532e-01
-1.88336372e-01 -8.91377687e-01 4.36775208e-01 1.06672931e+00
7.78641939e-01 1.10475302e+00 1.86964087e-02 2.26676807e-01
2.07795113e-01 -2.78714746e-01 9.02728856e-01 -1.13748908e+00
-3.41928095e-01 -5.72176993e-01 -5.34997247e-02 -8.65478992e-01
3.30024511e-01 -7.42984295e-01 2.63053685e-01 -1.86924577e+00
4.24736319e-03 -2.94237971e-01 -6.36855513e-02 8.71354997e-01
3.88771892e-01 -4.20147806e-01 1.96134359e-01 1.66366845e-01
-3.08226407e-01 7.21697649e-03 5.75475693e-01 -2.38625944e-01
-6.58501744e-01 -6.23245835e-02 -1.42606413e+00 8.33161831e-01
1.08901215e+00 -1.49287209e-01 -6.34155214e-01 -6.29557967e-01
8.59173462e-02 -2.09173396e-01 2.13419482e-01 -6.14208639e-01
5.29152036e-01 -1.76477984e-01 -1.32329851e-01 -2.53123403e-01
7.24271476e-01 -7.76288331e-01 -1.83509037e-01 5.30977082e-03
-6.80280745e-01 1.36746243e-01 4.71335769e-01 1.11772284e-01
6.13708682e-02 -2.86061883e-01 1.23190746e-01 -2.37858742e-01
-7.89905310e-01 -4.20255035e-01 -8.10541868e-01 -3.62531602e-01
8.35186362e-01 1.72199696e-01 -7.81538248e-01 -8.79494488e-01
-4.75088894e-01 3.16367984e-01 1.10861726e-01 8.65975261e-01
3.68942380e-01 -9.59202588e-01 -7.27873385e-01 8.68375748e-02
2.41267815e-01 -2.04548404e-01 1.19019918e-01 6.15133762e-01
-1.68273076e-01 8.91360819e-01 -4.80616279e-02 -2.61386335e-01
-1.21679688e+00 5.17863095e-01 -3.43609601e-02 -2.27830157e-01
-3.88793379e-01 7.75117278e-01 1.06603177e-02 -7.87858367e-01
5.84569812e-01 -1.65412962e-01 -1.19176693e-01 -7.25072846e-02
8.81333947e-01 2.83011287e-01 -9.36439782e-02 -2.07647860e-01
-4.20853227e-01 -7.37798633e-03 -6.03325605e-01 -4.70996827e-01
1.40154743e+00 1.00294828e-01 -5.22032641e-02 3.86140049e-01
7.18431890e-01 -2.42113948e-01 -5.79937100e-01 -1.90853566e-01
1.87232271e-01 7.67629743e-02 8.22190642e-02 -1.21739662e+00
-3.99557352e-01 4.66424555e-01 3.77527863e-01 4.28406984e-01
7.43941426e-01 2.71484286e-01 9.81300950e-01 9.12297785e-01
2.44306117e-01 -1.45424139e+00 -6.97696656e-02 6.77619338e-01
1.29206145e+00 -1.08902121e+00 -2.44473919e-01 -4.48615223e-01
-1.18908167e+00 7.64807880e-01 7.42701113e-01 3.50547582e-01
3.98124218e-01 6.93491220e-01 6.11452997e-01 -4.15269405e-01
-1.22884309e+00 1.24018663e-03 -2.74441421e-01 9.26835299e-01
7.81483173e-01 4.21805494e-03 -3.99925448e-02 1.06929135e+00
-1.95913404e-01 2.54407436e-01 1.95135817e-01 1.11353028e+00
-1.02016993e-01 -1.28200281e+00 -1.22333907e-01 2.16101065e-01
-1.57360863e-02 -1.72492489e-01 -9.28989410e-01 5.60663521e-01
-4.70025510e-01 1.32709682e+00 1.81477278e-01 -5.69491744e-01
6.51359856e-01 7.04350889e-01 8.90561268e-02 -4.74901378e-01
-8.69336784e-01 -1.36298895e-01 9.71689403e-01 -6.97710812e-01
-4.16880548e-02 -9.35104430e-01 -1.32012022e+00 -2.43134469e-01
2.10483134e-01 -1.68333165e-02 9.55365360e-01 6.44407868e-01
9.62821662e-01 3.90731901e-01 1.47407562e-01 -7.80866086e-01
-6.61691427e-01 -1.33299828e+00 3.48243900e-02 3.50056887e-01
-2.02555228e-02 -5.99561259e-02 -2.70043820e-01 1.92253254e-02] | [12.680472373962402, 7.904555797576904] |
c6fa281a-3790-422d-8e33-0cc29a725f7e | visual-speech-recognition-for-multiple | 2202.13084 | null | https://arxiv.org/abs/2202.13084v2 | https://arxiv.org/pdf/2202.13084v2.pdf | Visual Speech Recognition for Multiple Languages in the Wild | Visual speech recognition (VSR) aims to recognize the content of speech based on lip movements, without relying on the audio stream. Advances in deep learning and the availability of large audio-visual datasets have led to the development of much more accurate and robust VSR models than ever before. However, these advances are usually due to the larger training sets rather than the model design. Here we demonstrate that designing better models is equally as important as using larger training sets. We propose the addition of prediction-based auxiliary tasks to a VSR model, and highlight the importance of hyperparameter optimization and appropriate data augmentations. We show that such a model works for different languages and outperforms all previous methods trained on publicly available datasets by a large margin. It even outperforms models that were trained on non-publicly available datasets containing up to to 21 times more data. We show, furthermore, that using additional training data, even in other languages or with automatically generated transcriptions, results in further improvement. | ['Maja Pantic', 'Stavros Petridis', 'Pingchuan Ma'] | 2022-02-26 | null | null | null | null | ['lipreading'] | ['computer-vision'] | [ 2.59108454e-01 1.94281526e-03 -2.64958233e-01 -1.95423365e-01
-1.35149550e+00 -5.51681042e-01 6.36873603e-01 -1.24846004e-01
-3.74957681e-01 5.10631680e-01 6.65246606e-01 -3.78206819e-01
3.64227861e-01 1.16255134e-02 -5.73846698e-01 -4.47922140e-01
2.15810820e-01 4.52077955e-01 6.32986948e-02 -1.26915172e-01
5.11667095e-02 4.68966782e-01 -1.83874822e+00 2.97877192e-01
3.28710377e-01 9.72079635e-01 2.41784960e-01 8.45563650e-01
-1.80342607e-02 5.29142499e-01 -6.71467006e-01 -1.21920094e-01
1.50661051e-01 -2.50350446e-01 -6.08541012e-01 2.74914205e-01
4.73060369e-01 -1.62707880e-01 -3.79140139e-01 4.55217689e-01
9.46236610e-01 7.01703355e-02 3.85062128e-01 -1.21855783e+00
-7.73173809e-01 2.80778915e-01 -1.77878067e-01 1.81585811e-02
2.65129775e-01 3.41205835e-01 1.13631511e+00 -9.54872370e-01
5.44047713e-01 1.14602852e+00 6.09645486e-01 9.67111468e-01
-1.40477610e+00 -5.66762149e-01 4.19530971e-03 1.85760498e-01
-1.45397961e+00 -1.28706372e+00 7.04571426e-01 -2.89919406e-01
1.26972842e+00 4.41587508e-01 6.68398976e-01 1.50573039e+00
-7.04970241e-01 9.05768991e-01 1.00374830e+00 -5.31708479e-01
1.07142180e-01 4.59713727e-01 -4.12892878e-01 3.37715775e-01
-2.11880073e-01 2.53261983e-01 -7.22003758e-01 -1.13664433e-01
5.50485551e-01 -5.25407434e-01 -5.74153602e-01 -3.55120867e-01
-1.21702123e+00 8.26506972e-01 -1.46759406e-01 2.87499756e-01
-5.88356555e-02 -5.53119043e-03 5.18205583e-01 3.03571463e-01
5.06603181e-01 2.85851002e-01 -7.63316691e-01 -4.26384866e-01
-1.24151313e+00 -1.32058993e-01 7.32430339e-01 7.78107285e-01
3.40376616e-01 5.29307008e-01 4.62176166e-02 1.31433988e+00
4.79319364e-01 5.06836891e-01 7.68251717e-01 -1.00265789e+00
3.95267397e-01 2.11601913e-01 -7.24191517e-02 -4.77946728e-01
-3.30295086e-01 -5.06464720e-01 -4.47565496e-01 2.53569096e-01
4.42583323e-01 2.62799580e-02 -1.16807091e+00 1.88330173e+00
-2.32487321e-02 6.41771257e-02 1.73779801e-01 8.52518797e-01
1.02033484e+00 5.58213532e-01 -1.13945678e-02 -1.48125991e-01
1.12407565e+00 -8.26226890e-01 -7.32826710e-01 -2.81692594e-01
5.92237949e-01 -1.00018454e+00 1.40213287e+00 6.47159159e-01
-9.23932493e-01 -4.05042320e-01 -9.93000090e-01 -2.20524743e-01
-2.46339291e-01 5.08934438e-01 4.60454047e-01 8.73164296e-01
-1.45322490e+00 1.24815181e-01 -5.62880397e-01 -4.34693009e-01
3.30038428e-01 3.59309375e-01 -6.30327165e-01 3.10310513e-01
-8.92043889e-01 8.38609874e-01 -1.13178253e-01 -6.69923276e-02
-9.44633126e-01 -5.45575142e-01 -8.78365874e-01 -2.38784254e-02
2.63746113e-01 -2.93444574e-01 1.30855215e+00 -1.43728113e+00
-1.78666484e+00 1.06348538e+00 -3.77783626e-01 -4.96733725e-01
4.87991720e-01 6.83085248e-02 -3.99239331e-01 1.31071553e-01
-4.71570462e-01 8.84926558e-01 1.05350721e+00 -1.21646905e+00
-3.53346109e-01 -1.75760210e-01 -2.84391493e-01 1.01865269e-01
-3.53596926e-01 2.95367092e-01 -6.66856885e-01 -7.21717000e-01
-2.04766348e-01 -1.02811706e+00 2.60797795e-02 1.00677125e-02
-1.67362079e-01 -2.09949300e-01 6.06498480e-01 -1.08147192e+00
9.99072134e-01 -2.45286274e+00 -3.18550132e-02 -5.89535646e-02
3.03486794e-01 6.35504603e-01 -4.01868016e-01 2.24687725e-01
-2.64305949e-01 1.85098186e-01 -7.85507783e-02 -7.24904776e-01
2.00960729e-02 4.60162647e-02 -4.29870516e-01 4.83618766e-01
2.70822585e-01 7.04332471e-01 -4.65834349e-01 -3.79083991e-01
3.88435781e-01 8.70455682e-01 -6.12894297e-01 1.40601555e-02
-7.89519921e-02 3.87406617e-01 2.39465505e-01 5.88949919e-01
3.03246170e-01 -2.50019044e-01 4.26783413e-02 6.32339567e-02
-1.06601596e-01 6.31943107e-01 -9.85427320e-01 1.53832269e+00
-6.80504143e-01 1.27291918e+00 4.20350611e-01 -1.09115505e+00
8.87182295e-01 8.63406539e-01 4.46784675e-01 -5.99928021e-01
1.03822477e-01 2.66021639e-01 8.72948095e-02 -5.10818958e-01
2.68841296e-01 -1.14768229e-01 4.11185980e-01 3.23821127e-01
2.83598900e-01 -9.07322690e-02 -7.74415880e-02 7.86584243e-02
7.46142626e-01 -1.25539273e-01 -2.59181391e-02 1.30090177e-01
4.92703378e-01 -3.22486043e-01 5.19347787e-01 5.20036280e-01
-4.15913582e-01 8.85613501e-01 2.68401861e-01 -7.96910524e-02
-1.26998603e+00 -7.89998770e-01 -3.94834965e-01 1.15059197e+00
-3.71208459e-01 -6.78685844e-01 -5.27612448e-01 -3.66406441e-01
5.93197085e-02 6.55077100e-01 -4.73155290e-01 3.09444033e-02
-3.85959059e-01 -5.89200020e-01 8.28738809e-01 6.81965232e-01
-9.04762223e-02 -1.03690588e+00 -1.63328379e-01 9.94407013e-02
-2.21068144e-01 -1.33561707e+00 -3.67368907e-01 1.19797237e-01
-6.07248247e-01 -6.78034782e-01 -1.03149855e+00 -6.57490134e-01
2.53013581e-01 4.95165922e-02 8.75479043e-01 -2.17319932e-02
-3.77290517e-01 6.56095564e-01 -2.13419706e-01 -4.88336712e-01
-7.30790794e-01 1.18714094e-01 1.87360048e-01 2.01282073e-02
3.04225832e-01 -4.34476882e-01 -2.42625400e-01 1.81006998e-01
-5.96194744e-01 1.24063261e-01 6.33223116e-01 7.86012769e-01
3.30837488e-01 -5.47911644e-01 9.11271751e-01 -3.88297409e-01
5.80790102e-01 -1.45085320e-01 -6.67483687e-01 1.58838764e-01
-7.44225085e-01 1.57494798e-01 2.74405271e-01 -6.39822125e-01
-8.02677274e-01 9.63675156e-02 -6.19587421e-01 -6.42055154e-01
-4.34155792e-01 1.16588727e-01 -7.89445639e-02 -1.13685288e-01
4.87670481e-01 3.27789307e-01 2.44811296e-01 -7.41676331e-01
3.66896003e-01 1.15502870e+00 2.91425228e-01 -1.46083072e-01
5.55235803e-01 3.62076581e-01 -3.01121503e-01 -1.39668107e+00
-3.12211812e-01 -4.50350881e-01 -3.34435344e-01 -1.57281622e-01
6.99851513e-01 -9.66714740e-01 -6.35682464e-01 4.03164983e-01
-9.28781807e-01 -5.31390607e-01 -2.30247408e-01 6.18877709e-01
-6.18391931e-01 3.59070748e-01 -3.92217875e-01 -1.07202613e+00
-2.40920827e-01 -1.24666762e+00 1.17698753e+00 -2.88324833e-01
-2.27750406e-01 -8.67094874e-01 1.58160552e-01 8.07863772e-01
5.24701774e-01 -3.53669763e-01 6.25832617e-01 -1.00580537e+00
-4.08960819e-01 -1.44765720e-01 -1.19276255e-01 6.73299134e-01
2.79018104e-01 2.73120075e-01 -1.45195520e+00 -2.51155198e-01
-2.41025835e-01 -6.73392713e-01 1.06517935e+00 3.59010071e-01
1.17077529e+00 -3.29358101e-01 -1.54316917e-01 5.87473094e-01
9.26841676e-01 1.35250345e-01 4.88907546e-01 1.34958476e-01
5.16366661e-01 7.13927090e-01 1.37195677e-01 1.10495470e-01
4.32572961e-01 1.00403583e+00 1.44121379e-01 -2.28050888e-01
-6.33207023e-01 -3.90477687e-01 4.77270037e-01 1.00015879e+00
8.24079961e-02 -2.28185713e-01 -1.08249426e+00 8.14442337e-01
-1.41463971e+00 -9.03262615e-01 1.96253926e-01 2.47122908e+00
9.64082956e-01 -2.24971585e-02 3.68843347e-01 3.96184206e-01
5.68664312e-01 1.41764328e-01 -5.02408803e-01 -3.48621637e-01
-4.48655099e-01 2.50086904e-01 2.46344656e-01 7.39950836e-01
-9.85350192e-01 1.06799281e+00 7.66707611e+00 7.85282493e-01
-1.68121707e+00 5.60672730e-02 4.62731063e-01 -5.18839180e-01
-3.20577323e-01 -3.14909816e-01 -6.69744194e-01 2.83192843e-01
1.08523881e+00 1.46337286e-01 6.84662461e-01 6.52488172e-01
3.70346159e-01 1.97145224e-01 -9.35080230e-01 1.38140440e+00
3.94035041e-01 -1.23374748e+00 -2.02464134e-01 1.07994847e-01
3.47374737e-01 3.47784787e-01 3.40998650e-01 1.65376946e-01
-1.42231397e-02 -1.28069210e+00 8.47104311e-01 5.62375113e-02
8.95727277e-01 -5.15737355e-01 2.89677739e-01 2.23340720e-01
-1.04018903e+00 -6.19606525e-02 -9.82764885e-02 2.58206367e-01
-3.02813035e-02 9.72807556e-02 -1.20566618e+00 3.37215513e-02
6.15217090e-01 5.37605047e-01 -7.08675802e-01 1.07910645e+00
-1.56225115e-01 1.00271249e+00 -3.73841792e-01 -4.40930538e-02
-1.06205024e-01 2.17453137e-01 6.62303567e-01 1.28079259e+00
3.60644460e-02 -3.25493693e-01 -2.76418447e-01 5.13104320e-01
-8.04500207e-02 3.53623897e-01 -6.41119123e-01 -4.41251963e-01
4.64152545e-01 1.20369625e+00 -1.52207002e-01 -2.09203467e-01
-7.49745846e-01 6.32929981e-01 3.61038059e-01 4.96084452e-01
-4.83199388e-01 -1.71208590e-01 8.44343245e-01 2.23088965e-01
3.87451291e-01 -3.82625937e-01 -2.91624255e-02 -1.09218740e+00
-4.59259190e-02 -1.28384018e+00 1.12190634e-01 -8.88517678e-01
-1.01609719e+00 6.48560166e-01 -4.40791816e-01 -1.05967104e+00
-5.32785714e-01 -7.00491846e-01 -1.74069330e-01 9.31073368e-01
-1.68280077e+00 -1.07099187e+00 1.20090120e-01 5.52056849e-01
7.19464064e-01 -4.44020063e-01 9.05546188e-01 3.76602709e-01
-4.89107370e-01 8.69449794e-01 7.58598670e-02 2.77119994e-01
9.81404006e-01 -1.05721366e+00 4.25490797e-01 5.14143109e-01
6.38682723e-01 3.72957259e-01 6.71757996e-01 -1.68590665e-01
-1.18445802e+00 -5.69513798e-01 9.72703636e-01 -5.99617541e-01
6.47750616e-01 -8.15430403e-01 -9.15538311e-01 7.48065054e-01
2.53570855e-01 -4.27814201e-02 9.86910343e-01 2.81295866e-01
-5.80367684e-01 -1.35731921e-01 -8.77438247e-01 4.33032155e-01
9.49000716e-01 -1.01027882e+00 -5.64226031e-01 1.26601294e-01
7.55128264e-01 -2.30110407e-01 -6.01411581e-01 3.83099884e-01
5.99305570e-01 -6.75684154e-01 9.87665355e-01 -8.71685207e-01
5.34698553e-02 -1.71886131e-01 -2.69970804e-01 -1.19124401e+00
4.86583114e-02 -7.87616789e-01 -3.80856590e-03 1.37122273e+00
6.90136611e-01 -6.25557661e-01 7.48026550e-01 4.88083422e-01
-5.34319878e-02 -4.61264908e-01 -1.14888859e+00 -7.47419894e-01
-4.43818793e-02 -8.78879726e-01 3.74666393e-01 9.41587031e-01
-7.64949098e-02 3.88053536e-01 -5.17454982e-01 2.31031161e-02
4.44009185e-01 -1.79794922e-01 7.71573126e-01 -1.25193310e+00
-3.21684897e-01 -5.29343784e-01 -5.81535101e-01 -1.07716095e+00
4.04189438e-01 -8.71480763e-01 7.94603974e-02 -1.41278481e+00
-6.55814782e-02 -5.78191757e-01 -2.38154590e-01 8.00529420e-01
7.42165430e-04 4.85960424e-01 2.10756049e-01 7.42671192e-02
-2.54431725e-01 6.12387300e-01 9.61222351e-01 -9.19851437e-02
-2.38667533e-01 8.32197368e-02 -7.58256316e-01 6.55608416e-01
7.81812131e-01 -1.65036872e-01 -4.02778268e-01 -4.18596059e-01
1.21463224e-01 -5.19591905e-02 3.68875414e-01 -8.22842836e-01
1.30334020e-01 2.14679435e-01 3.44805270e-01 -3.47634763e-01
8.84735286e-01 -4.97853816e-01 -1.64573744e-01 3.37679833e-02
-6.56503439e-01 -2.07268268e-01 4.98679817e-01 3.64726126e-01
-1.74382582e-01 -2.08769143e-02 8.30014884e-01 1.98364958e-01
-4.65313613e-01 2.34195609e-02 -6.09207749e-01 1.65595546e-01
5.25797069e-01 -1.20556496e-01 -1.34309724e-01 -8.11962247e-01
-9.53972042e-01 -1.90896928e-01 4.28535938e-01 7.40104437e-01
5.37998319e-01 -1.26742518e+00 -8.00941527e-01 3.95780146e-01
2.32609034e-01 -4.44914073e-01 1.10833235e-02 7.87831187e-01
-6.13166094e-02 7.75226295e-01 1.29373744e-01 -6.65871203e-01
-1.69900358e+00 4.85301733e-01 3.11466396e-01 3.36786777e-01
-5.47667861e-01 8.08625281e-01 -1.36078242e-03 -4.35666054e-01
7.98861444e-01 -1.60802588e-01 -2.23108754e-01 7.28241280e-02
5.04843712e-01 2.23366126e-01 2.25084811e-01 -9.50430572e-01
-4.09608155e-01 3.41089666e-01 1.16433643e-01 -5.66120446e-01
1.44552982e+00 -1.71320830e-02 4.42095429e-01 5.97975492e-01
1.07662594e+00 5.38942814e-01 -1.31744123e+00 -1.80644810e-01
-2.03654930e-01 -4.56082195e-01 3.28931630e-01 -9.32102084e-01
-1.19410717e+00 1.21823382e+00 7.13448107e-01 2.11191058e-01
1.03429019e+00 2.61532217e-01 4.83158916e-01 1.78803459e-01
8.06736872e-02 -1.13394272e+00 1.59481645e-01 2.70356506e-01
1.05551529e+00 -1.29709077e+00 -3.54308814e-01 -7.56756589e-02
-8.89459491e-01 9.00433064e-01 2.69941032e-01 4.12426263e-01
4.70915735e-01 4.04892504e-01 5.13209820e-01 3.77081990e-01
-1.04702353e+00 -4.95183885e-01 4.65215981e-01 8.77019763e-01
6.96023941e-01 -1.12594217e-01 1.66792229e-01 4.69827592e-01
-2.74358034e-01 -2.94974428e-02 1.62175149e-01 4.81007427e-01
-2.85986781e-01 -1.38007164e+00 -3.36452842e-01 3.07082951e-01
-4.77677345e-01 -2.70798385e-01 -5.81330955e-01 7.31604755e-01
-2.02904716e-01 1.20783269e+00 6.06510937e-02 -2.37249851e-01
8.07103068e-02 3.60012323e-01 5.46681046e-01 -5.39247513e-01
-3.90750855e-01 2.81969905e-01 2.82932132e-01 -4.44567293e-01
-2.22989127e-01 -8.77027869e-01 -8.61728072e-01 -6.90563098e-02
-2.96404719e-01 -9.79705900e-02 9.88680005e-01 8.27369928e-01
6.31025434e-01 2.97128409e-01 5.67993104e-01 -7.48964012e-01
-5.41570127e-01 -1.06791246e+00 -3.33325714e-01 1.41184703e-01
8.38098466e-01 -5.34634054e-01 -7.83775270e-01 8.03469792e-02] | [14.368806838989258, 5.11421537399292] |
b7f5974b-448a-407a-860d-107d3471e432 | collaborative-regret-minimization-in-multi | 2301.11442 | null | https://arxiv.org/abs/2301.11442v1 | https://arxiv.org/pdf/2301.11442v1.pdf | Collaborative Regret Minimization in Multi-Armed Bandits | In this paper, we study the collaborative learning model, which concerns the tradeoff between parallelism and communication overhead in multi-agent reinforcement learning. For a fundamental problem in bandit theory, regret minimization in multi-armed bandits, we present the first and almost tight tradeoffs between the number of rounds of communication between the agents and the regret of the collaborative learning process. | ['Qin Zhang', 'Nikolai Karpov'] | 2023-01-26 | null | null | null | null | ['multi-armed-bandits'] | ['miscellaneous'] | [-6.32102847e-01 1.69485092e-01 -6.69515669e-01 -1.55626148e-01
-8.84177029e-01 -5.73538899e-01 4.22187716e-01 3.22641969e-01
-7.58698702e-01 1.25490415e+00 6.25266358e-02 -5.71079314e-01
-6.95613980e-01 -7.36004174e-01 -6.23400509e-01 -7.60970592e-01
-5.67371726e-01 7.78548658e-01 -7.50240311e-02 -3.46433856e-02
1.50521800e-01 3.99177998e-01 -8.89461696e-01 9.61776599e-02
6.25102937e-01 1.08344162e+00 -2.62662828e-01 1.05344427e+00
-1.68831348e-01 1.16004145e+00 -7.02097595e-01 -3.65350813e-01
4.54632670e-01 -6.28904164e-01 -9.46234345e-01 2.65657216e-01
-2.02675983e-01 -7.07231581e-01 -5.53067029e-01 1.01169634e+00
3.38742018e-01 2.65791208e-01 5.45607924e-01 -1.43470359e+00
-1.32368863e-01 1.24065506e+00 -1.17109096e+00 5.06470203e-01
-1.75964460e-02 -1.46205261e-01 1.22172368e+00 2.03774989e-01
2.16951951e-01 1.40557170e+00 2.15351567e-01 3.53639454e-01
-1.10116959e+00 -5.34387648e-01 3.97392184e-01 3.41399282e-01
-6.43222630e-01 -3.97300571e-01 3.68256569e-01 -1.99757278e-01
8.07783425e-01 1.13116749e-01 9.13391113e-01 4.20438170e-01
3.74911994e-01 1.24141097e+00 1.14068568e+00 -5.95907092e-01
7.83083320e-01 -1.50064602e-01 1.59676209e-01 6.67844236e-01
6.19743049e-01 6.06339395e-01 -6.76865339e-01 -6.92587256e-01
1.21838069e+00 1.63833220e-02 2.43432775e-01 -5.85254729e-01
-6.02069318e-01 1.22537327e+00 1.62712023e-01 -7.08208457e-02
-6.21238887e-01 7.85292089e-01 4.80277091e-01 1.01833355e+00
5.60914457e-01 2.65632570e-01 -4.14162248e-01 -2.62003899e-01
-6.24174118e-01 1.90741137e-01 1.11912215e+00 6.82304323e-01
7.06251323e-01 -1.65558979e-01 1.17685407e-01 7.15620816e-01
4.59028721e-01 4.49287683e-01 2.19024047e-01 -1.21818256e+00
6.03345633e-01 5.32657541e-02 7.23821223e-01 -2.76990712e-01
-5.21303296e-01 -2.52228171e-01 -5.36690593e-01 5.37499309e-01
5.03393412e-01 -1.08514607e+00 -1.52700752e-01 1.29514813e+00
6.11383557e-01 -2.91094929e-01 4.09593165e-01 8.49166334e-01
-7.67770633e-02 4.97640580e-01 -2.19206139e-01 -8.89423013e-01
7.74315834e-01 -1.47997117e+00 -7.26350129e-01 -4.24216628e-01
5.13008118e-01 -7.48996198e-01 1.74128875e-01 1.99471608e-01
-1.62605393e+00 4.82368581e-02 -8.55316997e-01 5.47629118e-01
1.84918612e-01 -3.32879663e-01 1.19367254e+00 7.31272459e-01
-6.77735090e-01 5.26174188e-01 -1.00750077e+00 -6.85507953e-02
2.71795809e-01 4.33132738e-01 2.73870617e-01 1.50394171e-01
-5.23371577e-01 6.98871791e-01 4.08021063e-01 -1.95299760e-01
-8.97445679e-01 -3.38986605e-01 -1.37607962e-01 1.07163593e-01
7.91016757e-01 -6.26764774e-01 1.88144982e+00 -1.23834622e+00
-2.04816937e+00 2.07683191e-01 2.30553791e-01 -8.08500648e-01
8.29402149e-01 -4.15877938e-01 1.50922954e-01 -1.26268372e-01
-2.03220233e-01 -3.58718820e-02 7.34299898e-01 -9.40505564e-01
-1.14521551e+00 -4.87841487e-01 4.79206592e-01 6.72402263e-01
-1.17541263e-02 -3.37171346e-01 7.74711370e-02 -3.06277066e-01
-2.08902583e-01 -1.10738671e+00 -6.40442729e-01 -4.35353458e-01
2.27841511e-01 -5.47608972e-01 2.01652363e-01 2.77621150e-02
5.59751153e-01 -2.25493884e+00 1.11352973e-01 2.94385076e-01
5.28328009e-02 -2.33779699e-01 -3.22624803e-01 8.28156888e-01
2.70648777e-01 -2.00229704e-01 5.92397749e-01 -7.50866979e-02
2.54955083e-01 4.56951708e-01 -3.05953532e-01 7.38339365e-01
-7.11662352e-01 7.65445232e-01 -9.06900346e-01 -2.26938173e-01
-1.00747406e-01 -7.16047764e-01 -5.45095801e-01 5.94425380e-01
-4.72194612e-01 4.35660966e-02 -8.65285456e-01 9.73032266e-02
5.56166589e-01 -4.05630767e-01 8.39633584e-01 5.64564645e-01
-1.77557752e-01 2.18784839e-01 -1.40072691e+00 1.36952174e+00
-5.27931154e-01 1.76860049e-01 5.90498447e-01 -9.56055105e-01
3.34169209e-01 3.89466614e-01 7.88299441e-01 -6.90896630e-01
-5.24567030e-02 6.48045242e-02 -1.77760497e-01 -1.83393463e-01
3.61596316e-01 -1.18637010e-01 -8.51297900e-02 1.47224021e+00
-1.75118431e-01 -1.49017438e-01 4.96940404e-01 6.44796044e-02
9.25706565e-01 -2.94250757e-01 7.05443144e-01 -2.16641083e-01
-1.89188998e-02 2.42547452e-01 2.87316263e-01 1.47598934e+00
-1.67261019e-01 -5.58389068e-01 6.57886267e-01 -8.62042546e-01
-8.85460615e-01 -8.79849434e-01 4.51139838e-01 1.72949111e+00
1.80798844e-01 -1.22996874e-01 -4.55150753e-01 -6.60665452e-01
4.76325423e-01 2.37150162e-01 -4.73667502e-01 3.26106369e-01
-2.86795795e-01 -8.52386594e-01 1.03503421e-01 3.12146187e-01
4.64028865e-01 -6.19902134e-01 -8.67614925e-01 5.94476521e-01
2.31217071e-01 -8.19541931e-01 -5.09600639e-01 3.79098207e-01
-9.31391001e-01 -1.36219299e+00 -3.57967734e-01 -3.49940181e-01
3.79733860e-01 4.64893490e-01 9.05074477e-01 1.18981972e-01
1.41148925e-01 7.31554449e-01 -3.50560606e-01 -7.32032955e-01
-2.49606878e-01 5.62923811e-02 -8.62009823e-02 -2.36542970e-01
-2.11110245e-02 -3.64137471e-01 -6.65868640e-01 1.03329524e-01
-7.86074519e-01 -6.27828529e-03 5.99321425e-01 9.24500704e-01
2.98830688e-01 5.28476052e-02 4.70652759e-01 -6.31696522e-01
1.14001238e+00 -2.53679961e-01 -1.23074687e+00 4.91073519e-01
-3.11462849e-01 2.31910303e-01 4.69328105e-01 -3.72292995e-01
-9.01533604e-01 -7.48364180e-02 3.59672487e-01 2.80013204e-01
3.43971729e-01 4.29308981e-01 6.29321396e-01 -2.39469498e-01
4.18706536e-01 3.18269245e-02 2.54842788e-01 -6.56708553e-02
5.10958552e-01 6.07105076e-01 -3.69188040e-01 -7.61403918e-01
1.51233852e-01 3.21476072e-01 2.49053501e-02 -5.19363046e-01
-1.02912128e+00 -3.04148346e-01 -1.72079727e-03 -8.24295953e-02
9.09125209e-02 -9.12554026e-01 -1.39094996e+00 1.01940438e-01
-1.02301836e+00 -5.92717290e-01 -6.54051542e-01 9.70578611e-01
-1.38854492e+00 2.57922351e-01 -7.61838198e-01 -1.25705135e+00
-3.74203682e-01 -7.11058736e-01 2.23104686e-01 2.45425165e-01
1.94839299e-01 -9.03117359e-01 6.61699593e-01 2.31552482e-01
2.69315600e-01 -3.16989720e-01 8.60659063e-01 -6.50889814e-01
-6.61038995e-01 1.43475652e-01 -1.15492873e-01 -6.79924041e-02
2.71729082e-02 -4.30698246e-01 -1.68021858e-01 -7.32713699e-01
3.80606093e-02 -7.39348531e-01 5.47037363e-01 7.17118800e-01
7.24296451e-01 -7.47817576e-01 -1.69256210e-01 9.08369720e-02
1.60885274e+00 3.75489056e-01 1.46044537e-01 5.05171299e-01
-3.90687436e-02 2.10288331e-01 7.39898264e-01 1.22386539e+00
1.00671284e-01 2.62235105e-01 3.84230494e-01 1.62588552e-01
4.79460150e-01 2.28324477e-02 2.29282126e-01 6.04599416e-01
-9.34032425e-02 -9.83345956e-02 -7.08342195e-01 3.11362147e-01
-2.62444043e+00 -9.08323884e-01 6.01363420e-01 2.19958615e+00
9.57419813e-01 -8.26710686e-02 6.37952089e-01 -2.58789212e-01
5.09104133e-01 1.78121522e-01 -9.20956552e-01 -7.41603196e-01
1.98706567e-01 -1.41420960e-01 1.05119777e+00 8.26186597e-01
-9.21372890e-01 8.71858418e-01 8.59833336e+00 8.95340383e-01
-5.95393479e-01 2.51676947e-01 4.65824991e-01 -8.18086028e-01
1.17076799e-01 -1.93755791e-01 -6.17876530e-01 1.44403204e-01
9.63705540e-01 -5.98039746e-01 1.15108848e+00 9.31422532e-01
1.51682794e-01 -3.71608436e-01 -1.25863540e+00 9.35554087e-01
-5.51572025e-01 -1.61251569e+00 -2.22521722e-01 2.46907249e-01
1.08660376e+00 4.50305074e-01 -2.80755043e-01 1.97369888e-01
1.52906621e+00 -6.05481684e-01 5.35512567e-01 7.45476559e-02
-3.12882289e-02 -1.25153685e+00 6.54458523e-01 5.98483801e-01
-6.86673164e-01 -6.55707121e-01 -4.54580486e-01 -5.42788386e-01
-1.34247364e-02 4.88791198e-01 -4.28858638e-01 4.14895624e-01
5.57407856e-01 1.84503734e-01 3.66225094e-01 1.26882184e+00
-5.12835803e-03 3.61180276e-01 -5.02375662e-01 -5.08607805e-01
6.75897896e-01 -5.67475796e-01 9.44307148e-02 8.51297617e-01
-7.23274574e-02 3.05406064e-01 6.70051754e-01 1.27099017e-02
-1.14641517e-01 2.20904186e-01 -2.03939840e-01 -1.73376545e-01
5.03139019e-01 8.22841108e-01 -5.17388284e-01 -4.09339458e-01
-4.12981898e-01 5.39230943e-01 8.54640484e-01 3.04229319e-01
-6.58167124e-01 2.90838510e-01 8.98894966e-01 -3.96042675e-01
6.28138304e-01 -4.31483924e-01 -1.64223358e-01 -8.25743496e-01
-2.57225454e-01 -8.39269221e-01 6.90985560e-01 7.65796602e-02
-1.63756430e+00 -1.76843610e-02 -2.85190314e-01 -6.13769710e-01
-5.29971123e-01 -2.64404893e-01 -3.21587473e-01 2.52142340e-01
-1.69661987e+00 -5.59834599e-01 5.24035931e-01 7.74561942e-01
5.22790134e-01 -3.53692025e-01 6.82902575e-01 -2.39484832e-01
-4.18717742e-01 4.26467627e-01 1.15443707e+00 -1.37447298e-01
3.50661397e-01 -1.18316543e+00 -1.00530647e-01 2.65035450e-01
1.00855920e-02 1.21745490e-01 7.08742201e-01 -3.06651503e-01
-1.86934769e+00 -7.13551521e-01 -1.16895989e-03 5.44021785e-01
1.27551937e+00 1.81932941e-01 -2.09207572e-02 8.88391912e-01
4.95555669e-01 -1.60990700e-01 9.40762520e-01 7.15881944e-01
-2.21190721e-01 -2.34695688e-01 -9.97956932e-01 4.73232925e-01
5.66754818e-01 -2.03323230e-01 -3.08559507e-01 7.24778950e-01
3.70677352e-01 -3.66065919e-01 -8.06747317e-01 -4.39447522e-01
7.06407964e-01 -7.67889738e-01 4.98759359e-01 -8.89210641e-01
7.42705837e-02 3.13854635e-01 -9.63545498e-03 -1.80227828e+00
-3.48041803e-01 -9.66589153e-01 -3.35220665e-01 3.60109746e-01
3.75928044e-01 -7.01483130e-01 1.16119289e+00 4.45376992e-01
4.77096111e-01 -3.82347584e-01 -1.32994258e+00 -1.06045365e+00
4.37284380e-01 2.55867600e-01 5.08260965e-01 5.99145651e-01
6.56123281e-01 4.07375187e-01 -7.19552815e-01 -1.24422684e-02
8.77852798e-01 6.46992922e-01 8.52013707e-01 -8.87147605e-01
-8.51396263e-01 -4.29576665e-01 3.47195178e-01 -1.41178775e+00
-6.73851520e-02 -2.26647392e-01 -1.28342330e-01 -1.44005370e+00
4.38182384e-01 -5.19990742e-01 -2.50794619e-01 2.53505826e-01
3.48941356e-01 -4.70158786e-01 4.65780228e-01 3.53163570e-01
-1.30431414e+00 4.55371946e-01 1.16770518e+00 -2.36334488e-01
-4.02165890e-01 2.71861464e-01 -5.22867858e-01 5.77600718e-01
9.57125843e-01 -7.34085441e-01 -1.67118832e-01 -7.88331270e-01
3.80758852e-01 9.90684390e-01 -2.58387595e-01 -3.94571662e-01
5.98354220e-01 -9.15950298e-01 2.49814708e-02 -5.63752830e-01
1.82721391e-01 -9.19019163e-01 -6.28188923e-02 9.30900872e-01
-6.62871003e-01 2.04769939e-01 -7.02435002e-02 8.12965989e-01
-9.04218182e-02 -4.93423045e-01 7.23839760e-01 -4.78321165e-01
-1.23544641e-01 2.55719364e-01 -8.73320103e-01 -3.97773124e-02
1.30254900e+00 3.01876336e-01 -4.97953743e-01 -5.49377620e-01
-8.99397790e-01 6.46279931e-01 3.89522426e-02 -8.56094342e-03
2.22653702e-01 -9.97356951e-01 -8.01334858e-01 -1.73879653e-01
-3.30250502e-01 -4.69947398e-01 1.61552384e-01 4.78946626e-01
-2.41732731e-01 3.75955462e-01 -4.26488608e-01 -7.07719475e-02
-1.31413305e+00 5.53134799e-01 5.88466585e-01 -8.14160764e-01
-1.53526664e-01 3.96023870e-01 -1.56702265e-01 1.15864165e-02
4.02832747e-01 2.81065404e-01 1.24839149e-01 5.13202772e-02
7.05299795e-01 7.21340716e-01 -2.12289661e-01 5.54214299e-01
-4.49007526e-02 -5.32037243e-02 -5.85216463e-01 -4.23259944e-01
1.45535934e+00 -4.12038863e-01 -2.62808353e-01 3.59619647e-01
6.20635867e-01 -2.09179223e-01 -1.40469241e+00 -5.93831003e-01
1.70040414e-01 -4.76314187e-01 3.73021960e-01 -6.85383022e-01
-1.09032750e+00 -7.58444890e-02 6.13115311e-01 7.26398945e-01
5.56014121e-01 3.80902775e-02 3.54873806e-01 8.76103818e-01
7.35819221e-01 -1.73800743e+00 8.27326104e-02 5.26197493e-01
6.36869967e-01 -1.14129817e+00 4.29610908e-01 4.62568663e-02
-5.65549016e-01 1.07380092e+00 2.42802575e-01 -4.96594042e-01
6.37920916e-01 4.76750374e-01 -2.60045640e-02 -1.07036680e-01
-1.21746325e+00 -1.83405310e-01 -6.22141302e-01 3.26062441e-01
7.32842907e-02 5.91281652e-01 -7.02856839e-01 2.20026955e-01
-1.47278458e-01 1.79829448e-01 6.26296103e-01 1.21507812e+00
-8.75089288e-01 -1.23064327e+00 -4.96416777e-01 4.62864876e-01
-6.32851541e-01 3.00902665e-01 -2.48883769e-01 5.28595746e-01
-5.62082231e-01 1.28518367e+00 3.38183343e-01 4.06580061e-01
1.21365435e-01 -4.73470777e-01 1.00454450e+00 -2.57381201e-01
-5.94686270e-01 3.04690540e-01 2.72947758e-01 -4.82443988e-01
-7.19962180e-01 -6.07850313e-01 -9.02148664e-01 -8.43178153e-01
-3.80412281e-01 8.90168071e-01 7.26948559e-01 1.10012352e+00
8.07771683e-02 2.92874515e-01 1.04224515e+00 -3.84833127e-01
-1.37996638e+00 -6.83231473e-01 -1.09546757e+00 -8.29292163e-02
6.53688848e-01 -2.52798110e-01 -1.68793231e-01 -7.14562774e-01] | [4.324129581451416, 3.0253093242645264] |
c9ccb27a-48a8-4f90-8796-240c14bee030 | shallow-optical-flow-three-stream-cnn-for | 2106.06489 | null | https://arxiv.org/abs/2106.06489v1 | https://arxiv.org/pdf/2106.06489v1.pdf | Shallow Optical Flow Three-Stream CNN for Macro- and Micro-Expression Spotting from Long Videos | Facial expressions vary from the visible to the subtle. In recent years, the analysis of micro-expressions $-$ a natural occurrence resulting from the suppression of one's true emotions, has drawn the attention of researchers with a broad range of potential applications. However, spotting microexpressions in long videos becomes increasingly challenging when intertwined with normal or macro-expressions. In this paper, we propose a shallow optical flow three-stream CNN (SOFTNet) model to predict a score that captures the likelihood of a frame being in an expression interval. By fashioning the spotting task as a regression problem, we introduce pseudo-labeling to facilitate the learning process. We demonstrate the efficacy and efficiency of the proposed approach on the recent MEGC 2020 benchmark, where state-of-the-art performance is achieved on CAS(ME)$^{2}$ with equally promising results on SAMM Long Videos. | ['Lai-Kuan Wong', 'John See', 'Gen-Bing Liong'] | 2021-06-11 | null | null | null | null | ['micro-expression-spotting'] | ['computer-vision'] | [ 2.68783927e-01 -1.98712558e-01 -2.26812512e-01 -8.00819755e-01
-4.18318480e-01 -2.29105651e-01 4.17401493e-01 -3.70941639e-01
-5.54247499e-01 6.24156177e-01 1.49840280e-01 1.05446413e-01
2.05464080e-01 -3.09778512e-01 -5.38481534e-01 -7.01923251e-01
-4.31136012e-01 -3.90846223e-01 -3.24922413e-01 -3.19736660e-01
1.00843012e-01 3.44671577e-01 -1.67293608e+00 5.78405678e-01
4.08942103e-01 1.64342868e+00 -3.89101803e-01 4.70553875e-01
-3.10702957e-02 1.49969161e+00 -7.36560702e-01 -7.69255996e-01
1.94946557e-01 -4.83254015e-01 -6.75934911e-01 1.41528353e-01
7.12870657e-01 -2.25125298e-01 -3.26789498e-01 1.26244509e+00
4.09626067e-01 1.88504934e-01 1.18629739e-01 -1.48388731e+00
-2.71703064e-01 1.53236046e-01 -9.72810626e-01 2.11765438e-01
4.87127453e-01 2.18227729e-01 9.50067341e-01 -9.00350869e-01
6.53619111e-01 1.37603927e+00 5.19594133e-01 6.13574207e-01
-8.72108340e-01 -1.09409380e+00 1.36200219e-01 4.65880990e-01
-1.42444432e+00 -7.55489886e-01 8.68279278e-01 -3.91179949e-01
8.05265307e-01 1.32879078e-01 6.77618980e-01 1.33787918e+00
-6.53393045e-02 9.34515476e-01 1.15870702e+00 -1.59086049e-01
1.24808162e-01 -1.90310270e-01 -3.09723109e-01 8.48014653e-01
-6.01805091e-01 1.85202047e-01 -1.06421626e+00 2.61714369e-01
4.03589904e-01 -2.81829864e-01 -3.28265995e-01 1.62954986e-01
-7.32783258e-01 5.08743346e-01 3.70379210e-01 6.11327514e-02
-3.64159077e-01 4.70962316e-01 6.66191041e-01 3.34089130e-01
7.46086717e-01 1.81818753e-01 -2.18412369e-01 -8.86558592e-01
-1.12175333e+00 1.86613515e-01 4.12447602e-01 7.36294329e-01
5.19000947e-01 1.76619142e-01 -2.31720477e-01 9.15854871e-01
8.85158181e-02 2.43462533e-01 9.94908661e-02 -1.18918014e+00
9.81303826e-02 5.08296967e-01 2.07946733e-01 -1.26172137e+00
-3.64007175e-01 -3.97721738e-01 -9.10766304e-01 4.49537426e-01
4.11402702e-01 -5.25760293e-01 -6.31037116e-01 2.15944886e+00
1.51270419e-01 5.86212218e-01 -2.66773403e-02 1.03122735e+00
6.70370281e-01 7.48392701e-01 2.53914773e-01 -4.28508818e-01
1.10482085e+00 -9.36740160e-01 -8.67391765e-01 -1.86260730e-01
3.98252279e-01 -6.57363832e-01 8.39405417e-01 5.38269222e-01
-9.72392261e-01 -5.73362529e-01 -9.30227518e-01 -5.79402335e-02
-9.12410840e-02 1.01759098e-01 6.77194059e-01 3.40475321e-01
-1.17747092e+00 5.75628936e-01 -6.84543192e-01 -1.39274657e-01
7.36730635e-01 3.76222342e-01 -5.45052946e-01 6.20110258e-02
-1.22682440e+00 7.25862384e-01 -1.49036050e-01 5.23009360e-01
-6.94647670e-01 -5.35695195e-01 -9.02329743e-01 1.09135304e-02
3.50304633e-01 -5.71250357e-02 1.12154341e+00 -1.70269763e+00
-1.68267012e+00 1.05707860e+00 -3.93650025e-01 -4.17362481e-01
6.59603536e-01 -4.73531872e-01 -6.65663302e-01 3.12801868e-01
-1.94630157e-02 9.48513091e-01 9.87056851e-01 -7.45437860e-01
-7.20335305e-01 -1.98202044e-01 2.48752654e-01 1.17695361e-01
-2.16633067e-01 5.37533343e-01 -4.93476987e-01 -5.77382922e-01
-5.18306017e-01 -9.19953227e-01 -1.10921068e-02 2.99257457e-01
-3.54016393e-01 -3.44994694e-01 1.02947736e+00 -4.43810284e-01
1.28047633e+00 -2.53822112e+00 1.09708510e-01 5.36768138e-02
1.90297648e-01 3.06586534e-01 -1.11024193e-01 -1.10700004e-01
-3.04957867e-01 1.38363456e-02 -3.87109332e-02 -4.87432420e-01
-4.41242196e-02 -1.86902732e-02 -9.58881080e-02 6.62503302e-01
5.07729948e-01 7.39121199e-01 -9.32270408e-01 -4.84681308e-01
5.09635694e-02 4.53651637e-01 -3.96150082e-01 3.91699493e-01
-1.49977207e-01 5.02633214e-01 -2.00775638e-01 7.48421967e-01
5.96314847e-01 -2.69644648e-01 -6.69609159e-02 -2.93256491e-01
-9.51177254e-02 -4.71464843e-01 -7.26238728e-01 1.71131575e+00
-3.36855739e-01 1.14896154e+00 3.06120455e-01 -1.00263202e+00
8.47199559e-01 4.09885675e-01 7.39022493e-01 -8.24283957e-01
2.97527671e-01 8.64979029e-02 6.75126016e-02 -8.20473731e-01
4.28196788e-01 -4.13424760e-01 -1.15934782e-01 8.28662440e-02
4.82678749e-02 1.96541727e-01 4.43042517e-01 -5.53564914e-02
9.83340025e-01 3.72980297e-01 1.72355041e-01 -4.06943373e-02
5.80816388e-01 -4.54675674e-01 7.77997077e-01 3.43763173e-01
-6.53003037e-01 4.61380094e-01 9.27611589e-01 -5.00875473e-01
-5.13702214e-01 -7.49771953e-01 1.30357921e-01 1.18067944e+00
7.21690208e-02 -4.57508266e-01 -6.72035336e-01 -7.35656142e-01
-2.47987837e-01 3.88357878e-01 -8.25874865e-01 -1.40194938e-01
-5.62235713e-01 -5.17688930e-01 6.32215142e-01 3.54792714e-01
7.55890369e-01 -1.28411770e+00 -8.46873462e-01 -1.08223790e-02
-4.03074622e-01 -1.54022574e+00 -4.20684725e-01 -1.78181157e-01
-1.58144921e-01 -9.24816370e-01 -6.65178597e-01 -6.12257838e-01
4.22086000e-01 -1.51077986e-01 1.13620126e+00 9.73439515e-02
-2.65827656e-01 2.38585453e-02 -3.91859502e-01 -2.44718537e-01
-1.10218994e-01 -3.42592120e-01 -1.33168444e-01 6.23967171e-01
5.17166317e-01 -4.45028543e-01 -7.84162879e-01 2.36611292e-01
-7.30173111e-01 1.04835406e-01 3.72645825e-01 6.09102726e-01
4.35795188e-01 -2.67904013e-01 6.67511940e-01 -6.49523139e-01
3.46201062e-01 -3.71159345e-01 -3.75897914e-01 9.57636088e-02
-1.31905004e-01 -2.34761536e-01 5.36738932e-01 -3.20094317e-01
-1.14668930e+00 1.60342887e-01 -3.13823938e-01 -7.40917861e-01
-2.21075982e-01 3.94373596e-01 -7.50721395e-02 -9.81111750e-02
3.22777689e-01 -1.55620612e-02 1.91962630e-01 5.01344651e-02
3.14414054e-01 4.81109828e-01 6.74324512e-01 -4.08585459e-01
3.39394242e-01 7.22103179e-01 2.31930941e-01 -9.62456703e-01
-1.19562173e+00 -4.99187112e-01 -2.68488824e-01 -8.19308996e-01
1.12173820e+00 -1.00492334e+00 -8.40298831e-01 6.81076825e-01
-1.12567747e+00 -2.88722306e-01 -5.84588163e-02 2.32169226e-01
-5.90898752e-01 1.92389339e-01 -5.26406884e-01 -7.03562379e-01
-1.83216587e-01 -1.30355597e+00 1.18409622e+00 3.85990113e-01
-5.37302136e-01 -6.83398902e-01 -3.39778289e-02 2.53116518e-01
3.55389804e-01 6.55631185e-01 4.88444448e-01 -4.46840189e-02
-2.88814515e-01 -2.63477802e-01 -3.93555939e-01 5.84413767e-01
5.83345965e-02 1.85288697e-01 -1.23324001e+00 -1.31284133e-01
-1.99183375e-01 -6.73710525e-01 7.74018884e-01 2.80008942e-01
1.52495468e+00 -1.11670315e-01 7.99550116e-02 9.30406868e-01
1.12356424e+00 2.00350061e-01 8.73040676e-01 1.26976177e-01
4.76446390e-01 7.37025619e-01 7.30695784e-01 7.70050228e-01
4.95733470e-02 7.93665051e-01 6.25131786e-01 -4.20298576e-01
5.45924790e-02 7.78928027e-02 6.10619485e-01 2.58807272e-01
-1.79501370e-01 -2.55091250e-01 -5.18796563e-01 2.74706304e-01
-1.80393600e+00 -1.26331055e+00 2.05912367e-01 1.71685469e+00
8.27171683e-01 1.24842711e-01 -7.54339248e-02 -7.42828473e-02
6.22082055e-01 6.75112486e-01 -4.30161417e-01 -4.85554010e-01
-4.13541853e-01 4.31183845e-01 -4.74884622e-02 3.23203504e-01
-1.25949574e+00 9.98896480e-01 5.89630032e+00 8.75995517e-01
-1.62129354e+00 -1.00893937e-01 1.32128501e+00 -4.78026450e-01
2.91683912e-01 -3.73706907e-01 -5.45015037e-01 4.47204947e-01
9.07141685e-01 -1.66593362e-02 1.63290739e-01 6.61034584e-01
6.03243113e-01 -2.16418460e-01 -1.07057667e+00 1.43508589e+00
2.59589463e-01 -1.22610164e+00 -2.75591224e-01 -3.27804774e-01
6.65997028e-01 -1.11423634e-01 2.61221081e-01 3.92017841e-01
-1.00577064e-01 -1.22446370e+00 9.52524602e-01 4.97407705e-01
1.15063882e+00 -7.80639946e-01 6.74844861e-01 -6.69759363e-02
-1.13611603e+00 -5.11024483e-02 2.07240567e-01 -3.75568569e-01
4.51963544e-01 3.24299127e-01 -2.68551350e-01 1.14885278e-01
8.60194147e-01 1.08503449e+00 -4.15718615e-01 7.05337107e-01
-3.55493546e-01 5.23690522e-01 -1.61244750e-01 -6.23212382e-03
2.87319273e-01 -1.77146137e-01 2.88955867e-01 1.44883013e+00
3.16447258e-01 1.75817311e-01 2.49524862e-02 6.59572899e-01
-2.96759665e-01 2.35523544e-02 -3.86895001e-01 -6.89044967e-02
-7.40328580e-02 1.34503341e+00 -5.09810150e-01 -3.20587069e-01
-5.82031250e-01 1.05977869e+00 4.07854617e-01 4.80746716e-01
-1.14255714e+00 -2.76928931e-01 9.90826309e-01 -1.06207035e-01
1.25495896e-01 9.11785290e-02 1.40009627e-01 -1.21484852e+00
1.85930908e-01 -8.49400640e-01 3.36734086e-01 -9.99504864e-01
-9.90123212e-01 7.62226760e-01 -2.02119350e-01 -1.32457519e+00
-4.05551612e-01 -6.82151794e-01 -5.93543947e-01 5.18819571e-01
-1.56535101e+00 -6.78937793e-01 -6.70662284e-01 5.82885325e-01
6.06988072e-01 1.67432174e-01 7.05380857e-01 5.37153184e-01
-6.89958632e-01 6.89853251e-01 -3.66492361e-01 4.81815845e-01
8.49762261e-01 -9.54842389e-01 -1.04935490e-01 9.92938280e-01
1.83714684e-02 1.66781038e-01 8.04235756e-01 -1.72895506e-01
-9.91267264e-01 -1.13032544e+00 7.28375256e-01 7.74533004e-02
6.79668725e-01 -3.40457290e-01 -5.69414258e-01 4.72881049e-01
3.48286390e-01 4.95410293e-01 7.08141625e-01 -3.91536281e-02
-2.78101206e-01 -3.38987142e-01 -9.84234095e-01 5.51530182e-01
9.28568304e-01 -6.80801451e-01 9.70148891e-02 3.80447321e-02
3.33211839e-01 -3.98837119e-01 -6.50722980e-01 5.42402565e-01
8.80869508e-01 -1.33844841e+00 7.23000109e-01 -6.78785205e-01
8.25446784e-01 -1.48998555e-02 -1.83735877e-01 -1.17691052e+00
7.11415559e-02 -9.42627430e-01 -8.84490162e-02 9.50524390e-01
1.96640611e-01 1.23866484e-01 9.74523187e-01 4.67901289e-01
2.64657084e-02 -1.02942204e+00 -1.17145383e+00 -3.46754342e-01
-3.63837570e-01 -6.67422295e-01 3.44877928e-01 9.28125083e-01
2.16321737e-01 2.36642525e-01 -7.66805470e-01 -1.27694339e-01
3.26317787e-01 3.21943499e-02 7.25174844e-01 -6.89773142e-01
-2.69563645e-01 -6.42008960e-01 -7.00977027e-01 -1.22659934e+00
5.78652680e-01 -4.83782768e-01 2.40293071e-01 -7.87181616e-01
1.19153947e-01 -1.17848010e-03 -5.57776213e-01 3.57342243e-01
-1.44712359e-01 5.18044293e-01 3.23939741e-01 -1.91314161e-01
-1.09133863e+00 6.08004570e-01 1.12178290e+00 -1.27757877e-01
2.29617640e-01 -1.69225186e-01 -5.05358219e-01 9.47206736e-01
5.05524039e-01 -6.83415383e-02 -2.79038012e-01 -4.01689321e-01
3.60452533e-01 2.97275782e-01 3.49298000e-01 -8.93033445e-01
-1.34899318e-01 -2.01782823e-01 1.79259419e-01 -2.94564426e-01
6.76037908e-01 -5.84854960e-01 -1.32686764e-01 1.27748445e-01
-5.31889558e-01 2.08398163e-01 1.44487172e-01 4.34663892e-01
-7.13170171e-01 1.75087541e-01 1.02665365e+00 4.78140078e-02
-1.21928287e+00 5.78885555e-01 -1.70894682e-01 2.15304181e-01
1.02165008e+00 -1.10639609e-01 -1.72624886e-01 -6.85733736e-01
-7.88435042e-01 1.99606836e-01 1.43192008e-01 5.50610483e-01
7.52507031e-01 -1.21235847e+00 -5.83899140e-01 8.94360766e-02
1.52740300e-01 -1.95569307e-01 4.21312630e-01 1.01040423e+00
-6.12688899e-01 -8.67073759e-02 -3.03910106e-01 -6.69351637e-01
-1.40458226e+00 8.71927589e-02 5.96422255e-01 -6.47600889e-02
-3.02307993e-01 1.20761561e+00 2.92346716e-01 3.24253500e-01
3.94336730e-01 -2.15683445e-01 -3.57264459e-01 1.21032521e-01
7.80808568e-01 2.04384550e-01 -2.55721718e-01 -1.10940850e+00
-3.59819800e-01 4.47817892e-01 5.77503853e-02 9.18770880e-02
1.27068150e+00 -2.03278974e-01 -1.77340969e-01 4.07850087e-01
1.63177681e+00 -2.26734608e-01 -1.60012805e+00 -1.33050233e-01
-5.22931926e-02 -6.31809413e-01 2.82855090e-02 -7.61803746e-01
-1.46048117e+00 9.54019308e-01 7.57280707e-01 -1.97187781e-01
1.33064413e+00 -8.56539086e-02 7.60986090e-01 1.40005410e-01
1.18516281e-01 -1.09012115e+00 3.59927893e-01 4.22801614e-01
8.20187807e-01 -1.47043121e+00 -3.31321687e-01 -3.09967250e-01
-5.63097537e-01 1.19715965e+00 7.32673049e-01 -7.92350844e-02
6.88812077e-01 3.25076193e-01 3.50984305e-01 -3.92917484e-01
-7.48245895e-01 4.50499989e-02 7.00676143e-02 1.91294581e-01
6.82655632e-01 -2.87164878e-02 5.27345017e-02 3.45115364e-01
-2.37110093e-01 3.51579219e-01 5.07469594e-01 8.03314388e-01
-9.96638909e-02 -5.26760459e-01 8.32086131e-02 2.64591396e-01
-9.71894801e-01 2.81497147e-02 -1.41838253e-01 6.73884273e-01
3.73164117e-01 1.02990544e+00 2.67485797e-01 -3.94956708e-01
2.26943642e-01 2.07847301e-02 3.58947903e-01 -2.45165691e-01
-2.69757271e-01 1.09630093e-01 3.13911796e-01 -1.09015346e+00
-9.03481901e-01 -6.56129897e-01 -1.14936733e+00 -2.89945126e-01
2.21793950e-02 7.39076582e-04 2.30645701e-01 1.08111000e+00
3.35740924e-01 4.58068997e-01 6.27304137e-01 -7.38137305e-01
-2.79545367e-01 -8.28115463e-01 -5.85885882e-01 9.73244429e-01
5.52353084e-01 -7.96427786e-01 -2.38013834e-01 2.32901633e-01] | [13.620864868164062, 1.8100894689559937] |
465a30eb-6291-40f0-9b9e-fb6a0250248c | comparison-of-evolving-granular-classifiers | 2005.04156 | null | https://arxiv.org/abs/2005.04156v1 | https://arxiv.org/pdf/2005.04156v1.pdf | Comparison of Evolving Granular Classifiers applied to Anomaly Detection for Predictive Maintenance in Computing Centers | Log-based predictive maintenance of computing centers is a main concern regarding the worldwide computing grid that supports the CERN (European Organization for Nuclear Research) physics experiments. A log, as event-oriented adhoc information, is quite often given as unstructured big data. Log data processing is a time-consuming computational task. The goal is to grab essential information from a continuously changeable grid environment to construct a classification model. Evolving granular classifiers are suited to learn from time-varying log streams and, therefore, perform online classification of the severity of anomalies. We formulated a 4-class online anomaly classification problem, and employed time windows between landmarks and two granular computing methods, namely, Fuzzy-set-Based evolving Modeling (FBeM) and evolving Granular Neural Network (eGNN), to model and monitor logging activity rate. The results of classification are of utmost importance for predictive maintenance because priority can be given to specific time intervals in which the classifier indicates the existence of high or medium severity anomalies. | ['Fabio Viola', 'Daniele Bonacorsi', 'Leticia Decker', 'Daniel Leite'] | 2020-04-08 | null | null | null | null | ['anomaly-classification'] | ['computer-vision'] | [ 6.04240596e-02 -1.98063701e-01 7.73746073e-02 -5.49206853e-01
-2.96633214e-01 -1.24451354e-01 5.36551714e-01 1.20661151e+00
-2.30836600e-01 8.17924798e-01 -3.43395889e-01 -3.09955716e-01
-6.28830612e-01 -1.18945837e+00 -1.03552721e-01 -6.85815573e-01
-7.61560977e-01 9.99954820e-01 3.75068754e-01 -2.38380674e-02
4.13450927e-01 1.33258116e+00 -2.05723214e+00 4.90983278e-01
9.05615687e-01 1.82193971e+00 -1.46154791e-01 7.07171679e-01
-1.84242055e-01 9.52713907e-01 -9.31252420e-01 5.87839425e-01
-1.04782142e-01 -3.24875474e-01 -5.12390792e-01 2.80953497e-02
-3.38936925e-01 3.85245293e-01 2.71497339e-01 9.70503211e-01
2.56565381e-02 4.66927350e-01 6.78172290e-01 -1.52804112e+00
9.60556492e-02 3.09478313e-01 6.98929504e-02 9.88829374e-01
2.56355792e-01 -3.93030256e-01 5.14050066e-01 -4.83179003e-01
4.09270793e-01 4.50500339e-01 4.50385123e-01 -1.64358869e-01
-6.84561610e-01 -3.06742196e-03 2.41177827e-01 9.97722447e-01
-1.23758030e+00 -2.69152839e-02 7.86401212e-01 -5.26732624e-01
9.76424575e-01 5.61485767e-01 9.41547871e-01 3.44938368e-01
7.10370302e-01 2.39552688e-02 9.94751811e-01 -5.58673799e-01
1.00319040e+00 -5.36256209e-02 -4.49150568e-03 4.01065707e-01
3.71541679e-01 3.89936492e-02 -5.41565955e-01 -3.64688963e-01
7.14227110e-02 2.14941099e-01 2.32498571e-02 7.53953531e-02
-5.97499251e-01 5.83038628e-01 1.36119395e-01 9.82561409e-01
-9.82679904e-01 -2.43641734e-01 5.86139441e-01 4.94985431e-01
8.28479767e-01 2.13588059e-01 -2.26734236e-01 -3.50933433e-01
-1.05582082e+00 1.39815405e-01 9.68608260e-01 4.67796147e-01
5.17483234e-01 3.71443510e-01 2.02779546e-02 2.30563238e-01
2.50517018e-02 1.56178907e-01 7.14784622e-01 -4.56672519e-01
-1.03995726e-01 8.48898530e-01 9.56409425e-02 -9.71998155e-01
-5.95010757e-01 -3.60589594e-01 -7.80790091e-01 4.20200378e-01
2.23618567e-01 3.68680447e-01 -3.95047635e-01 8.73463154e-01
5.62155783e-01 2.28574023e-01 -2.81479597e-01 3.89161497e-01
-2.41091073e-01 7.30014086e-01 4.28172052e-02 -8.18744063e-01
1.32865763e+00 -1.20821707e-01 -9.47166502e-01 4.43280309e-01
3.63589168e-01 -2.76680291e-01 4.12993610e-01 7.20608592e-01
-8.91136765e-01 -3.03395331e-01 -1.21285725e+00 8.29095066e-01
-1.06503010e+00 -5.02210319e-01 5.63765287e-01 5.18002629e-01
-7.61696637e-01 1.00114262e+00 -1.21191645e+00 -2.61636764e-01
8.91284570e-02 1.49140000e-01 3.37928496e-02 1.08411215e-01
-1.11299360e+00 9.95140851e-01 8.47022831e-01 3.25604051e-01
-7.13650584e-01 -3.84089828e-01 -6.23194754e-01 3.02121013e-01
1.93511128e-01 1.70014426e-01 8.90162051e-01 -4.57393348e-01
-8.92119646e-01 6.79257810e-01 7.94705227e-02 -9.03411090e-01
4.08155411e-01 3.96615684e-01 -1.16725481e+00 3.38088989e-01
-2.55083948e-01 -7.04661310e-01 8.80050957e-01 -9.78655040e-01
-1.12076962e+00 -6.39503837e-01 -5.12721837e-01 -2.02739030e-01
-4.45082903e-01 9.90417451e-02 5.12368917e-01 -3.45819890e-01
1.56175330e-01 -4.18361813e-01 -2.60380283e-02 -5.93316793e-01
1.67336851e-01 -5.65222263e-01 1.19772732e+00 -7.51466990e-01
1.72322893e+00 -1.73468232e+00 -2.15164244e-01 7.06902027e-01
5.01507856e-02 -7.52758607e-02 8.82601619e-01 4.55228925e-01
-1.99807659e-01 -3.06301653e-01 -4.25752550e-01 2.14057922e-01
-1.13112964e-01 3.56097132e-01 -2.48692900e-01 4.80605036e-01
5.82792461e-02 3.91094051e-02 -7.34915614e-01 -6.01346672e-01
2.45360568e-01 -2.50843000e-02 4.39853175e-03 3.61365288e-01
-6.31204665e-01 1.82580456e-01 -4.91548896e-01 1.17690837e+00
2.08951458e-01 -1.21776201e-01 1.01080276e-02 8.07147250e-02
-5.71718931e-01 -4.01601344e-01 -1.11512911e+00 1.09859705e+00
-3.73553306e-01 3.49593520e-01 -1.36753976e-01 -1.49034595e+00
1.17354441e+00 4.84734416e-01 1.09414184e+00 -9.25278544e-01
1.17158473e-01 3.59322727e-01 -4.47351456e-01 -5.66620767e-01
6.56430662e-01 -1.27620697e-01 -1.27707630e-01 5.89782000e-01
-2.03418612e-01 1.06812060e-01 4.65158075e-01 -2.02804103e-01
1.54897916e+00 -1.82347223e-01 4.07497495e-01 -2.47123301e-01
7.13407278e-01 1.42583713e-01 6.95072949e-01 1.67561278e-01
-2.69972861e-01 -3.01344246e-02 6.53283238e-01 -9.85375643e-01
-7.56497324e-01 -7.70290673e-01 -2.59983838e-01 1.00701952e+00
-1.63168684e-01 -4.64497581e-02 -4.42887723e-01 -4.64596063e-01
4.93615009e-02 1.04713452e+00 -5.46985388e-01 -3.66034120e-01
-5.44296086e-01 -1.10869694e+00 -6.63273931e-02 2.09896281e-01
1.89052103e-03 -1.54180455e+00 -1.17746961e+00 7.90140390e-01
6.11949787e-02 -7.01200724e-01 1.33915961e-01 7.24874556e-01
-1.06852520e+00 -1.20754409e+00 2.04415023e-01 -2.37830624e-01
6.36702001e-01 -4.99523014e-01 1.14574659e+00 2.77927756e-01
-7.47472942e-01 4.64615434e-01 -5.18560886e-01 -7.03039467e-01
-5.55295944e-01 -2.63074160e-01 3.39155048e-01 3.31375957e-01
4.45467800e-01 -7.22172678e-01 -2.61455059e-01 2.26968437e-01
-1.16979587e+00 -7.89618373e-01 7.56514519e-02 6.27481937e-01
9.75469530e-01 9.56160188e-01 9.32380140e-01 -9.41795170e-01
7.25901008e-01 -6.60266161e-01 -1.02006364e+00 5.04510045e-01
-1.22156477e+00 -1.09776534e-01 9.03576970e-01 -1.22459814e-01
-9.19456482e-01 -2.57746994e-01 1.29020840e-01 -3.57500166e-01
-2.70870984e-01 5.95767081e-01 -2.08958909e-01 2.80852038e-02
5.76307178e-01 1.54610485e-01 -4.50972825e-01 -3.70369077e-01
-3.51846248e-01 7.97415614e-01 7.71802604e-01 -8.91932845e-01
4.01977509e-01 5.06812692e-01 2.49175072e-01 -6.58876598e-01
-5.92326105e-01 -6.78928196e-01 -6.83671713e-01 -8.55819881e-01
7.90577829e-01 -5.04488647e-02 -7.56793439e-01 3.25718850e-01
-8.78847957e-01 1.43805176e-01 -7.95145392e-01 2.07622960e-01
-7.22809851e-01 -2.65164804e-02 -4.04682159e-01 -1.23749709e+00
-2.40045965e-01 -3.23843598e-01 4.03761715e-01 2.91481704e-01
-1.16473790e-02 -1.19660962e+00 1.79969102e-01 1.16293551e-02
5.95523596e-01 9.42531645e-01 1.31569016e+00 -1.21912301e+00
-3.71680021e-01 -9.44752514e-01 2.68938452e-01 1.84081689e-01
1.32681414e-01 2.59978399e-02 -8.91122103e-01 -3.00835222e-01
5.53481817e-01 1.84536934e-01 1.86985075e-01 6.11215830e-02
1.65326953e+00 -1.91577584e-01 -2.79539764e-01 2.49400437e-01
1.49312437e+00 9.29978549e-01 4.81049359e-01 5.04246950e-01
2.70950366e-02 4.62066948e-01 1.07890153e+00 1.14919257e+00
2.59262379e-02 5.48912525e-01 7.21567988e-01 4.10237819e-01
5.89287817e-01 1.53227851e-01 8.10373202e-02 8.50986123e-01
-3.83010417e-01 -5.35389073e-02 -8.70527327e-01 5.21931887e-01
-1.69392574e+00 -1.29795551e+00 1.58707783e-01 2.47792935e+00
5.31662464e-01 7.38362432e-01 1.39674410e-01 8.46153080e-01
8.99540722e-01 -1.24816835e-01 -6.10091865e-01 -4.42842633e-01
1.17870659e-01 1.83255196e-01 5.83173856e-02 1.45705476e-01
-8.54332447e-01 8.11849236e-02 4.36733627e+00 6.52408004e-01
-9.91914392e-01 2.01704443e-01 5.82370341e-01 -1.40040755e-01
2.44606156e-02 -1.03400275e-01 -6.42987013e-01 9.53735054e-01
1.67754090e+00 -5.69000483e-01 3.72527122e-01 9.24091697e-01
3.21380019e-01 -3.75677615e-01 -1.08382630e+00 6.13212407e-01
1.11116983e-01 -1.24520910e+00 -3.71028900e-01 6.30319864e-02
4.93275791e-01 -2.89965630e-01 -4.61852461e-01 1.46280140e-01
-2.89583392e-02 -8.07940423e-01 5.80122888e-01 9.65156198e-01
6.15808010e-01 -9.64502096e-01 7.80112028e-01 2.89844155e-01
-1.40795839e+00 -3.75358194e-01 -2.21197248e-01 1.70952260e-01
4.56203103e-01 1.04445863e+00 -6.07217729e-01 7.49652386e-01
8.90313089e-01 3.19722474e-01 -4.44052786e-01 1.21131241e+00
2.96658754e-01 5.85898459e-01 -7.04413176e-01 1.08301230e-01
-3.78552824e-02 -1.14799477e-01 4.75251228e-01 7.83205748e-01
5.59316874e-01 -6.11112975e-02 2.58099109e-01 5.23610294e-01
3.74139488e-01 -3.49192927e-03 -4.88453299e-01 -1.48505062e-01
6.64343417e-01 1.27916479e+00 -1.49074829e+00 -2.48997018e-01
-1.59298889e-02 5.80910802e-01 5.09337708e-02 -1.23105735e-01
-5.42165995e-01 -5.69858253e-01 2.70641237e-01 4.46433514e-01
6.69211447e-02 -1.35655358e-01 -2.49220356e-01 -7.77048707e-01
5.68521991e-02 -3.92902076e-01 1.05188847e+00 -2.63924181e-01
-1.24635279e+00 8.68262410e-01 3.48235928e-02 -1.51372313e+00
-5.29278278e-01 -5.82256913e-01 -8.65288138e-01 6.09618902e-01
-1.46133411e+00 -6.92154706e-01 -5.31927466e-01 8.33404005e-01
3.19469690e-01 -2.51191616e-01 1.00017858e+00 1.86207026e-01
-3.38212937e-01 -5.17213196e-02 2.26741165e-01 -3.58488798e-01
2.34789118e-01 -1.60515773e+00 -3.17567855e-01 8.12018812e-01
-5.47266975e-02 -2.97455549e-01 8.69127810e-01 -8.25263739e-01
-1.09190476e+00 -9.75003660e-01 1.02824175e+00 -2.82624811e-01
1.04127359e+00 6.91504925e-02 -1.30251932e+00 3.03843200e-01
-2.72840321e-01 5.35776973e-01 6.87354982e-01 -1.04029663e-01
2.89473116e-01 -4.17526662e-01 -1.48254704e+00 -3.31930548e-01
4.23568159e-01 -4.44970369e-01 -4.76175189e-01 5.03031433e-01
2.88953662e-01 -1.16682403e-01 -1.40978336e+00 2.50680000e-01
7.44131207e-02 -1.07147992e+00 5.32144070e-01 -8.03034961e-01
-1.83129311e-01 -2.54022419e-01 -2.19549447e-01 -1.40716672e+00
-2.95020342e-02 -4.22959626e-01 -7.00838923e-01 8.69475842e-01
-1.20469816e-01 -8.43696296e-01 7.21916914e-01 4.52328414e-01
-2.97618508e-01 -7.16838360e-01 -1.40703297e+00 -9.65199888e-01
-5.90609312e-01 -4.05965805e-01 8.72454703e-01 9.76691425e-01
3.09323281e-01 -4.57020342e-01 3.95177789e-02 4.29481596e-01
8.31477761e-01 3.64957511e-01 -2.44717106e-01 -1.92427552e+00
-1.00599632e-01 -3.90622526e-01 -9.79380727e-01 5.52997708e-01
4.03963886e-02 -7.24884033e-01 -4.99948598e-02 -1.15098643e+00
-7.16679096e-01 -7.16307521e-01 -1.01100338e+00 2.95443535e-01
2.54044980e-01 -2.25411803e-01 -1.04249649e-01 2.50373691e-01
-7.64964044e-01 4.53262597e-01 4.31310952e-01 -3.63814179e-03
-5.85975535e-02 5.96678019e-01 9.45887268e-02 5.12770414e-01
7.68356800e-01 -5.69818020e-01 -2.20671654e-01 4.62829143e-01
5.07890701e-01 5.93635738e-01 5.02724349e-02 -1.63054478e+00
6.35076702e-01 -3.29047084e-01 5.49939811e-01 -8.38752270e-01
-2.61093620e-02 -1.41683912e+00 5.17725885e-01 7.21617758e-01
-6.33487999e-02 5.65089703e-01 -5.76286297e-03 7.36304164e-01
-6.42536342e-01 -5.31263947e-01 7.27632523e-01 -6.01783954e-02
-7.41373003e-01 5.49933016e-01 -5.71852446e-01 -6.08571619e-02
1.48839390e+00 -2.08763331e-01 -2.29911376e-02 6.64732477e-04
-1.19737864e+00 1.86873928e-01 1.66737974e-01 1.11726828e-01
4.87044573e-01 -1.30051875e+00 -4.86727983e-01 3.60143393e-01
3.35831523e-01 1.73747875e-02 3.78846735e-01 5.58974326e-01
-6.86781943e-01 1.34385198e-01 -3.75869155e-01 -4.50286865e-01
-9.69814897e-01 7.25378811e-01 3.65677685e-01 -5.93324661e-01
-7.26038516e-01 2.47922838e-01 -1.12492156e+00 3.00642669e-01
2.83013433e-01 -2.77469665e-01 -4.93318558e-01 7.17378914e-01
7.90263534e-01 9.14908886e-01 1.01651001e+00 -8.38500708e-02
-2.00034976e-01 -6.44090027e-02 -2.90781166e-02 2.65893161e-01
1.51459539e+00 2.70575553e-01 -5.85853279e-01 9.81975794e-01
5.21633387e-01 -4.38911706e-01 -1.03026342e+00 -1.36528924e-01
9.14399922e-01 -2.25473359e-01 6.76411092e-02 -6.42239094e-01
-6.90921605e-01 5.71409881e-01 7.62527823e-01 1.28393316e+00
1.49868929e+00 -1.46626323e-01 3.99399042e-01 3.00363064e-01
6.04818344e-01 -1.67357910e+00 -3.40925068e-01 2.68630385e-01
6.04203582e-01 -8.40783358e-01 6.69502243e-02 1.17473602e-02
-1.93379283e-01 1.29702663e+00 4.49681878e-01 1.72626957e-01
1.03584051e+00 6.24175966e-01 -3.67596507e-01 -4.54187423e-01
-9.96338606e-01 1.06508233e-01 -1.55449554e-01 6.11232281e-01
2.17856170e-04 3.13291997e-01 -2.34921783e-01 4.53686863e-01
-1.13374993e-01 2.82279789e-01 3.98578405e-01 1.33209074e+00
-8.67353201e-01 -7.83219576e-01 -7.40876913e-01 8.98264170e-01
-2.97840357e-01 5.48793197e-01 2.45185658e-01 3.82694066e-01
2.48273551e-01 7.48293400e-01 5.72407007e-01 -9.87134576e-02
3.75334471e-01 4.97178108e-01 1.98354587e-01 -3.23981851e-01
-5.28127372e-01 -5.12333095e-01 -2.40655631e-01 -5.51110625e-01
-8.26640725e-02 -7.89946139e-01 -1.43883276e+00 -2.37653047e-01
-7.56012797e-02 7.10306108e-01 1.13276160e+00 9.90242302e-01
-9.02743952e-04 8.30442488e-01 1.05899203e+00 -7.37636089e-01
-5.62721074e-01 -9.15496647e-01 -1.20157146e+00 3.27188730e-01
1.17884785e-01 -8.27387393e-01 -5.46814203e-01 1.44574478e-01] | [7.021764278411865, 2.594040632247925] |
06500746-bb21-4bd1-84fd-46a20185308a | adaptive-conformal-predictions-for-time | 2202.07282 | null | https://arxiv.org/abs/2202.07282v1 | https://arxiv.org/pdf/2202.07282v1.pdf | Adaptive Conformal Predictions for Time Series | Uncertainty quantification of predictive models is crucial in decision-making problems. Conformal prediction is a general and theoretically sound answer. However, it requires exchangeable data, excluding time series. While recent works tackled this issue, we argue that Adaptive Conformal Inference (ACI, Gibbs and Cand{\`e}s, 2021), developed for distribution-shift time series, is a good procedure for time series with general dependency. We theoretically analyse the impact of the learning rate on its efficiency in the exchangeable and auto-regressive case. We propose a parameter-free method, AgACI, that adaptively builds upon ACI based on online expert aggregation. We lead extensive fair simulations against competing methods that advocate for ACI's use in time series. We conduct a real case study: electricity price forecasting. The proposed aggregation algorithm provides efficient prediction intervals for day-ahead forecasting. All the code and data to reproduce the experiments is made available. | ['Julie Josse', 'Yannig Goude', 'Olivier Féron', 'Aymeric Dieuleveut', 'Margaux Zaffran'] | 2022-02-15 | null | null | null | null | ['prediction-intervals'] | ['miscellaneous'] | [ 7.91696683e-02 2.42322251e-01 1.21813446e-01 -5.91204822e-01
-8.45677614e-01 -6.22788250e-01 8.75610411e-01 -1.39518231e-02
-1.42781681e-03 1.27841675e+00 -8.05174559e-02 -7.90337563e-01
-8.68716717e-01 -1.05858290e+00 -3.19745690e-01 -7.81093121e-01
-4.58440155e-01 8.76462698e-01 -7.55933151e-02 1.71216771e-01
3.13061506e-01 2.74017990e-01 -1.42378223e+00 -1.85856164e-01
1.44488204e+00 1.36712420e+00 5.87466545e-02 4.44926947e-01
1.20019354e-01 5.39200246e-01 -3.90225321e-01 -7.09162891e-01
4.05889660e-01 -3.39990348e-01 -3.66333574e-01 -6.05567515e-01
-3.48384351e-01 -2.53570944e-01 3.51494789e-01 9.19441164e-01
4.46962446e-01 1.89937189e-01 1.12693059e+00 -1.53961372e+00
-3.51273656e-01 1.03838527e+00 -6.80769458e-02 2.77672321e-01
-2.07750481e-02 -2.45743111e-01 8.85393620e-01 -6.37924373e-01
-5.11823148e-02 7.89737642e-01 9.87727821e-01 2.49571949e-02
-1.04036999e+00 -7.33819723e-01 -2.78832972e-01 3.75953823e-01
-1.39251626e+00 -8.96030888e-02 7.61921346e-01 -4.40294176e-01
7.86157548e-01 6.16593301e-01 7.35349238e-01 1.02433205e+00
6.44498408e-01 4.56186146e-01 1.48035955e+00 -5.98817170e-01
8.49837840e-01 1.21911138e-01 1.82223748e-02 -2.03398824e-01
2.65399694e-01 5.75419366e-01 -1.42339617e-01 -6.98651969e-01
4.09434438e-01 -6.82296753e-02 -2.31006742e-02 8.27996805e-02
-7.20086098e-01 1.04921854e+00 -2.44315788e-01 2.87142277e-01
-7.36135304e-01 1.15074612e-01 5.32004721e-02 5.61206996e-01
1.18749297e+00 6.34324774e-02 -7.94957697e-01 -3.44121486e-01
-1.26430237e+00 3.65799516e-01 1.08880281e+00 9.44792509e-01
1.24204673e-01 2.25275993e-01 -4.66673017e-01 4.35721785e-01
4.28843021e-01 8.19422781e-01 2.42025509e-01 -9.78737175e-01
2.03562930e-01 2.62465756e-02 4.93492514e-01 -6.66305602e-01
-5.11759698e-01 -5.17650545e-01 -9.46365356e-01 2.96999991e-01
3.78761321e-01 -6.38407588e-01 -4.54510242e-01 1.47000587e+00
1.01274014e-01 5.22795200e-01 -1.13787455e-02 3.24983716e-01
8.17136541e-02 8.20683002e-01 3.03208143e-01 -6.97129250e-01
9.59670305e-01 -3.34831148e-01 -7.64933288e-01 3.64056081e-01
2.72828847e-01 -7.02069104e-01 4.02185768e-01 5.02340615e-01
-1.06804144e+00 -2.73659587e-01 -7.17114151e-01 7.34832704e-01
-4.25193638e-01 -5.27404808e-02 5.82916319e-01 9.26285565e-01
-9.56437767e-01 1.15561402e+00 -7.26195574e-01 5.70417456e-02
2.93545902e-01 -5.82841337e-02 3.15431088e-01 5.61804056e-01
-1.67903805e+00 1.21410763e+00 4.40971404e-01 6.27233312e-02
-5.03939211e-01 -8.52216899e-01 -2.73617715e-01 3.43230724e-01
1.30568281e-01 -5.50702870e-01 1.24350357e+00 -7.51585960e-01
-1.65549624e+00 -1.56211751e-02 3.04714829e-01 -1.15704167e+00
8.94857705e-01 -1.22864045e-01 -7.36194372e-01 -2.18446493e-01
-1.27346680e-01 4.35049944e-02 1.08038199e+00 -8.32542419e-01
-7.89406478e-01 -2.54162878e-01 -3.50850374e-01 -1.53495714e-01
1.31297544e-01 -2.42561519e-01 5.84930122e-01 -8.91952455e-01
-1.38247728e-01 -1.05089092e+00 -2.28782654e-01 -6.90703332e-01
-6.13031201e-02 -6.60771906e-01 3.44944954e-01 -9.48870480e-01
1.51460493e+00 -1.86666679e+00 -3.06764275e-01 8.24270248e-01
-5.15114009e-01 -2.85412669e-01 4.41623837e-01 6.59012854e-01
-1.82947233e-01 1.02994941e-01 -6.88497841e-01 -1.49918184e-01
5.91820478e-01 2.07942978e-01 -7.58272171e-01 4.46634054e-01
-1.07848354e-01 7.14487076e-01 -6.74155712e-01 -3.32304686e-01
4.16814327e-01 4.34736103e-01 -1.86321080e-01 -1.01253390e-01
-3.86008888e-01 3.63347381e-01 -3.55622917e-01 4.27096277e-01
8.03360522e-01 -3.87713760e-01 9.77078080e-02 6.02313578e-02
-3.44833165e-01 -9.05102789e-02 -1.13945580e+00 1.10418832e+00
-3.90414596e-01 3.61754864e-01 -6.07015073e-01 -1.02843797e+00
9.36368763e-01 5.32396495e-01 4.93590504e-01 -4.23862845e-01
-6.23981981e-03 4.62741226e-01 2.46414244e-02 1.59389211e-03
2.46797845e-01 -2.60259688e-01 -8.41331258e-02 5.54938257e-01
-7.41455927e-02 -3.85294199e-01 6.67979196e-02 -2.50597775e-01
7.87110865e-01 5.50635993e-01 5.80075920e-01 -8.51870120e-01
3.38578135e-01 -3.31222326e-01 5.38843393e-01 9.10231888e-01
9.29795504e-02 3.01031947e-01 3.08943510e-01 -5.72106898e-01
-1.19925940e+00 -1.08416414e+00 -8.03505242e-01 7.60471344e-01
-3.45736176e-01 -1.26476735e-01 -5.17915189e-01 -2.90730029e-01
1.17003739e-01 1.92489922e+00 -7.31925786e-01 3.39344352e-01
7.10390210e-02 -1.13074207e+00 -4.61287014e-02 3.77165198e-01
3.16199094e-01 -9.51240718e-01 -8.78316700e-01 5.89905918e-01
5.57355918e-02 -3.96286100e-01 1.36565268e-01 2.50698090e-01
-8.33772600e-01 -6.93005264e-01 -7.78928757e-01 2.56898373e-01
3.01093787e-01 -5.56250453e-01 1.39557993e+00 -5.11699021e-01
1.75766960e-01 7.81848729e-01 -4.20895100e-01 -9.80440617e-01
-4.72120881e-01 -5.98578006e-02 -1.14798971e-01 -1.34509698e-01
4.67898220e-01 -9.95912075e-01 -7.16437340e-01 2.44400084e-01
-5.87206900e-01 -6.04240708e-02 3.63666862e-01 7.59810150e-01
5.90451539e-01 2.89820760e-01 1.02708292e+00 -9.92301941e-01
7.92228222e-01 -7.02994525e-01 -1.22715628e+00 7.17565596e-01
-1.39728153e+00 8.44867453e-02 3.68759066e-01 4.88226451e-02
-1.36963618e+00 -2.60846257e-01 7.82075748e-02 -1.82268009e-01
5.74938245e-02 8.19440782e-01 4.57403064e-01 1.90306306e-01
4.18927908e-01 1.68056384e-01 -3.16845089e-01 -3.30144763e-01
9.69062969e-02 4.46919829e-01 1.87566832e-01 -7.14619875e-01
6.95002377e-01 1.72516599e-01 2.58508205e-01 -2.81670839e-01
-6.63949966e-01 1.28759339e-01 -4.24000949e-01 -3.88050437e-01
6.37392819e-01 -7.72752285e-01 -6.39685333e-01 1.37868807e-01
-9.38998520e-01 -3.63306314e-01 -5.87504208e-01 7.73917973e-01
-9.25592959e-01 1.31552830e-01 -1.48452476e-01 -1.57752836e+00
-5.75510323e-01 -5.04392564e-01 6.37576044e-01 9.90605727e-02
-3.45705986e-01 -1.41845989e+00 4.86152411e-01 -2.64127165e-01
7.55586088e-01 5.40361464e-01 8.16582859e-01 -1.07240343e+00
-3.26099157e-01 -2.24083304e-01 1.08849443e-01 3.22558045e-01
-1.94609165e-01 1.85196117e-01 -1.12670565e+00 -1.76042989e-01
2.71162778e-01 2.23961696e-01 4.78344530e-01 7.62650788e-01
1.27328503e+00 -5.64350426e-01 -1.99823186e-01 1.71148524e-01
1.42368209e+00 4.88839239e-01 7.74561107e-01 2.51705647e-01
-1.48325175e-01 6.61738932e-01 6.95817173e-01 1.14468217e+00
3.26982230e-01 2.97154576e-01 2.80506670e-01 4.82713223e-01
7.36645699e-01 -5.55821471e-02 6.44637421e-02 1.04730856e+00
-7.36427784e-01 -2.56230742e-01 -9.97922242e-01 4.08802599e-01
-1.92055380e+00 -1.30617702e+00 -6.32660836e-02 2.48042750e+00
7.03069985e-01 3.04425448e-01 -1.35058565e-02 1.84363261e-01
5.34444869e-01 -3.07290316e-01 -4.49312538e-01 -3.06578875e-01
-1.44784302e-01 4.40222174e-01 6.84986532e-01 5.83242238e-01
-9.91332531e-01 3.99941159e-03 6.56923056e+00 1.19189203e+00
-5.10289550e-01 2.32499704e-01 9.46394145e-01 3.00288647e-01
-8.08504283e-01 2.19489932e-01 -4.59933877e-01 1.01922691e+00
1.63782215e+00 -8.89229834e-01 3.56941223e-01 9.01529312e-01
1.41388610e-01 -1.57032087e-01 -1.02713954e+00 6.84264839e-01
-4.54052776e-01 -1.21672118e+00 -2.61375695e-01 -1.15912177e-01
1.05229950e+00 -8.16905126e-02 1.55412242e-01 4.82804418e-01
5.37616134e-01 -9.60713446e-01 6.84484780e-01 1.33904397e+00
4.65991735e-01 -1.10173309e+00 1.25286806e+00 2.90231258e-01
-9.19080794e-01 -6.22444451e-02 -2.42866695e-01 -1.54954614e-02
2.46996760e-01 1.04714215e+00 -5.24261236e-01 1.08183324e+00
7.79958725e-01 5.11612475e-01 -3.37207884e-01 1.00386977e+00
3.41026448e-02 1.12186062e+00 -6.64645255e-01 -8.58913437e-02
4.23408635e-02 -7.61865914e-01 4.57913488e-01 8.52142751e-01
1.17389309e+00 3.12158585e-01 -4.06753868e-01 7.96924591e-01
6.62900031e-01 -1.62249189e-02 -4.94948298e-01 2.10230276e-01
8.55924904e-01 9.28009689e-01 -6.28223479e-01 -3.45510334e-01
-5.28748810e-01 4.42753106e-01 -3.87698323e-01 4.37298685e-01
-1.07227027e+00 -2.35625640e-01 -1.45995066e-01 -8.97963271e-02
4.11885291e-01 9.70359147e-02 -3.66402239e-01 -7.34795868e-01
-1.49841055e-01 -4.78516489e-01 7.38810539e-01 -8.88996780e-01
-1.95526373e+00 4.52517539e-01 5.72289944e-01 -1.56974316e+00
-8.39390457e-01 -4.05283839e-01 -7.86490619e-01 1.08888149e+00
-1.35523856e+00 -7.93717563e-01 1.42857358e-01 4.01920468e-01
2.88143605e-01 -3.56891543e-01 1.01423562e+00 -5.62745593e-02
-1.28415646e-02 2.66881794e-01 8.25304329e-01 -5.84634304e-01
3.33863854e-01 -1.40138817e+00 2.21315458e-01 4.23734933e-01
-1.32317618e-02 2.91539103e-01 1.08006942e+00 -7.54457951e-01
-8.07973623e-01 -9.42902029e-01 8.91268730e-01 -7.21901596e-01
8.77866089e-01 1.60269961e-01 -7.90966690e-01 7.29065716e-01
5.53634167e-01 -4.90825564e-01 9.75075722e-01 7.07777664e-02
2.83713043e-01 -2.50603765e-01 -1.28692591e+00 2.57130176e-01
4.86158043e-01 -2.00129375e-01 -8.17799389e-01 4.86800671e-01
3.54897648e-01 -1.71134640e-02 -1.52385867e+00 6.63586199e-01
5.72101653e-01 -1.06023920e+00 6.70175731e-01 -7.39078224e-02
9.33378935e-02 -9.14374292e-02 -2.35334456e-01 -1.38086069e+00
-4.01829630e-01 -9.61076736e-01 -2.42784217e-01 1.11517656e+00
6.82068348e-01 -1.28300011e+00 2.01756835e-01 1.06921744e+00
6.95075095e-02 -4.87786323e-01 -1.53590333e+00 -9.26243603e-01
4.38330412e-01 -6.94148839e-01 1.02137041e+00 1.01510036e+00
1.92982450e-04 -6.76075444e-02 -4.19868708e-01 7.27327168e-02
1.05819988e+00 4.25426245e-01 2.26853058e-01 -1.83064604e+00
-1.95115730e-01 -3.09654653e-01 -2.70456672e-02 -2.32524231e-01
1.60540029e-01 -5.82263470e-01 -2.93245077e-01 -1.10273182e+00
-1.85086027e-01 -7.04598606e-01 -4.98004943e-01 1.80995122e-01
5.14844619e-03 -3.08656663e-01 -2.43843589e-02 2.36019790e-01
-2.72443831e-01 7.19664514e-01 7.31336772e-01 1.53195843e-01
1.01531766e-01 5.31714916e-01 -9.62814912e-02 6.96871996e-01
1.14717972e+00 -3.82329077e-01 -7.49767721e-01 3.23378891e-01
6.66718423e-01 4.61073905e-01 4.07012194e-01 -9.99042690e-01
2.30629012e-01 -2.87567258e-01 4.28008407e-01 -1.18628013e+00
-5.51358908e-02 -1.19280767e+00 1.14382553e+00 4.16750878e-01
-2.93703109e-01 3.10970932e-01 3.07179421e-01 8.38135898e-01
-9.38698873e-02 -6.29727483e-01 2.79089451e-01 8.68330598e-02
-3.66357088e-01 3.04358929e-01 -2.61752635e-01 -1.16992444e-01
1.13239849e+00 1.75570205e-01 -1.32694736e-01 -7.57642746e-01
-9.69865561e-01 2.01710626e-01 -8.31041764e-03 -9.29444432e-02
6.52432963e-02 -1.42342758e+00 -1.21655166e+00 -8.93220231e-02
-2.84565240e-01 -4.81717467e-01 3.69177490e-01 8.43620002e-01
-2.26661086e-01 6.54555500e-01 -9.34638381e-02 -4.31092530e-01
-3.68796587e-01 4.98477221e-01 1.52199030e-01 -3.17250252e-01
-4.05443996e-01 3.39566737e-01 -2.20574260e-01 -2.10717484e-01
-8.42245445e-02 -3.61660808e-01 -2.50278205e-01 2.04472408e-01
2.87359983e-01 8.13875675e-01 5.29615348e-03 6.58633783e-02
-2.16979370e-01 2.42821112e-01 4.36082363e-01 -3.64763469e-01
1.52013254e+00 -3.85462761e-01 -7.23414347e-02 9.57226276e-01
3.00556451e-01 -2.31748030e-01 -1.38416862e+00 6.66086301e-02
2.95931429e-01 -2.30906725e-01 -3.89338136e-02 -1.01142788e+00
-4.48969007e-01 6.00827932e-01 8.72153461e-01 1.06890535e+00
1.11218274e+00 -5.10417044e-01 1.12567186e-01 1.66432798e-01
7.23461330e-01 -1.45521700e+00 -9.93218362e-01 1.39230683e-01
1.04594135e+00 -1.25752199e+00 3.80356878e-01 8.53674412e-02
-5.73489964e-01 1.18962836e+00 -1.27893984e-01 -9.67440382e-02
1.24177551e+00 2.38068983e-01 -4.07467693e-01 2.34208658e-01
-1.12461042e+00 1.69629216e-01 3.34388584e-01 4.57497627e-01
3.66037846e-01 5.78546286e-01 -6.89828932e-01 1.00876105e+00
-4.46718127e-01 2.34772041e-01 1.57833993e-01 5.80710709e-01
-1.66611552e-01 -8.16134453e-01 -3.15330654e-01 7.79690206e-01
-3.84571075e-01 -3.36073309e-01 2.83687532e-01 8.70235145e-01
-1.33907393e-01 8.13630104e-01 3.30086917e-01 1.51999280e-01
1.27293793e-02 5.07555127e-01 2.93770313e-01 2.37589166e-01
-2.56461352e-01 6.11259528e-02 1.34013340e-01 -1.44884229e-01
-6.43449128e-01 -1.13170278e+00 -5.93689680e-01 -6.14296377e-01
-3.41496557e-01 6.44257128e-01 5.55780470e-01 8.75066578e-01
3.89812320e-01 4.20880586e-01 7.66305029e-01 -6.59459293e-01
-1.08644855e+00 -1.20633233e+00 -8.01395118e-01 -8.10655281e-02
-1.18780173e-01 -8.26752663e-01 -8.83215725e-01 -4.45874780e-02] | [6.4068684577941895, 3.4103095531463623] |
896ddd6b-66ef-436c-a11c-4664bb543a01 | pens-a-dataset-and-generic-framework-for | null | null | https://aclanthology.org/2021.acl-long.7 | https://aclanthology.org/2021.acl-long.7.pdf | PENS: A Dataset and Generic Framework for Personalized News Headline Generation | In this paper, we formulate the personalized news headline generation problem whose goal is to output a user-specific title based on both a user{'}s reading interests and a candidate news body to be exposed to her. To build up a benchmark for this problem, we publicize a large-scale dataset named PENS (PErsonalized News headlineS). The training set is collected from user impressions logs of Microsoft News, and the test set is manually created by hundreds of native speakers to enable a fair testbed for evaluating models in an offline mode. We propose a generic framework as a preparatory solution to our problem. At its heart, user preference is learned by leveraging the user behavioral data, and three kinds of user preference injections are proposed to personalize a text generator and establish personalized headlines. We investigate our dataset by implementing several state-of-the-art user modeling methods in our framework to demonstrate a benchmark score for the proposed dataset. The dataset is available at https://msnews.github.io/pens.html. | ['Xing Xie', 'Qing He', 'Ying Qiao', 'Ling Luo', 'Xiting Wang', 'Xiang Ao'] | 2021-08-01 | null | null | null | acl-2021-5 | ['headline-generation'] | ['natural-language-processing'] | [ 2.22034827e-01 4.68989648e-02 -4.73125964e-01 -7.82923818e-01
-9.46119010e-01 -5.56362391e-01 8.39826584e-01 3.86223383e-03
-3.23360115e-01 6.19890034e-01 6.44278109e-01 -2.30734020e-01
2.73374557e-01 -6.77547038e-01 -6.08793080e-01 -1.06769189e-01
3.34057570e-01 7.42590249e-01 1.59565762e-01 -3.96534234e-01
3.19656253e-01 -8.11143368e-02 -1.56877792e+00 5.51310182e-01
1.04763877e+00 9.55709457e-01 5.54974020e-01 7.70204604e-01
6.89158365e-02 4.77946371e-01 -6.28689647e-01 -5.77841818e-01
-2.42707250e-03 -5.47775447e-01 -8.83840799e-01 1.89674005e-01
2.99197882e-01 -3.13080460e-01 -4.82411355e-01 8.07293236e-01
9.40069497e-01 3.93222451e-01 6.18960202e-01 -1.15813577e+00
-1.08852398e+00 1.25280857e+00 -1.98306158e-01 1.24342538e-01
5.92351735e-01 -3.22815515e-02 1.31030476e+00 -7.19440937e-01
5.83713949e-01 1.05579889e+00 6.85617253e-02 8.59208345e-01
-9.13192689e-01 -5.16476870e-01 4.49436158e-01 7.12280944e-02
-9.73634720e-01 -6.63768709e-01 6.48083687e-01 -2.82275289e-01
1.02274194e-01 8.66841793e-01 4.57888901e-01 1.80029166e+00
-2.83230364e-01 1.41795051e+00 8.27104867e-01 -3.30398768e-01
2.11891189e-01 7.12656379e-01 5.59357405e-01 3.17500323e-01
-2.15318561e-01 -2.15321019e-01 -7.22305954e-01 -2.83744663e-01
4.17234361e-01 1.99466452e-01 -5.16962171e-01 1.42684907e-01
-9.77262855e-01 7.69519985e-01 2.62180269e-01 -1.18270796e-02
-7.77052045e-02 -4.23658252e-01 5.92852198e-02 1.02120489e-01
6.41602576e-01 7.08535194e-01 -6.26184642e-01 -1.58865646e-01
-7.97049999e-01 6.27779424e-01 1.11024404e+00 1.42278111e+00
5.11293828e-01 -6.47902787e-01 -1.01184058e+00 1.23892605e+00
3.67223263e-01 6.12039030e-01 8.97042215e-01 -5.36915720e-01
5.71302652e-01 2.93629020e-01 4.10430849e-01 -7.73706615e-01
-7.92461410e-02 -7.65739501e-01 -5.42692423e-01 -6.83914721e-01
1.16001099e-01 -4.92651284e-01 -7.41577268e-01 1.71474957e+00
-1.04270387e-03 1.56575412e-01 -1.97921291e-01 7.10220993e-01
1.00924063e+00 9.96510327e-01 -1.41965196e-01 -2.21701831e-01
1.50774455e+00 -1.38009322e+00 -5.45564711e-01 -3.09175670e-01
3.18028957e-01 -8.09057295e-01 1.75484943e+00 1.63624987e-01
-9.15438533e-01 -4.05877680e-01 -7.42515683e-01 3.76118049e-02
-2.02058077e-01 4.94915515e-01 3.74981493e-01 6.35141969e-01
-7.36243844e-01 4.07558411e-01 -4.21334326e-01 -5.31598747e-01
1.32657140e-01 -1.01668417e-01 3.01088959e-01 7.59235397e-02
-1.46362603e+00 3.06994468e-01 -5.82087552e-03 -3.94141585e-01
-8.16364884e-01 -7.26453722e-01 -5.32834888e-01 2.33452201e-01
2.87639290e-01 -7.84551859e-01 2.01007605e+00 -6.26759350e-01
-1.74942946e+00 6.03470743e-01 -2.25469559e-01 -2.25399822e-01
8.00425768e-01 -3.85071576e-01 -5.74375868e-01 -4.49113250e-01
2.23762885e-01 2.48153716e-01 6.84823334e-01 -1.21669054e+00
-8.81958842e-01 -1.49563059e-01 2.08498508e-01 2.84578830e-01
-7.72447586e-01 1.51692927e-01 -1.11673045e+00 -8.68284166e-01
-4.90824133e-01 -1.07536972e+00 -2.70781666e-01 -7.92714596e-01
-1.06798422e+00 -2.68195182e-01 4.07582730e-01 -5.15783906e-01
1.67240858e+00 -1.87409401e+00 -2.11167723e-01 2.41529405e-01
8.72964412e-02 -4.84993607e-02 -9.09603164e-02 5.30383289e-01
3.41287524e-01 2.12785795e-01 2.74283230e-01 -7.74317443e-01
3.01710248e-01 -3.91507715e-01 -4.79915351e-01 -6.19533956e-02
-6.39064372e-01 6.03974104e-01 -9.16312516e-01 -3.40987802e-01
-4.22077447e-01 1.49582043e-01 -6.97554886e-01 7.53369451e-01
-5.22374988e-01 2.63679266e-01 -8.33213568e-01 4.12040263e-01
3.49188447e-01 -7.21096754e-01 -7.82753229e-02 -5.74932657e-02
-4.34936285e-02 7.19606757e-01 -9.67472017e-01 1.43295157e+00
-5.89218318e-01 3.94492537e-01 -4.04207021e-01 -2.12669164e-01
7.49253094e-01 1.82239309e-01 2.45399997e-01 -5.30324459e-01
2.72142768e-01 -1.64934427e-01 -4.26076502e-01 -4.73580182e-01
8.65042210e-01 5.18504322e-01 -2.41598755e-01 8.79229665e-01
2.57064980e-02 3.79578054e-01 4.58571792e-01 5.37317812e-01
1.03113782e+00 -2.76899219e-01 -9.16699506e-03 -1.61268458e-01
3.79404038e-01 -4.17995751e-01 2.87986755e-01 1.25343966e+00
1.94199041e-01 7.21697688e-01 3.16806495e-01 8.14002752e-02
-6.83725834e-01 -8.50395083e-01 2.88109984e-02 1.82721961e+00
1.29105777e-01 -8.91238987e-01 -9.11783159e-01 -8.45829785e-01
-3.75795007e-01 1.08503437e+00 -6.84126914e-01 3.65004204e-02
-8.68248940e-02 -9.19649899e-01 5.40364161e-02 2.33739123e-01
3.94154131e-01 -1.05910790e+00 -8.08470026e-02 1.93892270e-02
-4.38532889e-01 -8.53862524e-01 -1.24548781e+00 -3.41412157e-01
-2.04999104e-01 -6.31744921e-01 -1.07370389e+00 -7.07106411e-01
7.14839697e-01 2.96198696e-01 1.16444504e+00 -1.49849400e-01
1.48517057e-01 3.61089021e-01 -7.47327387e-01 -6.69127166e-01
-4.17865306e-01 7.90036380e-01 2.91493498e-02 3.74820143e-01
5.27766980e-02 -3.75706762e-01 -8.21346283e-01 6.23592257e-01
-7.69397795e-01 4.69797313e-01 6.11196756e-01 6.59690559e-01
3.43410522e-01 -1.48632720e-01 5.51660061e-01 -1.61972570e+00
1.15470898e+00 -7.72352815e-01 -3.54684085e-01 5.49808860e-01
-7.27558672e-01 -1.21906899e-01 6.81799829e-01 -5.26085913e-01
-1.23450041e+00 -1.75620615e-01 -2.10516647e-01 2.48502806e-01
-1.12474874e-01 7.14882910e-01 -4.41569209e-01 5.84968865e-01
8.89780343e-01 3.07666600e-01 -4.87134069e-01 -8.50107193e-01
4.49433506e-01 1.13908339e+00 3.99447560e-01 -6.40769064e-01
7.53145039e-01 -1.63056135e-01 -1.07285237e+00 -5.65746725e-01
-1.39997351e+00 -6.20350361e-01 -1.39804095e-01 -1.49676591e-01
4.14240181e-01 -7.50850677e-01 -5.47765076e-01 3.71899098e-01
-1.01757920e+00 -4.19828743e-01 -2.77210362e-02 2.49732718e-01
-4.42495167e-01 -7.58562759e-02 -6.20688021e-01 -6.67992771e-01
-5.02771020e-01 -1.10003674e+00 8.10341239e-01 5.60593724e-01
-3.61331075e-01 -9.38725889e-01 2.72247136e-01 5.19137740e-01
4.74219441e-01 -2.82328427e-01 4.93413538e-01 -1.23980153e+00
-3.30701590e-01 -5.74610114e-01 5.68889380e-02 1.64058477e-01
1.86135739e-01 1.48159591e-02 -1.07495618e+00 -1.70416698e-01
-3.05891961e-01 -2.51309067e-01 5.87952018e-01 6.43862709e-02
1.66122425e+00 -9.12117481e-01 -5.29426813e-01 4.65553820e-01
7.62072444e-01 6.73838034e-02 2.77829438e-01 3.33072364e-01
5.17988741e-01 3.92927200e-01 5.65399885e-01 9.06655967e-01
6.79579496e-01 7.81634271e-01 -1.09051242e-01 7.87108485e-03
2.24948943e-01 -8.18348825e-01 4.43387866e-01 7.52573311e-01
2.36511752e-02 -1.00360024e+00 -6.12634361e-01 2.00497434e-01
-1.85682952e+00 -1.07766926e+00 2.28920341e-01 2.10734344e+00
1.25664949e+00 1.84945524e-01 4.51810718e-01 -2.77828634e-01
9.49316263e-01 1.41125053e-01 -5.21885633e-01 1.75086126e-01
1.97058469e-01 -1.83501422e-01 4.13835198e-01 5.04288435e-01
-1.15061593e+00 8.69511485e-01 6.00660849e+00 8.49649608e-01
-1.10286379e+00 2.98132412e-02 1.00572979e+00 -1.69047698e-01
-6.22108400e-01 -2.70458162e-01 -1.39344382e+00 1.02983570e+00
1.05761278e+00 -8.83536756e-01 3.93801093e-01 1.05366409e+00
2.69582480e-01 5.04806578e-01 -1.18848407e+00 8.83947372e-01
2.42147624e-01 -1.41875291e+00 3.01469296e-01 -1.91382781e-01
7.64376462e-01 2.09567770e-02 4.74571824e-01 6.45409286e-01
5.98898232e-01 -6.97699547e-01 7.63272047e-01 4.52381670e-01
7.01139271e-01 -4.18346494e-01 4.36826050e-01 6.96419418e-01
-7.04423964e-01 3.95838683e-03 -1.30493477e-01 2.27787107e-01
4.24308598e-01 5.60310066e-01 -1.10386252e+00 2.25891963e-01
5.32697916e-01 6.18203640e-01 -7.46311188e-01 1.23469985e+00
-2.93361038e-01 8.82011771e-01 -1.57680288e-01 -4.36346173e-01
-8.61733221e-03 1.23366371e-01 3.11179101e-01 1.41034889e+00
4.37443495e-01 1.18896291e-01 2.88031071e-01 8.19682717e-01
-6.70624793e-01 5.03597438e-01 6.11760980e-03 -8.11733454e-02
7.89771736e-01 1.54088485e+00 -3.18322331e-01 -4.69851226e-01
-3.46944600e-01 1.11665750e+00 1.81193799e-01 4.60348666e-01
-9.61095512e-01 -5.27811170e-01 3.66664022e-01 4.19084609e-01
-1.40904665e-01 3.62667322e-01 -3.97023335e-02 -1.48589981e+00
-1.96840256e-01 -1.00986600e+00 4.19095755e-01 -6.03175223e-01
-1.55260694e+00 8.89163554e-01 -7.72899315e-02 -1.19650769e+00
-2.88286030e-01 -2.32972458e-01 -9.91697609e-01 9.28737998e-01
-1.13547397e+00 -1.01388264e+00 -6.21934175e-01 5.46356499e-01
8.25891435e-01 -3.42810422e-01 7.30136037e-01 5.00956833e-01
-8.49147856e-01 1.16833258e+00 1.13491021e-01 1.47991017e-01
1.05214930e+00 -1.32069767e+00 8.19342136e-01 7.06604660e-01
1.61382437e-01 8.03249419e-01 1.05111396e+00 -5.95552504e-01
-1.04150677e+00 -1.31570280e+00 1.03735030e+00 -5.72674572e-01
5.82608581e-01 -5.42170048e-01 -4.77891028e-01 9.09096301e-01
2.59746879e-01 -4.35985476e-01 1.09908295e+00 3.21773201e-01
-6.99955523e-02 5.69015369e-02 -7.88115442e-01 8.40321958e-01
1.13321567e+00 -1.69924691e-01 -2.55369604e-01 7.65953779e-01
9.41618621e-01 -6.69995308e-01 -6.68100059e-01 -7.11160824e-02
5.30325770e-01 -6.16156459e-01 6.34885848e-01 -8.29413414e-01
5.53310513e-01 3.06747463e-02 3.77982855e-02 -1.68348980e+00
-5.44627547e-01 -1.10453618e+00 -8.75126794e-02 1.70553815e+00
1.17044938e+00 -4.43693042e-01 7.54193783e-01 9.77666557e-01
-2.72288024e-01 -8.55968833e-01 -2.06614062e-02 -4.47468698e-01
-5.55442333e-01 -2.21089944e-01 8.74087274e-01 8.38488579e-01
1.92030594e-01 8.85405004e-01 -6.94130182e-01 8.68591815e-02
3.58150989e-01 2.14432612e-01 1.04245508e+00 -1.00595021e+00
-7.42843390e-01 -4.59012926e-01 4.50970381e-01 -1.76706982e+00
5.28465733e-02 -9.12942410e-01 1.27462670e-01 -1.43556654e+00
5.21143675e-01 -6.81292832e-01 -2.94377595e-01 3.19712549e-01
-4.53574508e-01 -1.71979383e-01 1.28386840e-01 2.65987456e-01
-9.53986108e-01 6.33613825e-01 1.15377355e+00 -1.47857174e-01
-4.81226265e-01 8.95118833e-01 -1.36921585e+00 5.68350852e-01
9.09740031e-01 -2.14926824e-01 -8.13929319e-01 -3.77566069e-01
2.72807747e-01 1.01869367e-01 -1.18656032e-01 -7.11594164e-01
3.73688519e-01 -4.99746114e-01 2.21818864e-01 -3.86942804e-01
1.74981147e-01 -2.46915266e-01 -1.70962080e-01 -1.66814178e-01
-1.20788431e+00 5.23864897e-03 -4.44619149e-01 4.96381283e-01
4.29059304e-02 -2.01893255e-01 6.06945634e-01 -5.77326640e-02
-3.82908970e-01 1.02605271e+00 -1.57218874e-01 3.37478578e-01
6.18158579e-01 6.22733533e-01 -6.83724403e-01 -9.35689628e-01
-7.28168845e-01 4.64390695e-01 3.80012393e-01 6.79140151e-01
4.22058731e-01 -1.34144104e+00 -7.39408553e-01 1.24883736e-02
5.30579925e-01 -4.45643097e-01 2.19315082e-01 3.51882100e-01
-9.86704528e-02 3.05251926e-01 -1.60626008e-03 8.23287386e-03
-1.23581016e+00 3.17058504e-01 3.20582166e-02 -2.19270766e-01
-2.75229454e-01 1.23076999e+00 1.57854915e-01 -3.66986394e-01
5.29200494e-01 -1.03479519e-01 -5.95799744e-01 -5.98289818e-02
1.04934323e+00 2.82038987e-01 -8.86773691e-02 -3.30082804e-01
1.21714965e-01 -4.04761702e-01 -5.38097322e-01 -4.12197381e-01
1.26270545e+00 -4.51876372e-01 3.16167027e-01 5.05861759e-01
1.11296618e+00 4.62213755e-01 -9.33343589e-01 -5.28708875e-01
-1.94833845e-01 -4.69315708e-01 -1.40120775e-01 -1.05698931e+00
-8.44941199e-01 3.08600008e-01 2.03458652e-01 5.57664454e-01
7.90512025e-01 1.73180819e-01 1.20695281e+00 4.31410819e-01
1.75952822e-01 -1.22152483e+00 2.84704380e-02 6.10024154e-01
1.01217175e+00 -1.19953716e+00 -4.00152683e-01 -3.23587030e-01
-1.09082019e+00 6.68236852e-01 7.58799493e-01 3.07657897e-01
8.30698788e-01 -1.24548651e-01 1.20675541e-01 1.79091170e-01
-9.46729541e-01 1.75787676e-02 6.11027002e-01 2.72533089e-01
8.41093719e-01 2.54098803e-01 -2.47013524e-01 1.43785548e+00
-8.60245824e-01 -1.39972195e-01 6.04472339e-01 5.86304307e-01
-4.75658298e-01 -1.23058140e+00 1.13788731e-01 8.38688433e-01
-4.01080519e-01 -1.11421630e-01 -3.31947595e-01 7.40935504e-02
-2.79932618e-01 9.15505528e-01 -3.20929319e-01 -7.04094231e-01
5.42260587e-01 -1.38165072e-01 -6.87042847e-02 -9.71076846e-01
-4.29788023e-01 -7.04499409e-02 2.92928100e-01 -3.79099697e-01
9.64833945e-02 -7.11428165e-01 -8.01135242e-01 -4.10653323e-01
-1.00561224e-01 5.36324859e-01 5.57460189e-01 6.74177825e-01
4.74794030e-01 3.93855423e-01 1.07809949e+00 -8.70294571e-01
-7.33773530e-01 -1.17916489e+00 -6.42450929e-01 6.35150313e-01
1.25252351e-01 -3.51539195e-01 -2.23976985e-01 2.56302744e-01] | [10.406988143920898, 5.911024570465088] |
e3e45955-2cad-4e35-bae6-3cb0c68b5f3b | ra-sr-using-a-ranking-algorithm-to | null | null | https://aclanthology.org/W13-1613 | https://aclanthology.org/W13-1613.pdf | RA-SR: Using a ranking algorithm to automatically building resources for subjectivity analysis over annotated corpora | null | ['Rafael Mu{\\~n}oz', "Andr{\\'e}s Montoyo", "Antonio Fern{\\'a}ndez", "Andy Gonz{\\'a}lez", "Yoan Guti{\\'e}rrez"] | 2013-06-01 | null | null | null | ws-2013-6 | ['subjectivity-analysis'] | ['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] | [-7.276517391204834, 3.7236030101776123] |
a60643f7-8b70-4936-9722-2c80cdb18b02 | illumination-normalization-via-merging | 1905.03904 | null | https://arxiv.org/abs/1905.03904v1 | https://arxiv.org/pdf/1905.03904v1.pdf | Illumination Normalization via Merging Locally Enhanced Textures for Robust Face Recognition | In order to improve the accuracy of face recognition under varying illumination conditions, a local texture enhanced illumination normalization method based on fusion of differential filtering images (FDFI-LTEIN) is proposed to weaken the influence caused by illumination changes. Firstly, the dynamic range of the face image in dark or shadowed regions is expanded by logarithmic transformation. Then, the global contrast enhanced face image is convolved with difference of Gaussian filters and difference of bilateral filters, and the filtered images are weighted and merged using a coefficient selection rule based on the standard deviation (SD) of image, which can enhance image texture information while filtering out most noise. Finally, the local contrast equalization (LCE) is performed on the fused face image to reduce the influence caused by over or under saturated pixel values in highlight or dark regions. Experimental results on the Extended Yale B face database and CMU PIE face database demonstrate that the proposed method is more robust to illumination changes and achieve higher recognition accuracy when compared with other illumination normalization methods and a deep CNNs based illumination invariant face recognition method | ['Chaobing Zheng', 'Wangming Xu', 'Shoulie Xie', 'Shiqian Wu'] | 2019-05-10 | null | null | null | null | ['robust-face-recognition'] | ['computer-vision'] | [ 4.01726812e-01 -7.25632429e-01 3.38327795e-01 -6.56223655e-01
1.61891252e-01 -1.90978751e-01 1.84972838e-01 -7.61690199e-01
-5.32486916e-01 6.28814697e-01 4.60788347e-02 1.42282978e-01
-6.05941527e-02 -9.69571173e-01 -2.84003466e-01 -1.42118144e+00
3.04346651e-01 -4.14471567e-01 8.74344781e-02 -1.08930767e-01
2.27545679e-01 9.65260983e-01 -1.89564443e+00 2.16678292e-01
7.27566719e-01 1.22063458e+00 -2.19206307e-02 8.69183093e-02
-1.24975890e-01 2.81311601e-01 -6.62368059e-01 -9.06874835e-02
4.04247075e-01 -4.26652580e-01 -3.12500358e-01 4.18967187e-01
4.28975165e-01 -4.97265011e-01 -3.98986191e-01 1.43850803e+00
8.62385929e-01 3.27916890e-01 4.37534899e-01 -1.00183058e+00
-7.29822814e-01 -1.58716485e-01 -8.18021953e-01 5.79733670e-01
-2.06467748e-01 4.10194770e-02 -2.06146076e-01 -1.26089418e+00
4.27453101e-01 1.67490041e+00 2.36311853e-01 5.03207862e-01
-1.13771987e+00 -9.69273090e-01 -2.44600885e-02 4.22716767e-01
-1.47088945e+00 -4.80894446e-01 7.81939089e-01 1.24746278e-01
5.87207675e-01 1.82956710e-01 3.28606546e-01 5.22129118e-01
4.07234907e-01 1.05009496e-01 1.53154266e+00 -5.75468898e-01
-1.59551427e-01 -1.45852178e-01 -2.19239309e-01 8.44123006e-01
1.89366326e-01 2.21914276e-01 -3.99193913e-01 9.84359682e-02
9.63130891e-01 2.21914798e-01 -4.32486117e-01 1.75721660e-01
-8.78940344e-01 3.71989340e-01 4.58414614e-01 4.35994744e-01
-3.87047529e-01 -5.37338018e-01 2.55417764e-01 3.77426714e-01
5.16167402e-01 -3.93410593e-01 -3.82994801e-01 4.24535453e-01
-7.39533126e-01 -2.18342364e-01 4.21463877e-01 2.64854670e-01
8.96063924e-01 3.43642086e-01 -2.84824938e-01 1.23454642e+00
4.41950589e-01 8.41798604e-01 4.47085619e-01 -6.36136830e-01
-3.74085419e-02 7.01114893e-01 -5.53698130e-02 -1.10888767e+00
-7.91548416e-02 -2.46759251e-01 -1.01288688e+00 5.50473690e-01
3.24150920e-01 -7.67769217e-02 -1.19129157e+00 1.48464549e+00
5.06112576e-01 8.84306729e-02 1.87699478e-02 1.04739964e+00
9.72118258e-01 4.63310540e-01 1.25591112e-02 -5.74063301e-01
1.42245889e+00 -5.10303617e-01 -1.01174092e+00 1.66174799e-01
-2.09588081e-01 -1.10237038e+00 7.98318565e-01 2.29564607e-01
-8.63390148e-01 -8.62727761e-01 -1.14244092e+00 1.11479923e-01
-5.00814080e-01 2.16221184e-01 1.69294536e-01 9.05314088e-01
-9.62180972e-01 3.85802001e-01 -5.78405797e-01 -2.74860501e-01
5.36342323e-01 5.32527328e-01 -5.64768076e-01 -4.13819164e-01
-1.02382469e+00 7.88229048e-01 2.59455085e-01 6.74460649e-01
-4.86817211e-01 -4.09505814e-01 -8.32848907e-01 1.26229674e-01
3.19488943e-02 -9.84185264e-02 5.08218348e-01 -1.37342370e+00
-1.78806710e+00 8.06031704e-01 -6.03183389e-01 3.17842156e-01
1.14174336e-01 2.70516634e-01 -8.46633911e-01 5.37954867e-02
-3.71133745e-01 3.34650666e-01 9.89148617e-01 -8.92625213e-01
-4.21814352e-01 -7.33682036e-01 -4.59919214e-01 3.69577885e-01
-2.09707871e-01 5.37848651e-01 -5.24141908e-01 -4.28728908e-01
3.53761286e-01 -4.23864126e-01 2.65079677e-01 5.86946867e-02
3.02466601e-01 -3.87494713e-02 1.64300859e+00 -1.01029885e+00
9.21950638e-01 -2.52236223e+00 -3.45460355e-01 6.59177542e-01
-2.60897219e-01 4.60921407e-01 -4.75233912e-01 -2.88396388e-01
-3.06945235e-01 -2.86746591e-01 -1.09149873e-01 3.08583826e-01
-2.89820403e-01 3.40217590e-01 2.20018759e-01 6.82339072e-01
1.26189709e-01 4.93432313e-01 -2.65095145e-01 -5.68355858e-01
4.63904858e-01 1.14362836e+00 3.14680114e-02 -5.45293465e-02
4.05327648e-01 5.65209746e-01 -8.06175619e-02 8.08716953e-01
1.50166297e+00 4.60600019e-01 2.93173771e-02 -4.68582004e-01
-2.80660778e-01 -3.83250117e-01 -1.55381799e+00 7.40879118e-01
-1.97956368e-01 3.90507162e-01 6.28227592e-01 -7.63634861e-01
1.31605756e+00 2.67399907e-01 2.83679217e-01 -1.09274435e+00
6.73464358e-01 3.24147008e-02 8.39781074e-04 -5.34263492e-01
-9.53116789e-02 -6.06876500e-02 7.68793404e-01 2.23757342e-01
-8.16722289e-02 3.48753095e-01 9.37108621e-02 -4.86807197e-01
2.78600216e-01 4.56456505e-02 -1.28175495e-02 -5.79458714e-01
1.19130838e+00 -8.13373923e-01 8.94932508e-01 1.54042259e-01
-4.68569160e-01 2.59556502e-01 7.81499818e-02 -6.54276788e-01
-7.13879108e-01 -7.31153011e-01 -5.53660095e-01 9.18613076e-01
3.83993983e-01 3.69355470e-01 -1.01189792e+00 -3.92116219e-01
-5.16353734e-02 4.00844328e-02 -6.48452342e-01 -1.58148557e-01
-4.95819241e-01 -1.09002805e+00 2.00392187e-01 3.43842626e-01
1.43044269e+00 -1.25185382e+00 -1.77677438e-01 3.29011157e-02
-1.81061476e-01 -7.37257063e-01 -5.35747528e-01 -1.47982568e-01
-5.75467110e-01 -1.03308308e+00 -5.86956263e-01 -1.04993284e+00
1.01743400e+00 4.56116438e-01 4.15325731e-01 3.65738153e-01
-5.98946750e-01 8.80998597e-02 -7.46982023e-02 -3.33405554e-01
-2.71343510e-03 -7.78227627e-01 -1.35412421e-02 5.38752258e-01
6.16619468e-01 -1.56778231e-01 -8.70977640e-01 7.72233844e-01
-1.04627931e+00 -5.09584308e-01 4.28512931e-01 1.08258116e+00
5.85265517e-01 8.03225994e-01 1.80359676e-01 -2.08726302e-01
4.44977880e-01 2.91360825e-01 -7.29497969e-01 3.82977456e-01
-4.01464283e-01 -2.25437850e-01 2.96108574e-01 -5.57709455e-01
-1.93844211e+00 -2.09479019e-01 1.80874303e-01 -1.00292869e-01
-3.53444070e-01 2.10685551e-01 -7.00932801e-01 -8.71621132e-01
3.76066566e-01 4.77948666e-01 3.79975796e-01 -2.55023539e-01
-8.51137415e-02 7.60504961e-01 6.23993754e-01 -2.33064935e-01
7.00818777e-01 6.57826841e-01 2.20492572e-01 -1.00802112e+00
-2.34289810e-01 -1.05258636e-01 -6.58388317e-01 -4.73940521e-01
8.76761138e-01 -6.97135627e-01 -9.42262471e-01 1.07178664e+00
-7.56621599e-01 1.10345371e-02 3.74684960e-01 5.78919053e-01
2.55552024e-01 5.11397541e-01 -7.27503061e-01 -6.97417080e-01
-5.13966739e-01 -1.15314209e+00 5.03814399e-01 1.01452529e+00
5.88327944e-01 -7.15001404e-01 -5.38730323e-01 2.88782090e-01
8.99701655e-01 1.48438126e-01 8.95108223e-01 4.41848338e-02
-3.78012568e-01 -9.76485312e-02 -5.42130291e-01 8.35766315e-01
5.69440544e-01 3.41899902e-01 -9.90073800e-01 -3.88856322e-01
2.48575017e-01 2.34947041e-01 7.70806968e-01 4.79655713e-01
1.26381350e+00 -1.17264330e-01 -8.95364508e-02 7.61679173e-01
1.55998373e+00 5.63374281e-01 1.04149413e+00 3.02084357e-01
2.00129926e-01 4.20837075e-01 5.16321242e-01 2.73118228e-01
-1.98839381e-01 2.83848882e-01 2.29154736e-01 -7.49358356e-01
-3.21459979e-01 5.04188359e-01 1.96415275e-01 1.71701923e-01
-3.37957561e-01 1.23324014e-01 -1.61748201e-01 1.38272807e-01
-1.28403175e+00 -1.08440936e+00 3.15252468e-02 2.20636153e+00
7.57146239e-01 -3.57043654e-01 -3.13289344e-01 1.26562536e-01
1.20128441e+00 -1.64142698e-01 -4.69491839e-01 -3.73393863e-01
-7.09984660e-01 7.23539233e-01 5.04506886e-01 2.83556879e-01
-1.02083302e+00 7.87599325e-01 5.81211758e+00 8.43539238e-01
-1.64632809e+00 -1.31535996e-02 7.83786654e-01 1.97523475e-01
2.81864703e-01 -3.97194207e-01 -6.75318003e-01 5.25256872e-01
2.35238999e-01 2.20996425e-01 8.45603168e-01 4.98010308e-01
3.38611245e-01 -4.01304871e-01 -1.53163597e-01 1.14964938e+00
2.19656795e-01 -4.79987025e-01 -1.33227333e-01 4.57775556e-02
7.03572392e-01 -3.57269287e-01 4.35406744e-01 4.61606719e-02
-8.57079849e-02 -9.61523890e-01 1.38214633e-01 6.92044973e-01
7.12532640e-01 -1.01105595e+00 1.09931016e+00 -1.88687876e-01
-1.22650051e+00 -1.58308372e-01 -6.79038882e-01 4.04194891e-02
-4.73753303e-01 5.59669852e-01 -2.47313648e-01 4.43841755e-01
9.82610285e-01 1.36234000e-01 -4.44282591e-01 8.11066270e-01
-1.56213060e-01 3.04934919e-01 -4.70042259e-01 3.22818786e-01
-1.72251642e-01 -4.82860535e-01 1.99991852e-01 1.08551085e+00
4.03112739e-01 4.11318392e-01 -1.87840268e-01 6.82714224e-01
-1.27086088e-01 2.40688503e-01 -1.02671914e-01 2.85768569e-01
3.67193371e-01 1.66596353e+00 -9.09569919e-01 -3.92201900e-01
-5.83159328e-01 1.03793478e+00 -2.87647635e-01 8.01058650e-01
-6.46006107e-01 -5.75256288e-01 4.33372647e-01 -2.62995720e-01
5.89749694e-01 5.52300178e-02 -1.25598311e-01 -8.98831129e-01
2.80499458e-04 -9.03062940e-01 3.78404588e-01 -6.76100373e-01
-8.93460929e-01 4.78813320e-01 -3.07530105e-01 -7.82150388e-01
7.09694743e-01 -7.75224447e-01 -7.55297244e-01 1.35455060e+00
-1.74839616e+00 -9.50675249e-01 -5.99717498e-01 9.71594989e-01
2.18583122e-01 -2.33629689e-01 5.81056714e-01 5.12368381e-01
-6.92301631e-01 6.03265226e-01 2.58440495e-01 1.12195231e-01
9.92480874e-01 -4.81584072e-01 -4.68480587e-01 1.06089258e+00
-7.20834672e-01 7.25156665e-01 3.10051262e-01 -6.12377286e-01
-1.11080110e+00 -9.98561263e-01 7.00459659e-01 2.63311088e-01
-2.76205182e-01 -1.12079218e-01 -1.08298385e+00 2.11476266e-01
3.26597095e-01 2.28550747e-01 5.43816745e-01 -4.28910106e-01
-3.37146461e-01 -5.83341718e-01 -1.71410871e+00 5.01591980e-01
4.74201411e-01 -3.54086667e-01 -2.10444540e-01 2.95875724e-02
-5.80200516e-02 -2.45312363e-01 -9.87654328e-01 8.74130845e-01
8.14832509e-01 -9.32013333e-01 8.15940559e-01 -4.56138812e-02
-2.04896465e-01 -8.58674705e-01 -1.69299215e-01 -9.07452106e-01
-6.00411057e-01 -2.49143571e-01 8.41631532e-01 1.43756878e+00
-6.44348785e-02 -1.00591087e+00 3.36020857e-01 5.46308935e-01
1.78217039e-01 -3.47057849e-01 -8.43434155e-01 -4.86713529e-01
-5.06139219e-01 5.55944622e-01 9.03217971e-01 8.04593086e-01
-4.62426960e-01 -2.55271763e-01 -4.36877972e-03 4.27175671e-01
7.14476943e-01 3.90287526e-02 4.89672542e-01 -1.03647137e+00
4.95836973e-01 -3.84541780e-01 -2.73268193e-01 -2.64814526e-01
5.31384870e-02 -5.00776410e-01 -1.39688596e-01 -1.08350885e+00
3.09699595e-01 3.04492831e-01 -6.96493804e-01 5.87965071e-01
-2.10983276e-01 8.09702039e-01 -1.73587650e-01 -4.31050435e-02
-5.80598712e-02 5.22476256e-01 1.59169579e+00 -1.42889693e-01
-5.01768887e-02 -2.01107964e-01 -3.65934998e-01 5.53238332e-01
6.14327490e-01 7.03407004e-02 4.64416072e-02 -1.39569744e-01
-5.23005068e-01 -2.47341141e-01 3.29661757e-01 -1.00551105e+00
1.04333416e-01 -2.29023322e-01 1.21054077e+00 -4.99401242e-01
1.25093013e-01 -9.72223401e-01 1.75087988e-01 4.70415801e-01
1.87769189e-01 -1.79276198e-01 4.43527937e-01 5.89052290e-02
-3.98295701e-01 1.28758281e-01 1.41597915e+00 5.00509031e-02
-8.17723989e-01 4.06955659e-01 -4.81530458e-01 -8.58635545e-01
8.61521602e-01 -7.25836396e-01 -4.42312181e-01 -4.52909842e-02
-5.85239887e-01 -2.22715050e-01 3.53105068e-01 2.55479336e-01
8.20775926e-01 -1.42829990e+00 -6.19945049e-01 1.08789349e+00
-2.58652180e-01 -5.04400373e-01 7.42018580e-01 9.14625764e-01
-5.15768051e-01 1.11963429e-01 -7.44242966e-01 -2.82865196e-01
-1.94514978e+00 3.53316188e-01 6.17667735e-01 3.06983441e-01
-1.28196374e-01 8.79785776e-01 3.16037893e-01 -1.74895197e-01
1.79657683e-01 6.00450858e-02 -4.27267760e-01 -3.36263537e-01
8.65703106e-01 4.41322267e-01 3.21691453e-01 -1.09505737e+00
-4.43025470e-01 8.90734851e-01 -1.59076497e-01 1.76268816e-01
1.19835603e+00 -3.03774565e-01 -7.57135212e-01 -5.02816737e-01
1.18841410e+00 -6.19950034e-02 -1.15842319e+00 -3.94195169e-01
-7.37430274e-01 -8.10652912e-01 3.54252011e-01 -9.81096506e-01
-1.46149945e+00 7.28257835e-01 1.35028780e+00 -2.48322681e-01
1.75150514e+00 -6.81601167e-01 4.34177816e-01 8.44281316e-02
-1.24062486e-01 -1.20764482e+00 -7.84824267e-02 5.51472783e-01
7.90373802e-01 -9.44996357e-01 -4.78881560e-02 -5.09084523e-01
-1.72881439e-01 1.47276938e+00 7.41849959e-01 1.24690440e-02
7.88951099e-01 3.93248081e-01 3.39156955e-01 8.12865514e-03
-8.35975260e-02 -2.22417563e-01 3.28796327e-01 6.16753936e-01
2.62401432e-01 -3.85008484e-01 -4.90300387e-01 1.42775998e-01
2.19875857e-01 9.15118400e-03 -1.16575748e-01 7.38960266e-01
-6.33215606e-01 -9.06400979e-01 -8.83893728e-01 1.81943610e-01
-6.71528578e-01 1.18331179e-01 -9.72574577e-02 7.12494612e-01
6.04964793e-01 1.09049594e+00 2.03953460e-01 -1.19890869e-01
1.84588164e-01 2.88484544e-01 7.88011491e-01 6.12763539e-02
-3.50852221e-01 6.46272719e-01 -5.20294249e-01 -4.74044263e-01
-7.60335445e-01 -3.63725334e-01 -1.08973563e+00 -5.04546762e-01
-6.13340974e-01 -2.10324019e-01 8.23470354e-01 9.04964745e-01
1.22495547e-01 4.11071628e-01 7.80330181e-01 -7.23578274e-01
-1.33546889e-01 -1.02354574e+00 -9.83686566e-01 3.96785378e-01
1.34225070e-01 -8.30317676e-01 -5.90399206e-01 2.27978826e-01] | [13.047440528869629, 0.4177974760532379] |
abfb1aae-97d8-4eeb-95e1-10e40e0e5dc9 | learning-based-video-motion-magnification | 1804.02684 | null | http://arxiv.org/abs/1804.02684v3 | http://arxiv.org/pdf/1804.02684v3.pdf | Learning-based Video Motion Magnification | Video motion magnification techniques allow us to see small motions
previously invisible to the naked eyes, such as those of vibrating airplane
wings, or swaying buildings under the influence of the wind. Because the motion
is small, the magnification results are prone to noise or excessive blurring.
The state of the art relies on hand-designed filters to extract representations
that may not be optimal. In this paper, we seek to learn the filters directly
from examples using deep convolutional neural networks. To make training
tractable, we carefully design a synthetic dataset that captures small motion
well, and use two-frame input for training. We show that the learned filters
achieve high-quality results on real videos, with less ringing artifacts and
better noise characteristics than previous methods. While our model is not
trained with temporal filters, we found that the temporal filters can be used
with our extracted representations up to a moderate magnification, enabling a
frequency-based motion selection. Finally, we analyze the learned filters and
show that they behave similarly to the derivative filters used in previous
works. Our code, trained model, and datasets will be available online. | ['Frédo Durand', 'Tae-Hyun Oh', 'Ronnachai Jaroensri', 'Wojciech Matusik', 'William T. Freeman', 'Mohamed Elgharib', 'Changil Kim'] | 2018-04-08 | learning-based-video-motion-magnification-1 | http://openaccess.thecvf.com/content_ECCV_2018/html/Tae-Hyun_Oh_Learning-based_Video_Motion_ECCV_2018_paper.html | http://openaccess.thecvf.com/content_ECCV_2018/papers/Tae-Hyun_Oh_Learning-based_Video_Motion_ECCV_2018_paper.pdf | eccv-2018-9 | ['motion-magnification'] | ['computer-vision'] | [ 3.31189334e-01 -2.58088320e-01 7.62746036e-02 2.09743991e-01
-2.06736177e-01 -7.91248083e-01 5.30436993e-01 -5.49469054e-01
-3.30588996e-01 6.29938960e-01 2.22543702e-01 -1.89436421e-01
4.59575094e-02 -6.79230034e-01 -7.86732495e-01 -6.70496583e-01
-4.72346812e-01 -4.98717517e-01 6.54394627e-01 -6.31503388e-02
2.18475714e-01 3.75737250e-01 -1.69756627e+00 3.28256845e-01
4.40570623e-01 7.96111166e-01 2.83382386e-01 1.08320177e+00
5.34524202e-01 7.80420184e-01 -8.14816058e-01 -1.52385207e-02
4.76568073e-01 -4.05114681e-01 -6.47472084e-01 1.39518753e-01
8.19126666e-01 -7.91822970e-01 -6.01610243e-01 8.92782032e-01
2.75383949e-01 3.23989451e-01 3.96087319e-01 -6.36300206e-01
-7.59142637e-01 3.18845809e-02 -5.24774551e-01 5.99830925e-01
4.10338193e-01 5.11815369e-01 6.44807577e-01 -7.92020559e-01
7.86375880e-01 1.13537383e+00 7.45531440e-01 7.99692333e-01
-1.27745306e+00 -3.54032934e-01 4.23098505e-02 1.21204615e-01
-1.08287358e+00 -5.57613909e-01 6.44818366e-01 -4.31422621e-01
9.30613697e-01 3.74204487e-01 7.69459546e-01 1.26279783e+00
1.94637820e-01 5.73387027e-01 6.65093780e-01 -4.13484186e-01
1.07870281e-01 -1.24410279e-01 -2.96518683e-01 6.87874913e-01
2.71514565e-01 3.30768406e-01 -3.95816088e-01 -6.83389083e-02
1.13382709e+00 -6.05009459e-02 -9.79502082e-01 -2.55377054e-01
-1.51879168e+00 4.63068724e-01 4.92621481e-01 3.05244148e-01
-1.18058339e-01 6.06949866e-01 1.49501801e-01 4.23493862e-01
3.37706596e-01 5.54507971e-01 -3.55648816e-01 -2.21156836e-01
-1.08025062e+00 1.62228808e-01 4.80146825e-01 6.87951863e-01
4.83041853e-01 1.96514532e-01 -9.11058206e-03 4.68265921e-01
-1.03489660e-01 4.71099347e-01 4.48237300e-01 -1.44825244e+00
2.06895545e-01 -5.83442077e-02 4.95285571e-01 -1.10314465e+00
-4.30701524e-01 -3.62124503e-01 -6.70802414e-01 6.33442819e-01
8.06479335e-01 -2.05735594e-01 -8.47981930e-01 1.64641058e+00
8.75162147e-03 2.96685040e-01 -1.11894779e-01 1.22225749e+00
4.60956901e-01 7.12200403e-01 -3.12771916e-01 -2.36408278e-01
1.23142970e+00 -1.01949000e+00 -7.94593632e-01 -2.82038420e-01
4.68130618e-01 -7.99808204e-01 1.38416100e+00 5.05000293e-01
-1.10286736e+00 -6.48584962e-01 -1.27196193e+00 -1.33072790e-02
2.86771241e-03 2.67747551e-01 4.98470157e-01 5.26128411e-01
-1.13738799e+00 1.20539427e+00 -1.05360150e+00 -4.43892747e-01
1.81776091e-01 1.63683608e-01 -2.17110500e-01 1.62802830e-01
-1.09990644e+00 7.38430023e-01 -3.52651253e-02 4.28532325e-02
-9.24158037e-01 -5.04271150e-01 -7.68657744e-01 1.25357181e-01
3.09252739e-01 -8.82116437e-01 1.32899320e+00 -1.18788493e+00
-1.59776723e+00 4.65817183e-01 -1.94006816e-01 -4.50438380e-01
6.74846828e-01 -5.26495993e-01 -4.01452422e-01 6.83065712e-01
-3.66389304e-01 6.87281609e-01 1.33086193e+00 -1.20441616e+00
-4.79250282e-01 2.23370835e-01 4.60533947e-01 -1.25884145e-01
-5.82294405e-01 -1.70483202e-01 -3.18499357e-01 -1.08000350e+00
-2.51289517e-01 -9.96769726e-01 -1.66874677e-01 4.56226885e-01
3.78682762e-02 5.69267035e-01 1.11014009e+00 -6.15780532e-01
1.44604313e+00 -2.32841039e+00 5.53024933e-02 -2.33384609e-01
1.33303627e-01 3.89812946e-01 -4.31824476e-02 2.34583110e-01
-1.72516257e-02 1.41599208e-01 -2.78126448e-01 -1.02862477e-01
-4.06351209e-01 2.80715059e-02 -3.79216313e-01 5.45416415e-01
2.77024001e-01 5.47297478e-01 -1.07297385e+00 -1.34198412e-01
2.63225675e-01 6.46077096e-01 -7.23832726e-01 1.10131375e-01
-7.30502605e-02 4.29798245e-01 -6.93686306e-02 4.55979556e-01
5.22909224e-01 -4.69514847e-01 2.11023781e-02 -4.39988166e-01
-3.12427823e-02 1.19760387e-01 -1.07940662e+00 1.46977508e+00
-4.54471290e-01 1.19355881e+00 1.30708829e-01 -4.96747911e-01
5.10909140e-01 3.18689197e-01 3.31835508e-01 -4.45339978e-01
-6.57242388e-02 1.38667122e-01 1.37645956e-02 -8.37180018e-01
5.08839846e-01 -2.42879000e-02 3.99050236e-01 3.18841815e-01
-1.01865917e-01 -3.13977934e-02 2.63098180e-01 -2.32665576e-02
1.21617413e+00 2.64199197e-01 1.32858843e-01 -2.27774456e-01
2.25787655e-01 -6.12542816e-02 4.21391666e-01 5.06400347e-01
-1.18283838e-01 9.75645781e-01 2.74778843e-01 -7.06349075e-01
-1.07642353e+00 -9.46543396e-01 -6.36482760e-02 6.87472463e-01
2.79099941e-01 -4.09485698e-01 -6.66049242e-01 -4.35149729e-01
-1.54908657e-01 3.71787459e-01 -5.78227282e-01 -1.23767138e-01
-9.08240676e-01 -4.69815910e-01 3.60932112e-01 6.32810175e-01
3.76460731e-01 -7.76805580e-01 -1.21560311e+00 9.55059901e-02
-4.29799408e-02 -1.16176295e+00 -7.35454917e-01 -1.38986588e-01
-1.03808272e+00 -1.08632863e+00 -1.12875450e+00 -5.21759212e-01
6.16548717e-01 6.35236084e-01 9.82051611e-01 1.66013539e-01
-3.18290621e-01 2.85373777e-01 -1.85269579e-01 6.24366216e-02
-2.27489486e-01 -2.09469035e-01 3.49135511e-02 -1.47484317e-01
-1.46906227e-01 -7.10609853e-01 -1.11577463e+00 3.61061066e-01
-1.06980622e+00 6.35600388e-02 3.33042592e-01 8.49748075e-01
8.00764486e-02 1.27648413e-01 5.95878139e-02 -4.44922268e-01
4.94633108e-01 -1.53598428e-01 -6.41598523e-01 -1.72302753e-01
-1.89945728e-01 4.12676111e-02 9.63572502e-01 -8.65847468e-01
-8.82365584e-01 2.99633574e-02 2.51317710e-01 -7.15257585e-01
-4.03736681e-02 7.22085908e-02 4.20367092e-01 -2.50700414e-01
9.61034417e-01 -1.07699960e-01 -1.63959354e-01 -4.57696795e-01
2.83555657e-01 3.95534217e-01 6.43140614e-01 -1.50704309e-01
8.57505620e-01 9.42303061e-01 -1.06575824e-01 -1.01730514e+00
-4.54066783e-01 -5.96178025e-02 -3.82824570e-01 -3.37367743e-01
7.23131239e-01 -8.42007637e-01 -6.45111263e-01 2.43439212e-01
-1.16674018e+00 -4.87637997e-01 -2.06631705e-01 6.91029251e-01
-5.21106303e-01 4.82483059e-01 -9.86465812e-01 -7.28443623e-01
2.35429760e-02 -1.02816343e+00 9.89939511e-01 3.48711133e-01
-3.96660358e-01 -8.99171114e-01 -5.47676943e-02 -1.19178087e-01
5.97784698e-01 3.64364207e-01 6.94854856e-01 3.16143066e-01
-8.60735774e-01 1.22547142e-01 7.86569640e-02 3.63278061e-01
3.74586612e-01 5.44227242e-01 -1.22390640e+00 -4.88407105e-01
1.19696192e-01 -3.70600894e-02 1.19256830e+00 5.76882005e-01
1.18685436e+00 -4.95241553e-01 -2.54240453e-01 9.71022785e-01
1.33973873e+00 1.97601482e-01 6.65413141e-01 3.87900800e-01
5.56336462e-01 4.90872264e-01 3.40761721e-01 4.02390540e-01
-2.06338912e-01 6.74886048e-01 3.60861301e-01 -2.04845145e-01
-3.11381251e-01 -3.95266712e-02 4.39912260e-01 4.52412903e-01
-4.88923430e-01 -2.54389256e-01 -6.29991055e-01 5.94412148e-01
-1.62487388e+00 -1.15469372e+00 6.08358569e-02 2.22200227e+00
4.97785836e-01 3.56309801e-01 1.42249569e-01 2.84391254e-01
4.88128453e-01 3.92554998e-01 -3.62783521e-01 -2.19271511e-01
-4.09006402e-02 1.18317865e-01 5.48102975e-01 5.79667687e-01
-1.39116609e+00 6.04005516e-01 7.29356575e+00 3.22870910e-01
-1.45415318e+00 -2.07100257e-01 4.36055988e-01 -5.93242228e-01
-3.59678268e-01 1.21874344e-02 -3.82182956e-01 5.21177769e-01
8.06714535e-01 2.61879060e-02 5.99627554e-01 6.35853708e-01
5.78984857e-01 -1.45888284e-01 -1.00923908e+00 9.38324511e-01
-2.37043038e-01 -1.50722897e+00 -1.58949904e-02 -9.86750200e-02
7.17550039e-01 -4.02962804e-01 2.12605596e-01 -1.94853589e-01
-6.62345067e-02 -1.12179101e+00 8.18331301e-01 5.84996045e-01
7.67090380e-01 -5.45066833e-01 3.07660669e-01 9.65866521e-02
-1.15674293e+00 -2.55107015e-01 -4.05999213e-01 -2.88715094e-01
2.60408372e-01 6.44862592e-01 -4.44171220e-01 2.76477486e-01
7.86956966e-01 7.23639488e-01 -5.00174403e-01 1.06708884e+00
-6.50622249e-02 6.31938696e-01 -2.45768726e-01 -6.01584744e-03
2.22233087e-01 7.65496269e-02 6.77272141e-01 1.20522165e+00
5.82351506e-01 -7.93815926e-02 -1.65944576e-01 7.52263606e-01
5.63786831e-03 -3.48814636e-01 -8.42847347e-01 -4.48743068e-02
4.66329530e-02 1.07801783e+00 -6.61253393e-01 -3.37711126e-01
-5.74046075e-01 1.11655962e+00 5.27949668e-02 7.89293170e-01
-8.71856689e-01 -4.83258188e-01 9.47759748e-01 3.85889351e-01
5.92795193e-01 -5.94315469e-01 1.01828963e-01 -1.42020369e+00
1.36743739e-01 -8.72186184e-01 1.46446973e-02 -8.72130156e-01
-7.89769530e-01 5.81764579e-01 -1.53732095e-02 -1.63447034e+00
-3.35192740e-01 -8.73544574e-01 -7.34608293e-01 5.12879789e-01
-1.20751548e+00 -4.40419704e-01 -3.53672653e-01 2.03456774e-01
5.41816413e-01 3.35191250e-01 5.63233972e-01 4.26666856e-01
-3.24431330e-01 2.20200866e-01 9.13845822e-02 7.58696571e-02
7.65902102e-01 -9.96823430e-01 6.00194693e-01 1.13381207e+00
1.16368257e-01 6.54283464e-01 1.02649522e+00 -3.24855506e-01
-1.36731112e+00 -8.25383425e-01 3.88851017e-01 -2.62152225e-01
5.80133736e-01 -3.06008756e-01 -1.04438889e+00 4.71711338e-01
3.05797875e-01 4.07692611e-01 2.61003256e-01 -4.27145749e-01
-1.20445706e-01 -3.08271591e-02 -9.07222152e-01 8.37845385e-01
1.25349009e+00 -4.03767020e-01 -3.82678300e-01 1.06298730e-01
7.67292202e-01 -4.15220797e-01 -4.62420881e-01 4.14275974e-01
7.89602458e-01 -1.21621549e+00 1.05616879e+00 -5.66143036e-01
5.42128980e-01 -6.40956640e-01 8.42732042e-02 -1.38654435e+00
-5.27987242e-01 -8.09301972e-01 -4.67381060e-01 6.15374506e-01
4.01115835e-01 -3.74298632e-01 6.45256460e-01 3.30232948e-01
1.44176066e-01 -6.97085261e-01 -6.09861374e-01 -9.57450986e-01
-2.26986021e-01 -1.07057579e-01 2.19471484e-01 8.42271388e-01
-4.21125144e-02 6.72692582e-02 -6.44754887e-01 2.85030454e-01
5.18994153e-01 1.06464610e-01 6.45950496e-01 -6.35541677e-01
-6.96439207e-01 -4.02458578e-01 -3.47617179e-01 -1.41721678e+00
-3.66348356e-01 -2.28856072e-01 -1.63547114e-01 -1.19747770e+00
-2.50195801e-01 2.19818398e-01 -4.77366857e-02 1.63894176e-01
-2.67100155e-01 4.07395780e-01 3.23128909e-01 1.12636574e-01
-1.73929751e-01 4.34830010e-01 1.65821505e+00 -1.45221666e-01
-1.50095552e-01 -1.35214403e-01 -3.95114124e-01 1.00107872e+00
6.91156805e-01 -2.23783031e-01 -6.00188553e-01 -7.10312486e-01
2.16853842e-01 9.33961198e-03 4.77741361e-01 -1.28526950e+00
3.33409844e-04 -5.54448441e-02 7.91295707e-01 -1.65612355e-01
3.86878878e-01 -6.89840496e-01 2.87003011e-01 7.50294864e-01
-3.18724364e-01 1.58885345e-01 3.28490734e-01 5.57069302e-01
-2.09580958e-01 -1.26783401e-01 1.01392210e+00 -2.78931618e-01
-6.47692680e-01 -7.78422598e-03 -5.99325657e-01 -1.80540353e-01
7.66517758e-01 -2.43221477e-01 -4.43455666e-01 -7.63224542e-01
-6.81399465e-01 -1.22010998e-01 8.21629882e-01 2.48939857e-01
6.91635907e-01 -1.31857729e+00 -3.27293158e-01 1.59486994e-01
-2.77067214e-01 -1.07655011e-01 1.23620495e-01 7.44748950e-01
-9.18027997e-01 1.79839686e-01 -1.33880049e-01 -6.20610356e-01
-1.14048421e+00 8.82838666e-01 4.10158485e-01 3.93189490e-02
-8.74185443e-01 5.90007663e-01 3.04408371e-01 3.56628805e-01
1.98022693e-01 -8.03521156e-01 -9.97442603e-02 -1.93373010e-01
9.10685539e-01 4.02454764e-01 -1.37042865e-01 -1.87128231e-01
-4.19185728e-01 8.15041780e-01 3.18177164e-01 -2.07386434e-01
1.19130778e+00 -2.26930976e-01 4.91041064e-01 3.25645149e-01
1.24889159e+00 2.88981676e-01 -1.78529787e+00 1.89252019e-01
-2.37840012e-01 -8.90459120e-01 -2.31170416e-01 -2.50565588e-01
-1.21617866e+00 9.99479592e-01 6.27741694e-01 4.72332627e-01
1.43279660e+00 -4.19582993e-01 8.60736907e-01 3.48627061e-01
2.71424294e-01 -8.18857968e-01 3.95469576e-01 3.47362876e-01
7.38738358e-01 -9.57750082e-01 -1.11582458e-01 -3.81077677e-01
-3.08032304e-01 1.48627281e+00 5.13910472e-01 -3.36375594e-01
4.44158822e-01 5.46614528e-01 1.43815726e-01 1.23037778e-01
-1.00346351e+00 -2.07500800e-01 2.89708287e-01 5.51801562e-01
5.57585776e-01 -3.53108019e-01 -6.90303445e-02 -8.13384876e-02
-1.39451325e-01 1.29444331e-01 7.87739813e-01 9.47467864e-01
-5.67114353e-01 -6.48746908e-01 -5.05839229e-01 1.87065154e-01
-6.21854782e-01 -1.81479137e-02 -1.78236276e-01 7.42666960e-01
4.70319279e-02 8.38217139e-01 2.02553436e-01 -5.25249302e-01
3.26651871e-01 -2.33090594e-01 7.25056708e-01 -1.63677499e-01
-2.71029860e-01 1.80664763e-01 2.00705335e-01 -7.92477310e-01
-5.63294649e-01 -2.26240948e-01 -1.02956867e+00 -3.67129415e-01
-1.18077487e-01 -3.04052532e-01 1.24826476e-01 5.50469339e-01
3.00096482e-01 6.83863342e-01 5.64409912e-01 -1.21918190e+00
-3.06748331e-01 -7.90909410e-01 -3.23033839e-01 6.52123034e-01
9.82324421e-01 -5.93196094e-01 -7.78080285e-01 4.46636021e-01] | [10.82657527923584, -1.4254724979400635] |
c2dbb11f-5158-4d77-8bab-ff5b45800382 | learning-to-parse-wireframes-in-images-of-man-1 | 2007.07527 | null | https://arxiv.org/abs/2007.07527v1 | https://arxiv.org/pdf/2007.07527v1.pdf | Learning to Parse Wireframes in Images of Man-Made Environments | In this paper, we propose a learning-based approach to the task of automatically extracting a "wireframe" representation for images of cluttered man-made environments. The wireframe (see Fig. 1) contains all salient straight lines and their junctions of the scene that encode efficiently and accurately large-scale geometry and object shapes. To this end, we have built a very large new dataset of over 5,000 images with wireframes thoroughly labelled by humans. We have proposed two convolutional neural networks that are suitable for extracting junctions and lines with large spatial support, respectively. The networks trained on our dataset have achieved significantly better performance than state-of-the-art methods for junction detection and line segment detection, respectively. We have conducted extensive experiments to evaluate quantitatively and qualitatively the wireframes obtained by our method, and have convincingly shown that effectively and efficiently parsing wireframes for images of man-made environments is a feasible goal within reach. Such wireframes could benefit many important visual tasks such as feature correspondence, 3D reconstruction, vision-based mapping, localization, and navigation. The data and source code are available at https://github.com/huangkuns/wireframe. | ['Shenghua Gao', 'Kun Huang', 'Tianjiao Ding', 'Zihan Zhou', 'Yi Ma', 'Yifan Wang'] | 2020-07-15 | learning-to-parse-wireframes-in-images-of-man | http://openaccess.thecvf.com/content_cvpr_2018/html/Huang_Learning_to_Parse_CVPR_2018_paper.html | http://openaccess.thecvf.com/content_cvpr_2018/papers/Huang_Learning_to_Parse_CVPR_2018_paper.pdf | cvpr-2018-6 | ['junction-detection', 'line-segment-detection'] | ['computer-vision', 'computer-vision'] | [-9.90286320e-02 1.01409733e-01 2.61073589e-01 -3.54377240e-01
-7.09675431e-01 -5.65559566e-01 6.34849131e-01 1.96431980e-01
-4.60752517e-01 1.95090011e-01 -1.62757814e-01 -3.76744688e-01
1.64081022e-01 -8.03771973e-01 -1.18479860e+00 7.58900680e-03
-2.65438706e-01 4.20675218e-01 5.30935526e-01 -3.97634745e-01
4.16177213e-01 7.58934438e-01 -1.51419544e+00 1.99961960e-02
4.92184430e-01 1.16093373e+00 4.42598701e-01 8.21405649e-01
-3.24016102e-02 4.28661197e-01 -1.25259161e-01 -3.52939487e-01
4.87686664e-01 -4.35986146e-02 -8.21312487e-01 1.29460454e-01
7.74424255e-01 -4.42781270e-01 -4.60352898e-01 9.15808260e-01
2.58734822e-01 1.02517940e-01 5.11274934e-01 -8.75332832e-01
-4.93395776e-01 1.10189475e-01 -7.22764194e-01 4.28033387e-03
4.88107800e-01 -1.38460472e-01 8.28584671e-01 -1.32723916e+00
9.00741756e-01 1.11396217e+00 9.46600497e-01 2.45339543e-01
-8.11986327e-01 -3.34750295e-01 -3.88215259e-02 1.21007815e-01
-1.49797976e+00 -5.32144785e-01 9.44294751e-01 -5.82301855e-01
1.10568273e+00 3.28186482e-01 8.93010497e-01 7.51861572e-01
9.97174159e-02 8.32260370e-01 6.08910084e-01 -8.89452457e-01
1.55423209e-02 -4.62413281e-02 -2.95450655e-03 1.28611076e+00
-2.43617576e-02 1.28180444e-01 -3.19728434e-01 1.76242903e-01
1.40147817e+00 -9.04376954e-02 -2.35662624e-01 -9.35410321e-01
-1.51568055e+00 6.50713623e-01 7.64791548e-01 1.69895038e-01
-1.08974114e-01 4.09804344e-01 2.35970169e-01 -1.05659448e-01
4.23589379e-01 3.86596680e-01 -2.43734792e-01 3.76085266e-02
-6.04349017e-01 1.75224513e-01 5.15157402e-01 1.33174551e+00
8.34320188e-01 -1.98249191e-01 4.16901439e-01 8.49876225e-01
-1.23667475e-02 6.54592752e-01 7.33880103e-02 -1.29095459e+00
4.83284593e-01 4.21870112e-01 3.68025154e-01 -1.30450296e+00
-7.52851844e-01 -2.09449589e-01 -7.42326796e-01 3.76342654e-01
4.97975320e-01 -2.78719533e-02 -9.53511238e-01 1.25056159e+00
4.99534845e-01 -8.18335414e-02 -2.58801579e-01 8.53939116e-01
1.00165927e+00 5.41275918e-01 -4.21913922e-01 4.82479066e-01
1.42915523e+00 -1.00283754e+00 -2.22848296e-01 -3.05690199e-01
6.25461161e-01 -1.00666797e+00 1.08069682e+00 1.47231460e-01
-1.18256521e+00 -7.50266850e-01 -1.02231371e+00 -4.92931157e-01
-3.60690087e-01 4.45793688e-01 7.71858633e-01 8.84705875e-03
-1.15742171e+00 4.66291904e-01 -7.89577961e-01 -6.14649832e-01
4.29703832e-01 5.64807542e-02 -6.69659317e-01 1.21091999e-01
-8.07614446e-01 8.37251723e-01 1.33589119e-01 5.05270958e-01
-6.19172513e-01 -3.27994019e-01 -1.34146082e+00 2.53456761e-03
3.65763694e-01 -7.95968056e-01 1.44440699e+00 -5.26089787e-01
-9.85250890e-01 1.31335044e+00 -1.97872698e-01 -4.16358978e-01
6.86423182e-01 -5.82543910e-01 1.05007596e-01 3.40970635e-01
1.87631488e-01 8.56567621e-01 5.24757266e-01 -1.40555274e+00
-7.41490245e-01 -2.60294527e-01 2.45596305e-01 1.31122172e-01
7.83542618e-02 -4.75840829e-02 -7.02513337e-01 -5.75994730e-01
3.68673742e-01 -8.23090374e-01 -4.16101754e-01 4.43006605e-01
-7.21658230e-01 -1.57149449e-01 5.12230814e-01 -7.23811746e-01
7.21300423e-01 -1.84094501e+00 -5.99057823e-02 2.50773460e-01
1.59447253e-01 1.14300661e-01 -8.17864984e-02 5.09234905e-01
1.20507590e-01 -3.72478506e-04 -1.38229147e-01 -4.70646113e-01
-1.46895930e-01 -1.00008279e-01 -3.15072715e-01 5.16683042e-01
-3.88714075e-02 9.98430669e-01 -7.20918894e-01 -4.06471699e-01
7.32615769e-01 4.77815628e-01 -2.62966037e-01 1.53322861e-01
-1.15563698e-01 3.37224364e-01 -4.50597763e-01 6.56561196e-01
5.91449440e-01 -1.87057778e-01 -3.77087742e-01 -4.48250175e-01
-4.48994517e-01 -2.04053313e-01 -9.54335630e-01 2.15147328e+00
-5.77742219e-01 1.17925823e+00 -1.77988723e-01 -6.81237519e-01
1.07596219e+00 -2.37884864e-01 3.67930353e-01 -9.40052688e-01
1.46141157e-01 2.20303729e-01 -4.02874917e-01 -6.17762327e-01
6.29064441e-01 6.03853703e-01 -1.16492800e-01 1.23937264e-01
-1.76055536e-01 -3.73500437e-01 1.21112809e-01 8.16411451e-02
9.36477900e-01 3.60713929e-01 3.66697371e-01 -3.00796509e-01
4.33081359e-01 3.52391183e-01 1.30382121e-01 6.79452121e-01
1.55143477e-02 9.35283899e-01 1.16799548e-01 -1.08227074e+00
-1.62113309e+00 -1.18093610e+00 -8.97460505e-02 6.19823933e-01
5.75997531e-01 -6.30676925e-01 -8.90606701e-01 -1.99869663e-01
-1.73033625e-01 3.00181061e-01 -6.67289495e-01 2.84834623e-01
-9.19320762e-01 1.11872621e-01 3.37718904e-01 8.52229357e-01
7.88330197e-01 -9.23150957e-01 -1.15475523e+00 -3.83986235e-02
-2.48164356e-01 -1.37217093e+00 -4.03253049e-01 9.26631689e-02
-4.31131870e-01 -1.33865166e+00 -7.90350080e-01 -1.14510691e+00
9.08430934e-01 5.60353637e-01 1.13801610e+00 1.50689617e-01
-8.70469689e-01 4.84819084e-01 -3.22010815e-01 -4.04284447e-01
1.07080616e-01 -1.80208057e-01 -2.41170377e-01 -3.85277987e-01
-7.74115277e-03 -4.35578436e-01 -8.13820601e-01 3.96888882e-01
-3.10485929e-01 5.14965117e-01 3.73696864e-01 5.76565146e-01
8.10023308e-01 -3.75948131e-01 -2.06976116e-01 -7.20020533e-01
2.42286250e-01 5.06874360e-02 -9.04481471e-01 2.46229991e-01
4.34866585e-02 -1.61325365e-01 4.88154262e-01 9.75775868e-02
-8.41107666e-01 3.88164341e-01 -4.78503883e-01 -3.36400211e-01
-4.66940790e-01 1.02999873e-01 2.40810499e-01 -2.67851144e-01
7.22904563e-01 5.99594712e-02 -3.37311476e-01 -4.35720623e-01
6.16470039e-01 4.53600019e-01 8.66247118e-01 -4.89498377e-01
9.49679375e-01 8.98635030e-01 5.39023988e-02 -1.02060378e+00
-7.59991944e-01 -5.83712339e-01 -1.15807247e+00 -2.79965937e-01
9.76270258e-01 -9.21736300e-01 -4.48725373e-01 5.59377551e-01
-1.54827034e+00 -5.70254624e-01 -1.65049419e-01 3.50995421e-01
-9.07326579e-01 3.32750082e-01 -6.35505021e-01 -5.83334088e-01
-3.50527465e-01 -9.70292687e-01 1.35540152e+00 2.08074749e-01
-2.38897637e-01 -1.00986803e+00 9.89987776e-02 4.09550518e-02
7.16109052e-02 7.43266642e-01 7.83168197e-01 -3.83323282e-02
-8.80915463e-01 -1.29499957e-01 -3.68421257e-01 7.51138339e-03
-7.29014426e-02 1.31202623e-01 -1.06272852e+00 -1.03384309e-01
-4.60958362e-01 -2.15974316e-01 7.75928497e-01 5.04413962e-01
1.27673912e+00 1.12148486e-01 -5.29341698e-01 1.06515539e+00
1.42498910e+00 2.82949507e-01 4.12357539e-01 7.41127133e-01
8.43228817e-01 6.84496284e-01 6.80098295e-01 3.40727061e-01
6.17605150e-01 6.54038668e-01 4.66545761e-01 -6.64893150e-01
-2.20022634e-01 -4.11474466e-01 -2.27579474e-01 4.58595932e-01
-3.44791651e-01 -1.50745451e-01 -1.20988667e+00 6.10451162e-01
-1.95455647e+00 -7.31726348e-01 -4.22066152e-01 1.96296203e+00
1.49120569e-01 2.31856897e-01 -4.62425388e-02 -7.02048987e-02
6.55368447e-01 2.84415390e-02 -4.24346507e-01 -3.07123780e-01
2.08648052e-02 -1.66262127e-02 6.63591743e-01 5.45858383e-01
-1.39445412e+00 1.08617389e+00 6.31755018e+00 5.04132688e-01
-8.41689229e-01 -3.06972235e-01 5.90205252e-01 4.59237069e-01
-1.89453393e-01 -8.36811587e-02 -5.67585528e-01 -2.37882882e-02
3.39270979e-01 7.69848153e-02 8.49593803e-02 1.05910993e+00
2.48272359e-01 -2.03300685e-01 -1.12202764e+00 1.30545187e+00
4.22125310e-02 -1.70571804e+00 -1.87322795e-01 -1.23727709e-01
6.11495256e-01 1.90690368e-01 -3.85364629e-02 -2.09475383e-01
2.43062943e-01 -9.27600980e-01 1.06126773e+00 5.98758161e-01
1.01125765e+00 -6.79676950e-01 4.90694761e-01 3.61118972e-01
-1.53318620e+00 1.63911730e-01 -5.47816455e-01 -3.18767652e-02
2.09182411e-01 5.05044580e-01 -7.00512946e-01 4.82219309e-01
9.83071625e-01 7.21065164e-01 -7.87321508e-01 1.35809851e+00
-2.99198568e-01 -1.93483550e-02 -4.54317123e-01 1.05872765e-01
3.06006312e-01 -1.86605841e-01 2.64887631e-01 1.25828969e+00
4.76133585e-01 1.26568694e-02 1.43460408e-01 7.91386604e-01
9.50111672e-02 8.60373303e-02 -1.04352534e+00 3.53235781e-01
4.97098863e-01 1.30958509e+00 -1.28070462e+00 -2.78235465e-01
-4.73260581e-01 1.01181924e+00 5.79317093e-01 3.15994501e-01
-7.68936574e-01 -7.86120832e-01 4.13433433e-01 3.02544504e-01
3.14029813e-01 -8.04812849e-01 -2.35200912e-01 -1.09171176e+00
3.49580437e-01 -1.36729956e-01 5.65187745e-02 -1.18436980e+00
-9.02461529e-01 9.90628779e-01 -1.15705756e-02 -1.31373680e+00
-2.84954160e-01 -7.38804519e-01 -6.51679754e-01 5.53692460e-01
-1.44427454e+00 -1.35787082e+00 -6.99430823e-01 7.02821553e-01
9.04192030e-01 1.09818153e-01 7.92837083e-01 5.71140647e-03
-2.44911402e-01 2.75270879e-01 1.96422845e-01 5.73436499e-01
3.51645768e-01 -1.21393001e+00 1.15807498e+00 9.66846347e-01
4.25782204e-01 6.61139488e-01 5.15737891e-01 -4.47845489e-01
-1.22625673e+00 -9.70807612e-01 5.76158762e-01 -5.07346928e-01
3.95334512e-01 -7.90525079e-01 -6.01916432e-01 8.05615306e-01
6.99150935e-02 1.48505360e-01 2.30822876e-01 -1.49693921e-01
-3.39656442e-01 1.13079764e-01 -8.72640073e-01 6.17490649e-01
1.57418680e+00 -4.87276256e-01 -4.60935801e-01 5.20670474e-01
6.92375302e-01 -8.72222900e-01 -5.30996144e-01 3.11878532e-01
6.58103645e-01 -1.30214238e+00 1.36766398e+00 -2.49300867e-01
5.25984645e-01 -2.29516834e-01 -6.87777996e-02 -1.23639262e+00
-2.33537063e-01 -4.86006439e-01 8.73170197e-02 8.55219364e-01
2.07039744e-01 -3.75534028e-01 8.13892782e-01 4.41873342e-01
-4.61428583e-01 -9.13721204e-01 -7.07156479e-01 -5.39279401e-01
-2.31423870e-01 -4.76004064e-01 3.45109046e-01 5.32302558e-01
-2.35363990e-01 1.33907631e-01 -2.98580289e-01 1.48665279e-01
7.29053974e-01 4.97869134e-01 1.10367668e+00 -1.07998848e+00
5.51195145e-02 -3.05784941e-01 -6.11605823e-01 -1.44832122e+00
3.04110926e-02 -6.33832514e-01 1.48147985e-01 -2.00327110e+00
-1.86713830e-01 -6.67843938e-01 3.15429628e-01 2.58953780e-01
2.22561970e-01 3.89565915e-01 2.03742728e-01 2.52142817e-01
-5.45798004e-01 3.57699513e-01 1.33228278e+00 6.84255511e-02
1.00262985e-01 -1.78653002e-01 -3.26569051e-01 1.32028210e+00
7.27399409e-01 8.36444944e-02 -2.83502907e-01 -8.72630954e-01
9.73736122e-02 -2.94695683e-02 6.39034092e-01 -1.05226052e+00
3.22282463e-01 -6.26603467e-03 8.84404123e-01 -8.45823348e-01
6.41993761e-01 -8.16537619e-01 -4.80535291e-02 3.46977681e-01
-2.72432685e-01 3.83902848e-01 2.24730790e-01 3.97301853e-01
-8.10400397e-02 -1.34033978e-01 5.97094476e-01 -5.50816655e-01
-1.33446658e+00 2.87018001e-01 -3.21334153e-02 -1.18067957e-01
1.20498037e+00 -3.51144463e-01 -3.16442609e-01 -3.26285481e-01
-6.73797131e-01 2.60161728e-01 6.30308807e-01 3.40471119e-01
1.24683249e+00 -1.21875560e+00 -4.77679461e-01 4.61170584e-01
2.54136443e-01 4.60673720e-01 2.30244830e-01 4.17889535e-01
-1.35367084e+00 5.35347164e-01 -4.30057108e-01 -8.23076367e-01
-1.21358287e+00 4.95062441e-01 3.89891833e-01 2.68719792e-01
-1.15579617e+00 1.03998649e+00 4.43543345e-01 -4.35246736e-01
2.83657253e-01 -6.66561365e-01 -8.24434403e-03 -4.12962019e-01
4.47755784e-01 3.15058708e-01 3.36235203e-03 -7.44797587e-01
-2.55784988e-01 1.31420195e+00 8.84878039e-02 1.38243496e-01
1.37310469e+00 -3.09780180e-01 1.09112680e-01 2.33753711e-01
1.27867138e+00 5.19702807e-02 -1.33708036e+00 -1.58097565e-01
-1.42872557e-01 -7.54396915e-01 -3.07891965e-01 -3.35241020e-01
-7.67693818e-01 1.05293500e+00 4.57376659e-01 1.83082610e-01
7.79181063e-01 2.15536639e-01 1.00262868e+00 6.52057707e-01
8.58252585e-01 -8.47720444e-01 1.19166769e-01 4.96086687e-01
1.00856662e+00 -1.26051486e+00 -2.82647580e-01 -7.50609219e-01
-2.43840069e-01 1.53880990e+00 6.98485374e-01 -4.30704236e-01
4.86474723e-01 3.79770339e-01 1.89887077e-01 -4.51505482e-01
-2.36096352e-01 -3.04241866e-01 3.36463302e-01 8.62705767e-01
9.87707153e-02 6.58484455e-03 2.72640169e-01 7.61451852e-03
-5.16277432e-01 -3.98744255e-01 4.76851791e-01 9.39018488e-01
-7.30428398e-01 -7.54843354e-01 -4.80161428e-01 5.47185652e-02
7.31497183e-02 4.21920717e-02 -4.17167395e-01 1.12097299e+00
9.11261439e-02 6.69226885e-01 2.07086936e-01 -3.64377320e-01
6.35074556e-01 -4.34207439e-01 6.48104370e-01 -4.74036396e-01
-1.93137266e-02 6.75405338e-02 2.78514415e-01 -9.47982073e-01
-3.31635475e-01 -4.00538594e-01 -1.43548357e+00 -4.67144363e-02
-6.75735772e-02 -4.12039682e-02 7.59355724e-01 6.50134504e-01
2.33147740e-01 3.04375917e-01 4.11068708e-01 -1.25387609e+00
-9.63163301e-02 -5.17317772e-01 -2.92149365e-01 4.09584254e-01
4.77602959e-01 -6.69205129e-01 -4.16365825e-02 1.92855299e-01] | [8.218770980834961, -1.8900768756866455] |
42af29f1-be81-464e-8423-b71b3693f84d | ccharmony-color-checker-based-image | 2206.00800 | null | https://arxiv.org/abs/2206.00800v1 | https://arxiv.org/pdf/2206.00800v1.pdf | CcHarmony: Color-checker based Image Harmonization Dataset | Image harmonization targets at adjusting the foreground in a composite image to make it compatible with the background, producing a more realistic and harmonious image. Training deep image harmonization network requires abundant training data, but it is extremely difficult to acquire training pairs of composite images and ground-truth harmonious images. Therefore, existing works turn to adjust the foreground appearance in a real image to create a synthetic composite image. However, such adjustment may not faithfully reflect the natural illumination change of foreground. In this work, we explore a novel transitive way to construct image harmonization dataset. Specifically, based on the existing datasets with recorded illumination information, we first convert the foreground in a real image to the standard illumination condition, and then convert it to another illumination condition, which is combined with the original background to form a synthetic composite image. In this manner, we construct an image harmonization dataset called ccHarmony, which is named after color checker (cc). The dataset is available at https://github.com/bcmi/Image-Harmonization-Dataset-ccHarmony. | ['Li Niu', 'Haoxu Huang'] | 2022-06-01 | null | null | null | null | ['image-harmonization'] | ['computer-vision'] | [ 4.56769079e-01 -3.76517117e-01 2.70217597e-01 -1.41126364e-01
-2.07683548e-01 -5.03304124e-01 3.93717051e-01 -4.06175077e-01
-2.04118490e-01 6.32540047e-01 9.40805208e-03 -4.18526754e-02
3.39412779e-01 -9.53540623e-01 -6.82159424e-01 -1.03137374e+00
8.87739897e-01 -1.14868261e-01 1.17599852e-01 -2.21869648e-01
1.38621509e-01 2.06182495e-01 -1.42084193e+00 1.68673381e-01
1.14452326e+00 7.99855173e-01 3.57706130e-01 4.30558950e-01
-1.66687220e-01 4.29008752e-01 -8.67387891e-01 -3.59150916e-01
5.56942642e-01 -1.01515043e+00 -3.89763713e-01 3.23137701e-01
3.99671555e-01 -2.12998658e-01 -3.11425656e-01 1.45518374e+00
6.39906645e-01 9.75018516e-02 2.11756885e-01 -1.47569311e+00
-1.02836990e+00 4.38189268e-01 -8.59776258e-01 -1.49260908e-01
-5.81882000e-02 3.92684489e-01 3.93523335e-01 -4.85608608e-01
5.15006542e-01 1.14120209e+00 3.66554201e-01 5.39753497e-01
-1.27441764e+00 -9.62058365e-01 -1.16978057e-01 1.89109996e-01
-1.21904910e+00 -3.76577556e-01 1.10318494e+00 3.04025151e-02
-1.15376569e-01 5.10131180e-01 1.14061141e+00 8.21549296e-01
1.40069753e-01 6.82297945e-01 1.46463692e+00 -4.41858560e-01
3.62980412e-03 3.95980524e-03 -3.88364613e-01 3.24366301e-01
3.54355812e-01 6.14859276e-02 -1.80300400e-01 3.57462615e-01
7.44404674e-01 2.67480426e-02 -7.39456296e-01 -2.99345821e-01
-1.32245457e+00 2.95470148e-01 6.58066571e-01 4.16122198e-01
-2.25201875e-01 -1.48611456e-01 1.14713527e-01 2.11295679e-01
1.07864812e-01 2.48455837e-01 8.91224295e-02 1.54117942e-01
-7.59347022e-01 9.01456773e-02 4.47621286e-01 5.74683070e-01
9.64682579e-01 2.77451783e-01 -1.02588393e-01 9.74894226e-01
1.29474506e-01 7.12645233e-01 5.59347630e-01 -1.09691513e+00
2.97238111e-01 8.08256209e-01 5.92796952e-02 -1.25825334e+00
-1.11352041e-01 -4.80811775e-01 -1.29713905e+00 4.28524792e-01
2.77741164e-01 -1.36235341e-01 -7.70285606e-01 1.75278497e+00
5.53105116e-01 1.94901586e-01 2.35709697e-01 1.16395450e+00
9.32139397e-01 7.95614362e-01 -2.58491695e-01 -2.98135608e-01
1.08966422e+00 -9.77833450e-01 -7.44306326e-01 -1.20054320e-01
-1.04504511e-01 -1.19230437e+00 1.28618169e+00 5.09859800e-01
-1.10060132e+00 -8.80861878e-01 -1.23976099e+00 5.21305315e-02
-1.57850832e-01 7.77925327e-02 3.42097968e-01 5.55716336e-01
-9.85246599e-01 1.49492979e-01 -3.89049560e-01 -1.01020195e-01
2.11094588e-01 -1.34033740e-01 -3.52975011e-01 -8.53322670e-02
-1.22266924e+00 6.35383010e-01 6.51422441e-01 4.44012880e-01
-7.30347514e-01 -6.39437139e-01 -7.06769586e-01 -2.64877081e-01
3.10990047e-02 -5.80414116e-01 1.04871833e+00 -1.62098527e+00
-1.61434865e+00 9.59509432e-01 1.08911522e-01 3.67913581e-02
6.85402751e-01 1.88223556e-01 -6.57926738e-01 -9.12428126e-02
-1.07229739e-01 6.34346664e-01 9.02577102e-01 -1.98294473e+00
-4.14397955e-01 -1.87562376e-01 2.67615337e-02 3.21573675e-01
-1.06311791e-01 -1.83189809e-01 -9.43300784e-01 -6.94706082e-01
2.08307840e-02 -9.05932367e-01 -1.02986151e-03 -2.01262951e-01
-8.15336287e-01 5.30831575e-01 7.93699265e-01 -6.75057232e-01
1.16011345e+00 -2.17879844e+00 -5.88594424e-03 3.36062819e-01
2.24156484e-01 3.98065686e-01 -4.35929924e-01 5.80552742e-02
-3.70685488e-01 -2.94756424e-02 -5.18584192e-01 -7.30115995e-02
-2.25125849e-01 1.25362054e-01 -2.77094375e-02 4.44564879e-01
5.69438711e-02 8.57208848e-01 -8.49619269e-01 -7.45841801e-01
2.81003028e-01 5.17104387e-01 -4.39661115e-01 3.98665488e-01
-6.54945597e-02 7.94632018e-01 -1.43197104e-01 5.42887509e-01
1.22887242e+00 1.12003215e-01 1.75243452e-01 -6.80171728e-01
-2.41980314e-01 -5.04462659e-01 -1.28498530e+00 1.36158526e+00
-3.20025772e-01 7.05923498e-01 1.68582171e-01 -9.33642089e-01
1.19789100e+00 7.11312424e-03 5.87872386e-01 -1.24564052e+00
3.18984836e-01 9.06754807e-02 1.08081385e-01 -2.87225574e-01
4.43718433e-01 -1.06473990e-01 2.06289992e-01 4.15882915e-01
-5.78746319e-01 -5.85482061e-01 2.73014486e-01 -1.22125462e-01
3.96551281e-01 1.88097060e-01 -2.14947052e-02 -7.83244963e-04
5.99526644e-01 3.04395929e-02 8.88227344e-01 3.84255171e-01
-7.18813539e-02 9.95148659e-01 4.13435221e-01 -5.11569917e-01
-1.23622596e+00 -1.26355839e+00 1.40780369e-02 3.95966262e-01
7.89437413e-01 -3.08994744e-02 -1.12462842e+00 2.47397833e-02
-3.68449152e-01 4.94357765e-01 -5.32373548e-01 -4.43132281e-01
-6.37413323e-01 -9.61550832e-01 4.39058214e-01 -3.48138921e-02
1.44338191e+00 -1.24287438e+00 -4.28271621e-01 6.05347054e-03
-5.09354770e-01 -6.44942045e-01 -6.79606318e-01 -3.98333371e-01
-2.10400268e-01 -1.23952055e+00 -7.93295741e-01 -8.96907628e-01
7.96881378e-01 5.12019813e-01 1.10701478e+00 3.00125152e-01
-3.62255335e-01 -2.07528159e-01 -2.49547705e-01 -3.68072331e-01
-6.23082161e-01 -4.64569837e-01 -1.90106899e-01 4.19169962e-01
-9.33160037e-02 -4.78329450e-01 -9.97416019e-01 4.70092207e-01
-1.43399537e+00 6.68196440e-01 5.71964443e-01 7.80679941e-01
9.21027303e-01 4.98066127e-01 1.69100627e-01 -6.87953889e-01
6.55281425e-01 -1.54384181e-01 -5.92360497e-01 4.68742460e-01
-3.60661149e-01 -3.85720879e-01 6.73409581e-01 -5.48511028e-01
-1.46368945e+00 -3.85889858e-02 5.36181666e-02 -6.00138247e-01
9.63765830e-02 3.04087460e-01 -5.87604344e-01 -8.46274346e-02
5.08251190e-01 4.90273356e-01 1.01279743e-01 -1.40863672e-01
4.30909693e-01 6.28343284e-01 1.09775507e+00 -3.51783305e-01
9.93196785e-01 4.66036916e-01 -2.95229077e-01 -4.80573952e-01
-5.80989778e-01 1.50694862e-01 -4.45213079e-01 -3.21805298e-01
9.68880951e-01 -7.01689422e-01 -4.23502445e-01 1.02014220e+00
-1.07463360e+00 -7.32279837e-01 -1.46161079e-01 2.23831981e-01
-2.89803356e-01 3.51962715e-01 -3.35329086e-01 -2.74494588e-01
-3.66668135e-01 -1.11174428e+00 8.17838430e-01 8.18201125e-01
1.72602341e-01 -6.27020121e-01 2.55458921e-01 3.56833041e-01
5.48490465e-01 6.21339023e-01 7.15624094e-01 2.64417857e-01
-4.28212464e-01 -3.63958813e-02 -3.53152514e-01 4.93291616e-01
4.97302443e-01 6.88278913e-01 -5.98609626e-01 -1.16217203e-01
4.03000042e-02 4.69017141e-02 5.26601732e-01 2.46514603e-01
1.24956000e+00 -2.31455535e-01 1.69131175e-01 9.23599422e-01
1.50105882e+00 5.82024753e-01 1.13834393e+00 5.75124860e-01
7.77185500e-01 4.21730965e-01 6.20251060e-01 2.48122811e-01
3.41549903e-01 7.19661951e-01 3.59053820e-01 -8.46890152e-01
-4.38689053e-01 -3.49814296e-01 3.38375941e-03 7.99135208e-01
2.40227934e-02 -3.60878885e-01 -7.82235861e-01 2.46887520e-01
-1.55693591e+00 -9.68730390e-01 -5.31786419e-02 2.02004743e+00
1.11221075e+00 -1.70247138e-01 -1.86638564e-01 3.23684700e-02
1.18266606e+00 9.27405208e-02 -6.71910346e-01 -1.81631938e-01
-5.61787009e-01 8.59913975e-03 1.65565580e-01 2.98715025e-01
-9.34680104e-01 7.60318577e-01 5.22291374e+00 8.06965351e-01
-1.65614998e+00 1.45514514e-02 9.42761779e-01 -9.15248245e-02
-6.28164411e-01 -5.37038222e-02 -6.53021857e-02 9.02473986e-01
3.05010289e-01 -4.16413307e-01 7.39841521e-01 3.23653340e-01
4.43290263e-01 -3.77001688e-02 -3.57400268e-01 1.36679769e+00
7.29866177e-02 -1.09563744e+00 2.94678628e-01 -3.17702383e-01
1.12547410e+00 -3.41215998e-01 1.30218357e-01 -9.43413675e-02
3.26926440e-01 -8.59756291e-01 7.14849710e-01 6.67447031e-01
9.34360027e-01 -6.46994591e-01 7.03398943e-01 -1.48931900e-02
-1.26871538e+00 2.08640546e-01 -2.84702599e-01 4.14625406e-01
9.98007655e-02 4.80807841e-01 -3.13538350e-02 6.83599889e-01
6.93971992e-01 5.28159261e-01 -7.26672232e-01 1.13804245e+00
-2.44680479e-01 2.82751828e-01 -6.29650727e-02 4.30465519e-01
-9.91465896e-02 -7.85792470e-01 2.46786430e-01 7.82401741e-01
3.93164188e-01 4.31670621e-02 5.36463112e-02 1.02047896e+00
-2.09006399e-01 1.20558217e-01 -3.63154799e-01 8.78051743e-02
5.89565635e-01 1.34643364e+00 -7.02118218e-01 -3.08001846e-01
-2.62591302e-01 1.21462727e+00 -1.17068604e-01 5.94039857e-01
-1.13502955e+00 -4.03498769e-01 3.60035837e-01 -2.79218405e-01
-2.70669609e-01 1.65149570e-01 -2.86295801e-01 -1.25410485e+00
6.08781613e-02 -1.31543481e+00 1.04548350e-01 -1.18415332e+00
-1.00758862e+00 7.14875519e-01 -8.67397711e-02 -1.50349438e+00
2.41415739e-01 -9.49569270e-02 -1.14560461e+00 9.66819227e-01
-1.19033301e+00 -1.24538088e+00 -1.19940484e+00 8.02208662e-01
1.14431262e-01 -3.19359377e-02 4.06177908e-01 4.94679719e-01
-9.47028816e-01 5.35806239e-01 2.20847353e-01 9.60417688e-02
8.73292923e-01 -9.65992928e-01 2.88623512e-01 1.02538681e+00
-8.15952793e-02 4.63057369e-01 7.46013641e-01 -4.90185082e-01
-1.24576271e+00 -1.07786918e+00 8.87266323e-02 6.22957200e-02
2.87279516e-01 3.22190608e-04 -1.01535976e+00 2.24005416e-01
5.54004729e-01 -9.01143700e-02 3.32290411e-01 -7.50224233e-01
-1.83189437e-01 -5.64812601e-01 -1.10420001e+00 1.03599918e+00
8.26124847e-01 -2.89730430e-01 -3.09938133e-01 1.02110490e-01
7.43806779e-01 -5.77553511e-01 -7.37836063e-01 4.86991227e-01
6.69746995e-01 -1.17693841e+00 8.66424680e-01 1.08984264e-03
6.92747533e-01 -8.98885310e-01 4.26408276e-02 -1.55993378e+00
-1.22901984e-01 -4.95535403e-01 5.77111006e-01 1.36435091e+00
1.38674349e-01 -8.25345576e-01 5.24172127e-01 4.39341694e-01
-1.21808827e-01 -5.17731309e-01 -4.88388121e-01 -4.38932389e-01
-1.48732252e-02 9.03308317e-02 1.03710544e+00 1.10079968e+00
-5.22760451e-01 -4.74865213e-02 -3.70858043e-01 4.83820178e-02
6.11944616e-01 5.58106899e-01 1.17597651e+00 -6.79755807e-01
-1.56273991e-01 -6.09295607e-01 -1.34096533e-01 -3.92269492e-01
4.73450217e-03 -4.78802443e-01 9.15716887e-02 -1.55970955e+00
3.23006392e-01 -5.16963005e-01 -2.56653100e-01 3.90806794e-01
-5.15444398e-01 9.19291079e-01 3.46086115e-01 3.02645594e-01
-2.57720709e-01 7.04526603e-01 1.82333803e+00 -3.83330226e-01
-3.01837087e-01 -3.29180926e-01 -8.44312906e-01 4.68709409e-01
1.33200502e+00 2.71340795e-02 -4.56031889e-01 -5.73767781e-01
-1.15258090e-01 -8.77757818e-02 3.38970482e-01 -9.88615036e-01
7.46456683e-02 -4.29280698e-01 5.99880993e-01 -4.06389892e-01
4.70802672e-02 -7.65006244e-01 1.01817989e+00 6.31843269e-01
-9.61681753e-02 1.30914539e-01 3.04436654e-01 1.44613460e-01
-5.62168479e-01 -7.16677606e-02 1.15196824e+00 -1.42153800e-01
-8.43512774e-01 3.07092488e-01 2.26752505e-01 7.26979077e-02
1.09749305e+00 -4.40056950e-01 -6.37574434e-01 -3.66750240e-01
-1.75432950e-01 1.13494702e-01 1.04195726e+00 4.33260053e-01
5.77782035e-01 -1.60932648e+00 -7.84010828e-01 3.81846696e-01
6.56104321e-03 -8.28786846e-03 6.21710062e-01 6.79988384e-01
-1.10150182e+00 -3.33961993e-01 -7.55327404e-01 -2.86665320e-01
-1.30395448e+00 6.02321982e-01 5.90230405e-01 2.00486511e-01
-4.78548050e-01 3.44687939e-01 5.17395318e-01 -3.81414115e-01
-1.03630722e-01 8.05964880e-03 -1.83293551e-01 -1.30107746e-01
5.27245939e-01 7.11779902e-03 -1.77246273e-01 -6.95875287e-01
-4.89064269e-02 8.20375621e-01 4.19211686e-01 4.65969592e-02
1.03027689e+00 -3.46332788e-01 -6.05464756e-01 1.21577546e-01
1.18656409e+00 2.33705699e-01 -1.25089777e+00 -1.19286738e-01
-5.69685996e-01 -7.49091446e-01 -1.35127455e-01 -6.23624206e-01
-1.74591005e+00 5.89076817e-01 9.07516301e-01 2.11537272e-01
1.69438672e+00 -5.10754585e-01 8.43157887e-01 7.24608302e-02
6.29238263e-02 -9.49773192e-01 1.36621833e-01 1.27126751e-02
1.12004256e+00 -1.11329281e+00 -6.96853846e-02 -1.44172654e-01
-8.05886507e-01 9.36174393e-01 8.76040041e-01 1.69389527e-02
2.65353411e-01 1.33557990e-01 5.67257404e-01 -5.67421429e-02
-1.85214549e-01 -7.55224377e-02 1.61889255e-01 6.01409018e-01
1.46301508e-01 7.89929032e-02 -1.93627745e-01 1.31382853e-01
-4.38230246e-01 -1.58297747e-01 3.89000237e-01 4.05686527e-01
-1.74988911e-01 -1.08197439e+00 -6.24007404e-01 1.32662550e-01
-1.92479745e-01 -1.07912518e-01 -5.05400121e-01 7.43200660e-01
5.33196449e-01 9.86298740e-01 1.87719315e-01 -4.26473290e-01
3.68799210e-01 -4.99435574e-01 2.62278289e-01 -1.68619454e-01
-3.01670134e-01 4.36527990e-02 -3.24725062e-01 -5.16841829e-01
-7.01242507e-01 -1.98740140e-01 -1.08255541e+00 -5.55932403e-01
8.27871263e-02 7.55524561e-02 4.50818002e-01 4.97804463e-01
1.28546590e-03 5.98509490e-01 8.34041119e-01 -1.04256403e+00
1.48231968e-01 -6.28413081e-01 -5.57774425e-01 8.07124078e-01
1.81736723e-01 -4.05348271e-01 -2.73307502e-01 3.12299222e-01] | [11.256548881530762, -1.2056082487106323] |
b68574f3-7dc9-49d7-9f3a-6424b25b3d5f | bloomberggpt-a-large-language-model-for | 2303.17564 | null | https://arxiv.org/abs/2303.17564v2 | https://arxiv.org/pdf/2303.17564v2.pdf | BloombergGPT: A Large Language Model for Finance | The use of NLP in the realm of financial technology is broad and complex, with applications ranging from sentiment analysis and named entity recognition to question answering. Large Language Models (LLMs) have been shown to be effective on a variety of tasks; however, no LLM specialized for the financial domain has been reported in literature. In this work, we present BloombergGPT, a 50 billion parameter language model that is trained on a wide range of financial data. We construct a 363 billion token dataset based on Bloomberg's extensive data sources, perhaps the largest domain-specific dataset yet, augmented with 345 billion tokens from general purpose datasets. We validate BloombergGPT on standard LLM benchmarks, open financial benchmarks, and a suite of internal benchmarks that most accurately reflect our intended usage. Our mixed dataset training leads to a model that outperforms existing models on financial tasks by significant margins without sacrificing performance on general LLM benchmarks. Additionally, we explain our modeling choices, training process, and evaluation methodology. We release Training Chronicles (Appendix C) detailing our experience in training BloombergGPT. | ['Gideon Mann', 'David Rosenberg', 'Prabhanjan Kambadur', 'Sebastian Gehrmann', 'Mark Dredze', 'Vadim Dabravolski', 'Steven Lu', 'Ozan Irsoy', 'Shijie Wu'] | 2023-03-30 | null | null | null | null | ['multi-task-language-understanding', 'movie-recommendation', 'sports-understanding', 'multiple-choice-qa', 'sarcasm-detection', 'reading-comprehension', 'disambiguation-q', 'ruin-names', 'snarks', 'formal-fallacies-syllogisms-negation', 'hyperbaton', 'penguins-in-a-table', 'temporal-sequences', 'logical-reasoning', 'common-sense-reasoning', 'reasoning-about-colored-objects', 'causal-judgment', 'date-understanding'] | ['methodology', 'miscellaneous', 'miscellaneous', 'natural-language-processing', 'natural-language-processing', 'natural-language-processing', 'natural-language-processing', 'natural-language-processing', 'natural-language-processing', 'natural-language-processing', 'natural-language-processing', 'reasoning', 'reasoning', 'reasoning', 'reasoning', 'reasoning', 'reasoning', 'reasoning'] | [-6.57735825e-01 -8.88952613e-02 -5.64925849e-01 -5.38596332e-01
-1.06723726e+00 -1.01167607e+00 8.03085744e-01 3.76304328e-01
-6.13521338e-01 6.76384270e-01 5.05244553e-01 -7.78989553e-01
1.17673308e-01 -7.27637053e-01 -5.86949468e-01 7.65339509e-02
-1.87977865e-01 7.25002050e-01 -8.27194303e-02 -9.41867605e-02
5.38940251e-01 3.47185671e-01 -9.29163396e-01 7.64258385e-01
3.72047216e-01 1.24577022e+00 -3.49839389e-01 4.37823802e-01
-8.40954244e-01 1.46512294e+00 -7.58261681e-01 -7.96146929e-01
4.77526933e-01 3.99255365e-01 -1.06338120e+00 -2.68414676e-01
3.77262384e-01 -1.69827744e-01 1.15997069e-01 5.13043165e-01
4.58549678e-01 1.01292599e-02 6.06771290e-01 -1.00820994e+00
-9.63101387e-01 1.16251814e+00 -5.38133025e-01 2.74758995e-01
4.03625339e-01 -1.33354589e-02 1.46413457e+00 -1.30579686e+00
5.28563976e-01 1.30308950e+00 8.43024790e-01 1.96672961e-01
-8.24393153e-01 -7.89418399e-01 2.07548991e-01 -2.79035300e-01
-1.02132726e+00 -2.67102540e-01 2.86689520e-01 -4.85675961e-01
1.81019878e+00 -1.13730496e-02 5.80731332e-01 1.00661051e+00
4.85642076e-01 1.12653041e+00 1.03675902e+00 -3.95912170e-01
1.31413653e-01 4.07904595e-01 4.80778128e-01 3.88247669e-01
3.72047454e-01 -2.06515938e-01 -8.95724297e-01 -5.25090396e-01
3.18176955e-01 -4.82324287e-02 2.78536826e-01 9.52206105e-02
-1.16704047e+00 1.14840484e+00 -1.43286914e-01 3.82940412e-01
-2.67350256e-01 1.59332678e-01 5.99566460e-01 5.79030752e-01
8.45390499e-01 1.03908873e+00 -1.25443554e+00 -3.34346682e-01
-1.06514883e+00 6.13988161e-01 1.59354424e+00 8.69240582e-01
5.56803703e-01 6.27729222e-02 1.46114612e-02 8.18539500e-01
4.27274853e-01 2.14977846e-01 9.52940285e-01 -9.22201037e-01
9.02859807e-01 8.10004652e-01 1.84455618e-01 -6.19527340e-01
-4.25478846e-01 -3.37112218e-01 -2.35766977e-01 -1.56082138e-01
5.30805409e-01 -4.24806178e-01 -5.38945496e-01 1.23113847e+00
-1.38843030e-01 -6.67968094e-02 4.00521219e-01 3.39391142e-01
9.68824685e-01 7.63822973e-01 2.65641928e-01 2.34811842e-01
1.41711783e+00 -9.05170679e-01 -1.68642744e-01 -6.85105443e-01
1.16417897e+00 -1.16486979e+00 1.01834917e+00 5.34860015e-01
-1.06290209e+00 1.49600049e-02 -6.49146318e-01 -1.53985411e-01
-8.09712052e-01 -2.06150964e-01 1.26274872e+00 6.97764277e-01
-9.61108446e-01 3.86235684e-01 -5.28346539e-01 -2.47547626e-01
2.66185284e-01 1.50302902e-01 -2.15859979e-01 1.08827159e-01
-1.35324764e+00 9.27422881e-01 4.72953737e-01 -4.67532158e-01
-4.31724995e-01 -1.10401440e+00 -7.97892511e-01 3.34652402e-02
1.95250645e-01 -5.01083255e-01 1.68118465e+00 -4.38208461e-01
-1.31848037e+00 1.10301089e+00 6.40448555e-02 -1.01429689e+00
4.77126926e-01 -5.90900898e-01 -4.14063931e-01 -2.67883480e-01
3.14942896e-01 4.32735562e-01 2.12776378e-01 -5.54769516e-01
-4.65273499e-01 9.76726785e-02 1.09910285e-02 -4.40274402e-02
-6.59138680e-01 5.94571233e-01 -3.55061203e-01 -8.36613834e-01
-3.10952038e-01 -5.05319476e-01 -3.00092489e-01 -7.79280186e-01
-3.93231511e-01 -4.13788259e-01 4.66609418e-01 -4.97593790e-01
1.30682302e+00 -1.86868978e+00 -8.14499319e-01 -1.22996615e-02
1.16183944e-01 1.48121104e-01 -1.94155097e-01 7.92864442e-01
-2.49449357e-01 7.65276551e-01 1.46519423e-01 -7.42794871e-01
3.31545353e-01 6.18451536e-02 -6.54734910e-01 -7.74909109e-02
1.11116730e-01 1.21363473e+00 -5.50848842e-01 -5.73477924e-01
2.95611210e-02 -7.53225526e-03 -6.95996821e-01 8.22720826e-02
-3.83635700e-01 -2.30097324e-01 -5.94899237e-01 1.08550346e+00
3.45383435e-01 -6.71617687e-01 1.29571063e-02 1.99604169e-01
-2.17144772e-01 8.63377035e-01 -9.64571953e-01 1.52959108e+00
-5.15122354e-01 3.72164160e-01 -2.96650857e-01 -7.59712398e-01
9.76192892e-01 3.58081847e-01 5.82712293e-01 -5.87737203e-01
-4.78463434e-02 5.38525939e-01 -3.24513614e-01 -3.55092347e-01
8.33473563e-01 -2.89013505e-01 -5.44384539e-01 9.45952177e-01
1.16985608e-02 -1.34084418e-01 4.11342442e-01 3.20635647e-01
1.16051674e+00 -3.51921618e-02 4.79664028e-01 -2.60553807e-01
7.60994479e-02 1.26368210e-01 5.12895346e-01 9.04932857e-01
1.02485819e-02 3.63969564e-01 5.41407406e-01 -6.49152815e-01
-8.51550996e-01 -8.39632034e-01 -3.05444479e-01 1.16757655e+00
-6.07312799e-01 -6.02166831e-01 -3.46924901e-01 -6.44041836e-01
6.09930515e-01 6.71206474e-01 -3.60428363e-01 4.15887743e-01
-5.39513826e-01 -1.41756630e+00 7.93283761e-01 5.56681931e-01
4.91480708e-01 -1.32044542e+00 -4.38447803e-01 4.83595669e-01
7.95444548e-02 -1.43807673e+00 -2.94335663e-01 2.79441029e-01
-8.88749480e-01 -1.09797168e+00 -4.65957612e-01 -6.78417683e-01
-3.35021690e-02 -2.21879318e-01 1.90737724e+00 -3.89162600e-01
-1.98805869e-01 3.49551678e-01 -2.59918749e-01 -6.40254259e-01
-4.78363574e-01 5.68775833e-01 -1.57768399e-01 -4.08740908e-01
9.26429510e-01 -2.05514014e-01 -2.44618967e-01 -1.37301266e-01
-8.35618913e-01 -2.45131165e-01 7.31087267e-01 4.64448720e-01
2.10855037e-01 -1.88747898e-01 9.18274999e-01 -1.38765323e+00
8.18353832e-01 -8.79459500e-01 -5.66075683e-01 3.43266398e-01
-8.37991953e-01 -1.66169796e-02 5.50616026e-01 -3.00918996e-01
-8.59491765e-01 -5.44984579e-01 -1.83039650e-01 3.86331528e-02
4.50015105e-02 1.13366473e+00 1.03202373e-01 5.43620484e-03
5.20800233e-01 -1.64335415e-01 -4.11069125e-01 -8.46316755e-01
2.59439260e-01 5.38855374e-01 2.50640601e-01 -9.66757357e-01
4.07366812e-01 1.37266561e-01 -3.29554915e-01 -4.62277025e-01
-1.25033975e+00 -6.37555420e-01 -3.41884941e-01 3.19582105e-01
4.01943088e-01 -9.77789581e-01 -6.04135752e-01 7.31956244e-01
-9.16607380e-01 -4.69549835e-01 -4.33981389e-01 5.93351603e-01
-1.73017040e-01 3.59863967e-01 -1.32505190e+00 -6.45193875e-01
-7.04449892e-01 -6.75062656e-01 8.54601800e-01 -8.06524977e-02
-4.59098846e-01 -1.59329534e+00 2.04394400e-01 4.90446061e-01
4.46206868e-01 2.74031639e-01 9.89626050e-01 -1.25681353e+00
-3.15496176e-01 -5.48446298e-01 -1.69366688e-01 5.00697494e-01
4.60495912e-02 -1.43205971e-01 -8.39517117e-01 -1.52307868e-01
2.05592699e-02 -7.66444147e-01 1.08084333e+00 2.66995847e-01
8.80482435e-01 -3.43892217e-01 -1.12451695e-01 4.11127597e-01
1.44261229e+00 -8.78504738e-02 2.25311413e-01 1.00323212e+00
4.63957727e-01 6.14775956e-01 5.26364267e-01 7.35528886e-01
6.86456144e-01 -3.37693617e-02 8.26144814e-02 5.02091739e-03
5.11921823e-01 -2.43360057e-01 7.36083210e-01 7.73096502e-01
3.67586702e-01 -4.00436163e-01 -1.32220185e+00 6.54574335e-01
-1.62656820e+00 -6.46851778e-01 3.69495824e-02 1.90968072e+00
1.19567466e+00 5.23628294e-01 -1.39553994e-01 -2.89388299e-01
5.90513162e-02 3.20034474e-01 -5.35504997e-01 -5.04997849e-01
-4.62642998e-01 4.81690377e-01 8.66918921e-01 3.72573435e-01
-1.19910264e+00 1.09086621e+00 7.77911663e+00 5.72698414e-01
-8.89756680e-01 5.33341318e-02 8.87106001e-01 -5.01750827e-01
-4.01616961e-01 1.62188008e-01 -1.43360877e+00 3.47867966e-01
1.39652872e+00 -4.59221691e-01 -1.60568759e-01 9.69576955e-01
1.10186480e-01 8.33267793e-02 -1.21511817e+00 6.80092931e-01
-3.07139065e-02 -1.52403426e+00 3.04671705e-01 2.49926493e-01
8.55672717e-01 4.97866780e-01 1.73755303e-01 7.72143424e-01
8.19193661e-01 -1.21440780e+00 8.87254238e-01 3.78659397e-01
3.37639570e-01 -6.81461930e-01 9.80541766e-01 2.95795768e-01
-1.09830499e+00 -1.21353328e-01 -3.41415167e-01 -2.97190517e-01
2.28895709e-01 9.01769400e-01 -8.27816188e-01 4.16112006e-01
8.10130060e-01 1.03240776e+00 -6.70579970e-01 7.00760305e-01
1.18382916e-01 8.93027723e-01 -4.06537890e-01 7.09317103e-02
5.37940204e-01 4.00908887e-02 2.80730315e-02 1.54220104e+00
2.35359594e-01 -4.14252132e-02 4.50630218e-01 7.18821764e-01
-5.30780494e-01 5.84689915e-01 -3.99786234e-01 -5.09477198e-01
5.19140363e-01 1.12619877e+00 -3.14660639e-01 -7.16305614e-01
-1.01041210e+00 3.11932474e-01 2.77697474e-01 3.06448251e-01
-5.53886473e-01 -2.00047195e-01 7.68648565e-01 1.57023281e-01
1.98719231e-03 -2.19151676e-01 -3.58198106e-01 -1.57464433e+00
2.11342070e-02 -9.46286917e-01 7.64570057e-01 -4.54653382e-01
-1.79434085e+00 3.65656346e-01 8.50216597e-02 -7.25007296e-01
-6.79592907e-01 -9.99962628e-01 -4.92944717e-01 9.67685819e-01
-1.85219920e+00 -1.08600128e+00 2.63539791e-01 3.39701146e-01
3.08054000e-01 -6.11018479e-01 8.87398839e-01 4.42813486e-01
-5.84037125e-01 5.48742652e-01 1.12081587e-01 5.59293091e-01
1.06640410e+00 -1.39497495e+00 1.04139340e+00 5.63311696e-01
9.96064097e-02 8.89860094e-01 3.71131480e-01 -5.49821138e-01
-1.16870284e+00 -1.20753229e+00 1.46493661e+00 -8.93541336e-01
1.23470163e+00 -3.43562156e-01 -9.11413372e-01 1.18117726e+00
1.49962559e-01 -2.34185711e-01 1.02479565e+00 3.51930887e-01
-6.56001031e-01 1.27914026e-01 -1.01913691e+00 1.64920852e-01
4.25239176e-01 -5.26215792e-01 -7.62167573e-01 4.41575110e-01
8.03602219e-01 -3.28127742e-01 -1.36449289e+00 1.99622631e-01
5.27566791e-01 -8.45713854e-01 1.06402171e+00 -9.42191005e-01
6.28647983e-01 3.34288388e-01 -2.61010379e-01 -7.54227459e-01
-3.36012617e-02 -5.95169663e-01 -3.26604009e-01 1.53167939e+00
8.25355232e-01 -1.19048882e+00 9.24254894e-01 1.02284515e+00
3.65404189e-02 -9.39085424e-01 -5.66911638e-01 -9.01129544e-01
7.78079093e-01 -7.48205245e-01 8.91788900e-01 1.25637186e+00
5.99476360e-02 8.20389763e-02 -1.74623698e-01 -3.26878488e-01
4.50785130e-01 4.31967974e-01 6.40623391e-01 -1.13493955e+00
-6.61709309e-01 -6.01307690e-01 1.05892368e-01 -1.16567051e+00
3.84793639e-01 -9.34728205e-01 -5.32829881e-01 -1.52261567e+00
3.03994447e-01 -6.10428929e-01 -3.43321472e-01 8.16876233e-01
6.51304722e-02 8.59193429e-02 2.37045452e-01 4.52407986e-01
-5.19473076e-01 3.59056145e-02 6.14298046e-01 -6.71036094e-02
-4.95901844e-03 -1.79799497e-01 -1.12563634e+00 8.69821489e-01
8.42346072e-01 -3.96235168e-01 -6.42123595e-02 -5.62804282e-01
7.36612320e-01 -9.45744142e-02 -3.62301208e-02 -6.13233507e-01
1.42827276e-02 -3.39507580e-01 3.25179577e-01 -6.35095060e-01
1.58077687e-01 -1.63322106e-01 -3.65526319e-01 2.40031198e-01
-3.98295969e-01 3.79545897e-01 2.82382995e-01 -3.45934629e-02
-5.34662962e-01 -3.66215169e-01 4.77148563e-01 -6.09558582e-01
-9.03572857e-01 3.84285033e-01 -3.49079460e-01 7.30962455e-01
5.97811937e-01 1.26624912e-01 -3.33880574e-01 -2.69078702e-01
-3.97995234e-01 4.37377483e-01 4.09516275e-01 4.97821122e-01
3.25861692e-01 -1.07683337e+00 -9.54153180e-01 -4.16554213e-02
9.77139026e-02 -8.42323899e-02 -3.56362104e-01 4.87977892e-01
-5.43010592e-01 1.07316029e+00 4.23975229e-01 5.51340617e-02
-8.05019557e-01 3.43400627e-01 3.68498266e-01 -1.05750966e+00
-4.49426472e-01 9.29243445e-01 1.80950955e-01 -6.07283592e-01
2.29009446e-02 -9.50771749e-01 -2.58941501e-01 2.29684696e-01
5.67656577e-01 1.46153674e-01 1.77785203e-01 -1.49929583e-01
-4.01053488e-01 3.60449821e-01 -2.90891856e-01 -1.37173936e-01
1.44722986e+00 9.54697132e-02 -3.07082206e-01 6.64701760e-01
1.35322320e+00 1.58478692e-01 -7.58170128e-01 -5.10034442e-01
6.55870497e-01 -9.89852622e-02 -2.90777385e-01 -8.01807344e-01
-9.92370427e-01 5.80220938e-01 -2.38260448e-01 3.11663628e-01
7.37156987e-01 -9.41565167e-03 1.14187205e+00 6.17897809e-01
5.97197175e-01 -1.09756553e+00 1.65803075e-01 1.00139797e+00
5.63140571e-01 -1.26283646e+00 1.08720124e-01 -2.34786514e-02
-6.78568006e-01 1.06653631e+00 2.92639166e-01 -1.59571141e-01
1.00420654e+00 6.60064399e-01 2.66946614e-01 -2.82431006e-01
-1.21739554e+00 4.11699533e-01 -3.11334431e-02 -2.62649953e-02
1.03373897e+00 -7.02986419e-02 -1.12020366e-01 1.06501293e+00
-7.50704467e-01 1.10248253e-01 5.65434813e-01 1.12379408e+00
-4.01094437e-01 -1.50756645e+00 -3.60284328e-01 8.13768446e-01
-1.34640336e+00 -7.24987268e-01 -4.82348979e-01 9.40145791e-01
-2.19637394e-01 8.80737305e-01 8.55149999e-02 2.11813767e-03
1.90917805e-01 4.71419722e-01 -5.50730266e-02 -1.06085753e+00
-1.06818008e+00 -2.86482632e-01 4.27480549e-01 -4.04024959e-01
-3.43803227e-01 -9.70350742e-01 -1.32686412e+00 -7.57008076e-01
1.51031569e-01 4.37573105e-01 4.52464521e-01 9.80906367e-01
3.08423609e-01 6.64023310e-02 3.35494906e-01 -3.08645755e-01
-7.99336731e-01 -1.06032944e+00 -8.97909045e-01 1.98809221e-01
5.61415590e-02 -2.01402903e-01 -4.08893555e-01 7.05048861e-03] | [10.766013145446777, 8.258549690246582] |
4fac476e-cfe6-44d2-9ce1-0bcf9e365e9c | a-textless-metric-for-speech-to-speech | 2210.11835 | null | https://arxiv.org/abs/2210.11835v1 | https://arxiv.org/pdf/2210.11835v1.pdf | A Textless Metric for Speech-to-Speech Comparison | This paper proposes a textless speech-to-speech comparison metric that allows comparing a speech hypothesis with a speech reference without falling-back to their text transcripts. We leverage recently proposed speech2unit encoders (such as HuBERT) to pseudo-transcribe the speech utterances into discrete acoustic units and propose a simple neural architecture that learns a speech-based metric which correlates well with its text-based counterpart. Such a textless metric could ultimately be interesting for speech-to-speech translation evaluation (for oral languages or languages with no reliable ASR system available). | ['Ioan Calapodescu', 'Olivier Galibert', 'Swen Ribeiro', 'Laurent Besacier'] | 2022-10-21 | null | null | null | null | ['speech-to-speech-translation'] | ['speech'] | [ 4.84904408e-01 5.13085783e-01 -8.14038292e-02 -7.79112935e-01
-1.35197294e+00 -5.80319345e-01 7.96714544e-01 -5.16601354e-02
-4.77231592e-01 6.04116440e-01 3.99044305e-01 -7.94095635e-01
3.80461663e-01 -2.64810354e-01 -5.91603398e-01 -3.55134130e-01
2.62450695e-01 7.25808263e-01 -6.64092079e-02 -3.03210378e-01
-3.27107936e-01 3.24952692e-01 -1.15133381e+00 2.31112316e-01
4.49128330e-01 9.21060324e-01 3.87327254e-01 1.19995928e+00
-2.76261289e-02 6.72721922e-01 -1.04004776e+00 -5.81233621e-01
2.19508305e-01 -7.84817278e-01 -8.48079205e-01 -1.52950332e-01
5.06013513e-01 -3.11747909e-01 -6.62299991e-01 1.04562461e+00
6.22808635e-01 3.74825656e-01 6.54586911e-01 -7.22326517e-01
-1.08602774e+00 8.15249085e-01 4.17451978e-01 2.79549778e-01
5.11246145e-01 -9.81422961e-02 1.04934180e+00 -1.03490961e+00
4.89896715e-01 1.44813669e+00 2.64119178e-01 7.61849761e-01
-9.99384582e-01 -2.28739619e-01 -3.81259978e-01 1.92146108e-01
-1.27869940e+00 -1.29561448e+00 5.50692618e-01 7.67927840e-02
1.39182508e+00 5.94811201e-01 1.71998635e-01 1.70914161e+00
-6.37914464e-02 5.95725298e-01 8.11631918e-01 -5.96231699e-01
3.28685343e-01 -2.31252834e-02 -3.02933246e-01 4.23119903e-01
-3.87882352e-01 2.55619884e-01 -7.37291813e-01 3.29529583e-01
4.45268095e-01 -4.97810036e-01 -3.43688965e-01 4.87435162e-01
-1.23885322e+00 6.15232289e-01 1.68582067e-01 6.00137234e-01
-2.77367830e-01 2.49756023e-01 6.13652289e-01 9.30990398e-01
4.78503257e-01 3.92378509e-01 -3.47233027e-01 -4.32811558e-01
-1.16394067e+00 -3.11595261e-01 7.33983278e-01 1.01590526e+00
4.75810021e-01 6.78230941e-01 -1.95282251e-01 1.03616679e+00
4.84553337e-01 8.62319887e-01 1.06243002e+00 -8.29367399e-01
5.56307495e-01 -2.29917303e-01 -1.87219843e-01 -3.43775064e-01
-6.64644539e-02 -2.74849802e-01 -6.67457163e-01 -5.05299419e-02
7.39871934e-02 -6.71481043e-02 -8.35510790e-01 1.54592443e+00
-7.01727793e-02 3.88998985e-02 5.35440803e-01 9.73520875e-01
9.23456013e-01 9.23662603e-01 -3.65741193e-01 -3.58401597e-01
1.03414583e+00 -1.07375669e+00 -9.49409485e-01 -3.09191793e-01
5.62494397e-01 -8.99105847e-01 1.43472505e+00 -2.46498901e-02
-1.35614276e+00 -6.56451404e-01 -1.26212418e+00 -3.45184267e-01
-3.78819913e-01 4.61094715e-02 -2.37095639e-01 7.31377304e-01
-1.58889723e+00 5.41078508e-01 -9.55984354e-01 -4.99693424e-01
-2.94513255e-01 1.37392893e-01 -4.77405041e-01 4.73770410e-01
-1.42310810e+00 1.18291104e+00 3.73570621e-01 1.24169879e-01
-1.02083385e+00 -9.17108804e-02 -9.71659839e-01 3.00662331e-02
7.49099813e-03 -5.72651446e-01 1.97939193e+00 -1.04262590e+00
-2.43985462e+00 8.24374497e-01 -4.49257612e-01 -7.17057347e-01
3.34146112e-01 1.69329986e-01 -7.55284429e-01 1.39435723e-01
-1.80546343e-01 5.47119200e-01 1.10686100e+00 -7.07229257e-01
-2.79173493e-01 -9.43250954e-02 -1.24650627e-01 3.56296420e-01
-2.22150177e-01 3.77212971e-01 -1.79086439e-02 -8.40206504e-01
2.05189258e-01 -8.19954455e-01 1.07936673e-01 -3.16814072e-02
-5.77913880e-01 -3.30050558e-01 6.60835087e-01 -8.49377990e-01
8.99028003e-01 -1.87331235e+00 1.70853421e-01 -2.26817966e-01
-8.59156996e-02 4.57684875e-01 -5.12871802e-01 5.56755960e-01
-7.46409222e-02 2.04136387e-01 -3.05274934e-01 -9.31346774e-01
1.69855684e-01 4.34428662e-01 -4.89400119e-01 4.39863324e-01
4.30968910e-01 9.77402091e-01 -8.03099096e-01 -2.78249294e-01
4.43891585e-01 6.83030248e-01 -7.03873411e-02 7.03818738e-01
-2.89794542e-02 3.22032660e-01 1.75576776e-01 4.91424382e-01
1.51410908e-01 3.04244339e-01 3.52749377e-02 5.79187796e-02
3.60410213e-02 1.28005338e+00 -3.70079160e-01 1.69586968e+00
-1.02361465e+00 1.10772991e+00 1.91084698e-01 -9.53906775e-01
1.23661792e+00 8.90269458e-01 -9.46028978e-02 -8.56681049e-01
2.36249313e-01 4.72421765e-01 2.32592449e-02 -2.16752872e-01
5.42521060e-01 -4.17456865e-01 1.59571037e-01 2.86777049e-01
6.00540936e-01 -5.80980241e-01 -1.38027564e-01 -1.96578488e-01
1.21311605e+00 -1.91685677e-01 3.22321594e-01 -6.05630726e-02
5.32001853e-01 -5.50927699e-01 4.90382090e-02 6.17757201e-01
-3.83586347e-01 1.01296246e+00 -8.43586251e-02 -1.57979041e-01
-1.20460665e+00 -1.39949143e+00 6.43175766e-02 1.14411664e+00
-3.14472169e-01 -4.00815517e-01 -9.02873218e-01 -4.81946081e-01
-5.36072314e-01 1.05379581e+00 -2.44211107e-01 -2.41902247e-01
-5.61510980e-01 2.23670810e-01 1.16584921e+00 4.79971051e-01
-3.09546124e-02 -1.02096224e+00 1.19685270e-01 4.59937274e-01
-1.82670325e-01 -1.33414948e+00 -7.49957204e-01 6.59448981e-01
-5.97170711e-01 -4.05117124e-01 -1.10438323e+00 -1.04485583e+00
1.54317796e-01 4.68793698e-02 9.03764665e-01 -1.77873999e-01
5.17398238e-01 8.16927627e-02 -5.23746252e-01 -2.53623307e-01
-1.56369734e+00 9.98288095e-02 5.97644567e-01 -8.91073719e-02
1.11791365e-01 -4.82855052e-01 -1.15991838e-01 3.74029726e-01
-7.75822282e-01 -2.08930537e-01 3.81307811e-01 7.88146436e-01
4.21540886e-01 -6.43932521e-01 7.33231723e-01 -1.74108803e-01
8.86949301e-01 -2.23781228e-01 -3.86565417e-01 3.76544923e-01
-3.39315593e-01 2.12540656e-01 1.01379275e+00 -3.52163941e-01
-6.62508607e-01 -1.33731216e-01 -8.63710940e-01 -6.79511845e-01
-2.25905478e-01 3.69305313e-01 -1.68955758e-01 3.08978587e-01
7.65751421e-01 4.17412788e-01 -7.23703727e-02 -5.33564329e-01
6.04308844e-01 1.48464286e+00 9.65510845e-01 -3.19043905e-01
6.92114711e-01 -2.10507438e-01 -4.60073411e-01 -1.20946252e+00
-5.38351238e-01 -4.00366604e-01 -7.09764063e-01 1.15816280e-01
8.55157912e-01 -7.48514354e-01 -6.04735732e-01 7.45292455e-02
-1.78479683e+00 -2.86790580e-01 -4.01769012e-01 8.07426870e-01
-1.01272845e+00 2.53347218e-01 -7.23320425e-01 -9.54651415e-01
-5.40672719e-01 -1.24565661e+00 1.37327552e+00 -1.20767251e-01
-5.06293356e-01 -9.32119310e-01 2.96589822e-01 1.48656100e-01
6.32951021e-01 -6.94440603e-01 5.47943890e-01 -1.18438697e+00
-4.24149372e-02 3.04340534e-02 7.35718608e-02 8.99279833e-01
3.78209412e-01 9.80318338e-02 -1.41613817e+00 -1.99248120e-01
2.75424302e-01 -3.17186892e-01 5.29455006e-01 2.33990356e-01
6.67069972e-01 -7.87674189e-01 1.92106292e-01 5.57284117e-01
5.57063222e-01 4.41763490e-01 4.91190135e-01 -9.83486921e-02
4.56457525e-01 4.42059696e-01 -3.81914005e-02 -1.92967847e-01
2.19511092e-01 9.87601340e-01 -7.72177726e-02 -5.96857965e-02
-6.37795269e-01 -3.13948750e-01 1.00019217e+00 2.02663970e+00
4.08864051e-01 -8.28904510e-01 -7.40667343e-01 4.43579733e-01
-1.39972317e+00 -8.32536876e-01 3.94894868e-01 2.26856637e+00
9.33677435e-01 1.81682616e-01 -1.18756481e-01 6.86491430e-02
8.05815578e-01 4.01683539e-01 -4.33728099e-01 -1.03087771e+00
-3.52503419e-01 5.27354002e-01 1.35327980e-01 8.39336932e-01
-7.98600137e-01 1.03322697e+00 7.62257862e+00 7.40878403e-01
-1.38705182e+00 4.78516012e-01 5.18863678e-01 7.46265501e-02
-3.62371236e-01 -1.81439787e-01 -3.69662464e-01 3.17841619e-01
2.10726571e+00 -4.85456109e-01 6.50459170e-01 7.58420885e-01
3.79976213e-01 6.51906610e-01 -1.65300381e+00 1.26555991e+00
1.25459179e-01 -1.18236673e+00 1.64110720e-01 -3.35221916e-01
3.86989921e-01 5.59017301e-01 -1.67313851e-02 3.51746053e-01
1.82828963e-01 -1.15067565e+00 1.08634579e+00 9.51155499e-02
1.28170991e+00 -3.43809426e-01 5.92569530e-01 3.34340967e-02
-1.17013979e+00 4.46595162e-01 -3.34479958e-01 8.97085592e-02
3.91848743e-01 2.16611564e-01 -1.77583003e+00 4.57881719e-01
1.92794144e-01 4.20638263e-01 -1.34288341e-01 4.70717072e-01
-3.02533060e-01 1.06103528e+00 -4.19119626e-01 -2.52159625e-01
2.51768410e-01 -1.67876612e-02 9.28507864e-01 1.37411416e+00
7.38446057e-01 -3.56245100e-01 7.30133243e-03 8.63566101e-01
-4.90343690e-01 2.90170282e-01 -8.45053196e-01 -5.89355350e-01
7.74263442e-01 7.27288425e-01 -4.31976587e-01 -4.76810753e-01
-3.39193165e-01 1.39365268e+00 2.42761046e-01 4.76188391e-01
-3.86124700e-01 -6.05233073e-01 9.56727862e-01 -1.98480025e-01
-9.36724097e-02 -6.53753519e-01 5.03484532e-02 -1.00862992e+00
2.56219298e-01 -8.86484623e-01 -3.16179842e-01 -1.01055443e+00
-1.27871943e+00 1.35317147e+00 -5.00490487e-01 -1.19717038e+00
-1.05142319e+00 -6.17227733e-01 -5.41858315e-01 9.82153833e-01
-1.29890680e+00 -9.87204432e-01 3.85020196e-01 2.32921496e-01
1.09274483e+00 -5.23074865e-01 1.21419752e+00 1.19021960e-01
-3.08382690e-01 9.15701628e-01 2.66711563e-01 1.91268533e-01
7.34620690e-01 -1.30264258e+00 1.21566594e+00 8.14535081e-01
8.46332788e-01 6.69383168e-01 7.04133868e-01 -3.20720762e-01
-1.36550736e+00 -9.60625291e-01 1.21378720e+00 -3.99927437e-01
8.79459262e-01 -5.39383471e-01 -1.01705778e+00 5.14896512e-01
5.10769188e-01 -7.76143894e-02 5.11538625e-01 -1.57579947e-02
-3.67099315e-01 -2.35090554e-01 -7.47262061e-01 7.73664653e-01
1.00758684e+00 -1.53936648e+00 -1.03398180e+00 3.95775497e-01
1.46214318e+00 -3.57676327e-01 -6.78010762e-01 6.01369999e-02
3.84673864e-01 -5.59236586e-01 5.72609842e-01 -6.07601941e-01
3.76869901e-03 -2.41401643e-01 -7.92248309e-01 -1.67227185e+00
1.23072840e-01 -1.29905963e+00 -7.03616664e-02 1.02448726e+00
5.80561221e-01 -5.88490844e-01 3.54564995e-01 1.28883392e-01
-7.53544033e-01 -2.99532801e-01 -1.71107435e+00 -1.37624395e+00
1.07986420e-01 -6.50270939e-01 5.83643496e-01 7.55204141e-01
1.07431628e-01 6.26018941e-01 -1.58692449e-01 1.81445345e-01
-1.16518708e-02 -5.97610533e-01 2.85341889e-01 -8.55003238e-01
-2.51065522e-01 -5.15863061e-01 -7.06114411e-01 -1.32482696e+00
5.46226680e-01 -1.30289149e+00 5.75819731e-01 -1.33753622e+00
-5.59484839e-01 -7.22813308e-02 -3.48817289e-01 5.35856128e-01
2.19683483e-01 8.20261016e-02 1.91406086e-01 4.58287857e-02
-3.07594091e-01 9.91365910e-01 8.39189887e-01 -4.20488417e-01
-8.34244862e-02 -6.00912981e-02 -2.15858236e-01 3.14860106e-01
5.54456532e-01 -5.37155330e-01 -3.03903192e-01 -6.08795226e-01
-2.88694084e-01 4.17776644e-01 4.38704975e-02 -1.12106085e+00
2.78339267e-01 -7.56902946e-03 -2.64495194e-01 -2.65247166e-01
6.12016618e-01 -5.24075091e-01 -8.34607929e-02 4.72418703e-02
-7.88893878e-01 2.46740535e-01 -1.16120309e-01 1.51670784e-01
-6.65652871e-01 -4.22092557e-01 5.73041201e-01 1.48903206e-01
-2.73992300e-01 -1.12048246e-01 -6.48249567e-01 4.63969633e-02
3.27500254e-01 -6.91070706e-02 -3.19491267e-01 -9.29236650e-01
-6.79679275e-01 -4.31483775e-01 3.02873611e-01 8.45175803e-01
6.81066334e-01 -1.38885725e+00 -1.00510311e+00 3.81858855e-01
8.67510512e-02 -3.61331165e-01 -3.82806867e-01 3.52299988e-01
-2.51210123e-01 8.09736967e-01 1.84685141e-01 -6.13256216e-01
-9.81159449e-01 2.57983595e-01 5.70289552e-01 4.55918014e-01
-3.52443367e-01 8.74599040e-01 -3.50447744e-02 -8.15836012e-01
2.99214989e-01 -7.03732431e-01 5.12604654e-01 -1.77625805e-01
6.36654973e-01 1.84116751e-01 7.81755805e-01 -1.06525826e+00
-5.02326787e-01 7.29347626e-03 1.46932736e-01 -8.05462122e-01
1.06047726e+00 -4.47525024e-01 3.78772527e-01 9.43346083e-01
1.59690249e+00 -1.11149594e-01 -8.03756177e-01 -2.90857524e-01
1.91071689e-01 3.47189605e-02 2.14410916e-01 -6.42960787e-01
-5.66032887e-01 1.21312106e+00 5.58101237e-01 2.91931093e-01
6.95890486e-01 1.62508458e-01 1.07853103e+00 9.39124167e-01
1.98867142e-01 -1.16073430e+00 2.26277933e-02 9.73289311e-01
1.20105064e+00 -1.23570621e+00 -7.38205850e-01 1.26414552e-01
-4.65680093e-01 1.34254742e+00 -5.97978309e-02 2.22140044e-01
4.36545134e-01 4.56950575e-01 5.88803411e-01 2.63751954e-01
-9.42200303e-01 -3.73163342e-01 5.86064756e-01 9.01129246e-01
7.25634456e-01 2.03888834e-01 6.90235645e-02 3.76184821e-01
-8.17850828e-01 -3.74655098e-01 5.57825148e-01 4.14820850e-01
-4.28957164e-01 -1.20658278e+00 -1.74441233e-01 8.72942507e-02
-2.31922805e-01 -4.59156662e-01 -6.30579352e-01 1.27839029e-01
-6.25747800e-01 1.32215345e+00 2.74989486e-01 -5.39040565e-01
2.78016269e-01 4.08865094e-01 2.69856632e-01 -9.16415930e-01
-5.09317756e-01 1.82581216e-01 3.33921671e-01 -5.05812049e-01
-1.56542912e-01 -2.38936156e-01 -1.26453471e+00 5.58892190e-02
-3.08745414e-01 8.89344886e-02 1.24847913e+00 1.06134522e+00
2.68311679e-01 4.69971955e-01 9.13312912e-01 -9.72836614e-01
-8.78886402e-01 -1.32045066e+00 -3.32021654e-01 2.00361609e-01
9.35060859e-01 -1.64344370e-01 -3.25925499e-01 1.38479203e-01] | [14.620027542114258, 7.151122570037842] |
14b3009d-c056-457a-b707-8b96306b2484 | towards-more-generalizable-one-shot-visual | 2110.13423 | null | https://arxiv.org/abs/2110.13423v2 | https://arxiv.org/pdf/2110.13423v2.pdf | Towards More Generalizable One-shot Visual Imitation Learning | A general-purpose robot should be able to master a wide range of tasks and quickly learn a novel one by leveraging past experiences. One-shot imitation learning (OSIL) approaches this goal by training an agent with (pairs of) expert demonstrations, such that at test time, it can directly execute a new task from just one demonstration. However, so far this framework has been limited to training on many variations of one task, and testing on other unseen but similar variations of the same task. In this work, we push for a higher level of generalization ability by investigating a more ambitious multi-task setup. We introduce a diverse suite of vision-based robot manipulation tasks, consisting of 7 tasks, a total of 61 variations, and a continuum of instances within each variation. For consistency and comparison purposes, we first train and evaluate single-task agents (as done in prior few-shot imitation work). We then study the multi-task setting, where multi-task training is followed by (i) one-shot imitation on variations within the training tasks, (ii) one-shot imitation on new tasks, and (iii) fine-tuning on new tasks. Prior state-of-the-art, while performing well within some single tasks, struggles in these harder multi-task settings. To address these limitations, we propose MOSAIC (Multi-task One-Shot Imitation with self-Attention and Contrastive learning), which integrates a self-attention model architecture and a temporal contrastive module to enable better task disambiguation and more robust representation learning. Our experiments show that MOSAIC outperforms prior state of the art in learning efficiency, final performance, and learns a multi-task policy with promising generalization ability via fine-tuning on novel tasks. | ['Pieter Abbeel', 'Kimin Lee', 'Fangchen Liu', 'Zhao Mandi'] | 2021-10-26 | null | null | null | null | ['robot-manipulation'] | ['robots'] | [ 4.56971824e-01 -1.41516209e-01 -1.56636268e-03 -1.58901848e-02
-6.91517949e-01 -6.24496698e-01 1.01750183e+00 -4.23432678e-01
-7.04418957e-01 9.12429273e-01 -2.14360893e-01 -9.30481404e-02
-3.84097576e-01 -1.36234000e-01 -8.25904667e-01 -7.25437462e-01
-2.71892339e-01 7.65066087e-01 6.10666096e-01 -3.81057769e-01
1.46999136e-01 2.84656554e-01 -1.58410680e+00 9.13413018e-02
9.59183276e-01 4.64894354e-01 1.01216531e+00 7.48438716e-01
4.60594654e-01 6.25408709e-01 -7.38803029e-01 6.35934621e-02
3.60273331e-01 -3.56869310e-01 -9.49144542e-01 1.16120860e-01
1.13112427e-01 -4.68478203e-01 -3.09960216e-01 7.79800355e-01
4.70947772e-01 6.11275911e-01 6.98472142e-01 -1.47436810e+00
-7.01333880e-01 2.74577796e-01 -2.05005527e-01 1.09952442e-01
4.22440082e-01 9.11483109e-01 5.56818008e-01 -6.38420463e-01
6.66086257e-01 1.25554931e+00 4.80783105e-01 9.06217694e-01
-1.22447121e+00 -5.65818667e-01 3.23486596e-01 1.91395804e-01
-7.62944579e-01 -3.26683551e-01 3.97883385e-01 -4.19166118e-01
1.49265492e+00 -3.67321700e-01 4.90424752e-01 1.82131541e+00
3.47141564e-01 8.77811551e-01 1.41751444e+00 -2.11291119e-01
3.68065745e-01 -1.72735378e-01 -2.10147873e-01 5.56651592e-01
5.26225343e-02 6.25920355e-01 -3.45409274e-01 6.45416677e-02
7.43165851e-01 3.77707690e-01 -2.75900662e-01 -7.79587686e-01
-1.68739712e+00 5.58853984e-01 4.91197944e-01 5.81152439e-01
-6.45957530e-01 3.36415768e-01 4.71139818e-01 9.40967441e-01
-9.98179317e-02 1.01299918e+00 -7.10906029e-01 -3.01310867e-01
-4.30085540e-01 3.66744697e-01 8.15423727e-01 1.01764154e+00
8.92121017e-01 1.92773774e-01 -3.72969270e-01 6.85869277e-01
-1.99464500e-01 5.18180549e-01 8.91750216e-01 -1.28242743e+00
3.93468618e-01 1.66844770e-01 2.05114663e-01 -9.87567455e-02
-4.29355860e-01 -2.84082532e-01 -3.07818770e-01 7.86880136e-01
3.82608712e-01 -2.94010550e-01 -1.28725910e+00 2.03216815e+00
1.00906871e-01 1.91172153e-01 3.62197787e-01 8.00864935e-01
2.34332278e-01 4.74550873e-01 6.57827258e-02 -4.97365259e-02
1.08186281e+00 -1.62383687e+00 -3.05531353e-01 -5.55137455e-01
7.02723920e-01 -3.46612364e-01 1.25110877e+00 4.36796397e-01
-8.51022065e-01 -8.85817945e-01 -1.07041156e+00 1.73331276e-01
-4.90759850e-01 -2.32430652e-01 4.42022324e-01 7.70134553e-02
-1.15715492e+00 8.69986057e-01 -9.21374738e-01 -7.85668194e-01
2.87095994e-01 4.25665468e-01 -5.56855261e-01 -2.87678152e-01
-9.89658415e-01 1.42373848e+00 5.47434390e-01 -4.03424770e-01
-1.82264757e+00 -5.65185189e-01 -9.43025231e-01 -5.57391122e-02
9.20272470e-01 -1.03489721e+00 1.69437528e+00 -1.01565099e+00
-1.92220926e+00 5.70696592e-01 3.59062731e-01 -4.45452273e-01
5.10280848e-01 -3.46314311e-01 -9.89844054e-02 2.33653728e-02
4.43052739e-01 8.98655474e-01 1.22337091e+00 -1.29356229e+00
-5.71687222e-01 -2.77839690e-01 4.79600459e-01 3.79142612e-01
1.01093836e-01 -1.28358275e-01 -1.29352659e-01 -4.59813148e-01
-5.13062656e-01 -1.34068131e+00 -3.48572880e-02 -1.45106137e-01
3.37516703e-02 -4.02067542e-01 1.01980937e+00 -2.49639437e-01
3.33448082e-01 -2.24584246e+00 6.74949169e-01 -4.68829513e-01
2.79905181e-02 4.84992236e-01 -6.84206069e-01 5.85373819e-01
-7.40703195e-02 -3.19094449e-01 -3.54681373e-01 -5.16446471e-01
1.36120260e-01 5.31726182e-01 -1.64680332e-01 1.92569330e-01
3.89313042e-01 1.24502468e+00 -1.34032345e+00 -5.51813319e-02
2.02566415e-01 1.16476789e-01 -3.70364308e-01 3.60546112e-01
-3.95186722e-01 9.38593328e-01 -4.20870841e-01 5.12855709e-01
-2.74397084e-03 -2.44250238e-01 4.99356054e-02 2.61612475e-01
4.47622202e-02 3.31984013e-02 -7.45345414e-01 2.37418962e+00
-8.12007606e-01 4.96489793e-01 1.65907741e-01 -1.16851497e+00
6.39624596e-01 3.62218857e-01 2.71747589e-01 -6.42157555e-01
4.32455586e-03 3.20123732e-01 5.36296189e-01 -8.32084477e-01
2.55385786e-01 -1.75990611e-01 -1.40956521e-01 6.14837289e-01
5.46193242e-01 -5.65750003e-01 2.80251771e-01 1.57588138e-03
1.56919467e+00 6.73046947e-01 4.58283663e-01 -2.12593097e-02
4.86856140e-02 1.60725981e-01 2.62513369e-01 1.19484222e+00
-6.17063105e-01 3.46756846e-01 1.07456274e-01 -2.29908273e-01
-1.19754219e+00 -1.02913105e+00 3.38664144e-01 1.59650373e+00
1.55079469e-01 4.06436399e-02 -3.29630882e-01 -9.08608735e-01
3.26030254e-01 7.96899974e-01 -8.82893026e-01 -3.82522553e-01
-7.51863301e-01 -1.98221415e-01 2.70956576e-01 5.50350785e-01
6.14886880e-01 -1.80271280e+00 -1.35435212e+00 2.03404739e-01
2.93409437e-01 -9.08300936e-01 -4.10752416e-01 6.16466463e-01
-6.36901557e-01 -1.08395684e+00 -1.14515531e+00 -1.06930447e+00
3.66672128e-01 5.55567145e-01 9.57924187e-01 -2.07677931e-01
-2.09039763e-01 8.91499400e-01 -4.11648691e-01 -1.89230457e-01
-5.69198489e-01 1.44476756e-01 4.04705346e-01 -5.16861916e-01
-2.17370838e-01 -8.79241884e-01 -3.27881813e-01 3.55546236e-01
-8.31798732e-01 -1.59581229e-01 1.28255403e+00 1.36682129e+00
1.19535305e-01 -3.43836814e-01 8.04209113e-01 -4.73740339e-01
8.35993767e-01 -6.66409433e-01 -3.34095091e-01 3.20375115e-01
-5.12163937e-01 2.95675546e-01 7.54339814e-01 -1.09014273e+00
-1.02422082e+00 -1.22190662e-01 4.13639098e-01 -8.97012830e-01
-2.96520263e-01 3.56641412e-01 3.30199361e-01 -1.91012979e-01
8.20525885e-01 4.97483790e-01 2.58491844e-01 -2.69067168e-01
4.84561324e-01 3.98082018e-01 6.89832389e-01 -8.16892982e-01
7.47449160e-01 1.93148673e-01 -2.19315469e-01 -4.95101124e-01
-5.14179766e-01 -3.28961611e-01 -7.21880555e-01 -5.74597493e-02
7.72715330e-01 -8.59455705e-01 -7.32288718e-01 6.95814073e-01
-9.82367754e-01 -1.38730156e+00 -4.57551986e-01 5.68764985e-01
-1.07934606e+00 1.98078781e-01 -3.10688406e-01 -5.27705789e-01
1.44520730e-01 -1.50238740e+00 1.23298943e+00 2.17435986e-01
-1.89332128e-01 -7.92845845e-01 3.48134398e-01 1.11504374e-02
6.77706659e-01 -1.98002867e-02 6.05995417e-01 -9.53634739e-01
-3.94457996e-01 2.46443212e-01 -8.08569714e-02 2.80061215e-01
1.21382996e-01 -6.85794532e-01 -7.13150382e-01 -9.17432606e-01
1.85370713e-01 -1.10792482e+00 9.82524455e-01 1.08889841e-01
5.97686172e-01 4.12752479e-02 -5.45157135e-01 3.56737494e-01
1.03240192e+00 4.36282635e-01 4.05849069e-01 6.25473201e-01
4.19323236e-01 4.71722424e-01 7.20799565e-01 1.83959156e-01
3.36019695e-01 8.37037742e-01 6.04276836e-01 3.63405466e-01
-2.68645465e-01 6.11432493e-02 6.85142994e-01 3.84888202e-01
-3.62687819e-02 -5.60939545e-03 -7.42732882e-01 8.14032793e-01
-2.06747079e+00 -1.00656259e+00 7.67738163e-01 1.96348596e+00
7.05825150e-01 2.01842338e-01 3.11324567e-01 -3.54572684e-01
4.36343193e-01 2.58276969e-01 -1.22629011e+00 -3.14064443e-01
2.41628885e-01 2.14475676e-01 1.53082348e-02 1.96003079e-01
-1.04788136e+00 1.11012077e+00 5.86701918e+00 5.35053015e-01
-1.10214972e+00 3.05020601e-01 -1.31671667e-01 -1.77808151e-01
3.30088735e-01 5.70933148e-02 -3.70421976e-01 3.63950372e-01
5.90515673e-01 -2.29856730e-01 8.79616797e-01 9.69474733e-01
-3.07024419e-01 -2.85212427e-01 -1.55815613e+00 8.14534843e-01
2.14649603e-01 -7.11003423e-01 -2.61342525e-01 -1.12901039e-01
9.48689699e-01 4.79727447e-01 1.73949748e-01 1.30681241e+00
7.60069430e-01 -1.03092349e+00 4.76534128e-01 3.52797717e-01
7.80218065e-01 -1.09748803e-02 4.55174237e-01 7.92198896e-01
-8.63192499e-01 -5.15953243e-01 -4.21776772e-02 -3.54567528e-01
2.15516668e-02 -3.48155260e-01 -7.89820850e-01 6.31527126e-01
7.83164680e-01 7.80929923e-01 -2.29896590e-01 8.70273471e-01
-4.29003209e-01 -4.60070185e-02 -1.41173393e-01 -1.55510843e-01
5.35611987e-01 1.52489215e-01 7.28725731e-01 8.94686341e-01
2.76108176e-01 6.99543431e-02 6.58041537e-01 8.58902872e-01
8.71133804e-02 -5.85398555e-01 -1.05885351e+00 1.83433965e-01
4.67312843e-01 1.03793073e+00 -2.43871287e-01 -4.72344756e-01
-3.69065493e-01 1.31789351e+00 7.49343812e-01 6.89325094e-01
-6.85742617e-01 -3.58098298e-01 5.79284728e-01 -5.04034579e-01
5.84037900e-01 -5.59774816e-01 2.65193105e-01 -1.13797724e+00
4.08422127e-02 -1.06831062e+00 6.41414747e-02 -9.10086572e-01
-1.34166145e+00 4.63979036e-01 2.50366360e-01 -1.24640393e+00
-7.87580132e-01 -6.26972675e-01 -6.50867879e-01 7.26990581e-01
-1.74222243e+00 -1.04304290e+00 -4.10245925e-01 6.55234873e-01
1.18931723e+00 -4.54086274e-01 8.22524786e-01 -2.19957396e-01
-4.37614411e-01 3.77243131e-01 1.63764760e-01 -3.48781824e-01
9.75711584e-01 -1.15929401e+00 4.01118010e-01 4.81693268e-01
-1.51514053e-01 5.24524748e-01 6.58609748e-01 -6.15356147e-01
-1.57406664e+00 -1.00684452e+00 1.86122313e-01 -4.47621077e-01
8.71667147e-01 -2.70142913e-01 -9.86658514e-01 1.03443325e+00
3.45887482e-01 9.24803037e-03 -9.48299766e-02 -9.83358994e-02
-2.98586577e-01 3.13587844e-01 -9.91147161e-01 7.21823394e-01
1.38038933e+00 -5.08673012e-01 -1.20748615e+00 4.68837291e-01
7.65853405e-01 -3.39055777e-01 -7.29386806e-01 4.89121348e-01
5.82828522e-01 -8.32944572e-01 8.50732386e-01 -7.84343123e-01
4.06318724e-01 -1.85315594e-01 6.82462081e-02 -2.03430963e+00
-4.26053613e-01 -8.76358151e-01 -8.79162997e-02 4.63656306e-01
2.03850180e-01 -9.04873013e-01 1.54867917e-01 4.23618183e-02
-6.35054469e-01 -7.38075495e-01 -8.48988593e-01 -1.64099288e+00
1.61737502e-01 -7.98949376e-02 2.54813135e-01 7.90867209e-01
2.34645754e-01 4.95183170e-01 -3.57749403e-01 -1.42388105e-01
4.23127800e-01 1.48477703e-01 1.25865054e+00 -1.07543314e+00
-6.53200269e-01 -6.02509022e-01 -3.16784494e-02 -1.24023581e+00
5.03941715e-01 -7.90676713e-01 5.22484004e-01 -1.37911487e+00
2.65190393e-01 -2.98074543e-01 -2.10792139e-01 7.97353208e-01
-2.00482801e-01 -2.05663025e-01 5.87251127e-01 6.22182190e-01
-1.05010509e+00 9.09089506e-01 1.49905622e+00 -1.83552504e-01
-3.09928089e-01 -1.13602795e-01 -3.33878547e-01 4.49117631e-01
6.38228714e-01 -4.01438743e-01 -4.75271493e-01 -6.59565032e-01
-3.20733637e-01 1.22251697e-01 4.96243984e-01 -1.30671942e+00
3.16443592e-01 -1.35176614e-01 1.50658131e-01 7.97034800e-02
5.73633254e-01 -6.33299053e-01 -1.14250332e-01 8.14569235e-01
-3.39405596e-01 1.47319764e-01 2.49316797e-01 7.68658102e-01
9.99699235e-02 -2.97698647e-01 6.63832486e-01 -3.99883032e-01
-1.25703752e+00 8.91942009e-02 -3.20110828e-01 9.12839994e-02
1.47079480e+00 -3.49943101e-01 -6.94639802e-01 -2.23169968e-01
-1.01600778e+00 6.41226947e-01 7.25346088e-01 7.07559705e-01
4.70291317e-01 -1.05311978e+00 -7.27849007e-01 5.87160885e-02
3.43497396e-01 -6.35070354e-02 4.11344469e-02 9.22768891e-01
2.17100635e-01 3.14963251e-01 -7.02021182e-01 -6.91327751e-01
-8.04303527e-01 8.66868973e-01 2.24583343e-01 -5.33247948e-01
-6.54432118e-01 6.65450990e-01 4.98412907e-01 -6.82950974e-01
2.23678753e-01 -2.65799433e-01 -1.34156391e-01 -3.48080993e-01
1.45701051e-01 3.17174703e-01 -3.05364490e-01 -2.47539997e-01
-2.32170686e-01 5.48104346e-01 -2.88696736e-01 -2.65217960e-01
1.23281181e+00 4.02277000e-02 3.41088146e-01 7.30296433e-01
9.62089837e-01 -7.82670557e-01 -1.97674036e+00 -4.65024292e-01
1.27524780e-02 -2.32938915e-01 -5.11202335e-01 -1.21082103e+00
-5.20181715e-01 7.28055716e-01 3.85229290e-01 6.75419420e-02
8.75284135e-01 1.44745529e-01 6.70625746e-01 9.88032341e-01
7.38613784e-01 -9.58205223e-01 1.02254426e+00 9.63750839e-01
1.18602359e+00 -1.59993017e+00 -1.90592796e-01 2.83174336e-01
-8.57080996e-01 8.92152190e-01 8.45183790e-01 -4.96672809e-01
1.86514884e-01 -6.55579893e-03 -2.73836076e-01 -2.49457866e-01
-1.03680885e+00 -4.42394823e-01 5.11913672e-02 9.94505644e-01
-3.21361542e-01 -1.04128614e-01 2.92456478e-01 2.59946376e-01
9.22765210e-02 2.13584110e-01 3.18769395e-01 1.40818942e+00
-5.16972721e-01 -7.99248695e-01 -5.61703593e-02 2.52174109e-01
1.34082407e-01 1.56853363e-01 -1.93759799e-01 1.09779656e+00
-7.00153410e-02 7.64296770e-01 -2.19509244e-01 -3.57042074e-01
5.09836912e-01 2.33784348e-01 9.38032746e-01 -1.02391207e+00
-6.53184414e-01 -3.74559939e-01 -7.56693706e-02 -7.44400501e-01
-4.87977028e-01 -8.56410921e-01 -1.01746082e+00 1.95852369e-01
-7.96315670e-02 -2.70684302e-01 6.66827917e-01 1.24742103e+00
4.89981174e-01 8.44132423e-01 5.71870804e-01 -1.45452535e+00
-1.23882854e+00 -1.36316931e+00 -3.35027277e-01 6.19560063e-01
5.79102397e-01 -1.25015521e+00 -3.97384673e-01 -1.31864473e-01] | [4.369646072387695, 1.0891437530517578] |
b9bed89a-3882-494c-86fa-3a02fa3e8575 | metaheuristic-approach-to-solve-portfolio | 2211.17193 | null | https://arxiv.org/abs/2211.17193v1 | https://arxiv.org/pdf/2211.17193v1.pdf | Metaheuristic Approach to Solve Portfolio Selection Problem | In this paper, a heuristic method based on TabuSearch and TokenRing Search is being used in order to solve the Portfolio Optimization Problem. The seminal mean-variance model of Markowitz is being considered with the addition of cardinality and quantity constraints to better capture the dynamics of the trading procedure, the model becomes an NP-hard problem that can not be solved using an exact method. The combination of three different neighborhood relations is being explored with Tabu Search. In addition, a new constructive method for the initial solution is proposed. Finally, I show how the proposed techniques perform on public benchmarks | ['Taylan Kabbani'] | 2022-11-10 | null | null | null | null | ['portfolio-optimization'] | ['time-series'] | [-1.31594598e-01 1.24175347e-01 -3.03354293e-01 -1.19091146e-01
-2.38203824e-01 -9.24477994e-01 3.93419802e-01 -4.46513519e-02
-3.44627202e-01 1.22183943e+00 -1.59969360e-01 -4.03028995e-01
-7.23611534e-01 -8.52426112e-01 -2.29500517e-01 -7.05384672e-01
9.27300379e-03 9.07325149e-01 2.81021714e-01 -3.23527724e-01
7.19311118e-01 5.57735801e-01 -1.35346925e+00 3.96046750e-02
6.79745138e-01 1.38259935e+00 2.60235611e-02 1.41180947e-01
-6.09598935e-01 4.16229397e-01 -4.45277989e-01 -6.01866663e-01
1.03260708e+00 -4.52102989e-01 -6.10281169e-01 1.15789518e-01
-3.09947938e-01 -3.52643542e-02 1.16273344e-01 1.23531544e+00
2.56331682e-01 2.89118797e-01 4.81644958e-01 -1.05411196e+00
-9.30427983e-02 9.13602293e-01 -8.04113448e-01 4.35633838e-01
7.45910034e-03 -2.63945162e-01 1.16564596e+00 -3.33575904e-01
6.35711610e-01 1.06816280e+00 -2.53016651e-02 -1.87102240e-02
-9.83835578e-01 -4.49716240e-01 1.03361644e-01 3.01130980e-01
-1.14271128e+00 1.40683442e-01 8.22970688e-01 -1.71779729e-02
6.93673670e-01 8.31100762e-01 1.00544322e+00 3.87823075e-01
4.27256107e-01 5.72741330e-01 1.43538284e+00 -4.60866153e-01
4.60525125e-01 3.92724305e-01 1.31985694e-01 9.50998887e-02
7.55141973e-01 3.85420352e-01 -1.24842480e-01 -5.55045128e-01
4.38419372e-01 -2.65637040e-02 -1.82214350e-01 -5.65148175e-01
-5.12130797e-01 1.10829628e+00 9.23751220e-02 5.34854352e-01
-4.02307540e-01 -1.18835740e-01 2.52875268e-01 4.25437033e-01
5.76076746e-01 8.28033686e-01 -3.72605950e-01 -2.29426533e-01
-1.18621492e+00 5.43129921e-01 1.21675348e+00 7.74972618e-01
2.76564449e-01 9.59081650e-02 -1.71344146e-01 3.30908000e-01
2.50560224e-01 2.82494426e-01 2.74096161e-01 -1.06024110e+00
5.15561223e-01 5.46907961e-01 4.55176204e-01 -8.02067757e-01
-2.67906904e-01 -7.73977458e-01 -3.10724348e-01 2.86574215e-01
4.87375796e-01 -1.09974474e-01 -5.07509887e-01 9.50512409e-01
3.63187701e-01 9.68491882e-02 -1.34744808e-01 6.92928016e-01
6.31894590e-03 7.06296504e-01 -7.69950271e-01 -8.82022381e-01
1.09296942e+00 -9.45997179e-01 -9.46602881e-01 1.96160242e-01
1.49952531e-01 -7.59570956e-01 1.38607219e-01 8.21692765e-01
-1.34680462e+00 2.00425759e-01 -1.05281615e+00 6.58712208e-01
-5.00446618e-01 -6.26336038e-01 6.07660711e-01 1.07356870e+00
-9.81292367e-01 5.66948116e-01 -2.84418494e-01 -8.40406418e-02
1.64682306e-02 4.69614476e-01 2.74330199e-01 1.46374822e-01
-1.02137733e+00 9.23800468e-01 7.88820803e-01 4.53120589e-01
-3.08478117e-01 -3.13334137e-01 -2.85285592e-01 2.54047930e-01
9.67100382e-01 -3.18599135e-01 8.17399740e-01 -7.85136878e-01
-1.81154191e+00 4.54559654e-01 2.35308170e-01 -7.68682122e-01
7.72564530e-01 1.53386772e-01 -5.74842468e-02 -3.34252254e-03
-4.45420325e-01 -1.90715104e-01 5.12546957e-01 -1.22481072e+00
-8.33467901e-01 -1.71767771e-01 2.10225940e-01 4.01792586e-01
-7.95908347e-02 2.52506107e-01 -1.53877974e-01 -6.44671440e-01
1.83949038e-01 -9.48117316e-01 -4.21278238e-01 -7.50602841e-01
-7.53212944e-02 1.07401296e-01 3.05004150e-01 -3.32746625e-01
1.53390229e+00 -1.78814721e+00 1.26868382e-01 1.06958473e+00
-1.68760717e-01 1.32465758e-03 3.65190297e-01 7.03214347e-01
7.03811869e-02 2.62619674e-01 -1.74934939e-01 1.18486904e-01
2.13646978e-01 3.99617046e-01 -3.69587541e-01 4.36942458e-01
-6.81018651e-01 6.84300125e-01 -4.87833261e-01 -2.52592266e-01
-1.10379547e-01 -4.29831564e-01 -5.38555682e-01 -2.65006796e-02
-3.71926904e-01 -1.46455526e-01 -5.19718230e-01 7.79763818e-01
8.98723722e-01 3.53929326e-02 4.77626890e-01 2.81996518e-01
-5.93768716e-01 -1.02930106e-01 -1.62552452e+00 1.17199504e+00
-1.29257157e-01 -3.37479226e-02 3.20994675e-01 -1.05084550e+00
9.21596229e-01 9.87673625e-02 6.37164593e-01 -5.75713217e-01
3.70702147e-01 4.09663111e-01 1.75300583e-01 5.82326427e-02
4.38146472e-01 -1.55244485e-01 3.62647977e-03 7.79025674e-01
-4.08186585e-01 -6.68232217e-02 5.89122236e-01 -3.75830323e-01
7.02632844e-01 4.96621020e-02 3.42748344e-01 -8.78360271e-01
6.50112450e-01 3.71532023e-01 7.76727200e-01 8.39189410e-01
2.48141468e-01 2.18359858e-01 8.83031547e-01 -4.61163402e-01
-7.63223708e-01 -4.90568966e-01 -3.11440468e-01 5.24510622e-01
1.73816592e-01 1.34476572e-01 -7.73740351e-01 -3.72493207e-01
1.54963866e-01 9.81157303e-01 -6.84301019e-01 3.02747995e-01
-4.11369324e-01 -1.07524049e+00 -1.70189291e-01 -1.24360994e-01
5.00784099e-01 -8.26218367e-01 -9.04127061e-01 5.18960118e-01
2.62858540e-01 -7.10529864e-01 -4.22739118e-01 3.93439233e-01
-1.01050723e+00 -1.08820879e+00 -6.88567221e-01 -1.49740621e-01
4.23114270e-01 2.12956779e-02 8.70621085e-01 1.91933122e-02
-1.70426995e-01 1.51087284e-01 -5.99135995e-01 -5.44833779e-01
-1.23315603e-01 2.11296733e-02 -3.03257346e-01 2.14621603e-01
2.98847035e-02 -3.68169755e-01 -5.90698004e-01 3.27237725e-01
-9.46892619e-01 -3.24602753e-01 5.25689662e-01 1.09121335e+00
5.66102207e-01 6.66728437e-01 4.54673201e-01 -9.72431123e-01
8.30990553e-01 -4.60278094e-01 -1.41539216e+00 6.71810627e-01
-1.09414184e+00 3.89470905e-01 4.07629371e-01 -3.51564348e-01
-1.21869564e+00 -1.65113330e-01 4.32108968e-01 -4.44635093e-01
5.45727372e-01 6.12324715e-01 -2.27815270e-01 -3.38703901e-01
-1.23929843e-01 1.74979299e-01 5.87442070e-02 -5.27107179e-01
2.88983919e-02 4.24990773e-01 -1.01984724e-01 -5.18275559e-01
6.61655247e-01 4.55329329e-01 3.89144123e-01 -2.24381179e-01
-4.90877360e-01 -2.80156881e-01 -1.37064055e-01 -4.30293940e-02
5.83342493e-01 -1.53918788e-02 -8.81929934e-01 1.99805751e-01
-8.37536395e-01 6.54832795e-02 -3.31921548e-01 4.55210954e-01
-7.15487838e-01 2.68242389e-01 -2.92366683e-01 -1.20846248e+00
-3.43573242e-01 -1.13107860e+00 2.05437124e-01 1.87084645e-01
1.56815544e-01 -8.84249032e-01 3.64808798e-01 5.01630008e-01
4.69230533e-01 4.06694174e-01 7.99419940e-01 -1.07216406e+00
-8.43600035e-01 -1.71902284e-01 -1.49978814e-03 -2.78068129e-02
-1.76590346e-02 -2.10251927e-01 -3.16218495e-01 -2.90056974e-01
5.88242888e-01 4.64614689e-01 3.37844461e-01 3.34796935e-01
6.85903609e-01 -6.63450718e-01 -1.82247937e-01 6.09054923e-01
1.72244942e+00 9.91838813e-01 3.75602543e-01 9.95493710e-01
-9.59884599e-02 7.31224656e-01 1.13334596e+00 8.02499890e-01
-2.01655030e-02 7.53674805e-01 6.63120091e-01 3.71653765e-01
8.37521851e-01 2.92479366e-01 -1.38213605e-01 2.11331606e-01
-2.03141883e-01 -4.99127418e-01 -9.38418210e-01 1.50228530e-01
-2.20222902e+00 -1.04608345e+00 3.19303870e-02 2.36385155e+00
4.93013710e-01 4.59941089e-01 1.99011415e-01 6.12531900e-02
7.22601891e-01 -8.95265564e-02 -3.04404825e-01 -9.93669629e-01
-1.77163482e-01 3.01795512e-01 1.16728425e+00 5.98976254e-01
-7.18098283e-01 6.49390221e-01 6.62492514e+00 8.21265876e-01
-9.39467549e-01 -7.84839243e-02 6.61221683e-01 -4.24857974e-01
-5.13464868e-01 3.64805341e-01 -8.18636715e-01 7.27151453e-01
8.19656372e-01 -5.56795120e-01 7.22202420e-01 5.55633008e-01
1.29123315e-01 -4.23492074e-01 -4.55260724e-01 7.53768146e-01
-9.60975438e-02 -1.28618157e+00 -1.63446441e-01 5.77176154e-01
8.80253077e-01 -6.30583167e-01 1.59473330e-01 -1.73620924e-01
2.55536228e-01 -7.84653664e-01 7.75336862e-01 8.82295966e-01
4.85431142e-02 -1.36249256e+00 1.18398154e+00 2.81161070e-01
-9.47397292e-01 -5.30622661e-01 -4.29567039e-01 -2.45228526e-03
3.31415057e-01 5.02928734e-01 -6.51016474e-01 1.08122480e+00
6.68468416e-01 -1.41061410e-01 -1.67074159e-01 1.55001140e+00
1.36333331e-01 4.08000767e-01 -5.20141065e-01 -4.04796094e-01
4.84405041e-01 -7.97871947e-01 7.33879387e-01 5.38189590e-01
6.54823601e-01 2.75469154e-01 -1.32840425e-01 9.15771604e-01
4.04107094e-01 4.03100878e-01 -3.90604675e-01 3.09563475e-03
2.36134797e-01 9.20743823e-01 -1.12375402e+00 -2.73512118e-02
-3.49470198e-01 5.17306089e-01 -2.94821173e-01 2.70390928e-01
-7.91373551e-01 -2.05543444e-01 2.45124415e-01 2.76806742e-01
5.56793213e-01 6.23272732e-02 -5.20635962e-01 -7.71754920e-01
2.55131796e-02 -9.01318967e-01 5.90814471e-01 -1.70640647e-01
-8.38260531e-01 6.17877245e-01 5.21555662e-01 -1.00562847e+00
-4.87696171e-01 -5.46729207e-01 -7.20179975e-01 8.81807268e-01
-1.34272254e+00 -3.55975509e-01 4.10496563e-01 2.83055812e-01
2.95869887e-01 -4.99111712e-01 2.87744612e-01 -5.42858355e-02
-5.25310159e-01 1.69560149e-01 7.75323451e-01 -6.15703881e-01
1.70084596e-01 -1.36982381e+00 -1.34992898e-01 7.55589306e-01
1.67687004e-03 4.96689171e-01 1.11785257e+00 -7.83956707e-01
-1.37813783e+00 -4.07103032e-01 6.63323820e-01 1.17614537e-01
1.08225739e+00 -1.78429455e-01 -6.79507852e-01 2.76404798e-01
6.43404663e-01 -3.75757307e-01 4.76408124e-01 -1.12801589e-01
2.57160813e-01 -5.11840761e-01 -1.26157904e+00 2.21672863e-01
6.02659702e-01 1.96440756e-01 -6.39199138e-01 1.70482695e-01
3.09087098e-01 -3.07443559e-01 -5.68033218e-01 4.16103721e-01
3.86236131e-01 -9.82373536e-01 6.32001221e-01 -1.84075549e-01
-3.73237133e-01 -8.50750953e-02 5.56784980e-02 -9.53863561e-01
-1.72931001e-01 -1.09780133e+00 -1.06867582e-01 1.15991950e+00
6.12378776e-01 -1.13832831e+00 8.01693141e-01 7.89428949e-01
3.76297265e-01 -1.05344212e+00 -1.28024650e+00 -1.01190364e+00
8.63781720e-02 3.16849768e-01 8.24049115e-01 5.46741664e-01
5.65767363e-02 -3.76734465e-01 -3.93625855e-01 -4.40086961e-01
8.60723197e-01 7.64679313e-01 1.15221851e-01 -1.11445284e+00
-6.19706631e-01 -8.51398408e-01 -8.61371011e-02 -5.20087421e-01
-4.83955666e-02 -6.56239331e-01 -2.37966955e-01 -1.14001679e+00
8.45015869e-02 -3.34348440e-01 -5.77040136e-01 -6.76311255e-02
3.01768899e-01 -3.45994323e-01 5.32974005e-01 7.62671605e-02
-3.76945555e-01 3.60748470e-01 9.62189257e-01 -7.28479922e-02
-2.80869365e-01 4.96021122e-01 -6.22648418e-01 3.72704029e-01
9.81024325e-01 -7.14883745e-01 -5.20151556e-01 3.05611640e-01
6.04828835e-01 6.58656538e-01 -2.61510551e-01 -5.56851923e-01
3.05484354e-01 -5.40686727e-01 -2.15276986e-01 -8.83933008e-01
2.04511121e-01 -1.10747671e+00 8.07381988e-01 6.20078206e-01
-8.00116882e-02 5.89608192e-01 3.01461387e-02 5.15663266e-01
-3.47038597e-01 -9.50781286e-01 5.46625853e-01 -3.46004575e-01
-2.23387226e-01 1.18944407e-01 -1.26231119e-01 -1.42539084e-01
1.48101425e+00 -3.52871120e-01 2.90071853e-02 -3.33506852e-01
-7.59862542e-01 6.59458041e-01 3.27737778e-01 4.07065302e-02
3.15891355e-01 -1.03823400e+00 -5.16149700e-01 -5.03666282e-01
-5.41007340e-01 -4.33606416e-01 8.14854875e-02 7.39296973e-01
-9.31357741e-01 7.73193121e-01 -4.28924054e-01 4.17579003e-02
-1.10506296e+00 8.53071570e-01 3.89178813e-01 -9.31516230e-01
-2.05127165e-01 4.13127929e-01 1.42434025e-02 -1.15617001e-02
1.47332415e-01 4.55720387e-02 -4.55002755e-01 4.82961357e-01
2.25712746e-01 7.25055218e-01 -1.77974686e-01 -2.64185488e-01
-2.01123834e-01 3.32968801e-01 7.50401467e-02 -4.76203740e-01
1.48548663e+00 -3.29180241e-01 -5.73954940e-01 3.10054630e-01
6.53806806e-01 1.25746503e-01 -7.38875031e-01 -3.44005763e-03
5.93717158e-01 -4.45504963e-01 -6.00829422e-02 -6.90788746e-01
-1.16128385e+00 3.41059536e-01 3.62142235e-01 6.75825834e-01
1.23584390e+00 -5.54467618e-01 5.34413755e-01 3.48731607e-01
7.16700494e-01 -1.47480202e+00 -6.54759824e-01 4.07574415e-01
9.57971454e-01 -6.15984797e-01 1.55285522e-01 -2.91392982e-01
-4.39548731e-01 1.22358322e+00 2.92160600e-01 -2.15474725e-01
7.65725017e-01 5.38408935e-01 -2.81060368e-01 1.49690002e-01
-7.59099841e-01 -2.00695202e-01 2.82931536e-01 -2.23368913e-01
-1.24704227e-01 -6.01628572e-02 -1.41941106e+00 7.63725579e-01
-2.52784133e-01 -2.37204999e-01 6.23840451e-01 1.02265489e+00
-4.96108294e-01 -1.65265739e+00 -7.26560056e-01 4.11207885e-01
-8.86827469e-01 -1.15664355e-01 -4.79604840e-01 9.24564838e-01
2.02031452e-02 8.19851458e-01 -9.98570174e-02 -4.77133803e-02
1.80732265e-01 -1.03755347e-01 2.70798862e-01 -3.01228493e-01
-9.29549813e-01 3.83981794e-01 4.76487055e-02 -5.67556024e-01
-1.47885814e-01 -9.31418300e-01 -9.60238039e-01 -2.42147714e-01
-4.00671095e-01 8.67783725e-01 5.50681174e-01 7.82864094e-01
-1.64135650e-01 3.12157631e-01 8.30262840e-01 -1.99879974e-01
-1.06364381e+00 -4.47591275e-01 -9.82355952e-01 -2.86110461e-01
-3.03633362e-01 -7.33795404e-01 -5.96999645e-01 -8.61223221e-01] | [5.014044761657715, 3.952735662460327] |
5b17dac5-2aab-4875-970d-81f935830f72 | self-supervised-3d-keypoint-learning-for-ego | 1912.03426 | null | https://arxiv.org/abs/1912.03426v3 | https://arxiv.org/pdf/1912.03426v3.pdf | Self-Supervised 3D Keypoint Learning for Ego-motion Estimation | Detecting and matching robust viewpoint-invariant keypoints is critical for visual SLAM and Structure-from-Motion. State-of-the-art learning-based methods generate training samples via homography adaptation to create 2D synthetic views with known keypoint matches from a single image. This approach, however, does not generalize to non-planar 3D scenes with illumination variations commonly seen in real-world videos. In this work, we propose self-supervised learning of depth-aware keypoints directly from unlabeled videos. We jointly learn keypoint and depth estimation networks by combining appearance and geometric matching via a differentiable structure-from-motion module based on Procrustean residual pose correction. We describe how our self-supervised keypoints can be integrated into state-of-the-art visual odometry frameworks for robust and accurate ego-motion estimation of autonomous vehicles in real-world conditions. | ['Patric Jensfelt', 'Rares Ambrus', 'Adrien Gaidon', 'Sudeep Pillai', 'Jiexiong Tang', 'Vitor Guizilini', 'Hanme Kim'] | 2019-12-07 | null | null | null | null | ['geometric-matching'] | ['computer-vision'] | [-1.71535090e-02 1.17823355e-01 -3.94118458e-01 -5.90004146e-01
-6.76556587e-01 -6.46265745e-01 7.82362223e-01 -4.04243976e-01
-2.91711271e-01 5.23412585e-01 -3.11612278e-01 -6.37443960e-02
2.10058212e-01 -5.67987919e-01 -1.23997390e+00 -3.16687346e-01
1.49112806e-01 9.53347325e-01 3.73255312e-01 -1.66358531e-01
3.50159883e-01 9.63982165e-01 -1.71071076e+00 -2.77148694e-01
5.68628252e-01 4.58512068e-01 1.89618826e-01 6.81189060e-01
1.25051796e-01 6.93404734e-01 -5.89505620e-02 -5.80500811e-02
6.82343781e-01 -2.87572443e-01 -2.15300411e-01 4.30363774e-01
1.48237205e+00 -6.52916610e-01 -8.79870951e-01 1.15331137e+00
2.85994215e-03 7.15204999e-02 5.26793957e-01 -1.54924500e+00
-1.81813002e-01 -5.16937017e-01 -5.56268036e-01 -1.79434583e-01
6.97625399e-01 3.24813664e-01 6.68157101e-01 -1.05275309e+00
1.37049413e+00 1.42153788e+00 7.52565503e-01 5.04860997e-01
-1.25427496e+00 -5.86383164e-01 -2.55111512e-02 3.39043349e-01
-1.31682277e+00 -7.91625619e-01 8.89048755e-01 -5.96311510e-01
1.03189349e+00 -3.47839177e-01 7.76074708e-01 8.34844649e-01
4.65764016e-01 5.98077118e-01 7.99251139e-01 -2.68117875e-01
1.59415990e-01 -3.85656059e-02 -5.24432123e-01 1.17304802e+00
6.27554178e-01 5.35163164e-01 -7.30718255e-01 -1.86866187e-02
1.24177504e+00 3.46708030e-01 -7.02518821e-02 -1.62810922e+00
-1.48250329e+00 8.69506896e-01 4.58614320e-01 -2.82024920e-01
-2.86175758e-01 5.62657595e-01 -1.25017822e-01 3.24185193e-01
2.96991587e-01 1.85930237e-01 -2.99861908e-01 8.52511302e-02
-9.40212071e-01 4.28601116e-01 6.03283346e-01 1.54432118e+00
1.60945201e+00 3.83174717e-01 5.41431606e-01 1.04930572e-01
7.46633768e-01 1.26308572e+00 2.58854419e-01 -1.66594768e+00
2.70842731e-01 4.67804879e-01 2.89872736e-01 -9.22349274e-01
-3.16697955e-01 -1.73029259e-01 -2.40333140e-01 8.26806128e-01
2.35593498e-01 1.02787890e-01 -1.15759230e+00 1.57765400e+00
6.04128897e-01 3.33564460e-01 3.51287633e-01 8.56347680e-01
7.92230666e-01 2.37677738e-01 -7.83118665e-01 4.75361757e-02
7.03629136e-01 -9.80996728e-01 -4.92486686e-01 -8.78293514e-01
4.24646348e-01 -6.24938190e-01 3.27317894e-01 -3.09983734e-02
-8.83018494e-01 -4.91446167e-01 -1.29401278e+00 -1.89767078e-01
-1.95277445e-02 -6.96966797e-02 4.47122782e-01 3.05580318e-01
-1.10767543e+00 2.09183335e-01 -1.01774979e+00 -6.86278224e-01
-6.62269220e-02 2.65280247e-01 -9.96193945e-01 -3.41864020e-01
-7.50939786e-01 1.34518278e+00 1.73053801e-01 -3.21788400e-01
-1.48712432e+00 -6.50508523e-01 -1.51425338e+00 -7.32859313e-01
3.33427310e-01 -1.22832394e+00 1.22434866e+00 -9.10774767e-01
-1.53083229e+00 1.37864923e+00 -6.13891065e-01 -4.64883268e-01
6.78094983e-01 -2.31224403e-01 1.26089836e-02 3.91619712e-01
3.69189382e-01 1.02570450e+00 1.08555603e+00 -1.85063612e+00
-7.24068940e-01 -6.53460085e-01 -1.41748354e-01 4.97451454e-01
7.53100753e-01 -6.48420632e-01 -5.18721163e-01 8.42777565e-02
8.90304804e-01 -1.18973076e+00 -2.86835551e-01 5.48661649e-01
7.38078505e-02 5.66072226e-01 1.06918669e+00 -2.67295063e-01
-7.86971077e-02 -1.62783253e+00 2.11819544e-01 7.75181651e-02
1.95392221e-01 -9.26839635e-02 -2.10947335e-01 4.29010779e-01
2.94118911e-01 -8.41413379e-01 2.04830587e-01 -5.86058915e-01
-1.23538688e-01 5.57990968e-01 -2.39341050e-01 1.18626046e+00
5.26952334e-02 9.49831545e-01 -1.17417979e+00 -1.35639817e-01
8.30787838e-01 4.61617440e-01 -4.72255975e-01 4.13112968e-01
-2.22130433e-01 6.15063369e-01 -3.15917969e-01 7.75400639e-01
8.31907392e-01 2.12758556e-01 -1.52840778e-01 -2.33623698e-01
-2.53196627e-01 1.13971107e-01 -1.13263333e+00 2.26036286e+00
-3.37062120e-01 9.67196822e-01 1.58722550e-01 -6.21777177e-01
9.12992835e-01 -1.77722916e-01 5.37791967e-01 -5.75122356e-01
1.67355267e-03 3.22692335e-01 -4.73737836e-01 -3.28244627e-01
7.50773609e-01 -2.73586214e-02 2.55054295e-01 6.88583851e-02
3.97129655e-01 -8.49402547e-01 -6.83008507e-02 3.15178335e-01
9.35090303e-01 6.20608211e-01 5.73295057e-01 4.62163649e-02
3.83871883e-01 1.79860517e-01 7.04213560e-01 4.69189107e-01
-1.78521857e-01 7.60529339e-01 -1.67249531e-01 -6.50005698e-01
-1.56337142e+00 -1.28289294e+00 3.52950655e-02 2.42343962e-01
6.68139040e-01 -1.10586867e-01 -3.24512303e-01 -5.12184322e-01
5.83481193e-01 2.64633000e-01 -4.38300192e-01 -4.60551903e-02
-4.86930966e-01 1.90124929e-01 3.11040252e-01 1.31726250e-01
4.29494381e-01 -5.17129958e-01 -1.00662911e+00 8.19581375e-02
2.17140839e-03 -1.48118281e+00 -2.74423569e-01 5.24113327e-02
-8.71675611e-01 -1.08638299e+00 -6.34014606e-01 -6.62636876e-01
8.47556710e-01 9.88278031e-01 1.04202223e+00 -2.77307928e-01
-1.02147702e-02 9.50062037e-01 -3.70981358e-02 -2.11303547e-01
-6.59659743e-01 -4.12404388e-01 5.07434130e-01 -3.09588648e-02
4.02104825e-01 -5.70576370e-01 -5.67959905e-01 4.98525411e-01
-4.76855695e-01 2.01346636e-01 6.65424228e-01 6.09737158e-01
8.92695129e-01 -7.49283195e-01 -6.56415746e-02 -3.47991884e-01
-4.71740097e-01 -1.06893189e-01 -1.07466555e+00 -1.15811855e-01
-5.28944314e-01 1.91140428e-01 1.42184362e-01 -1.66277990e-01
-7.03977764e-01 7.21690536e-01 2.79936045e-01 -1.02303421e+00
-2.85031110e-01 -1.43581882e-01 -1.91242516e-01 -8.18535388e-01
6.91092014e-01 2.47776195e-01 4.91464645e-01 1.16048716e-01
6.94296002e-01 3.01609337e-01 7.71479607e-01 -1.80222064e-01
1.51348078e+00 1.24019980e+00 3.95147711e-01 -9.19343233e-01
-4.96263295e-01 -1.01153576e+00 -1.06817973e+00 -2.69194454e-01
8.98708761e-01 -1.53170705e+00 -3.77249837e-01 4.75951016e-01
-1.18723369e+00 -3.31750572e-01 -1.53355256e-01 8.04565847e-01
-1.21371615e+00 6.01082742e-01 -9.57448781e-02 -5.76232314e-01
-4.44672778e-02 -1.19683707e+00 1.45302129e+00 2.24078730e-01
6.61280975e-02 -9.65448856e-01 4.74933147e-01 1.93244845e-01
-5.33330776e-02 2.94232100e-01 -8.34828690e-02 -3.91957089e-02
-1.35662007e+00 -2.40469292e-01 -4.77455780e-02 2.84903180e-02
3.49435098e-02 -1.03778481e-01 -8.87630701e-01 -5.54759383e-01
-8.31597671e-02 -3.88722450e-01 8.73169780e-01 3.93338084e-01
-2.72018462e-01 6.66728690e-02 -4.79517698e-01 1.22812092e+00
1.54291904e+00 7.72148669e-02 5.31562567e-01 6.25299811e-01
1.06487310e+00 5.81468821e-01 8.21070790e-01 1.67239279e-01
8.49682629e-01 7.71063566e-01 8.66277933e-01 -5.35277985e-02
-4.09286432e-02 -6.98525727e-01 5.72532058e-01 2.58114606e-01
2.30286092e-01 4.03896242e-01 -8.59426558e-01 7.04300284e-01
-1.91571522e+00 -9.38989401e-01 -6.88700750e-02 2.35673499e+00
2.62049317e-01 -1.83499809e-02 -3.42007697e-01 -4.27503914e-01
3.64608586e-01 2.67037243e-01 -8.22302401e-01 2.18167260e-01
-2.11163715e-01 -5.72060168e-01 1.09786558e+00 8.73213470e-01
-9.23013985e-01 1.33146119e+00 5.61985970e+00 -2.42331401e-02
-1.06658626e+00 2.67914105e-02 -1.26576826e-01 2.67498046e-01
-6.12103522e-01 4.93533462e-01 -1.09865451e+00 -2.84327328e-01
5.62371254e-01 -3.49329673e-02 2.84349024e-01 1.08444679e+00
8.17117617e-02 -2.64582217e-01 -1.38904822e+00 1.44603848e+00
6.62038743e-01 -1.36521745e+00 3.08810920e-02 2.97353059e-01
1.27780664e+00 7.51337886e-01 -1.63103506e-01 -1.98230401e-01
4.26243961e-01 -4.35254514e-01 9.30472136e-01 4.37651604e-01
6.89256310e-01 -4.91963446e-01 3.60190839e-01 4.22298700e-01
-1.21897173e+00 1.80811733e-01 -5.19347847e-01 5.86813763e-02
5.04063725e-01 3.16308498e-01 -9.01685417e-01 7.08437800e-01
4.43482339e-01 1.25927448e+00 -5.09942472e-01 1.01527226e+00
-3.69170129e-01 -3.65283877e-01 -3.71654928e-01 4.90859985e-01
3.05402964e-01 -3.34390491e-01 9.06595886e-01 7.38885641e-01
5.11724293e-01 -2.78741509e-01 1.61375389e-01 9.30449426e-01
1.66124597e-01 -4.55833852e-01 -1.31193864e+00 5.26239455e-01
4.82763797e-01 1.16937506e+00 -4.41222221e-01 -5.03503919e-01
-4.69290793e-01 1.03542960e+00 1.21229552e-01 4.93904620e-01
-4.78459597e-01 1.01073887e-02 1.05338359e+00 3.41973573e-01
4.12072957e-01 -8.94125879e-01 2.44880229e-01 -1.62271166e+00
-6.35043308e-02 -5.80094337e-01 -2.11952969e-01 -1.12378263e+00
-8.55694830e-01 2.50509202e-01 2.17529759e-01 -1.70959842e+00
-7.85481691e-01 -5.87283134e-01 -3.17820907e-01 5.93804836e-01
-1.90903103e+00 -1.47689044e+00 -8.24019432e-01 6.21384323e-01
6.11218572e-01 -1.77026510e-01 5.43093443e-01 -2.64515132e-01
3.99546742e-01 1.45391867e-01 4.10888344e-01 -1.07083917e-01
9.62252378e-01 -1.12440503e+00 1.05337131e+00 1.00067580e+00
4.38036352e-01 4.63176548e-01 8.87162805e-01 -6.09876156e-01
-2.03670239e+00 -1.03434682e+00 6.09926581e-01 -7.35607684e-01
4.20030564e-01 -3.49865705e-01 -6.39381349e-01 9.92498279e-01
-2.52769627e-02 4.37176168e-01 -2.41689131e-01 -5.76936185e-01
-4.81336117e-01 -2.78288364e-01 -1.08543324e+00 4.47340041e-01
1.18171489e+00 -8.70197237e-01 -6.02869272e-01 8.05561468e-02
5.44201851e-01 -8.72524321e-01 -4.19102490e-01 5.49923956e-01
8.26329708e-01 -1.00732291e+00 1.25953233e+00 -1.45146132e-01
-1.29336864e-01 -7.88585722e-01 -2.70013660e-01 -1.35455394e+00
-5.46401069e-02 -7.69519210e-01 -8.98537710e-02 6.43824637e-01
-5.57843298e-02 -7.36518025e-01 1.02170134e+00 1.42691910e-01
-2.50657827e-01 5.99921159e-02 -1.01378620e+00 -8.67324889e-01
-2.51355499e-01 -2.74425328e-01 1.33827016e-01 8.24758410e-01
-4.02573526e-01 2.44254112e-01 -4.92782503e-01 4.01926577e-01
1.34157753e+00 1.00346841e-01 1.76381636e+00 -1.20123696e+00
-1.15567289e-01 1.11185856e-01 -1.30674624e+00 -1.52835894e+00
5.56203902e-01 -5.75549245e-01 3.47239226e-01 -1.56809545e+00
-8.44652578e-02 -7.29975477e-02 5.15955865e-01 4.09415970e-03
2.64059186e-01 3.53162944e-01 1.49330705e-01 3.97472560e-01
-6.34089649e-01 6.63374186e-01 1.18562603e+00 -5.90004921e-02
-1.67206421e-01 -1.41244218e-01 2.20361963e-01 8.38384688e-01
3.69582623e-01 -3.93547028e-01 -6.00973368e-01 -5.58509767e-01
2.02468306e-01 3.03703994e-01 5.92262924e-01 -1.12067819e+00
4.44959253e-01 -3.43214005e-01 4.11171138e-01 -9.31649804e-01
7.08557725e-01 -9.31293249e-01 4.33725029e-01 4.16972190e-01
1.57601580e-01 2.58394420e-01 4.45345603e-02 7.80481100e-01
-1.08684361e-01 -7.50121772e-02 6.84551895e-01 -4.69141215e-01
-1.26709461e+00 5.47662973e-01 -2.40484491e-01 -4.58369777e-02
1.06822741e+00 -6.75461769e-01 -2.90760368e-01 -7.19400883e-01
-1.31789342e-01 4.95966077e-02 1.43586683e+00 5.50461590e-01
1.14381921e+00 -1.41206443e+00 -6.29978836e-01 8.46059740e-01
6.61952078e-01 3.30023021e-01 2.16599315e-01 6.30979180e-01
-1.05863059e+00 5.79605758e-01 -5.02196014e-01 -1.35512805e+00
-1.28683281e+00 4.38842773e-01 6.08650029e-01 3.99131447e-01
-7.17153072e-01 3.44185203e-01 4.03576493e-01 -8.77904594e-01
7.76494890e-02 -3.53470773e-01 4.00577366e-01 -4.51581985e-01
3.61795545e-01 2.02266634e-01 -2.36724373e-02 -1.39122438e+00
-4.07688141e-01 1.22904050e+00 1.03945524e-01 -4.71719712e-01
9.32685316e-01 -5.13863266e-01 1.89103484e-01 5.00098109e-01
1.41749132e+00 -8.96604955e-02 -1.91092193e+00 -3.94533634e-01
-2.48114839e-01 -7.69956708e-01 -1.20579377e-01 3.11754551e-02
-7.38398075e-01 7.14009941e-01 6.28967106e-01 -7.13288903e-01
3.57186735e-01 5.38757481e-02 5.19308448e-01 8.35315704e-01
9.12324786e-01 -7.42429018e-01 1.67941198e-01 7.74370670e-01
6.69789732e-01 -1.73703980e+00 2.48143554e-01 -1.38224587e-01
-4.00881350e-01 1.17384613e+00 6.52831018e-01 -4.18694466e-01
2.99620599e-01 6.16884865e-02 4.70063180e-01 -1.60814241e-01
-6.54527783e-01 -3.41901213e-01 3.25605512e-01 1.08428001e+00
-2.22525492e-01 -2.84323424e-01 3.40558976e-01 -7.75366902e-01
-1.06863317e-03 -3.05746913e-01 7.90198088e-01 1.01858783e+00
-6.93118274e-01 -9.74532723e-01 -4.36598152e-01 -3.00307900e-01
2.31155232e-01 1.33663923e-01 -3.34166676e-01 8.92372727e-01
-1.44145414e-01 6.18696809e-01 1.36298984e-01 -4.51386184e-01
3.25228125e-01 -1.40573785e-01 9.93853569e-01 -5.79166353e-01
2.62457579e-01 -1.96392033e-02 -5.22496849e-02 -9.68379021e-01
-7.97118783e-01 -7.79814124e-01 -1.40367198e+00 -1.63965240e-01
-3.91868576e-02 -2.74974674e-01 8.81787956e-01 8.14835131e-01
3.62406075e-01 -1.88505858e-01 5.61922431e-01 -1.73990226e+00
-3.80897135e-01 -6.40104413e-01 -4.25247669e-01 4.35376316e-01
9.65681970e-01 -7.74967790e-01 -5.86552739e-01 1.05476372e-01] | [8.088513374328613, -2.235980749130249] |
0a90f7c2-5018-4a64-ab2b-7ade1de2d5d9 | uvstyle-net-unsupervised-few-shot-learning-of | 2105.02961 | null | https://arxiv.org/abs/2105.02961v3 | https://arxiv.org/pdf/2105.02961v3.pdf | UVStyle-Net: Unsupervised Few-shot Learning of 3D Style Similarity Measure for B-Reps | Boundary Representations (B-Reps) are the industry standard in 3D Computer Aided Design/Manufacturing (CAD/CAM) and industrial design due to their fidelity in representing stylistic details. However, they have been ignored in the 3D style research. Existing 3D style metrics typically operate on meshes or pointclouds, and fail to account for end-user subjectivity by adopting fixed definitions of style, either through crowd-sourcing for style labels or hand-crafted features. We propose UVStyle-Net, a style similarity measure for B-Reps that leverages the style signals in the second order statistics of the activations in a pre-trained (unsupervised) 3D encoder, and learns their relative importance to a subjective end-user through few-shot learning. Our approach differs from all existing data-driven 3D style methods since it may be used in completely unsupervised settings, which is desirable given the lack of publicly available labelled B-Rep datasets. More importantly, the few-shot learning accounts for the inherent subjectivity associated with style. We show quantitatively that our proposed method with B-Reps is able to capture stronger style signals than alternative methods on meshes and pointclouds despite its significantly greater computational efficiency. We also show it is able to generate meaningful style gradients with respect to the input shape, and that few-shot learning with as few as two positive examples selected by an end-user is sufficient to significantly improve the style measure. Finally, we demonstrate its efficacy on a large unlabeled public dataset of CAD models. Source code and data will be released in the future. | ['Joseph Lambourne', 'Aditya Sanghi', 'Pradeep Kumar Jayaraman', 'Amir Khasahmadi', 'Hooman Shayani', 'Peter Meltzer'] | 2021-04-28 | null | http://openaccess.thecvf.com//content/ICCV2021/html/Meltzer_UVStyle-Net_Unsupervised_Few-Shot_Learning_of_3D_Style_Similarity_Measure_for_ICCV_2021_paper.html | http://openaccess.thecvf.com//content/ICCV2021/papers/Meltzer_UVStyle-Net_Unsupervised_Few-Shot_Learning_of_3D_Style_Similarity_Measure_for_ICCV_2021_paper.pdf | iccv-2021-1 | ['unsupervised-few-shot-learning'] | ['computer-vision'] | [-2.40288535e-03 7.36072361e-02 -9.38867256e-02 -5.94143212e-01
-7.34757602e-01 -5.01893580e-01 7.63999462e-01 -1.95968114e-02
5.86239956e-02 3.02018434e-01 2.54878491e-01 4.86662164e-02
-4.88376953e-02 -8.34204614e-01 -4.72132444e-01 -4.21782315e-01
1.97734073e-01 6.09474361e-01 4.85566296e-02 -4.10236120e-01
2.68923253e-01 6.45922244e-01 -1.50959444e+00 2.61270963e-02
6.70653343e-01 1.14385712e+00 1.53418913e-01 4.52209860e-01
-3.11067671e-01 4.26319838e-01 -4.43803817e-01 -3.51009369e-01
4.09503996e-01 -4.01659697e-01 -5.83378434e-01 4.10547137e-01
5.44610679e-01 -9.85862687e-02 3.98982614e-02 9.86557126e-01
7.15031445e-01 2.44895771e-01 9.32779491e-01 -1.07170975e+00
-7.23757923e-01 1.81354657e-02 -4.55591232e-01 -2.12326422e-01
3.25813115e-01 4.24851686e-01 1.13016260e+00 -1.32244706e+00
1.04023623e+00 1.41776073e+00 7.40510643e-01 6.46663547e-01
-1.37635410e+00 -4.61848974e-01 -9.53694060e-02 -1.95975527e-01
-1.20257437e+00 -3.79821092e-01 1.47512162e+00 -6.87391043e-01
7.38557518e-01 6.63664639e-02 6.58975661e-01 1.50243258e+00
1.15895554e-01 7.94066250e-01 1.14745426e+00 -3.11224043e-01
7.88985670e-01 7.36072883e-02 -3.64932604e-03 6.53258145e-01
-1.07689001e-01 2.39272669e-01 -6.04192853e-01 -2.40124881e-01
1.13056648e+00 -5.41282585e-03 2.87941415e-02 -9.92981315e-01
-1.17133558e+00 8.55467975e-01 5.18433452e-01 1.91734478e-01
-1.58852994e-01 -8.75377208e-02 4.18313622e-01 5.32901406e-01
9.54195917e-01 6.97119653e-01 -4.37768728e-01 -4.11760956e-01
-9.88660455e-01 3.16068709e-01 7.39609301e-01 1.36578178e+00
9.47008669e-01 1.96092933e-01 -2.35715896e-01 1.14940929e+00
6.94403797e-02 4.09915447e-01 2.55572051e-01 -1.14345443e+00
-1.79449171e-02 7.06336439e-01 8.89782906e-02 -9.41386163e-01
-1.63127482e-01 -1.78522423e-01 -7.48096347e-01 7.99489498e-01
3.17055315e-01 -1.23749837e-01 -1.09266305e+00 1.58577251e+00
2.92342961e-01 1.18495598e-01 -3.01410079e-01 1.00717521e+00
7.33999074e-01 1.71989426e-01 -9.79192480e-02 2.33145609e-01
1.15161645e+00 -6.69786990e-01 -5.28132737e-01 -1.59703463e-01
6.89500928e-01 -8.60550761e-01 1.50658631e+00 2.70092189e-01
-1.05783319e+00 -7.06848979e-01 -1.11386371e+00 -1.44766569e-01
-4.27869409e-01 -1.01466246e-01 5.42650938e-01 6.46243393e-01
-8.09937298e-01 1.07209253e+00 -7.67320216e-01 -5.29916227e-01
9.19682264e-01 8.65322426e-02 -2.08707079e-01 -9.21063870e-02
-7.89402902e-01 8.74935746e-01 -2.76539594e-01 -3.39352727e-01
-8.34436893e-01 -1.07447624e+00 -1.07693994e+00 -2.85562247e-01
3.89924467e-01 -8.84743750e-01 1.21644402e+00 -1.01589417e+00
-1.89063632e+00 1.15283275e+00 -9.03323740e-02 8.48696679e-02
7.10866570e-01 -2.94231027e-01 -2.23715797e-01 -1.43966109e-01
2.95937568e-01 6.45738482e-01 9.29354191e-01 -1.48898160e+00
-2.41979957e-01 -3.72731864e-01 1.47800982e-01 -7.71711469e-02
-1.85418278e-01 -2.41305262e-01 -3.83057833e-01 -1.12985981e+00
2.90631372e-02 -9.03500199e-01 -3.91996324e-01 6.00052238e-01
-1.81668907e-01 -1.71235919e-01 9.16125655e-01 -8.16324055e-02
7.29915082e-01 -2.33763003e+00 3.70609671e-01 1.92917034e-01
1.63280219e-01 2.20170822e-02 -1.67496353e-01 2.82992750e-01
2.08818465e-01 1.09563909e-01 -4.52065915e-01 -5.59412360e-01
1.31670326e-01 3.06248337e-01 7.51870498e-02 4.08462435e-01
6.13879561e-01 9.65251327e-01 -1.28716898e+00 -3.97200465e-01
5.70141077e-01 3.95377785e-01 -6.53571725e-01 2.36151949e-01
-3.48673433e-01 5.14231384e-01 -3.79181892e-01 8.32804263e-01
5.11387348e-01 -2.29672611e-01 -1.19317092e-01 -2.42225647e-01
2.40782887e-01 -1.77289933e-01 -1.26578176e+00 2.32928705e+00
-8.03654432e-01 4.26344931e-01 4.15749662e-03 -5.38461208e-01
1.13491762e+00 3.91310215e-01 5.29221654e-01 -5.06839633e-01
4.03709650e-01 5.17042339e-01 -2.92100459e-01 -2.59520531e-01
2.81169951e-01 -4.44103301e-01 -2.28936791e-01 3.73485953e-01
4.93790001e-01 -8.24189365e-01 -1.03609487e-01 -4.35792021e-02
9.85729992e-01 2.30789989e-01 1.97670057e-01 -5.43569982e-01
3.11432034e-01 -1.01422578e-01 6.28692508e-01 5.24810731e-01
-8.88059437e-02 1.09802544e+00 2.87258863e-01 -5.94507873e-01
-1.36711025e+00 -1.14359009e+00 -3.63942832e-02 8.38455021e-01
3.29940498e-01 -3.66154611e-01 -5.18823624e-01 -7.32883692e-01
2.05838799e-01 7.46418774e-01 -8.75087798e-01 -1.44489706e-01
-3.81108731e-01 -9.63306986e-03 8.84761885e-02 6.54888809e-01
1.68193400e-01 -1.07501519e+00 -5.09982169e-01 2.23290861e-01
5.87598801e-01 -9.14238155e-01 -4.58169907e-01 1.60949498e-01
-1.04556668e+00 -1.01672947e+00 -1.08688688e+00 -6.63528442e-01
6.86279178e-01 1.24972746e-01 1.30993044e+00 -1.03847057e-01
-1.67823702e-01 4.98316258e-01 -3.16098094e-01 -5.00050128e-01
-4.00372386e-01 1.02830365e-01 5.79401068e-02 6.07652999e-02
4.22522366e-01 -9.26127136e-01 -4.86294746e-01 3.93116057e-01
-6.61816001e-01 -5.90008385e-02 3.11243594e-01 9.90812123e-01
5.42644083e-01 -3.70849669e-01 5.13765275e-01 -1.26506770e+00
6.14464283e-01 -2.89024055e-01 -1.95492491e-01 -1.95162922e-01
-5.47648370e-01 7.08780661e-02 4.06221509e-01 -4.70171750e-01
-1.04238176e+00 -6.39017746e-02 -1.10990442e-01 -1.02836728e+00
-4.70444292e-01 1.02704227e-01 -1.85832545e-01 -2.65293390e-01
8.43368173e-01 -3.90297055e-01 1.65737197e-01 -7.70463824e-01
5.88711739e-01 4.01335835e-01 2.55814344e-01 -8.74368906e-01
1.04994690e+00 5.59027374e-01 2.31900532e-02 -9.24612999e-01
-7.95913041e-01 -4.72363710e-01 -7.85607636e-01 -2.18925342e-01
6.24451697e-01 -8.50217164e-01 -1.58384800e-01 4.25285488e-01
-1.05644619e+00 -2.96693623e-01 -7.53256977e-01 1.42959684e-01
-9.88015175e-01 1.35027796e-01 -6.23567581e-01 -6.71345770e-01
-2.24136591e-01 -1.18264699e+00 1.50686991e+00 -3.61732617e-02
-8.15558374e-01 -1.12489545e+00 -1.65416561e-02 2.51269434e-02
4.54322815e-01 5.55156171e-01 1.19262886e+00 -2.97729731e-01
-1.12319380e-01 -1.88267037e-01 -2.62195747e-02 3.88345778e-01
6.16060495e-01 -9.46159810e-02 -1.27814019e+00 -1.77472025e-01
-3.84326419e-03 -3.74281704e-01 4.87804085e-01 2.72094339e-01
1.12663066e+00 1.59523994e-01 -4.54522073e-02 4.64385718e-01
1.37081528e+00 -1.32018417e-01 2.66387910e-01 -5.33281602e-02
9.91269171e-01 6.28707707e-01 6.11388087e-01 4.64617312e-01
-9.46373492e-02 7.10666060e-01 2.64923930e-01 -2.98436344e-01
-3.24379355e-01 -2.70434707e-01 1.87961198e-02 7.64817357e-01
-2.94278592e-01 2.20268011e-01 -8.20723176e-01 6.56773686e-01
-1.64861441e+00 -6.56959832e-01 7.41298720e-02 2.10548902e+00
8.39606881e-01 4.80678797e-01 1.39449418e-01 2.44780719e-01
3.71873438e-01 2.57169366e-01 -6.24788344e-01 -5.36624789e-01
-1.01652734e-01 5.83546698e-01 2.30537653e-01 3.72306436e-01
-9.06217039e-01 9.91905987e-01 5.72040033e+00 8.68175745e-01
-1.05045092e+00 1.90303594e-01 5.60596764e-01 -3.32300454e-01
-5.60632706e-01 -2.07725167e-01 -2.95899689e-01 3.89856398e-01
4.45267379e-01 -6.85178814e-03 1.68891817e-01 1.21454608e+00
3.12018126e-01 2.24860623e-01 -1.56325209e+00 1.29474437e+00
-6.22218214e-02 -1.46692801e+00 5.64656183e-02 4.00672555e-02
9.70736086e-01 -1.65083677e-01 -3.78927325e-05 2.34510779e-01
3.64229977e-01 -8.78016055e-01 9.87651110e-01 4.76909310e-01
9.32912171e-01 -7.37005651e-01 5.19397616e-01 1.01148911e-01
-9.24018741e-01 1.07488535e-01 -2.66788900e-01 -2.07508415e-01
3.08389843e-01 8.87598276e-01 -5.61746001e-01 4.74029630e-01
4.92257565e-01 1.05622661e+00 -5.13190389e-01 7.24587023e-01
9.62356292e-03 3.65038991e-01 -2.34560087e-01 -1.36249717e-02
2.92056292e-01 -9.39004421e-02 7.91282952e-01 9.26392615e-01
3.38052034e-01 -2.18362242e-01 2.53325135e-01 1.25619650e+00
-6.35876134e-02 9.19047669e-02 -9.56320703e-01 5.66828400e-02
2.78422683e-01 1.19411933e+00 -7.00774193e-01 -3.47779512e-01
-4.41493094e-01 1.11684334e+00 1.89716205e-01 3.27719808e-01
-2.66707480e-01 -3.84146690e-01 1.11229813e+00 2.88189322e-01
3.70360762e-01 -3.91885757e-01 -9.69332278e-01 -9.59035218e-01
-3.62073109e-02 -8.20532501e-01 -1.32030860e-01 -6.62589014e-01
-1.86660206e+00 4.19851422e-01 -1.51580185e-01 -1.51857972e+00
-1.65145591e-01 -8.48722935e-01 -6.95283115e-01 7.80404091e-01
-1.24743998e+00 -1.38355935e+00 -4.00502890e-01 3.22959453e-01
1.00975037e+00 -1.78753018e-01 9.77587223e-01 2.59166151e-01
-2.24868238e-01 4.27308500e-01 -2.02992991e-01 3.38503905e-03
7.36377895e-01 -1.45807552e+00 7.33158410e-01 3.68136615e-01
2.35993773e-01 4.63251293e-01 9.02861893e-01 -6.55188084e-01
-1.45988870e+00 -9.87549782e-01 5.81126213e-01 -6.58551037e-01
5.03774405e-01 -6.27641439e-01 -8.77911210e-01 2.40757033e-01
-7.75154158e-02 2.81680971e-01 5.67329347e-01 3.11700284e-01
-3.03669751e-01 2.02709809e-01 -1.24646556e+00 8.23005021e-01
1.50163782e+00 -7.33390391e-01 -7.01237917e-01 2.01266304e-01
5.05149066e-01 -3.77864927e-01 -1.14668691e+00 5.32157123e-01
4.23060864e-01 -8.92925978e-01 8.01632345e-01 -5.67773342e-01
7.60824919e-01 -1.25327975e-01 -2.05036253e-01 -1.48084247e+00
-3.99609119e-01 -5.73627055e-01 -9.17085707e-02 1.29421711e+00
2.67451555e-01 -9.44221616e-02 9.21112418e-01 5.22032976e-01
-2.45880812e-01 -9.86139834e-01 -8.75154853e-01 -8.34805071e-01
2.43398353e-01 -5.77767730e-01 7.24912465e-01 1.03177345e+00
-2.55813628e-01 4.05731857e-01 -3.37685823e-01 -2.45863780e-01
6.36553526e-01 1.61708549e-01 9.23937380e-01 -1.58678448e+00
-2.08160594e-01 -5.80467880e-01 -5.80211282e-01 -8.16166997e-01
1.35688171e-01 -8.81721675e-01 7.69102499e-02 -1.18558013e+00
-3.07454884e-01 -7.07780898e-01 -1.28244147e-01 1.55744523e-01
-5.90260364e-02 3.50856394e-01 4.26727952e-03 3.44566479e-02
-2.40591779e-01 8.13368618e-01 1.51734865e+00 -4.17022616e-01
-2.78775960e-01 -6.99050128e-02 -5.83890557e-01 8.89108539e-01
5.87211788e-01 -2.54513174e-01 -4.41589624e-01 -3.92710418e-01
-2.64731515e-02 -2.94729352e-01 3.99246544e-01 -9.91723120e-01
-1.23363093e-01 -1.09733231e-01 4.45263147e-01 -2.31368005e-01
4.16474044e-01 -8.86181116e-01 8.74816254e-02 -4.45319489e-02
-2.19350323e-01 -1.84223920e-01 -1.30378939e-02 6.66963935e-01
-1.12724774e-01 -1.72900513e-01 9.54937756e-01 -4.10089672e-01
-7.75061548e-01 4.60469663e-01 -9.44149047e-02 1.35485634e-01
8.15830171e-01 -5.33712208e-01 3.76489192e-01 -1.72770277e-01
-4.88645673e-01 -4.97259498e-02 1.09944773e+00 7.28376210e-01
5.37561536e-01 -1.81907034e+00 -4.11209822e-01 4.19243693e-01
5.91568172e-01 2.99945563e-01 6.89588264e-02 4.37533140e-01
-3.22834074e-01 -1.28841475e-01 -3.34437251e-01 -7.81695724e-01
-8.41300488e-01 4.69357193e-01 2.16677636e-01 1.02136768e-01
-9.17524457e-01 8.82051647e-01 1.01228394e-01 -5.07948160e-01
1.47842288e-01 -4.32603300e-01 2.48699144e-01 2.70518735e-02
2.76225865e-01 4.12828982e-01 2.28200749e-01 -5.89732766e-01
-2.39710465e-01 8.87477160e-01 1.00000173e-01 -7.15890676e-02
1.46740556e+00 9.49236676e-02 3.71293902e-01 9.66319025e-01
1.32320547e+00 -8.06227699e-02 -1.43804538e+00 -3.59070778e-01
1.33934617e-01 -6.32379532e-01 2.05430910e-01 -4.31218237e-01
-1.10604751e+00 1.02331495e+00 6.33751690e-01 -1.55001000e-01
7.44618058e-01 5.88037595e-02 8.44576776e-01 1.53024957e-01
7.09999561e-01 -1.32737315e+00 3.33559960e-01 4.05470192e-01
9.63922977e-01 -1.46731973e+00 -1.28694579e-01 -4.79283780e-01
-6.73869967e-01 9.74973202e-01 4.51606035e-01 -4.32215244e-01
9.45923209e-01 2.05009878e-01 3.39956790e-01 -5.06482542e-01
-3.93860251e-01 -5.22282533e-03 3.71746212e-01 7.56132245e-01
4.93479371e-01 6.02956824e-02 6.83054626e-02 6.21113718e-01
-2.80916989e-01 1.31035164e-01 1.36509538e-01 1.19215810e+00
-1.13266416e-01 -1.23911262e+00 -5.75892590e-02 5.57861865e-01
-1.04653724e-01 1.96021557e-01 -4.89809096e-01 3.65805089e-01
1.47929102e-01 7.67688036e-01 2.80649010e-02 -4.81937736e-01
5.96948206e-01 1.01796620e-01 5.93145609e-01 -1.02048850e+00
-5.63658416e-01 -4.41741645e-02 -9.42905340e-03 -7.44801700e-01
-3.66079926e-01 -4.78786051e-01 -8.28469753e-01 -3.10664326e-01
-1.36565492e-01 -2.86532074e-01 3.73376161e-01 8.70018661e-01
5.01598597e-01 4.97689337e-01 7.95208752e-01 -1.47082973e+00
-3.59415323e-01 -1.02884245e+00 -8.57001424e-01 9.70723867e-01
1.15525752e-01 -1.23697555e+00 -4.17344660e-01 -8.42952053e-04] | [8.539926528930664, -3.078861713409424] |
4bd7d492-22bc-41be-aa0f-b9969d5a6d81 | malleable-2-5d-convolution-learning-receptive | 2007.09365 | null | https://arxiv.org/abs/2007.09365v1 | https://arxiv.org/pdf/2007.09365v1.pdf | Malleable 2.5D Convolution: Learning Receptive Fields along the Depth-axis for RGB-D Scene Parsing | Depth data provide geometric information that can bring progress in RGB-D scene parsing tasks. Several recent works propose RGB-D convolution operators that construct receptive fields along the depth-axis to handle 3D neighborhood relations between pixels. However, these methods pre-define depth receptive fields by hyperparameters, making them rely on parameter selection. In this paper, we propose a novel operator called malleable 2.5D convolution to learn the receptive field along the depth-axis. A malleable 2.5D convolution has one or more 2D convolution kernels. Our method assigns each pixel to one of the kernels or none of them according to their relative depth differences, and the assigning process is formulated as a differentiable form so that it can be learnt by gradient descent. The proposed operator runs on standard 2D feature maps and can be seamlessly incorporated into pre-trained CNNs. We conduct extensive experiments on two challenging RGB-D semantic segmentation dataset NYUDv2 and Cityscapes to validate the effectiveness and the generalization ability of our method. | ['Yajie Xing', 'Jingbo Wang', 'Gang Zeng'] | 2020-07-18 | null | https://www.ecva.net/papers/eccv_2020/papers_ECCV/html/3295_ECCV_2020_paper.php | https://www.ecva.net/papers/eccv_2020/papers_ECCV/papers/123640545.pdf | eccv-2020-8 | ['scene-parsing'] | ['computer-vision'] | [ 6.55019656e-02 2.09191084e-01 -7.49978051e-02 -9.04930949e-01
-8.51815045e-02 -6.44823313e-01 4.17404383e-01 -6.43055961e-02
-5.83562255e-01 1.09368190e-01 7.83266872e-02 -6.04835451e-01
-4.24820371e-02 -1.24775553e+00 -7.04889774e-01 -5.44053912e-01
1.08003631e-01 2.21416190e-01 5.51241398e-01 -2.38190770e-01
1.91627830e-01 9.53459084e-01 -1.34015703e+00 5.99371940e-02
8.33784640e-01 1.26087320e+00 3.56023282e-01 6.52381480e-01
-6.03410244e-01 4.63033199e-01 -2.09463269e-01 -1.65882692e-01
4.75936830e-01 -5.58556989e-02 -9.77893174e-01 1.49866834e-01
3.76845032e-01 -5.38142502e-01 -2.97822535e-01 9.52024937e-01
2.90138960e-01 1.21391870e-01 5.87332368e-01 -1.01301873e+00
-8.10283601e-01 1.94875106e-01 -4.76444542e-01 1.21171653e-01
1.40470579e-01 -1.05347686e-01 7.94143438e-01 -5.57273030e-01
3.62945914e-01 1.25112486e+00 6.73107028e-01 6.17724717e-01
-7.90909410e-01 -4.33741540e-01 3.60685229e-01 -1.86310500e-01
-1.09480715e+00 1.60041198e-01 9.45752323e-01 -3.75531048e-01
1.00348747e+00 3.85025479e-02 7.14435518e-01 6.31971478e-01
-2.32051849e-01 7.33896494e-01 8.99232566e-01 -2.91206986e-01
3.36771488e-01 -1.83411419e-01 7.45107457e-02 8.84084761e-01
4.90969885e-03 -2.49621198e-01 -1.56833962e-01 2.48963535e-01
1.52484906e+00 2.54157752e-01 -8.80187601e-02 -5.28521895e-01
-8.69370520e-01 8.95227194e-01 1.06795430e+00 9.80645642e-02
-2.17881486e-01 2.46213853e-01 -3.24399071e-03 -8.89871046e-02
4.99297887e-01 1.99209422e-01 -5.90823412e-01 4.88170050e-02
-3.10462147e-01 1.20096682e-02 3.66759151e-01 8.60090315e-01
1.15948451e+00 -5.53089142e-01 7.53921643e-02 7.57429540e-01
2.47153550e-01 2.98520952e-01 4.17735934e-01 -9.67208862e-01
6.89394534e-01 1.04733574e+00 -6.29110145e-04 -7.79386103e-01
-6.68057859e-01 1.12834498e-01 -7.17695415e-01 3.50658059e-01
4.53997076e-01 -6.02796488e-02 -1.34848464e+00 1.37669599e+00
5.81746042e-01 1.89691365e-01 1.39245257e-01 1.16340244e+00
1.09911513e+00 5.00175357e-01 1.51747027e-02 5.94917059e-01
1.02655768e+00 -8.51132810e-01 -2.45831773e-01 -3.85974169e-01
6.58221722e-01 -4.67494816e-01 1.18023992e+00 -2.39438973e-02
-8.86849523e-01 -6.34024441e-01 -1.01083279e+00 -5.02857804e-01
-5.62568963e-01 1.73813760e-01 1.09085214e+00 6.32529914e-01
-1.08717954e+00 4.83362019e-01 -1.10942757e+00 -1.72706679e-01
5.85653424e-01 4.88244683e-01 -4.16349024e-01 -1.37928650e-02
-1.10073817e+00 5.68336725e-01 3.83012980e-01 2.67448783e-01
-4.06881958e-01 -3.82466108e-01 -1.10868323e+00 -1.09754145e-01
3.30514042e-03 -6.51388764e-01 1.14227486e+00 -6.72833502e-01
-1.52379525e+00 1.14374375e+00 1.41960829e-02 -8.98515582e-02
3.32913667e-01 -2.08742112e-01 -3.44362706e-02 4.13412750e-01
3.63212973e-02 9.94304538e-01 5.89455247e-01 -1.08620536e+00
-8.50249946e-01 -6.17177129e-01 7.47306526e-01 3.98337781e-01
-3.25456858e-01 -4.48684275e-01 -8.52439821e-01 -3.80271852e-01
9.40381348e-01 -5.94705641e-01 -3.52659643e-01 2.48015240e-01
-4.96430516e-01 -3.47898096e-01 6.24101579e-01 -2.81387448e-01
8.88243914e-01 -2.19579005e+00 3.48185711e-02 2.72758633e-01
2.40182519e-01 1.26729563e-01 1.01155872e-02 -7.54723698e-02
1.89176708e-01 2.00981334e-01 -3.40313196e-01 -2.79955596e-01
-3.42818126e-02 4.44219738e-01 -3.20963375e-02 3.46587509e-01
5.03798962e-01 8.98045242e-01 -8.81690383e-01 -3.67686123e-01
4.01282370e-01 7.28745937e-01 -5.02241910e-01 2.75812566e-01
-1.01454481e-01 4.00262475e-01 -1.02608156e+00 6.16084993e-01
1.05568933e+00 -2.00908676e-01 -2.71940410e-01 -2.47485578e-01
-2.20899403e-01 4.78776693e-01 -9.37688410e-01 2.02650881e+00
-5.67519784e-01 4.03610528e-01 -2.03446105e-01 -9.89633381e-01
1.19900632e+00 -2.48616278e-01 2.53127217e-01 -8.31112683e-01
1.92625239e-01 1.75884753e-01 -4.01453763e-01 -5.38475692e-01
4.19488728e-01 3.08923684e-02 -4.82689701e-02 2.15638191e-01
-2.13623624e-02 -5.30291557e-01 -1.23559602e-01 -2.25308657e-01
9.86169219e-01 2.36698002e-01 -9.86265838e-02 -2.28443686e-02
4.48787123e-01 -3.20369489e-02 3.58561248e-01 7.05850303e-01
-1.19185910e-01 8.63020241e-01 5.78783989e-01 -7.65040994e-01
-8.53896081e-01 -1.13728118e+00 -3.58616382e-01 8.15408528e-01
5.56624651e-01 -3.84499729e-02 -8.74136865e-01 -8.89888525e-01
5.09208292e-02 2.62969643e-01 -8.95307720e-01 1.39505267e-02
-6.09531164e-01 -4.89906371e-01 4.68588561e-01 8.83148313e-01
1.06563091e+00 -9.54512477e-01 -1.01176834e+00 -1.25919059e-02
1.81704015e-01 -1.20385480e+00 -2.70120591e-01 6.15794837e-01
-1.21849525e+00 -1.02759767e+00 -7.10067749e-01 -1.13141930e+00
1.00091207e+00 3.63834351e-01 8.24200451e-01 -7.01673403e-02
-6.76584765e-02 1.99531019e-01 -5.54673433e-01 -1.04316652e-01
1.21120021e-01 3.91067743e-01 -3.95296276e-01 -1.91413507e-01
5.89810550e-01 -4.91452426e-01 -8.41868401e-01 3.22164536e-01
-9.18510437e-01 2.02322796e-01 4.36952978e-01 5.03578484e-01
9.37903941e-01 1.47800982e-01 -6.19252957e-02 -8.44782710e-01
2.84869134e-01 -1.29910320e-01 -6.67532444e-01 1.09234802e-01
-1.89957500e-01 3.23979169e-01 3.24418187e-01 -1.92639634e-01
-1.00789225e+00 4.73862201e-01 -4.17215168e-01 -3.64498615e-01
-4.93979722e-01 2.74341613e-01 -2.77980477e-01 -1.23829015e-01
3.31680000e-01 -2.23907847e-02 -2.70633370e-01 -5.57594657e-01
5.21921098e-01 7.16650903e-01 5.54007232e-01 -5.57861924e-01
6.57207251e-01 6.71415210e-01 -1.23311067e-02 -5.97029030e-01
-9.58915830e-01 -3.85268837e-01 -1.01390243e+00 7.04367906e-02
1.29737735e+00 -7.84762561e-01 -6.27574086e-01 6.53446794e-01
-1.08551216e+00 -7.20614433e-01 -1.15460396e-01 4.03459013e-01
-3.97339940e-01 -6.29988909e-02 -6.31148875e-01 -5.16981781e-01
-3.29776593e-02 -1.25026345e+00 1.34501994e+00 5.95598936e-01
2.21893772e-01 -1.04184580e+00 -6.90659136e-02 1.07269160e-01
1.71114951e-01 1.99091792e-01 9.64295864e-01 -2.02633500e-01
-4.21383768e-01 -1.84380621e-01 -5.51390767e-01 3.45151544e-01
1.87145323e-01 -1.31376028e-01 -1.07096958e+00 2.58533090e-01
-2.29180962e-01 -1.51758730e-01 1.01750326e+00 5.60965061e-01
1.64630246e+00 1.50082275e-01 -1.95261285e-01 1.17190456e+00
1.47098470e+00 1.97124168e-01 7.60765791e-01 6.31893635e-01
9.54538047e-01 4.85105693e-01 2.88419962e-01 2.99206883e-01
6.74663901e-01 3.79964501e-01 7.79675305e-01 -4.68735427e-01
1.31031141e-01 -4.06165212e-01 -2.40738198e-01 3.21539611e-01
-2.01642320e-01 -1.25448210e-02 -8.73161316e-01 2.08584353e-01
-1.64903975e+00 -4.25591081e-01 -6.39594644e-02 1.89824736e+00
6.41396582e-01 4.84872758e-01 -1.05261102e-01 7.53983781e-02
5.26306808e-01 6.31533489e-02 -7.05597222e-01 -5.45364141e-01
-1.75570488e-01 5.30176938e-01 7.98340261e-01 5.49304545e-01
-1.35557568e+00 1.19950986e+00 5.28833818e+00 4.37975973e-01
-1.45239663e+00 -1.86540142e-01 7.40228891e-01 2.34051794e-01
-4.43143725e-01 -8.94844383e-02 -8.27641785e-01 6.62750797e-03
1.29481733e-01 5.83503425e-01 1.76190034e-01 9.47853029e-01
-2.54460443e-02 -2.20495641e-01 -8.72273266e-01 1.08296561e+00
-2.88183540e-01 -1.22206545e+00 -1.19976431e-01 -1.10065244e-01
6.38554335e-01 4.33214195e-02 5.14521301e-02 8.55845883e-02
2.63595253e-01 -1.04183400e+00 7.29337752e-01 2.89603472e-01
7.66242325e-01 -8.15151453e-01 5.93533158e-01 1.71715721e-01
-1.24981773e+00 -3.28014791e-02 -6.44683421e-01 -2.28132397e-01
-7.11304024e-02 4.78328377e-01 -8.01105201e-01 2.32568145e-01
1.06357896e+00 7.52828658e-01 -5.52268624e-01 8.77819538e-01
-6.12123311e-01 2.61985093e-01 -4.49706882e-01 -1.97213978e-01
6.34958088e-01 -2.34033838e-01 -2.51883715e-01 1.01523566e+00
3.58449757e-01 3.91409427e-01 -1.15559995e-03 9.29450750e-01
-6.69904053e-02 -3.73139977e-02 -4.44671690e-01 1.31136283e-01
4.79122162e-01 1.12538624e+00 -1.14688623e+00 -1.35566115e-01
-4.50355947e-01 1.33121240e+00 5.62506855e-01 5.10827899e-01
-7.13829517e-01 -7.47096241e-01 9.00249004e-01 9.04643461e-02
6.48805439e-01 -3.50947678e-01 -6.10325992e-01 -9.43387806e-01
6.48301095e-02 2.77252160e-02 1.65616959e-01 -8.34808946e-01
-1.21479714e+00 6.81718946e-01 -4.87111025e-02 -1.03988659e+00
1.68166906e-01 -1.12003553e+00 -5.12146592e-01 9.87558246e-01
-1.76071227e+00 -9.90041196e-01 -6.19606674e-01 6.27640247e-01
1.57109708e-01 3.88561934e-01 6.74347281e-01 1.08241387e-01
-5.33011019e-01 3.33695620e-01 -3.20756763e-01 4.43342149e-01
1.32705003e-01 -1.35360038e+00 6.32751763e-01 5.61890721e-01
-2.06456423e-01 5.04686296e-01 9.02482793e-02 -3.99630845e-01
-1.05301642e+00 -1.11793184e+00 8.48471165e-01 -2.77828783e-01
2.65934169e-01 -5.27772844e-01 -7.73369253e-01 5.96281171e-01
-3.46577406e-01 4.21162844e-01 4.27560180e-01 -1.63988676e-02
-4.28424984e-01 -6.73815459e-02 -1.23296368e+00 3.82032514e-01
1.36133444e+00 -6.36910081e-01 -4.84017998e-01 5.73868118e-02
8.75986576e-01 -9.47312415e-01 -7.67583013e-01 4.98729706e-01
3.42144579e-01 -1.18768227e+00 1.20593619e+00 -3.91784281e-01
5.85972667e-01 -4.26244199e-01 -2.54082441e-01 -1.03227854e+00
-6.26351759e-02 -1.61391087e-02 2.20088169e-01 8.06887805e-01
3.70596468e-01 -5.81038535e-01 9.67495024e-01 6.63069427e-01
-4.28210884e-01 -8.77512157e-01 -8.37008953e-01 -3.98173481e-01
2.62068361e-01 -7.13641346e-01 1.03273439e+00 7.88800240e-01
-2.99920827e-01 3.31689529e-02 3.21937740e-01 4.52755511e-01
2.80231863e-01 3.18956614e-01 7.10565031e-01 -1.16867352e+00
-1.65751174e-01 -7.43738711e-01 -6.47637129e-01 -1.73132205e+00
-3.42136137e-02 -8.58811438e-01 1.18506357e-01 -1.88566101e+00
-2.51098245e-01 -9.91140127e-01 -2.03457847e-01 8.45029294e-01
-2.30473712e-01 2.55758762e-01 -3.31444800e-01 8.19467194e-03
-3.11360985e-01 4.56758171e-01 1.66657269e+00 -1.54533640e-01
-6.41449094e-01 3.97072732e-02 -4.57943171e-01 8.22039545e-01
9.51934814e-01 -8.22935179e-02 -6.01828873e-01 -8.94560099e-01
2.47777000e-01 -1.40341833e-01 4.50608730e-01 -8.50694418e-01
1.17897630e-01 -1.64440364e-01 6.34858012e-01 -6.80966079e-01
2.89648205e-01 -6.54268503e-01 -3.51239949e-01 8.02358389e-02
-2.26043001e-01 -8.88976008e-02 1.38026506e-01 9.76972505e-02
-1.85133263e-01 -2.26922497e-01 5.33961594e-01 -3.08957726e-01
-1.13919497e+00 5.24906814e-01 9.28231552e-02 -3.00270505e-02
8.10043693e-01 -5.06543040e-01 3.19571011e-02 -3.92734967e-02
-6.99960649e-01 1.38530806e-01 5.13793826e-01 3.43828738e-01
8.92691553e-01 -1.24824464e+00 -9.76973474e-02 5.13641655e-01
1.70238331e-01 8.08104038e-01 1.63349643e-01 3.69072169e-01
-1.03581512e+00 2.46071547e-01 -3.17797691e-01 -7.96476126e-01
-6.08235359e-01 2.77316153e-01 6.66330159e-01 1.64929822e-01
-8.15043330e-01 1.31997275e+00 4.24679548e-01 -7.46336639e-01
3.71223658e-01 -8.79657447e-01 -2.72903234e-01 -2.18568832e-01
3.39693576e-01 -9.79479477e-02 6.08269759e-02 -5.36747158e-01
-3.96738470e-01 9.95181620e-01 1.74535647e-01 3.55712771e-02
1.42954338e+00 -1.40747234e-01 -1.29445791e-02 2.32387543e-01
1.39037883e+00 -4.56305444e-01 -1.73514485e+00 -2.28256863e-02
-2.77027249e-01 -4.57240015e-01 2.48455256e-01 -3.93816978e-01
-1.44212949e+00 1.03092587e+00 5.69726527e-01 2.31439143e-01
1.35584879e+00 3.21378738e-01 8.91321480e-01 3.84381354e-01
2.21158117e-01 -9.02592838e-01 5.36769629e-03 7.07040548e-01
4.49982405e-01 -1.23540926e+00 -3.42021406e-01 -6.19798720e-01
-3.46342832e-01 1.51689279e+00 7.90599763e-01 -2.61062086e-01
9.07229125e-01 1.88694209e-01 3.05740505e-01 -3.71482819e-01
1.12106070e-01 -5.26852965e-01 1.99868470e-01 7.41924226e-01
3.43205303e-01 3.34795900e-02 -2.76140310e-02 4.24029887e-01
-3.35629255e-01 -7.05045089e-02 5.35682030e-02 1.08187401e+00
-5.58528185e-01 -1.02109075e+00 1.60541553e-02 1.15725264e-01
7.00573102e-02 3.39147933e-02 -1.95431113e-01 7.48290181e-01
5.60262620e-01 6.78483009e-01 5.72302580e-01 -4.77297932e-01
5.92414200e-01 -2.11254925e-01 6.44897044e-01 -5.40071964e-01
-2.39695534e-01 -5.88590614e-02 -3.65909517e-01 -5.96157551e-01
-5.39259076e-01 -3.97964299e-01 -1.97427630e+00 -8.35588053e-02
-5.90458550e-02 -1.80102170e-01 8.65635753e-01 9.05864418e-01
1.19303748e-01 4.69835460e-01 6.55835807e-01 -8.48988891e-01
-6.16364740e-02 -6.72098279e-01 -5.20058751e-01 4.36537623e-01
3.65305990e-01 -6.61680400e-01 -1.86955720e-01 -7.31524378e-02] | [9.19113826751709, -1.2473522424697876] |
2835b29c-5b9b-4b2e-acdc-f47970021e9b | recontrast-domain-specific-anomaly-detection | 2306.02602 | null | https://arxiv.org/abs/2306.02602v1 | https://arxiv.org/pdf/2306.02602v1.pdf | ReContrast: Domain-Specific Anomaly Detection via Contrastive Reconstruction | Most advanced unsupervised anomaly detection (UAD) methods rely on modeling feature representations of frozen encoder networks pre-trained on large-scale datasets, e.g. ImageNet. However, the features extracted from the encoders that are borrowed from natural image domains coincide little with the features required in the target UAD domain, such as industrial inspection and medical imaging. In this paper, we propose a novel epistemic UAD method, namely ReContrast, which optimizes the entire network to reduce biases towards the pre-trained image domain and orients the network in the target domain. We start with a feature reconstruction approach that detects anomalies from errors. Essentially, the elements of contrastive learning are elegantly embedded in feature reconstruction to prevent the network from training instability, pattern collapse, and identical shortcut, while simultaneously optimizing both the encoder and decoder on the target domain. To demonstrate our transfer ability on various image domains, we conduct extensive experiments across two popular industrial defect detection benchmarks and three medical image UAD tasks, which shows our superiority over current state-of-the-art methods. | ['Huiqi Li', 'Weihang Zhang', 'Lize Jia', 'Shuai Lu', 'Jia Guo'] | 2023-06-05 | null | null | null | null | ['defect-detection', 'unsupervised-anomaly-detection'] | ['computer-vision', 'methodology'] | [ 5.29707730e-01 3.11632067e-01 2.28716969e-01 -5.16967773e-01
-4.61800814e-01 3.00499313e-02 1.56284794e-01 1.31894618e-01
-2.40516588e-01 3.15559864e-01 -3.09415519e-01 -1.38381764e-01
2.02310272e-02 -7.66847908e-01 -1.16064823e+00 -6.71033740e-01
-1.86595201e-01 1.21925704e-01 1.33034840e-01 -8.47759172e-02
2.09922478e-01 3.16862285e-01 -1.48627496e+00 2.79270947e-01
9.49746609e-01 1.48082221e+00 1.48602659e-02 4.28788424e-01
9.80647281e-02 1.13071859e+00 -6.11168027e-01 -3.50755930e-01
3.23454589e-01 -4.98570889e-01 -7.67646194e-01 5.17790079e-01
4.36829656e-01 -5.51669061e-01 -7.95121193e-01 1.32294118e+00
2.86668450e-01 -1.61264777e-01 6.85067177e-01 -1.24454343e+00
-9.30190861e-01 2.55008012e-01 -4.07329589e-01 2.82830447e-01
1.62727013e-02 4.12075400e-01 9.23462927e-01 -9.58607554e-01
6.83357239e-01 1.04183722e+00 6.69348001e-01 6.26544356e-01
-1.20849288e+00 -2.60189146e-01 3.02403897e-01 1.32621780e-01
-9.93237138e-01 -4.23178524e-01 1.05825198e+00 -5.05714297e-01
9.34688747e-01 -2.15129077e-01 3.25215429e-01 1.17871976e+00
4.55040962e-01 9.47794557e-01 4.00938809e-01 -2.76871800e-01
1.80808023e-01 -1.13388203e-01 1.53266624e-01 1.13742435e+00
3.37829381e-01 2.12121129e-01 -4.53712314e-01 -3.16870846e-02
7.18972206e-01 3.23470443e-01 -2.11590886e-01 -6.97421432e-01
-1.09637392e+00 8.57862949e-01 5.86820066e-01 1.42607778e-01
-5.46451151e-01 -1.06600486e-01 6.68427050e-01 7.85347402e-01
5.12794912e-01 6.39985502e-01 -5.38391590e-01 2.01704323e-01
-5.65046132e-01 -8.45456719e-02 4.28106755e-01 9.43108320e-01
9.26356792e-01 2.10122779e-01 -2.07361415e-01 7.86959708e-01
2.43161738e-01 1.92710951e-01 5.59976459e-01 -8.04929554e-01
3.79469693e-01 1.09498703e+00 -3.69314641e-01 -8.88725698e-01
-2.13069260e-01 -3.48012745e-01 -1.08975840e+00 3.33160400e-01
4.49129581e-01 -6.23574741e-02 -1.30290318e+00 1.61917019e+00
1.56286269e-01 2.81612843e-01 2.31825307e-01 7.46938646e-01
4.59170192e-01 4.10774678e-01 -2.86449790e-01 1.88876241e-01
9.96960580e-01 -1.02709079e+00 -4.78778780e-01 -5.35400450e-01
9.33044732e-01 -3.33133757e-01 1.13927662e+00 4.03706223e-01
-1.01379609e+00 -5.80485582e-01 -1.39075625e+00 -7.98966512e-02
-2.63162524e-01 2.23118454e-01 2.69532442e-01 8.20235759e-02
-7.55514920e-01 9.23202574e-01 -9.73290384e-01 -3.01390767e-01
8.94930243e-01 2.36812353e-01 -4.60787684e-01 -3.76689404e-01
-1.02207077e+00 6.69006467e-01 3.77434254e-01 2.91702330e-01
-1.33422077e+00 -6.32949412e-01 -1.30858517e+00 3.96121629e-02
3.53227913e-01 -5.20367444e-01 1.09141195e+00 -1.32761860e+00
-1.29524422e+00 1.08523750e+00 1.73552975e-01 -6.89965248e-01
3.84485632e-01 -5.61774313e-01 -3.81645977e-01 2.31639728e-01
2.12762013e-01 5.34137249e-01 1.16716671e+00 -1.05907893e+00
-4.11568373e-01 -3.45357418e-01 -9.55683962e-02 -1.94683701e-01
-4.39323008e-01 -2.22484618e-01 -3.47062618e-01 -6.60958111e-01
2.23290145e-01 -6.23980165e-01 -4.17349249e-01 2.76682496e-01
-8.04989755e-01 -1.03407532e-01 9.25326109e-01 -5.86378038e-01
1.00386822e+00 -2.64798379e+00 1.48491576e-01 4.18013901e-01
2.97671378e-01 1.83675379e-01 -3.09295267e-01 -8.13985988e-02
-4.55295771e-01 -2.14435294e-01 -6.92113698e-01 -3.02529842e-01
-1.87893122e-01 4.50488418e-01 -2.12886289e-01 6.34242535e-01
9.35885251e-01 7.97642469e-01 -9.46424186e-01 -5.15315235e-01
1.42786458e-01 3.62079814e-02 -7.88140118e-01 5.63187242e-01
-1.73930362e-01 4.46616024e-01 -5.81851423e-01 1.00391531e+00
5.17085135e-01 -4.57911164e-01 -2.80410759e-02 -5.09463847e-02
2.72793382e-01 4.01565969e-01 -8.31574619e-01 2.10701728e+00
-2.10966349e-01 4.23136562e-01 -6.95727617e-02 -1.50893497e+00
9.47055042e-01 1.49175227e-01 6.28097534e-01 -8.74528229e-01
6.52232841e-02 3.69823635e-01 1.19776674e-01 -6.14797473e-01
6.47678599e-02 2.14361683e-01 4.35029119e-02 3.91400158e-01
6.53822839e-01 2.24950060e-01 2.18118683e-01 2.54454851e-01
1.48107207e+00 3.90263498e-02 1.66477650e-01 -2.74467826e-01
6.41908526e-01 -9.15725827e-02 7.69385278e-01 8.58767569e-01
-3.30520868e-01 7.90668488e-01 9.13356185e-01 -7.68523276e-01
-1.06704974e+00 -1.27239764e+00 -1.67309999e-01 6.93347752e-01
1.22628510e-01 -3.61199200e-01 -6.39773309e-01 -1.47009087e+00
4.01199833e-02 4.50239331e-01 -8.40488791e-01 -7.74238169e-01
-6.05348110e-01 -6.45311594e-01 3.87284011e-01 5.33338845e-01
5.68066359e-01 -1.29168200e+00 -5.15231907e-01 1.91920564e-01
1.17983408e-01 -8.73662353e-01 -3.23111475e-01 5.20285666e-01
-9.61381793e-01 -1.41682112e+00 -4.74884003e-01 -9.94966388e-01
1.13997364e+00 -1.76544532e-01 1.48096669e+00 4.20516908e-01
-5.31205773e-01 2.98648596e-01 -3.11173528e-01 -3.52731258e-01
-4.89216715e-01 2.38712519e-01 9.62625316e-04 6.59556910e-02
4.85064685e-01 -5.23816288e-01 -4.99196827e-01 2.72235155e-01
-1.00639760e+00 -3.24329227e-01 8.17044079e-01 1.25077307e+00
6.44351602e-01 -4.06405367e-02 5.04402459e-01 -1.19334805e+00
2.83589542e-01 -6.60074770e-01 -4.37919825e-01 1.97882473e-01
-6.92233741e-01 3.00491750e-01 6.37232959e-01 -2.71966010e-01
-8.30101788e-01 2.13758767e-01 -3.25662404e-01 -7.51747727e-01
-2.75564164e-01 3.13981831e-01 -3.22320342e-01 1.96317494e-01
7.84862995e-01 3.58315200e-01 3.51586998e-01 -4.37568247e-01
-1.13391437e-01 3.90054494e-01 7.13896453e-01 -5.50676942e-01
9.03253913e-01 3.25147539e-01 -3.05758327e-01 -5.80042243e-01
-1.01439869e+00 -9.62608159e-02 -6.72322512e-01 4.83907796e-02
8.18167210e-01 -8.68454099e-01 -3.03134143e-01 8.03042829e-01
-1.05989802e+00 -2.05327600e-01 -8.97797585e-01 2.52869010e-01
-5.11466444e-01 2.56759673e-01 -6.59251809e-01 -3.41846108e-01
-2.69408703e-01 -1.22076023e+00 1.06169295e+00 5.17399237e-02
3.84864509e-02 -7.04711497e-01 -1.93243288e-02 -2.57098436e-01
1.86948329e-01 2.39740223e-01 1.21631527e+00 -9.10313785e-01
-4.60856915e-01 -1.43701345e-01 -1.87918097e-01 9.95294750e-01
3.09984177e-01 -1.46129057e-01 -1.04743791e+00 -2.74807841e-01
1.92645684e-01 -6.57768965e-01 1.22209942e+00 1.88487887e-01
1.70141697e+00 -2.03066871e-01 -2.49345779e-01 7.77428985e-01
1.15194237e+00 1.50441393e-01 8.83630395e-01 5.51086068e-01
6.35315657e-01 4.85846996e-01 5.55062830e-01 1.75819010e-01
2.06675138e-02 6.76020682e-02 8.40738177e-01 -2.71451712e-01
2.01422021e-01 -3.23820561e-01 5.26638389e-01 6.36774361e-01
3.49907994e-01 -1.14411078e-01 -6.99843645e-01 6.72235370e-01
-1.88921690e+00 -5.48330069e-01 3.56157273e-01 1.91381061e+00
7.17549980e-01 5.36852002e-01 -2.26560697e-01 1.96516275e-01
6.26760840e-01 1.86243862e-01 -1.00545716e+00 -2.60040879e-01
-1.05041020e-01 1.73870593e-01 1.90821797e-01 -3.93416658e-02
-1.40550673e+00 6.00477278e-01 6.28346395e+00 2.92041004e-01
-9.99101162e-01 -8.60754997e-02 9.51940179e-01 8.87894183e-02
-7.37160221e-02 -3.00654501e-01 -1.79794848e-01 4.32151645e-01
7.04386353e-01 3.02131176e-01 1.23949587e-01 1.22501636e+00
-3.43726069e-01 3.10236454e-01 -1.55740941e+00 8.29201639e-01
2.27089152e-02 -1.22031176e+00 5.41303717e-02 -8.75573419e-03
6.06615961e-01 1.36479288e-01 1.49754867e-01 3.34164679e-01
3.31183285e-01 -9.54029262e-01 5.50269842e-01 3.56948018e-01
7.67141998e-01 -4.91170198e-01 9.15691197e-01 -4.87428419e-02
-6.69047534e-01 -2.16579124e-01 -5.76223254e-01 1.50448188e-01
-8.57621208e-02 9.52930927e-01 -6.37984395e-01 6.03820801e-01
9.09886003e-01 1.19261670e+00 -6.58284962e-01 7.19639778e-01
-3.14770460e-01 6.33063257e-01 -1.52018309e-01 6.03401780e-01
3.25470120e-01 -1.48003459e-01 5.10753870e-01 8.58447194e-01
3.07191402e-01 -4.67124641e-01 3.89914997e-02 1.02830577e+00
-2.95130134e-01 -2.01602757e-01 -1.07086456e+00 -1.69541508e-01
6.84913099e-02 9.93395805e-01 -3.22790831e-01 -2.40834832e-01
-7.16952801e-01 1.20110118e+00 4.33606207e-01 2.83075362e-01
-7.86977291e-01 -4.99649942e-01 9.29466367e-01 -1.33898454e-02
3.63953710e-01 5.41730374e-02 -2.36333922e-01 -1.34504640e+00
2.01667294e-01 -1.03222823e+00 6.18053377e-01 -3.83609772e-01
-1.63178158e+00 6.96461082e-01 -5.90295255e-01 -1.44457150e+00
-2.04784915e-01 -1.01519775e+00 -8.64593565e-01 3.74783009e-01
-1.62522924e+00 -7.53058851e-01 -1.64090797e-01 9.03743386e-01
5.30011356e-01 -4.32008326e-01 6.77770019e-01 3.98853302e-01
-9.50782776e-01 6.45397007e-01 8.43699649e-03 5.61394334e-01
6.73943520e-01 -1.38765395e+00 5.31268895e-01 1.15872598e+00
6.83197230e-02 3.63372326e-01 3.17766279e-01 -4.62513387e-01
-1.11154068e+00 -1.34314930e+00 4.41733837e-01 -1.96147054e-01
5.23028314e-01 -4.74370331e-01 -1.15882289e+00 9.64102685e-01
4.65893093e-03 5.26501954e-01 4.49566513e-01 4.68380339e-02
-2.99794674e-01 -1.71930194e-01 -1.16554546e+00 3.14099878e-01
1.25257397e+00 -5.23849964e-01 -6.01191759e-01 2.78008610e-01
6.70455635e-01 -4.87963378e-01 -6.99098289e-01 5.03996074e-01
9.20115933e-02 -9.39699709e-01 9.20126736e-01 -9.71727014e-01
8.86823595e-01 -1.89250454e-01 -1.68296292e-01 -1.43695796e+00
-2.97325402e-01 -3.89695615e-01 -4.07816589e-01 1.14472318e+00
5.22798896e-01 -6.78035140e-01 6.87540948e-01 3.29736918e-01
-6.52080178e-01 -8.98179114e-01 -8.38659823e-01 -6.47855103e-01
-1.24070019e-01 -3.72061402e-01 5.83620131e-01 8.16446245e-01
-3.04705858e-01 1.85933635e-01 -2.60785013e-01 3.09964478e-01
5.03407300e-01 -9.49461386e-03 5.33039451e-01 -1.29087496e+00
-4.10307534e-02 -3.10298707e-02 -6.70786083e-01 -9.78779852e-01
1.71454042e-01 -7.98738778e-01 2.83027649e-01 -9.89990354e-01
-8.05841386e-02 -3.37671936e-01 -8.13297868e-01 7.11806536e-01
-1.77573711e-01 4.10463251e-02 -2.57271647e-01 1.06475361e-01
-6.91361070e-01 7.63692141e-01 1.20535672e+00 -4.77953136e-01
-4.87896241e-02 -2.10148454e-01 -7.54509926e-01 8.75596046e-01
6.81856751e-01 -7.14287102e-01 -4.63603556e-01 -6.38740003e-01
1.38149112e-01 -4.76485521e-01 5.02717674e-01 -1.08532083e+00
7.87545964e-02 2.02067241e-01 6.98365152e-01 -1.50977746e-01
-1.64257646e-01 -8.52919579e-01 -6.03764176e-01 6.04713619e-01
-2.68380195e-01 1.09655753e-01 4.68632393e-02 6.67428017e-01
-5.45003593e-01 -2.89287835e-01 8.98340404e-01 -2.57314473e-01
-8.74678850e-01 5.16721845e-01 -1.73227221e-01 2.23311633e-01
1.04364371e+00 -1.19661056e-01 -3.53940248e-01 -5.23803122e-02
-7.04452515e-01 3.58060747e-01 5.98494649e-01 4.68734086e-01
9.17707384e-01 -1.34885526e+00 -5.62333763e-01 8.95554066e-01
3.81599993e-01 6.07557952e-01 2.54723072e-01 8.47919166e-01
-4.91318226e-01 4.80811968e-02 -4.10408586e-01 -8.55882168e-01
-6.06664181e-01 7.71394193e-01 6.93156600e-01 -4.02886480e-01
-9.62233126e-01 8.14123392e-01 4.71227676e-01 -5.46287835e-01
2.12763593e-01 -3.86863768e-01 1.86418384e-01 -3.22898358e-01
3.47949833e-01 6.25100508e-02 3.05510134e-01 -1.97114557e-01
-4.06643182e-01 1.31443456e-01 -6.03309512e-01 5.68667889e-01
1.56209385e+00 6.32319972e-02 -9.12129804e-02 2.89872795e-01
1.24674225e+00 -5.61696649e-01 -1.72850966e+00 -4.27400827e-01
2.20339447e-01 -3.55199337e-01 -1.24314517e-01 -4.95889217e-01
-1.62784123e+00 8.90304506e-01 7.70820439e-01 2.22550854e-01
1.40774429e+00 1.83057897e-02 8.02789927e-01 6.22417748e-01
5.99148050e-02 -1.23559582e+00 5.97863495e-01 5.21465659e-01
6.84476674e-01 -1.43710291e+00 -3.47836107e-01 -1.64598435e-01
-6.28067315e-01 1.09511673e+00 1.08553624e+00 -5.11578202e-01
7.76664257e-01 1.70886293e-01 5.31764850e-02 -6.55920625e-01
-6.14179134e-01 1.64356008e-02 9.84907150e-02 5.54126859e-01
1.95486575e-01 -5.23257673e-01 3.28569800e-01 6.78173661e-01
3.03087384e-01 1.12367738e-02 2.93005973e-01 1.16470647e+00
-2.84110993e-01 -7.94659436e-01 -2.42953300e-02 9.32703316e-01
-5.13031542e-01 5.45167401e-02 -2.90060371e-01 7.29801953e-01
2.82064021e-01 5.15431285e-01 4.06972855e-01 -4.48416412e-01
4.86096740e-01 2.93103665e-01 2.63151169e-01 -6.48782611e-01
-3.11157852e-01 -3.93445998e-01 -2.13840470e-01 -9.79965508e-01
-1.08619861e-01 -6.22895539e-01 -1.12276077e+00 1.85372606e-01
-2.39888534e-01 -7.42404088e-02 1.45674035e-01 9.74281788e-01
5.80240846e-01 9.20882642e-01 8.08162332e-01 -5.04333138e-01
-5.86096704e-01 -1.02830064e+00 -7.84548998e-01 7.55800545e-01
6.62074625e-01 -7.75617838e-01 -4.20330405e-01 1.46426171e-01] | [7.642261981964111, 2.0537490844726562] |
54ac60f9-771e-4696-8d00-5fb0f199f3cc | working-paper-improved-stock-price | 1902.08938 | null | http://arxiv.org/abs/1902.08938v1 | http://arxiv.org/pdf/1902.08938v1.pdf | Working Paper: Improved Stock Price Forecasting Algorithm based on Feature-weighed Support Vector Regression by using Grey Correlation Degree | With the widespread engineering applications ranging from artificial
intelligence and big data decision-making, originally a lot of tedious
financial data processing, processing and analysis have become more and more
convenient and effective. This paper aims to improve the accuracy of stock
price forecasting. It improves the support vector machine regression algorithm
by using grey correlation analysis (GCA) and improves the accuracy of stock
prediction. This article first divides the factors affecting the stock price
movement into behavioral factors and technical factors. The behavioral factors
mainly include weather indicators and emotional indicators. The technical
factors mainly include the daily closing data and the HS 300 Index, and then
measure relation through the method of grey correlation analysis. The
relationship between the stock price and its impact factors during the trading
day, and this relationship is transformed into the characteristic weight of
each impact factor. The weight of the impact factors of all trading days is
weighted by the feature weight, and finally the support vector regression (SVR)
is used. The forecast of the revised stock trading data was compared based on
the forecast results of technical indicators (MSE, MAE, SCC, and DS) and
unmodified transaction data, and it was found that the forecast results were
significantly improved. | [] | 2019-02-24 | null | null | null | null | ['stock-prediction'] | ['time-series'] | [-9.88902628e-01 -7.65980303e-01 -9.48967859e-02 -1.24233291e-01
3.39432925e-01 -3.69333744e-01 4.49938267e-01 -1.19545400e-01
-3.40736479e-01 6.61001444e-01 2.11507976e-01 -3.33572000e-01
-1.92460388e-01 -1.15533650e+00 2.55445033e-01 -9.08940196e-01
-3.35485458e-01 1.20981641e-01 2.07358152e-01 -8.49035740e-01
9.76209939e-01 3.81836087e-01 -1.42634296e+00 -2.68953711e-01
8.13931704e-01 1.73986554e+00 -8.73469636e-02 -1.36239618e-01
-4.51269358e-01 1.19215202e+00 -8.58094573e-01 -3.48780632e-01
5.80771267e-01 -3.72020930e-01 2.17147753e-01 -1.11958571e-01
-9.51836169e-01 -3.91177595e-01 7.43855685e-02 1.04534602e+00
2.87605375e-01 2.04317257e-01 3.12706292e-01 -1.31481946e+00
-8.13065052e-01 6.25703454e-01 -8.12209368e-01 5.16460121e-01
-2.34239444e-01 -1.79766983e-01 7.53121734e-01 -1.03804553e+00
1.36858433e-01 1.10191607e+00 4.70460415e-01 -3.12669307e-01
-4.99246836e-01 -9.97711360e-01 6.96594790e-02 4.46596742e-01
-9.44143176e-01 3.22714418e-01 8.07376564e-01 -5.05263388e-01
9.54021275e-01 2.58987248e-01 1.33383310e+00 -1.13369245e-03
6.84666514e-01 3.81336749e-01 1.41345668e+00 -1.68534011e-01
1.88333884e-01 4.27317053e-01 4.04719651e-01 -7.44257122e-02
6.40305817e-01 5.03142059e-01 -1.78288534e-01 -3.09429243e-02
6.27078533e-01 6.08634353e-01 -1.39173746e-01 6.22512341e-01
-1.02575791e+00 1.17078900e+00 2.07600445e-01 7.70685017e-01
-5.84715962e-01 -5.88289559e-01 3.62096101e-01 9.20299649e-01
7.31075346e-01 2.89696604e-01 -8.43109727e-01 -2.93175101e-01
-9.56925809e-01 7.21515194e-02 1.01170480e+00 2.86922395e-01
4.65880305e-01 8.05663228e-01 2.26492479e-01 4.24562186e-01
3.26370984e-01 9.61293936e-01 1.05073786e+00 -4.76348162e-01
2.05924734e-01 7.65366077e-01 1.21094830e-01 -1.55865693e+00
-4.85388011e-01 -5.25137067e-01 -9.07187700e-01 5.67933023e-01
-3.95633690e-02 -3.87261122e-01 -2.71576703e-01 8.65157008e-01
-1.15666173e-01 1.02436036e-01 3.16161782e-01 8.20394278e-01
4.90227103e-01 1.13259995e+00 -3.33856523e-01 -1.09899795e+00
1.10652936e+00 -5.02403915e-01 -1.31038749e+00 3.23265791e-01
2.61511058e-01 -8.68952215e-01 3.99386585e-01 6.25376880e-01
-1.04866219e+00 -5.47576785e-01 -1.07508147e+00 7.44861960e-01
-6.26117945e-01 -1.42321199e-01 5.41859686e-01 3.40606034e-01
-6.59426689e-01 8.48359585e-01 -4.02234375e-01 5.58903456e-01
-3.61428916e-01 2.03405067e-01 1.27621457e-01 6.90161645e-01
-1.77143776e+00 1.38054824e+00 6.80166245e-01 3.48346442e-01
3.54604721e-01 -5.92546463e-01 -4.47177380e-01 -3.35097909e-02
1.27896011e-01 1.86381694e-02 7.64881015e-01 -8.98533165e-01
-1.42930734e+00 1.29796013e-01 3.46447170e-01 -4.69960481e-01
2.23549619e-01 -3.34227867e-02 -1.13924515e+00 -2.05824971e-01
-2.08077490e-01 -4.23265845e-01 6.62271380e-01 -5.76022625e-01
-8.35888803e-01 -3.50287318e-01 -7.23526597e-01 -1.00124136e-01
-1.41236991e-01 3.52434248e-01 1.72825783e-01 -1.07601166e+00
2.45038971e-01 -6.31384075e-01 -1.75789550e-01 -8.26761246e-01
3.51426125e-01 -1.31902069e-01 7.75757730e-01 -1.17743754e+00
1.50527894e+00 -2.24015307e+00 -2.05834642e-01 9.58472133e-01
-1.67647794e-01 1.64906323e-01 5.23184180e-01 5.17164707e-01
-4.14624631e-01 9.68847498e-02 -2.01221630e-02 7.61697710e-01
-6.87605813e-02 1.46735400e-01 -8.20260823e-01 2.61161178e-01
1.97529435e-01 7.13650227e-01 -4.58993375e-01 -1.37337893e-01
2.47156471e-01 7.44154379e-02 2.67866794e-02 1.19888037e-01
2.97078907e-01 -1.00274310e-01 -4.54662919e-01 7.15279937e-01
8.57229352e-01 8.59285891e-03 -4.65661943e-01 -9.30536911e-02
-7.79654026e-01 -1.48616821e-01 -1.52119064e+00 3.34998429e-01
7.75330281e-03 2.69050062e-01 -2.68006533e-01 -7.92146444e-01
1.84642303e+00 3.56469154e-01 5.96777797e-01 -9.99095917e-01
2.19792530e-01 6.54601872e-01 7.85727650e-02 -5.65513670e-01
4.51700896e-01 -6.23522997e-01 1.04833193e-01 7.35227585e-01
-5.36527991e-01 -4.64852415e-02 2.41165176e-01 -3.26115522e-03
1.92960262e-01 -2.89280027e-01 4.48917687e-01 -3.14358681e-01
5.06587267e-01 3.82823385e-02 9.08948720e-01 -4.29934204e-01
-9.76528320e-03 6.27646074e-02 6.64528608e-01 -5.63978434e-01
-6.83742821e-01 -4.10324931e-01 -1.85243636e-01 5.77995837e-01
1.42529503e-01 1.22966290e-01 -1.23257553e-02 7.33281001e-02
5.54106653e-01 9.88164723e-01 -6.68046594e-01 -2.87423700e-01
-6.20805547e-02 -1.20371521e+00 -2.21173599e-01 3.95498872e-01
8.54387999e-01 -1.15795696e+00 -4.39608991e-01 4.81664896e-01
4.91108507e-01 -3.05637717e-01 1.73025820e-02 -6.37195408e-02
-9.22463477e-01 -7.99499333e-01 -7.14523673e-01 -2.62472659e-01
3.83320749e-01 1.28172010e-01 9.38055158e-01 4.94339377e-01
1.66777819e-01 -3.06561232e-01 -4.51147705e-01 -9.03729141e-01
8.23037177e-02 -6.93558574e-01 2.85010159e-01 8.82120356e-02
4.52563077e-01 -3.87954831e-01 -4.75484520e-01 4.55302268e-01
-7.99831867e-01 -3.54174793e-01 6.77464843e-01 6.82610631e-01
1.92545965e-01 8.02565575e-01 7.59192348e-01 -3.13882381e-01
9.42849278e-01 -6.08973265e-01 -1.14031351e+00 1.33484736e-01
-1.50197017e+00 7.35954987e-03 3.20099652e-01 -3.16921502e-01
-1.10231757e+00 -8.51023018e-01 2.26842389e-01 -2.69192785e-01
6.00273013e-01 1.11280489e+00 2.64986724e-01 4.81196530e-02
-7.58827329e-02 4.02880758e-01 4.03968245e-01 -5.49398303e-01
-1.50222868e-01 5.10459244e-01 2.24135295e-01 2.08397940e-01
1.06596458e+00 9.37151834e-02 1.55279830e-01 -3.16008300e-01
-3.73918623e-01 -2.58094788e-01 -4.97205377e-01 -5.27839065e-01
5.83131552e-01 -6.90334022e-01 -8.74454439e-01 9.12085950e-01
-6.45892322e-01 3.90012354e-01 -1.48514926e-01 1.10201192e+00
1.20088071e-01 -1.39518939e-02 -8.77297938e-01 -1.18969834e+00
-4.03464973e-01 -7.34473646e-01 1.04255013e-01 4.14829046e-01
-9.46615487e-02 -1.09458399e+00 2.67249018e-01 -1.66421846e-01
6.50192559e-01 5.56687057e-01 7.93709517e-01 -8.44374478e-01
-2.19488427e-01 -6.55577719e-01 -1.40025154e-01 5.70625246e-01
7.11855888e-02 6.38320744e-01 -3.09336662e-01 -8.38660225e-02
7.41468251e-01 9.80541930e-02 6.59243822e-01 2.97430009e-01
1.75741181e-01 -4.47720438e-01 2.83809960e-01 4.40844357e-01
1.46516466e+00 9.96625662e-01 7.25749075e-01 1.01394916e+00
1.20910175e-01 7.40645647e-01 1.03285122e+00 8.15384507e-01
6.29248023e-02 -1.92560069e-02 9.49259847e-02 2.49779020e-02
9.47734058e-01 2.10374773e-01 6.22411013e-01 1.62804329e+00
-7.78710127e-01 5.18195868e-01 -5.88385284e-01 -4.52596918e-02
-1.65747094e+00 -1.51140738e+00 -2.94245094e-01 1.87293410e+00
6.86578035e-01 2.87465543e-01 1.37336731e-01 5.86825848e-01
5.63758135e-01 1.90539658e-01 -5.39062679e-01 -5.11815310e-01
-5.49029469e-01 -8.72438177e-02 6.80340588e-01 2.89160252e-01
-6.04505062e-01 2.15410978e-01 6.04783010e+00 5.42877436e-01
-1.54922700e+00 -4.53729689e-01 7.32653379e-01 -1.49041623e-01
-4.86175805e-01 -1.31705895e-01 -8.14164340e-01 1.04322362e+00
9.18268263e-01 -8.49708259e-01 3.25873554e-01 8.91399860e-01
4.13484514e-01 -2.05884680e-01 -2.56217957e-01 7.07469821e-01
-2.09156811e-01 -1.10198784e+00 -2.17988089e-01 2.99638510e-01
7.29698241e-01 -2.52131850e-01 1.90295339e-01 4.88737017e-01
3.88136581e-02 -5.04720390e-01 8.09575677e-01 1.15513301e+00
2.55990047e-02 -1.03644776e+00 1.44046319e+00 2.93350667e-01
-1.24677551e+00 -2.06432804e-01 -5.00240803e-01 -7.23385513e-01
1.55952513e-01 8.98627281e-01 -5.91157116e-02 7.23939121e-01
7.68425941e-01 8.67483854e-01 -2.56368786e-01 6.33761644e-01
-4.25214097e-02 4.70808327e-01 -1.59457341e-01 -4.86806750e-01
1.52793303e-01 -1.09008813e+00 2.80167237e-02 4.33630526e-01
7.27092385e-01 6.05863869e-01 -2.46539310e-01 5.06582975e-01
6.41402543e-01 4.92556423e-01 -1.12431362e-01 -4.79163416e-02
3.95702541e-01 1.06572235e+00 -7.41937757e-01 -5.84444404e-01
-7.55126834e-01 1.33772403e-01 -5.05353689e-01 1.21375874e-01
-4.99323219e-01 -6.09106243e-01 3.96345109e-01 -1.33836865e-01
4.22241390e-01 -7.24760443e-02 -8.35908413e-01 -1.11782694e+00
1.95768207e-01 -5.83934784e-01 5.00874639e-01 -7.45903552e-01
-1.50578451e+00 3.98580998e-01 -1.20877944e-01 -1.56510210e+00
-5.65233886e-01 -6.03985250e-01 -1.12324667e+00 1.31985009e+00
-1.53289568e+00 5.91308856e-03 2.41435081e-01 6.86084211e-01
-7.91346952e-02 -9.70205903e-01 6.76088214e-01 -2.16497798e-02
-7.79294431e-01 -1.39682442e-01 7.39972055e-01 2.44279206e-01
2.36449867e-01 -9.83718216e-01 1.24719650e-01 7.11411834e-01
-4.48212713e-01 3.43462884e-01 5.35201728e-01 -1.12384188e+00
-9.24685776e-01 -4.38370109e-01 1.07464099e+00 2.68162489e-01
1.24799490e+00 3.61704707e-01 -1.16642082e+00 1.41190350e-01
2.92253107e-01 -2.82382876e-01 6.15834892e-01 -2.45518729e-01
-1.25067338e-01 -5.03373682e-01 -1.23435414e+00 2.84548998e-01
-3.34302843e-01 5.71468137e-02 -1.06790030e+00 -1.42283008e-01
6.88741386e-01 2.21317947e-01 -1.35766578e+00 3.00699621e-01
5.97987592e-01 -1.06259370e+00 6.95578337e-01 -4.22027081e-01
1.51927605e-01 -2.97923088e-01 1.38588399e-01 -1.41591632e+00
-6.18708670e-01 -3.84500563e-01 6.88235238e-02 1.14186144e+00
5.81064045e-01 -1.35304308e+00 9.42031443e-02 1.09890771e+00
2.43641719e-01 -8.17245543e-01 -7.35718489e-01 -8.23029220e-01
-1.75561205e-01 -2.93271453e-03 1.12911665e+00 1.30393696e+00
4.17887807e-01 -2.91949380e-02 -4.35241133e-01 -2.10044309e-01
4.61806476e-01 7.01954246e-01 1.48649007e-01 -1.47872186e+00
-1.48721114e-01 -8.51415515e-01 -2.58114100e-01 -9.36448500e-02
-3.89250606e-01 -2.81316996e-01 -6.96525216e-01 -1.11236036e+00
-3.81676197e-01 -4.98994328e-02 -8.15388083e-01 9.85570848e-02
-2.09439725e-01 -1.61737457e-01 4.65080947e-01 9.26865101e-01
3.07894409e-01 5.21692097e-01 1.35107994e+00 1.31629750e-01
-3.40086192e-01 2.81563103e-01 -4.37619418e-01 5.31781256e-01
9.79573309e-01 -5.62647469e-02 -5.64378547e-03 2.16720790e-01
6.39984906e-01 3.84144396e-01 -3.57742816e-01 -5.76129198e-01
8.42304677e-02 -5.14307976e-01 6.73546374e-01 -1.04517317e+00
-7.05269054e-02 -1.01199937e+00 4.94493842e-01 1.09607553e+00
1.67259108e-02 5.68526924e-01 5.39501868e-02 1.04309291e-01
-8.19625020e-01 -2.30967790e-01 3.92532408e-01 -1.06962360e-01
-7.81079650e-01 1.51988387e-01 -3.17104787e-01 -2.22216412e-01
1.32543635e+00 -2.42509067e-01 -2.41302773e-01 -5.00498772e-01
-5.15255749e-01 3.11411917e-01 -3.71068008e-02 2.98950225e-01
7.21223295e-01 -1.63736939e+00 -8.76514316e-01 4.28709000e-01
-5.69821179e-01 -5.22760868e-01 7.31238723e-02 1.06142735e+00
-7.67497778e-01 4.70504075e-01 -6.03159845e-01 2.63201594e-01
-8.92200828e-01 6.85021341e-01 1.53682843e-01 -3.05149406e-01
-1.77705050e-01 5.39768755e-01 -3.18795681e-01 4.25568730e-01
-7.31005445e-02 -4.68801111e-01 -9.04837608e-01 9.81555164e-01
9.21205938e-01 8.61405432e-01 2.69932300e-02 -7.97876775e-01
-1.23440914e-01 9.52755690e-01 2.28669181e-01 -3.07246923e-01
1.87933767e+00 2.96089170e-03 -7.12449789e-01 9.30852473e-01
9.50177789e-01 3.06813680e-02 -7.91797400e-01 -1.17756322e-01
3.78895849e-01 -4.93844628e-01 3.02034587e-01 -8.08546484e-01
-1.60371983e+00 5.30107737e-01 2.62730896e-01 9.09787476e-01
1.25969744e+00 -5.78236103e-01 1.07076430e+00 3.62426430e-01
3.93786490e-01 -1.51057529e+00 -4.62842405e-01 5.67408681e-01
1.10666060e+00 -1.18258679e+00 2.58390814e-01 2.96374019e-02
-1.02297878e+00 1.42537558e+00 -7.80626461e-02 -5.35001457e-01
1.39993811e+00 4.14612651e-01 4.42433298e-01 -8.54696780e-02
-8.01748753e-01 2.21480995e-01 2.01268926e-01 -4.25341316e-02
2.68457174e-01 5.52461520e-02 -7.02932119e-01 1.22422647e+00
-6.32088006e-01 3.79872583e-02 3.18712980e-01 8.68315518e-01
-7.50107408e-01 -7.82547295e-01 -9.48577285e-01 6.50892019e-01
-7.46346533e-01 -1.70330420e-01 -1.82355449e-01 8.79627466e-01
-1.24026306e-01 8.12324345e-01 4.90484446e-01 -7.12630153e-01
4.65994895e-01 3.94492187e-02 -4.39568073e-01 8.02274272e-02
-7.94591844e-01 4.11209196e-01 -1.98186114e-01 -1.24467894e-01
-5.20620227e-01 -9.67591643e-01 -1.48933518e+00 -9.44345117e-01
-4.56088364e-01 7.13080943e-01 8.62940192e-01 9.12482917e-01
-2.63007373e-01 2.96475053e-01 1.60149276e+00 -5.38119614e-01
-8.65250170e-01 -1.03853285e+00 -1.53631568e+00 2.37511128e-01
5.90632483e-02 -6.84210539e-01 -8.76803219e-01 -1.01574875e-01] | [4.647156238555908, 4.122612476348877] |
af136fe9-3b6c-4cf4-873e-7bad0637e546 | continual-learning-for-abdominal-multi-organ | 2306.00988 | null | https://arxiv.org/abs/2306.00988v1 | https://arxiv.org/pdf/2306.00988v1.pdf | Continual Learning for Abdominal Multi-Organ and Tumor Segmentation | The ability to dynamically extend a model to new data and classes is critical for multiple organ and tumor segmentation. However, due to privacy regulations, accessing previous data and annotations can be problematic in the medical domain. This poses a significant barrier to preserving the high segmentation accuracy of the old classes when learning from new classes because of the catastrophic forgetting problem. In this paper, we first empirically demonstrate that simply using high-quality pseudo labels can fairly mitigate this problem in the setting of organ segmentation. Furthermore, we put forward an innovative architecture designed specifically for continuous organ and tumor segmentation, which incurs minimal computational overhead. Our proposed design involves replacing the conventional output layer with a suite of lightweight, class-specific heads, thereby offering the flexibility to accommodate newly emerging classes. These heads enable independent predictions for newly introduced and previously learned classes, effectively minimizing the impact of new classes on old ones during the course of continual learning. We further propose incorporating Contrastive Language-Image Pretraining (CLIP) embeddings into the organ-specific heads. These embeddings encapsulate the semantic information of each class, informed by extensive image-text co-training. The proposed method is evaluated on both in-house and public abdominal CT datasets under organ and tumor segmentation tasks. Empirical results suggest that the proposed design improves the segmentation performance of a baseline neural network on newly-introduced and previously-learned classes along the learning trajectory. | ['Zongwei Zhou', 'Yaoyao Liu', 'Alan Yuille', 'Huimiao Chen', 'Xinyi Li', 'Yixiao Zhang'] | 2023-06-01 | null | null | null | null | ['tumor-segmentation'] | ['computer-vision'] | [ 4.60709989e-01 5.42861879e-01 -5.32743812e-01 -5.00720263e-01
-7.95756698e-01 -5.57001650e-01 4.07359034e-01 7.14213908e-01
-7.99476981e-01 7.26141155e-01 2.21801654e-01 -2.60217994e-01
9.32424292e-02 -7.43599057e-01 -7.77725816e-01 -7.30835855e-01
-1.50309846e-01 4.53076541e-01 2.62814552e-01 1.84142768e-01
-3.01766694e-01 5.17464280e-01 -1.04459000e+00 2.54830360e-01
9.95075405e-01 9.58275139e-01 2.72152781e-01 4.47102427e-01
-3.45154494e-01 5.29454410e-01 -2.69586802e-01 -3.02748978e-01
3.30980897e-01 7.66785592e-02 -9.95842636e-01 4.41682398e-01
3.48103285e-01 -3.68338853e-01 -2.36803889e-01 7.72152781e-01
3.26492816e-01 9.52783152e-02 4.36917573e-01 -8.88594151e-01
-6.86062157e-01 7.47583449e-01 -4.07877028e-01 1.11059025e-01
-1.73369050e-01 2.19046384e-01 7.58429885e-01 -9.56461489e-01
8.47205222e-01 6.05225265e-01 8.63853216e-01 8.25119913e-01
-1.21971631e+00 -5.91312170e-01 4.72903937e-01 -3.75387996e-01
-1.22873795e+00 -3.10752422e-01 5.28186083e-01 -5.02720416e-01
6.16983294e-01 1.51323810e-01 7.97682226e-01 9.53905344e-01
3.46536666e-01 9.63470578e-01 6.98553205e-01 -2.00466126e-01
1.33726656e-01 7.06105649e-01 3.73170972e-01 8.17958951e-01
4.58887339e-01 -2.30562657e-01 -1.85022518e-01 -2.01309677e-02
4.37754422e-01 4.28971559e-01 -3.34937096e-01 -7.43386865e-01
-1.14078784e+00 6.64872050e-01 8.68499339e-01 3.55693132e-01
-2.10019737e-01 -3.61869968e-02 5.40990710e-01 -7.71757141e-02
4.83126611e-01 2.19098896e-01 -6.57507896e-01 4.08030480e-01
-9.25481975e-01 -2.27447480e-01 6.78985178e-01 1.02876401e+00
7.90742815e-01 -2.16177478e-01 -2.07131714e-01 7.39070296e-01
1.29045427e-01 5.14396392e-02 8.58886063e-01 -3.27155590e-01
2.61853307e-01 7.87435532e-01 -3.13529462e-01 -4.67111468e-01
-6.31979823e-01 -1.03671360e+00 -9.38846171e-01 -1.84855059e-01
1.66130647e-01 -8.73211995e-02 -1.54048955e+00 1.90321445e+00
5.98689258e-01 1.63916141e-01 1.42795861e-01 4.79995281e-01
7.19049573e-01 3.15003693e-01 4.94384974e-01 -1.46110848e-01
1.24204886e+00 -1.27663875e+00 -5.77383101e-01 -3.33860964e-01
7.44665325e-01 -3.77422661e-01 9.17422354e-01 -4.23030816e-02
-7.22379267e-01 -4.99582499e-01 -9.01447892e-01 -9.41895917e-02
-4.51434314e-01 6.86029568e-02 7.95619547e-01 5.46900928e-01
-1.08399200e+00 3.85238469e-01 -1.03978872e+00 -4.64219719e-01
7.92949259e-01 6.89030826e-01 -5.08731067e-01 -2.13228330e-01
-9.41757560e-01 7.16692209e-01 6.60672545e-01 -2.72370894e-02
-1.01928794e+00 -1.08713543e+00 -8.61385643e-01 1.60810769e-01
5.02758324e-01 -6.82824790e-01 1.16296279e+00 -1.08533442e+00
-1.16132772e+00 8.70957911e-01 3.05083804e-02 -5.12982488e-01
5.84097385e-01 5.57333454e-02 -2.29303285e-01 1.32125631e-01
1.43696710e-01 1.24243319e+00 8.04871440e-01 -1.34603953e+00
-6.65757239e-01 -2.67068774e-01 1.45455846e-03 3.48373234e-01
-8.68710041e-01 -9.50165868e-01 -6.75271928e-01 -6.76292419e-01
1.66124940e-01 -1.12521815e+00 -5.22310078e-01 4.78123605e-01
-3.49636286e-01 4.21944484e-02 6.28835618e-01 -4.59292829e-01
1.14752984e+00 -2.25321603e+00 -8.87343735e-02 1.97852015e-01
3.69100749e-01 1.61761820e-01 -5.68963541e-03 -9.49204490e-02
-7.72552192e-02 1.27430379e-01 -4.37253833e-01 -5.32790482e-01
-3.41307372e-01 6.77223265e-01 1.37163550e-01 3.22788566e-01
2.19688654e-01 9.02514756e-01 -9.14210916e-01 -6.54922783e-01
1.27015218e-01 3.81305009e-01 -7.82815039e-01 8.04394037e-02
-1.32987425e-01 5.32172859e-01 -3.45288694e-01 7.72350848e-01
6.22289598e-01 -5.73609293e-01 2.80857861e-01 -1.55985981e-01
1.36801556e-01 -8.71583447e-02 -7.74665177e-01 1.93448460e+00
-5.21932542e-01 9.86242369e-02 -4.81588729e-02 -8.85324478e-01
5.49418151e-01 2.58112192e-01 6.63912058e-01 -3.35781246e-01
7.97343627e-02 1.94490999e-01 2.04301737e-02 -3.43838871e-01
4.71318454e-01 -2.79778093e-01 -1.54377148e-01 1.39557049e-01
3.39200914e-01 1.32033333e-01 6.70840889e-02 2.44276509e-01
1.04515529e+00 -1.94450021e-01 3.74579251e-01 -2.83172339e-01
5.32724321e-01 1.32778093e-01 7.92887628e-01 7.36862719e-01
-4.10922587e-01 4.26952034e-01 3.02482732e-02 -6.08441889e-01
-7.09059715e-01 -1.25609374e+00 -3.43140960e-01 1.01783144e+00
7.57321790e-02 -1.32322777e-02 -5.70206285e-01 -1.17487967e+00
2.16565013e-01 5.20733714e-01 -9.48242188e-01 -2.64322490e-01
-6.31295502e-01 -8.38672578e-01 3.03991437e-01 6.56873167e-01
2.59770721e-01 -7.37139165e-01 -5.67661643e-01 4.67215091e-01
-7.95231108e-03 -1.09506178e+00 -6.26560569e-01 5.64323485e-01
-1.17056823e+00 -9.80085254e-01 -8.69088352e-01 -1.00142252e+00
1.20356750e+00 1.35059729e-01 7.41587520e-01 1.68599695e-01
-5.08446932e-01 6.16136968e-01 -1.64725959e-01 -3.20978075e-01
-5.75147867e-01 5.37992656e-01 -1.06978506e-01 1.33960798e-01
9.62750539e-02 -2.19379604e-01 -6.56014025e-01 -3.42637412e-02
-1.24945331e+00 3.50585997e-01 6.83583736e-01 1.10816002e+00
8.25148761e-01 -1.98635638e-01 5.26382804e-01 -1.26570833e+00
2.05955759e-01 -6.48072481e-01 -2.59346187e-01 3.93400133e-01
-8.06821346e-01 4.00858745e-02 7.16047525e-01 -5.97033918e-01
-1.05417192e+00 4.33746696e-01 -7.62397870e-02 -2.83286184e-01
5.50904796e-02 5.18068910e-01 1.95344463e-01 -1.60026401e-01
4.73049819e-01 1.49152786e-01 7.32215792e-02 -4.07270849e-01
3.00523609e-01 4.69519109e-01 5.31149328e-01 -4.76345390e-01
5.86398244e-01 5.78994870e-01 -1.59174159e-01 -5.28012633e-01
-9.65099037e-01 -4.65601265e-01 -8.23208272e-01 -3.99704985e-02
8.12276065e-01 -1.02106392e+00 -1.34489566e-01 2.99496055e-01
-6.83587432e-01 -2.49815002e-01 -7.01741397e-01 3.30959976e-01
-1.41776219e-01 3.72730970e-01 -7.02773035e-01 -2.47180924e-01
-5.01758218e-01 -1.40774310e+00 7.60847092e-01 4.72311586e-01
-1.14069894e-01 -1.24164450e+00 -1.31447554e-01 1.87875763e-01
4.01616037e-01 2.91053861e-01 1.06754947e+00 -9.54311788e-01
-7.05921471e-01 -3.35738629e-01 -1.31935388e-01 2.75485307e-01
4.92834270e-01 -4.21046823e-01 -9.73754883e-01 -7.43720949e-01
-2.63114303e-01 -2.98975557e-01 1.07021379e+00 2.46465787e-01
1.24568725e+00 -3.13168645e-01 -8.84864509e-01 6.39424443e-01
1.55781806e+00 1.77499615e-02 5.81326382e-03 2.29551315e-01
7.05049157e-01 2.63212740e-01 3.29579890e-01 4.39859748e-01
4.39945549e-01 2.01745480e-01 3.31215858e-01 -4.22650725e-01
-2.81135052e-01 -2.62972981e-01 -7.07668364e-02 8.49743605e-01
5.41893363e-01 -3.10927242e-01 -1.13026822e+00 9.43224192e-01
-1.49324679e+00 -2.40997210e-01 4.60590005e-01 2.14499426e+00
1.18372357e+00 3.27234417e-01 -4.38421339e-01 -2.53182918e-01
4.55100894e-01 -1.19801499e-01 -7.69977093e-01 -1.26783445e-01
2.34628439e-01 3.84274125e-01 7.20014989e-01 4.35277313e-01
-1.30278838e+00 7.49340534e-01 5.89290524e+00 6.24855816e-01
-1.19919944e+00 2.91932166e-01 7.57935762e-01 -2.92928424e-02
-3.49211872e-01 -2.21967801e-01 -8.40438783e-01 1.73286632e-01
7.75723279e-01 -1.82384044e-01 -1.28493950e-01 8.87237549e-01
-2.12872550e-01 -1.30185604e-01 -1.37209189e+00 5.98282635e-01
1.20145477e-01 -1.52545655e+00 3.59666407e-01 -1.46480307e-01
7.35797226e-01 1.19796395e-01 2.74428457e-01 4.75878924e-01
-9.75954253e-03 -7.25277066e-01 4.49006498e-01 3.43679786e-01
9.68154967e-01 -4.07382041e-01 7.80181766e-01 2.86346883e-01
-1.09027290e+00 -3.32786798e-01 -1.25596777e-01 5.58018029e-01
-3.76651287e-02 4.87193406e-01 -1.39768481e+00 5.99768162e-01
4.60017383e-01 7.51787961e-01 -8.02459598e-01 1.02504790e+00
2.09119588e-01 4.33808148e-01 -3.69583219e-01 3.45894754e-01
4.22504991e-01 3.83771390e-01 2.77285129e-01 1.29381013e+00
1.79313287e-01 1.75510328e-02 5.17185807e-01 5.43400049e-01
-3.59744728e-01 9.05440152e-02 -3.45498919e-01 5.81378117e-02
2.93584913e-01 1.19437170e+00 -1.03082883e+00 -5.74683189e-01
-4.25321370e-01 9.67821360e-01 3.68293196e-01 1.20820291e-01
-7.32861042e-01 3.96049470e-02 3.98145199e-01 1.44622907e-01
2.20931783e-01 1.39453158e-01 -4.12241220e-01 -1.11029518e+00
-8.07831883e-02 -5.14892757e-01 8.25432897e-01 -7.20392391e-02
-1.25990915e+00 7.17044413e-01 -1.15401112e-01 -1.13603485e+00
1.73115149e-01 -4.20821458e-01 -1.55168578e-01 3.75166267e-01
-1.80437768e+00 -1.33102608e+00 -4.02577877e-01 4.79131222e-01
8.51520300e-01 4.91535403e-02 1.00454903e+00 5.05817592e-01
-5.40150166e-01 9.82198536e-01 2.42611185e-01 3.16092730e-01
7.71697938e-01 -1.15405810e+00 -3.53440717e-02 5.68634331e-01
-1.21417753e-01 5.20214498e-01 3.30447823e-01 -7.00763524e-01
-9.65074897e-01 -1.49202180e+00 6.73196673e-01 -2.62002110e-01
2.34466940e-01 -4.29851174e-01 -1.07111490e+00 8.73853683e-01
8.02120412e-05 3.20017397e-01 9.53061581e-01 -8.38652551e-02
-1.43854573e-01 -2.71242321e-01 -1.45008802e+00 4.39077616e-01
7.67776906e-01 -1.83977872e-01 -5.43133378e-01 4.55665022e-01
9.49358761e-01 -6.65877640e-01 -9.20838058e-01 6.03112161e-01
4.71240968e-01 -4.60575074e-01 9.49307680e-01 -5.02094090e-01
1.87579662e-01 -1.37746945e-01 1.14652611e-01 -1.12428117e+00
-2.94205070e-01 -2.05028459e-01 9.66660306e-02 9.28217173e-01
6.18802011e-01 -6.53288245e-01 1.00093150e+00 9.59700167e-01
-4.67833161e-01 -9.67727482e-01 -9.25277889e-01 -5.70355177e-01
2.62974977e-01 -4.15689461e-02 4.35282767e-01 1.15938842e+00
-1.88145399e-01 -2.23177765e-02 -3.23102847e-02 3.00104946e-01
5.43494403e-01 -1.79046032e-03 4.84965384e-01 -1.01894164e+00
-2.61217386e-01 -1.33945510e-01 -4.14512992e-01 -9.04554605e-01
-6.96198121e-02 -1.36393237e+00 2.95908768e-02 -1.46638823e+00
5.20110428e-01 -9.49728608e-01 -8.46292198e-01 9.09828186e-01
-2.99729884e-01 1.40026793e-01 1.70955867e-01 1.45237982e-01
-5.35183012e-01 4.34235692e-01 1.39973676e+00 -4.29581344e-01
-2.29905382e-01 2.20512189e-02 -5.93551040e-01 6.54837251e-01
6.37094378e-01 -7.05844045e-01 -4.63370860e-01 -6.07350707e-01
-3.37008774e-01 -1.38649076e-01 8.91183689e-02 -1.10175681e+00
4.01215196e-01 1.35410558e-02 6.33270621e-01 -5.06515920e-01
1.52400851e-01 -1.17713654e+00 9.64654610e-02 7.22573876e-01
-5.61898410e-01 -3.23469967e-01 5.19261181e-01 9.12239313e-01
-7.23183900e-02 -2.47844741e-01 9.89644170e-01 -3.63283634e-01
-7.63381660e-01 6.70225263e-01 -2.36817002e-01 -2.96805315e-02
1.28439665e+00 -3.15623432e-01 1.92443296e-01 1.46450445e-01
-1.30324781e+00 5.83724201e-01 4.08980101e-01 4.57856536e-01
5.76482296e-01 -1.00285876e+00 -3.13037336e-01 3.84884149e-01
3.22744161e-01 3.56992453e-01 4.47801858e-01 8.45703781e-01
-5.52121699e-01 1.99565932e-01 -7.25638792e-02 -8.08813393e-01
-1.12128997e+00 7.02432573e-01 3.89006078e-01 -7.00549901e-01
-7.32736230e-01 1.04868734e+00 5.04604697e-01 -6.11715019e-01
5.48453271e-01 -6.95041895e-01 -9.20397695e-03 -9.54888761e-03
1.74122199e-01 -2.14861169e-01 4.45550419e-02 -2.55765647e-01
-2.80991793e-01 2.41124168e-01 -7.62725532e-01 2.90371925e-01
1.36156917e+00 -3.60931814e-01 1.34957701e-01 4.15505022e-01
1.24436927e+00 -2.15681687e-01 -1.17275691e+00 -5.84198117e-01
4.16418985e-02 -1.81804553e-01 2.16565188e-02 -1.03612864e+00
-1.11686969e+00 7.99934328e-01 8.71201277e-01 -4.28424239e-01
1.18164182e+00 -4.33083251e-02 1.04225910e+00 2.79048115e-01
2.89757729e-01 -9.56834376e-01 2.40013804e-02 2.95143485e-01
2.61308581e-01 -1.26475024e+00 8.85328576e-02 -4.24385577e-01
-5.78350604e-01 1.09850240e+00 8.47498715e-01 2.75734931e-01
8.00203085e-01 1.84843883e-01 7.47344419e-02 1.23702567e-02
-6.45254970e-01 8.11218694e-02 9.95857641e-02 3.15414101e-01
1.69661582e-01 2.22385600e-01 -3.87275219e-02 6.36399329e-01
2.67129898e-01 2.60924041e-01 5.60723603e-01 1.21983528e+00
-4.13929880e-01 -1.22481573e+00 2.41576191e-02 6.97703063e-01
-4.75900948e-01 -8.25038180e-02 1.04317985e-01 1.02913451e+00
5.19867837e-01 2.17498645e-01 8.54236931e-02 1.71032604e-02
2.53116250e-01 2.73449659e-01 3.10495287e-01 -1.03362167e+00
-7.03744948e-01 -1.37660857e-02 -2.80276626e-01 -2.07904667e-01
-2.64075726e-01 -5.28052330e-01 -1.42156327e+00 2.90509552e-01
-3.61031592e-01 4.58664708e-02 3.67925823e-01 7.83646941e-01
2.70445704e-01 7.87736297e-01 4.60575938e-01 -4.12356675e-01
-6.35419726e-01 -5.79840779e-01 -2.87684411e-01 4.41441029e-01
6.43312633e-01 -6.00302100e-01 -7.57040605e-02 1.94985554e-01] | [14.73132038116455, -2.29110050201416] |
6d02b1dc-58fe-4280-a7cc-a584d353dd53 | spsg-self-supervised-photometric-scene | 2006.14660 | null | https://arxiv.org/abs/2006.14660v2 | https://arxiv.org/pdf/2006.14660v2.pdf | SPSG: Self-Supervised Photometric Scene Generation from RGB-D Scans | We present SPSG, a novel approach to generate high-quality, colored 3D models of scenes from RGB-D scan observations by learning to infer unobserved scene geometry and color in a self-supervised fashion. Our self-supervised approach learns to jointly inpaint geometry and color by correlating an incomplete RGB-D scan with a more complete version of that scan. Notably, rather than relying on 3D reconstruction losses to inform our 3D geometry and color reconstruction, we propose adversarial and perceptual losses operating on 2D renderings in order to achieve high-resolution, high-quality colored reconstructions of scenes. This exploits the high-resolution, self-consistent signal from individual raw RGB-D frames, in contrast to fused 3D reconstructions of the frames which exhibit inconsistencies from view-dependent effects, such as color balancing or pose inconsistencies. Thus, by informing our 3D scene generation directly through 2D signal, we produce high-quality colored reconstructions of 3D scenes, outperforming state of the art on both synthetic and real data. | ['Matthias Nießner', 'Julien Valentin', 'Yawar Siddiqui', 'Justus Thies', 'Angela Dai'] | 2020-06-25 | null | http://openaccess.thecvf.com//content/CVPR2021/html/Dai_SPSG_Self-Supervised_Photometric_Scene_Generation_From_RGB-D_Scans_CVPR_2021_paper.html | http://openaccess.thecvf.com//content/CVPR2021/papers/Dai_SPSG_Self-Supervised_Photometric_Scene_Generation_From_RGB-D_Scans_CVPR_2021_paper.pdf | cvpr-2021-1 | ['scene-generation'] | ['computer-vision'] | [ 6.14221632e-01 3.58307570e-01 4.24115300e-01 -4.60207820e-01
-1.12331128e+00 -8.39977145e-01 4.54027861e-01 -2.62870401e-01
6.04536757e-02 6.09743118e-01 1.08109422e-01 3.95857543e-02
1.80528164e-01 -9.68271315e-01 -1.12578166e+00 -5.76017857e-01
-3.57264467e-02 5.41115105e-01 -6.27034670e-03 3.24812904e-02
-1.03421502e-01 7.96530604e-01 -1.69509828e+00 3.60068798e-01
6.58038795e-01 1.03399074e+00 1.96781871e-03 9.77321327e-01
-1.92645460e-01 9.05793965e-01 -2.16505185e-01 -3.29795718e-01
7.63692439e-01 -7.08345413e-01 -3.38530034e-01 6.54150724e-01
7.80314505e-01 -7.88086176e-01 -5.06016076e-01 9.31096733e-01
1.21016316e-01 -2.20787048e-01 5.15328407e-01 -1.05116248e+00
-3.60464603e-01 -1.66026086e-01 -5.66819549e-01 -6.32438242e-01
8.92459214e-01 4.15064961e-01 7.44533598e-01 -7.68570364e-01
8.30622017e-01 1.29363084e+00 6.46293342e-01 5.04613161e-01
-1.84707570e+00 -5.18568456e-01 -1.90997007e-03 -4.21736807e-01
-1.18961966e+00 -5.26577890e-01 1.46434271e+00 -3.48809034e-01
2.55845994e-01 2.92860776e-01 9.48423684e-01 1.32086134e+00
3.24515961e-02 3.17444801e-01 1.62471044e+00 -3.19680631e-01
4.90488648e-01 -1.36531129e-01 -9.49975789e-01 8.45197082e-01
-1.32258353e-03 4.46105748e-01 -8.06924522e-01 -1.58555523e-01
1.30801487e+00 1.19084110e-02 -3.81436825e-01 -8.31357837e-01
-1.34317434e+00 3.25600922e-01 4.30570930e-01 -5.17549217e-01
-5.75807452e-01 4.08572882e-01 -3.42236131e-01 1.71779290e-01
6.95713282e-01 3.78395110e-01 -4.71645445e-01 -4.50516753e-02
-8.56379330e-01 2.39581063e-01 4.72711354e-01 9.59295630e-01
1.12267327e+00 4.11665469e-01 2.43666425e-01 2.15393513e-01
4.34569508e-01 1.22728801e+00 -3.17951828e-01 -1.80777609e+00
3.10453117e-01 3.04057926e-01 4.11063910e-01 -7.88807154e-01
-5.35420142e-02 -1.58214375e-01 -7.48525620e-01 9.68575001e-01
3.31187576e-01 2.23606210e-02 -1.17915618e+00 1.72633338e+00
4.83845055e-01 1.03140846e-01 9.65268090e-02 1.06543362e+00
6.79612875e-01 4.95557100e-01 -4.34190869e-01 -1.12804778e-01
7.62327671e-01 -3.24635863e-01 -5.10790288e-01 -2.81289905e-01
-2.17435867e-01 -8.80271018e-01 1.03651130e+00 5.25477588e-01
-1.44458532e+00 -4.98280019e-01 -1.01128459e+00 -1.13392420e-01
1.54159188e-01 -3.95356953e-01 3.97886515e-01 5.33428073e-01
-1.04349577e+00 5.30196011e-01 -8.97269726e-01 1.01449527e-01
4.58663136e-01 -2.37356305e-01 -5.55033326e-01 -6.22234821e-01
-7.52074242e-01 5.99527359e-01 -1.27236068e-01 -2.05866322e-01
-1.23805726e+00 -9.82732236e-01 -9.48903799e-01 -5.13754368e-01
4.06303167e-01 -7.75367260e-01 8.88424695e-01 -9.71103609e-01
-1.61056221e+00 1.11511648e+00 -3.91761437e-02 -1.63602438e-02
8.70943308e-01 -1.24974348e-01 -2.25784406e-01 5.70218801e-01
1.20618805e-01 6.82218492e-01 1.09428990e+00 -2.16802716e+00
-1.17351145e-01 -3.12928379e-01 5.96191697e-02 2.04498813e-01
6.34570062e-01 -7.12842584e-01 -4.66075003e-01 -4.96159464e-01
7.87106633e-01 -7.04048216e-01 -3.21491271e-01 8.08070958e-01
-6.41978502e-01 9.40484226e-01 6.59792066e-01 -7.86502481e-01
-9.35370475e-02 -2.02436948e+00 2.03209177e-01 1.54094979e-01
1.80325493e-01 -4.62491810e-01 -2.08614290e-01 3.20920020e-01
-9.13306326e-02 -6.63652942e-02 -6.52471304e-01 -9.12238836e-01
-1.00884788e-01 5.59741437e-01 -5.19071937e-01 6.61034763e-01
4.87885416e-01 8.21442783e-01 -1.19928372e+00 -2.58357286e-01
6.68405533e-01 8.15865338e-01 -6.59879863e-01 6.53690040e-01
-6.55052722e-01 1.29117882e+00 -3.54250252e-01 8.09254706e-01
9.68010783e-01 -8.44606981e-02 8.75531957e-02 -3.86953861e-01
4.44055162e-02 1.52328923e-01 -1.11542714e+00 2.48246121e+00
-5.79325676e-01 4.09141660e-01 3.73162270e-01 -5.46625555e-01
9.03643012e-01 1.62976891e-01 7.14297593e-01 -1.05505836e+00
-2.12271258e-01 8.12840313e-02 -7.89520621e-01 -2.42341250e-01
3.44525009e-01 -5.44243574e-01 -1.08849712e-01 5.44288933e-01
-6.28545061e-02 -1.30410671e+00 -6.16048753e-01 4.38502371e-01
1.21022856e+00 7.68218756e-01 -2.90115356e-01 3.52819711e-01
9.09601450e-02 6.18341453e-02 4.91388917e-01 6.13122761e-01
2.51255661e-01 1.30253255e+00 3.92720163e-01 -3.37917298e-01
-1.68524361e+00 -1.89804208e+00 1.08586125e-01 -1.57898255e-02
2.59610981e-01 -5.27634099e-03 -3.24173927e-01 -3.58892381e-01
1.28501415e-01 7.84511030e-01 -6.88352406e-01 -1.34420007e-01
-4.68438238e-01 -1.14294440e-01 1.96689233e-01 4.15064543e-02
4.58503217e-01 -6.18753850e-01 -6.85664117e-01 1.87447071e-01
-1.47631913e-01 -1.40185118e+00 -1.67253673e-01 3.42727304e-01
-9.91405189e-01 -1.32354867e+00 -5.58588982e-01 -2.83715166e-02
8.54928136e-01 2.28429824e-01 1.59456372e+00 -1.48193091e-01
-4.78978366e-01 8.59111309e-01 -3.89032334e-01 2.76923925e-02
-7.96600044e-01 -8.95790160e-01 -1.35607928e-01 2.44290009e-01
-6.06031775e-01 -1.04548872e+00 -7.20265031e-01 1.08546160e-01
-1.07028699e+00 5.58091998e-01 4.15007591e-01 5.57616532e-01
9.93221819e-01 -2.88922875e-03 -1.50962591e-01 -9.43637311e-01
-2.97157675e-01 -1.29184946e-01 -7.59784698e-01 -7.48468861e-02
-3.38317037e-01 1.82823792e-01 7.05981672e-01 -5.34335412e-02
-1.33951950e+00 4.34556574e-01 3.19668930e-03 -9.23819005e-01
-3.48814696e-01 -2.68356442e-01 -2.93465853e-01 -1.01088040e-01
5.85644960e-01 4.21576440e-01 1.05899565e-01 -5.21772146e-01
6.74478114e-01 4.78976816e-02 9.00530219e-01 -7.82451153e-01
1.51267421e+00 1.13630891e+00 3.92675459e-01 -4.64912027e-01
-8.98198128e-01 2.64680777e-02 -6.81350768e-01 -3.73504311e-01
1.04227316e+00 -1.32412899e+00 -3.54981631e-01 5.34767985e-01
-1.11011541e+00 -6.65141165e-01 -6.89646721e-01 2.84928858e-01
-9.70120192e-01 3.67459327e-01 -4.64418739e-01 -8.73213947e-01
1.83323994e-01 -9.61929977e-01 1.61854243e+00 -1.25997141e-01
1.30270824e-01 -7.36562550e-01 1.21255694e-02 3.99886370e-01
1.35629937e-01 1.03962111e+00 8.65563989e-01 5.90643644e-01
-1.14974141e+00 5.50399385e-02 -7.10816681e-02 5.27424872e-01
3.75294656e-01 -9.50599089e-02 -1.18233371e+00 -1.07442901e-01
-8.45598355e-02 -6.01618528e-01 3.38762790e-01 6.47658706e-02
9.21153545e-01 -1.03454247e-01 2.53468007e-01 1.17569554e+00
1.91695869e+00 -8.00902620e-02 6.51383698e-01 -1.56875432e-01
9.44505155e-01 5.89652896e-01 3.20075423e-01 5.24903119e-01
3.63682538e-01 5.81433117e-01 9.68288302e-01 -5.08986831e-01
-5.99574566e-01 -8.13814282e-01 8.70695859e-02 4.43186969e-01
-7.54899085e-02 -5.89961000e-02 -6.10772431e-01 1.15832768e-01
-1.33836353e+00 -7.11503506e-01 -1.99905723e-01 2.37196350e+00
8.97781074e-01 2.26005958e-03 -3.58056307e-01 1.28291622e-01
2.78252035e-01 3.94021451e-01 -9.71179426e-01 5.31347841e-02
-6.38116777e-01 4.87287700e-01 7.05015004e-01 6.94077194e-01
-5.02859414e-01 6.81997895e-01 6.21040964e+00 2.00330839e-01
-9.78944182e-01 1.25960216e-01 7.18510926e-01 -2.61906803e-01
-1.21181786e+00 1.58513024e-01 2.19595477e-01 2.56180048e-01
4.43018109e-01 3.88541281e-01 8.65676999e-01 4.58051711e-01
2.57796854e-01 -2.90480703e-01 -1.15813494e+00 1.21108103e+00
1.23616986e-01 -1.42008507e+00 1.18255049e-01 1.80031862e-02
1.10797274e+00 -8.78118351e-02 -7.45028555e-02 -4.91316199e-01
8.89769495e-01 -8.57960045e-01 1.23602319e+00 8.34461153e-01
1.28574836e+00 -5.47098577e-01 8.26592185e-03 2.82451570e-01
-8.23803782e-01 3.96071315e-01 -4.14107814e-02 1.43272743e-01
6.34933352e-01 9.92291987e-01 -2.82814711e-01 8.20881605e-01
7.09676325e-01 7.98391342e-01 -3.05015236e-01 5.11456728e-01
-7.26834238e-01 2.81428814e-01 -3.04153144e-01 8.37322533e-01
-5.50818481e-02 -4.16980624e-01 5.66383123e-01 4.33979779e-01
2.84095645e-01 3.62544715e-01 9.87771302e-02 1.41766989e+00
4.57231479e-04 -5.97882688e-01 -7.19007075e-01 3.39929581e-01
4.40254450e-01 9.27931428e-01 -4.83859152e-01 -2.05601245e-01
-2.64645368e-01 1.49925458e+00 4.84900214e-02 6.79692984e-01
-7.55109191e-01 4.94232625e-02 8.68368268e-01 3.74794155e-01
2.43394464e-01 -5.29248416e-01 -4.87981886e-01 -1.19391918e+00
1.60323560e-01 -5.65831423e-01 -2.39084542e-01 -1.55053246e+00
-1.27500653e+00 4.54783648e-01 -1.93470746e-01 -1.65277708e+00
-2.61895716e-01 -2.46200651e-01 -3.28469008e-01 1.00236368e+00
-1.75299215e+00 -1.16961920e+00 -5.69848895e-01 8.14481497e-01
9.86243710e-02 3.27371895e-01 8.91856432e-01 -8.73762369e-03
2.99383730e-01 4.56288457e-02 3.24652717e-02 -6.96663484e-02
7.26884604e-01 -1.24915040e+00 5.36461771e-01 8.81167591e-01
2.17172921e-01 -1.13339707e-01 7.85393894e-01 -3.90159637e-01
-1.97790551e+00 -1.03925347e+00 4.73257229e-02 -5.84258437e-01
-2.14930661e-02 -4.72807854e-01 -5.86252749e-01 4.54050839e-01
-1.99974343e-01 5.93954086e-01 1.88453078e-01 -6.22156680e-01
-7.60484397e-01 -2.18729496e-01 -1.62039721e+00 3.76291782e-01
1.42878842e+00 -8.98824155e-01 -3.14292610e-01 1.25362217e-01
9.06929255e-01 -7.40735412e-01 -8.36393118e-01 2.17079580e-01
4.02009368e-01 -1.50089765e+00 1.40484822e+00 -2.48732045e-01
9.75566208e-01 -6.92678154e-01 -5.42318046e-01 -1.40282106e+00
1.36258960e-01 -7.00352192e-01 -1.71098914e-02 9.33900237e-01
8.14538226e-02 -3.88611794e-01 9.93094623e-01 7.32731521e-01
-2.43539661e-01 -2.04058871e-01 -9.74490941e-01 -5.16447484e-01
-9.25823748e-02 -7.66064882e-01 7.14612424e-01 9.64218378e-01
-8.64292085e-01 -2.10635960e-01 -5.47490716e-01 5.23096323e-01
1.43253565e+00 5.05874693e-01 1.10846031e+00 -9.35046971e-01
-5.57893753e-01 2.21243292e-01 -4.16436762e-01 -1.04926777e+00
-4.57315072e-02 -4.75297570e-01 2.93824643e-01 -1.41389120e+00
-1.48151472e-01 -7.09940135e-01 1.21013768e-01 1.82549849e-01
2.59852648e-01 7.46163487e-01 2.89517343e-01 1.21442780e-01
-2.99108535e-01 6.47637546e-01 1.59288847e+00 -1.46445408e-01
-2.38041375e-02 -4.38463032e-01 -4.33752716e-01 6.69400632e-01
2.55676508e-01 -4.69565213e-01 -4.17453796e-01 -7.80324757e-01
3.53184015e-01 6.00022435e-01 7.90169895e-01 -9.97104168e-01
-1.11276187e-01 -2.56683707e-01 1.11369503e+00 -6.51296616e-01
8.63466978e-01 -1.09865224e+00 6.66454315e-01 7.14543760e-02
-1.62980855e-01 -2.82162666e-01 -2.21185721e-02 8.05101216e-01
-7.45161548e-02 5.33615530e-01 9.06910539e-01 -6.37828827e-01
-5.24492025e-01 5.42621851e-01 9.94461179e-02 1.49900228e-01
7.37295568e-01 -4.18554693e-01 8.28836262e-02 -8.10431361e-01
-6.34626210e-01 -2.56926924e-01 1.33963311e+00 1.12677805e-01
8.94917011e-01 -1.42237508e+00 -6.44127727e-01 7.14748442e-01
2.74060041e-01 8.16449642e-01 4.41609025e-01 1.51956156e-01
-8.45397830e-01 -2.34542400e-01 -2.99288273e-01 -8.40337455e-01
-6.99741960e-01 3.08562934e-01 2.58149028e-01 1.49167091e-01
-9.99284983e-01 7.30876386e-01 1.80062249e-01 -7.13185251e-01
5.19207865e-02 -2.37043872e-01 8.67388606e-01 -5.99778712e-01
2.23603889e-01 -1.20853029e-01 -7.98624940e-03 -6.08057082e-01
-2.27530822e-01 9.35110569e-01 8.00341845e-01 -4.85803336e-01
1.50059652e+00 -3.38657588e-01 1.00167252e-01 4.89102751e-01
1.35448110e+00 4.44391221e-01 -2.32852626e+00 -3.35063040e-01
-6.95485473e-01 -1.00805271e+00 2.03312010e-01 -9.66677964e-01
-1.36157572e+00 6.06342435e-01 4.28346545e-01 -2.90993571e-01
1.21928430e+00 1.11531407e-01 7.43194342e-01 -5.93253039e-02
8.53101134e-01 -6.81820214e-01 2.61192441e-01 8.49867016e-02
8.19142699e-01 -1.24762630e+00 2.96174675e-01 -4.72520947e-01
-4.02550399e-01 1.05100656e+00 2.71903753e-01 -2.05682740e-01
4.41706747e-01 2.80572146e-01 3.20676476e-01 -1.29203364e-01
-4.73953098e-01 3.59448195e-02 8.14392492e-02 8.88650119e-01
6.67788507e-03 2.34230720e-02 7.57131040e-01 -1.18632123e-01
-1.21724963e-01 -2.84094363e-01 5.91999531e-01 8.77550542e-01
1.08780198e-01 -1.16765070e+00 -7.60599017e-01 5.73937669e-02
2.01422662e-01 5.24652675e-02 -3.06889385e-01 6.30058110e-01
1.24955550e-01 7.45508730e-01 3.37356716e-01 -4.14355248e-01
4.20583427e-01 -2.29874507e-01 8.71160567e-01 -4.83667225e-01
2.57475317e-01 5.59288301e-02 7.11999312e-02 -1.25799942e+00
-7.07831800e-01 -5.46219528e-01 -1.14054382e+00 -2.97765106e-01
3.80899996e-01 -3.21198583e-01 8.71439576e-01 5.37120342e-01
2.85587400e-01 3.73836279e-01 1.22386074e+00 -1.33647060e+00
-3.28191705e-02 -3.46925557e-01 -8.42212856e-01 8.89997423e-01
8.31904233e-01 -5.53244948e-01 -7.56344199e-01 3.39913934e-01] | [9.308293342590332, -3.1939940452575684] |
ddad88db-f05c-4f2d-99eb-88071707e022 | continuous-time-audiovisual-fusion-with | 2203.13285 | null | https://arxiv.org/abs/2203.13285v2 | https://arxiv.org/pdf/2203.13285v2.pdf | Continuous-Time Audiovisual Fusion with Recurrence vs. Attention for In-The-Wild Affect Recognition | In this paper, we present our submission to 3rd Affective Behavior Analysis in-the-wild (ABAW) challenge. Learningcomplex interactions among multimodal sequences is critical to recognise dimensional affect from in-the-wild audiovisual data. Recurrence and attention are the two widely used sequence modelling mechanisms in the literature. To clearly understand the performance differences between recurrent and attention models in audiovisual affect recognition, we present a comprehensive evaluation of fusion models based on LSTM-RNNs, self-attention and cross-modal attention, trained for valence and arousal estimation. Particularly, we study the impact of some key design choices: the modelling complexity of CNN backbones that provide features to the the temporal models, with and without end-to-end learning. We trained the audiovisual affect recognition models on in-the-wild ABAW corpus by systematically tuning the hyper-parameters involved in the network architecture design and training optimisation. Our extensive evaluation of the audiovisual fusion models shows that LSTM-RNNs can outperform the attention models when coupled with low-complex CNN backbones and trained in an end-to-end fashion, implying that attention models may not necessarily be the optimal choice for continuous-time multimodal emotion recognition. | ['Björn W. Schuller', 'Michel Valstar', 'Adria Mallol-Ragolta', 'Mani Kumar Tellamekala', 'Vincent Karas'] | 2022-03-24 | null | null | null | null | ['multimodal-emotion-recognition', 'multimodal-emotion-recognition'] | ['computer-vision', 'speech'] | [ 2.14454502e-01 -1.28817692e-01 2.60957152e-01 -4.46736485e-01
-8.14692557e-01 -4.85483915e-01 5.12831569e-01 1.28645990e-02
-7.17076898e-01 2.68156260e-01 4.15603638e-01 1.33613214e-01
-1.20673299e-01 1.39459958e-02 -2.60933727e-01 -7.68359601e-01
-4.77920771e-01 3.65801781e-01 -5.16246557e-01 -5.90916336e-01
-2.23896414e-01 3.93446594e-01 -1.90184915e+00 5.45371294e-01
2.95526206e-01 1.63408196e+00 -3.89782995e-01 1.36667800e+00
-1.08172625e-01 1.21169496e+00 -6.16237879e-01 -5.94080329e-01
-3.01277936e-01 -4.51391578e-01 -8.56594384e-01 -1.73034951e-01
1.46224692e-01 1.88436180e-01 -7.39744231e-02 4.38574106e-01
9.54743087e-01 4.31082159e-01 6.05197608e-01 -1.41176534e+00
-1.00792319e-01 4.97074753e-01 -2.87209332e-01 4.32849437e-01
3.63034487e-01 2.58637309e-01 1.06156397e+00 -8.36906850e-01
3.19156766e-01 1.32595098e+00 8.20150971e-01 4.07969236e-01
-1.31667256e+00 -3.66620779e-01 2.91189820e-01 5.86105824e-01
-1.04207635e+00 -7.66371787e-01 8.49144399e-01 -3.37844849e-01
1.57281113e+00 5.40834010e-01 8.67992818e-01 1.70602930e+00
1.49047732e-01 9.99193907e-01 6.99423909e-01 -2.14821056e-01
1.29559547e-01 -6.93945438e-02 -5.91177605e-02 1.55263841e-01
-9.99724984e-01 1.74219400e-01 -9.34226394e-01 -4.28290248e-01
1.36134624e-01 -4.96187150e-01 -4.52929921e-02 8.58034641e-02
-8.84497821e-01 8.44327569e-01 3.38443816e-01 4.35339540e-01
-7.89099216e-01 6.96355700e-01 1.21085250e+00 7.09297359e-01
6.68795705e-01 5.95133603e-01 -6.66292012e-01 -8.38472366e-01
-7.96191037e-01 -1.38163920e-02 5.28277338e-01 2.36462548e-01
2.90503353e-01 5.24628520e-01 -4.15375203e-01 1.20462894e+00
5.64305633e-02 2.23349109e-01 6.50124550e-01 -8.19806397e-01
9.29401517e-02 1.59684479e-01 -2.85283476e-01 -7.16279089e-01
-7.68185139e-01 -2.85052270e-01 -8.93351138e-01 -2.99288034e-02
1.26219103e-02 -6.01112068e-01 -8.76087070e-01 1.91129482e+00
1.45978257e-01 -1.37788832e-01 1.34307981e-01 9.08364952e-01
8.50808322e-01 7.91711509e-01 4.73267049e-01 -5.09467304e-01
1.33672547e+00 -8.12683284e-01 -1.19566512e+00 -2.13387176e-01
6.62761867e-01 -8.36124480e-01 8.45474303e-01 5.33068240e-01
-1.42424119e+00 -6.29789233e-01 -8.67931604e-01 4.01565880e-02
-4.77112263e-01 1.08228447e-02 5.28687537e-01 4.69127595e-01
-1.12329376e+00 4.84774083e-01 -5.78232646e-01 -3.80345136e-01
-2.36285739e-02 6.40140116e-01 -2.23733425e-01 7.59203315e-01
-1.50353229e+00 1.00766551e+00 2.75950730e-01 4.86712992e-01
-8.80601346e-01 -6.24726832e-01 -8.43579769e-01 1.42592102e-01
1.57071486e-01 -4.42656964e-01 1.50073731e+00 -1.80100286e+00
-1.85080564e+00 6.37966394e-01 -1.91335510e-02 -3.76710117e-01
1.04016334e-01 -2.97702998e-01 -5.07571399e-01 1.53697610e-01
-7.22415447e-01 7.93913364e-01 1.00858140e+00 -9.05095637e-01
-3.23289931e-01 -2.74861634e-01 -4.05016690e-01 3.07878554e-01
-2.96751499e-01 4.12632436e-01 -2.54469723e-01 -6.19481027e-01
-4.53827113e-01 -9.77449834e-01 -7.66672641e-02 -4.91991937e-01
-1.80620253e-01 -2.39585280e-01 6.70848191e-01 -4.61307645e-01
1.20581508e+00 -2.28829026e+00 5.93647540e-01 2.96617448e-01
-8.75878930e-02 1.65933520e-01 -5.63318372e-01 6.08521581e-01
-6.68054163e-01 -1.44491166e-01 3.35776359e-02 -7.41267383e-01
2.08869636e-01 2.65029203e-02 -4.60426867e-01 3.38581562e-01
5.36117613e-01 1.03354645e+00 -4.84379560e-01 -3.67316276e-01
1.77453637e-01 9.25173402e-01 -3.03239375e-01 6.76413178e-01
-2.06770673e-01 2.39274427e-01 4.82087173e-02 4.84318763e-01
-1.48082925e-02 3.41075718e-01 2.22188625e-02 -2.66488969e-01
-6.03106171e-02 6.54917434e-02 -6.20731056e-01 1.62913370e+00
-6.43588006e-01 9.98276711e-01 2.45741770e-01 -5.91822684e-01
8.46543550e-01 9.42035019e-01 5.80551684e-01 -1.04261088e+00
6.68981671e-01 -6.98353723e-03 1.62820533e-01 -7.58083820e-01
5.74457645e-01 -4.23935235e-01 -4.20383543e-01 3.14130872e-01
5.07471740e-01 1.69944372e-02 1.06788121e-01 -8.34290609e-02
8.68975222e-01 4.14217934e-02 6.58470914e-02 9.57781300e-02
5.36485195e-01 -3.58562112e-01 2.51562029e-01 4.36225057e-01
-2.84849405e-01 4.91597325e-01 8.97471428e-01 -3.68663937e-01
-7.91266561e-01 -4.98209983e-01 2.41742969e-01 1.96208715e+00
-7.00811863e-01 -4.70588475e-01 -5.54577887e-01 -2.62327671e-01
-4.13593203e-01 6.06589258e-01 -1.26361394e+00 -4.35999155e-01
-1.45705491e-01 -7.05639303e-01 1.01557183e+00 5.46129823e-01
2.23918799e-02 -1.52380836e+00 -1.03286064e+00 3.16538483e-01
-3.17216754e-01 -8.91785145e-01 -2.01523289e-01 9.15184319e-01
-3.14555973e-01 -4.77181733e-01 -6.72535002e-01 -3.47808748e-01
-2.23890066e-01 -5.02155662e-01 1.13838220e+00 -1.44496486e-01
-3.00960690e-01 8.03502202e-01 -5.23485065e-01 -5.64450502e-01
-2.18854442e-01 2.54705638e-01 -1.64007097e-01 3.39660078e-01
1.88616782e-01 -7.87038386e-01 -3.14987987e-01 -4.63161878e-02
-8.75960767e-01 -3.30091268e-01 5.16809344e-01 7.50821948e-01
2.86363006e-01 -5.76725066e-01 5.49244523e-01 -2.98672974e-01
9.33241129e-01 -4.21098113e-01 -7.80577362e-02 1.42838210e-01
-2.65351273e-02 1.88521117e-01 3.82017791e-01 -6.98425531e-01
-9.81936216e-01 2.18327701e-01 -5.54383636e-01 -8.58110249e-01
-1.65820062e-01 6.64457321e-01 -3.59509289e-02 1.68056443e-01
4.05572563e-01 -9.08106267e-02 -1.50158098e-02 -3.91948484e-02
5.41629970e-01 4.82231706e-01 5.12023389e-01 -3.09982389e-01
-2.12611798e-02 6.12268113e-02 -1.79589629e-01 -9.46453929e-01
-5.38726509e-01 -4.78436857e-01 -3.50565016e-01 -6.47250175e-01
1.15613949e+00 -7.92554319e-01 -1.38186753e+00 4.40927625e-01
-1.24626029e+00 -5.95187485e-01 -2.11387068e-01 3.01397771e-01
-8.95256400e-01 -2.39629999e-01 -8.29014003e-01 -1.47973859e+00
-7.14439273e-01 -1.01601136e+00 1.21920145e+00 -1.52846068e-01
-8.51367414e-01 -1.04418123e+00 5.40848732e-01 -1.42249629e-01
6.70186996e-01 4.32585686e-01 9.79310513e-01 -7.15356886e-01
4.11899805e-01 -3.73608023e-02 2.43707344e-01 1.36218101e-01
-4.77460593e-01 3.15692574e-01 -1.72026563e+00 6.19336776e-02
-6.79312125e-02 -1.01349640e+00 1.06729341e+00 4.17867601e-01
1.01224983e+00 -1.16093040e-01 3.22480679e-01 4.34676141e-01
9.50261533e-01 3.36840123e-01 7.81299591e-01 7.33867511e-02
5.26509285e-01 9.42838609e-01 4.95508105e-01 6.92136109e-01
1.42588124e-01 7.24063933e-01 6.43708110e-01 -3.46478015e-01
3.73321742e-01 2.51783848e-01 7.42842376e-01 9.44319069e-01
-4.69958559e-02 -2.95710087e-01 -8.41608107e-01 5.20178497e-01
-1.89864194e+00 -1.12915838e+00 2.22522318e-02 1.91652393e+00
7.34923720e-01 -1.36619881e-01 6.29775345e-01 2.72741258e-01
1.64654419e-01 3.98740798e-01 -3.36573571e-01 -1.26581752e+00
-2.40958899e-01 2.64054507e-01 -4.81401086e-02 5.15067101e-01
-1.00076938e+00 7.03158140e-01 6.17495012e+00 7.85114050e-01
-1.52317309e+00 5.53590767e-02 7.85245955e-01 -8.58477712e-01
-1.86524510e-01 -4.57167327e-01 -1.08688712e-01 1.69824764e-01
1.66907632e+00 3.50144744e-01 4.76392567e-01 4.09355551e-01
3.15425545e-01 1.14532180e-01 -1.17701745e+00 1.27378309e+00
1.34042352e-01 -6.28892541e-01 -3.36595803e-01 -2.22534567e-01
2.34822676e-01 1.39779478e-01 3.45685393e-01 7.12875366e-01
-1.07193710e-02 -1.32977092e+00 9.12127793e-01 7.17095554e-01
5.67569017e-01 -1.21636760e+00 9.16500926e-01 -2.16432810e-01
-1.01112103e+00 -2.54141837e-01 8.20879787e-02 6.56719729e-02
1.57256424e-01 1.54922768e-01 -5.40995181e-01 2.17620164e-01
9.35034931e-01 5.24383426e-01 -4.52381879e-01 4.95355815e-01
1.93255022e-01 5.57949066e-01 -1.86106443e-01 -2.66231179e-01
5.39545119e-01 1.31334856e-01 3.97215724e-01 1.73477173e+00
1.40941143e-01 1.27528757e-02 -3.03312272e-01 4.71327424e-01
1.70945212e-01 1.76693469e-01 -4.28286701e-01 -4.97508973e-01
-8.16114768e-02 1.49292147e+00 -3.53528887e-01 -8.57933685e-02
6.81595951e-02 1.05967796e+00 4.61251378e-01 5.65028369e-01
-1.01535010e+00 -3.27105701e-01 1.08672297e+00 -5.02383888e-01
5.77998877e-01 1.86067879e-01 6.26737326e-02 -5.92061996e-01
-3.46561164e-01 -1.06240559e+00 6.42476439e-01 -1.15249014e+00
-1.19567811e+00 1.16084051e+00 -1.98898286e-01 -7.31631160e-01
-7.09679186e-01 -5.20959854e-01 -7.33531535e-01 6.17695689e-01
-1.24487031e+00 -1.01227915e+00 -1.23471860e-02 6.42898262e-01
4.70002145e-01 1.23491652e-01 1.09856069e+00 2.29537368e-01
-6.90365911e-01 6.47267222e-01 -2.12653771e-01 -2.59604990e-01
6.98203146e-01 -1.09023964e+00 3.88511121e-02 2.57672101e-01
1.02747278e-02 1.53641194e-01 1.06580257e+00 -4.81319427e-02
-1.42881513e+00 -6.92931712e-01 7.27448285e-01 -2.72897899e-01
7.31932759e-01 -5.64390242e-01 -8.89473319e-01 5.50827861e-01
9.91012871e-01 -4.63163048e-01 1.06885779e+00 6.39448702e-01
-4.13215339e-01 7.26875290e-02 -6.24031425e-01 5.38164139e-01
4.39633548e-01 -7.71004677e-01 -2.83045977e-01 -1.32177263e-01
7.21641719e-01 -1.61423951e-01 -1.01735926e+00 4.09646183e-01
7.86967754e-01 -1.07908642e+00 6.73824370e-01 -7.94585943e-01
5.42041183e-01 1.65714100e-01 -2.54997015e-01 -1.52295589e+00
-1.51176736e-01 -8.84065628e-01 -7.35060722e-02 1.27719378e+00
5.11078537e-01 -2.15558365e-01 3.03964347e-01 6.36369765e-01
-5.26601896e-02 -9.28749561e-01 -1.10394657e+00 -1.45585507e-01
-1.35027543e-01 -1.02716148e+00 3.04181725e-01 7.65865922e-01
3.23028624e-01 9.21813905e-01 -6.36935592e-01 -3.26370358e-01
-2.02595234e-01 -2.44763225e-01 4.95426595e-01 -1.00570321e+00
-3.00309122e-01 -8.92911673e-01 -4.80361372e-01 -2.98979580e-01
2.97751039e-01 -4.80699986e-01 2.15624630e-01 -8.46692979e-01
-8.04768652e-02 3.08370382e-01 -4.70771015e-01 6.55989707e-01
3.76916751e-02 2.19225481e-01 4.34156477e-01 -3.65279555e-01
-8.51013064e-01 1.20445991e+00 5.86973310e-01 -1.34561688e-01
-3.10505778e-01 -3.46662641e-01 -3.10197920e-01 4.68944281e-01
6.64778173e-01 -2.71051556e-01 -4.62997288e-01 -1.21897414e-01
6.68661714e-01 2.74479270e-01 2.70374775e-01 -7.21348882e-01
1.07151054e-01 1.86271474e-01 3.95841509e-01 -6.63334608e-01
1.13391876e+00 -9.11768079e-01 1.26249027e-02 2.42575072e-02
-8.65501583e-01 3.27880174e-01 7.94507504e-01 1.92375258e-01
-3.51741552e-01 6.66045770e-02 5.73548436e-01 1.86564088e-01
-4.50485110e-01 9.62060988e-02 -9.68178451e-01 8.59163795e-03
6.51572108e-01 7.15466589e-02 7.83768296e-02 -9.63864684e-01
-1.14460146e+00 2.98199475e-01 -2.05582321e-01 5.55139124e-01
5.78277647e-01 -1.40614331e+00 -6.76876664e-01 1.20591395e-01
3.10072899e-01 -7.17053711e-01 6.76268160e-01 1.25316215e+00
1.47961423e-01 3.83938462e-01 -2.65923649e-01 -6.00257814e-01
-1.67029631e+00 6.06070399e-01 7.01948702e-01 -2.25138992e-01
2.83292800e-01 9.51291203e-01 1.19374192e-03 -5.20253837e-01
6.71988785e-01 -1.98453844e-01 -4.03987765e-01 6.74575329e-01
5.09208918e-01 1.67167768e-01 2.23051369e-01 -8.22265923e-01
-4.20923084e-01 2.78840989e-01 3.70167345e-02 -6.07235551e-01
1.30401838e+00 -1.30992070e-01 -8.11757743e-02 1.19735432e+00
1.53526783e+00 -6.15470588e-01 -1.03852379e+00 7.35102072e-02
-9.53916535e-02 1.96099237e-01 3.29212695e-01 -9.16621625e-01
-1.13330448e+00 1.26848507e+00 8.10438156e-01 3.57927710e-01
1.40602303e+00 -1.90835983e-01 4.50243324e-01 3.53756368e-01
-3.90526235e-01 -1.42913163e+00 3.08363676e-01 8.36839557e-01
1.21740377e+00 -9.14370000e-01 -4.26194906e-01 3.09127361e-01
-1.34939337e+00 1.19473529e+00 5.19392371e-01 2.80277520e-01
4.56859738e-01 2.68895924e-01 4.24701065e-01 -5.01268327e-01
-1.66923618e+00 -4.76698399e-01 4.72734183e-01 3.14410955e-01
7.82642066e-01 -2.48108327e-01 4.10150200e-01 6.04752839e-01
-2.56736308e-01 -4.27374572e-01 8.94072503e-02 6.14204586e-01
3.23170871e-02 -7.08619058e-01 -3.89969617e-01 1.68525815e-01
-5.51170170e-01 -8.06607306e-02 -9.08177555e-01 6.49584711e-01
-9.21299011e-02 9.53595817e-01 3.37786496e-01 -6.47164464e-01
5.53107500e-01 6.28400803e-01 3.25926542e-01 1.30727902e-01
-1.45446682e+00 4.34729308e-01 4.21579301e-01 -7.81073630e-01
-5.60186565e-01 -6.91229880e-01 -1.04459953e+00 -1.83319837e-01
-1.44427672e-01 1.19662307e-01 7.33359873e-01 1.12388325e+00
4.60720509e-01 9.50875401e-01 5.94357967e-01 -1.20158148e+00
-1.42540410e-01 -1.25482714e+00 -4.11279291e-01 1.86722979e-01
6.63612843e-01 -2.91833997e-01 -3.37483734e-01 -1.92346722e-01] | [13.377293586730957, 5.241767406463623] |
3e81e878-3e24-47e1-9ca9-8a63068197f3 | modelling-latent-translations-for-cross | 2107.11353 | null | https://arxiv.org/abs/2107.11353v1 | https://arxiv.org/pdf/2107.11353v1.pdf | Modelling Latent Translations for Cross-Lingual Transfer | While achieving state-of-the-art results in multiple tasks and languages, translation-based cross-lingual transfer is often overlooked in favour of massively multilingual pre-trained encoders. Arguably, this is due to its main limitations: 1) translation errors percolating to the classification phase and 2) the insufficient expressiveness of the maximum-likelihood translation. To remedy this, we propose a new technique that integrates both steps of the traditional pipeline (translation and classification) into a single model, by treating the intermediate translations as a latent random variable. As a result, 1) the neural machine translation system can be fine-tuned with a variant of Minimum Risk Training where the reward is the accuracy of the downstream task classifier. Moreover, 2) multiple samples can be drawn to approximate the expected loss across all possible translations during inference. We evaluate our novel latent translation-based model on a series of multilingual NLU tasks, including commonsense reasoning, paraphrase identification, and natural language inference. We report gains for both zero-shot and few-shot learning setups, up to 2.7 accuracy points on average, which are even more prominent for low-resource languages (e.g., Haitian Creole). Finally, we carry out in-depth analyses comparing different underlying NMT models and assessing the impact of alternative translations on the downstream performance. | ['Siva Reddy', 'Ivan Vulić', 'Julia Kreutzer', 'Edoardo Maria Ponti'] | 2021-07-23 | null | null | null | null | ['paraphrase-identification'] | ['natural-language-processing'] | [ 4.43879485e-01 1.55177116e-01 -4.88585174e-01 -2.70182967e-01
-1.58308339e+00 -7.02798903e-01 9.56246734e-01 6.67496696e-02
-6.59330606e-01 9.75659907e-01 3.49090099e-01 -5.28638422e-01
1.74501717e-01 -6.64318025e-01 -1.00459898e+00 -3.87123197e-01
4.57497746e-01 9.10465062e-01 -1.97701320e-01 -2.19092250e-01
3.12658958e-02 -1.40933797e-03 -1.21779728e+00 3.51663291e-01
1.36048377e+00 6.56853855e-01 1.32334813e-01 2.97199756e-01
-9.88594145e-02 6.45276606e-01 -3.15668017e-01 -1.15670526e+00
1.76775888e-01 -5.99031568e-01 -7.30176568e-01 -1.68003321e-01
4.14155453e-01 -2.19147682e-01 -8.60709697e-03 1.10081995e+00
5.28374076e-01 6.50054310e-03 7.59136379e-01 -8.98426890e-01
-8.49306583e-01 9.84671652e-01 -3.06074977e-01 1.58531331e-02
2.43864074e-01 4.16707218e-01 1.34012878e+00 -1.01036382e+00
8.38885009e-01 1.25756025e+00 6.44433439e-01 4.44596529e-01
-1.58152223e+00 -5.78391492e-01 -1.28662467e-01 3.70809108e-01
-1.11408687e+00 -8.49307716e-01 3.76595259e-01 -4.41355109e-01
1.29170895e+00 -8.37736949e-02 4.07466561e-01 1.65981698e+00
2.98135757e-01 8.89107227e-01 1.21915400e+00 -5.98267257e-01
2.37408414e-01 2.32808590e-01 -1.31100938e-01 6.52989984e-01
6.93428656e-03 1.27190277e-02 -6.30706549e-01 -1.13380656e-01
1.66107088e-01 -3.16535860e-01 -1.00194201e-01 -1.03531502e-01
-1.28547657e+00 9.88664985e-01 8.11767578e-02 2.36371055e-01
-3.42138916e-01 6.92488924e-02 5.95529199e-01 4.98304635e-01
8.12327266e-01 5.13712347e-01 -4.93369699e-01 -4.63168114e-01
-1.19810939e+00 5.23814596e-02 6.92669094e-01 9.82272804e-01
6.90172553e-01 -1.78920031e-01 -4.75987762e-01 1.12947488e+00
-1.00224964e-01 4.19685215e-01 6.30870700e-01 -6.54573083e-01
8.58681977e-01 3.31000835e-01 -8.63126814e-02 -2.84854949e-01
1.16647974e-01 -5.27499616e-01 -5.93326449e-01 -1.94515541e-01
6.00123882e-01 -1.50857955e-01 -6.64365590e-01 1.95989251e+00
-1.14082703e-02 -7.40279704e-02 9.18669775e-02 6.68017924e-01
6.01079948e-02 5.27345717e-01 2.21172757e-02 -2.56895125e-01
1.43937087e+00 -1.23737156e+00 -6.55419350e-01 -5.94532847e-01
7.84827292e-01 -8.64561439e-01 1.53881156e+00 1.31390318e-01
-1.24823904e+00 -2.67811716e-01 -9.54941094e-01 -5.23574531e-01
-2.87321270e-01 1.25741825e-01 4.47396785e-01 3.32164764e-01
-7.05957532e-01 7.94038236e-01 -8.48372340e-01 -3.86645079e-01
3.45170557e-01 -3.94729897e-02 -8.78328308e-02 -4.52314407e-01
-1.51102531e+00 1.42862761e+00 1.57939360e-01 -9.26721096e-02
-7.08826661e-01 -7.49560177e-01 -7.29234815e-01 1.48148611e-01
3.05329800e-01 -7.90359855e-01 1.29882061e+00 -1.04471278e+00
-1.55609679e+00 1.04431665e+00 -3.00521940e-01 -5.29530048e-01
1.03658414e+00 -3.24908555e-01 -1.07301019e-01 -2.38310128e-01
4.16331947e-01 3.95604670e-01 5.64544559e-01 -6.04136586e-01
-3.75382721e-01 -3.29371929e-01 -1.67650685e-01 3.15661967e-01
-1.46789223e-01 1.41823083e-01 -3.03618401e-01 -6.82318866e-01
-3.62798482e-01 -9.66447592e-01 1.33405566e-01 -1.35869399e-01
-2.70337820e-01 -3.06917071e-01 5.16593158e-02 -9.55885589e-01
1.09330082e+00 -1.97142661e+00 4.72405225e-01 -4.03202832e-01
-3.78526926e-01 1.32973522e-01 -1.20734021e-01 4.91350174e-01
2.84897178e-01 9.74966586e-02 -4.18536603e-01 -7.22065091e-01
2.67916322e-01 8.71531889e-02 -4.28940922e-01 3.23408127e-01
4.09976602e-01 1.25202918e+00 -9.72744167e-01 -4.41490263e-01
-1.08012110e-01 3.63400847e-01 -5.00951707e-01 -1.37880549e-01
-4.04595882e-01 3.96133840e-01 -3.86535972e-02 6.42400026e-01
2.17551365e-01 -1.68176740e-01 2.29994252e-01 5.12506887e-02
-8.24796036e-02 9.46445465e-01 -5.25103927e-01 2.10058475e+00
-8.32158267e-01 4.81371850e-01 -3.03268254e-01 -8.85679781e-01
5.82717419e-01 4.43229169e-01 -3.44120972e-02 -7.62543380e-01
8.35132152e-02 6.82302594e-01 1.59394681e-01 -4.08047587e-01
2.82201737e-01 -5.32118738e-01 -2.66245693e-01 7.76158750e-01
2.76948839e-01 5.39861284e-02 1.04543358e-01 5.57493754e-02
8.78440082e-01 5.64865053e-01 3.67578954e-01 -3.44828248e-01
1.50953725e-01 1.03851356e-01 5.67999840e-01 6.98322475e-01
-1.55879900e-01 3.51820648e-01 5.78779459e-01 -3.73122171e-02
-1.28147113e+00 -1.16407776e+00 -2.13747010e-01 1.13962257e+00
-3.90900820e-01 -1.68320656e-01 -7.25743771e-01 -6.71142161e-01
-7.05721825e-02 1.20616782e+00 -5.11688590e-01 -3.70942295e-01
-5.22718251e-01 -7.37997413e-01 8.86216044e-01 3.04303795e-01
2.54384130e-01 -9.50794160e-01 -3.29949260e-01 2.85120666e-01
-6.74585044e-01 -1.29690230e+00 -5.30856133e-01 1.70319498e-01
-7.92931914e-01 -6.12715721e-01 -6.36447489e-01 -5.34059465e-01
2.54432648e-01 -1.60645619e-01 1.20379508e+00 -3.83091748e-01
2.36130673e-02 -2.14835390e-01 -2.09688216e-01 -1.20475739e-01
-6.96788847e-01 4.05953407e-01 7.19403774e-02 -1.33039311e-01
6.20107055e-01 -5.79093635e-01 -2.30410010e-01 -1.08248517e-01
-3.65506619e-01 2.53926873e-01 8.99275124e-01 1.20847273e+00
4.75140810e-01 -5.16099513e-01 7.44761646e-01 -9.10603523e-01
6.30200446e-01 -5.96460283e-01 -4.53380644e-01 6.46049559e-01
-7.87526131e-01 3.45834345e-01 6.99328125e-01 -5.97843468e-01
-1.12572014e+00 -2.70242959e-01 4.19420786e-02 -2.56513745e-01
1.04178168e-01 5.75119197e-01 -5.25447577e-02 5.12970209e-01
6.91124618e-01 4.05938983e-01 1.13535754e-01 -4.74974602e-01
6.04457378e-01 8.21527541e-01 3.04777801e-01 -7.41954982e-01
6.33344293e-01 1.90516654e-02 -3.81783515e-01 -4.93224442e-01
-1.00402176e+00 1.09219983e-01 -5.77518880e-01 3.28057818e-02
8.37569892e-01 -1.14880812e+00 -2.84104288e-01 3.14143836e-01
-1.39202964e+00 -5.21216512e-01 -1.77483395e-01 6.29129469e-01
-6.89924717e-01 1.74959406e-01 -1.06937754e+00 -5.95914364e-01
-5.33637285e-01 -1.40380144e+00 1.11007988e+00 -3.27880353e-01
-4.63657826e-01 -9.83154714e-01 4.35977429e-02 7.03875005e-01
4.98773605e-01 -1.66927055e-01 1.43728602e+00 -7.87783802e-01
-6.10764980e-01 -6.01252355e-02 -1.34616241e-01 2.20423251e-01
-2.33907811e-02 -2.34914139e-01 -1.05454695e+00 -2.73480982e-01
3.83335389e-02 -7.51580298e-01 7.24827170e-01 -3.19703075e-04
5.38999975e-01 -4.32859093e-01 -1.11667022e-01 5.24470568e-01
1.25052202e+00 -2.64329165e-01 4.05671656e-01 2.47290447e-01
4.88190234e-01 6.17525458e-01 5.63601553e-01 -4.47392724e-02
4.23663020e-01 1.02103519e+00 -1.68316856e-01 2.86256641e-01
-1.70805573e-01 -5.79165339e-01 7.51216352e-01 1.09819484e+00
-3.33630317e-03 -1.17434794e-02 -9.17919874e-01 5.47615051e-01
-1.83124566e+00 -8.93299401e-01 1.85859591e-01 2.44448733e+00
1.36123514e+00 2.76861995e-01 -6.49713501e-02 -1.69293582e-01
6.09054327e-01 -9.24650058e-02 -6.59509242e-01 -4.37221527e-01
-1.02368891e-01 3.56092036e-01 2.99827725e-01 7.11551785e-01
-6.31245315e-01 1.31144977e+00 5.67061043e+00 1.05731273e+00
-1.17992175e+00 5.49668610e-01 5.25749207e-01 -4.04622406e-01
-4.30204213e-01 1.95680708e-01 -7.66845107e-01 6.87467694e-01
1.10941219e+00 -3.12555224e-01 8.36219072e-01 5.00340641e-01
6.12145215e-02 5.73338382e-02 -1.39415240e+00 6.27634466e-01
1.16091326e-01 -1.09770155e+00 1.03146940e-01 1.00479238e-01
7.16212988e-01 4.31209773e-01 -8.43242463e-03 7.21666038e-01
3.26647311e-01 -8.49359751e-01 8.83382082e-01 1.53897822e-01
1.12625575e+00 -7.08734691e-01 4.78105932e-01 6.79529965e-01
-6.16474688e-01 6.73564225e-02 -3.13648880e-01 -1.20085672e-01
2.30149552e-01 6.10651374e-01 -7.18772113e-01 5.14146090e-01
1.19087435e-01 5.86644828e-01 -3.59766752e-01 5.15231907e-01
-6.07557833e-01 6.84940755e-01 -1.22373767e-01 -2.62296144e-02
2.54781902e-01 -2.98634082e-01 5.28016925e-01 1.29374754e+00
3.62504542e-01 -3.87290597e-01 -1.58920929e-01 1.17875135e+00
-3.84028137e-01 2.39735544e-01 -5.76323032e-01 -2.39034817e-01
6.51938379e-01 1.01007307e+00 -3.05173397e-01 -4.95275289e-01
-5.24826646e-01 1.36507714e+00 9.15046155e-01 3.50852281e-01
-9.39154744e-01 -2.20950201e-01 6.11445606e-01 -8.69395733e-02
7.42255449e-02 -4.78106700e-02 -3.30279559e-01 -1.59967744e+00
2.81316847e-01 -9.99902844e-01 1.03268623e-01 -4.88602012e-01
-1.51937926e+00 4.35791522e-01 -1.80729315e-01 -1.06160438e+00
-6.33176208e-01 -3.89202148e-01 -4.66207862e-01 1.14838445e+00
-1.54298687e+00 -1.25507486e+00 3.65555823e-01 2.22880080e-01
8.30170751e-01 -8.82943124e-02 1.03946352e+00 6.10154033e-01
-6.46048486e-01 9.63102818e-01 1.17454775e-01 1.59005612e-01
1.01580834e+00 -1.03261638e+00 6.93683028e-01 9.64921176e-01
1.06190600e-01 6.13259673e-01 5.51909864e-01 -5.88318706e-01
-1.33539796e+00 -9.71772432e-01 1.63890851e+00 -5.81479788e-01
1.01443839e+00 -7.99547195e-01 -8.48904490e-01 8.95360768e-01
2.11507469e-01 -2.74515450e-01 5.92953384e-01 5.06225407e-01
-6.22720838e-01 1.92908540e-01 -9.02967215e-01 8.56269002e-01
1.05101967e+00 -1.00355113e+00 -7.56626487e-01 4.20551956e-01
8.23372960e-01 -1.33565381e-01 -8.08497369e-01 1.17434129e-01
7.28201568e-01 -6.91645443e-01 6.41642153e-01 -7.34073937e-01
1.00224149e+00 2.58401185e-01 -2.32782617e-01 -1.49617147e+00
-2.04814240e-01 -5.49258292e-01 -7.92355090e-02 1.25874877e+00
8.35417151e-01 -6.14812136e-01 3.15167218e-01 3.84231627e-01
-1.20355496e-02 -8.57424855e-01 -1.17444539e+00 -9.26946282e-01
5.01538515e-01 -2.96509773e-01 3.43775690e-01 1.30534291e+00
2.91073531e-01 1.05438149e+00 -4.86925304e-01 -3.29702497e-01
6.73249960e-01 1.08988456e-01 3.65917057e-01 -8.98239851e-01
-8.41495872e-01 -5.36246419e-01 5.44762835e-02 -8.89620304e-01
5.40609598e-01 -1.44831061e+00 1.31911457e-01 -1.42749965e+00
4.75452393e-01 -3.49889070e-01 -6.80067986e-02 5.11050284e-01
-4.96209025e-01 5.93744926e-02 1.31299093e-01 4.15431678e-01
-2.19791368e-01 6.06890976e-01 1.04933977e+00 -7.77790770e-02
4.79356833e-02 -5.07414974e-02 -5.76766193e-01 3.97787988e-01
6.42792284e-01 -5.54570258e-01 -2.98319638e-01 -8.92399549e-01
2.71940053e-01 2.40049377e-01 2.14172527e-01 -4.60858315e-01
7.68181011e-02 -2.03050245e-02 -2.89327111e-02 -1.40614092e-01
4.41634625e-01 -3.83955270e-01 -1.08995840e-01 4.63037312e-01
-6.68934345e-01 -7.72905163e-03 -5.33976704e-02 3.76264721e-01
-7.56683797e-02 -1.43835306e-01 7.46972978e-01 -2.00888827e-01
-1.10480860e-01 -9.03362408e-03 -2.29315430e-01 4.04975802e-01
6.73040688e-01 5.50533570e-02 -4.54943955e-01 -1.55898750e-01
-4.35148656e-01 1.85779445e-02 5.56235373e-01 5.93950391e-01
1.23648355e-02 -1.35682797e+00 -1.04952025e+00 8.47999007e-02
3.25329959e-01 -2.83175498e-01 -6.33887574e-02 1.29855180e+00
-3.62049788e-02 4.52528656e-01 -1.25078157e-01 -3.51461649e-01
-7.10565567e-01 4.15620953e-01 3.00364524e-01 -4.64578122e-01
-5.64589858e-01 6.17143393e-01 -2.01273441e-01 -6.35975480e-01
1.03746243e-01 -4.91150916e-02 4.75171596e-01 2.69098580e-01
2.51101375e-01 4.08970863e-01 2.40566820e-01 -4.15506929e-01
-2.98684806e-01 2.75668740e-01 -2.32614607e-01 -5.05472004e-01
1.16970134e+00 -8.48117918e-02 -1.62946820e-01 8.50298464e-01
1.24469280e+00 5.42421304e-02 -1.09227419e+00 -4.98163790e-01
1.50995865e-01 -1.76589370e-01 -2.00280741e-01 -1.11407387e+00
-4.77965653e-01 1.11849248e+00 1.29392892e-01 -4.62244064e-01
7.11519718e-01 8.55202880e-03 9.50619042e-01 3.59533727e-01
4.68633622e-01 -1.18041885e+00 -2.79197663e-01 6.95578516e-01
5.81530690e-01 -1.31825638e+00 -1.95979491e-01 -6.19906969e-02
-7.51683354e-01 7.43231952e-01 2.83681720e-01 1.12494878e-01
4.13458049e-03 1.34580716e-01 -1.64643358e-02 2.00803414e-01
-1.19501328e+00 3.24380435e-02 1.87281653e-01 8.28552544e-02
7.72835374e-01 1.32170126e-01 -5.40936112e-01 5.37562370e-01
-3.11051279e-01 5.81961535e-02 2.00145736e-01 6.58198833e-01
-2.10160255e-01 -1.17106581e+00 3.12652364e-02 3.88462692e-01
-5.94212651e-01 -6.42217577e-01 -2.90633142e-01 4.01168704e-01
-7.34164640e-02 7.84829438e-01 3.15692537e-02 -2.17146620e-01
1.95285752e-01 6.19902015e-01 6.29317284e-01 -6.75610065e-01
-5.58520913e-01 -2.28665844e-02 2.43867338e-01 -4.37049031e-01
6.50731027e-02 -7.84020185e-01 -8.47103834e-01 -2.26927370e-01
-3.72854918e-01 -5.48556559e-02 7.01188087e-01 1.26499534e+00
4.81970459e-01 3.18624347e-01 2.57115036e-01 -5.81064522e-01
-1.21501637e+00 -1.28321958e+00 -6.32574707e-02 4.76569444e-01
1.20213605e-01 -4.79465157e-01 -3.92114311e-01 -8.27608407e-02] | [11.605688095092773, 10.176007270812988] |
b2ed562f-255c-4d87-ad15-32e0894d3647 | self-supervised-multi-frame-monocular-scene | 2105.02216 | null | https://arxiv.org/abs/2105.02216v1 | https://arxiv.org/pdf/2105.02216v1.pdf | Self-Supervised Multi-Frame Monocular Scene Flow | Estimating 3D scene flow from a sequence of monocular images has been gaining increased attention due to the simple, economical capture setup. Owing to the severe ill-posedness of the problem, the accuracy of current methods has been limited, especially that of efficient, real-time approaches. In this paper, we introduce a multi-frame monocular scene flow network based on self-supervised learning, improving the accuracy over previous networks while retaining real-time efficiency. Based on an advanced two-frame baseline with a split-decoder design, we propose (i) a multi-frame model using a triple frame input and convolutional LSTM connections, (ii) an occlusion-aware census loss for better accuracy, and (iii) a gradient detaching strategy to improve training stability. On the KITTI dataset, we observe state-of-the-art accuracy among monocular scene flow methods based on self-supervised learning. | ['Stefan Roth', 'Junhwa Hur'] | 2021-05-05 | null | http://openaccess.thecvf.com//content/CVPR2021/html/Hur_Self-Supervised_Multi-Frame_Monocular_Scene_Flow_CVPR_2021_paper.html | http://openaccess.thecvf.com//content/CVPR2021/papers/Hur_Self-Supervised_Multi-Frame_Monocular_Scene_Flow_CVPR_2021_paper.pdf | cvpr-2021-1 | ['scene-flow-estimation'] | ['computer-vision'] | [ 1.29021466e-01 -4.32222188e-01 -1.75219715e-01 -3.23468655e-01
-4.92575854e-01 -3.83814156e-01 6.95430815e-01 -6.26028299e-01
-5.16095221e-01 8.45757365e-01 4.08786118e-01 -2.37895548e-01
2.75240451e-01 -4.69598174e-01 -7.65622139e-01 -4.92817283e-01
4.82932664e-02 8.50956216e-02 5.62041640e-01 1.35309055e-01
1.75503641e-01 5.15498161e-01 -1.58136868e+00 3.48453224e-01
6.42672658e-01 1.03076863e+00 4.47646528e-01 9.45562899e-01
2.85234302e-03 1.52333450e+00 -2.19575301e-01 -2.60356486e-01
4.71691877e-01 -6.36565328e-01 -1.01280129e+00 3.53125259e-02
1.17132962e+00 -9.88757372e-01 -8.72319043e-01 6.55255556e-01
5.23294270e-01 1.76710606e-01 2.55813777e-01 -1.04713821e+00
-2.16709942e-01 1.06687620e-01 -4.26802367e-01 4.50545818e-01
5.60644865e-01 4.38954055e-01 8.00586820e-01 -7.35589623e-01
9.67176855e-01 1.23608363e+00 5.56605041e-01 5.54606199e-01
-1.23659611e+00 -5.42354167e-01 2.21479625e-01 4.38287228e-01
-9.21075463e-01 -7.10336804e-01 8.69864285e-01 -5.64935446e-01
1.36795998e+00 -3.09635520e-01 8.05035591e-01 1.08278322e+00
2.10078001e-01 8.81103694e-01 9.10849273e-01 -2.76434302e-01
1.97903030e-02 -1.63866594e-01 -4.18775737e-01 8.66800725e-01
5.06621040e-02 5.42800486e-01 -5.83704352e-01 3.33752781e-01
8.95260155e-01 -6.55625686e-02 -5.14213145e-01 -7.60683298e-01
-1.06867731e+00 6.53037846e-01 6.74912632e-01 2.89626539e-01
-1.64981231e-01 4.65631545e-01 4.47958589e-01 1.18997976e-01
7.76645064e-01 4.08375561e-01 -4.70887423e-01 -4.66737062e-01
-1.32261658e+00 2.78006464e-01 7.31450915e-01 7.64433324e-01
8.29687297e-01 5.38332641e-01 -4.33055572e-02 5.82098544e-01
2.37424135e-01 4.11669075e-01 3.72893184e-01 -1.20920074e+00
5.59076786e-01 3.60471874e-01 -6.91797677e-03 -9.91714418e-01
-5.52860856e-01 -6.07490361e-01 -6.24440074e-01 4.43910092e-01
6.74428821e-01 -1.06038809e-01 -7.06272900e-01 1.82418907e+00
2.50217855e-01 4.83133405e-01 -1.22243725e-01 1.24882579e+00
7.29803860e-01 6.07158065e-01 -1.55391306e-01 -1.17476741e-02
8.04228067e-01 -1.23435187e+00 -4.56374168e-01 -3.38883847e-01
7.24278629e-01 -8.12972009e-01 6.66560411e-01 2.90175885e-01
-1.18014336e+00 -7.80560374e-01 -9.41458285e-01 -4.81526434e-01
-1.33900225e-01 7.99785107e-02 6.91722393e-01 5.79689801e-01
-1.20852065e+00 6.39084697e-01 -9.70899522e-01 -3.85123998e-01
5.79845309e-01 2.10792586e-01 -5.70157945e-01 -3.31512094e-01
-8.81697297e-01 9.03911293e-01 2.60378629e-01 5.42477109e-02
-8.57063949e-01 -1.01431763e+00 -1.32643032e+00 -1.47341872e-02
2.45854348e-01 -1.04060447e+00 1.05394149e+00 -9.57288742e-01
-1.99988258e+00 7.98772514e-01 -2.90662527e-01 -6.04420960e-01
8.94709587e-01 -5.51592588e-01 7.15803653e-02 6.11551881e-01
3.96400318e-02 1.15767181e+00 8.77897680e-01 -1.01941133e+00
-7.53256917e-01 1.22591285e-02 3.10460418e-01 2.89123356e-01
1.55263208e-02 -2.30777860e-01 -1.56755358e-01 -3.88889849e-01
-1.49047181e-01 -9.64436710e-01 -6.18664846e-02 2.84071147e-01
-9.64520946e-02 2.95771450e-01 1.08011985e+00 -7.29778647e-01
1.00360715e+00 -1.90966761e+00 1.81758299e-01 -6.50587082e-01
2.59623915e-01 5.59818327e-01 2.66590472e-02 2.01953873e-01
-5.89841604e-02 -3.91247332e-01 -2.55647093e-01 -6.42990589e-01
-4.03781652e-01 4.98787463e-02 -3.01085562e-01 6.72735989e-01
4.87503082e-01 1.01001763e+00 -1.11522317e+00 -4.38192189e-01
9.44733799e-01 4.95726556e-01 -9.10887539e-01 4.58527356e-01
-2.42497437e-02 6.39744818e-01 1.27297016e-02 4.53304589e-01
6.82821810e-01 -3.24222207e-01 -2.64912963e-01 -4.17788386e-01
-5.15404344e-01 4.18096483e-01 -1.12428272e+00 2.19802022e+00
-5.92297137e-01 1.09742248e+00 3.05945072e-02 -9.06721473e-01
6.68978810e-01 2.07412302e-01 7.31562793e-01 -7.23729432e-01
6.99696913e-02 2.54539073e-01 -1.16746247e-01 -5.44351578e-01
5.99315166e-01 -2.15213105e-01 4.66722339e-01 2.68748552e-01
3.41221184e-01 -4.06136513e-01 4.23738778e-01 1.67934865e-01
1.11866593e+00 9.26175654e-01 1.95418239e-01 -2.33767390e-01
7.24965394e-01 -5.77836558e-02 4.74821836e-01 6.71170950e-01
-5.28608680e-01 1.15991008e+00 1.74112260e-01 -7.75623798e-01
-1.12485492e+00 -7.26830423e-01 -4.24757935e-02 4.89286274e-01
2.67455488e-01 -9.69048962e-02 -4.02128637e-01 -5.56038916e-01
-7.27541223e-02 3.45251232e-01 -3.28498214e-01 -1.04706384e-01
-8.78073275e-01 -4.47277069e-01 4.71746534e-01 4.70951527e-01
1.02132356e+00 -9.15445328e-01 -1.31267476e+00 4.47623968e-01
-4.03171569e-01 -1.75677347e+00 -4.92823362e-01 7.24083558e-02
-9.18473303e-01 -1.03405750e+00 -8.62381876e-01 -5.76351523e-01
1.41278416e-01 5.86709976e-01 1.09484553e+00 -3.09949160e-01
-2.64362514e-01 2.92328417e-01 -1.39667630e-01 9.07877833e-02
-1.25475541e-01 3.27961951e-01 -1.45665705e-01 9.15110707e-02
-5.14226109e-02 -5.59868574e-01 -7.71861732e-01 1.08697869e-01
-8.89457941e-01 4.43902701e-01 3.65223944e-01 8.98332059e-01
-1.09402798e-01 -5.31648099e-01 4.01635319e-01 -5.92859209e-01
-1.77726686e-01 -1.86254382e-01 -7.66699076e-01 -3.45274031e-01
-3.66571903e-01 3.79751697e-02 7.63953447e-01 -2.13649705e-01
-1.55709708e+00 3.88827622e-01 -1.57476410e-01 -7.29145169e-01
-1.78246096e-01 7.68506154e-02 1.68461680e-01 -3.79140615e-01
5.68382263e-01 1.53717786e-01 -7.69190043e-02 -1.59250155e-01
4.46128935e-01 3.73952389e-01 6.77894413e-01 -3.44720155e-01
6.34786069e-01 8.87946486e-01 9.25920010e-02 -9.91388440e-01
-1.01833463e+00 -6.17374837e-01 -9.41727340e-01 -5.04237592e-01
1.05571032e+00 -1.37962770e+00 -5.57875574e-01 8.24239731e-01
-1.27948797e+00 -6.63937867e-01 -1.48162529e-01 9.98395920e-01
-7.81499445e-01 4.64445859e-01 -9.46816564e-01 -6.69912755e-01
-1.00847758e-01 -1.22834575e+00 1.19807434e+00 5.75160533e-02
-4.75193933e-02 -1.10145605e+00 2.48251721e-01 3.72138470e-01
5.51238358e-01 3.36025864e-01 3.41371298e-01 2.09291399e-01
-1.00794470e+00 1.80673912e-01 -6.58704877e-01 3.47900748e-01
3.83504368e-02 -8.71022940e-02 -1.40211475e+00 -2.91783124e-01
-5.18877842e-02 -5.93169451e-01 9.93777454e-01 6.79053843e-01
6.76999092e-01 7.98522308e-02 -1.85445577e-01 1.18659282e+00
1.44507933e+00 1.88755944e-01 6.60055280e-01 4.24903870e-01
9.50932682e-01 4.76939827e-01 1.68993205e-01 2.53357291e-01
5.03635228e-01 7.65994072e-01 3.55000615e-01 -2.46139735e-01
-6.98704123e-01 -3.49343687e-01 3.53107899e-01 6.14600778e-01
1.06682755e-01 -2.43345693e-01 -6.42956078e-01 6.93802357e-01
-1.90882969e+00 -1.25684452e+00 -1.12954594e-01 1.94099152e+00
5.59699118e-01 1.65160164e-01 1.06955085e-04 2.86638848e-02
2.77877271e-01 6.00370049e-01 -5.39032578e-01 -2.19461694e-01
-4.23412085e-01 -8.55169892e-02 5.69470227e-01 7.60772645e-01
-1.37500727e+00 1.18173134e+00 6.09314108e+00 3.96636784e-01
-1.59331095e+00 -2.35276835e-04 6.57025874e-01 -2.06901610e-01
1.91571057e-01 1.03142850e-01 -8.14388454e-01 3.51732403e-01
5.44910133e-01 2.43239492e-01 3.20465326e-01 5.75717986e-01
2.51001924e-01 -3.58032674e-01 -1.04070425e+00 1.20859039e+00
3.80915701e-01 -1.40886974e+00 -1.90998361e-01 -1.71444743e-04
8.77221167e-01 3.44256818e-01 -3.48643720e-01 -1.93033200e-02
6.55492246e-02 -6.75503612e-01 8.65721524e-01 3.86771053e-01
8.35464001e-01 -2.54494220e-01 5.54829776e-01 1.38576910e-01
-1.26319718e+00 -1.08427808e-01 -1.57259434e-01 -6.35291040e-01
6.66395366e-01 6.52825117e-01 -3.39903295e-01 7.66482472e-01
8.28760445e-01 1.46386456e+00 -3.66826385e-01 1.23335886e+00
-1.26375780e-01 4.74135399e-01 -4.59606051e-01 2.64879078e-01
5.12647271e-01 -1.16753586e-01 6.17024362e-01 1.31125128e+00
1.04150422e-01 -8.01277235e-02 1.46857828e-01 9.78103638e-01
6.53537810e-02 -3.33807439e-01 -8.94517779e-01 5.12806058e-01
-4.90399748e-02 1.13544798e+00 -5.53841114e-01 -2.57919699e-01
-7.90983200e-01 9.37811553e-01 4.75730151e-01 3.98151189e-01
-5.91550112e-01 -1.19586967e-01 6.81004703e-01 8.26457366e-02
4.81320947e-01 -4.38125789e-01 -3.44697475e-01 -1.57011402e+00
6.63013160e-02 -3.62597317e-01 2.04507932e-01 -9.90057349e-01
-9.16731417e-01 7.04058468e-01 7.61981234e-02 -1.41203928e+00
-5.03522635e-01 -5.91343760e-01 -3.16779763e-01 6.37700737e-01
-2.03573084e+00 -1.05799472e+00 -4.63340968e-01 4.30410862e-01
6.54227257e-01 -1.32292405e-01 2.62321502e-01 7.08802342e-01
-3.88553858e-01 2.38893926e-01 -2.39307359e-01 5.93053922e-02
8.13978612e-01 -1.08853495e+00 3.67239684e-01 1.29652178e+00
4.52200994e-02 3.28500867e-02 5.74458063e-01 -3.68716657e-01
-1.30923760e+00 -1.10450077e+00 9.73883212e-01 -4.05350149e-01
4.72220868e-01 -3.70944321e-01 -7.90958047e-01 7.97021508e-01
2.50093818e-01 2.46979684e-01 -6.38086647e-02 -4.55896139e-01
-1.23584419e-01 -2.25347504e-01 -9.26115453e-01 5.36767721e-01
1.32383633e+00 -5.27531445e-01 -3.88144463e-01 -6.97970949e-03
5.81209838e-01 -6.87739134e-01 -4.06655759e-01 4.06850517e-01
7.19106555e-01 -1.35705483e+00 1.03468120e+00 -1.66133791e-01
8.49543989e-01 -2.64794588e-01 -3.66279185e-02 -1.22490573e+00
-3.43978375e-01 -6.45981491e-01 -1.85658485e-01 8.20750833e-01
-1.02788866e-01 -5.41526437e-01 9.41731930e-01 4.95078593e-01
-3.08339655e-01 -5.09161413e-01 -1.12561810e+00 -7.34005809e-01
-8.96679685e-02 -3.90224665e-01 2.11871061e-02 7.29129672e-01
-1.93449825e-01 5.88751853e-01 -8.53850186e-01 -2.23143667e-01
5.93123496e-01 1.91225469e-01 1.00932240e+00 -9.35216188e-01
-3.20796221e-01 -4.22341466e-01 -5.84977031e-01 -1.76416194e+00
3.72745752e-01 -6.50040805e-01 3.55987549e-02 -1.30780923e+00
-1.81545764e-02 -9.74254832e-02 2.84379452e-01 2.31237158e-01
-1.24158427e-01 4.39567477e-01 4.31429535e-01 2.05927894e-01
-6.24957621e-01 9.14697707e-01 1.20023799e+00 1.65029347e-01
-2.21225083e-01 -3.63899797e-01 -3.92823480e-02 6.02660894e-01
4.05503303e-01 -3.25505376e-01 -3.98194104e-01 -8.40438604e-01
-1.41051263e-01 2.99210221e-01 5.01103699e-01 -1.34437656e+00
3.12412262e-01 3.08505505e-01 4.74036634e-01 -5.07593632e-01
5.90206861e-01 -7.08927512e-01 -1.75396875e-01 7.52689302e-01
-1.94160119e-01 1.19198970e-01 2.73803174e-01 3.85078341e-01
-3.80477041e-01 1.03013374e-01 9.29249585e-01 -1.14013858e-01
-8.73598754e-01 3.84604573e-01 -2.68902034e-01 2.60941744e-01
6.30768776e-01 -3.45530242e-01 -4.16739970e-01 -5.60442924e-01
-9.28664282e-02 2.97171269e-02 4.45970058e-01 4.99699414e-01
5.95164239e-01 -1.13625062e+00 -7.37679899e-01 4.09380436e-01
1.49232661e-03 -2.32849070e-05 2.58500636e-01 7.75672019e-01
-9.90394950e-01 7.53909051e-01 -4.50215161e-01 -9.36877012e-01
-8.96518648e-01 3.40204090e-01 6.59167111e-01 -3.63219053e-01
-8.23960006e-01 6.19730294e-01 5.34551680e-01 -4.66237873e-01
8.65040068e-03 -3.28774244e-01 5.05730277e-03 -3.01548839e-01
3.64930928e-01 5.81404924e-01 2.05978267e-02 -7.44792998e-01
-3.20324481e-01 8.27113271e-01 2.30533540e-01 -2.27270931e-01
1.11618507e+00 -3.84270549e-01 2.96795309e-01 4.22776550e-01
1.50202072e+00 -4.46418971e-01 -2.06834006e+00 -1.88780665e-01
-3.64649773e-01 -8.42411637e-01 3.17314327e-01 -5.46214163e-01
-1.30412078e+00 1.09990251e+00 5.10184526e-01 -2.78318048e-01
1.12255752e+00 -4.34401035e-01 9.31646049e-01 9.11924541e-02
3.41076881e-01 -6.17266417e-01 4.25904579e-02 7.95950174e-01
4.72790301e-01 -1.51037133e+00 2.70251613e-02 -3.15857142e-01
-2.79368013e-01 1.33283722e+00 6.96922958e-01 -2.19633371e-01
6.74135268e-01 1.71326458e-01 2.22241849e-01 1.26057670e-01
-8.37366998e-01 -2.74871826e-01 2.04425529e-01 5.30139387e-01
3.80181104e-01 -4.73496318e-01 2.14537289e-02 -3.19793522e-01
-1.71476919e-02 4.24570054e-01 6.12362683e-01 7.40354419e-01
-7.88475424e-02 -7.21345663e-01 2.85874307e-02 1.59256592e-01
-3.98467153e-01 -6.39045462e-02 3.94371077e-02 8.14165592e-01
-1.28487140e-01 1.01635301e+00 1.07926153e-01 -2.14545667e-01
4.15415436e-01 -2.17253208e-01 6.83393896e-01 -2.38691941e-01
-5.63147485e-01 -3.92071046e-02 1.52931390e-02 -9.29419398e-01
-8.03701580e-01 -7.70339489e-01 -7.32562840e-01 -3.00930113e-01
-1.32406443e-01 -3.47058356e-01 3.96535635e-01 1.20773387e+00
5.03650367e-01 3.96299124e-01 5.58808208e-01 -1.34836030e+00
-7.19747245e-02 -8.52413237e-01 -1.44586280e-01 5.58759391e-01
6.32347584e-01 -6.51203036e-01 -3.41926664e-01 2.54999310e-01] | [8.684260368347168, -1.8766149282455444] |
c74ab011-3a71-4831-9c63-feb04274dadf | implementing-a-reverse-dictionary-based-on | 1606.00025 | null | http://arxiv.org/abs/1606.00025v5 | http://arxiv.org/pdf/1606.00025v5.pdf | Implementing a Reverse Dictionary, based on word definitions, using a Node-Graph Architecture | In this paper, we outline an approach to build graph-based reverse
dictionaries using word definitions. A reverse dictionary takes a phrase as an
input and outputs a list of words semantically similar to that phrase. It is a
solution to the Tip-of-the-Tongue problem. We use a distance-based similarity
measure, computed on a graph, to assess the similarity between a word and the
input phrase. We compare the performance of our approach with the Onelook
Reverse Dictionary and a distributional semantics method based on word2vec, and
show that our approach is much better than the distributional semantics method,
and as good as Onelook, on a 3k lexicon. This simple approach sets a new
performance baseline for reverse dictionaries. | ['Sushrut Thorat', 'Varad Choudhari'] | 2016-05-31 | implementing-a-reverse-dictionary-based-on-1 | https://aclanthology.org/C16-1263 | https://aclanthology.org/C16-1263.pdf | coling-2016-12 | ['reverse-dictionary'] | ['natural-language-processing'] | [-1.94531828e-01 2.44575609e-02 -6.10291779e-01 -4.26622272e-01
-5.14829040e-01 -1.17190754e+00 7.70257056e-01 5.33254623e-01
-6.34837508e-01 2.34792754e-01 6.88363671e-01 -6.11932158e-01
6.45767301e-02 -9.41620588e-01 -1.54713526e-01 -4.02746856e-01
2.77197957e-01 8.56733680e-01 2.20290646e-01 -9.14353430e-01
3.77778739e-01 7.26226717e-02 -9.65568125e-01 1.57351568e-02
4.22246456e-01 5.56829393e-01 2.38363832e-01 2.62657404e-01
-6.21816933e-01 3.74372721e-01 -4.59360182e-01 -7.59403288e-01
1.00302309e-01 -5.25464892e-01 -7.66872644e-01 -3.62009108e-01
4.96630281e-01 4.43820387e-01 -3.30898106e-01 1.32439053e+00
4.80809689e-01 2.77485102e-01 9.07884121e-01 -8.67380738e-01
-1.23617899e+00 1.01125824e+00 -2.13141263e-01 5.25023580e-01
7.31203675e-01 -2.04068482e-01 1.80319238e+00 -1.15926099e+00
8.51684511e-01 1.15933251e+00 9.50195909e-01 3.14044327e-01
-1.15819967e+00 -6.07962012e-01 9.14485082e-02 1.73583534e-02
-1.54303622e+00 -2.07630575e-01 5.51970482e-01 -3.11626017e-01
1.41472328e+00 1.84213948e-02 7.66534984e-01 8.96216393e-01
1.83235377e-01 3.30807894e-01 1.05374825e+00 -7.06405938e-01
1.01871014e-01 2.94360183e-02 6.40799761e-01 8.18027139e-01
2.33204916e-01 3.97868231e-02 -3.01261693e-01 -3.38356674e-01
4.75003690e-01 -2.75663435e-01 -2.11837798e-01 -1.42616734e-01
-1.06813419e+00 1.27267969e+00 3.67065728e-01 7.62042880e-01
-4.01974171e-02 1.68634057e-01 5.27025819e-01 6.65873110e-01
5.53790510e-01 7.34750271e-01 -5.99639535e-01 -1.42057827e-02
-8.26086640e-01 4.71018143e-02 1.05653346e+00 1.03471911e+00
6.40291393e-01 3.16378735e-02 8.27057734e-02 8.77012610e-01
4.98006165e-01 7.60195255e-01 1.12673044e+00 -2.18135789e-01
2.32191831e-01 4.71439779e-01 -6.22256756e-01 -1.21434236e+00
-4.27994132e-01 -4.46781576e-01 -3.17248225e-01 -2.49305010e-01
-1.12043202e-01 1.31039351e-01 -1.21048987e+00 1.66369390e+00
1.20639503e-01 2.37073228e-01 9.48918518e-03 6.80622101e-01
1.19617689e+00 4.79282439e-01 7.91169032e-02 -1.53869614e-01
1.44250989e+00 -9.24075544e-01 -8.35114121e-01 -2.39321455e-01
9.62440908e-01 -8.04201603e-01 1.33795702e+00 4.12525088e-02
-7.05857038e-01 -4.53268707e-01 -9.89212036e-01 -1.51562616e-01
-9.05025363e-01 -6.30257189e-01 4.76466566e-01 8.81945014e-01
-1.39512539e+00 6.14331543e-01 -3.70630264e-01 -6.76745653e-01
-2.12790072e-01 2.21817583e-01 -5.70537448e-01 -2.88120694e-02
-1.47992766e+00 1.25181615e+00 4.77446258e-01 -8.51806045e-01
-4.47339028e-01 -6.22574270e-01 -1.36556280e+00 -1.83450058e-01
1.76350866e-02 -7.54803300e-01 1.02999938e+00 -6.53430223e-01
-1.13157773e+00 1.24113584e+00 -1.59112647e-01 -3.50243539e-01
-4.65378523e-01 2.27082893e-01 -7.08302677e-01 -3.50429326e-01
5.22101283e-01 1.46502167e-01 7.28694737e-01 -1.18159652e+00
-5.00038385e-01 -4.07070905e-01 1.53743714e-01 2.33831227e-01
-3.23873073e-01 1.32484764e-01 -7.37171412e-01 -1.16816211e+00
6.12823367e-02 -9.82133210e-01 -3.07160586e-01 -6.25256002e-01
-2.82307535e-01 -4.83964443e-01 2.76671618e-01 -5.16925514e-01
1.65222704e+00 -2.17708421e+00 3.31124187e-01 6.54083133e-01
2.64003664e-01 1.20120876e-01 -4.31208313e-01 6.31012142e-01
-4.06142920e-01 4.06762600e-01 -2.99753278e-01 -2.94000328e-01
1.48930311e-01 6.07644856e-01 -2.08656177e-01 5.08468509e-01
-3.98237735e-01 1.16177821e+00 -1.32546639e+00 -5.01667202e-01
-1.18501633e-02 4.12219942e-01 -5.38579881e-01 -6.34743199e-02
-2.55750835e-01 -3.08540821e-01 -1.72957599e-01 4.25473332e-01
1.98731989e-01 7.53886160e-03 4.96417791e-01 -3.08425814e-01
1.58297852e-01 6.87309146e-01 -7.74442017e-01 1.95030355e+00
-7.76894152e-01 5.15516460e-01 -4.97757375e-01 -7.69863188e-01
9.90911663e-01 3.62341374e-01 2.20604822e-01 -7.69627571e-01
3.10036361e-01 3.42232943e-01 -1.76764816e-01 -3.17267925e-01
6.10827148e-01 -7.64763236e-01 -4.77307200e-01 7.50721633e-01
5.70578516e-01 -6.51824534e-01 2.76952446e-01 3.49720597e-01
1.19323564e+00 -7.92521685e-02 9.16010976e-01 -6.01354241e-01
2.92146295e-01 2.17043027e-01 1.95004895e-01 3.71271640e-01
2.22201377e-01 5.69210351e-01 1.19585760e-01 -3.79414856e-01
-6.24270201e-01 -1.23773253e+00 9.07772332e-02 1.44777119e+00
1.60348281e-01 -1.18065977e+00 -5.72505534e-01 -1.10057414e+00
2.11168945e-01 9.97706234e-01 -7.60403275e-01 -1.77849174e-01
-5.56810141e-01 -3.43248427e-01 7.37800062e-01 4.42298889e-01
-1.63659245e-01 -9.46701467e-01 1.85249239e-01 9.84452963e-02
-7.87392631e-02 -8.02947819e-01 -1.13006127e+00 3.47527027e-01
-5.39976120e-01 -9.50433969e-01 -4.88234609e-01 -1.45318949e+00
5.08379638e-01 2.53821194e-01 1.86518753e+00 3.11966300e-01
1.44615233e-01 3.41706753e-01 -7.18618214e-01 -1.75979286e-01
-4.62715656e-01 2.17533007e-01 4.07320231e-01 -4.32016015e-01
8.04103136e-01 -6.03102803e-01 -1.56539738e-01 -1.97864808e-02
-1.00901711e+00 -6.89212561e-01 1.62335277e-01 6.65553629e-01
7.42907703e-01 -4.03778739e-02 3.27898830e-01 -1.22849238e+00
1.09714770e+00 -7.68965542e-01 -1.76356673e-01 2.23445535e-01
-9.01171923e-01 4.17562008e-01 7.37798870e-01 -3.39007229e-01
-2.62117565e-01 -1.59324370e-02 -6.09351158e-01 -2.65765488e-01
2.15014532e-01 9.86547709e-01 4.85883020e-02 -1.71217665e-01
8.02794099e-01 1.91994056e-01 -3.37579906e-01 -6.11182511e-01
1.20325577e+00 6.89462841e-01 4.76240933e-01 -3.46812636e-01
9.40456271e-01 2.38319144e-01 -3.02948207e-01 -7.71353543e-01
-1.06610024e+00 -9.58397329e-01 -8.69419217e-01 2.47280583e-01
8.22640777e-01 -8.31158280e-01 -3.49612236e-02 -3.16175252e-01
-1.04012573e+00 -1.20674945e-01 -5.99963605e-01 2.65257597e-01
-3.21677059e-01 3.96694690e-01 -2.42841974e-01 -4.31110151e-02
-4.64247674e-01 -6.46727085e-01 8.52196872e-01 -2.11736679e-01
-6.51076436e-01 -1.95138943e+00 7.17474341e-01 -1.33944824e-01
5.05456090e-01 -3.52854803e-02 1.19268751e+00 -1.08280730e+00
4.78036612e-01 1.49343694e-02 9.85716358e-02 2.50307947e-01
6.09850407e-01 -3.80465508e-01 -4.60039496e-01 -3.79312217e-01
-3.02101411e-02 5.13304174e-02 8.90374064e-01 3.59508619e-02
5.44948399e-01 -2.08274886e-01 -3.85081172e-01 7.82499790e-01
1.78022647e+00 -5.82957454e-02 3.25836271e-01 3.21125686e-01
8.89781535e-01 1.55722633e-01 4.73726362e-01 1.38037801e-01
8.77226353e-01 7.08316565e-01 1.29219025e-01 -1.28334999e-01
-6.12065673e-01 -6.17371559e-01 4.42553043e-01 1.79244745e+00
2.59842455e-01 -4.20996606e-01 -1.12663555e+00 1.07028091e+00
-1.38938677e+00 -7.06390679e-01 -2.87443608e-01 1.89012492e+00
1.02696598e+00 5.65683246e-02 1.22749932e-01 -7.33882040e-02
5.34971476e-01 3.00913721e-01 -4.27273996e-02 -8.41069400e-01
-3.07154745e-01 1.04369771e+00 6.83302820e-01 9.23955262e-01
-8.08589518e-01 1.58938074e+00 7.67493010e+00 9.32206452e-01
-1.09852970e+00 3.43308121e-01 -3.02958310e-01 4.04469132e-01
-9.46988702e-01 1.49796670e-02 -5.96152306e-01 3.69546324e-01
9.78745520e-01 -7.08217382e-01 3.85447890e-01 7.22305834e-01
-3.07181925e-01 5.53272367e-01 -1.05165696e+00 1.13060892e+00
7.76315689e-01 -1.20607448e+00 5.10370135e-01 -1.73422486e-01
6.15923524e-01 5.20168602e-01 7.48123229e-02 3.47747684e-01
7.43093252e-01 -1.10501981e+00 6.35550857e-01 4.90745567e-02
1.03770983e+00 -6.34138286e-01 6.92338765e-01 -7.42575675e-02
-1.63560593e+00 4.24257070e-01 -3.40946257e-01 -7.67545477e-02
2.30478108e-01 4.83856916e-01 -7.31638551e-01 5.15913725e-01
3.69701684e-01 1.02459753e+00 -6.50291026e-01 6.59627259e-01
-7.09012628e-01 6.60913110e-01 -8.23377147e-02 -2.41443381e-01
3.74515027e-01 -2.09154159e-01 4.80624110e-01 1.66070724e+00
2.03822151e-01 1.00616843e-01 3.05666476e-01 3.59290928e-01
-3.17771912e-01 6.23581469e-01 -1.09039032e+00 -1.50624260e-01
4.06484097e-01 1.10908973e+00 -6.60161793e-01 -6.21197164e-01
-5.46455920e-01 1.15310407e+00 4.68747944e-01 -6.04363857e-03
-6.80465877e-01 -5.71997643e-01 9.24197614e-01 8.60561952e-02
4.45488960e-01 -5.33187568e-01 -2.41074428e-01 -1.15919793e+00
-3.43021959e-01 -9.39719141e-01 6.21223390e-01 -6.08665347e-01
-1.69125152e+00 8.76010358e-01 -2.64824063e-01 -9.72890377e-01
-1.63208112e-01 -8.07140172e-01 -4.48732316e-01 8.12466919e-01
-1.48453832e+00 -9.83311832e-01 1.55705556e-01 8.36396694e-01
5.91941833e-01 -2.80129969e-01 1.38584697e+00 3.10223877e-01
1.97468087e-01 7.09297597e-01 1.55358342e-02 3.24770451e-01
6.39947176e-01 -1.69783664e+00 1.09683096e+00 4.99100447e-01
1.01102841e+00 9.54520881e-01 6.68131948e-01 -7.30605245e-01
-1.44530106e+00 -1.16823947e+00 1.46628737e+00 -8.10458422e-01
1.22624898e+00 -1.99017748e-01 -6.82913303e-01 9.31922078e-01
5.84076583e-01 -6.30765483e-02 1.25286341e+00 4.03960705e-01
-8.29020977e-01 4.78269666e-01 -9.55336511e-01 5.98754048e-01
1.37148881e+00 -8.99060428e-01 -1.55800974e+00 5.05288780e-01
1.08041644e+00 -1.12454906e-01 -1.00956500e+00 -3.75903361e-02
3.82434905e-01 -3.63398939e-01 8.94281745e-01 -7.27510870e-01
3.41194309e-02 -2.79299259e-01 -4.35378104e-01 -2.11315584e+00
-6.93809509e-01 -5.00915110e-01 1.97375268e-01 9.51088786e-01
8.45155716e-01 -6.95695460e-01 4.77505773e-01 -1.03293173e-01
-1.10578254e-01 -4.33121860e-01 -7.73873806e-01 -1.18707025e+00
5.30070603e-01 -6.63812280e-01 7.10704684e-01 1.46726680e+00
4.52603668e-01 9.90703046e-01 1.11251883e-01 -3.36096846e-02
3.88755113e-01 1.79909661e-01 3.40361297e-01 -1.05393791e+00
-2.05230325e-01 -4.14239824e-01 -7.64352083e-01 -1.13586342e+00
6.33078694e-01 -1.73997521e+00 -1.33409068e-01 -1.75012124e+00
-2.26296708e-02 -5.57966530e-01 -5.87507725e-01 6.01924181e-01
-1.01461336e-01 7.09890187e-01 1.58818156e-01 -3.06212511e-02
-5.17175078e-01 2.70648032e-01 9.30643857e-01 -3.87627512e-01
-3.18821073e-02 -5.99345148e-01 -1.00458324e+00 6.86107218e-01
7.69893050e-01 -8.60037446e-01 -4.19463754e-01 -8.31200063e-01
5.14462590e-01 -3.14249754e-01 -2.78146058e-01 -5.46578884e-01
2.83290446e-01 -3.93201932e-02 -3.91538948e-01 -1.89097181e-01
9.42041278e-02 -7.57984757e-01 -4.84465323e-02 4.64695066e-01
-1.42432436e-01 7.97555506e-01 2.73438524e-02 3.58530879e-01
-4.60445106e-01 -2.87170261e-01 5.07500648e-01 -3.43371153e-01
-9.98135865e-01 2.98481464e-01 -1.53748482e-01 5.37177324e-01
6.42722368e-01 -1.45686623e-02 3.43905352e-02 -3.54574710e-01
-8.31654608e-01 -3.58935818e-02 6.49295092e-01 8.33302677e-01
6.88967884e-01 -1.66692185e+00 -6.99255228e-01 3.23198795e-01
5.12070239e-01 -8.34481537e-01 -7.97457099e-01 4.02939737e-01
-4.00606990e-01 3.77054214e-01 1.39608115e-01 -1.28616512e-01
-1.44358027e+00 7.84589112e-01 2.54153699e-01 -1.37121573e-01
-5.68321407e-01 1.00654697e+00 -3.01706530e-02 -8.18085015e-01
-1.51988626e-01 -5.18933773e-01 -4.99308407e-01 3.19043458e-01
4.59376454e-01 -1.26724005e-01 1.84499666e-01 -1.27345598e+00
-7.30001092e-01 1.13104367e+00 2.09448591e-01 -3.56072724e-01
1.17625451e+00 -4.03518379e-02 -4.48697060e-01 6.14556193e-01
1.59049606e+00 4.32192326e-01 2.83967167e-01 -6.06739223e-01
3.21270227e-01 -4.61748898e-01 2.15892553e-01 -6.67268038e-01
-1.13240981e+00 4.03854668e-01 2.81801701e-01 4.47786748e-01
8.04751515e-01 2.61154473e-01 9.70171869e-01 2.04314381e-01
5.68027973e-01 -9.39852536e-01 -9.76032615e-02 1.10597694e+00
7.20622063e-01 -1.01188540e+00 -7.87367597e-02 -3.42916906e-01
-6.47951305e-01 9.81094539e-01 -1.58627763e-01 -4.04468387e-01
1.20300746e+00 2.98608452e-01 4.59459037e-01 -7.54878581e-01
-5.53110778e-01 -8.37595344e-01 5.57279229e-01 7.67005742e-01
7.87729084e-01 2.74235904e-01 -6.10396624e-01 5.31030595e-01
-8.05996239e-01 -3.37602347e-01 2.03286141e-01 7.99895883e-01
-4.43410397e-01 -1.47748101e+00 9.73765105e-02 3.95856351e-01
-5.54787338e-01 -8.44287872e-01 -7.02564120e-01 8.01893830e-01
2.00389057e-01 1.06762350e+00 6.22359011e-03 -7.86207199e-01
6.57332718e-01 3.32719907e-02 3.95631492e-01 -1.41988254e+00
-9.06535566e-01 -3.83293629e-01 2.33831763e-01 -5.52979827e-01
-3.91134441e-01 -3.49053234e-01 -1.51428902e+00 -2.99302876e-01
-5.21738827e-01 4.72220659e-01 8.31419945e-01 9.91459846e-01
1.14638224e-01 1.81875288e-01 5.94185889e-01 -2.90829420e-01
-3.01447928e-01 -7.79760182e-01 -7.97740698e-01 7.87610233e-01
1.63169712e-01 -5.28257310e-01 -5.08176982e-01 1.34446202e-02] | [10.478796005249023, 8.977665901184082] |
d8ab98e0-e4ae-4e5e-93cf-6a9de1c2773b | separate-what-you-describe-language-queried | 2203.15147 | null | https://arxiv.org/abs/2203.15147v1 | https://arxiv.org/pdf/2203.15147v1.pdf | Separate What You Describe: Language-Queried Audio Source Separation | In this paper, we introduce the task of language-queried audio source separation (LASS), which aims to separate a target source from an audio mixture based on a natural language query of the target source (e.g., "a man tells a joke followed by people laughing"). A unique challenge in LASS is associated with the complexity of natural language description and its relation with the audio sources. To address this issue, we proposed LASS-Net, an end-to-end neural network that is learned to jointly process acoustic and linguistic information, and separate the target source that is consistent with the language query from an audio mixture. We evaluate the performance of our proposed system with a dataset created from the AudioCaps dataset. Experimental results show that LASS-Net achieves considerable improvements over baseline methods. Furthermore, we observe that LASS-Net achieves promising generalization results when using diverse human-annotated descriptions as queries, indicating its potential use in real-world scenarios. The separated audio samples and source code are available at https://liuxubo717.github.io/LASS-demopage. | ['Wenwu Wang', 'Mark D. Plumbley', 'Qiushi Huang', 'Jinzheng Zhao', 'Xinhao Mei', 'Qiuqiang Kong', 'Haohe Liu', 'Xubo Liu'] | 2022-03-28 | null | null | null | null | ['audio-source-separation'] | ['audio'] | [-5.85293546e-02 -2.65960366e-01 -8.73121619e-02 -4.36707199e-01
-1.79738736e+00 -5.70435524e-01 2.72234946e-01 -5.31418622e-03
-2.42451519e-01 4.04592276e-01 5.72874010e-01 3.11779350e-01
9.70484614e-02 -1.80157721e-01 -6.81030273e-01 -5.48187792e-01
-1.63942158e-01 3.84776592e-01 -1.04573868e-01 1.10620363e-02
-2.02840969e-01 2.12056302e-02 -1.50532484e+00 5.12832999e-01
4.14835691e-01 1.31629562e+00 2.20729545e-01 7.05419481e-01
5.10054156e-02 9.92534041e-01 -7.26940989e-01 -1.01324894e-01
-8.11899081e-03 -5.32378554e-01 -8.29936028e-01 -4.37204726e-02
5.34893394e-01 -1.61754891e-01 -3.35359454e-01 1.10120785e+00
8.72581303e-01 2.16145903e-01 4.66664106e-01 -1.49161446e+00
-4.36427236e-01 1.08325887e+00 -2.27234334e-01 1.89482108e-01
5.89148819e-01 2.34090015e-02 1.37453926e+00 -1.19854176e+00
1.03792757e-01 1.47327983e+00 5.57041168e-01 4.56099957e-01
-1.29763353e+00 -1.16355467e+00 1.47785708e-01 2.35920280e-01
-1.86364377e+00 -1.16761470e+00 1.03680384e+00 -4.35824901e-01
5.23290753e-01 2.15281337e-01 1.94452241e-01 1.46675456e+00
-3.80820274e-01 1.17999792e+00 6.07298076e-01 -1.47018373e-01
2.57075608e-01 2.29413316e-01 1.53351694e-01 2.26553291e-01
-6.02561414e-01 -1.49887353e-01 -9.17512357e-01 -3.87323081e-01
2.61708468e-01 -3.89446825e-01 -5.22576511e-01 1.18511826e-01
-1.24754012e+00 6.18141353e-01 2.43818626e-01 1.28425792e-01
-3.65470022e-01 3.40930820e-01 6.24080956e-01 1.17529020e-01
4.68503505e-01 2.44674236e-01 -1.50795504e-01 -2.81814069e-01
-9.27047014e-01 3.54914010e-01 7.90936768e-01 1.08231616e+00
4.87443417e-01 3.40495855e-01 -1.17052965e-01 1.34892893e+00
3.32585245e-01 7.22489715e-01 4.20292348e-01 -1.14797139e+00
4.53730345e-01 -1.47075588e-02 4.67958860e-02 -7.59289622e-01
-1.75796866e-01 -5.37514806e-01 -8.36923242e-01 -4.28099543e-01
5.10404967e-02 -1.66030824e-01 -5.41362762e-01 2.02770424e+00
1.48680970e-01 4.95174885e-01 3.14129323e-01 1.09411371e+00
1.30854023e+00 9.09893870e-01 6.75684586e-02 -3.87745798e-02
1.33245242e+00 -1.01644683e+00 -7.62355447e-01 -2.88860112e-01
-8.63676239e-03 -7.87700593e-01 1.28996897e+00 5.46328008e-01
-1.02907491e+00 -7.79811144e-01 -6.36014521e-01 -2.98136435e-02
9.53433365e-02 4.01549965e-01 3.37859839e-02 2.21455526e-02
-8.99722159e-01 1.12381347e-01 -6.84098661e-01 -2.82881975e-01
9.76884365e-02 6.69513410e-03 -2.25829378e-01 1.80321828e-01
-1.43734550e+00 -8.54494199e-02 4.22394782e-01 2.05136076e-01
-1.41066623e+00 -6.05348408e-01 -9.81935441e-01 2.49852777e-01
6.36306405e-01 -3.18529308e-01 1.76310873e+00 -9.56277013e-01
-1.41260791e+00 6.05690062e-01 -4.66353565e-01 -4.71542835e-01
2.36445636e-01 -3.58193934e-01 -8.08783352e-01 4.27936018e-01
4.31870103e-01 6.71250343e-01 8.77049148e-01 -1.37647688e+00
-7.92192459e-01 -1.20453477e-01 -2.51319766e-01 1.39885604e-01
-4.06831115e-01 2.93036491e-01 -7.25093126e-01 -8.13030422e-01
-2.88657602e-02 -9.08252776e-01 2.81493813e-01 -2.04183161e-01
-8.05911899e-01 -4.78219539e-01 5.65414250e-01 -5.53778052e-01
1.23491502e+00 -2.76743174e+00 -5.14379553e-02 -4.53769267e-02
9.74075496e-02 -8.35680664e-02 -3.92926961e-01 6.48953855e-01
-8.28777030e-02 -3.87037136e-02 -1.94717899e-01 -5.80623627e-01
3.48472893e-01 -2.26826280e-01 -8.75910461e-01 3.11336577e-01
1.11446217e-01 4.22955930e-01 -1.00087988e+00 -4.34805274e-01
-4.36314017e-01 6.70164406e-01 -3.31296712e-01 6.20782614e-01
-1.56079456e-01 5.21587014e-01 -2.05780268e-01 7.06382632e-01
3.87643725e-01 -4.89901677e-02 -2.42223561e-01 -3.65967989e-01
1.74490154e-01 7.22129107e-01 -1.34842193e+00 1.74122286e+00
-4.07789797e-01 7.56497264e-01 6.47398055e-01 -6.71080470e-01
9.26088035e-01 6.91139042e-01 4.58204865e-01 -4.01140243e-01
6.81777438e-03 2.75580376e-01 -3.20364743e-01 -5.15663087e-01
4.23476577e-01 -9.21831653e-02 -4.40414667e-01 3.94372463e-01
2.76090503e-01 -2.39683807e-01 2.09714219e-01 2.42013276e-01
8.43308568e-01 -2.73062348e-01 4.89833057e-02 -3.01080348e-04
5.32981157e-01 -3.21616173e-01 7.12661803e-01 7.01397717e-01
-3.44487816e-01 7.68114507e-01 2.80791968e-01 7.63058662e-02
-4.81505483e-01 -1.42604113e+00 4.08159532e-02 1.43176770e+00
6.51612505e-02 -6.92841232e-01 -4.32534546e-01 -2.72559911e-01
-1.36540249e-01 7.08059251e-01 -1.14789568e-01 -2.23375872e-01
-3.20873648e-01 -1.88296393e-01 1.09131253e+00 3.67488325e-01
3.39913517e-01 -1.04905558e+00 -1.76756456e-01 2.58014441e-01
-8.36042225e-01 -1.41357052e+00 -7.32323110e-01 1.90769061e-01
-1.96976930e-01 -6.82261705e-01 -7.19139218e-01 -9.49198961e-01
8.28064829e-02 2.41302788e-01 1.15774858e+00 -4.79975045e-01
1.64421245e-01 5.26862860e-01 -3.24959129e-01 -5.42536438e-01
-5.37098825e-01 1.42795086e-01 2.86943495e-01 3.90695870e-01
2.92060673e-01 -7.49223590e-01 -3.63747478e-01 3.07615340e-01
-7.76918352e-01 -1.60570323e-01 2.90324360e-01 5.46992362e-01
7.48524427e-01 2.01770335e-01 8.58762920e-01 -3.99718583e-01
9.97938216e-01 -6.93593562e-01 -1.73005074e-01 -7.63132796e-02
-9.37118679e-02 -3.38612735e-01 8.28936696e-01 -7.44517982e-01
-7.69061029e-01 1.88253030e-01 -3.03992927e-01 -6.80991888e-01
-5.14016747e-01 6.15819871e-01 -3.08130771e-01 6.93814099e-01
4.80474591e-01 4.17844981e-01 -3.73338759e-01 -6.60126746e-01
1.60438538e-01 1.17910969e+00 1.06683064e+00 -6.78129375e-01
6.05103195e-01 3.46370876e-01 -7.05280781e-01 -1.03942716e+00
-1.07348943e+00 -8.32253873e-01 -1.49799168e-01 -1.80426046e-01
5.98358572e-01 -1.52238667e+00 -6.94524169e-01 4.17832255e-01
-1.31727135e+00 -2.74787128e-01 -1.56399116e-01 6.27629638e-01
-5.99208176e-01 4.96579297e-02 -6.28262222e-01 -1.09762168e+00
-3.73780817e-01 -9.79663551e-01 1.41239262e+00 3.50442939e-02
-4.72800463e-01 -5.09344101e-01 1.87457323e-01 3.90761226e-01
1.21702909e-01 -6.45129457e-02 5.98438084e-01 -1.08946371e+00
-3.25671315e-01 -1.12768248e-01 -3.04557756e-02 4.71848071e-01
2.33903557e-01 -2.12251246e-01 -1.53231430e+00 -2.99253047e-01
7.38321915e-02 -7.03666389e-01 8.00410688e-01 8.81705508e-02
1.18624258e+00 -5.43556154e-01 5.13129402e-03 5.57099462e-01
1.05499136e+00 2.25187480e-01 2.71397736e-02 -1.69610128e-01
6.55166030e-01 5.64034879e-01 3.88953835e-01 5.11544108e-01
4.81542408e-01 6.68139815e-01 2.10930571e-01 -1.06009379e-01
-2.32445225e-01 -5.75851440e-01 6.51211679e-01 1.15290415e+00
7.25646973e-01 -3.66874605e-01 -1.00374246e+00 7.81212807e-01
-1.64458942e+00 -9.65514898e-01 2.89710402e-01 1.87196505e+00
1.06379735e+00 7.11456984e-02 5.06432235e-01 2.28335634e-01
7.61521459e-01 2.66997188e-01 -5.96624970e-01 7.16280341e-02
-8.01068917e-02 -2.20711187e-01 -1.96521059e-01 5.81001043e-01
-1.22184622e+00 6.12615287e-01 5.56871939e+00 1.09657156e+00
-1.24499142e+00 2.23715514e-01 3.45093697e-01 -3.51956576e-01
-1.69591770e-01 -3.61533374e-01 -7.79126227e-01 5.44435382e-01
1.19034302e+00 -3.94534618e-01 5.04742384e-01 5.83335340e-01
5.24948299e-01 1.18797764e-01 -1.47470605e+00 1.23922801e+00
2.63350010e-01 -7.25332141e-01 8.47317427e-02 -4.07457620e-01
1.99952200e-01 2.40843594e-01 5.33382446e-02 5.28869510e-01
1.59548253e-01 -8.95987153e-01 1.26408923e+00 3.08024257e-01
8.07963490e-01 -6.73988461e-01 3.92545372e-01 5.15115261e-01
-1.50389767e+00 -1.76408961e-01 -5.47399372e-02 1.54922724e-01
2.75202960e-01 5.28569221e-01 -1.08968735e+00 3.82708848e-01
9.42467868e-01 6.71941042e-01 -2.79886603e-01 9.64119673e-01
-2.77509332e-01 1.05276835e+00 -3.21474344e-01 1.71540648e-01
5.43454140e-02 2.16733366e-01 1.00360548e+00 1.60767543e+00
1.93774730e-01 -2.37332150e-01 4.65102434e-01 1.26914334e+00
-3.49817991e-01 1.85029626e-01 -4.95524734e-01 -2.92203337e-01
8.21884513e-01 1.10687840e+00 -3.07046890e-01 -2.48241395e-01
-4.05911356e-02 7.36032367e-01 1.66646644e-01 7.00974941e-01
-9.37367916e-01 -6.63514495e-01 6.41627431e-01 3.18839028e-02
1.24840595e-01 -4.90959622e-02 3.19657087e-01 -1.02558553e+00
2.51989037e-01 -1.00868154e+00 3.65688831e-01 -1.02013552e+00
-1.47454202e+00 1.00557852e+00 7.47280344e-02 -1.56845427e+00
-3.88230920e-01 8.57047886e-02 -5.44375360e-01 8.37805688e-01
-1.39347208e+00 -1.00486660e+00 -2.11015284e-01 6.76881731e-01
8.57180417e-01 -3.02377582e-01 8.13775182e-01 5.95107794e-01
-4.51654911e-01 7.18561411e-01 -6.70419727e-03 4.80657786e-01
1.04596281e+00 -1.02101457e+00 1.48564056e-01 7.10505009e-01
5.25110722e-01 3.97031546e-01 7.19111204e-01 -2.89602339e-01
-1.15902758e+00 -1.40700352e+00 9.54784811e-01 -1.20186590e-01
8.59050512e-01 -7.93199122e-01 -9.25022006e-01 5.12445509e-01
2.29719952e-01 4.85804416e-02 8.82828534e-01 -7.06046820e-02
-5.40647745e-01 -4.91973341e-01 -5.99626958e-01 3.53522360e-01
7.44636714e-01 -1.20307088e+00 -4.62783843e-01 2.94446886e-01
1.11373675e+00 -2.93192565e-01 -5.29009461e-01 2.84376770e-01
4.23660517e-01 -7.54758179e-01 9.33091223e-01 -3.61964047e-01
2.92920500e-01 -4.08742994e-01 -4.69186574e-01 -1.36631608e+00
-6.57105595e-02 -7.39419997e-01 3.48124802e-02 1.78058505e+00
4.17478025e-01 -3.87165159e-01 1.95377558e-01 2.02133656e-01
-1.05025865e-01 -4.67459500e-01 -9.69056368e-01 -9.99834657e-01
-2.34619588e-01 -8.67418051e-01 4.51168388e-01 7.43227482e-01
2.26963460e-02 7.00831413e-01 -5.85480154e-01 5.03708363e-01
5.06055772e-01 1.73165336e-01 8.60180795e-01 -1.00529468e+00
-6.09240115e-01 -3.10254395e-01 -7.64899254e-02 -1.23364401e+00
5.81128299e-01 -9.98306036e-01 5.81495285e-01 -1.21055388e+00
1.02534004e-01 -1.38525754e-01 -3.88361722e-01 3.76422614e-01
3.69732752e-02 2.18678221e-01 3.43881220e-01 4.45290834e-01
-8.93881798e-01 7.91301906e-01 6.31416976e-01 -5.65451443e-01
-3.90615106e-01 1.72021136e-01 -8.26451540e-01 6.73508346e-01
6.57644272e-01 -7.10470378e-01 -4.92968202e-01 -5.74019134e-01
2.92390864e-02 3.73469114e-01 5.38783014e-01 -1.03645587e+00
4.46083903e-01 9.03426930e-02 -2.69461542e-01 -5.85887194e-01
7.50360072e-01 -7.23748803e-01 6.03526160e-02 -2.94263270e-02
-9.53475773e-01 -1.65406600e-01 2.72317231e-01 5.69651842e-01
-8.46152842e-01 -3.60547639e-02 5.19363821e-01 1.65703088e-01
-4.25360024e-01 2.18901500e-01 -4.29392636e-01 4.24692392e-01
3.67005199e-01 3.65549922e-01 -1.20608553e-01 -9.95602548e-01
-7.84360349e-01 3.54673177e-01 -1.95000753e-01 5.75116038e-01
6.60342813e-01 -1.70838511e+00 -9.68562365e-01 7.85680413e-02
3.68896544e-01 1.54464528e-01 5.46273440e-02 7.69396484e-01
1.56629935e-01 3.28047127e-01 3.78301173e-01 -8.09630156e-01
-1.21624279e+00 3.37938547e-01 3.33475620e-01 3.44325781e-01
-3.93561929e-01 9.42462146e-01 4.39140499e-01 -3.37555617e-01
1.02317595e+00 -4.07508582e-01 -9.78502035e-02 8.37109312e-02
6.85706675e-01 2.70318627e-01 -1.71589985e-01 -8.87268424e-01
-4.95925367e-01 2.75284231e-01 2.49197930e-01 -4.48489040e-01
1.24024868e+00 -4.14308161e-01 -4.00186069e-02 1.16272175e+00
1.52773488e+00 3.47965270e-01 -9.48474944e-01 -6.50326788e-01
-1.48500219e-01 -1.56616509e-01 -8.12419578e-02 -6.38768733e-01
-9.38771129e-01 8.63030136e-01 5.41449368e-01 3.49570721e-01
1.28009903e+00 4.39372182e-01 9.97235298e-01 5.27632296e-01
2.31511258e-02 -8.86049449e-01 2.14001313e-01 5.02569497e-01
1.19713175e+00 -1.28218639e+00 -4.91484940e-01 -2.25086272e-01
-6.84575975e-01 8.94804597e-01 4.81397897e-01 2.60060318e-02
7.02849150e-01 3.47463697e-01 4.65286255e-01 -1.40998974e-01
-8.40623379e-01 -1.78599298e-01 4.64650631e-01 4.32735056e-01
3.87522042e-01 8.80924761e-02 5.09548545e-01 1.20569229e+00
-4.70184386e-01 -2.06913739e-01 3.36197436e-01 5.18479168e-01
-2.92428225e-01 -6.10959053e-01 -5.72614849e-01 -7.91535154e-02
-5.96165419e-01 -1.14344478e-01 -6.72862113e-01 2.68068254e-01
-3.81015502e-02 1.37142003e+00 2.16278490e-02 -5.30552745e-01
4.94306952e-01 1.96134225e-01 -1.62592724e-01 -7.04100728e-01
-3.47214788e-01 8.17875564e-01 1.22332439e-01 -4.89460260e-01
-4.84407276e-01 -4.39310580e-01 -1.34668672e+00 3.08066130e-01
1.46453539e-02 6.23030484e-01 4.01950330e-01 6.29998684e-01
3.58937591e-01 5.15507579e-01 8.50852013e-01 -8.60158443e-01
-6.33442998e-01 -1.01259291e+00 -6.32478595e-01 4.44740921e-01
9.57378089e-01 -3.09431463e-01 -6.13098145e-01 3.81881446e-01] | [15.258419036865234, 5.188126087188721] |
b30b9110-7779-409f-a841-dd6a301d3ad2 | completionformer-depth-completion-with | 2304.13030 | null | https://arxiv.org/abs/2304.13030v1 | https://arxiv.org/pdf/2304.13030v1.pdf | CompletionFormer: Depth Completion with Convolutions and Vision Transformers | Given sparse depths and the corresponding RGB images, depth completion aims at spatially propagating the sparse measurements throughout the whole image to get a dense depth prediction. Despite the tremendous progress of deep-learning-based depth completion methods, the locality of the convolutional layer or graph model makes it hard for the network to model the long-range relationship between pixels. While recent fully Transformer-based architecture has reported encouraging results with the global receptive field, the performance and efficiency gaps to the well-developed CNN models still exist because of its deteriorative local feature details. This paper proposes a Joint Convolutional Attention and Transformer block (JCAT), which deeply couples the convolutional attention layer and Vision Transformer into one block, as the basic unit to construct our depth completion model in a pyramidal structure. This hybrid architecture naturally benefits both the local connectivity of convolutions and the global context of the Transformer in one single model. As a result, our CompletionFormer outperforms state-of-the-art CNNs-based methods on the outdoor KITTI Depth Completion benchmark and indoor NYUv2 dataset, achieving significantly higher efficiency (nearly 1/3 FLOPs) compared to pure Transformer-based methods. Code is available at \url{https://github.com/youmi-zym/CompletionFormer}. | ['Mattoccia Stefano', 'Huang Guan', 'Zhu Zheng', 'Poggi Matteo', 'Guo Xianda', 'Zhang Youmin'] | 2023-04-25 | null | http://openaccess.thecvf.com//content/CVPR2023/html/Zhang_CompletionFormer_Depth_Completion_With_Convolutions_and_Vision_Transformers_CVPR_2023_paper.html | http://openaccess.thecvf.com//content/CVPR2023/papers/Zhang_CompletionFormer_Depth_Completion_With_Convolutions_and_Vision_Transformers_CVPR_2023_paper.pdf | cvpr-2023-1 | ['depth-completion'] | ['computer-vision'] | [ 1.33519284e-02 2.19239578e-01 6.80740550e-02 -4.48431611e-01
-4.97625589e-01 -9.64428782e-02 4.84488577e-01 -2.27586642e-01
-1.53356493e-01 4.24307615e-01 4.62164938e-01 -2.10360885e-01
1.29742667e-01 -1.05840039e+00 -9.44126368e-01 -6.36617661e-01
1.48111910e-01 1.36129633e-01 3.06292027e-01 -7.13375807e-02
1.15296222e-01 3.73472393e-01 -1.39467359e+00 3.05634379e-01
8.86253357e-01 1.36811066e+00 5.61374068e-01 6.62179112e-01
-1.99084222e-01 1.32242346e+00 -1.41320061e-02 -1.83218852e-01
2.81082839e-01 -1.79403767e-01 -6.34763062e-01 -8.66653100e-02
6.73695922e-01 -8.97622049e-01 -1.06722963e+00 8.87216449e-01
3.63482565e-01 1.51749365e-02 1.19953498e-01 -8.71916115e-01
-8.07439744e-01 2.92280763e-01 -6.16316617e-01 -1.85968392e-02
3.11953515e-01 1.68938175e-01 1.02452242e+00 -1.01095080e+00
3.93086672e-01 1.06625283e+00 6.39764965e-01 5.31312048e-01
-9.38720107e-01 -6.10372424e-01 3.76413941e-01 3.77566308e-01
-1.36185515e+00 -3.54900599e-01 7.83361375e-01 -1.21587068e-01
1.34536850e+00 -7.64916465e-02 1.04721272e+00 9.77106869e-01
3.76378775e-01 7.81337619e-01 8.29880893e-01 -4.52063791e-02
1.13373257e-01 -4.75019187e-01 -7.69009516e-02 8.24060619e-01
1.22754589e-01 6.48568347e-02 -6.48084998e-01 3.70711207e-01
1.21614587e+00 5.11013567e-01 -4.75782007e-01 -3.69378418e-01
-1.07832348e+00 6.42715514e-01 1.26138628e+00 6.21603243e-02
-4.59458560e-01 7.71108627e-01 -9.57109481e-02 -6.36558980e-02
5.65584779e-01 -1.22000771e-02 -3.59882325e-01 -1.52658112e-02
-8.14377725e-01 1.56579077e-01 6.05533540e-01 1.06198478e+00
1.31069052e+00 -9.32282954e-02 4.96650077e-02 5.18723071e-01
5.49831092e-01 2.68656671e-01 2.13465154e-01 -9.93193030e-01
6.26233339e-01 8.23810220e-01 -2.35884145e-01 -8.32531333e-01
-2.48096898e-01 -6.05810106e-01 -1.09830678e+00 2.67097980e-01
2.11801648e-01 4.46950225e-03 -1.07147753e+00 1.45462310e+00
2.13680044e-01 3.70773584e-01 -8.53194743e-02 9.99209821e-01
1.06531739e+00 7.02117920e-01 -3.17219406e-01 3.70403558e-01
9.29276168e-01 -1.41564214e+00 -4.85958904e-01 -7.30232894e-01
4.75959152e-01 -5.24498820e-01 9.53493178e-01 3.12898397e-01
-1.16837108e+00 -6.70518398e-01 -1.27848589e+00 -1.01386249e+00
-1.33438751e-01 -6.17158562e-02 1.03392541e+00 8.38729963e-02
-1.44757307e+00 6.30605638e-01 -1.13271475e+00 -2.62881488e-01
7.15277016e-01 4.01096970e-01 -3.86039585e-01 -7.93475330e-01
-7.93443978e-01 4.27463859e-01 -1.42978147e-01 4.84326929e-01
-1.14248228e+00 -8.91632318e-01 -1.09940040e+00 2.08605424e-01
-8.28087702e-02 -9.83176470e-01 1.21426678e+00 -5.95763385e-01
-1.28955317e+00 5.55762827e-01 -3.62166494e-01 -3.21613163e-01
4.04266417e-01 -3.83981913e-01 3.08251113e-01 1.37150660e-01
9.89181474e-02 1.04747367e+00 5.42418778e-01 -1.00742567e+00
-6.65855467e-01 -6.57043576e-01 3.52338046e-01 3.31799984e-01
-1.69387996e-01 -6.90996826e-01 -8.49671781e-01 -3.58933389e-01
6.30957544e-01 -6.26948059e-01 -4.31624413e-01 3.51006657e-01
-2.68383414e-01 1.08619584e-02 7.30641007e-01 -5.04362345e-01
1.07788575e+00 -2.14275455e+00 3.52827221e-01 -2.40605876e-01
6.83625281e-01 -1.74031734e-01 -9.23692286e-02 4.82214719e-01
-3.35403532e-02 -2.03862071e-01 -1.57671571e-01 -1.13130331e+00
-1.33075640e-01 4.00629699e-01 -2.60252148e-01 5.39595544e-01
3.57011296e-02 1.11786127e+00 -9.27314579e-01 -4.34935875e-02
5.37453473e-01 9.40570593e-01 -8.50178123e-01 3.93768460e-01
-1.50991902e-01 4.38515157e-01 -3.88412058e-01 8.19363415e-01
8.58560920e-01 -4.07161832e-01 -1.83719713e-02 -4.21205163e-01
-1.20697908e-01 6.08013391e-01 -7.13343501e-01 2.44954252e+00
-5.50991595e-01 8.54760528e-01 3.60374041e-02 -8.07593644e-01
8.28024387e-01 1.61323138e-02 4.12854880e-01 -1.05682790e+00
4.50220667e-02 1.96579263e-01 -2.15383187e-01 -2.20996082e-01
5.25015473e-01 1.34954423e-01 2.03335881e-01 5.36112301e-02
1.18020706e-01 -3.43524069e-01 -2.21315354e-01 4.01570618e-01
1.36777031e+00 4.76202995e-01 -4.76429313e-02 -9.91689786e-02
1.53472781e-01 -3.31727415e-01 5.30365348e-01 4.02438015e-01
-6.73972769e-03 1.10993421e+00 4.22997892e-01 -6.60420954e-01
-9.43172514e-01 -1.09326363e+00 4.02391478e-02 5.79386711e-01
3.57389480e-01 -5.89371622e-01 -6.37988150e-01 -2.95194924e-01
-1.04202949e-01 6.89560324e-02 -9.44566190e-01 -7.72372484e-02
-4.47080642e-01 -2.31702447e-01 2.94732094e-01 8.22062373e-01
1.07978213e+00 -8.03136289e-01 -5.01687706e-01 1.82326123e-01
-3.18196148e-01 -1.30741358e+00 -3.02107334e-01 3.35687160e-01
-1.08991420e+00 -9.68050241e-01 -6.48042917e-01 -7.32808709e-01
7.10481286e-01 5.58661759e-01 1.14666724e+00 2.42653340e-01
-1.53103605e-01 2.25359082e-01 -3.18924844e-01 -1.90055147e-01
4.13497359e-01 9.17883590e-02 -5.27438581e-01 -1.58681408e-01
3.22272390e-01 -1.09398091e+00 -1.27086496e+00 -6.49019629e-02
-8.65878999e-01 4.67788041e-01 6.28003836e-01 8.22636902e-01
7.23122001e-01 -2.65141636e-01 -6.61206469e-02 -5.52985370e-01
-2.59469636e-02 -4.45107996e-01 -5.58690727e-01 -1.42581135e-01
-5.02634287e-01 -1.05724446e-01 3.34764808e-01 1.47789106e-01
-8.63735199e-01 2.62058616e-01 -4.46709722e-01 -6.92761242e-01
6.67607039e-02 4.03951347e-01 -3.20679992e-01 -2.47520685e-01
4.29090500e-01 3.05689335e-01 -1.18688859e-01 -5.49779713e-01
1.37234285e-01 2.09240273e-01 5.20746052e-01 -3.08893383e-01
6.17878973e-01 7.84945190e-01 -2.36467347e-02 -5.97606182e-01
-1.06212187e+00 -4.44300532e-01 -6.01768553e-01 -7.29387403e-02
9.34395671e-01 -1.46001875e+00 -6.02862358e-01 8.21012795e-01
-1.33135402e+00 -8.64915013e-01 -2.57906824e-01 3.14146996e-01
-3.70101601e-01 3.26271683e-01 -7.70420969e-01 -3.54782581e-01
-2.71282762e-01 -1.21682274e+00 1.27850068e+00 2.48201177e-01
2.79933155e-01 -1.18455803e+00 -6.09632730e-02 4.49026525e-01
5.03312886e-01 1.71770066e-01 4.55320150e-01 3.69057953e-01
-1.44510794e+00 -3.92897688e-02 -5.93216121e-01 4.55335438e-01
1.23444997e-01 -2.93382645e-01 -1.40811801e+00 -2.63907641e-01
9.51766148e-02 -3.13308924e-01 1.32130337e+00 5.51066995e-01
1.28966033e+00 -8.96108150e-02 -2.04940081e-01 1.38113236e+00
1.77862942e+00 -1.88732132e-01 1.21020675e+00 2.13454023e-01
1.26489174e+00 3.29296738e-01 1.85973272e-01 4.84547466e-01
8.98597658e-01 2.62348086e-01 1.13014984e+00 -4.08294052e-01
-4.88265216e-01 -5.88353574e-01 1.42173260e-01 6.95201933e-01
-2.02545345e-01 -1.93243995e-01 -8.16880047e-01 5.93441606e-01
-1.88564265e+00 -6.64187074e-01 -7.16312975e-02 1.89527428e+00
4.66786385e-01 7.89547265e-02 -5.12478232e-01 2.09923804e-01
5.52531257e-02 5.19056737e-01 -6.22321069e-01 -1.60176605e-01
-4.88141067e-02 4.00079846e-01 5.85787296e-01 8.16748559e-01
-8.52644622e-01 8.66875947e-01 5.43096495e+00 5.83126128e-01
-1.07966149e+00 4.48366627e-02 7.88436174e-01 -1.97427318e-01
-5.77158630e-01 1.53408289e-01 -8.56118739e-01 6.75223917e-02
4.39266831e-01 2.85394520e-01 5.39223850e-01 7.58952618e-01
1.10522933e-01 -3.39069963e-01 -1.24558842e+00 1.21267951e+00
5.13990223e-03 -1.56048226e+00 1.07212223e-01 4.03474420e-01
9.84041929e-01 6.80882454e-01 1.11894429e-01 8.98828655e-02
1.67231470e-01 -1.28637218e+00 8.31079543e-01 4.48997885e-01
8.30526948e-01 -5.17861426e-01 6.44561887e-01 2.61198908e-01
-1.54709280e+00 -1.69043496e-01 -7.01162696e-01 -6.62481546e-01
-3.73922698e-02 8.56840253e-01 -2.81185180e-01 5.53570986e-01
1.05929303e+00 1.51132572e+00 -4.91568416e-01 1.00228846e+00
-4.10598725e-01 2.57330865e-01 -2.48428464e-01 3.30183357e-01
5.23179591e-01 -1.76141426e-01 1.86765119e-02 8.72257113e-01
5.07444918e-01 2.18674332e-01 -1.58629864e-01 1.00247443e+00
-2.12343127e-01 -3.32966745e-01 -6.53867483e-01 1.55250877e-01
1.83217376e-01 1.26015329e+00 -4.55875695e-01 -1.94348842e-01
-7.61194468e-01 1.10921288e+00 7.08617985e-01 5.80866635e-01
-7.74805129e-01 -2.61816025e-01 1.12036884e+00 3.54027212e-01
5.67973018e-01 -4.71020430e-01 -7.04987586e-01 -1.23961699e+00
2.41942212e-01 -4.74590689e-01 3.66104320e-02 -9.83064413e-01
-9.48397577e-01 6.80014014e-01 -5.40298164e-01 -1.10995138e+00
8.06959420e-02 -5.91410220e-01 -5.73342443e-01 1.07359612e+00
-2.00905657e+00 -1.45424318e+00 -1.10922229e+00 9.04815018e-01
5.51838338e-01 3.85390222e-01 7.35397339e-01 4.74087328e-01
-4.60040122e-01 3.47556531e-01 -1.37755468e-01 1.80658087e-01
3.61988187e-01 -1.15818250e+00 6.36999428e-01 9.60046351e-01
-1.12159722e-01 4.79259759e-01 1.72390327e-01 -5.12709737e-01
-1.65041089e+00 -1.10544682e+00 7.73019016e-01 -3.34078759e-01
5.21481216e-01 -5.37306249e-01 -7.12829471e-01 8.81721079e-01
1.95222273e-01 4.64134037e-01 1.38903305e-01 -2.24534646e-02
-5.51460505e-01 -4.10508007e-01 -7.01474369e-01 5.46655655e-01
1.44019413e+00 -6.88137889e-01 -6.20228238e-02 1.26594812e-01
9.22862172e-01 -6.78521395e-01 -7.75576532e-01 4.04643059e-01
6.19916737e-01 -1.62561953e+00 1.08510888e+00 3.15011173e-01
1.00763977e+00 -2.74565369e-01 -4.57850337e-01 -9.55554426e-01
-4.24911320e-01 -4.23067153e-01 -2.62540728e-01 7.51949012e-01
1.41868562e-01 -6.01047933e-01 1.06824386e+00 4.85092342e-01
-6.23355329e-01 -1.10202765e+00 -9.04558718e-01 -1.50140002e-01
8.23093392e-03 -6.77048802e-01 6.65800273e-01 5.40880024e-01
-2.09853306e-01 3.75506550e-01 -2.61670083e-01 2.82183379e-01
7.94674575e-01 -8.15448985e-02 7.69386709e-01 -8.62313330e-01
-4.58791405e-01 -3.32649618e-01 -5.47760248e-01 -1.93158472e+00
-2.84616649e-01 -6.34789526e-01 8.58122930e-02 -2.23073602e+00
8.15771818e-02 -5.65807819e-01 -1.08260497e-01 6.24444723e-01
-7.21397400e-02 5.55576324e-01 4.58662957e-02 1.07911125e-01
-4.68179852e-01 8.78618419e-01 1.66321313e+00 -1.91038743e-01
-2.46322434e-02 -3.01173627e-01 -7.48705208e-01 5.09190857e-01
5.86439490e-01 -1.37883425e-01 -7.66697943e-01 -1.15085387e+00
3.01886141e-01 4.53505293e-02 6.29345775e-01 -1.37250233e+00
5.06116986e-01 1.84837073e-01 6.11612976e-01 -7.85493553e-01
7.25270629e-01 -7.45377243e-01 7.31273293e-02 4.13210243e-01
1.05965652e-01 -2.57433858e-02 1.74801096e-01 5.27517200e-01
-4.32760030e-01 3.73654157e-01 6.33233130e-01 -3.10547888e-01
-9.11378622e-01 9.80094910e-01 5.97364567e-02 -1.55463055e-01
6.55754626e-01 -4.86431062e-01 -3.68466586e-01 -4.11881804e-01
-4.50256109e-01 1.48005947e-01 6.86677098e-01 2.22029790e-01
1.06926274e+00 -1.08379257e+00 -5.29880941e-01 5.23500085e-01
-1.50265209e-02 9.37175155e-01 5.82760870e-01 8.84809911e-01
-8.61765325e-01 3.26115072e-01 -2.65822560e-01 -7.59321630e-01
-6.94889486e-01 2.41653532e-01 5.79612851e-01 -2.02337369e-01
-9.91812348e-01 1.27593410e+00 9.07679260e-01 -1.32529452e-01
4.60699588e-01 -8.41080248e-01 1.59910053e-01 -3.86628151e-01
5.61573386e-01 1.04867049e-01 1.09779909e-01 -2.03590214e-01
-2.64330208e-01 7.42376626e-01 4.07350585e-02 1.53916076e-01
1.48076737e+00 -2.93903410e-01 -3.09483349e-01 1.53368711e-01
1.39186108e+00 -2.49872327e-01 -1.90944457e+00 -3.45502764e-01
-6.75490677e-01 -5.86981952e-01 4.28616017e-01 -3.92989874e-01
-1.46880722e+00 1.31785810e+00 3.30406040e-01 -2.42355287e-01
1.36968482e+00 -8.35943297e-02 9.28335428e-01 2.55997509e-01
4.09152925e-01 -5.88779032e-01 3.44071120e-01 7.07877994e-01
9.87667382e-01 -1.29269564e+00 2.10189037e-02 -4.35537577e-01
-1.48394391e-01 9.31317091e-01 7.97752559e-01 -4.04058039e-01
9.19812262e-01 3.94208461e-01 -9.87339392e-02 -2.87926108e-01
-7.73034275e-01 -1.58813760e-01 7.27604777e-02 7.61151195e-01
4.12208170e-01 -1.85612366e-01 2.46409133e-01 3.63721818e-01
-1.49885565e-01 2.00614750e-01 4.00722086e-01 8.40367794e-01
-3.64986449e-01 -8.77821386e-01 5.50485328e-02 2.32900113e-01
-1.18225917e-01 -4.93144780e-01 -5.53929619e-02 5.94603121e-01
1.73066258e-01 7.77777135e-01 2.17477605e-01 -5.72451651e-01
2.35504478e-01 -5.64349651e-01 5.60578167e-01 -6.87999368e-01
-3.50828141e-01 5.53296953e-02 -1.09907158e-01 -1.22506058e+00
-3.84478003e-01 -3.77252162e-01 -1.23695147e+00 -6.72247767e-01
1.98544055e-01 -3.56452793e-01 6.08620167e-01 8.00817966e-01
3.60142112e-01 8.03928554e-01 5.27799726e-01 -1.29994893e+00
5.22576421e-02 -9.49466884e-01 -5.62758148e-01 -9.28777549e-03
5.69029689e-01 -4.27711099e-01 -3.20953548e-01 -2.47786939e-01] | [8.82699203491211, -2.5334582328796387] |
d049bf0c-aeb9-403f-ba6c-7f6a960e4a4a | pp-structurev2-a-stronger-document-analysis | 2210.05391 | null | https://arxiv.org/abs/2210.05391v2 | https://arxiv.org/pdf/2210.05391v2.pdf | PP-StructureV2: A Stronger Document Analysis System | A large amount of document data exists in unstructured form such as raw images without any text information. Designing a practical document image analysis system is a meaningful but challenging task. In previous work, we proposed an intelligent document analysis system PP-Structure. In order to further upgrade the function and performance of PP-Structure, we propose PP-StructureV2 in this work, which contains two subsystems: Layout Information Extraction and Key Information Extraction. Firstly, we integrate Image Direction Correction module and Layout Restoration module to enhance the functionality of the system. Secondly, 8 practical strategies are utilized in PP-StructureV2 for better performance. For Layout Analysis model, we introduce ultra light-weight detector PP-PicoDet and knowledge distillation algorithm FGD for model lightweighting, which increased the inference speed by 11 times with comparable mAP. For Table Recognition model, we utilize PP-LCNet, CSP-PAN and SLAHead to optimize the backbone module, feature fusion module and decoding module, respectively, which improved the table structure accuracy by 6\% with comparable inference speed. For Key Information Extraction model, we introduce VI-LayoutXLM which is a visual-feature independent LayoutXLM architecture, TB-YX sorting algorithm and U-DML knowledge distillation algorithm, which brought 2.8\% and 9.1\% improvement respectively on the Hmean of Semantic Entity Recognition and Relation Extraction tasks. All the above mentioned models and code are open-sourced in the GitHub repository PaddleOCR. | ['dianhai yu', 'Xiaoguang Hu', 'Yi Liu', 'Lingfeng Zhu', 'Yuning Du', 'Mengtao An', 'Jun Zhou', 'Ruoyu Guo', 'Chenxia Li'] | 2022-10-11 | null | null | null | null | ['table-recognition', 'key-information-extraction'] | ['computer-vision', 'natural-language-processing'] | [ 2.13912472e-01 -9.05184969e-02 -5.62655665e-02 -1.87816054e-01
-2.31524408e-01 -5.55983663e-01 3.68150324e-01 3.09139401e-01
-2.94106305e-01 4.58518058e-01 -5.45400456e-02 -4.36282277e-01
-2.96658218e-01 -1.11297441e+00 -6.36642277e-01 -2.69844979e-01
4.32020098e-01 5.04872501e-01 5.59213161e-01 -9.07660928e-03
7.61479676e-01 6.23083472e-01 -1.42702913e+00 7.36126006e-01
9.20258284e-01 1.11707056e+00 6.80512965e-01 7.90732145e-01
-8.98105383e-01 8.53637278e-01 -6.61863744e-01 -3.60016704e-01
-4.62358594e-02 -7.72606488e-03 -9.57811177e-01 1.31824136e-01
2.45532081e-01 -4.57412571e-01 -4.81900245e-01 1.08113480e+00
6.49142683e-01 -1.84244663e-01 5.97469747e-01 -1.17988884e+00
-6.37987494e-01 8.46284449e-01 -7.67735302e-01 1.68974311e-05
3.84415627e-01 -1.10366821e-01 6.45346820e-01 -9.10963416e-01
8.95909190e-01 1.43396723e+00 5.24368763e-01 -7.39960407e-04
-6.26871884e-01 -5.86338758e-01 8.62239748e-02 7.82025635e-01
-1.62866533e+00 -3.86025250e-01 5.93241751e-01 -6.28583059e-02
1.44840467e+00 4.71755028e-01 3.24564755e-01 3.35820466e-01
2.59654343e-01 1.25770199e+00 6.76648736e-01 -6.73213959e-01
-1.92462727e-01 2.75786519e-01 5.29150963e-01 1.09665799e+00
4.07042921e-01 -7.56771624e-01 -4.96128291e-01 4.16383147e-01
6.30135477e-01 -9.51236635e-02 1.02175097e-03 -8.36576223e-02
-9.54606771e-01 2.38937125e-01 4.85618979e-01 2.76473641e-01
-2.99370848e-02 -5.02917804e-02 4.87069994e-01 -7.37830345e-03
-2.70749450e-01 1.67706430e-01 -4.39340711e-01 -8.91090706e-02
-7.96427310e-01 5.84997237e-02 5.45116782e-01 1.58178139e+00
5.23389101e-01 -2.53068238e-01 -3.51296723e-01 8.73310626e-01
4.76374626e-01 8.13190043e-01 3.68425637e-01 -6.02524459e-01
9.89450455e-01 1.35294998e+00 -4.15906817e-01 -1.32826173e+00
-7.23330259e-01 -1.30906865e-01 -9.53506529e-01 -2.30741322e-01
-6.74489886e-02 1.47710949e-01 -1.20503914e+00 8.69027495e-01
1.55726492e-01 -3.63533854e-01 8.42345878e-02 4.89880890e-01
1.13656008e+00 8.77114832e-01 -1.88638270e-01 -1.49736241e-01
1.86928928e+00 -1.03519082e+00 -1.10027909e+00 -3.48906890e-02
8.08002174e-01 -1.22928262e+00 9.62519944e-01 6.22723699e-01
-1.18713129e+00 -6.69823229e-01 -1.31079423e+00 -5.32044113e-01
-7.47506678e-01 6.02792442e-01 6.59954786e-01 4.03682709e-01
-7.00755239e-01 3.23072493e-01 -4.63350862e-01 -5.77893972e-01
6.33463621e-01 5.46333313e-01 -3.29641581e-01 -4.15754199e-01
-7.55140305e-01 6.36739850e-01 9.53876436e-01 4.24339056e-01
-2.00557023e-01 -3.08531523e-01 -6.05132163e-01 2.58283317e-01
6.73384249e-01 -4.44828182e-01 9.74626184e-01 -8.97812992e-02
-1.07064247e+00 5.01479328e-01 -2.06465766e-01 -3.77032943e-02
1.47685289e-01 -1.48841605e-01 -7.17736363e-01 2.97702342e-01
5.80101311e-02 5.86173296e-01 3.46754521e-01 -9.41728950e-01
-8.89845252e-01 -5.97754359e-01 -4.06655312e-01 5.11735082e-01
-3.93047899e-01 -8.40052515e-02 -1.45166373e+00 -3.83382082e-01
3.41947645e-01 -4.56309378e-01 1.21005267e-01 -2.06942722e-01
-6.44377768e-01 -5.53874038e-02 1.28814042e+00 -9.24767435e-01
1.83949220e+00 -1.96457219e+00 -1.42937988e-01 5.84232152e-01
7.51353428e-02 6.98628783e-01 1.34188369e-01 4.33697224e-01
4.04449284e-01 5.45831285e-02 -5.69189005e-02 7.21575171e-02
-1.34734455e-02 1.45904437e-01 -6.82368949e-02 -8.10389519e-02
-7.43503198e-02 1.13236678e+00 -2.54956812e-01 -1.00992763e+00
5.20360231e-01 2.01395348e-01 -4.77968901e-01 -1.73307344e-01
-2.47983873e-01 -3.45824987e-01 -5.01399934e-01 8.94948483e-01
1.10166395e+00 -3.09166938e-01 4.83442664e-01 -7.97398984e-01
-2.25211620e-01 -1.27954528e-01 -1.81429088e+00 1.85529375e+00
-1.64523885e-01 5.67731798e-01 3.90200838e-02 -7.74095356e-01
1.02149069e+00 3.49760503e-02 7.35926777e-02 -1.15325212e+00
4.88628328e-01 3.82957831e-02 -2.86745101e-01 -7.54294932e-01
8.27167511e-01 8.49652290e-01 5.46146557e-02 8.59241262e-02
-1.90983899e-02 8.00449997e-02 4.06509578e-01 5.09319067e-01
1.07867908e+00 1.79772198e-01 1.50881335e-01 -3.05293620e-01
1.00879180e+00 3.63395423e-01 2.63633192e-01 6.57182634e-01
3.06673825e-01 3.15054178e-01 5.45461893e-01 -1.45570695e-01
-9.58292305e-01 -8.97994757e-01 -2.93339670e-01 5.34217238e-01
5.23207664e-01 -8.36746037e-01 -9.26718831e-01 -4.55690056e-01
-3.23214382e-02 5.57803988e-01 1.52496342e-02 1.12913921e-01
-7.51306772e-01 -9.40625310e-01 6.03996873e-01 7.59936035e-01
1.25396276e+00 -1.03884256e+00 -3.01874459e-01 2.42244065e-01
7.80148953e-02 -1.12765658e+00 -2.43177429e-01 2.11409882e-01
-7.31310606e-01 -1.12038004e+00 -4.09101322e-02 -1.10202003e+00
9.95966434e-01 1.92556813e-01 7.51059771e-01 2.67303109e-01
-6.90138698e-01 -8.04487839e-02 -3.34063113e-01 -3.07075173e-01
1.46638706e-01 2.77991116e-01 -2.81054765e-01 -4.49071586e-01
4.80762035e-01 -1.84295818e-01 -6.10665202e-01 1.90496162e-01
-8.27049971e-01 4.05075431e-01 9.56214726e-01 5.27243614e-01
6.56301022e-01 4.50834453e-01 1.14356438e-02 -1.04210925e+00
5.76173484e-01 1.91526383e-01 -7.41117835e-01 5.99640071e-01
-1.12079287e+00 2.72942543e-01 7.08151042e-01 1.22058608e-01
-1.32623541e+00 8.24742988e-02 -1.85583308e-01 7.18517378e-02
-3.81806791e-02 2.76498884e-01 -9.19172823e-01 1.58971354e-01
2.15911165e-01 3.18662554e-01 -2.24181637e-01 -7.61046171e-01
5.37828624e-01 1.16422355e+00 7.74045050e-01 -4.14685369e-01
5.09208977e-01 2.08941102e-01 1.14899971e-01 -8.05330873e-01
-3.66718173e-01 -3.50382090e-01 -7.71449983e-01 3.70007604e-02
9.71330166e-01 -7.25721955e-01 -1.20085239e+00 5.36291242e-01
-1.26226103e+00 9.27792639e-02 2.01246426e-01 1.90630272e-01
2.85416767e-02 4.76349473e-01 -6.86648250e-01 -4.96163875e-01
-6.93023384e-01 -1.22603011e+00 9.55892742e-01 3.76250327e-01
1.42457247e-01 -5.18065512e-01 -6.03265345e-01 5.45582712e-01
-8.93062388e-04 -2.77491093e-01 1.15525115e+00 -3.20059597e-01
-9.43522692e-01 -3.42591703e-02 -9.39721406e-01 2.33505398e-01
-1.57355189e-01 4.75198701e-02 -6.84335172e-01 6.14802763e-02
-5.10743678e-01 2.78763294e-01 8.65739882e-01 1.87243402e-01
1.41826570e+00 -2.26668820e-01 -8.56893241e-01 8.95538747e-01
1.70089614e+00 9.16828692e-01 1.00713253e+00 5.20384133e-01
1.10984647e+00 2.73798734e-01 7.05536783e-01 4.05798972e-01
3.84343058e-01 4.99535769e-01 1.15639329e-01 5.01239002e-02
-3.69662493e-01 -3.94856662e-01 -4.30389494e-02 1.09461701e+00
2.27619514e-01 -8.17724705e-01 -9.91470158e-01 2.00416878e-01
-1.97101331e+00 -6.61987484e-01 -6.18980169e-01 1.54337549e+00
5.80722213e-01 4.82105464e-01 -4.91575658e-01 4.75508004e-01
7.19186068e-01 2.86229816e-03 -3.63189995e-01 -5.66281021e-01
-2.75707096e-01 8.70553404e-02 8.08487713e-01 4.43297148e-01
-9.37977135e-01 1.18206882e+00 5.42712927e+00 1.30420566e+00
-6.38424575e-01 -2.31719449e-01 4.83441085e-01 6.94306195e-02
-1.06288299e-01 -1.88473277e-02 -1.40992916e+00 4.47734922e-01
6.02809310e-01 2.29071870e-01 2.06185162e-01 7.17550457e-01
-1.56822592e-01 -4.26237911e-01 -7.73007333e-01 1.41021597e+00
2.85354465e-01 -1.63327134e+00 3.59123409e-01 -1.27954841e-01
1.82517707e-01 -6.46149278e-01 -2.03254953e-01 1.02338232e-01
-1.92789435e-01 -8.48413110e-01 4.49708819e-01 6.10198200e-01
8.08232903e-01 -1.09838843e+00 7.63450980e-01 1.35835052e-01
-1.49307466e+00 -1.04303092e-01 -4.02514189e-01 1.83518976e-01
1.41240492e-01 5.12661040e-01 -8.82790744e-01 7.64268219e-01
8.63616347e-01 4.63509023e-01 -9.18540537e-01 7.18483865e-01
-1.38827920e-01 1.19633950e-01 -3.30147326e-01 -2.33237028e-01
9.38643217e-02 -2.20543206e-01 7.00918585e-02 1.44942307e+00
1.50025457e-01 1.39764339e-01 -1.90109119e-01 4.44053560e-01
-1.21629886e-01 2.54723698e-01 -1.93569914e-01 6.53797016e-02
6.51235163e-01 1.48151708e+00 -1.16343307e+00 -6.13798678e-01
-4.19967949e-01 1.20621407e+00 -3.67817469e-02 -3.42178121e-02
-7.77498245e-01 -1.14810526e+00 5.53788543e-02 8.45058486e-02
4.50661421e-01 -1.57564163e-01 -6.49124324e-01 -9.27720189e-01
3.48228395e-01 -7.21599638e-01 2.37571612e-01 -1.08697200e+00
-5.47016561e-01 4.51023817e-01 -5.64012825e-02 -7.80155122e-01
1.31615505e-01 -1.08742893e+00 -1.16884187e-01 7.21136808e-01
-1.21649611e+00 -1.22306573e+00 -5.48663199e-01 5.74162483e-01
5.26161909e-01 -3.24956775e-01 5.63823462e-01 6.03181779e-01
-1.05472505e+00 6.76034033e-01 4.05266911e-01 2.73778170e-01
6.19735241e-01 -1.05110586e+00 3.59574258e-01 1.05027127e+00
6.41588569e-02 6.91896617e-01 8.73117223e-02 -7.62293935e-01
-1.87357891e+00 -1.09407842e+00 6.21065736e-01 -4.09584135e-01
2.50757962e-01 -6.32128596e-01 -7.30686069e-01 5.07205188e-01
1.54202536e-01 -3.60445499e-01 2.06561029e-01 -2.50783205e-01
-7.51078725e-02 -4.55643833e-01 -1.15582061e+00 6.61406100e-01
1.16272366e+00 -3.11976194e-01 -5.04923344e-01 3.61222327e-01
7.35755980e-01 -5.25053859e-01 -9.14483964e-01 3.22840303e-01
5.80490947e-01 -6.41671836e-01 1.03041768e+00 7.57709891e-02
1.79578468e-01 -7.40517855e-01 -2.70441979e-01 -3.42350721e-01
-3.60025316e-01 -1.73796460e-01 -4.48611319e-01 1.65139890e+00
4.37899202e-01 -2.80409575e-01 8.92711699e-01 2.34432057e-01
-2.22748056e-01 -7.29058564e-01 -4.58389193e-01 -5.12415409e-01
-6.06231987e-01 -5.41039884e-01 8.46253037e-01 4.10157084e-01
-9.28623676e-02 7.05746651e-01 4.07413878e-02 2.33040333e-01
3.36931556e-01 1.75407872e-01 6.15090013e-01 -8.95698845e-01
-1.27977014e-01 -3.64728093e-01 -4.23056096e-01 -1.39225423e+00
-4.13677603e-01 -1.06110609e+00 -3.63395661e-01 -2.15486598e+00
3.00729573e-01 -1.85488820e-01 -4.89147194e-02 5.05666554e-01
6.84827268e-02 5.97787984e-02 2.62265146e-01 1.47346124e-01
-7.94135094e-01 1.18842483e-01 1.21358287e+00 -3.81052732e-01
-2.05587715e-01 -5.97078681e-01 -6.48519278e-01 6.18654847e-01
5.41442692e-01 -4.85906489e-02 -7.01409996e-01 -7.32902646e-01
3.21838170e-01 -1.23803578e-01 5.09047545e-02 -1.11839390e+00
6.18534207e-01 2.54778743e-01 1.14532924e+00 -1.53182042e+00
1.88947544e-01 -9.08459663e-01 -5.32455705e-02 4.51199412e-01
1.52645260e-01 3.30193967e-01 5.45656085e-01 2.31937692e-01
-1.26690120e-01 -4.78750914e-01 3.20812136e-01 -7.07550421e-02
-1.35257828e+00 8.64747241e-02 -1.63778499e-01 -2.28285432e-01
9.91349518e-01 -4.11916375e-01 -8.06689560e-01 3.87818277e-01
-3.60487103e-01 4.61779863e-01 1.72026515e-01 2.54928291e-01
7.89448678e-01 -1.11816013e+00 -1.38872847e-01 4.31113571e-01
-5.42640360e-03 3.17146093e-01 5.27294159e-01 4.62764174e-01
-1.27249217e+00 8.52372766e-01 -2.63720602e-01 -3.87563258e-01
-1.44207704e+00 8.08386266e-01 -2.57282078e-01 -2.58180439e-01
-6.42634928e-01 6.62441194e-01 -1.56607807e-01 -3.92706990e-01
2.52150089e-01 -4.57274050e-01 -3.45538676e-01 1.11172006e-01
6.61230147e-01 6.32220685e-01 3.96494627e-01 -3.81491929e-01
-7.10657358e-01 8.09062600e-01 -4.37612563e-01 6.65572360e-02
1.06320643e+00 -2.58741111e-01 -3.91884565e-01 -1.61080122e-01
1.27490711e+00 -3.08674425e-02 -6.74420655e-01 1.31633893e-01
5.20014949e-02 -2.08225295e-01 2.65663594e-01 -1.13932002e+00
-9.34230924e-01 6.86361730e-01 7.74018347e-01 -1.83426321e-01
1.48053241e+00 -2.98757344e-01 9.12023902e-01 6.74608052e-01
1.32009596e-01 -1.49396491e+00 -2.14097112e-01 4.22964126e-01
5.33063293e-01 -9.58687663e-01 5.24425924e-01 -6.03671908e-01
-4.51131254e-01 1.20509934e+00 8.88435304e-01 2.44930252e-01
5.57989895e-01 7.44496882e-01 -3.11846554e-01 -3.16449344e-01
-5.12592554e-01 -9.08649340e-02 3.26626718e-01 4.54899490e-01
1.90522745e-01 -9.81332138e-02 -4.76650327e-01 7.55744636e-01
-2.16502413e-01 2.44869236e-02 1.90204576e-01 1.20352995e+00
-5.55538297e-01 -1.17842698e+00 -4.83186394e-01 6.84931338e-01
-1.29362673e-01 -2.44171515e-01 -2.01343119e-01 8.34686279e-01
4.91392851e-01 8.51243198e-01 3.19564641e-01 -6.69663250e-01
6.09548092e-01 7.70504167e-03 4.88541901e-01 -2.11263210e-01
-3.46989095e-01 1.30509272e-01 7.96879455e-02 -6.21427774e-01
8.78615491e-03 -3.74191672e-01 -1.70856106e+00 -5.44969738e-01
-4.01496708e-01 -9.56854746e-02 8.73174727e-01 6.46630824e-01
7.00972259e-01 1.08188629e+00 9.90812555e-02 -3.14337611e-02
8.52412209e-02 -7.83225060e-01 -5.98828077e-01 -7.26131722e-02
-4.03122663e-01 -2.89319426e-01 3.34271520e-01 2.41912052e-01] | [11.693243026733398, 2.682088613510132] |
90bb23f9-56a2-4622-9727-f3637f5b79a6 | temporal-context-aggregation-network-for | 2103.13141 | null | https://arxiv.org/abs/2103.13141v1 | https://arxiv.org/pdf/2103.13141v1.pdf | Temporal Context Aggregation Network for Temporal Action Proposal Refinement | Temporal action proposal generation aims to estimate temporal intervals of actions in untrimmed videos, which is a challenging yet important task in the video understanding field. The proposals generated by current methods still suffer from inaccurate temporal boundaries and inferior confidence used for retrieval owing to the lack of efficient temporal modeling and effective boundary context utilization. In this paper, we propose Temporal Context Aggregation Network (TCANet) to generate high-quality action proposals through "local and global" temporal context aggregation and complementary as well as progressive boundary refinement. Specifically, we first design a Local-Global Temporal Encoder (LGTE), which adopts the channel grouping strategy to efficiently encode both "local and global" temporal inter-dependencies. Furthermore, both the boundary and internal context of proposals are adopted for frame-level and segment-level boundary regressions, respectively. Temporal Boundary Regressor (TBR) is designed to combine these two regression granularities in an end-to-end fashion, which achieves the precise boundaries and reliable confidence of proposals through progressive refinement. Extensive experiments are conducted on three challenging datasets: HACS, ActivityNet-v1.3, and THUMOS-14, where TCANet can generate proposals with high precision and recall. By combining with the existing action classifier, TCANet can obtain remarkable temporal action detection performance compared with other methods. Not surprisingly, the proposed TCANet won the 1$^{st}$ place in the CVPR 2020 - HACS challenge leaderboard on temporal action localization task. | ['Nong Sang', 'Changxin Gao', 'Junjie Yan', 'Yu Qiao', 'Xiang Wang', 'Wei Wu', 'Dongliang Wang', 'Weihao Gan', 'Haisheng Su', 'Zhiwu Qing'] | 2021-03-24 | null | http://openaccess.thecvf.com//content/CVPR2021/html/Qing_Temporal_Context_Aggregation_Network_for_Temporal_Action_Proposal_Refinement_CVPR_2021_paper.html | http://openaccess.thecvf.com//content/CVPR2021/papers/Qing_Temporal_Context_Aggregation_Network_for_Temporal_Action_Proposal_Refinement_CVPR_2021_paper.pdf | cvpr-2021-1 | ['temporal-action-proposal-generation'] | ['computer-vision'] | [ 2.89782047e-01 -2.97226399e-01 -6.51701689e-01 -3.12522918e-01
-8.51481378e-01 -6.45891726e-02 5.41832447e-01 -1.76377907e-01
-4.44045454e-01 7.15980828e-01 5.24012029e-01 1.13541491e-01
1.65858865e-02 -3.03757548e-01 -4.19727236e-01 -5.44441521e-01
-2.44062722e-01 -5.82127050e-02 8.73802900e-01 9.61082354e-02
1.53767109e-01 2.02586371e-02 -1.39825785e+00 5.07658482e-01
7.88265347e-01 1.53652120e+00 2.18010515e-01 5.76780319e-01
4.21678908e-02 1.11936116e+00 -3.82991016e-01 -1.12831309e-01
2.74866134e-01 -5.67610204e-01 -6.58772051e-01 5.53988665e-02
4.28866923e-01 -6.44138277e-01 -4.64704037e-01 6.42596781e-01
3.90462726e-01 2.97420710e-01 3.60857397e-01 -1.18604815e+00
-1.74369007e-01 5.34059942e-01 -7.69278288e-01 5.84571362e-01
3.76484722e-01 4.86867458e-01 1.08493721e+00 -8.60923588e-01
7.69842148e-01 1.07277381e+00 4.99175429e-01 5.98855972e-01
-7.63329685e-01 -7.62684226e-01 6.96039438e-01 6.40070200e-01
-1.52761674e+00 -4.82472688e-01 5.87852061e-01 -2.95138448e-01
1.01381373e+00 4.50632945e-02 7.10220993e-01 1.19319582e+00
1.30184069e-01 1.28873110e+00 7.50593483e-01 -3.69993858e-02
2.48649999e-01 -5.39490283e-01 -3.78423572e-01 6.09351933e-01
-4.84055370e-01 1.63996607e-01 -8.58585119e-01 2.00403258e-01
8.98797393e-01 -1.23626525e-02 -4.48950917e-01 -2.08570957e-02
-1.53610849e+00 4.10164118e-01 3.10429931e-01 2.18604907e-01
-4.93328363e-01 5.24157822e-01 6.26092553e-01 2.15787608e-02
4.14503723e-01 -1.79947633e-02 -4.57284331e-01 -6.27040803e-01
-1.25684571e+00 1.97026163e-01 1.89980164e-01 1.13573515e+00
2.27947637e-01 4.28270735e-02 -9.11515296e-01 6.89744353e-01
2.70536423e-01 3.08210701e-01 3.67917478e-01 -1.23284566e+00
6.68524981e-01 4.16209847e-01 2.80770868e-01 -9.68166530e-01
-1.99924797e-01 -3.32843989e-01 -6.53840840e-01 -3.05056661e-01
2.28705421e-01 1.11505307e-01 -1.14425421e+00 1.86181808e+00
4.06869203e-01 9.26613688e-01 -2.67184705e-01 1.11449158e+00
6.67234540e-01 8.05456936e-01 5.21332979e-01 -5.56609869e-01
1.23875880e+00 -1.17472410e+00 -8.93637300e-01 -1.39200374e-01
7.57097304e-01 -6.18310869e-01 7.32901037e-01 4.13561732e-01
-1.24455976e+00 -7.74322510e-01 -1.04905498e+00 -5.19743674e-02
2.41599888e-01 3.78962517e-01 5.08555889e-01 -3.28988843e-02
-9.00065303e-01 4.55254257e-01 -1.02632654e+00 -6.77809119e-02
7.11333632e-01 2.43945673e-01 -1.67836279e-01 -1.41776711e-01
-1.27645421e+00 4.47224945e-01 4.79403794e-01 3.05052519e-01
-1.01423991e+00 -6.79360986e-01 -8.03631008e-01 -9.87286791e-02
8.46719742e-01 -4.55937803e-01 1.33363795e+00 -8.52136314e-01
-1.31470311e+00 2.96665728e-01 -4.01275158e-01 -8.04787099e-01
6.61795735e-01 -3.22424769e-01 -5.42833686e-01 5.01957119e-01
3.36774498e-01 1.17992306e+00 9.08582151e-01 -6.28973484e-01
-1.20126951e+00 -4.38870303e-02 -2.92010847e-02 3.19203377e-01
-1.34617314e-01 3.14439051e-02 -1.09793866e+00 -1.01340389e+00
2.76964933e-01 -7.25750148e-01 -3.75855088e-01 1.47746563e-01
-5.79653047e-02 -4.49650377e-01 8.51140559e-01 -7.49074519e-01
1.78004217e+00 -2.14867401e+00 2.38783211e-01 -1.39638027e-02
1.71153978e-01 3.66841227e-01 -1.55298620e-01 -9.38606635e-02
2.31190905e-01 -9.45607759e-03 -1.48977488e-01 -3.53654385e-01
-6.48777783e-02 1.85862705e-02 -2.38839269e-01 2.77501792e-01
2.61025369e-01 9.59233820e-01 -1.06425583e+00 -8.90968204e-01
3.56988907e-01 3.50918591e-01 -6.30774736e-01 1.06634945e-01
-4.75061655e-01 5.10845542e-01 -7.07871795e-01 8.44606459e-01
3.49893779e-01 -2.25708306e-01 9.49927792e-02 -4.39766437e-01
-1.10444181e-01 3.55315626e-01 -1.17221498e+00 1.94880092e+00
-2.11847037e-01 5.38323641e-01 -3.07293594e-01 -7.02059507e-01
7.36691594e-01 4.54702348e-01 7.91925430e-01 -1.10850513e+00
-9.73813329e-03 6.88670352e-02 -2.30027825e-01 -4.07761782e-01
6.28425479e-01 1.20826960e-01 -9.71023589e-02 1.71844617e-01
-2.08880808e-02 4.87065434e-01 3.83329839e-01 2.73840904e-01
1.36703837e+00 7.33977377e-01 2.62024343e-01 1.83870658e-01
7.25507975e-01 -2.07410425e-01 1.22927523e+00 4.18849140e-01
-6.38612151e-01 7.45379806e-01 5.18123090e-01 -4.70386237e-01
-7.45689511e-01 -8.77066016e-01 1.54864654e-01 9.67664719e-01
3.88843775e-01 -7.16807425e-01 -4.86871630e-01 -1.03247619e+00
-2.64990389e-01 3.99902195e-01 -5.61071277e-01 -1.04079485e-01
-9.05039549e-01 -2.14052886e-01 4.96652871e-01 9.91362929e-01
1.01432192e+00 -1.27901518e+00 -6.86840057e-01 5.80168724e-01
-6.71268642e-01 -1.53421271e+00 -9.76039648e-01 -3.28788161e-01
-8.41343999e-01 -1.03484082e+00 -7.08329082e-01 -3.34505856e-01
2.00467929e-01 2.87466913e-01 8.70460808e-01 9.28892046e-02
-1.90674514e-01 1.56560645e-01 -6.77409947e-01 9.40671042e-02
8.20963308e-02 1.82995643e-03 -1.35069773e-01 1.32105529e-01
3.50001186e-01 -4.50143188e-01 -1.08081460e+00 7.89730966e-01
-7.80340612e-01 2.21350402e-01 5.84754646e-01 7.01694250e-01
7.86813259e-01 -1.95478737e-01 7.55270123e-01 -3.22831392e-01
1.41579688e-01 -2.63747454e-01 -5.53033233e-01 3.61982822e-01
-4.39124465e-01 -1.73152655e-01 3.01374227e-01 -3.44167441e-01
-1.22771573e+00 -4.50731032e-02 -7.67543018e-02 -7.77066171e-01
4.05054949e-02 3.79618824e-01 -6.25117421e-02 3.01279992e-01
3.35663944e-01 3.25176656e-01 -3.46281379e-01 -2.83560842e-01
2.51186550e-01 3.37804615e-01 6.97529614e-01 -4.97278959e-01
3.00506383e-01 4.99138594e-01 -1.98759153e-01 -4.52808529e-01
-7.47577727e-01 -6.76196516e-01 -5.07982969e-01 -4.93042409e-01
1.10687172e+00 -1.15382230e+00 -6.97796881e-01 4.56433117e-01
-1.06244278e+00 -4.43221450e-01 -2.01429442e-01 6.93123281e-01
-6.62196696e-01 5.38740516e-01 -6.57586038e-01 -7.23812759e-01
-3.50599468e-01 -1.18340206e+00 1.22534359e+00 2.77403921e-01
-1.51068524e-01 -4.77313846e-01 -2.16217116e-01 4.66540426e-01
2.81680912e-01 1.89583525e-01 3.00099641e-01 -1.43980030e-02
-1.05839157e+00 6.28361031e-02 -4.36502069e-01 2.48447821e-01
-7.71278888e-02 5.35166673e-02 -6.34379447e-01 -1.35844931e-01
-4.39192653e-01 -3.36932719e-01 1.19137800e+00 6.47517383e-01
1.27293980e+00 -1.22154474e-01 -4.05266762e-01 5.45476258e-01
1.02399290e+00 4.81762707e-01 8.52471530e-01 7.58353695e-02
6.16519570e-01 2.54110634e-01 1.40281951e+00 7.24945843e-01
3.36596906e-01 1.15417826e+00 3.17486793e-01 1.83096901e-01
-2.64355719e-01 -3.32915306e-01 5.82157314e-01 6.06575847e-01
-3.97408426e-01 -3.02954197e-01 -6.05459392e-01 6.57543540e-01
-2.26099849e+00 -1.18395650e+00 6.92352355e-02 2.17162037e+00
8.05904746e-01 4.75419521e-01 1.90861315e-01 -4.14452925e-02
6.07570946e-01 4.83692020e-01 -4.78361607e-01 4.58017364e-02
-2.21921746e-02 7.99098238e-02 3.47914636e-01 2.01081112e-01
-1.40108156e+00 1.18977880e+00 4.88726568e+00 1.29931509e+00
-9.30716336e-01 1.85569972e-01 6.22907937e-01 -2.34751329e-01
7.10087046e-02 4.48238887e-02 -8.93448472e-01 6.54197156e-01
7.67921984e-01 2.23801225e-01 -2.18345597e-02 6.54300869e-01
6.43007994e-01 -3.95019531e-01 -9.53576684e-01 1.01444650e+00
-9.79578644e-02 -1.50234056e+00 -1.94266990e-01 -1.58788428e-01
6.90370381e-01 -1.15790866e-01 -7.68067315e-02 5.12799621e-01
-6.21253252e-02 -8.21112573e-01 9.68952417e-01 6.40252590e-01
8.67424846e-01 -6.51274741e-01 7.06361711e-01 4.85257879e-02
-1.86277258e+00 -3.28892112e-01 6.59170449e-02 2.70272762e-01
7.08962381e-01 2.51464576e-01 -6.46389782e-01 5.52533150e-01
7.69836247e-01 1.20664716e+00 -3.80100250e-01 1.13527489e+00
-2.65744269e-01 6.99163854e-01 -2.56174654e-01 1.93706572e-01
5.17916679e-01 5.72053641e-02 4.29282010e-01 1.15224266e+00
3.15561116e-01 5.06763458e-01 3.74932110e-01 6.12610996e-01
9.08500794e-03 -1.12272896e-01 -2.29219086e-02 3.15523595e-02
4.62141633e-01 1.01210141e+00 -8.16589355e-01 -5.54953337e-01
-3.49218428e-01 9.90652442e-01 -9.19334870e-03 3.81231934e-01
-1.46552432e+00 3.58469933e-02 5.38500547e-01 7.34530985e-02
5.74962139e-01 -2.31564075e-01 6.59457594e-02 -1.08968651e+00
2.32935891e-01 -7.93466985e-01 5.35716474e-01 -8.10232937e-01
-7.53173769e-01 3.96282792e-01 2.09746242e-01 -1.73512757e+00
-1.25217050e-01 -6.85859323e-02 -3.75146270e-01 4.49719340e-01
-1.34226072e+00 -1.18953919e+00 -3.43107343e-01 6.12533569e-01
1.06628728e+00 8.19178149e-02 2.18232110e-01 6.12474144e-01
-6.50864065e-01 5.93826115e-01 -6.16650343e-01 4.76778112e-02
7.32044816e-01 -8.75740349e-01 2.89866805e-01 9.32219625e-01
5.52685224e-02 1.27123743e-01 4.00920212e-01 -8.67062569e-01
-8.84494007e-01 -1.33830106e+00 8.89009595e-01 -2.02723891e-01
3.90937358e-01 -3.11138909e-02 -7.76481569e-01 4.76862848e-01
-7.80618042e-02 3.48121196e-01 1.91324711e-01 -2.67919451e-01
-1.33786485e-01 -2.25701243e-01 -7.12942421e-01 7.22906530e-01
1.44242537e+00 -3.80973160e-01 -3.26472580e-01 1.43077031e-01
8.86590183e-01 -4.49533492e-01 -8.62447798e-01 7.95675993e-01
7.63148367e-01 -1.01406062e+00 8.83462906e-01 -4.26064193e-01
6.31844461e-01 -5.99070072e-01 -1.76620305e-01 -5.12135804e-01
-2.00826705e-01 -7.53603458e-01 -4.58077997e-01 1.15628052e+00
2.74905711e-01 9.29251872e-03 8.90838325e-01 4.13557589e-01
-4.18448061e-01 -1.15987790e+00 -1.15792811e+00 -6.66102171e-01
-7.06977308e-01 -8.04736674e-01 3.21050674e-01 5.69257021e-01
-1.44620180e-01 8.87029544e-02 -6.96316302e-01 -9.22355801e-02
2.88288474e-01 -1.23179071e-01 6.57322705e-01 -5.81422150e-01
-2.89213300e-01 -4.61028993e-01 -5.68411469e-01 -1.64256501e+00
-1.02889240e-01 -4.16501880e-01 2.70478249e-01 -1.32845664e+00
3.85217033e-02 -5.65265477e-01 -5.90900898e-01 5.08532822e-01
-1.87962219e-01 2.86770076e-01 3.15551758e-01 1.79024652e-01
-1.34585726e+00 8.30938697e-01 1.36914241e+00 -1.93419114e-01
-3.77007186e-01 4.04017279e-03 -7.86961522e-03 6.31947756e-01
4.46590155e-01 -2.76695609e-01 -6.25647962e-01 -2.20607981e-01
3.43238842e-03 5.13493121e-01 4.33545530e-01 -1.22512925e+00
2.61422813e-01 -3.82751942e-01 3.20193946e-01 -9.84715164e-01
5.45773208e-01 -5.22591829e-01 5.63621633e-02 1.63287729e-01
-3.95845950e-01 -2.26352271e-02 6.11786284e-02 9.74804401e-01
-3.53985935e-01 3.29724163e-01 5.55347264e-01 3.41426097e-02
-1.29408705e+00 7.13505685e-01 -1.45329699e-01 1.15981512e-01
1.18251812e+00 -3.17920029e-01 -2.04332292e-01 -3.01644564e-01
-7.90302157e-01 6.02171719e-01 9.96866226e-02 6.00061536e-01
9.35181081e-01 -1.50171769e+00 -6.33319378e-01 -4.46140505e-02
2.51006931e-01 -6.07057437e-02 7.03075886e-01 1.46634531e+00
-4.06865329e-01 4.05357659e-01 -7.25895464e-02 -8.20011377e-01
-1.18591368e+00 3.81905884e-01 2.22661138e-01 -4.88827318e-01
-7.43352175e-01 1.03181493e+00 2.31526867e-01 4.49426383e-01
3.99353176e-01 -5.72373211e-01 -2.47236118e-01 2.15978008e-02
6.76234126e-01 2.78193235e-01 -2.29470059e-01 -6.95050180e-01
-3.65569204e-01 4.95708108e-01 -2.44262189e-01 -1.44654483e-01
8.47825289e-01 -2.36162186e-01 3.74673188e-01 1.01025902e-01
9.44260836e-01 -4.66402799e-01 -1.79010880e+00 -3.87862623e-01
9.05471072e-02 -6.77839339e-01 8.71740505e-02 -6.76654994e-01
-1.21961784e+00 7.73530602e-01 6.30055845e-01 -2.83188671e-01
1.34283209e+00 -1.42352298e-01 1.26140201e+00 -7.28867948e-02
4.72244263e-01 -1.42875063e+00 4.19967502e-01 2.72628874e-01
7.88941920e-01 -1.15909386e+00 1.66283324e-01 -4.28953856e-01
-7.92483091e-01 8.26900542e-01 9.02192473e-01 2.06824005e-01
4.33419585e-01 3.51135135e-02 -3.53675544e-01 -1.80491041e-02
-1.18033075e+00 -3.52533758e-01 5.34943640e-01 2.81166136e-01
2.50552118e-01 -1.88898779e-02 -6.64044082e-01 5.68666399e-01
5.64087927e-01 4.59677726e-01 8.57601017e-02 9.52249527e-01
-4.05306995e-01 -9.72788990e-01 4.21535857e-02 3.28665614e-01
-4.75279629e-01 -4.51485999e-02 -2.91845705e-02 7.81676292e-01
4.70868528e-01 8.02319765e-01 -3.80206096e-04 -5.92737734e-01
2.26673484e-01 -1.85790017e-01 3.69517356e-01 -2.28975430e-01
-3.64886761e-01 5.48821926e-01 2.63777673e-01 -1.18343091e+00
-6.99797869e-01 -7.22917676e-01 -1.52299356e+00 6.06377609e-02
-2.59066701e-01 1.91477984e-02 1.36375129e-01 1.04603910e+00
4.61816043e-01 7.51388371e-01 4.59004611e-01 -7.69771814e-01
-2.31715247e-01 -9.87625718e-01 -2.76513666e-01 3.20412159e-01
1.66588098e-01 -8.95064116e-01 -5.74778430e-02 2.28838965e-01] | [8.414219856262207, 0.48251813650131226] |
938ee548-5f25-4a48-a771-0d2629eaebb7 | automatic-extraction-of-ranked-snp-phenotype | 2012.00902 | null | https://arxiv.org/abs/2012.00902v1 | https://arxiv.org/pdf/2012.00902v1.pdf | Automatic Extraction of Ranked SNP-Phenotype Associations from Literature through Detecting Neural Candidates, Negation and Modality Markers | Genome-wide association (GWA) constitutes a prominent portion of studies which have been conducted on personalized medicine and pharmacogenomics. Recently, very few methods have been developed for extracting mutation-diseases associations. However, there is no available method for extracting the association of SNP-phenotype from text which considers degree of confidence in associations. In this study, first a relation extraction method relying on linguistic-based negation detection and neutral candidates is proposed. The experiments show that negation cues and scope as well as detecting neutral candidates can be employed for implementing a superior relation extraction method which outperforms the kernel-based counterparts due to a uniform innate polarity of sentences and small number of complex sentences in the corpus. Moreover, a modality based approach is proposed to estimate the confidence level of the extracted association which can be used to assess the reliability of the reported association. Keywords: SNP, Phenotype, Biomedical Relation Extraction, Negation Detection. | ['Alberto Diaz', 'Behrouz Bokharaeian'] | 2020-12-02 | null | null | null | null | ['negation-detection'] | ['natural-language-processing'] | [ 4.04542863e-01 2.41282836e-01 -4.11067575e-01 -2.98683584e-01
-6.62457407e-01 -3.79688650e-01 4.37353700e-01 7.26960599e-01
-1.95182547e-01 1.21808982e+00 3.04485768e-01 -4.12200719e-01
-4.27372426e-01 -7.58841097e-01 -3.81053388e-01 -5.51643252e-01
-1.66499615e-02 5.42779826e-02 9.32583436e-02 -1.83401719e-01
2.23615140e-01 3.47378477e-02 -1.30051672e+00 2.25589871e-01
1.32434976e+00 5.24767518e-01 -8.61402825e-02 3.82717848e-01
-1.17152266e-01 1.24652483e-01 -7.14163661e-01 -6.34189487e-01
-2.98700422e-01 -7.76919663e-01 -4.49619681e-01 -6.80486143e-01
-1.91679597e-02 8.61097574e-02 2.78840184e-01 1.21307838e+00
8.65842521e-01 -3.10403407e-01 7.90034711e-01 -8.63736808e-01
-5.35755098e-01 9.39148128e-01 -6.06864989e-01 1.32718205e-01
8.75371814e-01 -2.72888243e-02 1.03756928e+00 -5.05758166e-01
8.20999146e-01 9.56919074e-01 5.56178570e-01 4.10063356e-01
-9.25423324e-01 -4.85312104e-01 -2.08075896e-01 1.34180814e-01
-1.30409241e+00 -1.82633162e-01 6.29269302e-01 -3.54832917e-01
1.08122468e+00 5.30972660e-01 7.13083148e-01 9.26680684e-01
3.44456315e-01 2.31281430e-01 1.20745897e+00 -8.79864812e-01
2.21391067e-01 3.94600421e-01 2.04229981e-01 8.13924074e-01
6.46707416e-01 -7.09430501e-02 -8.01375687e-01 -6.85045660e-01
-1.12652786e-01 -7.20716715e-01 -2.69088268e-01 2.74825096e-01
-9.82955933e-01 4.87919033e-01 -1.83162391e-01 5.76902866e-01
-6.14066601e-01 -3.58645767e-01 6.69018745e-01 4.09401618e-02
4.76908922e-01 2.61416793e-01 -6.81340396e-01 4.81573418e-02
-8.36652994e-01 2.65180320e-01 8.19278598e-01 8.05595160e-01
2.43677706e-01 -2.97114134e-01 -5.67787349e-01 5.03082216e-01
3.90106589e-01 4.72367167e-01 6.44918680e-01 -2.21949100e-01
1.64232343e-01 7.28767991e-01 -1.40253425e-01 -1.15217078e+00
-4.28146124e-01 -3.27114999e-01 -6.37595892e-01 -2.30658665e-01
-1.77745502e-02 -4.41714346e-01 -4.88824695e-01 1.92259598e+00
7.12104857e-01 2.92588949e-01 4.09133613e-01 4.99282926e-01
1.07398891e+00 2.31519401e-01 4.13217157e-01 -5.14969766e-01
1.99799836e+00 -9.96616036e-02 -1.20053506e+00 2.43276611e-01
7.23141491e-01 -7.97805786e-01 6.06368959e-01 3.51629645e-01
-8.45518112e-01 -8.27073604e-02 -1.14355564e+00 1.63435452e-02
-6.28742695e-01 4.17436868e-01 8.04486394e-01 1.01645899e+00
-4.69482005e-01 2.51135141e-01 -4.07229930e-01 -5.45383394e-01
2.96838939e-01 4.17617619e-01 -4.03450251e-01 3.21619719e-01
-1.61976373e+00 1.14883804e+00 7.46189475e-01 2.12439612e-01
2.73183151e-03 -6.95300877e-01 -5.85847378e-01 8.10137577e-03
2.36554161e-01 -8.96455348e-01 3.53199899e-01 -5.05265772e-01
-1.38535559e+00 7.99271703e-01 -2.51466274e-01 -4.45673317e-01
1.49001509e-01 -1.61943614e-01 -9.11191344e-01 1.11988634e-01
2.53371090e-01 2.22166419e-01 2.42786527e-01 -4.34333712e-01
-3.86843741e-01 -5.63016951e-01 -1.73387304e-01 -8.21587443e-03
-3.51294637e-01 3.04682553e-01 -1.07501343e-01 -5.63206375e-01
-1.72484979e-01 -6.33064449e-01 -3.93249057e-02 -2.37054884e-01
-8.56150687e-01 -4.35732245e-01 4.62905437e-01 -9.10013914e-01
1.56850123e+00 -2.06001091e+00 -1.94897205e-01 4.34473842e-01
-9.80019867e-02 4.76490229e-01 8.96758437e-02 5.76573372e-01
-1.18755184e-01 4.28954840e-01 -1.82718948e-01 5.84291399e-01
-2.73795966e-02 -3.62276472e-02 2.17006970e-02 3.70499820e-01
5.92684507e-01 8.25937212e-01 -7.57121801e-01 -6.80640399e-01
-1.99572951e-01 4.82102454e-01 -5.93959928e-01 -3.38412710e-02
-3.16140413e-01 3.07371318e-01 -7.95652449e-01 7.60750890e-01
7.11288929e-01 1.92809165e-01 7.89746106e-01 -6.63966835e-01
-1.32491723e-01 2.60110050e-01 -8.99254620e-01 1.44677031e+00
-7.27094486e-02 1.49996772e-01 -2.18188420e-01 -9.27762270e-01
9.15681779e-01 5.22139430e-01 1.04334474e-01 -2.38542452e-01
1.63860545e-01 2.31787294e-01 3.42908174e-01 -1.13691270e+00
2.48145372e-01 -2.70101368e-01 3.68216783e-02 3.68806301e-03
-1.12848610e-01 2.88704515e-01 2.93340147e-01 1.82401806e-01
1.12707591e+00 5.14017045e-01 9.07240570e-01 -2.46599555e-01
9.04386520e-01 -5.08944085e-03 8.76850665e-01 3.77102792e-01
-7.67889619e-02 4.73049544e-02 7.51249731e-01 6.06502481e-02
-5.57049870e-01 -5.33865809e-01 -3.04359645e-01 4.46966052e-01
-2.22013488e-01 -5.56037545e-01 -6.38248980e-01 -6.52798235e-01
-9.28238854e-02 5.86562276e-01 -5.89206934e-01 -1.15984585e-02
-3.29933912e-01 -1.48070741e+00 9.32929158e-01 9.34743583e-02
2.83011109e-01 -7.17450500e-01 -3.95429283e-01 1.60945542e-02
-2.14135647e-01 -9.92803454e-01 2.70404443e-02 2.15730682e-01
-6.92048550e-01 -1.28398788e+00 -2.38817051e-01 -4.72504199e-01
7.63301492e-01 -7.28712797e-01 6.35060132e-01 -1.73470065e-01
-4.16749358e-01 -2.51451015e-01 -3.67247194e-01 -6.65592492e-01
-4.51605082e-01 -2.54285455e-01 1.65163027e-03 -3.47740263e-01
7.88837790e-01 -5.06533027e-01 -5.07397413e-01 -4.37086225e-01
-8.15877736e-01 -3.25373888e-01 1.07097101e+00 7.53090501e-01
2.77723432e-01 -6.99587241e-02 1.29906058e+00 -1.33282483e+00
1.09322572e+00 -6.17385805e-01 -3.44361424e-01 4.95074809e-01
-7.55017042e-01 2.04673097e-01 4.42434669e-01 -2.94536710e-01
-1.16620302e+00 -1.28117904e-01 -4.00874972e-01 6.52534425e-01
-3.94129962e-01 1.01553869e+00 -3.87356132e-01 2.06944853e-01
5.13809621e-01 -2.44146623e-02 6.80133142e-03 -2.24247530e-01
4.46601450e-01 7.04994023e-01 2.38541156e-01 -4.33769137e-01
2.92538077e-01 1.13643058e-01 3.59991133e-01 -7.47417986e-01
-6.21886373e-01 -3.62363666e-01 -4.14735556e-01 1.99841931e-01
9.90720153e-01 -6.86916530e-01 -9.64904130e-01 -1.07226245e-01
-1.06622684e+00 5.26934922e-01 2.35350490e-01 9.17477846e-01
2.21605189e-02 7.22011447e-01 -3.37593198e-01 -9.87298727e-01
-7.98753321e-01 -6.90102100e-01 6.98587537e-01 9.77586284e-02
-6.05970919e-01 -7.13359296e-01 3.45006049e-01 2.61577129e-01
1.63970143e-02 4.14516419e-01 1.55757153e+00 -8.79456758e-01
6.68435032e-03 -5.08883774e-01 -1.20847091e-01 4.25902419e-02
3.55007678e-01 1.80412382e-01 -7.64544725e-01 3.92067283e-01
-2.34535217e-01 -6.36960045e-02 5.17100453e-01 3.71180356e-01
4.93105263e-01 -3.15059125e-01 -5.11172116e-01 4.53652814e-02
1.41696405e+00 1.06916860e-01 5.94775200e-01 1.03867918e-01
3.18741530e-01 6.56566083e-01 9.09077585e-01 3.81910563e-01
4.28083539e-02 6.68174744e-01 -1.44038498e-01 1.97944224e-01
-6.20574318e-02 6.44501522e-02 4.20000941e-01 7.24909186e-01
-1.68849513e-01 -3.82134080e-01 -6.21031702e-01 3.23039174e-01
-1.66150880e+00 -8.32182586e-01 -4.58798051e-01 2.00264263e+00
1.19187915e+00 1.24043308e-01 9.67240892e-03 1.65466100e-01
6.95267498e-01 -2.85334975e-01 -6.85245320e-02 -4.32679236e-01
-4.18534845e-01 3.01402152e-01 1.06932014e-01 2.59977579e-01
-8.29074502e-01 5.55989504e-01 6.51529169e+00 7.86409318e-01
-8.94668400e-01 -6.44025486e-03 3.31428528e-01 3.14102292e-01
-5.62306404e-01 5.87570257e-02 -8.29571426e-01 4.68126774e-01
9.14316297e-01 -1.79953903e-01 -5.27512550e-01 3.91062886e-01
4.09408748e-01 -4.36356753e-01 -9.64141667e-01 4.47180986e-01
2.16959734e-02 -1.10636199e+00 8.53660852e-02 -9.66240652e-03
4.11353052e-01 -5.09262979e-01 -1.63019285e-01 -1.26483813e-01
-1.95867434e-01 -6.32665992e-01 1.61010459e-01 8.03015053e-01
5.90164006e-01 -6.33588910e-01 1.31799889e+00 1.34833708e-01
-8.76211166e-01 2.68154830e-01 -3.00444037e-01 -4.70303297e-02
9.09557417e-02 1.23403323e+00 -1.10996878e+00 1.18896198e+00
3.92379850e-01 2.86667824e-01 -3.66775602e-01 7.38519669e-01
-3.93480897e-01 8.59241605e-01 -3.73235941e-01 -5.37917316e-01
-3.15898299e-01 -3.38819444e-01 5.31172633e-01 1.54725981e+00
5.47389865e-01 2.40859970e-01 -1.51778519e-01 9.53201950e-01
1.38539150e-01 8.20160389e-01 -6.42350316e-01 -2.74336725e-01
5.22525847e-01 1.20128286e+00 -6.51377678e-01 -3.66543114e-01
-5.67753196e-01 5.97606719e-01 1.50628060e-01 2.50435114e-01
-6.14925444e-01 -4.41902339e-01 2.24901468e-01 -1.50995702e-01
1.16434269e-01 3.30599725e-01 -3.84364396e-01 -8.94186258e-01
5.62060326e-02 -9.43818212e-01 5.89187264e-01 -3.47519517e-01
-1.25315642e+00 4.44841921e-01 -9.35697630e-02 -9.76124048e-01
-2.01869413e-01 -4.12614167e-01 -6.10465765e-01 1.07852423e+00
-1.20188498e+00 -1.12777388e+00 2.39054710e-01 -3.11098956e-02
-1.65540710e-01 -1.95555180e-01 1.09615934e+00 3.90793264e-01
-9.89753127e-01 5.66409767e-01 -3.77548248e-01 -9.75274518e-02
8.42178285e-01 -1.01078200e+00 -4.35003340e-01 8.41473818e-01
-4.35944110e-01 1.05506599e+00 8.09567809e-01 -1.21668947e+00
-1.31568205e+00 -7.04127073e-01 1.33039629e+00 -1.40177608e-01
5.38043201e-01 1.83726922e-02 -7.73954451e-01 2.16870815e-01
4.70393747e-01 -2.93511748e-01 1.24024069e+00 2.98095882e-01
-2.21102417e-01 -2.35215828e-01 -1.39188147e+00 6.08434558e-01
5.85359335e-01 -3.98104161e-01 -6.77061379e-01 1.60295844e-01
3.60843539e-01 -1.80141609e-02 -1.07219064e+00 6.98207796e-01
6.73215926e-01 -6.48380756e-01 8.42205584e-01 -6.03582382e-01
2.11445108e-01 -5.86374819e-01 1.36248127e-01 -8.29400957e-01
-8.16202611e-02 -3.70728552e-01 3.23910243e-03 1.61867332e+00
8.16989541e-01 -6.84890270e-01 5.52907944e-01 4.18111950e-01
2.20403835e-01 -8.14850867e-01 -7.46803224e-01 -3.44191194e-01
-4.05409098e-01 -1.94861546e-01 4.80033338e-01 1.10342133e+00
4.83183026e-01 5.00138879e-01 -1.79587439e-01 3.42834920e-01
4.41345721e-01 1.26332343e-01 2.39077076e-01 -1.14060450e+00
-2.01767594e-01 -2.18870074e-01 -5.94689190e-01 1.41493911e-02
3.28893214e-02 -9.03520107e-01 -8.34304839e-02 -1.37106287e+00
4.50476140e-01 -1.83894396e-01 -3.79224390e-01 4.38284576e-01
-5.28273761e-01 1.14052571e-01 -6.55578434e-01 -4.16821986e-01
-5.55245057e-02 3.54862601e-01 7.12946355e-01 7.20323175e-02
-2.19134659e-01 -2.06633687e-01 -8.13727558e-01 6.69318914e-01
9.19894040e-01 -4.49508995e-01 -5.21962762e-01 4.08930093e-01
8.01439464e-01 -2.09317327e-01 1.78552896e-01 -4.53431427e-01
7.65813366e-02 3.33488844e-02 3.15063238e-01 -6.35717034e-01
-2.31304821e-02 -2.89982796e-01 2.89424211e-01 7.05451906e-01
-4.58120078e-01 1.74116753e-02 1.87656671e-01 7.94595599e-01
-1.41267732e-01 -4.73009259e-01 2.46892661e-01 -6.03079144e-03
-2.71206200e-01 -3.63490611e-01 -3.31175804e-01 -7.92938247e-02
9.79041696e-01 3.75194885e-02 -4.09385145e-01 8.91730338e-02
-6.35057747e-01 -1.73993021e-01 -1.44796208e-01 -1.34622604e-01
6.12943769e-01 -9.72474217e-01 -1.02800763e+00 -1.89565495e-01
4.88292687e-02 -6.78178310e-01 3.64397526e-01 1.24094486e+00
-5.03307223e-01 5.49998283e-01 -1.22510707e-02 -6.27315342e-02
-1.63466871e+00 6.83263719e-01 -1.63225174e-01 -1.66118577e-01
-2.68781334e-01 6.77309573e-01 -2.28575066e-01 -9.67138484e-02
6.26219483e-03 -3.38497072e-01 -6.71071172e-01 1.94697887e-01
3.73214841e-01 2.26933956e-01 7.09049180e-02 -3.68798465e-01
-8.83178115e-01 3.97019267e-01 1.22791067e-01 2.48802025e-02
1.07806528e+00 1.43825755e-01 -7.22786725e-01 2.55565822e-01
6.59725845e-01 6.06932759e-01 7.11773336e-02 8.30092579e-02
4.28979903e-01 -1.69030726e-02 -2.19441503e-01 -1.16009653e+00
-3.86828750e-01 3.32881391e-01 5.39059341e-01 -2.92942277e-03
1.05647063e+00 -1.66438799e-02 3.03594232e-01 2.06700087e-01
-4.35143970e-02 -1.06822455e+00 -6.52428746e-01 -1.53455004e-01
6.88353062e-01 -1.07614315e+00 3.56930077e-01 -9.70456302e-01
-4.01935905e-01 1.04201162e+00 4.02730763e-01 4.22493577e-01
5.79333067e-01 6.09865487e-01 6.36875257e-03 -2.85824150e-01
-8.03697944e-01 -1.89760938e-01 5.70476413e-01 6.34683311e-01
1.10111785e+00 2.27565601e-01 -1.51237106e+00 1.02169836e+00
-1.32104501e-01 2.20430374e-01 2.16811016e-01 7.39947557e-01
-1.61249429e-01 -1.43426895e+00 -1.88354239e-01 5.68946183e-01
-1.00185752e+00 -5.75481117e-01 -5.51372647e-01 5.34442961e-01
6.52616918e-01 1.05711257e+00 -6.31378591e-01 -2.40879774e-01
1.30824283e-01 3.79803747e-01 5.82818449e-01 -4.85469490e-01
-6.12535000e-01 2.48206511e-01 8.00339937e-01 -2.80899912e-01
-9.53474104e-01 -4.31140453e-01 -1.20457900e+00 1.80952832e-01
-7.79776335e-01 6.03908479e-01 5.92075467e-01 1.07916379e+00
5.65127432e-01 6.31547511e-01 -6.25087470e-02 2.84889668e-01
-1.13852739e-01 -1.11066830e+00 -4.58839238e-01 -3.72002907e-02
-2.24247485e-01 -3.56428593e-01 1.35452613e-01 -9.08555239e-02] | [8.425073623657227, 8.66662883758545] |
d6715357-88cd-4d3c-a11e-697a66ac451c | covtinet-covid-text-identification-network | null | null | https://link.springer.com/article/10.1007/s00521-023-08442-y | https://link.springer.com/article/10.1007/s00521-023-08442-y | CovTiNet: Covid text identification network using attention-based positional embedding feature fusion | Covid text identification (CTI) is a crucial research concern in natural language processing (NLP). Social and electronic media are simultaneously adding a large volume of Covid-affiliated text on the World Wide Web due to the effortless access to the Internet, electronic gadgets and the Covid outbreak. Most of these texts are uninformative and contain misinformation, disinformation and malinformation that create an infodemic. Thus, Covid text identification is essential for controlling societal distrust and panic. Though very little Covid-related research (such as Covid disinformation, misinformation and fake news) has been reported in high-resource languages (e.g. English), CTI in low-resource languages (like Bengali) is in the preliminary stage to date. However, automatic CTI in Bengali text is challenging due to the deficit of benchmark corpora, complex linguistic constructs, immense verb inflexions and scarcity of NLP tools. On the other hand, the manual processing of Bengali Covid texts is arduous and costly due to their messy or unstructured forms. This research proposes a deep learning-based network (CovTiNet) to identify Covid text in Bengali. The CovTiNet incorporates an attention-based position embedding feature fusion for text-to-feature representation and attention-based CNN for Covid text identification. Experimental results show that the proposed CovTiNet achieved the highest accuracy of 96.61±.001% on the developed dataset (BCovC) compared to the other methods and baselines (i.e. BERT-M, IndicBERT, ELECTRA-Bengali, DistilBERT-M, BiLSTM, DCNN, CNN, LSTM, VDCNN and ACNN). | ['Nazmul Siddique & Iqbal H. Sarker', 'Mohammed Moshiul Hoque', 'Md. Rajib Hossain'] | 2023-03-14 | null | null | null | neural-computing-and-applications-2023-3 | ['misinformation'] | ['miscellaneous'] | [-3.19170713e-01 -3.18557441e-01 1.24772433e-02 1.15640491e-01
-3.77932876e-01 -7.83857286e-01 9.70741749e-01 5.26112914e-01
-5.97118974e-01 8.48387361e-01 6.38389766e-01 -3.45187336e-01
3.69951911e-02 -6.34017527e-01 -2.15940297e-01 -3.04700345e-01
3.49088460e-01 9.03968513e-01 -1.90974429e-01 -6.71395183e-01
5.51034570e-01 3.65044504e-01 -1.11282706e+00 2.98397452e-01
1.07743001e+00 7.27319002e-01 1.01357006e-01 7.40124404e-01
-3.35500658e-01 7.70099640e-01 -8.08154225e-01 -7.58016646e-01
-1.86890900e-01 -2.35766932e-01 -9.33954298e-01 -6.23530447e-01
1.37885824e-01 -2.30320886e-01 -3.26184154e-01 1.15883768e+00
8.44703317e-01 -3.22657585e-01 1.01172733e+00 -1.07201350e+00
-1.10222781e+00 8.06363225e-01 -7.85179973e-01 6.44585431e-01
3.32157671e-01 -3.16596329e-01 7.41341710e-01 -1.02583003e+00
7.09455252e-01 1.63861537e+00 8.52882862e-01 1.86959878e-01
-6.89393163e-01 -8.69511008e-01 -4.21771437e-01 2.68508106e-01
-1.31146157e+00 -3.25143225e-02 5.80059111e-01 -7.63400912e-01
1.28923190e+00 2.81949073e-01 4.74885076e-01 1.39461529e+00
5.69895685e-01 7.07383096e-01 9.19985890e-01 -4.90911990e-01
-3.60213697e-01 4.77396250e-01 5.29252231e-01 4.22875553e-01
5.59069872e-01 -2.10968032e-01 -1.95454076e-01 -8.65296796e-02
7.27963746e-02 -1.00703672e-01 -3.62823755e-02 7.32216537e-01
-1.02197349e+00 1.26054251e+00 2.19399393e-01 8.71859789e-01
-4.20719475e-01 -3.02012205e-01 1.17467999e+00 3.59345078e-01
7.40649343e-01 2.36112252e-01 -5.06284893e-01 -3.08771372e-01
-5.80683887e-01 2.12505430e-01 8.12935948e-01 6.50158703e-01
2.25220695e-01 4.20778573e-01 -4.02386338e-02 1.04298675e+00
2.14387998e-01 9.40178096e-01 1.11600459e+00 2.33880907e-01
9.20277476e-01 7.67236352e-01 8.01537335e-02 -2.08305788e+00
-6.08285010e-01 -7.09251016e-02 -1.27050579e+00 -4.39191997e-01
-2.89327968e-02 -4.48207915e-01 -8.35208595e-01 1.17209268e+00
1.19747862e-01 -2.82086343e-01 1.48389816e-01 7.10372448e-01
1.22094631e+00 1.09907103e+00 -2.63151508e-02 -8.63478035e-02
1.56413579e+00 -4.00558919e-01 -1.05678189e+00 -1.53770000e-01
5.48535228e-01 -1.16434574e+00 9.46811378e-01 3.51091117e-01
-5.25280178e-01 -3.29048157e-01 -9.98790920e-01 -1.32469013e-01
-1.11282670e+00 9.40504521e-02 2.18143627e-01 9.18090641e-01
-6.34063900e-01 -3.82939796e-03 -3.26378644e-01 -5.74162304e-01
5.45158565e-01 2.58872539e-01 -4.45667952e-01 1.22772262e-01
-1.55146682e+00 1.27446544e+00 7.91989028e-01 1.85411870e-01
-6.94108784e-01 -2.01568559e-01 -8.44151676e-01 -3.80460083e-01
1.21632889e-01 8.63882378e-02 7.84260690e-01 -1.06520748e+00
-8.96954238e-01 9.39239562e-01 2.69496024e-01 -4.87046778e-01
5.66705823e-01 -5.20160675e-01 -7.79514074e-01 -2.38372639e-01
2.99383074e-01 5.52815981e-02 9.66054261e-01 -8.42405915e-01
-3.98352474e-01 -4.67720598e-01 -3.55517924e-01 1.50260061e-01
-6.30308151e-01 8.33099961e-01 1.59807593e-01 -8.67899358e-01
-1.52671844e-01 -6.03958428e-01 4.37357396e-01 -8.63568962e-01
-8.08691084e-01 -4.85401988e-01 1.65181112e+00 -1.21042013e+00
1.56052375e+00 -1.93043435e+00 -1.21962667e-01 -1.34897396e-01
4.69050497e-01 8.08118522e-01 4.12677407e-01 8.36327434e-01
4.40563783e-02 6.87327325e-01 9.80245844e-02 7.48276487e-02
-2.73797344e-02 2.12938160e-01 -2.13059470e-01 8.40844750e-01
2.86728684e-02 1.09169352e+00 -7.03691900e-01 -6.75827086e-01
2.56002843e-01 4.63752568e-01 -2.19832793e-01 -6.18909374e-02
5.19447848e-02 9.48610678e-02 -5.46821952e-01 7.92848289e-01
8.31823587e-01 -4.96786572e-02 -1.02215201e-01 -3.25308859e-01
-5.40211916e-01 8.10618550e-02 -9.23809588e-01 6.55157626e-01
-3.73249710e-01 1.11662662e+00 7.01419171e-03 -1.27658641e+00
7.82418787e-01 4.40879673e-01 2.39241477e-02 -4.05457884e-01
8.77406955e-01 2.98493683e-01 -2.61628758e-02 -9.39549625e-01
9.05664086e-01 -2.51539778e-02 -5.01705945e-01 5.03336728e-01
1.39088007e-02 1.96636841e-01 -4.45339158e-02 3.54627252e-01
6.20087028e-01 -5.56878388e-01 5.86699545e-01 -3.48452955e-01
7.72990704e-01 1.06447443e-01 2.61496097e-01 5.15714884e-01
-3.81867886e-01 2.43493840e-01 3.49505186e-01 -5.02484798e-01
-1.23420238e+00 -4.44535524e-01 -3.55185270e-01 9.76068616e-01
-9.97686833e-02 -7.28856176e-02 -5.68950653e-01 -4.95058030e-01
-2.76991222e-02 7.51795113e-01 -6.47020221e-01 7.61300623e-02
-6.70627296e-01 -1.05228913e+00 1.11532450e+00 2.46401221e-01
7.51767993e-01 -1.39196146e+00 -3.72508347e-01 4.16995943e-01
-4.17433023e-01 -9.66649115e-01 -3.65751415e-01 9.40785483e-02
-2.30196938e-01 -7.50183761e-01 -2.56277472e-01 -9.75736618e-01
2.16827810e-01 3.22983339e-02 8.75186205e-01 -1.45226225e-01
-4.24602270e-01 -2.23913535e-01 -6.21755600e-01 -9.73049700e-01
-6.70798719e-01 -7.78290629e-02 1.70586467e-01 -1.41803458e-01
8.21835637e-01 7.83193782e-02 -4.75249859e-03 -1.64512053e-01
-8.56717050e-01 -1.19968392e-01 1.27464071e-01 1.13326466e+00
-6.51937500e-02 1.48000151e-01 7.06057668e-01 -1.02798641e+00
1.14449441e+00 -9.28908408e-01 -3.98950040e-01 -1.46274686e-01
-4.51217532e-01 -4.43524331e-01 7.58904934e-01 -5.66527009e-01
-1.01519382e+00 -7.24898934e-01 -2.34548301e-01 4.87859137e-02
-1.63406000e-01 9.43874598e-01 6.08861968e-02 2.41242632e-01
6.18703902e-01 2.39983171e-01 -1.92592233e-01 -4.23117369e-01
1.78026333e-01 1.54971004e+00 3.33324164e-01 -1.09956190e-01
7.36777604e-01 5.80821000e-02 -8.04106474e-01 -1.17160606e+00
-7.36256361e-01 -5.03244817e-01 -5.83066940e-01 -2.72340953e-01
1.22749579e+00 -7.51857281e-01 -7.29654551e-01 8.31284881e-01
-1.58844233e+00 5.16392887e-01 4.37079936e-01 4.52426434e-01
2.27966204e-01 3.87272745e-01 -7.90546060e-01 -9.40054178e-01
-1.01745474e+00 -9.57554162e-01 6.36909902e-01 -1.64448936e-02
-3.62333149e-01 -1.03852594e+00 1.23201169e-01 4.47714180e-01
6.07620120e-01 4.96436715e-01 1.16887593e+00 -1.41843963e+00
3.45254570e-01 -4.83276993e-01 -6.37941360e-01 4.25847143e-01
1.51345372e-01 -2.10513808e-05 -9.24506009e-01 5.10527529e-02
1.38121679e-01 -7.07022071e-01 4.17593837e-01 2.71136820e-01
4.71682817e-01 -9.86051142e-01 -1.47764251e-01 7.27129728e-02
1.68523252e+00 4.18417066e-01 4.59755182e-01 5.34511626e-01
9.58977699e-01 2.64049232e-01 4.51553583e-01 6.87591076e-01
4.16211426e-01 2.63153464e-01 3.81854296e-01 1.51426256e-01
1.44007370e-01 -9.88362804e-02 4.26272869e-01 1.41995668e+00
9.82092470e-02 -8.14098775e-01 -1.33954990e+00 6.81277514e-01
-1.45827734e+00 -1.12353051e+00 -6.65666997e-01 1.57441592e+00
8.48616242e-01 4.48896214e-02 -1.34098873e-01 5.30440927e-01
9.32995617e-01 7.76158497e-02 -1.59957752e-01 -1.10742450e+00
-4.22380090e-01 -5.08574136e-02 4.44612265e-01 3.17539036e-01
-1.32187021e+00 1.22325385e+00 4.59954119e+00 1.00408721e+00
-1.26819050e+00 4.39628631e-01 5.07906079e-01 6.90534353e-01
-3.01318392e-02 -5.86059272e-01 -1.00148630e+00 7.31699407e-01
1.00350761e+00 -2.47808129e-01 2.14911267e-01 6.77590549e-01
3.26323599e-01 2.52556086e-01 -3.34694564e-01 9.81121600e-01
5.81159592e-01 -1.18273926e+00 8.00328329e-02 -1.07566833e-01
6.63920701e-01 5.50489128e-01 2.06007119e-02 3.87631983e-01
3.69942784e-01 -1.23133171e+00 7.01589227e-01 -3.34196277e-02
7.41197824e-01 -9.37278628e-01 1.44920743e+00 5.56379497e-01
-7.59310961e-01 -1.50224298e-01 -7.18172491e-01 -3.24089639e-02
-3.77206807e-03 4.03747737e-01 -1.16643178e+00 4.37895894e-01
7.77009130e-01 8.82211685e-01 -4.95990336e-01 3.90036851e-01
1.57169908e-01 9.15995061e-01 -2.86675602e-01 -7.68612206e-01
5.70211768e-01 -1.12108923e-01 7.07560658e-01 1.67908657e+00
1.33907750e-01 5.24594449e-02 1.10021746e-02 5.61612725e-01
-2.79009670e-01 6.70554876e-01 -9.50273633e-01 -5.63037217e-01
4.54907060e-01 1.19141269e+00 -5.54619312e-01 -3.87582451e-01
-3.09440285e-01 8.83231401e-01 2.44230628e-01 4.82974248e-03
-7.64704049e-01 -5.26778221e-01 1.96571544e-01 -1.96799725e-01
1.88415617e-01 -8.91657397e-02 -2.64376819e-01 -9.83452380e-01
-3.60090703e-01 -1.17299199e+00 5.09565771e-01 -4.90899652e-01
-1.75914586e+00 9.63963032e-01 1.46358125e-02 -9.98013437e-01
-2.26714127e-02 -6.47725224e-01 -4.13772076e-01 8.36747348e-01
-1.12800968e+00 -1.35868454e+00 -1.93863481e-01 7.03962266e-01
8.43197167e-01 -5.65320849e-01 8.19939613e-01 4.84003603e-01
-7.66414762e-01 5.32645464e-01 5.01672685e-01 5.81736922e-01
5.59039772e-01 -9.86103952e-01 1.74068049e-01 6.34558320e-01
-3.17896038e-01 3.20815384e-01 6.95897996e-01 -9.33039308e-01
-1.46034372e+00 -1.09020281e+00 1.34187567e+00 -5.60911238e-01
1.16852367e+00 -5.52454352e-01 -8.14873457e-01 6.87705636e-01
6.23223484e-01 -5.19459009e-01 5.17442822e-01 -2.68624365e-01
-2.39539638e-01 2.01928124e-01 -1.33827746e+00 4.08051908e-01
3.70931417e-01 -3.95602405e-01 -8.76774549e-01 5.40823340e-01
7.51009643e-01 -8.64335075e-02 -6.58273280e-01 -1.21703990e-01
3.90731603e-01 -6.16597235e-01 8.07631731e-01 -7.97367156e-01
5.98290324e-01 8.68472010e-02 -1.74404159e-01 -1.12382662e+00
-1.34993404e-01 -3.04016888e-01 4.02633011e-01 1.69296169e+00
3.36521089e-01 -8.87409210e-01 1.81934088e-01 1.27032012e-01
-7.50149637e-02 -1.76929295e-01 -9.56484914e-01 -5.19287109e-01
2.55104929e-01 -6.17436767e-01 3.50164205e-01 1.54814291e+00
2.18915148e-03 6.21381760e-01 -7.82689273e-01 -8.09824616e-02
1.94509432e-01 -1.92635506e-01 5.44752121e-01 -1.23437738e+00
4.30091918e-01 -3.27371538e-01 -6.57734215e-01 -3.51137102e-01
1.17742673e-01 -9.52311575e-01 -2.54284739e-01 -1.32213163e+00
2.53569573e-01 -3.37255687e-01 2.03552917e-01 4.12433892e-01
-2.51120664e-02 3.00343812e-01 1.15668230e-01 1.48751721e-01
-1.96602687e-01 3.55278701e-01 1.05032253e+00 -3.69585484e-01
-1.41668901e-01 -4.88090545e-01 -6.09551251e-01 9.03311789e-01
1.01235271e+00 -7.10034788e-01 -1.01145409e-01 -3.53930861e-01
6.15124583e-01 -1.76147521e-01 5.24136834e-02 -3.90105128e-01
4.38943319e-02 7.47422781e-03 3.77776027e-01 -1.11312366e+00
5.06315529e-02 -6.63331389e-01 -1.47033185e-01 6.03082836e-01
-2.59387325e-02 5.72123826e-01 3.73122633e-01 3.56910437e-01
-3.23904216e-01 -4.05124784e-01 7.52033293e-01 -7.87503496e-02
-3.94924045e-01 6.66126013e-02 -9.39163685e-01 4.88328248e-01
7.06385374e-01 -2.81340480e-01 -5.83984852e-01 -3.83405834e-01
9.29084793e-03 -1.40391529e-01 -1.10676102e-01 5.28205097e-01
6.95987999e-01 -1.02315688e+00 -1.29910100e+00 1.71154216e-01
7.66299292e-02 -2.92828828e-01 3.78244333e-02 7.71878362e-01
-9.30963635e-01 5.61614752e-01 -2.82344848e-01 -2.78001398e-01
-1.33858371e+00 4.31000680e-01 -7.15080872e-02 -3.83685797e-01
-5.59840143e-01 7.60212123e-01 -1.73806906e-01 -5.42070508e-01
-2.76232008e-02 3.58880758e-02 -8.88716042e-01 4.80447352e-01
4.37744588e-01 4.35204744e-01 7.08298087e-02 -1.39660847e+00
-4.55238074e-01 3.03369254e-01 -4.32420045e-01 2.04669520e-01
1.33993447e+00 -1.75506055e-01 -5.31153262e-01 4.27897424e-01
1.55601525e+00 9.62856561e-02 1.10875554e-01 8.59972686e-02
8.77903551e-02 -4.24595743e-01 2.23356321e-01 -1.07207930e+00
-6.61256492e-01 9.01169062e-01 5.98985136e-01 4.16590512e-01
6.82725489e-01 -1.36213347e-01 9.29551661e-01 3.62251163e-01
-1.18062459e-01 -1.28101420e+00 7.45627582e-02 9.63377714e-01
1.17607868e+00 -1.45155716e+00 -6.70226663e-02 7.19826296e-02
-8.52619350e-01 9.66185272e-01 4.70517755e-01 2.18224347e-01
9.55333650e-01 3.80047679e-01 1.74870402e-01 -3.82868260e-01
-3.81965101e-01 3.89847040e-01 1.15369730e-01 6.17656112e-01
5.91399550e-01 1.56131968e-01 -8.36986363e-01 7.32002556e-01
-5.98300576e-01 -5.20474672e-01 7.84022748e-01 7.71831691e-01
-3.96917850e-01 -5.30480325e-01 -6.09957814e-01 8.13903213e-01
-1.23096120e+00 -3.85824382e-01 -5.97192526e-01 9.96951699e-01
1.33081213e-01 1.11739492e+00 -6.71707988e-02 -4.05687630e-01
-1.10485312e-02 -1.59320995e-01 -2.21568063e-01 -3.88404399e-01
-9.99688148e-01 4.90173744e-03 4.47628826e-01 1.23473659e-01
-3.51641804e-01 -5.85120618e-01 -1.05811656e+00 -6.93383873e-01
-7.45738745e-01 -5.41735776e-02 9.86975431e-01 9.69973624e-01
-4.48882319e-02 5.81496730e-02 3.78222167e-01 -4.52812880e-01
-4.45441425e-01 -1.46165323e+00 -4.25548822e-01 3.24116170e-01
2.41424605e-01 -5.68845630e-01 -2.46802881e-01 -8.79520550e-03] | [8.841241836547852, 10.399020195007324] |
86ba8008-cc0a-45e3-9a4d-9e5903500441 | an-overview-of-differentiable-particle | 2302.09639 | null | https://arxiv.org/abs/2302.09639v1 | https://arxiv.org/pdf/2302.09639v1.pdf | An overview of differentiable particle filters for data-adaptive sequential Bayesian inference | By approximating posterior distributions with weighted samples, particle filters (PFs) provide an efficient mechanism for solving non-linear sequential state estimation problems. While the effectiveness of particle filters has been recognised in various applications, the performance of particle filters relies on the knowledge of dynamic models and measurement models, and the construction of effective proposal distributions. An emerging trend in designing particle filters is the differentiable particle filters (DPFs). By constructing particle filters' components through neural networks and optimising them by gradient descent, differentiable particle filters are a promising computational tool to perform inference for sequence data in complex high-dimensional tasks such as vision-based robot localisation. In this paper, we provide a review of recent advances in differentiable particle filters and their applications. We place special emphasis on different design choices of key components of differentiable particle filters, including dynamic models, measurement models, proposal distributions, optimisation objectives, and differentiable resampling techniques. | ['Yunpeng Li', 'Xiongjie Chen'] | 2023-02-19 | null | null | null | null | ['sequential-bayesian-inference'] | ['time-series'] | [ 1.43513143e-01 -2.90658414e-01 -2.31024828e-02 -4.27739918e-02
-3.92526299e-01 -2.03099266e-01 1.14947331e+00 -9.69032291e-03
-7.50278175e-01 1.02727532e+00 8.06943700e-02 -1.89513192e-01
-4.98039991e-01 -6.85012460e-01 -7.40212083e-01 -7.73754239e-01
-4.51764792e-01 6.28355324e-01 4.26498324e-01 -1.20672673e-01
4.70643669e-01 8.97596240e-01 -1.90744197e+00 -4.76771504e-01
7.51366079e-01 7.69747734e-01 6.04786277e-01 1.25044882e+00
-4.61620800e-02 4.10851538e-01 -9.57721055e-01 -2.42589917e-02
-1.17737882e-01 -4.24252450e-01 -2.07118407e-01 -9.57986165e-04
-3.22129242e-02 -4.76590768e-02 -2.19303802e-01 1.10598350e+00
5.26740432e-01 7.69139409e-01 7.32173562e-01 -1.10025847e+00
-5.42275667e-01 1.05573677e-01 1.04977794e-01 3.36614251e-01
2.37640440e-01 1.02140196e-01 6.61996305e-01 -8.42324078e-01
2.03099638e-01 1.57730591e+00 1.18721676e+00 3.25798243e-01
-1.24886191e+00 -7.88944885e-02 2.46458396e-01 1.39320850e-01
-9.59636629e-01 -4.36620116e-01 2.19884261e-01 -4.42393720e-01
1.02743149e+00 3.51091594e-01 7.86006451e-01 9.26656544e-01
6.88565075e-01 9.48751867e-01 9.19127941e-01 -4.09425616e-01
5.87136149e-01 -4.46271114e-02 -5.19994795e-02 4.85097528e-01
2.58113474e-01 7.37187147e-01 -9.75486636e-02 -5.77722430e-01
9.92092431e-01 1.22323772e-02 -2.08721623e-01 -5.79643846e-01
-1.20677316e+00 9.87415016e-01 8.77628922e-02 -1.31641731e-01
-6.76698208e-01 5.10507286e-01 2.55622745e-01 6.42555282e-02
4.88414794e-01 4.81257379e-01 -4.68473315e-01 -2.02700943e-01
-5.01157284e-01 5.84928989e-01 1.01698172e+00 6.35909438e-01
4.68366951e-01 2.55924672e-01 -4.18690830e-01 7.18915820e-01
7.84545064e-01 1.15784073e+00 4.20200855e-01 -7.41507411e-01
-8.88527259e-02 -1.35044500e-01 8.27291787e-01 -7.39303589e-01
-4.37163293e-01 -4.93004024e-01 -7.88798392e-01 4.93385643e-01
2.57828325e-01 -2.62311965e-01 -9.13641512e-01 1.47161663e+00
5.94514787e-01 5.39673090e-01 6.71378151e-02 7.11161196e-01
1.72558203e-01 7.71174252e-01 7.60625601e-02 -2.56581247e-01
1.12129807e+00 -7.34363675e-01 -7.84121513e-01 -2.88261592e-01
-7.22910836e-02 -7.34601855e-01 4.51922357e-01 2.67337471e-01
-9.84984279e-01 -7.98075080e-01 -9.30052876e-01 3.38499963e-01
-3.08718264e-01 3.19818377e-01 4.73474056e-01 6.62327349e-01
-8.92913342e-01 1.14201963e+00 -1.33396494e+00 -2.05665588e-01
1.41974419e-01 5.33429086e-01 2.37696558e-01 3.69120449e-01
-1.00309277e+00 1.47188628e+00 3.95517707e-01 3.64148408e-01
-9.44194734e-01 -4.34674859e-01 -7.25030482e-01 -1.24162614e-01
-5.87571077e-02 -5.86801052e-01 1.46037054e+00 -2.20812768e-01
-2.22557950e+00 -2.43002605e-02 -2.58573294e-01 -8.30733895e-01
4.81901318e-01 -4.13183898e-01 -4.04283464e-01 -2.47905642e-01
-1.03907138e-01 1.80962533e-01 1.33946991e+00 -7.85063922e-01
-1.24223781e+00 2.06318736e-01 -1.52119845e-01 9.97519866e-02
3.81885320e-01 -2.38095656e-01 2.31661145e-02 -9.65363011e-02
-1.12885423e-01 -8.66787493e-01 -6.67562604e-01 -1.80873007e-01
3.17032412e-02 -4.84291762e-01 6.71356082e-01 -4.06647831e-01
9.35385764e-01 -1.97901869e+00 2.86887705e-01 1.53508663e-01
-7.47887120e-02 4.14312690e-01 -1.47646725e-01 4.93933797e-01
4.80354309e-01 -6.85834765e-01 -1.89282760e-01 -5.81753075e-01
5.32280385e-01 3.90895426e-01 -5.37366152e-01 9.94663179e-01
3.46338958e-01 8.35606158e-01 -1.27503896e+00 4.38646302e-02
9.63729084e-01 8.19315791e-01 -3.40591550e-01 1.11949503e-01
-3.04837853e-01 5.43708563e-01 -4.89725232e-01 4.98168431e-02
5.26907623e-01 -4.47443537e-02 -8.40621293e-02 -8.32195356e-02
-5.18236279e-01 2.56734252e-01 -1.48197484e+00 1.04235470e+00
-3.55924070e-01 6.14314079e-01 1.83851928e-01 -1.03022397e+00
1.00972307e+00 1.69330627e-01 4.03803855e-01 -3.97643119e-01
1.81528762e-01 2.25326195e-01 -2.46101797e-01 -1.21091634e-01
8.56929064e-01 -2.88809597e-01 6.74987957e-02 2.00003996e-01
8.38802233e-02 -4.90506142e-01 4.32417482e-01 -4.70441639e-01
8.39128435e-01 3.22381496e-01 4.11478817e-01 -3.37864518e-01
7.65297174e-01 -9.76783410e-02 2.94443548e-01 1.20535707e+00
-7.92737007e-02 2.26293787e-01 -1.39456868e-01 -4.90909696e-01
-1.15173697e+00 -1.18463016e+00 -3.56973588e-01 9.16670918e-01
1.10009916e-01 2.87024844e-02 -2.43948191e-01 -1.55870065e-01
3.84820879e-01 6.17233932e-01 -4.86764967e-01 -3.29942852e-01
-5.51091731e-01 -9.78210092e-01 2.23612860e-01 4.10705060e-01
1.95914969e-01 -1.08843446e+00 -8.93200696e-01 9.14787471e-01
3.72354269e-01 -5.31814754e-01 -2.65047878e-01 3.87834311e-01
-8.83453727e-01 -1.13862967e+00 -7.07871318e-01 -4.45344120e-01
5.12657285e-01 2.06361022e-02 9.91255522e-01 -3.62199962e-01
-1.87864780e-01 6.42490149e-01 -1.76972091e-01 -7.57535517e-01
-6.03462994e-01 -2.64740348e-01 4.24473256e-01 -2.28150770e-01
1.21134534e-01 -2.18351156e-01 -4.05480027e-01 3.70911479e-01
-5.82209408e-01 -5.18350780e-01 5.44012547e-01 1.12161958e+00
7.66827762e-01 3.99854213e-01 1.44981042e-01 -5.52106023e-01
1.22173917e+00 -3.82547975e-01 -1.19576359e+00 8.82076845e-02
-3.43410730e-01 2.79398263e-01 6.38196528e-01 -6.91550434e-01
-1.02353430e+00 -9.90338251e-02 -4.80613321e-01 -1.89510852e-01
-1.14532195e-01 4.12986994e-01 3.82455856e-01 -9.85913575e-02
5.97318828e-01 3.33794504e-01 3.99912655e-01 -3.85542452e-01
4.22880977e-01 3.28916878e-01 5.82110882e-01 -6.18764460e-01
4.91007119e-01 6.93330228e-01 1.57017112e-01 -1.08650458e+00
-6.85338438e-01 -5.59326649e-01 -3.90288293e-01 -1.59924492e-01
6.37140036e-01 -5.45603096e-01 -1.11980867e+00 4.80178624e-01
-8.89909208e-01 -3.03604037e-01 -8.19371641e-01 9.09880400e-01
-8.87024283e-01 3.95760596e-01 -4.00411963e-01 -1.26334238e+00
-1.98679283e-01 -1.16085660e+00 1.19595683e+00 5.39842844e-01
-1.92064509e-01 -1.41193569e+00 4.67956126e-01 -6.64103806e-01
5.18679798e-01 1.26375156e-02 2.37605363e-01 -3.73978466e-01
-3.48652303e-01 -5.95641017e-01 1.98489070e-01 4.20744270e-01
1.74239799e-01 4.22521867e-02 -5.40545881e-01 -6.44388616e-01
1.23324260e-01 2.77410895e-01 5.58913410e-01 1.09075141e+00
4.71187472e-01 4.77444530e-02 -6.97099388e-01 4.67074186e-01
1.20106328e+00 3.71570170e-01 4.70432729e-01 4.10075039e-01
1.31643802e-01 1.40600115e-01 5.27808964e-01 6.14052176e-01
4.25101183e-02 4.11882579e-01 3.43758196e-01 3.83182406e-01
6.66167811e-02 -1.68301970e-01 3.49669129e-01 5.71194887e-01
1.42363720e-02 -2.49606371e-01 -3.85895401e-01 3.80659491e-01
-2.03135586e+00 -1.06790459e+00 -1.55094162e-01 2.29702210e+00
2.42007628e-01 9.28140283e-02 1.00318491e-01 -5.31771854e-02
8.65400374e-01 -4.51850258e-02 -6.34094536e-01 -1.68076843e-01
9.26873758e-02 3.30809414e-01 7.58809209e-01 7.96944678e-01
-1.32308495e+00 5.79858243e-01 7.61108208e+00 6.39345586e-01
-7.44702458e-01 -3.81863154e-02 -4.27885711e-01 3.13756764e-01
1.11804292e-01 -1.83545232e-01 -1.38578749e+00 8.16084802e-01
1.17999530e+00 -6.49335533e-02 4.04329628e-01 8.00120771e-01
3.88718247e-01 -3.67364287e-01 -8.26903462e-01 7.43785143e-01
-3.93585712e-01 -1.49677658e+00 -3.68481487e-01 1.37430295e-01
7.24384546e-01 3.58323574e-01 1.38286188e-01 5.24959922e-01
9.53619480e-01 -7.42117226e-01 8.47687542e-01 1.03508425e+00
8.56051072e-02 -8.57428730e-01 8.51069689e-01 4.24062043e-01
-1.14504766e+00 -3.13667059e-01 -9.66440737e-01 -3.50094020e-01
7.91642427e-01 8.86943579e-01 -7.82592297e-01 1.81766883e-01
4.19951797e-01 7.04607427e-01 1.21982113e-01 1.73723698e+00
-2.15778306e-01 3.16907138e-01 -7.67875731e-01 -6.70235813e-01
3.38055044e-01 -4.75199193e-01 8.77438366e-01 1.11917365e+00
3.73552442e-01 -1.86451465e-01 4.08886909e-01 7.12169290e-01
7.35156536e-01 -5.16248047e-01 -1.54294357e-01 -2.95668185e-01
5.83645761e-01 9.15385306e-01 -8.48400772e-01 -3.46548051e-01
-2.60643423e-01 5.79097033e-01 1.18351296e-01 1.97615609e-01
-7.22272038e-01 -4.02145803e-01 1.27905858e+00 -3.98627549e-01
6.99469805e-01 -7.80581951e-01 3.25379580e-01 -7.86600292e-01
-4.92520094e-01 -5.67694247e-01 2.77515687e-02 -4.13878560e-01
-1.40471041e+00 3.27732921e-01 1.60631850e-01 -1.03013599e+00
-4.63748634e-01 -9.22791123e-01 -5.49917340e-01 1.06396210e+00
-1.31165707e+00 -6.23884201e-01 1.81324095e-01 1.59540936e-01
4.94340420e-01 -3.82922962e-02 7.52017856e-01 -1.31414071e-01
-1.26499861e-01 -1.11718565e-01 1.03437877e+00 -3.08291554e-01
1.12482585e-01 -1.31428528e+00 1.03614104e+00 7.42765903e-01
8.79395008e-02 6.85678601e-01 1.34704661e+00 -8.76431286e-01
-1.67311990e+00 -8.22467148e-01 3.71487677e-01 -6.17377400e-01
8.52437258e-01 -3.65410969e-02 -6.38079703e-01 4.93588120e-01
-4.06028003e-01 -1.23925641e-01 8.03121999e-02 7.75940120e-02
6.09419942e-01 2.15862066e-01 -1.04743636e+00 4.43908244e-01
6.92202389e-01 -3.35844636e-01 -6.71698153e-01 4.23124969e-01
2.29532763e-01 -7.59728253e-01 -8.24575484e-01 2.66227096e-01
6.42804742e-01 -7.89962769e-01 1.38305712e+00 -5.03081381e-01
-5.55483043e-01 -5.86054564e-01 8.90471935e-02 -1.79884124e+00
-8.39361012e-01 -4.81144547e-01 -5.01387715e-01 5.08807898e-01
-5.92792518e-02 -1.21196234e+00 8.56165707e-01 1.51626259e-01
-3.71979773e-01 -5.08223891e-01 -8.52926493e-01 -9.79081869e-01
-2.43073180e-01 -4.97881383e-01 6.07556462e-01 7.29518430e-03
-4.37332064e-01 8.38770419e-02 -4.68406498e-01 5.59623778e-01
1.01144123e+00 -1.48596600e-01 7.86409557e-01 -1.46398056e+00
-4.76162344e-01 -4.33471769e-01 -4.27860975e-01 -1.49962020e+00
2.13565171e-01 -2.94734001e-01 6.62720978e-01 -1.82878280e+00
-6.43646657e-01 -3.61267924e-01 -4.60755490e-02 -3.58494759e-01
-2.15966195e-01 -6.43111765e-02 -7.93251842e-02 1.09002396e-01
-4.27379221e-01 7.52561986e-01 1.07858074e+00 5.17331846e-02
-2.38912106e-01 7.58121848e-01 -1.31222485e-02 6.79821372e-01
7.27544069e-01 -2.79176384e-01 -4.71893728e-01 -2.85837770e-01
9.57752690e-02 -1.54359251e-01 4.69827294e-01 -1.18318498e+00
4.55698162e-01 -1.86850488e-01 5.80295026e-01 -8.89785707e-01
5.42405605e-01 -6.27757132e-01 5.51512718e-01 8.55704010e-01
5.41040823e-02 1.64082855e-01 2.07052678e-01 1.00377631e+00
-4.33591940e-02 -4.20997500e-01 8.85320246e-01 -2.75842637e-01
-1.08692694e+00 2.55923480e-01 -8.01224589e-01 -2.62254477e-01
7.06859231e-01 -3.28806937e-01 6.52528033e-02 -1.84901372e-01
-9.30724263e-01 2.31017172e-01 1.78885192e-01 3.24295193e-01
6.11839175e-01 -1.13390648e+00 -5.08726537e-01 4.66953814e-01
-4.77044284e-01 -2.06077341e-02 4.90199514e-02 8.44470024e-01
-4.38280076e-01 6.03545249e-01 -4.13117893e-02 -7.92392969e-01
-8.57041597e-01 3.33017677e-01 3.99570972e-01 -7.53905848e-02
-8.85040283e-01 1.21040189e+00 -2.43180156e-01 -7.14430988e-01
1.56507850e-01 -6.41707957e-01 -1.04080155e-01 -2.19959140e-01
7.59901762e-01 7.42625713e-01 -1.68161720e-01 -4.09341365e-01
-2.23849535e-01 2.74990976e-01 2.02288598e-01 -1.11982316e-01
1.25290012e+00 -3.40753943e-01 6.75447956e-02 4.32188421e-01
7.52977788e-01 -3.46934319e-01 -1.91079473e+00 -9.26230103e-02
1.70350805e-01 -3.44803274e-01 8.16482604e-02 -3.72893542e-01
-1.99475974e-01 4.42515016e-01 5.89594603e-01 4.41655785e-01
3.86875600e-01 -2.36284062e-01 5.64280808e-01 4.08161253e-01
6.80246949e-01 -1.11563540e+00 -2.44293839e-01 1.24210680e+00
5.95389366e-01 -7.87711859e-01 1.71110272e-01 1.17794015e-01
-1.35049418e-01 1.01222813e+00 -7.96435326e-02 -7.65222728e-01
9.27943647e-01 3.74952883e-01 -3.19281369e-01 6.49447516e-02
-5.87580025e-01 -3.34647179e-01 4.66160357e-01 1.04390168e+00
8.58476087e-02 1.43304318e-01 -1.44808650e-01 -4.17650677e-02
-3.62439871e-01 -8.05182159e-02 2.41650715e-01 1.10539627e+00
-1.06377113e+00 -1.00025928e+00 -7.15021491e-01 7.18673229e-01
-8.09976272e-03 2.03076184e-01 5.24374008e-01 5.64045846e-01
-1.87193498e-01 7.30552614e-01 3.39423984e-01 1.36005998e-01
5.05374193e-01 -1.19854301e-01 7.78572500e-01 -5.11989117e-01
-1.25386998e-01 -1.69257894e-01 1.77810248e-02 -4.22327429e-01
-4.19843912e-01 -9.99967694e-01 -9.57806289e-01 -2.74800390e-01
-5.99166155e-01 7.38065124e-01 1.02964938e+00 1.22729897e+00
2.48129502e-01 6.62641943e-01 2.21877038e-01 -1.52795815e+00
-1.20004511e+00 -9.00262833e-01 -7.39657998e-01 -1.24089211e-01
5.65025985e-01 -1.25973868e+00 -4.56081003e-01 -2.99234748e-01] | [6.573317050933838, 3.591362237930298] |
f8dda1a0-383d-48c1-b36a-5fcb3d6b57b3 | pars-pseudo-label-aware-robust-sample-1 | 2201.10836 | null | https://arxiv.org/abs/2201.10836v1 | https://arxiv.org/pdf/2201.10836v1.pdf | PARS: Pseudo-Label Aware Robust Sample Selection for Learning with Noisy Labels | Acquiring accurate labels on large-scale datasets is both time consuming and expensive. To reduce the dependency of deep learning models on learning from clean labeled data, several recent research efforts are focused on learning with noisy labels. These methods typically fall into three design categories to learn a noise robust model: sample selection approaches, noise robust loss functions, or label correction methods. In this paper, we propose PARS: Pseudo-Label Aware Robust Sample Selection, a hybrid approach that combines the best from all three worlds in a joint-training framework to achieve robustness to noisy labels. Specifically, PARS exploits all training samples using both the raw/noisy labels and estimated/refurbished pseudo-labels via self-training, divides samples into an ambiguous and a noisy subset via loss analysis, and designs label-dependent noise-aware loss functions for both sets of filtered labels. Results show that PARS significantly outperforms the state of the art on extensive studies on the noisy CIFAR-10 and CIFAR-100 datasets, particularly on challenging high-noise and low-resource settings. In particular, PARS achieved an absolute 12% improvement in test accuracy on the CIFAR-100 dataset with 90% symmetric label noise, and an absolute 27% improvement in test accuracy when only 1/5 of the noisy labels are available during training as an additional restriction. On a real-world noisy dataset, Clothing1M, PARS achieves competitive results to the state of the art. | ['Jordan Massiah', 'Yunlong Jiao', 'Arushi Goel'] | 2022-01-26 | pars-pseudo-label-aware-robust-sample | https://openreview.net/forum?id=ovRQmeVFbrC | https://openreview.net/pdf?id=ovRQmeVFbrC | null | ['learning-with-noisy-labels', 'learning-with-noisy-labels'] | ['computer-vision', 'natural-language-processing'] | [ 1.24782048e-01 -3.27568173e-01 4.39932197e-02 -7.92291522e-01
-1.91881919e+00 -4.77021754e-01 2.76520580e-01 2.00270507e-02
-8.75779033e-01 9.54873502e-01 7.53249228e-02 2.57467121e-01
-8.15027282e-02 -5.14778018e-01 -6.17160678e-01 -8.99362981e-01
2.89573312e-01 4.55839992e-01 -7.57747218e-02 2.86431253e-01
-3.56645942e-01 1.27379015e-01 -1.56579101e+00 2.74955541e-01
7.25344181e-01 1.43952727e+00 -1.07420467e-01 3.82554054e-01
-1.57887697e-01 8.76959980e-01 -7.10401237e-01 -5.14876485e-01
4.48863924e-01 -2.57344037e-01 -8.07078481e-01 -1.03488239e-02
7.13318169e-01 -5.38487919e-02 -1.68013811e-01 1.38592362e+00
1.06288826e+00 2.35835969e-01 4.49645668e-01 -9.57473516e-01
-4.93289530e-01 8.19386780e-01 -3.73457074e-01 9.13756937e-02
-2.97139704e-01 1.32057369e-01 9.74213719e-01 -1.14345062e+00
3.71532321e-01 1.41596055e+00 9.83265758e-01 7.13455260e-01
-1.51764154e+00 -1.05471361e+00 2.24677831e-01 -2.23041419e-02
-1.60366929e+00 -4.90334719e-01 4.23357040e-01 -3.65724742e-01
3.96294564e-01 1.64942250e-01 -1.17064767e-01 1.28947806e+00
-4.04202104e-01 8.57950270e-01 1.45417273e+00 -3.95859748e-01
4.48009372e-01 1.08264394e-01 6.18206084e-01 2.94314355e-01
2.20495015e-01 1.86202750e-01 -4.92447227e-01 -2.75095820e-01
9.02507678e-02 -2.12744981e-01 -2.55092084e-01 1.56128466e-01
-9.56693232e-01 7.07937717e-01 4.36410904e-01 1.68399757e-03
-8.10657442e-02 1.79248154e-01 5.93215704e-01 3.72960836e-01
9.46120799e-01 3.43924999e-01 -8.02069545e-01 6.33947402e-02
-9.69565630e-01 1.87005177e-01 4.97532308e-01 8.04021418e-01
8.18431199e-01 8.83754119e-02 -7.31783152e-01 1.29116774e+00
2.89845258e-01 8.39027107e-01 3.83528352e-01 -8.65262032e-01
6.96913123e-01 1.58264205e-01 1.80987060e-01 -4.80961382e-01
-5.48467159e-01 -1.13490951e+00 -9.64953840e-01 5.65250851e-02
3.89375895e-01 -3.18571895e-01 -1.28230715e+00 2.09414387e+00
5.09213328e-01 3.85539681e-01 5.45405522e-02 8.57519031e-01
1.11160851e+00 3.76712948e-01 4.41082060e-01 -1.72846630e-01
1.11872721e+00 -1.03084803e+00 -8.85217249e-01 -3.59253556e-01
8.43937993e-01 -7.71807432e-01 1.30238271e+00 6.70110941e-01
-6.53580606e-01 -6.60948753e-01 -8.07310343e-01 -5.18462621e-02
-2.49465778e-01 4.47844177e-01 1.61407948e-01 8.06339145e-01
-6.76238716e-01 6.27148628e-01 -6.17085695e-01 2.22722188e-01
8.56515408e-01 1.67100564e-01 -3.05982649e-01 -6.79680884e-01
-1.22003174e+00 4.73529816e-01 2.32679605e-01 1.06196627e-01
-1.09005404e+00 -8.73385370e-01 -7.75953293e-01 1.62970982e-02
5.62845469e-01 -3.32664222e-01 1.49361837e+00 -8.07567000e-01
-1.13107693e+00 8.42352331e-01 -1.30673521e-03 -4.63387489e-01
7.36725092e-01 -5.42053461e-01 -4.61561412e-01 -2.49176532e-01
2.75477916e-01 5.24704635e-01 6.36176944e-01 -1.52670646e+00
-7.90233731e-01 -3.18224967e-01 -3.71541768e-01 -3.19125988e-02
-2.01204836e-01 1.08042918e-02 -3.24273735e-01 -9.29914057e-01
2.33215153e-01 -9.68819499e-01 -2.40374967e-01 -2.79765695e-01
-5.94769061e-01 -2.95833677e-01 6.46674097e-01 -4.51498151e-01
1.09965587e+00 -2.16548181e+00 -4.45747435e-01 1.23929039e-01
1.83480665e-01 6.59894824e-01 -4.42228049e-01 -1.71820089e-01
-1.58628061e-01 4.35285032e-01 -1.52712673e-01 -8.62883270e-01
1.43114761e-01 1.10742278e-01 -1.95786655e-01 4.61988777e-01
1.00313276e-01 7.34891534e-01 -1.10542381e+00 -2.22951949e-01
-3.57420444e-02 5.86085260e-01 -8.59485418e-02 1.18960962e-01
-2.41839029e-02 5.25722086e-01 -3.32082123e-01 6.87701881e-01
1.00836670e+00 -7.05297515e-02 -2.20930904e-01 -3.26215416e-01
5.11069357e-01 2.71087915e-01 -1.66464853e+00 1.38290811e+00
-5.66790044e-01 2.14377269e-01 3.42241406e-01 -7.41870761e-01
9.02687311e-01 5.06041288e-01 2.08270073e-01 -6.86977088e-01
3.20380419e-01 4.16884750e-01 -4.88408417e-01 -4.98460352e-01
3.28806639e-02 -3.36620212e-01 -1.24857426e-01 3.25766891e-01
2.56530225e-01 -2.69519165e-02 2.79796362e-01 -6.03065342e-02
1.11341405e+00 -9.96282548e-02 -2.81086136e-02 -7.44111091e-02
2.23662108e-01 -4.18892652e-01 1.06326997e+00 1.18785346e+00
-4.72474247e-01 1.02582335e+00 1.30291760e-01 -2.91612983e-01
-6.22518837e-01 -1.03784692e+00 -3.97182852e-01 1.18915832e+00
-1.48383915e-01 -2.61014104e-01 -8.42429817e-01 -1.12967515e+00
-1.39375404e-01 7.58112609e-01 -5.12443841e-01 -1.51345834e-01
-3.18386137e-01 -1.28658211e+00 7.12581813e-01 4.66447830e-01
6.44671559e-01 -8.38528335e-01 2.88801491e-01 1.51533902e-01
-5.01880765e-01 -1.12609303e+00 -4.68933463e-01 7.49465227e-01
-4.68052804e-01 -1.04407287e+00 -3.15645725e-01 -8.22433472e-01
5.72604299e-01 2.22649574e-01 1.27410889e+00 -1.05475001e-01
-1.16180420e-01 5.80218770e-02 -6.39481187e-01 -5.06037474e-01
-2.53581405e-01 2.47861017e-02 1.17882386e-01 2.14239433e-01
4.54772115e-01 -2.40801126e-01 -3.14956337e-01 4.73044872e-01
-9.47774231e-01 -5.01125157e-01 3.66837919e-01 1.14916539e+00
1.15367651e+00 2.66727418e-01 9.44409490e-01 -1.22792661e+00
3.44118476e-01 -6.04624987e-01 -5.07347882e-01 2.50617385e-01
-5.33023775e-01 7.31973723e-02 6.54877543e-01 -4.77485716e-01
-1.15221834e+00 2.93380439e-01 -4.29667383e-01 -3.36804837e-01
-3.03787231e-01 2.81946450e-01 -4.90628660e-01 1.08860053e-01
1.03127861e+00 -1.87330782e-01 -6.63400412e-01 -9.80585814e-01
3.99993926e-01 9.41329002e-01 5.43430507e-01 -7.77888179e-01
7.53794789e-01 3.82861465e-01 -2.22929582e-01 -2.18246639e-01
-1.72220969e+00 -6.46865129e-01 -3.85457724e-01 1.11792192e-01
5.41792572e-01 -1.20213342e+00 -1.85973153e-01 7.39404500e-01
-7.46154070e-01 -3.80030513e-01 -6.41693413e-01 5.74017525e-01
-6.84900284e-02 1.21660471e-01 -7.73331404e-01 -8.22189093e-01
-4.91933554e-01 -1.36188555e+00 1.25253916e+00 1.28095254e-01
1.58489496e-01 -5.64202547e-01 -9.51225385e-02 5.21685541e-01
4.02415544e-01 1.21596783e-01 5.64169645e-01 -1.08228302e+00
-3.26526642e-01 -4.91086632e-01 -3.72411460e-01 1.05975938e+00
6.96748644e-02 -4.48925227e-01 -1.55849898e+00 -4.63686198e-01
5.45820743e-02 -1.02902615e+00 1.31246197e+00 3.58724952e-01
1.42633080e+00 -8.90157465e-03 -1.06355324e-01 6.05920136e-01
1.49615145e+00 -1.89343318e-01 2.77804255e-01 1.17178097e-01
7.21859455e-01 2.70886540e-01 7.52948642e-01 3.24500084e-01
1.23188868e-01 4.60258007e-01 4.07024711e-01 -3.91133688e-02
-5.56553185e-01 -1.40570089e-01 -4.06937450e-02 6.75886452e-01
5.84759474e-01 -3.47526401e-01 -8.02229822e-01 5.27439356e-01
-1.73280358e+00 -6.23122275e-01 -2.24052161e-01 2.30692291e+00
1.19131029e+00 1.52705982e-01 -3.28099668e-01 4.14335042e-01
8.88887703e-01 -1.04914546e-01 -6.24982834e-01 3.10521960e-01
-3.79434198e-01 3.80954921e-01 5.60729861e-01 4.89139974e-01
-1.75275004e+00 8.44812930e-01 6.02377176e+00 1.54613006e+00
-9.33499575e-01 6.20712161e-01 1.24716055e+00 -4.57134396e-01
2.09142398e-02 -4.59948689e-01 -1.06138444e+00 6.10711515e-01
9.42182958e-01 3.81230593e-01 2.56322235e-01 9.74074125e-01
2.75565952e-01 -1.22050576e-01 -9.90766823e-01 1.22206736e+00
-1.07689649e-01 -8.85780096e-01 -3.43021870e-01 -2.77276814e-01
1.22691786e+00 4.40504968e-01 1.95127934e-01 6.35220349e-01
8.23094606e-01 -1.14810455e+00 8.69635761e-01 5.03768265e-01
9.04252589e-01 -7.96554744e-01 1.32592940e+00 4.23634529e-01
-1.05238152e+00 -3.20637375e-01 -4.07641202e-01 1.16993360e-01
3.57530713e-02 1.54337704e+00 -3.46216708e-01 3.49230468e-01
9.98332560e-01 4.62493777e-01 -6.49283230e-01 1.34489322e+00
-4.58936006e-01 1.28170907e+00 -5.62580645e-01 4.24811304e-01
4.72016782e-02 9.73130316e-02 2.81806082e-01 1.26513827e+00
1.48303926e-01 -4.43696827e-02 5.74708641e-01 5.55412769e-01
-6.96818292e-01 1.65113583e-02 -2.48300447e-03 5.74994385e-01
7.51533389e-01 1.46827769e+00 -6.65078282e-01 -4.27028805e-01
-1.92953154e-01 5.19796133e-01 4.99969810e-01 5.71174681e-01
-7.66025543e-01 -3.45772684e-01 6.01086378e-01 -1.90047190e-01
2.15466663e-01 3.12560886e-01 -6.44463778e-01 -1.06032825e+00
7.27348700e-02 -8.98854256e-01 4.11835551e-01 -5.25609374e-01
-1.83687866e+00 8.01017344e-01 -3.36637825e-01 -1.07567608e+00
3.47212166e-01 -2.11726665e-01 -1.63567916e-01 9.45582390e-01
-1.66111076e+00 -1.04949784e+00 -3.12920868e-01 2.04753503e-01
5.70782423e-01 1.17822990e-01 7.69552112e-01 8.57788205e-01
-8.37519169e-01 1.07925189e+00 6.38069868e-01 3.13872755e-01
1.17081201e+00 -1.27629352e+00 3.60588878e-01 9.85129058e-01
2.46773884e-01 1.21840745e-01 4.00705189e-01 -5.21980703e-01
-4.45395678e-01 -1.81440187e+00 9.16985214e-01 -4.21254218e-01
1.49415299e-01 -5.50922990e-01 -1.06295407e+00 5.21856606e-01
-4.03045744e-01 8.31477046e-01 6.16245866e-01 5.84758185e-02
-7.43081510e-01 -5.27381778e-01 -1.51193976e+00 2.12357998e-01
1.15961730e+00 -3.90412331e-01 -2.00209796e-01 6.34429276e-01
8.59404683e-01 -4.78126854e-01 -8.16874921e-01 6.08236372e-01
-6.77016079e-02 -6.44344151e-01 8.57498825e-01 -4.10990834e-01
-6.98217228e-02 -3.95126790e-01 -4.40012813e-01 -1.54169691e+00
-3.62525105e-01 -3.09303343e-01 2.60417819e-01 1.82927656e+00
5.37078202e-01 -4.36293602e-01 7.36552775e-01 5.63200653e-01
-2.09153816e-01 -7.92691469e-01 -1.07478476e+00 -7.92094946e-01
4.93977591e-02 -9.20806050e-01 5.61075747e-01 8.97948802e-01
-7.76602447e-01 3.42254579e-01 -5.06704748e-01 1.03500873e-01
8.07139158e-01 -4.57854480e-01 4.89057869e-01 -1.31841838e+00
-2.17674226e-02 -1.47829026e-01 1.98928919e-02 -7.41138458e-01
3.23949218e-01 -1.04211342e+00 5.20534813e-01 -1.41410756e+00
1.76484972e-01 -9.33902323e-01 -6.57037079e-01 8.93165827e-01
-5.74957550e-01 6.27533376e-01 3.93083245e-02 1.44776165e-01
-1.02697182e+00 7.43266404e-01 8.80690813e-01 -2.85457224e-01
1.13325410e-01 2.87694842e-01 -7.93901622e-01 6.89872801e-01
7.60996163e-01 -1.09570074e+00 -3.09852213e-01 -4.58365440e-01
1.36288583e-01 -6.36724710e-01 1.15347013e-01 -1.20954466e+00
-1.59286320e-01 5.58032244e-02 2.56152004e-01 -3.27532560e-01
1.81567833e-01 -8.05438280e-01 -8.10341462e-02 8.56750384e-02
-8.13164055e-01 -5.61012149e-01 -3.28141451e-02 8.59130800e-01
-3.12250346e-01 -4.02681738e-01 1.31555116e+00 -3.37604247e-02
-5.20521879e-01 3.21817040e-01 1.74689889e-02 6.79045677e-01
6.00857317e-01 3.43747109e-01 -4.55660403e-01 -1.17243089e-01
-9.31019068e-01 3.13675761e-01 -7.39257187e-02 3.11695457e-01
1.57663047e-01 -1.47378302e+00 -9.12652016e-01 2.27385368e-02
1.30475417e-01 3.86724353e-01 3.84727001e-01 5.70011020e-01
-9.91785824e-02 7.94166103e-02 5.66720545e-01 -6.14569962e-01
-1.20808041e+00 4.05572295e-01 4.69337732e-01 -6.27383471e-01
-2.33670652e-01 1.40411294e+00 -2.21499847e-03 -8.91212881e-01
7.87444055e-01 -3.28562677e-01 -2.01457180e-02 3.99455093e-02
7.89287090e-01 5.76797724e-01 6.68898404e-01 -7.01140165e-01
-1.36863977e-01 5.32379925e-01 2.86094937e-02 2.50167549e-01
1.15947998e+00 -3.36426735e-01 1.21689186e-01 4.41881180e-01
1.38099217e+00 -8.26678202e-02 -1.46830583e+00 -1.02948499e+00
8.20636898e-02 -5.16343772e-01 3.63072753e-01 -1.43449879e+00
-1.49149132e+00 6.15265727e-01 9.98157918e-01 -7.54120797e-02
1.18522978e+00 -4.93390411e-02 8.94639969e-01 3.89333010e-01
2.15164691e-01 -1.34868860e+00 8.09019282e-02 6.25142097e-01
5.31890273e-01 -1.65672112e+00 -2.66271859e-01 -5.10545611e-01
-5.99807799e-01 4.34391439e-01 5.95299721e-01 4.86496277e-02
9.53659356e-01 3.47406447e-01 5.38527966e-01 1.06543504e-01
-5.72484136e-01 -3.57656777e-01 2.47087374e-01 6.31723404e-01
2.32070178e-01 9.15947333e-02 -2.02070057e-01 1.21000040e+00
-2.20659599e-02 -1.07805759e-01 1.68210536e-01 6.68804407e-01
-5.33286095e-01 -1.00652099e+00 -6.36425436e-01 8.02819490e-01
-7.55982637e-01 -2.63380259e-01 1.45894391e-02 3.76833320e-01
7.83622980e-01 1.40203285e+00 -3.50990266e-01 -3.16660672e-01
6.22756958e-01 3.29521060e-01 -4.24505286e-02 -8.16540718e-01
-7.37348974e-01 1.68547422e-01 1.43314824e-01 -4.81243879e-01
-3.41735512e-01 -3.50659668e-01 -1.09588516e+00 2.11121440e-02
-6.68061495e-01 1.89464465e-01 6.45682573e-01 1.02935326e+00
3.39093685e-01 5.94031215e-01 5.82415998e-01 -8.12646866e-01
-9.78111267e-01 -1.24569702e+00 -7.99351990e-01 7.96579421e-01
4.65944409e-01 -7.73286164e-01 -7.49804795e-01 -1.00243337e-01] | [9.377847671508789, 3.877126932144165] |
86f81758-cc42-4ca6-ab21-631634669fcd | scalable-randomized-kernel-methods-for | 2304.04692 | null | https://arxiv.org/abs/2304.04692v1 | https://arxiv.org/pdf/2304.04692v1.pdf | Scalable Randomized Kernel Methods for Multiview Data Integration and Prediction | We develop scalable randomized kernel methods for jointly associating data from multiple sources and simultaneously predicting an outcome or classifying a unit into one of two or more classes. The proposed methods model nonlinear relationships in multiview data together with predicting a clinical outcome and are capable of identifying variables or groups of variables that best contribute to the relationships among the views. We use the idea that random Fourier bases can approximate shift-invariant kernel functions to construct nonlinear mappings of each view and we use these mappings and the outcome variable to learn view-independent low-dimensional representations. Through simulation studies, we show that the proposed methods outperform several other linear and nonlinear methods for multiview data integration. When the proposed methods were applied to gene expression, metabolomics, proteomics, and lipidomics data pertaining to COVID-19, we identified several molecular signatures forCOVID-19 status and severity. Results from our real data application and simulations with small sample sizes suggest that the proposed methods may be useful for small sample size problems. Availability: Our algorithms are implemented in Pytorch and interfaced in R and would be made available at: https://github.com/lasandrall/RandMVLearn. | ['Han Lu', 'Sandra E. Safo'] | 2023-04-10 | null | null | null | null | ['data-integration'] | ['knowledge-base'] | [-2.5306368e-01 -2.8042260e-01 -5.3392631e-01 -5.5959409e-01
-7.3315227e-01 -5.7549566e-01 1.9622612e-01 3.8677517e-01
1.3209301e-01 6.6471785e-01 3.2520479e-01 -1.7000689e-01
-4.7331974e-01 -4.2099518e-01 -5.7331836e-01 -7.3338103e-01
-5.9388220e-01 5.1503927e-01 -3.1362054e-01 3.2618362e-01
-7.3872708e-02 4.0422660e-01 -1.2409782e+00 3.8752142e-01
9.1792703e-01 7.3896736e-01 4.9617067e-02 7.8443754e-01
1.7600822e-01 5.7265109e-01 -1.0369691e-01 -4.1680891e-02
2.0959450e-01 -2.4102554e-01 -3.6455011e-01 -1.4893381e-01
3.0802646e-01 -8.5602537e-02 -9.1586761e-02 9.5777649e-01
6.0425079e-01 -2.0585643e-02 8.3920097e-01 -1.5224882e+00
-8.1531960e-01 3.5942453e-01 -7.5537401e-01 2.4874385e-01
8.5086532e-02 -2.1216662e-01 7.4129856e-01 -1.3060577e+00
6.1492318e-01 1.2010788e+00 6.6611886e-01 1.9847561e-01
-1.4680127e+00 -7.2806811e-01 5.1499985e-02 4.2808607e-01
-1.4044757e+00 -5.2229494e-01 6.3222480e-01 -8.2504082e-01
6.7310697e-01 3.5677192e-01 3.4032100e-01 9.0759748e-01
6.4954954e-01 3.9162904e-01 1.2388633e+00 -4.5493588e-02
1.7334992e-01 3.3191875e-01 3.8326702e-01 9.7945917e-01
2.9367772e-01 -5.1287055e-02 -6.4862603e-01 -9.5758611e-01
5.2909195e-01 5.6192452e-01 -3.1190345e-01 -7.6228803e-01
-1.4453883e+00 1.0172945e+00 2.3532931e-02 -8.7260343e-02
-5.4485154e-01 -1.4851414e-01 4.7225398e-01 2.8506893e-01
7.4259603e-01 -1.4568223e-01 -5.1001161e-01 3.6747238e-01
-6.7068112e-01 2.0530710e-02 7.5788432e-01 7.6986808e-01
6.3036615e-01 -6.9694802e-02 1.3995329e-01 8.6486721e-01
1.9550635e-01 6.5975916e-01 5.6433296e-01 -9.7967911e-01
2.2828239e-01 4.7718075e-01 3.6741011e-02 -1.1076692e+00
-7.1187520e-01 -5.5391420e-02 -1.0509249e+00 8.9318722e-02
2.4984609e-01 -1.8499027e-01 -5.9775430e-01 1.7490761e+00
5.5367470e-01 5.5457419e-01 2.0372669e-01 6.9492185e-01
8.4058791e-01 4.3121096e-01 8.8509604e-02 -5.2919459e-01
1.4109554e+00 -5.7532388e-01 -8.5079676e-01 3.7452850e-01
4.8609906e-01 -6.9478291e-01 4.9021497e-01 4.0836057e-01
-1.0757444e+00 -4.3052515e-01 -7.1877569e-01 3.0800244e-01
-4.0701348e-01 4.7921774e-01 6.0250014e-01 1.4668354e-01
-1.0225756e+00 5.1660156e-01 -1.1748286e+00 -2.6606634e-01
2.4466695e-01 1.8901037e-01 -6.2674278e-01 -7.4719749e-02
-1.0102872e+00 6.3885093e-01 -6.7933418e-02 4.2487796e-02
-6.7396802e-01 -1.0232923e+00 -7.3893392e-01 -1.1130798e-02
6.3027069e-02 -9.3694264e-01 5.3637308e-01 -6.0226768e-01
-9.1868365e-01 5.9822845e-01 -4.4419944e-01 -2.3673154e-01
1.4268151e-01 -1.3156834e-01 -4.6541595e-01 3.0870575e-01
1.8104701e-01 8.4527932e-02 7.1268845e-01 -9.2763418e-01
-3.3078849e-01 -7.6470459e-01 -4.6384183e-01 1.8899076e-01
-2.9786414e-01 8.5555196e-02 -1.2653780e-01 -6.0618466e-01
2.4281718e-01 -8.7715107e-01 -3.1692353e-01 2.2511149e-02
-2.6855662e-01 -9.1745645e-02 8.3399409e-01 -9.1547817e-01
8.4856808e-01 -1.9865296e+00 5.6508130e-01 2.0044641e-01
3.6131552e-01 -8.6822428e-02 9.6284650e-02 6.2362379e-01
-5.7495171e-01 6.7635901e-02 2.4075959e-02 -1.0989003e-01
-4.1117403e-01 -6.2696204e-02 -2.8631598e-01 9.5324224e-01
-1.2222827e-01 7.3367274e-01 -7.5491035e-01 -3.3999321e-01
1.9738007e-01 5.3818446e-01 -3.8122210e-01 3.5056129e-01
2.3964304e-01 3.3282235e-01 -4.2515782e-01 8.2113647e-01
7.2193414e-01 -6.6031688e-01 4.8788077e-01 -4.9953091e-01
1.3938172e-01 -4.9224856e-01 -1.3388481e+00 1.4026058e+00
-6.4280011e-02 2.7912846e-01 2.5613075e-01 -1.2730049e+00
7.0986974e-01 5.8831227e-01 8.7305129e-01 -4.3819588e-02
-1.4778665e-01 -2.0055065e-01 -3.5431042e-01 -7.2061831e-01
-1.8226437e-01 -1.6516815e-01 8.9845464e-02 4.7976455e-01
2.8364542e-01 4.2368585e-01 -3.4151962e-01 2.3871660e-01
8.9303714e-01 -1.8817876e-01 8.6557376e-01 -1.9262695e-01
5.2549899e-01 -2.7685341e-01 7.9525477e-01 5.6730169e-01
-1.2552077e-01 2.4288763e-01 6.5706867e-01 -4.1172507e-01
-8.9499265e-01 -1.4037614e+00 -4.2636201e-01 9.6586239e-01
-2.2189219e-01 -3.5396177e-01 -1.7682591e-01 -5.2410448e-01
3.5192406e-01 2.4550091e-01 -8.0899328e-01 -2.3883397e-02
-5.3991031e-02 -9.2400092e-01 2.8392190e-01 5.1800573e-01
-2.9555595e-01 -3.7560049e-01 -1.7497173e-02 -7.2247177e-02
-1.8411154e-01 -7.7208227e-01 -4.7025207e-01 1.3121009e-01
-9.4861168e-01 -1.4147142e+00 -6.3293159e-01 -5.4144496e-01
7.1578991e-01 3.2461515e-01 5.2269113e-01 -4.1526660e-01
-5.0615305e-01 8.2372206e-01 -6.8193562e-02 -4.9785653e-01
-1.9256653e-01 -4.7022998e-01 6.2444407e-01 1.6208354e-01
1.4881372e-01 -5.9059030e-01 -8.4945267e-01 3.4181908e-01
-7.4586612e-01 -3.5889409e-02 2.0275132e-01 1.0006747e+00
7.3261058e-01 -3.1120604e-01 9.2462629e-01 -9.9697906e-01
7.0452970e-01 -1.1310035e+00 -6.2249500e-01 4.2269313e-01
-8.2591236e-01 -1.1578909e-01 5.8651757e-01 -5.9996325e-01
-8.0639595e-01 2.9173577e-02 6.4875287e-01 -1.0160986e+00
-1.6167374e-02 7.5392801e-01 3.1826340e-02 2.3749141e-01
5.4640269e-01 1.6700098e-01 6.2211251e-01 -4.8059934e-01
5.2254230e-01 5.0245887e-01 2.4452694e-01 -3.6773688e-01
3.9972800e-01 7.5604373e-01 1.7093316e-01 -6.4667356e-01
-3.5614789e-01 -6.6134948e-01 -5.8047652e-01 -1.5093397e-01
8.5929507e-01 -1.3647294e+00 -8.9960480e-01 3.3591396e-01
-5.9380990e-01 1.6924597e-01 7.9828739e-02 9.1992480e-01
-6.6252136e-01 3.4778959e-01 -5.5606574e-01 -3.9415652e-01
-4.6641663e-01 -1.1234815e+00 7.4337238e-01 7.0787683e-02
-1.3759322e-01 -1.2660812e+00 3.8259971e-01 2.7423942e-01
3.1476542e-01 3.9815500e-01 1.1410732e+00 -1.0403600e+00
-2.3122945e-01 -3.0884483e-01 5.3617181e-03 3.0909342e-01
5.3540057e-01 5.4923344e-02 -7.7281511e-01 -4.5657665e-01
1.6699690e-01 -3.1971201e-01 5.9224522e-01 7.1893984e-01
1.3113873e+00 -3.8015765e-01 -5.6902921e-01 8.3004606e-01
1.3711189e+00 1.0940577e-01 3.7652157e-02 -2.5119728e-01
7.8420514e-01 6.4919227e-01 4.6805158e-01 8.4395880e-01
5.1840162e-01 6.2329000e-01 6.1028425e-02 -2.2861885e-01
2.2558722e-01 1.4120159e-01 3.5334164e-01 1.0674447e+00
-6.7913845e-02 1.8374471e-01 -8.3551449e-01 3.6621386e-01
-2.1562510e+00 -1.0790064e+00 -1.8979126e-01 2.1704071e+00
6.5574604e-01 -7.3212600e-01 1.5540054e-01 -7.2636229e-01
7.8798419e-01 -1.9122387e-01 -7.8285980e-01 -2.8458670e-01
2.6593495e-02 6.4909592e-02 4.8621592e-01 5.1520944e-01
-1.0619785e+00 3.4000427e-01 6.8596969e+00 3.9569861e-01
-1.0874550e+00 2.0911767e-01 7.0545542e-01 -3.3785260e-01
-1.6319853e-01 -8.6451754e-02 -5.3659427e-01 1.1646800e-01
1.2844223e+00 -6.2136257e-01 4.2132977e-01 7.1614617e-01
5.7929379e-01 9.1433354e-02 -1.1052228e+00 7.8854436e-01
1.7256027e-01 -1.5037242e+00 -1.3547918e-01 7.5243756e-02
6.7293608e-01 1.2900645e-01 1.6961294e-01 -3.0390410e-02
5.1583850e-01 -9.9222940e-01 -7.1812749e-02 1.0941806e+00
6.4210075e-01 -7.3291856e-01 4.8870024e-01 2.7911943e-01
-1.0944520e+00 -5.1324375e-02 -4.5752877e-01 4.8236084e-01
-1.6501633e-02 6.5073842e-01 -1.2479856e+00 9.0559375e-01
5.7675046e-01 9.9031150e-01 -3.8725725e-01 7.4071354e-01
3.2954174e-01 5.9852463e-01 4.9752064e-02 5.7874870e-01
-4.5560026e-01 -4.2017177e-01 4.5304745e-01 9.8330891e-01
4.8414907e-01 1.7525466e-01 3.4570563e-01 7.1723592e-01
2.5126633e-01 4.5403901e-01 -8.1883532e-01 -6.8274438e-03
3.7549922e-01 1.4287479e+00 -5.1131171e-01 -5.3389841e-01
-8.7145108e-01 7.1122324e-01 4.8827252e-01 6.1543947e-01
-9.0649921e-01 -4.7422744e-02 1.0168897e+00 -1.7474766e-01
3.1932265e-01 2.8503109e-02 -9.4808675e-02 -1.5433393e+00
-2.3636802e-01 -1.0263695e+00 8.8701171e-01 -6.9576722e-01
-1.6122136e+00 1.5033065e-01 3.0414316e-01 -1.4401149e+00
-2.8673866e-01 -4.8415902e-01 -4.0366566e-01 1.1941518e+00
-1.0341321e+00 -1.1142110e+00 -6.8443529e-02 8.3435464e-01
1.8215790e-01 -2.9992491e-01 1.1683060e+00 3.6562867e-02
-5.3709263e-01 2.9650196e-01 6.1150730e-01 -9.7585283e-02
1.1906914e+00 -1.1429731e+00 -2.7227935e-01 3.8163808e-01
-1.7024361e-01 9.7515845e-01 5.2091330e-01 -7.7775836e-01
-1.5632290e+00 -1.1746700e+00 3.6292848e-01 -3.9518362e-01
8.2531202e-01 -3.2042825e-01 -1.0798389e+00 9.6390176e-01
2.5114632e-01 4.1918716e-01 1.4797747e+00 1.8851098e-01
-4.3432426e-01 -2.1138430e-01 -1.1666405e+00 2.5449881e-01
4.9497783e-01 -4.3663689e-01 -3.9210656e-01 5.6877857e-01
4.8436287e-01 -2.4183232e-01 -1.5279927e+00 4.1861889e-01
6.7060918e-01 -5.8690560e-01 1.2080948e+00 -1.0921556e+00
1.6904049e-01 -3.4874475e-01 -3.4949970e-01 -1.5290101e+00
-4.2691785e-01 -2.4057871e-01 -2.7687404e-01 1.0356736e+00
1.9576867e-01 -9.5912492e-01 3.8840008e-01 5.3499424e-01
2.5346601e-01 -8.2127756e-01 -9.7479033e-01 -3.9364174e-01
1.9862343e-02 -9.7146826e-03 2.8704458e-01 1.3617551e+00
3.8911214e-01 3.1610529e-04 -6.3769132e-01 6.7846411e-01
8.9597690e-01 4.7265416e-01 6.6230500e-01 -1.0421486e+00
-4.9452579e-01 -3.1244997e-02 -5.6482226e-01 2.3083918e-02
3.1036568e-01 -1.2550863e+00 -6.1586463e-01 -1.2272955e+00
7.0878547e-01 -3.4203753e-01 -6.7973655e-01 5.4087812e-01
-3.1252345e-01 -2.3836750e-01 -1.5036266e-01 2.2375982e-01
-2.6001379e-01 2.8598422e-01 8.0631858e-01 -8.0721386e-02
-2.7514404e-01 8.9615107e-02 -7.6562309e-01 6.6911042e-01
6.7172617e-01 -3.1774348e-01 -6.3800633e-01 3.6871383e-01
-1.0855858e-01 7.0549148e-01 4.0606347e-01 -3.7946165e-01
2.2497074e-01 -4.2386785e-01 8.0744398e-01 -5.2030987e-01
2.9323500e-01 -7.6256937e-01 6.5694642e-01 4.2895338e-01
-3.1519306e-01 3.8688022e-01 1.3465104e-01 8.2596320e-01
3.6898531e-02 2.6887724e-01 5.2788401e-01 -2.0533094e-02
-2.6902637e-01 3.7988073e-01 -3.9092788e-01 -6.4502150e-02
1.2696490e+00 3.0324504e-01 -5.5628496e-01 -5.1349324e-01
-1.2540915e+00 4.4152996e-01 2.6060867e-01 4.6182448e-01
8.7301672e-01 -1.4648335e+00 -8.4080571e-01 3.8017425e-01
2.1483877e-01 -5.8075273e-01 5.3674436e-01 1.1962935e+00
-4.1542690e-02 3.5635930e-01 -3.1050348e-01 -7.6894164e-01
-1.7245711e+00 9.2830098e-01 3.1303927e-01 -1.0193745e-01
-5.6965774e-01 3.0031136e-01 4.8569331e-01 -5.8435851e-01
4.7464702e-02 -2.1019289e-01 -3.0551678e-01 4.1333836e-01
6.9870573e-01 5.8637679e-01 -1.2756926e-01 -6.1085182e-01
-5.8598673e-01 3.8062242e-01 -1.5015517e-01 1.5553966e-01
1.6338071e+00 -1.9875070e-01 -4.2386910e-01 8.4793180e-01
1.4345683e+00 -3.4195807e-02 -7.2014773e-01 -3.0310869e-01
-3.7508538e-01 -2.5185975e-01 -1.6607456e-01 -5.4388231e-01
-9.5003980e-01 4.3164426e-01 9.5405763e-01 -1.1698410e-01
1.1141683e+00 2.6502737e-01 -2.9351749e-02 1.1212775e-01
1.0426092e-01 -6.3496399e-01 -8.6393170e-02 -2.4728868e-02
8.7685776e-01 -1.1308919e+00 2.6855057e-01 -3.2204369e-01
-8.0374914e-01 1.2841470e+00 3.9014834e-01 -2.0226236e-02
1.1429105e+00 2.0129544e-01 1.8709959e-01 -3.8615230e-01
-1.2699203e+00 2.8960702e-01 4.1303337e-01 4.7758606e-01
5.1839560e-01 4.4714507e-01 -3.6474425e-01 6.0102367e-01
3.5538641e-01 1.3430093e-01 5.4641598e-01 6.7435986e-01
-1.1532892e-01 -1.0709724e+00 -6.0311764e-01 8.1793338e-01
-5.1523882e-01 3.5190146e-02 -8.1874570e-03 5.8940923e-01
-2.3257595e-01 5.9220326e-01 -9.2634819e-02 -2.9292992e-01
2.5912818e-01 6.0149980e-01 2.2556788e-01 -4.0191156e-01
-6.7399472e-02 3.5551795e-01 -1.7564116e-01 -8.8041788e-01
-3.9900249e-01 -1.1601282e+00 -1.1828609e+00 -1.7205077e-01
-1.0010341e-01 5.0361823e-02 4.3586534e-01 5.2400964e-01
7.3061424e-01 3.2969442e-01 7.9202461e-01 -5.2574545e-01
-6.9096148e-01 -8.1147885e-01 -7.9716575e-01 4.4065449e-01
5.1668060e-01 -6.8870765e-01 -3.1429508e-01 3.0920774e-01] | [6.769935607910156, 5.376345634460449] |
238790be-9926-45cf-8da7-293869a9e025 | multi-span-acoustic-modelling-using-raw | 1906.11047 | null | https://arxiv.org/abs/1906.11047v2 | https://arxiv.org/pdf/1906.11047v2.pdf | Multi-Span Acoustic Modelling using Raw Waveform Signals | Traditional automatic speech recognition (ASR) systems often use an acoustic model (AM) built on handcrafted acoustic features, such as log Mel-filter bank (FBANK) values. Recent studies found that AMs with convolutional neural networks (CNNs) can directly use the raw waveform signal as input. Given sufficient training data, these AMs can yield a competitive word error rate (WER) to those built on FBANK features. This paper proposes a novel multi-span structure for acoustic modelling based on the raw waveform with multiple streams of CNN input layers, each processing a different span of the raw waveform signal. Evaluation on both the single channel CHiME4 and AMI data sets show that multi-span AMs give a lower WER than FBANK AMs by an average of about 5% (relative). Analysis of the trained multi-span model reveals that the CNNs can learn filters that are rather different to the log Mel filters. Furthermore, the paper shows that a widely used single span raw waveform AM can be improved by using a smaller CNN kernel size and increased stride to yield improved WERs. | ['Philip Woodland', 'Patrick von Platen', 'Chao Zhang'] | 2019-06-21 | null | null | null | null | ['acoustic-modelling'] | ['speech'] | [ 2.50369072e-01 -7.87010267e-02 4.90367532e-01 -4.79343623e-01
-1.05953586e+00 -3.15146595e-01 3.14914674e-01 6.30906150e-02
-9.47542548e-01 2.94656098e-01 1.54375225e-01 -6.10674739e-01
1.92854047e-01 -5.37453771e-01 -6.36396110e-01 -4.77651119e-01
-7.62289315e-02 -1.13021761e-01 2.14022353e-01 -4.91178811e-01
-1.57507405e-01 4.78030682e-01 -1.54321885e+00 4.42513883e-01
3.04358065e-01 1.28445005e+00 2.65354127e-01 1.29342127e+00
-8.83740261e-02 4.83643621e-01 -1.28132212e+00 -3.40764374e-01
3.55979830e-01 -3.28446746e-01 -4.57145542e-01 -3.41164976e-01
5.95798671e-01 -2.74015099e-01 -7.42489398e-01 8.09851885e-01
8.77328932e-01 4.08751756e-01 1.60114020e-01 -7.41367280e-01
-5.65166652e-01 1.18662488e+00 1.55773714e-01 4.27313328e-01
5.29657863e-02 2.40876660e-01 1.10606050e+00 -1.15966606e+00
-3.16804498e-01 1.19152224e+00 1.00248504e+00 4.41642761e-01
-1.12077832e+00 -8.40354443e-01 -1.77090406e-01 3.64428043e-01
-1.56827855e+00 -8.73450160e-01 5.44313431e-01 3.06542460e-02
1.58590138e+00 4.45863605e-01 4.95727926e-01 1.00989985e+00
-6.13710396e-02 5.11203766e-01 6.32109284e-01 -6.39349699e-01
1.77978411e-01 -2.07943827e-01 8.17254484e-02 3.47660333e-01
-9.65023786e-02 1.12437062e-01 -8.56180191e-01 9.15956497e-02
4.25370663e-01 -6.20659292e-01 -3.60829711e-01 7.90690958e-01
-1.05703259e+00 8.13438892e-01 2.43707061e-01 3.42540085e-01
-2.88901746e-01 4.53917712e-01 5.12593150e-01 8.34240913e-01
3.02504539e-01 5.72018027e-01 -6.91357791e-01 -4.26188588e-01
-1.23272705e+00 8.93182606e-02 7.83523023e-01 7.35844433e-01
4.57203001e-01 1.02491343e+00 -8.59883949e-02 1.16511428e+00
2.69286007e-01 7.59349525e-01 7.10287690e-01 -4.77907360e-01
4.59914118e-01 -1.52013540e-01 -1.77472323e-01 -5.47508001e-01
-3.16238791e-01 -6.23939693e-01 -6.99810445e-01 1.76186204e-01
5.68519056e-01 -2.58101583e-01 -1.37029123e+00 1.64762318e+00
-3.04020345e-01 1.02953196e-01 1.27680749e-01 6.42198801e-01
8.85260940e-01 9.46425855e-01 -1.78088829e-01 1.68222919e-01
1.25911319e+00 -8.24155152e-01 -9.74055111e-01 -3.65060687e-01
4.59474176e-01 -1.03252149e+00 1.19945395e+00 7.14166045e-01
-1.26723468e+00 -8.87954354e-01 -1.50805163e+00 -3.95368971e-02
-6.14809513e-01 3.86413991e-01 -3.57264243e-02 1.20335209e+00
-1.36937606e+00 7.35561013e-01 -6.21316612e-01 -2.55242232e-02
6.81456029e-02 4.80078846e-01 -1.69329584e-01 2.37003461e-01
-1.46964216e+00 9.82536256e-01 6.10473156e-01 2.49347016e-01
-8.34359527e-01 -9.20006216e-01 -9.74622309e-01 2.93500096e-01
-5.23481146e-02 7.21334992e-03 1.63894272e+00 -7.97486424e-01
-2.14829350e+00 1.66389555e-01 -1.64256785e-02 -1.01203907e+00
1.82705224e-01 -4.24518198e-01 -9.87677157e-01 1.59642115e-01
-6.68940842e-01 4.87205863e-01 1.06900108e+00 -5.10921538e-01
-5.17882526e-01 2.86956370e-01 -2.62304097e-01 -3.26745808e-01
-6.63139939e-01 1.12266354e-01 8.92080516e-02 -8.88278186e-01
-8.04220662e-02 -6.89827979e-01 -3.04881558e-02 -3.85443896e-01
-2.44829699e-01 -2.14337841e-01 7.85359621e-01 -9.70902681e-01
1.53118634e+00 -2.31342053e+00 -3.77756417e-01 2.91382432e-01
-2.31941238e-01 8.37806106e-01 -3.65350455e-01 4.57679510e-01
-2.37494454e-01 2.08597541e-01 -2.28293151e-01 -4.17157292e-01
2.02556532e-02 5.42187952e-02 -3.64595801e-01 3.01867992e-01
6.87434971e-01 6.34651124e-01 -5.48008800e-01 2.48902723e-01
3.00154597e-01 7.26009071e-01 -4.13270622e-01 3.24205339e-01
-1.78425033e-02 8.45477823e-03 4.88286942e-01 6.13989718e-02
5.54950535e-01 2.40159273e-01 -2.38491505e-01 -2.11214319e-01
-2.02911794e-01 8.28341484e-01 -1.01890242e+00 1.51903248e+00
-8.33175361e-01 1.20067573e+00 7.06745163e-02 -7.61607051e-01
1.28829658e+00 7.81782150e-01 -4.99049611e-02 -6.88221514e-01
1.31950691e-01 5.21834135e-01 6.01737678e-01 -7.94207007e-02
5.27021706e-01 -8.48177001e-02 1.98334306e-01 1.95914015e-01
3.70044053e-01 -3.58485729e-01 -1.54815093e-01 -2.03582689e-01
1.32508016e+00 -6.22154415e-01 -1.14561534e-02 7.07217492e-03
5.89031160e-01 -4.69081432e-01 1.10870294e-01 9.00651932e-01
1.10918574e-01 9.39083278e-01 -5.64666279e-02 -3.92310441e-01
-1.52159131e+00 -1.10802245e+00 -2.48973668e-01 9.94369984e-01
-7.17072964e-01 -6.53814793e-01 -7.40393400e-01 4.34162393e-02
-1.87392294e-01 8.35479379e-01 -2.70656526e-01 -2.48653620e-01
-8.72190416e-01 -5.28593242e-01 1.45224380e+00 7.91698515e-01
3.89474094e-01 -1.14827538e+00 -4.15371180e-01 5.94869852e-01
1.35164529e-01 -1.36117649e+00 -4.35614944e-01 6.66899323e-01
-4.51435238e-01 -3.21878850e-01 -8.17362249e-01 -5.91563880e-01
-4.58784029e-03 -7.24816620e-02 8.15236568e-01 1.07556894e-01
-3.91198024e-02 9.13065225e-02 -7.63006926e-01 -7.19637096e-01
-6.84168339e-01 2.18879372e-01 3.59977931e-01 1.56343341e-01
2.94153035e-01 -5.90645313e-01 -3.79281878e-01 3.62666957e-02
-9.72692549e-01 -1.93477541e-01 6.03589594e-01 1.02700949e+00
7.11146668e-02 -9.57310051e-02 1.07547891e+00 -2.05712810e-01
6.99406683e-01 -1.86110929e-01 -3.96298438e-01 -1.75843701e-01
-4.61589217e-01 -2.30396856e-02 9.39666271e-01 -7.68548071e-01
-7.55567372e-01 -1.31973803e-01 -9.14107740e-01 -5.65956533e-01
-1.77155808e-01 5.36515951e-01 1.12198867e-01 5.17145023e-02
8.26875269e-01 2.80783385e-01 4.12859060e-02 -7.22572923e-01
3.57147872e-01 1.19027686e+00 5.37133694e-01 -2.13878945e-01
7.26278722e-01 2.46600546e-02 -5.42361259e-01 -1.31047451e+00
-5.62140167e-01 -3.52278143e-01 -3.69221330e-01 -8.44895542e-02
6.87436104e-01 -8.07459950e-01 -5.84222317e-01 5.91047943e-01
-1.26317704e+00 -4.94705737e-01 -3.42885643e-01 8.51226270e-01
-3.50474060e-01 3.90903987e-02 -1.00080657e+00 -9.94960546e-01
-6.01455152e-01 -1.01868403e+00 8.69773269e-01 5.00997491e-02
-4.29008633e-01 -7.45584369e-01 -2.92046070e-01 2.86229756e-02
1.08705533e+00 -3.35312635e-01 9.26179886e-01 -1.07836652e+00
-1.98731609e-02 -4.39038396e-01 7.19817132e-02 9.78979051e-01
6.78102598e-02 5.04279360e-02 -1.59594965e+00 -3.40192825e-01
3.08373850e-02 -7.33327642e-02 8.62655520e-01 3.16557854e-01
1.22787571e+00 -2.28698850e-01 5.71778953e-01 5.11645317e-01
1.05318069e+00 5.98467469e-01 8.74364972e-01 1.74631745e-01
4.36239630e-01 -1.55017488e-02 1.91669594e-02 4.26278412e-01
-2.74392605e-01 6.79298222e-01 8.90787989e-02 -5.57125807e-02
-5.91097832e-01 1.73735097e-02 8.51217985e-01 1.67492676e+00
1.79712042e-01 -2.99246311e-01 -8.87137055e-01 6.76886678e-01
-1.26833236e+00 -8.47630560e-01 -1.23470493e-01 2.22010517e+00
1.01214230e+00 3.86010259e-01 2.49673184e-02 7.25485981e-01
6.09844625e-01 1.33562848e-01 -1.31327555e-01 -9.74859536e-01
-2.67316878e-01 1.11697400e+00 4.87887859e-01 4.95820522e-01
-6.89723670e-01 7.58685708e-01 6.80729437e+00 1.13333118e+00
-1.46732938e+00 1.89477995e-01 2.98057497e-01 -1.75102100e-01
-4.90701608e-02 -5.34935296e-01 -7.26404786e-01 2.80909538e-01
1.99248803e+00 -2.63575613e-02 5.91986716e-01 5.51898181e-01
3.15696359e-01 3.11429113e-01 -1.16513002e+00 1.05695379e+00
-2.51281653e-02 -1.22791028e+00 6.76755533e-02 -1.27662197e-01
2.51268864e-01 2.53577083e-01 3.13958287e-01 4.34877962e-01
1.45932809e-01 -1.45367622e+00 1.11820841e+00 2.57330030e-01
1.03824878e+00 -9.35995817e-01 9.00562167e-01 -2.75871567e-02
-1.41511893e+00 -1.61321074e-01 -4.52167690e-01 -1.67052358e-01
1.74118392e-02 4.96498466e-01 -1.36274266e+00 3.65447551e-01
7.47003913e-01 1.24899991e-01 -4.31559891e-01 1.14589107e+00
8.96823332e-02 1.27565372e+00 -5.20244241e-01 -9.87900421e-02
4.17444050e-01 4.19797599e-01 5.33450127e-01 1.79504514e+00
6.17553592e-01 -1.72273368e-01 -2.95946538e-01 6.38325751e-01
-5.14658242e-02 -1.43480795e-02 -3.49615425e-01 -3.43167931e-01
6.49159551e-01 9.54460919e-01 -2.35758230e-01 -3.32817584e-01
-4.98535991e-01 5.49897552e-01 1.38907945e-02 4.44991857e-01
-6.73440337e-01 -9.04289722e-01 8.45906019e-01 -4.25250866e-02
6.26670718e-01 -5.70407569e-01 -3.32847565e-01 -5.87069690e-01
-1.73168927e-01 -9.86513972e-01 -6.90567717e-02 -7.58007705e-01
-1.28399026e+00 1.04369617e+00 -3.57802749e-01 -1.10820973e+00
-2.29563013e-01 -9.88196850e-01 -7.36882389e-01 1.30691266e+00
-1.40585899e+00 -7.52237082e-01 3.43715884e-02 4.99464750e-01
9.88966107e-01 -4.34412152e-01 1.14852571e+00 5.54796636e-01
-3.63800526e-01 1.01385117e+00 -1.07012972e-01 5.67286491e-01
3.88650507e-01 -1.24310827e+00 1.01151085e+00 8.59312713e-01
4.54479218e-01 5.32140553e-01 6.11665487e-01 -2.10140303e-01
-1.14885378e+00 -1.05882609e+00 7.07935989e-01 -1.76192924e-01
6.73390865e-01 -4.80935752e-01 -1.15949416e+00 4.07996267e-01
4.85086322e-01 -9.15609673e-02 6.96843624e-01 -2.49738246e-02
-5.57701886e-01 -2.55747765e-01 -7.59338021e-01 5.90868294e-01
5.99027455e-01 -8.93579841e-01 -6.85438216e-01 -2.12625146e-01
1.21295238e+00 -3.36857706e-01 -7.77940512e-01 3.51479858e-01
5.16648650e-01 -5.93524158e-01 9.75631297e-01 -5.57538331e-01
6.33235350e-02 -3.37345630e-01 -4.67262894e-01 -1.78382909e+00
-1.28331602e-01 -7.68950701e-01 -2.30922967e-01 9.71495748e-01
8.24676991e-01 -5.40388167e-01 2.04938233e-01 2.08971724e-01
-7.07002699e-01 -5.48898876e-01 -1.13039911e+00 -1.16417205e+00
5.79704903e-02 -1.22971654e+00 6.73295021e-01 4.16534364e-01
-1.65570959e-01 1.24394283e-01 -3.32530886e-01 4.04089928e-01
4.91286702e-02 -8.45709383e-01 4.66450542e-01 -1.02420521e+00
-2.58840621e-01 -4.80327815e-01 -5.12397707e-01 -8.61824095e-01
-1.63837701e-01 -7.70257950e-01 3.94237757e-01 -9.67571676e-01
-8.18805873e-01 -2.43310675e-01 -6.53711200e-01 5.16703963e-01
1.11196429e-01 3.64106745e-01 3.60826164e-01 -4.58970666e-01
3.28871645e-02 4.81217533e-01 7.86262512e-01 -1.94163382e-01
-1.37376934e-01 1.30384611e-02 -8.02343041e-02 5.99014401e-01
9.68954682e-01 -4.67832386e-01 -2.03800455e-01 -5.20181417e-01
1.02158494e-01 -3.65216881e-02 1.64862111e-01 -1.53287458e+00
3.90039265e-01 3.39123011e-01 5.43302476e-01 -5.20970941e-01
8.72573972e-01 -4.73783195e-01 1.11429216e-02 3.71572852e-01
-6.32592201e-01 2.88594604e-01 6.69472516e-01 2.70703554e-01
-3.64562184e-01 -3.45825136e-01 7.79199362e-01 2.16425881e-02
-5.30230880e-01 -1.00638561e-01 -7.66355097e-01 -1.90968946e-01
1.74096659e-01 -1.78707883e-01 -9.20572057e-02 -6.31442666e-01
-8.26589525e-01 -4.57291335e-01 -3.78036976e-01 4.32691216e-01
8.46264660e-01 -1.35916638e+00 -9.36593771e-01 5.53541601e-01
-1.45514116e-01 -2.07201988e-01 4.16740552e-02 6.17437601e-01
-5.25838256e-01 4.74739254e-01 -2.43137572e-02 -4.56555486e-01
-1.15730095e+00 -5.27861528e-03 5.22958279e-01 2.67349452e-01
-6.61632419e-01 1.17367494e+00 -3.14795524e-01 -2.43360281e-01
4.73016769e-01 -6.70272410e-01 5.20978868e-02 7.65780266e-03
9.46449459e-01 3.26469421e-01 8.04855227e-01 -5.17511666e-01
-2.23971158e-01 7.83897266e-02 1.33631309e-03 -6.59981668e-01
1.39212227e+00 1.30129591e-01 3.62927079e-01 6.98027253e-01
1.28013217e+00 1.90616980e-01 -9.07166839e-01 -3.65436047e-01
1.42396733e-01 -2.40694121e-01 3.20277512e-01 -8.67910445e-01
-9.66860414e-01 1.18127990e+00 6.57578826e-01 5.64780295e-01
1.04195964e+00 -3.22300404e-01 1.15557337e+00 5.71124732e-01
1.04349911e-01 -1.20697629e+00 1.57317847e-01 1.03915262e+00
1.06285846e+00 -6.61936998e-01 -6.43594027e-01 5.12676239e-02
-3.38712513e-01 1.61029434e+00 3.21999282e-01 -1.58769011e-01
8.36631596e-01 7.01246977e-01 4.57056969e-01 1.57225803e-01
-7.52336264e-01 -4.05379146e-01 3.31351489e-01 5.04887164e-01
5.50940573e-01 4.61092517e-02 1.23056568e-01 6.26527846e-01
-9.14526403e-01 -4.84428763e-01 6.72114491e-01 5.62240243e-01
-6.62628293e-01 -1.00847054e+00 -6.43503129e-01 5.24288118e-01
-6.06781006e-01 -5.60315430e-01 -3.07008047e-02 4.53975588e-01
-1.20528162e-01 1.32259381e+00 4.23393458e-01 -8.45274389e-01
4.09390360e-01 5.13041198e-01 1.96935132e-01 -7.08682239e-01
-9.82760310e-01 9.67039913e-02 1.99648932e-01 -2.06790179e-01
6.26152754e-02 -2.25325301e-01 -1.20373011e+00 -2.44013548e-01
-5.09332120e-01 -4.84743156e-03 1.02990270e+00 9.53280389e-01
2.59222575e-02 1.01629519e+00 5.67991734e-01 -8.06393862e-01
-7.69572198e-01 -1.44522786e+00 -4.93089318e-01 1.31386518e-01
8.38973582e-01 -1.13576837e-01 -4.52333659e-01 9.18029249e-02] | [14.816458702087402, 6.0371599197387695] |
236f7ea2-3678-44be-b2f3-2d958303f18f | convolutional-sequence-to-sequence-learning | 1705.03122 | null | http://arxiv.org/abs/1705.03122v3 | http://arxiv.org/pdf/1705.03122v3.pdf | Convolutional Sequence to Sequence Learning | The prevalent approach to sequence to sequence learning maps an input
sequence to a variable length output sequence via recurrent neural networks. We
introduce an architecture based entirely on convolutional neural networks.
Compared to recurrent models, computations over all elements can be fully
parallelized during training and optimization is easier since the number of
non-linearities is fixed and independent of the input length. Our use of gated
linear units eases gradient propagation and we equip each decoder layer with a
separate attention module. We outperform the accuracy of the deep LSTM setup of
Wu et al. (2016) on both WMT'14 English-German and WMT'14 English-French
translation at an order of magnitude faster speed, both on GPU and CPU. | ['Jonas Gehring', 'Michael Auli', 'David Grangier', 'Yann N. Dauphin', 'Denis Yarats'] | 2017-05-08 | convolutional-sequence-to-sequence-learning-1 | https://icml.cc/Conferences/2017/Schedule?showEvent=806 | http://proceedings.mlr.press/v70/gehring17a/gehring17a.pdf | icml-2017-8 | ['bangla-spelling-error-correction'] | ['natural-language-processing'] | [ 5.91902494e-01 -8.05346891e-02 -1.81169763e-01 -9.14514884e-02
-9.23355341e-01 -8.81646931e-01 5.78266978e-01 -9.43939239e-02
-8.27929795e-01 9.81100559e-01 1.28575549e-01 -9.34114993e-01
6.32211506e-01 -5.92869937e-01 -9.89373386e-01 -3.42509538e-01
-4.07073274e-02 3.98319215e-01 -1.51867986e-01 -2.90574491e-01
1.85783520e-01 1.35535657e-01 -9.38351035e-01 5.78987360e-01
3.79293144e-01 7.96770692e-01 4.36959535e-01 1.30529010e+00
-1.30338281e-01 1.00746441e+00 -4.00313079e-01 -4.29920971e-01
3.67280781e-01 -4.73207235e-01 -9.14505720e-01 -3.10584754e-01
3.98643583e-01 -1.98193982e-01 -4.04711694e-01 7.60934770e-01
6.80119932e-01 1.81997135e-01 4.71662521e-01 -5.22318006e-01
-8.70263159e-01 8.37520838e-01 -2.55149454e-01 3.82718503e-01
2.46885389e-01 3.35836202e-01 1.04567933e+00 -1.09477103e+00
7.18441188e-01 1.10596704e+00 1.07110572e+00 4.58549023e-01
-1.40072381e+00 -4.27237719e-01 1.39275372e-01 9.16054398e-02
-1.05797517e+00 -5.83749831e-01 -2.53553148e-02 -2.33477756e-01
2.00796318e+00 1.54136956e-01 6.88211083e-01 1.23861361e+00
5.39778113e-01 6.89864039e-01 9.40831244e-01 -5.06773591e-01
-2.07619160e-01 -2.49195486e-01 -1.39587764e-02 8.10074687e-01
-3.44341010e-01 1.23467438e-01 -3.40215325e-01 4.87641990e-02
7.50249326e-01 -2.10685972e-02 1.91350803e-02 2.09860936e-01
-1.35122108e+00 9.38060701e-01 3.68552536e-01 1.70120403e-01
-3.04689646e-01 7.26335943e-01 9.91943777e-01 7.91617334e-01
5.63945055e-01 4.39728051e-01 -8.32275689e-01 -5.64902663e-01
-9.62926507e-01 5.99321239e-02 8.45739305e-01 1.14464390e+00
7.17117012e-01 2.87290007e-01 -2.55206764e-01 7.70955503e-01
-1.32996246e-01 4.98010576e-01 7.73768663e-01 -6.12928092e-01
6.80866838e-01 1.00307517e-01 -1.25865281e-01 -5.26313603e-01
-4.56858188e-01 -5.67510486e-01 -7.24255502e-01 -2.02327128e-02
2.36095503e-01 -5.56057632e-01 -1.04458487e+00 1.54484069e+00
-4.55912322e-01 1.09044321e-01 9.80512425e-02 7.23107994e-01
4.80983913e-01 8.67956221e-01 -1.12610891e-01 -1.42636418e-01
1.07149494e+00 -1.44361317e+00 -5.36095679e-01 -5.24895310e-01
1.00822675e+00 -1.16087794e+00 1.18371451e+00 -1.01904742e-01
-1.35934269e+00 -3.57440412e-01 -1.05284441e+00 -5.67321062e-01
-3.43093336e-01 2.93161869e-01 5.66515923e-01 4.60717827e-01
-1.54505765e+00 7.69860446e-01 -7.82306731e-01 -2.52102554e-01
2.18521535e-01 5.62035084e-01 -2.09039658e-01 2.00917363e-01
-1.31497693e+00 1.46987498e+00 5.72983325e-01 2.25894064e-01
-7.14511931e-01 -7.50450730e-01 -8.30412090e-01 -3.03451661e-02
-1.03909925e-01 -8.71276021e-01 1.74036860e+00 -1.13311434e+00
-1.97285867e+00 8.71730328e-01 -5.31045735e-01 -1.04139388e+00
5.87106764e-01 -2.53909290e-01 -1.35307372e-01 -2.39763767e-01
-2.21454412e-01 6.62103355e-01 6.96184278e-01 -2.60197014e-01
-4.27327663e-01 2.33353496e-01 -1.71075016e-01 2.05305144e-01
5.60616609e-05 2.41319850e-01 -2.49989524e-01 -6.01214886e-01
-2.39824340e-01 -1.26026678e+00 -4.20453668e-01 -3.48831952e-01
-2.18510181e-01 -2.44273641e-03 5.66370368e-01 -8.98653209e-01
1.25157392e+00 -1.71324825e+00 2.96333998e-01 -1.16463259e-01
4.84814793e-02 3.50139827e-01 -3.03148210e-01 6.56481564e-01
-1.35631248e-01 2.15087757e-01 -5.46693444e-01 -4.20215428e-01
-2.63275541e-02 3.04267555e-01 -5.00139892e-01 4.17975485e-01
4.26335990e-01 1.50012076e+00 -8.80913436e-01 -1.48107097e-01
5.87005801e-02 4.51240897e-01 -5.10917008e-01 7.08723143e-02
-4.44494009e-01 3.01751435e-01 1.57647833e-01 2.23634124e-01
2.79044151e-01 -4.02626514e-01 3.59356731e-01 4.07834560e-01
-3.99055123e-01 1.10125518e+00 -4.08079833e-01 2.04063869e+00
-1.03593242e+00 1.15313935e+00 -4.28421408e-01 -6.62600517e-01
8.43766212e-01 5.77440381e-01 -6.64806142e-02 -7.31505930e-01
8.38847086e-02 5.23373008e-01 1.21994898e-01 -2.79772162e-01
6.70926213e-01 -2.49802396e-01 -8.75913724e-02 7.97033072e-01
1.35436565e-01 4.33063433e-02 2.49864295e-01 -8.99568722e-02
1.15276325e+00 4.41159993e-01 1.87926874e-01 -2.77944982e-01
2.19599128e-01 -9.97451395e-02 2.66914427e-01 8.35086465e-01
3.29978198e-01 6.42030835e-01 5.30146122e-01 -7.11221099e-01
-1.68142259e+00 -7.68838108e-01 9.09232199e-02 1.51359951e+00
-6.47281110e-01 -5.50217748e-01 -6.25029802e-01 -3.01194459e-01
-2.54460961e-01 6.48826063e-01 -6.91786230e-01 1.33810222e-01
-1.21359253e+00 -6.15275860e-01 9.41105306e-01 5.79135120e-01
2.60020971e-01 -1.29857528e+00 -7.38524675e-01 7.33996391e-01
-2.09156618e-01 -1.05155504e+00 -7.73045003e-01 7.04689264e-01
-9.89896357e-01 -3.93602163e-01 -8.11342716e-01 -9.71846163e-01
3.46301347e-01 -1.20301165e-01 1.56870472e+00 9.90462229e-02
-1.12798154e-01 -3.15013647e-01 -5.74523397e-02 -2.77372241e-01
-6.78724587e-01 8.44908476e-01 -2.67054170e-01 -5.49180210e-01
2.03089744e-01 -5.74935734e-01 -3.47400218e-01 -3.19698244e-01
-5.53469658e-01 4.49047953e-01 6.33445084e-01 1.11405051e+00
3.55755538e-01 -7.50738144e-01 2.47113779e-01 -1.04654551e+00
7.28260159e-01 -4.62911040e-01 -5.67928016e-01 1.86373055e-01
-4.27291602e-01 4.21446979e-01 9.09860969e-01 -4.99558538e-01
-6.62191331e-01 6.39319420e-02 -4.76743400e-01 -2.85000294e-01
3.81596893e-01 6.28969133e-01 5.36005259e-01 -2.31729671e-02
5.26090384e-01 4.17321086e-01 2.92390138e-02 -3.17281038e-01
5.62121570e-01 3.95914286e-01 3.48170012e-01 -3.55047464e-01
3.42520207e-01 -1.79817989e-01 -2.15912059e-01 -3.79382193e-01
-5.24030566e-01 2.86571179e-02 -8.14754486e-01 1.80796742e-01
6.52969003e-01 -9.63019013e-01 -9.07470703e-01 3.54281396e-01
-1.55129457e+00 -7.58459508e-01 -7.62230456e-02 3.46116483e-01
-8.05989027e-01 -5.22498414e-03 -1.22651184e+00 -4.13353086e-01
-8.62040579e-01 -1.22021425e+00 9.63963985e-01 -2.36620098e-01
-5.00180066e-01 -1.27954900e+00 8.25152174e-02 -4.38416153e-01
7.86624432e-01 -8.61451551e-02 9.53593910e-01 -3.62635493e-01
-6.74566805e-01 -1.01913199e-01 -2.92140990e-01 1.98450223e-01
-1.61657736e-01 -8.65446702e-02 -7.67109990e-01 -5.59072495e-01
-1.07976958e-01 -4.67142671e-01 1.09864736e+00 1.87244281e-01
8.37769508e-01 -5.73327661e-01 3.76676931e-03 8.75189424e-01
1.45872235e+00 -9.92083848e-02 7.21378148e-01 5.95722556e-01
6.83517992e-01 1.35550842e-01 6.72287345e-02 1.76346719e-01
1.04712181e-01 4.52362806e-01 2.61570923e-02 -2.50921845e-01
-1.91023737e-01 -3.27146620e-01 7.52451062e-01 1.32547081e+00
1.40112303e-02 -1.89214274e-01 -9.61811125e-01 4.32098597e-01
-1.82581866e+00 -1.01854026e+00 -4.17687073e-02 1.95974541e+00
1.05247188e+00 2.76495069e-01 3.42155397e-02 -2.02727884e-01
3.85161877e-01 1.15611926e-01 -4.02606905e-01 -1.28797150e+00
-2.32451603e-01 6.73143029e-01 1.06377602e+00 8.04618657e-01
-7.68432379e-01 1.36755455e+00 7.82143736e+00 6.81141794e-01
-1.51881063e+00 2.55638182e-01 6.27621889e-01 -4.23682809e-01
-3.60526830e-01 3.10271922e-02 -7.87193060e-01 5.01074612e-01
1.63272822e+00 -8.31061527e-02 6.46531463e-01 4.99071956e-01
5.20911813e-02 2.81815141e-01 -1.10266030e+00 7.72166133e-01
-1.17289551e-01 -1.63512099e+00 -8.05527940e-02 -1.19397484e-01
7.72565126e-01 7.30208933e-01 2.40636051e-01 4.48668599e-01
5.35086393e-01 -1.53715360e+00 8.11466813e-01 5.03148019e-01
1.07336974e+00 -8.86070013e-01 5.32553375e-01 1.72137588e-01
-1.08864272e+00 3.72512918e-03 -5.19321322e-01 -5.42550087e-01
2.05062389e-01 3.21263403e-01 -1.24387002e+00 1.92796364e-01
4.16220874e-01 8.69449914e-01 -4.05873328e-01 4.96548772e-01
-3.87847662e-01 8.16806555e-01 -3.20524961e-01 -2.40330026e-01
8.17444384e-01 -2.18449324e-01 3.48071486e-01 1.97283173e+00
3.73315811e-01 -2.64124155e-01 -7.02474685e-03 6.42010808e-01
-2.33647838e-01 9.93570834e-02 -6.74987555e-01 -1.92942277e-01
2.30207086e-01 8.93177629e-01 -5.38398981e-01 -4.78222817e-01
-4.13785934e-01 1.23366821e+00 8.08014929e-01 5.36973596e-01
-8.22890580e-01 -5.33425152e-01 7.72190988e-01 -2.79302746e-01
6.83536470e-01 -6.66973650e-01 -6.91153228e-01 -1.19160759e+00
1.19626001e-02 -1.01922226e+00 8.22680816e-02 -5.99727154e-01
-8.18081439e-01 9.42576885e-01 -7.00540483e-01 -9.58181262e-01
-8.47320795e-01 -6.84064090e-01 -4.52072024e-01 1.53571033e+00
-1.37984717e+00 -9.61817026e-01 5.00226557e-01 2.02358007e-01
6.41508102e-01 -1.50295213e-01 1.21633935e+00 2.54442334e-01
-4.29949075e-01 8.50515783e-01 3.30327451e-01 2.00699732e-01
3.93598408e-01 -1.08254611e+00 1.47501493e+00 9.07498896e-01
9.73057747e-02 9.94671762e-01 6.36797428e-01 -4.22625899e-01
-1.55665994e+00 -1.04471099e+00 1.37283063e+00 -4.53841746e-01
8.79727721e-01 -7.08074391e-01 -8.94073069e-01 1.13405406e+00
6.75669730e-01 -2.01044604e-01 4.30148363e-01 1.67083904e-01
-3.75776112e-01 5.22548676e-01 -4.31548655e-01 8.29106331e-01
1.09542263e+00 -1.02067852e+00 -2.71761239e-01 3.86683762e-01
1.06524467e+00 -8.62314165e-01 -7.37090409e-01 1.78669944e-01
5.84938049e-01 -5.97548068e-01 7.92123556e-01 -7.60044098e-01
5.74480593e-01 -5.08982278e-02 -4.76451404e-02 -1.29173458e+00
-3.80940497e-01 -9.52880621e-01 -8.73634443e-02 4.35888320e-01
9.44298148e-01 -7.39014208e-01 5.73838174e-01 9.30912942e-02
-3.48326534e-01 -1.00927186e+00 -7.74898767e-01 -7.06901491e-01
4.14689898e-01 -3.99536401e-01 4.97153968e-01 7.00498819e-01
-9.93204042e-02 7.15117872e-01 -7.02915609e-01 -3.65997076e-01
-8.95485878e-02 1.13961317e-01 4.21497256e-01 -5.25893331e-01
-5.21023870e-01 -6.71585798e-01 -9.73392874e-02 -1.43462646e+00
2.77307659e-01 -1.41042578e+00 7.50648230e-02 -1.46049571e+00
-1.66762788e-02 -1.87529206e-01 -2.37145126e-01 7.49812782e-01
-1.04261987e-01 4.86349851e-01 3.19103062e-01 1.89664379e-01
-2.93675482e-01 3.57724309e-01 1.00180960e+00 -5.39419800e-02
-4.94448794e-03 -2.00397193e-01 -2.67578572e-01 2.68878967e-01
1.00127757e+00 -5.82048476e-01 -9.31515917e-02 -1.08929658e+00
4.48283970e-01 1.56578198e-01 5.48404232e-02 -6.28420413e-01
1.71257809e-01 2.27453321e-01 3.77421290e-01 -3.97550285e-01
1.77049473e-01 -1.99644893e-01 6.68213516e-02 6.72502577e-01
-7.30637014e-01 8.10182691e-01 6.14840090e-01 1.02027901e-01
-7.29478663e-04 -2.69036025e-01 5.86475790e-01 -3.28394443e-01
-4.10871953e-01 1.17365763e-01 -8.43197405e-01 -7.06733689e-02
3.53914976e-01 -1.50033414e-01 -1.65139899e-01 -3.58428121e-01
-5.52332044e-01 3.41114104e-02 4.73450065e-01 3.88298094e-01
4.18965340e-01 -1.17223883e+00 -9.36218798e-01 3.03278863e-01
-3.33157718e-01 -4.60735649e-01 -1.85271710e-01 9.05940175e-01
-8.80945444e-01 8.88689339e-01 -1.99765861e-01 -5.57479143e-01
-1.05271137e+00 5.28524935e-01 4.35592353e-01 -5.20362914e-01
-7.32901216e-01 1.01838708e+00 -2.71592200e-01 -6.29635811e-01
7.94986188e-02 -3.08068991e-01 3.64593863e-01 -1.82099968e-01
5.03891528e-01 2.19701931e-01 3.15103143e-01 -4.26642239e-01
-1.62659764e-01 2.95808673e-01 -4.08067435e-01 -3.61992121e-01
1.19986105e+00 -3.64346392e-02 -3.93783212e-01 7.28485525e-01
1.73655307e+00 -3.27955842e-01 -1.19999588e+00 -3.18077445e-01
3.87868911e-01 5.03649935e-02 -2.59007722e-01 -8.25589240e-01
-6.62728548e-01 1.15096796e+00 4.63273257e-01 -3.05527359e-01
8.80306065e-01 -5.90259075e-01 1.14846003e+00 7.17924476e-01
1.62301481e-01 -9.05726552e-01 -2.02690125e-01 1.49687457e+00
7.29812145e-01 -9.97853875e-01 -3.26334178e-01 1.78794980e-01
-4.52008516e-01 1.45419097e+00 1.40094817e-01 -3.12155038e-01
1.21848628e-01 7.16323197e-01 2.67208844e-01 3.24224859e-01
-1.31487894e+00 -7.37747326e-02 1.21029072e-01 1.22112378e-01
1.10021794e+00 5.32814339e-02 -3.15340161e-01 -1.13871291e-01
-6.04605436e-01 1.71104312e-01 4.02352899e-01 9.27561164e-01
-2.78654397e-01 -1.39541090e+00 -2.53701843e-02 4.62995857e-01
-6.35161221e-01 -8.66768777e-01 -6.80492371e-02 4.71451014e-01
-3.07307690e-01 5.63289940e-01 2.65103608e-01 -3.46311122e-01
6.07288769e-03 3.65769416e-01 6.73874736e-01 -7.59392083e-01
-1.12819612e+00 -5.21339960e-02 2.47807890e-01 -3.45801532e-01
-7.34850438e-03 -5.09712875e-01 -1.39614737e+00 -7.07520902e-01
3.07855997e-02 -1.31403700e-01 7.41391301e-01 9.38066900e-01
4.34830368e-01 6.69554889e-01 4.34303701e-01 -8.55721653e-01
-5.97784221e-01 -1.11963415e+00 1.23968162e-01 -6.20099045e-02
5.29135108e-01 6.25453815e-02 1.57913685e-01 1.02810308e-01] | [10.769718170166016, 7.249521255493164] |
af2f9ddb-6dd1-4656-9bef-76e70a01a5ea | the-food-recognition-benchmark-using | 2106.14977 | null | https://arxiv.org/abs/2106.14977v2 | https://arxiv.org/pdf/2106.14977v2.pdf | The Food Recognition Benchmark: Using DeepLearning to Recognize Food on Images | The automatic recognition of food on images has numerous interesting applications, including nutritional tracking in medical cohorts. The problem has received significant research attention, but an ongoing public benchmark to develop open and reproducible algorithms has been missing. Here, we report on the setup of such a benchmark using publicly available food images sourced through the mobile MyFoodRepo app. Through four rounds, the benchmark released the MyFoodRepo-273 dataset constituting 24,119 images and a total of 39,325 segmented polygons categorized in 273 different classes. Models were evaluated on private tests sets from the same platform with 5,000 images and 7,865 annotations in the final round. Top-performing models on the 273 food categories reached a mean average precision of 0.568 (round 4) and a mean average recall of 0.885 (round 3). We present experimental validation of round 4 results, and discuss implications of the benchmark setup designed to increase the size and diversity of the dataset for future rounds. | ['Marcel Salathé', 'Victor Boulanger', 'Harris Héritier', 'Djilani Kebaili', 'Eric Antoine Scuccimarra', 'Gaurav Singhal', 'Sharada Prasanna Mohanty'] | 2021-06-28 | null | null | null | null | ['food-recognition'] | ['computer-vision'] | [ 5.16794398e-02 -1.79471802e-02 -3.24967265e-01 -4.98837113e-01
-1.11697435e+00 -7.78327167e-01 1.59695491e-01 8.47139418e-01
-5.21579206e-01 2.72040248e-01 3.79009753e-01 7.91385174e-02
1.30027622e-01 -7.10756958e-01 -1.06821287e+00 -6.07644200e-01
-5.20975888e-01 3.47957492e-01 3.38712424e-01 1.74940050e-01
-2.40632787e-01 -3.12532425e-01 -1.43004000e+00 7.99414515e-01
5.24468541e-01 9.74925041e-01 1.49269700e-01 1.01681292e+00
3.38876486e-01 3.22841674e-01 -4.18162853e-01 -7.91847169e-01
4.26088542e-01 -6.28154650e-02 -5.59877694e-01 3.32594067e-02
5.09894550e-01 5.47501445e-03 3.27889293e-01 1.07517958e+00
6.48549080e-01 -3.21471184e-01 5.46397924e-01 -9.10082936e-01
-9.53059256e-01 8.92666638e-01 -6.70141876e-01 3.02446634e-01
6.13337517e-01 3.02438706e-01 8.80712807e-01 -6.95982039e-01
7.08455682e-01 9.08544600e-01 1.38959563e+00 1.76780075e-01
-1.16668022e+00 -6.14566922e-01 -6.13055453e-02 -1.58867612e-01
-1.57351840e+00 -4.33910280e-01 -9.30985063e-02 -4.87676680e-01
9.11670923e-01 6.06318474e-01 8.43873739e-01 9.17018771e-01
1.92658588e-01 7.48248279e-01 1.25984240e+00 -2.93073326e-01
2.89421845e-02 4.44861412e-01 4.74310309e-01 9.83832538e-01
8.62916887e-01 -9.54016857e-03 -3.52966309e-01 -2.99799830e-01
1.73151165e-01 -1.91629976e-01 -3.72844040e-02 -2.65681297e-01
-1.34330940e+00 1.02629006e+00 6.82884395e-01 8.19250345e-02
-3.51097226e-01 -3.74173641e-01 4.95135397e-01 2.41814554e-01
8.86541188e-01 1.98398575e-01 -6.61048412e-01 3.69624466e-01
-6.60573602e-01 3.06715518e-01 8.75020504e-01 8.27218592e-01
2.84543633e-01 -7.06863880e-01 -2.64141023e-01 8.04967880e-01
5.05833447e-01 9.05413032e-01 3.23688269e-01 -4.32561100e-01
3.07887524e-01 5.66484451e-01 2.44748052e-02 -1.03948104e+00
-9.80910957e-01 -4.70716029e-01 -5.44201612e-01 -2.75701463e-01
7.26592124e-01 -3.24245781e-01 -9.76139188e-01 1.29383636e+00
6.18914485e-01 2.48517562e-02 7.83933252e-02 7.56234288e-01
1.56604612e+00 4.57292467e-01 4.44390893e-01 1.54262856e-01
1.88371062e+00 -1.15716445e+00 -3.64849597e-01 -2.48728693e-02
7.72670150e-01 -9.51785862e-01 8.00111353e-01 4.89929438e-01
-8.43641460e-01 -3.42497557e-01 -1.03718328e+00 8.07643682e-03
-7.38410294e-01 2.87718236e-01 7.38082528e-01 1.01678932e+00
-1.06076121e+00 1.31705463e-01 -9.21304941e-01 -8.73583734e-01
6.74593747e-01 3.83419245e-01 -3.81004184e-01 -2.65653908e-01
-7.09956527e-01 3.38840544e-01 2.92157441e-01 -1.50814846e-01
-6.98440909e-01 -1.19217765e+00 -9.29343104e-01 -3.64972621e-01
1.14601597e-01 -7.82859683e-01 1.13749981e+00 -7.48917997e-01
-8.94649982e-01 1.42944622e+00 4.54977989e-01 -5.07215977e-01
4.42698777e-01 -8.00791681e-02 -5.63222647e-01 1.60119548e-01
5.05222917e-01 1.08089054e+00 2.09996551e-02 -8.32165003e-01
-8.75140190e-01 -5.15504301e-01 -5.94808459e-02 -1.61081880e-01
-1.07625917e-01 1.82160452e-01 -4.72303540e-01 -6.72856748e-01
-1.96457997e-01 -1.27903152e+00 -2.46301845e-01 -7.22581670e-02
-5.74163020e-01 -1.08865350e-01 -1.57139719e-01 -8.88978362e-01
8.72710526e-01 -2.07867408e+00 -4.43551600e-01 6.65656254e-02
-1.16169162e-01 -7.08224550e-02 -1.50879204e-01 -1.62960187e-01
-1.03949517e-01 1.77250355e-02 -2.42689386e-01 -1.15867518e-01
-1.42595991e-01 -8.15104768e-02 5.34651220e-01 7.26033390e-01
1.17960632e-01 1.01015198e+00 -1.05549276e+00 -5.48057795e-01
7.42677425e-04 5.00814676e-01 -6.55840695e-01 -2.91591048e-01
-3.46027493e-01 2.11174011e-01 -4.81000125e-01 1.09939694e+00
7.30335593e-01 -3.87329072e-01 1.78788558e-01 -3.12321246e-01
-2.43988330e-03 -2.88699448e-01 -1.06529701e+00 1.80017996e+00
1.40710965e-01 1.15276568e-01 2.15445966e-01 -4.19747323e-01
4.50283080e-01 9.45812985e-02 9.05636907e-01 -6.03423059e-01
3.59772474e-01 -3.16602327e-02 -1.78041235e-02 -7.22826898e-01
3.16048831e-01 2.90301859e-01 -3.40437561e-01 3.70128274e-01
-1.12222090e-01 4.58536267e-01 6.17414296e-01 -1.67791262e-01
1.23712611e+00 2.36731157e-01 3.71498913e-01 -7.64852047e-01
7.07417578e-02 7.74645090e-01 1.89803287e-01 6.50074601e-01
-5.58164656e-01 7.67911553e-01 1.40900552e-01 -7.44935751e-01
-7.72462666e-01 -8.64384413e-01 -5.79908669e-01 1.26956868e+00
-2.71338850e-01 -4.19803500e-01 -9.35418606e-01 -9.35378373e-01
1.83902681e-01 8.35544467e-02 -1.02506900e+00 2.18437329e-01
-3.17520052e-01 -1.64839089e+00 6.77984416e-01 3.56729925e-01
4.23940212e-01 -1.05125713e+00 -8.21060836e-01 1.70493498e-01
-3.84909034e-01 -1.02933848e+00 -5.31097054e-01 1.28076240e-01
-5.69590032e-01 -1.70530605e+00 -1.04603481e+00 -8.92703831e-01
6.44304037e-01 4.87077832e-01 1.62521493e+00 1.30627736e-01
-6.78940058e-01 4.47485507e-01 -6.35860860e-01 -8.99743855e-01
-4.63922858e-01 4.26129282e-01 -3.26889366e-01 -2.21092805e-01
7.49743581e-01 3.40263546e-01 -9.36602235e-01 4.01057243e-01
-8.65629137e-01 -5.28464802e-02 4.06765014e-01 4.94144052e-01
8.93443227e-01 -3.35766196e-01 5.58788180e-01 -1.00636446e+00
5.74526638e-02 -8.43746483e-01 -5.96675336e-01 4.44931626e-01
-5.15516222e-01 -1.77513286e-01 1.50277559e-02 -2.73723662e-01
-7.70432711e-01 6.23655140e-01 -4.10512507e-01 2.98823476e-01
-1.64306581e-01 4.53766286e-01 2.17952847e-01 3.53880860e-02
9.77782845e-01 -4.02139783e-01 1.46467566e-01 -6.40058458e-01
3.28438908e-01 5.23142576e-01 3.96387577e-01 -3.13885659e-01
1.52116209e-01 4.77853566e-01 -3.63772005e-01 -8.08983803e-01
-9.98375058e-01 -7.77996778e-01 -4.29421753e-01 1.49775267e-01
1.22510588e+00 -1.29064715e+00 -4.74372208e-01 4.61266667e-01
-5.19283533e-01 -7.31385708e-01 -2.36758992e-01 4.88257080e-01
-6.79549947e-02 4.44813333e-02 -5.31926692e-01 -1.54136494e-01
-7.33959675e-01 -1.22016561e+00 1.20884967e+00 1.65162638e-01
-2.50828534e-01 -8.14745247e-01 3.64549667e-01 6.16435409e-01
7.63991773e-02 7.12548494e-01 4.90613937e-01 -5.58986664e-01
-1.37293890e-01 -1.57874241e-01 -2.94591337e-01 -1.36422753e-01
1.70096219e-01 -9.76127014e-02 -1.02177560e+00 -2.72669017e-01
-5.75079918e-01 -3.04417312e-01 1.18739104e+00 7.20513821e-01
8.89498115e-01 8.61421451e-02 -6.86329424e-01 3.65551084e-01
1.42872083e+00 -9.61827263e-02 3.20400268e-01 5.37533879e-01
4.55882281e-01 5.01670361e-01 7.22121954e-01 2.12277696e-01
7.27784395e-01 5.76759100e-01 4.23681676e-01 -4.60670590e-01
-4.58206922e-01 1.90399885e-01 3.54854941e-01 3.94864202e-01
-5.24380803e-03 -7.11304322e-02 -7.97541499e-01 6.66056454e-01
-1.41813445e+00 -6.06406868e-01 -4.90135312e-01 2.23597836e+00
7.23831356e-01 -1.18206851e-01 6.30577862e-01 7.91939441e-03
5.89486480e-01 -1.75974086e-01 -4.38685209e-01 -4.89621237e-02
-1.02452099e-01 1.85825199e-01 9.06087160e-01 1.84870899e-01
-1.64293492e+00 3.61750901e-01 7.29377031e+00 3.96003157e-01
-6.98006153e-01 4.45042700e-01 1.11508572e+00 -2.53519088e-01
2.11266398e-01 -7.75333643e-01 -9.51421559e-01 4.10188198e-01
1.19722617e+00 4.02786016e-01 -8.66926238e-02 7.70408988e-01
-5.31899408e-02 -2.85645545e-01 -9.23024952e-01 5.31160176e-01
2.86455780e-01 -1.06177866e+00 -4.18281585e-01 1.47067815e-01
9.69661057e-01 5.54518402e-01 6.84317350e-02 3.52313370e-01
3.44568521e-01 -9.06113684e-01 7.76178479e-01 2.61913896e-01
8.11454356e-01 -5.36194026e-01 8.03633213e-01 -2.84445621e-02
-1.44042623e+00 -1.21889427e-01 -4.21027809e-01 2.36914217e-01
-3.18099171e-01 7.49918461e-01 -9.47331548e-01 7.11571395e-01
1.32682168e+00 7.05422640e-01 -1.19430947e+00 1.25863159e+00
3.52123082e-01 4.25934643e-01 -6.60362184e-01 -1.12798400e-01
1.17799163e-01 1.95863008e-01 1.77921560e-02 1.62395036e+00
1.67086706e-01 1.34671196e-01 4.14948791e-01 5.55974096e-02
-5.24168424e-02 5.56680501e-01 -3.79599184e-01 1.03089429e-01
-2.23124713e-01 1.53561747e+00 -1.34152675e+00 -2.57877499e-01
-7.14623272e-01 5.28316796e-01 9.69523797e-04 -2.90637344e-01
-1.06588268e+00 2.72821695e-01 4.58560884e-01 1.62657842e-01
3.48130107e-01 3.89088869e-01 -1.19000211e-01 -8.29707801e-01
-2.26087466e-01 -9.78164494e-01 1.17356634e+00 -4.28720355e-01
-1.45276880e+00 3.92185718e-01 1.65208161e-01 -7.95789123e-01
3.47906470e-01 -7.27433264e-01 3.43101509e-02 3.06974411e-01
-1.41831040e+00 -1.40410483e+00 -4.89248842e-01 3.20865661e-01
5.52253127e-01 1.37294069e-01 1.17862868e+00 7.86982596e-01
-6.64555192e-01 7.67671585e-01 1.93595096e-01 1.18540883e-01
8.61419380e-01 -1.26439691e+00 2.95616150e-01 2.30372071e-01
3.88410777e-01 2.38832429e-01 6.37972236e-01 -7.84480631e-01
-1.39192009e+00 -1.63163853e+00 6.96423590e-01 -8.76621366e-01
3.40265214e-01 -2.32221723e-01 -5.57503998e-01 7.45022118e-01
2.47960776e-01 1.90402552e-01 1.15908778e+00 1.58870295e-01
-2.33530566e-01 -1.78340554e-01 -1.58513629e+00 1.10259049e-01
1.07148623e+00 2.42266268e-01 -8.08478892e-02 9.08293843e-01
6.82750702e-01 -8.30763459e-01 -1.20845973e+00 4.98907626e-01
8.99274826e-01 -7.88710713e-01 1.23918664e+00 -2.30907574e-01
2.22979680e-01 -3.17784026e-02 -1.75096780e-01 -1.10086000e+00
-3.20340067e-01 5.09450845e-02 2.83886433e-01 9.73900080e-01
8.54589939e-01 -4.54570383e-01 8.73890877e-01 1.06836595e-01
-2.70128995e-01 -9.13531423e-01 -5.58512747e-01 -3.27972412e-01
1.50364593e-01 -2.30744526e-01 6.97325766e-01 8.12354982e-01
-3.72709513e-01 -1.32378548e-01 9.92526934e-02 2.08535507e-01
5.54742396e-01 7.64271840e-02 5.81314683e-01 -1.06798804e+00
-2.72416353e-01 -1.15267359e-01 -1.62462115e-01 -4.09350574e-01
-2.59553671e-01 -1.31918371e+00 1.38956204e-01 -1.61847460e+00
8.48326266e-01 -6.08189583e-01 -3.49030763e-01 7.63402164e-01
-2.63935119e-01 1.14523935e+00 -1.96128245e-02 -8.49194229e-02
-7.56701887e-01 -3.72206658e-01 1.07296526e+00 -4.79491591e-01
-2.32302308e-01 -1.11943811e-01 -1.08468854e+00 7.57332087e-01
8.38282168e-01 -6.13580048e-01 -1.31807640e-01 -3.46331239e-01
2.60635883e-01 -5.67699134e-01 4.48910356e-01 -1.12571681e+00
-2.30707988e-01 1.62290320e-01 7.82367826e-01 -8.53415191e-01
-3.83681431e-02 -7.01800883e-01 6.18367732e-01 1.04561877e+00
-8.07227269e-02 2.48061076e-01 3.45595628e-01 5.62812328e-01
4.71190006e-01 -4.83139232e-02 6.31073833e-01 -3.53743494e-01
-5.25804341e-01 1.67548075e-01 -1.30771948e-02 -7.25675896e-02
1.31497753e+00 -6.79698354e-03 -3.81498188e-01 4.40204740e-01
-9.77374554e-01 3.31907421e-01 2.63585150e-01 5.64678669e-01
4.48001735e-02 -1.08526158e+00 -1.02539289e+00 2.33883187e-01
4.43197340e-01 -2.33266637e-01 2.39264891e-01 8.37444723e-01
-7.54075587e-01 6.53472364e-01 -2.92648468e-02 -1.03044057e+00
-1.60666728e+00 7.05211878e-01 2.46108636e-01 -5.55094719e-01
-9.76718724e-01 5.93382120e-01 1.75834090e-01 -5.17872989e-01
8.94250572e-02 -6.55258536e-01 -4.62356955e-01 2.22374097e-01
8.56469154e-01 3.91875386e-01 5.02731383e-01 -8.34702015e-01
-5.20400226e-01 5.28603017e-01 1.05329208e-01 4.40331578e-01
1.43170893e+00 3.81224183e-03 2.44288653e-01 4.98839691e-02
9.87914562e-01 -1.41944557e-01 -9.45911765e-01 1.18575044e-01
3.93905602e-02 -1.21890053e-01 -2.36490503e-01 -1.12297368e+00
-1.27760303e+00 8.11750665e-02 1.37559938e+00 3.89977485e-01
9.30067658e-01 2.35986233e-01 8.35329413e-01 -6.12365864e-02
4.37540233e-01 -5.78045607e-01 -1.17240794e-01 1.82967573e-01
6.32706463e-01 -1.52394569e+00 3.49984467e-01 -8.27039719e-01
-3.48609090e-01 6.64045632e-01 2.28914708e-01 -5.62601499e-02
8.17544043e-01 3.46524507e-01 3.55134726e-01 -5.67616403e-01
-3.14035416e-01 -7.46099697e-03 3.68573785e-01 8.18486989e-01
6.28452182e-01 5.22448659e-01 -4.14548934e-01 8.07242751e-01
-3.09446156e-01 1.89974204e-01 1.82114139e-01 8.80372167e-01
-2.56325006e-01 -9.68590438e-01 -7.15288103e-01 6.49574101e-01
-9.64426816e-01 8.43819305e-02 -9.80554447e-02 7.38528550e-01
4.84479845e-01 1.22971940e+00 6.02427032e-03 1.05814904e-01
4.74027872e-01 -1.39021456e-01 6.70897663e-01 -7.10545957e-01
-1.17191148e+00 1.29053488e-01 4.73104179e-01 -5.68643510e-01
-8.48872244e-01 -1.02744448e+00 -1.06486690e+00 -2.10060000e-01
-8.71473774e-02 -5.94252683e-02 7.07078934e-01 5.16452074e-01
2.94007689e-01 5.08519530e-01 1.63810089e-01 -5.76043248e-01
-3.56932610e-01 -9.21124220e-01 -3.89075369e-01 4.74467874e-01
-3.27962302e-02 -5.70718288e-01 1.12567350e-01 4.71412688e-01] | [11.566028594970703, 4.3730387687683105] |
74cc0638-d0cc-4285-95f0-c0b64e6dff25 | relational-program-synthesis-with-numerical | 2210.00764 | null | https://arxiv.org/abs/2210.00764v2 | https://arxiv.org/pdf/2210.00764v2.pdf | Relational program synthesis with numerical reasoning | Program synthesis approaches struggle to learn programs with numerical values. An especially difficult problem is learning continuous values over multiple examples, such as intervals. To overcome this limitation, we introduce an inductive logic programming approach which combines relational learning with numerical reasoning. Our approach, which we call NUMSYNTH, uses satisfiability modulo theories solvers to efficiently learn programs with numerical values. Our approach can identify numerical values in linear arithmetic fragments, such as real difference logic, and from infinite domains, such as real numbers or integers. Our experiments on four diverse domains, including game playing and program synthesis, show that our approach can (i) learn programs with numerical values from linear arithmetical reasoning, and (ii) outperform existing approaches in terms of predictive accuracies and learning times. | ['Andrew Cropper', 'Céline Hocquette'] | 2022-10-03 | null | null | null | null | ['program-synthesis', 'inductive-logic-programming', 'relational-reasoning'] | ['computer-code', 'methodology', 'natural-language-processing'] | [ 9.31817293e-03 2.12327167e-01 -8.69023323e-01 -3.17776799e-01
-9.07580733e-01 -8.64364803e-01 1.87406495e-01 4.01832879e-01
-3.27016674e-02 1.28062165e+00 -3.41753989e-01 -9.98680890e-01
-3.13384414e-01 -1.67657840e+00 -1.04535258e+00 -1.33575827e-01
-5.56836069e-01 6.64698482e-01 3.80548388e-01 -2.97656029e-01
2.37045035e-01 2.44375765e-01 -1.61795735e+00 8.18946064e-01
9.42099273e-01 1.18137968e+00 -7.81549513e-01 5.46684802e-01
-5.97703457e-01 1.81677973e+00 -5.84658206e-01 -4.23767537e-01
1.54312909e-01 -1.56991214e-01 -8.41162324e-01 -6.85517907e-01
1.87081724e-01 -4.93508041e-01 -1.16465434e-01 1.20137823e+00
-1.99399084e-01 -4.06025946e-02 5.55816770e-01 -1.69136012e+00
-5.40255249e-01 1.33161819e+00 -3.37515384e-01 -3.67080748e-01
8.52564633e-01 1.52628154e-01 1.47872233e+00 -3.69731128e-01
3.14527601e-01 1.47610903e+00 1.01905286e+00 2.79372960e-01
-1.62564695e+00 -7.00734317e-01 -1.20892785e-01 6.24298342e-02
-1.39950740e+00 1.08458728e-01 4.94056046e-01 -4.67973530e-01
1.16514158e+00 3.14570248e-01 7.11062729e-01 2.73642629e-01
1.23899728e-01 8.18874419e-01 1.10661888e+00 -6.57962203e-01
5.31357527e-01 1.14404589e-01 2.33689487e-01 7.77756810e-01
3.57692599e-01 3.44814420e-01 -1.31856248e-01 -3.22980851e-01
9.10791874e-01 -3.42762738e-01 1.78584874e-01 -3.73585433e-01
-1.30940759e+00 1.14966178e+00 2.28969291e-01 9.22297910e-02
1.45274743e-01 6.44372880e-01 6.63146913e-01 7.00487137e-01
-7.66491666e-02 7.87437618e-01 -7.16230035e-01 -3.45377773e-01
-8.04158568e-01 7.51587391e-01 1.40182459e+00 1.30940247e+00
7.14894593e-01 1.07208401e-01 8.89605582e-02 3.46354753e-01
1.13345504e-01 4.58434552e-01 1.26849502e-01 -1.06901419e+00
7.06398427e-01 6.64606452e-01 1.40228420e-01 -8.03705812e-01
-1.83768690e-01 5.40520191e-01 -5.84047079e-01 4.68814373e-01
7.26254106e-01 -5.17090082e-01 -5.72523832e-01 1.63266468e+00
-7.27936924e-02 6.52753264e-02 4.45918620e-01 2.38726363e-01
9.95814025e-01 9.42909062e-01 6.87338114e-02 -2.78498769e-01
1.03207195e+00 -4.85016584e-01 -5.43444991e-01 6.05766922e-02
1.17777181e+00 1.39216240e-03 1.01104069e+00 8.02229226e-01
-1.34698761e+00 -3.24408233e-01 -1.20532525e+00 6.70077326e-03
-6.52874649e-01 -2.30617523e-01 1.34332764e+00 6.85927391e-01
-8.59791636e-01 6.01881504e-01 -6.43312454e-01 6.08975172e-01
4.40198839e-01 7.53544271e-01 -1.46081522e-02 2.93945633e-02
-1.48567593e+00 6.90750360e-01 1.31008220e+00 -6.38369501e-01
-4.34563130e-01 -1.22295046e+00 -1.23520887e+00 2.25592449e-01
7.02581048e-01 -3.33797127e-01 1.55756593e+00 -8.20561767e-01
-1.50095546e+00 6.03120923e-01 2.72523493e-01 -9.05827403e-01
3.93371433e-01 2.30204999e-01 -2.58597523e-01 -2.54593313e-01
-3.33193094e-01 5.64604402e-01 5.73521033e-02 -1.03628516e+00
-6.43286228e-01 6.67709708e-02 9.78351712e-01 -2.56195813e-01
3.34743932e-02 -1.22168005e-01 3.75596136e-01 -2.75374353e-01
1.06888928e-01 -6.70436144e-01 -4.10950571e-01 -3.74242454e-03
-3.30948383e-01 -5.95034659e-01 2.31380209e-01 -1.94922417e-01
1.52028787e+00 -1.81427240e+00 1.33111417e-01 7.74770677e-01
1.81610391e-01 -1.04286954e-01 4.39126283e-01 2.79098451e-01
-2.43383795e-01 3.25083196e-01 -6.63749725e-02 7.86646783e-01
4.18981940e-01 3.42006505e-01 -6.38187230e-01 -8.71248692e-02
4.54417109e-01 1.06118131e+00 -1.06842732e+00 -7.87971973e-01
7.51671940e-02 -2.98401445e-01 -8.98013175e-01 1.11336976e-01
-1.22936833e+00 -3.44599187e-01 -3.18219155e-01 7.51381576e-01
5.31923294e-01 -2.15387255e-01 5.23968101e-01 1.99205384e-01
-3.31363603e-02 4.11842853e-01 -1.56692147e+00 1.20338976e+00
-5.24147034e-01 4.21531379e-01 -5.52465141e-01 -1.03084600e+00
1.10406637e+00 1.64696142e-01 2.52346158e-01 -5.25623560e-01
1.38475969e-02 3.86792064e-01 1.91326607e-02 -5.30802965e-01
3.27451050e-01 -4.48214501e-01 -6.36180103e-01 2.72057027e-01
-1.93139464e-01 -8.50333929e-01 5.48451483e-01 -4.27936614e-02
1.07265949e+00 1.67527452e-01 7.59284616e-01 -1.18601806e-01
7.70828068e-01 4.74758923e-01 5.55006385e-01 7.54247785e-01
3.52550387e-01 -5.05955517e-02 1.64156604e+00 -7.29696751e-01
-8.40673268e-01 -1.32346916e+00 -4.82913442e-02 1.07572150e+00
-1.39042765e-01 -6.92775190e-01 -3.91002893e-01 -5.42125642e-01
2.90616184e-01 9.35427964e-01 -2.97922641e-01 -3.72096002e-02
-8.61284256e-01 -2.60152459e-01 6.88320935e-01 1.08552980e+00
2.30678275e-01 -1.20909393e+00 -3.15141499e-01 9.61434171e-02
1.20908268e-01 -8.77300918e-01 4.11097914e-01 6.04121745e-01
-1.02071500e+00 -1.12235057e+00 1.22584641e-01 -8.75916541e-01
3.34448934e-01 -6.74056530e-01 1.44888830e+00 -9.17768478e-03
4.98125628e-02 1.07945673e-01 5.57035431e-02 -4.66885686e-01
-7.17020392e-01 3.15110646e-02 -2.50481606e-01 -8.73822331e-01
5.94755948e-01 -5.24714172e-01 4.90115553e-01 -2.19933316e-01
-9.40104783e-01 -9.26863551e-02 2.58564115e-01 8.82297158e-01
6.64687276e-01 8.12369227e-01 6.04438543e-01 -1.34904993e+00
7.40180492e-01 -5.71770191e-01 -1.36750960e+00 6.61518991e-01
-2.90434629e-01 4.70052242e-01 1.12480998e+00 -5.75680375e-01
-6.27457142e-01 4.36974205e-02 9.30279940e-02 -1.92678586e-01
8.42406452e-02 9.64539766e-01 -1.82553723e-01 -5.27589731e-02
1.06153476e+00 1.47131989e-02 -6.38809577e-02 3.95759732e-01
3.93435866e-01 3.40181798e-01 6.39222503e-01 -1.64762318e+00
8.01361084e-01 -7.99215883e-02 4.13773656e-01 -2.89177477e-01
-6.91657066e-01 2.10119531e-01 -5.34285367e-01 1.13660291e-01
4.15126383e-01 -9.22282517e-01 -1.26125658e+00 8.42326880e-02
-1.00763202e+00 -8.63237202e-01 -3.90276164e-01 2.99180776e-01
-1.02490258e+00 -1.29974961e-01 -6.91130877e-01 -1.02242899e+00
3.15646827e-02 -9.51290488e-01 6.00191951e-01 2.14364722e-01
-4.91615236e-01 -1.30327117e+00 -1.10979192e-03 -2.74731606e-01
3.05616800e-02 8.48554194e-01 1.72398734e+00 -5.98902225e-01
-6.79442704e-01 -2.82356948e-01 -2.68835366e-01 3.95902812e-01
-1.25533849e-01 2.84522265e-01 -4.88055497e-01 8.90072286e-02
-5.56275606e-01 -9.87933218e-01 2.28637189e-01 1.97984487e-01
1.65939891e+00 -5.97256005e-01 -1.36989310e-01 4.35166627e-01
1.59251475e+00 3.73117596e-01 1.02148700e+00 3.01945239e-01
2.42277935e-01 3.80002797e-01 8.62836421e-01 6.60734057e-01
3.70956004e-01 2.49392807e-01 2.46542573e-01 3.18478078e-01
4.01534468e-01 -4.04175371e-01 1.15523197e-01 1.05237529e-01
-1.78690121e-01 4.04867530e-01 -1.20218647e+00 3.53891075e-01
-1.75714028e+00 -1.12260079e+00 7.43675455e-02 2.08562708e+00
1.79000592e+00 7.23212123e-01 4.80689257e-01 4.66069818e-01
2.67616481e-01 -2.03060761e-01 -4.74585414e-01 -7.97429264e-01
1.43345818e-01 9.43134010e-01 3.30700099e-01 5.84713995e-01
-1.12283492e+00 1.06603348e+00 6.96341133e+00 8.24281096e-01
-8.73308182e-01 -5.26270270e-01 4.34869975e-01 3.05447429e-02
-5.36733270e-01 -4.74248268e-02 -6.18309617e-01 3.47469360e-01
1.11488044e+00 -6.87489927e-01 6.42989933e-01 1.39598298e+00
-6.03942990e-01 6.47071823e-02 -1.71932638e+00 8.95693243e-01
-4.83908236e-01 -1.50536501e+00 -1.50023460e-01 -3.30473632e-01
8.67177606e-01 -7.72847414e-01 1.36490583e-01 1.14173460e+00
9.76055622e-01 -1.68657172e+00 6.31619692e-01 4.10694242e-01
1.05267036e+00 -1.28184152e+00 7.85809875e-01 2.11965799e-01
-1.20041573e+00 -3.19224060e-01 -3.74816537e-01 -6.06600344e-01
-5.60072720e-01 4.98593569e-01 -8.57859313e-01 2.77312368e-01
3.81345034e-01 8.26143980e-01 -4.28106755e-01 8.21349084e-01
-4.58058894e-01 3.73817682e-01 -3.73412371e-01 -4.18596983e-01
1.08775117e-01 1.94148421e-02 -3.49867672e-01 1.02066720e+00
1.54274657e-01 3.97418916e-01 1.89534947e-01 1.31078732e+00
2.92774737e-02 -2.89609790e-01 -8.32972109e-01 -1.77513406e-01
7.14497268e-01 6.37369096e-01 -4.08570856e-01 -5.22309840e-01
-6.16040289e-01 -3.05913776e-01 2.81379461e-01 2.28247225e-01
-1.20253444e+00 -8.22828829e-01 5.53188920e-01 -5.84422499e-02
1.09241836e-01 -2.17765257e-01 -9.09203529e-01 -9.04720068e-01
6.43609762e-02 -1.36457372e+00 5.37582219e-01 -4.92222369e-01
-9.60391819e-01 1.72773167e-01 5.91459870e-01 -1.19809258e+00
-7.30015814e-01 -1.06125844e+00 -4.05942917e-01 6.14777803e-01
-1.16849625e+00 -7.78151691e-01 -8.69154707e-02 4.89906818e-01
-1.80704489e-01 -2.30498761e-01 8.32482517e-01 -7.43147507e-02
-9.62122381e-02 8.33288848e-01 -3.33234340e-01 3.34738940e-01
2.08175704e-01 -1.62217784e+00 1.33727873e-02 1.37521043e-01
-6.15115762e-02 6.43953562e-01 4.67092603e-01 -3.38523865e-01
-1.69320238e+00 -9.68270600e-01 6.69132292e-01 -2.97346145e-01
1.03154421e+00 -9.94192213e-02 -9.04877901e-01 1.08506823e+00
-4.11562890e-01 1.28236145e-01 6.82421446e-01 3.86868000e-01
-8.84519875e-01 -2.17105746e-01 -1.32868648e+00 7.44957149e-01
7.11623609e-01 -5.41063905e-01 -9.79476631e-01 4.53633994e-01
8.76754761e-01 -1.03048885e+00 -1.43761754e+00 7.43340790e-01
6.98676169e-01 -8.96671414e-01 1.31965160e+00 -1.00660360e+00
1.04742634e+00 -2.49069721e-01 -3.57447654e-01 -8.70087981e-01
5.14597781e-02 -3.58273923e-01 -6.43907666e-01 9.55801964e-01
5.10270894e-01 -6.58478379e-01 8.46301079e-01 9.47717130e-01
2.98237503e-01 -1.04391336e+00 -6.07817531e-01 -9.79507148e-01
8.28380585e-01 -7.26797998e-01 1.14662600e+00 9.38232720e-01
6.43093407e-01 1.01989299e-01 6.35824427e-02 -5.23769222e-02
2.89532572e-01 6.02763176e-01 7.40075946e-01 -1.42467201e+00
-6.04562163e-01 -7.07488537e-01 -7.02453196e-01 -5.20700336e-01
6.78049326e-01 -9.76911306e-01 -4.77986373e-02 -1.15139806e+00
-1.29220784e-01 -8.41096818e-01 -2.88565326e-02 9.69441772e-01
2.67644107e-01 -1.79984316e-01 1.16156517e-02 -4.45132643e-01
-7.14676499e-01 9.32224933e-03 1.01404345e+00 -4.27981347e-01
-3.51533651e-01 1.99558049e-01 -7.27622449e-01 8.69286060e-01
7.19151974e-01 -1.74235687e-01 -6.31479144e-01 1.22374602e-01
1.03622973e+00 7.32399166e-01 2.42784619e-01 -1.34919691e+00
1.84739083e-01 -8.51004720e-01 2.90912390e-01 -5.23942649e-01
-1.29859805e-01 -6.51900828e-01 5.41236205e-03 5.56538701e-01
-7.79217362e-01 -8.01149607e-02 6.18266106e-01 -1.67699996e-03
-3.88119668e-01 -6.48355842e-01 6.68206394e-01 -2.29242772e-01
-7.87676692e-01 -1.68908820e-01 -1.46709710e-01 5.27346313e-01
1.19687295e+00 8.11272785e-02 -3.30556363e-01 -3.01360071e-01
-6.39920473e-01 5.05297422e-01 2.45960861e-01 1.06677751e-03
6.11117065e-01 -1.55122793e+00 -4.66312140e-01 1.96348727e-01
1.00943685e-01 4.48023498e-01 -3.59896064e-01 4.49883372e-01
-9.70562041e-01 5.42155206e-01 -1.68573290e-01 -4.36376214e-01
-1.00360203e+00 7.09184408e-01 3.18890661e-01 -5.91352642e-01
-1.44398272e-01 5.70272207e-01 -2.16790259e-01 -7.40520656e-01
6.21738434e-01 -1.11538851e+00 -1.39467986e-02 -9.44203660e-02
7.53171325e-01 3.17468435e-01 -2.13350534e-01 4.05264288e-01
-2.59517372e-01 3.55639338e-01 1.81171238e-01 1.05761625e-01
1.21870124e+00 8.90846789e-01 -6.59233272e-01 7.62050986e-01
1.01605821e+00 -2.43835270e-01 -8.47714126e-01 -4.52060521e-01
4.62318271e-01 -3.47710758e-01 -3.85807961e-01 -8.34309518e-01
-6.03837550e-01 5.09457827e-01 -1.15088075e-01 4.62199628e-01
1.11776149e+00 -1.28992841e-01 5.27854860e-01 1.07291150e+00
8.99715185e-01 -8.98058951e-01 1.68634877e-01 8.65797222e-01
5.21447718e-01 -1.33142328e+00 2.04280198e-01 -4.65232790e-01
-4.87101674e-01 1.59573770e+00 8.53293538e-01 -2.75085598e-01
4.17896509e-01 1.05268884e+00 -3.50267708e-01 1.61682010e-01
-1.00343406e+00 -8.02911445e-02 6.65249676e-02 6.46264434e-01
6.81671619e-01 3.07147413e-01 5.93669415e-02 8.54080319e-01
-7.97221601e-01 5.65857947e-01 5.53632140e-01 1.08585501e+00
-3.82468194e-01 -1.15469456e+00 -6.02297425e-01 5.85559249e-01
-3.22010934e-01 -1.51743814e-01 -1.35518938e-01 1.26705265e+00
1.24763578e-01 5.64987004e-01 4.75858711e-03 -4.55710113e-01
1.79440349e-01 1.95403546e-01 9.57099855e-01 -7.18875587e-01
-4.06942517e-01 -5.95293641e-01 2.59857744e-01 -3.29107255e-01
-4.97464091e-02 -4.11655724e-01 -1.70052803e+00 -6.98717415e-01
1.20695733e-01 1.50532424e-01 5.26404493e-02 7.10858166e-01
-6.06842399e-01 6.36590898e-01 3.40874791e-01 -2.28669837e-01
-9.94763732e-01 -3.99355263e-01 -6.22986019e-01 -1.03054121e-01
3.44695985e-01 -4.22188401e-01 -3.13053250e-01 -1.95147589e-01] | [8.809698104858398, 7.154491424560547] |
5d414f17-6208-456d-80e3-7faeaee23dac | bsdar-beam-search-decoding-with-attention | 1909.09485 | null | https://arxiv.org/abs/1909.09485v1 | https://arxiv.org/pdf/1909.09485v1.pdf | BSDAR: Beam Search Decoding with Attention Reward in Neural Keyphrase Generation | This study mainly investigates two decoding problems in neural keyphrase generation: sequence length bias and beam diversity. We introduce an extension of beam search inference based on word-level and n-gram level attention score to adjust and constrain Seq2Seq prediction at test time. Results show that our proposed solution can overcome the algorithm bias to shorter and nearly identical sequences, resulting in a significant improvement of the decoding performance on generating keyphrases that are present and absent in source text. | ["Iftitahu Ni'mah", 'Mykola Pechenizkiy', 'Vlado Menkovski'] | 2019-09-17 | null | null | null | null | ['keyphrase-generation'] | ['natural-language-processing'] | [ 6.13431633e-01 -2.00891539e-01 -2.73087561e-01 -7.51723498e-02
-9.79842663e-01 -8.26165974e-01 6.66219294e-01 9.88472253e-02
-8.73042107e-01 1.28619015e+00 8.31783712e-01 -7.42523015e-01
8.20031464e-02 -7.88231373e-01 -6.05453968e-01 -5.59183061e-01
5.68778664e-02 1.39551014e-01 4.91116047e-02 -6.16245449e-01
7.34146893e-01 3.01972747e-01 -1.45708549e+00 5.98507762e-01
6.45650685e-01 5.08927345e-01 4.95012671e-01 1.25128472e+00
-2.90617704e-01 6.89502120e-01 -1.06017709e+00 -4.26951170e-01
2.84239352e-01 -9.05749440e-01 -6.28262281e-01 -5.47720551e-01
5.07324278e-01 -2.61761457e-01 -3.72518331e-01 1.12973642e+00
1.07071877e+00 2.76128054e-01 4.49625283e-01 -6.61652327e-01
-6.32034540e-01 1.54010808e+00 -3.78525794e-01 8.65964532e-01
6.39046013e-01 2.11613268e-01 1.38655376e+00 -7.97314465e-01
4.66687948e-01 1.07844245e+00 4.00399655e-01 7.00552821e-01
-1.00495589e+00 -9.82624888e-01 1.21263348e-01 3.73917788e-01
-1.42229760e+00 -3.42647284e-01 3.23605120e-01 7.89092504e-04
1.39673042e+00 6.66301191e-01 6.76855922e-01 1.47321022e+00
6.34362400e-01 8.28754008e-01 7.93024361e-01 -5.93226790e-01
1.61332324e-01 -3.36408019e-01 1.01988852e-01 3.97488028e-01
2.41983995e-01 2.29127839e-01 -1.07872736e+00 -2.71186292e-01
7.36523151e-01 -8.33588243e-01 -4.19137746e-01 6.00237608e-01
-1.49235618e+00 1.04168761e+00 1.30805168e-02 3.67702872e-01
-5.54490924e-01 1.79511368e-01 3.18471998e-01 3.58909905e-01
2.01936528e-01 9.79066730e-01 -5.72455466e-01 -6.00061774e-01
-1.13639617e+00 7.77166247e-01 8.00443351e-01 9.37396944e-01
2.72747546e-01 3.59548837e-01 -9.80813622e-01 6.72480702e-01
-1.40412807e-01 8.16512764e-01 1.02018976e+00 -2.87884057e-01
6.80658221e-01 -1.97379678e-01 -3.48793864e-02 -8.74848366e-01
-1.13562934e-01 -8.16939294e-01 -9.04208302e-01 -4.92109656e-01
-3.51024978e-02 -5.97018480e-01 -1.22325754e+00 1.96663427e+00
-1.86128736e-01 1.29260436e-01 5.89830764e-02 7.04820931e-01
8.47539604e-01 9.96194184e-01 -3.84369344e-02 -4.19531882e-01
1.30701971e+00 -8.79098237e-01 -1.16760516e+00 -2.88778454e-01
5.99186838e-01 -9.52418864e-01 1.08053541e+00 5.22868037e-01
-1.25199115e+00 -5.49667776e-01 -1.02866709e+00 2.93833494e-01
-2.53492445e-01 -2.12397873e-01 5.52863121e-01 7.80802965e-01
-9.08353388e-01 4.99381244e-01 -8.00316781e-02 6.63176924e-02
5.10522500e-02 1.45867199e-01 6.92673773e-02 2.18112588e-01
-1.88674831e+00 7.90033996e-01 1.06495249e+00 -1.17853291e-01
-6.83937371e-01 -8.29569757e-01 -5.77845573e-01 1.33212395e-02
1.37484014e-01 -6.87789917e-01 1.47282875e+00 -6.42345428e-01
-1.53502965e+00 2.82870650e-01 -2.65950888e-01 -7.99475074e-01
2.00701401e-01 -4.37147558e-01 -5.31855345e-01 -2.30992258e-01
-8.79036263e-02 8.67131472e-01 8.78611982e-01 -5.88736296e-01
-9.48660612e-01 2.27623492e-01 -3.54170263e-01 5.87540090e-01
-3.36258858e-01 7.63867125e-02 -3.94029975e-01 -1.54212332e+00
-2.31527478e-01 -6.65086389e-01 -4.08047616e-01 -1.14647520e+00
-7.16792524e-01 -1.86199397e-01 1.70185640e-01 -8.55947554e-01
1.90954423e+00 -1.57286072e+00 4.23940659e-01 3.30562562e-01
-2.63735861e-01 2.56094903e-01 -5.54049134e-01 7.29214489e-01
3.64743173e-02 3.72238427e-01 -2.03209281e-01 3.54929179e-01
-1.96470529e-01 5.44451997e-02 -7.52215922e-01 -9.84062478e-02
-1.80735695e-03 1.11886048e+00 -1.00632322e+00 -2.84869552e-01
-4.31811929e-01 -2.05112007e-02 -6.46326482e-01 4.57721949e-01
-5.71508706e-01 2.81914026e-01 -1.25422671e-01 3.69015932e-01
4.40944374e-01 6.02012575e-02 1.25578836e-01 4.46959697e-02
-2.06775572e-02 8.71361971e-01 -9.52337444e-01 1.76936436e+00
-3.41953367e-01 5.91630757e-01 -5.66114843e-01 -2.66995341e-01
7.26819277e-01 3.09758484e-01 -1.67116672e-01 -9.18532550e-01
9.31695849e-02 1.51346577e-02 5.53107977e-01 -1.84561178e-01
1.10811710e+00 -4.46944498e-02 -1.29106671e-01 6.28775835e-01
9.26203653e-03 -2.65130281e-01 4.53981519e-01 4.46085632e-01
1.10293436e+00 -1.37216926e-01 5.28421998e-01 -3.85523260e-01
5.03036320e-01 -2.29249194e-01 2.26970598e-01 1.46713865e+00
4.38647449e-01 3.00173312e-01 3.20796967e-01 -5.97859882e-02
-1.24939382e+00 -8.68253887e-01 8.15701410e-02 1.59426069e+00
-2.35594243e-01 -7.84468532e-01 -7.15233803e-01 -7.56148040e-01
-9.21805799e-02 1.23681390e+00 -6.58930421e-01 -2.45350033e-01
-9.23033118e-01 -9.05193090e-01 1.04977286e+00 3.18555236e-01
-2.01520920e-02 -1.35755110e+00 -4.22545522e-01 3.59351635e-01
-4.45948064e-01 -8.12163830e-01 -8.04926515e-01 5.25156498e-01
-7.19023407e-01 -4.66937870e-01 -1.08279073e+00 -5.35370648e-01
4.44779605e-01 -1.36148572e-01 1.28681195e+00 -2.18039774e-03
-9.00340918e-03 -1.77139714e-01 -8.86942804e-01 -4.77651626e-01
-5.77990949e-01 6.83106005e-01 1.56169030e-04 -2.83737600e-01
2.12866932e-01 -3.66868317e-01 -4.12244141e-01 -2.00807959e-01
-1.02403975e+00 1.85437277e-01 9.16082084e-01 9.19874847e-01
1.82934806e-01 -3.42600159e-02 5.22505105e-01 -8.92185748e-01
1.49078989e+00 -5.27169526e-01 -3.75597239e-01 5.49048744e-02
-8.44886363e-01 4.48379040e-01 5.96692324e-01 -6.02761567e-01
-7.64087200e-01 -4.62364972e-01 -4.80892718e-01 7.89950714e-02
1.97397172e-01 6.94706500e-01 1.21678203e-01 1.99220896e-01
8.02985668e-01 8.87600780e-01 -2.81376511e-01 -2.64606386e-01
5.16614735e-01 5.55068254e-01 3.43452722e-01 -4.01443630e-01
7.53989697e-01 -3.37727964e-01 -2.13324532e-01 -6.35106206e-01
-7.82666862e-01 -2.95441389e-01 -4.43166137e-01 1.85548645e-02
6.05242491e-01 -7.42351592e-01 -1.56030104e-01 3.94590855e-01
-1.27528906e+00 -2.91612595e-01 -9.95929465e-02 4.06106859e-01
-2.54703790e-01 4.92431492e-01 -5.92885017e-01 -7.29685783e-01
-9.72496808e-01 -1.06491899e+00 8.21659803e-01 -2.03101970e-02
-5.87013721e-01 -5.58422029e-01 2.42583379e-01 -2.71370769e-01
4.64659542e-01 -4.47751731e-01 1.10833645e+00 -9.35516596e-01
-3.97938401e-01 -6.72057495e-02 1.41366869e-01 2.14287907e-01
-9.00374055e-02 -3.54279816e-01 -5.37545085e-01 -4.28235590e-01
-4.03375089e-01 -3.07740539e-01 1.11942148e+00 2.66122699e-01
1.20574296e+00 -6.86916173e-01 -6.61093816e-02 5.78886926e-01
1.27121639e+00 3.85158569e-01 7.51452446e-01 3.19765598e-01
4.32959259e-01 3.54836524e-01 5.82234323e-01 6.43051624e-01
-1.12227827e-01 5.69125891e-01 1.36076272e-01 3.40390742e-01
-1.33797020e-01 -6.53468072e-01 4.44191396e-01 1.06354749e+00
1.79650560e-01 -9.50155318e-01 -7.65904188e-01 5.41504860e-01
-1.59657240e+00 -1.07159328e+00 -1.68706551e-02 2.01914907e+00
1.45080984e+00 4.23536152e-01 -1.09012090e-02 1.93564042e-01
5.89893639e-01 3.60519916e-01 -2.24786431e-01 -6.80551708e-01
-5.12364447e-01 7.85416067e-01 9.26474571e-01 7.56416619e-01
-7.64143348e-01 1.35130107e+00 7.99390125e+00 1.33621943e+00
-8.89980137e-01 -1.56242475e-01 3.34240258e-01 -3.90396357e-01
-8.07383537e-01 -3.92607898e-01 -1.23558700e+00 5.80273628e-01
1.29139423e+00 -2.65856832e-01 4.84069318e-01 3.01914483e-01
-2.34756246e-02 7.73907974e-02 -1.00291276e+00 9.85269129e-01
2.37838566e-01 -1.50850582e+00 8.08608770e-01 -1.49566293e-01
7.44277000e-01 -1.59319252e-01 1.10665537e-01 3.56388628e-01
5.62813282e-01 -1.26993632e+00 8.51387084e-01 5.13276219e-01
7.74832845e-01 -1.06800950e+00 6.33634984e-01 4.88420993e-01
-7.54232764e-01 -9.03102532e-02 -2.94416159e-01 3.20765488e-02
2.81982273e-01 4.98616427e-01 -1.16900086e+00 2.44933814e-01
2.63049334e-01 2.80036420e-01 -4.40622687e-01 8.13977897e-01
-6.15362585e-01 9.87657428e-01 6.15696535e-02 -8.48095536e-01
4.76452410e-01 2.47734666e-01 9.19875026e-01 1.90941608e+00
3.46703678e-01 1.84636995e-01 -1.16890147e-01 4.14458632e-01
-5.84097393e-02 3.06333989e-01 -3.70110661e-01 -2.35326976e-01
7.39965916e-01 9.10832465e-01 -5.16082585e-01 -4.35030788e-01
1.72891513e-01 1.23154998e+00 1.21832378e-01 3.43063235e-01
-8.45019341e-01 -6.42587304e-01 7.04601109e-01 -2.07450986e-01
6.03851497e-01 -1.85850784e-01 -4.19571728e-01 -8.81548584e-01
-4.40072477e-01 -1.55823636e+00 3.03016514e-01 -6.08492076e-01
-9.53558147e-01 7.06619263e-01 -5.51444851e-02 -7.42242992e-01
-6.48251772e-01 -2.25907564e-01 -2.52315760e-01 1.25396943e+00
-1.22111547e+00 -6.69782519e-01 2.00296521e-01 2.67736524e-01
9.25257623e-01 -6.96468294e-01 8.48553419e-01 6.21366948e-02
-2.18694225e-01 1.20274854e+00 6.29988611e-02 9.57147852e-02
4.91221219e-01 -1.28285527e+00 1.03119421e+00 1.01754701e+00
4.58975345e-01 8.45614076e-01 8.85349333e-01 -1.02032435e+00
-1.41467845e+00 -8.66855681e-01 1.12362611e+00 -3.87315869e-01
5.38951874e-01 -5.83850861e-01 -6.45727634e-01 3.14254165e-01
5.59889019e-01 -9.27419066e-01 8.21813464e-01 1.01539763e-02
-3.53777200e-01 3.40184003e-01 -5.59594631e-01 1.09561098e+00
1.09942687e+00 -3.87322485e-01 -7.25972831e-01 2.08924666e-01
1.12930393e+00 -7.21332014e-01 -4.61664230e-01 6.68344796e-01
6.41055644e-01 -6.26890004e-01 8.83382857e-01 -1.01127350e+00
6.02723777e-01 1.16690286e-01 -3.13514918e-01 -1.53468871e+00
-7.51317024e-01 -1.00223827e+00 -4.46703881e-01 7.88980842e-01
8.34359467e-01 -2.29645744e-01 8.21329236e-01 -3.59085619e-01
-1.13065466e-02 -7.27515578e-01 -8.42689693e-01 -5.00807464e-01
2.31015846e-01 -4.43634152e-01 7.42452502e-01 6.31244600e-01
-1.73940286e-01 7.46015608e-01 -7.45082021e-01 -5.30858934e-02
2.58534402e-01 -2.41206497e-01 5.94140112e-01 -5.99535584e-01
-5.57612956e-01 -7.53428817e-01 1.24400668e-01 -1.61012733e+00
6.36611879e-02 -1.01541662e+00 4.28624302e-01 -1.13116455e+00
1.36311069e-01 -8.37336183e-02 -5.93890965e-01 2.18892753e-01
-7.37819672e-01 3.14008832e-01 1.11896791e-01 -1.86090022e-02
-1.21508069e-01 5.40291667e-01 1.23135972e+00 -8.41969103e-02
1.69608407e-02 -7.13621154e-02 -8.10319841e-01 4.98656891e-02
7.97130167e-01 -5.70689440e-01 -5.20969450e-01 -4.11615252e-01
7.87852228e-01 7.04448000e-02 -1.76062211e-01 -7.44203687e-01
9.62759107e-02 -3.11482400e-01 3.49452883e-01 -8.97350729e-01
2.54104938e-02 -3.58574778e-01 -1.07171856e-01 7.74351299e-01
-1.11108637e+00 4.70419317e-01 5.20002186e-01 6.04352117e-01
1.74196899e-01 -6.00546777e-01 4.51759845e-01 -3.49150985e-01
-5.77180564e-01 9.61507484e-02 -8.51485908e-01 2.38176465e-01
5.55659890e-01 -1.09160736e-01 1.13729555e-02 -6.45232856e-01
-1.15525343e-01 -9.76411924e-02 -2.77893305e-01 7.58667827e-01
8.45537484e-01 -1.22095203e+00 -1.35839450e+00 3.83713305e-01
4.27022874e-02 -4.44031388e-01 -2.23166764e-01 2.57539809e-01
-4.86243188e-01 9.89736199e-01 -7.72935599e-02 -7.82748014e-02
-1.38761294e+00 1.67900354e-01 -3.90290003e-03 -4.81323242e-01
-2.76835859e-01 1.59284055e+00 -2.66858995e-01 -2.60875791e-01
5.33990383e-01 -2.80145258e-01 -3.45545352e-01 1.42619625e-01
8.62132192e-01 1.18884563e-01 2.35480040e-01 -2.77016282e-01
-3.22919972e-02 6.82231262e-02 -5.01714349e-01 -5.07686317e-01
6.68825030e-01 1.90307736e-01 -2.40993619e-01 4.18237656e-01
7.51187384e-01 5.28149486e-01 -4.92893308e-01 -2.10702896e-01
3.96751203e-02 -4.19683099e-01 2.05018684e-01 -1.24698949e+00
-4.33234036e-01 4.14461523e-01 4.37951744e-01 6.49618208e-02
9.98137832e-01 -2.47918621e-01 1.25196993e+00 5.48086464e-01
1.34258345e-01 -9.84597981e-01 1.14481203e-01 1.14116418e+00
9.44305182e-01 -6.44550681e-01 4.07625176e-02 1.32630676e-01
-5.50642431e-01 9.10543799e-01 5.22594750e-01 2.45074958e-01
2.11758986e-01 3.37129384e-01 3.43009867e-02 2.55837232e-01
-1.08306444e+00 -1.20811947e-01 6.38666749e-01 4.88899082e-01
7.43208885e-01 -8.00386071e-02 -8.78097057e-01 3.14266622e-01
-8.69530320e-01 -4.94835168e-01 3.17817599e-01 7.34169424e-01
-6.79044425e-01 -1.18987370e+00 -3.59093189e-01 7.37477839e-01
-8.71673167e-01 -1.24526703e+00 -5.80822051e-01 2.65843958e-01
2.25457363e-02 6.99423492e-01 1.60998777e-01 -7.15110958e-01
2.23837141e-02 2.25368500e-01 5.87060332e-01 -6.84432387e-01
-9.79707897e-01 5.39770462e-02 2.97772557e-01 -1.32009760e-01
1.07875071e-01 -3.92942965e-01 -9.04947102e-01 -2.92495847e-01
-5.48363924e-01 5.33426642e-01 5.51960409e-01 6.35938346e-01
2.43306875e-01 8.08474183e-01 5.49504459e-01 -4.89439934e-01
-9.25509691e-01 -1.42864680e+00 -3.22684109e-01 8.64398107e-02
3.74912262e-01 -8.16812331e-04 -1.01043597e-01 -1.03749156e-01] | [12.15931224822998, 8.989936828613281] |
9413d928-30e8-4e5f-b589-4d6f8b2a2b06 | robustness-benchmark-of-road-user-trajectory | 2304.01895 | null | https://arxiv.org/abs/2304.01895v1 | https://arxiv.org/pdf/2304.01895v1.pdf | Robustness Benchmark of Road User Trajectory Prediction Models for Automated Driving | Accurate and robust trajectory predictions of road users are needed to enable safe automated driving. To do this, machine learning models are often used, which can show erratic behavior when presented with previously unseen inputs. In this work, two environment-aware models (MotionCNN and MultiPath++) and two common baselines (Constant Velocity and an LSTM) are benchmarked for robustness against various perturbations that simulate functional insufficiencies observed during model deployment in a vehicle: unavailability of road information, late detections, and noise. Results show significant performance degradation under the presence of these perturbations, with errors increasing up to +1444.8\% in commonly used trajectory prediction evaluation metrics. Training the models with similar perturbations effectively reduces performance degradation, with error increases of up to +87.5\%. We argue that despite being an effective mitigation strategy, data augmentation through perturbations during training does not guarantee robustness towards unforeseen perturbations, since identification of all possible on-road complications is unfeasible. Furthermore, degrading the inputs sometimes leads to more accurate predictions, suggesting that the models are unable to learn the true relationships between the different elements in the data. | ['René van de Molengraft', 'Jos Elfring', 'Emilia Silvas', 'Manuel Muñoz Sánchez'] | 2023-04-04 | null | null | null | null | ['trajectory-prediction'] | ['computer-vision'] | [ 3.49589109e-01 1.69729844e-01 -4.88830283e-02 -1.82966858e-01
-5.59485376e-01 -5.39435744e-01 5.49598634e-01 2.62412250e-01
-5.19121289e-01 7.49895275e-01 2.12474599e-01 -8.12539577e-01
-1.28071204e-01 -7.65655637e-01 -1.07136655e+00 -5.60487509e-01
-2.10435480e-01 2.41101339e-01 5.10675669e-01 -4.01344597e-01
2.11207509e-01 7.49196291e-01 -1.99694157e+00 1.63720593e-01
8.54362786e-01 5.38920999e-01 5.16886152e-02 8.29139948e-01
1.26082540e-01 6.51549280e-01 -6.86472595e-01 -1.12356357e-01
2.98296362e-01 1.59453511e-01 -3.78188193e-01 -4.49950486e-01
5.27181089e-01 -5.42248845e-01 -5.53297281e-01 6.65256202e-01
3.44255507e-01 3.45401824e-01 4.87509519e-01 -1.44617248e+00
1.62628934e-01 2.39524588e-01 -1.11047894e-01 3.10239106e-01
7.73522481e-02 5.60302854e-01 3.65858495e-01 -4.94081378e-01
3.99490744e-01 1.12812102e+00 1.05548048e+00 5.52590430e-01
-1.25403023e+00 -7.33147383e-01 2.72058785e-01 3.75967264e-01
-1.12941492e+00 -7.52452254e-01 4.54438776e-01 -5.98966300e-01
1.15993941e+00 5.21633625e-01 1.93049580e-01 1.50041151e+00
2.59645402e-01 2.67141819e-01 5.67499101e-01 8.42011049e-02
1.37985364e-01 3.67652446e-01 4.50132340e-02 -4.36901152e-02
4.83099788e-01 7.48697460e-01 -2.38420069e-01 -1.03488602e-02
1.14622183e-01 -2.17556015e-01 -4.16638181e-02 3.81834544e-02
-9.49578047e-01 2.10343927e-01 5.13615847e-01 -1.19446535e-02
-5.32580853e-01 2.55708188e-01 5.62430799e-01 2.69526064e-01
1.71793580e-01 6.17056608e-01 -5.28578103e-01 -5.61681151e-01
-7.44596720e-01 4.90071177e-01 3.86281282e-01 8.76443624e-01
7.00559974e-01 4.09161270e-01 -2.54501879e-01 3.76079947e-01
-1.13435276e-02 6.92505002e-01 1.86962098e-01 -8.71814251e-01
7.25901365e-01 4.65303302e-01 4.96727198e-01 -1.33450186e+00
-7.99090862e-01 -6.10627890e-01 -5.40390670e-01 3.35760415e-01
5.35676479e-01 -4.59469557e-01 -9.73462462e-01 1.72192800e+00
1.19922347e-01 5.04963398e-01 1.83189452e-01 5.23499370e-01
4.04719472e-01 6.20397508e-01 4.50575858e-01 5.82445674e-02
4.85632151e-01 -6.41130865e-01 -7.07697570e-01 -7.05398440e-01
9.50808883e-01 -4.80943054e-01 9.45759296e-01 1.48859784e-01
-6.36253774e-01 -7.08046019e-01 -1.10308194e+00 4.81391490e-01
-5.75655639e-01 -4.45360929e-01 1.25809148e-01 7.10970819e-01
-9.12451029e-01 8.74025881e-01 -6.63546443e-01 -3.42333764e-01
1.61266625e-01 5.59425771e-01 -3.40468854e-01 -1.88062131e-01
-1.23284304e+00 1.42167938e+00 3.64318967e-01 4.09328133e-01
-9.47912574e-01 -1.02614367e+00 -7.49508500e-01 -1.45140529e-01
2.27398828e-01 -5.60253933e-02 1.08089054e+00 -5.33292890e-01
-8.69832695e-01 5.01962788e-02 -2.33584389e-01 -7.62443781e-01
9.32500303e-01 -5.12971640e-01 -9.31974888e-01 -5.31321466e-01
-1.18862279e-01 5.29332936e-01 5.63644111e-01 -1.47941196e+00
-6.53475106e-01 -4.66420911e-02 -2.33182326e-01 8.43418911e-02
-1.00120962e-01 -2.13878632e-01 3.41169238e-02 -2.58361340e-01
-2.35489503e-01 -1.06430042e+00 -4.19568658e-01 -3.45423341e-01
-4.41959769e-01 4.30248469e-01 1.05565369e+00 -8.34441900e-01
1.39636719e+00 -1.93012404e+00 -5.28240263e-01 3.42884928e-01
-5.15643507e-02 7.63281584e-01 -2.91038454e-01 6.12038374e-01
-7.76632205e-02 2.11114496e-01 -7.38443956e-02 -9.56106558e-02
-1.85490534e-01 3.99891019e-01 -3.10481310e-01 4.09214139e-01
2.39790469e-01 7.59502232e-01 -9.35663164e-01 1.61561772e-01
5.78999639e-01 4.45633560e-01 -3.05668861e-01 -4.19062190e-02
-9.77539271e-02 6.79154754e-01 -9.55201685e-02 3.70401084e-01
6.51341677e-01 5.44278920e-01 -1.48921281e-01 1.48289636e-01
-2.86644965e-01 2.36045301e-01 -9.20364499e-01 7.44183958e-01
-4.33242291e-01 9.90447342e-01 -3.38001341e-01 -6.28035724e-01
8.87591720e-01 2.43096724e-01 3.13398361e-01 -1.05429530e+00
6.54199570e-02 2.53091842e-01 3.34136307e-01 -7.91287065e-01
7.55870402e-01 1.73780084e-01 8.23862776e-02 -7.49342665e-02
-6.36692524e-01 2.66184181e-01 -9.91463140e-02 -2.70161897e-01
1.33379459e+00 -8.75307396e-02 -3.60336959e-01 3.87012474e-02
3.25582951e-01 2.82926261e-01 7.55596638e-01 8.43918741e-01
-3.85075629e-01 3.60925019e-01 3.25043648e-01 -5.35732508e-01
-1.25867343e+00 -7.60795057e-01 -1.04128849e-02 7.81296551e-01
1.21366084e-01 -2.72730619e-01 -5.35758257e-01 -5.51753998e-01
2.52322465e-01 1.35358918e+00 -6.38119340e-01 -8.09654713e-01
-7.72581995e-01 -7.09008276e-01 8.34300518e-01 7.66618013e-01
2.85012454e-01 -7.41801918e-01 -7.34309077e-01 3.83174628e-01
-5.16096540e-02 -1.29737282e+00 1.31893352e-01 -1.47347689e-01
-6.66580319e-01 -1.07312405e+00 -1.45070478e-01 -9.50227156e-02
6.63220882e-01 3.80332202e-01 6.72264636e-01 3.15615177e-01
1.14854939e-01 2.91592646e-02 -2.11537182e-01 -6.49777174e-01
-8.87305140e-01 -5.28114736e-02 3.08082610e-01 -8.91039446e-02
2.49598891e-01 -4.17541385e-01 -5.45720994e-01 6.78948164e-01
-4.97577727e-01 -1.30639419e-01 4.49809432e-01 5.40634990e-01
1.17508546e-01 1.62306800e-01 7.61776865e-01 -7.09556460e-01
4.20749068e-01 -7.89900720e-01 -3.46454144e-01 -8.86083543e-02
-8.55917335e-01 -4.43752408e-02 7.08822489e-01 -5.35335302e-01
-9.20096159e-01 8.33927765e-02 -3.62829328e-01 -3.99089038e-01
-6.59190953e-01 3.61318290e-01 -8.50236192e-02 -1.04757935e-01
1.03833580e+00 5.54428995e-02 -4.81398636e-03 -4.28567171e-01
8.58780295e-02 4.95598346e-01 5.84424675e-01 -7.57365227e-02
1.09344923e+00 3.84678811e-01 8.35930780e-02 -1.01335943e+00
-3.99964154e-01 -3.75494599e-01 -4.13930416e-01 -6.13413393e-01
4.34257448e-01 -8.26347232e-01 -4.69554484e-01 5.22735000e-01
-1.01242054e+00 -7.66958654e-01 -5.34717701e-02 4.67426181e-01
-2.90141344e-01 1.09599523e-01 -3.36636486e-03 -1.06353056e+00
1.12573661e-01 -1.10247099e+00 5.40125906e-01 1.67065620e-01
-4.95515525e-01 -8.77824664e-01 -1.33248284e-01 2.42412984e-01
9.33237016e-01 5.73106170e-01 7.66544044e-01 -8.03829253e-01
-2.88782090e-01 -6.28488064e-01 7.78884999e-03 2.97835529e-01
5.36535680e-02 2.56953686e-01 -1.20036805e+00 -2.70118088e-01
-6.13024831e-01 2.65935361e-01 6.70451164e-01 4.38417159e-02
9.15207088e-01 -4.51019347e-01 -6.46579802e-01 1.85752541e-01
1.03410828e+00 4.67385709e-01 9.03289139e-01 5.65249920e-01
6.39923334e-01 9.59106326e-01 7.25627959e-01 2.43845563e-02
3.79196405e-01 7.63711214e-01 9.81069744e-01 -9.58793312e-02
-4.44517471e-02 -4.76000249e-01 4.79040265e-01 5.42600174e-03
1.01366982e-01 -3.52410406e-01 -1.32216156e+00 7.84347296e-01
-1.86139798e+00 -1.02002656e+00 -7.31243908e-01 2.67248225e+00
9.79976803e-02 7.12147593e-01 2.85988569e-01 3.66721451e-01
5.91884911e-01 -8.72835070e-02 -7.59980917e-01 -5.26941419e-01
1.03972323e-01 -4.86182749e-01 1.09774208e+00 4.87626880e-01
-9.63631868e-01 8.42141211e-01 6.72816610e+00 2.61969924e-01
-1.39365232e+00 -1.54576197e-01 3.32116783e-01 -1.44091189e-01
-1.85356498e-01 -1.06483936e-01 -7.26408362e-01 7.58293808e-01
1.74427772e+00 3.88097279e-02 1.27901524e-01 6.15934074e-01
8.73706937e-01 -1.45742059e-01 -7.92174637e-01 4.21933800e-01
-2.42111698e-01 -1.18337989e+00 -2.52969593e-01 8.48169401e-02
4.80436087e-01 3.95258665e-01 1.46494150e-01 5.44312716e-01
1.21323735e-01 -1.18326247e+00 6.94613993e-01 6.64907396e-01
4.99660879e-01 -9.49447513e-01 8.75438988e-01 6.64591491e-01
-7.82331645e-01 -3.57618362e-01 -1.84931189e-01 -2.87622809e-01
3.71227980e-01 2.08680362e-01 -1.24109948e+00 3.77038747e-01
5.55370748e-01 4.23117429e-01 -7.42045879e-01 1.36809194e+00
-8.61370042e-02 9.73838389e-01 -4.14559782e-01 1.89348146e-01
3.04128885e-01 4.31676060e-01 9.03016567e-01 1.17624390e+00
2.60599375e-01 -3.78956228e-01 -9.85599384e-02 3.75027955e-01
4.09583688e-01 -3.01003158e-01 -1.03370512e+00 2.44941682e-01
7.25984633e-01 7.32026160e-01 -2.23697916e-01 -9.83147509e-03
-2.76442409e-01 4.49017465e-01 -2.70394702e-02 6.19479299e-01
-1.03486228e+00 -1.89427719e-01 1.15919125e+00 5.43729901e-01
-1.09602332e-01 -2.83726662e-01 -4.34407383e-01 -3.94959033e-01
7.92957749e-03 -5.50412059e-01 -9.63974074e-02 -4.59037989e-01
-5.87597489e-01 5.43208003e-01 7.89782479e-02 -1.39119065e+00
-3.76805335e-01 -4.42872465e-01 -8.95828068e-01 7.94393659e-01
-1.52968097e+00 -8.91012013e-01 -3.28800917e-01 1.30584510e-02
3.43979716e-01 6.55060932e-02 5.97314715e-01 5.81281483e-01
-8.20267797e-01 7.74633110e-01 9.58652422e-02 -2.20654503e-01
6.09859347e-01 -8.14503908e-01 9.33376372e-01 1.08716547e+00
-4.47394162e-01 3.18641126e-01 1.40968990e+00 -8.07467043e-01
-1.19167089e+00 -1.73832440e+00 8.83739591e-01 -6.63614571e-01
4.63354796e-01 -2.64839798e-01 -1.28355205e+00 6.94209337e-01
-3.21157396e-01 -1.74919918e-01 3.77493709e-01 -1.17257342e-01
-1.27007499e-01 -1.85129985e-01 -1.14234042e+00 9.79087830e-01
1.03524101e+00 -1.69975251e-01 -1.39428064e-01 1.94213495e-01
7.27163434e-01 -4.24047887e-01 -5.24067760e-01 8.49903703e-01
5.52724183e-01 -8.70011091e-01 9.71214175e-01 -8.74280095e-01
-7.70622864e-02 -2.89579540e-01 -1.58196986e-01 -1.32802212e+00
-1.60926431e-01 -5.21437109e-01 -5.97932711e-02 1.10413682e+00
8.61162066e-01 -8.21145535e-01 7.53547788e-01 9.87832129e-01
-4.85736519e-01 -6.66351199e-01 -9.01306570e-01 -1.00900471e+00
2.16395304e-01 -1.13717699e+00 7.17686772e-01 7.21563101e-01
-2.26429924e-01 -4.83099557e-02 -6.38210118e-01 6.37765408e-01
2.77093798e-01 -8.27003360e-01 9.92740154e-01 -1.08131206e+00
2.73529768e-01 -4.67433333e-01 -6.77438021e-01 -5.17341435e-01
1.22331262e-01 -3.73945266e-01 2.93304116e-01 -1.39164317e+00
-4.73815799e-01 -6.56305254e-01 -1.45227253e-01 4.68993694e-01
-1.56575963e-01 -1.18377902e-01 4.22832519e-02 -4.85486872e-02
-7.22917765e-02 3.75950158e-01 6.54524088e-01 -4.92024682e-02
-3.51883709e-01 4.96480197e-01 -4.62123811e-01 6.82376206e-01
1.06557095e+00 -4.70812738e-01 -5.17786086e-01 -4.88329351e-01
2.47358844e-01 -1.77886710e-01 7.06948876e-01 -1.46662021e+00
3.16303194e-01 -3.25192988e-01 2.05465198e-01 -7.65924633e-01
4.26035166e-01 -9.38836336e-01 5.44923306e-01 5.28762221e-01
-2.35984996e-01 2.31882349e-01 9.09811258e-01 6.79216981e-01
1.79566920e-01 6.46353364e-02 4.99495268e-01 3.09680551e-01
-9.38198149e-01 4.52368855e-02 -7.57611036e-01 -2.49745339e-01
1.23144221e+00 -4.75091785e-01 -4.93508518e-01 -5.30988157e-01
-5.30510426e-01 5.67782283e-01 4.28380847e-01 8.78448009e-01
4.32843149e-01 -9.44460332e-01 -5.42760313e-01 2.29243129e-01
1.53665379e-01 -2.76324719e-01 4.96614426e-01 8.60014319e-01
-4.31745052e-01 5.03502131e-01 -1.27337754e-01 -3.72595906e-01
-1.15529060e+00 4.49530065e-01 5.53928852e-01 2.13201880e-01
-5.98793745e-01 4.97763097e-01 -2.96656579e-01 -5.85641682e-01
4.29017603e-01 -2.98571169e-01 -2.66115189e-01 -1.33947536e-01
5.40340841e-01 8.45212579e-01 4.44645435e-01 -9.59771514e-01
-3.68545592e-01 1.87237948e-01 4.00094092e-02 7.43215680e-02
1.02686059e+00 -2.61987031e-01 7.28127718e-01 3.78786564e-01
8.40351701e-01 -1.85789630e-01 -1.57915032e+00 7.17590749e-02
2.58282721e-01 -4.45700169e-01 7.88860023e-02 -9.93867517e-01
-6.32809520e-01 9.04211521e-01 7.55071878e-01 7.98938796e-02
6.72739089e-01 -5.96519709e-01 9.37961578e-01 3.99523318e-01
3.35637033e-01 -9.67552304e-01 -5.45348883e-01 5.04959345e-01
7.94205368e-01 -1.16977370e+00 -3.64331335e-01 -1.49137154e-01
-6.74170911e-01 7.84706354e-01 8.82052183e-01 2.61332840e-01
4.81380880e-01 1.72623783e-01 2.13569388e-01 8.56354088e-02
-8.84632409e-01 -2.19929084e-01 1.69392377e-01 7.88093030e-01
-1.75886869e-01 1.53312832e-01 2.41308510e-01 1.92480415e-01
-2.35133350e-01 -4.04331625e-01 6.76014543e-01 8.00722659e-01
-4.68116462e-01 -7.39925206e-01 -4.41252768e-01 6.17218316e-01
-7.30824545e-02 2.65403718e-01 -1.81469381e-01 9.32559788e-01
1.95617318e-01 1.24538970e+00 1.45480111e-01 -9.87021685e-01
8.71108592e-01 1.39457375e-01 -3.03776413e-01 -3.88383716e-01
-4.49399084e-01 -5.97774386e-01 5.18734515e-01 -6.83988452e-01
1.04414038e-01 -8.83975089e-01 -1.28736305e+00 -5.18069923e-01
-1.16780646e-01 -2.55507022e-01 7.67153859e-01 1.09566665e+00
6.54348850e-01 7.92855084e-01 4.65028673e-01 -9.11686182e-01
-5.86022675e-01 -7.89047301e-01 4.22954820e-02 2.53175288e-01
7.57464767e-01 -9.14342582e-01 -4.75861698e-01 -3.57451022e-01] | [5.780621528625488, 1.0670394897460938] |
1aae8bb9-9b4a-49af-9c5d-c9b987d14d5a | a-novel-multimodal-music-genre-classifier | 2011.11970 | null | https://arxiv.org/abs/2011.11970v1 | https://arxiv.org/pdf/2011.11970v1.pdf | A Novel Multimodal Music Genre Classifier using Hierarchical Attention and Convolutional Neural Network | Music genre classification is one of the trending topics in regards to the current Music Information Retrieval (MIR) Research. Since, the dependency of genre is not only limited to the audio profile, we also make use of textual content provided as lyrics of the corresponding song. We implemented a CNN based feature extractor for spectrograms in order to incorporate the acoustic features and a Hierarchical Attention Network based feature extractor for lyrics. We then go on to classify the music track based upon the resulting fused feature vector. | ['Abhilash Nandy', 'Manish Agrawal'] | 2020-11-24 | null | null | null | null | ['genre-classification'] | ['computer-vision'] | [ 2.89723992e-01 -4.44284439e-01 -8.53919610e-02 -8.35056230e-02
-8.15225720e-01 -6.22353435e-01 5.92911363e-01 1.05602995e-01
-5.25759459e-01 3.85770649e-01 6.61518872e-01 1.30579278e-01
-4.41665590e-01 -5.82610488e-01 -2.37390772e-01 -7.38008022e-01
2.41757229e-01 -1.86563283e-02 -2.44489878e-01 -8.24049115e-02
4.71851736e-01 4.49655771e-01 -1.90694058e+00 3.21280897e-01
2.97502369e-01 1.51016152e+00 -2.10758559e-02 8.53411674e-01
1.27400542e-02 6.28708482e-01 -6.90868616e-01 -6.29459918e-02
1.62307560e-01 -4.95961159e-01 -5.32713354e-01 -1.00871800e-02
1.88429460e-01 -1.16485104e-01 -4.54404086e-01 8.20062160e-01
6.70800209e-01 4.30786639e-01 7.38975525e-01 -8.55946720e-01
-7.65207857e-02 9.19031203e-01 -2.06365317e-01 4.64115769e-01
4.65443403e-01 -2.18103990e-01 1.29380631e+00 -8.18063140e-01
5.86601317e-01 6.40493572e-01 3.64355445e-01 1.03271917e-01
-7.36176908e-01 -6.86316013e-01 -8.05036724e-02 4.79809791e-01
-1.49192524e+00 -6.27300203e-01 1.32387102e+00 -5.41326344e-01
7.71457434e-01 3.13567907e-01 8.84922147e-01 1.05067241e+00
1.81852490e-01 7.21888006e-01 7.21538842e-01 -5.04283905e-01
9.47310403e-02 -1.11105219e-01 1.36260450e-01 1.76814720e-01
-4.77280825e-01 -1.24179319e-01 -7.77183235e-01 -1.06985599e-01
5.57945609e-01 -3.69405076e-02 -2.70102561e-01 2.35488281e-01
-1.12043166e+00 8.06616008e-01 3.53805393e-01 7.13714540e-01
-6.27922475e-01 2.40793511e-01 6.01042867e-01 4.16138858e-01
2.51333773e-01 7.62505829e-01 -1.97519228e-01 -4.42357272e-01
-1.33857810e+00 3.25053513e-01 6.99868381e-01 2.18098432e-01
4.54682976e-01 3.22963238e-01 -1.38381004e-01 1.00235057e+00
3.74374896e-01 1.23615831e-01 9.04627264e-01 -9.15550470e-01
1.18435770e-01 4.68563497e-01 -2.29460806e-01 -1.19502807e+00
-3.66184473e-01 -7.88149536e-01 -5.65891922e-01 -1.97781727e-01
1.72594041e-01 8.07847641e-03 -4.55472469e-01 1.50504327e+00
1.24982692e-01 2.44267091e-01 -1.94368467e-01 1.03005195e+00
9.48646665e-01 4.37093049e-01 -2.87308902e-01 -1.69105500e-01
1.34623969e+00 -6.30497277e-01 -7.20659792e-01 3.80355418e-01
1.66153938e-01 -1.04225481e+00 8.26602936e-01 7.41486549e-01
-7.92241633e-01 -6.29577935e-01 -1.39187479e+00 -1.22704701e-02
-2.67661691e-01 3.74579728e-01 3.58161211e-01 2.57725865e-01
-5.82392037e-01 8.37842643e-01 -5.29810727e-01 -1.56260833e-01
2.26182655e-01 3.43284726e-01 -2.45735481e-01 3.83626193e-01
-1.12955368e+00 4.34609026e-01 3.99457425e-01 2.09287971e-01
-7.09745467e-01 -3.96536857e-01 -4.80650187e-01 2.77305901e-01
1.54566318e-01 -3.77849609e-01 1.21131086e+00 -1.05555129e+00
-1.82133710e+00 4.96670932e-01 1.37003466e-01 -5.23528337e-01
2.99621616e-02 -2.98807025e-01 -5.64972520e-01 3.41666758e-01
-9.41539779e-02 1.10044733e-01 1.05853987e+00 -7.48491824e-01
-6.68727219e-01 -2.31918737e-01 -7.69711882e-02 4.18874085e-01
-6.34560406e-01 1.77638650e-01 -2.85286427e-01 -8.94878387e-01
8.80789608e-02 -8.14325631e-01 2.10281953e-01 -5.75750589e-01
-6.54057503e-01 -3.49615335e-01 7.19025135e-01 -7.28337705e-01
1.70144284e+00 -2.43822598e+00 3.68920356e-01 2.10266575e-01
1.02197327e-01 5.10769114e-02 2.19974741e-02 7.19277263e-01
7.25933984e-02 -2.61126697e-01 1.13808200e-01 -5.88047914e-02
1.64098948e-01 -3.69377315e-01 -4.84796107e-01 3.06721270e-01
-7.38667650e-03 6.28989935e-01 -5.75849473e-01 -3.62064719e-01
2.58255631e-01 4.90462273e-01 -5.90213358e-01 2.70172924e-01
-5.79302013e-02 5.83984971e-01 -5.29980540e-01 4.66984868e-01
1.00080363e-01 2.38476560e-01 -1.06538303e-01 -5.07969856e-01
-2.58394957e-01 6.75959766e-01 -1.14209449e+00 2.04599953e+00
-4.68431532e-01 7.14577436e-01 -3.85973155e-02 -6.98961377e-01
1.03396010e+00 7.24047244e-01 8.37993145e-01 -2.83983499e-01
5.46485424e-01 2.46613622e-01 3.59499663e-01 -4.27733958e-01
7.35535800e-01 -9.02925432e-02 -1.97657317e-01 5.49810350e-01
3.89686167e-01 1.48591578e-01 2.65878979e-02 -2.17491478e-01
1.08203149e+00 1.67819694e-01 3.49515527e-01 -3.35578993e-02
6.41868472e-01 -1.49741650e-01 2.98081309e-01 5.52254498e-01
3.51934135e-02 7.25625694e-01 2.97187548e-02 -1.81527138e-01
-9.47253466e-01 -7.53508210e-01 -2.85296321e-01 1.34080839e+00
-2.98795879e-01 -5.60015440e-01 -5.03197968e-01 -2.71992534e-01
-2.49514595e-01 2.67826468e-01 -4.65186656e-01 -1.03095748e-01
-5.10770500e-01 -3.30989480e-01 6.61094069e-01 2.38608092e-01
1.47449046e-01 -1.62478614e+00 -4.68168765e-01 5.57957709e-01
-7.87890628e-02 -8.03235233e-01 -5.07476568e-01 5.45306087e-01
-5.33605695e-01 -1.03194439e+00 -5.46147823e-01 -5.72889686e-01
-2.97616273e-01 -1.81580395e-01 8.18386316e-01 -2.00275064e-01
-3.34485710e-01 4.05832142e-01 -6.09344840e-01 -3.86920273e-01
-2.59288937e-01 5.50823927e-01 6.53755963e-02 4.62514222e-01
3.65845591e-01 -9.55456793e-01 -6.67597651e-01 -1.30887643e-01
-8.35010290e-01 -2.25204006e-01 3.62053335e-01 7.08148181e-01
4.70663190e-01 9.97998640e-02 7.93444097e-01 -5.74731827e-01
7.13060558e-01 -3.44184667e-01 -1.52697057e-01 -3.45104784e-01
-2.30892256e-01 -3.24581265e-01 5.69729984e-01 -6.54639363e-01
-7.75686443e-01 8.46698657e-02 -4.21578169e-01 -4.19433385e-01
-4.01254475e-01 8.67836595e-01 -1.86669514e-01 2.49426216e-01
4.50454354e-01 2.80938894e-01 -5.66375911e-01 -8.42564106e-01
2.45343789e-01 1.14892411e+00 7.27385402e-01 -4.10302788e-01
5.79227209e-01 2.23628640e-01 -1.65263802e-01 -1.02358842e+00
-9.07054543e-01 -6.89144075e-01 -7.13428855e-01 -5.76772273e-01
9.08107698e-01 -6.67290330e-01 -9.21810389e-01 2.32718378e-01
-1.00032854e+00 3.53421986e-01 -2.34930113e-01 8.78079057e-01
-6.21861994e-01 2.05681860e-01 -7.53729939e-01 -9.84043062e-01
-6.28572285e-01 -9.77881908e-01 9.19866323e-01 4.29423116e-02
-5.42991638e-01 -4.81445253e-01 3.28211457e-01 3.62258911e-01
2.48351693e-01 1.48079500e-01 1.00939059e+00 -1.00059259e+00
-2.53000557e-01 -4.64805663e-01 6.35524541e-02 3.09581608e-01
2.05085829e-01 -1.06732026e-01 -1.42061317e+00 -4.28620838e-02
2.07628250e-01 -2.36134842e-01 9.73039448e-01 4.48943973e-01
1.04765463e+00 -2.89997876e-01 2.60521084e-01 6.27538860e-01
1.31502581e+00 3.01163644e-01 3.18306386e-01 4.74371254e-01
8.84465098e-01 3.15858096e-01 3.37481648e-01 5.39428771e-01
2.11839490e-02 1.04862893e+00 2.87881821e-01 3.10078084e-01
-1.73899233e-01 -3.63311738e-01 3.19966465e-01 1.46352267e+00
-6.96947724e-02 -2.98090111e-02 -6.40494466e-01 4.97639120e-01
-1.86179602e+00 -9.90283370e-01 1.96972340e-01 2.19015336e+00
6.75435603e-01 -1.42733082e-01 2.00717166e-01 7.92975962e-01
5.06718338e-01 1.75796598e-01 -3.80466402e-01 -2.85105199e-01
8.92225951e-02 3.53274882e-01 9.14949775e-02 4.18632701e-02
-1.33226752e+00 7.09602654e-01 5.97350740e+00 9.18487668e-01
-1.40021980e+00 2.74678990e-02 -1.28607884e-01 -3.05202484e-01
2.49914769e-02 -1.72613651e-01 -4.30357724e-01 4.42834109e-01
9.52882946e-01 7.39777908e-02 7.74927378e-01 6.61109805e-01
3.27336460e-01 1.05835736e-01 -1.12535560e+00 1.06628168e+00
2.22934082e-01 -1.05148673e+00 1.27729654e-01 1.32729769e-01
4.92473543e-01 1.04505837e-01 1.04448840e-01 2.25520462e-01
-2.84030855e-01 -1.03572869e+00 8.15611362e-01 7.31187344e-01
4.59361672e-01 -1.00984836e+00 8.18705440e-01 2.75852591e-01
-1.35827470e+00 -1.81753755e-01 -2.45442227e-01 -2.28497073e-01
1.54796494e-02 2.43290663e-01 -9.77637529e-01 6.05325043e-01
6.90201819e-01 9.47520852e-01 -3.43815058e-01 1.30536115e+00
4.23633382e-02 8.74784231e-01 -1.22401662e-01 6.51493371e-02
3.66191822e-03 -1.80177540e-01 7.98390687e-01 1.26607668e+00
4.34467524e-01 -3.08473796e-01 1.25832155e-01 5.77627718e-01
-2.61908412e-01 5.76720297e-01 -5.10666490e-01 -4.72561955e-01
4.16834861e-01 1.44604361e+00 -6.30134106e-01 -1.96973816e-01
-1.39145285e-01 7.71781564e-01 6.58117682e-02 3.45814407e-01
-2.69634068e-01 -5.76412439e-01 5.82873225e-01 -2.54511628e-02
4.65325207e-01 -1.08314656e-01 -9.42870602e-02 -1.08048749e+00
-2.31978759e-01 -9.16971326e-01 3.90318245e-01 -7.66543567e-01
-1.41100323e+00 7.92245924e-01 -4.21230108e-01 -1.54812336e+00
-5.04297435e-01 -3.24144930e-01 -5.50381899e-01 8.49578440e-01
-1.25366724e+00 -1.06612039e+00 1.63587295e-02 6.03293419e-01
5.97199798e-01 -4.74511683e-01 1.11894929e+00 6.84953570e-01
-3.51794899e-01 3.86012673e-01 1.82796270e-01 3.92331839e-01
7.49857366e-01 -1.02595127e+00 -1.33071020e-01 3.67849350e-01
7.74580240e-01 5.99862278e-01 6.48286223e-01 -2.79093295e-01
-1.34787476e+00 -9.41276073e-01 8.96845460e-01 -1.89100236e-01
9.46881115e-01 -2.57736027e-01 -7.14068890e-01 2.37903252e-01
2.33834028e-01 -5.18764913e-01 9.70501304e-01 3.21316004e-01
-3.18908900e-01 -2.43060738e-01 -3.97295982e-01 3.31149936e-01
6.06243610e-01 -1.07894659e+00 -6.87595963e-01 -1.06007397e-01
6.31969869e-01 -2.18202218e-01 -1.06981266e+00 2.26067111e-01
1.02146482e+00 -5.92687249e-01 7.45576620e-01 -4.77135122e-01
4.78487194e-01 -3.50129366e-01 -4.10276085e-01 -1.15749121e+00
-2.52701998e-01 -3.76379132e-01 -3.90164256e-02 1.37125540e+00
3.02138984e-01 -1.03076182e-01 6.71819866e-01 -2.41878405e-01
-2.64219671e-01 -4.28716183e-01 -1.02193165e+00 -3.36314738e-01
-4.12778705e-01 -8.74382854e-01 6.19968891e-01 7.43416131e-01
1.64932430e-01 5.90803266e-01 -6.84114397e-01 -2.07074612e-01
2.46478885e-01 6.01648569e-01 6.56471968e-01 -1.52740753e+00
-6.64521396e-01 -6.69609189e-01 -6.11937344e-01 -6.58161163e-01
-8.99567828e-03 -1.25167155e+00 1.31319970e-01 -1.22586560e+00
2.92105258e-01 5.67498654e-02 -1.07468474e+00 2.10339859e-01
3.07048678e-01 8.49572480e-01 4.23342258e-01 3.85009825e-01
-6.00723743e-01 5.85060894e-01 9.57040012e-01 -2.51827419e-01
-4.98796791e-01 3.48838568e-01 -5.42131126e-01 7.52964258e-01
6.74767792e-01 -5.66656411e-01 -8.82101059e-02 -3.03479508e-02
3.99530292e-01 2.33189851e-01 1.86717823e-01 -1.22849941e+00
3.15798968e-01 1.59330517e-01 2.91266143e-01 -7.98627317e-01
7.05952764e-01 -7.84063756e-01 1.27745420e-01 6.42654225e-02
-7.28178620e-01 -1.89104363e-01 6.40902594e-02 4.26182508e-01
-6.37649834e-01 -4.20425415e-01 3.68152618e-01 1.20390922e-01
-2.27898285e-01 2.70567834e-01 -6.12811327e-01 -2.51033485e-01
8.15074146e-02 -2.15431467e-01 1.51376799e-01 -4.52903420e-01
-9.27173376e-01 -5.19217968e-01 -3.20198461e-02 5.81056118e-01
3.38742584e-01 -1.58633959e+00 -5.79277098e-01 1.78900123e-01
2.88515061e-01 -6.64935291e-01 2.88665056e-01 5.72050571e-01
-6.20698482e-02 5.46277940e-01 -2.70926565e-01 -3.54691595e-01
-1.33222592e+00 3.89081031e-01 3.70561033e-02 -5.82203381e-02
-5.48761785e-01 5.66397309e-01 5.87594183e-03 -6.31195530e-02
5.33334017e-01 -1.08700790e-01 -8.90685618e-01 3.94282788e-01
5.32057047e-01 2.38240227e-01 2.30710357e-01 -8.83713782e-01
-2.42404193e-01 6.46031380e-01 2.02572376e-01 -5.47876239e-01
1.58818841e+00 -6.27809465e-02 2.25312673e-02 9.56888735e-01
1.27357423e+00 1.68991297e-01 -8.06587934e-01 -3.94170374e-01
1.70021072e-01 -2.64541328e-01 3.60901058e-01 -7.38698184e-01
-8.12611401e-01 8.13415110e-01 6.05267882e-01 3.30573052e-01
1.18172598e+00 -2.16972321e-01 7.60026455e-01 3.69262457e-01
1.52263895e-01 -1.16226554e+00 -1.29316047e-01 6.13823593e-01
9.14672792e-01 -7.58790791e-01 -4.60609794e-02 1.76133677e-01
-3.60912651e-01 1.22091651e+00 -4.68309522e-02 -4.73257571e-01
1.00391972e+00 8.95428807e-02 3.56971264e-01 -2.41201505e-01
-5.63993156e-01 -3.09526831e-01 8.67688894e-01 3.42901528e-01
8.04509878e-01 -6.54458106e-02 -3.59736979e-01 1.03238273e+00
-8.53498042e-01 -4.93033184e-03 8.09024423e-02 3.62343699e-01
-5.94052017e-01 -1.11750913e+00 -3.55698198e-01 6.27547801e-01
-8.74776721e-01 -2.09759295e-01 -5.01193225e-01 1.73519462e-01
2.52224028e-01 1.15344763e+00 -8.00244212e-02 -6.73712552e-01
3.38964835e-02 2.62674272e-01 3.10526311e-01 -6.96659029e-01
-1.19649839e+00 7.04807639e-01 -6.34847283e-02 -3.74614239e-01
-3.16829175e-01 -4.66240585e-01 -9.96162295e-01 1.46657690e-01
-3.87746841e-01 2.07311779e-01 1.10379851e+00 1.08005023e+00
9.88250524e-02 1.04186463e+00 8.80838513e-01 -1.19856369e+00
-2.76993841e-01 -1.33684456e+00 -8.54730368e-01 2.69494027e-01
3.43117595e-01 -5.24403751e-01 -2.58340627e-01 -7.32501298e-02] | [15.857216835021973, 5.1790876388549805] |
dac8c475-60f3-4988-b3bd-60fab2b08556 | hyperspectral-unmixing-with-endmember-1 | 1609.03500 | null | http://arxiv.org/abs/1609.03500v1 | http://arxiv.org/pdf/1609.03500v1.pdf | Hyperspectral Unmixing with Endmember Variability using Partial Membership Latent Dirichlet Allocation | The application of Partial Membership Latent Dirichlet Allocation(PM-LDA) for
hyperspectral endmember estimation and spectral unmixing is presented. PM-LDA
provides a model for a hyperspectral image analysis that accounts for spectral
variability and incorporates spatial information through the use of
superpixel-based 'documents.' In our application of PM-LDA, we employ the
Normal Compositional Model in which endmembers are represented as Normal
distributions to account for spectral variability and proportion vectors are
modeled as random variables governed by a Dirichlet distribution. The use of
the Dirichlet distribution enforces positivity and sum-to-one constraints on
the proportion values. Algorithm results on real hyperspectral data indicate
that PM-LDA produces endmember distributions that represent the ground truth
classes and their associated variability. | ['Alina Zare', 'Sheng Zou'] | 2016-09-12 | null | null | null | null | ['hyperspectral-unmixing'] | ['computer-vision'] | [ 4.90170985e-01 -3.16700101e-01 -3.35164219e-01 -1.81309819e-01
-3.24770182e-01 -7.60081708e-01 7.69248545e-01 -2.29012966e-01
2.58517504e-01 8.84638608e-01 4.07326818e-01 -3.42413843e-01
-1.07330002e-01 -1.00437462e+00 -1.95149347e-01 -1.30084682e+00
1.38806105e-01 6.22930825e-01 -4.35335070e-01 2.57711142e-01
-5.79370372e-02 5.68831146e-01 -1.39357436e+00 3.56069416e-01
8.63831758e-01 6.32106364e-01 1.21546812e-01 3.94475460e-01
-1.59523353e-01 4.53541130e-01 -4.08417284e-01 4.99572963e-01
4.57367837e-01 -3.27595562e-01 -5.44850230e-01 7.51342654e-01
1.75535232e-01 -3.39376301e-01 1.52122006e-01 1.36467874e+00
2.54629582e-01 2.90456116e-01 1.52547610e+00 -1.24794865e+00
-7.37051070e-01 4.72558886e-01 -9.42887545e-01 -3.23654503e-01
-8.26078877e-02 2.09915116e-02 9.12139118e-01 -8.48166764e-01
3.62369627e-01 1.34903681e+00 7.99340665e-01 -1.25859052e-01
-1.86116707e+00 -4.92318332e-01 -2.49189734e-02 -1.18702576e-01
-1.80405533e+00 -1.52050853e-01 6.60300255e-01 -1.07551348e+00
8.17896128e-01 5.27393222e-01 8.96986425e-01 3.09966654e-01
-4.97474372e-02 5.95791519e-01 1.27327597e+00 -6.42808378e-01
4.63183284e-01 7.47976825e-02 3.71509194e-01 8.28688070e-02
6.14969075e-01 2.64693443e-02 -1.40402228e-01 -7.39061594e-01
5.31468272e-01 1.80100396e-01 -2.83783257e-01 -5.28135538e-01
-9.89402890e-01 9.91968453e-01 1.48271397e-01 3.98297794e-03
-8.60849321e-01 -1.05646119e-01 -8.78277868e-02 -2.01791450e-01
1.19973099e+00 1.20938271e-01 -5.18816933e-02 7.19695330e-01
-1.39002967e+00 1.33770093e-01 6.56999826e-01 7.87808359e-01
1.18907773e+00 2.99737155e-01 -1.82031631e-01 1.10700452e+00
8.67998064e-01 1.18446755e+00 2.98247546e-01 -1.02651763e+00
-4.38558280e-01 4.03562754e-01 6.24669552e-01 -7.88259864e-01
8.29151422e-02 -2.37067536e-01 -7.50283062e-01 4.07233059e-01
9.97780859e-02 -9.00202990e-02 -1.32974672e+00 1.29873538e+00
4.30231877e-02 -3.80753055e-02 1.20623820e-01 9.21056509e-01
4.40575242e-01 1.21418762e+00 6.57374203e-01 -6.98250890e-01
9.35197771e-01 -4.07982469e-01 -8.46413374e-01 3.46106030e-02
2.46385679e-01 -6.25044525e-01 7.37184107e-01 4.33279306e-01
-6.01809621e-01 -1.01222567e-01 -9.33910429e-01 2.95492142e-01
-5.28776109e-01 1.65773898e-01 4.27195847e-01 8.19060028e-01
-9.54330444e-01 2.35732153e-01 -8.17190945e-01 -2.88243651e-01
4.07492995e-01 -2.91394815e-02 8.76665264e-02 3.49008478e-02
-7.75186956e-01 6.32976890e-01 4.81591314e-01 -3.93306371e-03
-1.06675971e+00 -6.23868942e-01 -6.83994830e-01 1.20904729e-01
-4.13954794e-01 -3.82909894e-01 8.36635470e-01 -1.26005626e+00
-1.33215868e+00 9.56233859e-01 -3.86810839e-01 -1.02591999e-01
4.20394950e-02 2.36620426e-01 -4.00875717e-01 3.18010777e-01
6.49189875e-02 5.73517859e-01 1.09634674e+00 -1.80770779e+00
-2.30861783e-01 -3.16724032e-01 -6.03522718e-01 3.35984111e-01
-3.37664902e-01 -1.60808370e-01 5.36871672e-01 -6.28542125e-01
7.96808839e-01 -1.03871810e+00 -1.40879139e-01 -1.01787068e-01
-3.49537015e-01 1.02583468e-01 9.58002090e-01 -8.00795257e-01
1.06083679e+00 -2.39597416e+00 3.43176275e-02 5.87623477e-01
8.00022408e-02 2.98326425e-02 -1.16310352e-02 2.33471677e-01
-5.32250285e-01 1.30011141e-01 -9.55759764e-01 -3.99994962e-02
2.46366486e-01 2.26567253e-01 -3.36793810e-01 9.92277265e-01
-1.13213100e-01 2.17576697e-01 -9.46123183e-01 -5.87862246e-02
4.31968749e-01 4.96717304e-01 -1.87442675e-01 -8.43795761e-02
-4.86745477e-01 2.18590698e-03 1.95484862e-01 6.71909332e-01
1.27938485e+00 -1.13881908e-01 4.38728422e-01 -1.55564591e-01
-5.84900320e-01 -6.64998367e-02 -1.16618216e+00 1.14728606e+00
2.04871461e-01 5.99497736e-01 4.93399203e-01 -6.09676063e-01
8.23064685e-01 3.87993306e-01 5.76734126e-01 3.74007225e-01
-2.22317040e-01 1.26250014e-01 -3.37112635e-01 -1.15365081e-01
5.72377086e-01 -7.12441564e-01 5.65364420e-01 6.31702662e-01
-2.35726722e-02 -3.73449177e-01 -5.79635054e-02 -7.14685544e-02
1.41541019e-01 1.62508413e-01 4.29475933e-01 -9.84555364e-01
2.69726276e-01 5.79096079e-01 3.05821955e-01 4.18362975e-01
4.34052013e-02 4.39784795e-01 1.86528802e-01 -9.25507918e-02
-1.15572178e+00 -1.36301315e+00 -6.00645304e-01 1.13552535e+00
-1.97674334e-01 3.53998095e-01 -3.75119984e-01 2.02255890e-01
1.35496810e-01 1.17836595e+00 -2.89800972e-01 3.17536086e-01
5.19998789e-01 -1.80535614e+00 7.38336166e-05 1.16737455e-01
2.49691427e-01 -6.37866139e-01 -6.12449348e-02 1.62974536e-01
-2.87673414e-01 -2.80231267e-01 6.44321665e-02 6.48645237e-02
-7.73097694e-01 -1.03583860e+00 -1.04425275e+00 -5.08064747e-01
9.65626359e-01 4.62119430e-01 7.56118357e-01 -8.55202734e-01
-2.74417728e-01 5.74193358e-01 -3.66687834e-01 -4.98973042e-01
-5.47580421e-01 -4.38268840e-01 7.64570683e-02 1.70019120e-01
8.72981906e-01 -7.23662734e-01 -5.31668901e-01 6.30581379e-02
-1.13515985e+00 -2.94422396e-02 1.80873066e-01 6.31558299e-01
7.11960852e-01 3.02573681e-01 2.75435358e-01 -8.51912856e-01
5.17823160e-01 -7.07724392e-01 -7.59004831e-01 1.97065860e-01
-8.58073473e-01 -3.63829345e-01 -3.16755950e-01 -3.91654938e-01
-1.46646631e+00 1.40438467e-01 6.38665795e-01 -3.25804949e-01
-5.27685046e-01 8.76607776e-01 -2.29711428e-01 6.32844120e-02
1.02335346e+00 3.41099560e-01 8.85393322e-02 -3.03600967e-01
5.96438587e-01 1.08483076e+00 3.19374353e-01 -4.09819663e-01
4.31260645e-01 7.71506310e-01 -2.75957659e-02 -1.56381202e+00
-5.32533526e-01 -1.04484367e+00 -7.22053766e-01 -2.81286806e-01
9.09217775e-01 -1.35369909e+00 -9.75248367e-02 7.11542249e-01
-8.40009928e-01 -3.78372818e-01 -5.31799018e-01 7.30130076e-01
-3.51891816e-01 5.15400589e-01 -3.56265306e-01 -1.29582846e+00
-2.71045119e-01 -7.37013519e-01 6.49529457e-01 -1.64310470e-01
-1.86811283e-01 -1.12741935e+00 1.56245351e-01 6.69099167e-02
3.51834536e-01 3.74911278e-01 1.13999915e+00 -3.23723972e-01
-2.90661961e-01 -3.96664906e-03 -4.36813593e-01 7.90754080e-01
4.87724930e-01 1.59860253e-01 -1.36740649e+00 -2.78115660e-01
8.16218704e-02 2.59671599e-01 1.02435851e+00 1.04085457e+00
8.85230660e-01 -4.96110529e-01 -1.64303869e-01 5.09729326e-01
1.76558292e+00 3.85447703e-02 6.92775786e-01 1.19568473e-02
7.84409404e-01 6.97446048e-01 1.47033140e-01 8.55728745e-01
-6.97446018e-02 -1.61544010e-01 5.53076744e-01 -2.37580582e-01
1.19951628e-01 1.32480443e-01 2.28267476e-01 4.49838847e-01
1.76111162e-02 -2.68499076e-01 -1.14758039e+00 9.30020511e-01
-1.52145886e+00 -1.20788884e+00 -8.21686327e-01 2.19192648e+00
8.51989269e-01 -7.08538353e-01 -2.12263111e-02 1.31439015e-01
1.04822338e+00 3.42828482e-01 -2.94114351e-01 1.27078101e-01
-4.88818169e-01 -8.69516656e-02 8.70608985e-01 8.98892581e-01
-1.22126627e+00 8.35563779e-01 7.57297611e+00 4.33523059e-01
-9.56419528e-01 1.88170373e-01 4.55799222e-01 8.60569403e-02
-5.87505639e-01 2.52562851e-01 -4.14296389e-01 2.94982225e-01
5.86733937e-01 -2.02699915e-01 5.45649350e-01 5.43298423e-01
6.28158450e-01 -5.02976954e-01 -3.99515927e-01 8.28534305e-01
9.16706696e-02 -8.87048542e-01 4.38342422e-01 1.41878873e-01
1.37982512e+00 1.11019701e-01 2.18543932e-01 -4.68943745e-01
6.62847459e-01 -7.79228926e-01 5.30329943e-01 6.97459161e-01
1.02685010e+00 -6.53407753e-01 5.19170821e-01 1.02784261e-01
-7.76839614e-01 8.37902501e-02 -7.72562802e-01 -2.32793003e-01
-6.50458932e-02 1.19751275e+00 -1.10026658e+00 7.72048831e-02
3.21537673e-01 7.81379580e-01 1.19354211e-01 9.42956150e-01
-2.98309952e-01 9.64417100e-01 -4.78839308e-01 4.39441949e-01
1.11068211e-01 -8.88958454e-01 8.30866873e-01 1.39164257e+00
5.81979096e-01 3.42421561e-01 2.27270216e-01 1.36220205e+00
3.85345757e-01 1.83443248e-01 -4.71607000e-01 -5.51792800e-01
5.91119409e-01 9.87419069e-01 -6.70475125e-01 -5.23562312e-01
-4.51041572e-02 7.01087177e-01 -7.50450015e-01 1.00737536e+00
-2.13174433e-01 1.09015189e-01 7.79747963e-01 3.35253179e-01
1.40982255e-01 -4.16521221e-01 -5.88691950e-01 -9.80379939e-01
-6.83451772e-01 -5.49327731e-01 2.73533493e-01 -1.22113216e+00
-1.67175925e+00 1.44322410e-01 3.93164665e-01 -1.19867313e+00
5.45314923e-02 -8.03374171e-01 -3.13533604e-01 1.50485659e+00
-1.60326850e+00 -1.43477869e+00 -3.79806012e-01 5.38283825e-01
-4.42273682e-03 -2.48099938e-01 1.27991462e+00 -2.39595622e-01
-2.65938133e-01 -4.45592225e-01 9.91264284e-01 -2.37304792e-01
6.22142196e-01 -1.33608150e+00 -4.25305128e-01 8.39807332e-01
-4.18953776e-01 5.16498268e-01 8.71085107e-01 -9.80980158e-01
-6.18010283e-01 -1.37438715e+00 5.61687469e-01 1.05902269e-01
6.72880888e-01 3.79215777e-01 -7.38970101e-01 6.60183430e-01
1.98708951e-01 -3.49186122e-01 1.55032146e+00 1.10099599e-01
-1.82522193e-01 8.25268105e-02 -1.25919151e+00 6.46300986e-02
5.72043620e-02 -6.59548163e-01 -3.80926818e-01 8.71454239e-01
-6.49424503e-04 1.99985579e-01 -8.94578874e-01 1.33579910e-01
3.49360943e-01 -4.61113602e-01 8.49850535e-01 -2.82457262e-01
7.71138817e-02 -6.97559595e-01 -4.79191840e-01 -1.47067106e+00
-8.81936371e-01 -2.33936429e-01 1.95723876e-01 9.57165122e-01
3.76755446e-01 -5.45761049e-01 4.04830694e-01 4.88636434e-01
-1.12950198e-01 4.94894058e-01 -5.03738105e-01 -5.42151451e-01
1.46737427e-01 -1.62231937e-01 6.23460412e-01 1.50408781e+00
1.86889041e-02 -2.59064466e-01 -2.25341797e-01 6.20702863e-01
1.13176966e+00 1.86122686e-01 1.92417815e-01 -1.64468515e+00
5.58148474e-02 -3.41662914e-01 -4.69912514e-02 -6.46221101e-01
5.18878877e-01 -9.06271040e-01 3.25523376e-01 -1.55979633e+00
4.57656354e-01 -2.95906603e-01 -1.10022843e-01 6.33939803e-01
1.75048500e-01 4.56342936e-01 -4.08545732e-01 6.21805310e-01
3.05573404e-01 3.81107122e-01 7.26505578e-01 -5.93166649e-01
-6.11461699e-01 -1.82684049e-01 -5.36648393e-01 8.42658639e-01
8.93059492e-01 -4.94185477e-01 -4.17852104e-01 -2.85804749e-01
2.64263362e-01 -3.63924176e-01 4.97490972e-01 -6.77242160e-01
-1.62715614e-01 -5.60857892e-01 3.62251729e-01 -8.58957410e-01
3.84382814e-01 -8.73656631e-01 7.11100042e-01 1.16261676e-01
-1.32390723e-01 -8.29633951e-01 2.79459238e-01 6.03283823e-01
-8.58192071e-02 -1.86223567e-01 1.01924324e+00 -2.20379248e-01
-5.58524668e-01 1.20367922e-01 -1.11526942e+00 -7.37004340e-01
8.54171276e-01 -3.68046582e-01 -2.76860833e-01 -2.90897995e-01
-1.20166802e+00 -1.95509478e-01 6.11301661e-01 -4.93749201e-01
4.32145149e-01 -1.41670883e+00 -1.12310088e+00 2.14240119e-01
1.25117600e-01 -5.29649295e-02 2.16390371e-01 6.26274049e-01
-9.72455800e-01 2.72633314e-01 -3.21205050e-01 -6.97655380e-01
-1.08250856e+00 3.40476841e-01 5.16444147e-01 3.35250348e-01
-5.16762398e-02 7.31418073e-01 3.86345506e-01 -2.74313003e-01
-3.31028640e-01 -4.55632769e-02 -1.91671595e-01 5.12564242e-01
3.60041380e-01 7.41669476e-01 -3.48633766e-01 -9.38771784e-01
-1.36487618e-01 -3.32088433e-02 5.95440745e-01 -3.93622816e-01
1.50117064e+00 -3.93000126e-01 -1.13591862e+00 7.68375456e-01
7.24753737e-01 -1.82270050e-01 -1.28778148e+00 -1.95541367e-01
-3.32193196e-01 -4.19716924e-01 6.89504027e-01 -1.00815964e+00
-4.97906208e-01 8.62962067e-01 7.76714742e-01 3.26939553e-01
1.16477406e+00 -5.12396514e-01 -2.85722643e-01 5.79305403e-02
-1.46425486e-01 -1.12180102e+00 -6.62025034e-01 5.01949668e-01
7.72886097e-01 -1.02811944e+00 5.05761802e-01 -5.30871570e-01
-4.43099827e-01 9.43635583e-01 -1.23461522e-01 1.72603935e-01
1.13764477e+00 -4.71471325e-02 9.39294770e-02 -2.37286821e-01
3.25202830e-02 -1.66672230e-01 4.22539115e-01 9.09698904e-01
6.16395235e-01 8.40654433e-01 -3.37815553e-01 2.90944785e-01
1.41152531e-01 -1.55322537e-01 4.97535646e-01 6.41575992e-01
-8.92906249e-01 -6.59368515e-01 -1.05698133e+00 5.33809125e-01
-1.70535788e-01 -4.08598185e-01 -3.06187779e-01 8.57572705e-02
1.83461547e-01 1.08417726e+00 1.15435965e-01 2.77791619e-01
-4.03832406e-01 5.97117066e-01 8.94401670e-02 -8.88146400e-01
1.77569136e-01 7.46857762e-01 -1.03052892e-01 2.18895257e-01
-9.02949750e-01 -9.73839164e-01 -9.36130822e-01 -4.25173938e-01
-5.77466190e-01 2.97671318e-01 8.38095784e-01 7.63441026e-01
-8.78023356e-02 2.59283692e-01 4.40357536e-01 -1.04626524e+00
-5.28128326e-01 -1.39462066e+00 -1.59231770e+00 2.60757841e-02
2.67272592e-01 -3.42182368e-01 -7.82763124e-01 7.32675791e-01] | [9.98596477508545, -1.9871857166290283] |
f580251a-9e0a-4166-81f3-b816a9c73f8a | using-scene-graph-context-to-improve-image | 1901.03762 | null | http://arxiv.org/abs/1901.03762v2 | http://arxiv.org/pdf/1901.03762v2.pdf | Using Scene Graph Context to Improve Image Generation | Generating realistic images from scene graphs asks neural networks to be able
to reason about object relationships and compositionality. As a relatively new
task, how to properly ensure the generated images comply with scene graphs or
how to measure task performance remains an open question. In this paper, we
propose to harness scene graph context to improve image generation from scene
graphs. We introduce a scene graph context network that pools features
generated by a graph convolutional neural network that are then provided to
both the image generation network and the adversarial loss. With the context
network, our model is trained to not only generate realistic looking images,
but also to better preserve non-spatial object relationships. We also define
two novel evaluation metrics, the relation score and the mean opinion relation
score, for this task that directly evaluate scene graph compliance. We use both
quantitative and qualitative studies to demonstrate that our pro-posed model
outperforms the state-of-the-art on this challenging task. | ['Subarna Tripathi', 'Hanlin Tang', 'Anahita Bhiwandiwalla', 'Alexei Bastidas'] | 2019-01-11 | null | null | null | null | ['image-generation-from-scene-graphs'] | ['computer-vision'] | [ 7.07305610e-01 3.48456860e-01 3.26082468e-01 -4.64830279e-01
-6.31035388e-01 -5.88539064e-01 7.99153090e-01 -5.54917306e-02
-2.24638253e-01 5.00181913e-01 1.74751073e-01 -6.50798380e-02
1.62012309e-01 -1.13689053e+00 -1.13871920e+00 -2.74109185e-01
2.95592487e-01 2.36771643e-01 3.17227840e-01 -2.60083705e-01
8.51865932e-02 5.93781173e-01 -1.40705931e+00 6.02432311e-01
9.01777387e-01 1.05765629e+00 1.12791896e-01 7.35033751e-01
1.54814944e-01 1.37643337e+00 -5.76171219e-01 -8.15539122e-01
4.30511147e-01 -6.24279082e-01 -8.28583062e-01 1.85506999e-01
9.86596763e-01 -3.43911231e-01 -4.34753984e-01 1.22666109e+00
4.89591390e-01 2.68692255e-01 7.38653958e-01 -1.39930379e+00
-1.03268480e+00 5.72421789e-01 -2.05405161e-01 -8.96928832e-02
4.70273286e-01 6.42132938e-01 1.08436716e+00 -7.18454421e-01
9.58639562e-01 1.40470719e+00 4.52550232e-01 4.92943168e-01
-1.36498666e+00 -5.73697448e-01 2.78776139e-01 2.64545143e-01
-1.07392383e+00 -4.49033588e-01 1.05146194e+00 -5.58617115e-01
6.98734999e-01 2.47935817e-01 4.21863556e-01 1.24863601e+00
5.54643795e-02 6.57570601e-01 1.05106080e+00 -2.76356101e-01
1.84375688e-01 1.83700114e-01 -3.23027194e-01 6.69141948e-01
1.27632096e-01 2.27745667e-01 -4.27465528e-01 4.02125418e-01
7.06147194e-01 -3.00746262e-01 -4.42876905e-01 -5.53670526e-01
-1.11504793e+00 8.88226211e-01 9.74256337e-01 -4.32101302e-02
-4.82090920e-01 5.21756053e-01 -1.74599129e-03 9.70813483e-02
3.47009808e-01 8.99140239e-01 2.31763851e-02 3.25485528e-01
-7.85291851e-01 6.03332579e-01 7.12295234e-01 1.06027627e+00
7.57043660e-01 2.00198889e-01 -7.07155287e-01 5.76184213e-01
1.05155304e-01 5.26207149e-01 -1.17174149e-01 -1.03980112e+00
5.69920778e-01 7.69007087e-01 -1.09493107e-01 -1.33187521e+00
-1.17748097e-01 -4.48749065e-01 -8.57138097e-01 3.57873052e-01
4.94104058e-01 2.86309514e-02 -8.74521434e-01 1.99289691e+00
1.64359555e-01 8.92949328e-02 7.89199620e-02 1.10780871e+00
1.12761986e+00 5.92559218e-01 1.96228236e-01 2.92953461e-01
1.09008622e+00 -1.16891718e+00 -5.43757975e-01 -5.20342588e-01
2.22957030e-01 -7.93762445e-01 1.28663170e+00 3.99239846e-02
-1.44443285e+00 -7.46062100e-01 -8.69453907e-01 -2.77772039e-01
-4.30336237e-01 -3.00662611e-02 4.53843474e-01 2.73036748e-01
-1.14442003e+00 6.45268738e-01 -4.26891804e-01 -2.52772301e-01
6.94710493e-01 5.31358831e-02 -4.32460040e-01 -3.04038703e-01
-1.21673524e+00 8.69535863e-01 3.43723297e-01 -1.04731679e-01
-1.00501502e+00 -6.54283226e-01 -1.13810897e+00 1.84144586e-01
3.19755673e-01 -1.08890676e+00 1.00545275e+00 -1.22347426e+00
-1.25498605e+00 9.13404405e-01 2.38503605e-01 -4.57409471e-01
7.08037734e-01 7.10125342e-02 -1.03922851e-01 2.68735170e-01
2.15704486e-01 1.03219295e+00 7.61794984e-01 -1.66014242e+00
-3.52906972e-01 -9.05120447e-02 5.39249837e-01 2.71854877e-01
-5.44913076e-02 -2.10881963e-01 -5.78627408e-01 -8.35687757e-01
-1.94872484e-01 -9.57164168e-01 -2.37930819e-01 1.71526119e-01
-7.67540872e-01 1.67500943e-01 6.10841572e-01 -6.98392034e-01
6.49954855e-01 -2.00954151e+00 1.21587984e-01 7.96928108e-02
1.23959847e-01 1.57226145e-01 -6.40605509e-01 3.49004447e-01
-2.71637976e-01 1.96803197e-01 -4.39842135e-01 -4.17213291e-01
4.66085300e-02 -1.02424175e-02 -6.08110487e-01 1.36209950e-01
7.31487334e-01 1.38102949e+00 -1.03097999e+00 -3.20397407e-01
4.49737251e-01 5.68524599e-01 -7.67769575e-01 5.93186378e-01
-5.58391154e-01 4.72330451e-01 -2.26048589e-01 2.71188855e-01
8.27602804e-01 -3.81488144e-01 -1.07139640e-01 -6.05772972e-01
3.78010929e-01 -2.95064915e-02 -9.59863245e-01 1.68129683e+00
-4.65902179e-01 5.78028321e-01 -3.35282892e-01 -8.44015062e-01
7.80755460e-01 -1.48466051e-01 1.21935256e-01 -1.14685464e+00
5.73009066e-02 -2.44463742e-01 -1.74958751e-01 -5.10552168e-01
5.42892516e-01 4.20786738e-02 -8.60513598e-02 2.41408110e-01
-1.37163121e-02 -6.18443012e-01 1.50002524e-01 3.99907291e-01
1.16378379e+00 1.75132245e-01 -1.27560087e-02 1.61049596e-03
4.50132161e-01 -3.69088985e-02 1.90051511e-01 8.26702833e-01
2.21654046e-02 1.13948917e+00 7.19548225e-01 -3.79137009e-01
-1.01613069e+00 -1.40801787e+00 4.39388603e-01 8.40428233e-01
5.13195992e-01 -6.62227049e-02 -8.05572510e-01 -8.79354417e-01
-1.35789976e-01 9.88661230e-01 -7.89502323e-01 -4.30590898e-01
-5.09694993e-01 -2.36968622e-01 3.39905381e-01 4.56339449e-01
7.37735331e-01 -1.40346694e+00 -4.00433928e-01 -4.50994149e-02
-3.73044372e-01 -1.62600338e+00 -6.14939868e-01 -3.23853463e-01
-2.69371241e-01 -1.19244397e+00 -4.61624622e-01 -7.79196560e-01
8.55263054e-01 6.43125698e-02 1.65572929e+00 9.60239545e-02
-1.61713704e-01 2.89688408e-01 -3.73847067e-01 -2.84054160e-01
-6.00625515e-01 3.96238863e-02 -5.91090858e-01 2.31539100e-01
-1.86456233e-01 -6.07315004e-01 -8.40679467e-01 3.08403522e-01
-1.29366636e+00 5.77318192e-01 6.59725547e-01 6.70793831e-01
6.00956798e-01 -3.05575430e-02 2.51677543e-01 -1.15880835e+00
5.66724718e-01 -9.34274718e-02 -6.19709969e-01 2.86593139e-01
-3.16132754e-01 1.98711917e-01 7.62053132e-01 -3.47570211e-01
-9.67679262e-01 2.54636168e-01 -6.32627755e-02 -5.33956468e-01
-3.16026285e-02 1.90546498e-01 -6.15515709e-01 -6.94268942e-02
8.78034890e-01 1.15899697e-01 -2.51475543e-01 1.47651032e-01
8.94256592e-01 -1.53816701e-03 8.18656623e-01 -5.37116528e-01
1.23663449e+00 5.97000957e-01 1.01699159e-01 -4.63821262e-01
-1.01999044e+00 -4.72450294e-02 -3.76837701e-01 -3.21102172e-01
1.19073129e+00 -8.57688487e-01 -4.31808561e-01 2.44890139e-01
-1.48882425e+00 -7.42813587e-01 -4.84438539e-01 -5.48530445e-02
-8.30152929e-01 9.41579193e-02 -2.86935717e-01 -5.61177850e-01
-3.22264880e-01 -1.15007663e+00 1.38405263e+00 6.30761608e-02
-4.27712090e-02 -9.48015571e-01 -1.64954022e-01 4.19243187e-01
5.39362192e-01 9.13957775e-01 8.19633663e-01 -1.96567550e-01
-1.12812173e+00 -1.70416702e-02 -8.08904767e-01 4.94204491e-01
3.34415473e-02 -1.40919447e-01 -1.07585514e+00 -1.43402562e-01
-2.06883758e-01 -4.44043189e-01 9.09006238e-01 2.02406496e-01
1.36205018e+00 -6.38888240e-01 8.77873674e-02 8.76817584e-01
1.45042241e+00 -1.41773373e-01 1.11880124e+00 1.04621939e-01
1.09034991e+00 9.41292286e-01 3.19022775e-01 -1.45262862e-02
4.92662519e-01 8.07267964e-01 7.86569953e-01 -4.61098373e-01
-7.60025501e-01 -5.65312207e-01 1.53560549e-01 1.66569054e-01
2.60930866e-01 -7.06961274e-01 -7.25000262e-01 5.11203587e-01
-1.89595973e+00 -1.18240035e+00 -1.27968296e-01 1.81659889e+00
7.81388283e-01 1.49912626e-01 -1.84951216e-01 -1.52982697e-01
6.00500166e-01 3.26825559e-01 -3.93100768e-01 -2.99481928e-01
-3.33268613e-01 2.05564618e-01 3.67329866e-01 5.92113793e-01
-1.12831199e+00 1.19127071e+00 5.37278271e+00 6.29942656e-01
-1.16299665e+00 -1.25395879e-01 8.89511287e-01 1.49235860e-01
-4.81804341e-01 6.81308806e-02 -2.19858825e-01 3.52717012e-01
3.78195912e-01 -1.47916004e-02 5.70045531e-01 8.45189214e-01
1.77416280e-01 1.07268356e-01 -1.31000197e+00 9.52893376e-01
4.02941674e-01 -1.48675859e+00 5.83451807e-01 -5.38789146e-02
1.10376740e+00 -2.44478106e-01 3.85904431e-01 1.75108686e-01
5.19740343e-01 -1.41466284e+00 1.09144187e+00 7.56770909e-01
8.95366609e-01 -4.61109519e-01 6.05271935e-01 5.29943127e-03
-1.14800870e+00 3.17006469e-01 -1.46879375e-01 1.83706731e-01
2.23560289e-01 6.37617886e-01 -8.99166822e-01 6.07450426e-01
4.41629320e-01 6.14176750e-01 -9.33819294e-01 8.18728030e-01
-6.43490970e-01 1.84673101e-01 6.05476759e-02 2.23295510e-01
2.24539295e-01 5.89744374e-02 3.56582433e-01 1.08377469e+00
1.31515846e-01 -1.12147696e-01 1.47580981e-01 1.47281146e+00
-3.66622180e-01 1.72069091e-02 -1.00038207e+00 2.11103894e-02
8.51988271e-02 1.29508889e+00 -5.45414746e-01 -1.74784660e-01
-1.82755917e-01 1.19478238e+00 5.51120341e-01 6.07078969e-01
-8.92021656e-01 -2.96221584e-01 5.47425210e-01 4.41627085e-01
2.06937134e-01 -9.22602192e-02 -3.67899746e-01 -1.05848861e+00
1.56763166e-01 -9.16572094e-01 1.72476530e-01 -1.26098120e+00
-1.52774692e+00 8.08920681e-01 -2.02082805e-02 -1.21342862e+00
-3.61176282e-01 -5.97101450e-01 -7.32286155e-01 8.32567275e-01
-1.71098733e+00 -1.57997251e+00 -9.28745687e-01 6.01830125e-01
3.06684196e-01 4.75270487e-02 5.31939387e-01 1.06773943e-01
-2.35474601e-01 6.40646577e-01 -5.91994762e-01 4.17987913e-01
5.82809210e-01 -1.31576049e+00 7.69483864e-01 1.15490675e+00
2.04887509e-01 3.02117556e-01 6.73920751e-01 -5.55997431e-01
-1.22326529e+00 -1.86705375e+00 5.77500939e-01 -7.52939820e-01
2.59382248e-01 -5.75262368e-01 -7.80648589e-01 4.43540484e-01
2.14729697e-01 4.04772103e-01 2.03223214e-01 -5.06263137e-01
-6.64122641e-01 -2.52708972e-01 -1.06726956e+00 9.36229885e-01
1.39079154e+00 -5.51045418e-01 -2.23473847e-01 3.20657223e-01
9.44141805e-01 -4.46850955e-01 -5.79767346e-01 6.55526817e-01
3.48467529e-01 -1.20361161e+00 1.16392815e+00 -5.36596596e-01
1.02281880e+00 -3.65749598e-01 -3.32073838e-01 -1.45791948e+00
-2.68768847e-01 -4.50372934e-01 2.94888556e-01 1.25165141e+00
6.01026118e-01 -3.88452232e-01 7.60016322e-01 6.24161720e-01
-6.86043799e-02 -5.06243467e-01 -5.07611692e-01 -6.89536452e-01
2.25043595e-02 -3.46134394e-01 6.88915670e-01 8.56574416e-01
-6.99553490e-01 6.75908327e-01 -2.26732016e-01 1.40851095e-01
6.35397613e-01 2.66826868e-01 1.13716602e+00 -7.59719372e-01
-3.15793991e-01 -5.44168890e-01 -5.25731862e-01 -7.58561075e-01
2.96798617e-01 -1.03238904e+00 1.78316116e-01 -1.87822902e+00
1.02191821e-01 -2.85555363e-01 -2.91651413e-02 3.64573091e-01
-3.21585953e-01 5.60256481e-01 4.38392669e-01 -2.74174273e-01
-7.24861205e-01 8.11634064e-01 1.65912092e+00 -4.06031013e-01
9.76911113e-02 -3.49364132e-01 -9.15525794e-01 4.32152569e-01
5.96349418e-01 -2.43838325e-01 -8.35712850e-01 -7.71397054e-01
3.97390813e-01 -1.58374205e-01 8.65642726e-01 -1.15657508e+00
-2.92751994e-02 -3.27908546e-01 5.28786302e-01 -2.58316517e-01
3.36541444e-01 -7.73688614e-01 2.73254186e-01 1.74903467e-01
-5.51954150e-01 2.12608539e-02 1.23460524e-01 4.85669851e-01
-1.37016550e-01 2.60435373e-01 1.05453479e+00 -1.35743469e-01
-5.85229576e-01 5.68951547e-01 2.82601774e-01 4.11321431e-01
1.12946296e+00 -2.92407610e-02 -5.91376543e-01 -7.79835641e-01
-4.98755604e-01 1.51965693e-01 7.88333833e-01 5.73012710e-01
9.68070924e-01 -1.59225655e+00 -9.04805422e-01 1.07300177e-01
4.83122110e-01 1.65834799e-01 2.10543692e-01 2.65002519e-01
-7.42484748e-01 2.52054632e-02 -1.97966412e-01 -5.56317687e-01
-9.76885915e-01 7.30220139e-01 4.75272804e-01 -3.68431509e-01
-3.96363914e-01 8.56439412e-01 8.51163030e-01 -5.88278174e-01
1.47838175e-01 -5.26042402e-01 -9.89067629e-02 -3.17453206e-01
3.01804274e-01 -1.74404100e-01 3.52599472e-02 -7.20693946e-01
-1.98186263e-01 5.45419693e-01 3.00849169e-01 -1.34906322e-01
1.14563310e+00 9.74797457e-02 1.16664907e-02 -1.32387370e-01
1.31880045e+00 -2.30667502e-01 -1.47609079e+00 -1.90651938e-01
-3.76704842e-01 -6.02779806e-01 -1.48983777e-01 -1.00167286e+00
-1.30906272e+00 8.06054533e-01 5.20340562e-01 2.88045526e-01
1.22532785e+00 7.64718577e-02 6.83560491e-01 1.57569706e-01
7.84277767e-02 -6.67406678e-01 7.27297187e-01 3.98601443e-01
1.49439704e+00 -1.38967884e+00 -1.60432443e-01 -7.24446237e-01
-8.20758879e-01 7.20348775e-01 9.63371277e-01 -3.08847904e-01
3.57543170e-01 5.24376184e-02 7.49504194e-03 -3.29119653e-01
-5.38062871e-01 -4.77046847e-01 6.71435118e-01 8.09455276e-01
1.87223881e-01 -3.72894816e-02 3.35391074e-01 3.38440806e-01
-6.66664064e-01 -2.35753238e-01 3.25579613e-01 4.83558744e-01
1.89946685e-03 -8.48321617e-01 -1.11260325e-01 3.42167646e-01
-2.15341002e-01 -1.59443364e-01 -8.17790151e-01 5.46093225e-01
2.05971256e-01 1.10825706e+00 8.32928866e-02 -4.41794127e-01
7.99004555e-01 -4.81124550e-01 5.19157887e-01 -7.03662694e-01
-4.02037203e-01 -5.05457520e-01 1.37020081e-01 -8.68667126e-01
-3.00698698e-01 -1.63657814e-01 -8.88516188e-01 -1.64193645e-01
-1.81982040e-01 -2.70862192e-01 5.50954998e-01 5.91497064e-01
3.01414251e-01 9.27773595e-01 7.20964134e-01 -8.86840522e-01
-1.79991290e-01 -6.66630387e-01 -9.70685408e-02 1.11881876e+00
2.33986557e-01 -3.78841162e-01 -1.64962500e-01 3.24929386e-01] | [11.395779609680176, -0.27312785387039185] |
66dd9ba1-376d-47c1-b4be-438f2e85d1dd | videotrack-learning-to-track-objects-via | null | null | http://openaccess.thecvf.com//content/CVPR2023/html/Xie_VideoTrack_Learning_To_Track_Objects_via_Video_Transformer_CVPR_2023_paper.html | http://openaccess.thecvf.com//content/CVPR2023/papers/Xie_VideoTrack_Learning_To_Track_Objects_via_Video_Transformer_CVPR_2023_paper.pdf | VideoTrack: Learning To Track Objects via Video Transformer | Existing Siamese tracking methods, which are built on pair-wise matching between two single frames, heavily rely on additional sophisticated mechanism to exploit temporal information among successive video frames, hindering them from high efficiency and industrial deployments. In this work, we resort to sequence-level target matching that can encode temporal contexts into the spatial features through a neat feedforward video model. Specifically, we adapt the standard video transformer architecture to visual tracking by enabling spatiotemporal feature learning directly from frame-level patch sequences. To better adapt to the tracking task, we carefully blend the spatiotemporal information in the video clips through sequential multi-branch triplet blocks, which formulates a video transformer backbone. Our experimental study compares different model variants, such as tokenization strategies, hierarchical structures, and video attention schemes. Then, we propose a disentangled dual-template mechanism that decouples static and dynamic appearance changes over time, and reduces the temporal redundancy in video frames. Extensive experiments show that our method, named as VideoTrack, achieves state-of-the-art results while running in real-time. | ['Chao Ma', 'Yan Lu', 'Jiahao Li', 'Lei Chu', 'Fei Xie'] | 2023-01-01 | null | null | null | cvpr-2023-1 | ['visual-tracking'] | ['computer-vision'] | [ 2.62453109e-01 -6.32385612e-01 -6.54401839e-01 -1.09506413e-01
-6.10755086e-01 -7.34400153e-01 6.88537657e-01 -4.89396065e-01
-3.05676341e-01 3.79008293e-01 3.07038188e-01 -1.19819835e-01
-7.32261464e-02 -3.19864333e-01 -8.93266201e-01 -8.04188490e-01
-9.71147511e-03 -2.35666275e-01 4.79162723e-01 -3.80131393e-03
1.38929173e-01 1.48527518e-01 -1.41240525e+00 2.66912997e-01
5.82994223e-01 1.15907240e+00 2.02547073e-01 4.70650226e-01
1.02008641e-01 9.71289098e-01 -1.02674931e-01 -5.14104009e-01
3.64316523e-01 -3.61710101e-01 -2.51534432e-01 1.85774937e-01
8.11486959e-01 -3.72772276e-01 -1.02466404e+00 9.24916863e-01
1.12374924e-01 1.30693555e-01 2.06038401e-01 -1.41275299e+00
-8.17992866e-01 3.18087071e-01 -7.29919732e-01 4.45717216e-01
3.71467799e-01 4.71860617e-01 1.17955184e+00 -8.72728765e-01
5.71607113e-01 1.17979228e+00 6.66334510e-01 4.99307603e-01
-1.35778821e+00 -8.73142302e-01 8.45713198e-01 6.09479487e-01
-1.05831993e+00 -6.44184053e-01 1.01807845e+00 -5.99167109e-01
6.60675764e-01 2.09465221e-01 9.21615720e-01 1.33375680e+00
6.56741634e-02 1.16875851e+00 7.78354228e-01 -6.75332695e-02
-2.16042057e-01 -5.96885681e-01 -1.40001580e-01 8.52226973e-01
1.73221320e-01 5.53904414e-01 -8.87439191e-01 1.30879685e-01
9.00256217e-01 4.59905595e-01 -5.62087893e-01 -8.28130782e-01
-1.66290760e+00 5.65456033e-01 3.67402971e-01 2.85906255e-01
-1.81141496e-01 3.20919454e-01 4.22420293e-01 1.72664285e-01
2.67574549e-01 -1.09661045e-02 -3.46069485e-01 -1.67665973e-01
-1.05350864e+00 1.57094970e-01 2.49656394e-01 1.11626351e+00
7.00873911e-01 9.42831710e-02 -5.99974394e-01 2.24885762e-01
3.93911868e-01 5.75078547e-01 4.24866349e-01 -1.17817140e+00
5.36010742e-01 4.43739921e-01 1.19556777e-01 -1.26451182e+00
3.01056616e-02 -3.35005224e-01 -6.08929574e-01 -1.92109704e-01
3.95655602e-01 9.50751677e-02 -8.41695547e-01 2.08779407e+00
4.12848294e-01 8.73379648e-01 -2.60472506e-01 1.04578304e+00
4.91029054e-01 3.89140099e-01 6.33468404e-02 -2.90475279e-01
1.28663123e+00 -1.38658607e+00 -6.09266162e-01 2.75627468e-02
2.33706772e-01 -5.32842696e-01 9.70731080e-01 1.50315955e-01
-1.09979141e+00 -6.52561307e-01 -9.48028028e-01 -1.45475268e-01
-5.27635179e-02 5.98131903e-02 4.80622292e-01 3.47715557e-01
-7.44095862e-01 5.98593056e-01 -1.26740026e+00 -9.55006480e-02
3.98313582e-01 3.10841382e-01 -4.51470017e-01 -9.72365066e-02
-1.08214486e+00 4.21126932e-01 1.27586216e-01 3.34000707e-01
-9.79813695e-01 -9.38008070e-01 -1.09589958e+00 1.19679183e-01
7.26409078e-01 -9.10911083e-01 1.15647912e+00 -1.07839465e+00
-1.57603681e+00 6.36494517e-01 -4.52676982e-01 -2.66365647e-01
5.80732107e-01 -3.70529324e-01 -2.91027278e-01 2.44419649e-01
1.15085274e-01 5.31639338e-01 1.22686481e+00 -1.07312334e+00
-8.64638627e-01 -2.59590387e-01 2.27204517e-01 7.03376383e-02
-5.00516772e-01 5.32725900e-02 -9.60444748e-01 -1.06817043e+00
-1.35743216e-01 -9.35797930e-01 -9.79747251e-02 3.23062390e-01
-2.03968361e-01 -9.75754261e-02 1.00294352e+00 -4.59337622e-01
1.40130591e+00 -2.29029131e+00 5.67407310e-01 -5.66230454e-02
5.62796891e-01 2.25269109e-01 -3.03996414e-01 7.76138306e-02
6.32135272e-02 -2.34109700e-01 8.08541626e-02 -4.33381051e-01
1.47019163e-01 1.30361199e-01 -2.95201212e-01 5.79702079e-01
2.58759916e-01 1.07970059e+00 -1.21551871e+00 -6.84934974e-01
2.01526806e-01 5.18605471e-01 -7.13987648e-01 2.98213601e-01
-2.33643398e-01 6.54427230e-01 -5.82595348e-01 7.04140544e-01
4.30572152e-01 -4.37678874e-01 1.97838783e-01 -5.63473284e-01
-2.78141350e-01 1.07281029e-01 -6.68043077e-01 2.17677355e+00
-2.40966618e-01 6.62611783e-01 1.06156312e-01 -1.03070343e+00
4.53060389e-01 2.74491966e-01 7.67375469e-01 -7.98603535e-01
-4.67095114e-02 -2.71751136e-02 -6.44066632e-02 -4.93142128e-01
3.82467419e-01 1.20384820e-01 2.81822239e-03 1.46628022e-01
1.06569856e-01 6.50523782e-01 4.28073436e-01 1.52250573e-01
1.12357056e+00 8.65576446e-01 1.34598643e-01 2.55407523e-02
5.99009454e-01 -2.41592407e-01 1.02768195e+00 5.55077612e-01
-4.76531535e-01 5.87590098e-01 5.16795456e-01 -4.68254328e-01
-9.26252127e-01 -1.13002992e+00 2.96242446e-01 1.19295287e+00
4.64276493e-01 -5.85323215e-01 -4.24989492e-01 -6.66322231e-01
-3.55603062e-02 -3.10145058e-02 -7.77009845e-01 -1.89285740e-01
-9.20954823e-01 -2.85179406e-01 2.62445867e-01 6.59253836e-01
3.74226809e-01 -7.54891992e-01 -7.67838240e-01 3.03503633e-01
-2.13606358e-01 -1.32488203e+00 -1.19970691e+00 -1.58373728e-01
-6.96033061e-01 -1.18722916e+00 -8.98661852e-01 -6.96138442e-01
1.64311767e-01 7.82102346e-01 1.02526963e+00 4.92155962e-02
-1.50083184e-01 4.78407323e-01 -4.45025712e-01 3.45755905e-01
2.95011282e-01 2.03766245e-02 4.49116416e-02 3.72174025e-01
1.67077646e-01 -7.54157484e-01 -9.32096124e-01 5.01316428e-01
-7.83879578e-01 2.39292532e-01 6.40605390e-01 1.09713769e+00
6.43062055e-01 -5.41808069e-01 7.31458589e-02 -5.14590383e-01
-7.96465725e-02 -4.37655658e-01 -7.82086551e-01 5.99352717e-01
-2.78932542e-01 1.87309682e-01 7.01259732e-01 -7.22666562e-01
-8.94115210e-01 2.22545207e-01 3.11793745e-01 -1.39610147e+00
1.29285499e-01 2.67304271e-01 -3.48938555e-01 -2.29351118e-01
3.81149240e-02 5.41986823e-01 -5.00573069e-02 -3.30378413e-01
4.40513194e-01 3.23388427e-02 7.71785200e-01 -7.31835186e-01
1.02451861e+00 5.70650458e-01 -1.27214164e-01 -3.14690769e-01
-9.77870524e-01 -4.29330468e-01 -6.29603684e-01 -2.90572017e-01
8.74706566e-01 -1.04646862e+00 -1.00162292e+00 3.51703882e-01
-1.02813387e+00 -2.55274594e-01 -1.71180386e-02 4.97198522e-01
-6.73324466e-01 5.60970187e-01 -6.30337775e-01 -3.88618320e-01
1.47935618e-02 -1.28252637e+00 1.33452046e+00 2.32395649e-01
1.67757675e-01 -8.02023649e-01 1.45531774e-01 2.27026671e-01
2.84385622e-01 2.43443742e-01 5.89089513e-01 -2.13161722e-01
-1.16917634e+00 1.63545683e-01 -3.27456295e-01 -5.38025349e-02
1.92489877e-01 6.23536073e-02 -6.49661720e-01 -4.26034957e-01
-1.83588848e-01 -1.05223589e-01 1.00554979e+00 2.81861067e-01
1.22274876e+00 -4.18669105e-01 -5.07864237e-01 1.00095069e+00
1.14617527e+00 1.22064434e-01 3.27910960e-01 3.18192422e-01
1.10629559e+00 3.99969697e-01 5.76967895e-01 3.16741228e-01
4.22994345e-01 1.03709233e+00 4.18812722e-01 5.24048172e-02
-1.34033009e-01 -5.30646741e-01 6.93002343e-01 9.07195330e-01
-1.65943116e-01 -1.07925266e-01 -5.01520038e-01 4.34838474e-01
-2.20303965e+00 -1.45273423e+00 2.85380691e-01 2.10648417e+00
5.56189775e-01 6.93367943e-02 3.75504911e-01 -1.92012012e-01
7.78918207e-01 6.28483832e-01 -8.48166406e-01 5.47795892e-01
-2.50134729e-02 -1.53822392e-01 3.81041408e-01 2.15070054e-01
-1.35640323e+00 8.52870822e-01 5.85571527e+00 7.80762553e-01
-1.19541991e+00 1.91371620e-01 2.66562462e-01 -4.68840271e-01
-3.66995364e-01 6.81455657e-02 -6.13144636e-01 7.62465596e-01
6.29641235e-01 -2.19305068e-01 4.59592551e-01 5.60604811e-01
2.13793874e-01 4.73447829e-01 -1.30837643e+00 1.11234462e+00
5.12904413e-02 -1.52441275e+00 9.93384272e-02 3.71229984e-02
6.05575025e-01 -1.33589730e-01 3.55362386e-01 2.29365170e-01
1.17728047e-01 -6.40322804e-01 1.16719449e+00 5.26168048e-01
6.58347249e-01 -3.43156397e-01 1.10317856e-01 -2.63516065e-02
-1.79104185e+00 -1.65406957e-01 7.89870173e-02 1.26959935e-01
2.23490745e-01 2.50766445e-02 2.19117701e-01 7.36461282e-01
8.04206312e-01 1.48300600e+00 -4.27922696e-01 1.17424285e+00
1.69197714e-03 4.58064526e-01 -1.87749267e-01 2.47281075e-01
3.50625455e-01 -1.72143251e-01 6.46891534e-01 1.00863469e+00
2.17169896e-01 1.03849269e-01 4.36943859e-01 7.08803058e-01
8.59035924e-02 -2.72843599e-01 -4.66687173e-01 9.61014256e-03
4.12521362e-01 1.10215080e+00 -4.90464807e-01 -1.90698907e-01
-1.04146755e+00 9.59605753e-01 3.09525281e-01 5.14504611e-01
-1.17736173e+00 -6.20177686e-02 1.02156842e+00 -1.33045033e-01
7.34331667e-01 -3.20326567e-01 2.57047087e-01 -1.78648829e+00
3.74579459e-01 -9.41570520e-01 3.72865945e-01 -4.77186233e-01
-1.34091353e+00 5.21867514e-01 -7.61567429e-02 -1.72725356e+00
-1.82088211e-01 -6.27891898e-01 -6.55237675e-01 2.91331649e-01
-1.67561769e+00 -1.43997490e+00 -3.14424515e-01 9.88776743e-01
5.62961817e-01 2.67901216e-02 3.54597151e-01 6.21477604e-01
-8.05891454e-01 8.74386251e-01 1.27182290e-01 2.82014251e-01
6.77065134e-01 -9.38129485e-01 3.42212290e-01 1.08744764e+00
4.23159450e-01 6.88140571e-01 4.93145108e-01 -3.60184282e-01
-1.81576908e+00 -1.02694070e+00 5.60454905e-01 -3.10970277e-01
1.02631092e+00 -4.41240251e-01 -8.39393616e-01 7.17073500e-01
2.61986285e-01 3.44848573e-01 5.89160442e-01 -1.15145580e-03
-7.21908331e-01 -2.10594386e-01 -5.32673895e-01 8.06326985e-01
1.50792360e+00 -5.99100232e-01 -4.67545182e-01 7.12232990e-03
9.88198757e-01 -4.58146781e-01 -7.91968584e-01 4.25650090e-01
9.62307870e-01 -9.38652754e-01 1.08041739e+00 -8.50057960e-01
4.16197449e-01 -6.17197871e-01 -2.02116460e-01 -7.73869097e-01
-5.26720822e-01 -9.57882226e-01 -6.52160645e-01 1.12599552e+00
1.25633314e-01 -2.66511828e-01 9.27906275e-01 5.51203430e-01
-1.39845818e-01 -8.24853003e-01 -1.01267242e+00 -8.43222857e-01
-2.39038393e-01 -2.35044762e-01 5.27909517e-01 9.09659922e-01
-8.18093270e-02 2.27201357e-01 -8.21019769e-01 4.30096537e-02
7.23442912e-01 5.90413809e-01 6.51229560e-01 -9.56046939e-01
-5.69160402e-01 -6.87234402e-01 -5.39802074e-01 -1.60654926e+00
3.80808145e-01 -4.92124230e-01 6.35759309e-02 -7.91253030e-01
4.26434457e-01 -1.79590210e-01 -6.07902110e-01 4.66993749e-01
-3.93326014e-01 3.08594674e-01 5.95793962e-01 4.52082813e-01
-1.04154932e+00 9.68385339e-01 1.33833539e+00 -3.23444784e-01
1.10480733e-01 -2.44521260e-01 -4.73197520e-01 4.99164253e-01
2.54764497e-01 -3.47263873e-01 -3.90916467e-01 -7.99273729e-01
-1.13395922e-01 3.86885315e-01 5.46203852e-01 -9.00951505e-01
4.01735038e-01 -4.03053850e-01 3.09725791e-01 -4.32764620e-01
4.84671801e-01 -9.48562503e-01 1.55517966e-01 3.48089844e-01
-4.21773642e-01 3.88299227e-01 -1.09173544e-01 1.02518320e+00
-3.96624237e-01 3.46251994e-01 5.27139366e-01 8.90064910e-02
-8.94528151e-01 9.51136827e-01 -1.39998659e-01 4.48486470e-02
1.09486413e+00 -3.23265702e-01 -2.14450255e-01 -1.47138119e-01
-7.55708754e-01 5.22504926e-01 7.01018453e-01 6.83161676e-01
4.33687627e-01 -1.64872944e+00 -3.13608974e-01 1.97265208e-01
1.23260200e-01 -3.31541121e-01 4.42515910e-01 1.18560851e+00
6.03668764e-03 5.37052095e-01 -2.34464929e-01 -7.97841728e-01
-1.17177594e+00 9.86019075e-01 2.91571617e-01 -2.57555276e-01
-8.15622568e-01 6.33365512e-01 6.25573874e-01 1.63955793e-01
3.75219673e-01 -2.63287574e-01 2.74641253e-03 -9.52425376e-02
5.46213269e-01 3.21700014e-02 -3.54270220e-01 -8.52318108e-01
-3.99446607e-01 8.94194663e-01 -1.30320325e-01 -3.29015478e-02
1.08452022e+00 -2.81510770e-01 2.36506745e-01 3.24226350e-01
1.27379429e+00 -1.05743602e-01 -1.98615050e+00 -5.69337010e-01
-5.53226285e-02 -9.39504206e-01 -1.46585345e-01 -1.86534420e-01
-1.52288544e+00 8.51567328e-01 3.80938977e-01 1.45981293e-02
1.25551689e+00 -1.43428117e-01 9.70519125e-01 2.01148063e-01
3.68285328e-01 -5.21564841e-01 3.07376653e-01 3.29911619e-01
4.34864879e-01 -1.13303924e+00 -2.62792468e-01 -3.09941173e-01
-4.02323127e-01 9.68736589e-01 7.92592645e-01 -2.45519564e-01
5.82207501e-01 9.09197628e-02 -2.39831328e-01 4.07120399e-02
-1.03975999e+00 -2.65014350e-01 5.02754927e-01 5.51078737e-01
3.34761441e-01 -2.75359541e-01 -1.62640750e-01 5.71135402e-01
4.27075624e-01 5.06159402e-02 1.26811843e-02 8.77968490e-01
-7.94284791e-02 -1.23330629e+00 -1.59184381e-01 1.58175573e-01
-3.66901755e-01 -1.22225761e-01 -7.93353543e-02 8.39905441e-01
1.26115412e-01 6.46971405e-01 9.27014798e-02 -5.23250580e-01
2.84071326e-01 -2.05812320e-01 7.07399368e-01 -1.82317525e-01
-4.18886155e-01 3.36469442e-01 -3.33525419e-01 -1.05221581e+00
-9.96802628e-01 -9.93357062e-01 -7.61799216e-01 -3.69117498e-01
-1.44980028e-01 7.68356249e-02 -2.24365573e-02 9.48647738e-01
5.73982418e-01 4.33664709e-01 6.24745548e-01 -1.26526392e+00
-3.90093416e-01 -5.48795104e-01 -2.15310276e-01 4.15682852e-01
7.47132599e-01 -1.11126268e+00 -8.55623782e-02 3.60338271e-01] | [8.907095909118652, 0.31437185406684875] |
918a949c-4c16-4e25-b3b8-94728ef1b44a | learning-semantics-enriched-representation | 2007.06959 | null | https://arxiv.org/abs/2007.06959v1 | https://arxiv.org/pdf/2007.06959v1.pdf | Learning Semantics-enriched Representation via Self-discovery, Self-classification, and Self-restoration | Medical images are naturally associated with rich semantics about the human anatomy, reflected in an abundance of recurring anatomical patterns, offering unique potential to foster deep semantic representation learning and yield semantically more powerful models for different medical applications. But how exactly such strong yet free semantics embedded in medical images can be harnessed for self-supervised learning remains largely unexplored. To this end, we train deep models to learn semantically enriched visual representation by self-discovery, self-classification, and self-restoration of the anatomy underneath medical images, resulting in a semantics-enriched, general-purpose, pre-trained 3D model, named Semantic Genesis. We examine our Semantic Genesis with all the publicly-available pre-trained models, by either self-supervision or fully supervision, on the six distinct target tasks, covering both classification and segmentation in various medical modalities (i.e.,CT, MRI, and X-ray). Our extensive experiments demonstrate that Semantic Genesis significantly exceeds all of its 3D counterparts as well as the de facto ImageNet-based transfer learning in 2D. This performance is attributed to our novel self-supervised learning framework, encouraging deep models to learn compelling semantic representation from abundant anatomical patterns resulting from consistent anatomies embedded in medical images. Code and pre-trained Semantic Genesis are available at https://github.com/JLiangLab/SemanticGenesis . | ['Zongwei Zhou', 'Michael B. Gotway', 'Jianming Liang', 'Fatemeh Haghighi', 'Mohammad Reza Hosseinzadeh Taher'] | 2020-07-14 | null | null | null | null | ['lung-nodule-detection', 'lung-nodule-segmentation', 'liver-segmentation'] | ['medical', 'medical', 'medical'] | [ 3.35746557e-01 6.83367848e-01 -3.83067489e-01 -6.38563335e-01
-7.26467609e-01 -3.82392853e-01 4.40127075e-01 2.94933885e-01
-1.72782183e-01 6.26737714e-01 6.16798937e-01 2.13999432e-02
-1.73248470e-01 -6.11700058e-01 -7.51159072e-01 -5.72388113e-01
-2.35795975e-01 4.52383101e-01 -1.31323477e-02 -2.43373826e-01
-2.82727480e-01 2.36055583e-01 -1.38076532e+00 3.48355591e-01
8.62819672e-01 9.89926159e-01 4.40517128e-01 3.87161642e-01
-2.06141010e-01 7.23729670e-01 -1.48826733e-01 -3.30575198e-01
1.09208740e-01 -5.36478162e-01 -1.13942885e+00 1.65673733e-01
3.63761991e-01 -4.31350060e-02 -4.37543184e-01 1.17898500e+00
4.31295961e-01 -2.02667400e-01 5.87811291e-01 -9.12049949e-01
-9.07052934e-01 4.63174343e-01 -3.95492166e-01 3.22615832e-01
1.52276382e-01 2.41180182e-01 9.71024394e-01 -4.85019684e-01
1.04357147e+00 9.11993146e-01 6.63051188e-01 9.31210339e-01
-1.18646431e+00 -4.44715619e-01 -2.66262472e-01 -1.62818402e-01
-1.09243703e+00 -3.20608497e-01 8.59780729e-01 -4.19234097e-01
3.97571623e-01 1.95287496e-01 8.56607795e-01 1.42507207e+00
2.69500434e-01 8.62191737e-01 1.42389691e+00 -2.25709066e-01
2.08460018e-01 1.03122681e-01 -6.17454108e-03 1.19308162e+00
1.85490921e-01 1.25198096e-01 -3.39994371e-01 -1.12179682e-01
1.00002110e+00 3.01486403e-01 -3.75220746e-01 -6.33053780e-01
-1.38419735e+00 6.07549727e-01 1.17589331e+00 4.72210974e-01
-3.78518313e-01 2.64943987e-01 4.64485794e-01 -5.48152626e-02
4.76788133e-01 5.81281602e-01 -4.22037631e-01 3.34254414e-01
-7.54532695e-01 -1.08166635e-01 4.51191008e-01 7.43125916e-01
7.15944052e-01 1.58105224e-01 4.03661840e-02 8.87903810e-01
2.06510112e-01 1.87372848e-01 9.85945821e-01 -6.93983734e-01
-2.12825492e-01 7.48955965e-01 -6.46876335e-01 -7.04336882e-01
-6.50607169e-01 -1.02862239e+00 -1.04410887e+00 2.56083220e-01
2.56024420e-01 1.76195920e-01 -1.24001157e+00 2.04898643e+00
4.54489142e-01 1.55780196e-01 3.23967010e-01 1.09359252e+00
1.32366896e+00 -3.70787606e-02 5.89048028e-01 2.57661015e-01
1.56457734e+00 -9.14201319e-01 -5.02561092e-01 -2.72106826e-01
6.32078886e-01 -2.97259361e-01 1.21718073e+00 -1.40300468e-01
-7.62172937e-01 -3.69830132e-01 -9.46988881e-01 -6.46198122e-03
-1.45719945e-01 -6.72943965e-02 9.99672651e-01 1.78921700e-01
-1.13658881e+00 7.04577684e-01 -9.96104956e-01 -4.42717522e-01
1.22282064e+00 9.70941037e-02 -9.26998019e-01 -2.97050681e-02
-1.06495905e+00 8.14522743e-01 4.12351698e-01 -2.68222213e-01
-1.20708132e+00 -1.14759243e+00 -1.09631550e+00 -1.42764598e-01
3.14280540e-01 -1.11523533e+00 9.69659984e-01 -1.26907337e+00
-8.56016994e-01 1.80802810e+00 1.16609283e-01 -4.92623329e-01
3.93548936e-01 1.17915034e-01 -4.51584935e-01 6.61947548e-01
5.83985984e-01 9.24516797e-01 6.66908801e-01 -1.57588685e+00
-1.09845988e-01 -7.96823204e-01 -1.35375544e-01 2.00310856e-01
-2.92518854e-01 -4.68398780e-01 -2.84927189e-01 -1.05749071e+00
2.38209814e-01 -8.16877306e-01 -5.51844418e-01 4.43177909e-01
-5.39945066e-01 2.00289294e-01 3.39540094e-01 -6.87867582e-01
7.05944777e-01 -2.28057146e+00 1.10502973e-01 7.48367086e-02
6.25962496e-01 -5.68194017e-02 -3.06275934e-02 9.43167433e-02
-4.10759062e-01 4.76447567e-02 -7.30269372e-01 -1.79072037e-01
-3.58006865e-01 5.22928596e-01 -7.44408742e-02 5.32509387e-01
1.29684091e-01 1.39123595e+00 -1.26016283e+00 -7.28446186e-01
2.35193253e-01 3.51742089e-01 -4.05575782e-01 2.14325383e-01
-9.00126770e-02 9.79882717e-01 -6.12014592e-01 8.73128891e-01
2.78955311e-01 -8.55075061e-01 2.07881585e-01 -3.97784799e-01
5.19683003e-01 8.82576406e-02 -2.88466424e-01 2.69338250e+00
-4.01042134e-01 9.58963335e-02 7.66117349e-02 -1.38670707e+00
7.15498090e-01 3.22703630e-01 9.11210597e-01 -8.97915602e-01
1.98760889e-02 3.93311977e-01 -8.91451985e-02 -6.29752517e-01
-5.16590662e-02 -6.51610732e-01 1.36091663e-02 2.82231987e-01
6.00504577e-01 -6.74521849e-02 -2.31505454e-01 4.59121585e-01
1.02654183e+00 1.28896639e-01 3.77366245e-01 -6.99310601e-01
1.76808640e-01 3.39070767e-01 3.83193821e-01 4.66755271e-01
-4.43304747e-01 6.96510315e-01 1.67834625e-01 -3.75077546e-01
-8.48685682e-01 -1.56880534e+00 -5.14553428e-01 6.14801943e-01
4.01481181e-01 -2.30976790e-01 -6.25189424e-01 -8.70374501e-01
1.28418878e-01 2.10083723e-01 -1.08647871e+00 -4.09893095e-01
-2.93296874e-01 -6.02356374e-01 5.73996365e-01 6.54872954e-01
4.69812572e-01 -1.17494130e+00 -6.78637326e-01 1.49189278e-01
-9.34528187e-02 -1.07267225e+00 -2.37301320e-01 1.54146388e-01
-1.18121755e+00 -1.21521008e+00 -9.38598335e-01 -1.03077376e+00
8.53635848e-01 7.24984035e-02 1.24038601e+00 2.25836337e-01
-9.48675394e-01 6.56903386e-01 -2.79232532e-01 -1.61239594e-01
-4.24208343e-01 -8.84028152e-02 -3.38178843e-01 -2.11251557e-01
-1.58909798e-01 -8.19254935e-01 -9.56574440e-01 -2.01587453e-01
-8.96906793e-01 3.50650787e-01 7.24024534e-01 1.03118718e+00
8.55472445e-01 -4.52402204e-01 7.13495672e-01 -1.25426793e+00
1.71985999e-01 -8.52927208e-01 1.44233644e-01 2.52600670e-01
-5.49889863e-01 2.14011148e-01 3.34762424e-01 -1.71382297e-02
-9.07396376e-01 3.18396538e-02 -4.60506231e-01 -5.67373991e-01
-4.98023957e-01 4.81958896e-01 1.66435003e-01 3.65937017e-02
8.63785923e-01 3.66771072e-01 5.74968874e-01 -4.47658211e-01
4.20411527e-01 1.92189738e-01 1.05233705e+00 -5.70345640e-01
5.34027994e-01 9.98816133e-01 9.86108184e-02 -4.93234932e-01
-1.30849183e+00 -5.36837578e-01 -4.54479486e-01 -4.06946652e-02
1.21924675e+00 -1.02436864e+00 -2.46897966e-01 1.47327080e-01
-5.55378497e-01 -3.60446453e-01 -6.86621785e-01 3.49494994e-01
-7.11599350e-01 4.07400221e-01 -6.35369003e-01 -1.29361063e-01
-6.24808848e-01 -1.12239456e+00 1.26007247e+00 1.61304817e-01
-4.20087844e-01 -1.47719002e+00 9.55966935e-02 5.12865484e-01
3.45523536e-01 7.19162285e-01 1.13617492e+00 -6.24814034e-01
-2.10621879e-01 1.57015488e-01 -2.98891813e-01 2.33081132e-01
5.05429983e-01 -6.47511661e-01 -9.90791619e-01 -2.27535054e-01
-7.62963220e-02 -5.86034179e-01 9.35750723e-01 4.32829082e-01
1.53474152e+00 -1.91830188e-01 -4.37144399e-01 1.05160451e+00
1.56326735e+00 -3.98794144e-01 3.34098995e-01 2.99172670e-01
7.30543613e-01 7.49582231e-01 1.91281706e-01 1.15043901e-01
5.54359972e-01 3.48578453e-01 7.34068096e-01 -7.10132718e-01
-6.65048063e-01 -4.04283971e-01 -4.32052791e-01 5.88149250e-01
1.30468249e-01 3.13772410e-01 -1.07834804e+00 6.84889436e-01
-1.58634210e+00 -6.92304611e-01 1.01032980e-01 1.80032253e+00
1.12225401e+00 -1.76989362e-01 -9.52409208e-02 -2.20490694e-01
5.89101136e-01 1.50750712e-01 -7.17839658e-01 2.03831449e-01
-1.89550817e-01 6.25077665e-01 3.97941828e-01 1.34685561e-01
-1.29943502e+00 1.00924826e+00 5.96836948e+00 7.15754509e-01
-1.29047871e+00 4.63194132e-01 7.89936364e-01 1.50620475e-01
-6.13071978e-01 -1.30217403e-01 -4.36917581e-02 3.09686512e-01
4.80686188e-01 -1.43813446e-01 4.34637396e-03 1.04822838e+00
-2.57506460e-01 1.46483704e-01 -8.67209792e-01 1.15295315e+00
6.68574721e-02 -1.73290825e+00 2.84155577e-01 -1.03276581e-01
7.15720952e-01 1.12445816e-01 1.35224998e-01 1.37913274e-02
3.09582561e-01 -1.18069732e+00 5.69305182e-01 3.38353276e-01
1.23199773e+00 -1.34553194e-01 3.43044400e-01 -8.15217495e-02
-1.00921690e+00 6.78479224e-02 2.08199937e-02 4.63965744e-01
-6.69234758e-03 5.15452623e-01 -6.63577259e-01 8.72771978e-01
9.50334013e-01 1.29801762e+00 -4.75338966e-01 8.30756962e-01
-2.28912979e-01 5.06343901e-01 1.08708642e-01 5.65089107e-01
4.51137751e-01 2.49412835e-01 5.23643076e-01 1.08491290e+00
-9.90644321e-02 9.23630372e-02 2.78992623e-01 1.07839251e+00
-1.77265137e-01 2.60215610e-01 -4.95417327e-01 9.01737511e-02
1.60810817e-02 1.45698977e+00 -9.23872769e-01 -5.63251317e-01
-1.64260790e-01 1.08093417e+00 4.14818436e-01 3.11250687e-01
-7.06380963e-01 1.47013173e-01 7.51338363e-01 2.60922909e-01
-2.42599055e-01 8.68638754e-02 -5.76501131e-01 -1.36459100e+00
-1.33230820e-01 -5.70311606e-01 8.13353360e-01 -8.28519762e-01
-1.68839538e+00 7.53031611e-01 -1.84384912e-01 -1.18508923e+00
1.17734196e-02 -5.78957438e-01 -5.11946142e-01 6.51189446e-01
-1.65177810e+00 -1.39189267e+00 -6.09500051e-01 9.48552489e-01
3.38677287e-01 -1.65657043e-01 1.08658564e+00 3.04888844e-01
-1.57648399e-01 4.53388423e-01 -2.57145524e-01 2.58361757e-01
7.82014251e-01 -1.49582660e+00 -9.69112152e-04 3.23112667e-01
1.89555928e-01 5.47291696e-01 3.29352766e-01 -7.41477132e-01
-1.14150333e+00 -1.13563311e+00 2.05695614e-01 -3.98697615e-01
7.42677927e-01 -3.81581746e-02 -1.06055975e+00 5.75652182e-01
-4.26185653e-02 5.38552046e-01 9.89546239e-01 -5.41629978e-02
-4.79850262e-01 1.48505136e-01 -1.19932842e+00 5.17257214e-01
1.54073155e+00 -5.93041956e-01 -9.13717449e-01 5.94923258e-01
8.03187549e-01 -5.04641414e-01 -1.18252254e+00 6.62856221e-01
3.41369182e-01 -7.22577333e-01 1.23959076e+00 -8.74289095e-01
8.25987339e-01 -4.71064029e-03 -8.34951252e-02 -1.18333566e+00
-1.35696754e-01 -1.42001748e-01 1.28553674e-01 7.00064301e-01
2.40784973e-01 -6.88732624e-01 8.65865588e-01 3.71609390e-01
-6.97700500e-01 -1.04976714e+00 -9.29078043e-01 -6.19679689e-01
2.26201758e-01 -2.12586045e-01 4.16879058e-01 1.44905496e+00
5.68330698e-02 5.34598120e-02 -3.33919413e-02 -1.19911388e-01
9.59926486e-01 3.44451696e-01 4.04085040e-01 -1.08493555e+00
-3.85022640e-01 -5.92585385e-01 -8.14143836e-01 -5.33679247e-01
3.01468730e-01 -1.85949183e+00 -3.08060855e-01 -1.72366691e+00
5.22002280e-01 -8.11418951e-01 -5.09582520e-01 8.52264702e-01
-2.38911644e-01 4.90966946e-01 -2.04686522e-01 5.40951788e-01
-5.62270522e-01 6.80274785e-01 1.78823006e+00 -1.42873973e-01
2.27555946e-01 -3.87685359e-01 -1.13615024e+00 7.04158247e-01
7.02415943e-01 -3.61837089e-01 -4.26718324e-01 -4.03998107e-01
-3.62385482e-01 1.13906644e-01 8.76313329e-01 -9.33795094e-01
-1.78406425e-02 2.18516380e-01 5.66524327e-01 9.94945541e-02
7.29756504e-02 -6.24743938e-01 7.24813119e-02 6.89186394e-01
-4.83997017e-01 -3.31615299e-01 3.02081317e-01 5.24842978e-01
-4.35352653e-01 1.36537194e-01 8.74380112e-01 -7.12665617e-01
-9.91317391e-01 6.71266794e-01 1.39161184e-01 4.76558477e-01
9.79544342e-01 -3.62274826e-01 -1.76990211e-01 -3.70711535e-02
-1.45867729e+00 1.25194997e-01 4.59993571e-01 4.36570406e-01
7.42059410e-01 -1.25406468e+00 -6.01653159e-01 1.78325370e-01
4.95989501e-01 4.51523177e-02 8.65345597e-01 8.35326254e-01
-5.44621706e-01 1.12359516e-01 -6.49418712e-01 -9.92666423e-01
-8.20574224e-01 4.72487807e-01 4.23419476e-01 -2.70843714e-01
-1.27750540e+00 7.26852238e-01 7.18718648e-01 -6.17437065e-01
-4.48967069e-02 -1.03083447e-01 -1.79450959e-03 -5.43267906e-01
1.50317430e-01 -3.26755673e-01 -1.36229217e-01 -6.48385704e-01
-4.55080658e-01 5.08422196e-01 8.06026906e-02 1.83109596e-01
1.41765666e+00 7.92583451e-02 -1.08613431e-01 1.38695031e-01
1.36133492e+00 -3.58665407e-01 -1.09375608e+00 -4.20826256e-01
1.13486629e-02 -3.56792539e-01 2.56771799e-02 -1.02178717e+00
-1.42038548e+00 8.03527415e-01 6.80322409e-01 -3.81928027e-01
1.09255636e+00 6.01412296e-01 8.19076061e-01 -1.76296160e-01
5.07939339e-01 -6.16857231e-01 3.60219747e-01 -1.52059406e-01
9.65021431e-01 -1.53000951e+00 -1.74710289e-01 -5.57122469e-01
-9.44752514e-01 7.05357373e-01 6.93263173e-01 -2.84782648e-01
8.13566029e-01 1.53195903e-01 2.40742102e-01 -6.92836165e-01
-3.92588347e-01 -3.91799152e-01 4.46056962e-01 6.72557414e-01
4.61235553e-01 3.02111894e-01 -1.03924703e-02 6.80600882e-01
-3.18313211e-01 -9.58102643e-02 8.61335248e-02 9.30934608e-01
-1.54699683e-01 -8.36776316e-01 5.74720800e-02 6.32694960e-01
-5.04312634e-01 -1.82709068e-01 -1.31713971e-01 7.29671359e-01
-2.33390313e-02 3.29627961e-01 -4.07838374e-02 -6.20408393e-02
9.08690095e-02 -6.23655505e-02 5.61151683e-01 -8.84397447e-01
-5.51012993e-01 -1.53663144e-01 -1.77652434e-01 -7.45942771e-01
-4.16340142e-01 -2.68163562e-01 -1.67158484e+00 2.01666862e-01
1.91521183e-01 -4.57170345e-02 5.68707645e-01 8.26382160e-01
3.77153605e-01 7.59289563e-01 2.22978994e-01 -4.41670388e-01
-3.23320597e-01 -5.83238959e-01 -5.62471330e-01 1.02741349e+00
2.50511527e-01 -7.93406665e-01 -8.05908367e-02 8.62378329e-02] | [14.674354553222656, -2.1915102005004883] |
284d28db-b7b1-4b02-ab66-ed7b54c8b927 | normalized-human-pose-features-for-human | null | null | http://openaccess.thecvf.com//content/ICCV2021/html/Liu_Normalized_Human_Pose_Features_for_Human_Action_Video_Alignment_ICCV_2021_paper.html | http://openaccess.thecvf.com//content/ICCV2021/papers/Liu_Normalized_Human_Pose_Features_for_Human_Action_Video_Alignment_ICCV_2021_paper.pdf | Normalized Human Pose Features for Human Action Video Alignment | We present a novel approach for extracting human pose features from human action videos. The goal is to let the pose features capture only the poses of the action while being invariant to other factors, including video backgrounds, the video subject's anthropometric characteristics and viewpoints. Such human pose features facilitate the comparison of pose similarity and can be used for down-stream tasks, such as human action video alignment and pose retrieval. The key to our approach is to first normalize the poses in the video frames by retargeting the poses onto a pre-defined 3D skeleton to not only disentangle subject physical features, such as bone lengths and ratios, but also to unify global orientations of the poses. Then the normalized poses are mapped to a pose embedding space of high-level features, learned via unsupervised metric learning. We evaluate the effectiveness of our normalized features both qualitatively by visualizations, and quantitatively by a video alignment task on the Human3.6M dataset and an action recognition task on the Penn Action dataset. | ['Chiew-Lan Tai', 'Hongbo Fu', 'Qifeng Chen', 'Mingyi Shi', 'Jingyuan Liu'] | 2021-01-01 | null | null | null | iccv-2021-1 | ['pose-retrieval', 'video-alignment'] | ['computer-vision', 'computer-vision'] | [ 2.65595257e-01 -1.15576826e-01 -2.73997784e-01 -5.06762564e-01
-5.98560035e-01 -5.37298620e-01 6.88481510e-01 1.58634767e-01
-4.45255041e-01 2.81187356e-01 8.29166055e-01 4.52326655e-01
-1.87249318e-01 -2.89471179e-01 -4.34822232e-01 -5.67564130e-01
-2.53912985e-01 3.99205178e-01 1.43572584e-01 -9.25284401e-02
2.94995040e-01 6.48750126e-01 -1.54975057e+00 5.14087453e-02
7.47940987e-02 8.78301442e-01 -4.34582204e-01 7.73920953e-01
5.08936524e-01 5.86961746e-01 -4.80872214e-01 -1.32275775e-01
4.87199575e-01 -5.77781618e-01 -7.03464508e-01 3.96446794e-01
8.99106860e-01 -4.28662956e-01 -5.40690541e-01 9.37664151e-01
4.20771658e-01 3.77013862e-01 7.69012094e-01 -1.41153550e+00
4.05144058e-02 5.09392470e-02 -5.14917731e-01 2.72967905e-01
1.01898444e+00 -7.04285726e-02 1.01877737e+00 -6.77988827e-01
9.42567945e-01 1.45459080e+00 6.06075704e-01 4.12656873e-01
-1.21244919e+00 -4.07390684e-01 -1.54024899e-01 4.09093857e-01
-1.37593746e+00 -5.21871567e-01 8.11699808e-01 -8.25931013e-01
5.04330039e-01 4.05121386e-01 8.84784102e-01 9.32584405e-01
2.75784284e-01 4.11504894e-01 7.32725680e-01 -3.90991122e-01
1.07713707e-01 -4.79846120e-01 -2.43793994e-01 9.00116742e-01
1.00124434e-01 -1.20742731e-01 -8.44093978e-01 -6.98010027e-02
9.82281625e-01 2.85940737e-01 -3.29782575e-01 -1.13228571e+00
-1.61104345e+00 7.28587389e-01 2.70022899e-01 4.78247665e-02
-2.75338680e-01 1.66994855e-01 2.99607605e-01 9.61142480e-02
2.47237951e-01 5.76958239e-01 -4.62295294e-01 -3.56865257e-01
-5.12133241e-01 3.61621350e-01 4.91894811e-01 7.73253918e-01
7.87571192e-01 -3.81442428e-01 -1.87439829e-01 3.37940693e-01
2.93141067e-01 5.35510600e-01 5.52621365e-01 -1.28965437e+00
4.69814122e-01 5.95550478e-01 2.69933348e-03 -1.52727222e+00
-5.74888349e-01 3.00025165e-01 -2.15667665e-01 6.51592016e-02
7.11539745e-01 1.73690677e-01 -7.11388350e-01 1.79278600e+00
5.31183660e-01 -1.68241560e-01 -3.11945766e-01 9.95723486e-01
5.53192973e-01 8.47828761e-02 -6.11660024e-03 3.48118842e-02
1.47885883e+00 -6.10123336e-01 -3.75556618e-01 -9.85204242e-04
7.84659564e-01 -6.18893862e-01 8.91335785e-01 -2.70373989e-02
-1.01753533e+00 -5.16738117e-01 -9.62067604e-01 -1.64130360e-01
-1.64202422e-01 6.45022243e-02 1.20725349e-01 3.79804313e-01
-5.26714146e-01 7.82893419e-01 -1.21824396e+00 -7.06367850e-01
1.00605756e-01 3.96055132e-01 -7.91521430e-01 2.50819057e-01
-8.41465116e-01 9.57423985e-01 1.74939767e-01 -1.46845952e-01
-5.72363615e-01 -4.87612128e-01 -1.27814984e+00 -1.87055692e-01
4.59860444e-01 -5.30474484e-01 9.89235163e-01 -7.38652885e-01
-1.55884194e+00 9.93038893e-01 8.51381645e-02 -6.79212734e-02
5.36809862e-01 -4.39246148e-01 -2.04111546e-01 6.36987627e-01
7.63825178e-02 5.87525010e-01 7.94205785e-01 -6.06793582e-01
-4.29718435e-01 -7.59653568e-01 2.12518960e-01 6.18653238e-01
-2.95025140e-01 -2.35261973e-02 -5.69880068e-01 -7.96184003e-01
3.97234648e-01 -1.23963523e+00 -1.60707742e-01 6.07774913e-01
-1.40418723e-01 1.51986584e-01 6.83778048e-01 -9.07999754e-01
8.93064737e-01 -2.27860928e+00 9.35663223e-01 4.19183701e-01
3.63071918e-01 -2.78610975e-01 -1.95386350e-01 1.72288865e-01
-2.21960679e-01 -2.79921055e-01 7.38193169e-02 -8.41637254e-02
-1.60737514e-01 1.80690676e-01 3.39544505e-01 1.00737965e+00
1.02126896e-01 8.62032413e-01 -1.03700256e+00 -7.50805020e-01
4.40751672e-01 3.56207699e-01 -6.51904345e-01 2.45194644e-01
1.40557006e-01 6.94832563e-01 -5.50438404e-01 4.92745876e-01
-6.43164217e-02 8.18856508e-02 2.61886567e-01 -6.63214028e-01
2.54427046e-01 2.03419238e-01 -1.26471412e+00 2.07571220e+00
5.07815890e-02 6.44475937e-01 -2.74397254e-01 -9.00789559e-01
6.31459653e-01 1.70403183e-01 1.11052394e+00 -3.74726176e-01
3.24452817e-01 -1.49363667e-01 -7.22588226e-03 -5.40135503e-01
2.39892364e-01 2.89786249e-01 -9.02906582e-02 5.07020175e-01
1.54278740e-01 9.79620740e-02 2.02552378e-01 4.90481853e-02
9.83799815e-01 3.00090134e-01 5.98946691e-01 -4.16284561e-01
3.27226341e-01 -3.86311501e-01 2.57455498e-01 1.57710277e-02
-2.33153597e-01 8.28961432e-01 4.99484122e-01 -5.17680943e-01
-1.06669164e+00 -1.33006096e+00 1.63002819e-01 8.63911748e-01
6.02750555e-02 -8.61614287e-01 -7.93045938e-01 -6.69448853e-01
1.15876175e-01 -5.49381860e-02 -8.62765431e-01 -4.36048925e-01
-6.18532658e-01 -1.42273813e-01 2.65302658e-01 8.05423319e-01
6.56743050e-02 -6.61406875e-01 -1.12369812e+00 -2.26237029e-01
-2.80304104e-01 -1.25103796e+00 -8.47907543e-01 -1.71493754e-01
-8.76193166e-01 -1.42750657e+00 -5.05611718e-01 -3.82390678e-01
8.00197661e-01 2.22776961e-02 7.78744996e-01 -8.95634294e-02
-6.12938762e-01 8.58375132e-01 -4.15097952e-01 6.89268811e-03
-1.58208653e-01 -2.46455953e-01 6.18537962e-01 2.26836517e-01
2.09863365e-01 -6.35721922e-01 -7.75607407e-01 5.49503982e-01
-7.71181703e-01 1.38367442e-02 2.47325122e-01 4.59966063e-01
5.97273052e-01 -3.82019430e-01 -2.82308429e-01 -3.11898559e-01
8.11626911e-02 1.33122012e-01 -1.68169364e-01 2.74962097e-01
-1.23990206e-02 -3.67386825e-03 1.14347778e-01 -5.18805802e-01
-2.89503962e-01 3.49425405e-01 3.70456934e-01 -5.99707186e-01
-2.16106232e-02 3.26772034e-01 -5.67205667e-01 -5.09180799e-02
7.34250665e-01 -1.84953809e-01 2.57402271e-01 -3.49491596e-01
4.30828422e-01 3.08806092e-01 7.98876524e-01 -6.36302829e-01
8.38924110e-01 5.14807463e-01 3.72368306e-01 -1.02353179e+00
-6.37004316e-01 -5.40987790e-01 -1.45742691e+00 -2.88787097e-01
1.24451709e+00 -7.56446779e-01 -6.75812960e-01 2.43722245e-01
-8.72616112e-01 -1.41117722e-03 -3.77886325e-01 7.76773155e-01
-1.03015888e+00 6.65389359e-01 -3.13536227e-01 -2.78001368e-01
1.56559587e-01 -1.08239651e+00 1.38357031e+00 -8.28720257e-02
-7.96836197e-01 -8.43148351e-01 1.96787417e-01 3.15906733e-01
-2.49710947e-01 7.18067348e-01 9.09286141e-01 -5.77767730e-01
-2.34117970e-01 -3.77840042e-01 2.48886451e-01 2.53985167e-01
5.70732832e-01 1.83326807e-02 -5.59236109e-01 -4.55110610e-01
-1.84499875e-01 -1.53550148e-01 3.37354094e-01 2.59204626e-01
9.14685547e-01 -1.85383230e-01 -2.82288998e-01 7.59694517e-01
9.05053079e-01 -9.91828591e-02 4.83931959e-01 2.24566862e-01
1.01745057e+00 7.44417012e-01 6.07769549e-01 4.56595391e-01
9.28533971e-02 1.13385570e+00 1.15878619e-01 -4.19155024e-02
-3.75736728e-02 -4.48007137e-01 4.27582532e-01 8.29705775e-01
-4.71296638e-01 4.04902786e-01 -7.62855411e-01 5.15806861e-03
-1.83291876e+00 -8.03548038e-01 3.29470098e-01 2.45843768e+00
6.45814717e-01 -2.73248274e-02 5.48282862e-01 2.27518827e-01
6.10915959e-01 1.85634568e-01 -4.36454445e-01 9.21378210e-02
2.70748764e-01 1.69470504e-01 3.35238367e-01 3.83454114e-01
-1.24457264e+00 5.06730258e-01 6.27188110e+00 1.29106328e-01
-9.45864618e-01 -3.09448630e-01 1.33440211e-01 -4.84152794e-01
2.02765360e-01 -9.69667658e-02 -3.46036226e-01 1.25586346e-01
6.46173477e-01 -2.63211399e-01 1.31264031e-01 7.48992741e-01
4.55078110e-02 1.33349998e-02 -1.73568201e+00 1.15223014e+00
4.33899105e-01 -9.13611054e-01 1.49629936e-01 1.12163037e-01
3.64744276e-01 -3.61851275e-01 -1.98439971e-01 -1.35606423e-01
-3.32394391e-01 -9.70167696e-01 7.48579800e-01 6.37032628e-01
8.38855743e-01 -5.86526155e-01 2.84002036e-01 -8.83660242e-02
-1.25503314e+00 5.69093935e-02 -1.28891766e-01 -2.17895731e-01
4.02438529e-02 -1.60134107e-01 -5.56536019e-01 4.76163656e-01
7.20476747e-01 1.05448043e+00 -8.04737628e-01 8.61471415e-01
-5.15328497e-02 6.57828106e-03 -3.47630471e-01 2.04911053e-01
-4.02351946e-01 -2.11140916e-01 5.91475666e-01 8.96665812e-01
2.13504240e-01 2.46224463e-01 4.34143901e-01 3.20657283e-01
1.77445248e-01 3.49548429e-01 -5.21308661e-01 -2.48442069e-01
1.01067416e-01 1.13058436e+00 -8.14946651e-01 -1.16771884e-01
-4.62122411e-01 9.06655431e-01 1.51110310e-02 2.71594554e-01
-8.02309453e-01 -2.00789049e-01 1.08593297e+00 3.99095595e-01
1.21101201e-01 -6.21004283e-01 2.36445874e-01 -1.35602736e+00
1.91250548e-01 -1.14247882e+00 3.63076001e-01 -7.16585636e-01
-9.40754056e-01 2.79584169e-01 4.80269194e-01 -1.54371107e+00
-4.91555691e-01 -7.49651015e-01 -2.37199828e-01 4.63531733e-01
-4.26479846e-01 -9.43083644e-01 -4.36207891e-01 7.63823986e-01
4.51463908e-01 -8.94552190e-03 8.87818575e-01 2.09139600e-01
-4.36938494e-01 5.41802347e-01 -2.13680819e-01 3.99978936e-01
7.92185724e-01 -1.06070626e+00 2.84875005e-01 6.04238749e-01
6.51417077e-01 6.30455732e-01 8.26996326e-01 -4.17324811e-01
-1.44660330e+00 -6.72230959e-01 5.15884042e-01 -9.61184800e-01
6.11719608e-01 -5.73870778e-01 -4.51383352e-01 6.74993038e-01
-4.65211123e-01 2.29927853e-01 7.48889267e-01 2.42432561e-02
-3.87381196e-01 -1.32611021e-01 -8.63268495e-01 8.15987587e-01
1.29075944e+00 -6.19674087e-01 -7.55038261e-01 3.10439676e-01
3.57000709e-01 -5.84405422e-01 -1.05707395e+00 3.37576866e-01
1.11033392e+00 -7.40604579e-01 1.13774371e+00 -1.11627245e+00
5.65108359e-01 -2.86462039e-01 -6.03412032e-01 -1.03952217e+00
-3.50404650e-01 -4.88266528e-01 2.82178167e-02 8.45835090e-01
2.31585489e-03 -9.69358012e-02 9.24859464e-01 6.38382316e-01
3.85019034e-01 -6.93918765e-01 -1.01049531e+00 -6.48726463e-01
-2.86342502e-01 -9.66052711e-02 4.17860597e-01 7.40885794e-01
1.21582128e-01 1.51951611e-01 -3.13416362e-01 -2.90751904e-02
3.80944848e-01 2.67933886e-02 1.03117800e+00 -1.37105179e+00
-2.58382350e-01 -2.55664200e-01 -1.36626232e+00 -7.80030072e-01
4.54659127e-02 -7.14803815e-01 -3.43345068e-02 -9.82853115e-01
4.07977849e-01 3.14112782e-01 -3.64030957e-01 3.88252974e-01
4.15633954e-02 3.83334041e-01 3.26502770e-01 2.25848198e-01
-5.44034183e-01 5.09362757e-01 1.07492077e+00 2.41649225e-02
-1.26090124e-01 -1.88553289e-01 -2.73925245e-01 1.01586640e+00
4.21072870e-01 -3.70540082e-01 -4.38124627e-01 -2.51235247e-01
-4.28698696e-02 1.49286821e-01 4.81891662e-01 -1.16126454e+00
-1.63255364e-01 -3.74595731e-01 6.64097846e-01 -2.74081618e-01
4.78309602e-01 -7.82520890e-01 1.15994193e-01 5.34191251e-01
-4.71979380e-01 3.79963011e-01 -9.47294831e-02 5.47658563e-01
-7.23220315e-03 1.89726800e-01 5.82293212e-01 -8.97649676e-02
-4.78663713e-01 4.13754135e-01 3.94901782e-02 2.69159377e-01
1.09949672e+00 -5.90458632e-01 1.49494395e-01 -6.11495316e-01
-9.05558169e-01 -9.71858799e-02 8.74273658e-01 7.10100174e-01
4.98310357e-01 -1.56307185e+00 -4.76386994e-01 3.67820501e-01
3.55836689e-01 -2.71920204e-01 -5.37008271e-02 1.02035618e+00
-6.11736655e-01 2.66982138e-01 -7.18849957e-01 -9.66398478e-01
-1.71719182e+00 4.46877480e-01 3.18230659e-01 1.65727392e-01
-7.51689732e-01 5.42087436e-01 3.57986361e-01 -2.67048448e-01
3.83909494e-01 -4.29301262e-01 -1.51113302e-01 1.91517159e-01
4.00262803e-01 5.67424834e-01 -5.13929315e-02 -1.20843339e+00
-6.82178378e-01 9.90166962e-01 2.90911704e-01 -1.53009340e-01
1.21733391e+00 -1.39247403e-01 1.36850715e-01 5.02946794e-01
1.61334074e+00 1.17752053e-01 -1.26280570e+00 -2.45531648e-01
1.09953126e-02 -8.46039712e-01 -4.04854149e-01 -2.66996890e-01
-1.06338000e+00 7.00211227e-01 6.72371268e-01 -3.24609548e-01
9.08677638e-01 1.10714510e-01 4.83849555e-01 4.89740580e-01
3.79233539e-01 -1.00760269e+00 5.44691682e-01 1.68536544e-01
1.01614141e+00 -9.47374403e-01 4.94579583e-01 -4.29173827e-01
-4.61140275e-01 1.16331589e+00 3.52659702e-01 -1.12199642e-01
6.58132493e-01 -6.04703277e-02 2.73488194e-01 -1.69174671e-01
-1.15622759e-01 -5.19541055e-02 8.26600373e-01 5.82680583e-01
5.35961866e-01 7.66733289e-02 -3.27618957e-01 1.59395754e-01
-5.06798863e-01 -2.46864632e-01 2.74785876e-01 1.04405284e+00
-1.94605291e-01 -1.07937026e+00 -2.27050155e-01 3.33918124e-01
-2.82635063e-01 4.33639824e-01 -5.82301855e-01 8.12479556e-01
-6.56588152e-02 4.69559908e-01 1.63070872e-01 -6.05007529e-01
6.57649398e-01 8.47315416e-02 9.78793681e-01 -7.77907848e-01
4.94008921e-02 1.67461317e-02 -9.21924785e-02 -1.04854906e+00
-5.35847187e-01 -1.03763378e+00 -1.04037058e+00 1.12119086e-01
1.93925912e-03 -1.37591317e-01 5.42897761e-01 9.57995653e-01
1.43104643e-01 1.23779453e-01 4.67887759e-01 -1.19311762e+00
-6.41791582e-01 -7.76157856e-01 -6.84502840e-01 1.07717800e+00
2.52766281e-01 -1.09380329e+00 -1.70639277e-01 3.97625655e-01] | [7.129944324493408, -0.6922248601913452] |
6ed9e226-2ea6-45ff-b676-ad8df1e70103 | unconstrained-face-detection-and-open-set | 1708.02337 | null | http://arxiv.org/abs/1708.02337v3 | http://arxiv.org/pdf/1708.02337v3.pdf | Unconstrained Face Detection and Open-Set Face Recognition Challenge | Face detection and recognition benchmarks have shifted toward more difficult
environments. The challenge presented in this paper addresses the next step in
the direction of automatic detection and identification of people from outdoor
surveillance cameras. While face detection has shown remarkable success in
images collected from the web, surveillance cameras include more diverse
occlusions, poses, weather conditions and image blur. Although face
verification or closed-set face identification have surpassed human
capabilities on some datasets, open-set identification is much more complex as
it needs to reject both unknown identities and false accepts from the face
detector. We show that unconstrained face detection can approach high detection
rates albeit with moderate false accept rates. By contrast, open-set face
recognition is currently weak and requires much more attention. | ['Cezary Stankiewicz', 'Mohammad Iqbal Nouyed', 'Jürgen Beyerer', 'Deva Ramanan', 'Manuel Günther', 'Akshay Raj Dhamija', 'Mohamad Al Jazaery', 'Josef Kittler', 'Guodong Guo', 'Chi Ho Chan', 'Shufan Yang', 'Peiyun Hu', 'Terrance E. Boult', 'Min Jiang', 'Christian Herrmann'] | 2017-08-08 | null | null | null | null | ['robust-face-recognition'] | ['computer-vision'] | [ 4.51736510e-01 -2.89439261e-01 1.00499779e-01 -7.33828306e-01
-5.85284829e-01 -9.77986336e-01 7.28419542e-01 -5.26953459e-01
-5.27209044e-01 7.11254179e-01 -2.77388424e-01 -8.69222879e-02
1.57156140e-01 -2.32444763e-01 -3.47718000e-01 -7.17225552e-01
-7.22033605e-02 4.33324099e-01 -9.23009515e-02 2.86978367e-03
8.60830098e-02 6.81547523e-01 -1.99063814e+00 1.89411715e-01
1.89453810e-01 9.42019105e-01 -4.69979644e-01 1.01624966e+00
2.59261727e-01 7.46752918e-02 -7.10066736e-01 -6.24491394e-01
6.30833507e-01 -1.59986764e-01 -3.93470317e-01 4.28725481e-01
1.54360437e+00 -8.61808836e-01 -3.61339748e-01 1.29119706e+00
7.64249980e-01 -9.84183550e-02 5.61162412e-01 -1.60057116e+00
-5.56928277e-01 -3.62564981e-01 -7.19980299e-01 3.65838468e-01
8.54304552e-01 4.36388403e-01 2.34502986e-01 -1.17763186e+00
4.26763624e-01 1.60995126e+00 7.57783115e-01 8.80738437e-01
-1.32458293e+00 -1.03149521e+00 6.42561587e-03 -9.04061422e-02
-1.73088193e+00 -1.14052594e+00 -1.58032309e-02 -3.89274657e-01
9.03127015e-01 4.63914305e-01 3.66243273e-01 1.35342085e+00
-4.02065188e-01 2.31922194e-01 1.11802316e+00 -3.29163402e-01
-1.07690997e-01 3.18946272e-01 3.51833463e-01 6.70142949e-01
7.32222855e-01 2.36454859e-01 -3.99065197e-01 -4.65024740e-01
8.26421618e-01 2.74674237e-01 -4.00848597e-01 -5.36467284e-02
-8.51407826e-01 6.16926968e-01 -4.66830805e-02 -1.54573172e-01
-6.30795062e-02 -4.16771293e-01 2.47322187e-01 4.63340402e-01
2.04828650e-01 2.86132358e-02 -1.98249534e-01 -3.14997844e-02
-1.04823792e+00 1.53924674e-01 1.12745070e+00 1.05335391e+00
6.20023608e-01 -1.45031447e-02 -1.78810507e-01 7.81782866e-01
9.14464146e-02 1.01617289e+00 -4.67987172e-02 -9.54655647e-01
1.20399058e-01 3.72301370e-01 3.79089266e-01 -8.93657565e-01
-9.28696692e-02 -1.03978686e-01 -5.99487662e-01 6.80286825e-01
8.96515071e-01 -3.82805765e-01 -8.67378235e-01 1.52978957e+00
4.19147789e-01 2.26964295e-01 -5.07839248e-02 9.38363850e-01
7.33628094e-01 3.08676422e-01 -1.24266841e-01 -3.25787425e-01
1.55518663e+00 -4.10262823e-01 -7.73959756e-01 -5.55730581e-01
8.02161265e-03 -8.21883142e-01 3.91477525e-01 4.00947630e-01
-8.31577957e-01 -3.39843154e-01 -8.82814705e-01 1.38156474e-01
-2.97936797e-01 1.02862634e-01 4.35698986e-01 1.49333847e+00
-1.35179305e+00 -7.94493109e-02 -3.42744201e-01 -9.02646840e-01
7.45025933e-01 8.68568659e-01 -9.42276120e-01 -5.08944690e-01
-6.15226090e-01 9.64228809e-01 -8.14383477e-02 1.67967111e-01
-8.12155306e-01 -4.71546978e-01 -9.76706445e-01 -1.31595850e-01
6.32746279e-01 -3.29841882e-01 1.05450380e+00 -1.20125496e+00
-1.07643139e+00 1.39541018e+00 -5.33788562e-01 -3.06120843e-01
5.11410177e-01 -1.84087098e-01 -6.86322451e-01 9.27382037e-02
-3.29451784e-02 6.72491789e-01 1.16267145e+00 -9.79217410e-01
-6.09181225e-01 -7.49472916e-01 -4.48492020e-02 4.08837497e-02
-4.60092932e-01 9.43443358e-01 -3.87688190e-01 -3.10332775e-01
-2.33846545e-01 -1.06665075e+00 2.21798494e-01 4.16371435e-01
-4.28864099e-02 -6.20381422e-02 1.28421247e+00 -5.88738143e-01
7.24088728e-01 -2.05137920e+00 -3.85001361e-01 6.78705648e-02
1.54762670e-01 7.13143170e-01 -6.24095602e-03 4.06181477e-02
-1.07411847e-01 8.17468464e-02 2.06664950e-01 -3.76152456e-01
-1.42404810e-01 1.13811418e-01 -1.81418926e-01 6.79820061e-01
4.19393122e-01 4.42911148e-01 -5.81765890e-01 -4.92461860e-01
2.48389736e-01 6.84534431e-01 -5.29167354e-01 3.32500152e-02
4.62935150e-01 1.56973779e-01 1.10668145e-01 1.24203348e+00
9.47494388e-01 5.90465916e-03 1.95425883e-01 3.76332961e-02
-3.04195769e-02 -3.49786252e-01 -1.57993031e+00 8.13568592e-01
1.93028644e-01 8.53937924e-01 7.93221951e-01 -5.78500867e-01
8.09549809e-01 4.55878794e-01 2.36165941e-01 -2.15368375e-01
2.25259557e-01 4.01043259e-02 1.81450740e-01 -5.03323078e-01
4.19426233e-01 2.07841322e-02 3.45970064e-01 2.69758821e-01
6.01481013e-02 1.95604578e-01 2.98843682e-01 -8.40344355e-02
9.61670339e-01 -6.75569624e-02 3.57848495e-01 -2.37042278e-01
7.42311358e-01 7.17252959e-03 5.54215372e-01 9.49856937e-01
-8.01484942e-01 6.83129191e-01 2.82647144e-02 -5.94600558e-01
-7.04840124e-01 -1.17674303e+00 -4.30354744e-01 1.01063740e+00
-6.15366325e-02 -1.46079257e-01 -1.03672826e+00 -6.90199018e-01
2.16263890e-01 -4.65563424e-02 -4.80801642e-01 7.10339323e-02
-4.61747557e-01 -9.20416474e-01 8.76310289e-01 3.14352959e-01
3.94695789e-01 -6.73069775e-01 -3.69114727e-01 -3.40122312e-01
1.01708561e-01 -1.34561574e+00 -6.69618130e-01 -3.93465936e-01
-2.87353814e-01 -1.43058860e+00 -8.68746519e-01 -9.03012037e-01
9.81301010e-01 5.34381032e-01 1.10245037e+00 3.02927017e-01
-8.66815448e-01 7.46901870e-01 -2.12976485e-02 -6.35616720e-01
-2.29659900e-01 -5.66253245e-01 7.47518301e-01 4.35521901e-01
8.64554346e-01 1.30787209e-01 -6.62023306e-01 7.26074219e-01
-4.81473744e-01 -7.77627409e-01 2.10098311e-01 7.38044143e-01
1.01336300e-01 6.48733526e-02 4.96526390e-01 -4.85050112e-01
3.15873355e-01 -2.11984888e-01 -7.28446782e-01 3.28071505e-01
-1.94407701e-01 -6.72621965e-01 2.75173604e-01 -5.59906542e-01
-1.06559145e+00 2.50863522e-01 1.12881124e-01 -4.54290539e-01
-5.34522653e-01 -5.51261485e-01 -3.19531083e-01 -5.75828552e-01
7.98295081e-01 1.13682095e-02 3.46971869e-01 -2.18488425e-01
-2.32869044e-01 8.47576499e-01 7.64631808e-01 -1.79762945e-01
1.01704848e+00 6.53897941e-01 -3.27259749e-01 -1.34030044e+00
-5.25609255e-01 -4.96875107e-01 -7.20926702e-01 -3.88443708e-01
6.63608968e-01 -1.08622265e+00 -9.22481000e-01 6.85492098e-01
-9.79877472e-01 2.98027128e-01 2.49997377e-02 2.93823153e-01
9.93155595e-03 5.42156994e-01 -3.07646185e-01 -1.39254653e+00
-2.21238017e-01 -1.23499143e+00 1.21443236e+00 3.41845810e-01
-3.71393532e-01 -6.07426643e-01 -4.65032667e-01 4.97075140e-01
2.81786919e-01 2.04727635e-01 -4.33261916e-02 -7.00637341e-01
-3.99593443e-01 -8.48810971e-01 -3.02417517e-01 4.16550577e-01
4.57525253e-01 2.46666580e-01 -1.38833642e+00 -7.14691579e-01
-6.40485883e-02 -5.20356357e-01 6.93841040e-01 2.81898797e-01
7.54987359e-01 -2.17553779e-01 -4.29131299e-01 6.16952062e-01
1.15403473e+00 2.83388495e-01 4.54317361e-01 -3.68817151e-02
3.46151948e-01 8.10843349e-01 3.37704748e-01 4.02885705e-01
1.74084622e-02 6.12976551e-01 1.46553278e-01 -1.20676830e-01
-1.44661501e-01 2.22313344e-01 6.34691536e-01 -4.61600989e-01
3.39332670e-02 -2.57845879e-01 -9.79608655e-01 7.43846148e-02
-1.43612301e+00 -1.42848086e+00 -1.70045406e-01 2.61700273e+00
4.92918044e-01 -4.42953743e-02 3.71905327e-01 2.16981158e-01
1.12050152e+00 -2.19808668e-01 -5.13925433e-01 -2.18790229e-02
-4.23330307e-01 2.69844551e-02 3.94517392e-01 4.94097650e-01
-1.28348875e+00 5.91229320e-01 7.75420189e+00 2.81720370e-01
-8.52797031e-01 -1.24178194e-01 8.24573874e-01 -4.84935313e-01
6.16087079e-01 -2.55159378e-01 -1.47855449e+00 5.03673553e-01
6.26729548e-01 2.00904727e-01 4.84496653e-01 6.32989645e-01
-7.12157041e-02 -2.18579486e-01 -1.46935701e+00 1.52215815e+00
6.19299054e-01 -8.84437740e-01 -1.02066785e-01 2.97866017e-01
4.92079079e-01 -2.16099456e-01 2.21772075e-01 2.62043893e-01
1.15082867e-01 -1.28461123e+00 3.29301685e-01 3.83425772e-01
1.07678998e+00 -5.04622400e-01 6.41437829e-01 3.44403088e-01
-9.88863766e-01 -2.49617070e-01 -2.96004951e-01 -4.85176027e-01
-1.62889421e-01 3.39679904e-02 -7.85408020e-01 -1.87334388e-01
8.25532019e-01 2.94575423e-01 -7.35055566e-01 8.79373670e-01
3.82783800e-01 3.06564182e-01 -6.49355233e-01 1.04268923e-01
-1.36378407e-01 -1.81957096e-01 3.91106427e-01 1.24475944e+00
1.85636595e-01 2.45010018e-01 4.83311772e-01 3.97676170e-01
-9.61123034e-02 -3.29526573e-01 -1.02243721e+00 1.05603531e-01
5.28640330e-01 1.32407105e+00 -5.93843520e-01 -4.34554398e-01
-6.50126338e-01 9.77157712e-01 1.07936092e-01 3.59840304e-01
-7.21189320e-01 1.24663927e-01 1.33300328e+00 2.15643644e-01
9.93379429e-02 -1.49661481e-01 2.12048665e-01 -1.21956074e+00
7.91388527e-02 -1.10342205e+00 6.21606350e-01 -3.87471884e-01
-1.20244551e+00 3.54597539e-01 -3.14964503e-02 -9.66455340e-01
-3.83425236e-01 -1.04782581e+00 -5.90216935e-01 9.00494516e-01
-1.14349675e+00 -9.32816505e-01 -4.84826058e-01 7.86421597e-01
6.00479782e-01 -3.09740335e-01 9.48609829e-01 4.49532866e-01
-9.72480834e-01 9.47986424e-01 9.53531265e-02 4.44661349e-01
9.89128590e-01 -7.80485928e-01 3.50983381e-01 9.70441878e-01
4.24430473e-03 6.26625299e-01 6.36017382e-01 -5.10205030e-01
-1.83440399e+00 -7.42434263e-01 6.00756943e-01 -9.59663033e-01
2.55039722e-01 -6.76230907e-01 -5.44676781e-01 7.05148041e-01
8.55914652e-02 5.06009340e-01 7.11085498e-01 5.90126254e-02
-5.31492710e-01 -7.88117796e-02 -1.49669349e+00 4.25928116e-01
1.12710476e+00 -6.21232092e-01 -1.76615015e-01 5.34897089e-01
-1.04913235e-01 -1.30049184e-01 -5.80809951e-01 2.98797369e-01
8.50648522e-01 -1.01132369e+00 1.32330513e+00 -5.35756648e-01
-4.25626546e-01 -3.00503641e-01 -3.01824659e-01 -5.81890583e-01
-2.34006301e-01 -7.83949137e-01 6.48267418e-02 1.53234720e+00
-2.96939611e-02 -7.99051821e-01 8.65503788e-01 1.09466434e+00
4.83768374e-01 -1.65810451e-01 -9.38298881e-01 -9.50045526e-01
-3.50432932e-01 -1.00193091e-01 2.79055655e-01 7.93037415e-01
-2.66337812e-01 1.66977167e-01 -4.11280781e-01 4.37874615e-01
1.02253628e+00 -1.62989512e-01 9.41944361e-01 -1.43454003e+00
7.41684884e-02 -3.51625502e-01 -5.92829585e-01 -5.41635096e-01
3.14705819e-01 -4.76398230e-01 -1.11845404e-01 -8.61982286e-01
3.81242037e-01 4.29478809e-02 1.47022113e-01 3.80666733e-01
-1.77523419e-01 9.84032273e-01 1.73295766e-01 5.97804599e-02
-5.62255681e-01 -2.50676703e-02 5.48944116e-01 -2.09502116e-01
2.21128657e-01 4.83103916e-02 -5.90704143e-01 8.79964173e-01
5.68880856e-01 -1.12795226e-01 -9.39589888e-02 -3.99067730e-01
-3.62462819e-01 -2.93657109e-02 8.04208219e-01 -1.05852509e+00
3.67966533e-01 5.96180335e-02 1.02877057e+00 -2.28356957e-01
8.29881549e-01 -9.46914554e-01 7.38867298e-02 5.05898952e-01
1.63107872e-01 2.19811544e-01 3.68016571e-01 4.97052968e-01
1.01491123e-01 -1.15274362e-01 8.95371854e-01 -2.66412139e-01
-7.06141651e-01 4.79996771e-01 -3.88596892e-01 8.28907192e-02
1.13748538e+00 -8.99662852e-01 -3.57518345e-01 -2.63525367e-01
-7.80593336e-01 2.22212866e-01 6.78552508e-01 5.63566983e-01
7.08840907e-01 -9.60813642e-01 -9.71426010e-01 7.78103411e-01
5.95951788e-02 -5.55007398e-01 -5.23297861e-02 6.37848616e-01
-1.35383293e-01 2.00235099e-01 -1.63780153e-01 -7.25528300e-01
-2.09900236e+00 4.53794271e-01 5.15619338e-01 5.42967319e-01
-2.89191276e-01 9.61035907e-01 2.50393599e-01 -2.16527492e-01
4.52371240e-01 4.45334613e-01 -9.94042903e-02 1.71735376e-01
1.13282490e+00 5.48433781e-01 -6.42247498e-02 -1.15247965e+00
-6.84842587e-01 4.38980162e-01 -3.32186788e-01 9.89049301e-02
8.70856583e-01 5.15952893e-02 -1.56710725e-02 -1.64211705e-01
1.01804435e+00 -2.43760556e-01 -1.45512819e+00 -4.67280336e-02
-2.41402373e-01 -1.05993974e+00 -1.17011972e-01 -6.63151979e-01
-7.88483620e-01 5.90477705e-01 1.02226186e+00 1.03602119e-01
8.99221838e-01 -2.57911563e-01 3.37895721e-01 6.91718578e-01
4.80562478e-01 -9.01262999e-01 -2.09809810e-01 4.97259438e-01
7.02168465e-01 -2.00378251e+00 7.52081200e-02 -5.90922117e-01
-1.48168027e-01 9.63185012e-01 8.82514775e-01 2.41317749e-01
6.66239440e-01 4.84772235e-01 -2.39492338e-02 -3.28535028e-02
-4.21062261e-01 -1.40228674e-01 1.30744845e-01 9.92119610e-01
3.99490207e-01 -1.54733986e-01 4.07354563e-01 5.26231714e-02
1.17809363e-02 -1.69644549e-01 3.08195025e-01 8.97556663e-01
-3.54835659e-01 -7.85817325e-01 -1.20761704e+00 6.38354480e-01
-5.80613434e-01 6.59652650e-02 -8.05515647e-01 6.75932944e-01
1.17915072e-01 1.36867189e+00 2.61615396e-01 -7.74144009e-02
1.83578804e-01 2.80738205e-01 7.01963067e-01 -6.70845509e-01
-6.34155810e-01 -2.46017218e-01 1.29236728e-01 -4.68920350e-01
-3.03462654e-01 -1.09783065e+00 -6.30050063e-01 -4.70948786e-01
-4.15525585e-01 -2.56398141e-01 5.81593335e-01 7.00292289e-01
2.33635545e-01 -1.82344988e-01 4.67528403e-01 -1.18184531e+00
-7.95468807e-01 -7.75599658e-01 -4.79166776e-01 4.46604192e-01
6.74714148e-01 -7.02385128e-01 -4.14595932e-01 2.31395766e-01] | [13.340620994567871, 0.7946226596832275] |
f7f4fe64-17b8-41e9-b686-6fcf5b8029fb | se-md-a-single-encoder-multiple-decoder-deep | 2106.15325 | null | https://arxiv.org/abs/2106.15325v1 | https://arxiv.org/pdf/2106.15325v1.pdf | SE-MD: A Single-encoder multiple-decoder deep network for point cloud generation from 2D images | 3D model generation from single 2D RGB images is a challenging and actively researched computer vision task. Various techniques using conventional network architectures have been proposed for the same. However, the body of research work is limited and there are various issues like using inefficient 3D representation formats, weak 3D model generation backbones, inability to generate dense point clouds, dependence of post-processing for generation of dense point clouds, and dependence on silhouettes in RGB images. In this paper, a novel 2D RGB image to point cloud conversion technique is proposed, which improves the state of art in the field due to its efficient, robust and simple model by using the concept of parallelization in network architecture. It not only uses the efficient and rich 3D representation of point clouds, but also uses a novel and robust point cloud generation backbone in order to address the prevalent issues. This involves using a single-encoder multiple-decoder deep network architecture wherein each decoder generates certain fixed viewpoints. This is followed by fusing all the viewpoints to generate a dense point cloud. Various experiments are conducted on the technique and its performance is compared with those of other state of the art techniques and impressive gains in performance are demonstrated. Code is available at https://github.com/mueedhafiz1982/ | ['M. Hassaballah', 'Shabir Ahmad Parah', 'Rouf Ul Alam Bhat', 'Abdul Mueed Hafiz'] | 2021-06-17 | null | null | null | null | ['point-cloud-generation'] | ['computer-vision'] | [ 1.71646494e-02 -2.27004766e-01 2.92667389e-01 -2.96747088e-01
-5.15829921e-01 -4.20729488e-01 5.19933641e-01 -3.31645042e-01
-1.31523445e-01 5.00920713e-01 -1.50122896e-01 -1.76739320e-01
1.44002825e-01 -9.10808980e-01 -8.14564943e-01 -5.85655749e-01
3.35663468e-01 4.66418952e-01 4.23208177e-01 -4.37696755e-01
4.08016413e-01 8.65760505e-01 -1.59933114e+00 4.84031141e-02
4.62593853e-01 1.04948616e+00 4.29619908e-01 5.73324800e-01
-3.43602389e-01 3.66919249e-01 -3.82282972e-01 -3.44146252e-01
8.63835812e-01 -3.21007073e-01 -4.69715148e-01 2.03375444e-01
4.93548363e-01 -5.72474599e-01 -2.08343655e-01 9.78062868e-01
7.48388231e-01 -1.36294737e-01 4.38993335e-01 -1.56845737e+00
-5.94627261e-01 -6.48286343e-02 -1.00304174e+00 -2.70115416e-02
1.96494132e-01 7.34988600e-02 3.84371281e-01 -9.30305898e-01
4.85102713e-01 1.28485715e+00 7.95637906e-01 5.43751061e-01
-7.04291761e-01 -9.26466823e-01 -2.55550534e-01 1.81450412e-01
-1.61551642e+00 -3.30850571e-01 1.05210829e+00 -2.78686494e-01
1.15985572e+00 2.67626578e-03 9.27746356e-01 6.79076850e-01
2.47264609e-01 2.34950483e-01 1.14779222e+00 -5.68482578e-01
1.67509124e-01 -2.97049303e-02 -1.85702294e-01 8.12434494e-01
2.75854468e-01 2.69294590e-01 -4.57563221e-01 4.23360988e-02
1.27067006e+00 3.08914393e-01 -4.18081461e-03 -3.46656621e-01
-1.04419076e+00 6.40617371e-01 7.74621367e-01 1.73062552e-02
-5.23563862e-01 4.16852474e-01 1.55230120e-01 1.73064917e-01
3.80204707e-01 -2.04045787e-01 -3.77203226e-01 -1.03316449e-01
-9.13597167e-01 2.57612735e-01 4.93627906e-01 1.50022459e+00
8.65232170e-01 2.01261312e-01 5.15714526e-01 3.98439586e-01
7.87757516e-01 6.58068955e-01 1.71474874e-01 -1.13675380e+00
6.46276176e-01 9.25180674e-01 -1.35859460e-01 -1.11539912e+00
-2.01936781e-01 -1.83737114e-01 -1.03983831e+00 7.56534040e-01
-2.35497430e-01 -8.59296322e-02 -1.24113762e+00 1.14977837e+00
6.75981104e-01 2.34293222e-01 1.01010017e-01 9.94091630e-01
9.34144437e-01 8.68722320e-01 -3.69936824e-01 1.82118490e-01
1.19180715e+00 -8.73706877e-01 -2.47812480e-01 2.19626166e-02
1.31655797e-01 -1.07977772e+00 6.79798067e-01 3.90935212e-01
-1.20669246e+00 -6.58817410e-01 -1.01957965e+00 -4.95815545e-01
-3.74276817e-01 -8.35350342e-03 5.02394080e-01 5.78563511e-01
-1.45793939e+00 3.17817420e-01 -9.95465636e-01 -4.04964626e-01
5.29406905e-01 6.44053400e-01 -4.63776588e-01 -1.82247311e-01
-6.68386579e-01 9.43702400e-01 4.30219710e-01 3.26854050e-01
-6.36710405e-01 -4.68121409e-01 -7.65906990e-01 -2.63699025e-01
-7.80630037e-02 -1.21515238e+00 1.15138388e+00 -8.21006775e-01
-1.70356929e+00 6.76330805e-01 -1.78477526e-01 -2.11423218e-01
5.00566959e-01 -6.03616312e-02 3.67891565e-02 2.48323157e-01
-1.30431056e-02 1.00955021e+00 7.10546672e-01 -1.48309672e+00
-7.19527304e-01 -5.75962424e-01 1.14781179e-01 4.54954147e-01
2.31992528e-01 -6.71350360e-02 -5.42724609e-01 -1.89182013e-01
5.19117892e-01 -9.83583033e-01 -3.23873341e-01 3.53211224e-01
-3.31625313e-01 -4.75978181e-02 1.07227218e+00 -3.80369753e-01
5.90394795e-01 -2.07938695e+00 -3.03266440e-02 1.17187850e-01
3.29469979e-01 2.61098355e-01 2.08720118e-01 6.75122440e-01
1.15278158e-02 1.54087141e-01 -1.98685303e-01 -5.77688813e-01
-1.38150454e-01 1.61770165e-01 -1.80392593e-01 5.48316658e-01
1.44962326e-01 6.71076775e-01 -4.72608447e-01 -5.78453720e-01
6.78021193e-01 1.04883683e+00 -4.51363862e-01 8.26531202e-02
-9.78231505e-02 4.38541949e-01 -2.79209495e-01 8.52644920e-01
1.14706552e+00 -2.06245661e-01 -3.56180996e-01 -4.74151969e-01
-4.15966362e-01 -6.02501035e-02 -1.29704416e+00 1.99768019e+00
-3.07290584e-01 3.21704417e-01 -4.11982797e-02 -6.78712428e-01
1.05432820e+00 4.49321091e-01 3.75282139e-01 -5.01426697e-01
3.99242580e-01 3.49905252e-01 -1.80143207e-01 -2.99880445e-01
3.92898500e-01 -4.13518578e-01 2.33752742e-01 2.52151042e-01
-2.40080357e-02 -5.48779666e-01 -2.00935766e-01 7.81886876e-02
8.68451178e-01 6.62995636e-01 1.44535685e-02 2.28602037e-01
2.87529528e-01 4.33090925e-01 3.21246117e-01 1.90845892e-01
-3.72958668e-02 1.00218618e+00 -3.02031133e-02 -4.45386350e-01
-1.46100318e+00 -1.07101011e+00 1.53143570e-01 1.18883431e-01
4.78798747e-01 -3.20937335e-01 -3.27061802e-01 -2.11037144e-01
-2.43246675e-01 5.47886968e-01 -3.02205950e-01 1.20105885e-01
-6.24532700e-01 -4.71127093e-01 3.62152636e-01 4.85870034e-01
9.34715271e-01 -8.05173337e-01 -8.58329892e-01 -3.19635645e-02
6.00413121e-02 -1.06828880e+00 1.69984072e-01 1.39103383e-02
-1.46145296e+00 -9.21180964e-01 -5.93649387e-01 -7.13814616e-01
8.48761976e-01 7.97303319e-01 1.10593843e+00 1.94943726e-01
-8.58001485e-02 1.23024218e-01 -4.16944146e-01 -7.94293761e-01
-1.52619988e-01 -1.48783341e-01 -9.62913707e-02 -4.41916794e-01
5.19377768e-01 -7.69048095e-01 -7.76134670e-01 -4.43449914e-02
-9.39712107e-01 4.56353605e-01 8.19508016e-01 4.25301611e-01
8.76079500e-01 8.27570558e-02 9.75707695e-02 -4.59772408e-01
2.68179208e-01 -3.66491288e-01 -8.30009758e-01 -1.33724555e-01
-4.54408854e-01 -2.17842907e-01 4.98170614e-01 2.30929241e-01
-9.30661738e-01 4.22504336e-01 -3.03579599e-01 -9.20647323e-01
-3.39731246e-01 2.87092347e-02 -1.49192009e-02 -2.79747695e-01
3.49017888e-01 3.42441112e-01 2.61229128e-01 -4.84477937e-01
3.94127458e-01 7.13781059e-01 1.72353715e-01 -1.45229608e-01
1.08865488e+00 6.85232937e-01 2.65551329e-01 -6.45831466e-01
-3.62522960e-01 -4.53187078e-01 -8.47151160e-01 -2.50610530e-01
8.43576550e-01 -1.10216200e+00 -6.00500166e-01 4.79808271e-01
-1.59812808e+00 6.40235692e-02 -5.88091165e-02 4.70635325e-01
-5.64311922e-01 3.18161130e-01 -4.84792560e-01 -7.05905080e-01
-6.93707049e-01 -1.28492653e+00 1.32057476e+00 3.20343137e-01
2.14298517e-01 -5.99043548e-01 3.46770771e-02 3.33674282e-01
4.19229358e-01 5.41927338e-01 6.85729742e-01 2.12992597e-02
-1.20936894e+00 -2.39819288e-01 -3.75958353e-01 3.72524142e-01
1.58517182e-01 1.83220014e-01 -7.71283984e-01 -3.74995396e-02
2.55730361e-01 -1.30655691e-01 2.00100392e-01 2.47636229e-01
7.06136405e-01 -6.11833073e-02 -3.17641050e-01 8.31349313e-01
2.11220765e+00 3.61854613e-01 6.90132618e-01 4.54038262e-01
9.72501457e-01 1.23008929e-01 2.46866614e-01 2.98324138e-01
6.78762138e-01 4.40526754e-01 9.04084623e-01 -1.38883233e-01
-2.20994100e-01 -1.09164193e-01 8.58903676e-02 1.29216766e+00
-4.59728479e-01 -1.21659547e-01 -9.84152317e-01 4.01680082e-01
-1.57865059e+00 -7.68486023e-01 -4.57494766e-01 2.02882743e+00
5.03686011e-01 -5.69084287e-02 -1.03759237e-01 2.58360237e-01
6.96212471e-01 -8.94307494e-02 -3.70958626e-01 -3.65815073e-01
1.58581689e-01 4.27987397e-01 5.93008816e-01 3.44232887e-01
-6.70793831e-01 9.13054705e-01 5.85147667e+00 4.25671637e-01
-1.27924168e+00 2.88743466e-01 2.40744352e-01 -2.12981462e-01
-6.19642660e-02 4.47193757e-02 -8.31605792e-01 5.19663274e-01
7.11045384e-01 1.09456852e-01 2.15713874e-01 8.00166368e-01
2.26389408e-01 -8.19007009e-02 -7.99762130e-01 1.38654149e+00
2.46141672e-01 -1.42332387e+00 2.10389718e-01 1.02946728e-01
6.41636014e-01 4.85939503e-01 -1.85061678e-01 -1.22414045e-01
2.49170572e-01 -8.06610048e-01 7.66483426e-01 5.16075313e-01
7.26112783e-01 -8.64317179e-01 7.88160443e-01 5.45596540e-01
-1.04878414e+00 2.44588181e-01 -6.62154973e-01 -2.14032024e-01
3.70596707e-01 4.89875257e-01 -8.69935274e-01 7.95682132e-01
9.21642840e-01 5.88899910e-01 -4.23997641e-01 1.22746718e+00
8.94662365e-03 1.47422880e-01 -8.21992993e-01 7.09597170e-02
4.08207446e-01 -3.65080059e-01 3.25317234e-01 6.99434936e-01
8.47434700e-01 3.33741337e-01 -1.79835051e-01 8.52726221e-01
5.19563109e-02 -6.45396709e-02 -1.05180562e+00 4.82869536e-01
4.53510344e-01 1.18383765e+00 -8.26327384e-01 -3.40964258e-01
-5.41853786e-01 9.24241006e-01 4.81262282e-02 -2.14605872e-02
-8.09742033e-01 -3.91303122e-01 2.57354677e-01 1.83437154e-01
3.43451142e-01 -7.47069716e-01 -3.91969442e-01 -8.64918649e-01
-2.69013760e-03 -5.97048342e-01 -1.39067665e-01 -1.32319510e+00
-1.05839539e+00 1.02941597e+00 1.86883241e-01 -1.76337731e+00
-1.67565018e-01 -5.74653327e-01 -4.34555948e-01 1.12426996e+00
-1.68146157e+00 -1.44479525e+00 -7.90412784e-01 9.43473220e-01
6.83243334e-01 -1.49058059e-01 5.46032012e-01 4.91502315e-01
-2.47464836e-01 9.43063274e-02 -8.93420503e-02 -1.47343591e-01
2.74493068e-01 -9.63570952e-01 5.96248090e-01 9.60690320e-01
-3.70387323e-02 4.88156050e-01 3.12765479e-01 -7.67067790e-01
-1.85967565e+00 -1.10136306e+00 6.80703938e-01 -4.64148104e-01
3.24008614e-02 -2.47126758e-01 -4.85412359e-01 6.51169717e-01
4.78951156e-01 2.01314706e-02 5.39165974e-01 -5.95156729e-01
-6.66133463e-02 -1.41052589e-01 -1.34507322e+00 4.16708767e-01
1.04981184e+00 -1.14373393e-01 -4.41217750e-01 3.16131145e-01
7.44903445e-01 -8.26745331e-01 -5.66198826e-01 2.27872446e-01
2.62399882e-01 -1.31590521e+00 9.81785178e-01 2.04854105e-02
7.26838708e-01 -7.15818644e-01 -4.42461073e-01 -1.07522404e+00
-3.12072694e-01 -3.34348202e-01 -1.32222459e-01 9.57856178e-01
7.18477815e-02 -5.57782054e-01 1.06455326e+00 4.96203393e-01
-4.40170407e-01 -7.96861470e-01 -9.73479152e-01 -4.31706458e-01
-1.96342960e-01 -3.93703610e-01 8.13032269e-01 7.25243032e-01
-7.80957222e-01 4.83567089e-01 -2.48651430e-01 4.56211120e-01
8.11015069e-01 2.17399389e-01 9.64170277e-01 -1.09488320e+00
1.02945380e-01 -8.45247060e-02 -7.63384700e-01 -9.26289558e-01
-4.57933962e-01 -8.15237284e-01 -1.80095017e-01 -2.00842619e+00
-2.17271999e-01 -6.98287070e-01 2.76326239e-01 3.10128838e-01
4.06882703e-01 6.31932497e-01 5.66938758e-01 4.39321488e-01
-2.29903013e-01 4.45434988e-01 1.33384573e+00 1.96624056e-01
-5.74749932e-02 -6.57412037e-02 -4.80340391e-01 6.46989524e-01
9.18177366e-01 -5.14717877e-01 -6.87691689e-01 -1.10934436e+00
2.74223179e-01 6.94133900e-03 6.27096891e-01 -1.34634018e+00
3.19700450e-01 7.81309754e-02 7.26920366e-01 -1.26034784e+00
8.71620178e-01 -1.41362286e+00 7.18337834e-01 3.43665808e-01
5.11191070e-01 8.00520480e-01 1.91672236e-01 4.71475303e-01
-2.09222645e-01 -2.34265536e-01 6.03974223e-01 -6.66771352e-01
-7.72581100e-01 5.00228703e-01 2.30857849e-01 -5.20295680e-01
1.29188335e+00 -8.06684017e-01 -9.95781049e-02 -1.00322597e-01
-3.21482956e-01 -1.44278094e-01 7.46427000e-01 3.04746211e-01
9.06027794e-01 -1.46571672e+00 -6.10076904e-01 3.84320438e-01
-2.43680894e-01 7.84868538e-01 1.97645500e-01 5.05632699e-01
-1.39753699e+00 4.19363588e-01 -5.62258244e-01 -9.40497398e-01
-1.12826574e+00 4.42015707e-01 2.43240073e-01 2.29666546e-01
-8.42023611e-01 7.88369417e-01 -1.55869946e-01 -5.16020894e-01
-4.04303521e-03 -3.90870422e-01 1.52631059e-01 -3.69477481e-01
9.30732414e-02 3.36445093e-01 2.11746618e-01 -9.52437818e-01
-3.20187777e-01 1.02697027e+00 1.95038885e-01 -8.31100941e-02
1.45182753e+00 -2.90888965e-01 -2.01682240e-01 2.92888045e-01
1.29377258e+00 -3.83356243e-01 -1.17824173e+00 1.52232801e-03
-5.20898640e-01 -7.69286036e-01 1.92824453e-01 -6.25517309e-01
-1.43242502e+00 1.15362597e+00 8.82125139e-01 7.46896565e-02
1.21464026e+00 -2.31601670e-01 1.00292313e+00 1.21107757e-01
7.39450991e-01 -6.34131372e-01 -2.52283275e-01 4.35937077e-01
8.92767608e-01 -1.22653210e+00 1.81448996e-01 -4.82034385e-01
-3.91215444e-01 1.23922253e+00 7.40906239e-01 -4.35672998e-01
8.28520000e-01 3.82948905e-01 2.68024534e-01 -4.94947195e-01
-5.76670349e-01 1.34261355e-01 -2.75256783e-01 7.33722150e-01
2.74004966e-01 -2.32329279e-01 3.64332385e-02 -7.09346756e-02
-4.97304887e-01 3.84241551e-01 4.64512974e-01 1.27891386e+00
-1.50097802e-01 -1.08970368e+00 -5.79221904e-01 3.75448227e-01
-3.23777497e-01 -1.65525138e-01 -1.32421508e-01 9.03568745e-01
3.59190047e-01 6.61721230e-01 2.93781102e-01 -4.67219889e-01
3.47138733e-01 -1.82879180e-01 6.13212645e-01 -5.77938557e-01
-5.37587762e-01 1.00328743e-01 -3.81269902e-01 -4.76339847e-01
-8.44191492e-01 -3.97853851e-01 -1.12100697e+00 -5.98039865e-01
-3.91855747e-01 -1.45094350e-01 1.14040053e+00 6.16829991e-01
8.02059710e-01 2.37962604e-01 5.75182319e-01 -1.47823071e+00
-2.69465536e-01 -8.06585789e-01 -2.82240480e-01 1.20492496e-01
3.07719499e-01 -7.63863385e-01 -1.04997471e-01 2.68586904e-01] | [8.293181419372559, -3.242771625518799] |
95acce13-76b5-4033-ac06-1e7c01398656 | pruning-the-way-to-reliable-policies-a-multi | 2306.08044 | null | https://arxiv.org/abs/2306.08044v1 | https://arxiv.org/pdf/2306.08044v1.pdf | Pruning the Way to Reliable Policies: A Multi-Objective Deep Q-Learning Approach to Critical Care | Most medical treatment decisions are sequential in nature. Hence, there is substantial hope that reinforcement learning may make it possible to formulate precise data-driven treatment plans. However, a key challenge for most applications in this field is the sparse nature of primarily mortality-based reward functions, leading to decreased stability of offline estimates. In this work, we introduce a deep Q-learning approach able to obtain more reliable critical care policies. This method integrates relevant but noisy intermediate biomarker signals into the reward specification, without compromising the optimization of the main outcome of interest (e.g. patient survival). We achieve this by first pruning the action set based on all available rewards, and second training a final model based on the sparse main reward but with a restricted action set. By disentangling accurate and approximated rewards through action pruning, potential distortions of the main objective are minimized, all while enabling the extraction of valuable information from intermediate signals that can guide the learning process. We evaluate our method in both off-policy and offline settings using simulated environments and real health records of patients in intensive care units. Our empirical results indicate that pruning significantly reduces the size of the action space while staying mostly consistent with the actions taken by physicians, outperforming the current state-of-the-art offline reinforcement learning method conservative Q-learning. Our work is a step towards developing reliable policies by effectively harnessing the wealth of available information in data-intensive critical care environments. | ['Ahmed Alaa', 'Alexander Schubert', 'Ali Shirali'] | 2023-06-13 | null | null | null | null | ['q-learning'] | ['methodology'] | [ 3.65499139e-01 3.54510784e-01 -5.30798852e-01 7.36440299e-03
-9.74329233e-01 -3.04350793e-01 1.35255426e-01 6.75136805e-01
-7.40504384e-01 1.22741199e+00 3.27319086e-01 -3.90304893e-01
-6.61219120e-01 -6.52962089e-01 -5.22387981e-01 -7.89817929e-01
-4.06756163e-01 7.24027991e-01 -2.44230554e-01 -3.22080478e-02
1.16695866e-01 3.36819470e-01 -1.01470876e+00 1.53458908e-01
9.99405146e-01 1.02144027e+00 1.64033145e-01 6.52674317e-01
3.16621214e-01 9.22241449e-01 -4.12875831e-01 -1.84488580e-01
3.66631031e-01 -6.31217241e-01 -4.72621828e-01 -8.30511525e-02
-2.25566626e-01 -5.84043682e-01 -1.40744358e-01 6.95000887e-01
7.97629416e-01 2.49530405e-01 3.69404763e-01 -7.13345826e-01
1.41233206e-01 6.73079967e-01 -1.36244216e-03 5.13883792e-02
7.66987801e-02 6.80133104e-01 1.04799724e+00 -8.96533579e-02
5.63834012e-01 1.07589769e+00 4.92023975e-01 6.43050969e-01
-1.43695664e+00 -2.78040767e-01 2.09518746e-01 -2.91045886e-02
-8.96590352e-01 -3.09431523e-01 3.89384925e-01 -3.66509050e-01
7.97099590e-01 2.52865404e-01 9.28931594e-01 1.00045288e+00
4.71073329e-01 6.46353602e-01 1.17583787e+00 -2.31045216e-01
6.40228808e-01 6.33065253e-02 -3.27228665e-01 6.25128031e-01
1.59752235e-01 7.23686635e-01 -2.74411798e-01 -4.49547380e-01
7.85745084e-01 3.27745557e-01 -2.99037963e-01 -4.03224498e-01
-1.09012032e+00 9.20502245e-01 3.26466918e-01 5.18037081e-02
-1.01443756e+00 3.44776720e-01 2.95447290e-01 2.00530574e-01
2.14496538e-01 8.37909698e-01 -5.94953537e-01 -5.49514115e-01
-9.34373081e-01 3.55356842e-01 6.48721755e-01 3.20271760e-01
2.88913071e-01 1.02893777e-01 -7.43694782e-01 5.29381692e-01
9.63113680e-02 5.40317535e-01 3.87468040e-01 -1.21309161e+00
4.65370744e-01 4.13746238e-01 5.27031362e-01 -5.02729118e-01
-5.98054707e-01 -7.76107550e-01 -5.79372287e-01 2.87511945e-01
5.94386756e-01 -7.82932520e-01 -7.71318078e-01 1.72691715e+00
3.49194229e-01 8.81717633e-03 1.87988058e-02 8.27013016e-01
-5.97560890e-02 1.60354137e-01 3.97419751e-01 -6.68502092e-01
1.18905473e+00 -4.90154505e-01 -5.44061601e-01 -1.57113567e-01
5.94324410e-01 -1.07607171e-01 7.38935888e-01 5.85120678e-01
-1.24513805e+00 -2.05738410e-01 -6.31551147e-01 6.30354166e-01
2.35938817e-01 -1.26476698e-02 5.66414118e-01 5.10536849e-01
-7.24257648e-01 1.22055507e+00 -1.10255313e+00 6.63422793e-02
8.75549614e-01 7.46813953e-01 4.55772020e-02 -1.25702098e-01
-1.18785667e+00 9.40384150e-01 2.20132649e-01 2.91393176e-02
-1.23997426e+00 -9.02172029e-01 -5.53159416e-01 3.85347396e-01
8.99092257e-01 -9.42738295e-01 1.38007700e+00 -9.29789901e-01
-1.56201565e+00 4.43346612e-02 2.36663133e-01 -1.02596915e+00
9.66173351e-01 -2.03909725e-01 8.73353556e-02 3.48338902e-01
-2.39501789e-01 6.81041658e-01 9.26088333e-01 -8.47456396e-01
-6.32205069e-01 -3.19054365e-01 -3.97168733e-02 2.93495774e-01
-2.16145292e-01 -3.73040140e-01 1.21184804e-01 -4.42426056e-01
-5.65912068e-01 -8.61239851e-01 -8.77221465e-01 -1.25306651e-01
-1.64842486e-01 1.25060067e-01 1.29989237e-01 -6.62828445e-01
1.27246332e+00 -1.86214268e+00 1.83218956e-01 2.93938369e-01
1.75186336e-01 2.10811824e-01 -6.79160208e-02 5.98033547e-01
2.14139521e-01 -5.40356338e-02 -3.17263216e-01 -1.91002131e-01
-2.92727530e-01 4.31587994e-01 -2.29625806e-01 4.25109863e-01
4.16056782e-01 1.04978538e+00 -1.17869473e+00 -3.31865400e-01
2.22994924e-01 2.29642272e-01 -8.20306480e-01 3.95600140e-01
-4.92202759e-01 9.04315174e-01 -7.94619739e-01 5.21784604e-01
-4.85974504e-03 -1.35593921e-01 3.94222081e-01 3.89419675e-01
8.34036320e-02 1.64304748e-01 -1.02259266e+00 1.31243622e+00
-5.66312671e-01 -8.01791251e-02 7.80891255e-02 -1.37866163e+00
6.17813945e-01 3.44845474e-01 1.14313829e+00 -6.71491086e-01
3.04579705e-01 6.75401762e-02 3.55929345e-01 -5.02821267e-01
-8.46275687e-02 -5.82958817e-01 8.42181519e-02 2.53950089e-01
-2.40985211e-02 -2.37884015e-01 8.32444131e-02 -7.26219863e-02
1.36050284e+00 8.71154219e-02 4.25195634e-01 -1.66997716e-01
2.43601575e-01 2.01679002e-02 7.11338103e-01 9.89056587e-01
-2.74205059e-01 2.89842993e-01 9.59739447e-01 -2.22289488e-01
-7.50901163e-01 -8.77112329e-01 -5.78684583e-02 5.47172368e-01
-3.54316771e-01 2.04441454e-02 -4.44185555e-01 -8.99980426e-01
4.25545245e-01 7.77287781e-01 -8.11147332e-01 -4.82297868e-01
-5.90424478e-01 -8.38309050e-01 1.53703004e-01 4.13400233e-01
-1.31368175e-01 -1.22556841e+00 -1.21373963e+00 8.07499826e-01
1.91067353e-01 -7.74652123e-01 -3.50411296e-01 5.14779031e-01
-1.30081189e+00 -1.14954770e+00 -9.13467944e-01 1.24307700e-01
6.27061725e-01 -3.81555170e-01 1.05356765e+00 -1.18004292e-01
-4.05096978e-01 4.90715176e-01 -1.64326027e-01 -5.56457400e-01
-4.85369623e-01 -1.01958498e-01 -2.43761465e-02 -5.06553128e-02
-1.14096940e-01 -3.02333713e-01 -9.30612087e-01 -1.02166317e-01
-7.63737798e-01 -2.58444488e-01 7.64256895e-01 1.24782288e+00
7.10750997e-01 -1.45996988e-01 9.53012824e-01 -1.06144667e+00
8.06453884e-01 -6.26644135e-01 -8.32256675e-01 7.25083351e-02
-1.05717826e+00 4.29553330e-01 9.29801166e-01 -5.34765065e-01
-8.22901845e-01 1.46736011e-01 3.01453918e-02 -4.81579334e-01
8.99313092e-02 5.68678617e-01 3.57746929e-01 2.92252183e-01
6.31988525e-01 6.11953065e-02 3.24324280e-01 -1.84082299e-01
9.86170769e-02 3.38826418e-01 1.19085483e-01 -6.83238924e-01
1.46136478e-01 2.98579395e-01 4.19337243e-01 -4.73875999e-01
-7.37281799e-01 -3.76084477e-01 -1.06849447e-01 -1.96272001e-01
5.07011056e-01 -7.62691855e-01 -1.07815862e+00 -6.58971220e-02
-4.32206303e-01 -6.64751649e-01 -1.02206123e+00 7.19663143e-01
-9.74509835e-01 8.70299786e-02 -3.76676679e-01 -1.15858555e+00
-3.60777348e-01 -1.30866683e+00 9.05958414e-01 7.65419379e-02
-1.68470219e-01 -7.86376595e-01 3.61615717e-01 1.51156455e-01
2.65239060e-01 3.72439772e-01 8.62668812e-01 -5.90848804e-01
-4.74189818e-01 -1.76329687e-02 2.99267560e-01 3.85385394e-01
1.53878555e-01 -3.08530301e-01 -7.20360041e-01 -3.47296089e-01
1.57594718e-02 -4.36290145e-01 9.06793654e-01 9.48525786e-01
1.09603310e+00 -4.13493931e-01 -2.20730856e-01 3.46596986e-01
1.39077616e+00 4.48844492e-01 3.19634587e-01 8.57917368e-02
1.37806624e-01 6.24197900e-01 9.38674808e-01 9.99268234e-01
2.24100471e-01 4.69375879e-01 5.94879091e-01 -3.37527357e-02
3.03067297e-01 -2.49892458e-01 2.57509410e-01 7.13123828e-02
-7.89350644e-02 -3.05632353e-02 -8.10907602e-01 5.12582123e-01
-2.09657383e+00 -8.81320119e-01 4.85158980e-01 2.55991936e+00
1.13770390e+00 3.08069110e-01 4.47190791e-01 -4.16033715e-02
1.42480448e-01 -3.39306951e-01 -9.55216587e-01 -5.49904168e-01
3.57962996e-01 4.70536500e-01 7.69038856e-01 4.45115447e-01
-8.98234427e-01 4.59985048e-01 5.98323011e+00 3.64795357e-01
-1.07097018e+00 -2.81384408e-01 1.07043910e+00 -3.98015559e-01
-7.33711347e-02 -1.68064222e-01 -4.86521751e-01 3.91708106e-01
1.37177372e+00 -6.12026863e-02 6.53623462e-01 7.49547899e-01
9.64475870e-01 -4.31514174e-01 -1.17311525e+00 4.72225010e-01
-5.56872845e-01 -1.27272165e+00 -4.48904783e-01 1.49794951e-01
7.46168673e-01 -1.09921293e-02 -1.49374893e-02 4.35039580e-01
5.16323090e-01 -1.13128817e+00 4.96767670e-01 8.61747503e-01
7.33992636e-01 -9.46440101e-01 8.90375912e-01 5.55002987e-01
-4.86183643e-01 -6.51693642e-01 -2.34248843e-02 -4.59241681e-02
5.78318238e-02 7.57772803e-01 -1.17435431e+00 4.31011647e-01
2.34982267e-01 6.10802174e-01 -7.04300329e-02 1.11110878e+00
-9.98496935e-02 8.66976321e-01 -2.48303354e-01 -1.22606188e-01
3.30002844e-01 -1.69504806e-01 3.56595665e-01 8.46771896e-01
1.76244408e-01 2.54805833e-01 4.03585225e-01 7.16075480e-01
-2.92135235e-02 6.29808428e-03 -4.89300221e-01 -2.72085786e-01
2.21244305e-01 9.70394135e-01 -5.43185174e-01 -2.90815294e-01
-1.04319118e-01 4.16630358e-01 2.49987110e-01 1.86385944e-01
-6.49563372e-01 6.20008670e-02 5.97400188e-01 6.85589239e-02
3.37095708e-01 -2.37404332e-02 -3.89379591e-01 -8.90559733e-01
-2.99415231e-01 -1.17045033e+00 5.66790819e-01 -5.64738959e-02
-9.81658518e-01 2.04028800e-01 -3.41345407e-02 -1.13196552e+00
-7.96810508e-01 -2.95086533e-01 -2.91980803e-01 9.75907326e-01
-1.58034790e+00 -3.84357691e-01 2.42560819e-01 3.18637818e-01
3.93155187e-01 6.67241290e-02 7.65698671e-01 2.12778710e-03
-5.79456151e-01 3.26324016e-01 3.66508365e-01 -3.77925754e-01
5.04810929e-01 -1.27063751e+00 -2.70288199e-01 2.80918449e-01
-1.54834062e-01 2.10666478e-01 6.36428833e-01 -7.79304266e-01
-1.41948330e+00 -8.75264227e-01 2.06042737e-01 -1.98709011e-01
4.96899575e-01 1.07014263e-02 -5.98763645e-01 8.65151361e-02
-2.81396031e-01 -1.96701996e-02 3.87028962e-01 -1.54757172e-01
4.83772397e-01 -2.38193125e-01 -1.28779507e+00 5.59104145e-01
5.33535540e-01 -4.11265939e-02 -3.16696495e-01 2.14185357e-01
4.77712810e-01 -3.01001340e-01 -9.23838019e-01 4.03877735e-01
4.83325392e-01 -6.78070486e-01 9.14380491e-01 -8.98928642e-01
5.27313888e-01 2.41257980e-01 2.57283330e-01 -1.52857184e+00
-1.03671923e-01 -8.15910101e-01 -4.49283570e-01 5.93570650e-01
3.26938927e-01 -5.72608709e-01 8.93257916e-01 7.46728897e-01
1.47857055e-01 -1.34268498e+00 -1.00435960e+00 -5.35708368e-01
6.90102801e-02 -1.92448393e-01 2.75943637e-01 5.55779576e-01
5.07005677e-02 2.14213543e-02 -4.25681621e-01 -1.95892841e-01
5.35431206e-01 7.12233931e-02 3.52051735e-01 -9.76707101e-01
-8.81674528e-01 -4.79336023e-01 2.29584090e-02 -6.34277284e-01
-2.01807335e-01 -5.20376563e-01 2.14345366e-01 -1.41220975e+00
6.55690134e-02 -6.44412816e-01 -6.33667231e-01 5.95806956e-01
-4.26133305e-01 -4.48438436e-01 2.66395241e-01 -9.41937417e-02
-2.68226951e-01 7.62425661e-01 1.51460481e+00 -9.07928497e-02
-5.41915059e-01 4.52384681e-01 -7.23162293e-01 3.90758902e-01
8.72035086e-01 -7.72921503e-01 -5.70317745e-01 1.39174685e-01
-1.23450525e-01 9.35340643e-01 2.20775425e-01 -7.12818623e-01
-2.32237712e-01 -7.69291818e-01 3.77761394e-01 -1.19018003e-01
1.63764670e-01 -9.11442101e-01 -1.68423951e-01 1.13307345e+00
-6.02287948e-01 -1.43332705e-01 2.93331861e-01 7.29290247e-01
2.56662462e-02 -1.68338850e-01 7.97127903e-01 -3.11863393e-01
-4.40844521e-02 2.88199365e-01 -2.61296391e-01 3.66269350e-01
1.04174399e+00 1.52123749e-01 1.19871221e-01 -5.72015882e-01
-8.90160918e-01 5.38955629e-01 -3.43761290e-03 3.63795012e-02
4.80577230e-01 -7.51332760e-01 -8.78768265e-01 5.85900731e-02
-3.44954610e-01 -1.04986377e-01 1.90346286e-01 9.96009350e-01
-1.77112341e-01 5.43479145e-01 -1.57243446e-01 -3.43745887e-01
-8.57589304e-01 5.71504831e-01 5.52518845e-01 -7.91484356e-01
-5.30893862e-01 3.51826012e-01 -7.18018785e-02 2.53430735e-02
2.78209358e-01 -4.96073693e-01 -1.95234105e-01 1.02421947e-01
3.96064579e-01 5.28608739e-01 1.49282023e-01 8.56728479e-02
-2.54166901e-01 9.37698316e-03 3.34587656e-02 -2.33998686e-01
1.61621833e+00 2.00116456e-01 4.48910326e-01 1.30360082e-01
7.18130767e-01 -1.71358064e-01 -1.71958423e+00 6.56857416e-02
9.46268439e-02 -3.82375240e-01 2.78379351e-01 -1.19001591e+00
-9.10422683e-01 7.80672073e-01 7.52637565e-01 -3.76673415e-02
1.33305585e+00 -3.30941349e-01 6.08397961e-01 3.83027196e-01
3.63687277e-01 -1.17173064e+00 -2.31687538e-02 6.82777017e-02
6.66864276e-01 -1.21127975e+00 1.65420562e-01 2.64639914e-01
-9.74144220e-01 8.87806952e-01 3.36880773e-01 -2.15019748e-01
4.78349358e-01 1.93317786e-01 -8.23354349e-03 4.27744910e-02
-1.12453997e+00 -2.53151566e-01 8.81134719e-02 3.69834542e-01
3.83312516e-02 2.90298402e-01 -4.65467244e-01 5.94348788e-01
3.35261643e-01 3.17472160e-01 5.38698316e-01 1.00324130e+00
-3.47612441e-01 -1.38141751e+00 -2.79579252e-01 9.40401256e-01
-7.15026557e-01 -1.64444029e-01 -8.51221010e-03 5.72069287e-01
-1.45380888e-02 7.31258035e-01 -2.92813152e-01 9.52947959e-02
5.39085865e-01 -4.55697849e-02 4.48050976e-01 -6.57474816e-01
-8.14620972e-01 2.35847205e-01 1.42476335e-01 -8.65617752e-01
-7.17862917e-04 -8.42580020e-01 -1.25958967e+00 2.08054274e-01
-1.74767241e-01 2.09330931e-01 4.72529858e-01 8.75487506e-01
4.81875181e-01 9.07876611e-01 8.28877628e-01 -5.83016038e-01
-1.43410385e+00 -6.29236341e-01 -4.04976219e-01 2.02971950e-01
7.47715712e-01 -6.81531906e-01 -5.21639176e-02 -4.04433459e-01] | [4.018851280212402, 2.730570077896118] |
f8e129e6-9ec4-4976-8168-462edc86ff61 | video-super-resolution-via-deep-draft | null | null | http://openaccess.thecvf.com/content_iccv_2015/html/Liao_Video_Super-Resolution_via_ICCV_2015_paper.html | http://openaccess.thecvf.com/content_iccv_2015/papers/Liao_Video_Super-Resolution_via_ICCV_2015_paper.pdf | Video Super-Resolution via Deep Draft-Ensemble Learning | We propose a new direction for fast video super-resolution (VideoSR) via a SR draft ensemble, which is defined as the set of high-resolution patch candidates before final image deconvolution. Our method contains two main components -- i.e., SR draft ensemble generation and its optimal reconstruction. The first component is to renovate traditional feedforward reconstruction pipeline and greatly enhance its ability to compute different super resolution results considering large motion variation and possible errors arising in this process. Then we combine SR drafts through the nonlinear process in a deep convolutional neural network (CNN). We analyze why this framework is proposed and explain its unique advantages compared to previous iterative methods to update different modules in passes. Promising experimental results are shown on natural video sequences. | ['Ziyang Ma', 'Ruiyu Li', 'Xin Tao', 'Jiaya Jia', 'Renjie Liao'] | 2015-12-01 | null | null | null | iccv-2015-12 | ['image-deconvolution'] | ['computer-vision'] | [ 4.70558971e-01 -1.63448572e-01 1.58431321e-01 -1.20348014e-01
-8.51829886e-01 -1.54756382e-01 3.39078844e-01 -9.91636157e-01
-1.71222731e-01 1.01362419e+00 4.33590025e-01 2.10130453e-01
-8.08222219e-02 -7.21127629e-01 -9.30190623e-01 -6.84633911e-01
4.27861363e-02 -9.60082412e-02 5.05383372e-01 -3.62570792e-01
3.85707378e-01 5.72635591e-01 -1.51217639e+00 7.12310612e-01
9.35956955e-01 7.55366266e-01 6.51505053e-01 6.42062604e-01
8.94100312e-03 1.07137311e+00 -3.41616392e-01 1.52447209e-01
3.41875345e-01 -6.73346877e-01 -9.21619058e-01 2.44870827e-01
5.38700998e-01 -9.12898779e-01 -3.61179531e-01 1.12022102e+00
4.73214298e-01 3.57886523e-01 1.68248549e-01 -2.64372319e-01
-7.92475104e-01 9.32200551e-01 -6.75116420e-01 6.45807922e-01
3.77623647e-01 1.10199675e-02 4.20321882e-01 -1.33652961e+00
9.81183767e-01 1.24426770e+00 8.35260868e-01 8.13569725e-01
-1.24692261e+00 -4.37546700e-01 -3.43127102e-02 3.54718745e-01
-1.46929240e+00 -6.37009382e-01 7.89588571e-01 -3.61528277e-01
5.69313645e-01 8.64829794e-02 4.26174343e-01 9.63974714e-01
1.43703166e-03 6.06980622e-01 9.97686327e-01 -2.60527153e-02
2.92069744e-02 -3.95543158e-01 -3.29606142e-03 4.39929694e-01
-6.93692490e-02 3.94302130e-01 -7.56084263e-01 3.15118521e-01
1.75372314e+00 -3.65864545e-01 -8.60949814e-01 2.49107733e-01
-1.35233295e+00 3.71801466e-01 3.31811905e-01 3.76078665e-01
-5.99072576e-01 4.31769677e-02 -8.86748955e-02 1.28299415e-01
4.37151074e-01 3.67851168e-01 -2.47355178e-01 1.60933241e-01
-1.30600381e+00 3.18221897e-01 1.11600496e-01 5.87101400e-01
6.85812414e-01 6.20430470e-01 -2.60335952e-01 1.06525350e+00
1.49284909e-02 -9.67260152e-02 3.66287440e-01 -1.41425586e+00
1.38810277e-01 6.62435964e-02 4.87941563e-01 -6.30534112e-01
-1.90304160e-01 -8.86678338e-01 -1.06180835e+00 3.13727230e-01
6.53965101e-02 -1.49153933e-01 -7.83311069e-01 1.37204695e+00
2.48254001e-01 8.72827590e-01 6.18010648e-02 1.46068347e+00
1.17382789e+00 7.10086882e-01 -4.13751096e-01 -5.07411301e-01
9.40490305e-01 -1.10348606e+00 -1.01062930e+00 2.35683694e-01
-1.88079432e-01 -7.57551670e-01 4.29414749e-01 5.52329183e-01
-1.76255405e+00 -1.17958748e+00 -1.17506135e+00 -2.81054884e-01
2.79641896e-01 3.06271911e-01 3.51188719e-01 -5.38571328e-02
-1.49636471e+00 1.31070018e+00 -6.06080174e-01 2.62056798e-01
5.16116440e-01 1.59081474e-01 -1.38600335e-01 8.95895362e-02
-1.29627097e+00 8.49817634e-01 3.71442229e-01 5.36422849e-01
-1.08044732e+00 -8.87901664e-01 -6.89991176e-01 -1.16549991e-01
2.13150725e-01 -1.01532340e+00 1.14828026e+00 -1.20494485e+00
-1.81993961e+00 4.85511214e-01 -3.81174862e-01 -4.98821169e-01
5.38872302e-01 -4.36293840e-01 -4.21339035e-01 4.06584084e-01
-1.02403514e-01 4.57627058e-01 1.27202332e+00 -1.56165242e+00
-8.48620296e-01 7.62064084e-02 -2.61580944e-01 2.85497844e-01
2.90115625e-01 1.95846468e-01 -5.20888209e-01 -8.31208289e-01
4.16935354e-01 -4.66437697e-01 -5.31459451e-01 -2.09272191e-01
1.92019381e-02 1.60454482e-01 5.86030960e-01 -1.13384449e+00
1.20836103e+00 -1.79036105e+00 6.37773991e-01 -3.01120162e-01
3.90640557e-01 5.15272558e-01 -2.11700320e-01 -1.45714924e-01
-4.57588434e-01 -2.09113136e-02 -2.86083758e-01 -2.64403135e-01
-6.59648657e-01 -1.01659007e-01 -4.21820611e-01 3.40280592e-01
5.87409616e-01 8.73538673e-01 -8.13896000e-01 -2.35298961e-01
3.46001834e-01 1.03809404e+00 -3.05937111e-01 2.48973623e-01
-2.12630462e-02 1.00406492e+00 -3.33054960e-01 4.16618675e-01
1.20695281e+00 -3.88913780e-01 -7.85878003e-02 -6.79362476e-01
-6.50889814e-01 1.94312051e-01 -1.58696342e+00 1.72507584e+00
-6.80360273e-02 5.21526456e-01 3.82140517e-01 -6.61003649e-01
8.94582808e-01 2.89709389e-01 6.75061464e-01 -4.58326280e-01
1.59749500e-02 2.04312414e-01 -3.62804145e-01 -5.23034275e-01
1.01169181e+00 -7.12647364e-02 8.95830691e-01 4.59253713e-02
1.19566485e-01 4.06427443e-01 5.70107661e-02 1.04456134e-01
8.11876476e-01 7.81674206e-01 -3.48790991e-03 -2.54996926e-01
8.63190532e-01 -1.14385240e-01 8.61509919e-01 5.13039649e-01
2.61290427e-02 1.26057470e+00 9.50859115e-02 -6.27579212e-01
-1.36770022e+00 -1.18105733e+00 -2.34975800e-01 3.76646638e-01
3.38936239e-01 -1.29327759e-01 -6.84845328e-01 -1.83918282e-01
-5.87779522e-01 3.02739680e-01 -6.23763204e-01 1.72325671e-01
-1.15318274e+00 -7.78413177e-01 1.02816097e-01 5.95209897e-01
7.90340424e-01 -1.01244366e+00 -3.95176917e-01 3.11101794e-01
-5.37120581e-01 -1.33166349e+00 -3.24013114e-01 -2.66189128e-01
-1.17511570e+00 -8.84884357e-01 -9.33483779e-01 -7.28547454e-01
4.46607977e-01 5.78111827e-01 1.25444627e+00 1.55468792e-01
-1.03083216e-01 9.43392292e-02 -3.48975360e-01 4.52284753e-01
-4.20565248e-01 -4.14408952e-01 -9.20011252e-02 3.06684017e-01
-2.25148767e-01 -7.15645134e-01 -8.05649817e-01 1.57669276e-01
-9.07278121e-01 5.06446660e-01 6.13188922e-01 8.99944305e-01
1.11994040e+00 2.29390204e-01 4.95245934e-01 -5.76471448e-01
2.90386677e-01 -3.48466873e-01 -6.66067064e-01 1.56423841e-02
-3.83569032e-01 1.20329261e-02 6.34326637e-01 -4.89109963e-01
-1.60873079e+00 7.30717927e-02 -2.26582035e-01 -7.98020720e-01
-8.97558779e-03 2.03541055e-01 1.24586120e-01 -1.93403304e-01
5.24356484e-01 6.33353710e-01 -4.91946302e-02 -7.78037608e-01
3.12922180e-01 1.68998674e-01 9.69546914e-01 -2.67175525e-01
1.10276318e+00 7.18132138e-01 -1.48173138e-01 -8.53443086e-01
-7.91290402e-01 -2.98049212e-01 -9.28934038e-01 -5.03100574e-01
1.01620305e+00 -1.26227641e+00 -4.37997490e-01 4.74597812e-01
-1.31069243e+00 -2.42089942e-01 -2.98376858e-01 5.50024390e-01
-6.20424926e-01 5.25064111e-01 -1.07490957e+00 -6.73736274e-01
-5.61059892e-01 -1.11709023e+00 9.83717561e-01 6.27843380e-01
4.32719648e-01 -5.76087475e-01 1.78400781e-02 4.54442471e-01
7.12592423e-01 2.00952530e-01 8.34816694e-02 2.78186381e-01
-1.26508725e+00 6.60759151e-01 -4.59418535e-01 5.61846912e-01
-1.35054335e-01 2.52817757e-02 -8.07289600e-01 -2.81594694e-01
1.63446188e-01 1.29375532e-01 1.21879005e+00 7.90163875e-01
1.19195044e+00 7.97414556e-02 -7.39238709e-02 1.23722076e+00
1.88137186e+00 7.63435140e-02 1.20382977e+00 4.03951794e-01
7.72946715e-01 2.40476474e-01 5.98434806e-01 2.48955324e-01
8.33325759e-02 9.06976342e-01 2.92907923e-01 -1.39039800e-01
-7.47794986e-01 -2.39053704e-02 5.98993361e-01 9.25854623e-01
-8.38073075e-01 2.00332671e-01 -2.57596910e-01 4.55810189e-01
-1.89896989e+00 -1.34332371e+00 -4.32520896e-01 2.01681495e+00
7.89530635e-01 -2.61856645e-01 -2.08446831e-01 -2.42389634e-01
9.04370010e-01 1.83629557e-01 -3.98809731e-01 3.23452912e-02
-4.58078712e-01 4.39575404e-01 3.10767204e-01 6.74204528e-01
-9.29340065e-01 1.05380809e+00 6.53708410e+00 7.06144214e-01
-1.00592625e+00 1.71003625e-01 4.85755265e-01 -8.77195671e-02
-2.25907072e-01 -2.16955572e-01 -8.97139788e-01 2.72778332e-01
7.35846996e-01 7.64734894e-02 8.01062942e-01 4.41434532e-01
6.80176914e-01 1.91373900e-02 -7.54287779e-01 1.09299898e+00
1.12699963e-01 -1.94150805e+00 1.33076593e-01 -3.37427080e-01
9.79767263e-01 1.28845468e-01 -9.46693346e-02 -1.00230180e-01
2.42025256e-01 -1.07744515e+00 7.47654021e-01 1.03830564e+00
8.68059754e-01 -7.53563881e-01 5.39449394e-01 4.44459617e-02
-1.39117050e+00 -1.67708118e-02 -5.87420404e-01 2.23261148e-01
6.07984304e-01 5.48713207e-01 -2.33304307e-01 9.86755371e-01
9.48645413e-01 1.00211132e+00 -2.49469176e-01 1.05636990e+00
-1.20618634e-01 3.39161158e-01 2.54469305e-01 8.67120922e-01
-1.33215502e-01 -4.75647420e-01 1.01327467e+00 1.07136428e+00
4.94863033e-01 4.66668427e-01 -1.57154456e-01 1.40835583e+00
1.02548704e-01 -3.58091623e-01 -1.66066661e-01 4.73915011e-01
2.67054409e-01 1.37844992e+00 -5.69838345e-01 -4.28032339e-01
-3.46624464e-01 1.21563649e+00 2.50029296e-01 6.61448419e-01
-9.74962473e-01 1.30265042e-01 5.90224087e-01 1.89720988e-01
6.45124197e-01 -2.63784230e-01 -2.56838322e-01 -1.54490995e+00
5.16460985e-02 -7.81354606e-01 1.49417117e-01 -1.24218833e+00
-9.38755870e-01 8.35926771e-01 -2.83447504e-01 -1.56292725e+00
-6.52495176e-02 -2.35677809e-01 -3.30550849e-01 1.10568261e+00
-1.99738240e+00 -9.70140994e-01 -4.25877541e-01 6.51816964e-01
9.88068044e-01 9.49859545e-02 2.75089025e-01 4.87866431e-01
-6.22032583e-01 1.95112392e-01 1.23834722e-01 6.37556165e-02
4.85483438e-01 -9.83085692e-01 2.06292614e-01 1.41311240e+00
-2.82151133e-01 5.89289308e-01 7.07684755e-01 -7.49129117e-01
-1.29109108e+00 -1.09129667e+00 4.06305790e-01 -3.04354250e-01
3.84426117e-01 2.23642766e-01 -1.20805657e+00 5.71742475e-01
1.77487805e-01 4.29260358e-02 -1.97683088e-02 -4.85207975e-01
-8.00561458e-02 -6.29450008e-02 -1.07258940e+00 6.24112964e-01
1.29128706e+00 -2.16060683e-01 -5.35213828e-01 -1.39470190e-01
9.81371939e-01 -7.35618949e-01 -1.19789636e+00 6.91374183e-01
3.16355824e-01 -1.28687227e+00 1.43026769e+00 -2.94576406e-01
9.55221415e-01 -8.94436181e-01 -2.35940386e-02 -1.20335162e+00
-8.04949522e-01 -9.02965903e-01 -3.94622594e-01 9.76327658e-01
1.21312335e-01 -2.28230134e-01 5.72445929e-01 1.70269802e-01
-4.50763136e-01 -5.06876051e-01 -7.20584571e-01 -4.74114686e-01
-1.73341334e-01 -9.95252505e-02 4.93676603e-01 1.01472259e+00
-6.40989602e-01 3.69490176e-01 -7.97554851e-01 5.32424331e-01
1.20756233e+00 1.44131094e-01 5.00211477e-01 -9.44399893e-01
-3.11458558e-01 -3.73162121e-01 1.45144284e-01 -1.28055239e+00
-2.73093492e-01 -4.77235556e-01 1.05072476e-01 -1.60466397e+00
3.23239118e-01 -1.97222847e-02 -3.46429735e-01 -1.60056025e-01
-3.58484477e-01 4.11628634e-01 2.22775359e-02 4.10334617e-01
-6.24856055e-01 5.32089651e-01 1.81374943e+00 4.37402189e-01
-4.57603604e-01 -1.99688002e-01 -5.42072535e-01 6.28463030e-01
5.77721834e-01 -1.34911407e-02 -1.92240685e-01 -5.82988739e-01
5.01192026e-02 6.39695585e-01 5.76166391e-01 -1.06475186e+00
1.19084194e-02 -6.07822239e-02 7.53091156e-01 -9.97353673e-01
3.07844430e-01 -6.47671580e-01 6.04836643e-01 1.93099111e-01
-3.74206632e-01 -1.43306524e-01 3.38833556e-02 4.15801555e-01
-2.97275305e-01 -1.12632930e-01 1.26529729e+00 -4.64035124e-01
-1.04909575e+00 3.81596684e-01 -1.89555973e-01 -4.29115534e-01
5.58168828e-01 -4.62434888e-01 -3.60702097e-01 -1.35866642e-01
-9.10261154e-01 -2.63741184e-02 4.46101308e-01 5.16890585e-01
1.10870075e+00 -1.36699474e+00 -1.14082456e+00 1.68994218e-01
-6.08669877e-01 2.96867281e-01 8.07992280e-01 8.80935848e-01
-5.52089334e-01 3.13465321e-03 -3.73763084e-01 -5.64599574e-01
-1.15643120e+00 4.34823960e-01 6.68070078e-01 -1.71854675e-01
-1.22592163e+00 8.73687923e-01 1.40954018e-01 8.83589610e-02
-5.74644841e-02 -6.61454350e-02 -7.87161946e-01 -3.16645354e-01
1.25270998e+00 4.73730922e-01 -2.77401596e-01 -8.48313987e-01
4.36230153e-02 8.61916542e-01 6.59686103e-02 -4.29072380e-02
1.77879643e+00 -5.76802611e-01 -3.79026443e-01 1.05669580e-01
7.69187450e-01 -2.37327471e-01 -1.72780180e+00 -2.85331219e-01
-2.93164432e-01 -6.12801135e-01 4.03402120e-01 -6.65894389e-01
-1.47066522e+00 4.16588068e-01 6.95121348e-01 -2.89173245e-01
1.38682187e+00 -2.18104705e-01 9.11982954e-01 -2.22240910e-01
3.73446375e-01 -1.06002474e+00 5.30571351e-03 4.58508730e-01
1.07138264e+00 -9.93305564e-01 2.71254390e-01 -5.24607897e-01
-4.02050108e-01 1.45491529e+00 6.65784895e-01 -3.36388767e-01
3.21054667e-01 3.62450063e-01 -1.15625523e-01 -7.96121284e-02
-8.36288035e-01 -2.22510681e-01 3.06967288e-01 6.86236143e-01
4.24317777e-01 -3.45009208e-01 -5.56492150e-01 6.54431045e-01
1.54146120e-01 6.02614284e-01 9.02913690e-01 4.04879272e-01
-6.09654427e-01 -7.50099242e-01 -6.07975662e-01 1.03840940e-01
-6.04921460e-01 -1.37464032e-01 3.57083440e-01 2.97896564e-01
2.15045393e-01 7.50969410e-01 -9.81102288e-02 -4.97583061e-01
3.38025689e-02 -3.83897513e-01 5.50924897e-01 -2.43415430e-01
-4.13533062e-01 4.83479828e-01 7.33195022e-02 -1.07535756e+00
-8.31206620e-01 -5.96965671e-01 -1.30082262e+00 -3.96806486e-02
-3.01813722e-01 -9.61368829e-02 3.74020517e-01 7.72353590e-01
2.84971416e-01 9.75609779e-01 5.67408502e-01 -1.34077156e+00
-2.35273153e-01 -8.41858685e-01 -3.87812674e-01 1.31325349e-01
4.73975569e-01 -5.70238054e-01 -3.59113663e-01 5.06942630e-01] | [11.029964447021484, -1.9616544246673584] |
5e7e3d67-bb65-4909-bdd4-23fc0babbeb1 | flexlip-a-controllable-text-to-lip-system | 2206.03206 | null | https://arxiv.org/abs/2206.03206v1 | https://arxiv.org/pdf/2206.03206v1.pdf | FlexLip: A Controllable Text-to-Lip System | The task of converting text input into video content is becoming an important topic for synthetic media generation. Several methods have been proposed with some of them reaching close-to-natural performances in constrained tasks. In this paper, we tackle a subissue of the text-to-video generation problem, by converting the text into lip landmarks. However, we do this using a modular, controllable system architecture and evaluate each of its individual components. Our system, entitled FlexLip, is split into two separate modules: text-to-speech and speech-to-lip, both having underlying controllable deep neural network architectures. This modularity enables the easy replacement of each of its components, while also ensuring the fast adaptation to new speaker identities by disentangling or projecting the input features. We show that by using as little as 20 min of data for the audio generation component, and as little as 5 min for the speech-to-lip component, the objective measures of the generated lip landmarks are comparable with those obtained when using a larger set of training samples. We also introduce a series of objective evaluation measures over the complete flow of our system by taking into consideration several aspects of the data and system configuration. These aspects pertain to the quality and amount of training data, the use of pretrained models, and the data contained therein, as well as the identity of the target speaker; with regard to the latter, we show that we can perform zero-shot lip adaptation to an unseen identity by simply updating the shape of the lips in our model. | ['Horia Cucu', 'Adriana Stan', 'Beata Lorincz', 'Dan Oneata'] | 2022-06-07 | null | null | null | null | ['audio-generation', 'text-to-video-generation'] | ['audio', 'natural-language-processing'] | [ 3.03010136e-01 5.51858246e-01 3.07447854e-02 1.68723483e-02
-1.01725459e+00 -5.55357218e-01 7.92643607e-01 -1.59604147e-01
-2.72546977e-01 7.62664616e-01 2.05863535e-01 2.31712535e-01
2.25784913e-01 -5.03414571e-01 -6.94033682e-01 -8.70010793e-01
2.99987853e-01 5.56806445e-01 2.41316751e-01 -9.62087288e-02
-2.07919683e-02 5.13992190e-01 -2.00127554e+00 1.83317825e-01
5.88150442e-01 1.13980460e+00 6.45744652e-02 8.20124745e-01
-4.38251980e-02 3.89474392e-01 -6.31880462e-01 -5.22565901e-01
2.83065170e-01 -5.09333313e-01 -5.14710724e-01 3.04734588e-01
5.16617298e-01 -1.81390792e-01 -4.76388969e-02 8.25585306e-01
9.03510571e-01 6.85023656e-03 7.12713778e-01 -1.32481790e+00
-2.66979426e-01 4.85550016e-01 -6.34382591e-02 -4.32028651e-01
4.43149269e-01 1.37283340e-01 8.26017499e-01 -1.04450738e+00
7.41123855e-01 9.52202320e-01 4.72922355e-01 8.22560310e-01
-1.33623409e+00 -3.67509335e-01 -1.35589063e-01 -2.11954080e-02
-1.35358644e+00 -1.32800424e+00 7.19755828e-01 -6.53976917e-01
3.54345918e-01 1.44579858e-01 4.82593149e-01 1.27285683e+00
-2.94022977e-01 7.40264773e-01 7.77915418e-01 -7.01390266e-01
3.10650676e-01 6.48104906e-01 -4.53049034e-01 4.01619703e-01
-1.29151225e-01 4.08161618e-02 -5.45285344e-01 5.30863293e-02
5.56196809e-01 -6.14862382e-01 -5.12444198e-01 -6.51273668e-01
-1.16620719e+00 6.92038119e-01 -3.40888873e-02 3.84606987e-01
-3.76299256e-03 -1.92903150e-02 3.54865015e-01 7.82610700e-02
4.06244665e-01 2.16046184e-01 -2.03202680e-01 1.55825112e-02
-1.13660979e+00 1.83350727e-01 7.70619810e-01 1.14042211e+00
5.41524231e-01 3.48684788e-01 -2.51933992e-01 8.07695508e-01
3.83091457e-02 4.80644614e-01 7.35564172e-01 -8.60737503e-01
3.93294811e-01 7.05680847e-02 8.56218860e-02 -6.72529519e-01
-1.03363574e-01 -8.02490637e-02 -7.23251820e-01 4.17087257e-01
5.72757602e-01 -4.11252767e-01 -8.18227887e-01 2.06220508e+00
3.85635704e-01 1.85619537e-02 1.47466566e-02 4.89910334e-01
6.85630620e-01 4.73355234e-01 -1.65613681e-01 -4.67654705e-01
1.15421939e+00 -8.47398818e-01 -7.48937666e-01 3.68225537e-02
1.60470739e-01 -9.47619200e-01 1.02793634e+00 2.12782100e-01
-1.46652198e+00 -8.34856689e-01 -8.61236989e-01 -9.51774642e-02
-4.13305521e-01 5.55360913e-01 -2.11427417e-02 6.67402625e-01
-1.33573997e+00 4.44266260e-01 -4.33424443e-01 -3.29420149e-01
5.07717207e-02 4.78017479e-01 -5.15988946e-01 3.68638664e-01
-1.01665795e+00 6.59461379e-01 1.08913749e-01 -2.36007654e-05
-7.04974532e-01 -4.35019374e-01 -9.48985934e-01 9.30907503e-02
3.57745171e-01 -7.65939355e-01 1.24434793e+00 -1.16878045e+00
-2.11086893e+00 8.78619075e-01 -2.65018046e-01 -1.48401842e-01
9.89151418e-01 7.10131414e-03 -1.65177315e-01 2.69502312e-01
-8.43462255e-03 8.99255395e-01 1.22999966e+00 -1.23254061e+00
-4.20646727e-01 -2.22020783e-02 -1.80561557e-01 1.07109047e-01
-2.51461357e-01 -1.13111131e-01 -5.45808375e-01 -7.70104766e-01
-2.22837225e-01 -9.40306664e-01 2.38047436e-01 3.31095666e-01
-3.97218853e-01 -3.11595667e-03 5.77890635e-01 -5.18481553e-01
7.87291408e-01 -2.24753666e+00 4.36579913e-01 -1.32234201e-01
-3.89116467e-03 3.05010766e-01 -2.36138314e-01 3.03223252e-01
-1.66772351e-01 8.40385780e-02 -1.87823266e-01 -9.47715282e-01
8.99613351e-02 -2.52505034e-01 -3.21606129e-01 3.43758941e-01
1.92221537e-01 6.40975296e-01 -5.71259201e-01 -6.28744543e-01
3.41826171e-01 7.64239848e-01 -4.60794747e-01 3.55265528e-01
-2.01948896e-01 3.73196185e-01 4.80820537e-02 3.47926259e-01
4.23754066e-01 3.78361911e-01 1.05143584e-01 -2.34248966e-01
-1.19030684e-01 1.41948521e-01 -1.27251863e+00 1.47650206e+00
-6.50045753e-01 8.97265732e-01 5.11337101e-01 -7.05745697e-01
8.48937988e-01 8.39953899e-01 5.74251950e-01 -3.94285440e-01
2.06675023e-01 2.96960771e-01 -2.23499984e-01 -5.36626995e-01
4.23866481e-01 -2.58079171e-01 8.35971683e-02 4.25600201e-01
4.22432840e-01 -3.62944603e-01 2.48754531e-01 -1.47833437e-01
3.85065973e-01 1.77800477e-01 2.03174666e-01 -1.21787630e-01
6.62478566e-01 -6.29162908e-01 1.21414408e-01 1.56833276e-01
-8.69053081e-02 1.01590157e+00 5.17042041e-01 1.28583297e-01
-1.08419585e+00 -1.07046318e+00 -3.32765505e-02 8.19053113e-01
-1.73882291e-01 -2.33814016e-01 -1.19103229e+00 -4.11294490e-01
-2.46252105e-01 6.49455190e-01 -6.28662765e-01 -1.38689950e-02
-4.83276814e-01 -4.61212456e-01 5.95061839e-01 2.00717464e-01
2.25259632e-01 -1.12081337e+00 -3.77765685e-01 7.66156986e-02
-3.68434310e-01 -1.29169273e+00 -6.71091795e-01 -8.07124600e-02
-4.52922910e-01 -8.21953595e-01 -1.09651089e+00 -8.10914338e-01
6.24041438e-01 -1.69817924e-01 8.21422100e-01 -3.25718194e-01
-1.82837337e-01 3.84775370e-01 1.49499148e-03 -3.45124274e-01
-7.98995495e-01 2.05997169e-01 1.36343420e-01 5.44836164e-01
-3.09601307e-01 -6.46948278e-01 -2.84753293e-01 4.62459564e-01
-8.71887624e-01 1.52794078e-01 3.48494709e-01 5.57282805e-01
5.32078981e-01 -2.69232392e-01 6.03713930e-01 -4.12212163e-01
5.79930425e-01 -1.96882725e-01 -5.43598890e-01 1.30278066e-01
-3.55049729e-01 1.15289114e-01 6.62910223e-01 -6.21452630e-01
-8.35476816e-01 2.90008783e-01 -2.92892754e-01 -4.61895466e-01
-2.48369828e-01 -1.72201917e-01 -5.70792139e-01 -1.18788956e-02
4.89940643e-01 2.71989942e-01 2.94047087e-01 -5.06119192e-01
5.76483428e-01 9.54422951e-01 5.45055985e-01 -4.58686292e-01
7.47172356e-01 3.49066257e-01 -9.09912363e-02 -1.11095178e+00
-3.17190439e-01 -1.76332623e-01 -7.75476277e-01 -3.81757289e-01
7.18923032e-01 -6.59175217e-01 -8.01733077e-01 7.07184732e-01
-1.13250327e+00 -5.41803420e-01 -6.79438949e-01 2.88577706e-01
-1.03634763e+00 2.09064290e-01 -3.23362917e-01 -7.91254342e-01
-2.56486326e-01 -1.30637658e+00 1.29161906e+00 4.86280769e-03
-1.75334588e-01 -9.13572192e-01 1.62945092e-01 3.84559840e-01
4.14842129e-01 1.94415584e-01 7.16535211e-01 -5.14562964e-01
-4.41337675e-01 -3.27343851e-01 3.51759717e-02 5.67210853e-01
1.66003183e-01 2.56911099e-01 -1.45087862e+00 -1.96103022e-01
7.29970559e-02 -3.03469449e-01 7.24487603e-01 3.79476130e-01
7.79734313e-01 -4.04524237e-01 -1.95260629e-01 5.56336284e-01
1.05484283e+00 -4.19811755e-02 5.65691948e-01 -2.02587500e-01
4.00150895e-01 9.74415302e-01 2.90711552e-01 1.96992069e-01
1.72223002e-01 1.16250479e+00 1.63439870e-01 -8.10759589e-02
-5.74206412e-01 -3.98924559e-01 5.87673008e-01 7.00892985e-01
-7.80538842e-02 -3.90454471e-01 -4.65619981e-01 5.98025322e-01
-1.47070372e+00 -1.01715767e+00 3.03084731e-01 2.48024464e+00
7.84987807e-01 8.76427628e-03 3.55236799e-01 4.60397869e-01
8.95173073e-01 1.39654458e-01 -3.01937252e-01 -2.60675728e-01
-3.37258801e-02 1.61036119e-01 1.04466956e-02 9.45218205e-01
-9.04564679e-01 8.03274453e-01 6.71263790e+00 8.53322566e-01
-1.51028776e+00 2.98534762e-02 4.54220593e-01 -3.62277776e-01
-1.74639046e-01 -2.90937513e-01 -7.99632072e-01 4.99543697e-01
7.95313418e-01 -2.31348991e-01 5.43279469e-01 7.70536721e-01
3.63062233e-01 -2.96143387e-02 -1.25513327e+00 8.72097135e-01
4.46522146e-01 -1.09074461e+00 2.03502551e-01 -7.37266541e-02
4.51565713e-01 -2.71319091e-01 1.71283960e-01 2.01153811e-02
-3.26971143e-01 -8.76664162e-01 1.24132073e+00 4.90396649e-01
1.32071292e+00 -4.55616146e-01 3.23431075e-01 2.92598128e-01
-1.13452220e+00 6.25132993e-02 -4.08744253e-02 4.00059819e-01
2.49272987e-01 4.38168406e-01 -8.73656154e-01 3.81978869e-01
1.27367049e-01 1.65301621e-01 -3.54844242e-01 8.86514902e-01
-1.69896707e-01 2.48383477e-01 -3.57382208e-01 1.30114183e-01
-2.66633898e-01 5.92196956e-02 6.92678511e-01 9.98956263e-01
5.44407070e-01 -5.51847875e-01 -2.75422603e-01 9.51431274e-01
-9.19758454e-02 2.99908161e-01 -6.34566963e-01 1.57620043e-01
2.30815202e-01 1.24434626e+00 -3.95928055e-01 -2.29095429e-01
-8.95814132e-03 8.82969379e-01 1.30620345e-01 4.03519064e-01
-6.95913851e-01 -5.55621326e-01 6.61128283e-01 5.18431067e-01
2.58523196e-01 -8.45567659e-02 -5.37180416e-02 -1.14206803e+00
2.76323438e-01 -8.19333613e-01 -1.35484532e-01 -9.36877906e-01
-7.39846528e-01 9.58398283e-01 6.11004643e-02 -1.18726754e+00
-8.27346027e-01 -5.07312775e-01 -6.01586699e-01 9.23082769e-01
-1.42045331e+00 -1.15166759e+00 -2.60744482e-01 6.64406896e-01
6.49702191e-01 -3.06072861e-01 9.21938598e-01 3.80282611e-01
-4.27487403e-01 8.03176999e-01 3.08981556e-02 1.67126264e-02
8.43513072e-01 -8.95532489e-01 2.63654470e-01 7.11513400e-01
2.18824387e-01 2.20799863e-01 6.60875022e-01 -1.33210942e-01
-8.34593177e-01 -9.06179547e-01 1.09160435e+00 -2.98255414e-01
3.79607856e-01 -7.54725397e-01 -4.34286058e-01 4.73587245e-01
1.42180219e-01 -9.00223553e-02 4.28301871e-01 -2.38896027e-01
-3.68094556e-02 -3.21074069e-01 -1.12178934e+00 7.14709222e-01
8.33136916e-01 -6.08647466e-01 -3.12781006e-01 1.56489000e-01
5.95509410e-01 -2.97382414e-01 -4.78695512e-01 1.90707818e-01
7.14175224e-01 -1.21188104e+00 7.96387196e-01 -1.86381027e-01
1.83856204e-01 -1.79388314e-01 5.51597355e-03 -1.11499298e+00
1.59392774e-01 -1.10796058e+00 6.10303767e-02 1.48124421e+00
4.65177387e-01 -5.47642052e-01 7.59874463e-01 4.05527413e-01
-3.07103787e-02 -6.25609696e-01 -1.12155938e+00 -5.30636549e-01
1.87408794e-02 -3.66037458e-01 4.45692837e-01 4.98099506e-01
5.76044247e-02 4.70825791e-01 -4.29887831e-01 -1.55345947e-01
4.78970498e-01 -6.65579513e-02 1.00556183e+00 -1.04248011e+00
-3.46693367e-01 -5.31866491e-01 -2.65133351e-01 -1.00785255e+00
1.54754877e-01 -7.18442976e-01 1.82116628e-01 -1.21488547e+00
-2.49343827e-01 -1.75933912e-01 2.22962424e-01 1.75007552e-01
1.28777891e-01 3.79024804e-01 3.92862111e-01 1.02123685e-01
2.57494114e-02 6.05661809e-01 1.21365547e+00 7.83237920e-04
-2.75828689e-01 4.62313443e-01 -3.70211393e-01 7.45269001e-01
6.13374591e-01 -2.38615543e-01 -4.88406450e-01 -3.04739624e-01
-1.25392303e-01 2.89105028e-01 3.10873330e-01 -1.16760147e+00
1.48018792e-01 9.30888802e-02 -3.44332214e-03 -1.18921399e-01
8.73928547e-01 -7.15703487e-01 1.78277865e-01 1.75716609e-01
-5.13830304e-01 -3.44269753e-01 2.20563516e-01 2.12787122e-01
-2.88614184e-01 -4.04960245e-01 1.10316133e+00 1.62787199e-01
-8.10976401e-02 1.76312253e-01 -4.42292422e-01 7.07051381e-02
1.04322708e+00 -3.23343068e-01 7.13250488e-02 -7.71026790e-01
-1.11803377e+00 -2.24373206e-01 5.00348568e-01 4.13206786e-01
2.95773298e-01 -1.16579962e+00 -6.68918312e-01 5.77606082e-01
-1.05894618e-01 -6.95947111e-02 6.24974146e-02 7.74237096e-01
-2.36556903e-01 4.42540050e-01 -1.82407945e-01 -5.01105845e-01
-1.41419864e+00 6.02307618e-01 5.75823903e-01 -2.96266507e-02
-1.85699195e-01 5.33882797e-01 2.72163987e-01 -1.49724200e-01
5.03108084e-01 -2.64068544e-01 -1.27827942e-01 4.08024877e-01
4.07980710e-01 3.51639301e-01 -7.99418464e-02 -7.71082878e-01
2.24686973e-02 8.55343461e-01 3.93356949e-01 -5.20759463e-01
1.04821110e+00 -2.50329763e-01 1.53635129e-01 4.77790236e-01
1.08289683e+00 4.53974158e-01 -1.26863742e+00 9.38048363e-02
-3.88813853e-01 -1.63693205e-01 -1.94012120e-01 -4.88733530e-01
-1.02975655e+00 1.21973133e+00 4.37662184e-01 2.39091679e-01
1.10673892e+00 -6.96645379e-02 5.59474468e-01 -9.99152660e-03
1.01450667e-01 -1.05134130e+00 1.93116292e-01 1.06376424e-01
1.17323089e+00 -9.56745744e-01 -3.37999940e-01 -4.73058224e-01
-4.99618530e-01 1.09557796e+00 2.03273281e-01 1.76401034e-01
5.20252109e-01 3.95238966e-01 2.06678659e-01 4.14371014e-01
-6.51292682e-01 -2.82922506e-01 2.97496259e-01 7.98529446e-01
2.93824017e-01 -1.69253990e-01 2.99890749e-02 2.62865782e-01
-3.09419572e-01 1.19145758e-01 3.95640582e-01 2.92056203e-01
-2.59051859e-01 -1.24414885e+00 -2.94458061e-01 -6.76944330e-02
-3.71666759e-01 -7.41618648e-02 -4.81905609e-01 8.33138406e-01
2.88803458e-01 8.17841470e-01 3.82542163e-02 -1.60920113e-01
3.48768443e-01 3.62163931e-01 4.37922448e-01 -5.71167707e-01
-3.11652690e-01 2.61526316e-01 8.07770118e-02 -3.14787328e-01
-2.92809367e-01 -7.64575541e-01 -7.89585114e-01 -1.27517164e-01
-3.71300966e-01 5.86069860e-02 8.39792967e-01 7.43726194e-01
3.34086359e-01 2.05155596e-01 6.58738971e-01 -1.35861540e+00
-4.70610112e-01 -1.06460547e+00 -5.04700899e-01 2.75915623e-01
4.59612131e-01 -5.45813024e-01 -6.09684467e-01 4.33989644e-01] | [13.21837043762207, -0.3604690730571747] |
3359c645-8e10-4b95-8a27-0010a90bec84 | exploiting-knowledge-graphs-for-facilitating | 2010.05213 | null | https://arxiv.org/abs/2010.05213v1 | https://arxiv.org/pdf/2010.05213v1.pdf | Exploiting Knowledge Graphs for Facilitating Product/Service Discovery | Most of the existing techniques to product discovery rely on syntactic approaches, thus ignoring valuable and specific semantic information of the underlying standards during the process. The product data comes from different heterogeneous sources and formats giving rise to the problem of interoperability. Above all, due to the continuously increasing influx of data, the manual labeling is getting costlier. Integrating the descriptions of different products into a single representation requires organizing all the products across vendors in a single taxonomy. Practically relevant and quality product categorization standards are still limited in number; and that too in academic research projects where we can majorly see only prototypes as compared to industry. This work presents a cost-effective solution for e-commerce on the Data Web by employing an unsupervised approach for data classification and exploiting the knowledge graphs for matching. The proposed architecture describes available products in web ontology language OWL and stores them in a triple store. User input specifications for certain products are matched against the available product categories to generate a knowledge graph. This mullti-phased top-down approach to develop and improve existing, if any, tailored product recommendations will be able to connect users with the exact product/service of their choice. | ['Sarika Jain'] | 2020-10-11 | null | null | null | null | ['product-categorization'] | ['miscellaneous'] | [-4.70995232e-02 2.19397880e-02 -5.15542626e-01 -6.31790817e-01
-1.82398647e-01 -7.89128363e-01 3.79021019e-01 7.37069547e-01
-9.08944383e-02 3.85647327e-01 1.01387672e-01 -2.00119391e-01
-8.43886018e-01 -1.14001918e+00 -3.85886803e-02 -1.82827637e-01
4.13168639e-01 7.84629226e-01 4.90101069e-01 -6.81041420e-01
2.34054744e-01 4.69313383e-01 -2.33274364e+00 6.89076483e-01
8.91067922e-01 1.22498178e+00 4.27556306e-01 -1.81731139e-03
-1.05704916e+00 4.36479360e-01 -2.44933367e-01 -5.41718423e-01
3.64558250e-01 -1.98126107e-01 -8.57843220e-01 2.93255270e-01
3.43019307e-01 3.23265970e-01 4.67469066e-01 1.38095868e+00
-5.84979355e-02 -1.66127205e-01 2.90557742e-01 -1.35759544e+00
-6.33301735e-01 7.67572641e-01 1.74730644e-02 -3.81036788e-01
6.55859530e-01 -4.92002636e-01 1.12443602e+00 -5.57120860e-01
1.08402550e+00 1.05180871e+00 4.49238539e-01 1.69142693e-01
-1.06398892e+00 -2.89651096e-01 7.77132511e-02 4.19462025e-01
-1.40311182e+00 -1.23234950e-01 6.67656124e-01 -5.67498982e-01
8.77607346e-01 4.96131927e-01 7.08362877e-01 4.86048251e-01
-3.65872353e-01 -1.05351038e-01 9.57162499e-01 -7.44625509e-01
2.81003892e-01 1.17863691e+00 6.74007177e-01 3.89849812e-01
9.24061120e-01 -2.61116087e-01 -2.76456654e-01 2.44917329e-02
1.29022688e-01 2.08116740e-01 1.93725713e-02 -7.46756434e-01
-4.22003418e-01 6.87065840e-01 7.56873414e-02 1.01114631e+00
-4.35762763e-01 -6.42279685e-01 4.95723754e-01 5.44555306e-01
-1.80152170e-02 3.93380731e-01 -7.62566328e-01 1.68907925e-01
-5.42309403e-01 3.24481577e-01 1.17570281e+00 1.46131551e+00
8.94779205e-01 -3.76147479e-01 7.77349591e-01 7.37590373e-01
7.55984485e-01 1.30219713e-01 5.65537035e-01 -5.24778545e-01
2.62856185e-01 1.41339278e+00 2.29109198e-01 -1.19153619e+00
-3.37559909e-01 -4.55149859e-01 -5.03565744e-03 5.28657675e-01
3.66331011e-01 4.12605256e-01 -5.09118497e-01 1.15786505e+00
5.96897840e-01 -7.56233692e-01 1.87036827e-01 6.57469928e-01
8.18953097e-01 4.01624709e-01 3.21125150e-01 -2.03086331e-01
1.67776167e+00 -5.04456043e-01 -1.05341637e+00 1.65465683e-01
7.36883342e-01 -1.10631299e+00 7.44751871e-01 6.45209432e-01
-7.40611434e-01 -6.39122188e-01 -1.32496428e+00 3.42751928e-02
-1.39328825e+00 -3.03638756e-01 7.73339510e-01 9.04961169e-01
-5.68360269e-01 7.60896802e-01 -8.90760943e-02 -9.76691127e-01
2.32870487e-04 3.94181162e-01 -4.25453454e-01 -1.91602364e-01
-1.12587047e+00 1.13394547e+00 9.78291392e-01 -1.51142567e-01
-1.17759816e-01 -6.21694863e-01 -6.54941678e-01 -3.23792472e-02
5.45453072e-01 -5.90527833e-01 9.67294335e-01 -1.14683115e+00
-1.08018017e+00 5.83985031e-01 2.42085114e-01 -2.94308454e-01
2.10359991e-01 2.59332418e-01 -1.31749415e+00 -3.62343907e-01
1.47030160e-01 1.20347850e-02 2.95080155e-01 -1.50860178e+00
-1.41591656e+00 -7.28210449e-01 2.49462247e-01 -2.46731509e-02
-4.08351660e-01 1.92733407e-01 -1.36661202e-01 1.32702421e-02
2.04599127e-01 -3.26801211e-01 -4.55121472e-02 -3.29501420e-01
1.38129920e-01 -2.56652802e-01 8.69764447e-01 -4.26344752e-01
1.52757382e+00 -1.85257232e+00 -1.17740333e-01 5.60092986e-01
9.77857858e-02 1.50012106e-01 1.92108974e-01 9.71894860e-01
-4.81688380e-02 1.54730022e-01 3.36990356e-01 3.99975389e-01
4.70262915e-01 4.25050348e-01 1.01054177e-01 -6.97914287e-02
-4.49536324e-01 2.90234685e-01 -7.64312863e-01 -6.11993730e-01
3.38273615e-01 4.58828896e-01 -2.70193368e-01 -1.28767893e-01
-4.01206404e-01 -6.72062114e-02 -6.49228513e-01 8.09540749e-01
4.85401541e-01 -6.00549728e-02 6.73827887e-01 -7.33326018e-01
-4.39561069e-01 2.95213550e-01 -1.71222830e+00 1.69039261e+00
-7.47732520e-01 -1.48796603e-01 -3.93278599e-02 -9.81228888e-01
1.17946553e+00 6.34131253e-01 5.83369195e-01 -7.88550138e-01
4.07262117e-01 8.88986230e-01 -6.23654388e-02 -6.64015114e-01
4.43350315e-01 -5.68484664e-02 9.37013552e-02 7.14011490e-02
2.77552247e-01 1.32699251e-01 8.80028248e-01 -1.45960540e-01
6.40863776e-01 5.48011303e-01 5.78849435e-01 -3.33928674e-01
8.23461056e-01 4.38349128e-01 3.08681220e-01 2.60259956e-01
3.09058249e-01 -5.71031310e-03 4.73109223e-02 -7.83631504e-01
-1.13235795e+00 -6.80787444e-01 -4.08737957e-01 8.49253833e-01
6.50633201e-02 -7.52884567e-01 -6.34811103e-01 -6.42126918e-01
8.69774073e-02 6.80033207e-01 -2.56692171e-01 2.16773823e-01
-5.97962271e-03 -1.72509283e-01 -3.43280494e-01 -3.25081609e-02
1.92597494e-01 -1.00125670e+00 -6.69683456e-01 5.13365984e-01
2.65139252e-01 -9.51959789e-01 2.06122205e-01 2.85438485e-02
-8.14500809e-01 -1.17994571e+00 4.98419367e-02 -7.79480577e-01
5.14413476e-01 8.09216350e-02 1.13792992e+00 7.92618096e-02
-3.83819848e-01 1.12430476e-01 -8.01420987e-01 -6.10516191e-01
-6.37968183e-01 1.24268591e-01 -1.17945500e-01 7.22667649e-02
9.34900820e-01 -5.87451160e-01 -2.57657319e-01 4.03203309e-01
-9.86186683e-01 -2.08934799e-01 3.32780838e-01 2.19470575e-01
4.59864527e-01 7.16207564e-01 7.64571190e-01 -1.30013800e+00
4.64558572e-01 -5.23973584e-01 -9.35852826e-01 4.47303116e-01
-1.23188937e+00 1.11436643e-01 5.66741407e-01 -4.98804785e-02
-1.17049670e+00 2.51174241e-01 -2.13163108e-01 3.04972112e-01
-5.31559169e-01 8.03410530e-01 -5.74643314e-01 5.32562379e-03
5.70446372e-01 -3.33990127e-01 -2.11703464e-01 -9.83388186e-01
5.32929838e-01 1.00272977e+00 2.06950128e-01 -3.25507820e-01
6.28580093e-01 1.60035193e-01 -1.92256570e-02 -5.74850857e-01
-6.99345469e-01 -7.83890903e-01 -7.68230557e-01 -1.94348872e-01
6.93971157e-01 -3.67784768e-01 -7.96922266e-01 -2.73562282e-01
-9.86081898e-01 4.80913937e-01 -6.75779343e-01 3.85588557e-01
-3.21868181e-01 1.52143225e-01 6.52209390e-03 -7.93107569e-01
-1.57018930e-01 -9.51035678e-01 4.02485490e-01 1.62286535e-01
-4.59370464e-01 -9.24221635e-01 1.01024276e-02 5.20576239e-01
5.61094344e-01 -5.31519018e-02 1.26664984e+00 -1.07226288e+00
-4.31445807e-01 -6.77274108e-01 -1.54126525e-01 3.12337577e-01
5.06946981e-01 -1.44824684e-01 -5.90728045e-01 1.50050297e-01
-7.43204728e-02 2.98233300e-01 -1.34560287e-01 -2.80665308e-01
3.38101923e-01 -2.21506670e-01 -3.65601778e-01 -1.94950458e-02
2.11021185e+00 8.18954587e-01 5.54674029e-01 7.18779743e-01
4.37728614e-01 1.21245146e+00 9.06420410e-01 3.91097039e-01
2.23091096e-01 1.07320428e+00 3.13907981e-01 2.11657017e-01
-1.18272170e-01 -2.16903999e-01 -2.16316387e-01 7.40539432e-01
-1.27675191e-01 -4.83690910e-02 -5.38806856e-01 4.66652483e-01
-1.97014046e+00 -1.00693786e+00 -3.44223469e-01 2.47972322e+00
5.18435121e-01 1.42393410e-01 1.92826331e-01 4.18686956e-01
5.26622176e-01 -4.35598016e-01 -9.01975334e-02 -6.00040734e-01
1.52349338e-01 2.07794324e-01 5.92001855e-01 3.95481884e-01
-6.46768212e-01 4.79860902e-01 5.32685518e+00 5.60207903e-01
-7.31681466e-01 3.11356217e-01 -2.79715210e-01 3.09903651e-01
-5.76972127e-01 3.04694653e-01 -1.02161705e+00 3.72289360e-01
1.05194855e+00 -6.34404600e-01 3.84068310e-01 1.05555975e+00
1.46993056e-01 1.39322475e-01 -9.57869709e-01 9.24479425e-01
-5.71841598e-02 -1.42362094e+00 1.96832210e-01 2.94430137e-01
4.30842608e-01 -4.18421686e-01 -4.33599293e-01 -2.63625253e-02
3.34941685e-01 -5.13182878e-01 8.37680459e-01 5.10068357e-01
2.31888831e-01 -7.30696797e-01 1.00699139e+00 3.99936549e-02
-1.43668842e+00 -2.62652010e-01 -3.64803404e-01 2.09117547e-01
2.90547609e-01 3.50706369e-01 -4.90352601e-01 1.13515425e+00
7.60102689e-01 5.41266859e-01 -3.01649153e-01 9.90957618e-01
2.81639338e-01 -3.04756194e-01 -2.96007901e-01 -6.15603179e-02
-7.86163211e-02 -5.81752241e-01 5.87122664e-02 9.85674143e-01
7.50523329e-01 -2.03774020e-01 9.45649892e-02 6.84210956e-01
2.46777534e-01 9.11353767e-01 -8.08137774e-01 -2.80186981e-01
3.20889324e-01 1.44779801e+00 -6.41348124e-01 -2.98000544e-01
-9.27600443e-01 5.39602757e-01 -2.18745023e-01 -4.47599627e-02
-3.28458816e-01 -7.48579443e-01 6.75615430e-01 6.25827253e-01
3.82959217e-01 3.71942781e-02 2.02981438e-02 -7.76219487e-01
9.93323922e-02 -7.88813174e-01 5.91280580e-01 -5.28549433e-01
-1.35853863e+00 8.61851990e-01 2.40616888e-01 -1.43476915e+00
-3.19096714e-01 -7.69182205e-01 3.11804026e-01 8.65183294e-01
-1.31944239e+00 -1.24135101e+00 -2.06119090e-01 4.60398406e-01
4.39344049e-01 -2.17170343e-01 1.12362480e+00 9.86676633e-01
2.31018979e-02 -8.30893144e-02 -9.07172784e-02 -3.55369985e-01
4.15081769e-01 -1.07178879e+00 -2.31424794e-01 5.15191436e-01
2.50260025e-01 7.70574749e-01 8.04702759e-01 -7.58183122e-01
-1.39131999e+00 -5.74590802e-01 1.48603010e+00 -2.36584976e-01
9.76344943e-01 -1.94770202e-01 -9.35048878e-01 4.70194519e-01
4.71406221e-01 -1.88320473e-01 9.42508698e-01 1.34727910e-01
-5.95025837e-01 -6.81872368e-01 -1.41052926e+00 2.06780925e-01
1.04198694e+00 -1.66480362e-01 -7.13487744e-01 3.34330738e-01
5.32948196e-01 2.71234155e-01 -1.48126864e+00 1.28434673e-02
7.75383890e-01 -1.05828023e+00 6.28229976e-01 -6.86760485e-01
-2.68295109e-01 -7.30272949e-01 -4.11424696e-01 -9.80868638e-01
-1.61908597e-01 -2.69603372e-01 4.62425649e-01 1.66446328e+00
7.53630996e-01 -7.26060569e-01 6.94022357e-01 8.54013085e-01
-7.77718797e-02 -8.98700431e-02 -6.04496717e-01 -8.07917237e-01
-6.23752475e-01 -4.63358968e-01 1.17039740e+00 9.58353221e-01
5.37940502e-01 2.89207101e-01 3.05720530e-02 9.49401185e-02
5.98657966e-01 3.69600028e-01 3.83424252e-01 -1.89958596e+00
-2.78796077e-01 -4.82802898e-01 -8.45747292e-01 -1.17520571e-01
-4.12307948e-01 -1.03982782e+00 -4.32541460e-01 -2.00516152e+00
-2.33155116e-01 -6.81379437e-01 -3.10308278e-01 2.90251255e-01
1.00872457e+00 4.73822802e-02 2.31806681e-01 1.19631335e-01
-3.94475222e-01 -3.77352297e-01 7.70014048e-01 -5.13733327e-02
-1.85361713e-01 -1.15051180e-01 -8.08766127e-01 6.90689743e-01
6.32154822e-01 -6.06490791e-01 -6.69060171e-01 1.71620976e-02
7.69518077e-01 -3.55701685e-01 -2.67032564e-01 -1.14819741e+00
2.66606390e-01 -2.99673587e-01 -1.12300619e-01 -3.46200258e-01
-4.06913608e-02 -1.92619693e+00 1.07821608e+00 2.77043909e-01
-8.72698277e-02 -8.33185464e-02 -1.12918928e-01 1.02800034e-01
-4.06561106e-01 -9.07921910e-01 5.13803780e-01 -3.85874808e-01
-1.19544137e+00 1.04407713e-01 -1.57493487e-01 -6.09065950e-01
1.06157506e+00 -4.80279416e-01 -2.50610054e-01 9.97782722e-02
-1.10362267e+00 -2.43554279e-01 5.05931437e-01 7.77343273e-01
1.76741835e-02 -1.24134874e+00 -1.12418287e-01 2.19968408e-01
5.90694249e-01 -6.47381425e-01 1.57039002e-01 2.66517520e-01
-5.94061673e-01 7.15954721e-01 -5.86771250e-01 -2.22274438e-02
-1.22340572e+00 1.19556022e+00 2.04746038e-01 -1.37526914e-01
-3.60839397e-01 -1.11203484e-01 -4.26999658e-01 -2.83139020e-01
1.26780927e-01 -1.58470705e-01 -8.51929486e-01 4.49342579e-01
6.25659585e-01 3.55736375e-01 5.24037004e-01 -8.75340044e-01
-2.41939470e-01 8.11383963e-01 4.03346214e-03 -5.69616891e-02
1.31526542e+00 -4.30684239e-01 -2.80493468e-01 4.19203520e-01
1.09298456e+00 1.26802206e-01 -3.32771569e-01 -1.37474298e-01
7.31532991e-01 -6.08482122e-01 -1.34755820e-01 -1.00689447e+00
-1.06475794e+00 3.08782279e-01 9.14526820e-01 9.98341322e-01
1.03381133e+00 1.80262417e-01 4.14246887e-01 2.91924894e-01
9.38588381e-01 -1.31840706e+00 -8.79312634e-01 -9.12895352e-02
6.15508437e-01 -8.70971859e-01 1.18151575e-01 -9.22104836e-01
-4.06827211e-01 1.38014054e+00 1.86965927e-01 2.57799745e-01
8.87564182e-01 3.34656209e-01 2.80029565e-01 -6.39963329e-01
-4.23720181e-01 -6.00741267e-01 1.78637251e-01 9.50754344e-01
6.91396952e-01 -1.70573249e-01 -1.01427007e+00 6.99997783e-01
-1.60950031e-02 4.50808972e-01 2.25727856e-01 9.74837840e-01
-6.94318235e-01 -2.11355495e+00 -1.06174603e-01 2.73438126e-01
-4.16510791e-01 2.67007589e-01 -1.05327643e-01 9.71561730e-01
7.79134452e-01 1.05872703e+00 -3.89359123e-03 -2.44923532e-01
1.06761456e+00 2.54200667e-01 4.25936133e-01 -7.84964263e-01
-6.57094479e-01 4.76129130e-02 5.32068372e-01 -4.41723555e-01
-6.33261442e-01 -5.93565404e-01 -1.22643936e+00 -5.82773872e-02
-2.92374700e-01 4.62749451e-01 1.41527152e+00 7.56516337e-01
4.88781899e-01 3.45856786e-01 3.60824376e-01 -3.51390481e-01
-2.87903756e-01 -5.24723411e-01 -8.99611294e-01 8.45068693e-01
-2.96006680e-01 -9.07960713e-01 -8.98247138e-02 3.26659024e-01] | [9.11515998840332, 7.822516918182373] |
2fb4dd9d-19eb-4c52-b5df-aea7e785918e | complementary-learning-subnetworks-for | 2306.11967 | null | https://arxiv.org/abs/2306.11967v1 | https://arxiv.org/pdf/2306.11967v1.pdf | Complementary Learning Subnetworks for Parameter-Efficient Class-Incremental Learning | In the scenario of class-incremental learning (CIL), deep neural networks have to adapt their model parameters to non-stationary data distributions, e.g., the emergence of new classes over time. However, CIL models are challenged by the well-known catastrophic forgetting phenomenon. Typical methods such as rehearsal-based ones rely on storing exemplars of old classes to mitigate catastrophic forgetting, which limits real-world applications considering memory resources and privacy issues. In this paper, we propose a novel rehearsal-free CIL approach that learns continually via the synergy between two Complementary Learning Subnetworks. Our approach involves jointly optimizing a plastic CNN feature extractor and an analytical feed-forward classifier. The inaccessibility of historical data is tackled by holistically controlling the parameters of a well-trained model, ensuring that the decision boundary learned fits new classes while retaining recognition of previously learned classes. Specifically, the trainable CNN feature extractor provides task-dependent knowledge separately without interference; and the final classifier integrates task-specific knowledge incrementally for decision-making without forgetting. In each CIL session, it accommodates new tasks by attaching a tiny set of declarative parameters to its backbone, in which only one matrix per task or one vector per class is kept for knowledge retention. Extensive experiments on a variety of task sequences show that our method achieves competitive results against state-of-the-art methods, especially in accuracy gain, memory cost, training efficiency, and task-order robustness. Furthermore, to make the non-growing backbone (i.e., a model with limited network capacity) suffice to train on more incoming tasks, a graceful forgetting implementation on previously learned trivial tasks is empirically investigated. | ['Zhigang Zeng', 'Depeng Li'] | 2023-06-21 | null | null | null | null | ['class-incremental-learning', 'incremental-learning'] | ['computer-vision', 'methodology'] | [ 3.17635179e-01 -9.52491770e-04 4.50725481e-02 -2.99511641e-01
-2.05443546e-01 -4.66623753e-01 3.91618043e-01 2.85577625e-01
-8.78770590e-01 9.93228257e-01 -3.84760737e-01 -2.27811590e-01
-1.83249518e-01 -8.35775316e-01 -1.12487614e+00 -9.42011893e-01
3.15251462e-02 4.02027577e-01 4.08968866e-01 1.19354866e-01
2.16591835e-01 5.21685243e-01 -1.94492602e+00 3.08505505e-01
9.52463567e-01 1.17364562e+00 3.64619881e-01 2.79961765e-01
-1.37884412e-02 7.91225255e-01 -4.62858528e-01 -5.02901375e-01
1.76014960e-01 -9.60639641e-02 -7.25027978e-01 2.74539348e-02
4.28140879e-01 -1.94875821e-01 -3.43819797e-01 7.85233676e-01
3.38409066e-01 3.68137658e-01 3.42074484e-01 -1.02842057e+00
-7.38476753e-01 6.51843309e-01 -1.39178768e-01 2.10181922e-01
-2.93690920e-01 2.77157992e-01 4.51158017e-01 -1.24132800e+00
5.52468061e-01 5.23316026e-01 8.13888490e-01 6.35828137e-01
-1.38379884e+00 -8.60917449e-01 4.68680501e-01 3.60780329e-01
-1.50813127e+00 -8.20532382e-01 6.47689641e-01 -3.75867397e-01
1.11216629e+00 3.81355509e-02 8.23337555e-01 1.20128930e+00
6.30784512e-01 6.73565745e-01 8.67408633e-01 -2.80075043e-01
7.29793191e-01 3.68173927e-01 3.55105460e-01 7.47534275e-01
5.22525728e-01 -2.76008118e-02 -1.01316118e+00 4.63400632e-02
4.43989605e-01 4.91142482e-01 -2.47880489e-01 -8.31108928e-01
-1.00027883e+00 3.80825847e-01 3.73105764e-01 4.18037891e-01
-3.36425811e-01 -1.61650285e-01 3.65043283e-01 6.75206184e-01
4.58489507e-01 3.55336905e-01 -7.18812406e-01 1.52257219e-01
-1.15340877e+00 7.11490214e-02 7.16590226e-01 8.35780740e-01
1.02966225e+00 1.28178194e-01 -2.05434546e-01 6.96463168e-01
-3.22274119e-01 2.68046886e-01 9.30139601e-01 -5.00669062e-01
2.35197365e-01 6.03380203e-01 -1.07407682e-01 -7.24421203e-01
-4.91620928e-01 -1.10024118e+00 -1.10701561e+00 -5.19667345e-04
3.05770904e-01 9.12495181e-02 -1.04986107e+00 2.03550267e+00
1.77783534e-01 1.17583200e-01 6.95003420e-02 4.61036205e-01
3.20895702e-01 3.75129253e-01 -2.82314848e-02 -5.72770715e-01
1.09008384e+00 -8.53471935e-01 -4.78035569e-01 -5.78432381e-01
4.43135083e-01 -1.33186668e-01 1.15892720e+00 5.55607378e-01
-1.16781735e+00 -6.82601750e-01 -1.18729317e+00 -7.43450522e-02
-5.12262285e-01 -2.65022554e-02 3.88285428e-01 3.76850158e-01
-1.00157630e+00 6.60464764e-01 -8.04259539e-01 -1.01936005e-01
7.49534547e-01 5.42488217e-01 -5.03233790e-01 -1.86274454e-01
-1.05821478e+00 8.18902910e-01 7.43226290e-01 2.54096478e-01
-1.13215363e+00 -8.61152530e-01 -6.48212910e-01 4.58541691e-01
5.28936386e-01 -9.42127585e-01 1.12904763e+00 -9.65903640e-01
-1.50546336e+00 7.72181571e-01 -1.35731801e-01 -7.32743442e-01
5.70406735e-01 -3.65080774e-01 -1.66138828e-01 -3.25406730e-01
-2.48712525e-01 5.36441505e-01 1.44661975e+00 -1.13166070e+00
-5.10788321e-01 -4.69291240e-01 -2.63255358e-01 -2.24406049e-02
-8.99549365e-01 -6.93438292e-01 -1.46059200e-01 -4.94060934e-01
2.07143813e-01 -9.41765308e-01 9.52069983e-02 3.21877673e-02
-1.60332814e-01 8.11276659e-02 6.19935393e-01 -3.67120445e-01
1.36295044e+00 -2.39883113e+00 2.70136535e-01 -3.22799347e-02
4.37132478e-01 3.84071231e-01 -1.37467876e-01 5.99056147e-02
-4.28375378e-02 -2.15188682e-01 -4.45763230e-01 -6.85964465e-01
-2.76130408e-01 3.57655168e-01 -5.26182592e-01 3.48471999e-01
2.38661483e-01 9.74517822e-01 -8.30100417e-01 -8.32430497e-02
-2.43784636e-01 3.71411055e-01 -5.42492151e-01 1.76192522e-01
-1.89150259e-01 1.95664391e-01 5.56910336e-02 4.51520383e-01
7.09511340e-01 -3.44438314e-01 2.49061242e-01 1.00310273e-01
-1.27615720e-01 1.39852673e-01 -9.92549360e-01 1.85088730e+00
-5.62654495e-01 3.21168900e-01 -6.87664226e-02 -1.04843903e+00
9.23016071e-01 3.50969471e-02 1.37935253e-02 -7.61302769e-01
5.63440807e-02 2.77068406e-01 -3.32531519e-02 -1.49988323e-01
5.07291973e-01 -1.11086771e-01 7.69509152e-02 5.31336606e-01
4.19543535e-01 4.09669310e-01 -6.27361983e-02 3.90257351e-02
1.31215858e+00 -1.17937200e-01 2.67490864e-01 -3.06606174e-01
4.24769461e-01 -3.12861085e-01 7.95120299e-01 9.95074809e-01
-1.16090871e-01 3.98017108e-01 8.82509649e-02 -8.75070632e-01
-8.35343719e-01 -1.22998941e+00 -4.57360446e-02 1.32992744e+00
-2.09483996e-01 -1.48598835e-01 -5.76641381e-01 -6.05474591e-01
9.30121467e-02 8.16906750e-01 -8.25529099e-01 -9.73530650e-01
-6.90252960e-01 -6.33086801e-01 1.92976162e-01 2.95949727e-01
5.79410374e-01 -1.24462330e+00 -1.02027726e+00 4.14002329e-01
1.37711689e-01 -5.76070189e-01 -3.40241998e-01 8.69028866e-01
-1.21421933e+00 -7.94420004e-01 -6.53252482e-01 -8.04489672e-01
7.89518595e-01 3.52908969e-01 9.41264212e-01 1.35374278e-01
-2.99474865e-01 2.66406000e-01 -3.09063718e-02 -2.65214473e-01
-1.86029170e-02 8.23185265e-01 3.76989126e-01 3.28632921e-01
2.61874378e-01 -9.51522887e-01 -5.84537983e-01 1.14960121e-02
-1.05019999e+00 1.06757767e-01 8.32467914e-01 1.23112833e+00
7.01537490e-01 7.59675056e-02 7.76706874e-01 -9.23583984e-01
4.99268711e-01 -5.52436054e-01 -4.23311144e-01 3.35025817e-01
-9.43805039e-01 2.89371908e-01 7.71140814e-01 -9.76923108e-01
-1.06162167e+00 3.05014998e-01 3.73285085e-01 -6.11697912e-01
1.68194652e-01 4.75759655e-01 -7.08853155e-02 -1.34470193e-02
6.58707857e-01 8.55915368e-01 4.46399339e-02 -4.51425523e-01
2.64247090e-01 2.16621742e-01 6.79188848e-01 -5.19090116e-01
6.71126962e-01 2.30573550e-01 -3.16454291e-01 -5.24741113e-01
-9.57797706e-01 6.42039701e-02 -8.30515504e-01 -1.33947760e-01
2.90002674e-01 -9.25984383e-01 -7.01929212e-01 8.78953516e-01
-1.12029493e+00 -5.13662279e-01 -7.51470327e-01 1.88148111e-01
-4.63124126e-01 -2.19515234e-01 -6.07651412e-01 -6.69545114e-01
-3.96954507e-01 -8.47937703e-01 5.78767240e-01 9.02756900e-02
-3.47991660e-02 -6.63139284e-01 1.20433263e-01 -4.02330384e-02
7.20123529e-01 -3.54060352e-01 1.20168304e+00 -6.22190118e-01
-7.53371596e-01 -1.00128613e-01 2.07951255e-02 4.21375334e-01
-5.33582643e-02 -5.87389708e-01 -1.23837852e+00 -7.32253432e-01
3.99702936e-01 -4.18854922e-01 1.36162627e+00 -5.00942357e-02
1.35734642e+00 -6.17507517e-01 -3.42117608e-01 5.19871056e-01
1.10328305e+00 2.01874048e-01 5.78023851e-01 3.77436429e-01
3.82624865e-01 3.95078987e-01 1.11276619e-01 4.67001945e-01
2.67338574e-01 2.64426351e-01 3.99369419e-01 4.67946947e-01
-5.12399450e-02 -2.20495909e-01 3.77672434e-01 1.01739514e+00
1.63604498e-01 -7.46148825e-02 -8.90926659e-01 6.14017725e-01
-1.98318434e+00 -8.38516831e-01 6.32317483e-01 2.59589386e+00
1.25178993e+00 3.77141207e-01 -3.80440295e-01 1.77384213e-01
3.39313656e-01 3.06639746e-02 -1.18686414e+00 6.69361465e-03
-1.51100412e-01 3.13810557e-01 1.81476220e-01 2.78687924e-01
-9.06244159e-01 9.33093250e-01 5.56152153e+00 6.79883003e-01
-1.29279542e+00 4.05844420e-01 6.22382820e-01 -5.45121312e-01
-1.71859115e-01 -1.05764270e-01 -9.06058192e-01 3.56160551e-01
1.01232255e+00 -2.84252882e-01 6.10289037e-01 8.85029972e-01
-4.81869370e-01 -1.89171627e-01 -1.24553871e+00 1.00407946e+00
1.08778499e-01 -1.44790196e+00 1.34534150e-01 -1.20504491e-01
4.60273385e-01 -4.62571457e-02 4.93938327e-01 7.09034264e-01
-5.61275259e-02 -7.15209782e-01 1.01533985e+00 1.03774750e+00
7.18905926e-01 -6.64340556e-01 5.56281626e-01 6.85721755e-01
-7.94646859e-01 -5.14700592e-01 -4.21748102e-01 -2.28309333e-01
-2.73510844e-01 8.69968891e-01 -5.59505820e-01 2.31415585e-01
8.47044349e-01 5.11783719e-01 -9.26381111e-01 9.13850605e-01
1.45234346e-01 4.84740525e-01 -2.32526854e-01 2.83442348e-01
-2.13967890e-01 1.92574769e-01 2.94505656e-01 9.02445436e-01
4.23630148e-01 -6.77549541e-02 -5.60191125e-02 7.37330496e-01
-3.55661303e-01 -1.03597432e-01 -6.28171980e-01 2.09928498e-01
6.88665330e-01 9.84180152e-01 -6.27002299e-01 -2.13339016e-01
-1.14195123e-02 1.13875508e+00 1.08511734e+00 3.42488199e-01
-4.49075937e-01 -2.34210506e-01 4.99423921e-01 2.74807304e-01
5.61895072e-01 -3.92364025e-01 -4.49545056e-01 -1.36528957e+00
3.59591633e-01 -6.36792839e-01 3.08239222e-01 -5.00342309e-01
-1.14348376e+00 9.01481926e-01 -2.90804654e-01 -9.07537162e-01
-1.08376004e-01 -3.88740361e-01 -4.18692678e-01 4.12685812e-01
-1.44612408e+00 -8.90290022e-01 -2.82151490e-01 6.64737999e-01
4.72944260e-01 -3.79248559e-01 8.14598799e-01 3.34890217e-01
-8.83253813e-01 1.02998245e+00 2.24799678e-01 -3.77248704e-01
8.48019958e-01 -7.38848031e-01 8.43729451e-03 7.69764125e-01
1.44418046e-01 9.60646510e-01 4.58075225e-01 -6.52218640e-01
-1.54948640e+00 -1.26807988e+00 1.02658844e+00 -2.73092479e-01
4.44819421e-01 -9.26002026e-01 -1.57838643e+00 7.62163758e-01
-1.61411360e-01 9.89876017e-02 5.86897492e-01 2.26538450e-01
-6.70572579e-01 -5.84105313e-01 -9.46976781e-01 4.59579557e-01
1.29088759e+00 -5.42006731e-01 -5.49099565e-01 1.45361364e-01
8.74372184e-01 -1.97638795e-01 -5.10036945e-01 1.84616417e-01
6.39752388e-01 -8.92776430e-01 7.34897971e-01 -5.96430361e-01
1.45163648e-02 -1.60870880e-01 6.21766448e-02 -1.18369341e+00
-5.04951000e-01 -5.77560484e-01 -6.27176106e-01 1.02206063e+00
4.09160674e-01 -8.45154047e-01 7.17349231e-01 5.62231541e-01
-1.59492090e-01 -8.76657724e-01 -1.24645996e+00 -9.63776767e-01
9.96612981e-02 -2.65954375e-01 5.65995276e-01 9.22528505e-01
-2.56182611e-01 5.12635350e-01 -4.44898248e-01 -4.61794101e-02
4.76506561e-01 3.40977386e-02 5.57438552e-01 -1.41874111e+00
-3.13233256e-01 -3.43962789e-01 -3.96088995e-02 -7.12731004e-01
2.85185158e-01 -1.00990117e+00 1.56716064e-01 -8.08939397e-01
3.37017298e-01 -5.98706126e-01 -8.10693562e-01 9.26342070e-01
-7.51468092e-02 -6.56654462e-02 2.93855727e-01 4.60143864e-01
-8.52758467e-01 9.13142085e-01 8.22156966e-01 -2.33838886e-01
-4.62465852e-01 -2.07888875e-02 -6.47862852e-01 4.59614158e-01
7.42613018e-01 -8.31902742e-01 -6.38430834e-01 -6.53483629e-01
4.73050892e-01 -2.90263146e-01 4.25533623e-01 -1.43367910e+00
8.07636201e-01 6.42449707e-02 5.44345856e-01 -4.13380384e-01
3.14708918e-01 -7.98329473e-01 4.31240797e-01 6.98869467e-01
-3.74551654e-01 3.34793217e-02 3.25062662e-01 8.06009769e-01
2.79603489e-02 -2.98266917e-01 7.40465522e-01 -1.30937219e-01
-5.06139636e-01 4.70648468e-01 -4.10424620e-01 -1.53897464e-01
9.86612201e-01 -1.87594339e-01 -3.33467156e-01 1.30178362e-01
-9.72698212e-01 5.46216369e-02 4.70793754e-01 4.63204771e-01
7.81379223e-01 -1.13123786e+00 -3.50512564e-01 6.77963614e-01
1.45011157e-01 2.22456649e-01 5.80713272e-01 7.19745576e-01
1.85710147e-01 2.58373260e-01 -3.64979714e-01 -4.35963839e-01
-8.11077178e-01 9.35956836e-01 2.26334229e-01 -4.03668314e-01
-5.58714926e-01 8.92339647e-01 2.87644684e-01 -2.87537456e-01
5.08166850e-01 -2.56580889e-01 2.01968864e-01 3.48188072e-01
8.01131964e-01 1.71226263e-01 5.08402169e-01 2.79382151e-03
-3.31734657e-01 5.01999892e-02 -8.32707345e-01 1.94037557e-01
1.49532676e+00 -1.02406353e-01 -2.18804076e-01 8.78894389e-01
9.43117499e-01 -4.58520502e-01 -1.54472899e+00 -7.34782636e-01
2.62067514e-03 -8.06558579e-02 -8.41794815e-03 -7.98173547e-01
-1.15643728e+00 1.00859213e+00 7.15111792e-01 -1.92188933e-01
1.25652254e+00 -3.16063225e-01 7.52360761e-01 8.70786309e-01
7.55902469e-01 -1.07932866e+00 3.25407892e-01 9.02195334e-01
8.79729509e-01 -8.82313073e-01 -6.23516776e-02 2.03568637e-01
-2.98064739e-01 1.08533835e+00 8.16785395e-01 3.06966230e-02
8.34148884e-01 1.57503664e-01 -4.65603083e-01 2.30649449e-02
-1.14069510e+00 1.76687166e-01 7.94637501e-02 3.57072115e-01
-8.59292373e-02 -1.98470026e-01 1.72185414e-02 1.02804041e+00
-6.69101998e-02 3.52060586e-01 3.01332772e-01 1.25820291e+00
-6.58418238e-01 -7.94083059e-01 3.31594944e-02 5.84038198e-01
3.30946408e-02 -2.98665196e-01 -1.11941449e-01 4.88197714e-01
2.64945447e-01 3.58531564e-01 2.51905352e-01 -5.12173712e-01
2.79597938e-01 6.36120915e-01 4.21395361e-01 -6.72404230e-01
-7.80815065e-01 -4.29837912e-01 -4.30511028e-01 -4.86875087e-01
9.84934345e-02 -6.71632171e-01 -8.62637758e-01 -4.07187670e-01
-2.05630988e-01 -1.03788957e-01 3.02815020e-01 8.63324523e-01
8.42265964e-01 5.46783268e-01 3.89915437e-01 -7.28721023e-01
-7.58850276e-01 -7.07300305e-01 -5.95565081e-01 9.56681743e-02
4.38239753e-01 -8.46589983e-01 -2.46025413e-01 9.91460755e-02] | [9.817081451416016, 3.4233343601226807] |
c24b9ea1-1fa9-4f93-b7be-627e8482f1cc | a-novel-workflow-for-accurately-and | null | null | https://aclanthology.org/2020.findings-emnlp.38 | https://aclanthology.org/2020.findings-emnlp.38.pdf | A Novel Workflow for Accurately and Efficiently Crowdsourcing Predicate Senses and Argument Labels | Resources for Semantic Role Labeling (SRL) are typically annotated by experts at great expense. Prior attempts to develop crowdsourcing methods have either had low accuracy or required substantial expert annotation. We propose a new multi-stage crowd workflow that substantially reduces expert involvement without sacrificing accuracy. In particular, we introduce a unique filter stage based on the key observation that crowd workers are able to almost perfectly filter out incorrect options for labels. Our three-stage workflow produces annotations with 95{\%} accuracy for predicate labels and 93{\%} for argument labels, which is comparable to expert agreement. Compared to prior work on crowdsourcing for SRL, we decrease expert effort by 4x, from 56{\%} to 14{\%} of cases. Our approach enables more scalable annotation of SRL, and could enable annotation of NLP tasks that have previously been considered too complex to effectively crowdsource. | ['Walter Lasecki', 'Yunyao Li', 'Jonathan K. Kummerfeld', 'Huaiyu Zhu', 'Youxuan Jiang'] | 2020-11-01 | null | null | null | findings-of-the-association-for-computational | ['semantic-role-labeling'] | ['natural-language-processing'] | [ 4.69558090e-01 8.36752295e-01 1.08055361e-01 -4.28099036e-01
-1.27179527e+00 -1.35925627e+00 5.96901238e-01 6.63213193e-01
-8.67858171e-01 1.15417981e+00 4.20129389e-01 -2.70269722e-01
2.64740318e-01 -3.69508892e-01 -4.70743001e-01 2.78624855e-02
6.81607425e-01 9.54937756e-01 7.02858031e-01 -2.58568645e-01
8.28521922e-02 1.49834514e-01 -1.51326632e+00 5.24954677e-01
7.61457205e-01 6.33891940e-01 -1.88445434e-01 5.76714575e-01
-4.92423385e-01 1.24046421e+00 -9.75690842e-01 -7.66901314e-01
2.02679500e-01 -1.97949767e-01 -1.37469828e+00 -5.39468490e-02
4.33169782e-01 -2.62716800e-01 3.67178708e-01 8.57033193e-01
5.71470797e-01 3.16654652e-01 2.70785779e-01 -1.24901521e+00
-6.29667401e-01 6.82736278e-01 -3.21888179e-01 -2.61509754e-02
9.67924714e-01 1.13866702e-01 1.21674109e+00 -8.78865719e-01
8.14796507e-01 1.14566267e+00 9.15725231e-01 7.61307240e-01
-1.18272221e+00 -5.04003882e-01 1.69432148e-01 -4.12750006e-01
-1.31317079e+00 -5.34470439e-01 1.95392072e-01 -8.39811087e-01
9.45350826e-01 3.92197937e-01 3.40014577e-01 5.43805659e-01
-8.57719183e-01 6.96642876e-01 1.29829752e+00 -4.90173638e-01
3.26250851e-01 3.03535968e-01 -7.20782205e-02 5.00034511e-01
4.83677655e-01 -8.75115275e-01 -7.97264934e-01 -6.08766139e-01
5.31560183e-01 -4.68054414e-01 -6.36741817e-02 2.90755451e-01
-1.08873689e+00 5.88229597e-01 -1.82712391e-01 4.05407399e-01
-1.13182090e-01 1.41905144e-01 3.01771879e-01 -1.96446981e-02
6.62483692e-01 8.91216636e-01 -4.89235759e-01 -3.63666445e-01
-7.83468485e-01 5.63941240e-01 1.15207481e+00 1.26745725e+00
6.98083878e-01 -2.89407969e-01 -3.64080012e-01 8.58083427e-01
1.10463992e-01 3.53562742e-01 -1.93482097e-02 -1.49263728e+00
4.63849962e-01 1.16141164e+00 1.35769105e+00 -7.24614382e-01
-3.94471854e-01 3.07268947e-01 1.09275118e-01 1.15684696e-01
1.01429915e+00 -2.92803705e-01 -3.83800745e-01 1.40856409e+00
5.76440394e-01 -3.72409701e-01 -1.73292056e-01 8.75310361e-01
8.43308747e-01 -7.07356781e-02 7.25903392e-01 1.12090427e-02
1.69109154e+00 -7.53166258e-01 -7.63247013e-01 -2.51112759e-01
9.31779087e-01 -9.11052406e-01 1.50691080e+00 8.71417448e-02
-1.31171894e+00 -2.28068650e-01 -5.71112871e-01 -3.29217702e-01
-2.41105333e-01 1.02464832e-01 6.14351332e-01 8.91938090e-01
-1.17904890e+00 5.83188355e-01 -5.72220981e-01 -4.15492982e-01
4.90134358e-01 3.56937200e-01 -3.87336075e-01 9.79844760e-03
-1.12058723e+00 9.19527531e-01 2.30221048e-01 -2.30970651e-01
-5.89468837e-01 -7.15397835e-01 -7.88713694e-01 -3.23758334e-01
9.11533892e-01 -1.55732617e-01 1.83257854e+00 -8.24697614e-01
-1.20930207e+00 1.23777509e+00 -5.36607862e-01 -1.65966317e-01
8.58773768e-01 -3.87383252e-01 -1.98351219e-01 3.20071429e-01
8.02070439e-01 6.90654397e-01 1.47974730e-01 -1.28467572e+00
-8.85585964e-01 2.14569718e-02 6.05636418e-01 3.71618092e-01
-2.68434942e-01 7.66973734e-01 -1.59422040e-01 -5.86168170e-01
-2.44469941e-01 -9.14916635e-01 -3.09090376e-01 9.40555334e-02
4.24647257e-02 -6.33435071e-01 2.79836148e-01 -8.05457413e-01
1.02004051e+00 -1.63151729e+00 -4.42479700e-01 -2.62617350e-01
5.86096108e-01 5.91271579e-01 1.39022380e-01 4.78723437e-01
3.82923037e-01 7.17353821e-01 -3.34314972e-01 -4.31280524e-01
2.02160582e-01 2.52136111e-01 -1.44984841e-01 2.73549139e-01
4.04698044e-01 8.54679227e-01 -1.42987573e+00 -7.94652820e-01
-1.34999260e-01 1.61233202e-01 -2.87446886e-01 6.65735379e-02
-5.37224770e-01 7.12532878e-01 -3.83277863e-01 8.78658175e-01
3.71248633e-01 -5.38057089e-01 4.72834259e-01 4.39135104e-01
-2.90322512e-01 3.71025771e-01 -1.31508505e+00 1.48995984e+00
-1.26411706e-01 3.78561109e-01 3.52744102e-01 -1.40064552e-01
7.93893278e-01 7.11794198e-01 4.98408318e-01 -2.01295182e-01
-9.39674824e-02 5.97520947e-01 -3.72953117e-01 -4.53718036e-01
6.59035742e-01 -1.73358038e-01 -2.18917102e-01 9.00819838e-01
-1.94999754e-01 -3.21643800e-01 3.03368300e-01 1.94563136e-01
1.33725584e+00 3.68389219e-01 4.00996208e-01 -2.52034754e-01
5.84885240e-01 6.60732985e-01 7.31701195e-01 9.60561693e-01
-5.97536862e-01 4.29451227e-01 5.41676283e-01 -6.44056857e-01
-1.06774342e+00 -3.70424122e-01 3.10394675e-01 1.61746562e+00
7.74885938e-02 -4.14775342e-01 -1.04987073e+00 -1.07676923e+00
-8.76864344e-02 6.23289466e-01 -3.61200631e-01 6.85385704e-01
-7.26913035e-01 -4.25850570e-01 1.13966286e+00 6.91602647e-01
3.87184173e-01 -1.11374629e+00 -7.77224123e-01 4.25381780e-01
-4.07080203e-01 -1.56133080e+00 -2.59564847e-01 -3.01007122e-01
-1.03621922e-01 -1.30272031e+00 -6.17189825e-01 -4.99016255e-01
7.46379554e-01 1.82134241e-01 1.35040355e+00 4.76963252e-01
-3.39329764e-02 3.45945865e-01 -6.25285923e-01 -7.80403078e-01
-3.39493662e-01 -2.92005073e-02 -1.00727193e-01 -3.21533203e-01
8.63917053e-01 -1.14331298e-01 -5.68546712e-01 6.10606790e-01
-8.00898790e-01 -2.83172995e-01 -2.30113287e-02 2.93347806e-01
5.06495953e-01 -4.08008754e-01 7.65788972e-01 -1.36648941e+00
8.45786631e-01 -3.83403569e-01 -5.17094135e-01 3.02586645e-01
-6.21645212e-01 -3.35521609e-01 4.67518508e-01 -2.11647153e-01
-1.23375738e+00 3.08854103e-01 -1.81999132e-02 2.15470344e-01
-5.18140674e-01 -7.11132661e-02 1.40131535e-02 -1.77600428e-01
1.03315032e+00 -5.97477794e-01 -2.53216356e-01 -5.58546603e-01
2.50149786e-01 5.58788717e-01 2.19006523e-01 -8.87936234e-01
6.27764285e-01 8.24250281e-01 -4.09209311e-01 -3.50250572e-01
-1.27600741e+00 -5.89425325e-01 -6.44578993e-01 -5.68316057e-02
9.77180541e-01 -1.16491914e+00 -1.05864918e+00 -5.93064725e-02
-1.25343585e+00 -7.42212951e-01 -5.88181794e-01 6.85181469e-02
-2.05745324e-01 4.89308685e-01 -4.92567003e-01 -1.13048601e+00
-2.15716496e-01 -8.91983628e-01 1.36526394e+00 2.27489591e-01
-1.02238977e+00 -8.21212590e-01 -1.22551382e-01 9.22875822e-01
3.49492639e-01 3.65867704e-01 2.20996112e-01 -1.21264911e+00
-1.62045896e-01 -4.75319326e-01 -3.56523097e-01 4.46073003e-02
-1.64934576e-01 -2.77872771e-01 -1.04442585e+00 2.17707828e-02
-3.98139209e-01 -7.43805170e-01 8.90857652e-02 -2.84138799e-01
6.74270451e-01 -2.68210292e-01 -3.21436077e-01 -4.13865954e-01
9.59094942e-01 1.25416428e-01 1.56749591e-01 4.33227837e-01
6.72719359e-01 9.66084063e-01 9.43364024e-01 4.92619991e-01
7.21453249e-01 4.74924177e-01 -4.76865359e-02 -6.70047700e-02
1.24631099e-01 -2.70193994e-01 -8.75838920e-02 -5.88529604e-03
-3.83468717e-01 -4.41761136e-01 -1.41019547e+00 8.00283551e-01
-2.07612729e+00 -5.75658321e-01 -2.69158810e-01 1.71007943e+00
1.01456320e+00 1.84263464e-03 3.07623476e-01 1.39715150e-01
1.01614022e+00 8.24802890e-02 -2.07979023e-01 -1.43734396e-01
-7.19140694e-02 3.98726314e-01 5.87051809e-01 7.34742343e-01
-9.38412726e-01 1.37458873e+00 6.56298637e+00 4.87016380e-01
-2.99358189e-01 7.30066419e-01 2.89546967e-01 -1.37695760e-01
-5.29344916e-01 3.97055626e-01 -1.04276156e+00 4.74687070e-01
6.20627224e-01 -9.89868306e-03 3.78693372e-01 8.55764985e-01
1.27758190e-01 -3.13813925e-01 -9.18461382e-01 6.65757596e-01
-1.42041510e-02 -1.34613705e+00 -3.11793387e-01 -1.56962216e-01
9.07888293e-01 -2.29009673e-01 -6.89023077e-01 1.92707896e-01
1.11985600e+00 -1.15000641e+00 9.94317234e-01 2.98205703e-01
1.02690983e+00 -3.93298149e-01 8.82411480e-01 5.65096557e-01
-1.09537888e+00 1.43883433e-02 -1.92329034e-01 -4.86368746e-01
5.37397623e-01 4.68428552e-01 -1.24870300e+00 1.47766382e-01
7.77556539e-01 -3.19181047e-02 -5.20443916e-01 5.03248751e-01
-7.85236835e-01 5.21569014e-01 -2.63075501e-01 -4.31586355e-01
9.79922116e-02 2.97970772e-01 4.97975677e-01 1.27648330e+00
-2.37404972e-01 4.99117136e-01 6.26716614e-01 5.02272666e-01
-5.17728329e-01 1.93435460e-01 -4.00926620e-01 -1.69177726e-01
1.05815721e+00 1.21225309e+00 -1.02043486e+00 -4.95779008e-01
-2.17563674e-01 7.94682801e-01 4.34281290e-01 2.03335091e-01
-5.79488695e-01 -2.74709076e-01 3.59022737e-01 4.49186832e-01
-8.17651227e-02 -1.38201639e-01 -5.35097420e-01 -8.27113390e-01
3.74887943e-01 -8.88067782e-01 3.96633506e-01 -7.17557371e-01
-9.37240243e-01 6.22213781e-01 -3.72255556e-02 -7.66820252e-01
-1.93201974e-01 -4.66508687e-01 -9.53932703e-02 1.02474606e+00
-1.20851123e+00 -1.25382853e+00 -5.17627180e-01 2.55097330e-01
3.25694621e-01 1.00621037e-01 9.55900371e-01 2.39112109e-01
-1.53105184e-01 5.48725426e-01 -7.85963178e-01 2.54984200e-01
9.34211671e-01 -1.41888499e+00 6.00976110e-01 7.53871381e-01
-2.68209875e-01 7.57615507e-01 6.94754481e-01 -9.89881694e-01
-7.03660727e-01 -1.08724332e+00 1.47886753e+00 -1.33772850e+00
6.63968325e-01 -3.93879026e-01 -6.93430662e-01 8.21663141e-01
3.44162248e-02 2.05007330e-01 1.05261910e+00 -1.14185333e-01
-3.31677347e-01 4.96172756e-01 -1.52066958e+00 5.08015454e-01
1.35628617e+00 -7.87284851e-01 -5.44805169e-01 7.86693513e-01
9.37396467e-01 -5.49188197e-01 -9.33468699e-01 1.93019450e-01
3.61335933e-01 -5.10660887e-01 5.56688726e-01 -5.89323342e-01
1.91000566e-01 -7.72463262e-01 -6.40767217e-02 -6.90899074e-01
-1.43265694e-01 -9.02420104e-01 2.03193322e-01 1.53978574e+00
6.81078851e-01 -6.14664078e-01 6.90185905e-01 1.66278148e+00
-8.79830196e-02 -3.65091145e-01 -6.65109515e-01 -4.46774065e-01
-2.84077317e-01 -4.69169796e-01 7.70967960e-01 1.19872940e+00
-6.62781671e-02 1.67518407e-01 -1.52581245e-01 9.13809687e-02
1.96166649e-01 -4.15356129e-01 8.08733702e-01 -1.40040278e+00
-1.22545496e-01 -1.00970320e-01 9.99954790e-02 -3.90258610e-01
3.14674675e-01 -6.45667553e-01 1.11595564e-01 -2.03070784e+00
-5.55045940e-02 -7.25116134e-01 2.27723405e-01 1.05287695e+00
-4.90595937e-01 7.45250106e-01 1.11028068e-01 4.85540748e-01
-1.15019095e+00 -2.89628088e-01 8.91173065e-01 3.57273757e-01
-2.24720865e-01 -3.43563318e-01 -1.35891151e+00 1.08474600e+00
6.72841072e-01 -6.51435792e-01 -2.13191122e-01 -8.22835803e-01
8.14077556e-01 -4.66232747e-01 2.85158515e-01 -6.26555383e-01
3.63496840e-01 -2.72908270e-01 1.50207445e-01 1.72976649e-03
9.21891928e-02 -4.87808138e-01 1.53147787e-01 -7.10638091e-02
-5.46507776e-01 -5.85581400e-02 1.10239215e-01 5.25985301e-01
-9.83557552e-02 -6.54419959e-01 4.39990044e-01 -8.68293762e-01
-6.21210635e-01 -2.06935570e-01 -4.15413916e-01 5.41904271e-01
1.14437962e+00 -2.73260802e-01 -5.30800521e-01 -2.76386887e-01
-7.23770201e-01 3.53751808e-01 8.84250283e-01 -1.09285079e-01
-1.58879712e-01 -9.11437571e-01 -6.70988023e-01 -6.74654126e-01
1.26179904e-01 4.17930394e-01 -8.46872255e-02 4.82242465e-01
-7.59159207e-01 1.44313052e-01 1.71527684e-01 -1.14059195e-01
-1.07937407e+00 1.92195565e-01 -1.19541781e-02 -3.74117464e-01
-2.77688742e-01 1.17374349e+00 -4.93918747e-01 -6.46890402e-01
2.92841028e-02 1.33561939e-01 -3.66233855e-01 4.10644740e-01
7.07736373e-01 7.94391930e-01 4.05987315e-02 -6.97947621e-01
-5.92923760e-01 2.35464901e-01 2.02855587e-01 -6.50137603e-01
1.14502251e+00 -1.12441154e-02 -2.71763295e-01 2.86455512e-01
3.65868151e-01 5.69077373e-01 -1.16341281e+00 -1.42712787e-01
5.18428624e-01 -4.77899700e-01 -6.96770430e-01 -1.05640686e+00
-5.86306751e-02 4.39138114e-01 -6.49528280e-02 5.34183443e-01
4.89310592e-01 1.68344185e-01 6.79268777e-01 3.39636534e-01
6.85284078e-01 -1.50111639e+00 -1.70784052e-02 4.13881212e-01
6.91831946e-01 -1.42998970e+00 1.27813950e-01 -7.45753229e-01
-9.55532610e-01 5.94879210e-01 7.51678050e-01 3.08293626e-02
-1.78764444e-02 3.39593977e-01 2.64134109e-01 -4.00801957e-01
-3.61605406e-01 -3.32831889e-01 -8.06990191e-02 6.52630329e-01
7.34920800e-01 1.55638978e-01 -6.22444093e-01 8.46156478e-01
6.70231180e-04 1.00128427e-01 4.56039757e-01 1.47117889e+00
-4.65290099e-01 -1.14037657e+00 -5.25711000e-01 1.84570268e-01
-9.52785313e-01 -2.02601366e-02 -9.62698817e-01 6.28608704e-01
4.63013530e-01 1.48030400e+00 -1.69141278e-01 7.76005611e-02
6.16454244e-01 3.88569236e-01 5.63133620e-02 -1.17390835e+00
-1.13808978e+00 -1.55739501e-01 9.32635427e-01 -5.73426545e-01
-8.76536787e-01 -6.58575296e-01 -1.46561980e+00 1.77959912e-02
-1.64651766e-01 2.75356501e-01 3.47038358e-01 1.23505497e+00
4.40230012e-01 -3.02687776e-03 -1.64672866e-01 -6.28069818e-01
-3.01500678e-01 -8.93079102e-01 -1.01883188e-01 7.30075002e-01
-1.79660708e-01 -6.73864603e-01 -1.50937736e-01 2.47725800e-01] | [9.705842018127441, 4.916252136230469] |
54d8323f-2449-4962-b84e-95d672fb2b38 | multi-view-graph-representation-learning-for | 2305.03458 | null | https://arxiv.org/abs/2305.03458v1 | https://arxiv.org/pdf/2305.03458v1.pdf | Multi-View Graph Representation Learning for Answering Hybrid Numerical Reasoning Question | Hybrid question answering (HybridQA) over the financial report contains both textual and tabular data, and requires the model to select the appropriate evidence for the numerical reasoning task. Existing methods based on encoder-decoder framework employ a expression tree-based decoder to solve numerical reasoning problems. However, encoders rely more on Machine Reading Comprehension (MRC) methods, which take table serialization and text splicing as input, damaging the granularity relationship between table and text as well as the spatial structure information of table itself. In order to solve these problems, the paper proposes a Multi-View Graph (MVG) Encoder to take the relations among the granularity into account and capture the relations from multiple view. By utilizing MVGE as a module, we constuct Tabular View, Relation View and Numerical View which aim to retain the original characteristics of the hybrid data. We validate our model on the publicly available table-text hybrid QA benchmark (TAT-QA) and outperform the state-of-the-art model. | ['Kang Liu', 'Jun Zhao', 'Yuanzhe Zhang', 'Fangyu Lei', 'Yifan Wei'] | 2023-05-05 | null | null | null | null | ['reading-comprehension', 'machine-reading-comprehension'] | ['natural-language-processing', 'natural-language-processing'] | [-1.76287144e-01 3.11299801e-01 -1.59200788e-01 -3.68253469e-01
-1.13068044e+00 -7.44671106e-01 4.17904019e-01 4.47624445e-01
-1.21058807e-01 6.31737530e-01 4.81850564e-01 -6.30026877e-01
-9.92041305e-02 -1.44646585e+00 -9.18046713e-01 2.59797312e-02
5.30044079e-01 8.36767375e-01 3.10907811e-01 -6.68829441e-01
3.01569581e-01 -1.32312909e-01 -1.18491673e+00 1.01337016e+00
1.26575387e+00 1.28275168e+00 2.48977486e-02 6.17484987e-01
-6.91465497e-01 1.51972210e+00 -5.15267372e-01 -1.31030893e+00
6.40982538e-02 -5.64720213e-01 -1.04859030e+00 -1.33435041e-01
3.87396961e-01 -4.52274024e-01 -3.14444244e-01 1.04991961e+00
2.86230683e-01 -3.01264197e-01 7.74986327e-01 -1.11911845e+00
-7.84209609e-01 9.43228364e-01 -3.69874716e-01 1.96692437e-01
7.39168167e-01 -9.27172005e-02 1.59349072e+00 -6.55008197e-01
7.32101440e-01 1.37767255e+00 4.74517852e-01 1.91499054e-01
-6.79694772e-01 -1.50184005e-01 1.07315838e-01 9.85234916e-01
-1.13196576e+00 -1.79740459e-01 7.61452019e-01 -1.99000791e-01
1.30454636e+00 5.25785148e-01 5.25749803e-01 8.10805976e-01
5.65319955e-01 8.09426188e-01 1.07927239e+00 -3.45471114e-01
5.55788465e-02 1.62806306e-02 1.47736728e-01 9.74471450e-01
2.65908897e-01 -5.54436505e-01 -6.31589830e-01 1.83766738e-01
5.12174904e-01 -4.88867432e-01 -2.78964549e-01 -3.11024189e-01
-1.15452182e+00 7.55482793e-01 2.54982591e-01 -6.10267818e-02
-2.99589783e-01 -7.75982812e-02 6.96473300e-01 4.15673584e-01
4.08194028e-02 3.53836596e-01 -7.09837198e-01 -2.60806710e-01
-7.77673125e-01 4.37000841e-01 1.24167240e+00 1.37852550e+00
5.74982762e-01 -5.31836450e-01 -6.61319792e-01 5.23754179e-01
3.69935542e-01 5.26782632e-01 2.64457732e-01 -8.26922774e-01
1.54932189e+00 1.29301047e+00 -2.08447337e-01 -1.06674135e+00
-1.79905072e-01 -3.98551404e-01 -8.26238930e-01 -2.98423082e-01
3.54697227e-01 1.37130603e-01 -6.55420899e-01 1.31959403e+00
2.58235693e-01 -5.58224022e-01 4.14630324e-01 7.33626723e-01
1.17782009e+00 6.64720893e-01 -3.37450325e-01 -1.30760133e-01
1.93417811e+00 -1.03582418e+00 -1.24646688e+00 -2.27642395e-02
1.09175575e+00 -3.63119096e-01 1.26373756e+00 5.33255219e-01
-1.35703290e+00 -3.04002225e-01 -1.31258345e+00 -8.82319391e-01
-7.93911636e-01 1.95791245e-01 3.62109244e-01 5.58107734e-01
-6.32544041e-01 1.44872472e-01 -4.74122494e-01 1.80283919e-01
1.81020945e-01 -4.07187566e-02 -2.93840528e-01 -3.69848341e-01
-1.74215877e+00 1.04628229e+00 5.97853303e-01 4.09495354e-01
-3.34681749e-01 -5.56627333e-01 -1.11386716e+00 5.45406938e-01
1.08753359e+00 -1.16678882e+00 1.21370828e+00 -2.20557332e-01
-1.37789190e+00 4.52862203e-01 -1.23937786e-01 -4.79022801e-01
6.84499919e-01 -1.09062418e-01 -4.20203328e-01 3.43847483e-01
1.81494460e-01 6.46844655e-02 1.39176607e-01 -9.76916492e-01
-4.55128819e-01 -5.60042858e-01 5.38586974e-01 4.75034505e-01
2.13831946e-01 -1.94668949e-01 -7.05247998e-01 -4.07242507e-01
2.11740434e-01 -3.09321702e-01 4.26462680e-01 -3.42233628e-01
-6.67490542e-01 -4.01323944e-01 3.10829699e-01 -1.32026184e+00
1.73967111e+00 -1.63910544e+00 6.30331099e-01 1.39124349e-01
2.44487122e-01 -3.04469168e-01 2.58301646e-01 7.38648415e-01
2.88698018e-01 1.18190311e-01 -3.46950948e-01 -1.66603178e-01
4.30608839e-01 5.07870138e-01 -2.92982727e-01 -1.47955403e-01
1.84008867e-01 1.26362610e+00 -6.06387615e-01 -1.08774686e+00
-1.91404864e-01 -2.73060143e-01 -7.80285299e-01 3.65412146e-01
-7.92085409e-01 -7.11463913e-02 -4.48563993e-01 7.76441157e-01
8.10159504e-01 -4.63820100e-01 4.31956768e-01 -5.35949886e-01
2.43101522e-01 6.75608873e-01 -1.27697718e+00 1.86623335e+00
-4.23899472e-01 5.17123677e-02 -2.26204902e-01 -6.12698555e-01
6.71826720e-01 2.54155785e-01 -5.03898934e-02 -9.87005115e-01
-8.44472572e-02 3.17507088e-01 4.22732942e-02 -8.47358644e-01
5.30396700e-01 -4.49034013e-02 -2.42031470e-01 2.31914535e-01
1.66954458e-01 -3.39890122e-01 5.85370779e-01 6.98522568e-01
1.02617371e+00 3.08308125e-01 3.77253771e-01 1.83661237e-01
9.73727942e-01 2.50986159e-01 3.37647140e-01 4.85885620e-01
6.17362797e-01 3.67536515e-01 1.36075103e+00 -1.98296949e-01
-8.20842028e-01 -8.40184748e-01 1.36152297e-01 7.96902061e-01
1.33787587e-01 -9.88284409e-01 -9.32492495e-01 -1.03327584e+00
-2.69982547e-01 1.09315956e+00 -7.80414701e-01 3.21795270e-02
-7.80544758e-01 -4.79630619e-01 6.15694463e-01 6.93651557e-01
9.33022678e-01 -6.51694119e-01 -6.53430998e-01 2.07384452e-01
-8.15273643e-01 -1.42248309e+00 -1.65049896e-01 1.61191434e-01
-6.62352026e-01 -1.32731342e+00 -9.94065553e-02 -3.25115472e-01
2.75259107e-01 -4.85020787e-01 1.49675488e+00 9.97844040e-02
2.07331151e-01 2.17908651e-01 -5.40504098e-01 -2.92693436e-01
-4.62328464e-01 2.90428489e-01 -7.71117091e-01 -2.42220089e-01
1.03854880e-01 -1.67269379e-01 -2.35538065e-01 7.75441974e-02
-1.03492463e+00 5.71485221e-01 5.91120958e-01 8.15406144e-01
6.11354649e-01 1.01012230e-01 2.77878463e-01 -1.16168642e+00
7.73919940e-01 -4.29503918e-01 -7.89748013e-01 1.08719230e+00
-7.44508803e-01 5.77111781e-01 8.77167225e-01 2.87634581e-01
-1.26774371e+00 -4.57238466e-01 -1.74327239e-01 -1.97313745e-02
3.90984744e-01 1.07716858e+00 -8.50952327e-01 6.25890255e-01
2.27987990e-01 3.48978817e-01 -2.26013169e-01 -3.47833693e-01
6.23431265e-01 4.28975821e-01 4.02937323e-01 -8.94646645e-01
4.52341884e-01 8.68807510e-02 2.98202723e-01 2.27278113e-01
-1.02948916e+00 1.49542484e-02 -8.44352484e-01 -2.11329618e-03
1.05754769e+00 -8.62415373e-01 -8.38540077e-01 -3.69826481e-02
-1.49927962e+00 4.11842540e-02 4.34386618e-02 2.02144399e-01
-7.13236451e-01 5.12046993e-01 -8.05870295e-01 -6.81500852e-01
-3.36515486e-01 -1.19716036e+00 1.17486894e+00 -2.53687501e-01
1.68779582e-01 -1.02947986e+00 -9.44286883e-02 8.12065661e-01
5.27991019e-02 9.87101793e-02 1.80221343e+00 -6.71564102e-01
-1.08821011e+00 6.00573383e-02 -4.83066082e-01 1.16622791e-01
-2.08383232e-01 -2.41259620e-01 -5.18048048e-01 3.98832828e-01
1.55906200e-01 -4.05543745e-01 6.33353591e-01 -9.95203331e-02
1.18503952e+00 -5.99628866e-01 1.29521653e-01 4.30924773e-01
1.51749957e+00 5.72636388e-02 1.00751078e+00 4.86255080e-01
7.62415946e-01 6.04818523e-01 6.13230348e-01 2.90197998e-01
1.31398416e+00 5.76053143e-01 6.15734875e-01 3.82729024e-01
6.53353194e-03 -6.00667536e-01 6.13538288e-02 1.24478340e+00
-1.09156832e-01 -6.25187576e-01 -9.53006864e-01 1.09358080e-01
-1.91448188e+00 -9.73101914e-01 -3.17536622e-01 1.70847559e+00
1.09091103e+00 2.31736779e-01 -2.33352929e-01 3.32576573e-01
1.29888967e-01 -1.47652365e-02 -3.18595976e-01 -5.24911582e-01
-2.40499005e-01 -3.65674086e-02 4.59520131e-01 5.10147035e-01
-6.04112089e-01 6.86415434e-01 5.55141163e+00 9.81260419e-01
-4.65719253e-01 -4.85493764e-02 4.92486030e-01 1.93915337e-01
-8.41936409e-01 2.17064121e-03 -7.14889824e-01 1.74921662e-01
8.86226058e-01 -4.51724082e-02 6.37964070e-01 3.79399747e-01
-3.90851468e-01 -1.44490749e-01 -1.40401959e+00 1.00375354e+00
2.29823202e-01 -1.40112686e+00 4.79443759e-01 -5.75665310e-02
1.39040217e-01 -6.19521201e-01 -7.29261115e-02 7.01119840e-01
9.50316375e-04 -1.21382070e+00 9.26376641e-01 7.72041500e-01
7.38089681e-01 -5.43496966e-01 1.10262549e+00 5.90887487e-01
-1.29346192e+00 -1.10771082e-01 -3.12717348e-01 1.40845282e-02
8.10142383e-02 2.82886177e-01 -5.59796274e-01 1.59915066e+00
3.94382447e-01 5.03188491e-01 -9.14668083e-01 3.27651829e-01
-3.75421822e-01 3.11920494e-01 -9.08589736e-03 -9.99327898e-02
1.19527645e-01 -3.46160650e-01 6.95378631e-02 9.03846681e-01
3.47049683e-01 2.99165756e-01 -3.50788534e-01 1.10159075e+00
-2.65739799e-01 2.48806342e-01 -2.69310087e-01 -2.35890709e-02
2.16340050e-01 8.64520967e-01 -4.73551661e-01 -6.64887965e-01
-6.83696926e-01 9.08495247e-01 6.99584484e-01 1.47701949e-01
-1.03950357e+00 -3.46417129e-01 -2.22456187e-01 -1.14761963e-01
4.23744172e-01 -2.06001773e-01 -7.17862248e-01 -1.53159606e+00
5.73570430e-01 -1.30200696e+00 9.42418635e-01 -1.21102285e+00
-9.87518489e-01 5.97744882e-01 4.05863196e-01 -1.08457232e+00
-4.50982094e-01 -7.13186860e-01 -5.44576868e-02 9.16037381e-01
-1.53719914e+00 -1.08012676e+00 -2.15779737e-01 5.54296792e-01
5.03902555e-01 1.02056963e-02 7.79583454e-01 3.05285245e-01
-4.04117614e-01 6.05921626e-01 -3.61941993e-01 3.60339731e-01
3.02342862e-01 -1.60445833e+00 2.61177450e-01 6.63516819e-01
2.67339200e-01 5.44302225e-01 4.13936138e-01 -6.73251927e-01
-1.99287701e+00 -6.10181808e-01 1.16388690e+00 -8.35810125e-01
6.35261297e-01 -5.28972208e-01 -1.04048729e+00 8.35220098e-01
4.98879075e-01 -3.58719856e-01 4.07830864e-01 4.10414748e-02
-7.55544245e-01 -1.31372333e-01 -9.59834635e-01 5.23410916e-01
9.54007566e-01 -6.55765355e-01 -1.04418528e+00 2.18942359e-01
8.94983709e-01 -1.09541035e+00 -1.08332098e+00 3.89664590e-01
3.79027277e-01 -1.12315929e+00 7.23721027e-01 -5.03068626e-01
1.10711741e+00 -3.31090957e-01 -4.29409891e-01 -1.09643674e+00
1.31325409e-01 -5.23174368e-02 -6.28302217e-01 1.13562429e+00
7.71741450e-01 -4.88778770e-01 4.20427442e-01 5.83183885e-01
-7.92994425e-02 -1.15264428e+00 -1.06934714e+00 -4.36949432e-01
2.63478495e-02 -4.50382501e-01 1.11578584e+00 6.62460566e-01
3.25340271e-01 7.55151927e-01 -1.72216132e-01 2.08312541e-01
1.69982404e-01 4.72495943e-01 5.54901183e-01 -9.27745998e-01
-4.11604881e-01 -1.91590130e-01 -1.16878606e-01 -1.17390537e+00
-1.20269153e-02 -9.68094468e-01 -3.40075493e-01 -2.25452375e+00
1.30977079e-01 8.63395929e-02 5.56133948e-02 2.16682494e-01
-8.30182359e-02 -4.75254178e-01 2.93773532e-01 -7.72136524e-02
-8.49623144e-01 6.93958521e-01 1.65429389e+00 -2.30501235e-01
4.04693693e-01 -4.25789118e-01 -7.78993309e-01 3.14072609e-01
3.85219991e-01 -1.83566108e-01 -5.85307300e-01 -7.64627337e-01
1.05850708e+00 9.10246313e-01 2.01564640e-01 -8.04973662e-01
5.11454225e-01 -2.46169716e-01 2.80332893e-01 -9.55067635e-01
3.76085252e-01 -1.08488560e+00 -1.97587416e-01 2.58109450e-01
-4.99612302e-01 7.16393948e-01 6.57397658e-02 5.07899344e-01
-5.02697766e-01 -3.02308530e-01 6.93651214e-02 -4.11853611e-01
-2.95964599e-01 9.81367603e-02 -9.16570649e-02 5.09245753e-01
4.41798747e-01 8.30218047e-02 -8.34217250e-01 -4.55263346e-01
-2.50094652e-01 4.47669983e-01 1.25149667e-01 7.37600625e-02
6.25308335e-01 -1.17213154e+00 -6.23901367e-01 6.19305894e-02
3.54934454e-01 3.45698059e-01 2.22841173e-01 8.18429947e-01
-8.34525108e-01 6.11720383e-01 -1.91817105e-01 -1.07518122e-01
-9.15835738e-01 7.93050706e-01 5.09530902e-01 -8.79460990e-01
-4.51081932e-01 4.93012071e-01 -8.60560983e-02 -4.71537173e-01
1.25005152e-02 -1.05672646e+00 -5.14152050e-01 1.24265976e-01
2.35384792e-01 2.60505587e-01 4.84115601e-01 -2.78263956e-01
-1.73486382e-01 4.75295722e-01 -1.77404494e-03 -2.99810022e-01
8.44743013e-01 -2.05904990e-01 -5.65448701e-01 4.84741658e-01
9.51176465e-01 2.26763174e-01 -5.56004941e-01 -1.96570829e-01
1.82962716e-01 -1.68110043e-01 -2.47773349e-01 -1.17590380e+00
-5.94531894e-01 1.03196073e+00 -1.24221124e-01 3.65824670e-01
1.14271510e+00 -1.54696524e-01 9.00187254e-01 5.41350067e-01
2.99678385e-01 -1.03831017e+00 -5.14850318e-02 7.92959511e-01
1.16302168e+00 -9.69871879e-01 2.00039715e-01 -8.45647275e-01
-9.16065633e-01 1.42447174e+00 8.06019247e-01 3.47565353e-01
2.09991112e-01 1.81719825e-01 -7.92988688e-02 -6.03723288e-01
-1.19758213e+00 -1.07878760e-01 5.56554914e-01 8.84917304e-02
4.21551019e-01 -8.31374526e-02 -3.60069990e-01 1.01345921e+00
-6.09739423e-01 9.90690291e-03 4.86549139e-01 8.32311511e-01
-9.56286266e-02 -9.83554900e-01 -3.12655926e-01 5.07336080e-01
-4.48287070e-01 -4.46985543e-01 -2.77595043e-01 7.75070965e-01
2.40886196e-01 8.18509996e-01 -8.97627696e-02 -3.01023602e-01
6.05695605e-01 3.72631341e-01 8.26357186e-01 -3.88305515e-01
-7.60031223e-01 -3.02469701e-01 5.22929192e-01 -4.10943180e-01
-1.57761171e-01 -4.01359558e-01 -1.37183452e+00 -3.36781859e-01
-9.09647867e-02 2.33852267e-01 4.72496718e-01 1.22382808e+00
1.81763738e-01 9.20096219e-01 1.13085255e-01 3.15741986e-01
-7.92250454e-01 -9.73352253e-01 -5.19605100e-01 2.23955423e-01
6.24251887e-02 -4.46141124e-01 -7.44857341e-02 -2.71266662e-02] | [10.614324569702148, 7.863439083099365] |
8b83bc11-dcee-4a2e-b452-8f19874eb0ee | robustness-testing-for-multi-agent | 2306.06136 | null | https://arxiv.org/abs/2306.06136v1 | https://arxiv.org/pdf/2306.06136v1.pdf | Robustness Testing for Multi-Agent Reinforcement Learning: State Perturbations on Critical Agents | Multi-Agent Reinforcement Learning (MARL) has been widely applied in many fields such as smart traffic and unmanned aerial vehicles. However, most MARL algorithms are vulnerable to adversarial perturbations on agent states. Robustness testing for a trained model is an essential step for confirming the trustworthiness of the model against unexpected perturbations. This work proposes a novel Robustness Testing framework for MARL that attacks states of Critical Agents (RTCA). The RTCA has two innovations: 1) a Differential Evolution (DE) based method to select critical agents as victims and to advise the worst-case joint actions on them; and 2) a team cooperation policy evaluation method employed as the objective function for the optimization of DE. Then, adversarial state perturbations of the critical agents are generated based on the worst-case joint actions. This is the first robustness testing framework with varying victim agents. RTCA demonstrates outstanding performance in terms of the number of victim agents and destroying cooperation policies. | ['Guanjun Liu', 'Ziyuan Zhou'] | 2023-06-09 | null | null | null | null | ['multi-agent-reinforcement-learning'] | ['methodology'] | [ 3.70848514e-02 2.57778652e-02 1.84457183e-01 3.27672064e-01
-4.31455016e-01 -6.75308943e-01 8.76590014e-01 2.06050664e-01
-3.41131449e-01 1.04261494e+00 -3.32509547e-01 -2.98731297e-01
-3.92452329e-01 -9.01833475e-01 -5.46512723e-01 -1.18251228e+00
-5.15131474e-01 4.72539753e-01 6.14016533e-01 -6.39475346e-01
1.73685804e-01 5.84860563e-01 -1.38466740e+00 -3.62675071e-01
8.51738334e-01 6.29590333e-01 -4.02002215e-01 7.11104631e-01
5.61918020e-01 1.08608186e+00 -1.00346303e+00 -5.76263899e-03
4.16107476e-01 -5.84405899e-01 -4.95652497e-01 -4.59855460e-02
-5.31077802e-01 -1.74274579e-01 -2.20702723e-01 1.08204782e+00
5.74068010e-01 5.83548427e-01 4.39966500e-01 -2.02886438e+00
-2.51476802e-02 7.15381086e-01 -4.01160538e-01 5.77121377e-02
4.40546334e-01 6.93961442e-01 4.01565194e-01 -2.23302022e-02
2.60782778e-01 1.47739732e+00 3.28728527e-01 7.51942575e-01
-7.58628666e-01 -6.75788939e-01 3.54765385e-01 3.05159450e-01
-1.08151484e+00 -1.89690098e-01 6.36698604e-01 -3.78757149e-01
6.59563065e-01 5.03461584e-02 3.56752574e-01 1.06140661e+00
7.39105165e-01 2.20078990e-01 1.15359330e+00 -2.40957871e-01
7.66740024e-01 -1.30606934e-01 -3.66420776e-01 6.69810057e-01
3.49450588e-01 6.52953804e-01 -1.90584213e-01 -5.96404195e-01
4.00673181e-01 -3.93906772e-01 5.87241948e-02 -7.95086659e-03
-1.13616931e+00 7.09431231e-01 4.93510179e-02 -6.05305210e-02
-6.63875818e-01 9.67611745e-02 6.56744719e-01 6.73564553e-01
1.38752470e-02 6.89553916e-01 -3.07806790e-01 -2.28433069e-02
-9.00716148e-03 3.70150328e-01 7.64912844e-01 2.76624203e-01
3.44012886e-01 7.33722508e-01 -1.94889575e-01 3.35314900e-01
2.52270669e-01 6.82533562e-01 5.17416537e-01 -9.25960720e-01
1.16815612e-01 6.91404760e-01 3.06317687e-01 -9.60713267e-01
-3.27547073e-01 -1.95100829e-01 -5.66043854e-01 1.07126665e+00
-1.01417445e-01 -8.18620384e-01 -4.94283617e-01 1.90465784e+00
7.22154438e-01 5.05374253e-01 6.95097327e-01 6.48282349e-01
2.12942556e-01 5.48654556e-01 -4.63486537e-02 -5.34152567e-01
7.56001115e-01 -6.39248848e-01 -5.92080414e-01 -9.76132136e-03
3.43385249e-01 -3.33888203e-01 3.63582939e-01 4.32743907e-01
-9.62265551e-01 -2.45262221e-01 -1.17242229e+00 1.33910728e+00
-3.20791572e-01 -4.74935859e-01 2.08272919e-01 6.15809500e-01
-7.21299529e-01 5.07670820e-01 -7.75054574e-01 -7.25768581e-02
-1.31862581e-01 6.46194935e-01 -2.21451014e-01 4.06855464e-01
-1.38596761e+00 1.18048608e+00 5.37137330e-01 3.13772969e-02
-1.75119996e+00 -2.76744425e-01 -9.92362916e-01 -2.61624515e-01
6.99762166e-01 -2.28338480e-01 1.04463756e+00 -9.10396039e-01
-1.95758069e+00 1.85872465e-01 5.65614879e-01 -7.44837582e-01
7.72129416e-01 1.26497254e-01 -6.38279557e-01 1.04737043e-01
5.69399931e-02 3.41159590e-02 1.10178530e+00 -1.22836578e+00
-6.26928985e-01 -3.18321139e-02 3.06992471e-01 2.83677459e-01
8.92959721e-03 1.28589859e-02 4.08750772e-01 -6.39230430e-01
-5.08768499e-01 -1.27022374e+00 -4.19841707e-01 -5.38449466e-01
-2.73353398e-01 -2.00856164e-01 9.98571694e-01 -1.70144424e-01
8.75589490e-01 -1.86121988e+00 4.51700479e-01 4.65797931e-01
-2.16031149e-01 6.41163647e-01 -4.60143268e-01 8.63332808e-01
3.00441682e-02 -1.09799072e-01 -1.39770389e-01 1.88991934e-01
1.68754190e-01 6.08391725e-02 -4.95451182e-01 4.82265443e-01
1.94885135e-01 3.54342669e-01 -8.21385622e-01 -1.20306157e-01
2.68538922e-01 1.22064859e-01 -2.87377506e-01 5.83242655e-01
-3.58016074e-01 6.53167546e-01 -6.44184232e-01 5.27035177e-01
3.78866047e-01 5.83339930e-01 1.07242554e-01 2.41072491e-01
-2.40934968e-01 -4.51820552e-01 -1.28789055e+00 6.27487540e-01
-1.46368116e-01 2.85897646e-02 1.24992795e-01 -8.23841453e-01
1.14851189e+00 6.03947997e-01 6.96199894e-01 -4.60948735e-01
4.89940882e-01 -2.04263441e-02 3.50655079e-01 -4.39463347e-01
1.12094916e-01 1.21389769e-01 -3.89375806e-01 7.85489380e-01
-1.91464677e-01 -3.07092279e-01 3.63645762e-01 8.00366402e-02
1.32357776e+00 -1.24549933e-01 2.44973004e-01 3.91282403e-04
9.83141720e-01 -1.01837605e-01 9.81825948e-01 8.16086948e-01
-4.91762757e-01 -3.31336111e-01 5.72794318e-01 -4.68618274e-01
-7.66673267e-01 -7.65973091e-01 5.35154104e-01 7.36765742e-01
3.47391188e-01 -5.16722426e-02 -6.99010015e-01 -8.41063082e-01
1.16153862e-02 8.85495424e-01 -7.30099618e-01 -9.28953707e-01
-5.01192868e-01 -6.82110488e-01 7.56935298e-01 2.60473732e-02
7.14711010e-01 -1.38130403e+00 -9.59066331e-01 2.36240640e-01
3.10480297e-01 -9.65831757e-01 -1.28616765e-01 6.15549535e-02
-2.40892127e-01 -1.53230894e+00 -9.62252915e-02 -4.93446380e-01
7.34200597e-01 -1.45745754e-01 4.22201604e-01 3.28987241e-01
-3.67288925e-02 5.28683662e-01 -4.75111991e-01 -5.62583327e-01
-1.18373358e+00 -4.04634506e-01 8.15823555e-01 2.04401448e-01
-1.97399274e-01 -3.88358563e-01 -1.92514494e-01 6.94697976e-01
-1.04244852e+00 -5.47286689e-01 4.16782916e-01 7.35067606e-01
2.38607571e-01 6.19745970e-01 9.70574260e-01 -4.33211118e-01
1.18559074e+00 -4.37244862e-01 -1.04421651e+00 6.13270283e-01
-4.33763772e-01 9.77802500e-02 1.04322124e+00 -7.10695207e-01
-7.13109374e-01 -2.94507265e-01 1.48503274e-01 -3.27763170e-01
-1.56835541e-01 2.92583227e-01 -4.57644872e-02 -3.11314464e-01
5.81656873e-01 1.63688749e-01 3.80194455e-01 2.13566110e-01
-1.08563803e-01 3.51627171e-01 2.87480921e-01 -7.08864331e-01
1.17398262e+00 4.35592867e-02 4.27278697e-01 -5.41893840e-01
-2.53132403e-01 2.78410405e-01 -2.41247803e-01 -6.94198430e-01
4.60828871e-01 -4.61446255e-01 -1.28968847e+00 8.53510976e-01
-7.58874655e-01 -5.91716886e-01 -1.31899476e-01 3.93602639e-01
-6.50972724e-01 3.24601680e-01 -2.29313612e-01 -1.05329692e+00
-3.61174077e-01 -1.40094101e+00 3.60227257e-01 4.18954074e-01
1.94401771e-01 -8.17333698e-01 4.93419826e-01 -9.07968357e-03
3.00098807e-01 7.96927989e-01 8.31751168e-01 -7.63789535e-01
-2.03747064e-01 -4.29392189e-01 6.94217145e-01 3.64078999e-01
1.68351263e-01 4.03645396e-01 -3.89510155e-01 -8.41499686e-01
-4.31830063e-02 -4.79478240e-01 6.39503077e-02 1.81777030e-01
3.13519865e-01 -4.88673866e-01 -1.63306549e-01 3.27604380e-03
1.09983587e+00 9.14087355e-01 3.99974853e-01 6.50185287e-01
1.66461214e-01 4.22802299e-01 9.44488525e-01 6.17662072e-01
4.14186418e-02 4.63572145e-01 1.00476575e+00 2.57163912e-01
5.01369894e-01 1.89206079e-02 9.82021332e-01 4.19952691e-01
-4.31610532e-02 -3.21493506e-01 -7.67648637e-01 8.02480802e-02
-2.08347106e+00 -1.07514656e+00 5.23009002e-01 2.39787245e+00
4.66352135e-01 3.94117236e-01 3.68574381e-01 2.48804778e-01
8.98787379e-01 -6.78366050e-02 -9.67898667e-01 -7.09991455e-01
-9.86125022e-02 -2.47572452e-01 2.50494957e-01 5.69129348e-01
-1.06242585e+00 9.32662427e-01 5.35829687e+00 4.82777625e-01
-1.09546781e+00 -2.34988913e-01 2.17434719e-01 3.19266617e-01
1.67103007e-01 6.66972250e-02 -5.57502627e-01 4.76534069e-01
5.89043319e-01 -4.81768578e-01 6.11755192e-01 6.16599739e-01
5.29262722e-01 -3.12183440e-01 -5.80339789e-01 2.83989519e-01
6.80956617e-02 -8.88934612e-01 1.51213557e-02 -1.91485301e-01
9.33245659e-01 -2.56724834e-01 -2.33714394e-02 4.30965990e-01
9.17629182e-01 -7.74387956e-01 5.43970048e-01 5.15713692e-01
2.67905623e-01 -1.31833112e+00 9.11082923e-01 3.65609735e-01
-1.00276661e+00 -5.03096998e-01 -4.75506075e-02 -1.14359766e-01
1.18720653e-02 -8.27767029e-02 -7.27746964e-01 6.02848947e-01
3.55460614e-01 2.12691024e-01 -5.32328010e-01 8.89905930e-01
-5.57389617e-01 3.77426684e-01 -3.27657610e-02 -1.77533567e-01
3.13744932e-01 -2.83575416e-01 1.08815861e+00 3.69471878e-01
5.01901396e-02 3.95598933e-02 6.40079319e-01 3.30528766e-01
3.32752913e-01 -2.85067022e-01 -8.22344542e-01 8.05723965e-02
7.39469707e-01 9.88858581e-01 -6.45172954e-01 -3.70744541e-02
2.35979497e-01 5.37845612e-01 1.83877781e-01 3.32048833e-01
-1.05223453e+00 -4.54223871e-01 7.91514993e-01 -4.02707666e-01
9.57109500e-03 -2.92154223e-01 5.16330242e-01 -6.54680431e-01
-2.86163837e-01 -1.32950830e+00 4.43617314e-01 -3.53073806e-01
-1.08820415e+00 9.78478134e-01 -3.85420397e-02 -1.51635027e+00
-5.58396280e-01 -1.97747707e-01 -1.11719763e+00 4.18932676e-01
-1.16950691e+00 -9.57521915e-01 -5.41538484e-02 8.73961151e-01
3.67828995e-01 -9.67474282e-01 8.43444645e-01 -3.56526405e-01
-1.28948903e+00 4.87715483e-01 -7.53504261e-02 -3.96337621e-02
4.70365971e-01 -8.76102865e-01 -3.29004414e-02 1.24119055e+00
-6.05503857e-01 1.29295155e-01 1.01447833e+00 -8.82212579e-01
-1.45376790e+00 -1.17922854e+00 4.82507283e-03 3.35067436e-02
8.32968235e-01 8.40443447e-02 -4.87962842e-01 5.08313894e-01
4.07292753e-01 -2.71584690e-01 2.62815952e-01 -5.51718116e-01
5.01083955e-02 -1.77830577e-01 -1.40891910e+00 1.01962125e+00
1.97246820e-01 -4.50375266e-02 -3.03889722e-01 2.81508863e-01
6.64672852e-01 -2.05485433e-01 -8.09409022e-01 6.92278862e-01
1.80849895e-01 -7.93908596e-01 6.05474651e-01 -7.05070853e-01
-1.26284882e-01 -8.23400259e-01 8.91383663e-02 -1.75027943e+00
-5.29633164e-02 -1.04918849e+00 -5.37454262e-02 1.13001180e+00
7.14301839e-02 -1.01015806e+00 2.85982728e-01 4.26301509e-01
-1.01452298e-01 -6.58639073e-01 -1.02614582e+00 -1.05176151e+00
-5.72416447e-02 1.37390658e-01 6.20957494e-01 8.87044191e-01
-2.46628840e-02 -3.09061091e-02 -5.11793673e-01 6.40888512e-01
6.27624452e-01 -3.28124762e-01 8.85795414e-01 -8.33097517e-01
-3.22040141e-01 -3.93828273e-01 -5.15370786e-01 2.85453022e-01
5.02476931e-01 -2.61972934e-01 1.63238436e-01 -1.00882328e+00
-3.63709271e-01 -3.36128116e-01 -3.54241222e-01 6.64640427e-01
-1.35955900e-01 -1.76876456e-01 2.35147506e-01 -9.73179787e-02
-7.01041579e-01 7.40958214e-01 9.54222560e-01 -1.68522790e-01
-3.81500155e-01 3.65860552e-01 -1.23754784e-01 7.06710577e-01
1.28961110e+00 -5.19482970e-01 -5.89867055e-01 1.35364652e-01
1.02481417e-01 3.48252773e-01 4.14902508e-01 -1.37522459e+00
2.87655801e-01 -7.72033095e-01 -2.88640708e-01 -2.01091826e-01
1.04874767e-01 -7.77397990e-01 2.78787285e-01 1.04027355e+00
-2.73440957e-01 5.44878185e-01 3.21205616e-01 5.93128324e-01
-1.30576193e-01 -2.25121513e-01 1.01159739e+00 -6.25054119e-03
-6.03681982e-01 2.57022321e-01 -8.59464407e-01 -3.05417217e-02
1.80738044e+00 3.14194970e-02 -5.72338462e-01 -4.12529886e-01
-3.64276826e-01 7.08632171e-01 4.11154658e-01 4.69220310e-01
6.62688911e-01 -1.01486528e+00 -8.38714421e-01 2.76460052e-01
-2.05572218e-01 -3.94744009e-01 1.23523280e-01 6.54963434e-01
-4.00130928e-01 -1.08749218e-01 -5.78370214e-01 -3.78549062e-02
-1.37713838e+00 6.70545101e-01 7.38435984e-01 -2.60712624e-01
-2.14287698e-01 4.86707568e-01 -3.44084620e-01 -2.93424666e-01
3.07186216e-01 4.75502551e-01 -3.82006645e-01 -1.13173202e-01
4.68441457e-01 5.96076965e-01 9.78018157e-03 -6.21702492e-01
-4.80816215e-01 3.91876459e-01 -8.81367624e-02 -3.65005225e-01
1.16895807e+00 2.01576903e-01 -3.18395644e-01 1.46505386e-01
3.75354409e-01 -1.91447183e-01 -1.47868836e+00 1.28995076e-01
-1.00745723e-01 1.72002856e-02 -1.17151581e-01 -8.79704177e-01
-8.99626613e-01 1.57774866e-01 4.70136881e-01 4.24032211e-01
1.13685679e+00 -5.90336144e-01 4.31399345e-01 5.91047943e-01
5.93446374e-01 -1.29021633e+00 3.68153363e-01 6.75829411e-01
1.02735686e+00 -1.04331779e+00 -6.57819435e-02 8.64442959e-02
-1.03633189e+00 9.30447638e-01 9.41103458e-01 -4.08353478e-01
5.86125314e-01 3.86789143e-01 2.61993110e-01 -4.26500570e-04
-1.02942002e+00 -5.07053547e-02 -1.93824142e-01 8.55803967e-01
-4.48026627e-01 -4.37039882e-03 -1.91185534e-01 6.22274913e-02
1.21534310e-01 -3.83660376e-01 8.35420549e-01 1.13706493e+00
-3.88743997e-01 -1.05220354e+00 -5.89545965e-01 4.77493601e-03
-9.42834541e-02 4.27309543e-01 -4.08612639e-01 6.28288448e-01
8.49688351e-02 1.38840199e+00 -3.23193997e-01 -7.15883851e-01
4.51316446e-01 -2.41276950e-01 8.76238123e-02 -2.58367628e-01
-8.98160696e-01 -4.03231323e-01 -3.66821773e-02 -4.60080832e-01
-4.19237316e-01 -4.80725586e-01 -1.27649450e+00 -4.25271511e-01
-2.08789870e-01 5.70863605e-01 3.15417618e-01 1.01648605e+00
2.23577425e-01 6.80634499e-01 1.42289484e+00 -7.44629979e-01
-8.03526342e-01 -6.61472380e-01 -3.88273031e-01 2.11831182e-01
4.86672729e-01 -9.72763300e-01 -5.12856364e-01 -4.09494311e-01] | [3.842705488204956, 2.299281597137451] |
0416f346-2610-4111-9f7a-cf7d4a067cf7 | single-multi-source-cross-lingual-ner-via | 2004.12440 | null | https://arxiv.org/abs/2004.12440v2 | https://arxiv.org/pdf/2004.12440v2.pdf | Single-/Multi-Source Cross-Lingual NER via Teacher-Student Learning on Unlabeled Data in Target Language | To better tackle the named entity recognition (NER) problem on languages with little/no labeled data, cross-lingual NER must effectively leverage knowledge learned from source languages with rich labeled data. Previous works on cross-lingual NER are mostly based on label projection with pairwise texts or direct model transfer. However, such methods either are not applicable if the labeled data in the source languages is unavailable, or do not leverage information contained in unlabeled data in the target language. In this paper, we propose a teacher-student learning method to address such limitations, where NER models in the source languages are used as teachers to train a student model on unlabeled data in the target language. The proposed method works for both single-source and multi-source cross-lingual NER. For the latter, we further propose a similarity measuring method to better weight the supervision from different teacher models. Extensive experiments for 3 target languages on benchmark datasets well demonstrate that our method outperforms existing state-of-the-art methods for both single-source and multi-source cross-lingual NER. | ['Jian-Guang Lou', 'Börje F. Karlsson', 'Qianhui Wu', 'Zijia Lin', 'Biqing Huang'] | 2020-04-26 | single-multi-source-cross-lingual-ner-via-1 | https://aclanthology.org/2020.acl-main.581 | https://aclanthology.org/2020.acl-main.581.pdf | acl-2020-6 | ['cross-lingual-ner'] | ['natural-language-processing'] | [-2.5322151e-01 -1.5716591e-01 -6.5975314e-01 -6.3391769e-01
-1.2213655e+00 -9.0067279e-01 5.9822559e-01 7.5434588e-02
-8.2970852e-01 8.2002097e-01 2.7284211e-01 -2.7425820e-01
2.9369015e-01 -4.9048555e-01 -5.8204728e-01 -4.5754239e-01
5.3929734e-01 4.6376088e-01 2.2835745e-01 -2.6647097e-01
-1.8309678e-01 2.2377016e-01 -9.5443231e-01 1.2730579e-01
1.2901986e+00 5.2312285e-01 1.3143215e-01 7.7971287e-02
-7.3575652e-01 8.1667221e-01 -3.4986976e-01 -7.3444945e-01
2.3828636e-01 -4.4306624e-01 -8.1935465e-01 -1.9179244e-01
5.0391757e-01 2.9094157e-01 1.5227876e-02 1.2748982e+00
5.5509770e-01 9.2349581e-02 7.0281988e-01 -1.0355644e+00
-8.0478233e-01 1.0210544e+00 -7.1182853e-01 -8.8559784e-02
-6.6331071e-03 -3.0951360e-01 9.4244522e-01 -1.1719614e+00
5.5707681e-01 1.2520795e+00 7.7720821e-01 6.8303561e-01
-1.1300923e+00 -9.4393438e-01 1.7008506e-01 -5.5976845e-02
-1.4190962e+00 -5.1896012e-01 7.7322733e-01 -2.6460820e-01
5.3660017e-01 -2.5445393e-01 -8.9305611e-03 1.0138874e+00
-3.9719898e-01 8.4702885e-01 1.5046676e+00 -7.7304548e-01
-6.0879659e-02 5.6639475e-01 1.7620750e-01 6.4352530e-01
5.0381724e-02 -2.2060420e-02 -5.7936805e-01 -6.1367203e-02
3.3687294e-01 -1.2049325e-01 -3.0495641e-01 -4.3501908e-01
-1.2915509e+00 8.3327889e-01 2.4224076e-01 6.6868246e-01
-1.1788289e-01 -4.5974019e-01 4.9679950e-01 3.2899210e-01
6.4829540e-01 3.3629867e-01 -1.0425220e+00 3.1756449e-02
-9.6867788e-01 -4.3473271e-01 1.0366558e+00 1.2309260e+00
1.0785372e+00 5.4172233e-02 5.4647196e-02 1.1022273e+00
4.5895362e-01 7.5647128e-01 7.1351928e-01 -3.7436709e-01
8.1058323e-01 7.2142762e-01 -7.3122352e-02 -3.7029725e-01
-4.3474534e-03 -1.6762546e-01 -7.2794998e-01 -1.8123053e-01
3.9362198e-01 -4.9079898e-01 -9.5303005e-01 1.8908120e+00
6.0303587e-01 2.7414092e-01 6.1200845e-01 4.6095747e-01
1.0124385e+00 5.3236979e-01 3.4157401e-01 -2.2174111e-01
1.2925413e+00 -1.4829799e+00 -6.5989918e-01 -1.9643487e-01
1.1254736e+00 -9.5683944e-01 1.1330626e+00 -4.8670623e-02
-6.1852908e-01 -6.2882370e-01 -6.5713847e-01 -1.8290269e-01
-7.6020992e-01 5.4870558e-01 1.0583776e-01 6.9787222e-01
-7.7751386e-01 2.6974830e-01 -7.4133790e-01 -4.5741087e-01
-2.6271978e-02 1.8736854e-01 -5.7543135e-01 -3.9866793e-01
-1.2634383e+00 9.9315429e-01 7.5036037e-01 -1.5414688e-01
-8.8163167e-01 -7.9811031e-01 -1.0533530e+00 4.7438104e-02
4.4327840e-01 -1.0698831e-01 1.2513865e+00 -9.0586025e-01
-1.4404103e+00 9.0817332e-01 -8.8610411e-02 -7.4554734e-02
2.2674920e-01 -2.6345396e-01 -5.7671213e-01 -3.5448891e-01
4.4778964e-01 6.4608544e-01 2.5251102e-01 -1.2799870e+00
-8.2268751e-01 -3.4735411e-01 -1.6506597e-01 3.3039936e-01
-7.0976436e-01 2.1753354e-01 -3.7851766e-01 -5.0652778e-01
-2.1176484e-01 -8.6300206e-01 -1.9529442e-01 -5.8961821e-01
-4.0280864e-01 -7.9869086e-01 8.0653238e-01 -4.8864633e-01
1.0539166e+00 -1.9542123e+00 -1.3176899e-01 -1.6090673e-01
-2.2319868e-01 5.8783942e-01 -3.4059843e-01 6.3755888e-01
-1.7343102e-02 3.0235559e-01 -1.0873914e-01 -5.0142294e-01
-5.2676685e-02 2.9044607e-01 -2.3523359e-01 2.1038850e-01
1.9394119e-01 6.0762912e-01 -1.0660104e+00 -9.0339273e-01
-7.8146487e-02 5.8819205e-01 -4.3852039e-02 6.0775912e-01
1.5166144e-01 7.2329158e-01 -6.3415515e-01 5.5741286e-01
4.6667013e-01 -5.6791428e-02 3.7193117e-01 -2.5783831e-01
-4.0456471e-01 5.6406987e-01 -1.1229906e+00 1.9989096e+00
-8.7255293e-01 -1.1920375e-03 2.1682104e-02 -8.9824367e-01
9.7713012e-01 7.0368886e-01 2.5141323e-01 -4.5653918e-01
-2.6055634e-02 6.1745149e-01 -1.4060512e-01 -2.7637881e-01
3.2749417e-01 -4.0862036e-01 -3.1049830e-01 5.9448469e-01
5.6503642e-01 7.9746433e-02 3.9337361e-01 -2.8991750e-02
7.2445369e-01 2.7118289e-01 5.2278453e-01 -3.1178817e-01
7.3998713e-01 1.2016206e-01 1.0247606e+00 5.1569104e-01
-3.6191359e-01 2.1725072e-01 -6.8180032e-02 -9.3644321e-02
-7.1993524e-01 -7.8973532e-01 -1.7235945e-01 1.5296406e+00
2.3784351e-02 -4.5894629e-01 -7.7579844e-01 -1.4397122e+00
-4.3188307e-01 6.5141046e-01 -2.9348335e-01 1.6901675e-01
-5.3443938e-01 -4.2107293e-01 8.2478851e-01 5.0160837e-01
4.6052679e-01 -1.0884342e+00 6.3570410e-02 2.6622289e-01
-2.7322552e-01 -1.4300103e+00 -7.1166211e-01 3.7133527e-01
-7.6528037e-01 -9.7515118e-01 -8.1043065e-01 -1.2446730e+00
6.5183866e-01 1.3964789e-01 1.2194034e+00 -3.5264558e-01
3.8701841e-01 4.1692209e-01 -4.5438996e-01 -2.9443213e-01
-4.3714273e-01 5.8161765e-01 3.5534069e-01 6.8961099e-02
8.4672910e-01 -3.7316957e-01 7.3732570e-02 3.8164005e-01
-7.1223855e-01 -1.5602776e-02 6.1879236e-01 1.0561025e+00
7.9486781e-01 2.8084498e-02 5.8407784e-01 -1.2908002e+00
2.5790662e-01 -5.8716887e-01 -5.6961995e-01 8.5022765e-01
-5.4718995e-01 3.7576279e-01 9.2572755e-01 -7.4075437e-01
-1.4866138e+00 2.8563324e-01 3.4618978e-03 -4.4124696e-01
-5.5275077e-01 7.1449131e-01 -4.5773742e-01 -1.3431800e-02
4.3511236e-01 1.2539750e-01 -7.7413642e-01 -1.0049828e+00
5.4663450e-01 7.9140198e-01 4.2912516e-01 -9.3760657e-01
8.3643854e-01 1.5246202e-01 -4.3274301e-01 -5.3487587e-01
-1.4904542e+00 -7.1937382e-01 -1.2095895e+00 3.5146952e-01
9.2871547e-01 -1.2797400e+00 -1.3343454e-02 2.3806305e-01
-1.2831001e+00 -2.2947931e-01 -2.3903675e-01 8.9157516e-01
-6.6203043e-02 5.9255645e-02 -6.0120660e-01 -6.1213344e-01
-2.3129787e-01 -1.1929200e+00 8.7583786e-01 5.2984029e-01
3.2205036e-01 -1.4897388e+00 5.6910938e-01 2.7471212e-01
2.6526126e-01 -3.1816158e-01 8.3093661e-01 -1.4806756e+00
-7.8442581e-02 3.9160945e-03 -9.7843461e-02 4.9819702e-01
4.9570876e-01 -3.7208310e-01 -9.1652375e-01 -3.1774172e-01
4.1734945e-02 -9.2613453e-01 4.5969117e-01 -2.4457571e-01
4.5647207e-01 -2.1231306e-01 -2.8132373e-01 4.7195902e-01
1.4617360e+00 -2.1905307e-01 -5.4296467e-02 2.6904318e-01
1.2183379e+00 9.6370858e-01 7.0906132e-01 -1.2699704e-01
8.4867960e-01 4.5855096e-01 -2.5598383e-01 -4.0461519e-01
-1.5029508e-01 -6.4232928e-01 7.4880612e-01 1.7751261e+00
2.5154597e-01 -1.5902741e-01 -1.1494763e+00 8.9796722e-01
-1.5802271e+00 -5.3551871e-01 -3.5865892e-02 2.1459808e+00
1.3171917e+00 -2.7368006e-01 -1.5716369e-01 -5.6080770e-01
9.9451995e-01 -6.2058441e-02 -4.4423842e-01 -8.9291580e-02
-1.4531770e-01 1.6685696e-01 5.1847428e-01 2.5838035e-01
-1.3328369e+00 1.4405776e+00 5.1515603e+00 1.1045982e+00
-1.1508876e+00 5.5976456e-01 3.6818826e-01 4.9752504e-01
-1.8298613e-01 3.2868719e-01 -1.4616807e+00 2.3443323e-01
1.1786299e+00 -1.7715713e-01 -3.2983474e-02 1.0287786e+00
-1.6534244e-01 3.8396940e-01 -1.1827379e+00 7.4000341e-01
1.5886031e-01 -6.3451767e-01 -1.2287296e-01 5.4840829e-02
1.2269636e+00 5.6572884e-01 -3.2678488e-01 7.4292481e-01
8.9434373e-01 -7.2678578e-01 3.3165643e-01 1.3088489e-01
9.1528660e-01 -7.2669733e-01 8.1702363e-01 7.3918074e-01
-1.4478769e+00 5.7204205e-01 -4.8294866e-01 4.6362677e-01
3.2228807e-01 3.3356750e-01 -7.8000361e-01 8.1463015e-01
6.8710619e-01 8.1637365e-01 -5.7657284e-01 7.0738214e-01
-7.9779750e-01 1.0462039e+00 -3.7900856e-01 2.5806823e-01
3.1123033e-01 -2.5982398e-01 1.0773065e-01 1.3817635e+00
3.8005120e-01 -5.8452331e-02 8.4887087e-01 6.1306161e-01
-4.8804477e-01 7.7715212e-01 -7.6870102e-01 -2.4434456e-01
6.1195552e-01 1.5247043e+00 -4.3774328e-01 -4.5196724e-01
-1.0266306e+00 7.8269398e-01 7.9219645e-01 2.6483881e-01
-4.7321472e-01 -4.9604699e-01 3.2985926e-01 -3.7039390e-01
3.6579615e-01 -2.8891179e-01 1.0427081e-01 -1.6032127e+00
-1.9073518e-01 -9.1086620e-01 6.2218356e-01 -2.8281140e-01
-1.9373175e+00 8.4258312e-01 -1.8995573e-01 -1.2293651e+00
-2.4486224e-01 -4.4657302e-01 -5.1780784e-01 8.6200905e-01
-2.1080372e+00 -1.6959859e+00 2.9310870e-01 8.4724754e-01
5.4004169e-01 -3.8221791e-01 9.7077000e-01 5.6412220e-01
-5.9295946e-01 7.7388752e-01 3.3471236e-01 6.4118636e-01
1.3846706e+00 -1.3138212e+00 1.8195315e-01 1.0043828e+00
6.8947744e-01 7.3907924e-01 7.8322761e-02 -6.9984508e-01
-1.2353673e+00 -1.2523563e+00 1.4065197e+00 -4.3927237e-01
9.0449810e-01 -3.1083474e-01 -1.1583377e+00 9.1596431e-01
5.2803910e-01 2.8685886e-01 1.2038805e+00 4.2480177e-01
-7.5857508e-01 6.8461314e-02 -9.7606826e-01 3.4359309e-01
6.1870563e-01 -7.3133749e-01 -9.6636409e-01 2.4904221e-01
8.3188003e-01 -2.4541458e-01 -1.2480756e+00 2.9600632e-01
1.5749956e-02 -4.7254887e-01 5.4845148e-01 -8.2499784e-01
1.6436832e-01 -4.7656983e-01 -1.3719371e-01 -1.5401521e+00
1.5517044e-01 -2.7378312e-01 3.0536315e-01 2.0621352e+00
6.1442381e-01 -4.2350194e-01 3.3035228e-01 4.3002594e-01
-3.1190971e-01 -3.4874070e-01 -8.2028979e-01 -9.4651753e-01
5.2967322e-01 -2.7817902e-01 5.8055502e-01 1.6075007e+00
-1.5104963e-01 8.0675399e-01 -4.5764211e-01 3.2875133e-01
6.5230316e-01 2.0718597e-01 6.9461668e-01 -1.3220096e+00
2.0167848e-02 4.0584151e-02 3.3523273e-02 -1.0516768e+00
8.5171366e-01 -1.1349174e+00 3.7982649e-01 -1.2021842e+00
3.2678246e-01 -9.1566849e-01 -5.8439934e-01 1.1019398e+00
-3.3614603e-01 1.5535469e-01 1.2998989e-01 2.7916062e-01
-8.3210731e-01 7.9633725e-01 9.1948169e-01 8.3545297e-02
-7.7675939e-02 -2.8384972e-01 -6.0454309e-01 9.5300663e-01
5.4134560e-01 -1.1279120e+00 -2.3168144e-01 -7.3885769e-01
-1.9791579e-01 -1.5476047e-01 -2.5460327e-01 -8.4232521e-01
5.5620283e-01 -3.5263658e-01 1.9097647e-02 -4.2154193e-01
-5.1298894e-02 -9.7801667e-01 -3.7222433e-01 -3.4655039e-03
-4.4380882e-01 9.1297152e-03 -7.8938186e-02 4.4779205e-01
-5.6901801e-01 -4.8774061e-01 8.4859294e-01 -5.5912390e-02
-7.1614695e-01 4.7132662e-01 9.4041184e-02 7.6486862e-01
7.1598333e-01 4.2597783e-01 -2.6104730e-01 4.4326819e-02
-4.5440567e-01 4.2102930e-01 5.9408337e-01 5.9807074e-01
1.6875001e-02 -1.6095326e+00 -9.6463823e-01 -2.0632101e-02
4.7549400e-01 -1.0032095e-01 -2.9764709e-03 7.6321411e-01
2.5274736e-01 4.5506841e-01 -1.8844509e-02 -5.0161207e-01
-1.1283243e+00 5.7226062e-01 2.1096738e-01 -7.2492415e-01
-2.3083266e-02 6.9420123e-01 3.3853072e-01 -1.4490424e+00
1.0785208e-01 1.3903847e-01 -4.2623228e-01 1.3204557e-01
3.3410034e-01 8.5985593e-02 3.4459401e-02 -1.1930536e+00
-1.9649330e-01 7.4312878e-01 -3.0005217e-01 -2.1505298e-01
1.2732009e+00 -3.0073592e-01 -5.6433700e-02 8.2845002e-01
1.2305005e+00 5.2743095e-01 -8.4789896e-01 -9.1603416e-01
5.0610030e-01 -7.0096269e-02 -4.2528007e-02 -8.1014669e-01
-1.0259080e+00 1.1392026e+00 4.9336436e-01 -2.2107698e-01
9.4936937e-01 8.5404985e-02 9.2393494e-01 6.0179347e-01
5.1513571e-01 -1.1147811e+00 -2.6447618e-01 8.9439851e-01
1.3311233e-01 -1.7308575e+00 -3.6342716e-01 -3.8327715e-01
-8.7079567e-01 9.8095036e-01 9.6598256e-01 3.1925353e-01
7.4427438e-01 2.0494692e-01 6.8490487e-01 3.7030578e-02
-6.4327770e-01 -5.1384991e-01 5.1662636e-01 4.5974454e-01
1.0737314e+00 6.3563906e-02 -3.4832498e-01 8.8729852e-01
-2.8862303e-02 -2.4270754e-01 2.7429399e-01 1.0442585e+00
-1.8091871e-01 -1.9170965e+00 -2.5824836e-01 -1.1749729e-01
-6.4025438e-01 -5.1123208e-01 -3.9068902e-01 7.7193683e-01
3.6844152e-01 1.0278177e+00 -4.2037424e-01 -6.8210967e-02
3.4218061e-01 4.6871477e-01 1.4087704e-01 -1.0549165e+00
-9.4181490e-01 1.5154102e-01 -2.9683257e-02 -2.0644397e-01
-9.0483868e-01 -4.1465884e-01 -1.1872263e+00 2.2277692e-01
-5.9791976e-01 7.9839194e-01 6.9043839e-01 1.1659650e+00
2.7423349e-01 3.4475997e-02 7.0644665e-01 -4.4718620e-01
-6.6614854e-01 -1.0497513e+00 -6.5657037e-01 3.3537313e-01
2.6358746e-02 -5.6463283e-01 -4.3609977e-01 2.4721515e-01] | [10.031351089477539, 9.619291305541992] |
2dff84a4-ab59-47fc-99da-5c1873fe1045 | stden-towards-physics-guided-neural-networks | 2209.00225 | null | https://arxiv.org/abs/2209.00225v1 | https://arxiv.org/pdf/2209.00225v1.pdf | STDEN: Towards Physics-Guided Neural Networks for Traffic Flow Prediction | High-performance traffic flow prediction model designing, a core technology of Intelligent Transportation System, is a long-standing but still challenging task for industrial and academic communities. The lack of integration between physical principles and data-driven models is an important reason for limiting the development of this field. In the literature, physics-based methods can usually provide a clear interpretation of the dynamic process of traffic flow systems but are with limited accuracy, while data-driven methods, especially deep learning with black-box structures, can achieve improved performance but can not be fully trusted due to lack of a reasonable physical basis. To bridge the gap between purely data-driven and physics-driven approaches, we propose a physics-guided deep learning model named Spatio-Temporal Differential Equation Network (STDEN), which casts the physical mechanism of traffic flow dynamics into a deep neural network framework. Specifically, we assume the traffic flow on road networks is driven by a latent potential energy field (like water flows are driven by the gravity field), and model the spatio-temporal dynamic process of the potential energy field as a differential equation network. STDEN absorbs both the performance advantage of data-driven models and the interpretability of physics-based models, so is named a physics-guided prediction model. Experiments on three real-world traffic datasets in Beijing show that our model outperforms state-of-the-art baselines by a significant margin. A case study further verifies that STDEN can capture the mechanism of urban traffic and generate accurate predictions with physical meaning. The proposed framework of differential equation network modeling may also cast light on other similar applications. | ['Hu Zhang', 'Jiawei Jiang', 'Zhe Jiang', 'Jingyuan Wang', 'Jiahao Ji'] | 2022-09-01 | null | null | null | null | ['physics-informed-machine-learning', 'spatio-temporal-forecasting'] | ['graphs', 'time-series'] | [-2.71217525e-01 -3.28316242e-01 -4.65468407e-01 -3.62705618e-01
4.74615060e-02 2.74522360e-02 7.07071185e-01 -3.84253830e-01
1.97787866e-01 8.02300990e-01 1.97698951e-01 -9.28490043e-01
-5.54974496e-01 -1.41162145e+00 -8.07964265e-01 -8.73730063e-01
1.91168472e-01 4.87957805e-01 3.76754373e-01 -6.98480070e-01
9.27326679e-02 6.81296825e-01 -1.51509476e+00 -2.21062794e-01
1.30772996e+00 9.35681939e-01 8.55878666e-02 3.15750539e-01
-6.77903771e-01 7.12948620e-01 5.93512431e-02 -4.37826626e-02
2.13331953e-01 -8.25416222e-02 -6.30307972e-01 -2.60307223e-01
2.22959831e-01 -4.47791219e-01 -1.16445172e+00 6.30379081e-01
3.80134806e-02 3.59248310e-01 7.53644049e-01 -1.66097689e+00
-7.25836873e-01 1.59329057e-01 -3.61920685e-01 2.07350925e-01
-2.77710974e-01 6.00882769e-01 8.93558443e-01 -4.82652068e-01
2.08552867e-01 1.48613107e+00 5.70783198e-01 6.28010452e-01
-1.04650652e+00 -6.92779839e-01 3.82637858e-01 3.31715584e-01
-1.01433182e+00 -4.85309251e-02 1.17586696e+00 -6.20168209e-01
6.37842059e-01 1.46463141e-01 7.73032606e-01 1.06881714e+00
5.81013262e-01 7.21830785e-01 9.04957175e-01 1.25702709e-01
1.91125408e-01 -7.28936642e-02 3.82947803e-01 6.71392441e-01
3.14245254e-01 6.32465482e-01 -3.12134802e-01 2.66539961e-01
8.05991471e-01 1.56266183e-01 5.94468713e-02 -2.40870774e-01
-5.83402932e-01 7.40274310e-01 9.27822232e-01 5.89141157e-03
-4.57958281e-01 5.92443526e-01 1.34851962e-01 -1.51954263e-01
5.64912021e-01 -7.29494765e-02 -4.38452274e-01 -7.43724927e-02
-7.34790683e-01 7.22876489e-01 4.76715833e-01 8.22696149e-01
1.01185453e+00 3.69676799e-01 -5.21931574e-02 4.25101668e-01
7.86607921e-01 9.07335818e-01 -1.65069327e-01 -9.50881004e-01
3.77623469e-01 8.78965378e-01 1.65593550e-01 -1.09631312e+00
-4.36895013e-01 -2.19299510e-01 -9.86863315e-01 2.07060397e-01
5.70295930e-01 -4.71041262e-01 -9.27567661e-01 1.75604963e+00
3.40803206e-01 5.00130475e-01 -2.22086728e-01 1.03305101e+00
8.61089528e-01 1.13153040e+00 2.37019807e-01 2.07669571e-01
1.09101439e+00 -9.72971618e-01 -6.59635961e-01 -7.34209120e-02
5.96878767e-01 -1.17966615e-01 9.09626901e-01 -2.45686620e-01
-9.24662113e-01 -6.94112301e-01 -7.10662782e-01 -1.46109968e-01
-7.02823579e-01 -2.51576573e-01 9.74381983e-01 5.49059033e-01
-8.38446081e-01 5.86105049e-01 -1.05003786e+00 -4.06410873e-01
5.64974248e-01 2.12702811e-01 1.20736487e-01 1.83835700e-01
-1.53265643e+00 9.71255243e-01 2.08395123e-02 4.57046926e-01
-7.75409579e-01 -1.29087400e+00 -6.33409381e-01 1.78377077e-01
2.76173115e-01 -1.02527869e+00 1.06452906e+00 -2.49975905e-01
-1.52259374e+00 6.49276674e-02 -5.00112712e-01 -5.39489329e-01
4.86957699e-01 -2.33782399e-02 -8.10580313e-01 -2.00740650e-01
-6.40337467e-02 5.48892617e-01 3.13279182e-01 -1.36669230e+00
-6.93907857e-01 6.46727309e-02 2.61439294e-01 -1.75232098e-01
-1.17274895e-01 -5.92423141e-01 -2.27947831e-01 -1.15189835e-01
-1.50405481e-01 -9.78907287e-01 -3.60077858e-01 3.25611234e-01
-3.51592183e-01 -4.50745821e-01 1.27621007e+00 -3.16274434e-01
1.33145773e+00 -1.60377336e+00 -3.50972503e-01 1.54531479e-01
3.31227601e-01 6.13394260e-01 7.72350505e-02 4.58463699e-01
3.47962789e-02 3.65344048e-01 -2.56149918e-01 1.16745315e-01
2.06239536e-01 6.39383495e-01 -6.68926358e-01 1.64658144e-01
2.86287487e-01 1.20012009e+00 -9.66756821e-01 -3.74388367e-01
7.88692534e-01 7.68652856e-01 -4.54679608e-01 -1.17319994e-01
-4.94617879e-01 6.70890152e-01 -1.03600478e+00 2.07570866e-01
1.01871729e+00 -1.65027678e-01 -3.06852072e-01 -3.12338322e-01
-4.77060288e-01 5.91800451e-01 -9.47928369e-01 1.12138212e+00
-7.35607862e-01 9.29912210e-01 -1.63619623e-01 -1.19751334e+00
9.78048801e-01 6.11626767e-02 8.40409994e-01 -1.17191660e+00
-8.18524510e-03 9.31122899e-02 2.70054042e-01 -6.46079957e-01
5.14221668e-01 -3.16155881e-01 1.29118577e-01 3.59008521e-01
-4.13681597e-01 9.76378322e-02 8.11532065e-02 8.69060010e-02
8.20557117e-01 3.61343473e-01 -5.33931196e-01 -3.97959530e-01
7.16247439e-01 3.28772545e-01 6.91480041e-01 5.27707875e-01
-3.83826494e-01 1.12863332e-01 3.33693057e-01 -8.23569596e-01
-1.10453939e+00 -1.01467288e+00 -4.12770152e-01 6.97195053e-01
5.75382769e-01 -1.52738318e-01 -6.47031009e-01 -4.28809851e-01
4.12337929e-01 8.42835128e-01 -6.12642467e-01 -4.33969110e-01
-9.02776241e-01 -9.55717504e-01 4.09442186e-01 6.14238501e-01
9.49648440e-01 -7.79223263e-01 -9.19366553e-02 4.11784708e-01
-2.29573309e-01 -1.14496589e+00 4.28221710e-02 -5.03245294e-01
-7.01955140e-01 -1.02496040e+00 -2.28500232e-01 -1.71120942e-01
4.25163001e-01 6.26545310e-01 9.74710345e-01 2.64018476e-01
1.45518556e-02 1.29262596e-01 6.91724718e-02 -5.29361367e-01
-5.34036875e-01 1.40088841e-01 1.38884351e-01 1.63646102e-01
5.92040658e-01 -7.71119952e-01 -9.46239889e-01 6.74257338e-01
-6.71652317e-01 3.20640922e-01 3.36346388e-01 2.94210464e-01
2.76379704e-01 4.02209848e-01 5.91266513e-01 -4.25048649e-01
3.74920130e-01 -6.99611068e-01 -6.32981777e-01 -6.39597774e-02
-6.84358299e-01 2.91146398e-01 7.06598461e-01 2.47349571e-02
-1.60176146e+00 -2.96659976e-01 -3.34425658e-01 -2.11345136e-01
-3.49789500e-01 4.26971704e-01 -3.09943914e-01 -9.53275785e-02
2.84401119e-01 3.15353483e-01 -8.26847777e-02 -5.37241459e-01
3.94751400e-01 4.51419592e-01 3.23955536e-01 -1.02882671e+00
1.16992867e+00 8.41614604e-01 6.82073116e-01 -9.01024342e-01
-7.71307290e-01 -2.01217249e-01 -6.27209067e-01 -4.89962548e-01
7.65404582e-01 -6.36711240e-01 -1.10483265e+00 5.83980501e-01
-1.21471608e+00 -4.61188197e-01 -2.72319138e-01 3.82442981e-01
-4.80236143e-01 7.90798888e-02 -2.80669302e-01 -9.93516326e-01
7.99905658e-02 -1.05896294e+00 8.35684896e-01 5.36855221e-01
3.21869344e-01 -1.58316910e+00 1.52878501e-02 3.23665559e-01
7.65535355e-01 4.10103023e-01 1.18999112e+00 9.86767784e-02
-1.02440274e+00 2.39916593e-02 -6.14686787e-01 -1.89266615e-02
1.76189572e-01 4.56786633e-01 -8.55743945e-01 4.55711097e-01
-2.33564407e-01 3.83087277e-01 9.74487126e-01 7.55484760e-01
1.28755069e+00 -7.86692500e-02 -6.87426090e-01 5.39113581e-01
1.34386015e+00 1.98077872e-01 7.68142343e-01 1.61233380e-01
1.01327348e+00 6.91585064e-01 3.64090770e-01 1.23246126e-01
8.59931469e-01 7.48100638e-01 5.54475784e-01 -2.44968042e-01
-2.53407747e-01 -5.96645713e-01 9.76490676e-02 6.41444325e-01
-2.89807677e-01 -4.29107666e-01 -1.15719187e+00 5.60247183e-01
-2.14590406e+00 -1.18204296e+00 -1.12804806e+00 1.81686115e+00
1.65276930e-01 2.89009929e-01 2.69746274e-01 -5.86750731e-02
6.33428574e-01 2.34144181e-01 -8.76763344e-01 -4.47366834e-01
-5.42477071e-02 -1.75667420e-01 6.62980914e-01 5.48091352e-01
-7.40067661e-01 9.89489615e-01 6.08224821e+00 7.97212780e-01
-1.43078268e+00 4.04850654e-02 4.64713335e-01 3.52530479e-01
-3.81311357e-01 1.74322039e-01 -9.98896658e-01 7.70029604e-01
1.34179580e+00 -3.55678618e-01 1.33725151e-01 6.13343537e-01
1.14119995e+00 7.40416870e-02 -9.06175315e-01 5.70385218e-01
-7.65948117e-01 -1.76189244e+00 2.71414518e-01 6.17805183e-01
6.02036476e-01 2.48498097e-01 7.06389025e-02 4.37051833e-01
3.27610165e-01 -8.96685779e-01 2.66680032e-01 8.66779208e-01
3.39682490e-01 -4.17714357e-01 3.91093403e-01 4.43389148e-01
-1.51376498e+00 2.52646171e-02 -3.96712512e-01 -4.06767428e-01
7.05364347e-01 7.31606066e-01 -1.49545252e-01 7.39577413e-01
4.05074894e-01 1.02753556e+00 -1.60896033e-01 1.17329657e+00
-3.38380821e-02 1.07697928e+00 -3.90848905e-01 2.50969287e-02
6.37239218e-01 -5.37002206e-01 5.74192762e-01 9.79244113e-01
1.88860893e-01 -6.03348529e-03 -1.05724856e-01 1.49830067e+00
1.43061772e-01 -3.19134027e-01 -7.59112179e-01 1.45943329e-01
1.86230302e-01 1.11083508e+00 -2.83238918e-01 -1.60148740e-01
-4.98240620e-01 1.10942172e-02 -4.71622087e-02 6.47896945e-01
-1.34946322e+00 -4.56823669e-02 1.19850671e+00 5.77888191e-01
4.46706600e-02 -5.79976380e-01 -5.33337414e-01 -9.57993686e-01
-1.50929883e-01 -4.81810905e-02 -2.93579668e-01 -7.40341306e-01
-1.33367622e+00 1.96987763e-02 2.73265332e-01 -1.25171053e+00
1.86514214e-01 -9.90633011e-01 -1.16984129e+00 1.10291088e+00
-2.04097152e+00 -1.14658141e+00 -4.75165129e-01 3.98540556e-01
4.12486792e-01 1.07318878e-01 3.42483371e-01 5.55908799e-01
-9.41491604e-01 -2.00500464e-04 2.78368324e-01 5.76924570e-02
6.62078336e-02 -1.06669652e+00 8.49088490e-01 6.16628110e-01
-4.66453373e-01 4.84460145e-01 7.54193604e-01 -5.91535330e-01
-1.62519765e+00 -1.43687606e+00 8.79985571e-01 -5.78955591e-01
8.51899683e-01 -3.04342210e-01 -1.16632891e+00 2.45362371e-01
-2.25897849e-01 2.50387132e-01 2.72629857e-01 2.29684711e-02
5.89845441e-02 -5.57544291e-01 -8.98058832e-01 6.25470519e-01
1.16100156e+00 -3.66782039e-01 -2.56615758e-01 5.01110375e-01
5.82598567e-01 -1.58568814e-01 -5.70878506e-01 3.43370706e-01
5.72400630e-01 -7.83225656e-01 1.03956342e+00 -8.23458910e-01
6.28476381e-01 -3.40413749e-01 1.41714066e-01 -1.24696147e+00
-4.28645581e-01 -5.87985873e-01 -4.46571410e-01 1.15625107e+00
2.32047543e-01 -7.78658807e-01 9.03497398e-01 1.10644865e+00
-2.10725576e-01 -8.29198956e-01 -9.22260523e-01 -9.33105826e-01
7.37354398e-01 -9.82486665e-01 1.08102334e+00 5.81655025e-01
-4.09612387e-01 2.90332735e-01 -3.40014786e-01 2.63200819e-01
5.92072964e-01 8.58947635e-03 8.97342443e-01 -1.59504533e+00
3.14781427e-01 -7.85497010e-01 -3.80383193e-01 -1.51500618e+00
5.42920530e-01 -5.64567387e-01 -1.67624459e-01 -1.94010246e+00
-1.47515610e-01 -8.36792946e-01 -1.72198772e-01 6.31429106e-02
3.53854187e-02 -2.86986232e-01 -1.01797715e-01 1.38050213e-01
-8.14509839e-02 9.78816867e-01 1.55806613e+00 -2.17324510e-01
-1.35969311e-01 1.54935241e-01 -6.49349093e-01 6.88967705e-01
8.56481850e-01 -2.12979019e-01 -5.78351498e-01 -4.98291671e-01
1.12543389e-01 -2.84684468e-02 8.09813917e-01 -9.37988818e-01
4.05759186e-01 -7.89532959e-01 -7.56806955e-02 -8.11577678e-01
2.37043008e-01 -1.01888299e+00 -5.57295009e-02 4.39610034e-01
-1.03730569e-02 -2.93301910e-01 4.13240075e-01 6.58225000e-01
1.20173246e-02 4.07825410e-01 5.70589125e-01 1.65433913e-01
-7.94910312e-01 9.39648449e-01 -6.18410766e-01 -9.70561579e-02
8.01528156e-01 -4.62961137e-01 -6.18099451e-01 -4.24159169e-01
-2.83742189e-01 5.98935902e-01 5.69051281e-02 6.38371825e-01
2.87750423e-01 -1.51369905e+00 -4.71424162e-01 1.76842779e-01
-1.59162313e-01 8.26077610e-02 5.02720892e-01 7.61507273e-01
-5.10292351e-01 9.43526924e-01 -8.08440223e-02 -6.79563344e-01
-2.88653493e-01 2.65442610e-01 8.48615646e-01 -5.84972650e-02
-8.05585980e-01 2.74403751e-01 5.48474789e-01 -7.70556629e-01
-2.89143890e-01 -5.35599887e-01 -6.97052926e-02 -3.95889670e-01
2.27720872e-01 8.15381467e-01 -1.26595601e-01 -1.13299894e+00
-3.52575839e-01 8.24574113e-01 3.28062236e-01 2.56537735e-01
1.19683909e+00 -3.51402760e-01 2.80168951e-01 3.15637946e-01
1.05423677e+00 -5.06953418e-01 -1.59599102e+00 -2.69173980e-01
-3.56315047e-01 -3.21145952e-01 6.41134381e-01 -6.84642792e-01
-1.47802389e+00 1.34266591e+00 3.40094119e-01 4.56149220e-01
6.49707913e-01 -5.26581466e-01 1.44675422e+00 3.96893263e-01
1.70643792e-01 -1.01791584e+00 -4.14044052e-01 6.91340566e-01
4.90416199e-01 -1.35677004e+00 -2.77727902e-01 -6.99686050e-01
-3.57425436e-02 1.16152310e+00 9.89975154e-01 -2.03855440e-01
1.25099790e+00 7.88914040e-02 -2.49066368e-01 -2.98607498e-01
-7.49577165e-01 -3.22758973e-01 4.28915620e-01 5.56493104e-01
1.85785621e-01 5.09880157e-03 -1.07013136e-02 3.54371011e-01
-5.74712269e-02 3.58774036e-01 2.49429375e-01 6.03399336e-01
-5.04183710e-01 -1.06944919e+00 -2.68010590e-02 2.60447949e-01
2.33284071e-01 2.27883697e-01 -2.07848451e-03 9.98652697e-01
3.05957824e-01 1.19398129e+00 3.66003096e-01 -3.04163158e-01
4.45886970e-01 -1.61302596e-01 6.19026460e-02 -7.72633180e-02
-5.52070066e-02 -4.59953576e-01 -1.11923873e-01 -5.59047103e-01
-4.95834023e-01 -4.88773406e-01 -1.50490105e+00 -1.11721122e+00
-9.42368899e-03 4.24446724e-02 8.27372611e-01 1.45879126e+00
6.21429980e-01 7.20725775e-01 7.45801091e-01 -9.05834973e-01
9.19618309e-02 -4.95517492e-01 -3.68724018e-01 3.30305368e-01
5.31385362e-01 -1.16862237e+00 -3.83963823e-01 -2.23846763e-01] | [6.396213054656982, 1.9749839305877686] |
dab54f0c-6ff2-4358-a46d-b5cdb16c34b9 | automatic-fado-music-classification | 1406.4447 | null | http://arxiv.org/abs/1406.4447v1 | http://arxiv.org/pdf/1406.4447v1.pdf | Automatic Fado Music Classification | In late 2011, Fado was elevated to the oral and intangible heritage of
humanity by UNESCO. This study aims to develop a tool for automatic detection
of Fado music based on the audio signal. To do this, frequency spectrum-related
characteristics were captured form the audio signal: in addition to the Mel
Frequency Cepstral Coefficients (MFCCs) and the energy of the signal, the
signal was further analysed in two frequency ranges, providing additional
information. Tests were run both in a 10-fold cross-validation setup (97.6%
accuracy), and in a traditional train/test setup (95.8% accuracy). The good
results reflect the fact that Fado is a very distinctive musical style. | ['Pedro Girão Antunes', 'David Martins de Matos', 'Isabel Trancoso', 'Ricardo Ribeiro'] | 2014-06-17 | null | null | null | null | ['music-classification'] | ['music'] | [ 2.24445269e-01 -1.96139425e-01 1.22992843e-01 2.39157304e-01
-6.48607612e-01 -8.89291525e-01 2.94159293e-01 3.66113156e-01
-4.19271469e-01 6.29649222e-01 3.86868179e-01 1.58934042e-01
-4.64052856e-01 -5.89382648e-01 -5.08886315e-02 -5.10066748e-01
-3.08365315e-01 -2.04273015e-01 9.18515697e-02 -1.52229220e-01
4.14370030e-01 4.30580914e-01 -1.84854269e+00 3.23624372e-01
6.21743798e-01 1.05066800e+00 1.14464022e-01 1.02287138e+00
1.70684218e-01 6.08015954e-01 -1.08675086e+00 -2.77637392e-01
-2.56485622e-02 -5.15922368e-01 -6.03451073e-01 -2.03223437e-01
1.17246829e-01 -7.47620240e-02 3.12725693e-01 1.01193953e+00
5.24887562e-01 1.54152691e-01 4.33144480e-01 -6.68971717e-01
8.86022858e-03 5.94073176e-01 5.01250178e-02 1.73631892e-01
8.53550613e-01 -1.29750088e-01 9.82859492e-01 -8.76690030e-01
6.21589124e-01 8.04362178e-01 7.39488900e-01 5.97024746e-02
-1.08667171e+00 -8.41883421e-01 -6.81727171e-01 2.27642566e-01
-1.35764122e+00 -1.78920612e-01 9.94251311e-01 -7.39651561e-01
6.27397895e-01 4.74084288e-01 1.04960358e+00 8.20659637e-01
2.55270064e-01 3.03581029e-01 1.51904047e+00 -7.89541304e-01
2.55376071e-01 1.48306936e-01 -2.27860168e-01 3.44015099e-02
-4.87951823e-02 6.94093928e-02 -6.76792324e-01 -2.47409225e-01
6.23857319e-01 -4.90815938e-01 -3.82219553e-01 2.77627051e-01
-9.75861609e-01 6.93059027e-01 -4.50007850e-03 7.99464285e-01
-5.93389153e-01 -1.89503625e-01 6.10595286e-01 5.01682460e-01
1.50250942e-01 6.67048693e-01 -2.66786575e-01 -6.37518227e-01
-1.10264111e+00 6.45992830e-02 9.09958899e-01 -1.21535666e-01
2.41579577e-01 2.81781554e-01 3.88713747e-01 9.41852450e-01
2.22537130e-01 3.35921884e-01 8.04592967e-01 -9.27613854e-01
3.70403426e-03 3.32195729e-01 -3.03677041e-02 -1.04468024e+00
-2.44763389e-01 -6.54430091e-01 -4.25417066e-01 3.70203108e-01
6.26279771e-01 -1.39559492e-01 -3.50989938e-01 1.52594364e+00
1.80307835e-01 -1.35741770e-01 -9.57440957e-02 7.73744702e-01
4.80985492e-01 7.63511419e-01 -1.12182923e-01 -5.18951058e-01
1.31400204e+00 -1.47643164e-01 -1.01140296e+00 3.40335101e-01
1.18308187e-01 -1.28883839e+00 1.11868715e+00 1.17142475e+00
-1.03757644e+00 -7.77240753e-01 -1.32100630e+00 4.23388243e-01
-2.21155033e-01 2.37199947e-01 3.52746218e-01 1.01911426e+00
-7.11403251e-01 6.53936028e-01 -3.30432624e-01 -1.38758883e-01
-1.93857118e-01 1.46256819e-01 -3.46406937e-01 4.72571760e-01
-1.22908175e+00 4.64339495e-01 5.07340729e-01 -1.05650220e-02
-5.66196382e-01 -5.52244067e-01 -5.65520287e-01 -4.90979329e-02
6.08457327e-02 1.70779645e-01 1.09878743e+00 -1.27159274e+00
-1.54601932e+00 9.16154206e-01 3.13408703e-01 -3.77982795e-01
5.98026574e-01 -4.73499179e-01 -9.22692120e-01 5.53552687e-01
-1.42064810e-01 -7.86443055e-02 7.93150544e-01 -7.74395525e-01
-5.71904242e-01 -1.80001527e-01 -7.76681155e-02 -2.42184013e-01
-3.05048823e-01 2.11233869e-01 7.79869184e-02 -1.11452162e+00
1.25036880e-01 -8.11876476e-01 4.85672206e-01 -5.76304317e-01
2.06226669e-02 -4.36462834e-02 4.71123606e-01 -1.20656157e+00
1.65620017e+00 -2.62596536e+00 -4.02464598e-01 4.64850724e-01
-1.72400922e-01 3.62964392e-01 2.75651306e-01 7.37073541e-01
-1.82060555e-01 -2.50499845e-01 -1.42905759e-02 4.12602276e-01
2.97696739e-02 -2.93768644e-01 -2.90408164e-01 3.43717128e-01
4.73538786e-02 1.90338224e-01 -8.92341554e-01 -2.53590941e-01
1.19455501e-01 5.05992293e-01 -3.02206039e-01 -1.05517149e-01
2.91711062e-01 3.77950132e-01 -1.43200681e-02 6.24322534e-01
4.43101823e-01 6.10059917e-01 4.46703970e-01 -1.88855037e-01
-3.45860720e-01 3.52174550e-01 -1.47399724e+00 1.34732735e+00
-3.84733707e-01 9.15826499e-01 1.80029884e-01 -6.63230002e-01
1.32418680e+00 6.60464585e-01 4.88635272e-01 -6.59853518e-01
1.64173350e-01 7.24111259e-01 2.13763639e-01 -6.88837230e-01
6.42746329e-01 -3.90869975e-01 -4.91351448e-02 5.05181067e-02
1.19558461e-01 -1.00154921e-01 2.26853952e-01 -1.63663134e-01
5.88602364e-01 3.06058794e-01 5.35813630e-01 -3.97045881e-01
6.90561175e-01 -3.00347656e-01 2.89569527e-01 1.57860294e-01
-1.02270737e-01 5.14622986e-01 4.04928267e-01 -2.62180567e-02
-6.84328258e-01 -9.35668051e-01 -2.81372309e-01 8.75133216e-01
-4.74040389e-01 -6.21721923e-01 -9.57909703e-01 -7.91651905e-02
-1.78580299e-01 4.97370154e-01 -5.60358584e-01 -1.39241770e-01
-5.38935423e-01 -2.78466314e-01 7.52192378e-01 2.28045732e-01
2.33346954e-01 -1.16663909e+00 -8.73768210e-01 5.89071393e-01
-1.51337177e-01 -6.51181519e-01 -3.24136019e-01 3.46792191e-01
-7.60447919e-01 -1.20806706e+00 -7.42697656e-01 -7.64912724e-01
-2.49833286e-01 -2.19814047e-01 7.38812983e-01 -2.97410756e-01
-4.70787495e-01 4.12258416e-01 -5.26177227e-01 -6.20436251e-01
-4.98365581e-01 -2.23943412e-01 -2.03251783e-02 2.22998530e-01
2.10424677e-01 -7.88124144e-01 -5.05722702e-01 2.84616709e-01
-9.45499003e-01 -6.73216224e-01 3.96234959e-01 5.23451447e-01
5.47125697e-01 2.48465419e-01 8.90521765e-01 -4.31212962e-01
7.81574965e-01 -1.95609301e-01 -4.52092886e-01 -1.95925310e-01
-2.34527841e-01 -5.39458632e-01 6.31218612e-01 -6.59160733e-01
-8.04368496e-01 -1.16757005e-01 -3.32546860e-01 -3.58508267e-02
-1.60206586e-01 7.06921935e-01 -1.94671601e-01 1.33923262e-01
7.55909383e-01 1.48610687e-02 -9.36422050e-02 -6.60047889e-01
-2.42746413e-01 1.03393328e+00 9.84751582e-01 -3.87646347e-01
4.75372463e-01 1.30884513e-01 -9.41926613e-02 -1.20105326e+00
-5.12384593e-01 -6.63117349e-01 -5.52598059e-01 -7.21152008e-01
7.60897934e-01 -6.60660565e-01 -6.21686459e-01 5.47316968e-01
-8.03066373e-01 -7.23755881e-02 -4.60458428e-01 1.07637239e+00
-4.91952360e-01 3.63230288e-01 -4.17411804e-01 -1.40711629e+00
-3.33114088e-01 -4.51945424e-01 4.09696996e-01 2.65612632e-01
-7.23059177e-01 -8.50760937e-01 4.15353358e-01 2.10481375e-01
1.24387778e-01 7.87793159e-01 7.23523736e-01 -5.17676651e-01
3.40815902e-01 -4.91393238e-01 3.21716934e-01 5.92914104e-01
1.54141724e-01 2.47115776e-01 -1.26282787e+00 -1.08259931e-01
4.46692646e-01 -1.91022426e-01 5.59518933e-01 2.52119064e-01
5.23128211e-01 -3.57314870e-02 4.53368008e-01 2.82958783e-02
1.40567970e+00 5.48036516e-01 6.77295506e-01 4.59654123e-01
-1.16532579e-01 7.70995378e-01 7.80094147e-01 4.90283698e-01
-3.55531216e-01 8.30372512e-01 2.97785550e-01 3.51332545e-01
-3.25433910e-01 -3.30322027e-01 7.70357966e-01 1.01401448e+00
-3.79975885e-01 2.08021849e-01 -6.71044469e-01 6.54896259e-01
-1.21274149e+00 -1.07346547e+00 -5.46433985e-01 2.58110785e+00
7.51406550e-01 3.48509938e-01 6.22177422e-01 1.24558413e+00
7.32127666e-01 -8.40677842e-02 1.79881960e-01 -7.06408799e-01
-2.71738708e-01 7.04494417e-01 4.99117747e-02 3.51453632e-01
-1.01448750e+00 1.25920177e-01 6.73584652e+00 9.03624475e-01
-1.49659622e+00 -1.40378952e-01 -7.87080079e-02 -1.03995882e-01
-1.37555217e-02 -3.39686275e-01 -1.90827310e-01 5.69161117e-01
1.27226818e+00 -3.75479683e-02 3.53177994e-01 5.33573210e-01
3.58339190e-01 -3.13525617e-01 -5.34326673e-01 1.00016999e+00
-2.69106962e-02 -8.56373787e-01 -3.44143569e-01 2.23846078e-01
4.18213040e-01 -4.49609548e-01 1.81457121e-02 9.19821560e-02
-5.78403115e-01 -9.23155487e-01 1.24725175e+00 5.98697722e-01
8.85752916e-01 -1.07244682e+00 9.36612964e-01 8.72282982e-02
-1.29587018e+00 -2.46393442e-01 1.04668908e-01 -2.65449077e-01
7.34662041e-02 7.13849008e-01 -7.76557028e-01 6.86037898e-01
7.23594606e-01 2.85354316e-01 -4.60636616e-01 1.36352801e+00
-1.75976232e-01 1.03965604e+00 -2.21700802e-01 8.29308257e-02
1.28451496e-01 -1.30140901e-01 7.38614976e-01 1.46457732e+00
6.10517561e-01 -2.42356986e-01 -3.77254397e-01 3.63551199e-01
2.65841305e-01 5.45314074e-01 -1.00726657e-01 -3.02031249e-01
3.43467265e-01 1.02196479e+00 -8.88982177e-01 -1.17418312e-01
2.07843315e-02 7.12293506e-01 -5.23103952e-01 -8.12589675e-02
-5.86151361e-01 -8.23074520e-01 1.65297732e-01 2.22927526e-01
1.95317850e-01 -7.57499561e-02 6.89750016e-02 -6.19677424e-01
1.00201897e-01 -9.28530276e-01 4.41773593e-01 -5.93647182e-01
-9.18043911e-01 4.55346018e-01 -1.23762287e-01 -1.32931447e+00
-3.51954162e-01 -5.62916577e-01 -7.27537394e-01 7.88646519e-01
-9.17393804e-01 -6.99488878e-01 -1.33720502e-01 3.65901470e-01
3.62900347e-01 -3.82693857e-02 1.15904129e+00 5.98660648e-01
1.05883822e-01 4.44911271e-01 2.26458251e-01 1.04399137e-01
6.78186774e-01 -1.46208847e+00 -1.93023309e-01 3.38285148e-01
2.22496897e-01 5.13239741e-01 9.12775874e-01 -4.71084595e-01
-9.56967354e-01 -5.20997822e-01 1.19196284e+00 1.20194912e-01
7.64614105e-01 -6.09842204e-02 -7.57544100e-01 -2.26722751e-02
4.77747470e-02 -5.76448441e-01 1.22304583e+00 -5.76292276e-02
-2.95283169e-01 -1.85890779e-01 -1.12127280e+00 -8.05198587e-03
3.38008493e-01 -9.23209369e-01 -8.36223006e-01 -2.49215916e-01
1.25828654e-01 -1.19383752e-01 -1.22736108e+00 1.65959954e-01
1.18843400e+00 -1.13061559e+00 6.79427207e-01 -4.72441874e-02
2.38291979e-01 -4.49138105e-01 -3.44465792e-01 -9.39295828e-01
-1.48086424e-03 -1.00570250e+00 2.08051473e-01 1.52938139e+00
1.95206150e-01 -4.69788969e-01 3.60110879e-01 -2.60870159e-01
2.68052481e-02 -3.28184187e-01 -1.00985920e+00 -9.30845439e-01
-3.05882514e-01 -7.13139772e-01 3.25349689e-01 9.48029399e-01
1.97931886e-01 1.15167476e-01 -3.59380186e-01 -3.98998529e-01
2.23801196e-01 -8.34307224e-02 5.32355189e-01 -1.64050543e+00
-4.14904058e-01 -5.09334922e-01 -6.27064288e-01 2.22322885e-02
-4.26400810e-01 -5.89682877e-01 -3.58753949e-01 -9.35211360e-01
-2.30139121e-01 -1.35404468e-02 -5.36439419e-01 -7.14433491e-02
9.05186012e-02 8.81380141e-01 4.83842582e-01 2.61971325e-01
1.61672682e-01 -1.08481408e-03 7.59105086e-01 -1.31308325e-02
-4.73564655e-01 2.78104633e-01 -2.46521324e-01 7.41216958e-01
7.75859833e-01 -3.60881358e-01 -1.82544038e-01 3.61255825e-01
3.76878709e-01 1.46452889e-01 3.08494180e-01 -1.30893469e+00
-3.34814280e-01 5.83654130e-03 2.56678641e-01 -6.08371198e-01
5.02194583e-01 -7.46931374e-01 8.28361452e-01 7.88991809e-01
-2.33369455e-01 -1.93498105e-01 2.65111178e-01 1.90116391e-01
-5.88134110e-01 -6.73847497e-01 7.69365251e-01 1.68200463e-01
-4.51292604e-01 -7.90963590e-01 -4.90263700e-01 -3.75408262e-01
7.73936391e-01 -6.08754635e-01 1.83671430e-01 -4.46340621e-01
-9.91398573e-01 -7.05035329e-01 4.12772417e-01 5.87244667e-02
2.31540143e-01 -1.21336031e+00 -6.61558330e-01 1.24715202e-01
-2.63586920e-02 -8.53619158e-01 3.69107723e-01 9.22261775e-01
-7.35513687e-01 2.32332736e-01 -6.30197704e-01 -1.78752437e-01
-1.56452250e+00 2.88625926e-01 1.70186982e-01 -1.29211172e-01
-4.57214147e-01 4.48514581e-01 -3.64321709e-01 1.94770116e-02
2.63185084e-01 -1.04805022e-01 -7.65935361e-01 9.27100062e-01
7.84912705e-01 7.01107562e-01 2.68534064e-01 -7.07678080e-01
-3.57194901e-01 6.54956043e-01 6.05786085e-01 -6.39788389e-01
1.26693153e+00 1.07511185e-01 -8.61247182e-02 1.05232489e+00
1.07794023e+00 1.02087319e+00 -8.13044071e-01 3.37561518e-01
2.74545401e-01 -4.39064950e-01 -1.35891452e-01 -9.48217809e-01
-4.86753881e-01 7.09324360e-01 7.52023399e-01 8.82783830e-01
1.33956659e+00 -5.32980978e-01 5.33571482e-01 -2.50633240e-01
2.47324228e-01 -1.43401325e+00 1.39861014e-02 2.20865801e-01
9.78233159e-01 -3.57491553e-01 -3.04619316e-02 -3.02755237e-01
-3.44833821e-01 1.55533886e+00 -3.71178955e-01 -1.40888631e-01
6.06903970e-01 -1.06521271e-01 2.59979904e-01 1.30137995e-01
-2.30982631e-01 -1.44307867e-01 5.14513969e-01 6.70249045e-01
7.93805063e-01 1.36279792e-01 -9.41391528e-01 9.09769952e-01
-8.43056500e-01 3.36674571e-01 5.16860425e-01 7.57295132e-01
-5.87605536e-01 -1.01008582e+00 -9.37243760e-01 1.51954353e-01
-1.07927811e+00 3.30731064e-01 -7.59054184e-01 8.16456258e-01
4.04581100e-01 1.22682559e+00 -9.54190567e-02 -5.16475201e-01
3.85381937e-01 3.30961764e-01 4.45596367e-01 -1.51191363e-02
-1.14301622e+00 6.70741975e-01 2.69698948e-01 -2.51975745e-01
-4.20233935e-01 -7.39961445e-01 -1.11898518e+00 -1.21466771e-01
-3.62417012e-01 7.22461879e-01 1.06903648e+00 6.05207801e-01
-3.26815158e-01 7.65874147e-01 8.69066596e-01 -6.22293890e-01
-2.69173920e-01 -1.23623121e+00 -9.91060793e-01 3.50688845e-01
1.43686771e-01 -2.10388482e-01 -4.42583531e-01 1.82677940e-01] | [15.751646995544434, 5.3428473472595215] |
97b111e7-4195-4661-96aa-5b7ced36b007 | joint-perceptual-learning-for-enhancement-and | 2307.03536 | null | https://arxiv.org/abs/2307.03536v1 | https://arxiv.org/pdf/2307.03536v1.pdf | Joint Perceptual Learning for Enhancement and Object Detection in Underwater Scenarios | Underwater degraded images greatly challenge existing algorithms to detect objects of interest. Recently, researchers attempt to adopt attention mechanisms or composite connections for improving the feature representation of detectors. However, this solution does \textit{not} eliminate the impact of degradation on image content such as color and texture, achieving minimal improvements. Another feasible solution for underwater object detection is to develop sophisticated deep architectures in order to enhance image quality or features. Nevertheless, the visually appealing output of these enhancement modules do \textit{not} necessarily generate high accuracy for deep detectors. More recently, some multi-task learning methods jointly learn underwater detection and image enhancement, accessing promising improvements. Typically, these methods invoke huge architecture and expensive computations, rendering inefficient inference. Definitely, underwater object detection and image enhancement are two interrelated tasks. Leveraging information coming from the two tasks can benefit each task. Based on these factual opinions, we propose a bilevel optimization formulation for jointly learning underwater object detection and image enhancement, and then unroll to a dual perception network (DPNet) for the two tasks. DPNet with one shared module and two task subnets learns from the two different tasks, seeking a shared representation. The shared representation provides more structural details for image enhancement and rich content information for object detection. Finally, we derive a cooperative training strategy to optimize parameters for DPNet. Extensive experiments on real-world and synthetic underwater datasets demonstrate that our method outputs visually favoring images and higher detection accuracy. | ['Xin Fan', 'Risheng Liu', 'Jiewen Xiao', 'Wanqi Yuan', 'Chenping Fu'] | 2023-07-07 | null | null | null | null | ['image-enhancement', 'object-detection', 'bilevel-optimization', 'multi-task-learning'] | ['computer-vision', 'computer-vision', 'methodology', 'methodology'] | [ 4.10767108e-01 1.40731081e-01 5.97412407e-01 -3.88854474e-01
-6.39866114e-01 -1.76433608e-01 5.80031872e-02 -7.59378821e-02
-9.41291749e-01 5.04958093e-01 3.27978767e-02 1.58828214e-01
-1.89408496e-01 -8.80244553e-01 -9.25259054e-01 -1.14368463e+00
-1.23675160e-01 -4.89321768e-01 4.86710995e-01 -3.98137838e-01
2.32924223e-01 5.92226870e-02 -1.71248126e+00 1.17302448e-01
1.04482388e+00 1.15960729e+00 9.25932467e-01 8.54918540e-01
1.90363064e-01 8.17437649e-01 -5.00560224e-01 -4.11653161e-01
4.53759134e-01 -1.29057899e-01 -8.54901895e-02 -5.11639081e-02
5.20535111e-01 -1.02958393e+00 -4.15828615e-01 1.50650311e+00
9.71777380e-01 2.05649868e-01 4.31555986e-01 -8.44144762e-01
-8.84953737e-01 7.79079616e-01 -7.96001911e-01 4.06826884e-01
-1.25745818e-01 2.10580736e-01 9.38188612e-01 -1.13285089e+00
-1.04156129e-01 1.28201020e+00 7.10579038e-01 4.82566774e-01
-6.60674572e-01 -6.82616055e-01 5.19510150e-01 9.41694826e-02
-9.18035626e-01 -5.10322213e-01 6.59982979e-01 -7.25581497e-02
5.72507679e-01 6.39152676e-02 5.94090283e-01 7.25235403e-01
1.72273606e-01 1.09974062e+00 1.00325489e+00 6.85312152e-02
-1.26689561e-02 1.65971220e-01 -3.06319185e-02 7.21251428e-01
6.10160649e-01 1.78769305e-01 -4.37976032e-01 2.85671830e-01
7.44298041e-01 3.31649154e-01 -5.10381281e-01 1.23657249e-01
-6.43219590e-01 6.19183362e-01 7.22229183e-01 -1.21173756e-02
-2.94372648e-01 3.20344537e-01 1.46020621e-01 3.22318554e-01
2.02356815e-01 3.15035969e-01 -5.04785061e-01 3.89408171e-01
-3.68260503e-01 3.79296802e-02 3.22352231e-01 7.66057074e-01
1.14878356e+00 2.86970735e-01 -1.84422992e-02 8.99447262e-01
8.67004335e-01 9.42653954e-01 3.44759613e-01 -8.77162576e-01
4.81989235e-01 4.15338486e-01 4.16375361e-02 -1.03772402e+00
-6.88662171e-01 -5.55697560e-01 -8.22421730e-01 6.33955836e-01
1.87968746e-01 -5.79370737e-01 -8.85788977e-01 1.65161288e+00
1.45901084e-01 4.41364571e-02 5.86648524e-01 1.22307348e+00
1.23801577e+00 7.67880678e-01 1.07796222e-01 -2.32321396e-02
1.27418196e+00 -9.79429901e-01 -6.83124542e-01 -5.82971990e-01
3.63910764e-01 -4.95119780e-01 6.70581043e-01 1.96822241e-01
-1.21309555e+00 -6.61938787e-01 -1.37326610e+00 -1.72684267e-01
-1.35477751e-01 4.31961447e-01 5.48930228e-01 5.14696002e-01
-1.00465286e+00 5.53513408e-01 -6.84493065e-01 1.66876912e-01
3.56796354e-01 4.95905608e-01 -1.08013175e-01 -1.99244514e-01
-1.09321713e+00 8.16845417e-01 3.42196763e-01 5.88268876e-01
-1.39274073e+00 -4.98131633e-01 -1.14093304e+00 1.33713469e-01
2.91011155e-01 -4.77569371e-01 1.23863566e+00 -8.53076994e-01
-1.39634180e+00 3.91239405e-01 3.63848537e-01 -2.29007587e-01
4.65202749e-01 -5.57870030e-01 -1.11753687e-01 2.28289187e-01
3.75013873e-02 6.35347724e-01 1.02948833e+00 -1.51766253e+00
-1.09787583e+00 -4.10306245e-01 3.69881392e-01 5.91152906e-01
-7.68957376e-01 1.98536236e-02 -4.77714270e-01 -5.03458440e-01
3.72529626e-01 -3.36239278e-01 -4.72353160e-01 6.21450603e-01
-1.96455300e-01 -3.13688517e-02 1.01988804e+00 -6.01514518e-01
6.06961131e-01 -2.28853893e+00 1.43920749e-01 -2.27852121e-01
3.27156991e-01 2.86254227e-01 -5.33505261e-01 -8.37075859e-02
4.61211026e-01 -1.05704524e-01 -3.58379513e-01 -6.12097919e-01
-1.51921332e-01 3.98788899e-01 -5.91329597e-02 6.78198040e-01
2.33788311e-01 6.88944578e-01 -1.07226110e+00 -5.37270367e-01
2.07731247e-01 4.07103449e-01 -5.61441243e-01 4.25580919e-01
1.32035241e-01 1.90607250e-01 -5.91458261e-01 8.79580319e-01
9.77689564e-01 4.53916714e-02 -1.13322526e-01 -4.95636702e-01
-4.05973643e-01 -3.64859492e-01 -1.37495375e+00 1.31488955e+00
-5.09053528e-01 6.27945781e-01 7.37148046e-01 -9.79145706e-01
1.10498631e+00 1.75935065e-06 2.31429562e-01 -1.01265311e+00
3.84699017e-01 3.21339160e-01 -5.82592450e-02 -9.83789325e-01
5.80032885e-01 -2.39930660e-01 2.53583223e-01 1.34422705e-01
3.58542278e-02 5.69455922e-02 -4.24628350e-04 4.21524867e-02
6.64855003e-01 3.29678133e-02 -1.80743597e-02 -2.69480050e-01
2.71152377e-01 -6.03712022e-01 8.41148734e-01 9.90904510e-01
-3.52145851e-01 5.23646235e-01 -1.87003762e-01 -3.11318219e-01
-8.97808075e-01 -9.06963408e-01 -1.48532733e-01 1.45731437e+00
1.08056784e+00 3.75049591e-01 -4.02687371e-01 -3.33592832e-01
-1.48694292e-01 -4.19146307e-02 -7.64622390e-01 -3.66151810e-01
-5.56945622e-01 -1.16876829e+00 6.08533919e-01 7.57381499e-01
7.50571966e-01 -1.00552118e+00 -8.61319721e-01 3.06982189e-01
-9.55268927e-03 -1.03058958e+00 -2.40131691e-01 4.79146987e-01
-7.22144485e-01 -9.39488649e-01 -1.08876574e+00 -1.10611701e+00
7.18639731e-01 8.34492862e-01 6.52929127e-01 4.63679194e-01
-2.70170182e-01 1.71447262e-01 -5.24384141e-01 -7.60588765e-01
-4.34454866e-02 -5.93218982e-01 8.99240300e-02 1.89565390e-01
-1.50610819e-01 -4.74038482e-01 -1.00997651e+00 2.71789283e-01
-1.17018664e+00 -1.40537873e-01 1.09241927e+00 8.47473264e-01
1.16214685e-01 -2.77683705e-01 5.68769753e-01 -1.60410136e-01
2.71206617e-01 -4.41403329e-01 -6.18845105e-01 1.69138342e-01
-2.40605384e-01 2.65040202e-03 2.87476093e-01 -4.39503163e-01
-1.33517349e+00 1.24174573e-01 -3.28590184e-01 -3.12381834e-01
1.91539422e-01 5.11604190e-01 -2.53588349e-01 -4.05851573e-01
4.62147981e-01 5.96871853e-01 1.29858136e-01 -4.10194069e-01
-4.15521227e-02 5.76227486e-01 5.38533032e-01 -1.66950032e-01
8.71548295e-01 7.69201398e-01 -4.11790192e-01 -8.87042403e-01
-1.09506595e+00 -3.82602572e-01 -1.34038001e-01 -4.31613058e-01
9.36601102e-01 -1.50353718e+00 -8.00134063e-01 8.84037256e-01
-1.13174915e+00 -3.86205882e-01 1.09406345e-01 5.56681275e-01
-8.69428515e-02 6.41105592e-01 -7.72067428e-01 -1.11452448e+00
-4.39291656e-01 -1.18624604e+00 1.19469547e+00 9.01141524e-01
9.53251004e-01 -8.07274878e-01 -2.29538321e-01 1.63743705e-01
6.09302998e-01 -7.89521411e-02 1.48273110e-01 -1.27803221e-01
-6.58921599e-01 1.19956478e-01 -8.09276283e-01 5.28915405e-01
2.71949153e-02 -5.39379716e-02 -1.21077764e+00 -5.81990898e-01
-1.80539656e-02 -4.90882665e-01 1.43392372e+00 4.47868705e-01
1.07957745e+00 -1.98030904e-01 -1.76435739e-01 9.35720205e-01
1.57445025e+00 1.73981652e-01 6.20431006e-01 4.70163137e-01
7.30940759e-01 7.63129652e-01 5.96383929e-01 7.84223795e-01
6.44043922e-01 2.23038107e-01 1.10541666e+00 -5.43209195e-01
-2.01809153e-01 4.94209230e-02 6.47103667e-01 5.17428935e-01
-3.61770630e-01 -1.51068375e-01 -3.51280659e-01 6.42802358e-01
-1.61395621e+00 -7.77913928e-01 -1.34913653e-01 1.55681479e+00
5.70447028e-01 -2.39781275e-01 -2.61690319e-01 -1.08876869e-01
7.11487472e-01 3.92387658e-02 -8.20775926e-01 3.03403109e-01
-6.19089842e-01 -4.34131235e-01 7.69171178e-01 2.35130519e-01
-1.26191819e+00 5.93602657e-01 5.69659376e+00 4.51343775e-01
-9.06319737e-01 7.08074123e-02 4.93390352e-01 1.60942256e-01
-3.45415831e-01 -2.77280360e-01 -1.08133972e+00 3.02514881e-01
1.55215681e-01 1.46814600e-01 -1.20134167e-02 8.68562400e-01
2.19662651e-01 -2.11742651e-02 -6.41865730e-01 1.00868273e+00
2.79282957e-01 -1.12378943e+00 4.88970615e-02 -5.65509275e-02
8.65928471e-01 2.37537980e-01 7.83182234e-02 4.42666501e-01
4.82708007e-01 -6.23959720e-01 1.06812775e+00 5.35482585e-01
3.68863434e-01 -2.63457417e-01 1.07945585e+00 2.60415524e-01
-1.34156430e+00 -8.26672912e-01 -8.66819501e-01 -2.86568701e-01
3.09376985e-01 3.39381963e-01 -1.00245802e-02 2.34440759e-01
1.28755844e+00 7.98714459e-01 -5.10355949e-01 1.43007147e+00
-2.61244386e-01 1.15018964e-01 -3.23875129e-01 -2.40262926e-01
3.30634505e-01 -2.27073669e-01 6.51137590e-01 1.21066737e+00
4.96235758e-01 4.95244175e-01 2.26788685e-01 6.09298050e-01
-1.38986468e-01 -6.16355501e-02 -5.15794754e-01 4.66564149e-01
2.52646685e-01 1.38160956e+00 -4.96362329e-01 -3.75381052e-01
-7.02443957e-01 7.77234733e-01 2.22158879e-01 3.42098534e-01
-8.42908561e-01 -5.97361028e-01 8.55883718e-01 -4.95267481e-01
3.97203416e-01 -1.63189650e-01 -2.80936301e-01 -1.05231643e+00
-4.64853048e-02 -6.22701108e-01 3.17935795e-01 -6.76562369e-01
-1.21406794e+00 5.75923145e-01 -4.55301911e-01 -1.37911510e+00
7.01650023e-01 -8.27046037e-01 -6.96726441e-01 5.91208816e-01
-2.27806282e+00 -1.03490365e+00 -8.95247161e-01 4.88646209e-01
7.48447776e-01 1.91398948e-01 1.48556530e-01 6.58247471e-01
-7.90943146e-01 6.84739769e-01 2.21737668e-01 3.58267337e-01
4.89966452e-01 -1.25041652e+00 -2.21882775e-01 1.21788013e+00
-2.79579252e-01 1.59213498e-01 6.54644370e-01 -4.13806379e-01
-1.61618447e+00 -1.22494066e+00 -4.96813888e-03 -3.74763273e-02
5.49664438e-01 2.37832516e-02 -1.06019628e+00 3.59630704e-01
1.96430176e-01 -1.36534069e-02 4.07675445e-01 -4.84670848e-01
-1.11113926e-02 -3.33064288e-01 -8.25067043e-01 4.30428296e-01
9.08117592e-01 -1.02568761e-01 -5.03428757e-01 1.85983613e-01
8.26978266e-01 -3.06967527e-01 -7.37758219e-01 7.39977777e-01
4.44895118e-01 -1.04107463e+00 1.01877308e+00 -1.01873383e-01
5.79944313e-01 -6.33842826e-01 -4.10256177e-01 -1.16706169e+00
-1.03644855e-01 -2.93096781e-01 7.65888169e-02 9.64599431e-01
6.04629338e-01 -5.34671009e-01 4.38108265e-01 3.31670851e-01
-8.36682737e-01 -4.94652778e-01 -5.73943913e-01 -5.88888109e-01
-9.18595344e-02 -2.68463910e-01 3.16253871e-01 6.07631445e-01
-3.49975586e-01 1.89933240e-01 -8.30449224e-01 9.37951326e-01
1.05766869e+00 3.15416336e-01 6.79157674e-01 -1.18569756e+00
-3.80216658e-01 -4.38895077e-01 -1.30482718e-01 -1.53963447e+00
-4.84594166e-01 -4.10159826e-01 7.67568469e-01 -1.90819526e+00
5.04314303e-01 -3.12515259e-01 -3.25048387e-01 7.24416733e-01
-6.83149278e-01 6.99377179e-01 2.24369153e-01 1.15800016e-01
-8.83681118e-01 1.04191351e+00 1.54869366e+00 -3.10764551e-01
-2.64486577e-02 -8.45448300e-02 -1.06844008e+00 8.18639755e-01
5.64693034e-01 -2.52259910e-01 -6.23237761e-03 -1.09650123e+00
2.97812790e-01 -8.20482373e-02 4.26134646e-01 -1.09511101e+00
5.87448657e-01 -2.42041871e-02 4.51097965e-01 -2.97490299e-01
5.25133431e-01 -6.05207086e-01 -5.03504097e-01 7.69930720e-01
3.83105827e-04 -4.14200187e-01 1.82797417e-01 8.90021086e-01
-3.37121159e-01 -4.36376452e-01 1.15263748e+00 -2.37775579e-01
-1.29037225e+00 4.26979333e-01 -4.35038775e-01 -3.39840502e-01
9.75715816e-01 -5.02866328e-01 -5.34858882e-01 -2.47565731e-01
-6.21222436e-01 7.32233465e-01 -3.73523459e-02 5.03005683e-01
1.09865355e+00 -6.04233623e-01 -1.14068091e+00 -6.45268410e-02
4.58320864e-02 1.19158141e-01 6.88816190e-01 8.10560763e-01
-4.08925325e-01 -4.41847503e-01 -2.99306601e-01 -5.56277335e-01
-1.10208917e+00 1.67178974e-01 6.10465825e-01 2.65087157e-01
-5.91254771e-01 1.30396748e+00 6.95136845e-01 -3.65633219e-01
2.71995872e-01 -2.61439204e-01 -6.32113695e-01 2.17518166e-01
9.69567955e-01 4.04605329e-01 -3.99400145e-01 -3.17614466e-01
-4.82450798e-02 7.62781441e-01 -2.42050570e-02 3.63971174e-01
1.76461875e+00 -6.91747785e-01 5.54791540e-02 -9.30569172e-02
9.79357898e-01 -3.09283674e-01 -2.00773978e+00 -3.75257432e-01
-4.09531981e-01 -3.59781623e-01 2.25445285e-01 -5.05451798e-01
-1.39245403e+00 9.57720280e-01 7.98513532e-01 1.62425682e-01
1.38349962e+00 7.98813850e-02 6.99463427e-01 6.72155678e-01
3.20335478e-01 -1.02217698e+00 6.56717539e-01 5.08737862e-01
8.67783725e-01 -1.85697675e+00 3.68195362e-02 -3.00444067e-01
-5.27721584e-01 1.07423949e+00 1.07637775e+00 -1.07973829e-01
3.45242023e-01 5.27939141e-01 4.74482417e-01 -2.22022563e-01
-3.11377078e-01 -5.68960845e-01 -1.00409940e-01 5.60454369e-01
-1.03118770e-01 -2.89249808e-01 -8.64577293e-02 7.92732477e-01
2.75252938e-01 -7.03185022e-01 6.94572926e-01 8.46918046e-01
-1.19160140e+00 -4.83729213e-01 -3.65975261e-01 3.59833121e-01
-5.17868221e-01 -2.62199163e-01 8.26748982e-02 5.07768929e-01
2.76327014e-01 1.05558145e+00 -8.84355456e-02 -3.41582060e-01
3.84510905e-01 -9.65861917e-01 1.66421123e-02 -3.68653536e-01
-3.31758022e-01 7.80826062e-02 -2.06648782e-01 -2.73385376e-01
-5.86203516e-01 -5.13133168e-01 -1.27587247e+00 3.88818830e-01
-7.45973885e-01 1.66673675e-01 6.81076705e-01 8.83582890e-01
-1.05054937e-01 7.51170874e-01 8.35130990e-01 -1.35922801e+00
-6.95734859e-01 -1.16478705e+00 -7.86946535e-01 7.54458755e-02
6.54773057e-01 -6.51348233e-01 -6.46059752e-01 -4.69449423e-02] | [10.674208641052246, -3.5138609409332275] |
5efcf9c1-852f-47cb-ae3f-304bc8cdb39f | boosting-adversarial-robustness-from-the | 2210.05118 | null | https://arxiv.org/abs/2210.05118v1 | https://arxiv.org/pdf/2210.05118v1.pdf | Boosting Adversarial Robustness From The Perspective of Effective Margin Regularization | The adversarial vulnerability of deep neural networks (DNNs) has been actively investigated in the past several years. This paper investigates the scale-variant property of cross-entropy loss, which is the most commonly used loss function in classification tasks, and its impact on the effective margin and adversarial robustness of deep neural networks. Since the loss function is not invariant to logit scaling, increasing the effective weight norm will make the loss approach zero and its gradient vanish while the effective margin is not adequately maximized. On typical DNNs, we demonstrate that, if not properly regularized, the standard training does not learn large effective margins and leads to adversarial vulnerability. To maximize the effective margins and learn a robust DNN, we propose to regularize the effective weight norm during training. Our empirical study on feedforward DNNs demonstrates that the proposed effective margin regularization (EMR) learns large effective margins and boosts the adversarial robustness in both standard and adversarial training. On large-scale models, we show that EMR outperforms basic adversarial training, TRADES and two regularization baselines with substantial improvement. Moreover, when combined with several strong adversarial defense methods (MART and MAIL), our EMR further boosts the robustness. | ['Antoni B. Chan', 'Ziquan Liu'] | 2022-10-11 | null | null | null | null | ['adversarial-defense'] | ['adversarial'] | [ 1.52021542e-01 2.58664429e-01 -2.74361670e-02 -3.39253783e-01
-6.62389576e-01 -9.43078518e-01 5.26284754e-01 -2.34979272e-01
-6.83002174e-01 7.68554211e-01 -1.24628358e-01 -5.24930418e-01
-3.51750903e-04 -8.86584103e-01 -1.07052362e+00 -8.47416937e-01
-1.46062389e-01 -2.09920928e-01 3.14535469e-01 -3.43830258e-01
-1.83410645e-01 8.75971138e-01 -8.01801801e-01 -2.42430791e-01
9.38518703e-01 1.14227867e+00 -3.80869269e-01 4.93068814e-01
3.56699884e-01 9.30244863e-01 -9.40645397e-01 -8.32934022e-01
7.10549057e-01 -5.69891147e-02 -6.04287684e-01 -4.64411646e-01
7.92185724e-01 -5.37150443e-01 -9.05863702e-01 1.47309685e+00
6.46840036e-01 1.63902566e-01 6.52010441e-01 -1.23993611e+00
-7.49555886e-01 8.37531269e-01 -3.28460991e-01 2.74520308e-01
-4.62524235e-01 1.37864143e-01 1.00083578e+00 -5.24042428e-01
3.04015130e-01 1.37121403e+00 9.77248490e-01 8.60441566e-01
-1.20149601e+00 -9.26808417e-01 3.13546807e-01 -2.35702783e-01
-1.18745995e+00 -2.28511319e-01 7.77432084e-01 -2.42395729e-01
5.90863585e-01 1.11496098e-01 6.94031194e-02 1.48590481e+00
3.80812436e-01 6.29913747e-01 7.42713869e-01 -9.56271663e-02
3.08891535e-01 1.27982974e-01 1.05423257e-01 5.10652661e-01
2.56929576e-01 5.25172174e-01 1.35188699e-01 -1.54386178e-01
7.99698353e-01 3.43221612e-02 -3.61435920e-01 -1.70509487e-01
-6.04370475e-01 8.56648088e-01 7.87398279e-01 7.85024241e-02
-2.80154757e-02 4.82645273e-01 6.83984756e-01 6.19529843e-01
4.93170083e-01 6.06863141e-01 -4.71306622e-01 1.97409421e-01
-5.71601272e-01 2.94480294e-01 6.94528461e-01 6.12717807e-01
5.55296957e-01 5.24547100e-01 -1.29923955e-01 9.83196735e-01
1.41376659e-01 7.20006287e-01 3.10510933e-01 -9.33938682e-01
7.42226899e-01 4.79339391e-01 -3.00801694e-01 -8.70219827e-01
-2.86813021e-01 -6.63027525e-01 -1.14235413e+00 8.41451406e-01
5.85150361e-01 -6.92442477e-01 -8.57373655e-01 2.24361539e+00
5.80154285e-02 9.30301554e-04 2.47353017e-01 6.31953657e-01
3.83174717e-01 4.77127731e-01 1.60482019e-01 2.08373651e-01
7.85053551e-01 -6.56266749e-01 -6.17140710e-01 -5.96040547e-01
4.09378678e-01 -5.13310730e-01 1.00577903e+00 1.97774604e-01
-1.12636387e+00 -3.10926378e-01 -1.42720282e+00 1.45718217e-01
-4.48926300e-01 -3.79339188e-01 2.46358812e-01 8.72666478e-01
-8.57369006e-01 1.11950076e+00 -8.87578309e-01 1.73292607e-01
7.27438569e-01 5.85498393e-01 -4.14506048e-01 8.52980018e-02
-1.69725561e+00 1.00860631e+00 4.28132266e-01 2.29524402e-03
-1.11855030e+00 -1.02649045e+00 -6.89227581e-01 5.75746633e-02
1.94530979e-01 -3.63245249e-01 9.60485935e-01 -1.16092300e+00
-1.50578344e+00 5.14920115e-01 7.28978813e-01 -8.62273812e-01
9.04778361e-01 -4.87579703e-01 -4.03385490e-01 8.29334334e-02
-5.28484404e-01 5.25973678e-01 1.00714052e+00 -1.10891461e+00
-2.14607324e-02 -3.16179186e-01 4.60887015e-01 -8.13153088e-02
-9.17546451e-01 -1.05564542e-01 3.33897062e-02 -1.12121093e+00
-4.04717267e-01 -1.01530159e+00 -2.33901784e-01 1.77692965e-01
-5.25269449e-01 1.41449586e-01 8.96071732e-01 -5.77288210e-01
1.16907012e+00 -2.28555274e+00 4.88615446e-02 3.68358970e-01
2.27215812e-01 7.46712506e-01 -2.46284425e-01 1.88059434e-02
-3.70884329e-01 4.27862614e-01 -4.95784581e-01 -5.12972996e-02
1.11081481e-01 4.24993157e-01 -5.68790197e-01 6.05931997e-01
4.35412765e-01 8.61827672e-01 -6.33535743e-01 9.45063382e-02
2.60249544e-02 7.28746295e-01 -6.93684101e-01 3.19087356e-01
6.01547584e-02 6.42999485e-02 -2.60884672e-01 3.37854773e-01
9.66203868e-01 1.09496906e-01 -4.22457419e-02 -9.88398716e-02
3.71024758e-01 -1.47220224e-01 -1.01443970e+00 1.03264153e+00
-4.92759317e-01 8.66013944e-01 1.17471896e-01 -1.08507705e+00
8.55976105e-01 1.48573995e-01 -4.33307141e-03 -4.87539977e-01
2.13313341e-01 2.13390306e-01 1.16601937e-01 5.02964444e-02
5.59832202e-03 -2.53719211e-01 5.30623039e-03 9.97376591e-02
1.08988911e-01 1.07811220e-01 -2.86770582e-01 1.81484848e-01
1.35573053e+00 -3.22182536e-01 -6.16162568e-02 -3.79418105e-01
5.21969199e-01 -6.07581854e-01 6.26837194e-01 6.65918589e-01
-3.85529011e-01 4.60327804e-01 8.33956301e-01 -3.08337927e-01
-1.07741714e+00 -1.36234474e+00 -4.64988559e-01 1.06581616e+00
-2.14246102e-02 2.30880342e-02 -1.13368821e+00 -1.13029623e+00
3.09781998e-01 5.04607856e-01 -8.38857293e-01 -8.60596716e-01
-7.18785644e-01 -8.65095496e-01 1.16691697e+00 7.94399142e-01
7.72546172e-01 -8.14968526e-01 6.42536283e-02 -5.71433082e-02
2.31706128e-01 -9.04287875e-01 -6.39619589e-01 4.44769740e-01
-8.16382408e-01 -8.34486723e-01 -7.19371259e-01 -4.83879358e-01
7.99241662e-01 -2.61558920e-01 9.96908724e-01 -7.17297569e-02
-1.02565117e-01 7.76406303e-02 -4.98649105e-02 -4.41028178e-01
-6.71569467e-01 7.00151771e-02 2.81305671e-01 -1.40929878e-01
-1.31872937e-01 -7.60206819e-01 -4.38610673e-01 4.36749160e-01
-1.31238627e+00 -8.35630238e-01 4.33457345e-01 1.02118564e+00
2.92637199e-01 7.19866380e-02 9.14294362e-01 -1.00822461e+00
6.55838788e-01 -3.42463195e-01 -7.53046632e-01 1.67674959e-01
-6.37957990e-01 4.20390666e-01 1.21369374e+00 -8.70939314e-01
-8.90161514e-01 -4.14837152e-01 -5.43809235e-01 -8.07932377e-01
1.94436103e-01 8.32464844e-02 -4.28907871e-01 -6.22285903e-01
1.05411565e+00 -3.33551317e-01 -8.82810429e-02 -3.01276088e-01
3.87707025e-01 1.95577249e-01 4.59136963e-01 -4.34726357e-01
1.29697490e+00 2.51285970e-01 1.58657208e-01 -4.11324263e-01
-8.53142858e-01 3.05194259e-01 -3.41175705e-01 2.34003384e-02
6.61307871e-01 -7.92568564e-01 -6.93774283e-01 7.12207735e-01
-9.09243345e-01 -5.95319271e-01 -4.36501563e-01 2.90071338e-01
-1.87102348e-01 3.00494105e-01 -7.52644777e-01 -6.63304985e-01
-4.57764596e-01 -1.08172190e+00 4.07284141e-01 1.16978772e-02
1.30334109e-01 -1.40708542e+00 -6.02459162e-02 -1.64094537e-01
6.37936234e-01 7.95117438e-01 9.26396728e-01 -9.05059814e-01
-1.77137867e-01 -4.55180973e-01 -1.20185360e-01 1.31415927e+00
-9.00316089e-02 7.81719759e-02 -1.18007028e+00 -4.30792272e-01
1.03550173e-01 -5.82564592e-01 1.08931279e+00 3.14766377e-01
1.13045216e+00 -6.64844215e-01 2.24527836e-01 1.05216193e+00
1.43124914e+00 6.06259368e-02 7.00244188e-01 4.68924254e-01
8.59094679e-01 3.70348245e-01 -5.88463573e-03 1.91445470e-01
-2.60825723e-01 3.11131775e-01 8.77942026e-01 -2.31301621e-01
1.30293608e-01 -7.47271255e-02 7.77667701e-01 4.23915476e-01
3.34138796e-02 -2.55916655e-01 -7.76864588e-01 1.13681622e-01
-1.55297554e+00 -8.87858689e-01 3.67854267e-01 2.17120743e+00
1.09399557e+00 5.42480350e-01 -1.24105588e-01 2.68391967e-01
6.78903878e-01 3.49917412e-01 -1.03244364e+00 -6.65874302e-01
-4.00679469e-01 2.32618138e-01 1.03738368e+00 6.70670807e-01
-1.30472112e+00 8.75691414e-01 6.72649813e+00 9.51134682e-01
-1.03630722e+00 5.49200699e-02 8.57386887e-01 -2.09148824e-01
-2.55754620e-01 -4.66102511e-01 -6.84551537e-01 3.50154817e-01
1.02580988e+00 -1.16064355e-01 5.38556457e-01 1.17886913e+00
-2.07333490e-01 6.91393971e-01 -1.12262142e+00 4.79575455e-01
-2.56820858e-01 -1.20855606e+00 1.28623063e-03 1.31048501e-01
8.84124517e-01 1.63511515e-01 5.56996584e-01 5.89436173e-01
6.34568334e-01 -1.04115474e+00 6.86123431e-01 1.72233805e-01
6.57773376e-01 -1.31344450e+00 8.88635993e-01 1.42698348e-01
-8.40971291e-01 -3.52656960e-01 -6.40467703e-01 3.40619653e-01
-1.14446558e-01 5.88268876e-01 -3.43823254e-01 1.29758403e-01
4.80946392e-01 4.36752647e-01 -5.03578722e-01 5.08994818e-01
-4.16266084e-01 6.68528914e-01 -1.56335175e-01 3.41435105e-01
3.39951932e-01 -3.12703885e-02 6.63681090e-01 1.15152085e+00
-1.22779012e-01 -2.53546774e-01 -1.40587226e-01 8.59876394e-01
-7.26480663e-01 -1.61704853e-01 -6.74052894e-01 -5.95669597e-02
4.90191132e-01 1.02052629e+00 -1.29465565e-01 -3.87540869e-02
-2.05368102e-01 9.47145224e-01 5.83947778e-01 6.11652851e-01
-1.11584747e+00 -7.66417086e-01 1.19422245e+00 -1.07095800e-01
1.23911574e-01 4.62805815e-02 -3.87118131e-01 -9.60065007e-01
8.75893831e-02 -9.51415062e-01 2.03589231e-01 -4.64133434e-02
-1.65540242e+00 8.25831831e-01 -3.01085442e-01 -1.06790221e+00
-1.05093241e-01 -9.45679963e-01 -8.42722893e-01 9.08714533e-01
-1.60536313e+00 -7.28982210e-01 2.27645069e-01 8.16445172e-01
3.32249999e-02 -3.65268886e-01 6.20995045e-01 5.36770165e-01
-1.10349488e+00 1.44914174e+00 5.14383554e-01 6.08105838e-01
7.89986074e-01 -1.35724437e+00 6.90222740e-01 1.06124306e+00
-3.93033177e-01 8.40565085e-01 6.78599119e-01 -3.45078677e-01
-1.09059739e+00 -1.47239363e+00 2.44311750e-01 -3.29108715e-01
1.08657658e+00 -5.04077494e-01 -1.11817026e+00 7.19122410e-01
1.52003653e-02 4.80896920e-01 6.00813746e-01 -3.30258697e-01
-8.26220453e-01 -4.81370866e-01 -1.51800597e+00 7.40050614e-01
8.96227777e-01 -6.79710507e-01 -4.12873298e-01 3.62265497e-01
1.17533076e+00 -2.56230235e-01 -1.08087432e+00 5.42726338e-01
4.76628095e-01 -8.69054317e-01 1.29489005e+00 -8.91246855e-01
4.00283277e-01 2.68574148e-01 -2.51546681e-01 -1.43639255e+00
-1.37356550e-01 -7.70664930e-01 -2.60415286e-01 1.36859751e+00
5.25955141e-01 -9.12446856e-01 6.78429544e-01 9.27764535e-01
-1.65155023e-01 -8.51031482e-01 -9.49302912e-01 -1.24939191e+00
7.49010801e-01 -3.88798356e-01 3.10420424e-01 8.62004757e-01
-3.58152360e-01 -2.11794659e-01 -4.19198900e-01 4.01229411e-01
7.87990868e-01 -7.52737045e-01 3.71822089e-01 -1.07010961e+00
-2.37387270e-01 -6.02712274e-01 -5.45955837e-01 -7.74151266e-01
6.14288092e-01 -8.97486508e-01 -1.50166392e-01 -8.40847015e-01
-3.09889823e-01 -3.11998397e-01 -6.42998219e-01 6.12276256e-01
-3.85430098e-01 2.59239167e-01 2.12544620e-01 -1.33207932e-01
-1.69060171e-01 4.71401185e-01 9.94978070e-01 -1.83046490e-01
-3.59150358e-02 1.36168525e-01 -8.10709417e-01 8.99781406e-01
8.68489921e-01 -5.90542316e-01 -4.37018275e-01 -3.36349845e-01
1.83036506e-01 -5.51127672e-01 3.06780308e-01 -1.07225060e+00
-6.36304095e-02 -5.47404513e-02 4.04736549e-01 1.42956510e-01
8.76586512e-02 -1.03653157e+00 -2.67827451e-01 6.52371526e-01
-6.35832071e-01 4.19345796e-02 4.57210809e-01 6.24052942e-01
-3.13586928e-02 -2.83457607e-01 1.44493568e+00 1.64838836e-01
-1.97109327e-01 5.62534690e-01 -1.32416129e-01 6.05347037e-01
9.38786328e-01 1.58876956e-01 -4.82045025e-01 -9.54177976e-02
-5.14977813e-01 1.20463997e-01 3.49533886e-01 3.42053205e-01
4.60101306e-01 -1.60265338e+00 -5.95597804e-01 1.74066514e-01
-2.54050732e-01 -1.14619054e-01 2.23140821e-01 3.35320771e-01
-5.66074729e-01 -4.13035043e-02 -2.53658175e-01 -1.46028429e-01
-1.00446105e+00 6.54178619e-01 7.16738343e-01 -4.25959140e-01
-4.13049370e-01 1.17756236e+00 4.49978173e-01 -4.26388770e-01
7.45389640e-01 -2.09726036e-01 -1.64151192e-02 -9.07905996e-02
5.91958582e-01 5.67557693e-01 1.29175205e-02 -3.35154355e-01
-3.72184813e-01 4.33603883e-01 -2.06517845e-01 6.70469254e-02
1.25950170e+00 2.21859276e-01 1.45143300e-01 5.61120063e-02
1.68614757e+00 -1.07080646e-01 -1.83003986e+00 -3.17935973e-01
-1.61567569e-01 -1.61262766e-01 1.76213905e-01 -6.83801711e-01
-1.54120553e+00 1.03418291e+00 6.87758088e-01 3.03590894e-01
1.16485822e+00 -4.42344755e-01 8.51473808e-01 6.28518760e-01
-1.05748631e-01 -1.02515554e+00 2.82820642e-01 7.65845001e-01
9.14645851e-01 -1.27580774e+00 -2.48873025e-01 -2.98891701e-02
-7.24762440e-01 1.06578779e+00 7.31954336e-01 -5.20037055e-01
7.90074050e-01 6.89837694e-01 1.31072059e-01 3.16297740e-01
-4.75729972e-01 3.90688926e-01 2.85351813e-01 5.50316930e-01
9.52664167e-02 -2.03470483e-01 1.35010600e-01 5.73468268e-01
-2.10720137e-01 -6.33367121e-01 2.15968624e-01 6.72565639e-01
-4.02272463e-01 -1.09601152e+00 -3.03072006e-01 2.04419166e-01
-1.02594268e+00 -2.20875695e-01 -4.05396193e-01 7.91391730e-01
-3.56818624e-02 6.79035485e-01 -1.15607575e-01 -5.58867097e-01
4.90789801e-01 3.00781615e-02 1.59310043e-01 -1.63226962e-01
-1.00965929e+00 -4.62475866e-01 -1.80634156e-01 -4.54809368e-01
8.30799490e-02 -1.64998427e-01 -1.16022098e+00 -6.22795165e-01
-4.40746665e-01 -1.73870504e-01 5.08745611e-01 7.79747725e-01
1.47977397e-02 7.21727610e-01 1.05845165e+00 -6.21023238e-01
-1.36749125e+00 -8.27429831e-01 -6.10104442e-01 5.44958591e-01
7.27437615e-01 -4.50320214e-01 -1.01938736e+00 -2.32198805e-01] | [5.661258697509766, 7.8669915199279785] |
8e117d1e-f389-4f64-94cb-cb66ab498937 | abmdrnet-adaptive-weighted-bi-directional | null | null | http://openaccess.thecvf.com//content/CVPR2021/html/Zhang_ABMDRNet_Adaptive-Weighted_Bi-Directional_Modality_Difference_Reduction_Network_for_RGB-T_Semantic_CVPR_2021_paper.html | http://openaccess.thecvf.com//content/CVPR2021/papers/Zhang_ABMDRNet_Adaptive-Weighted_Bi-Directional_Modality_Difference_Reduction_Network_for_RGB-T_Semantic_CVPR_2021_paper.pdf | ABMDRNet: Adaptive-Weighted Bi-Directional Modality Difference Reduction Network for RGB-T Semantic Segmentation | Semantic segmentation models gain robustness against poor lighting conditions by virtue of complementary information from visible (RGB) and thermal images. Despite its importance, most existing RGB-T semantic segmentation models perform primitive fusion strategies, such as concatenation, element-wise summation and weighted summation, to fuse features from different modalities. These strategies, unfortunately, overlook the modality differences due to different imaging mechanisms, so that they suffer from the reduced discriminability of the fused features. To address such an issue, we propose, for the first time, the strategy of bridging-then-fusing, where the innovation lies in a novel Adaptive-weighted Bi-directional Modality Difference Reduction Network (ABMDRNet). Concretely, a Modality Difference Reduction and Fusion (MDRF) subnetwork is designed, which first employs a bi-directional image-to-image translation based method to reduce the modality differences between RGB features and thermal features, and then adaptively selects those discriminative multi-modality features for RGB-T semantic segmentation in a channel-wise weighted fusion way. Furthermore, considering the importance of contextual information in semantic segmentation, a Multi-Scale Spatial Context (MSC) module and a Multi-Scale Channel Context (MCC) module are proposed to exploit the interactions among multi-scale contextual information of cross-modality features together with their long-range dependencies along spatial and channel dimensions, respectively. Comprehensive experiments on MFNet dataset demonstrate that our method achieves new state-of-the-art results. | ['Jungong Han', 'Nianchang Huang', 'Dingwen Zhang', 'Yongjiang Luo', 'Shenlu Zhao', 'Qiang Zhang'] | 2021-06-19 | null | null | null | cvpr-2021-1 | ['thermal-image-segmentation'] | ['computer-vision'] | [ 7.22345293e-01 -2.50878155e-01 -1.99831072e-02 -4.74400431e-01
-8.17377329e-01 -3.50220889e-01 6.42557025e-01 7.89698493e-03
-5.32544374e-01 4.41542953e-01 1.46311708e-02 2.70624757e-02
-4.08694983e-01 -7.56365955e-01 -3.64829391e-01 -1.15294528e+00
3.56010228e-01 -2.11257756e-01 4.72169727e-01 -4.06141311e-01
1.42797559e-01 3.93010646e-01 -1.83885682e+00 3.73608351e-01
1.14308572e+00 1.43819237e+00 3.62677962e-01 2.33409002e-01
-3.60473365e-01 2.73919374e-01 -1.59778878e-01 -1.52658388e-01
2.81134188e-01 -5.38977563e-01 -6.03092968e-01 1.85833424e-01
1.93317011e-01 9.14970413e-02 -1.37604177e-01 1.06822586e+00
5.04713595e-01 3.24685782e-01 3.40340018e-01 -1.25430405e+00
-3.96646738e-01 1.98325515e-01 -7.51549363e-01 1.33493887e-02
2.33822286e-01 2.83821195e-01 8.68766248e-01 -7.26629376e-01
4.38406229e-01 9.85532999e-01 4.66152728e-01 3.45750809e-01
-1.18106246e+00 -4.76302952e-01 3.25050980e-01 2.70099491e-01
-1.21551132e+00 -1.15742303e-01 1.19321287e+00 -2.15377063e-02
7.35552549e-01 4.96915549e-01 8.75122786e-01 9.45835650e-01
1.63770393e-01 6.92389846e-01 1.65725350e+00 -3.99788409e-01
1.28371209e-01 -7.90258646e-02 -2.06141338e-01 6.06556714e-01
-2.45376021e-01 1.86927151e-02 -8.15122783e-01 2.40080327e-01
6.55840397e-01 2.01157436e-01 -2.94108897e-01 -2.59696245e-01
-1.36150038e+00 6.02706313e-01 7.29298890e-01 7.92328954e-01
-3.31387728e-01 -5.83207002e-03 1.61525324e-01 7.87290856e-02
4.25324202e-01 8.50548744e-02 -5.97562075e-01 1.09955966e-01
-8.33580375e-01 -2.31485114e-01 1.84774026e-01 6.17239356e-01
1.17206490e+00 -1.58424735e-01 -2.48623550e-01 8.56133938e-01
4.38134551e-01 6.50118113e-01 5.57068706e-01 -7.32129276e-01
4.67041284e-01 7.34523416e-01 -4.67551708e-01 -9.50937092e-01
-6.15076363e-01 -3.55091900e-01 -9.17013466e-01 -1.52210947e-02
1.84379190e-01 3.18003505e-01 -1.19570410e+00 1.77621233e+00
5.99635243e-01 3.30246806e-01 -5.04724421e-02 1.17067420e+00
8.81677449e-01 2.49995694e-01 3.38100672e-01 -3.32341939e-01
1.43895304e+00 -8.40427339e-01 -4.39671010e-01 -1.18101977e-01
2.95806795e-01 -8.15171301e-01 9.90564466e-01 5.10845073e-02
-8.50323319e-01 -5.71377754e-01 -9.65745032e-01 -2.03846484e-01
-7.21470237e-01 -2.20026448e-01 8.48482609e-01 7.48836994e-01
-8.64711642e-01 3.64415109e-01 -5.18288970e-01 -2.96848595e-01
2.78110296e-01 3.01445365e-01 -3.82147282e-01 -1.03083722e-01
-1.33953941e+00 5.70483565e-01 5.04517138e-01 4.41317201e-01
-2.73558199e-01 -5.66811383e-01 -8.26108992e-01 -2.19340473e-01
6.11788452e-01 -7.41955042e-01 5.49083292e-01 -1.11667871e+00
-1.56559181e+00 5.81323624e-01 -1.03791162e-01 8.89859200e-02
2.69520342e-01 2.17422426e-01 -6.03554547e-01 6.03531778e-01
5.53689562e-02 7.31084108e-01 1.11925268e+00 -1.43640196e+00
-8.16665232e-01 -7.28941083e-01 3.28769945e-02 5.61650813e-01
-5.30657113e-01 -2.79040247e-01 -6.73704267e-01 -6.40999556e-01
6.61081433e-01 -7.68044353e-01 -2.21135408e-01 -1.38100922e-01
-3.88994634e-01 4.47572321e-02 8.09420824e-01 -5.60729802e-01
1.05002034e+00 -2.15932822e+00 3.86520892e-01 4.75019187e-01
1.35113627e-01 1.13333486e-01 -2.08893567e-01 6.12850264e-02
1.00009240e-01 -1.21809602e-01 -5.85465431e-01 -3.50066870e-01
-9.17169899e-02 3.03292662e-01 9.56422985e-02 4.98062551e-01
1.92513004e-01 1.00092900e+00 -9.35131967e-01 -7.77660847e-01
8.07557881e-01 7.48942554e-01 -1.96651593e-01 -1.95158198e-01
-6.52403682e-02 1.01304150e+00 -8.33261788e-01 9.31552410e-01
8.47729385e-01 7.49114975e-02 2.74845473e-02 -6.53671682e-01
-1.87949508e-01 -2.05140039e-01 -9.92771268e-01 2.31445312e+00
-6.36316359e-01 -6.27024770e-02 2.22256511e-01 -1.02504659e+00
7.37116575e-01 1.35239378e-01 9.32612181e-01 -1.29723275e+00
2.85807401e-01 5.19305408e-01 -3.52298826e-01 -4.03799146e-01
4.90242988e-01 -3.20155233e-01 -1.89947575e-01 8.75156745e-02
8.85786023e-03 -3.45010251e-01 -2.72350051e-02 -3.23163420e-02
6.39201045e-01 3.58814538e-01 -1.09315850e-01 7.92441070e-02
9.18211460e-01 -2.98901796e-01 5.36592960e-01 4.94651854e-01
-2.94714540e-01 8.12096655e-01 1.31128564e-01 -9.20200571e-02
-5.30303836e-01 -9.92709816e-01 -1.66493416e-01 1.04716635e+00
6.81169212e-01 -1.15939550e-01 -6.19707525e-01 -6.89919114e-01
-7.28761703e-02 4.27016705e-01 -6.15851104e-01 -3.36052388e-01
-4.83734339e-01 -9.01374221e-01 3.19651693e-01 4.41554159e-01
8.53571534e-01 -8.67483735e-01 -8.71960580e-01 2.12333053e-01
-2.42285177e-01 -1.30836272e+00 -3.27419460e-01 4.22566384e-01
-9.28874135e-01 -9.18567896e-01 -5.68577170e-01 -3.34896535e-01
4.23068672e-01 5.16668022e-01 5.70232391e-01 1.07316598e-01
-3.87017131e-01 5.21228254e-01 -5.50487876e-01 1.97012305e-01
1.09135866e-01 9.12995487e-02 -3.29171330e-01 4.37939346e-01
1.25929400e-01 -6.20072782e-01 -8.61294985e-01 2.60616601e-01
-1.34196627e+00 2.39150375e-01 7.78965414e-01 9.66774344e-01
8.34874153e-01 7.13631734e-02 3.67483526e-01 -4.30247486e-01
1.31048605e-01 -2.09615871e-01 -2.10857809e-01 5.67033648e-01
-5.55921972e-01 -9.54810157e-02 4.68965143e-01 -2.21178845e-01
-1.42624044e+00 2.09174857e-01 -1.49561569e-01 -3.51478189e-01
-2.67133027e-01 3.24348390e-01 -5.13208628e-01 -4.20045644e-01
7.63686523e-02 4.85172927e-01 -1.12038411e-01 -2.81431466e-01
7.96206415e-01 4.26210403e-01 4.65209663e-01 -6.33881032e-01
7.36724615e-01 8.95899355e-01 2.25925416e-01 -6.51031017e-01
-5.10361016e-01 -5.25460005e-01 -7.71857023e-01 -4.37209457e-01
1.24078846e+00 -7.97326624e-01 -5.47280967e-01 7.52846539e-01
-7.21590817e-01 5.56619018e-02 -3.08264375e-01 5.19664407e-01
-4.68723536e-01 4.14570719e-01 -4.69690919e-01 -5.91720343e-01
-1.27706677e-01 -1.45346355e+00 1.42621839e+00 5.49143612e-01
4.92697984e-01 -8.24047506e-01 -4.54623014e-01 7.22305119e-01
5.65155447e-01 3.85600954e-01 8.65272462e-01 -2.45118946e-01
-5.70512295e-01 5.07361889e-02 -6.53460622e-01 3.96735996e-01
2.47752994e-01 -3.15678954e-01 -1.12400639e+00 3.47929262e-02
-6.54776953e-03 -1.02590315e-01 1.10182512e+00 3.88063073e-01
1.27760994e+00 4.46556568e-01 -1.87490970e-01 8.61845970e-01
1.62456596e+00 2.27504857e-02 5.80483556e-01 3.35577637e-01
1.23727763e+00 5.37001133e-01 5.55869102e-01 3.04476321e-01
5.60013413e-01 7.97075927e-01 5.23936808e-01 -5.23596406e-01
-3.46022576e-01 1.23258606e-01 1.65786520e-02 7.96259940e-01
-2.94899017e-01 5.45071140e-02 -4.45164233e-01 1.97754458e-01
-1.82643306e+00 -5.07487476e-01 -1.95299268e-01 2.05609322e+00
6.61030889e-01 -1.06257170e-01 -4.98968959e-02 1.92204446e-01
6.46407604e-01 3.89613926e-01 -6.47766411e-01 -1.43091515e-01
-6.08269751e-01 3.08494776e-01 6.92995369e-01 1.59935877e-01
-1.01376426e+00 8.52466345e-01 4.85507441e+00 1.40967071e+00
-1.36767435e+00 5.55772722e-01 7.33557820e-01 -1.16179483e-02
-6.64278746e-01 1.04168899e-01 -1.72503173e-01 5.28064787e-01
3.10872555e-01 5.82338214e-01 5.20703137e-01 2.36219242e-01
2.50590183e-02 -4.34790403e-01 -5.10939658e-01 9.04797614e-01
1.20425552e-01 -7.84089386e-01 -2.05232263e-01 -4.42205444e-02
7.24436641e-01 -1.29806489e-01 1.39351100e-01 -2.35489793e-02
-3.04671347e-01 -6.72830641e-01 8.66659343e-01 6.75462246e-01
7.22822428e-01 -7.08621979e-01 5.49711287e-01 -1.39099464e-01
-1.61764956e+00 -2.52932340e-01 6.08853325e-02 3.96335065e-01
2.97784507e-01 8.79649460e-01 3.09012216e-02 1.24090612e+00
7.94340670e-01 7.13845730e-01 -7.11136878e-01 6.02107644e-01
-2.69318521e-01 1.37160406e-01 -3.89194995e-01 2.81684667e-01
3.84952933e-01 -4.03747678e-01 4.33018446e-01 1.06010699e+00
3.96966159e-01 5.43402843e-02 5.29436953e-02 7.49007463e-01
3.76481712e-01 -2.92219110e-02 -1.51062369e-01 2.77984887e-01
2.80437142e-01 1.59374416e+00 -1.28999758e+00 -1.83005571e-01
-5.45851767e-01 1.12023604e+00 -1.62106395e-01 6.14655137e-01
-8.81154180e-01 -1.05648346e-01 3.72052759e-01 -2.60389656e-01
4.78956461e-01 -2.76764393e-01 -6.45187080e-01 -1.07773662e+00
2.02515721e-01 -4.86879468e-01 3.73520613e-01 -7.30338991e-01
-1.18052602e+00 5.05196810e-01 1.06006101e-01 -1.30448377e+00
3.30721885e-01 -3.41176569e-01 -2.50243843e-01 8.44384789e-01
-2.01302409e+00 -1.72738993e+00 -3.63158852e-01 1.08052349e+00
3.30669492e-01 1.61565349e-01 3.76473844e-01 4.53248531e-01
-5.86336732e-01 4.30530369e-01 2.27838801e-03 -4.18496877e-01
5.45568645e-01 -1.09410870e+00 -2.74856150e-01 8.03059042e-01
-1.69416845e-01 4.97083068e-01 2.95597941e-01 -4.43208486e-01
-1.65029657e+00 -8.90181124e-01 3.57588053e-01 -5.48089035e-02
5.00423074e-01 -1.48923710e-01 -7.53129840e-01 -4.30696718e-02
-3.96498479e-02 3.33315790e-01 4.62315679e-01 -2.43030697e-01
-3.88823748e-01 -3.22409272e-01 -1.28091788e+00 3.45538169e-01
1.19674802e+00 -7.25747466e-01 -2.99699277e-01 8.29992369e-02
8.66992116e-01 -3.30378383e-01 -9.75763857e-01 6.97416306e-01
6.95711255e-01 -1.30877817e+00 1.25420463e+00 1.72286332e-01
3.21847618e-01 -6.54178977e-01 -4.17035997e-01 -1.14799464e+00
-5.53967291e-03 -2.13873208e-01 2.18430296e-01 1.31129766e+00
1.38352796e-01 -8.23012710e-01 3.23857009e-01 4.77560252e-01
-3.59419882e-01 -7.62043834e-01 -1.49030125e+00 -5.89997709e-01
-2.79809803e-01 -6.31788731e-01 8.42435777e-01 9.16094005e-01
-2.90984750e-01 -7.36080867e-04 -2.51463801e-01 5.28974161e-02
5.28991878e-01 4.13481832e-01 1.65392995e-01 -9.78090346e-01
-1.82027638e-01 -7.09403336e-01 -3.22663605e-01 -6.85921967e-01
-3.58123668e-02 -8.37670505e-01 2.05999181e-01 -1.41894245e+00
8.83730575e-02 -6.31806612e-01 -8.45010340e-01 4.62201059e-01
-3.58615637e-01 6.49168968e-01 4.12523448e-01 5.73279187e-02
-7.90299177e-01 7.75886178e-01 1.64001775e+00 -7.46265203e-02
-1.51903242e-01 -3.91408533e-01 -6.21416569e-01 5.56246579e-01
4.40162241e-01 -2.19450921e-01 -3.66068661e-01 -3.81483465e-01
1.14019513e-01 2.77597345e-02 6.14725769e-01 -9.51839626e-01
1.58833697e-01 -2.53172487e-01 3.95879418e-01 -4.15463954e-01
5.44009566e-01 -1.08589351e+00 5.35075320e-03 2.41847187e-01
-6.91924542e-02 -3.90867651e-01 -2.89359409e-02 6.89557612e-01
-4.23059195e-01 1.77230835e-01 5.97773552e-01 -8.71053562e-02
-1.12833047e+00 3.19037199e-01 -1.36205524e-01 -3.65674943e-01
8.88832867e-01 -6.20772183e-01 -2.85242528e-01 1.44241363e-01
-6.04935646e-01 3.36570702e-02 4.52540398e-01 5.13136625e-01
5.21884561e-01 -1.35919917e+00 -2.25665972e-01 3.38943005e-01
3.70862097e-01 -6.91468222e-03 8.99610877e-01 1.51326978e+00
4.82079908e-02 1.86897546e-01 -2.32226625e-01 -7.89655626e-01
-9.62455153e-01 2.48777747e-01 2.38539189e-01 -2.57311136e-01
-4.57283288e-01 9.42031980e-01 1.96565643e-01 -3.61199051e-01
-1.72835216e-01 -4.36673313e-01 -2.14935154e-01 2.75562048e-01
1.40176654e-01 1.91312641e-01 4.38742757e-01 -1.19362676e+00
-5.79564810e-01 1.09794021e+00 2.99444705e-01 -1.68947801e-01
1.17230666e+00 -5.97650230e-01 -2.55513400e-01 2.56454676e-01
1.28069961e+00 -3.20617378e-01 -1.36142051e+00 -3.96072894e-01
-3.94706011e-01 -5.56280613e-01 3.84352535e-01 -8.47349942e-01
-1.54197907e+00 8.02625477e-01 9.30390418e-01 1.86915979e-01
1.92296445e+00 1.96436830e-02 1.12397981e+00 -3.72914612e-01
4.56561536e-01 -1.44408250e+00 1.39974535e-01 2.22968325e-01
4.41586852e-01 -1.26678658e+00 4.31413427e-02 -7.16378331e-01
-6.29872143e-01 1.06335175e+00 4.32244867e-01 3.85141581e-01
5.38090229e-01 -8.93386230e-02 -2.50628311e-02 -2.12611631e-01
-1.16941176e-01 -6.62814677e-01 4.54879135e-01 5.97380579e-01
1.55011892e-01 2.32809544e-01 -5.32983065e-01 5.02371311e-01
3.02846581e-01 -1.59010872e-01 -2.36423641e-01 1.07787263e+00
-2.17506647e-01 -1.16180921e+00 -5.24824619e-01 2.25866154e-01
-2.14645937e-02 -1.46506146e-01 -6.17362373e-02 5.94378293e-01
8.52501690e-01 1.10341084e+00 -2.12015316e-01 -7.71246076e-01
1.93761095e-01 -9.46588069e-02 5.61980486e-01 -1.35161996e-01
-8.33047152e-01 4.77058560e-01 -3.52185220e-01 -8.98239315e-01
-1.10535371e+00 -5.94004750e-01 -1.39755416e+00 1.58039764e-01
-4.08944517e-01 -4.30203438e-01 1.04127908e+00 1.35120082e+00
2.25479171e-01 8.08987975e-01 8.51719141e-01 -1.17828345e+00
2.28332058e-02 -6.58337176e-01 -7.06031144e-01 6.30307615e-01
2.10991412e-01 -8.92080724e-01 -3.15845579e-01 -2.15986326e-01] | [9.493329048156738, -1.0513101816177368] |
09f3cd63-6b51-40ad-b666-a1bc0f96aca4 | a-closer-look-at-few-shot-out-of-distribution | null | null | https://aclanthology.org/2022.coling-1.36 | https://aclanthology.org/2022.coling-1.36.pdf | A Closer Look at Few-Shot Out-of-Distribution Intent Detection | We consider few-shot out-of-distribution (OOD) intent detection, a practical and important problem for the development of task-oriented dialogue systems. Despite its importance, this problem is seldom studied in the literature, let alone examined in a systematic way. In this work, we take a closer look at this problem and identify key issues for research. In our pilot study, we reveal the reason why existing OOD intent detection methods are not adequate in dealing with this problem. Based on the observation, we propose a promising approach to tackle this problem based on latent representation generation and self-supervision. Comprehensive experiments on three real-world intent detection benchmark datasets demonstrate the high effectiveness of our proposed approach and its great potential in improving state-of-the-art methods for few-shot OOD intent detection. | ['Albert Y.S. Lam', 'Xiao-Ming Wu', 'Lu Fan', 'Haowen Liang', 'Li-Ming Zhan'] | null | null | null | null | coling-2022-10 | ['intent-detection', 'task-oriented-dialogue-systems'] | ['natural-language-processing', 'natural-language-processing'] | [ 3.00167799e-01 2.24177524e-01 -3.31956774e-01 -2.67797858e-01
-5.19335926e-01 -2.28701994e-01 1.08616996e+00 1.94033206e-01
-2.12154061e-01 4.11119223e-01 6.34446919e-01 -1.58962429e-01
2.83597596e-02 -3.82804245e-01 2.55906194e-01 -4.36209559e-01
4.48433198e-02 4.36083347e-01 2.80619591e-01 -4.98203397e-01
7.66934454e-01 -1.02214282e-02 -1.71988785e+00 1.47826657e-01
9.98748422e-01 4.05555069e-01 1.66730821e-01 4.25550193e-01
-2.73989201e-01 1.18896520e+00 -9.83582497e-01 -3.57361674e-01
-8.64057243e-02 -5.78624427e-01 -9.57843781e-01 3.33741844e-01
1.10537462e-01 -4.78495240e-01 -2.76187867e-01 1.00661933e+00
6.51599467e-01 3.65556389e-01 9.57379162e-01 -1.28955126e+00
-5.13487458e-01 3.09586138e-01 -4.46244359e-01 5.08258402e-01
8.06034684e-01 4.37368043e-02 1.23944938e+00 -7.82742679e-01
6.72473311e-01 1.25955915e+00 3.13751221e-01 6.69270217e-01
-1.06108773e+00 -2.02094436e-01 6.12137020e-02 5.64745180e-02
-9.66980278e-01 -6.90770268e-01 9.08529043e-01 -4.63456303e-01
1.20410347e+00 2.43633553e-01 4.34829116e-01 1.20165646e+00
2.29107857e-01 1.26336229e+00 1.08994472e+00 -9.13718641e-01
3.60554814e-01 3.72656494e-01 7.35139787e-01 6.99409783e-01
1.61545798e-01 -3.20721388e-01 -7.92599022e-01 -3.67479503e-01
4.06884313e-01 -2.21993565e-03 -2.23921180e-01 -2.31155097e-01
-9.79380727e-01 1.15177405e+00 -2.05211386e-01 7.93157756e-01
-3.10036987e-01 -4.90057200e-01 7.27658093e-01 3.80724818e-01
9.35796857e-01 5.25437236e-01 1.77108377e-01 -7.08708763e-01
-8.62149954e-01 3.19228053e-01 1.14646971e+00 8.66732419e-01
3.63580912e-01 5.17492592e-02 -4.50854778e-01 9.72018182e-01
4.13331002e-01 -1.60557225e-01 6.94847047e-01 -5.06366670e-01
3.50507289e-01 9.14465725e-01 2.35333323e-01 -1.02628326e+00
-3.06808054e-01 1.68057621e-01 -5.15461147e-01 1.17520103e-02
1.93454191e-01 -1.01896897e-01 -6.69926584e-01 1.18996859e+00
2.62314826e-01 -1.70200884e-01 2.08218113e-01 7.68941581e-01
9.65892792e-01 6.99669659e-01 -6.26795227e-03 -5.25316894e-01
1.70158339e+00 -1.20624769e+00 -1.26028252e+00 -3.36701721e-01
8.24017167e-01 -8.71923089e-01 1.19334555e+00 2.49854416e-01
-5.79330564e-01 -3.42310607e-01 -9.03140008e-01 9.52964090e-03
-2.69981235e-01 1.59387559e-01 9.98239219e-01 8.56187761e-01
-6.53608680e-01 8.83445069e-02 -5.83595395e-01 -9.43554342e-01
-2.01337174e-01 -5.49506806e-02 -5.50846681e-02 8.37447047e-02
-1.27219009e+00 8.57196033e-01 3.13263208e-01 -2.14466095e-01
-7.82398462e-01 -2.73557156e-01 -9.74146128e-01 -1.43733382e-01
7.01199532e-01 -2.85548300e-01 1.57463050e+00 -4.32608396e-01
-1.69086933e+00 8.37020040e-01 -2.98175365e-01 -4.95810688e-01
2.87880182e-01 -3.72083575e-01 -3.13680530e-01 6.18483648e-02
3.18386436e-01 2.97227800e-01 9.52460468e-01 -9.42738414e-01
-6.78135157e-01 -2.45329678e-01 3.50990564e-01 5.53639054e-01
-7.45850444e-01 4.25264895e-01 -3.77824605e-01 -7.14790761e-01
-1.74320385e-01 -7.88164198e-01 -1.59009069e-01 -5.67404151e-01
-4.39961016e-01 -8.88495684e-01 9.76885974e-01 -2.59240597e-01
1.54371262e+00 -1.99885905e+00 -8.69681835e-02 -4.80565727e-01
2.88426250e-01 4.82577056e-01 2.83615619e-01 1.03284121e+00
2.83723027e-01 -1.65751427e-02 -8.49882290e-02 -5.58705211e-01
3.27879846e-01 5.57295121e-02 -4.61190850e-01 3.80155146e-01
-2.56855716e-03 6.80104375e-01 -1.10320485e+00 -6.36300027e-01
4.89417255e-01 2.85134435e-01 -2.47825801e-01 7.43679464e-01
-2.64317751e-01 -3.99529561e-02 -6.97397590e-01 5.50451040e-01
3.19496691e-01 -1.14732601e-01 1.68760702e-01 -1.69009920e-02
-4.14570391e-01 6.39848888e-01 -9.61921096e-01 1.45968521e+00
-3.36507022e-01 7.81313419e-01 -2.59967029e-01 -1.02106559e+00
1.07219732e+00 5.54777622e-01 2.88805962e-01 -5.42552352e-01
1.53188527e-01 1.00076068e-02 -8.19194689e-02 -5.65583110e-01
1.08978784e+00 -3.09535921e-01 -4.71624076e-01 8.27692568e-01
2.85339117e-01 2.00636685e-03 5.52621126e-01 2.69582450e-01
9.70142722e-01 -2.10596576e-01 8.71330380e-01 -2.09863812e-01
6.34763479e-01 1.92538679e-01 1.38322592e-01 9.73636150e-01
-5.96009672e-01 3.22495073e-01 6.46288455e-01 -3.66822898e-01
-6.74519658e-01 -3.50083798e-01 -3.61866802e-02 1.35164225e+00
1.31712593e-02 -8.07035863e-01 -8.66915226e-01 -9.46853995e-01
-3.20129663e-01 9.95006382e-01 -6.15951717e-01 -7.52820820e-02
-2.90762872e-01 -8.58910561e-01 6.44705653e-01 7.14484006e-02
5.67648053e-01 -1.27727318e+00 -8.74363661e-01 3.63747239e-01
-4.65772539e-01 -1.03595829e+00 -2.76517957e-01 1.53134704e-01
-6.67104423e-01 -7.62304068e-01 -7.99176574e-01 -7.41805613e-01
4.92280155e-01 9.36797559e-01 1.02442777e+00 1.10947557e-01
-4.29820120e-01 4.83427256e-01 -7.60301530e-01 -4.39450353e-01
-7.60573506e-01 2.10538164e-01 1.34678751e-01 -4.49307598e-02
8.65105569e-01 -1.01433747e-01 -1.85988203e-01 3.24702650e-01
-9.01190579e-01 -1.05828188e-01 3.06207240e-01 9.27414954e-01
-2.08660379e-01 2.96990097e-01 6.25505328e-01 -1.10174751e+00
1.45021248e+00 -3.90894026e-01 -3.17295820e-01 1.47599742e-01
-6.09416425e-01 1.47541435e-02 3.91010940e-01 -1.94490001e-01
-1.45999098e+00 -3.50556284e-01 -1.68136775e-01 -1.20321155e-01
-4.98168260e-01 5.20121932e-01 1.44063115e-01 1.51153281e-01
4.78229135e-01 3.96108717e-01 -9.69533026e-02 -5.69181204e-01
2.23465055e-01 1.20817149e+00 -2.13797987e-02 -1.69115767e-01
4.11036074e-01 4.08768803e-01 -4.81843442e-01 -1.40493429e+00
-9.65965331e-01 -1.16777015e+00 -6.10356271e-01 -2.45197415e-01
7.75306582e-01 -8.63893926e-01 -3.83583426e-01 3.58485043e-01
-1.13059676e+00 -5.68798706e-02 -1.85221776e-01 2.61162579e-01
-5.65051436e-01 8.27844799e-01 -6.29685402e-01 -1.45337355e+00
-5.68873107e-01 -9.40339506e-01 1.19566917e+00 1.80427119e-01
-6.98950231e-01 -1.19519150e+00 4.29605097e-01 7.23511040e-01
3.13927323e-01 -3.80379975e-01 7.35850573e-01 -1.05819726e+00
-1.29059970e-01 -3.33647400e-01 -4.71248589e-02 1.90681089e-02
5.16655207e-01 -2.36990720e-01 -9.38385367e-01 -1.26910955e-01
5.46321273e-01 -5.37430108e-01 7.55040884e-01 -6.25929162e-02
2.31008172e-01 -2.62384921e-01 -1.81686580e-01 -2.33914718e-01
1.07760298e+00 1.40790820e-01 4.30745304e-01 3.65406990e-01
3.62720758e-01 7.95914531e-01 1.30347717e+00 8.83547723e-01
3.72601748e-01 1.06977534e+00 8.93882364e-02 1.09474368e-01
-8.13302249e-02 -2.61390507e-01 5.25021911e-01 9.38100278e-01
-2.24609338e-02 -5.54512858e-01 -9.40124989e-01 5.23173630e-01
-2.02794123e+00 -1.06517613e+00 -1.28742054e-01 1.95428574e+00
6.90390050e-01 2.61547625e-01 2.67519444e-01 2.09560558e-01
6.52029812e-01 6.37445152e-01 7.75828958e-02 -5.72015762e-01
3.12449098e-01 -3.96109134e-01 -2.18413562e-01 3.74264985e-01
-1.30887163e+00 1.15757966e+00 6.45321655e+00 9.77130949e-01
-7.39303648e-01 1.44418225e-01 2.87132055e-01 4.60538656e-01
-1.74841791e-01 -2.89618000e-02 -1.14898658e+00 4.08989280e-01
7.59075522e-01 -2.29320690e-01 2.38808431e-03 1.04437435e+00
1.30521491e-01 -4.09492165e-01 -9.23784673e-01 7.25431323e-01
7.08363652e-01 -8.83377910e-01 -4.51838039e-03 4.36048582e-02
5.92706501e-01 -4.46525574e-01 -2.01703072e-01 6.22486293e-01
1.57195747e-01 -5.69513857e-01 2.52127945e-01 1.51237696e-01
2.05251470e-01 -5.23920417e-01 1.01825500e+00 8.12243938e-01
-9.17322695e-01 1.81569010e-01 -4.73652095e-01 -4.57168758e-01
3.15669745e-01 4.40265864e-01 -1.22886467e+00 3.92284036e-01
3.60655874e-01 6.45270169e-01 -2.93328404e-01 8.76648903e-01
-2.73380786e-01 7.41640747e-01 -4.00583667e-04 -5.35669208e-01
3.89540881e-01 -1.20316751e-01 7.90887892e-01 1.32450473e+00
-3.21122794e-03 -2.15549953e-02 2.38255307e-01 7.07611620e-01
1.83233932e-01 3.39488924e-01 -1.16247666e+00 -3.58609945e-01
9.95324776e-02 1.22265995e+00 -9.72628534e-01 -4.34726179e-01
-5.38425624e-01 1.18346667e+00 3.03219587e-01 -9.46200192e-02
-5.99770010e-01 -4.53102618e-01 4.51177835e-01 -2.48729065e-01
2.04223081e-01 -2.08672032e-01 2.88808830e-02 -1.39500046e+00
-2.19156489e-01 -1.06089747e+00 4.67616677e-01 -2.65698642e-01
-1.45580828e+00 4.89994079e-01 2.37517521e-01 -1.43144536e+00
-5.86752713e-01 -2.79541701e-01 -8.41221035e-01 3.85275692e-01
-1.39031732e+00 -9.63870883e-01 -1.80014104e-01 2.31220335e-01
1.33182859e+00 -2.57160395e-01 1.12671828e+00 8.07619169e-02
-6.75285280e-01 3.17300141e-01 -4.66470122e-02 -9.95597169e-02
7.51591444e-01 -1.27312386e+00 6.49291456e-01 7.62620986e-01
2.41191834e-01 8.89750898e-01 1.01806045e+00 -8.30937028e-01
-1.52573931e+00 -6.47581339e-01 1.32172573e+00 -4.22313333e-01
7.61585474e-01 -5.06369114e-01 -8.30566049e-01 3.45583916e-01
5.26184976e-01 -5.03163815e-01 9.48891759e-01 3.33554178e-01
-7.59859011e-02 5.74871719e-01 -9.74880338e-01 6.61476493e-01
7.88911581e-01 -4.89244044e-01 -1.39202487e+00 4.51655447e-01
6.12776220e-01 -1.58803314e-01 -6.11448348e-01 2.04166949e-01
2.94546217e-01 -1.23127329e+00 7.51590252e-01 -3.31474841e-01
2.47568190e-01 7.79666752e-02 2.98381262e-02 -1.09901619e+00
-6.40303865e-02 -8.96299720e-01 -3.40407670e-01 1.53663576e+00
-1.30534619e-01 -3.50672483e-01 8.28978181e-01 5.17829239e-01
-1.75799906e-01 -4.68034863e-01 -8.71729374e-01 -6.47538543e-01
-4.09705043e-01 -2.15564519e-01 3.41879949e-02 9.00585771e-01
7.01088071e-01 9.10154045e-01 -8.76817167e-01 -2.42143109e-01
6.28920376e-01 3.42309892e-01 1.10919523e+00 -1.12764633e+00
-6.53345138e-02 -1.57049149e-01 -3.38172913e-01 -1.29887915e+00
1.60534039e-01 -3.50407690e-01 3.04379374e-01 -1.68631327e+00
3.75740528e-01 -1.51722673e-02 9.23986062e-02 2.85545170e-01
-4.04625624e-01 1.92146808e-01 2.22283244e-01 3.81305009e-01
-1.05782068e+00 7.76556313e-01 8.02375734e-01 -4.23819721e-02
-3.56188506e-01 1.35536507e-01 -6.01162970e-01 8.89500499e-01
7.76116848e-01 -5.71011305e-01 -6.01398528e-01 2.36152560e-01
-2.06506342e-01 1.10488869e-01 -9.13962573e-02 -9.10800636e-01
1.36276051e-01 -8.96550715e-02 -5.26026785e-01 -7.45313823e-01
3.92295688e-01 -6.10246062e-01 -5.57017148e-01 5.31210005e-01
-3.32487881e-01 -3.13048422e-01 -1.37128055e-01 6.12170339e-01
-2.94364393e-01 -8.76632988e-01 3.89818639e-01 -2.16172233e-01
-1.14154696e+00 -2.84315050e-01 -1.06804240e+00 2.73238987e-01
1.03445113e+00 -2.25717053e-01 -3.86125863e-01 -5.90898275e-01
-8.80922824e-02 3.32989581e-02 3.56085479e-01 7.07283974e-01
6.53499186e-01 -8.82791877e-01 -4.72304732e-01 2.46912494e-01
6.26234174e-01 -6.19447827e-01 2.80963153e-01 6.53339922e-01
-1.59123600e-01 8.33520710e-01 -1.21991657e-01 -4.47434127e-01
-1.51482320e+00 4.62274641e-01 -2.33502150e-01 -6.83249891e-01
-7.31767714e-01 3.92655909e-01 1.79458648e-01 -4.44986790e-01
6.64205730e-01 1.93417445e-01 -7.57166564e-01 3.96765143e-01
7.09651947e-01 4.04413491e-01 -1.34901986e-01 -7.18379617e-01
-2.49215007e-01 1.18078038e-01 -4.39673364e-01 -1.09618224e-01
1.13222086e+00 -3.95725787e-01 -5.61454259e-02 1.02879417e+00
9.53041315e-01 -7.78930634e-02 -7.35388815e-01 -1.58627465e-01
2.37464413e-01 -6.63416147e-01 -1.33858174e-02 -4.40087557e-01
-2.13477001e-01 9.52369809e-01 4.20658767e-01 9.99646366e-01
6.65942311e-01 2.39018984e-02 6.90422893e-01 6.68665171e-01
4.71840680e-01 -1.10058975e+00 5.83353281e-01 8.49889398e-01
6.14879191e-01 -1.55737638e+00 1.41219884e-01 -5.05818009e-01
-1.00471294e+00 1.03337014e+00 6.79906785e-01 4.87139374e-02
2.81811804e-01 7.56527856e-03 1.44882128e-01 -4.68271375e-01
-9.58609998e-01 -3.90435278e-01 2.60299116e-01 4.62504297e-01
8.22080851e-01 -2.81541973e-01 -6.04089916e-01 3.48319530e-01
2.33675987e-01 -2.86456253e-02 7.13452816e-01 1.40089202e+00
-7.63790607e-01 -1.23297226e+00 -1.69889629e-01 4.80582505e-01
-5.13435662e-01 -8.89488682e-02 -5.81583917e-01 6.47697687e-01
-4.03011978e-01 1.24516308e+00 -1.80100650e-01 -2.88156807e-01
2.69658923e-01 2.67267227e-01 1.24154799e-02 -1.07801521e+00
-6.22107744e-01 2.15694815e-01 4.88374740e-01 -2.26372272e-01
-6.96813941e-01 -6.82751834e-01 -8.58747721e-01 -4.15691361e-02
-7.32663691e-01 4.07422185e-01 4.69770223e-01 9.67808366e-01
2.26202950e-01 3.13035071e-01 6.85940444e-01 -7.49868512e-01
-8.29515696e-01 -1.28615701e+00 -6.01044834e-01 5.11079848e-01
1.74984992e-01 -8.55622053e-01 -5.60047507e-01 -1.19917043e-01] | [12.417348861694336, 7.549142360687256] |
54722b42-7832-4bc5-afb0-08530ab9e423 | computationally-efficient-heart-rate | null | null | https://doi.org/10.23919/EUSIPCO.2017.8081656 | https://www.researchgate.net/publication/319176582_Computationally_Efficient_Heart_Rate_Estimation_During_Physical_Exercise_Using_Photoplethysmographic_Signals | Computationally efficient heart rate estimation during physical exercise using photoplethysmographic signals | Wearable devices that acquire photoplethysmographic (PPG) signals are becoming increasingly popular to monitor the heart rate during physical exercise. However, high accuracy and low computational complexity are conflicting requirements. We propose a method that provides highly accurate heart rate estimates at a very low computational cost in order to be implementable on wearables. To achieve the lowest possible complexity, only basic signal processing operations, i.e., correlation-based fundamental frequency estimation and spectral combination, harmonic noise damping and frequency domain tracking, are used. The proposed approach outperforms state-of-the-art methods on current benchmark data considerably in terms of computation time, while achieving a similar accuracy. | ['Abdelhak M. Zoubir', 'Tim Schäck', 'Michael Muma'] | 2017-08-28 | null | null | null | 2017-25th-european-signal-processing | ['photoplethysmography-ppg', 'heart-rate-estimation'] | ['medical', 'medical'] | [ 3.51258636e-01 -3.39962751e-01 -8.46719146e-02 -2.04709545e-01
-6.39071822e-01 -3.61063600e-01 -8.72500092e-02 1.44022346e-01
-4.36518133e-01 7.91833997e-01 -2.70336628e-01 -9.61484835e-02
-2.20617969e-02 -3.85644585e-01 9.97706503e-02 -7.40441382e-01
4.89026494e-03 -2.07821041e-01 4.51815827e-03 3.93636674e-01
1.64984152e-01 3.35446000e-01 -1.27621889e+00 -4.66256022e-01
1.10192537e+00 1.39340329e+00 -3.37329477e-01 7.79370904e-01
3.62247586e-01 2.08791360e-01 -6.74376309e-01 -2.12712541e-01
1.47289515e-01 -8.30558896e-01 -4.35689352e-02 -1.87392175e-01
2.90282249e-01 -2.11435318e-01 -7.32698441e-02 9.14492130e-01
1.00072110e+00 5.26932627e-02 2.43761256e-01 -6.35436237e-01
6.54842937e-04 1.48248196e-01 -4.89659131e-01 5.19666493e-01
4.49723244e-01 2.75549501e-01 4.03803706e-01 -7.15468943e-01
8.39081109e-02 3.18588912e-01 1.03500342e+00 3.12440485e-01
-1.61568904e+00 -5.91661811e-01 -7.10232556e-01 2.74379313e-01
-1.55537784e+00 -5.89076757e-01 1.23549259e+00 -1.13860466e-01
5.61267614e-01 7.27569461e-01 9.47974265e-01 6.62084877e-01
4.21288073e-01 -1.33293541e-02 1.58372712e+00 -2.85822004e-01
3.03248644e-01 2.08001330e-01 1.72164783e-01 2.41069123e-01
4.74478662e-01 -4.41813879e-02 -6.87466562e-01 -3.89384419e-01
8.18115115e-01 -2.31958702e-01 -6.80974841e-01 3.25483941e-02
-9.45418060e-01 3.72695208e-01 -2.14065343e-01 4.81483847e-01
-7.49669015e-01 3.02183092e-01 4.82475847e-01 1.88195929e-01
2.39576504e-01 3.90869379e-01 -2.12823108e-01 -8.23548734e-01
-1.01182044e+00 5.36726043e-02 1.03997195e+00 3.66514266e-01
2.28854284e-01 2.95267135e-01 -1.36570007e-01 6.22977853e-01
2.19853535e-01 5.94004273e-01 6.29886985e-01 -9.27106500e-01
7.74508938e-02 3.72811198e-01 3.04088831e-01 -1.12639081e+00
-6.85930908e-01 -5.51839232e-01 -9.01544094e-01 -2.36235008e-01
4.61301506e-01 -1.94991782e-01 -2.78609633e-01 1.28445101e+00
5.35361767e-01 6.50204599e-01 -1.32181287e-01 1.20569885e+00
7.99587369e-01 4.08516109e-01 2.69493103e-01 -8.40925634e-01
1.75928676e+00 -4.49462570e-02 -1.13371789e+00 -8.43589157e-02
2.10321173e-01 -8.91465783e-01 8.74770403e-01 6.51287973e-01
-1.21786785e+00 -7.83401072e-01 -1.13578403e+00 1.89860910e-01
3.03107798e-01 2.78446913e-01 5.83425999e-01 1.41475809e+00
-6.18664563e-01 9.99044776e-01 -8.22974443e-01 -9.85085815e-02
-5.52076884e-02 1.49385765e-01 1.52102588e-02 5.59571207e-01
-1.22540343e+00 7.30281353e-01 3.44149172e-02 3.42771918e-01
-6.81000575e-02 -1.15990353e+00 -7.05136657e-01 1.13145545e-01
1.32463396e-01 -5.27933002e-01 8.87076855e-01 -3.27103436e-01
-2.27627897e+00 6.83714807e-01 -1.71467513e-01 -3.79691243e-01
4.68429804e-01 -3.05436969e-01 -7.49785721e-01 5.49690902e-01
-6.79192066e-01 -3.87611896e-01 1.03793955e+00 -4.64750051e-01
-2.88983453e-02 -3.50588471e-01 -6.87177658e-01 2.91565835e-01
-5.15173495e-01 -1.04922712e-01 -1.32823780e-01 -2.68867612e-01
3.93005610e-01 -8.65870774e-01 3.96082662e-02 2.23411303e-02
1.03518665e-01 1.17766224e-01 4.51705933e-01 -7.74927378e-01
1.54771805e+00 -2.20126486e+00 -1.49344474e-01 3.39510083e-01
2.43661717e-01 6.92961395e-01 5.25436521e-01 2.68593073e-01
1.60528064e-01 -2.20472619e-01 -2.88273022e-02 -1.33723125e-01
-2.47129157e-01 -2.11655453e-01 8.44860300e-02 9.89430010e-01
-1.43021718e-01 7.35970318e-01 -6.35891199e-01 -4.44198728e-01
8.16611707e-01 6.62539542e-01 -8.44229981e-02 3.03615332e-01
5.66651285e-01 6.72538638e-01 -1.52710140e-01 4.49424297e-01
6.30113542e-01 -6.41893223e-02 6.43790960e-01 -7.57180393e-01
-2.09695622e-01 4.05398637e-01 -1.44199014e+00 1.46369731e+00
-4.87123102e-01 5.12861192e-01 1.67054720e-02 -8.03511143e-01
1.43283975e+00 6.96484745e-01 7.10414529e-01 -8.11268628e-01
2.89290905e-01 4.37037319e-01 2.65643150e-01 -7.49159336e-01
8.71809348e-02 -4.66748595e-01 1.81719512e-01 3.19824517e-01
-2.84766465e-01 -3.06381464e-01 -1.60087068e-02 -3.47550213e-01
7.53881872e-01 3.46017932e-03 7.59025037e-01 -6.84018910e-01
8.84574890e-01 -5.39689362e-01 6.19924903e-01 4.12710398e-01
-5.66992700e-01 3.97022963e-01 2.54424423e-01 -5.37245572e-01
-6.51191592e-01 -7.02703953e-01 -3.21561486e-01 1.96532652e-01
1.25156179e-01 -3.14014375e-01 -6.16661727e-01 1.39187217e-01
2.18324438e-02 2.72067159e-01 -2.10461974e-01 -9.21036303e-02
-7.67075181e-01 -7.65064359e-01 7.35809922e-01 3.59654278e-01
4.92019206e-01 -7.40297079e-01 -1.27886069e+00 5.46946526e-01
-1.87542439e-01 -1.16210759e+00 -5.14405072e-01 -2.18405589e-01
-1.27498138e+00 -1.00941920e+00 -8.93755615e-01 -2.65636891e-01
8.02195892e-02 -8.32805634e-02 9.99101877e-01 -1.84727058e-01
-8.42732251e-01 3.00050855e-01 2.57020462e-02 -3.10589075e-01
8.37449059e-02 -1.34638190e-01 1.40844703e-01 3.51304621e-01
4.65282887e-01 -9.61739063e-01 -1.30143678e+00 2.75535196e-01
-2.33155131e-01 -4.21097517e-01 4.72360373e-01 4.93738890e-01
5.70897937e-01 -1.81769148e-01 6.54481232e-01 -5.62024534e-01
9.03917670e-01 1.00411497e-01 -7.36696601e-01 -2.15885550e-01
-7.39369571e-01 -5.11434793e-01 8.14930737e-01 -6.04414999e-01
-9.02926922e-01 9.71583091e-03 -4.26407494e-02 -2.69470245e-01
3.08069903e-02 4.71253216e-01 3.11417848e-01 -4.45886731e-01
9.82179523e-01 6.50886297e-01 3.63478780e-01 -3.95319253e-01
-4.93755713e-02 5.07727027e-01 8.92136395e-01 -3.46095979e-01
4.82871503e-01 2.14484140e-01 5.35230517e-01 -1.23310304e+00
-2.58670390e-01 -6.94666803e-01 -4.22490448e-01 -4.95649904e-01
5.14841616e-01 -7.86141098e-01 -1.35681415e+00 6.25711083e-01
-7.66481757e-01 2.33568281e-01 -2.02513158e-01 1.02343142e+00
-5.13724387e-01 7.98478186e-01 -7.14353025e-01 -1.33173037e+00
-1.02351868e+00 -4.45360810e-01 4.12572473e-01 5.93862057e-01
-4.99160051e-01 -8.19766521e-01 1.31382555e-01 2.31830597e-01
8.19754303e-01 7.83133030e-01 3.01641345e-01 1.15996271e-01
-7.21545815e-02 -4.76893485e-01 5.26786223e-03 3.34907919e-01
3.49168450e-01 -2.40000382e-01 -1.09866738e+00 -3.01116943e-01
6.14123642e-01 5.41648529e-02 9.59862396e-02 6.33764625e-01
9.70266342e-01 -1.11428030e-01 -2.43888199e-02 6.07849061e-01
1.60132980e+00 3.02766532e-01 9.88765061e-01 -3.87778550e-01
2.44278178e-01 3.86868030e-01 4.42983687e-01 7.51780748e-01
-1.67289287e-01 6.03236496e-01 -1.52446240e-01 -2.38380313e-01
-1.92272458e-02 7.50406832e-02 -6.15871651e-03 9.80492651e-01
-4.21130687e-01 2.92941928e-01 -5.55570543e-01 2.28721067e-01
-1.51629078e+00 -7.56313145e-01 -4.46264982e-01 2.64371872e+00
9.40335512e-01 7.96305984e-02 3.34384829e-01 5.94814301e-01
6.59296811e-01 1.58601403e-01 -5.34299493e-01 -5.80867767e-01
3.44659358e-01 7.12368906e-01 4.63377357e-01 3.49340618e-01
-7.82386303e-01 6.89908043e-02 6.58472157e+00 2.72202820e-01
-1.34382725e+00 4.89424588e-03 4.06868964e-01 -3.83398384e-02
2.08209202e-01 -1.76614329e-01 -4.86813605e-01 7.04832077e-01
1.32514501e+00 -4.72230464e-01 1.80134669e-01 7.45339870e-01
5.35650373e-01 -3.41892868e-01 -6.63883328e-01 1.61658072e+00
1.33228913e-01 -1.06698108e+00 -7.73942351e-01 -1.90209091e-01
2.11848870e-01 -7.34147549e-01 -2.34230399e-01 -2.12911084e-01
-1.07420659e+00 -5.90302467e-01 -4.81481329e-02 7.41254032e-01
8.98511350e-01 -7.83492744e-01 6.67203367e-01 7.73537606e-02
-1.26079667e+00 2.82353520e-01 -2.58535266e-01 -3.45855772e-01
3.18045050e-01 1.39048290e+00 -2.62920648e-01 6.13218427e-01
2.29477718e-01 4.26207036e-01 2.03001276e-02 1.28183949e+00
-2.03456447e-01 7.53227532e-01 -5.93418062e-01 -2.25139439e-01
-4.27065164e-01 -3.56813073e-01 5.57614803e-01 8.87821853e-01
4.54484761e-01 6.78068817e-01 5.85049838e-02 9.04945612e-01
2.91267991e-01 3.95892203e-01 -2.02309132e-01 9.13001075e-02
5.34591436e-01 1.38107407e+00 -5.00619650e-01 -3.33291113e-01
-3.41082215e-01 5.58459997e-01 -3.97566676e-01 -7.59215653e-02
-9.04646456e-01 -9.24241543e-01 7.16924131e-01 4.09862071e-01
-3.99098575e-01 -3.52585703e-01 -5.66366673e-01 -9.22830343e-01
4.04436350e-01 -7.69032896e-01 4.17070568e-01 -1.42664075e-01
-9.84103143e-01 4.26126897e-01 -3.10430706e-01 -1.56672978e+00
-3.05610389e-01 -1.76361054e-01 -7.49308467e-01 1.37948322e+00
-1.57811224e+00 -2.13099927e-01 -7.59655476e-01 6.22517765e-01
1.43831581e-01 5.79092503e-01 1.04099619e+00 6.12536252e-01
-4.18174684e-01 5.78629017e-01 -4.96648222e-01 -4.20556635e-01
6.15359366e-01 -1.06196415e+00 1.25123963e-01 7.87253082e-01
-4.86698478e-01 7.43510962e-01 7.55407810e-01 -4.66545314e-01
-1.81294346e+00 -4.37373430e-01 1.07483375e+00 4.91008833e-02
3.51722717e-01 -1.93529069e-01 -1.01053858e+00 -3.87436479e-01
-1.54386684e-01 3.60443860e-01 8.12429249e-01 -7.98972324e-02
2.29396552e-01 -6.29956961e-01 -1.23397219e+00 2.66410530e-01
4.84520018e-01 -5.78285754e-01 -4.82284427e-01 -1.70062613e-02
-9.74768624e-02 -7.57879674e-01 -1.64727771e+00 4.06056464e-01
1.02409649e+00 -8.96997869e-01 7.71582365e-01 2.67343521e-01
-2.79749364e-01 -4.75071639e-01 5.83787560e-01 -8.99801135e-01
-2.70432204e-01 -1.54822636e+00 -3.91772956e-01 1.04455030e+00
2.65138354e-02 -1.13389945e+00 8.13752651e-01 1.08018363e+00
2.79733121e-01 -5.55244565e-01 -1.02949679e+00 -9.55422878e-01
-7.80311584e-01 -7.19017014e-02 2.35504508e-02 7.53449261e-01
6.63027465e-01 1.24543808e-01 -9.80776072e-01 -9.04066116e-02
1.00896049e+00 2.92172879e-01 5.51149905e-01 -1.32413125e+00
-3.64131153e-01 7.04771001e-03 -6.27963841e-01 -6.61805749e-01
-5.70751846e-01 -1.90497428e-01 -8.46821442e-02 -9.48444724e-01
-1.62403718e-01 -1.30285591e-01 -5.18461227e-01 5.80182858e-02
-4.81051385e-01 8.59344184e-01 6.32074836e-04 6.86312020e-02
8.52837339e-02 2.98765957e-01 9.01622236e-01 4.94862616e-01
-8.47831488e-01 1.94096237e-01 -3.54564488e-01 4.93348509e-01
8.44148219e-01 -4.35466498e-01 -3.78709465e-01 3.94866049e-01
-7.56093487e-02 4.79142696e-01 1.35126254e-02 -1.46534324e+00
2.29749009e-01 1.73149928e-01 5.21828175e-01 -2.58715838e-01
4.70100313e-01 -5.84966123e-01 7.56916940e-01 1.00189281e+00
-3.31136137e-02 -1.41463190e-01 1.49847820e-01 2.10705176e-01
-2.00433195e-01 -2.08714101e-02 1.07900369e+00 -2.63799503e-02
-9.57683697e-02 1.11876890e-01 -2.84129918e-01 -1.13713309e-01
8.34178627e-01 -5.23088932e-01 -3.71443033e-01 -2.97210187e-01
-6.53960049e-01 -2.64384866e-01 2.69283932e-02 1.16593897e-01
5.28069913e-01 -1.13825870e+00 -6.70067310e-01 2.98403800e-01
-6.45828545e-02 -7.89661527e-01 7.24617541e-01 1.38909757e+00
-6.11372352e-01 3.81232023e-01 -3.22557569e-01 -6.64079964e-01
-1.26506782e+00 1.97645307e-01 5.38388968e-01 -2.42446005e-01
-9.23913300e-01 1.78052649e-01 -4.34099436e-01 5.06708205e-01
-6.97901994e-02 -1.49102047e-01 -1.72335535e-01 -7.96437263e-02
8.87570083e-01 7.11597085e-01 3.24486166e-01 1.50981266e-02
-5.13013899e-01 9.26172793e-01 6.13252878e-01 2.25051478e-01
6.87523723e-01 -3.05983305e-01 9.35394391e-02 4.62132394e-01
8.59220386e-01 -2.57539973e-02 -8.70219767e-01 -1.13547646e-01
-1.28621578e-01 -8.36876154e-01 2.13214278e-01 -5.48063755e-01
-9.64918911e-01 8.20890963e-01 9.49607849e-01 3.95841032e-01
1.50269926e+00 -8.08971047e-01 8.43112469e-01 -8.91605578e-03
3.76043051e-01 -1.23188674e+00 -1.69604391e-01 -4.43795264e-01
4.78559881e-01 -6.64420962e-01 5.63792706e-01 -8.79346311e-01
-3.68362099e-01 1.26439941e+00 3.38814974e-01 -3.88362378e-01
7.56787777e-01 1.75309211e-01 2.35322729e-01 1.75206155e-01
-4.07779127e-01 -5.49437217e-02 4.26911622e-01 4.92672533e-01
7.28453755e-01 6.88846782e-02 -1.45915473e+00 2.80950785e-01
2.00939119e-01 4.69434440e-01 4.43115354e-01 6.72168255e-01
-1.42920747e-01 -8.64804864e-01 -2.68210322e-01 4.76256758e-01
-1.03490460e+00 2.69947816e-02 2.73169484e-02 5.49103916e-01
-4.51963395e-01 1.14294863e+00 -2.52376229e-01 -2.22470816e-02
5.93950331e-01 3.18347663e-01 5.07949948e-01 -2.11972594e-01
-9.27831054e-01 4.06604648e-01 1.69483006e-01 -8.28735888e-01
-5.03401458e-01 -3.77300531e-01 -7.30049372e-01 -2.23051667e-01
-3.48280907e-01 1.64787769e-01 9.02105570e-01 4.97174203e-01
5.60829461e-01 2.66778260e-01 7.78737009e-01 -6.07703686e-01
-6.31942749e-01 -1.06911039e+00 -8.97808313e-01 4.24813330e-01
1.47526681e-01 -2.33045414e-01 -4.65182215e-01 4.77148443e-02] | [13.924986839294434, 2.9539175033569336] |
7eab7df8-a444-45a0-8e45-07d4c47f3d5a | landmark-tracking-in-liver-us-images-using | 2209.06952 | null | https://arxiv.org/abs/2209.06952v1 | https://arxiv.org/pdf/2209.06952v1.pdf | Landmark Tracking in Liver US images Using Cascade Convolutional Neural Networks with Long Short-Term Memory | This study proposed a deep learning-based tracking method for ultrasound (US) image-guided radiation therapy. The proposed cascade deep learning model is composed of an attention network, a mask region-based convolutional neural network (mask R-CNN), and a long short-term memory (LSTM) network. The attention network learns a mapping from a US image to a suspected area of landmark motion in order to reduce the search region. The mask R-CNN then produces multiple region-of-interest (ROI) proposals in the reduced region and identifies the proposed landmark via three network heads: bounding box regression, proposal classification, and landmark segmentation. The LSTM network models the temporal relationship among the successive image frames for bounding box regression and proposal classification. To consolidate the final proposal, a selection method is designed according to the similarities between sequential frames. The proposed method was tested on the liver US tracking datasets used in the Medical Image Computing and Computer Assisted Interventions (MICCAI) 2015 challenges, where the landmarks were annotated by three experienced observers to obtain their mean positions. Five-fold cross-validation on the 24 given US sequences with ground truths shows that the mean tracking error for all landmarks is 0.65+/-0.56 mm, and the errors of all landmarks are within 2 mm. We further tested the proposed model on 69 landmarks from the testing dataset that has a similar image pattern to the training pattern, resulting in a mean tracking error of 0.94+/-0.83 mm. Our experimental results have demonstrated the feasibility and accuracy of our proposed method in tracking liver anatomic landmarks using US images, providing a potential solution for real-time liver tracking for active motion management during radiation therapy. | ['Xiaofeng Yang', 'Tian Liu', 'Yue Chen', 'Pretesh Patel', 'Jacob F. Wynne', 'Yang Lei', 'Zhen Tian', 'Xianjin Dai', 'Yupei Zhang'] | 2022-09-14 | null | null | null | null | ['landmark-tracking'] | ['computer-vision'] | [-5.94811179e-02 2.06857890e-01 -2.75226295e-01 -9.49518606e-02
-1.02794969e+00 -4.31957185e-01 2.08842829e-01 9.29297954e-02
-4.65194434e-01 2.49272734e-01 2.80026555e-01 -3.30700904e-01
-1.39986202e-01 -5.16146183e-01 -6.69431806e-01 -1.09004533e+00
-5.20686269e-01 3.02208513e-01 3.38311553e-01 2.79674619e-01
1.41817629e-01 6.26091957e-01 -4.44061249e-01 2.63250023e-01
5.42890012e-01 1.17968786e+00 4.24640864e-01 7.04522431e-01
8.25385749e-02 8.49827051e-01 -4.68239963e-01 2.96435535e-01
3.22386533e-01 -5.40584028e-01 -8.89238298e-01 4.03690115e-02
2.83831656e-01 -2.74273962e-01 -5.51833987e-01 9.47801530e-01
8.49430442e-01 3.26881200e-01 3.93247426e-01 -3.00693393e-01
-4.88415927e-01 6.10761940e-01 -6.14020705e-01 6.46983385e-01
1.43796444e-01 2.65117306e-02 2.46953249e-01 -8.47219467e-01
7.00205147e-01 5.70623219e-01 9.97912645e-01 6.43109918e-01
-6.34654641e-01 -5.88116407e-01 -1.99708074e-01 -7.21007660e-02
-1.48963916e+00 -8.35058317e-02 4.21852469e-01 -5.99656880e-01
6.97854877e-01 2.65864819e-01 6.41235590e-01 4.73467529e-01
9.18164015e-01 5.57994306e-01 6.32792830e-01 -4.87217098e-01
-1.14304781e-01 -1.31658450e-01 3.55919413e-02 1.25405443e+00
2.47910451e-02 4.19772118e-01 2.42597312e-01 3.87622160e-03
1.30182576e+00 1.81242257e-01 -6.99333191e-01 -2.73099422e-01
-1.57377815e+00 7.90663242e-01 1.24819314e+00 9.47554111e-01
-5.23248911e-01 1.83621407e-01 4.58904058e-01 -2.66462058e-01
2.91143239e-01 2.66327739e-01 -3.22995126e-01 3.61886442e-01
-7.98576593e-01 -6.02316320e-01 6.61586165e-01 8.20380926e-01
1.67336985e-02 8.72123316e-02 -5.86449623e-01 5.33923090e-01
4.36915666e-01 6.49049953e-02 1.07420421e+00 -3.75445545e-01
-2.46396568e-02 3.60220522e-01 8.73498768e-02 -9.17078972e-01
-8.65488410e-01 -7.66819596e-01 -1.08191919e+00 -2.03183293e-02
4.31626678e-01 -1.75410032e-01 -1.40811670e+00 1.42108607e+00
3.42868119e-01 4.69713897e-01 -3.26026887e-01 1.35307205e+00
1.48120224e+00 5.11299193e-01 2.85021007e-01 -5.01973391e-01
1.41667521e+00 -1.21269357e+00 -7.27743804e-01 2.13180959e-01
8.91332746e-01 -7.74867654e-01 5.66424787e-01 -1.93239152e-01
-1.10969675e+00 -6.88785255e-01 -9.36415553e-01 3.50533724e-01
3.91473472e-02 5.42438507e-01 3.63492399e-01 5.23467064e-01
-1.20391393e+00 6.51609302e-01 -1.25571680e+00 -3.39869827e-01
3.99607211e-01 7.08640397e-01 -2.49188632e-01 3.04542154e-01
-8.58407974e-01 1.08278000e+00 3.48918587e-01 6.31278396e-01
-9.65252936e-01 -8.25231194e-01 -9.92229998e-01 4.17028777e-02
1.16581231e-01 -6.24812245e-01 1.17958188e+00 -9.13688898e-01
-1.50633192e+00 1.00914192e+00 -1.37989566e-01 -3.70674521e-01
4.86184925e-01 2.17352107e-01 -2.89648890e-01 1.34367615e-01
7.56908730e-02 4.06529903e-01 6.49259508e-01 -1.10555339e+00
-5.87360203e-01 -1.67662367e-01 -2.15786800e-01 -8.54214802e-02
1.79089054e-01 7.66226351e-02 -4.94120330e-01 -9.58290219e-01
5.91787994e-01 -1.07733607e+00 -4.20873463e-01 -4.30518724e-02
-2.99619853e-01 -1.09120846e-01 5.93188465e-01 -9.10611212e-01
1.27215457e+00 -1.92458069e+00 -1.30219180e-02 3.28652978e-01
2.24961117e-01 1.53515786e-01 3.32596228e-02 -4.72964257e-01
-3.77282113e-01 5.38580567e-02 -4.68627689e-03 9.72612016e-03
-7.20160544e-01 -1.27481043e-01 2.79967070e-01 9.51728225e-01
-4.41105515e-01 1.14569855e+00 -9.72598910e-01 -6.72531962e-01
4.62572604e-01 4.04846013e-01 9.21420462e-04 4.13832158e-01
2.23085344e-01 1.00390959e+00 -3.32881600e-01 7.95174062e-01
4.55426753e-01 -4.27877933e-01 8.59229788e-02 -6.40332162e-01
-2.19298169e-01 -2.04335809e-01 -8.16794336e-01 2.09675694e+00
-5.19470334e-01 5.97834229e-01 -5.53648956e-02 -5.74605942e-01
7.74615169e-01 6.57003284e-01 1.01041830e+00 -5.41760564e-01
6.37346506e-01 2.00363308e-01 3.51298481e-01 -8.13094854e-01
-4.82148193e-02 -2.30935887e-01 2.69482613e-01 5.22318661e-01
9.26068276e-02 2.70049095e-01 -1.10487543e-01 -3.04838419e-01
8.61689925e-01 8.39924365e-02 3.30146074e-01 -4.94661987e-01
8.20573211e-01 6.89482465e-02 5.80729604e-01 8.02433670e-01
-6.36931539e-01 7.90571272e-01 -1.60424724e-01 -1.12211263e+00
-5.77381909e-01 -7.37026155e-01 -3.59956503e-01 6.47101939e-01
4.21735644e-01 7.75326565e-02 -5.58361471e-01 -1.03158844e+00
-2.84744829e-01 3.45826328e-01 -1.03573072e+00 1.97153538e-02
-1.09114802e+00 -6.26313567e-01 2.19281331e-01 5.81975877e-01
1.96070939e-01 -1.42169154e+00 -8.60074759e-01 3.02182674e-01
-9.79763344e-02 -7.93607831e-01 -8.00391376e-01 3.51617336e-02
-1.01482141e+00 -1.33667469e+00 -1.07154620e+00 -1.36226082e+00
1.18924618e+00 1.00487828e-01 8.00812721e-01 2.83547521e-01
-3.85270357e-01 1.19108982e-01 -1.10092297e-01 7.74393743e-03
-5.30234277e-01 -9.74904522e-02 -1.30150527e-01 -2.06058919e-01
-6.49549812e-02 1.56005070e-01 -9.97735262e-01 4.26494092e-01
-3.87626767e-01 -9.30580962e-03 7.33606875e-01 1.18493938e+00
6.75667882e-01 -3.29564959e-01 2.49740109e-01 -5.83639145e-01
1.78294018e-01 -3.94079030e-01 -7.24374533e-01 4.24658895e-01
-2.48170584e-01 -3.28311533e-01 3.98613602e-01 -5.22323549e-01
-8.39717984e-01 2.14803353e-01 2.77964808e-02 -7.15188146e-01
-4.39541824e-02 5.98629475e-01 7.62136638e-01 -6.52058780e-01
5.11987984e-01 2.25633964e-01 8.85276869e-03 -1.74492449e-02
4.04312499e-02 2.73927867e-01 6.20019972e-01 -4.10142317e-02
3.45843673e-01 3.06991965e-01 1.37603320e-02 -3.43144655e-01
-6.03020310e-01 -3.89958113e-01 -8.27656150e-01 -4.22751099e-01
1.12387133e+00 -6.76854074e-01 -8.17031622e-01 8.08795840e-02
-1.16841936e+00 -3.23545426e-01 -2.20314488e-01 9.55780447e-01
-3.30864370e-01 2.73263514e-01 -1.18279743e+00 -3.59936953e-01
-9.59081411e-01 -1.71685863e+00 7.58188486e-01 6.26237512e-01
-9.75395590e-02 -1.19387269e+00 -8.64481106e-02 -1.39532715e-01
6.27268016e-01 5.19847512e-01 6.65905356e-01 -7.37698615e-01
-6.95581436e-01 -4.95845914e-01 -2.88718522e-01 -1.73675165e-01
5.25747299e-01 -2.35194117e-01 -6.40625536e-01 -6.77596807e-01
2.30898604e-01 2.29045808e-01 6.55743003e-01 1.40130663e+00
1.28680062e+00 -4.45266664e-02 -5.72487414e-01 9.42994654e-01
1.42404699e+00 8.03190172e-01 2.64688253e-01 3.50415915e-01
7.42907107e-01 -8.96138325e-02 5.21925271e-01 2.32154518e-01
-9.23992246e-02 4.54948157e-01 4.49164450e-01 -6.58517003e-01
-2.69821465e-01 2.59681433e-01 -2.50902623e-01 7.58334279e-01
-1.98099852e-01 2.34057605e-01 -1.08149219e+00 5.08638442e-01
-1.62959242e+00 -6.96347654e-01 -6.08222224e-02 2.07951093e+00
5.77807486e-01 -1.28285497e-01 -3.14646214e-01 -4.81208980e-01
8.30504954e-01 1.65911198e-01 -2.64129609e-01 -8.31520334e-02
5.01300216e-01 2.97471911e-01 6.22092247e-01 6.19391441e-01
-1.54336178e+00 7.11374640e-01 5.83237839e+00 5.73491693e-01
-1.55510378e+00 2.76685983e-01 7.64809489e-01 2.74277389e-01
3.04473847e-01 -3.06051612e-01 -3.79269958e-01 3.03835094e-01
5.34041107e-01 1.34270731e-02 -7.54370466e-02 7.11826503e-01
3.27825010e-01 -2.01143757e-01 -1.00891876e+00 9.14128184e-01
1.77202702e-01 -1.67723727e+00 -2.11588174e-01 -3.36202353e-01
8.06579769e-01 1.86197385e-01 -6.33570701e-02 1.54462740e-01
-1.71583757e-01 -1.13085055e+00 4.20304269e-01 6.39401317e-01
9.72039878e-01 -3.74857038e-01 1.24654889e+00 2.58142680e-01
-1.33149910e+00 1.55643011e-02 -2.44044289e-01 5.16731739e-01
-4.86364327e-02 -2.44950578e-02 -1.20941901e+00 7.36044049e-01
6.21811807e-01 5.79016864e-01 -2.80403227e-01 1.35881197e+00
1.39062004e-02 2.22317949e-01 -9.13344249e-02 8.53049681e-02
4.60632771e-01 -1.35614052e-01 5.51330149e-01 1.32697535e+00
5.33676863e-01 3.73682976e-01 3.67520809e-01 6.92686558e-01
3.18430588e-02 1.06310494e-01 -3.07177514e-01 5.97251654e-01
2.55363882e-01 1.48840785e+00 -1.21158910e+00 -4.35645759e-01
-3.17631155e-01 6.07324660e-01 -1.71374410e-01 2.89597273e-01
-1.01133895e+00 -1.05344914e-01 -2.25838259e-01 -3.91606987e-02
1.99966207e-01 1.20396037e-02 -3.21113467e-01 -9.60370779e-01
-3.91334534e-01 -3.38137954e-01 7.84759939e-01 -5.67967713e-01
-8.51789713e-01 1.08578336e+00 -4.22040284e-01 -1.53163528e+00
-8.06951597e-02 -3.07813913e-01 -9.35682714e-01 1.06543732e+00
-1.20196688e+00 -1.18337715e+00 -7.89218962e-01 4.28954363e-01
7.35017836e-01 -1.03180006e-01 9.38431561e-01 1.83489919e-01
-5.93681574e-01 5.50916791e-01 -1.80109695e-01 6.62527919e-01
5.05592227e-01 -1.07888198e+00 -7.33538810e-03 8.47923696e-01
-1.15797386e-01 7.53677428e-01 2.42150992e-01 -7.46245742e-01
-1.34865093e+00 -1.25924754e+00 4.13076401e-01 -2.28429541e-01
4.11261499e-01 3.34684104e-01 -7.67555535e-01 9.96197283e-01
2.83933699e-01 8.14549625e-01 5.73150218e-01 -6.78650618e-01
5.68648994e-01 1.59450144e-01 -1.41432464e+00 1.70085296e-01
5.86516500e-01 -4.43083867e-02 -4.05744046e-01 4.97492224e-01
5.09198070e-01 -1.43935323e+00 -1.25063884e+00 8.70241642e-01
6.17487848e-01 -5.89427590e-01 9.79686379e-01 -4.39324260e-01
5.73535562e-02 -4.42870468e-01 3.51309001e-01 -1.12498713e+00
-7.44586289e-01 -4.14531648e-01 5.55688441e-02 3.72704297e-01
2.93156713e-01 -3.09853584e-01 8.63338292e-01 4.19960141e-01
-6.37850702e-01 -1.18868804e+00 -1.11553156e+00 -2.13182181e-01
-4.38088663e-02 -7.65136108e-02 1.66736498e-01 1.08311844e+00
-9.72386152e-02 -3.43871444e-01 5.04207890e-03 5.20272255e-01
5.00597417e-01 2.55447686e-01 1.75678313e-01 -5.48775494e-01
3.22814062e-02 -5.56698084e-01 -3.03629518e-01 -9.59656835e-01
-9.67862159e-02 -1.18118584e+00 1.30696893e-01 -1.72893846e+00
2.28683919e-01 -6.01431370e-01 -7.06835032e-01 4.23524439e-01
-1.87076747e-01 2.03998551e-01 4.79871184e-02 3.36003065e-01
-3.43165368e-01 1.65121913e-01 1.71891141e+00 -3.02663475e-01
-4.49518621e-01 4.44496632e-01 3.89715359e-02 7.30439126e-01
5.07781863e-01 -4.51140285e-01 1.73322707e-01 -4.14812356e-01
-7.69688189e-01 7.08244741e-01 1.61163658e-01 -9.73976791e-01
6.89355671e-01 2.16718093e-01 1.02942324e+00 -7.90795803e-01
-1.55538872e-01 -9.70362306e-01 1.41329706e-01 1.24465382e+00
-2.47797176e-01 -5.10916337e-02 2.54896879e-01 1.76878273e-01
-2.11951375e-01 -2.32435539e-01 1.01798773e+00 -3.76816243e-01
-6.99956119e-01 7.47223496e-01 -2.13152006e-01 -5.53632200e-01
1.34388506e+00 -2.42241308e-01 4.44732495e-02 -1.21965325e-02
-1.29709697e+00 1.13430001e-01 -1.80828050e-01 2.58476555e-01
9.44867551e-01 -1.36775661e+00 -6.91318870e-01 4.24093783e-01
-2.00104684e-01 1.76670834e-01 3.34901720e-01 1.47276223e+00
-9.48655486e-01 5.48000515e-01 1.28811691e-02 -1.13534737e+00
-1.15233231e+00 5.60638845e-01 1.14255512e+00 -2.88733840e-01
-9.28338706e-01 1.03586447e+00 2.83755779e-01 -1.67797834e-01
3.92197579e-01 -8.45714211e-01 -3.57956022e-01 -4.06887859e-01
3.85573864e-01 1.46380765e-02 1.79912433e-01 -7.92484641e-01
-4.53106165e-01 8.99955273e-01 -7.39724562e-02 4.22945589e-01
1.08113050e+00 -1.72421262e-02 -1.04576759e-01 -8.81719068e-02
1.10186088e+00 -2.48308167e-01 -9.57611442e-01 -3.23527962e-01
1.60811886e-01 -4.29107040e-01 3.19543928e-01 -8.10164988e-01
-1.59472919e+00 5.00668108e-01 1.16652787e+00 1.76422577e-02
1.12829232e+00 -2.41964888e-02 8.65887523e-01 -2.55293757e-01
3.60627621e-01 -2.32758939e-01 -6.64688051e-02 4.09513980e-01
9.94234562e-01 -1.42929256e+00 -5.54105602e-02 -1.88936427e-01
-4.45035905e-01 1.58293641e+00 7.31862664e-01 -3.25923622e-01
7.84752131e-01 2.48719722e-01 3.19239765e-01 -3.26216877e-01
-1.13962457e-01 2.59986483e-02 5.61199069e-01 2.32570097e-01
6.81897104e-01 7.34302402e-02 -2.95541853e-01 5.12948096e-01
3.96777123e-01 8.59690085e-02 3.50873679e-01 7.97586799e-01
-3.62836808e-01 -3.90401065e-01 -4.75018889e-01 3.62184614e-01
-6.63971782e-01 1.56360641e-01 5.37610710e-01 9.69141781e-01
1.24540538e-01 5.83424270e-01 3.15768346e-02 -1.76734582e-01
2.64390677e-01 -3.50430787e-01 5.93224406e-01 -5.19526064e-01
-1.07551253e+00 5.75476944e-01 -3.99822444e-01 -6.18033350e-01
-2.74254411e-01 -2.52104551e-01 -1.60150623e+00 5.51154278e-02
-5.68543494e-01 2.18423307e-01 6.27809048e-01 7.63005495e-01
-3.53406258e-02 9.28721249e-01 7.37911165e-01 -1.06058741e+00
-2.83837199e-01 -1.15015984e+00 -2.27964699e-01 2.51101255e-01
5.55403948e-01 -5.25922477e-01 -1.78922683e-01 -5.57268551e-03] | [14.42657470703125, -2.6494498252868652] |
69753a6a-b140-4ad5-a49a-117776fb5de9 | direction-aware-spatial-context-features-for-1 | 1805.04635 | null | https://arxiv.org/abs/1805.04635v2 | https://arxiv.org/pdf/1805.04635v2.pdf | Direction-aware Spatial Context Features for Shadow Detection and Removal | Shadow detection and shadow removal are fundamental and challenging tasks, requiring an understanding of the global image semantics. This paper presents a novel deep neural network design for shadow detection and removal by analyzing the spatial image context in a direction-aware manner. To achieve this, we first formulate the direction-aware attention mechanism in a spatial recurrent neural network (RNN) by introducing attention weights when aggregating spatial context features in the RNN. By learning these weights through training, we can recover direction-aware spatial context (DSC) for detecting and removing shadows. This design is developed into the DSC module and embedded in a convolutional neural network (CNN) to learn the DSC features at different levels. Moreover, we design a weighted cross entropy loss to make effective the training for shadow detection and further adopt the network for shadow removal by using a Euclidean loss function and formulating a color transfer function to address the color and luminosity inconsistencies in the training pairs. We employed two shadow detection benchmark datasets and two shadow removal benchmark datasets, and performed various experiments to evaluate our method. Experimental results show that our method performs favorably against the state-of-the-art methods for both shadow detection and shadow removal. | ['Pheng-Ann Heng', 'Chi-Wing Fu', 'Xiaowei Hu', 'Lei Zhu', 'Jing Qin'] | 2018-05-12 | null | null | null | null | ['shadow-removal', 'shadow-detection-and-removal', 'shadow-detection'] | ['computer-vision', 'computer-vision', 'computer-vision'] | [ 6.37583613e-01 -2.20912531e-01 3.42642128e-01 -7.21635342e-01
-1.45384178e-01 -1.62697256e-01 4.07608807e-01 -4.74074572e-01
-3.84085000e-01 6.86132729e-01 3.35789144e-01 -4.47274178e-01
2.57722437e-01 -6.33317649e-01 -6.96078897e-01 -1.11230111e+00
8.55326727e-02 -3.43520820e-01 7.17536509e-01 -1.91992417e-01
3.73154193e-01 6.81391776e-01 -1.39358914e+00 7.67936260e-02
1.12804687e+00 1.16934133e+00 7.40565181e-01 6.90310061e-01
-5.45182116e-02 1.16047943e+00 -9.10736978e-01 9.72104445e-02
2.64460981e-01 -5.50868154e-01 -2.45508358e-01 -1.28676146e-01
2.85196394e-01 -5.90753913e-01 -5.67335546e-01 6.26845241e-01
6.86966360e-01 3.47387135e-01 4.22467798e-01 -1.23556530e+00
-8.94927740e-01 1.60532317e-03 -5.54986179e-01 2.59296924e-01
-5.07134050e-02 3.12800705e-02 7.22154796e-01 -9.59953666e-01
2.51829952e-01 1.19281435e+00 5.90442955e-01 2.13753060e-01
-6.72396839e-01 -6.21984482e-01 5.05014420e-01 5.44266045e-01
-1.20629048e+00 -2.55916119e-01 1.11038446e+00 2.57963479e-01
6.46022379e-01 4.36606884e-01 6.72503948e-01 1.01752591e+00
4.08442289e-01 1.19329190e+00 1.17272699e+00 -2.88460970e-01
9.75181684e-02 2.09986921e-02 3.16319168e-02 9.09723103e-01
7.41274878e-02 1.55634612e-01 -5.22592545e-01 1.51594371e-01
6.24144018e-01 2.70246595e-01 -6.10992849e-01 -3.86679709e-01
-9.44038033e-01 6.02807105e-01 1.03246927e+00 -8.89848396e-02
-1.49235323e-01 4.77061450e-01 2.59194702e-01 -1.58752277e-01
4.09875661e-01 7.51008168e-02 -4.37641740e-01 4.68301237e-01
-7.03721404e-01 1.34744616e-02 5.64382195e-01 7.57526398e-01
7.40221322e-01 2.87168771e-01 -7.07699597e-01 7.74140120e-01
2.49951988e-01 9.12798464e-01 2.33192697e-01 -8.13955605e-01
2.75162637e-01 4.70202982e-01 1.57543540e-01 -1.22074652e+00
-4.64269757e-01 -5.62536538e-01 -7.80629039e-01 4.80403990e-01
-1.17689691e-01 -3.50021683e-02 -1.17447340e+00 1.66321003e+00
2.63413578e-01 5.51479876e-01 1.31864905e-01 1.13294327e+00
8.41253638e-01 6.87335670e-01 -3.34666558e-02 -1.73326373e-01
1.13192439e+00 -1.28481472e+00 -9.57049549e-01 -4.35257107e-01
1.67443380e-01 -6.98814869e-01 1.35965812e+00 1.48674874e-02
-6.00508034e-01 -5.01824915e-01 -1.43479276e+00 -4.04479831e-01
-7.90734708e-01 3.35347533e-01 6.57815099e-01 3.61905307e-01
-9.47667122e-01 3.59370619e-01 -6.10727489e-01 -1.59086362e-01
4.87704456e-01 2.06997693e-01 2.70026058e-01 -3.35403755e-02
-1.28245330e+00 8.68471801e-01 6.06328771e-02 6.94175839e-01
-9.30675685e-01 -4.37703460e-01 -8.50353420e-01 1.99400336e-01
5.79246283e-01 -4.98607665e-01 9.25926924e-01 -1.30333674e+00
-1.44762421e+00 2.72342473e-01 -3.68430108e-01 -3.07690501e-01
3.53531897e-01 -4.96948183e-01 -5.01299798e-01 -1.07081681e-01
9.40578729e-02 2.71887422e-01 1.04427338e+00 -1.81644547e+00
-7.19242871e-01 -6.17154278e-02 2.15037927e-01 5.61145306e-01
-2.34031960e-01 -1.46235198e-01 -7.15581656e-01 -9.39052820e-01
6.83846995e-02 -8.46527755e-01 -5.70524782e-02 2.35256180e-01
-6.83985114e-01 1.66334301e-01 1.62617731e+00 -9.18209076e-01
1.19087410e+00 -1.98527396e+00 9.72929895e-02 1.79491028e-01
2.25799337e-01 4.52912927e-01 -1.06057957e-01 -9.64820758e-02
2.30994865e-01 -3.77960026e-01 -8.17206144e-01 -6.08451247e-01
1.05502456e-02 4.41744268e-01 -6.14422739e-01 4.70722347e-01
8.04739073e-02 1.03225195e+00 -8.56429935e-01 -4.76176053e-01
4.77355331e-01 8.11544359e-01 -1.57012627e-01 5.70804238e-01
-2.00281307e-01 1.14625432e-01 -4.65518922e-01 8.60313058e-01
9.58276331e-01 1.02094002e-02 1.49795422e-02 -4.26613063e-01
-1.78227291e-01 1.00938268e-02 -8.30248594e-01 1.39625096e+00
-7.65331805e-01 1.08317769e+00 4.58246581e-02 -6.09520733e-01
7.87699819e-01 -3.83712262e-01 7.58980960e-02 -1.09363103e+00
2.51201242e-01 -3.10553741e-02 -2.19457105e-01 -4.19454753e-01
6.68955743e-01 1.92084983e-01 6.77283434e-03 4.41358119e-01
-4.95577365e-01 1.15722299e-01 -5.22269309e-01 1.61128759e-01
1.02593327e+00 2.56813258e-01 -1.27541289e-01 -2.25540355e-01
7.06923842e-01 -4.94947374e-01 7.59715617e-01 7.29877412e-01
-2.36231461e-01 6.97263539e-01 2.74752706e-01 -4.37403023e-01
-4.25296843e-01 -1.13140976e+00 2.13235840e-01 1.49341905e+00
7.47115195e-01 2.62938309e-02 -5.90605021e-01 -9.86777186e-01
1.46382734e-01 8.52510512e-01 -9.91207898e-01 -4.06836033e-01
-7.47023702e-01 -8.75344038e-01 3.15965086e-01 8.09196174e-01
1.10857356e+00 -1.38655841e+00 -1.07318270e+00 -1.55146420e-01
-1.70759723e-01 -9.71800208e-01 -5.61361909e-01 5.74796617e-01
-1.84789315e-01 -1.03508306e+00 -7.39865601e-01 -7.89783180e-01
6.33313000e-01 9.41148698e-01 8.95585358e-01 3.71892452e-01
-4.76467818e-01 2.20644936e-01 -3.33699763e-01 -7.47815251e-01
3.75536829e-01 -1.56334396e-02 -2.47170687e-01 1.80465117e-01
-1.46894023e-01 -5.68011165e-01 -1.18379247e+00 2.36219436e-01
-9.40981507e-01 1.89080819e-01 9.82187986e-01 7.77624130e-01
3.32873642e-01 -1.57404765e-01 2.15738252e-01 -9.74870265e-01
5.80039799e-01 -2.07478672e-01 -6.55670106e-01 5.11170506e-01
-6.20092034e-01 8.14487860e-02 6.65339172e-01 5.79633676e-02
-1.70277870e+00 5.02696894e-02 2.28504762e-01 -3.22610945e-01
2.82749534e-01 9.77439806e-03 -5.04074514e-01 -3.75865459e-01
2.27864608e-01 5.01680970e-01 -5.22014976e-01 -1.09122045e-01
4.94282573e-01 4.06267166e-01 5.39662182e-01 -1.68108597e-01
1.04525185e+00 8.50081325e-01 -1.46475690e-03 -7.33817518e-01
-1.23196244e+00 -2.53212810e-01 -6.71146750e-01 -2.89488018e-01
8.96061003e-01 -5.66974461e-01 -6.70225859e-01 7.39622235e-01
-1.11876941e+00 -8.01029384e-01 7.34717175e-02 -1.94752783e-01
-2.00731382e-01 3.71875077e-01 -8.08848590e-02 -9.24861372e-01
-3.87387067e-01 -9.44489777e-01 1.33443618e+00 4.14683312e-01
4.76698607e-01 -1.06567097e+00 -9.30437893e-02 -2.80918777e-02
6.91828251e-01 3.49103421e-01 8.62272918e-01 2.06802804e-02
-1.00843728e+00 1.35005429e-01 -9.53237891e-01 4.78158742e-01
3.26421559e-01 -2.95169562e-01 -1.40361762e+00 1.79351866e-02
5.82616851e-02 2.76324630e-04 1.49556947e+00 5.21499753e-01
1.84447956e+00 -2.20627323e-01 -4.47721928e-01 1.10151732e+00
1.48058116e+00 1.76370919e-01 7.99060524e-01 3.47531915e-01
1.28168344e+00 4.59546119e-01 7.22457230e-01 3.29368025e-01
4.59356338e-01 5.14553189e-01 6.43751860e-01 -7.05710351e-01
-4.56033498e-01 -4.58909795e-02 2.01045930e-01 3.69417876e-01
1.16565926e-02 -5.08513451e-01 -5.16663551e-01 2.86084980e-01
-2.07690597e+00 -7.35751748e-01 7.61805773e-02 1.88001943e+00
5.06087005e-01 1.44638613e-01 -5.88546693e-01 -1.31498724e-01
3.92463207e-01 8.38262618e-01 -8.74184847e-01 -1.99421331e-01
-4.99185324e-01 4.53739800e-02 9.28798139e-01 7.28359401e-01
-1.26252341e+00 1.41051507e+00 5.75642872e+00 6.61531866e-01
-1.37610841e+00 -5.30197434e-02 5.54887831e-01 6.72275946e-03
-4.57710564e-01 5.83581962e-02 -3.76645327e-01 5.44741869e-01
2.84874707e-01 4.88092154e-01 5.90791285e-01 5.75762689e-01
4.41622555e-01 -5.23197591e-01 -4.68340516e-01 8.52594793e-01
4.59629834e-01 -1.03281569e+00 -1.02048434e-01 -2.98383772e-01
7.33894408e-01 3.41063365e-02 2.12240785e-01 3.72299105e-01
1.67021558e-01 -1.04162455e+00 6.83113515e-01 8.64948392e-01
5.66154420e-01 -6.35562122e-01 7.42558539e-01 -1.22754820e-01
-1.48135865e+00 -3.78739804e-01 -4.14013147e-01 2.06962422e-01
-3.23854829e-03 6.05515480e-01 -6.89504981e-01 4.53657687e-01
7.87631154e-01 6.72881305e-01 -7.24984169e-01 6.77358985e-01
-5.53339422e-01 2.92500496e-01 -3.95383090e-02 -9.38873515e-02
1.87048569e-01 -1.96012199e-01 1.99915797e-01 1.36759293e+00
-5.31136207e-02 1.56726763e-01 -2.17781030e-02 7.41947591e-01
-1.23239808e-01 -3.15634310e-01 -3.86425704e-01 5.24645030e-01
4.69649047e-01 1.42513835e+00 -9.35857117e-01 -3.43929768e-01
-2.64013112e-01 1.66211343e+00 2.44449139e-01 8.29694510e-01
-1.32348049e+00 -8.67190182e-01 6.19162977e-01 -2.76219428e-01
2.60907531e-01 -1.94988102e-01 -5.37160873e-01 -7.72863805e-01
1.14291072e-01 -3.27733368e-01 -2.23732945e-02 -1.20605063e+00
-8.76496553e-01 5.19660890e-01 -1.88146159e-01 -8.06818008e-01
4.47408259e-01 -6.53745532e-01 -1.10373998e+00 8.55169475e-01
-2.30887699e+00 -1.48022127e+00 -1.12538266e+00 7.08052397e-01
5.49546838e-01 7.93593004e-02 3.45137894e-01 2.26205230e-01
-7.24735916e-01 4.91232961e-01 9.53663960e-02 -8.43149703e-03
8.79719257e-01 -1.33822548e+00 4.01556432e-01 9.59822655e-01
-4.61842626e-01 3.37524235e-01 5.97731292e-01 -7.05462515e-01
-1.37477207e+00 -1.41393614e+00 5.30229986e-01 -2.68699676e-01
2.95672357e-01 -5.88607371e-01 -9.26247716e-01 4.59812909e-01
2.08220184e-01 2.31874049e-01 2.89185196e-01 -1.70298278e-01
-3.04953665e-01 -3.95925552e-01 -8.93925309e-01 8.00572872e-01
1.49357736e+00 -6.44266188e-01 -3.29387069e-01 4.21056986e-01
1.01858687e+00 -4.19713378e-01 5.76888844e-02 5.83872676e-01
6.62367463e-01 -1.26556063e+00 1.17683077e+00 4.08084095e-02
1.93273216e-01 -5.67621350e-01 -3.55377972e-01 -1.17586792e+00
-2.12109894e-01 -2.13511229e-01 -2.53059119e-01 1.01670873e+00
1.61953747e-01 -7.13723361e-01 7.44506180e-01 2.72896349e-01
-5.50293744e-01 -1.12863100e+00 -6.35226905e-01 -3.75854164e-01
-4.89227176e-01 -2.81962425e-01 6.90087914e-01 5.25933802e-01
-8.30651879e-01 1.48934901e-01 -5.63036561e-01 5.00262856e-01
5.31611443e-01 4.86240029e-01 6.50614440e-01 -8.20265412e-01
1.07364751e-01 -4.55373466e-01 1.46367192e-01 -1.13404977e+00
3.94814789e-01 -4.63431567e-01 5.78455627e-01 -1.85478234e+00
2.62992799e-01 -4.63012367e-01 -5.90109825e-01 6.17049456e-01
-5.80534101e-01 2.26415366e-01 1.91520035e-01 6.79993108e-02
-7.72504747e-01 1.22660041e+00 1.18198872e+00 -1.04638726e-01
-2.51146257e-01 2.12395154e-02 -5.96555948e-01 7.89146483e-01
7.25500405e-01 -1.82217985e-01 -5.93683839e-01 -6.74065769e-01
-1.87783375e-01 -4.27841783e-01 7.71023631e-01 -1.04643929e+00
2.64853597e-01 -3.65072459e-01 8.02244782e-01 -9.00125623e-01
7.29141831e-01 -9.44442451e-01 -6.42528176e-01 4.48785305e-01
-2.33229160e-01 -2.38869548e-01 8.91372636e-02 7.86923230e-01
1.42921492e-01 3.65822315e-01 6.93627656e-01 6.24673441e-02
-1.13822711e+00 2.33058766e-01 -1.09454021e-01 -1.75063893e-01
9.12904501e-01 -2.00367972e-01 -3.74125153e-01 -3.96305650e-01
-8.56662393e-02 3.97229433e-01 3.00060093e-01 4.16698426e-01
1.08135235e+00 -1.21381819e+00 -2.45177045e-01 3.92915875e-01
-5.47611795e-04 -4.73307185e-02 2.58467138e-01 7.06337988e-01
-5.77624321e-01 2.84767181e-01 -1.52701408e-01 -3.19353014e-01
-1.19282365e+00 2.48596966e-01 5.25324881e-01 1.17442623e-01
-5.73425055e-01 1.05151236e+00 8.16260099e-01 -3.86967033e-01
5.86438119e-01 -5.19340158e-01 -9.05243158e-02 -2.70715594e-01
2.77906001e-01 3.78763825e-01 -9.74182189e-02 -4.81493354e-01
-4.34451431e-01 5.78931868e-01 3.23330641e-01 1.38670295e-01
1.10750425e+00 -3.80318671e-01 -1.47662163e-01 2.68794239e-01
1.34843493e+00 2.58999541e-02 -1.76560593e+00 -1.95777088e-01
-2.34803274e-01 -5.41702807e-01 3.16551626e-01 -1.12302518e+00
-1.35861778e+00 7.86726594e-01 9.44098771e-01 -8.27491656e-02
1.56838858e+00 -2.94735372e-01 1.11301768e+00 4.99172002e-01
-1.15125045e-01 -1.33727491e+00 2.74957389e-01 7.77999640e-01
1.04281437e+00 -1.46834552e+00 1.94264770e-01 -3.72310281e-01
-6.85078919e-01 9.77607250e-01 7.56511211e-01 -2.66156852e-01
7.78182805e-01 2.61427462e-01 1.93562716e-01 -2.03578442e-01
-2.30212092e-01 -3.50631654e-01 4.60805357e-01 5.88782728e-01
6.59224316e-02 1.04328685e-01 5.47775254e-02 4.92112726e-01
1.05009824e-01 -4.10763532e-01 2.08975628e-01 1.12664616e+00
-7.74833858e-01 -6.70210540e-01 -3.41349840e-01 2.97026753e-01
1.59786746e-01 -3.29017848e-01 -7.68111467e-01 6.68546259e-01
3.62813681e-01 6.80264890e-01 -1.20289057e-01 -4.91537482e-01
2.67368734e-01 -3.58566999e-01 1.31092340e-01 -2.94429362e-01
-1.77894950e-01 -1.75478190e-01 -2.14651957e-01 -7.93720603e-01
-3.67109537e-01 -2.51037389e-01 -1.50089765e+00 -9.35411155e-02
-2.69255370e-01 -2.36692384e-01 7.09842801e-01 1.12321186e+00
2.50490397e-01 1.22869563e+00 9.35942948e-01 -1.17833114e+00
-1.27827171e-02 -5.73686004e-01 -5.03638506e-01 2.50755008e-02
8.01898956e-01 -9.38143790e-01 -3.10661852e-01 -3.30661923e-01] | [10.856163024902344, -4.115509986877441] |
f11ab35c-bb67-4bdc-8b25-354aae0bd88d | multi-variate-probabilistic-time-series | 2002.06103 | null | https://arxiv.org/abs/2002.06103v3 | https://arxiv.org/pdf/2002.06103v3.pdf | Multivariate Probabilistic Time Series Forecasting via Conditioned Normalizing Flows | Time series forecasting is often fundamental to scientific and engineering problems and enables decision making. With ever increasing data set sizes, a trivial solution to scale up predictions is to assume independence between interacting time series. However, modeling statistical dependencies can improve accuracy and enable analysis of interaction effects. Deep learning methods are well suited for this problem, but multivariate models often assume a simple parametric distribution and do not scale to high dimensions. In this work we model the multivariate temporal dynamics of time series via an autoregressive deep learning model, where the data distribution is represented by a conditioned normalizing flow. This combination retains the power of autoregressive models, such as good performance in extrapolation into the future, with the flexibility of flows as a general purpose high-dimensional distribution model, while remaining computationally tractable. We show that it improves over the state-of-the-art for standard metrics on many real-world data sets with several thousand interacting time-series. | ['Abdul-Saboor Sheikh', 'Kashif Rasul', 'Urs Bergmann', 'Roland Vollgraf', 'Ingmar Schuster'] | 2020-02-14 | multivariate-probabilistic-time-series | https://openreview.net/forum?id=WiGQBFuVRv | https://openreview.net/pdf?id=WiGQBFuVRv | iclr-2021-1 | ['probabilistic-deep-learning', 'probabilistic-time-series-forecasting'] | ['computer-vision', 'time-series'] | [-1.77152202e-01 -5.07683337e-01 -2.23631430e-02 -3.57692689e-01
-3.45430344e-01 -7.14650989e-01 1.02400219e+00 1.47621468e-01
-2.13255003e-01 6.95421696e-01 3.38311315e-01 -6.00911975e-01
-4.28834081e-01 -7.87473798e-01 -8.13735783e-01 -9.85451758e-01
-8.32366347e-01 4.94723082e-01 -7.67756477e-02 -2.97916323e-01
3.56050618e-02 6.44002020e-01 -1.48504853e+00 -5.61828986e-02
4.69672501e-01 1.08039403e+00 -4.31869596e-01 7.71684051e-01
-1.65721297e-01 8.85201097e-01 -6.12589121e-01 -8.28659758e-02
1.10129789e-01 -8.84490386e-02 -2.65405506e-01 -4.78349328e-01
-6.61270367e-03 -3.86778355e-01 -4.77901727e-01 3.10221493e-01
3.42061073e-01 1.98667735e-01 8.86477113e-01 -1.42843187e+00
-4.98017579e-01 5.79941511e-01 -5.18690944e-01 3.27582389e-01
-1.98094040e-01 9.44713056e-02 8.99488807e-01 -2.84885079e-01
1.35165468e-01 1.21361911e+00 8.39988530e-01 -9.21658240e-03
-1.55291915e+00 -5.54626286e-01 3.03288877e-01 1.18171975e-01
-1.06961596e+00 -1.06954418e-01 8.87441456e-01 -7.29267955e-01
1.22582138e+00 1.76240399e-01 4.61469203e-01 1.26220882e+00
8.17794859e-01 6.75801635e-01 7.90152013e-01 -8.69897455e-02
3.87832195e-01 -4.83876914e-01 2.61131406e-01 -1.97715387e-01
-1.60656333e-01 3.62159073e-01 -2.79015660e-01 -2.60761082e-01
8.03073347e-01 4.60184991e-01 1.58489570e-01 -3.02136660e-01
-1.21534598e+00 7.52427697e-01 6.77755550e-02 6.45318151e-01
-6.53833628e-01 4.79837060e-01 6.06128812e-01 6.54501677e-01
9.45829153e-01 2.62841851e-01 -9.60239112e-01 -4.33970898e-01
-9.07349288e-01 5.24145424e-01 9.24975574e-01 4.12233472e-01
3.57127219e-01 3.38885307e-01 -9.86597985e-02 3.97317737e-01
3.28267142e-02 5.94689131e-01 3.91984820e-01 -9.38841820e-01
2.90467571e-02 1.04145624e-01 2.38022804e-01 -9.72551465e-01
-7.89184868e-01 -6.51816487e-01 -1.25174916e+00 2.76994318e-01
7.20472276e-01 -2.71289468e-01 -7.76718974e-01 1.82895708e+00
1.49075508e-01 4.36168313e-01 -1.09473407e-01 4.24281776e-01
1.39312357e-01 9.15801048e-01 1.55503362e-01 -3.98082584e-01
8.21783781e-01 -3.49834591e-01 -6.57880962e-01 1.24685779e-01
7.82532871e-01 -5.62792063e-01 7.63776481e-01 5.95548809e-01
-6.32174134e-01 -4.33107346e-01 -6.73235714e-01 1.22964039e-01
-4.90744233e-01 -3.78166407e-01 9.87684131e-01 5.26903093e-01
-1.09093177e+00 1.01983857e+00 -1.18391204e+00 -9.22967568e-02
1.03593193e-01 5.06691575e-01 -1.96504474e-01 2.89598644e-01
-1.23064375e+00 7.92794645e-01 -6.01404868e-02 -1.80012956e-01
-7.81820118e-01 -1.28420413e+00 -5.05288124e-01 3.08165997e-01
-1.93615749e-01 -5.50083935e-01 1.32915735e+00 -7.61191308e-01
-1.60496151e+00 1.93030745e-01 -1.25071704e-01 -7.66928613e-01
4.98183578e-01 -3.54575902e-01 -6.73921287e-01 -4.76197094e-01
-3.16268384e-01 4.58522290e-02 9.04315472e-01 -6.23532295e-01
-2.99808860e-01 -3.26444000e-01 1.25644524e-02 -4.35711205e-01
-4.84862477e-01 1.93977170e-02 2.51849651e-01 -7.16673195e-01
-1.98976561e-01 -8.64229918e-01 -5.59943557e-01 -1.67416468e-01
9.49324444e-02 -3.62729967e-01 9.29722905e-01 -7.10996568e-01
1.34881890e+00 -2.16436982e+00 1.23580717e-01 2.70596575e-02
2.92561889e-01 -1.73169106e-01 -1.53029338e-01 8.22268784e-01
-5.37237048e-01 2.32282560e-03 -1.25090167e-01 -4.68507826e-01
1.15257703e-01 3.60008687e-01 -7.93399751e-01 7.43946612e-01
1.84932321e-01 7.41869748e-01 -7.83243537e-01 1.42090812e-01
4.97811943e-01 7.44178176e-01 -3.59515220e-01 1.15042590e-01
-3.54078948e-01 6.52534902e-01 -2.16442943e-01 5.65866679e-02
5.43331385e-01 -4.71518695e-01 4.30671684e-02 8.72642994e-02
-2.88736284e-01 2.80050635e-01 -9.47625935e-01 1.20207775e+00
-6.26015067e-01 8.56112063e-01 -2.50169665e-01 -1.37050855e+00
9.75785375e-01 3.27245027e-01 1.14551139e+00 -6.23170257e-01
1.93644390e-01 1.20890796e-01 3.06360841e-01 -2.11373419e-01
1.87416524e-01 -2.46228993e-01 -1.48733392e-01 4.08069551e-01
9.06704366e-02 1.16531685e-01 9.01969373e-02 -2.37996370e-01
1.13232160e+00 7.39268363e-02 1.41889751e-01 -4.11341995e-01
1.03983060e-01 -4.14994597e-01 3.68513793e-01 5.74243605e-01
2.41315916e-01 1.07222371e-01 8.26266050e-01 -8.94337535e-01
-1.25061369e+00 -1.05158281e+00 -2.23771483e-01 1.18655515e+00
-4.94242460e-01 -3.34677398e-01 -9.53708217e-02 -1.86188519e-01
2.17779666e-01 1.02248418e+00 -8.07650626e-01 -9.47653949e-02
-5.99433184e-01 -9.53023076e-01 2.50153154e-01 7.50725389e-01
-1.83484077e-01 -6.88107908e-01 -3.77770215e-01 5.83353758e-01
1.91224575e-01 -7.03043401e-01 1.58446394e-02 4.15697515e-01
-8.78790379e-01 -6.33364439e-01 -7.12321818e-01 7.40427524e-03
-2.98108645e-02 -1.25035211e-01 1.23367488e+00 -5.55619299e-01
-7.91737586e-02 4.29961920e-01 6.25176076e-03 -6.90417945e-01
-3.40917379e-01 -1.46886185e-01 3.10417831e-01 5.02611957e-02
3.35868865e-01 -1.29089165e+00 -5.98639011e-01 1.10012859e-01
-1.03356147e+00 -3.72574061e-01 2.81578362e-01 7.30098605e-01
3.51745129e-01 2.53865868e-01 6.78930402e-01 -4.17943001e-01
4.52254325e-01 -9.09018993e-01 -9.13156450e-01 -5.20224236e-02
-4.30066854e-01 1.76022455e-01 1.14137638e+00 -8.03260505e-01
-6.43653035e-01 -2.36664057e-01 -6.25992864e-02 -6.00716412e-01
-1.12555467e-01 7.33657360e-01 2.19120741e-01 1.98752701e-01
5.49192727e-01 2.67037690e-01 2.51449570e-02 -6.34882867e-01
4.93424922e-01 2.83569187e-01 3.93676639e-01 -5.15994489e-01
4.72738504e-01 5.34329295e-01 5.63798368e-01 -9.33319926e-01
-4.93253082e-01 -3.35405082e-01 -6.85848653e-01 -2.44155765e-01
5.22564590e-01 -7.51151562e-01 -9.14616644e-01 5.84948778e-01
-1.07090735e+00 -5.03091633e-01 -1.83521062e-01 5.71243525e-01
-7.56549835e-01 8.37938264e-02 -5.94016731e-01 -1.02946413e+00
3.58587271e-03 -6.58450365e-01 1.10093057e+00 -2.06071898e-01
-3.88545513e-01 -1.59058273e+00 4.49370027e-01 -5.84169328e-01
7.38344431e-01 7.37083554e-01 1.05504513e+00 -8.01636815e-01
-2.64079183e-01 -5.39560199e-01 2.93011993e-01 2.56638914e-01
7.99754485e-02 3.63834739e-01 -8.60260904e-01 -1.27362385e-01
1.63583964e-01 3.19325715e-01 7.26891577e-01 8.37502360e-01
1.45715868e+00 -1.95498303e-01 -3.18320096e-01 5.01237094e-01
1.04972112e+00 2.90980995e-01 5.35014033e-01 -5.29008033e-03
6.82270586e-01 7.22882748e-01 1.11497872e-01 7.52749383e-01
3.42612594e-01 4.97923732e-01 3.96341085e-01 8.52007940e-02
4.81406301e-01 -2.22521741e-02 3.50927025e-01 8.58316183e-01
-8.23740810e-02 -5.39397955e-01 -1.25706720e+00 5.18458247e-01
-1.95494103e+00 -1.14096653e+00 -5.01822114e-01 2.36101127e+00
3.86292517e-01 2.59056836e-01 3.08520108e-01 2.34424651e-01
2.46984020e-01 2.58322001e-01 -7.53743708e-01 -1.30820170e-01
-3.36633682e-01 1.25769615e-01 5.37269771e-01 2.97356635e-01
-1.31373751e+00 3.29334348e-01 7.50434494e+00 5.54800570e-01
-1.53731763e+00 7.90552236e-03 7.99410105e-01 -1.69448346e-01
-3.81110221e-01 -1.68125778e-01 -5.94691694e-01 6.36547327e-01
1.68021917e+00 -5.41992962e-01 3.73809010e-01 6.86802685e-01
4.29509521e-01 3.27596724e-01 -1.39602709e+00 1.03021610e+00
-4.78454053e-01 -1.26676595e+00 -2.11705685e-01 3.68209839e-01
6.27085328e-01 3.82565260e-01 3.20759833e-01 4.69377816e-01
5.09612799e-01 -1.11106312e+00 3.81001711e-01 8.63844335e-01
4.07213598e-01 -7.43628860e-01 5.49199224e-01 4.88452107e-01
-9.17065978e-01 -2.73313642e-01 -1.47688121e-01 -6.15130782e-01
3.49575579e-01 9.71646726e-01 -3.00869703e-01 4.16439801e-01
6.91376984e-01 1.12900138e+00 -1.63212284e-01 1.05149245e+00
3.38645011e-01 1.02536225e+00 -9.12676811e-01 4.49135303e-02
1.47859007e-01 -1.20165043e-01 5.78090310e-01 9.20727491e-01
6.30424976e-01 -1.85904413e-01 -6.70181878e-04 4.89035249e-01
3.66408646e-01 -5.08958250e-02 -7.24367380e-01 -4.55627918e-01
1.51344180e-01 9.13329124e-01 -5.65551102e-01 -2.61040598e-01
-6.29183173e-01 4.01146859e-01 1.15059599e-01 4.61418033e-01
-8.69138002e-01 9.05950591e-02 9.59056497e-01 1.73182875e-01
3.49662155e-01 -6.67436302e-01 -4.02653337e-01 -1.07587349e+00
-9.16304961e-02 -6.59313440e-01 4.62132335e-01 -4.61203992e-01
-1.80669808e+00 3.44810188e-01 9.77235362e-02 -1.50197172e+00
-7.85887420e-01 -8.18298280e-01 -6.77930176e-01 8.63266885e-01
-1.06158543e+00 -1.12478852e+00 2.08320513e-01 3.77288014e-01
2.86130637e-01 5.53389033e-03 9.06498969e-01 2.22722471e-01
-2.36718327e-01 1.78653225e-01 7.47149646e-01 -3.79634380e-01
6.15605474e-01 -1.30320501e+00 6.53226614e-01 7.28919864e-01
-1.45843886e-02 5.38733959e-01 1.20424938e+00 -4.63178426e-01
-1.45472598e+00 -9.78490591e-01 7.65265644e-01 -5.64529181e-01
1.29633403e+00 -4.51678276e-01 -1.17737079e+00 6.29524291e-01
3.70992199e-02 -1.58117637e-02 9.50943232e-01 5.30935347e-01
-3.98116052e-01 -1.49387017e-01 -5.94744086e-01 5.02968907e-01
6.14384472e-01 -4.15969551e-01 -2.46360615e-01 3.89042258e-01
6.88055277e-01 -6.41432703e-02 -1.28628969e+00 3.79877955e-01
7.66791344e-01 -7.85140932e-01 9.21238720e-01 -8.15655231e-01
5.64281583e-01 -1.59896418e-01 -2.54508346e-01 -1.38066697e+00
-6.15161777e-01 -8.95179152e-01 -5.49875677e-01 1.03646278e+00
2.16526136e-01 -7.03402936e-01 3.89764965e-01 9.08160686e-01
1.16504252e-01 -5.46501696e-01 -9.57288504e-01 -1.07142162e+00
5.73245883e-01 -8.19390059e-01 9.29361463e-01 1.03081608e+00
-1.49297804e-01 1.89298227e-01 -6.44331634e-01 1.56826377e-01
5.72381020e-01 1.30415976e-01 8.43648791e-01 -1.66111588e+00
-5.85451186e-01 -8.58866453e-01 -5.23722231e-01 -1.00343919e+00
2.63103336e-01 -3.87284398e-01 -3.89939338e-01 -1.07417762e+00
-4.04271990e-01 -2.01916069e-01 -6.14208758e-01 2.27741465e-01
8.05334374e-02 -2.61253446e-01 -9.48847607e-02 1.03451833e-01
-2.41460621e-01 7.64663875e-01 7.97627509e-01 3.37692872e-02
-2.38802984e-01 3.05055112e-01 -2.65777677e-01 5.78872621e-01
8.51950109e-01 -2.39969715e-01 -4.83923227e-01 -3.43405873e-01
2.81733274e-01 3.97530138e-01 4.01495546e-01 -8.14464033e-01
1.06650718e-01 -3.10021222e-01 5.64224958e-01 -7.79189348e-01
2.63314873e-01 -9.35593247e-01 5.63565850e-01 1.82355672e-01
-3.28726947e-01 3.26356769e-01 4.37403619e-01 6.79774344e-01
-1.89240336e-01 5.60874879e-01 4.02533591e-01 3.10211033e-01
-5.23402929e-01 5.24169266e-01 -5.30324340e-01 -3.53629589e-01
8.99695873e-01 2.20405966e-01 -2.33819067e-01 -8.25409114e-01
-7.83666492e-01 1.50843412e-01 1.73288375e-01 7.25391567e-01
-9.01945308e-02 -1.23783219e+00 -6.11281872e-01 1.17067419e-01
-1.20546147e-01 -3.60903561e-01 5.13003230e-01 9.71627235e-01
-2.51168218e-02 4.31884110e-01 -1.73908472e-01 -8.09344828e-01
-7.42616534e-01 9.03777182e-01 3.33221227e-01 -4.38629448e-01
-4.93342519e-01 4.30670798e-01 4.97426778e-01 -2.50201613e-01
2.54232250e-02 -5.19041121e-01 -2.67398089e-01 1.26808867e-01
8.42010200e-01 3.73097956e-01 5.08772582e-02 -4.44941849e-01
-2.96443224e-01 2.45923534e-01 1.29857827e-02 2.06182096e-02
1.82430065e+00 3.75479041e-03 -3.16686749e-01 1.17363739e+00
1.35136223e+00 -3.54134798e-01 -1.46339893e+00 -1.25280768e-02
-5.86479297e-03 -8.80010575e-02 1.61111876e-01 -5.46894729e-01
-7.67559350e-01 1.06917834e+00 4.78640050e-01 8.54350209e-01
1.06365359e+00 -2.20298201e-01 4.26123887e-01 2.56905347e-01
2.88712949e-01 -4.77006555e-01 -3.56498927e-01 7.34625161e-01
8.45450640e-01 -9.39803421e-01 -2.96962142e-01 2.50185039e-02
-1.33316651e-01 1.35303140e+00 2.26827323e-01 -4.32719946e-01
1.24996436e+00 6.31302714e-01 -1.15999162e-01 1.09289728e-01
-1.16469097e+00 1.34866521e-01 2.87021875e-01 5.78477800e-01
8.16452622e-01 2.10729092e-01 -1.10630423e-01 6.46089196e-01
-3.24561536e-01 -9.37174074e-03 3.65996838e-01 5.72779298e-01
5.32455295e-02 -1.08775103e+00 -2.48219371e-01 7.16140270e-01
-6.11343205e-01 1.96079705e-02 1.97094560e-01 6.80643022e-01
-3.73972803e-01 8.99666429e-01 4.84265447e-01 -3.17334354e-01
1.56705394e-01 1.43258229e-01 2.27593616e-01 8.43725875e-02
-4.16127086e-01 3.07530046e-01 -1.61974192e-01 -6.14713848e-01
-3.21613997e-01 -8.74331415e-01 -7.62554646e-01 -7.22610235e-01
1.16383851e-01 -1.54189378e-01 6.27068043e-01 1.14266634e+00
5.37436008e-01 7.34601080e-01 7.06821859e-01 -1.14658523e+00
-6.13472939e-01 -9.69520032e-01 -6.10628843e-01 2.64153242e-01
7.96355724e-01 -7.29366064e-01 -4.94866341e-01 1.71323135e-01] | [7.000034332275391, 3.1934292316436768] |
0a14b92b-4db8-48f8-b5e2-4037a87a779b | quantum-artificial-life-in-an-ibm-quantum | 1711.09442 | null | http://arxiv.org/abs/1711.09442v2 | http://arxiv.org/pdf/1711.09442v2.pdf | Quantum Artificial Life in an IBM Quantum Computer | We present the first experimental realization of a quantum artificial life
algorithm in a quantum computer. The quantum biomimetic protocol encodes
tailored quantum behaviors belonging to living systems, namely,
self-replication, mutation, interaction between individuals, and death, into
the cloud quantum computer IBM ibmqx4. In this experiment, entanglement spreads
throughout generations of individuals, where genuine quantum information
features are inherited through genealogical networks. As a pioneering
proof-of-principle, experimental data fits the ideal model with accuracy.
Thereafter, these and other models of quantum artificial life, for which no
classical device may predict its quantum supremacy evolution, can be further
explored in novel generations of quantum computers. Quantum biomimetics,
quantum machine learning, and quantum artificial intelligence will move forward
hand in hand through more elaborate levels of quantum complexity. | ['U. Alvarez-Rodriguez', 'L. Lamata', 'E. Solano', 'M. Sanz'] | 2017-11-26 | null | null | null | null | ['artificial-life'] | ['miscellaneous'] | [ 2.07954824e-01 1.12377413e-01 3.25771719e-01 1.41615614e-01
2.23466188e-01 -5.19961178e-01 8.89621317e-01 -3.12419310e-02
-2.11471722e-01 1.05730939e+00 -4.42633152e-01 -2.70457864e-01
-5.97891770e-02 -1.40420973e+00 -3.91762614e-01 -1.01395249e+00
-5.81281722e-01 9.16894734e-01 5.81968240e-02 -7.35529542e-01
3.71057630e-01 1.74420848e-01 -1.46846426e+00 -1.55840412e-01
9.31742072e-01 2.12272346e-01 -4.66334850e-01 1.15191996e+00
3.16605121e-01 4.79429960e-01 -1.26145914e-01 -5.90483308e-01
3.27154070e-01 -1.17360115e+00 -6.39471531e-01 -3.61662567e-01
-5.07224619e-01 -1.46164685e-01 -9.45494056e-01 1.24785161e+00
2.26556167e-01 -2.16702849e-01 1.00942850e+00 -8.06468368e-01
-1.15566981e+00 7.32166171e-01 9.87953618e-02 -5.39786667e-02
4.05774862e-01 7.06359684e-01 1.10098779e+00 -3.29216532e-02
6.98037624e-01 9.51106608e-01 5.99298716e-01 8.75991285e-01
-1.40674734e+00 -4.72247124e-01 -1.59946620e+00 6.74641952e-02
-1.60855103e+00 -7.94264823e-02 1.36997983e-01 -2.46116176e-01
1.29027188e+00 7.11849704e-02 1.23802209e+00 6.19095325e-01
1.17701733e+00 5.66059165e-02 1.33955574e+00 -7.18020797e-01
6.53297007e-01 -3.39888126e-01 2.49561340e-01 1.12538803e+00
6.22498155e-01 1.29743934e+00 -6.97929263e-01 -4.79423016e-01
6.80358946e-01 -1.14539750e-01 1.57716736e-01 -5.96224628e-02
-1.31610334e+00 6.24212563e-01 3.81805360e-01 4.01887834e-01
-4.26567227e-01 8.08671355e-01 -5.52875288e-02 5.06145597e-01
-2.74490714e-01 6.79069221e-01 -2.40590557e-01 -4.24144715e-01
-4.59912002e-01 3.41184318e-01 9.74876165e-01 8.03963244e-01
1.25467479e+00 -5.02568364e-01 3.03611010e-01 -3.02289903e-01
2.74885744e-01 8.56296599e-01 4.01364267e-01 -1.19289601e+00
-6.66516364e-01 3.36422116e-01 3.46776098e-02 -2.32465297e-01
-6.25086725e-01 -1.20102018e-01 -1.01786554e+00 -8.06723535e-02
1.63240165e-01 -2.67887831e-01 -6.06873274e-01 1.32901108e+00
2.05677822e-01 3.68388981e-01 5.84771991e-01 5.12567341e-01
2.70646334e-01 5.25679886e-01 -3.25467974e-01 -3.75113159e-01
1.41560626e+00 -4.99593198e-01 -6.28285468e-01 4.57066864e-01
9.96112823e-01 -3.36892456e-01 3.82206112e-01 2.44332656e-01
-1.06817532e+00 -5.88890314e-02 -1.25200534e+00 3.89554113e-01
-4.13365871e-01 -9.81407404e-01 1.51298976e+00 1.38584757e+00
-1.02130377e+00 1.47265065e+00 -1.12675762e+00 -6.91209435e-01
9.07963291e-02 6.27792358e-01 -2.01840207e-01 1.91698164e-01
-1.63502324e+00 7.21110821e-01 4.94269669e-01 -2.20895573e-01
-8.90434325e-01 -7.60719832e-03 -2.14667946e-01 -1.48682848e-01
-2.31521025e-01 -1.74918842e+00 8.96748781e-01 -3.75081390e-01
-1.94116950e+00 9.73043799e-01 4.49652858e-02 -6.93426728e-01
-1.49869412e-01 6.01065040e-01 -5.01256704e-01 6.38210326e-02
-2.48154894e-01 3.77685487e-01 5.37270308e-01 -5.73982596e-01
-1.54913872e-01 -6.58180296e-01 -1.56290486e-01 -1.28818229e-01
-2.45800130e-02 -2.54574209e-01 3.69749844e-01 2.14599147e-01
1.43667579e-01 -1.46108377e+00 -4.44180489e-01 -6.24823153e-01
-1.38579607e-01 -2.99128652e-01 -1.12295233e-01 3.12873542e-01
1.11089242e+00 -1.87685347e+00 3.79005194e-01 1.08830675e-01
3.72489661e-01 -1.73382133e-01 -4.72069867e-02 1.11998749e+00
3.72811347e-01 3.73916835e-01 -4.23181802e-01 4.74382162e-01
2.55906016e-01 4.00335550e-01 2.06650630e-01 6.95670187e-01
2.64974147e-01 1.52258897e+00 -1.09024298e+00 -3.81006986e-01
-2.50403941e-01 1.80853710e-01 -9.08170760e-01 -1.88203618e-01
-9.24450979e-02 6.87927127e-01 -5.66708446e-01 5.77155292e-01
4.00929958e-01 -7.48557270e-01 -1.83691494e-02 5.20809710e-01
-2.31812730e-01 6.33707121e-02 -4.75790322e-01 1.73288536e+00
1.40228823e-01 2.29335710e-01 -4.13002878e-01 -5.66191316e-01
4.20550078e-01 1.40255049e-01 3.42314094e-01 -8.26382160e-01
2.35031545e-01 4.63104308e-01 9.06763077e-01 -3.60630333e-01
6.77018940e-01 -5.69343448e-01 -3.21133226e-01 1.02970338e+00
1.94628581e-01 -8.41971636e-01 4.15285200e-01 2.53020048e-01
1.59249747e+00 1.30858004e-01 4.56755161e-01 -2.25575402e-01
3.43690485e-01 1.87989965e-01 1.71003237e-01 1.18967104e+00
-6.75517023e-01 3.04711193e-01 -1.13959890e-03 -5.60734630e-01
-1.57151210e+00 -1.27922857e+00 -4.14615899e-01 5.23971677e-01
5.89612067e-01 -8.37055385e-01 -1.16097724e+00 3.48011494e-01
-1.15993015e-01 5.25199175e-01 -4.73631650e-01 -5.97176671e-01
-3.06696683e-01 -1.45675266e+00 8.97786260e-01 -5.21995425e-01
4.09082115e-01 -1.13063967e+00 -6.83725059e-01 4.11744267e-01
3.47451836e-01 -7.84770787e-01 1.72563821e-01 2.89350539e-01
-1.00353849e+00 -6.98829889e-01 -3.65739495e-01 -1.44511268e-01
3.98852348e-01 -2.71661848e-01 1.14015865e+00 5.66887677e-01
-8.05826426e-01 9.75544900e-02 -3.54853302e-01 1.07177518e-01
-1.18629277e+00 -9.73212421e-02 6.37358904e-01 -6.24378800e-01
6.96432710e-01 -8.28101516e-01 -7.24969447e-01 -2.27782875e-01
-8.80782604e-01 -9.42003652e-02 7.14756370e-01 1.30027616e+00
1.64955109e-01 4.68892276e-01 2.12902669e-02 -5.56991637e-01
5.99076152e-01 -3.26467097e-01 -4.48288023e-01 3.53351176e-01
-5.33332765e-01 4.19385135e-01 7.18418002e-01 1.77869871e-01
-5.65043509e-01 -2.19509766e-01 1.48750469e-01 5.13440311e-01
-7.42981136e-02 3.28887343e-01 5.01215577e-01 -3.58186185e-01
8.95659447e-01 6.23928189e-01 2.18540579e-01 3.88233304e-01
7.17181861e-01 8.60953212e-01 2.48778462e-01 -5.37836492e-01
8.44065368e-01 5.67336202e-01 7.68318713e-01 -1.17891276e+00
-1.51738390e-01 2.41600409e-01 -8.79738927e-01 9.18967351e-02
9.17126656e-01 -5.61596096e-01 -1.34475434e+00 6.63704693e-01
-1.07297635e+00 -2.05777347e-01 -4.92633045e-01 4.24923331e-01
-9.59362626e-01 6.00488186e-01 -1.12202144e+00 -1.20325923e+00
-1.95866302e-01 -6.47467792e-01 9.24597919e-01 6.63207889e-01
-2.03261167e-01 -8.13323975e-01 6.71807945e-01 1.74791485e-01
3.29506963e-01 -1.15100525e-01 9.68082666e-01 -2.81673640e-01
-1.16911411e+00 -3.19355160e-01 1.78714946e-01 -2.67608076e-01
-5.71476668e-02 5.20247042e-01 -4.54821408e-01 -3.27727437e-01
-2.67194569e-01 -3.92437190e-01 7.46090949e-01 -1.59353344e-03
2.35950306e-01 -1.95172764e-02 -3.43742073e-01 5.35780430e-01
1.20508301e+00 4.06400472e-01 7.53210247e-01 1.41785577e-01
8.11488852e-02 3.40719253e-01 -5.71249425e-02 4.86841112e-01
3.23147297e-01 3.37161094e-01 2.33565316e-01 7.55842984e-01
2.48603076e-01 -1.13388687e-01 4.88636672e-01 1.65513253e+00
-7.27076769e-01 -8.12478960e-02 -1.10485196e+00 -7.87895173e-02
-1.41940916e+00 -1.40370154e+00 -2.37665206e-01 2.40931296e+00
5.94532847e-01 -1.02081122e-02 4.09213454e-03 -1.88332662e-01
6.70016468e-01 -4.79611665e-01 -4.87458199e-01 -5.74106932e-01
-4.89558727e-01 5.16968429e-01 6.54256821e-01 3.31792057e-01
-7.78479815e-01 1.16868985e+00 7.45677805e+00 5.81287980e-01
-5.87400436e-01 2.63480365e-01 1.94547340e-01 4.81304467e-01
-5.59045792e-01 8.16994250e-01 -5.33776045e-01 5.67245603e-01
1.81319153e+00 -5.89224219e-01 1.43071067e+00 -1.88877255e-01
-2.54801840e-01 -2.51500160e-01 -1.07299161e+00 8.81603360e-01
-5.62534213e-01 -1.46586382e+00 -1.87021285e-01 5.48251152e-01
1.03117371e+00 3.69237840e-01 -1.61746427e-01 2.74915069e-01
5.11369348e-01 -1.01738489e+00 3.43509287e-01 8.09622169e-01
5.87278962e-01 -6.27208531e-01 5.45207620e-01 6.47637665e-01
-4.74359542e-01 -2.51981676e-01 -7.02602804e-01 -7.81179726e-01
3.45026791e-01 2.92117655e-01 -3.55239302e-01 5.89293778e-01
3.59892190e-01 5.23458838e-01 -3.99033576e-01 6.56469703e-01
-9.30587016e-03 3.84818077e-01 -4.67345357e-01 -8.33284318e-01
6.81806803e-02 -1.00001454e+00 5.37329435e-01 5.84628105e-01
5.34995794e-01 7.32609093e-01 -6.91355407e-01 1.35489094e+00
1.72759429e-01 -3.92111540e-01 -8.92356455e-01 -1.15107894e+00
3.69120151e-01 9.29648101e-01 -8.71621311e-01 -4.48471725e-01
7.17659593e-02 1.27444065e+00 -1.10003687e-01 -6.82459306e-03
-5.80426097e-01 -5.96644461e-01 5.64666748e-01 -2.01187551e-01
-1.46863172e-02 -4.57206994e-01 -4.08558786e-01 -1.56376803e+00
-8.11558366e-01 -5.65443456e-01 -3.03270787e-01 -4.73618716e-01
-1.36980927e+00 4.92435604e-01 -7.35034883e-01 -8.63403678e-01
-6.18567169e-01 -5.78775465e-01 -5.06083548e-01 5.36403298e-01
-5.91645300e-01 -5.39302289e-01 2.87007838e-01 1.29656106e-01
-7.00568736e-01 -3.79442245e-01 1.46070361e+00 -1.88241944e-01
-5.93191266e-01 2.21918702e-01 8.49167407e-01 -2.56000966e-01
-6.70841187e-02 -1.13590777e+00 9.40242708e-01 6.36899889e-01
4.11702484e-01 1.06269634e+00 9.55270469e-01 -7.17488050e-01
-2.27050686e+00 -2.94675529e-01 7.02297628e-01 -5.27563691e-01
1.18477178e+00 -4.37116921e-01 -3.44065338e-01 2.14929476e-01
3.94602269e-01 -1.07615985e-01 7.88959086e-01 -2.04992928e-02
-3.63088958e-02 3.63124579e-01 -1.01066411e+00 7.82517314e-01
1.32313156e+00 -1.16766012e+00 -5.60864925e-01 6.50857270e-01
8.10402691e-01 1.89763695e-01 -1.08580375e+00 2.81664163e-01
1.01959169e+00 -1.22312641e+00 6.63193583e-01 -6.91308320e-01
4.75094646e-01 -3.53359848e-01 -1.83108356e-03 -1.12892032e+00
-5.02695262e-01 -1.43457460e+00 8.68232027e-02 4.37391430e-01
3.02351475e-01 -9.95289743e-01 7.05110013e-01 3.40517879e-01
3.16774487e-01 -1.03791416e-01 -1.23826826e+00 -9.57198381e-01
6.31057382e-01 -6.85363933e-02 7.26971686e-01 7.26109445e-01
4.92462307e-01 4.55016941e-01 -1.86467856e-01 -1.85847238e-01
1.16627705e+00 2.02320948e-01 6.91911817e-01 -1.25976110e+00
-7.32810318e-01 -5.35936952e-01 -9.48924124e-01 -7.58693516e-01
7.92150758e-03 -9.98604894e-01 -4.70778421e-02 -7.20627785e-01
5.43890595e-01 -2.22186655e-01 -2.59539604e-01 -2.84385324e-01
3.40204313e-02 5.24463832e-01 7.88702965e-02 4.00518805e-01
-7.91154683e-01 7.55305827e-01 1.28259730e+00 -1.05388492e-01
4.84880339e-03 -2.85823166e-01 -1.36017427e-01 2.97199041e-01
5.21488965e-01 -6.65246785e-01 2.88855463e-01 8.37941542e-02
8.07449460e-01 4.66419429e-01 1.87112659e-01 -1.09702718e+00
4.46419865e-01 -5.55056036e-02 -1.76588699e-01 -1.45095527e-01
2.90740430e-01 -1.95215493e-01 6.75919950e-01 1.32186079e+00
1.67883813e-01 -1.60201237e-01 -5.94173253e-01 5.33544242e-01
1.52523115e-01 -2.88804591e-01 7.34450758e-01 -3.83377641e-01
-3.03603530e-01 4.06146884e-01 -6.79585516e-01 -2.28505448e-01
1.19865406e+00 -3.55706811e-01 -7.93678999e-01 -1.70075566e-01
-7.15518773e-01 -1.27868503e-01 1.24726760e+00 -2.02342644e-01
2.91940272e-01 -1.03765881e+00 -7.49670088e-01 3.93556178e-01
8.58713463e-02 -7.89856970e-01 3.65482092e-01 8.23112249e-01
-1.24736941e+00 6.80120528e-01 -5.40297925e-01 -6.92625165e-01
-3.59827787e-01 7.35413551e-01 5.24328947e-01 8.92453082e-03
-1.94320098e-01 6.48281693e-01 -2.93082297e-02 -6.38491035e-01
-9.91681457e-01 2.67969728e-01 3.82322758e-01 -5.56613326e-01
3.43132019e-01 2.79447019e-01 -4.01234746e-01 -7.79290140e-01
-2.89413542e-01 7.78649449e-01 3.87089282e-01 -2.08933562e-01
1.21708286e+00 -2.88843066e-01 -1.05555415e+00 6.97542965e-01
7.78072298e-01 -3.32343549e-01 -5.21136940e-01 1.63695574e-01
1.09856598e-01 1.24828741e-02 -3.40415865e-01 -3.98133337e-01
-5.72091825e-02 9.71118808e-01 5.62024474e-01 8.96825790e-01
7.36003637e-01 3.39890271e-02 8.71014953e-01 1.02944839e+00
1.37756824e+00 -7.88768888e-01 -3.76962930e-01 7.38704979e-01
-1.38710052e-01 -1.05714107e+00 -3.33918594e-02 1.49810284e-01
1.15819789e-01 1.35760438e+00 -4.11209650e-02 -4.48009968e-01
8.77717018e-01 1.16929322e-01 -5.90639591e-01 -2.97146529e-01
-1.18969548e+00 -3.22933137e-01 -3.55995506e-01 3.51529360e-01
4.06271636e-01 7.05204904e-01 -8.42619121e-01 2.44199529e-01
-4.97763604e-01 2.45646477e-01 1.06712997e+00 9.25122797e-01
-5.86204886e-01 -1.38645744e+00 -2.38487035e-01 2.14134455e-01
-2.63754696e-01 -3.82301658e-01 -4.16393071e-01 2.93855697e-01
2.89755374e-01 7.41134763e-01 2.24625051e-01 -6.96165979e-01
-3.95112842e-01 3.18722008e-03 1.04412031e+00 -3.88451040e-01
-3.83787811e-01 -5.27972043e-01 -3.09950501e-01 -3.17083180e-01
-4.85868454e-01 -7.53122509e-01 -1.65007174e+00 -1.15429854e+00
-6.32025957e-01 5.22107899e-01 5.74167371e-01 9.70751047e-01
4.81951058e-01 2.13242546e-01 6.00995600e-01 -6.48485959e-01
-7.81350136e-01 -6.44158363e-01 -1.18904769e+00 1.36313155e-01
1.53640313e-02 -1.96433902e-01 -5.68247676e-01 -1.78130716e-01] | [5.604348182678223, 4.8661417961120605] |
4ab93f2b-4826-4909-be36-267e9ef8c7a4 | contrastive-boundary-learning-for-point-cloud | 2203.05272 | null | https://arxiv.org/abs/2203.05272v2 | https://arxiv.org/pdf/2203.05272v2.pdf | Contrastive Boundary Learning for Point Cloud Segmentation | Point cloud segmentation is fundamental in understanding 3D environments. However, current 3D point cloud segmentation methods usually perform poorly on scene boundaries, which degenerates the overall segmentation performance. In this paper, we focus on the segmentation of scene boundaries. Accordingly, we first explore metrics to evaluate the segmentation performance on scene boundaries. To address the unsatisfactory performance on boundaries, we then propose a novel contrastive boundary learning (CBL) framework for point cloud segmentation. Specifically, the proposed CBL enhances feature discrimination between points across boundaries by contrasting their representations with the assistance of scene contexts at multiple scales. By applying CBL on three different baseline methods, we experimentally show that CBL consistently improves different baselines and assists them to achieve compelling performance on boundaries, as well as the overall performance, eg in mIoU. The experimental results demonstrate the effectiveness of our method and the importance of boundaries for 3D point cloud segmentation. Code and model will be made publicly available at https://github.com/LiyaoTang/contrastBoundary. | ['DaCheng Tao', 'Baosheng Yu', 'Zhe Chen', 'Yibing Zhan', 'Liyao Tang'] | 2022-03-10 | null | http://openaccess.thecvf.com//content/CVPR2022/html/Tang_Contrastive_Boundary_Learning_for_Point_Cloud_Segmentation_CVPR_2022_paper.html | http://openaccess.thecvf.com//content/CVPR2022/papers/Tang_Contrastive_Boundary_Learning_for_Point_Cloud_Segmentation_CVPR_2022_paper.pdf | cvpr-2022-1 | ['point-cloud-segmentation'] | ['computer-vision'] | [ 6.49597570e-02 -2.92329788e-01 -1.80413142e-01 -2.80638367e-01
-8.17840993e-01 -7.99278140e-01 4.98340577e-01 1.27267420e-01
1.32548204e-02 -8.16276018e-03 -1.77826539e-01 -3.84856552e-01
2.40482152e-01 -6.18787348e-01 -5.95892012e-01 -2.79248506e-01
1.06106684e-01 2.98118085e-01 5.92566848e-01 -3.21903871e-03
5.20648718e-01 9.31001604e-01 -1.39045715e+00 7.53865317e-02
1.20202088e+00 1.01321566e+00 1.98786587e-01 4.54153240e-01
-5.21303773e-01 -6.77748322e-02 -5.21659970e-01 -2.69294828e-02
6.80349112e-01 -1.46027401e-01 -7.94651687e-01 2.09257886e-01
7.73647249e-01 -3.01959485e-01 7.46245636e-03 1.04688215e+00
2.20904246e-01 3.40002961e-02 8.37057292e-01 -1.43258190e+00
-4.41001892e-01 6.19079918e-02 -9.06668484e-01 2.46829018e-01
2.63618052e-01 2.08915308e-01 1.11283898e+00 -1.09955454e+00
4.47406143e-01 1.24200523e+00 7.86421895e-01 1.77523032e-01
-1.07681155e+00 -7.18047798e-01 6.08657420e-01 -2.02849716e-01
-1.51327252e+00 -1.80067897e-01 1.00428069e+00 -6.26782119e-01
7.34461665e-01 2.65258312e-01 5.48121691e-01 5.76055467e-01
-1.55546278e-01 1.07537377e+00 9.60231125e-01 -1.81748226e-01
8.50041956e-02 -2.41649032e-01 3.28418732e-01 4.38055873e-01
2.79896885e-01 -3.02708782e-02 -1.89904332e-01 4.48505357e-02
1.03660285e+00 -1.45579344e-02 -1.87795714e-01 -6.32636189e-01
-1.10488033e+00 5.53291917e-01 6.53444111e-01 2.14143604e-01
-9.16065201e-02 1.08892828e-01 9.91475657e-02 -7.04799667e-02
7.39637256e-01 3.27343553e-01 -3.71886641e-01 5.20956367e-02
-1.08440590e+00 3.45937431e-01 4.53182012e-01 1.26902747e+00
8.32072616e-01 -3.41774166e-01 -1.91571787e-01 9.07438874e-01
3.07104796e-01 6.11520588e-01 -2.83327520e-01 -1.19381809e+00
4.05648440e-01 8.65428567e-01 6.86017871e-02 -9.31705236e-01
-4.53800261e-01 -1.59914553e-01 -4.24209446e-01 2.43461594e-01
3.91346812e-01 3.79012898e-02 -1.27407086e+00 1.22565091e+00
5.76104641e-01 4.76650029e-01 -2.91761458e-01 8.67258012e-01
1.17709386e+00 5.72370887e-01 9.63125229e-02 4.24746722e-01
1.08727932e+00 -9.41297352e-01 -2.27463022e-01 -3.16270798e-01
6.31824911e-01 -8.61587226e-01 1.30972075e+00 3.60612534e-02
-1.06770790e+00 -6.13584042e-01 -8.30402315e-01 -2.77148008e-01
-1.93236977e-01 8.30814987e-03 6.89298093e-01 4.10829067e-01
-1.06205976e+00 5.10709345e-01 -1.00615227e+00 -2.84929842e-01
7.31582761e-01 2.39921734e-01 7.42786601e-02 1.61852315e-02
-6.82953179e-01 4.54139709e-01 1.64886504e-01 -8.90345499e-02
-4.48119760e-01 -9.82169271e-01 -8.76250803e-01 -4.08470184e-02
2.81777352e-01 -6.77047014e-01 1.29573202e+00 -4.06931907e-01
-1.08443713e+00 1.09290195e+00 -2.67914295e-01 -8.41705427e-02
6.55066013e-01 -4.55867708e-01 6.64369762e-02 2.37031147e-01
4.11462247e-01 1.17984939e+00 4.05300200e-01 -1.75306010e+00
-8.77793610e-01 -3.13627124e-01 2.81104315e-02 3.07178020e-01
2.41348147e-01 -2.84496784e-01 -1.10704553e+00 -5.28328300e-01
6.95047319e-01 -8.37147534e-01 -3.15117240e-01 -1.10019641e-02
-5.27172446e-01 -4.01548922e-01 9.90644515e-01 -3.46224159e-01
9.56705511e-01 -2.38225436e+00 -2.91388929e-01 1.67794257e-01
1.48127437e-01 4.30400297e-02 -8.71462971e-02 8.65823478e-02
1.59067482e-01 5.90362310e-01 -6.87513947e-01 -4.33845848e-01
6.47171140e-02 -4.56369221e-02 -1.85562998e-01 3.49522740e-01
5.24932206e-01 1.05132043e+00 -7.00322568e-01 -6.70188844e-01
5.03600299e-01 4.16970998e-01 -7.08631396e-01 7.39266202e-02
-2.76975632e-01 7.21517026e-01 -6.12658620e-01 1.03072000e+00
1.14806294e+00 -4.27421540e-01 -4.19863433e-01 -1.87544152e-02
-2.36841604e-01 1.44733459e-01 -1.11799204e+00 1.78999400e+00
-1.50856823e-01 4.21027482e-01 1.51540950e-01 -5.35164297e-01
9.78597820e-01 -1.61676735e-01 7.02032566e-01 -5.74243307e-01
2.62582656e-02 2.84618586e-01 -2.41748020e-01 -1.43232584e-01
5.68010926e-01 1.15729123e-01 -7.82911032e-02 6.17366470e-02
-4.50819910e-01 -8.76352549e-01 6.13746829e-02 1.11832529e-01
7.86967516e-01 3.47455084e-01 -4.05093543e-02 -2.41322860e-01
3.72764230e-01 3.64883184e-01 6.23974800e-01 7.88769782e-01
-5.63112080e-01 1.06532347e+00 2.97979981e-01 -4.86629605e-02
-8.70337009e-01 -1.25354028e+00 -5.26514530e-01 6.69175804e-01
1.07719231e+00 -1.77387059e-01 -8.13045025e-01 -7.04022229e-01
4.22295183e-01 6.87985122e-01 -2.31366754e-01 1.87221676e-01
-6.09272480e-01 -4.01706636e-01 3.43568414e-01 6.16787314e-01
8.02536011e-01 -7.84066081e-01 -4.51436311e-01 -2.40288660e-01
-1.57805979e-01 -1.35004234e+00 -5.46995640e-01 -1.98003367e-01
-1.22027922e+00 -9.26078379e-01 -7.45648384e-01 -7.55013645e-01
4.82059836e-01 7.50992239e-01 1.26042891e+00 1.98465899e-01
-1.66785792e-01 4.79288787e-01 -3.73836219e-01 -3.85156453e-01
-6.00657202e-02 3.58925045e-01 -2.69623309e-01 -5.13276279e-01
5.05878210e-01 -5.62352538e-01 -8.23927701e-01 4.98526543e-01
-7.25187778e-01 1.15333736e-01 4.18206960e-01 1.77683592e-01
8.65972281e-01 -3.68950546e-01 2.72026479e-01 -9.59278047e-01
4.01766926e-01 -3.89235198e-01 -6.97504580e-01 -8.27533826e-02
-3.52310807e-01 -3.59977216e-01 6.26333356e-02 1.57181863e-02
-8.36179554e-01 -3.88266817e-02 -3.68444473e-01 -5.44924557e-01
-5.86739540e-01 1.21466197e-01 -2.69263506e-01 -2.64500529e-01
3.28670651e-01 -1.18780933e-01 -2.99170762e-01 -6.42946064e-01
4.31078106e-01 6.58768535e-01 3.84206027e-01 -7.54342735e-01
7.20914304e-01 7.04801738e-01 -2.01977357e-01 -8.14971209e-01
-7.68058658e-01 -9.20944929e-01 -9.17564452e-01 -2.10318401e-01
8.54603231e-01 -9.49716270e-01 -2.89635986e-01 3.36321414e-01
-1.06381321e+00 -4.50415581e-01 -2.17164740e-01 2.73471057e-01
-5.61109781e-01 5.87136090e-01 -5.21633089e-01 -6.04030669e-01
-9.77704749e-02 -1.32381713e+00 1.55559957e+00 3.26817065e-01
-5.14525659e-02 -8.79490852e-01 -2.85989027e-02 4.56415176e-01
7.37400129e-02 4.02628064e-01 8.51988673e-01 -5.00147283e-01
-9.07689631e-01 6.57234117e-02 -6.27303064e-01 8.42760280e-02
2.21566439e-01 2.50516534e-01 -1.10428143e+00 -7.30499551e-02
-1.59650922e-01 6.04740269e-02 9.48792338e-01 6.18890703e-01
1.26457334e+00 4.97480214e-01 -6.05599821e-01 9.70268130e-01
1.27072752e+00 1.71363935e-01 4.40875947e-01 3.97217691e-01
8.89849961e-01 4.48627830e-01 8.02509725e-01 2.52099782e-01
4.91870373e-01 4.69250530e-01 4.53641713e-01 -5.66273928e-01
-3.12915236e-01 -1.85461402e-01 -3.97005789e-02 6.54546082e-01
-1.31983869e-02 -5.91415353e-02 -1.37024963e+00 7.33272910e-01
-1.63638330e+00 -3.65621746e-01 -4.18860584e-01 1.91456139e+00
4.54913288e-01 3.21623921e-01 1.20759048e-01 -1.57243758e-01
9.17444170e-01 1.56313211e-01 -9.27507222e-01 -1.18817143e-01
1.59836859e-02 8.49679857e-02 5.39908230e-01 5.39992690e-01
-1.40766966e+00 1.31994426e+00 6.33406973e+00 8.14382076e-01
-1.06397247e+00 -3.39740142e-02 7.25048363e-01 8.57883841e-02
-3.74569148e-01 1.83831267e-02 -9.02037084e-01 2.84463406e-01
2.89662033e-01 1.00968063e-01 -8.09697583e-02 9.76120591e-01
2.46434301e-01 -1.20682359e-01 -1.06144834e+00 1.13391614e+00
-2.76940733e-01 -1.22841251e+00 1.12782747e-01 1.82051491e-02
9.32230949e-01 4.33283240e-01 6.04151599e-02 2.53253400e-01
2.02055141e-01 -8.60571504e-01 7.60442793e-01 7.28532970e-02
5.82849622e-01 -6.37588859e-01 4.44564611e-01 2.72854745e-01
-1.33712018e+00 2.94779837e-01 -2.59219110e-01 5.02260365e-02
2.04105809e-01 4.96217519e-01 -7.30456650e-01 6.30993128e-01
9.01443064e-01 8.79216909e-01 -6.91273987e-01 1.40855265e+00
-1.45946488e-01 5.10998368e-01 -5.88175297e-01 3.62061739e-01
4.72515613e-01 -4.21373129e-01 6.01719320e-01 1.35833371e+00
1.79631844e-01 2.20587462e-01 5.91965318e-01 1.21197796e+00
-8.29461440e-02 1.30881682e-01 -6.63121939e-01 9.86081958e-02
6.87241614e-01 1.06919265e+00 -1.25482595e+00 -2.22443268e-01
-5.02745450e-01 7.01862335e-01 1.68480515e-01 4.41011280e-01
-8.36285591e-01 -3.19513798e-01 1.00004017e+00 2.09224075e-01
3.42958689e-01 -5.58838189e-01 -8.55050921e-01 -1.05237603e+00
7.35138655e-02 -3.95492792e-01 2.68231064e-01 -6.37365818e-01
-1.34543526e+00 3.16930890e-01 3.05637002e-01 -1.31229556e+00
4.32202309e-01 -4.22293037e-01 -6.25722647e-01 6.99127078e-01
-1.78667426e+00 -1.14893186e+00 -5.75757563e-01 2.79275417e-01
7.20309079e-01 5.67281663e-01 1.41303003e-01 1.94911271e-01
-4.60067183e-01 3.68654579e-01 -3.85772660e-02 1.53295726e-01
4.84016240e-01 -1.26861513e+00 1.04410887e+00 9.39094543e-01
8.75197351e-02 4.90443170e-01 2.92041659e-01 -9.03231561e-01
-8.56515467e-01 -1.08422077e+00 2.68151850e-01 -6.43857419e-01
3.08837533e-01 -4.47454274e-01 -1.15547442e+00 5.18632770e-01
-8.38967636e-02 -3.95303257e-02 4.94300276e-01 1.08850896e-01
-2.09864154e-01 3.46499741e-01 -1.13092172e+00 7.39933014e-01
1.41668355e+00 -3.29255998e-01 -5.79022050e-01 2.06468731e-01
9.63165879e-01 -6.57720149e-01 -9.10858870e-01 9.20408428e-01
1.66430816e-01 -1.11139357e+00 1.17132008e+00 -2.97179788e-01
3.15649360e-01 -5.68406105e-01 -9.73218232e-02 -8.81264865e-01
-2.67781883e-01 -2.10724950e-01 2.61132896e-01 1.23614693e+00
3.56135756e-01 -5.46227813e-01 1.07814980e+00 7.26113796e-01
-4.30563450e-01 -7.30802417e-01 -8.32148433e-01 -8.53884578e-01
3.63673806e-01 -6.86313391e-01 8.23414326e-01 1.00673258e+00
-4.82071579e-01 1.65698290e-01 4.31756079e-01 5.06853878e-01
5.18242002e-01 6.93539143e-01 9.58599329e-01 -1.22178650e+00
3.09935093e-01 -8.19208920e-01 -4.04451191e-01 -1.65105557e+00
-9.09011345e-03 -1.11945617e+00 1.70193329e-01 -1.83444142e+00
3.63442898e-02 -9.22190785e-01 -1.01430103e-01 2.74038345e-01
-3.97006840e-01 3.41087759e-01 4.07899737e-01 6.05280042e-01
-6.10348701e-01 4.75765884e-01 1.54405475e+00 -1.27801880e-01
-5.84330320e-01 3.16600353e-02 -6.78653777e-01 9.62704062e-01
9.37899172e-01 -3.10342431e-01 -2.68947214e-01 -5.52476227e-01
-3.19609493e-01 -2.64132470e-01 3.42389047e-01 -1.05495882e+00
2.75275856e-02 -2.10667163e-01 4.34626132e-01 -1.12698209e+00
3.10057670e-01 -6.06757164e-01 -9.22972336e-02 1.37503937e-01
6.08602837e-02 -1.17959321e-01 5.00839293e-01 4.36110586e-01
-1.57770947e-01 -1.55013233e-01 7.14409113e-01 -1.82459712e-01
-9.73478019e-01 4.97982532e-01 1.48315400e-01 3.68390679e-01
1.02291548e+00 -6.20795608e-01 -8.94652158e-02 -6.82217884e-04
-5.69437265e-01 6.99041963e-01 1.09967256e+00 4.27292556e-01
6.35088384e-01 -1.11353052e+00 -4.87788320e-01 2.54632294e-01
2.43474603e-01 6.94604397e-01 1.58979058e-01 6.72684550e-01
-8.77228141e-01 3.25824380e-01 8.28749165e-02 -1.31611371e+00
-1.04979706e+00 2.84970850e-01 3.53750408e-01 2.10201353e-01
-8.83237779e-01 1.03817677e+00 6.92157865e-01 -6.15143538e-01
1.76854208e-01 -8.80540371e-01 -8.99475589e-02 -1.29689828e-01
-3.64632308e-02 2.24988684e-01 2.87641846e-02 -5.81209958e-01
-5.08392453e-01 1.11594868e+00 -7.42533803e-02 1.57069087e-01
1.05447030e+00 -1.42880112e-01 1.98176757e-01 4.97905582e-01
1.00656426e+00 1.01412393e-01 -1.60499156e+00 1.69258080e-02
1.80816576e-01 -8.03498447e-01 7.86566734e-02 -4.64523613e-01
-9.81923819e-01 9.10739779e-01 5.55938244e-01 2.07361504e-01
9.57048833e-01 3.06381792e-01 9.60018158e-01 -4.50184792e-02
4.09707487e-01 -8.22498381e-01 -2.12128371e-01 5.63859224e-01
6.27684474e-01 -1.50003493e+00 1.30462795e-01 -1.04381895e+00
-5.95201075e-01 7.87539959e-01 6.66228950e-01 -2.96731114e-01
7.23058760e-01 1.55422747e-01 3.06030780e-01 -4.16699171e-01
-1.62471011e-01 -5.44575453e-01 3.65755707e-01 4.92515504e-01
2.81027406e-01 2.97550887e-05 -1.08143941e-01 2.67536491e-01
-1.78144813e-01 -2.36233905e-01 2.07113340e-01 9.84020531e-01
-5.05822837e-01 -9.91570532e-01 -4.84251767e-01 3.47939640e-01
-2.42590383e-01 -2.23667715e-02 -7.52172768e-01 1.10680497e+00
6.28254339e-02 8.31858873e-01 3.16558659e-01 -2.50605851e-01
6.05630100e-01 -1.77786335e-01 2.81455129e-01 -7.99910009e-01
-4.57739353e-01 2.52368510e-01 -2.21927971e-01 -6.03117645e-01
-3.81178379e-01 -7.97584414e-01 -1.56635988e+00 -1.31391555e-01
-4.40530032e-01 6.94666952e-02 6.08190298e-01 5.86428642e-01
6.36108398e-01 2.79232442e-01 5.54843247e-01 -1.15174091e+00
-2.77693700e-02 -6.67005241e-01 -3.54419410e-01 5.28059661e-01
1.95619062e-01 -9.25825715e-01 -5.49477279e-01 -1.01760395e-01] | [7.969854831695557, -3.203348159790039] |
f0cbbe97-a0db-4704-b64d-da70634e2c83 | hpi-dhc-at-trec-2018-precision-medicine-track | null | null | https://trec.nist.gov/pubs/trec27/papers/hpi-dhc-PM.pdf | https://trec.nist.gov/pubs/trec27/papers/hpi-dhc-PM.pdf | HPI-DHC at TREC 2018 Precision Medicine Track | The TREC-PM challenge aims for advances in the field of information retrieval applied to precision medicine. Here we describe our experimental setup and the achieved results in its 2018 edition. We explored the use of unsupervised topic models, supervised document classification, and rule-based query-time search term boosting and expansion. We participated in the biomedical articles and clinical trials subtasks and were among the three highest-scoring teams. Our results showed that query expansion associated with hand-crafted rules contribute to better values of information retrieval metrics. However, the use of a precision medicine classifier did not show the expected improvement for the biomedical abstracts subtask. In the future, we plan to add different terminologies to replace hand-crafted rules and experiment with negation detection. | ['Jan-Philipp Sachs', 'Harry Freitas da Cruz', 'Ariane Morassi Sasso', 'Suparno Datta', 'Michel Oleynik', 'Erwin Bottinger', 'Benjamin Bergner', 'Erik Faessler', 'Arpita Kappattanavar'] | 2018-11-14 | null | null | null | notebook-papers-of-the-trec-2018-conference | ['negation-detection'] | ['natural-language-processing'] | [ 1.21025033e-01 2.87095219e-01 -3.22397858e-01 -2.23991722e-01
-1.06769598e+00 -3.90385032e-01 7.14032471e-01 7.76698411e-01
-9.02512908e-01 7.58934617e-01 4.31127489e-01 -4.85068381e-01
-6.37991071e-01 -5.18739700e-01 -4.38594133e-01 -3.45771670e-01
-2.01088414e-01 9.11098897e-01 3.99565995e-01 -3.34087461e-01
5.33904612e-01 1.99655984e-02 -1.40620410e+00 9.22857881e-01
8.27009439e-01 6.95798695e-01 7.74184614e-02 4.90602672e-01
-5.78460954e-02 4.92417574e-01 -5.40971816e-01 -3.17905486e-01
-5.83785288e-02 -1.03771061e-01 -9.12770331e-01 -6.50931001e-01
2.28613526e-01 7.75439572e-03 1.47689298e-01 6.40788794e-01
7.21161008e-01 -1.02434874e-01 6.72542751e-01 -8.15757573e-01
-3.04429531e-01 6.89854980e-01 -1.99359417e-01 3.82229984e-01
6.98277831e-01 -2.96144694e-01 8.33531737e-01 -7.60409832e-01
1.07251000e+00 1.09801984e+00 5.22893012e-01 3.60095143e-01
-9.46823120e-01 -5.80001235e-01 -7.09422007e-02 3.35669875e-01
-1.27705503e+00 -1.34888366e-01 3.78941029e-01 -5.62684894e-01
1.52364731e+00 3.36148947e-01 5.42207122e-01 1.00433171e+00
3.96467239e-01 3.94101918e-01 1.02638578e+00 -9.81205523e-01
4.23922241e-01 3.33215535e-01 5.03743351e-01 3.57794583e-01
4.01070744e-01 1.56497449e-01 -2.99216717e-01 -7.47432888e-01
-5.13269678e-02 3.03604044e-02 -4.00693715e-01 -4.05352823e-02
-1.15233850e+00 9.38137710e-01 3.28122616e-01 8.07004273e-01
-7.01318026e-01 -3.11590225e-01 6.76983595e-01 2.41791517e-01
6.82186425e-01 8.35542738e-01 -8.34265769e-01 -2.01983303e-01
-1.21683908e+00 3.75301808e-01 9.24167752e-01 8.50964546e-01
3.04319173e-01 -9.63234067e-01 -6.84498250e-01 8.52693439e-01
2.58452594e-01 2.10070074e-01 7.46130109e-01 -7.86982298e-01
3.72844130e-01 7.47434914e-01 1.73309505e-01 -7.84969091e-01
-8.23348224e-01 -3.89204651e-01 -3.88817370e-01 -3.96648854e-01
1.11529648e-01 -8.74704868e-02 -1.25913167e+00 1.14224970e+00
1.66207477e-01 -4.19607222e-01 -8.69295746e-02 6.74941540e-01
8.20388496e-01 2.80978501e-01 3.66204143e-01 -3.81942481e-01
1.93075716e+00 -7.46361673e-01 -1.06083882e+00 2.71608442e-01
1.18339121e+00 -8.13246071e-01 6.53470695e-01 6.62549078e-01
-9.33137298e-01 -2.51096040e-01 -9.83201444e-01 4.63797785e-02
-6.83763862e-01 3.44898105e-02 4.37348098e-01 6.52965963e-01
-1.05740118e+00 4.33624059e-01 -7.80832946e-01 -5.72800159e-01
2.14948982e-01 2.52150089e-01 -1.25094667e-01 -2.80780643e-01
-1.59528041e+00 1.18315911e+00 5.46272397e-01 -3.67673963e-01
-3.80671769e-01 -8.96968901e-01 -4.32924151e-01 1.24645904e-01
2.90068120e-01 -7.34431028e-01 1.24156046e+00 -8.86418447e-02
-9.67095792e-01 9.79438901e-01 -1.10756852e-01 -6.77120447e-01
5.31576499e-02 -2.65381247e-01 -3.41128945e-01 2.89904743e-01
2.10628957e-01 7.16283023e-01 5.65238632e-02 -5.59558809e-01
-6.56183243e-01 -6.29623175e-01 -1.78383574e-01 6.00650236e-02
-4.82428163e-01 3.15918952e-01 -4.39753532e-01 -5.11538386e-01
-9.09619480e-02 -1.05230892e+00 -2.67516732e-01 -6.99160814e-01
-9.61031243e-02 -6.62073433e-01 1.78138047e-01 -7.27185130e-01
1.54367661e+00 -1.89257693e+00 -2.17931300e-01 2.26595297e-01
-1.22016277e-02 2.60058403e-01 1.94139451e-01 8.16055059e-01
-3.71367596e-02 5.06877780e-01 4.52083528e-01 3.27188000e-02
-5.43817803e-02 -8.99126902e-02 -1.96058050e-01 -5.02365008e-02
-1.94921359e-01 8.05805564e-01 -7.78895199e-01 -8.30277801e-01
-1.01549394e-01 4.25264686e-01 -4.29536015e-01 -4.04080629e-01
-4.21007961e-01 -5.10631055e-02 -6.55095458e-01 6.16638958e-01
3.27949226e-01 -3.80328059e-01 2.49332383e-01 -5.08936979e-02
8.63983238e-04 6.28298759e-01 -5.16735017e-01 1.83271408e+00
-1.48592085e-01 2.68405229e-01 -2.97503799e-01 -7.22469747e-01
6.15703762e-01 7.67808080e-01 6.80988193e-01 -1.04978645e+00
-5.19582592e-02 4.12371635e-01 3.57426777e-02 -6.77598953e-01
3.33502829e-01 9.38224792e-02 1.62094027e-01 1.35492221e-01
5.62895127e-02 6.22346066e-02 3.60508621e-01 3.68878156e-01
1.33291948e+00 -1.35648265e-01 3.34485948e-01 -5.12490153e-01
4.27860200e-01 3.99881899e-01 1.35800198e-01 9.81363952e-01
-5.29851168e-02 4.95478779e-01 2.64052838e-01 -5.28955817e-01
-7.30493605e-01 -2.55937397e-01 -5.69232702e-01 1.30414677e+00
-5.52799404e-01 -9.93916214e-01 -5.80302000e-01 -7.66865492e-01
-1.39938712e-01 7.38269567e-01 -6.42682135e-01 -1.17889792e-01
-2.27872878e-01 -6.80858552e-01 5.74293554e-01 1.86257422e-01
8.70811939e-02 -1.14604092e+00 -6.14976287e-01 3.18935931e-01
-2.26690069e-01 -8.74083698e-01 -2.72556692e-01 4.66840386e-01
-1.10163260e+00 -1.03227592e+00 -1.18182993e+00 -6.59081578e-01
1.66529939e-01 -3.50408614e-01 1.04142940e+00 -3.21787409e-02
-4.60279733e-01 5.36102116e-01 -7.66748786e-01 -9.07619357e-01
-2.55955160e-01 5.22537887e-01 -3.38372558e-01 -8.46100450e-01
7.89556205e-01 -9.68676284e-02 -1.00944436e+00 6.86417595e-02
-1.06134009e+00 -4.52259809e-01 6.17083251e-01 7.70993114e-01
1.98923752e-01 -4.02128726e-01 4.85792041e-01 -9.52676177e-01
9.35249805e-01 -1.60475448e-01 -3.77751380e-01 5.75870514e-01
-1.38887954e+00 3.27018231e-01 -1.97417766e-01 -3.08467984e-01
-7.88080394e-01 -2.03905866e-01 -2.12737143e-01 1.16489399e-02
-1.12300083e-01 1.04273546e+00 4.19454396e-01 1.06338501e-01
1.15209150e+00 -2.10794032e-01 -2.91823804e-01 -6.36773825e-01
-1.35937352e-02 8.12558413e-01 -3.24692875e-01 -4.27229851e-01
-1.30971014e-01 1.57334790e-01 -5.35913184e-02 -4.78923649e-01
-5.89008391e-01 -8.46874475e-01 2.89415848e-02 3.63757834e-02
9.16305363e-01 -7.81438053e-01 -5.14606059e-01 -1.24322571e-01
-1.26358473e+00 7.41409361e-02 -1.77744806e-01 7.83773839e-01
-1.37198478e-01 4.05745119e-01 -5.88116348e-01 -6.24186039e-01
-8.98067892e-01 -1.01474524e+00 1.34412706e+00 -4.49554846e-02
-4.52244192e-01 -7.94681966e-01 7.03458130e-01 4.86591578e-01
6.12491727e-01 -2.49532446e-01 1.16937315e+00 -1.42708731e+00
1.57611698e-01 -6.08257353e-01 2.12705687e-01 -1.49427727e-02
-7.20735937e-02 -2.63462126e-01 -8.21367979e-01 -1.03638649e-01
-5.75074591e-02 -3.21590364e-01 1.16663158e+00 4.66736227e-01
9.61408317e-01 -1.47593170e-01 -8.24899971e-01 -2.43393391e-01
1.08460319e+00 6.33654892e-01 7.58630991e-01 7.15344250e-01
-1.21592723e-01 6.82029307e-01 8.28838766e-01 2.48321310e-01
2.26955146e-01 9.04127240e-01 -7.54955709e-02 1.85196891e-01
1.06008157e-01 1.17737167e-02 -9.84445736e-02 4.12894815e-01
-2.66055107e-01 -3.26952308e-01 -1.33943665e+00 5.16185343e-01
-1.80072057e+00 -4.67909783e-01 -9.14227664e-02 2.18421698e+00
1.15885079e+00 3.95760536e-01 -1.11107118e-01 5.09344898e-02
3.73140216e-01 -6.72229707e-01 -1.54290292e-02 -4.40805227e-01
2.61335135e-01 5.05522907e-01 3.49650979e-01 3.24632883e-01
-1.06573009e+00 5.85813522e-01 6.71831083e+00 1.01652801e+00
-7.21888065e-01 1.25487223e-01 4.66880590e-01 -1.28464296e-01
-1.26815349e-01 1.79163218e-01 -1.01597512e+00 4.44774270e-01
1.21919084e+00 5.88053614e-02 -2.35700130e-01 7.45660782e-01
3.92729901e-02 -3.18058401e-01 -1.03119159e+00 4.26794857e-01
-3.04282457e-02 -1.28398454e+00 1.00618582e-02 1.70173764e-01
5.58201849e-01 2.33798757e-01 -2.93085843e-01 7.18802989e-01
3.33535336e-02 -7.61694908e-01 6.83848634e-02 7.57518649e-01
2.57123888e-01 -4.48776856e-02 1.23937154e+00 3.63808215e-01
-2.46307567e-01 2.04057798e-01 -1.78117290e-01 6.39596507e-02
-2.26369262e-01 8.66281569e-01 -1.29818809e+00 8.48121524e-01
8.09829712e-01 2.83150166e-01 -7.44079769e-01 1.39726472e+00
1.72243521e-01 7.26372778e-01 -4.74223703e-01 -3.96045595e-01
3.21836444e-03 4.10704672e-01 4.46051151e-01 1.41682351e+00
2.40405634e-01 -1.06745161e-01 1.81938320e-01 1.38572961e-01
1.80505022e-01 6.03961527e-01 -3.14110100e-01 -7.32492805e-02
1.97743773e-01 1.03197610e+00 -8.31678629e-01 -5.85193515e-01
-2.00673461e-01 3.93154383e-01 -1.24561220e-01 4.52030718e-01
-5.01049459e-01 -4.40845549e-01 -3.21329057e-01 3.07844043e-01
3.35085064e-01 4.04517025e-01 -1.30937606e-01 -8.74123991e-01
-1.13053937e-02 -8.91321540e-01 9.32581306e-01 -7.78519690e-01
-1.13656354e+00 7.01406360e-01 5.81997395e-01 -8.41334760e-01
-4.55608308e-01 -4.64520752e-01 2.11838773e-03 6.52961910e-01
-1.23645294e+00 -7.98955858e-01 3.32183465e-02 4.42613661e-01
1.95993394e-01 -5.73395640e-02 1.22589850e+00 5.26895225e-01
8.15824345e-02 3.96040440e-01 1.59722984e-01 -1.46252215e-01
1.15401602e+00 -1.01782084e+00 -3.62957180e-01 -1.04321949e-01
1.15781361e-02 1.00009513e+00 8.54268432e-01 -9.15799320e-01
-8.73261392e-01 -5.70871890e-01 1.37663591e+00 -3.07759076e-01
4.59379464e-01 2.05027852e-02 -7.36467242e-01 3.80359590e-01
5.39033711e-01 -6.28235042e-01 8.30057621e-01 6.42984569e-01
-2.61439085e-01 -9.25419778e-02 -1.03539014e+00 2.50048816e-01
4.85112071e-01 -2.88865894e-01 -8.70973051e-01 9.62625384e-01
6.96977496e-01 -2.08003998e-01 -1.14294147e+00 9.40288782e-01
6.56536937e-01 -2.69120783e-01 9.49908674e-01 -9.24424887e-01
3.44077736e-01 2.43679971e-01 4.39871550e-02 -9.81476843e-01
-2.40473717e-01 -5.74519597e-02 3.98460366e-02 8.69619191e-01
9.92704093e-01 -5.32476068e-01 7.01285422e-01 7.97593594e-01
-1.10962458e-01 -9.16934252e-01 -9.60789561e-01 -5.87982893e-01
4.32958931e-01 -2.44379118e-01 -1.61886290e-02 5.46781778e-01
6.05341852e-01 4.84128624e-01 2.16609761e-01 -3.71702820e-01
4.55459729e-02 -7.59626999e-02 1.52591169e-01 -1.24651289e+00
-1.09236293e-01 -4.23410475e-01 -4.72855806e-01 -3.73751044e-01
-4.09219533e-01 -8.96560669e-01 1.74419694e-02 -1.74967241e+00
5.00178993e-01 -9.74255726e-02 -5.86076796e-01 7.44587123e-01
-3.00479472e-01 1.12016790e-01 -1.51596919e-01 2.48707861e-01
-8.92013371e-01 -2.24127853e-03 8.48073363e-01 -3.23938459e-01
-1.76056370e-01 8.26230720e-02 -7.24290550e-01 3.10067415e-01
7.27908731e-01 -8.95326912e-01 -2.66717255e-01 -4.34754491e-02
7.93640971e-01 -4.92968708e-02 1.83441460e-01 -7.29524851e-01
6.42681360e-01 2.08604932e-01 2.49838963e-01 -7.52005577e-01
1.45319685e-01 -7.46835887e-01 3.53669040e-02 6.70850456e-01
-7.29896426e-01 -2.25665849e-02 3.94453228e-01 3.84119749e-01
-2.69328564e-01 -5.14581740e-01 5.24910390e-02 -3.50577325e-01
-5.48416227e-02 -3.68033737e-01 -6.07497871e-01 -5.04165292e-02
7.47188687e-01 1.75649196e-01 -3.18121403e-01 -1.97677746e-01
-9.23748255e-01 4.33657944e-01 -9.69413221e-02 3.09715211e-01
3.57914031e-01 -7.34426141e-01 -7.76757360e-01 -3.33801776e-01
4.47576404e-01 -3.56602967e-01 1.50073215e-01 1.08275199e+00
-2.71269470e-01 1.36166406e+00 2.43927374e-01 -6.64794564e-01
-1.44132161e+00 8.37724686e-01 -5.61051555e-02 -1.22320855e+00
-2.15910181e-01 6.09671175e-01 -2.52284169e-01 -4.11754638e-01
6.27291441e-01 -5.05457222e-01 -4.76207286e-01 1.97811574e-01
5.83579659e-01 3.21391374e-01 1.00142360e+00 3.72189552e-01
-6.45317733e-01 3.00296992e-01 -5.87056994e-01 -4.06048864e-01
1.21471393e+00 2.96071708e-01 -1.69410154e-01 2.61496902e-01
1.12582660e+00 -1.89125016e-01 1.23434313e-01 -1.60507351e-01
6.26900852e-01 1.96051776e-01 2.56989926e-01 -1.69893932e+00
-4.45448250e-01 5.15890598e-01 8.74129176e-01 3.21566254e-01
1.17731392e+00 1.63824603e-01 2.54355550e-01 9.10790980e-01
4.92869377e-01 -1.09686828e+00 -3.35643113e-01 5.31025648e-01
9.17747796e-01 -1.17982101e+00 3.28225791e-01 -1.45625174e-01
-4.41573411e-01 9.98914838e-01 4.54203263e-02 3.19493532e-01
9.82076645e-01 1.74912810e-02 -1.55930966e-02 -6.92613304e-01
-1.00817072e+00 -1.19867697e-01 7.94663548e-01 2.56925225e-01
8.93264771e-01 -1.61517426e-01 -1.25973010e+00 6.84873521e-01
1.03223503e-01 6.57099724e-01 -2.65324593e-01 1.25845277e+00
-2.31350020e-01 -1.42328024e+00 -3.19137335e-01 7.45507181e-01
-9.90684032e-01 -5.55562317e-01 -5.94330549e-01 7.16508031e-01
1.87157691e-02 1.00982177e+00 -4.19935554e-01 -2.59637386e-01
3.46052408e-01 7.29999185e-01 5.06652713e-01 -5.96294641e-01
-1.05837405e+00 3.45634520e-01 2.85516798e-01 -4.67874408e-01
-5.47386885e-01 -7.08006859e-01 -9.02774394e-01 3.28619510e-01
-6.74230278e-01 8.03691745e-01 9.48652267e-01 8.37502003e-01
7.75791824e-01 4.80241895e-01 -1.12836085e-01 -1.63197026e-01
-6.10084951e-01 -1.74974513e+00 -1.06192142e-01 1.73870832e-01
1.86095331e-02 -5.66619456e-01 -3.25764358e-01 -4.48105261e-02] | [8.770298957824707, 8.658711433410645] |
0f4e018c-8a01-4911-85ff-124f87c416a0 | deep-iterative-and-adaptive-learning-for | 1912.07832 | null | https://arxiv.org/abs/1912.07832v1 | https://arxiv.org/pdf/1912.07832v1.pdf | Deep Iterative and Adaptive Learning for Graph Neural Networks | In this paper, we propose an end-to-end graph learning framework, namely Deep Iterative and Adaptive Learning for Graph Neural Networks (DIAL-GNN), for jointly learning the graph structure and graph embeddings simultaneously. We first cast the graph structure learning problem as a similarity metric learning problem and leverage an adapted graph regularization for controlling smoothness, connectivity and sparsity of the generated graph. We further propose a novel iterative method for searching for a hidden graph structure that augments the initial graph structure. Our iterative method dynamically stops when the learned graph structure approaches close enough to the optimal graph. Our extensive experiments demonstrate that the proposed DIAL-GNN model can consistently outperform or match state-of-the-art baselines in terms of both downstream task performance and computational time. The proposed approach can cope with both transductive learning and inductive learning. | ['Mohammed J. Zaki', 'Lingfei Wu', 'Yu Chen'] | 2019-12-17 | null | null | null | null | ['graph-structure-learning'] | ['graphs'] | [-4.21599112e-02 4.30778056e-01 -2.78044164e-01 -4.40204769e-01
-6.15432858e-01 -5.42621672e-01 6.39749825e-01 3.36686075e-01
-2.76159853e-01 3.00623477e-01 2.08637401e-01 -4.91444200e-01
-1.44059762e-01 -7.84393132e-01 -8.43574822e-01 -4.36016142e-01
-3.29709500e-01 4.92733479e-01 3.25149037e-02 3.43435034e-02
-5.51968440e-03 3.36042494e-01 -8.37506711e-01 -9.60432887e-02
7.77683258e-01 6.86090112e-01 2.25507200e-01 7.10184932e-01
1.49313118e-02 9.88365948e-01 7.33878314e-02 -4.81529176e-01
4.04564172e-01 -3.75492334e-01 -8.06294560e-01 3.24245185e-01
6.39699936e-01 -1.03410520e-01 -7.94693649e-01 1.11221385e+00
4.39617276e-01 2.96375424e-01 6.16923034e-01 -1.33389664e+00
-8.81564140e-01 8.38456213e-01 -6.29667401e-01 1.68638721e-01
-7.13591406e-04 1.30914718e-01 1.60824490e+00 -1.01380241e+00
6.37711942e-01 1.23433518e+00 7.55057156e-01 2.92054474e-01
-1.37352777e+00 -4.77596849e-01 5.78022957e-01 -6.85816929e-02
-1.30926931e+00 -2.50504553e-01 1.05763268e+00 -4.47372258e-01
7.95135975e-01 -3.49275231e-01 5.67039371e-01 1.06270123e+00
-5.49050234e-02 5.80473244e-01 7.59940326e-01 -4.94370729e-01
2.32238978e-01 -4.36927021e-01 2.96641022e-01 1.44042861e+00
3.14808518e-01 3.01821833e-03 -4.01535183e-01 8.14021826e-02
7.79905081e-01 -1.76441073e-01 -3.01547386e-02 -7.74106979e-01
-1.01193142e+00 9.12529469e-01 1.03150249e+00 1.98985338e-01
-3.31992120e-01 6.12951517e-01 4.25901204e-01 3.83993477e-01
6.11029625e-01 4.14554924e-01 -2.88117081e-01 2.82915324e-01
-5.76210558e-01 -3.53033654e-02 6.79808736e-01 1.04974234e+00
1.03210056e+00 1.29566208e-01 -2.15584844e-01 6.41021490e-01
6.63917482e-01 1.19042687e-01 -1.44023700e-02 -6.49400175e-01
7.55412102e-01 9.20766056e-01 -5.96220315e-01 -1.11131907e+00
-4.03053135e-01 -5.41824698e-01 -8.87187362e-01 -1.52833700e-01
3.71822774e-01 -2.45760649e-01 -1.07821727e+00 1.93508387e+00
3.30506623e-01 7.91241407e-01 -1.01690203e-01 6.95941448e-01
1.03767979e+00 3.13400984e-01 8.05987045e-02 2.02338219e-01
9.32491839e-01 -1.38298750e+00 -4.41315770e-01 -5.24104834e-01
1.02191460e+00 -2.38867715e-01 1.33764875e+00 -8.33875760e-02
-9.10515428e-01 -4.02280003e-01 -9.29449022e-01 -1.79409847e-01
-6.45061955e-02 -4.34606671e-02 5.98002851e-01 1.56461924e-01
-1.39547014e+00 7.07829356e-01 -8.42255354e-01 -3.34140360e-01
4.35780257e-01 2.33309984e-01 -4.71037567e-01 -3.78729731e-01
-8.47150803e-01 2.26528957e-01 6.21018827e-01 1.08515203e-01
-1.01423180e+00 -6.63091421e-01 -1.23315465e+00 9.80990902e-02
5.57810247e-01 -8.91444266e-01 8.76279235e-01 -7.61476219e-01
-1.25646758e+00 1.08737946e+00 1.06558442e-01 -5.71152627e-01
1.86548084e-01 -1.51652470e-01 -1.16075493e-01 1.04370974e-01
2.64150891e-02 3.28703344e-01 8.07422817e-01 -1.03940797e+00
-5.03699221e-02 -3.75383139e-01 8.84975027e-03 3.56674552e-01
-3.37339491e-01 -2.21692547e-01 -8.38059664e-01 -7.25069046e-01
5.85326292e-02 -1.06386006e+00 -4.92806613e-01 -6.02271520e-02
-6.47780478e-01 -3.12556177e-01 5.77353716e-01 -4.82976526e-01
1.25980091e+00 -1.98219001e+00 3.85658801e-01 5.02417207e-01
9.12067473e-01 1.98225796e-01 -5.34775615e-01 5.13128519e-01
-1.29606828e-01 7.36671034e-03 -2.51524508e-01 -6.46012485e-01
1.69679448e-02 2.51539886e-01 2.65762135e-02 5.89835882e-01
1.59029648e-01 1.43448448e+00 -1.20973277e+00 -2.69565791e-01
3.69130797e-03 4.37299013e-01 -6.56546533e-01 5.42567670e-01
-3.89778644e-01 3.82490039e-01 -4.00001466e-01 4.11876172e-01
5.22589445e-01 -7.78009117e-01 6.39097929e-01 -2.02766418e-01
4.03094590e-01 4.16062295e-01 -1.04116321e+00 1.99205101e+00
-3.65053773e-01 5.31316638e-01 1.78470001e-01 -1.24971354e+00
1.05939603e+00 -1.29896134e-01 1.72105893e-01 -5.60118437e-01
1.67417124e-01 -3.53607126e-02 3.07096560e-02 -2.37857565e-01
8.54035243e-02 1.97370514e-01 1.76431984e-02 5.12427270e-01
4.23189700e-01 2.54689872e-01 1.12656362e-01 6.43446028e-01
1.46343303e+00 1.33092478e-01 2.09321395e-01 -4.84418571e-01
3.49468321e-01 -5.06923378e-01 3.21565449e-01 6.81462407e-01
-6.38117790e-02 3.32770139e-01 7.83814192e-01 -2.50223428e-01
-8.89936090e-01 -1.22321582e+00 5.64442396e-01 1.29424334e+00
4.93882075e-02 -7.06070781e-01 -7.24974751e-01 -1.13594365e+00
3.74024734e-02 2.91954398e-01 -5.92215478e-01 -3.66871297e-01
-6.20636284e-01 -3.52366149e-01 3.57503086e-01 4.70784217e-01
4.17030782e-01 -9.40992355e-01 4.01275277e-01 1.91517979e-01
1.05062157e-01 -1.47561300e+00 -9.41966414e-01 -8.95047709e-02
-9.82743859e-01 -1.26336098e+00 -2.06952929e-01 -1.29332197e+00
9.67411995e-01 3.62813383e-01 1.35871732e+00 4.19453591e-01
2.13567470e-03 4.57858294e-01 -3.26988548e-01 1.48448616e-01
-3.83421510e-01 3.81675780e-01 -3.04033548e-01 2.55792469e-01
-1.05782144e-01 -7.46987402e-01 -6.84841812e-01 3.90067138e-02
-6.75532758e-01 8.40771571e-02 4.13504511e-01 6.03608966e-01
7.05995381e-01 -2.27306679e-01 5.99178195e-01 -1.28825808e+00
9.38998997e-01 -5.70923448e-01 -6.39283955e-01 2.72653162e-01
-9.99183714e-01 5.56702912e-01 6.44732952e-01 -1.34696260e-01
-4.71517205e-01 7.32747912e-02 7.09249005e-02 -7.68091321e-01
1.96714863e-01 9.93225813e-01 -1.99302837e-01 -2.62799650e-01
5.81668854e-01 -9.76561159e-02 1.46457806e-01 -4.56343532e-01
9.20011342e-01 1.12577111e-01 6.51608884e-01 -5.32738924e-01
1.11388111e+00 2.55914211e-01 3.24282020e-01 -4.62019235e-01
-1.00780189e+00 -5.61667442e-01 -7.96544909e-01 -1.63484767e-01
8.36103976e-01 -1.15254939e+00 -4.35592890e-01 5.59957922e-01
-9.04753029e-01 -7.64723718e-01 -3.81945632e-02 3.89857262e-01
-4.09415334e-01 6.54888809e-01 -7.36712575e-01 -3.89607996e-01
-6.14889741e-01 -8.17922711e-01 1.14510334e+00 -1.72142640e-01
2.18160927e-01 -1.73272383e+00 3.59064043e-01 1.53189823e-01
1.19323909e-01 5.53330898e-01 9.09221768e-01 -7.69647539e-01
-6.52919471e-01 -1.14269897e-01 -5.83011329e-01 2.69180059e-01
1.66323990e-01 -3.32462043e-01 -6.35462046e-01 -6.58614635e-01
-5.92072964e-01 -4.65447187e-01 1.05923557e+00 2.74533391e-01
9.71897840e-01 -4.55930263e-01 -2.33079806e-01 1.03417873e+00
1.57147086e+00 -4.48029280e-01 4.76629138e-01 8.07872508e-03
1.58788788e+00 3.93906057e-01 2.36536428e-01 1.96108758e-01
7.20879257e-01 4.64076757e-01 6.07109189e-01 -3.72657269e-01
-3.77139032e-01 -6.90775514e-01 3.45460922e-01 9.94286895e-01
1.82893962e-01 -3.93415630e-01 -9.83278990e-01 5.51165700e-01
-2.19126606e+00 -5.78923225e-01 -1.65849522e-01 2.07637453e+00
4.78596658e-01 2.46734455e-01 2.41919473e-01 -2.30506122e-01
9.06212509e-01 3.94222021e-01 -5.64112842e-01 -2.99594283e-01
9.54620987e-02 3.32478195e-01 5.25341809e-01 7.97320127e-01
-1.05107772e+00 1.42957687e+00 6.09806061e+00 4.30294156e-01
-8.96526456e-01 -1.57281861e-01 5.64967990e-01 1.74817160e-01
-3.66926104e-01 1.20987773e-01 -4.18488324e-01 1.75854370e-01
9.55580115e-01 -1.99069634e-01 8.20937276e-01 7.19470084e-01
3.57314385e-02 6.78990126e-01 -1.29848778e+00 9.94359136e-01
2.54075564e-02 -1.50790727e+00 8.53298753e-02 2.35973760e-01
7.36524343e-01 5.26885629e-01 -7.25595728e-02 4.75406080e-01
8.30986440e-01 -1.01933098e+00 3.56514484e-01 2.93332309e-01
7.32616425e-01 -6.09718740e-01 2.81874537e-01 1.68578625e-01
-1.69512475e+00 2.12216467e-01 -1.41337380e-01 -5.14955707e-02
2.98256278e-02 5.78403711e-01 -9.52234864e-01 6.15439892e-01
3.08397532e-01 1.16131318e+00 -8.44635725e-01 8.29999208e-01
-6.36337936e-01 8.57568145e-01 -1.37424126e-01 1.82353854e-01
6.08470857e-01 -5.41235209e-01 5.23768842e-01 1.11427677e+00
4.78394628e-02 -1.56162381e-01 6.01466954e-01 9.70259070e-01
-7.46205807e-01 1.90803275e-01 -1.01972425e+00 -3.78110051e-01
4.80810672e-01 1.38375497e+00 -7.95461237e-01 -2.25548744e-01
-5.47728777e-01 1.00230145e+00 1.26391494e+00 5.19722998e-01
-7.46879518e-01 -3.76415223e-01 4.85457927e-01 1.28396124e-01
6.00487828e-01 -4.85273361e-01 1.38897389e-01 -1.22743917e+00
8.08038414e-02 -6.67936444e-01 8.07791650e-01 -4.98231113e-01
-1.46392953e+00 6.44837618e-01 -1.82652339e-01 -7.19293714e-01
-1.77173257e-01 -5.90596378e-01 -9.18972373e-01 5.74860096e-01
-1.55117667e+00 -1.38707519e+00 -4.32994008e-01 8.02603066e-01
4.80846837e-02 -2.27649167e-01 4.68613356e-01 7.52545372e-02
-5.82204759e-01 7.88427651e-01 -6.88371062e-02 4.74074513e-01
5.10731637e-01 -1.53740823e+00 1.03435862e+00 1.10224807e+00
5.25616944e-01 4.43364471e-01 1.85130179e-01 -7.08348930e-01
-1.61674905e+00 -1.72995877e+00 7.28923202e-01 -2.09628001e-01
1.10468662e+00 -8.65866959e-01 -1.03321183e+00 9.38544273e-01
1.45294011e-01 4.87697899e-01 4.36824918e-01 3.31703544e-01
-7.80167818e-01 -1.17957378e-02 -8.77699912e-01 6.34975791e-01
1.56410480e+00 -7.50261009e-01 -1.44646466e-01 6.13304794e-01
1.16393065e+00 -4.19651270e-01 -8.03322792e-01 1.98534638e-01
-2.27302257e-02 -6.24439597e-01 8.51453364e-01 -8.98849845e-01
1.81736737e-01 -9.55724120e-02 -1.35420077e-02 -1.50351632e+00
-6.51763618e-01 -1.18290949e+00 -3.36207837e-01 1.11760116e+00
5.21675110e-01 -6.31238937e-01 9.66937184e-01 3.26692790e-01
-2.23223716e-01 -9.10678148e-01 -6.84420168e-01 -7.80548096e-01
-4.31810282e-02 -3.46334815e-01 3.24987739e-01 1.14472497e+00
-1.88196018e-01 9.19043720e-01 -2.44856298e-01 3.33907843e-01
1.03931725e+00 1.82209909e-01 8.82481158e-01 -1.26218069e+00
-5.75714946e-01 -3.75987947e-01 -6.75847888e-01 -1.05868280e+00
6.31824434e-01 -1.70080853e+00 -2.37274617e-01 -1.90282702e+00
1.18104897e-01 -2.44963109e-01 -5.08454442e-01 5.30164421e-01
-4.15731847e-01 3.98162790e-02 1.68148160e-01 -7.27874711e-02
-8.96701336e-01 5.38133442e-01 1.17838657e+00 -7.87830129e-02
-3.84333760e-01 -2.30402112e-01 -1.01888025e+00 4.73823398e-01
6.93475008e-01 -3.85344744e-01 -7.38311470e-01 -6.99512064e-01
4.11938012e-01 -1.10837430e-01 4.27217960e-01 -6.23749852e-01
2.22952381e-01 1.11369386e-01 -2.88588881e-01 -1.17180638e-01
-2.04538748e-01 -5.74299812e-01 -2.73058414e-01 2.49864817e-01
-6.04656935e-01 1.57184020e-01 5.83947301e-02 1.08676386e+00
1.16750367e-01 1.31568193e-01 8.04205656e-01 1.77874014e-01
-5.92771471e-01 9.76215601e-01 1.72501147e-01 6.38758004e-01
8.41868281e-01 2.32715309e-02 -3.19714844e-01 -4.45079237e-01
-8.46988678e-01 5.21757424e-01 2.60605305e-01 4.66734856e-01
7.01812923e-01 -1.59542215e+00 -8.72901320e-01 5.04046828e-02
3.38555694e-01 9.73344594e-02 -2.27007106e-01 7.10151553e-01
-4.65555906e-01 -7.88131356e-02 2.99465179e-01 -4.35303539e-01
-1.14499032e+00 6.80310965e-01 3.30671281e-01 -7.01805115e-01
-1.00259817e+00 9.82711434e-01 3.31431776e-01 -6.49339676e-01
4.02520537e-01 -4.99023259e-01 1.10829890e-01 -3.54709387e-01
3.21142823e-02 2.10738048e-01 1.64971198e-03 -5.74230075e-01
-2.68810928e-01 4.30854648e-01 -2.10479647e-01 2.90849775e-01
1.37141585e+00 -9.51294973e-02 -1.61018118e-01 1.53233528e-01
1.66292953e+00 -1.73566744e-01 -1.24846184e+00 -7.81148970e-01
1.71477169e-01 -2.13639706e-01 3.04626971e-01 -3.67806822e-01
-1.28662586e+00 6.14922464e-01 2.15623260e-01 3.03294957e-02
8.80699396e-01 3.33294153e-01 7.81189501e-01 6.45183146e-01
2.10894234e-02 -7.71486104e-01 5.04919887e-01 6.05331421e-01
8.43536377e-01 -1.20851755e+00 1.80456731e-02 -5.31196475e-01
-4.61187780e-01 8.30481529e-01 5.58269739e-01 -6.84008837e-01
8.72119069e-01 1.88847262e-04 -2.46136814e-01 -6.58620000e-01
-8.51130188e-01 -3.87131691e-01 7.11563647e-01 6.77798569e-01
3.85629833e-01 6.53588325e-02 1.60793796e-01 2.55139977e-01
-4.42670956e-02 -3.58861268e-01 2.54902512e-01 5.10043085e-01
-2.71182179e-01 -9.48654056e-01 2.63395488e-01 4.08266336e-01
-7.18390569e-02 -4.42181379e-01 -7.38035500e-01 7.92641461e-01
-4.61261004e-01 8.51856291e-01 1.10883296e-01 -5.22708952e-01
4.04978096e-01 -2.76074201e-01 6.51669025e-01 -8.59976530e-01
-3.42449993e-01 -4.02096510e-02 2.53711969e-01 -7.60131359e-01
-1.17020883e-01 -2.54870802e-01 -1.48795438e+00 -1.17203437e-01
-3.53966206e-01 -4.51466776e-02 2.70373881e-01 8.49008620e-01
6.77514970e-01 5.00636220e-01 7.30767965e-01 -6.69492483e-01
-3.70297551e-01 -7.89102435e-01 -6.13895476e-01 5.38376868e-01
2.50066966e-01 -3.79143268e-01 -3.73535097e-01 -1.07034750e-01] | [7.1616668701171875, 6.248135089874268] |
3d63b2a7-54bb-4563-9cfe-740c91a00d02 | neural-networks-with-manifold-learning-for | 1612.03961 | null | http://arxiv.org/abs/1612.03961v1 | http://arxiv.org/pdf/1612.03961v1.pdf | Neural Networks with Manifold Learning for Diabetic Retinopathy Detection | Widespread outreach programs using remote retinal imaging have proven to
decrease the risk from diabetic retinopathy, the leading cause of blindness in
the US. However, this process still requires manual verification of image
quality and grading of images for level of disease by a trained human grader
and will continue to be limited by the lack of such scarce resources.
Computer-aided diagnosis of retinal images have recently gained increasing
attention in the machine learning community. In this paper, we introduce a set
of neural networks for diabetic retinopathy classification of fundus retinal
images. We evaluate the efficiency of the proposed classifiers in combination
with preprocessing and augmentation steps on a sample dataset. Our experimental
results show that neural networks in combination with preprocessing on the
images can boost the classification accuracy on this dataset. Moreover the
proposed models are scalable and can be used in large scale datasets for
diabetic retinopathy detection. The models introduced in this paper can be used
to facilitate the diagnosis and speed up the detection process. | ['Christye Sisson', 'Rajeev Ramchandran', 'Arjun Raj Rajanna', 'Ali Shokoufandeh', 'Raymond Ptucha', 'Kamelia Aryafar'] | 2016-12-12 | null | null | null | null | ['diabetic-retinopathy-detection'] | ['medical'] | [ 1.56149283e-01 -3.71657722e-02 -1.01896068e-02 -6.79294884e-01
-3.99338514e-01 -1.58561975e-01 4.25393656e-02 7.42546245e-02
-5.61386526e-01 5.97154677e-01 6.58870786e-02 -6.33863032e-01
-2.89766610e-01 -8.47397327e-01 -2.63407350e-01 -6.65442288e-01
1.59622625e-01 1.07324481e-01 2.63357133e-01 1.02176517e-01
4.92250800e-01 8.87983739e-01 -1.93888688e+00 4.13916647e-01
1.30507898e+00 9.29769933e-01 1.56392738e-01 7.47263551e-01
1.45475373e-01 9.75818932e-01 -4.41121042e-01 -1.06732726e-01
4.63723153e-01 -4.56110924e-01 -5.52073717e-01 4.78594840e-01
9.53850746e-01 -9.10309732e-01 -1.22562997e-01 1.02887416e+00
9.32935357e-01 -2.11112648e-01 5.16685188e-01 -3.87997419e-01
-4.94426966e-01 -1.26061559e-01 -5.56023717e-01 4.13337737e-01
-1.25425741e-01 1.60667941e-01 2.24183798e-01 -5.64780354e-01
3.39804828e-01 8.20679188e-01 5.88222980e-01 2.80465662e-01
-7.89475441e-01 -3.83238137e-01 -2.41610959e-01 6.20034218e-01
-1.04678309e+00 -4.59437013e-01 2.51196891e-01 -7.80134559e-01
7.15296328e-01 1.39502153e-01 9.14547384e-01 -7.16919526e-02
-1.95114940e-01 2.67406106e-01 1.54850996e+00 -7.62223363e-01
1.35749020e-02 7.55602941e-02 3.93972039e-01 9.99140382e-01
6.03131413e-01 1.59149453e-01 1.09619588e-01 1.86593667e-01
9.35065150e-01 4.16783132e-02 -3.41603816e-01 -3.48559837e-03
-6.12839639e-01 8.60099137e-01 7.06823409e-01 -2.35516764e-02
-5.59812844e-01 -2.14612082e-01 1.71915382e-01 2.55225003e-01
5.59736550e-01 3.46273571e-01 -1.86973289e-01 1.96958333e-01
-8.21181417e-01 -1.93263784e-01 3.38339061e-01 3.52629721e-01
4.72277731e-01 -2.70056754e-01 -2.81267583e-01 1.11783028e+00
1.47359848e-01 3.74781668e-01 1.82082117e-01 -1.07140648e+00
1.08896545e-03 1.03112984e+00 2.39572987e-01 -7.19185770e-01
-5.23873210e-01 -3.84184808e-01 -7.59183466e-01 9.06403720e-01
7.17839837e-01 -3.28305274e-01 -1.30110347e+00 7.17187941e-01
3.10043305e-01 3.63589474e-03 -1.82181135e-01 1.06951308e+00
8.43513429e-01 2.71928638e-01 -9.78661403e-02 -2.32966602e-01
1.24212432e+00 -9.41674113e-01 -4.42128360e-01 3.30991633e-02
6.23726547e-01 -9.17374611e-01 7.16465056e-01 5.62128901e-01
-9.88705099e-01 -4.65708762e-01 -7.02395022e-01 -2.25580171e-01
6.38217246e-03 1.01985884e+00 7.46047676e-01 8.37811410e-01
-1.27804267e+00 2.46739775e-01 -7.76483238e-01 -8.15114260e-01
8.75244796e-01 5.22492647e-01 -3.18891943e-01 -4.92857516e-01
-3.35945815e-01 1.09507608e+00 2.16832422e-02 4.85485375e-01
-4.10780936e-01 -3.86496514e-01 -3.96149337e-01 -2.63482392e-01
1.55966850e-02 -6.59867108e-01 9.46307063e-01 -9.28093135e-01
-1.26713550e+00 1.21668553e+00 -2.45980471e-01 -4.29402947e-01
6.41411483e-01 -2.23600999e-01 -2.87966043e-01 3.96371812e-01
-1.23800553e-01 4.55077112e-01 4.92043793e-01 -7.93061376e-01
-1.19443822e+00 -6.18393838e-01 1.34638742e-01 6.50658682e-02
-2.62860447e-01 5.02530456e-01 -3.59836221e-01 3.93925272e-02
8.30792338e-02 -6.35333121e-01 -4.58083987e-01 4.18157488e-01
-2.20480248e-01 -1.66331843e-01 4.08313990e-01 -9.79676247e-01
9.28280771e-01 -1.93267775e+00 -4.43599015e-01 1.84777483e-01
3.00375402e-01 8.76747847e-01 -7.79905394e-02 -7.52833411e-02
1.49586082e-01 -3.08541488e-02 2.33375251e-01 1.80391744e-01
-7.52438009e-01 9.29606929e-02 1.84435576e-01 5.61896563e-01
3.08819920e-01 4.06454772e-01 -5.57238817e-01 -3.85524899e-01
5.18124700e-01 6.50854886e-01 -3.69787663e-01 1.95385471e-01
3.09668869e-01 4.19919640e-01 -2.79503733e-01 6.61375523e-01
5.36131024e-01 -3.67952436e-01 -3.66017781e-02 -3.19920003e-01
-3.77198726e-01 1.37926653e-01 -9.36052501e-01 8.90437603e-01
-2.08156303e-01 8.32405686e-01 -1.93222791e-01 -9.69898820e-01
9.53963399e-01 1.53143480e-01 7.44094849e-02 -5.43550313e-01
2.93997258e-01 1.77424610e-01 4.78702009e-01 -1.13123500e+00
6.06478713e-02 6.72721639e-02 9.63900805e-01 1.43550381e-01
-3.39854866e-01 3.49160612e-01 5.81984103e-01 -4.11384374e-01
7.73691356e-01 -8.53992850e-02 5.02380133e-01 1.92753792e-01
5.62988222e-01 2.09430650e-01 3.91346872e-01 6.04007423e-01
-3.52828175e-01 4.80022520e-01 3.74962419e-01 -8.03459227e-01
-1.08486176e+00 -3.85241151e-01 -4.81795937e-01 5.60878098e-01
-1.05073668e-01 3.18526179e-01 -6.16719484e-01 -4.07927573e-01
-8.07901025e-02 1.30030960e-01 -4.81975704e-01 2.71790713e-01
-2.41036683e-01 -1.13046253e+00 1.81891173e-01 4.98492867e-01
7.00126469e-01 -7.76540458e-01 -8.40977311e-01 -1.51082471e-01
2.53471397e-02 -8.04361880e-01 2.00680554e-01 -4.40566987e-01
-1.21170664e+00 -1.55195844e+00 -9.37759399e-01 -1.05832934e+00
1.17228639e+00 5.44235170e-01 6.72743857e-01 4.34931099e-01
-8.37900341e-01 -3.42402644e-02 -2.77232140e-01 -5.81101179e-01
-3.35493118e-01 -3.22045475e-01 -2.23268777e-01 2.07269102e-01
6.18528068e-01 -3.20814699e-01 -1.01745546e+00 2.84007192e-01
-4.53538328e-01 1.97142269e-02 9.77607965e-01 6.13890648e-01
5.65448582e-01 9.54166800e-02 3.96707892e-01 -9.86882806e-01
4.22093749e-01 1.09343126e-01 -1.07996786e+00 3.39155942e-01
-7.30478466e-01 -4.13595736e-01 -1.64143100e-01 -2.91867465e-01
-9.23371494e-01 1.46635901e-02 7.12210387e-02 -3.02591547e-02
-4.58693922e-01 6.17832780e-01 3.45981002e-01 -7.10327625e-01
9.27710891e-01 -2.34983400e-01 3.01146150e-01 -7.28672743e-01
4.59052995e-02 1.11440587e+00 3.97331089e-01 1.79787740e-01
2.53936738e-01 5.18159032e-01 2.84949780e-01 -9.51525033e-01
-9.89873350e-01 -7.23509908e-01 -4.35470581e-01 -4.25430834e-01
8.35000038e-01 -1.06486964e+00 -6.74031913e-01 8.28297377e-01
-9.05712187e-01 -1.48328856e-01 9.03979167e-02 8.38134050e-01
7.29239220e-03 3.49361390e-01 -4.81492430e-01 -9.71459985e-01
-2.01132730e-01 -1.04419577e+00 4.83494133e-01 7.50183940e-01
4.21412677e-01 -7.99818635e-01 -1.83964267e-01 9.27911341e-01
4.96550083e-01 1.53620958e-01 1.00150514e+00 -1.58883080e-01
-7.47920334e-01 -5.27658343e-01 -7.62813330e-01 9.14687932e-01
3.92663926e-01 2.97373295e-01 -1.01116216e+00 -6.43667951e-02
-3.87102008e-01 -1.83255345e-01 1.01401293e+00 8.96560907e-01
1.02415538e+00 -1.11130998e-01 -1.13533854e-01 5.62345088e-01
1.54495156e+00 3.66776645e-01 1.05035686e+00 4.66030240e-01
4.89810526e-01 7.93625593e-01 5.22221446e-01 8.70083719e-02
2.12626681e-01 2.77293921e-01 3.39860529e-01 -7.61681736e-01
-5.24065435e-01 4.04347509e-01 -1.76945806e-01 4.27550673e-01
-1.00409973e+00 1.71921059e-01 -1.07733071e+00 7.50073969e-01
-1.50994813e+00 -9.66178894e-01 -6.63624525e-01 2.36722922e+00
7.20109105e-01 -3.33348542e-01 1.52892172e-01 9.88220349e-02
8.93801451e-01 -6.92961693e-01 -3.27882349e-01 -2.27305904e-01
-5.28577752e-02 2.08121404e-01 6.76978886e-01 4.08425689e-01
-1.12722731e+00 6.15999520e-01 6.49533939e+00 -1.13724114e-03
-1.39933002e+00 -5.39639965e-02 7.46939301e-01 1.07732147e-01
6.70761168e-01 -1.95659474e-01 -7.31397092e-01 1.62708133e-01
5.76718271e-01 2.51141250e-01 2.59894311e-01 5.84936798e-01
6.95162892e-01 -4.48286057e-01 -7.35332787e-01 7.73106337e-01
5.06162755e-02 -1.32633567e+00 -1.30112739e-02 1.09586470e-01
8.79734635e-01 1.84762999e-01 3.73108604e-04 -2.92061299e-01
1.49534836e-01 -1.16275394e+00 -3.74385089e-01 8.65606248e-01
8.57576728e-01 -6.66089237e-01 1.22849619e+00 4.13430296e-02
-3.11807603e-01 -2.67674357e-01 -5.40244639e-01 -3.77772152e-01
-1.34912267e-01 6.43856645e-01 -1.08239555e+00 1.41521618e-01
7.88011432e-01 5.91405869e-01 -9.62603331e-01 2.06826258e+00
-3.73479396e-01 7.43618846e-01 -5.65209538e-02 4.94621396e-02
-1.24969460e-01 -5.02476573e-01 1.33991137e-01 6.61775887e-01
4.12916392e-01 2.44542167e-01 -2.26268619e-02 4.07724738e-01
2.77146310e-01 3.04187238e-01 -3.83339077e-01 -4.38273512e-02
8.00542086e-02 1.19148421e+00 -4.91368443e-01 -1.48950025e-01
-6.64205134e-01 3.41446519e-01 9.13356990e-02 4.80760902e-01
-1.82552859e-01 -5.08584380e-01 5.37847102e-01 4.24365759e-01
2.22237796e-01 1.39377862e-01 -3.68555993e-01 -8.23121905e-01
-3.50008942e-02 -7.80793607e-01 2.01651305e-01 -1.00502598e+00
-1.02121711e+00 5.64001501e-01 -5.04760683e-01 -1.12645268e+00
5.84553368e-02 -9.10653710e-01 -4.26090002e-01 1.19505525e+00
-2.08251309e+00 -1.11157608e+00 -8.86140943e-01 5.41560054e-01
2.17211451e-02 -4.27617669e-01 6.86404586e-01 5.28106987e-01
-6.71642542e-01 2.51413852e-01 7.79134482e-02 3.71520579e-01
8.47896874e-01 -1.03559375e+00 -3.56216997e-01 1.11050689e+00
-3.63502175e-01 5.40974498e-01 4.23619807e-01 -5.14288008e-01
-7.17239857e-01 -1.29823935e+00 9.07911122e-01 -1.03924409e-01
3.35123509e-01 5.85214019e-01 -6.79241240e-01 4.28896397e-01
-3.63255106e-02 1.42642960e-01 7.57021964e-01 2.21379608e-01
-6.93601221e-02 -4.30363804e-01 -1.16998398e+00 3.85263085e-01
5.64785898e-01 -3.20411474e-01 -3.65988076e-01 6.19605541e-01
-8.67297594e-03 -3.22750062e-01 -7.87167192e-01 3.94112974e-01
5.05808592e-01 -1.07838488e+00 6.94513679e-01 -7.73867667e-01
5.96758187e-01 -5.36485314e-01 1.85845584e-01 -8.93262029e-01
-8.00741985e-02 -1.35790497e-01 4.71449822e-01 1.04795027e+00
3.46614182e-01 -7.57973552e-01 6.55317724e-01 5.45568049e-01
1.08806387e-01 -5.32493174e-01 -6.24490738e-01 -2.29152456e-01
-2.29405686e-01 2.42694840e-02 -6.04348350e-03 5.80553234e-01
-4.70848531e-01 7.72902220e-02 -2.74439126e-01 5.63328981e-01
7.00107634e-01 1.92854881e-01 7.87039220e-01 -1.63275468e+00
-6.90092668e-02 -3.47112805e-01 -1.04071081e+00 -4.45784956e-01
-6.23670459e-01 -5.32280207e-01 -1.79181308e-01 -2.14399791e+00
3.34251761e-01 -3.83567661e-01 -2.12606668e-01 6.77758574e-01
-3.10950756e-01 6.77035511e-01 -1.48150250e-01 1.94973394e-01
-7.76904589e-03 -1.23638943e-01 1.46083105e+00 -1.82849485e-02
-3.68068606e-01 3.22744071e-01 -8.91357005e-01 8.22992384e-01
8.86558950e-01 -1.84175566e-01 -3.38145912e-01 -4.69681919e-01
1.67233869e-01 -1.19863816e-01 5.65744817e-01 -1.05062497e+00
1.87005877e-01 -1.78594574e-01 4.73288745e-01 -3.30681413e-01
1.82192586e-02 -6.73603296e-01 -2.11770639e-01 5.55988610e-01
-8.69380236e-02 -5.68054914e-01 1.32418945e-02 3.23533863e-01
-2.73077220e-01 -3.65647711e-02 1.13154709e+00 9.71187931e-03
-5.62269032e-01 1.20392404e-01 -2.39721820e-01 -4.53233182e-01
1.02547038e+00 -4.09675807e-01 -8.14436376e-01 -1.93301305e-01
-7.11171091e-01 1.31868288e-01 3.83240968e-01 3.30866268e-03
6.21151209e-01 -7.17412591e-01 -1.01925159e+00 1.49908379e-01
2.62957066e-01 -5.28293587e-02 2.41482511e-01 1.29733419e+00
-1.10020280e+00 2.69907504e-01 -6.20201290e-01 -6.01979375e-01
-1.82897782e+00 2.84541488e-01 6.85253084e-01 2.05362767e-01
-5.82708657e-01 6.96041048e-01 -1.70029625e-01 1.78826183e-01
3.95934075e-01 -5.45693934e-01 -9.03432310e-01 1.74741503e-02
9.96468782e-01 6.07166231e-01 1.76408336e-01 -3.23614001e-01
9.54571962e-02 7.11678505e-01 -3.19993526e-01 4.13902372e-01
1.51106763e+00 -1.99355960e-01 -5.07746458e-01 1.26075894e-01
4.60981697e-01 -1.13995954e-01 -9.58978236e-01 -2.08965927e-01
-1.83763102e-01 -7.00384736e-01 6.54570341e-01 -1.29626644e+00
-1.28547394e+00 1.03898478e+00 1.36560678e+00 4.39485200e-02
1.49661577e+00 -3.20298672e-01 2.49666587e-01 5.31592369e-01
2.55505502e-01 -7.68381834e-01 -4.82210696e-01 -5.35979010e-02
5.09845555e-01 -1.46880293e+00 1.28364131e-01 -6.79782629e-01
-4.57308054e-01 1.23632884e+00 3.91700506e-01 -1.76833197e-01
4.48896497e-01 -1.84397087e-01 7.58451819e-01 -4.11402173e-02
-3.66523415e-01 -7.51772285e-01 5.07360756e-01 9.15175557e-01
6.01536512e-01 2.66006291e-02 -6.64138913e-01 -4.22048345e-02
2.65978396e-01 6.68189287e-01 9.24985349e-01 6.59996152e-01
-8.70389581e-01 -1.11559284e+00 -3.02563995e-01 9.56330240e-01
-5.76101124e-01 -7.11540282e-02 -2.68940508e-01 5.37092566e-01
3.56237650e-01 1.25477970e+00 2.47235913e-02 1.46540016e-01
3.22406054e-01 -1.60259321e-01 6.62519395e-01 -6.51197135e-01
-3.72983962e-01 1.23906583e-01 3.78843337e-01 -4.34636712e-01
-8.50688338e-01 -5.02411723e-01 -6.84877753e-01 -2.86699664e-02
-2.68118769e-01 -3.32126021e-01 7.36253440e-01 8.40756357e-01
4.40609872e-01 4.70020741e-01 4.81361061e-01 -3.65902185e-01
-4.75454032e-01 -1.15955091e+00 -6.90159321e-01 2.30894033e-02
5.18890977e-01 -4.22071457e-01 -1.77832901e-01 5.07937968e-01] | [15.83666706085205, -4.002407550811768] |
0fc11582-9296-4036-9bf8-f7d258be7c2b | transfer-learning-from-partial-annotations | 1908.10851 | null | https://arxiv.org/abs/1908.10851v1 | https://arxiv.org/pdf/1908.10851v1.pdf | Transfer Learning from Partial Annotations for Whole Brain Segmentation | Brain MR image segmentation is a key task in neuroimaging studies. It is commonly conducted using standard computational tools, such as FSL, SPM, multi-atlas segmentation etc, which are often registration-based and suffer from expensive computation cost. Recently, there is an increased interest using deep neural networks for brain image segmentation, which have demonstrated advantages in both speed and performance. However, neural networks-based approaches normally require a large amount of manual annotations for optimising the massive amount of network parameters. For 3D networks used in volumetric image segmentation, this has become a particular challenge, as a 3D network consists of many more parameters compared to its 2D counterpart. Manual annotation of 3D brain images is extremely time-consuming and requires extensive involvement of trained experts. To address the challenge with limited manual annotations, here we propose a novel multi-task learning framework for brain image segmentation, which utilises a large amount of automatically generated partial annotations together with a small set of manually created full annotations for network training. Our method yields a high performance comparable to state-of-the-art methods for whole brain segmentation. | ['Yike Guo', 'Wenjia Bai', 'Chengliang Dai', 'Elsa Angelini', 'Yuanhan Mo'] | 2019-08-28 | null | null | null | null | ['brain-image-segmentation'] | ['medical'] | [ 3.32737982e-01 1.13578066e-01 3.29246610e-01 -5.78651845e-01
-7.93006778e-01 -3.55950981e-01 2.99320757e-01 2.37061217e-01
-1.01494801e+00 6.18489802e-01 -2.41112202e-01 -2.71666169e-01
-6.98444024e-02 -4.20791000e-01 -4.96737033e-01 -6.98795021e-01
-6.60651699e-02 9.91145968e-01 5.45437098e-01 -4.02938910e-02
2.51913756e-01 6.54159248e-01 -1.24937439e+00 -2.10455582e-01
9.39547837e-01 9.44900930e-01 3.06013584e-01 2.70915627e-01
-4.04681742e-01 2.51532882e-01 -4.40011293e-01 -1.72588557e-01
3.68588030e-01 -1.58281416e-01 -1.11622393e+00 -1.02096219e-02
1.47294015e-01 -1.59619361e-01 2.42158741e-01 1.17756963e+00
7.80858994e-01 2.21383333e-01 6.76189899e-01 -9.72281635e-01
-2.76997853e-02 4.66232866e-01 -6.67755663e-01 5.46607494e-01
-1.55167386e-01 5.84549494e-02 4.30648267e-01 -5.16799688e-01
3.83033216e-01 8.59966338e-01 6.47100866e-01 4.46530044e-01
-1.26589930e+00 -5.78922570e-01 -1.67977586e-01 7.02055171e-02
-1.33624530e+00 -3.03421497e-01 8.18347633e-01 -8.44123423e-01
7.80444086e-01 1.10854581e-02 7.52963126e-01 8.17818165e-01
1.85003519e-01 3.49942356e-01 1.48512018e+00 -2.45102003e-01
2.69767821e-01 -3.41330394e-02 1.94002971e-01 4.97085422e-01
1.46398395e-02 -3.99019867e-01 2.67689496e-01 7.54564777e-02
1.11909211e+00 -8.73412490e-02 -8.42959359e-02 -5.60103118e-01
-1.19125903e+00 8.78523707e-01 3.95508528e-01 6.26906276e-01
-5.69571733e-01 -4.46688123e-02 6.30119503e-01 -5.78681454e-02
6.38650417e-01 2.52256095e-01 -4.10506636e-01 8.55426267e-02
-1.38834059e+00 4.74326536e-02 4.87024754e-01 4.91829544e-01
6.34273231e-01 -2.33536452e-01 2.76059331e-03 1.14567125e+00
3.04043442e-01 -4.54287007e-02 7.88008988e-01 -6.22178018e-01
2.55250424e-01 7.06831515e-01 -2.06067532e-01 -8.42321694e-01
-9.70045507e-01 -3.28195602e-01 -1.29980361e+00 3.21386129e-01
6.21966958e-01 -2.02690482e-01 -1.09386241e+00 1.44589055e+00
6.09924316e-01 -6.32251278e-02 -3.35743695e-01 1.03991926e+00
7.40980208e-01 8.02450627e-02 2.95355856e-01 -2.11944789e-01
1.33269382e+00 -8.68646801e-01 -5.40842474e-01 -2.11601421e-01
6.27048492e-01 -6.34783864e-01 9.06355679e-01 1.91030502e-01
-1.11944640e+00 -4.12396431e-01 -7.90230334e-01 1.41877616e-02
-4.15196896e-01 -1.07992053e-01 6.35448635e-01 7.78473318e-01
-1.31487930e+00 5.25410056e-01 -9.73809898e-01 -1.96366414e-01
1.02763546e+00 8.33692253e-01 -5.63373148e-01 2.27885976e-01
-9.89812016e-01 1.25907218e+00 8.03292394e-01 3.34231377e-01
-6.65447056e-01 -6.40299618e-01 -8.46612334e-01 -1.71741724e-01
6.04612052e-01 -4.43963587e-01 1.15126479e+00 -8.51197362e-01
-1.53182304e+00 1.25725424e+00 2.56938070e-01 -4.60530370e-01
8.24009776e-01 1.41769899e-02 -6.19230606e-02 3.43359470e-01
6.37471452e-02 9.15578425e-01 8.79609525e-01 -9.62636530e-01
-5.62567301e-02 -6.23403132e-01 -8.02844986e-02 5.92947938e-02
4.89055272e-03 3.95473480e-01 -2.84203112e-01 -5.12377441e-01
1.45574734e-01 -7.79524863e-01 -5.96835256e-01 -1.12106904e-01
-3.83824676e-01 -3.69825423e-01 7.21623778e-01 -8.31495583e-01
7.24968374e-01 -1.69315147e+00 1.17349386e-01 2.91403860e-01
5.54960668e-01 3.92176956e-01 1.56338483e-01 -2.53577977e-01
-2.10618600e-01 7.29840025e-02 -6.60315931e-01 -2.39416644e-01
-1.90748945e-01 2.40576006e-02 5.10869801e-01 5.50243497e-01
9.85853523e-02 7.94771135e-01 -6.34586811e-01 -9.08699036e-01
3.45191449e-01 5.99990189e-01 -3.90493333e-01 1.92762345e-01
-9.03528556e-02 1.08567536e+00 -3.52909565e-01 3.63876253e-01
5.95848024e-01 -3.55543137e-01 -6.57855300e-03 -2.33741745e-01
-1.57917038e-01 -2.24532634e-01 -1.03376460e+00 1.93475246e+00
-3.50738049e-01 1.19183592e-01 1.97351918e-01 -1.65169525e+00
8.38336289e-01 5.69314837e-01 8.19611728e-01 -3.92769486e-01
8.25363159e-01 4.19131666e-01 3.75447541e-01 -7.44315326e-01
-2.94043552e-02 -2.39055499e-01 1.46236941e-01 6.82488978e-01
2.23622411e-01 -4.83764559e-01 2.23846748e-01 -1.86220884e-01
1.01880860e+00 1.86135933e-01 3.17718536e-01 -5.27895808e-01
8.31898570e-01 -5.75488172e-02 5.26092112e-01 2.96681792e-01
-4.76231992e-01 5.32372713e-01 4.45256561e-01 -5.75746119e-01
-1.33731043e+00 -5.06091952e-01 -2.92855322e-01 6.41610146e-01
-2.13310853e-01 2.11001426e-01 -1.35590875e+00 -7.14315832e-01
-4.02350634e-01 2.52535135e-01 -5.22033453e-01 2.33656332e-01
-7.28438497e-01 -1.21026671e+00 5.73765695e-01 3.30953270e-01
7.14397609e-01 -1.26473427e+00 -8.57174277e-01 4.73140985e-01
-2.71227490e-02 -1.39422011e+00 -4.10064280e-01 2.25992814e-01
-1.14568162e+00 -1.01586843e+00 -1.15938056e+00 -8.17401588e-01
9.03168976e-01 -2.02759504e-01 1.00488484e+00 2.21929714e-01
-5.78029752e-01 3.49978879e-02 -1.42258525e-01 -3.66655171e-01
-2.90580690e-01 3.34238827e-01 -6.02111928e-02 7.97452778e-02
2.00819403e-01 -9.58773553e-01 -6.25300825e-01 2.98637390e-01
-1.08302391e+00 2.31142655e-01 7.56773591e-01 6.10167801e-01
8.19841504e-01 2.25628074e-02 7.75058270e-01 -1.04518723e+00
7.11024225e-01 -2.96756953e-01 -7.68655598e-01 2.11879555e-02
-3.00553858e-01 -6.61716387e-02 4.66456085e-01 -4.92601544e-01
-7.73935199e-01 2.91120559e-01 -5.32211304e-01 -3.23237628e-01
-5.59002638e-01 4.67448652e-01 7.78065715e-03 -5.71568906e-01
5.04395187e-01 2.71449257e-02 3.88537407e-01 -5.65210521e-01
1.24135084e-01 4.45948750e-01 4.61440146e-01 -3.60285342e-01
5.61672807e-01 2.06792280e-01 2.36209527e-01 -7.49920309e-01
-7.94663966e-01 -4.28428859e-01 -1.53863645e+00 -3.33210796e-01
1.24020362e+00 -3.98794264e-01 -4.20728058e-01 6.18210495e-01
-1.19866288e+00 -4.99011219e-01 1.11457244e-01 4.41028208e-01
-5.59333086e-01 5.05734384e-01 -3.58028084e-01 -2.96732008e-01
-6.84824467e-01 -1.83548272e+00 8.00852597e-01 1.20031402e-01
-1.72636747e-01 -1.06474102e+00 -1.06010936e-01 6.84052289e-01
5.30597150e-01 4.57961261e-01 1.09206939e+00 -1.02894986e+00
-3.14421803e-01 -2.17770979e-01 -4.87799644e-01 3.17472875e-01
6.96556121e-02 -5.27213752e-01 -7.97322392e-01 -1.28869131e-01
1.33921295e-01 -3.88957500e-01 4.10918653e-01 6.59916937e-01
1.55387557e+00 1.80312917e-01 -2.03869328e-01 4.61385101e-01
1.27852094e+00 2.56338060e-01 3.24965477e-01 2.60849744e-01
8.70365620e-01 7.62063801e-01 9.56022590e-02 5.98097853e-02
3.54271144e-01 6.24135196e-01 3.60743195e-01 -2.49929100e-01
1.31702274e-01 7.59452522e-01 -4.31533366e-01 1.00870836e+00
-4.28830147e-01 2.42091730e-01 -1.23508048e+00 4.83440667e-01
-1.59918392e+00 -5.62702358e-01 -1.72647640e-01 2.09524775e+00
1.07189667e+00 1.14198089e-01 4.04736549e-01 2.89181083e-01
7.68938720e-01 -8.53067264e-02 -5.92338562e-01 -1.46900490e-01
2.10761160e-01 4.39201683e-01 3.30422819e-01 2.35947102e-01
-1.32701707e+00 8.53367150e-01 6.28618193e+00 7.84519076e-01
-1.13613808e+00 6.14849210e-01 6.98179245e-01 1.24448678e-02
3.32655430e-01 -3.61649096e-01 -4.79011536e-01 5.20218432e-01
9.57994580e-01 3.22848856e-01 2.97678083e-01 6.07629895e-01
2.80650586e-01 -3.60219449e-01 -9.96957481e-01 1.15145004e+00
4.36498001e-02 -1.18704522e+00 -3.25964034e-01 1.84648201e-01
5.57174683e-01 2.21228287e-01 -4.05632764e-01 9.37362313e-02
-3.47083248e-02 -1.06874406e+00 4.04087007e-01 3.07499468e-01
6.74316406e-01 -6.78707123e-01 8.28412652e-01 3.73035818e-01
-8.15306127e-01 3.71549070e-01 -2.57473409e-01 2.54334152e-01
4.37338352e-01 8.38759542e-01 -6.43496394e-01 3.88092607e-01
6.56084001e-01 1.14274047e-01 -4.48827565e-01 1.46329319e+00
-3.39876786e-02 1.64823815e-01 -3.52850378e-01 2.33320549e-01
4.31407213e-01 -4.49488014e-01 3.09203476e-01 1.07494843e+00
1.22420594e-01 8.32404941e-02 5.03741801e-01 8.91590238e-01
-1.30674884e-01 3.99917752e-01 -3.48060608e-01 7.63577595e-02
-2.40850914e-02 1.66297174e+00 -1.53198266e+00 -3.33863497e-01
-3.43983412e-01 8.98476303e-01 4.29081172e-01 -1.16660990e-01
-6.32259071e-01 -2.59747803e-01 6.44260943e-02 2.85650976e-02
1.88612808e-02 -3.84078562e-01 -3.04320544e-01 -8.15858781e-01
-1.35907426e-01 -6.33524954e-01 1.25708759e-01 -4.37805504e-01
-1.05263102e+00 8.80943596e-01 2.36176282e-01 -8.03284585e-01
-1.87835559e-01 -5.13504326e-01 -6.25440300e-01 9.15707409e-01
-1.41953945e+00 -9.45866883e-01 -3.73495370e-01 5.71303070e-01
5.28763115e-01 -9.87213701e-02 6.85337484e-01 7.05273211e-01
-7.34489858e-01 1.55666843e-01 -3.73825103e-01 2.35382393e-01
4.34153497e-01 -1.18232846e+00 1.03570586e-02 6.33646846e-01
-2.49238089e-01 3.26518357e-01 4.54778463e-01 -5.81686080e-01
-8.86500061e-01 -9.76942897e-01 5.90278745e-01 -1.33080587e-01
7.49017358e-01 -1.91800445e-01 -1.12358785e+00 6.02208138e-01
1.45093068e-01 3.08914393e-01 7.55595982e-01 -2.05085486e-01
3.12833995e-01 1.73941404e-01 -1.38715434e+00 3.94387245e-01
8.41814160e-01 -2.52197951e-01 -6.90553248e-01 5.76606750e-01
4.67907488e-01 -5.69200695e-01 -1.20154703e+00 3.64830196e-01
3.02231461e-01 -7.60279953e-01 9.39007461e-01 -2.23853216e-01
1.20771304e-01 -1.97216585e-01 5.01267433e-01 -1.38699460e+00
-1.87508464e-02 -2.60272115e-01 3.40911448e-01 1.02317119e+00
2.31431499e-01 -5.62486887e-01 7.13179171e-01 7.97049642e-01
-2.77898878e-01 -8.93710256e-01 -1.11325276e+00 -6.11968815e-01
1.76790074e-01 -2.99882323e-01 5.42517066e-01 9.18282628e-01
-4.97280836e-01 2.44124785e-01 1.01266289e-02 -1.71623584e-02
9.15172100e-01 -1.87240288e-01 4.11968440e-01 -1.77095819e+00
3.40513259e-01 -7.49322891e-01 -4.81576711e-01 -4.05169100e-01
3.23374957e-01 -1.10659516e+00 6.87782988e-02 -1.59898055e+00
5.94609119e-02 -5.51539302e-01 -1.52244225e-01 5.64246833e-01
9.10285711e-02 6.01593435e-01 -1.49489790e-01 2.42005467e-01
-4.57292497e-01 3.70252073e-01 1.49322641e+00 -1.11799072e-02
-3.80718142e-01 3.91075984e-02 -3.31424564e-01 1.03767574e+00
9.18573380e-01 -5.64620435e-01 -3.11870724e-01 -4.39704925e-01
-1.07003875e-01 -2.63879122e-03 4.01138037e-01 -9.47037697e-01
3.01611662e-01 1.71252921e-01 4.00807589e-01 -5.78353465e-01
3.04743469e-01 -9.21771586e-01 7.92912394e-02 3.05586606e-01
-6.32040054e-02 5.32665774e-02 1.63289532e-01 3.69050987e-02
-1.26062557e-01 -5.48811793e-01 1.14561737e+00 -5.89724004e-01
-3.64201277e-01 6.92284584e-01 -4.48891491e-01 6.86352924e-02
1.12849128e+00 -3.37348223e-01 3.73586416e-01 2.86612689e-01
-1.11459982e+00 1.73197880e-01 -4.08017300e-02 1.10283971e-01
3.05812985e-01 -1.02313340e+00 -4.11822468e-01 2.86738928e-02
-4.40499932e-01 6.04281723e-01 4.00594473e-01 1.46296060e+00
-6.74684405e-01 3.01052719e-01 -6.41549766e-01 -7.96722710e-01
-1.18909240e+00 3.16814691e-01 5.03525436e-01 -4.48834896e-01
-9.53959644e-01 6.62775993e-01 -4.66560721e-02 -6.02146268e-01
2.66686380e-02 -4.30981159e-01 -6.34737372e-01 3.20962459e-01
3.61495078e-01 2.38133326e-01 2.55824387e-01 -8.66011918e-01
-2.63648331e-01 7.16497719e-01 -2.26623937e-01 9.22477245e-03
1.67663455e+00 4.97225560e-02 -5.13958812e-01 1.51698396e-01
1.07118440e+00 -6.60895050e-01 -1.10323143e+00 -2.62048483e-01
3.52855533e-01 -9.58023220e-02 5.37087560e-01 -6.39704049e-01
-1.39338791e+00 1.03044307e+00 6.81961715e-01 1.28964737e-01
1.07202840e+00 -6.79646432e-02 8.55185032e-01 1.32802114e-01
6.01967990e-01 -1.15896869e+00 -2.30539531e-01 4.12948698e-01
8.53561044e-01 -1.53181446e+00 -1.58686608e-01 -2.97948420e-01
-3.68972301e-01 1.08099341e+00 5.52427232e-01 -1.94788903e-01
8.19584548e-01 2.20087260e-01 8.63318071e-02 -4.90706772e-01
1.88109264e-01 -1.21789165e-01 3.83374006e-01 5.18613696e-01
3.36208552e-01 -1.71090022e-01 -4.86801445e-01 5.05109429e-01
-1.56147763e-01 1.02148525e-01 1.82889611e-01 7.97337413e-01
-3.34656090e-01 -1.22608781e+00 -4.37682211e-01 8.14475060e-01
-7.93469012e-01 -6.46623150e-02 8.41430277e-02 7.74887025e-01
1.23875722e-01 4.86613661e-01 -4.05109301e-02 1.95681557e-01
1.72048882e-01 3.39109898e-01 6.66350663e-01 -7.01814353e-01
-7.95979917e-01 2.34942779e-01 -2.62339175e-01 -4.33613926e-01
-8.27219844e-01 -7.24400640e-01 -1.39755762e+00 1.85894117e-01
-1.82858706e-01 -5.02511598e-02 9.65010464e-01 1.50518119e+00
9.88722369e-02 6.71792865e-01 2.36637846e-01 -1.43284714e+00
-2.78792590e-01 -1.25032914e+00 -5.19519210e-01 2.48437405e-01
6.17151372e-02 -9.78868365e-01 -1.18041253e-02 2.80369259e-02] | [14.345670700073242, -2.418712615966797] |
ff4ce528-97c2-45ec-9b5f-152056e2e78d | contrastive-multiview-coding | 1906.05849 | null | https://arxiv.org/abs/1906.05849v5 | https://arxiv.org/pdf/1906.05849v5.pdf | Contrastive Multiview Coding | Humans view the world through many sensory channels, e.g., the long-wavelength light channel, viewed by the left eye, or the high-frequency vibrations channel, heard by the right ear. Each view is noisy and incomplete, but important factors, such as physics, geometry, and semantics, tend to be shared between all views (e.g., a "dog" can be seen, heard, and felt). We investigate the classic hypothesis that a powerful representation is one that models view-invariant factors. We study this hypothesis under the framework of multiview contrastive learning, where we learn a representation that aims to maximize mutual information between different views of the same scene but is otherwise compact. Our approach scales to any number of views, and is view-agnostic. We analyze key properties of the approach that make it work, finding that the contrastive loss outperforms a popular alternative based on cross-view prediction, and that the more views we learn from, the better the resulting representation captures underlying scene semantics. Our approach achieves state-of-the-art results on image and video unsupervised learning benchmarks. Code is released at: http://github.com/HobbitLong/CMC/. | ['Yonglong Tian', 'Phillip Isola', 'Dilip Krishnan'] | 2019-06-13 | null | https://www.ecva.net/papers/eccv_2020/papers_ECCV/html/1453_ECCV_2020_paper.php | https://www.ecva.net/papers/eccv_2020/papers_ECCV/papers/123560749.pdf | eccv-2020-8 | ['self-supervised-image-classification', 'self-supervised-action-recognition'] | ['computer-vision', 'computer-vision'] | [ 1.12905884e-02 -1.33666605e-01 -1.13544121e-01 -3.91586006e-01
-6.68723464e-01 -7.86657929e-01 6.22297645e-01 -6.92587346e-02
-6.79330602e-02 3.35202903e-01 5.91890275e-01 2.89755553e-01
-4.06243280e-02 -6.99046731e-01 -1.02980614e+00 -7.42851138e-01
-3.87295820e-02 1.07125960e-01 5.70591912e-02 -3.75688337e-02
1.95071772e-01 1.40892908e-01 -1.84438288e+00 6.19898438e-01
1.00830197e-01 9.70503330e-01 4.54211354e-01 7.44486570e-01
3.01851004e-01 8.96138549e-01 -1.39187083e-01 -3.33251595e-01
2.38770902e-01 -2.84231484e-01 -8.07153523e-01 3.41340125e-01
7.36013651e-01 -2.07316801e-01 -4.82134461e-01 1.02819526e+00
2.91364074e-01 8.46401080e-02 6.58019185e-01 -1.02327514e+00
-5.89684963e-01 1.80031806e-01 -7.58101285e-01 9.88842621e-02
6.32532060e-01 -5.79339825e-02 1.36323154e+00 -9.62481678e-01
9.65814352e-01 1.36893415e+00 2.61292547e-01 3.89238447e-01
-1.60345995e+00 -2.04286784e-01 4.65280086e-01 3.47015828e-01
-1.09333694e+00 -4.76203740e-01 9.90096807e-01 -6.28551245e-01
6.82366073e-01 3.37261349e-01 6.93707466e-01 1.47170675e+00
5.70533693e-01 7.78080940e-01 1.33764327e+00 -2.06809118e-01
2.10591778e-01 2.82837242e-01 -1.28101945e-01 6.79187894e-01
7.04654753e-02 2.16248229e-01 -1.05327690e+00 -1.54119357e-01
5.39515674e-01 2.82478929e-01 -5.92697859e-01 -8.29138160e-01
-1.24940372e+00 8.33766997e-01 2.88811207e-01 3.09082866e-02
-2.70035267e-01 3.01032979e-02 2.55212426e-01 4.48150754e-01
5.28421164e-01 2.63698310e-01 -3.64734173e-01 6.60016537e-02
-5.36894143e-01 2.85682499e-01 8.35016608e-01 7.09385276e-01
8.68987679e-01 -3.30720395e-01 4.84009832e-01 7.67847359e-01
3.51423472e-01 6.15720034e-01 1.29514948e-01 -1.38167393e+00
2.41432011e-01 9.97995585e-02 -1.01437859e-01 -1.00648785e+00
-2.65335768e-01 -2.67492741e-01 -8.15353215e-01 4.33853149e-01
2.81710356e-01 4.07528542e-02 -7.23767102e-01 1.83216584e+00
1.23782612e-01 -3.56718637e-02 -1.76626295e-02 9.50316012e-01
9.61732149e-01 5.25817335e-01 -3.22914243e-01 -1.35618806e-01
1.26465476e+00 -8.31681848e-01 -5.28855860e-01 -3.84862572e-01
7.55410045e-02 -9.02680933e-01 9.36888039e-01 8.72748613e-01
-1.16075730e+00 -7.46215940e-01 -9.52844083e-01 -1.61090910e-01
-2.58916199e-01 -2.91419923e-01 4.10774440e-01 1.70904145e-01
-9.18695092e-01 4.95924264e-01 -8.22818339e-01 -3.51537108e-01
1.25803173e-01 -2.19429776e-01 -7.39975214e-01 -3.81626129e-01
-7.68447220e-01 6.60307288e-01 -3.21487430e-04 -3.99850190e-01
-9.44199800e-01 -5.87613881e-01 -1.01802015e+00 -1.61923796e-01
3.75948668e-01 -9.78197396e-01 9.85179961e-01 -9.68781412e-01
-1.11158419e+00 1.09980416e+00 -3.91145557e-01 2.91391816e-02
4.36969429e-01 -3.78469497e-01 -2.51087397e-01 5.15216172e-01
1.22882143e-01 5.53193748e-01 9.37769234e-01 -1.94263983e+00
-3.23249102e-01 -7.73097038e-01 3.92100394e-01 3.55217457e-01
8.44000578e-02 -3.19877565e-01 -4.74857986e-01 -6.11701906e-01
4.30248111e-01 -1.08606422e+00 3.85781936e-02 1.69106320e-01
-3.39236975e-01 1.49068847e-01 6.52747631e-01 -6.19109869e-01
5.91708243e-01 -2.40526414e+00 5.97818315e-01 8.87511298e-02
3.39367509e-01 -6.60573542e-01 1.51194092e-02 6.49250150e-01
-1.67356923e-01 -5.94232492e-02 3.08264289e-02 -4.65847790e-01
-1.76590517e-01 2.94300824e-01 -3.60937804e-01 7.16508329e-01
-3.05501461e-01 5.49527705e-01 -7.58217573e-01 -2.65734673e-01
4.44479495e-01 5.74148774e-01 -6.34987295e-01 3.33956689e-01
-7.24197477e-02 7.67652333e-01 -2.28887618e-01 4.37841922e-01
6.70189142e-01 -7.21991599e-01 7.42141902e-02 -4.54413712e-01
-5.65795749e-02 1.46899953e-01 -1.30930567e+00 2.05055475e+00
-5.42857528e-01 6.28284395e-01 1.10784411e-01 -9.25412238e-01
5.29856920e-01 2.33475223e-01 5.11747062e-01 -6.44227266e-01
-1.34879425e-01 -1.16590530e-01 -4.10268486e-01 -7.05946684e-01
1.10565700e-01 -2.36287624e-01 4.82341945e-02 3.56129467e-01
3.24887991e-01 -2.32949570e-01 -1.84014738e-01 3.67484331e-01
8.34684193e-01 2.22691700e-01 5.92463315e-01 -1.01059705e-01
9.67687368e-02 -3.93492430e-01 4.07818794e-01 6.21857166e-01
8.00789986e-03 8.89291286e-01 3.77408296e-01 -4.89499807e-01
-9.05986428e-01 -1.56539905e+00 -1.62611470e-01 1.12345350e+00
5.14323115e-01 -4.56555605e-01 -2.48512045e-01 -3.26667041e-01
1.59405675e-02 6.67429149e-01 -6.20220840e-01 -1.97565798e-02
-1.06599718e-01 -1.78627014e-01 -6.87249228e-02 4.54492688e-01
2.27493495e-01 -7.67559230e-01 -8.30841780e-01 -1.59641206e-01
-5.47953963e-01 -1.25082028e+00 -2.17336833e-01 1.74163938e-01
-6.54957831e-01 -9.73215878e-01 -5.20738602e-01 -4.61867929e-01
3.54848534e-01 6.15323126e-01 1.52142942e+00 -2.85369992e-01
-3.63816947e-01 1.09709322e+00 -3.54274958e-01 -3.09168279e-01
-5.26302047e-02 -5.91948509e-01 1.08987456e-02 2.29085445e-01
9.17977914e-02 -9.18051660e-01 -7.08821654e-01 6.47223517e-02
-7.50074983e-01 1.27017885e-01 3.75135809e-01 7.12572396e-01
8.46241713e-01 -1.48099095e-01 -4.76434901e-02 -9.50378001e-01
-6.13871217e-02 -5.31451046e-01 -1.66398287e-01 1.95082694e-01
-9.04851183e-02 -2.91854255e-02 5.98253310e-01 -7.61888623e-02
-1.11946630e+00 1.59781560e-01 1.55006707e-01 -7.05471933e-01
-5.51082909e-01 3.67400587e-01 -2.23663241e-01 8.86573642e-02
6.35144293e-01 2.44140208e-01 -4.84292172e-02 -5.24429262e-01
5.93458354e-01 1.36423245e-01 3.86087477e-01 -3.78542513e-01
5.52975714e-01 1.04442966e+00 -2.10111612e-04 -1.18607354e+00
-9.96869862e-01 -7.69612432e-01 -5.76907992e-01 -4.47498769e-01
9.58414078e-01 -1.29264510e+00 -6.76452875e-01 3.35447609e-01
-1.17126441e+00 5.92326447e-02 -1.70136571e-01 7.71596253e-01
-8.50187004e-01 3.87800425e-01 -5.50379872e-01 -6.76572382e-01
2.48541504e-01 -9.49926019e-01 1.26200271e+00 2.25219186e-02
-1.79206923e-01 -1.02550554e+00 2.15150937e-01 5.32260120e-01
-7.25352541e-02 3.65251154e-01 7.59832501e-01 -3.16032290e-01
-7.51433253e-01 1.51777253e-01 2.89613120e-02 5.15998185e-01
1.26419932e-01 -1.49275765e-01 -1.51491165e+00 -4.43121761e-01
4.72580552e-01 -7.14801490e-01 1.30224156e+00 6.46483898e-01
1.41472518e+00 -2.57000536e-01 -1.84601441e-01 7.72213995e-01
1.66805518e+00 -1.56931892e-01 4.68679011e-01 -1.00113891e-01
8.14723611e-01 9.47439432e-01 1.18921362e-01 5.30370235e-01
5.36774278e-01 5.18577099e-01 6.66942060e-01 1.33840382e-01
1.46719087e-02 -1.71852946e-01 3.96407217e-01 1.03602219e+00
-4.50592577e-01 -4.18096483e-01 -6.26525939e-01 3.06972533e-01
-1.67887497e+00 -1.22071826e+00 1.82966366e-01 2.16205621e+00
2.75570184e-01 5.07250912e-02 2.92987358e-02 -1.76948339e-01
3.14108580e-01 6.65108740e-01 -6.37811303e-01 -2.07740560e-01
-3.54799867e-01 5.81058487e-03 2.45204911e-01 5.62402844e-01
-1.15552151e+00 2.68650889e-01 5.88398981e+00 3.86910170e-01
-1.11980367e+00 -6.70239842e-03 6.78064346e-01 -3.96498650e-01
-6.56434655e-01 -1.52705655e-01 -2.71631062e-01 2.45310247e-01
5.31947136e-01 5.54184392e-02 6.29039526e-01 7.77403414e-01
7.50864148e-02 -2.21809819e-01 -1.47911477e+00 1.21630502e+00
5.25233269e-01 -1.21054459e+00 1.24855682e-01 1.80700108e-01
7.45753884e-01 1.82805240e-01 2.55736381e-01 -2.89915293e-01
1.40613317e-01 -9.33694899e-01 8.46882761e-01 9.18932378e-01
4.78235364e-01 -5.08095622e-01 3.39524478e-01 4.11572814e-01
-1.29475510e+00 7.15408325e-02 -2.27772117e-01 1.28753092e-02
1.15966275e-01 5.24338722e-01 -3.80258411e-02 7.30353296e-01
1.05430996e+00 1.17755127e+00 -4.38674062e-01 6.70801222e-01
-5.20808734e-02 4.32726800e-01 -7.81214759e-02 3.63928288e-01
9.21650790e-03 -1.38440028e-01 8.26642394e-01 9.59604979e-01
1.84174597e-01 1.11804634e-01 4.26932842e-01 8.38540852e-01
-1.29547231e-02 -1.01172775e-01 -1.14687884e+00 3.77686471e-01
1.31052583e-01 1.05720973e+00 -4.07025516e-01 -1.43991217e-01
-9.71425056e-01 1.07087803e+00 3.58646870e-01 6.30360603e-01
-5.13132811e-01 2.45188013e-01 5.97685158e-01 1.76948249e-01
5.42235851e-01 -1.02250807e-01 -1.61832973e-01 -1.49125016e+00
3.90392393e-01 -9.55011964e-01 4.55080360e-01 -1.00775218e+00
-1.71809232e+00 2.66861558e-01 3.15964371e-02 -1.42482162e+00
-2.56662875e-01 -6.74394190e-01 -3.81521374e-01 6.53999388e-01
-1.28299415e+00 -1.00879776e+00 -3.15058231e-01 7.56662130e-01
7.18825817e-01 7.96346590e-02 8.26461315e-01 8.62414762e-02
1.32001221e-01 6.16770312e-02 8.55581388e-02 -7.38438666e-02
8.07062864e-01 -1.44010413e+00 8.96290615e-02 5.19775748e-01
6.26478612e-01 5.04063606e-01 7.93875277e-01 -2.93276101e-01
-1.62467372e+00 -7.78508186e-01 5.32802880e-01 -5.53060710e-01
4.91283089e-01 -4.07620490e-01 -8.10038149e-01 8.20513606e-01
3.96613508e-01 2.04762995e-01 9.00893450e-01 3.35420579e-01
-9.89446282e-01 -4.29059044e-02 -8.11957121e-01 3.65456104e-01
9.99106228e-01 -9.44208264e-01 -5.99740505e-01 2.48107567e-01
5.23230493e-01 -3.10462177e-01 -7.29128897e-01 2.47120038e-01
8.90424192e-01 -1.51984692e+00 1.19746292e+00 -5.15226483e-01
8.39644790e-01 8.51046592e-02 -6.73075378e-01 -1.60294914e+00
-3.95157546e-01 -2.64669508e-01 -1.33952394e-01 6.74136281e-01
2.61156380e-01 -6.06278062e-01 3.56762707e-01 2.24251434e-01
1.18049204e-01 -8.10616255e-01 -7.63969779e-01 -7.13674366e-01
-7.07013607e-02 -3.87433290e-01 -3.14452127e-02 9.43037808e-01
-1.78107724e-01 6.22878969e-01 -6.83945894e-01 3.01949859e-01
1.00129139e+00 4.33581531e-01 6.20795906e-01 -1.15571499e+00
-7.02989161e-01 -1.14924684e-01 -5.35215020e-01 -1.21287358e+00
-3.47830057e-02 -7.56580174e-01 -1.88775733e-01 -1.32434034e+00
5.71092427e-01 3.15978795e-01 -2.57583678e-01 -1.10248521e-01
9.22707841e-02 4.17898521e-02 4.43738580e-01 1.56546831e-01
-6.49165988e-01 4.38779056e-01 1.36966908e+00 -1.29663602e-01
2.32187137e-01 -4.02353145e-02 -5.95779002e-01 1.24896312e+00
6.12616301e-01 -2.13328347e-01 -5.03641784e-01 -4.90009427e-01
5.07052720e-01 2.95370847e-01 8.68994892e-01 -8.12321246e-01
2.23029077e-01 -6.10343106e-02 6.20763958e-01 -5.17974555e-01
7.60235012e-01 -7.47577906e-01 1.84711590e-01 1.53404772e-01
-4.85874474e-01 1.27032861e-01 -1.94769382e-01 1.06250191e+00
-4.02298957e-01 1.90311089e-01 7.83297956e-01 -6.54169202e-01
-7.16156244e-01 2.30751157e-01 -1.70816928e-01 3.57880056e-01
5.73881447e-01 -1.20065823e-01 -3.46528769e-01 -8.02822411e-01
-9.75925565e-01 -1.35612682e-01 5.65937757e-01 4.48262572e-01
6.41556680e-01 -1.25538290e+00 -6.60549343e-01 1.67283386e-01
5.28015077e-01 -1.96592927e-01 5.20370007e-01 6.17332995e-01
-1.43251851e-01 1.46104753e-01 -1.36206761e-01 -8.98789167e-01
-1.33997560e+00 5.28811634e-01 1.92690760e-01 1.11513563e-01
-8.77657235e-01 8.80882680e-01 9.02511835e-01 -2.85648376e-01
2.68708020e-01 -1.70550391e-01 -1.62066743e-01 1.17641471e-01
5.41876853e-01 4.34779346e-01 -1.89957961e-01 -7.99845397e-01
-2.76129812e-01 8.79356623e-01 1.50380924e-01 -1.76313579e-01
1.39875603e+00 -4.24308300e-01 -1.48746923e-01 1.27801275e+00
1.54905021e+00 1.02503926e-01 -1.23534620e+00 -4.25064147e-01
-6.49371445e-01 -6.36748075e-01 -1.44815678e-02 -6.63650095e-01
-1.02069759e+00 9.90924776e-01 5.62358022e-01 4.74027157e-01
1.08142948e+00 5.90086341e-01 4.34619993e-01 4.05611873e-01
4.93808001e-01 -9.57481027e-01 4.33430105e-01 2.64321893e-01
1.13170457e+00 -1.51874793e+00 1.24910168e-01 -4.71820801e-01
-8.34941626e-01 9.88543272e-01 4.20714051e-01 -1.60197452e-01
1.04728782e+00 1.45354792e-01 -3.73877287e-02 -5.87302744e-01
-1.01273394e+00 -2.66701411e-02 4.65726018e-01 2.90910810e-01
4.49958205e-01 3.94188493e-01 4.87342149e-01 2.36238614e-01
-2.27985993e-01 -4.47459370e-01 4.07513171e-01 7.32044458e-01
-3.47661018e-01 -5.40568709e-01 -2.27474824e-01 5.05589604e-01
-3.79868418e-01 9.37480405e-02 -4.72515464e-01 6.53708994e-01
2.37345606e-01 8.96997631e-01 6.73030391e-02 -3.46094668e-01
3.55539083e-01 -7.29950741e-02 8.00606191e-01 -6.38100564e-01
9.94706079e-02 4.22509819e-01 -9.36846808e-02 -1.01589799e+00
-9.74218130e-01 -8.43420386e-01 -7.16701686e-01 -2.58411855e-01
8.76185149e-02 -3.28506857e-01 2.82498181e-01 9.17783558e-01
1.08082838e-01 2.44126976e-01 8.36350083e-01 -9.24621165e-01
-3.12591761e-01 -5.93503118e-01 -8.90636325e-01 7.98016489e-01
8.07298243e-01 -7.69084334e-01 -6.91252470e-01 4.57249939e-01] | [9.826175689697266, 0.23607824742794037] |
c23dc3cf-539e-4625-a6e3-fdd5efc296c9 | multimathbf3net-segmenting-flooded-buildings | 1812.01756 | null | http://arxiv.org/abs/1812.01756v1 | http://arxiv.org/pdf/1812.01756v1.pdf | Multi$^{\mathbf{3}}$Net: Segmenting Flooded Buildings via Fusion of Multiresolution, Multisensor, and Multitemporal Satellite Imagery | We propose a novel approach for rapid segmentation of flooded buildings by
fusing multiresolution, multisensor, and multitemporal satellite imagery in a
convolutional neural network. Our model significantly expedites the generation
of satellite imagery-based flood maps, crucial for first responders and local
authorities in the early stages of flood events. By incorporating multitemporal
satellite imagery, our model allows for rapid and accurate post-disaster damage
assessment and can be used by governments to better coordinate medium- and
long-term financial assistance programs for affected areas. The network
consists of multiple streams of encoder-decoder architectures that extract
spatiotemporal information from medium-resolution images and spatial
information from high-resolution images before fusing the resulting
representations into a single medium-resolution segmentation map of flooded
buildings. We compare our model to state-of-the-art methods for building
footprint segmentation as well as to alternative fusion approaches for the
segmentation of flooded buildings and find that our model performs best on both
tasks. We also demonstrate that our model produces highly accurate segmentation
maps of flooded buildings using only publicly available medium-resolution data
instead of significantly more detailed but sparsely available very
high-resolution data. We release the first open-source dataset of fully
preprocessed and labeled multiresolution, multispectral, and multitemporal
satellite images of disaster sites along with our source code. | ['Marc Rußwurm', 'Tim G. J. Rudner', 'Piotr Bilinski', 'Veronika Kopackova', 'Benjamin Bischke', 'Ramona Pelich', 'Jakub Fil'] | 2018-12-05 | null | null | null | null | ['segmenting-flooded-buildings'] | ['computer-vision'] | [ 4.10780728e-01 -1.99250236e-01 2.31439576e-01 -4.99889433e-01
-1.48851740e+00 -5.66751838e-01 3.90726238e-01 4.87267613e-01
-7.78970778e-01 6.05712235e-01 6.42602265e-01 -5.88858128e-01
-1.20826162e-01 -1.66798472e+00 -6.72499597e-01 -5.46384692e-01
-5.73179483e-01 4.57811147e-01 2.13809475e-01 -7.80642092e-01
-1.59238502e-01 7.65509665e-01 -1.40112579e+00 5.06292880e-01
7.31128573e-01 7.54895329e-01 4.90556180e-01 8.04911852e-01
4.89740223e-02 2.93700844e-01 -2.01568231e-01 2.43275747e-01
3.40660661e-01 1.55725166e-01 -1.02972436e+00 1.38046503e-01
2.51352668e-01 -9.24226761e-01 -2.79004484e-01 6.97591662e-01
6.98450685e-01 -1.91780820e-01 4.15119976e-01 -6.03616714e-01
-4.77853805e-01 8.32153201e-01 -8.16242039e-01 6.43631399e-01
2.15217561e-01 1.82460606e-01 7.48019159e-01 -6.23450160e-01
1.25354916e-01 1.21280479e+00 9.81381297e-01 -4.17365164e-01
-1.17758930e+00 -5.19751668e-01 3.66764702e-02 -2.46707007e-01
-1.79474771e+00 -5.33727705e-01 2.68536210e-01 -5.26590943e-01
1.28296340e+00 3.13881248e-01 4.25781786e-01 2.53739119e-01
-8.97221044e-02 3.21024507e-01 8.38421166e-01 -1.87394559e-01
3.03180460e-02 -6.32578433e-01 -6.37122095e-02 5.45736492e-01
1.91472292e-01 2.45159611e-01 -1.50244877e-01 -2.45797634e-01
7.96778440e-01 1.34795085e-01 -2.33563870e-01 5.58857024e-01
-1.20836282e+00 9.44745302e-01 1.05988538e+00 5.76252162e-01
-1.12602246e+00 1.54780343e-01 -1.63572785e-02 -4.63369638e-02
9.39733028e-01 -7.05121756e-02 -2.88465291e-01 3.92252862e-01
-1.62868285e+00 2.67571419e-01 5.87930083e-02 3.01437706e-01
1.08451188e+00 8.95428061e-02 1.79305717e-01 7.59919882e-01
3.78092766e-01 1.03212667e+00 -1.46043614e-01 -9.00205553e-01
8.67856622e-01 5.21793187e-01 2.65138090e-01 -9.39413011e-01
-9.20133471e-01 -5.68685830e-02 -1.11006117e+00 9.72047821e-02
-1.11225490e-02 -2.83971816e-01 -1.02152753e+00 1.39959037e+00
2.11116433e-01 5.73926046e-02 9.91719812e-02 9.37630177e-01
6.96201861e-01 9.97441292e-01 4.71628517e-01 2.69336581e-01
1.37666774e+00 -4.93438058e-02 -3.10751289e-01 -6.64861858e-01
3.50369066e-01 -2.73178160e-01 5.67437947e-01 -2.91643977e-01
-7.43373215e-01 -2.32888818e-01 -8.06962192e-01 2.23219573e-01
-6.92336142e-01 -1.57276392e-01 5.79664230e-01 4.01645362e-01
-1.41261196e+00 5.20153046e-01 -1.00207293e+00 -5.81417024e-01
9.05436814e-01 2.58049548e-01 -2.18978643e-01 -1.72170788e-01
-1.39256883e+00 9.44573641e-01 5.32923043e-01 4.95454580e-01
-8.12837362e-01 -6.96154952e-01 -1.15289140e+00 6.35100380e-02
-1.42175987e-01 -2.67246366e-01 8.66679609e-01 -2.32331201e-01
-4.12625939e-01 7.57982850e-01 9.43429694e-02 -5.01693428e-01
3.78988832e-02 5.08164987e-02 -2.56979197e-01 2.51228243e-01
7.44764745e-01 9.53770578e-01 2.22561046e-01 -1.13321483e+00
-1.03466666e+00 -5.55019438e-01 1.75204091e-02 2.94914186e-01
-2.10604250e-01 3.98109972e-01 9.42072347e-02 -4.86928731e-01
3.66278529e-01 -4.22950596e-01 -7.82002866e-01 -1.49781615e-01
-9.86982509e-02 4.25825566e-01 7.34390914e-01 -1.16595173e+00
1.10964406e+00 -1.89369750e+00 -2.34684661e-01 1.02142334e-01
-1.35028169e-01 8.65547508e-02 -3.12014908e-01 3.97519767e-01
-3.02230455e-02 5.67582071e-01 -1.16847301e+00 9.56350863e-02
-1.81749791e-01 2.82857955e-01 -5.89522302e-01 4.78721440e-01
2.14460805e-01 8.55034351e-01 -8.72321963e-01 -5.00947893e-01
4.98395145e-01 6.90628529e-01 6.26289919e-02 2.07937822e-01
-3.61114703e-02 6.11653149e-01 -5.16783297e-01 7.66914010e-01
1.07323599e+00 7.25098923e-02 -1.76730677e-01 -1.42502338e-01
-5.85913479e-01 -1.08329365e-02 -1.06580293e+00 1.41965187e+00
-5.64566076e-01 4.09500301e-01 3.91903102e-01 -8.70573819e-01
7.59576380e-01 5.14246464e-01 7.50077903e-01 -9.60303903e-01
-2.08126560e-01 1.63302079e-01 -9.04310703e-01 -4.60836321e-01
9.03663278e-01 -3.81184220e-01 -5.28411150e-01 9.19520855e-01
-4.44856077e-01 -4.25126195e-01 1.00711010e-01 7.05853105e-02
8.61424267e-01 -1.17959946e-01 2.07905814e-01 -1.35380194e-01
7.42476508e-02 3.51046950e-01 5.93786538e-02 5.78329086e-01
6.11086376e-02 8.12735796e-01 -1.90831527e-01 -8.84998918e-01
-1.12864494e+00 -9.84636903e-01 -2.99876839e-01 1.16832078e+00
-2.63938665e-01 2.32916877e-01 -4.54835117e-01 -1.12102009e-01
-3.98163684e-03 6.24934435e-01 -4.02351290e-01 6.28359139e-01
-7.21114814e-01 -1.88378835e+00 1.13034809e+00 5.88477314e-01
9.56607640e-01 -9.50147510e-01 -9.68665600e-01 5.57978272e-01
-8.27193677e-01 -1.15290403e+00 2.20355421e-01 1.56519622e-01
-8.35704446e-01 -9.23606396e-01 -6.66377068e-01 -4.00114208e-01
3.38616252e-01 4.80301142e-01 1.05631614e+00 1.01524651e-01
-3.00825983e-01 1.36207327e-01 -3.44643444e-01 3.38480994e-03
-2.34985769e-01 4.74977717e-02 -4.40164089e-01 -2.15106755e-01
1.02134787e-01 -8.04592431e-01 -4.49926972e-01 1.25346005e-01
-1.35628748e+00 9.56834629e-02 4.16179597e-01 1.14060536e-01
6.48971856e-01 3.85762393e-01 3.70873243e-01 -1.05288401e-01
3.93719584e-01 -9.28715587e-01 -7.35109806e-01 1.97792232e-01
-9.08849239e-02 -1.53790027e-01 6.40901178e-02 5.07307827e-01
-1.02221262e+00 4.68063146e-01 -4.68703896e-01 5.86974621e-01
-6.86555803e-01 1.11481667e+00 2.78311729e-01 -1.21573107e-02
7.87102640e-01 1.97157905e-01 -6.12230897e-01 -6.38693333e-01
5.43957055e-01 1.23447311e+00 1.01519704e+00 -2.98804939e-01
8.71302605e-01 9.56525981e-01 -2.92181224e-01 -1.17247605e+00
-7.42178261e-01 -6.00042224e-01 -1.04408205e+00 -1.80504963e-01
1.03566420e+00 -1.37990272e+00 1.37858450e-01 7.81933904e-01
-1.21462536e+00 -5.88189900e-01 -3.12174764e-02 1.39649168e-01
-2.01831520e-01 3.70034963e-01 -3.73167783e-01 -8.20667803e-01
-6.49200261e-01 -7.38688946e-01 1.38826656e+00 2.60478761e-02
1.89409927e-01 -8.36145580e-01 2.50222355e-01 2.36675978e-01
8.96383703e-01 8.93319786e-01 5.17473638e-01 3.38007249e-02
-6.55630171e-01 -1.40375555e-01 -6.46556199e-01 -1.36243850e-01
2.30057269e-01 -1.91703826e-01 -8.92932415e-01 -2.22566947e-01
-4.03064668e-01 -3.12817752e-01 1.47503161e+00 7.44474828e-01
6.28327429e-01 -4.32795733e-01 -4.11761791e-01 6.49310350e-01
1.79311502e+00 -3.09720039e-01 8.10893595e-01 6.89341903e-01
5.59524894e-01 5.82712650e-01 4.45543259e-01 7.44979382e-01
1.01357722e+00 4.25938457e-01 7.99725235e-01 -4.71904129e-01
2.45439216e-01 3.37356567e-01 9.18707550e-02 7.71810189e-02
-1.58721820e-01 -2.94133157e-01 -1.63619864e+00 1.23229086e+00
-1.62862408e+00 -1.24813390e+00 -3.80397469e-01 2.13270688e+00
5.88938951e-01 -4.85337228e-01 -2.40214895e-02 5.25008738e-02
7.85828948e-01 6.85841560e-01 -1.14935420e-01 1.30727276e-01
-5.40140152e-01 4.95887309e-01 1.20512962e+00 8.47817063e-01
-1.86185944e+00 1.05550277e+00 6.75125408e+00 2.91800380e-01
-9.48008537e-01 3.30381364e-01 5.74211359e-01 1.38852103e-02
-3.99128199e-01 -6.85933828e-02 -5.38616598e-01 2.63079852e-02
1.39113510e+00 2.11023360e-01 2.94199020e-01 2.65330672e-01
6.90622628e-01 -5.45267284e-01 -2.13516906e-01 3.56220931e-01
-4.90923285e-01 -1.59687984e+00 -2.04733863e-01 1.50626570e-01
6.60277843e-01 1.00117707e+00 -2.04579949e-01 -2.59776443e-01
1.02517128e+00 -1.05737376e+00 6.03163898e-01 4.80538279e-01
8.86678874e-01 -7.28882849e-01 6.56672060e-01 4.05225754e-01
-1.82846701e+00 -2.56460428e-01 -2.44027689e-01 -5.62989675e-02
7.27861643e-01 6.95934296e-01 -5.03935874e-01 7.81858325e-01
9.77448165e-01 6.19467080e-01 -6.30560935e-01 9.85111892e-01
-2.13442788e-01 5.20133078e-01 -9.31488514e-01 9.51012611e-01
6.17451191e-01 1.63662061e-01 1.62455738e-01 1.29612219e+00
5.36666989e-01 7.89313376e-01 4.76766765e-01 7.45956182e-01
2.37154692e-01 -2.15815932e-01 -6.74697101e-01 2.33512938e-01
4.47292954e-01 1.06929839e+00 -7.48565793e-01 -3.33326548e-01
-1.81562930e-01 5.84940255e-01 -4.44074571e-02 3.03226471e-01
-6.72096908e-01 -1.26442015e-01 4.12295103e-01 1.01570554e-01
3.83132249e-01 -5.82262099e-01 -5.72553039e-01 -8.39529932e-01
-3.33825231e-01 -4.41295356e-01 4.91566002e-01 -8.66538107e-01
-7.65789747e-01 8.31436157e-01 3.08038980e-01 -8.77449930e-01
-2.52118915e-01 1.06946692e-01 -6.68749690e-01 1.22237277e+00
-2.23093820e+00 -1.71088099e+00 -6.10797763e-01 4.27282244e-01
3.17534447e-01 2.66211390e-01 9.93970931e-01 2.76920587e-01
-3.73814881e-01 -5.12160599e-01 9.26825255e-02 4.18408662e-01
7.33180121e-02 -8.99917543e-01 1.07762015e+00 1.14872336e+00
-1.44849122e-01 -2.69079179e-01 3.76367778e-01 -8.76358330e-01
-6.23469591e-01 -1.74481905e+00 1.12043965e+00 3.49743851e-02
4.05956566e-01 2.18461975e-01 -1.01739955e+00 7.35587597e-01
1.54803386e-02 -1.41781345e-01 4.40543324e-01 -3.31213981e-01
-1.08266369e-01 -5.60424440e-02 -1.28269625e+00 2.23668441e-02
4.76815969e-01 -6.41082227e-01 -4.85634774e-01 6.29191041e-01
5.09372592e-01 -1.52571067e-01 -9.28060830e-01 4.80038643e-01
3.38732719e-01 -1.00372374e+00 1.16285658e+00 8.49935636e-02
4.22183663e-01 -3.23626280e-01 -7.40108967e-01 -1.18157649e+00
-5.57548404e-01 2.08356887e-01 6.28678143e-01 8.66455853e-01
4.92901921e-01 -3.13638955e-01 4.39375579e-01 5.84934115e-01
-1.48009956e-01 -8.68014153e-03 -1.07670164e+00 -4.01887268e-01
3.25107008e-01 -7.80014217e-01 1.02880573e+00 6.60016000e-01
-4.97935295e-01 -3.13689619e-01 -1.54493645e-01 1.18686819e+00
8.04630339e-01 3.52306753e-01 2.24544302e-01 -1.26984274e+00
4.75411087e-01 -2.28500098e-01 -1.18345656e-02 -6.59260869e-01
-1.15859985e-01 -6.51871383e-01 1.56380653e-01 -2.12254095e+00
1.05602577e-01 -6.78773880e-01 -1.03662154e-02 1.23088050e+00
6.66781589e-02 7.76897609e-01 -1.58760384e-01 2.15766564e-01
-1.28683195e-01 4.09376919e-01 3.38163227e-01 -3.94977927e-01
-2.38492951e-01 -1.04348443e-01 -5.93381345e-01 5.68961382e-01
8.88115525e-01 -7.95279443e-01 2.66147673e-01 -1.18839395e+00
1.10329643e-01 3.69722486e-01 9.14461672e-01 -1.18159711e+00
5.83961643e-02 -4.92234290e-01 2.71589309e-01 -1.13174200e+00
9.43557546e-02 -6.69964790e-01 4.33993608e-01 3.43499124e-01
-2.22446006e-02 -4.88213487e-02 4.94831711e-01 2.90608943e-01
-2.01991543e-01 1.49922580e-01 8.37688386e-01 -4.93382245e-01
-9.79593098e-01 5.22721350e-01 -8.38621318e-01 -2.09228769e-01
6.92816257e-01 -2.21603010e-02 -3.91029358e-01 -4.30500329e-01
-6.66971922e-01 5.21471977e-01 3.41883600e-01 2.18982905e-01
6.04158223e-01 -1.00451529e+00 -1.49121845e+00 2.84833401e-01
-8.32852870e-02 3.85007948e-01 4.61038172e-01 4.87377733e-01
-8.60215068e-01 4.33084846e-01 -3.47210258e-01 -5.86595118e-01
-8.94901812e-01 -6.03251532e-02 5.44801176e-01 -3.28117490e-01
-5.36554515e-01 5.58877647e-01 -2.60357201e-01 -6.46689355e-01
-4.30472493e-01 -4.70903426e-01 -4.44115311e-01 3.59334856e-01
7.77703285e-01 1.78008586e-01 1.22226916e-01 -1.41522300e+00
-4.89019811e-01 5.96158862e-01 6.10710382e-01 -6.22599483e-01
2.04197621e+00 -5.21886230e-01 -2.73241997e-01 5.38072642e-03
9.02633727e-01 -6.29567921e-01 -1.18504107e+00 -3.65475953e-01
1.61084626e-02 -2.47739986e-01 8.04592073e-01 -8.02284479e-01
-1.36239302e+00 9.89424527e-01 7.81554461e-01 1.32476062e-01
1.44937956e+00 1.37336537e-01 1.03608239e+00 4.05779332e-01
4.41947520e-01 -8.89813244e-01 -5.41800678e-01 4.17596579e-01
9.40655947e-01 -1.24730992e+00 3.45599979e-01 -2.98243370e-02
-4.14168179e-01 9.62635100e-01 -5.79252243e-02 9.04613584e-02
8.87796521e-01 5.61063230e-01 9.85924155e-02 -4.15932566e-01
-2.71800041e-01 -8.01031649e-01 -2.41002783e-01 7.78774917e-01
8.20207819e-02 3.06221515e-01 3.37450296e-01 3.00871462e-01
6.87938482e-02 6.23763427e-02 4.55912739e-01 1.10369182e+00
-1.14802170e+00 -8.59927773e-01 -9.74016130e-01 3.80224377e-01
-2.46942371e-01 -5.65118253e-01 1.31588399e-01 3.80483180e-01
1.90586716e-01 1.09984732e+00 2.98273921e-01 -2.05554307e-01
2.36801624e-01 -2.13922784e-01 -1.26094535e-01 -4.38738406e-01
-5.94553113e-01 -2.23286971e-02 9.37257111e-02 -5.42818546e-01
-8.53679657e-01 -8.17779779e-01 -1.17983532e+00 -4.90180105e-01
1.42776454e-02 3.39243449e-02 8.57554078e-01 1.15319824e+00
2.60702103e-01 1.91389486e-01 8.11425865e-01 -1.65417445e+00
2.71063601e-03 -9.38750446e-01 -5.94804168e-01 -1.60357490e-01
6.53599739e-01 -2.38403007e-01 -8.57657287e-03 1.10836572e-03] | [9.444518089294434, -1.3576301336288452] |
dc5b39c1-dd72-4827-ae38-d632373ea453 | dropgnn-random-dropouts-increase-the | 2111.06283 | null | https://arxiv.org/abs/2111.06283v1 | https://arxiv.org/pdf/2111.06283v1.pdf | DropGNN: Random Dropouts Increase the Expressiveness of Graph Neural Networks | This paper studies Dropout Graph Neural Networks (DropGNNs), a new approach that aims to overcome the limitations of standard GNN frameworks. In DropGNNs, we execute multiple runs of a GNN on the input graph, with some of the nodes randomly and independently dropped in each of these runs. Then, we combine the results of these runs to obtain the final result. We prove that DropGNNs can distinguish various graph neighborhoods that cannot be separated by message passing GNNs. We derive theoretical bounds for the number of runs required to ensure a reliable distribution of dropouts, and we prove several properties regarding the expressive capabilities and limits of DropGNNs. We experimentally validate our theoretical findings on expressiveness. Furthermore, we show that DropGNNs perform competitively on established GNN benchmarks. | ['Roger Wattenhofer', 'Lukas Faber', 'Karolis Martinkus', 'Pál András Papp'] | 2021-11-11 | null | http://proceedings.neurips.cc/paper/2021/hash/b8b2926bd27d4307569ad119b6025f94-Abstract.html | http://proceedings.neurips.cc/paper/2021/file/b8b2926bd27d4307569ad119b6025f94-Paper.pdf | neurips-2021-12 | ['graph-regression'] | ['graphs'] | [ 1.32617399e-01 2.63696134e-01 -1.91255003e-01 -1.06087990e-01
-3.68689686e-01 -6.21864557e-01 1.59057766e-01 1.44073874e-01
-6.18250608e-01 8.80615413e-01 -2.21759543e-01 -5.04879355e-01
-4.38605547e-01 -1.19058728e+00 -1.02886152e+00 -6.38257146e-01
-4.72291470e-01 2.65058309e-01 3.94354403e-01 -3.35305780e-02
-2.70949513e-01 6.78349495e-01 -1.26625741e+00 -6.77193105e-02
6.32878244e-01 4.00248468e-01 -3.60407442e-01 9.01377261e-01
7.30499625e-02 9.81911063e-01 -7.19302833e-01 -3.02036732e-01
4.32766736e-01 -8.28361630e-01 -7.27201760e-01 -2.56060630e-01
4.73038286e-01 -3.80522281e-01 -1.04448283e+00 1.53027034e+00
3.86978209e-01 3.02174360e-01 2.76959449e-01 -1.62732804e+00
-6.96276307e-01 1.60149574e+00 -2.84752876e-01 2.94674516e-01
-4.43687402e-02 1.86567873e-01 1.04889536e+00 -1.57915428e-01
7.42826104e-01 1.19612002e+00 7.25813568e-01 8.68638575e-01
-1.41834724e+00 -8.06422234e-01 3.16041142e-01 -2.09039778e-01
-1.18600619e+00 -4.20596004e-01 4.82646883e-01 3.31081934e-02
9.25486803e-01 2.08665833e-01 6.10365987e-01 1.17330956e+00
-1.57842949e-01 5.21548510e-01 5.23198128e-01 -2.80499846e-01
4.52954024e-01 -4.63304996e-01 6.97506666e-01 8.54815722e-01
9.61941719e-01 -3.95614328e-03 -4.55266356e-01 -1.10340565e-01
8.05837810e-01 -4.74635176e-02 -4.54929322e-01 -1.90749168e-01
-6.53572798e-01 9.15058017e-01 6.91945493e-01 2.48345226e-01
-2.34847814e-02 7.15390861e-01 4.05793577e-01 7.27600753e-01
2.58876503e-01 1.21131137e-01 1.84634775e-02 2.07953081e-01
-6.22142851e-01 -7.98600018e-02 1.12951779e+00 1.07371223e+00
6.89571977e-01 2.93710649e-01 -6.78190514e-02 2.86414355e-01
-8.91352296e-02 4.25753713e-01 1.92867279e-01 -9.44860518e-01
4.56349581e-01 6.27457321e-01 -5.60975850e-01 -8.44792128e-01
-5.01684904e-01 -3.74277920e-01 -1.13907588e+00 1.47532746e-01
5.84513724e-01 -4.66806918e-01 -9.63053942e-01 2.11021590e+00
-5.08014448e-02 1.59289435e-01 -2.69833077e-02 5.96844673e-01
7.47087538e-01 4.93776411e-01 -1.71857208e-01 1.34202078e-01
4.32311833e-01 -9.47137177e-01 -2.75812387e-01 -1.21478677e-01
1.09239626e+00 2.58235067e-01 7.81988680e-01 2.41402760e-01
-1.25384164e+00 1.42836496e-02 -1.12157166e+00 -5.99527881e-02
-2.49950483e-01 -3.14752311e-01 5.69531858e-01 9.56491590e-01
-1.50185120e+00 1.04845810e+00 -1.08585584e+00 -2.28406727e-01
4.97896373e-01 7.91079044e-01 -3.53790939e-01 -2.58214384e-01
-9.13028598e-01 2.69518554e-01 6.85404778e-01 3.32127869e-01
-1.07945824e+00 -4.21859592e-01 -9.76414084e-01 2.50028193e-01
2.69275635e-01 -5.81700981e-01 1.01368380e+00 -1.07918131e+00
-1.14259791e+00 6.30881369e-01 -2.40517762e-02 -8.62860382e-01
7.28184938e-01 1.76398769e-01 -1.20796680e-01 1.92080513e-01
-2.91390210e-01 2.54181564e-01 1.99634150e-01 -9.22278166e-01
-4.28223699e-01 -2.46813729e-01 5.35737157e-01 -5.90092912e-02
-4.02867466e-01 -6.27143607e-02 -4.23703313e-01 -5.00617884e-02
-1.18468076e-01 -7.90759504e-01 -3.85149270e-01 -1.07601546e-01
-9.22291040e-01 -1.90536246e-01 4.85994756e-01 -3.92639786e-02
9.87690687e-01 -2.33981919e+00 2.80508608e-01 5.68303287e-01
1.10491407e+00 3.51955354e-01 -5.69947064e-01 2.08684102e-01
2.03448877e-01 5.39693773e-01 1.94106493e-02 -4.61144120e-01
2.49959111e-01 4.25841570e-01 -2.86867358e-02 8.57374191e-01
-2.47627422e-01 8.79614770e-01 -7.96288133e-01 2.27743555e-02
4.20643948e-02 3.21944267e-01 -6.14820480e-01 -4.10290845e-02
-3.79778177e-01 -1.11006290e-01 -2.24102631e-01 8.54868814e-02
8.75029385e-01 -4.78158146e-01 5.65338552e-01 4.56682682e-01
4.35073107e-01 4.28593695e-01 -1.12650549e+00 9.96704936e-01
-1.16067424e-01 7.35471189e-01 3.31242204e-01 -8.01062703e-01
5.59404492e-01 -9.11370367e-02 -4.68964763e-02 -4.85905379e-01
4.75887179e-01 2.94407487e-01 2.10726008e-01 -4.67863455e-02
2.08649457e-01 1.29522070e-01 9.59823430e-02 6.82880878e-01
3.11977595e-01 5.97658932e-01 6.64153755e-01 5.70916057e-01
1.63014388e+00 -4.73164350e-01 -2.20863432e-01 -1.24639757e-01
1.22545585e-01 -4.61369574e-01 4.97268260e-01 1.50091493e+00
-3.44417334e-01 5.12852848e-01 1.34917915e+00 -3.56572509e-01
-1.02321768e+00 -1.16960490e+00 6.45444095e-01 1.24125433e+00
1.42179862e-01 -4.93480325e-01 -1.15859222e+00 -9.24037695e-01
-9.65841860e-02 3.74756724e-01 -8.56032848e-01 -4.64738697e-01
-4.90032494e-01 -7.83983290e-01 1.19581211e+00 4.84714389e-01
4.14516240e-01 -1.22025406e+00 -1.93968013e-01 -9.61775798e-03
3.08996439e-01 -1.17255545e+00 -2.29439914e-01 5.43832839e-01
-8.54799628e-01 -1.40302980e+00 -3.42845231e-01 -1.11002326e+00
9.55466509e-01 2.54904091e-01 1.01462531e+00 7.47452378e-01
1.10441647e-01 1.16639644e-01 -1.00960478e-01 2.61753976e-01
-6.79097593e-01 7.93079615e-01 -1.36999130e-01 -3.72863382e-01
1.96634024e-01 -9.02922153e-01 -1.56518564e-01 -1.99389726e-01
-1.18590689e+00 -1.36374995e-01 2.65854508e-01 4.07372624e-01
3.56148422e-01 1.89886302e-01 3.33218217e-01 -1.26575124e+00
8.67644608e-01 -5.47754407e-01 -1.04326952e+00 1.93397582e-01
-2.23958135e-01 2.89730519e-01 1.15743482e+00 -5.61014295e-01
-1.70797929e-01 -2.55063295e-01 -6.54539838e-02 -2.97039717e-01
-1.17049269e-01 5.66553473e-01 -2.15203613e-01 -4.41949040e-01
6.81230724e-01 4.31440733e-02 -5.84738478e-02 -2.43666723e-01
4.19638723e-01 8.99206847e-02 4.60755050e-01 -5.46875417e-01
8.66203725e-01 2.78879404e-01 1.54736012e-01 -7.16800213e-01
-6.20147347e-01 3.62895355e-02 -2.58803427e-01 -5.50777316e-02
3.76417994e-01 -4.90407437e-01 -1.17734897e+00 6.68716729e-01
-1.11867976e+00 -9.40014362e-01 -3.36732864e-01 2.30560675e-01
-1.18973576e-01 1.87542245e-01 -1.27206886e+00 -7.83286393e-01
-3.20247978e-01 -9.43713665e-01 2.16845140e-01 4.38624799e-01
9.40748006e-02 -1.18671393e+00 3.81131917e-02 -4.44141030e-01
4.22064453e-01 3.40824097e-01 1.05850947e+00 -1.10913622e+00
-7.59495020e-01 -6.42626882e-02 -3.70847464e-01 2.46856377e-01
-2.30897605e-01 -3.60955000e-02 -6.87006354e-01 -5.35008192e-01
-3.10915172e-01 -5.07102251e-01 1.19480824e+00 3.37892324e-01
1.06663346e+00 -5.13103962e-01 -2.83854514e-01 1.05405402e+00
1.70007348e+00 -1.64629534e-01 7.22718358e-01 3.54760349e-01
8.91682506e-01 -2.40876246e-02 -7.00349092e-01 1.27145112e-01
-1.32127842e-02 -2.29599565e-01 6.86141372e-01 -3.21851000e-02
-5.73057197e-02 -2.73694277e-01 4.97651905e-01 8.25969636e-01
3.50676961e-02 -9.66783047e-01 -6.47487938e-01 4.34516668e-01
-1.76812816e+00 -5.89353502e-01 -2.28611022e-01 2.06519723e+00
5.78645051e-01 3.85150492e-01 2.76960492e-01 2.37634048e-01
1.01560557e+00 2.51424938e-01 -8.27813447e-01 -5.19106686e-01
-4.51877087e-01 5.47654510e-01 1.03049195e+00 6.69076204e-01
-8.00959229e-01 9.22270954e-01 6.98015261e+00 5.29164732e-01
-7.68839836e-01 -1.35906652e-01 5.34293771e-01 -4.68645424e-01
-3.89751941e-01 -1.90345291e-02 -7.09898114e-01 1.34858072e-01
1.32416439e+00 -5.82125902e-01 9.71203744e-01 5.97425163e-01
-2.65682787e-01 1.90829277e-01 -1.28588462e+00 4.69397396e-01
-5.58840223e-02 -1.29796791e+00 2.77452409e-01 1.20856300e-01
7.02235997e-01 3.89415115e-01 2.47302372e-02 3.11442703e-01
1.10310662e+00 -1.27954113e+00 5.07584870e-01 -4.37970497e-02
5.11671960e-01 -1.18268335e+00 6.66417539e-01 3.05327505e-01
-9.14972842e-01 -2.25132424e-02 -5.57175398e-01 -2.16469482e-01
-1.88786954e-01 5.81187367e-01 -3.74244869e-01 6.53904557e-01
3.77687216e-01 3.80936027e-01 -6.02019668e-01 1.36178732e+00
-3.78178418e-01 8.35130215e-01 -6.70461357e-01 -3.56977284e-01
3.54486257e-01 -2.23712072e-01 5.07566452e-01 9.86662090e-01
1.42442733e-01 -1.70431808e-01 1.47977233e-01 1.20247102e+00
-9.42287922e-01 -2.85929888e-01 -6.77920997e-01 -1.90906510e-01
5.21019399e-01 9.67420042e-01 -9.46602404e-01 -1.78096563e-01
-1.85361698e-01 1.03503120e+00 9.40247715e-01 6.82656765e-01
-8.87574196e-01 -9.54647005e-01 5.72355926e-01 -1.23311348e-01
4.09144938e-01 -2.11465001e-01 1.35227874e-01 -1.00110483e+00
1.22104481e-01 -9.47351515e-01 5.45854449e-01 -3.93255800e-01
-1.11456037e+00 8.38906229e-01 -3.22774857e-01 -2.74203897e-01
1.54263183e-01 -6.76564813e-01 -6.88900471e-01 5.68054199e-01
-1.16599417e+00 -6.74726188e-01 -1.57246083e-01 5.46621203e-01
-2.36248940e-01 3.81708235e-01 6.27978206e-01 2.18078643e-01
-9.57458854e-01 9.27342594e-01 3.05438638e-01 7.83975780e-01
1.26503021e-01 -1.22653830e+00 7.33254075e-01 1.27591932e+00
1.83440581e-01 5.98077655e-01 4.95363653e-01 -6.24566317e-01
-1.30004656e+00 -1.42915964e+00 6.35160267e-01 2.55504638e-01
6.45717502e-01 -7.57492185e-01 -8.50730836e-01 1.27608907e+00
1.04891248e-02 2.90495634e-01 3.07555109e-01 1.71336353e-01
-6.48845077e-01 1.40036702e-01 -1.14077628e+00 8.88368785e-01
1.49117315e+00 -4.83815253e-01 -4.34601530e-02 1.41080141e-01
1.04965019e+00 -6.18710458e-01 -4.04367715e-01 -5.14121093e-02
1.92868233e-01 -1.09687662e+00 5.50711453e-01 -9.41249311e-01
1.30636066e-01 -4.23297659e-02 8.44499692e-02 -1.28977954e+00
-2.63820529e-01 -1.01079237e+00 -1.66194543e-01 1.05759525e+00
5.62368929e-01 -1.02459633e+00 1.10293388e+00 4.15529996e-01
2.33037204e-01 -3.57954830e-01 -9.27754581e-01 -1.18029535e+00
5.24040580e-01 -3.59963983e-01 7.26458788e-01 5.25749743e-01
-7.24699497e-02 2.63629526e-01 -2.51423001e-01 2.21185252e-01
7.50040531e-01 -2.21140578e-01 5.53599179e-01 -1.28144169e+00
-1.55556336e-01 -6.72345757e-01 -5.38109541e-01 -8.21648240e-01
4.79957730e-01 -1.24815321e+00 -8.14234838e-02 -1.65723789e+00
4.16453868e-01 -8.11194181e-02 -4.32738394e-01 6.32136762e-01
9.00005475e-02 2.23556116e-01 1.47780895e-01 -4.07344282e-01
-1.05356991e+00 2.31530800e-01 8.10401201e-01 -6.40703440e-02
-4.01864767e-01 -1.02720581e-01 -7.30026066e-01 6.39321864e-01
1.01960123e+00 -8.89093220e-01 -4.88534927e-01 -5.24915874e-01
4.17940795e-01 -2.52298594e-01 5.42020738e-01 -1.18144345e+00
3.90709907e-01 1.16576746e-01 6.43410087e-02 -2.30324596e-01
-2.72484303e-01 -3.74351650e-01 3.71832773e-02 6.90242410e-01
-6.96863651e-01 1.58095527e-02 1.67494401e-01 5.74057043e-01
3.24030280e-01 -2.18661740e-01 7.07255185e-01 2.03445524e-01
-5.07786684e-02 6.25551105e-01 -4.68613178e-01 6.51886165e-01
7.44072795e-01 -2.92979404e-02 -7.97376573e-01 -4.45845127e-01
-7.41784394e-01 3.98852795e-01 5.35942018e-01 -9.53995064e-02
5.42308748e-01 -1.04979098e+00 -4.82818991e-01 2.42496803e-01
-2.84754962e-01 1.77323833e-01 1.54133672e-02 6.83255494e-01
-7.65625358e-01 -5.34552038e-02 -3.75654958e-02 2.59413105e-02
-1.20339835e+00 6.79668248e-01 8.28643799e-01 -3.12139571e-01
-1.00403273e+00 7.43179917e-01 -5.34704700e-02 -3.98502380e-01
8.12294722e-01 -5.70833862e-01 2.46282309e-01 -5.14818370e-01
4.79613692e-01 5.05778909e-01 -9.53077711e-03 -1.63763940e-01
-2.57251650e-01 -2.20707692e-02 -2.35293537e-01 -2.32560769e-01
1.45144188e+00 1.65804356e-01 -3.21656108e-01 2.06214309e-01
1.38655841e+00 -9.87649187e-02 -1.05305386e+00 -1.64569408e-01
-2.71045845e-02 1.17788330e-01 -2.66456038e-01 -3.50435942e-01
-1.42500615e+00 5.94817817e-01 1.05872020e-01 5.15398443e-01
1.11684561e+00 -9.22114961e-03 8.95377398e-01 7.57091165e-01
2.23370805e-01 -8.16752017e-01 -5.30753195e-01 7.78428018e-01
2.02125624e-01 -4.78895187e-01 -3.32774371e-01 -3.28576177e-01
2.13502347e-01 1.03465617e+00 5.38119555e-01 -5.48198342e-01
4.37289357e-01 4.18909580e-01 -3.94973576e-01 -1.13102742e-01
-8.79637063e-01 -1.64334938e-01 -3.19570303e-01 8.23558927e-01
-1.19530983e-01 -7.71525055e-02 5.50236029e-04 6.78287089e-01
-1.47104695e-01 4.21588868e-02 8.76471221e-01 7.44423032e-01
-4.68466580e-01 -9.65327263e-01 1.88776195e-01 3.47511917e-01
-6.24330103e-01 -3.41761738e-01 -7.64409482e-01 9.72869515e-01
-1.89348534e-01 8.85496259e-01 -4.09440733e-02 -4.74518955e-01
2.12597623e-01 -2.00756311e-01 7.24933445e-01 -3.93286586e-01
-8.20658028e-01 -4.66254950e-01 2.24227473e-01 -5.71495593e-01
1.24703035e-01 7.84780830e-02 -1.45193708e+00 -1.03898454e+00
-2.26893738e-01 2.64798433e-01 1.83157399e-01 7.39535272e-01
3.70636135e-01 6.84708476e-01 3.84409368e-01 -3.51026148e-01
-5.72097659e-01 -5.97758532e-01 -8.13973725e-01 1.90460965e-01
7.12918520e-01 9.28717777e-02 -1.10567844e+00 -4.71359342e-01] | [6.89182710647583, 6.195714473724365] |
4c6c275f-adf4-4327-af88-ada2bbd56ae5 | soloist-few-shot-task-oriented-dialog-with-a | 2005.05298 | null | https://arxiv.org/abs/2005.05298v4 | https://arxiv.org/pdf/2005.05298v4.pdf | SOLOIST: Building Task Bots at Scale with Transfer Learning and Machine Teaching | We present a new method SOLOIST that uses transfer learning and machine teaching to build task bots at scale. We parameterize classical modular task-oriented dialog systems using a Transformer-based auto-regressive language model, which subsumes different dialog modules into a single neural model. We pre-train, on heterogeneous dialog corpora, a task-grounded response generation model, which can generate dialog responses grounded in user goals and real-world knowledge for task completion. The pre-trained model can be efficiently adapted to accomplish new tasks with a handful of task-specific dialogs via machine teaching, where training samples are generated by human teachers interacting with the system. Experiments show that (i) SOLOIST creates new state-of-the-art on well-studied task-oriented dialog benchmarks, including CamRest676 and MultiWOZ; (ii) in the few-shot fine-tuning settings, SOLOIST significantly outperforms existing methods, and (iii) the use of machine teaching substantially reduces the labeling cost of fine-tuning. The pre-trained models and codes are available at https://aka.ms/soloist. | ['Chunyuan Li', 'Shahin Shayandeh', 'Lars Liden', 'Jianfeng Gao', 'Jinchao Li', 'Baolin Peng'] | 2020-05-11 | null | null | null | null | ['end-to-end-dialogue-modelling'] | ['natural-language-processing'] | [-1.98270883e-02 3.78165454e-01 7.11951032e-02 -5.39485574e-01
-8.22614133e-01 -1.00287914e+00 9.72981751e-01 -3.13840777e-01
-3.77351791e-01 7.83344209e-01 4.30663943e-01 -3.14183563e-01
1.03179090e-01 -7.38964796e-01 -3.25738817e-01 -2.04309300e-01
4.03903693e-01 1.40166020e+00 4.86932158e-01 -1.07978940e+00
2.95054652e-02 8.73767212e-02 -9.30971444e-01 4.84409094e-01
7.61976719e-01 6.24617517e-01 2.93442369e-01 9.18347061e-01
-2.00184420e-01 1.27259302e+00 -8.72045875e-01 -6.92173839e-01
8.26035216e-02 -3.65961730e-01 -1.56983352e+00 4.94472682e-02
-2.14732513e-02 -6.49543762e-01 -4.24833059e-01 4.51955825e-01
4.22303945e-01 5.26793063e-01 6.79891229e-01 -1.15332174e+00
-8.09953809e-01 9.26196098e-01 2.56607026e-01 -1.02058940e-01
1.51003376e-01 7.91992247e-01 1.10095203e+00 -9.33535874e-01
6.22205853e-01 1.66227257e+00 4.28789347e-01 1.25269997e+00
-1.55154777e+00 -4.88018155e-01 -1.13316946e-01 -2.21230298e-01
-6.51690245e-01 -5.57820618e-01 4.31628674e-01 -6.32344842e-01
1.40604508e+00 6.10811338e-02 1.28286436e-01 1.53465545e+00
6.03266172e-02 6.32403731e-01 1.04536712e+00 -2.33823419e-01
1.61876336e-01 2.41515994e-01 2.59401470e-01 7.51239061e-01
-6.54185593e-01 3.95371579e-02 -4.43321437e-01 -3.31700355e-01
5.58586895e-01 -1.35955676e-01 1.64315984e-01 -1.78969368e-01
-1.28045058e+00 1.31490779e+00 3.78204674e-01 2.90288270e-01
-1.69489443e-01 3.39510404e-02 6.49955094e-01 6.39925241e-01
5.52416265e-01 8.68261933e-01 -6.84130907e-01 -2.22332045e-01
-7.97578022e-02 3.50126803e-01 1.20895672e+00 1.14187515e+00
7.31284320e-01 2.05711171e-01 -7.84049153e-01 1.06089377e+00
-4.85195816e-02 4.19041961e-01 7.95620680e-01 -1.40702689e+00
6.46991253e-01 7.69288003e-01 2.62908876e-01 1.35390893e-01
-3.57991278e-01 2.55517542e-01 -4.38097477e-01 1.63692370e-01
6.55164540e-01 -7.30172217e-01 -5.88547349e-01 1.83488786e+00
2.64958233e-01 -3.32395315e-01 3.69229019e-01 5.35256684e-01
1.04698300e+00 6.40068173e-01 3.30795556e-01 5.62563278e-02
1.45327973e+00 -1.43773890e+00 -5.52871883e-01 -4.46982652e-01
5.55089533e-01 -5.49049616e-01 1.53124523e+00 9.64057073e-02
-1.23400795e+00 -6.65971518e-01 -5.04446507e-01 -3.14249188e-01
-4.05221522e-01 -1.08525835e-01 5.48249304e-01 4.62002128e-01
-1.36836159e+00 2.27157414e-01 -4.45679784e-01 -4.23935384e-01
3.30599323e-02 3.55203301e-01 -1.34335179e-02 3.86705041e-01
-1.41442478e+00 1.23870254e+00 2.35797063e-01 -4.85849291e-01
-1.52897131e+00 -7.11600482e-01 -7.67052233e-01 2.98021615e-01
6.87727809e-01 -9.07614946e-01 2.40579033e+00 -3.99888963e-01
-2.37144804e+00 8.71700287e-01 1.13444701e-01 -5.43740928e-01
4.35750812e-01 -5.65877557e-02 7.85499141e-02 -1.01018921e-01
-5.55997863e-02 1.17669153e+00 7.50331759e-01 -9.89718437e-01
-4.69211966e-01 1.19762681e-01 5.45827508e-01 3.47694129e-01
-4.50462133e-01 1.46814689e-01 4.97751683e-02 -2.91543961e-01
-8.81594121e-01 -1.01426327e+00 -4.36743826e-01 -4.95769620e-01
-5.23620009e-01 -9.96474624e-01 7.63907492e-01 -2.50760972e-01
6.79820418e-01 -1.61805856e+00 3.19848001e-01 -5.70057929e-01
4.75697935e-01 4.27346051e-01 -3.93717736e-01 9.10448730e-01
2.41954312e-01 -1.28295928e-01 -1.39450669e-01 -4.87755597e-01
3.19692016e-01 1.68907419e-01 -6.90788567e-01 -2.94896781e-01
3.03851336e-01 1.27307904e+00 -9.60334063e-01 -2.69670814e-01
4.57345665e-01 -6.59544766e-02 -6.94533765e-01 1.09554625e+00
-9.53318119e-01 8.04396093e-01 -5.33091009e-01 1.56436127e-03
-2.44905241e-02 -4.33773816e-01 1.28139034e-01 1.68253899e-01
6.51683938e-03 9.33432221e-01 -2.48199403e-01 1.96770358e+00
-9.21154797e-01 4.14267451e-01 1.42724469e-01 -7.67951310e-01
1.10012209e+00 6.44768178e-01 2.24296004e-01 -3.98207247e-01
3.49183559e-01 -1.38352245e-01 -2.02635676e-02 -5.04767597e-01
6.38575792e-01 -1.32086352e-01 -5.13192415e-01 9.90840435e-01
7.43408322e-01 -6.23512566e-01 2.23017275e-01 4.86446738e-01
1.25966191e+00 -1.53379411e-01 2.74319798e-01 -2.11800590e-01
4.74870324e-01 3.08347285e-01 3.71636674e-02 8.17190766e-01
-1.28160939e-01 -3.25272717e-02 4.10185605e-01 -4.02917266e-01
-7.63131022e-01 -1.09882677e+00 3.08248252e-01 2.19771385e+00
-3.50787222e-01 -3.16593409e-01 -1.09056950e+00 -9.79831398e-01
-1.19180985e-01 9.68699157e-01 -5.56240022e-01 -2.47221991e-01
-5.40025294e-01 -2.79342532e-01 6.99624658e-01 2.89414912e-01
6.88017130e-01 -1.75758660e+00 -6.49174690e-01 2.72744328e-01
-3.42030853e-01 -1.19274640e+00 -7.53117204e-01 3.49434406e-01
-8.01077545e-01 -8.92035425e-01 -6.15548551e-01 -9.01794195e-01
2.09037721e-01 3.37704152e-01 1.53666961e+00 -1.34037687e-02
3.27629372e-02 6.45852685e-01 -2.89049476e-01 -1.53549254e-01
-1.13233030e+00 5.16137660e-01 -7.32089058e-02 -4.91459996e-01
4.43845242e-01 -5.71507990e-01 -1.61097869e-01 5.81491649e-01
-4.61634904e-01 1.84274063e-01 2.95839608e-01 1.14453435e+00
-3.09067547e-01 -5.49228609e-01 1.01704490e+00 -1.28472567e+00
1.35116243e+00 -4.62366730e-01 -7.18318045e-01 3.12846482e-01
-4.44000453e-01 3.15230876e-01 7.82871962e-01 -6.60553932e-01
-1.75463188e+00 7.28249028e-02 -7.96815827e-02 -1.77355528e-01
-3.74763280e-01 5.05238958e-02 1.53559729e-01 2.19686702e-01
1.27453530e+00 1.14689648e-01 7.65187340e-03 -5.06947219e-01
9.81560290e-01 8.91940057e-01 6.13400102e-01 -1.02538168e+00
8.12798321e-01 -1.07442245e-01 -5.94668806e-01 -4.98064190e-01
-9.00238812e-01 -2.29199246e-01 -6.30683422e-01 -1.36376798e-01
1.05548275e+00 -8.33017170e-01 -1.31408179e+00 3.58873516e-01
-1.38246727e+00 -1.44881451e+00 -4.47190821e-01 -7.93866292e-02
-8.64938378e-01 -1.34239957e-01 -1.10069764e+00 -7.87686348e-01
-7.54047334e-01 -1.27758026e+00 1.02165353e+00 2.45439470e-01
-5.21975398e-01 -1.16469693e+00 4.64403808e-01 7.27641404e-01
7.10775197e-01 -6.01644754e-01 9.92230654e-01 -1.28088295e+00
-3.38742822e-01 3.13018739e-01 6.13798276e-02 1.52581573e-01
1.78368241e-01 -4.80673969e-01 -1.23198199e+00 -2.33834773e-01
1.77061826e-01 -1.41461492e+00 3.59416038e-01 -1.04540549e-01
7.40944564e-01 -5.75911939e-01 -4.45103236e-02 -2.53647715e-02
5.12718141e-01 2.91914195e-01 1.81687981e-01 -4.71503884e-02
2.83794671e-01 1.01307428e+00 5.46336591e-01 4.22113568e-01
6.94057405e-01 9.69124496e-01 1.63260639e-01 1.09525666e-01
-1.47194430e-01 -2.82304078e-01 6.54983819e-01 8.07990789e-01
2.07838714e-01 -2.27585614e-01 -8.80774796e-01 4.62593466e-01
-1.95748103e+00 -8.07535648e-01 2.20305875e-01 1.73567307e+00
1.52064013e+00 8.91720783e-03 6.42245710e-01 -7.76651323e-01
6.28997266e-01 -6.84029907e-02 -7.68333852e-01 -5.71652412e-01
4.07135904e-01 6.17395282e-01 1.20867388e-02 9.06797707e-01
-9.22970772e-01 1.72529209e+00 6.23928881e+00 6.37751579e-01
-7.28565395e-01 5.29539347e-01 5.53781807e-01 1.34315133e-01
-1.06445653e-02 3.37878533e-04 -9.09468830e-01 1.19163185e-01
1.18287170e+00 -3.25811595e-01 8.70088339e-01 1.05270255e+00
-7.93752745e-02 3.99160057e-01 -1.21085727e+00 2.65313923e-01
-2.81939477e-01 -1.22592461e+00 -1.54753715e-01 -1.36820093e-01
5.33034623e-01 1.99600801e-01 1.17568016e-01 1.21358085e+00
1.73365140e+00 -1.00242865e+00 3.07643741e-01 1.40347138e-01
7.17277408e-01 -2.11994395e-01 3.09510052e-01 6.70101881e-01
-8.46177936e-01 -1.18360423e-01 -3.52291703e-01 -1.13662995e-01
1.14117689e-01 -2.22139716e-01 -1.63684988e+00 5.27983792e-02
5.73473155e-01 4.24619526e-01 -4.92656767e-01 1.12464830e-01
-6.34685576e-01 6.38463318e-01 1.02455303e-01 -3.24645996e-01
1.86627850e-01 -1.11087091e-01 3.39199275e-01 1.25968659e+00
-1.26932919e-01 2.52090096e-01 4.41601634e-01 1.32375371e+00
-3.84246111e-01 -2.49997631e-01 -6.32891893e-01 -6.06069900e-02
7.08255589e-01 1.66589212e+00 -8.82394835e-02 -4.53863263e-01
-2.39198878e-01 8.19904268e-01 7.74111867e-01 3.38297278e-01
-6.39814973e-01 -1.37429610e-01 6.00808620e-01 -3.99833798e-01
6.79558814e-02 -1.21130496e-01 -1.19987644e-01 -1.04802036e+00
-7.18793452e-01 -1.21673763e+00 4.21733707e-01 -8.81401539e-01
-1.65890181e+00 9.31933641e-01 1.58694968e-01 -7.93684185e-01
-1.06909966e+00 -6.12989008e-01 -9.80993092e-01 8.57841790e-01
-7.65456438e-01 -1.17195237e+00 -3.59617352e-01 9.06460762e-01
1.32870197e+00 -6.83953762e-01 1.34966171e+00 -2.56902933e-01
-5.49060225e-01 5.50744116e-01 -4.69382346e-01 2.47058108e-01
1.14848244e+00 -1.50269639e+00 9.47597504e-01 1.81977347e-01
-1.19995676e-01 5.58813214e-01 6.90834165e-01 -6.09864354e-01
-1.03574884e+00 -1.13603806e+00 7.25708783e-01 -1.11209810e+00
9.57417011e-01 -8.27050924e-01 -8.59519541e-01 1.07338047e+00
9.56920564e-01 -5.14201403e-01 5.16071558e-01 3.34679365e-01
-5.38658381e-01 2.74473816e-01 -1.15834355e+00 7.81058490e-01
9.23577249e-01 -6.54793084e-01 -1.07932806e+00 7.89984047e-01
1.25101876e+00 -5.37026107e-01 -9.67120826e-01 4.04857919e-02
1.22242205e-01 -7.34407961e-01 1.06593859e+00 -1.05627716e+00
4.79301184e-01 3.68245333e-01 1.90889686e-01 -1.73229253e+00
-3.49266112e-01 -1.09438062e+00 -1.29433140e-01 1.27263427e+00
5.69228709e-01 -6.38968885e-01 4.68888789e-01 8.16760302e-01
-3.97034675e-01 -4.40198451e-01 -7.12299287e-01 -6.51462436e-01
5.99646091e-01 2.03837976e-01 5.67605257e-01 9.18558955e-01
5.57429254e-01 1.41818261e+00 -4.84740078e-01 -6.91162765e-01
1.59818232e-01 -1.26566766e-02 1.24793649e+00 -1.31080008e+00
-5.57587624e-01 -3.96100909e-01 6.78151608e-01 -1.35608673e+00
6.19976342e-01 -8.55907142e-01 3.93655360e-01 -1.22875929e+00
1.38640761e-01 -5.47716796e-01 2.06433356e-01 9.66609240e-01
-2.87743449e-01 -1.81429431e-01 1.97088242e-01 2.72873014e-01
-7.59824097e-01 9.12771821e-01 1.33126879e+00 -2.75665492e-01
-5.17393947e-01 2.85724431e-01 -7.01236129e-01 5.31389773e-01
1.00753748e+00 -4.17110890e-01 -7.03453600e-01 -2.97768623e-01
-2.12302715e-01 4.61834610e-01 1.36713564e-01 -6.10438466e-01
3.66471380e-01 -4.08616185e-01 -1.92829266e-01 -1.55619197e-02
7.61746168e-01 -4.19567078e-01 -3.91342252e-01 4.92733151e-01
-9.43313837e-01 4.15593013e-02 2.04032868e-01 2.51033753e-01
2.17620537e-01 -4.43398505e-01 7.61857748e-01 -2.79613107e-01
-4.60621625e-01 6.03163205e-02 -8.20971847e-01 4.11108762e-01
7.52096891e-01 5.95070779e-01 -1.13526094e+00 -8.65967035e-01
-4.54301119e-01 5.10646701e-01 1.40585303e-01 6.93428695e-01
2.59717792e-01 -1.01739442e+00 -7.50450015e-01 -1.89836398e-01
2.56045163e-01 -6.78749233e-02 1.04703724e-01 4.84186977e-01
4.69885357e-02 5.76724410e-01 -2.79088497e-01 -4.10048336e-01
-1.01477253e+00 5.79950213e-01 4.65347409e-01 -7.31196880e-01
-3.87612790e-01 8.34425867e-01 5.71152508e-01 -1.24619150e+00
1.58267289e-01 -4.35890071e-02 -3.06282401e-01 -2.89038241e-01
2.95859188e-01 1.07529052e-01 -2.08923593e-01 -1.99256659e-01
1.35899693e-01 -2.02769220e-01 -3.38186264e-01 -5.90199888e-01
1.09932756e+00 1.66911542e-01 1.03103317e-01 4.33534682e-01
5.89747012e-01 -2.95033783e-01 -1.42051828e+00 -7.19075859e-01
2.79458556e-02 1.86785519e-01 -5.42810500e-01 -1.22557950e+00
-5.66803694e-01 1.10889709e+00 1.38173833e-01 3.76820892e-01
7.37605453e-01 2.23295048e-01 9.09797668e-01 1.18018603e+00
3.54596674e-01 -9.31132495e-01 8.69493127e-01 1.26005864e+00
9.44878817e-01 -1.22308242e+00 -5.82031786e-01 -2.21804440e-01
-1.25692689e+00 9.36647952e-01 1.22475946e+00 -1.07607350e-01
1.55872807e-01 1.98342204e-01 2.46872976e-01 -2.87184834e-01
-1.38088310e+00 -2.28801951e-01 5.59453852e-02 7.26041913e-01
5.44686377e-01 1.21742506e-02 2.94002444e-01 8.38853896e-01
-5.00660956e-01 -2.16669276e-01 4.21221972e-01 6.38734460e-01
-6.80806875e-01 -1.37341917e+00 1.34605914e-02 1.65743008e-01
4.08329582e-03 -1.48223072e-01 -1.05031455e+00 6.64066315e-01
-5.56814134e-01 1.45427215e+00 -1.89743042e-01 -6.33090734e-01
3.80575866e-01 5.77503681e-01 2.88928628e-01 -1.23927510e+00
-1.43825436e+00 -3.08849931e-01 5.18138051e-01 -4.00837660e-01
6.18982762e-02 -7.36236721e-02 -1.10620260e+00 -3.50608498e-01
-1.27973765e-01 4.31574941e-01 1.78035840e-01 1.06929314e+00
4.33685720e-01 6.57999456e-01 6.99814737e-01 -8.98338199e-01
-1.26641321e+00 -1.67125070e+00 1.78216130e-01 4.33319807e-01
-4.69900295e-02 -6.65168643e-01 -1.78291932e-01 1.98352277e-01] | [12.821136474609375, 8.061004638671875] |
9ac72554-b1a8-48ed-8e0a-9d1831ec8ba2 | augmented-balanced-image-dataset-generator | null | null | https://www.ijrar.org/viewfull.php?&p_id=IJRAR22B1901 | https://www.ijrar.org/papers/IJRAR22B1901.pdf | Augmented Balanced Image Dataset Generator Using AugStatic Library | The mixed data consists of various structured and unstructured data. The exponential boom of the amount of data has made the datasets of varying samples. This paper focuses on the image dataset generator that balances an imbalanced dataset using the AugStatic augmentation library. The datasets, including various classes, are said to be balanced if the number of samples in the classes is equal. This gives a fair chance for the model to learn about all the classes. An augmented image dataset balances an imbalanced image dataset. It is useful when the data is less in a specific category, generating new data with it. There are multiple augmentation techniques supported by the AugStatic library that helps in developing the augmented balanced library by the iterative implementation of the augmentations on the generated dataset. It takes an input of the existing dataset, majority and minority classes sample count that returns a balanced image dataset by iteratively applying the augmentations on the generated augmented images in the minority class. This generator is efficient and can be used for any image dataset. | ['Dr. Venkateswara Rao Gurrala', 'Allena Venkata Sai Abhishek'] | 2022-05-01 | null | null | null | international-journal-of-research-and-1 | ['image-manipulation-detection', 'image-smoothing', 'image-matting', 'image-variation', 'image-stitching', 'image-augmentation', 'image-manipulation', 'image-morphing', 'roi-based-image-generation', 'image-cropping', 'detecting-image-manipulation', 'data-visualization', 'data-visualization'] | ['computer-vision', 'computer-vision', 'computer-vision', 'computer-vision', 'computer-vision', 'computer-vision', 'computer-vision', 'computer-vision', 'computer-vision', 'computer-vision', 'computer-vision', 'methodology', 'miscellaneous'] | [ 5.80710471e-01 2.57548600e-01 -2.51657695e-01 -5.88736892e-01
-4.72858548e-01 -3.92161667e-01 4.93640035e-01 3.63905847e-01
-5.26248962e-02 7.08662927e-01 1.72112063e-01 1.98820815e-03
7.92793930e-02 -1.05363429e+00 -4.90774214e-01 -7.28943169e-01
2.03802273e-01 1.01215518e+00 2.34886542e-01 -1.97344959e-01
4.39093679e-01 3.88909876e-01 -2.17620659e+00 9.20133591e-01
6.80379629e-01 8.93019795e-01 -2.76511516e-02 9.30382133e-01
-5.37599921e-01 1.14023817e+00 -1.01605058e+00 -9.76712406e-02
7.86054850e-01 -5.91123819e-01 -7.94388950e-01 4.97352421e-01
5.99780738e-01 -1.73339993e-01 1.96613804e-01 9.04882014e-01
5.25181532e-01 -2.58446455e-01 8.24057519e-01 -2.00603771e+00
-1.37652516e-01 5.59285641e-01 -1.11689448e+00 3.19632977e-01
1.72247171e-01 -1.39253577e-02 4.36888665e-01 -7.87572026e-01
8.46481681e-01 1.26855624e+00 4.03826058e-01 1.28651783e-01
-9.71859753e-01 -7.94346929e-01 -2.28730083e-01 3.39163803e-02
-8.67687225e-01 -8.24247580e-03 6.07242405e-01 -6.51735127e-01
6.08413398e-01 5.20925939e-01 7.81004846e-01 2.59092689e-01
-9.85164195e-02 6.95826232e-01 1.39852774e+00 -7.72211194e-01
1.25409782e-01 4.35605526e-01 4.04552281e-01 1.74894184e-01
5.93203187e-01 -9.48214158e-02 -4.49617445e-01 3.33183594e-02
2.67581731e-01 1.23893969e-01 1.99875459e-01 -4.17406768e-01
-9.75575745e-01 8.46163929e-01 3.93435568e-01 2.07491115e-01
-3.56056780e-01 -3.83648694e-01 6.57315314e-01 6.18567467e-01
5.58067322e-01 3.48047733e-01 -4.10701454e-01 2.30873272e-01
-1.06307030e+00 6.36308074e-01 5.61892092e-01 9.24805164e-01
1.14046204e+00 -1.58227667e-01 9.67688113e-02 7.71950543e-01
4.11383398e-02 4.42820013e-01 6.03460789e-01 -5.23477197e-01
7.68541992e-01 1.54855776e+00 -8.81315991e-02 -1.08421659e+00
-3.33277076e-01 -2.33466640e-01 -1.01442087e+00 8.49005461e-01
2.66410381e-01 1.20072022e-01 -1.45487356e+00 1.41514575e+00
7.85867810e-01 -4.05252934e-01 1.48176506e-01 3.75972003e-01
1.05155241e+00 8.41422260e-01 -2.64171332e-01 -1.37822449e-01
1.07025039e+00 -7.91676283e-01 -6.70777798e-01 -1.61251083e-01
5.49894810e-01 -1.07182050e+00 1.04209006e+00 6.01635098e-01
-9.75723505e-01 -8.69398057e-01 -1.21564555e+00 -3.34446430e-02
-7.22544074e-01 1.28213447e-02 1.70827955e-01 7.12692559e-01
-9.18534935e-01 4.38764021e-02 -1.96719378e-01 -1.98029101e-01
4.63201880e-01 4.70480382e-01 -5.10181665e-01 -4.83985059e-02
-7.02059090e-01 7.87192106e-01 7.74925411e-01 -3.35313678e-01
-7.56941438e-01 -7.17443585e-01 -8.32878768e-01 -4.13855106e-01
-1.09043598e-01 -4.67394084e-01 8.57758343e-01 -1.71421885e+00
-5.39535880e-01 1.35460615e+00 2.21740142e-01 -4.23012525e-01
8.15645814e-01 3.62415552e-01 -9.21978429e-02 -3.30465809e-02
3.37140620e-01 1.05822742e+00 8.36834669e-01 -1.68619502e+00
-9.31933641e-01 -6.68021560e-01 -2.04256535e-01 4.23549771e-01
-5.46324402e-02 -1.46682248e-01 -1.36059523e-01 -7.21532106e-01
1.94061190e-01 -8.04955244e-01 -1.50315106e-01 -1.87931895e-01
-5.96647799e-01 2.95487165e-01 1.40678453e+00 -4.71361160e-01
1.10310721e+00 -2.08341384e+00 -3.38059336e-01 5.37137687e-01
1.79997474e-01 1.82964459e-01 -1.82832077e-01 3.68289590e-01
-7.76947200e-01 2.30703458e-01 -3.00427377e-01 -1.61228254e-01
-3.56286883e-01 2.10167184e-01 -1.28489155e-02 2.93802947e-01
1.30117252e-01 3.42270136e-01 -5.58873653e-01 -5.94475389e-01
2.56630361e-01 1.18297830e-01 -6.64847076e-01 5.48674524e-01
-1.41109213e-01 3.27385336e-01 1.18335553e-01 4.91455466e-01
1.35393250e+00 1.88770488e-01 -5.00598736e-03 -3.10426772e-01
-1.26740932e-01 -5.42829894e-02 -1.73661172e+00 8.51699889e-01
5.14134355e-02 5.65790951e-01 -4.11202818e-01 -1.11330652e+00
1.32458770e+00 4.12098318e-02 4.95691866e-01 -8.59631002e-01
4.68207076e-02 2.51823008e-01 8.96007270e-02 -7.32419431e-01
7.08984375e-01 -3.24366659e-01 -8.43737498e-02 7.53639817e-01
-5.08384705e-02 -6.95513070e-01 7.84057319e-01 3.87623340e-01
6.34831131e-01 1.75878243e-03 4.38820869e-01 -1.66735932e-01
4.78599846e-01 3.94937932e-01 2.16687843e-01 6.15664661e-01
7.87603855e-02 8.28895450e-01 7.87026525e-01 -7.99437046e-01
-1.80999148e+00 -7.64216244e-01 3.10846865e-02 7.79405117e-01
-1.83224961e-01 -2.77298950e-02 -8.73720407e-01 -5.30694485e-01
-1.59076318e-01 -1.09256823e-02 -9.94856954e-01 -7.81285390e-02
-2.06901550e-01 -9.36953723e-01 2.42518619e-01 1.71425030e-01
8.50389302e-01 -1.34405065e+00 -7.18080282e-01 -2.67026871e-01
-1.57400906e-01 -5.38461804e-01 -5.48933782e-02 4.42302495e-01
-9.30439830e-01 -1.32308781e+00 -3.93921554e-01 -1.08850503e+00
1.24733698e+00 1.85229495e-01 1.38624144e+00 2.27791533e-01
-5.34398198e-01 -2.29204491e-01 -4.33009237e-01 -1.09777164e+00
-7.86608338e-01 9.93041545e-02 -3.45903814e-01 3.89974713e-02
4.09614325e-01 -1.87424347e-01 -3.68454635e-01 1.56382084e-01
-1.58835673e+00 4.53868210e-01 2.22667322e-01 6.01986766e-01
6.76291823e-01 5.58459818e-01 5.42024195e-01 -1.35241020e+00
3.81224155e-01 -6.06510758e-01 -4.07710433e-01 -1.20872043e-01
-2.88691103e-01 9.96156875e-03 3.60526592e-01 -2.93000638e-01
-7.72340357e-01 1.95488244e-01 1.48830041e-01 5.53079471e-02
-2.43013620e-01 3.43147814e-01 -5.07117212e-01 3.51867825e-01
8.26142728e-01 -2.42694214e-01 3.71562153e-01 -1.66302815e-01
2.84240156e-01 9.91475940e-01 3.69496733e-01 -1.73326321e-02
6.22202754e-01 4.92009372e-01 -1.20810382e-01 -6.41408801e-01
-6.55352294e-01 -5.11039078e-01 -6.54409826e-01 -3.93128544e-01
5.97913444e-01 -5.97345471e-01 4.60512340e-02 9.66133893e-01
-7.23157585e-01 -2.19862282e-01 -7.19944000e-01 -8.33532810e-02
-3.14914823e-01 -5.45655154e-02 -9.67118249e-04 -6.95272267e-01
-3.43294591e-01 -1.32512856e+00 7.80851305e-01 3.41287345e-01
-3.68493706e-01 -4.99553502e-01 2.83506215e-01 5.15517652e-01
2.84169316e-01 6.84743106e-01 1.15239823e+00 -9.20397162e-01
-3.94684374e-01 -5.09777308e-01 -2.51563221e-01 6.90478265e-01
3.70999277e-01 3.94051850e-01 -8.78464878e-01 -7.25477040e-02
3.84340659e-02 -5.15941143e-01 5.83644152e-01 3.01065385e-01
1.16276491e+00 -4.37058777e-01 -3.78821604e-02 2.41280898e-01
1.64277852e+00 4.99615759e-01 1.01335847e+00 5.82426190e-01
4.99650538e-01 8.14906478e-01 7.88999975e-01 4.38766778e-01
3.82731199e-01 2.59546310e-01 6.69957101e-01 -6.37702584e-01
-1.61444515e-01 1.23827063e-01 -1.48159191e-01 5.43178678e-01
3.50893378e-01 -2.57374763e-01 -9.37132418e-01 6.80545628e-01
-1.47076237e+00 -1.27834642e+00 -5.53508461e-01 2.33537030e+00
1.13348758e+00 2.47697100e-01 4.14372683e-01 9.73730564e-01
9.13228631e-01 6.02149107e-02 -4.63726521e-01 -7.98329890e-01
-2.67728955e-01 3.03557783e-01 3.63757998e-01 5.78293025e-01
-1.12758350e+00 2.90855706e-01 6.79828501e+00 5.69020510e-01
-1.03004706e+00 -1.63875356e-01 1.36873007e+00 -5.07613234e-02
-4.15631205e-01 -8.80516395e-02 -8.59621882e-01 5.76972246e-01
5.55336893e-01 -1.56066045e-01 -8.97239968e-02 7.99851537e-01
2.11983219e-01 -7.25021601e-01 -6.90819502e-01 8.36170495e-01
3.19130599e-01 -1.19149983e+00 4.85114336e-01 5.27586676e-02
1.31724977e+00 -2.35419616e-01 9.60180536e-02 7.93903768e-02
1.49970204e-01 -9.82155561e-01 7.03492641e-01 3.74965847e-01
7.47134030e-01 -1.03078949e+00 1.07422471e+00 4.79009151e-01
-6.64490521e-01 -1.09468482e-01 -2.69903779e-01 -1.92997321e-01
-4.15747523e-01 9.72060025e-01 -1.19163644e+00 1.91182569e-01
6.14099085e-01 2.50386506e-01 -9.25531924e-01 1.19128239e+00
3.39691758e-01 2.08047092e-01 -3.93967748e-01 4.23019856e-01
-5.90859391e-02 -2.11127400e-01 6.47352412e-02 1.04384243e+00
-2.38884408e-02 -3.98730755e-01 2.52261251e-01 1.93960026e-01
-1.15166843e-01 3.62134486e-01 -9.24815297e-01 2.43466809e-01
2.89023370e-01 1.33349764e+00 -9.70482051e-01 -8.47252309e-01
-5.89596294e-02 5.95568120e-01 -1.20258726e-01 -1.54490307e-01
-5.83813012e-01 -4.24343705e-01 1.13478169e-01 4.43629026e-01
-5.19794486e-02 3.91647756e-01 -8.63554597e-01 -5.29342651e-01
2.10937429e-02 -1.65386331e+00 7.50480294e-01 -9.87395465e-01
-9.50921953e-01 8.11944306e-01 2.37167999e-01 -1.44603539e+00
-2.47014135e-01 -3.57145727e-01 -2.88126349e-01 8.59429479e-01
-1.11055529e+00 -1.02534723e+00 -1.04180527e+00 5.08899510e-01
5.56838512e-01 -1.82715610e-01 6.02216959e-01 3.82087022e-01
-2.59172469e-01 1.22918911e-01 -2.49307364e-01 3.22265923e-02
6.97388291e-01 -1.42310750e+00 1.56088129e-01 8.31193745e-01
-1.54497549e-01 1.44465640e-01 8.19158316e-01 -9.14905012e-01
-5.47136247e-01 -1.06277835e+00 8.21407795e-01 -3.44161063e-01
-8.16601049e-03 -1.92713946e-01 -7.17764258e-01 5.79045951e-01
5.09509206e-01 7.50230700e-02 8.47450137e-01 -7.78037131e-01
-2.80913174e-01 -3.79219711e-01 -1.71368790e+00 2.00373933e-01
4.37650472e-01 -1.49679054e-02 -4.82817769e-01 3.51174951e-01
2.61638165e-01 -5.78187466e-01 -3.86325449e-01 4.12370145e-01
3.05503190e-01 -1.24772465e+00 6.53043687e-01 -5.46245337e-01
8.98383975e-01 -6.95638955e-01 -7.14146644e-02 -1.18068147e+00
1.01869628e-01 1.60005838e-02 3.14277828e-01 1.31935596e+00
5.53119957e-01 -4.38192040e-01 9.22505975e-01 3.21903408e-01
3.74772549e-01 -6.22149467e-01 -4.50929970e-01 -1.83589935e-01
-1.46688163e-01 -1.64351448e-01 9.25748825e-01 8.58069360e-01
-4.62195188e-01 8.33467543e-02 -2.56298900e-01 -1.65354833e-01
5.67189634e-01 2.64331907e-01 1.32322991e+00 -1.25826371e+00
1.55351117e-01 4.11043549e-03 -7.99213290e-01 -3.17622185e-01
-5.29453516e-01 -7.75245607e-01 -2.98593193e-01 -1.65353048e+00
6.20547116e-01 -8.74126673e-01 2.18895912e-01 6.64606929e-01
-1.66863397e-01 1.06365514e+00 1.30486861e-01 2.27263615e-01
-5.00372797e-02 -1.54296383e-01 1.04415131e+00 -1.81412935e-01
-2.31717348e-01 -1.32419840e-01 -8.13280106e-01 7.25865126e-01
1.19587290e+00 -7.10625768e-01 -6.81056559e-01 6.65520728e-02
5.42307079e-01 -5.12879610e-01 -8.46856534e-02 -1.14924836e+00
-2.34281480e-01 -2.91916728e-02 7.61957824e-01 -1.31388843e+00
-1.43985763e-01 -1.01594329e+00 5.53513587e-01 5.59654951e-01
-4.32747841e-01 4.79330182e-01 1.20718524e-01 -2.34960228e-01
-4.34093118e-01 -4.60599899e-01 1.32442510e+00 -5.45272887e-01
-4.46235001e-01 2.23851725e-01 -1.19973689e-01 2.88485318e-01
1.39189684e+00 -8.57596457e-01 -2.64612734e-01 -4.40405160e-01
-6.65298700e-01 2.61760727e-02 7.66496778e-01 3.87011170e-01
3.34668249e-01 -1.30838454e+00 -7.13391721e-01 7.78230727e-01
2.54682064e-01 3.67511123e-01 1.46435499e-01 4.37981516e-01
-1.11093867e+00 -2.69372195e-01 -6.95218503e-01 -7.30944574e-01
-1.67923462e+00 6.53314590e-01 4.36989099e-01 -4.61302131e-01
-2.62230247e-01 4.15475279e-01 6.81683347e-02 -7.23737299e-01
-5.26550738e-03 -3.09719205e-01 -5.66815197e-01 7.35693991e-01
8.84824395e-01 3.57607037e-01 3.68203551e-01 -7.05302060e-01
-7.35348165e-02 4.36707228e-01 -1.42841592e-01 -2.14813650e-01
1.55629480e+00 3.88149009e-03 -5.70140302e-01 5.18316448e-01
1.33657670e+00 1.02665827e-01 -8.61388981e-01 2.51604944e-01
-1.76211759e-01 -7.49523938e-01 -3.64092320e-01 -8.94264698e-01
-1.46820903e+00 4.56993520e-01 8.72378469e-01 5.20732045e-01
1.57125688e+00 -3.02424341e-01 5.01851849e-02 -3.80925804e-01
6.23152778e-02 -1.05949450e+00 3.00196469e-01 3.42998207e-01
9.10333514e-01 -1.20638359e+00 2.48962745e-01 -2.81023771e-01
-9.21823621e-01 1.02783227e+00 8.29368174e-01 -2.64528453e-01
5.46317458e-01 8.78775358e-01 5.98889410e-01 -3.40542644e-01
-5.75110495e-01 6.06426448e-02 -1.05983391e-01 1.00044036e+00
3.17619979e-01 -1.48050785e-01 -5.33619344e-01 -5.41524738e-02
-5.03058136e-01 2.25755677e-01 9.07791495e-01 1.06116152e+00
-6.22100651e-01 -1.32882071e+00 -8.40146065e-01 7.88758099e-01
-4.32645738e-01 9.21744481e-03 -5.88341773e-01 7.72557497e-01
8.67731512e-01 7.50294566e-01 5.55072069e-01 -1.17196657e-01
3.47070754e-01 -4.91589494e-02 1.80847734e-01 -5.35852134e-01
-8.09937000e-01 -3.66305918e-01 2.77407374e-02 -2.22562626e-01
-6.39287233e-01 -5.22820175e-01 -1.15094173e+00 -1.64168090e-01
-2.29791254e-01 -1.78397149e-02 7.61320949e-01 5.78015864e-01
1.58143699e-01 3.09708297e-01 9.37734127e-01 -8.47417355e-01
-4.39001918e-02 -9.79710340e-01 -4.54214662e-01 8.74457061e-01
3.95215958e-01 -3.46857548e-01 -4.55241203e-01 5.28839588e-01] | [8.7741117477417, 4.262900352478027] |
18bfd1a9-bcff-4288-8b47-20fc5af4dbff | deepgestalt-identifying-rare-genetic | 1801.07637 | null | http://arxiv.org/abs/1801.07637v1 | http://arxiv.org/pdf/1801.07637v1.pdf | DeepGestalt - Identifying Rare Genetic Syndromes Using Deep Learning | Facial analysis technologies have recently measured up to the capabilities of
expert clinicians in syndrome identification. To date, these technologies could
only identify phenotypes of a few diseases, limiting their role in clinical
settings where hundreds of diagnoses must be considered.
We developed a facial analysis framework, DeepGestalt, using computer vision
and deep learning algorithms, that quantifies similarities to hundreds of
genetic syndromes based on unconstrained 2D images. DeepGestalt is currently
trained with over 26,000 patient cases from a rapidly growing
phenotype-genotype database, consisting of tens of thousands of validated
clinical cases, curated through a community-driven platform. DeepGestalt
currently achieves 91% top-10-accuracy in identifying over 215 different
genetic syndromes and has outperformed clinical experts in three separate
experiments.
We suggest that this form of artificial intelligence is ready to support
medical genetics in clinical and laboratory practices and will play a key role
in the future of precision medicine. | ['Lina Basel-Salmon', 'Omri Bar', 'Nicole Fleischer', 'Martin Zenker', 'Dekel Gelbman', 'Yaron Gurovich', 'Peter Krawitz', 'Susanne B Kamphausen', 'Yair Hanani', 'Lynne M. Bird', 'Karen W. Gripp'] | 2018-01-23 | null | null | null | null | ['medical-genetics'] | ['miscellaneous'] | [ 2.77736276e-01 2.63110191e-01 4.28729616e-02 -6.65074468e-01
-9.42954421e-01 -4.77565855e-01 3.28600071e-02 3.40964645e-01
-1.17830865e-01 4.68493372e-01 5.11921234e-02 -1.02111325e-01
-3.38227749e-01 -5.40086985e-01 -1.90946385e-01 -4.47820216e-01
-3.91629249e-01 1.16861296e+00 -5.06299794e-01 7.28656426e-02
-3.78934056e-01 2.90674031e-01 -1.24663627e+00 5.11215150e-01
6.60594225e-01 9.32314157e-01 -4.55175132e-01 7.76245117e-01
7.77867809e-02 1.74189538e-01 -5.53981721e-01 -7.08507895e-01
-4.57302807e-03 -3.52415025e-01 -3.56701910e-01 -2.66120940e-01
5.18335044e-01 -4.68942434e-01 7.10030049e-02 8.02900910e-01
1.11997461e+00 -7.78027117e-01 2.52780139e-01 -9.82796192e-01
-5.76250494e-01 3.50351572e-01 -7.48267114e-01 9.40755457e-02
8.31669152e-01 4.38673407e-01 1.21811783e+00 -5.49314201e-01
9.63071108e-01 1.20506966e+00 8.62409174e-01 6.37587070e-01
-1.18662274e+00 -4.83944327e-01 -3.82702380e-01 -6.98758960e-02
-1.50818205e+00 -5.86613894e-01 5.72412834e-02 -6.26788855e-01
1.11954987e+00 1.17783271e-01 8.28608751e-01 7.14962423e-01
-9.45134237e-02 4.33131397e-01 4.36297596e-01 -1.61604717e-01
-2.83255037e-02 -6.65245533e-01 -3.66432905e-01 1.27770615e+00
3.55333388e-01 -1.02023736e-01 -7.12949157e-01 -7.22231627e-01
4.29847270e-01 -3.73795629e-01 3.04653831e-02 1.56466309e-02
-1.13677895e+00 6.37975991e-01 -1.36093825e-01 1.48531720e-01
-2.65232682e-01 1.24143027e-01 3.77060711e-01 3.26864898e-01
5.04955053e-01 4.94927496e-01 -5.99090517e-01 -3.55735153e-01
-7.93358207e-01 3.63928020e-01 3.94051254e-01 5.52078426e-01
2.88866460e-01 -3.94509435e-02 1.66017618e-02 9.60378766e-01
1.29482016e-01 4.70047861e-01 4.29626882e-01 -8.20649028e-01
-2.76965573e-02 1.01752102e+00 -1.25428200e-01 -7.75381565e-01
-1.07154524e+00 -2.97824860e-01 -4.70136166e-01 2.96521395e-01
4.95578617e-01 -5.83849549e-01 -8.02905083e-01 1.63178146e+00
5.79984426e-01 1.70805991e-01 -3.44284892e-01 7.81711340e-01
9.19359684e-01 -2.15450495e-01 -3.05487812e-02 2.55406290e-01
1.42128778e+00 4.51801829e-02 -3.40542018e-01 2.00505093e-01
8.59920382e-01 -4.84072536e-01 5.84323227e-01 8.24817717e-01
-1.33772027e+00 -1.16843814e-02 -8.27234328e-01 5.67512363e-02
-4.21728075e-01 3.75204414e-01 9.60587621e-01 9.91873264e-01
-1.24259400e+00 5.65645695e-01 -7.77501285e-01 -7.58443236e-01
9.23882246e-01 6.95548415e-01 -7.56535053e-01 -1.19757988e-01
-8.84247482e-01 6.14238024e-01 1.31866291e-01 8.18999112e-02
-6.33465767e-01 -1.22201419e+00 -6.45476401e-01 -1.07787706e-01
1.56587232e-02 -8.61408055e-01 8.96452188e-01 -5.72587967e-01
-9.59190249e-01 1.46521330e+00 -7.07377344e-02 -7.60044754e-02
3.74015182e-01 -1.41144246e-01 -9.06302810e-01 4.18703675e-01
-8.91712904e-02 7.63703048e-01 3.43659967e-01 -1.47434875e-01
-6.41694784e-01 -8.07848394e-01 -4.30344671e-01 -7.63847381e-02
-4.35780510e-02 4.69285220e-01 -2.51379073e-01 -2.12944075e-01
-8.30650628e-02 -7.62533426e-01 1.06300302e-01 7.36313283e-01
-3.28908235e-01 -2.20039353e-01 6.43425584e-01 -7.13903964e-01
6.97858751e-01 -2.01779056e+00 1.21285580e-01 2.29072288e-01
6.87957585e-01 5.73534429e-01 -3.77394527e-01 2.21443087e-01
-3.92299592e-01 1.76139340e-01 1.40612543e-01 3.66323069e-02
1.73574969e-01 -3.65501076e-01 4.63679701e-01 7.14786351e-01
5.08334875e-01 1.18286586e+00 -9.52171504e-01 -3.92025299e-02
1.30561173e-01 4.44057971e-01 -6.02087021e-01 3.21793854e-02
-3.52668136e-01 1.94541767e-01 -4.08619493e-01 1.24958873e+00
7.10465491e-01 -5.50190508e-01 5.72366655e-01 1.35926828e-01
3.93910676e-01 -3.20134640e-01 -5.30361891e-01 1.93402791e+00
3.33213836e-01 4.20786828e-01 3.32747102e-01 -6.61500573e-01
9.27912176e-01 4.28478420e-01 7.55964279e-01 -1.43066943e-01
2.67449468e-01 -1.87939610e-02 4.96529579e-01 -8.99351478e-01
-5.06264031e-01 -2.27351546e-01 2.31302783e-01 4.48460221e-01
-2.51907343e-03 1.31482378e-01 2.54345834e-01 -6.71316236e-02
1.84114146e+00 6.88143016e-04 3.61779034e-01 2.53979415e-01
1.77180693e-01 4.29666974e-02 7.76878536e-01 2.01538414e-01
-5.25483608e-01 8.23484659e-01 8.97767425e-01 -8.49778473e-01
-9.35125053e-01 -1.20182943e+00 -1.18752323e-01 8.24897051e-01
-7.81637430e-01 -3.91206682e-01 -5.44220626e-01 -5.55200160e-01
6.89871132e-01 -2.96714287e-02 -7.00702071e-01 -1.56147867e-01
1.32096678e-01 -1.20917678e+00 1.30189919e+00 2.30939135e-01
-3.00123100e-03 -5.82068205e-01 -3.22231352e-01 6.17385134e-02
2.70151526e-01 -1.00755310e+00 -1.99477561e-02 -4.06957030e-01
-1.39590934e-01 -1.61428189e+00 -7.80712545e-01 -5.38541138e-01
5.04403353e-01 -6.87133133e-01 9.36318040e-01 1.88055292e-01
-1.31173849e+00 1.71296999e-01 -2.76565313e-01 -7.82865644e-01
-3.17183137e-01 -1.93467915e-01 3.49908888e-01 2.36130580e-02
9.62814331e-01 -6.10810161e-01 -6.66890979e-01 -1.53641865e-01
-4.61423874e-01 -2.82058507e-01 7.57422626e-01 6.55369759e-01
2.09464580e-01 -3.99785131e-01 8.97005796e-01 -8.24595749e-01
5.51133335e-01 -5.36758482e-01 -6.48278058e-01 3.81090492e-01
-4.03913677e-01 -3.34677607e-01 3.70717317e-01 -3.44743975e-03
-8.26392174e-01 2.80398935e-01 -4.83656853e-01 -1.72976002e-01
-5.59598923e-01 7.53221989e-01 4.72343490e-02 -2.60193825e-01
4.08125192e-01 -1.73941538e-01 4.28005576e-01 -4.40773427e-01
1.18562080e-01 7.05301821e-01 3.97666276e-01 -1.91203013e-01
2.33226255e-01 6.13375485e-01 3.87279540e-01 -8.13982844e-01
-4.44698185e-01 -2.93035030e-01 -4.62945312e-01 -8.10385123e-02
8.34196508e-01 -8.27235520e-01 -1.37298548e+00 4.97691303e-01
-6.54873848e-01 9.24343243e-03 2.76998252e-01 3.41067195e-01
-2.30143696e-01 1.81957722e-01 -5.08602500e-01 -3.46640170e-01
-4.91144180e-01 -9.81614828e-01 1.46597505e+00 -6.89369589e-02
-7.89582312e-01 -7.82325566e-01 5.35134852e-01 3.87072951e-01
1.16514213e-01 5.50656796e-01 1.32813704e+00 -6.06192350e-01
-9.24833864e-02 -4.50079590e-01 -4.59763736e-01 -1.99352533e-01
5.00358701e-01 3.81135195e-01 -1.02169871e+00 -9.35539007e-02
-1.08174622e+00 -6.99414790e-01 6.58388376e-01 3.74902099e-01
1.07538497e+00 5.17619371e-01 -4.67972577e-01 9.09495592e-01
1.14811814e+00 3.52331161e-01 4.67021346e-01 -3.96848351e-01
6.02209747e-01 6.05226576e-01 2.90638596e-01 8.95232320e-01
1.77983403e-01 5.46264291e-01 2.82472432e-01 -2.71118373e-01
-8.60366002e-02 -8.21874440e-02 -2.01983288e-01 9.61591676e-02
-1.03795528e-01 -2.30569050e-01 -1.26529622e+00 3.27268809e-01
-1.38318086e+00 -7.99752533e-01 2.26318419e-01 1.89003539e+00
5.57816148e-01 -1.96882531e-01 4.38924342e-01 -3.24029654e-01
6.99932933e-01 -3.27050030e-01 -9.78712201e-01 -4.93576556e-01
-2.94170082e-01 3.19997609e-01 -1.40431181e-01 -1.75981387e-01
-8.97610486e-01 6.59674346e-01 8.17875290e+00 3.76304656e-01
-1.13769519e+00 -3.47133577e-01 1.02138710e+00 -4.52951640e-01
-1.32513836e-01 -5.76942086e-01 -3.78674477e-01 3.28718215e-01
9.12293136e-01 1.80898327e-02 2.28443503e-01 5.03390908e-01
4.00659800e-01 5.24014123e-02 -1.27755976e+00 1.19179070e+00
-7.10017756e-02 -1.60123289e+00 -1.91417336e-01 2.06849650e-01
3.95104110e-01 2.37749338e-01 4.06111509e-01 -2.65030235e-01
3.58556002e-01 -1.50592244e+00 -1.64364576e-01 6.70609653e-01
1.56939733e+00 -7.93422282e-01 6.65314317e-01 -2.68324018e-01
-7.55033851e-01 -1.07005261e-01 -2.61660010e-01 -2.68306471e-02
1.40378535e-01 8.47684979e-01 -1.41888165e+00 2.15022773e-01
7.75311947e-01 6.73774660e-01 -2.90574402e-01 8.07013929e-01
1.20687611e-01 3.81657481e-01 -2.31458142e-01 6.73910379e-02
-1.70783862e-01 -3.47408913e-02 1.82542458e-01 9.18754220e-01
4.25757587e-01 2.55209804e-01 -2.36527115e-01 8.24786901e-01
-6.38455451e-02 7.33718276e-02 -3.18660378e-01 -7.80120850e-01
2.42905572e-01 1.45554590e+00 -5.90337694e-01 -5.67128882e-02
-6.46629035e-01 6.37030125e-01 5.55258691e-01 9.86108556e-02
-5.39182484e-01 -4.65070337e-01 1.38895655e+00 6.38819300e-03
8.82634968e-02 3.82556438e-01 2.50918977e-02 -8.51909637e-01
-3.30521405e-01 -1.31201422e+00 7.25450039e-01 -9.87360954e-01
-1.29625309e+00 2.09473610e-01 -6.39620125e-01 -8.55364799e-01
-3.22891861e-01 -1.04498088e+00 -5.73767424e-01 7.30239928e-01
-9.03201044e-01 -1.23435342e+00 -9.71634537e-02 9.02924612e-02
-3.64159077e-01 -6.78668857e-01 1.33594084e+00 2.82878369e-01
-1.01979351e+00 7.85541832e-01 2.59092208e-02 2.70720333e-01
1.05569792e+00 -1.20752776e+00 6.54933810e-01 2.82810688e-01
-2.32780337e-01 5.68356872e-01 2.98707098e-01 -7.41387427e-01
-1.47849226e+00 -9.99244571e-01 7.69289613e-01 -2.97932744e-01
6.07465744e-01 -1.40450314e-01 -4.17284906e-01 8.52895856e-01
-2.07088426e-01 7.68952519e-02 1.44122267e+00 4.25942570e-01
-4.25948441e-01 -3.16799939e-01 -1.80894291e+00 3.07842165e-01
1.07374549e+00 -5.37748098e-01 -2.22685691e-02 5.70724070e-01
1.87773570e-01 -3.81222069e-01 -1.13880563e+00 3.86061132e-01
1.05873072e+00 -9.43747878e-01 8.41795444e-01 -1.02611542e+00
3.51303339e-01 -1.61549002e-01 2.52355367e-01 -1.43549967e+00
-4.20023233e-01 -7.74091244e-01 1.62702858e-01 8.82733464e-01
6.41147017e-01 -7.85413563e-01 1.24406004e+00 7.27824926e-01
5.20198494e-02 -8.67964149e-01 -9.57439601e-01 -2.25142285e-01
7.76805868e-03 -2.33120680e-01 1.11325002e+00 8.94506156e-01
4.83715117e-01 -2.12376237e-01 1.09434754e-01 1.54207945e-01
5.20225406e-01 -1.51868135e-01 5.34495234e-01 -1.38280201e+00
-3.90474141e-01 -5.10547757e-01 -1.18988824e+00 1.33705378e-01
-5.70759065e-02 -9.19906557e-01 -5.78535140e-01 -1.12172306e+00
3.47218782e-01 -3.01153451e-01 -3.49724144e-01 7.94464171e-01
-5.92836849e-02 4.89345282e-01 -4.41760987e-01 -8.15578997e-01
-3.44219118e-01 -1.17436819e-01 8.61847997e-01 -1.81338012e-01
2.20947891e-01 -2.39708006e-01 -9.95262206e-01 8.06157351e-01
7.42011487e-01 -1.02914244e-01 -3.81420314e-01 -3.92367959e-01
5.06271720e-01 -1.24865890e-01 1.39637306e-01 -7.49424934e-01
-1.98088154e-01 -3.25302370e-02 9.69130456e-01 -4.51010168e-01
5.01276970e-01 -1.32130057e-01 3.81592572e-01 5.79769254e-01
-3.60679179e-01 2.99083423e-02 1.68377519e-01 1.41971186e-01
8.42250437e-02 3.03396881e-01 6.24488950e-01 -1.96516290e-01
-6.08369529e-01 4.48063046e-01 -3.17429662e-01 -4.92998213e-02
1.14644659e+00 3.62819619e-02 -6.21584773e-01 -4.92056489e-01
-1.08218682e+00 2.85235912e-01 3.38251442e-01 4.16636541e-02
7.96345890e-01 -1.00161409e+00 -1.18851316e+00 3.95848632e-01
4.80225027e-01 -5.22867084e-01 5.76270103e-01 8.68251979e-01
-1.06151831e+00 4.50794697e-01 -2.75362283e-01 -6.26894653e-01
-1.76462078e+00 3.36278051e-01 3.47500354e-01 3.12696457e-01
-4.59909558e-01 1.22493100e+00 -4.94555011e-03 -4.82965916e-01
6.46844581e-02 -6.67338595e-02 8.93955827e-02 -1.54279275e-02
1.13290393e+00 5.59105814e-01 2.94420868e-01 -5.30114770e-01
-6.09509706e-01 5.56449294e-01 -6.74972087e-02 8.44641402e-02
1.33608568e+00 3.21414471e-01 -1.83013469e-01 -5.01211956e-02
1.09752357e+00 -1.46073312e-01 -6.65134609e-01 3.59668702e-01
-1.05383083e-01 -2.76343316e-01 -1.91669941e-01 -1.26984251e+00
-1.22678721e+00 7.51739860e-01 7.48303294e-01 -4.35255140e-01
1.23676872e+00 1.42145425e-01 6.83726728e-01 4.22218889e-01
2.81668842e-01 -7.85257041e-01 -4.79485333e-01 -1.02859557e-01
3.98913026e-01 -1.11789155e+00 1.40499502e-01 -3.92208010e-01
-2.33500749e-01 1.04773200e+00 3.54601979e-01 2.21648589e-01
4.00649279e-01 6.81281209e-01 3.97093892e-01 -4.84710366e-01
-1.15336895e+00 -2.22166419e-01 3.14026952e-01 1.09590650e+00
8.74967337e-01 3.88306290e-01 -7.65805840e-02 6.26026869e-01
-1.16571501e-01 3.53272974e-01 9.75875407e-02 7.28533626e-01
-3.15549046e-01 -1.34291816e+00 -2.69260705e-01 1.05395472e+00
-7.38822818e-01 5.81774209e-03 -9.44720447e-01 4.99102533e-01
6.54283822e-01 9.09746170e-01 2.14595273e-01 -3.89643461e-01
-4.72619105e-03 2.65145957e-01 7.19961286e-01 -6.00607455e-01
-2.77704149e-01 3.20569873e-01 4.91111249e-01 -8.57200503e-01
-2.06069946e-02 -1.04694688e+00 -1.19813085e+00 -3.71928483e-01
2.22220927e-01 -3.29179108e-01 4.12350684e-01 8.95115376e-01
9.82475758e-01 3.18352968e-01 4.92969202e-03 -2.60280192e-01
-1.41588435e-01 -7.03802884e-01 -1.04001212e+00 1.95128158e-01
1.59685060e-01 -5.71849585e-01 3.68798584e-01 -1.57947212e-01] | [6.213565826416016, 5.7072601318359375] |
c8fc343e-f169-439b-b75b-a7da7510cf62 | partialfed-cross-domain-personalized | null | null | http://proceedings.neurips.cc/paper/2021/hash/c429429bf1f2af051f2021dc92a8ebea-Abstract.html | http://proceedings.neurips.cc/paper/2021/file/c429429bf1f2af051f2021dc92a8ebea-Paper.pdf | PartialFed: Cross-Domain Personalized Federated Learning via Partial Initialization | The burst of applications empowered by massive data have aroused unprecedented privacy concerns in AI society. Currently, data confidentiality protection has been one core issue during deep model training. Federated Learning (FL), which enables privacy-preserving training across multiple silos, gained rising popularity for its parameter-only communication. However, previous works have shown that FL revealed a significant performance drop if the data distributions are heterogeneous among different clients, especially when the clients have cross-domain characteristic, such as traffic, aerial and in-door. To address this challenging problem, we propose a novel idea, PartialFed, which loads a subset of the global model’s parameters rather than loading the entire model used in most previous works. We first validate our algorithm with manually decided loading strategies inspired by various expert priors, named PartialFed-Fix. Then we develop PartialFed-Adaptive, which automatically selects personalized loading strategy for each client. The superiority of our algorithm is proved by demonstrating the new state-of-the-art results on cross-domain federated classification and detection. In particular, solely by initializing a small fraction of layers locally, we improve the performance of FedAvg on Office-Home and UODB by 4.88% and 2.65%, respectively. Further studies show that the adaptive strategy performs significantly better on domains with large deviation, e.g. improves AP50 by 4.03% and 4.89% on aerial and medical image detection compared to FedAvg. | ['Bo Bai', 'Yi Yang', 'Hongxing Huo', 'Benyuan Sun'] | 2021-12-01 | null | https://openreview.net/forum?id=lgDP84byd5 | https://openreview.net/pdf?id=lgDP84byd5 | neurips-2021-12 | ['medical-image-detection'] | ['computer-vision'] | [-5.33009470e-02 -1.03332350e-04 -2.52303332e-01 -5.55515587e-01
-7.75685430e-01 -7.50131249e-01 4.16379064e-01 -1.91945303e-02
-4.02773261e-01 9.81733739e-01 -1.36933446e-01 -4.04437631e-01
-3.61981243e-01 -6.88148022e-01 -8.20337236e-01 -9.34432387e-01
-2.11078212e-01 2.30641395e-01 1.00713588e-01 9.92037803e-02
-2.66844779e-01 6.36869252e-01 -1.35458744e+00 4.08869624e-01
9.62722778e-01 1.40614009e+00 -3.20605725e-01 1.98244620e-02
-7.16692433e-02 5.86047947e-01 -7.22211599e-01 -8.75727355e-01
5.87324619e-01 1.49748638e-01 -6.47682905e-01 -1.42591015e-01
3.39767307e-01 -6.90780818e-01 -4.16621596e-01 9.92681026e-01
6.30298972e-01 -2.23618090e-01 2.61015266e-01 -1.58634591e+00
-4.34843659e-01 7.77321696e-01 -6.33184075e-01 -1.36062950e-01
-1.35312274e-01 1.32596463e-01 7.67251790e-01 -3.57428998e-01
5.05077362e-01 8.49826932e-01 5.55186868e-01 6.78014874e-01
-1.16385829e+00 -1.25810218e+00 3.26489836e-01 2.93391705e-01
-1.47617543e+00 -3.66601080e-01 7.00183868e-01 -1.83989599e-01
4.31048512e-01 5.14074385e-01 4.53357063e-02 1.34729636e+00
3.12494189e-02 8.51768672e-01 9.97857511e-01 6.61588907e-02
3.70676666e-01 5.26366174e-01 1.36789948e-01 4.00378406e-01
4.93179977e-01 1.08083785e-01 -4.74445760e-01 -6.63231432e-01
1.82404742e-01 1.76763803e-01 -5.20267963e-01 -5.12247145e-01
-7.26726115e-01 6.93598151e-01 1.98088497e-01 -2.49987841e-02
-3.12514275e-01 -2.60613680e-01 5.50384223e-01 2.68919706e-01
4.19836402e-01 -4.99153696e-02 -7.73713827e-01 2.46341407e-01
-9.11162913e-01 3.08989614e-01 8.98304403e-01 1.15609896e+00
5.78504026e-01 -3.35283607e-01 -2.06291154e-01 7.55917370e-01
-6.92020915e-03 3.18170369e-01 2.65163839e-01 -8.40112507e-01
6.24061763e-01 6.04449928e-01 1.37651026e-01 -9.52675223e-01
-2.06501946e-01 -7.02617228e-01 -1.15859151e+00 -1.19696736e-01
4.72407162e-01 -5.57171762e-01 -5.53229034e-01 1.92042851e+00
5.78486741e-01 1.31569192e-01 9.54553336e-02 8.50302577e-01
4.20426995e-01 3.56587648e-01 2.24378824e-01 -9.74844992e-02
1.42877221e+00 -5.98977029e-01 -6.59758627e-01 6.73760325e-02
5.77357948e-01 -3.52340698e-01 8.12716961e-01 6.98112011e-01
-6.25989676e-01 -9.46262255e-02 -9.70548093e-01 2.91909754e-01
-4.96028274e-01 8.72321874e-02 6.89706147e-01 9.93200064e-01
-8.05139005e-01 4.50056762e-01 -5.62787056e-01 -1.67656794e-01
1.16025305e+00 4.87435997e-01 -6.44519031e-01 -3.15595776e-01
-1.39076829e+00 9.03117433e-02 1.35321721e-01 -4.02245820e-01
-9.76246834e-01 -8.94949198e-01 -4.49093252e-01 2.04746008e-01
4.98161077e-01 -6.17928624e-01 1.04676890e+00 -4.85470086e-01
-1.10512555e+00 8.60120654e-01 1.14789709e-01 -8.64024580e-01
1.01518500e+00 -2.73297429e-01 -5.32997608e-01 1.44996092e-01
-2.59733796e-01 3.17758352e-01 8.24777246e-01 -1.29114497e+00
-8.45833957e-01 -6.48963571e-01 1.31013304e-01 -2.94291526e-01
-1.00434911e+00 -8.79698899e-03 -4.86623198e-01 -5.86779654e-01
-3.21784228e-01 -6.19493008e-01 -2.10360438e-01 4.53610867e-01
-5.58135629e-01 -5.77926449e-02 1.26268971e+00 -5.59954822e-01
1.30369687e+00 -2.39235830e+00 -4.89222735e-01 2.39025488e-01
4.19927090e-01 4.59349364e-01 1.46248728e-01 3.08265686e-01
6.15611561e-02 1.99879572e-01 -2.82432407e-01 -5.22975206e-01
2.85676152e-01 2.10491747e-01 -4.98972058e-01 6.94520533e-01
-1.52363583e-01 4.60778147e-01 -4.35637683e-01 -6.34867847e-01
-6.30989522e-02 5.34923553e-01 -6.69726610e-01 2.92290688e-01
-1.90059811e-01 2.98377901e-01 -5.26573360e-01 8.32731962e-01
1.39273214e+00 -4.37390000e-01 3.42546374e-01 -2.14183882e-01
6.74934983e-02 -1.17163524e-01 -1.17627418e+00 1.49850070e+00
-2.21998557e-01 1.88897967e-01 6.17026329e-01 -9.57947671e-01
9.89424348e-01 4.94117588e-01 6.42893493e-01 -6.05159760e-01
3.45416993e-01 5.79092503e-02 -4.69121277e-01 -4.20856237e-01
1.77150831e-01 1.74725339e-01 -7.52542615e-02 4.27481860e-01
-1.50942713e-01 7.04926729e-01 -3.75493169e-01 2.20153451e-01
1.34812939e+00 -3.18413198e-01 6.06780574e-02 -1.67795151e-01
6.25033557e-01 -3.14288616e-01 6.82207108e-01 7.59487867e-01
-5.69803476e-01 4.78920221e-01 6.03737772e-01 -3.62004459e-01
-6.71989083e-01 -9.05114710e-01 -3.86823237e-01 9.78167474e-01
2.39990130e-01 -3.93329233e-01 -8.95883143e-01 -1.16439795e+00
4.41520840e-01 6.78245366e-01 -5.42710960e-01 -1.94514602e-01
-3.28986824e-01 -8.89984429e-01 8.07398140e-01 1.84002757e-01
9.69707370e-01 -5.89983225e-01 -4.79525238e-01 -8.69801547e-03
-9.98941958e-02 -1.31302369e+00 -3.26876998e-01 2.11068958e-01
-5.86775541e-01 -1.02822030e+00 -7.69684732e-01 -3.75673294e-01
5.90588927e-01 1.32389367e-01 6.61318600e-01 -2.39027008e-01
-2.58218288e-01 1.38474569e-01 -2.49072447e-01 -3.05393428e-01
-5.68039417e-02 3.27161193e-01 -5.84456325e-02 6.05735183e-01
3.52837861e-01 -7.36366749e-01 -7.03451753e-01 6.70359433e-01
-1.00669444e+00 -4.52464730e-01 5.39329529e-01 7.25844681e-01
3.58732909e-01 9.75695252e-02 5.21285415e-01 -1.30809212e+00
5.93220949e-01 -8.83109272e-01 -6.43658280e-01 3.75876814e-01
-7.13321686e-01 -1.26478001e-01 9.33864772e-01 -4.67865407e-01
-1.06968415e+00 1.76909700e-01 3.75906453e-02 -8.39936852e-01
-5.06723642e-01 -2.15435494e-02 -6.97643161e-01 -1.81757838e-01
5.74202776e-01 1.57339960e-01 -7.16873705e-02 -7.37415254e-01
1.38829023e-01 1.15804541e+00 5.22934437e-01 -7.09971666e-01
7.32145369e-01 7.02804148e-01 -2.45131508e-01 -5.47980011e-01
-4.02828962e-01 -1.93092316e-01 -6.50539994e-02 1.17543012e-01
4.93395984e-01 -9.49876189e-01 -1.14029217e+00 7.38433421e-01
-1.11352098e+00 -2.64699645e-02 -1.16940580e-01 2.34861568e-01
-1.39344320e-01 4.37273353e-01 -3.76663506e-01 -9.05154407e-01
-6.18947327e-01 -9.33205903e-01 7.98429906e-01 1.84647292e-02
7.98657611e-02 -6.91544473e-01 -3.40128690e-01 5.03917456e-01
5.15983701e-01 5.64014256e-01 7.33022213e-01 -1.05211008e+00
-6.51001096e-01 -3.52196306e-01 -3.25590312e-01 4.61336255e-01
8.26287419e-02 -3.99339229e-01 -1.38097537e+00 -4.89663333e-01
9.12758037e-02 -1.80940241e-01 4.52299923e-01 1.35440668e-02
1.79596150e+00 -6.83292747e-01 -5.96378446e-01 1.20548451e+00
1.33740008e+00 1.69089273e-01 4.69762027e-01 3.52621317e-01
3.64769012e-01 6.66763425e-01 5.94812453e-01 1.01844037e+00
1.71150506e-01 4.53889102e-01 7.54887819e-01 6.88385032e-03
3.82249087e-01 -1.61088496e-01 1.20217897e-01 -1.52120709e-01
5.42061031e-01 -5.06079495e-01 -7.08960652e-01 5.44141531e-01
-1.80905330e+00 -6.59259737e-01 2.52887845e-01 2.19200063e+00
7.86000252e-01 8.69068131e-02 1.08236112e-01 1.19964652e-01
7.74670005e-01 3.40571880e-01 -9.38020229e-01 -8.66389126e-02
-2.99563080e-01 1.88125119e-01 1.03145730e+00 -2.48071440e-02
-1.24637055e+00 5.96063077e-01 5.24355555e+00 1.03608608e+00
-1.16936672e+00 2.14187905e-01 8.86609197e-01 -3.41722935e-01
-1.56417355e-01 -3.33448499e-01 -9.48940873e-01 7.67502308e-01
8.28344762e-01 -4.33507651e-01 3.00123721e-01 1.20061290e+00
-1.29089713e-01 4.03851628e-01 -1.01191032e+00 1.06803322e+00
-2.82040715e-01 -1.32051194e+00 -5.44546871e-04 3.98895890e-01
4.29317862e-01 -5.51484227e-02 3.10328096e-01 1.10575467e-01
3.77552271e-01 -7.96676934e-01 4.63487178e-01 9.29063484e-02
9.83005881e-01 -9.42917049e-01 7.11741686e-01 6.21306896e-01
-8.28818738e-01 -3.72818738e-01 -3.31344187e-01 4.45857763e-01
-1.08696297e-01 6.88007295e-01 -6.09770179e-01 7.82077074e-01
1.06135464e+00 3.97546470e-01 -3.00640464e-01 9.45078790e-01
1.90546274e-01 7.44242907e-01 -5.22919178e-01 1.98545918e-01
4.52547371e-02 2.02090099e-01 4.14825141e-01 9.57330346e-01
1.72137529e-01 7.52331018e-02 -2.10401192e-01 7.18268752e-01
-3.91327947e-01 1.09818958e-01 -5.60346484e-01 1.76129952e-01
8.27745080e-01 1.21750736e+00 5.71107306e-02 -2.03221645e-02
-3.44722360e-01 8.43842566e-01 2.74116009e-01 3.73734951e-01
-1.05632842e+00 -5.22944331e-01 1.15196657e+00 2.90990859e-01
5.74962080e-01 1.81947276e-01 -2.12269112e-01 -1.01600432e+00
2.82378316e-01 -1.10127032e+00 8.08778167e-01 -1.58827648e-01
-1.52639282e+00 8.75744760e-01 2.38846000e-02 -1.25501060e+00
1.30094245e-01 -3.91949892e-01 -3.75133902e-01 7.82957554e-01
-1.52564836e+00 -1.12103283e+00 -3.91488373e-01 1.15095198e+00
-2.38521978e-01 -3.42986822e-01 8.82096469e-01 7.95285285e-01
-8.53630543e-01 1.56707418e+00 4.29166883e-01 2.18076289e-01
8.57529581e-01 -7.13874042e-01 2.37785235e-01 7.93703616e-01
-2.40507871e-01 6.32063448e-01 4.13161933e-01 -3.50341886e-01
-1.31253886e+00 -1.26078928e+00 6.98864460e-01 -1.03776082e-01
3.53581399e-01 -6.91429615e-01 -9.69869375e-01 6.17143035e-01
-3.20271105e-02 4.38736469e-01 8.18707705e-01 -2.01906964e-01
-6.02780402e-01 -7.40988910e-01 -1.99103105e+00 3.28408211e-01
1.05811095e+00 -2.22984776e-01 -1.70312747e-02 2.85560757e-01
9.57373619e-01 -3.93558949e-01 -8.59421492e-01 3.39589685e-01
4.22994345e-01 -1.28439629e+00 8.26419473e-01 -6.61703467e-01
-1.14984699e-01 -4.63572629e-02 -3.84948820e-01 -7.44291723e-01
-1.36993706e-01 -1.04414940e+00 -3.84463847e-01 1.58574915e+00
1.36482239e-01 -1.19772339e+00 1.34265733e+00 9.62754130e-01
2.32589871e-01 -8.67175162e-01 -1.11083639e+00 -8.00296307e-01
-9.87227634e-02 -3.70801270e-01 1.30832148e+00 1.08897936e+00
-3.77200395e-01 -4.35614288e-01 -5.39043725e-01 5.34559250e-01
9.90153551e-01 3.25334072e-02 9.28827941e-01 -1.14996040e+00
-2.85105169e-01 1.61885582e-02 -2.01993048e-01 -9.82541025e-01
7.33744353e-02 -6.21239066e-01 -5.33546805e-01 -8.88409972e-01
-2.05509543e-01 -6.22718990e-01 -5.64306974e-01 8.89996350e-01
2.11373255e-01 -1.36776850e-01 2.31878117e-01 1.23290248e-01
-4.01847482e-01 4.42901254e-01 9.50858831e-01 -1.12473734e-01
7.15435445e-02 3.35002005e-01 -1.01495314e+00 3.00774276e-01
8.66572917e-01 -6.46558285e-01 -4.60185260e-01 -4.13017184e-01
-3.79988223e-01 -4.66396250e-02 4.42779183e-01 -1.06540728e+00
3.00011754e-01 -1.11411490e-01 2.71633923e-01 -4.39970523e-01
2.09498554e-01 -1.48099720e+00 4.35913086e-01 3.96400392e-01
-1.88158557e-01 -3.50276053e-01 3.06098849e-01 6.94832385e-01
-1.47359684e-01 2.89326549e-01 7.76632845e-01 1.46213830e-01
-4.74814266e-01 7.07970023e-01 2.23393828e-01 1.82920620e-02
1.40408862e+00 -1.06749505e-01 -6.51810646e-01 -3.07718337e-01
-4.66084003e-01 5.29134870e-01 3.68240058e-01 2.15518013e-01
4.05185848e-01 -1.09748125e+00 -5.26713490e-01 5.43653727e-01
5.08504696e-02 1.51902229e-01 5.06434381e-01 6.26200020e-01
-1.68361574e-01 4.07831460e-01 -1.40631095e-01 -4.59745705e-01
-1.45838618e+00 6.81229472e-01 3.45702469e-01 -2.47291610e-01
-6.05912745e-01 9.72733021e-01 2.41749212e-01 -3.53625566e-01
8.63530457e-01 -4.07387093e-02 2.04048991e-01 2.09305733e-02
5.87571979e-01 4.27210838e-01 1.97436854e-01 -2.99426258e-01
-6.68859124e-01 1.71742246e-01 -3.85399103e-01 3.13918024e-01
1.43533707e+00 1.75830480e-02 1.42263755e-01 -3.00621957e-01
1.52368450e+00 1.95793584e-02 -1.54190326e+00 -4.00770754e-01
-3.23816746e-01 -7.35747337e-01 -6.02975488e-02 -9.07719314e-01
-1.48102450e+00 7.41309166e-01 7.64341295e-01 1.76200941e-01
1.45352376e+00 -1.47991464e-01 1.25078130e+00 1.11068167e-01
7.69646466e-01 -5.91200292e-01 -4.44324166e-01 3.64137162e-03
5.63939154e-01 -1.00758111e+00 -4.23988625e-02 -3.38452160e-01
-5.95839620e-01 6.62755549e-01 7.59972870e-01 2.87559152e-01
8.73235226e-01 3.45929742e-01 6.16814122e-02 6.58427551e-02
-8.36576641e-01 4.72722352e-01 -1.33036807e-01 5.99454641e-01
-2.35079840e-01 -6.96795732e-02 -1.06319353e-01 1.24084127e+00
-5.89431599e-02 1.96175203e-01 -2.80316398e-02 9.91000354e-01
3.92700806e-02 -1.28642058e+00 -4.05490279e-01 3.79800022e-01
-7.65965462e-01 2.55864739e-01 -3.38850588e-01 6.77028894e-01
3.75156939e-01 6.92773759e-01 -8.42362270e-02 -7.25583494e-01
3.64671528e-01 4.51547280e-02 -1.25775933e-01 -8.47838372e-02
-8.37342978e-01 -2.88328081e-01 -9.16177705e-02 -8.28163743e-01
9.29480046e-02 -5.95205605e-01 -1.06422555e+00 -6.64930046e-01
-1.43397167e-01 3.32063675e-01 6.15702868e-01 4.69008952e-01
9.07932520e-01 1.50364712e-01 8.82963896e-01 -2.10474759e-01
-1.06860578e+00 -3.83694381e-01 -8.37230861e-01 3.34312737e-01
5.41267991e-01 -4.77322102e-01 -6.65158749e-01 -2.58659631e-01] | [5.905886650085449, 6.408730983734131] |
68178e0b-da72-474c-b8bd-cfc011f866f5 | chatgpt-or-academic-scientist-distinguishing | 2303.16352 | null | https://arxiv.org/abs/2303.16352v1 | https://arxiv.org/pdf/2303.16352v1.pdf | ChatGPT or academic scientist? Distinguishing authorship with over 99% accuracy using off-the-shelf machine learning tools | ChatGPT has enabled access to AI-generated writing for the masses, and within just a few months, this product has disrupted the knowledge economy, initiating a culture shift in the way people work, learn, and write. The need to discriminate human writing from AI is now both critical and urgent, particularly in domains like higher education and academic writing, where AI had not been a significant threat or contributor to authorship. Addressing this need, we developed a method for discriminating text generated by ChatGPT from (human) academic scientists, relying on prevalent and accessible supervised classification methods. We focused on how a particular group of humans, academic scientists, write differently than ChatGPT, and this targeted approach led to the discovery of new features for discriminating (these) humans from AI; as examples, scientists write long paragraphs and have a penchant for equivocal language, frequently using words like but, however, and although. With a set of 20 features, including the aforementioned ones and others, we built a model that assigned the author, as human or AI, at well over 99% accuracy, resulting in 20 times fewer misclassified documents compared to the field-leading approach. This strategy for discriminating a particular set of humans writing from AI could be further adapted and developed by others with basic skills in supervised classification, enabling access to many highly accurate and targeted models for detecting AI usage in academic writing and beyond. | ['David Hua', 'Romana Jarosova', 'Madeline Isom', 'Aleesa E. Chua', 'Heather Desaire'] | 2023-03-28 | null | null | null | null | ['culture'] | ['speech'] | [ 8.31430256e-02 3.55154902e-01 -2.53946751e-01 2.91102022e-01
-3.00599009e-01 -8.70231390e-01 1.17299676e+00 4.35489267e-01
-3.91117096e-01 8.32350194e-01 2.77194142e-01 -4.77272481e-01
-3.90291452e-01 -8.42014492e-01 -1.33788854e-01 -4.17290539e-01
7.60017633e-01 8.79455864e-01 -1.38500422e-01 -2.00928852e-01
9.58040178e-01 5.71730912e-01 -1.30999351e+00 -6.87753558e-02
1.32347524e+00 3.84566844e-01 1.39262497e-01 7.83508599e-01
-6.65612757e-01 1.17247760e+00 -1.17089272e+00 -8.24890196e-01
-1.90159947e-01 -6.23306870e-01 -9.53523755e-01 -1.77921504e-01
7.95102894e-01 -4.24522832e-02 -2.25343928e-01 6.61313951e-01
3.22109163e-01 -7.03215003e-02 9.98917103e-01 -8.99583817e-01
-1.02324271e+00 9.29495573e-01 -2.04393223e-01 2.78924853e-01
6.44382775e-01 2.85853118e-01 8.09408903e-01 -5.58520973e-01
7.86185980e-01 1.20517755e+00 5.30645549e-01 3.73676211e-01
-1.02006280e+00 -7.06834972e-01 -1.74537584e-01 1.37251601e-01
-1.10468686e+00 -3.73574644e-01 6.01471603e-01 -1.02800596e+00
6.66739523e-01 3.63796622e-01 9.81179595e-01 1.52322459e+00
1.12980083e-01 5.66633105e-01 1.05681026e+00 -7.51133084e-01
-6.86788931e-02 7.22939670e-01 2.99264014e-01 5.97497702e-01
3.58803719e-01 -4.28355426e-01 -8.10597360e-01 -2.11519837e-01
1.85494870e-01 -1.27386332e-01 -2.10606009e-01 5.18775165e-01
-1.72003818e+00 8.04996729e-01 -1.92400098e-01 9.33348536e-01
-1.31189913e-01 -1.58886105e-01 2.27691144e-01 4.32143033e-01
4.52505708e-01 1.22135103e+00 1.16398279e-02 -7.81790793e-01
-1.28218758e+00 2.43623286e-01 1.21927619e+00 8.38284194e-01
3.60896826e-01 -1.18009105e-01 -4.05681014e-01 8.64696383e-01
-1.52376369e-01 5.96111655e-01 6.75259233e-01 -7.53181994e-01
3.92958313e-01 1.02005708e+00 -2.80751139e-02 -1.37458026e+00
-1.34056926e-01 -6.32486701e-01 -7.04348266e-01 -1.96298286e-01
8.32668364e-01 -5.89364134e-02 -5.32749355e-01 1.21192098e+00
-1.90391615e-01 -5.49463034e-01 -1.99907497e-01 7.33168542e-01
5.88593364e-01 6.18885100e-01 -2.15644445e-02 -6.27897978e-02
1.30706930e+00 -7.25013494e-01 -6.79611504e-01 -2.58828163e-01
7.69867063e-01 -9.30686295e-01 1.10817254e+00 7.27912843e-01
-1.08276212e+00 -3.69870931e-01 -8.74871910e-01 -1.53811321e-01
-7.45857596e-01 8.74334946e-02 4.84900922e-01 9.44251835e-01
-8.43923151e-01 7.58070350e-01 -1.17097460e-01 -8.04597795e-01
6.17435932e-01 -1.14920050e-01 -2.17947364e-01 1.08640440e-01
-1.06313097e+00 1.26112056e+00 -1.87874178e-03 -2.38197088e-01
-2.74994373e-01 -8.89699042e-01 -1.58961773e-01 -4.89160568e-02
3.05783272e-01 -5.67292869e-01 8.40520561e-01 -9.63595450e-01
-1.53684103e+00 1.24887824e+00 -8.71669129e-02 -2.04845443e-01
1.09471989e+00 -1.94322318e-01 -4.35871512e-01 7.41235446e-03
4.20288920e-01 7.27111325e-02 9.99641359e-01 -8.37962806e-01
-3.47315043e-01 -3.45774263e-01 -2.40280151e-01 -2.59075850e-01
-9.25360262e-01 3.24705392e-01 1.43471375e-01 -7.67030954e-01
-3.47864687e-01 -6.30706429e-01 3.64358783e-01 -2.73923296e-02
-3.46229166e-01 -6.47530496e-01 6.30904019e-01 -8.08522284e-01
1.39388275e+00 -1.95086408e+00 2.54742742e-01 4.22758460e-01
8.48833501e-01 2.70821780e-01 2.53002971e-01 7.34886587e-01
6.33318007e-01 5.85985839e-01 -1.15306284e-02 1.67619381e-02
2.06069171e-01 -8.92084762e-02 -3.86574298e-01 1.84788302e-01
4.90303561e-02 7.51827121e-01 -1.22590744e+00 -5.60304344e-01
-1.22648567e-01 2.10416541e-01 -2.46478263e-02 8.81390423e-02
-1.01274513e-02 4.05021816e-01 -5.08739650e-01 7.18152583e-01
1.57834776e-02 -3.90968561e-01 1.14907391e-01 6.39864683e-01
-5.85346103e-01 1.67009309e-01 -4.97095168e-01 1.01783848e+00
-5.27790427e-01 1.30069602e+00 -2.91944176e-01 -7.58238375e-01
1.27086997e+00 1.96850866e-01 -4.18656021e-02 -7.02808201e-01
1.86354622e-01 6.32464409e-01 3.41674119e-01 -4.22649175e-01
4.64436352e-01 1.88802734e-01 -3.33015062e-02 7.60443509e-01
-1.17661536e-01 -3.49156618e-01 3.82428348e-01 4.20259804e-01
1.22145057e+00 -2.08179966e-01 1.48634925e-01 -4.28379387e-01
6.60975397e-01 2.63587445e-01 -7.51104355e-02 1.22084832e+00
-1.48514017e-01 3.00092965e-01 7.35125482e-01 -3.02741289e-01
-1.01521146e+00 -7.90012777e-01 -1.47271052e-01 1.11990190e+00
-4.05744940e-01 -3.97088259e-01 -6.06433272e-01 -3.90311748e-01
1.52421907e-01 9.49394524e-01 -2.74649739e-01 -1.87867224e-01
-5.17994046e-01 -2.23968863e-01 8.92806172e-01 -6.45907084e-03
4.06987935e-01 -1.14482653e+00 -6.76231742e-01 2.01042473e-01
9.29134637e-02 -8.04243624e-01 -2.67466724e-01 1.00153089e-01
-4.03823316e-01 -1.04624355e+00 -1.28923059e+00 -5.85130334e-01
5.08706033e-01 -2.37238422e-01 1.13437200e+00 3.69034201e-01
-3.95368099e-01 3.57240885e-01 -3.65252286e-01 -8.37143958e-01
-9.70550418e-01 5.94396889e-01 9.23798420e-03 -2.86952913e-01
3.11375141e-01 -2.99720675e-01 -1.03442945e-01 -1.05392925e-01
-4.22094077e-01 6.98870942e-02 6.37087524e-01 5.72825193e-01
-6.20106399e-01 -4.16263431e-01 8.06015670e-01 -1.02535021e+00
9.03071105e-01 -4.39381987e-01 -1.91365629e-01 1.89204320e-01
-8.13568950e-01 -1.16205655e-01 9.53146994e-01 -7.55300224e-01
-7.78868556e-01 -7.11375773e-01 2.41782144e-01 -1.12626426e-01
-9.36824307e-02 4.83194530e-01 2.73140579e-01 -1.31224394e-01
6.88319147e-01 2.90543586e-01 1.40069693e-01 -3.92471582e-01
4.08475995e-02 1.14533961e+00 4.23977911e-01 -5.32067835e-01
9.15329278e-01 -8.16733539e-02 -9.51108336e-02 -1.23454440e+00
-7.75316894e-01 -2.56071448e-01 -5.47146738e-01 -4.97433603e-01
6.68511391e-01 -4.14631009e-01 -9.61446822e-01 6.19341850e-01
-1.21038973e+00 -3.13127548e-01 -1.46840364e-01 4.02085334e-01
-1.33100390e-01 3.61652583e-01 -5.21866024e-01 -8.94814074e-01
-3.07974488e-01 -7.08648205e-01 5.33087850e-01 3.82142395e-01
-1.03725278e+00 -1.15739977e+00 -2.52588652e-02 8.41381788e-01
6.66386664e-01 2.93701626e-02 1.30522549e+00 -1.11248422e+00
-2.53462315e-01 -2.80009091e-01 -2.18227476e-01 2.18747839e-01
-1.79395676e-02 2.51572251e-01 -7.16747403e-01 1.95516631e-01
-3.16138625e-01 -2.74259448e-01 5.53069532e-01 -4.37445760e-01
1.00607407e+00 -6.89501584e-01 -3.63314062e-01 3.11646033e-02
7.70364821e-01 2.17667893e-01 3.50900501e-01 5.93053043e-01
8.40071797e-01 5.88821054e-01 8.53930116e-02 3.71478975e-01
3.17334160e-02 5.29226005e-01 -3.31956655e-01 3.77099782e-01
2.44711600e-02 -2.23233059e-01 2.34455824e-01 9.61609304e-01
-4.40154016e-01 -4.13555056e-01 -1.37130570e+00 5.75082958e-01
-1.60540485e+00 -1.19804919e+00 -4.29557651e-01 1.97259927e+00
9.96235907e-01 2.54581422e-01 1.64973453e-01 1.97562680e-01
3.38396251e-01 -2.77414490e-02 -2.92781919e-01 -8.10097516e-01
-2.37028688e-01 4.04647827e-01 4.14184719e-01 2.01231211e-01
-5.00276685e-01 8.51713777e-01 6.61245966e+00 7.12141216e-01
-1.19917226e+00 -2.04717264e-01 5.03793657e-01 9.97920334e-02
-2.11140603e-01 -1.68935552e-01 -8.01061928e-01 9.46079850e-01
1.03182924e+00 -6.92584276e-01 5.12119710e-01 5.84203601e-01
1.88311890e-01 -1.05908521e-01 -1.02985537e+00 8.23460221e-01
5.28775930e-01 -1.33765161e+00 2.52901018e-01 3.00796896e-01
7.04221785e-01 -5.17575502e-01 -1.24376655e-01 1.90786839e-01
1.76989719e-01 -1.23877835e+00 9.10127997e-01 9.55887973e-01
4.29913729e-01 -5.10626435e-01 7.00324655e-01 6.86020911e-01
-1.56409144e-01 -1.56556875e-01 -8.25702026e-02 -5.04199564e-01
-5.05843759e-01 8.77454102e-01 -8.72850537e-01 2.17270702e-01
4.03681129e-01 7.34239340e-01 -8.73667657e-01 5.72770178e-01
-8.79304036e-02 7.79132009e-01 -7.01464666e-03 -6.31413877e-01
1.93706200e-01 -3.31824392e-01 7.51855195e-01 1.27894008e+00
4.82582003e-01 -2.85046130e-01 -2.27290124e-01 1.14662278e+00
-2.25077525e-01 1.60161614e-01 -6.94613636e-01 -9.11232471e-01
6.58350587e-01 1.27881110e+00 -6.64205968e-01 -4.08775091e-01
-2.23684683e-01 1.14053273e+00 3.67882758e-01 1.02320604e-01
-4.55422014e-01 -6.39063776e-01 2.94092894e-01 4.20034617e-01
-2.28100106e-01 -1.92054048e-01 -5.18338144e-01 -8.09166670e-01
-9.11182463e-02 -1.06865895e+00 -2.37362795e-02 -4.39001143e-01
-1.56760991e+00 2.29641035e-01 -3.55616659e-01 -7.03971863e-01
-1.95464656e-01 -7.28031218e-01 -7.08445489e-01 1.04630852e+00
-8.56428683e-01 -1.00006592e+00 -5.28524220e-01 3.30294855e-02
3.34687322e-01 -7.17300713e-01 6.54538333e-01 9.56694409e-02
-4.65372682e-01 6.49794400e-01 1.85666978e-01 2.56597012e-01
9.89928901e-01 -1.25812972e+00 -2.14338806e-02 2.85080343e-01
1.28115743e-01 8.68737757e-01 5.99384308e-01 -5.91850460e-01
-1.38846827e+00 -7.07737267e-01 1.54875553e+00 -9.99610841e-01
1.07168162e+00 -4.53068286e-01 -7.33481586e-01 3.72043103e-01
3.13192993e-01 -9.93000805e-01 6.82227612e-01 4.63091247e-02
-1.65694594e-01 3.55877757e-01 -7.79912472e-01 7.28994310e-01
1.07489431e+00 -6.05399907e-01 -8.48345101e-01 6.44018948e-01
2.90239751e-01 -5.70984185e-03 -9.39209402e-01 -3.87705743e-01
6.81033075e-01 -7.68508196e-01 6.36667192e-01 -5.36185563e-01
9.78052378e-01 1.61711395e-01 5.93104243e-01 -1.23748648e+00
-5.12243271e-01 -7.71137595e-01 -1.75534695e-01 1.54350495e+00
3.32995921e-01 -8.80613983e-01 5.90256512e-01 5.55369675e-01
-1.86151668e-01 -4.78679746e-01 -6.96564555e-01 -8.19114447e-01
4.86123174e-01 1.47101998e-01 1.88354030e-01 1.41589749e+00
2.30555236e-01 4.40915078e-01 -1.66932315e-01 -6.85496211e-01
1.78472757e-01 -3.49767357e-02 8.71548474e-01 -1.84713125e+00
8.02234411e-02 -1.15803874e+00 -5.41352630e-01 -4.11365569e-01
3.87949497e-01 -1.13762653e+00 -3.17542464e-01 -1.61826396e+00
2.90797949e-01 -3.07372749e-01 3.53747100e-01 4.52353001e-01
-1.07735477e-01 3.13835174e-01 2.42610514e-01 5.49828947e-01
-6.17059097e-02 -1.12565104e-02 1.30495024e+00 -3.50595325e-01
-1.64977565e-01 -2.25530416e-01 -1.02222705e+00 5.84851384e-01
6.31965697e-01 -2.76990503e-01 7.87289534e-03 -2.59525087e-02
7.23267138e-01 -3.96105230e-01 5.08193314e-01 -1.11396766e+00
2.92696446e-01 -3.45048964e-01 6.14010572e-01 -8.96233395e-02
-9.27267596e-02 -5.99566042e-01 2.40430851e-02 4.41993564e-01
-4.02537733e-01 -2.39255041e-01 -6.81767315e-02 7.65147358e-02
2.90534012e-02 -5.26117027e-01 6.18068576e-01 -3.13832462e-01
-2.30876833e-01 -3.23118716e-01 -8.48583519e-01 2.76874334e-01
8.88452232e-01 -4.63046402e-01 -7.16715336e-01 -3.60996574e-01
-4.01378304e-01 9.58126858e-02 5.61640561e-01 4.90150422e-01
1.19279616e-01 -7.22852051e-01 -8.28155816e-01 5.54253533e-02
1.34678483e-01 -3.28561157e-01 -2.81919658e-01 9.41443324e-01
-6.52150750e-01 6.79854929e-01 -2.56859988e-01 -1.85303465e-01
-1.04369283e+00 1.67356357e-01 -1.33514609e-02 -1.90502211e-01
-6.70436382e-01 5.36032021e-01 -3.31148297e-01 -2.47158363e-01
1.20524473e-01 1.10260695e-01 -3.11378121e-01 5.17311215e-01
4.87391859e-01 8.30055177e-01 -8.45051929e-02 -5.13257623e-01
-3.09071869e-01 2.96977937e-01 -2.16074899e-01 3.01486719e-02
1.04443979e+00 2.45298862e-01 -2.92575568e-01 1.01230836e+00
7.83925891e-01 5.78133106e-01 -1.99957937e-01 2.05814660e-01
2.38708422e-01 -6.39868498e-01 -1.52053863e-01 -1.08102274e+00
-3.42049688e-01 7.58129001e-01 7.29842484e-02 5.55800259e-01
3.49278599e-01 1.07845590e-01 6.00337565e-01 4.94118184e-01
2.00604737e-01 -1.31300402e+00 2.44100884e-01 7.34432280e-01
8.65054488e-01 -1.04008090e+00 6.00827374e-02 -1.35737449e-01
-4.41424757e-01 1.35125470e+00 4.85082299e-01 3.00120175e-01
-1.02224043e-02 -1.48443803e-01 5.36166094e-02 -2.75872737e-01
-4.96386468e-01 1.34002179e-01 3.52532864e-01 4.35851902e-01
8.37807894e-01 4.74541150e-02 -6.83243811e-01 4.41837549e-01
-6.93409324e-01 1.32287264e-01 8.02575767e-01 6.47004008e-01
-3.71189207e-01 -9.85586226e-01 -5.11530876e-01 9.63494956e-01
-4.27799374e-01 -2.66712811e-02 -1.25897956e+00 8.06382418e-01
4.81841229e-02 8.64073873e-01 3.25851828e-01 -1.42853722e-01
1.05014451e-01 5.68922162e-01 3.37928861e-01 -5.32945037e-01
-1.05963981e+00 -6.28710032e-01 3.47070694e-01 1.71056330e-01
-1.57553524e-01 -4.84621406e-01 -8.11384499e-01 -8.50470841e-01
-1.45389706e-01 2.86832929e-01 3.97442400e-01 1.16362441e+00
1.92452118e-01 3.10703933e-01 4.12350029e-01 -5.32381415e-01
-3.22537273e-01 -1.20095885e+00 -4.77436304e-01 2.48816341e-01
3.40230018e-02 -4.12337065e-01 -6.77883744e-01 -1.20068565e-01] | [9.605244636535645, 10.538883209228516] |
b9512126-047e-4313-a255-a6ff73b78da2 | an-embarrassingly-simple-model-for-dialogue | 2012.13873 | null | https://arxiv.org/abs/2012.13873v2 | https://arxiv.org/pdf/2012.13873v2.pdf | An Embarrassingly Simple Model for Dialogue Relation Extraction | Dialogue relation extraction (RE) is to predict the relation type of two entities mentioned in a dialogue. In this paper, we propose a simple yet effective model named SimpleRE for the RE task. SimpleRE captures the interrelations among multiple relations in a dialogue through a novel input format named BERT Relation Token Sequence (BRS). In BRS, multiple [CLS] tokens are used to capture possible relations between different pairs of entities mentioned in the dialogue. A Relation Refinement Gate (RRG) is then designed to extract relation-specific semantic representation in an adaptive manner. Experiments on the DialogRE dataset show that SimpleRE achieves the best performance, with much shorter training time. Further, SimpleRE outperforms all direct baselines on sentence-level RE without using external resources. | ['Eng Siong Chng', 'Jinjie Ni', 'Hao Zhang', 'Aixin Sun', 'Fuzhao Xue'] | 2020-12-27 | null | null | null | null | ['dialog-relation-extraction'] | ['natural-language-processing'] | [ 7.79437125e-02 9.69424665e-01 -2.70172179e-01 -5.61112761e-01
-6.13669336e-01 -6.08552039e-01 9.03434277e-01 4.04192865e-01
-4.81598854e-01 9.68043923e-01 7.35127866e-01 -4.33842391e-01
1.99946642e-01 -7.63827622e-01 -1.34052172e-01 2.09427282e-01
1.39852121e-01 8.37834954e-01 5.31656861e-01 -1.00353324e+00
-1.00689325e-02 1.46473810e-01 -5.17595291e-01 7.50928998e-01
8.33834469e-01 8.32426965e-01 4.03789319e-02 6.87410712e-01
-4.05698448e-01 1.22663832e+00 -8.96931648e-01 -7.80590177e-01
-1.63751751e-01 -5.92158377e-01 -1.75076735e+00 5.39136538e-03
-3.81013364e-01 -3.09533745e-01 -5.07605493e-01 5.47541916e-01
1.76268622e-01 4.65187162e-01 7.28565931e-01 -8.37965310e-01
-2.01638475e-01 1.24754572e+00 -1.62458166e-01 3.29866648e-01
1.01671624e+00 -2.50848442e-01 1.51176369e+00 -9.80109096e-01
8.34076524e-01 1.67888117e+00 4.58127916e-01 6.66588902e-01
-1.19538617e+00 -3.01878929e-01 9.95698273e-02 2.67247409e-01
-1.06048286e+00 -5.87724984e-01 4.96537149e-01 4.51677799e-04
1.79188001e+00 4.12022918e-01 3.55567157e-01 9.01901722e-01
1.04915679e-01 7.53553987e-01 8.79091740e-01 -4.75022256e-01
-2.08441451e-01 2.26975039e-01 6.67302787e-01 5.39052308e-01
-2.58869708e-01 -7.32134819e-01 -6.89110518e-01 -2.55824357e-01
5.61081707e-01 -6.22686088e-01 -3.68684828e-01 3.34586143e-01
-8.17960978e-01 9.39623654e-01 3.46840769e-01 3.46286237e-01
-8.59201699e-02 -5.61059892e-01 6.03697658e-01 5.09995103e-01
4.75735664e-01 9.61826384e-01 -8.56989264e-01 -2.64919281e-01
1.61825772e-02 4.01468754e-01 1.54442883e+00 1.26485717e+00
4.17356372e-01 -7.22352564e-01 -6.11520469e-01 1.19373071e+00
1.45371258e-01 -1.19798131e-01 3.22841734e-01 -8.90982091e-01
1.03939307e+00 1.03116930e+00 1.46014109e-01 -9.01772261e-01
-5.84072173e-01 1.04546897e-01 -6.89422548e-01 -6.67027712e-01
3.66627574e-01 -4.71064776e-01 -2.15401441e-01 1.34868002e+00
4.92047787e-01 -2.21209884e-01 6.58041835e-01 4.93868440e-01
1.63108325e+00 7.21247673e-01 2.02758849e-01 -2.61355102e-01
1.75446308e+00 -1.09502792e+00 -1.14807570e+00 -4.21651781e-01
8.88039351e-01 -5.18519342e-01 7.77810395e-01 -9.94795635e-02
-8.70465338e-01 -7.38771856e-02 -7.56604075e-01 -4.77435917e-01
-3.00461590e-01 -1.21166751e-01 7.66339302e-01 1.86192513e-01
-7.05370307e-01 5.00930130e-01 -2.51500964e-01 -3.48863959e-01
3.93743441e-02 2.56987423e-01 -3.59539539e-01 2.99080700e-01
-1.92463660e+00 1.44678736e+00 6.73196077e-01 2.56001025e-01
-1.53047293e-01 -3.92078489e-01 -1.49435723e+00 2.68702418e-01
1.01372957e+00 -5.22120297e-01 1.71461225e+00 -8.86538774e-02
-2.04911304e+00 7.46876955e-01 -4.71930951e-01 -7.11496174e-01
1.11086607e-01 -4.14929539e-01 -3.94885927e-01 5.50164431e-02
-1.22052453e-01 4.50851053e-01 1.05714448e-01 -1.14430678e+00
-6.43204510e-01 9.54498574e-02 5.61855853e-01 6.81716263e-01
8.49881619e-02 6.01387501e-01 -4.27455366e-01 -3.51840764e-01
-9.20155793e-02 -7.80140281e-01 -3.37698549e-01 -9.12774622e-01
-8.92312109e-01 -1.06108928e+00 4.83757764e-01 -7.05824614e-01
1.57589507e+00 -1.59253621e+00 1.81521967e-01 -1.23436667e-01
4.56703842e-01 5.26983440e-01 1.24383904e-02 8.53693604e-01
3.56800966e-02 9.82462540e-02 -1.60618156e-01 -2.10515350e-01
-8.53625908e-02 4.26888943e-01 -3.51083636e-01 -3.51032913e-01
5.76747656e-01 1.26303017e+00 -9.75017607e-01 -6.52423561e-01
4.41736951e-02 -2.66562346e-02 -2.93823659e-01 6.28569424e-01
-4.72694397e-01 5.06678820e-01 -6.00315869e-01 1.09946139e-01
2.45083496e-01 -3.78332347e-01 7.33820558e-01 -3.36542934e-01
3.11321557e-01 1.37136447e+00 -7.75133610e-01 1.37288904e+00
-6.67720139e-01 4.54678178e-01 -2.35289901e-01 -6.75449848e-01
1.00903761e+00 5.63777804e-01 1.99814439e-02 -4.93135810e-01
3.05327982e-01 -9.31907371e-02 2.12345004e-01 -4.44871932e-01
8.80185425e-01 1.77072678e-02 -4.37641352e-01 4.09713626e-01
2.58675605e-01 -2.10926875e-01 3.02301049e-01 6.59273267e-01
1.20023835e+00 -3.10695674e-02 1.04866397e+00 8.18051770e-02
8.04992557e-01 -6.65339381e-02 6.21144176e-01 5.37275434e-01
8.00938252e-03 -9.27903503e-03 8.32663298e-01 -4.09695804e-02
-4.80104804e-01 -6.35701537e-01 2.83976495e-02 9.62661326e-01
3.33657444e-01 -1.09715712e+00 -4.69286442e-01 -1.26019669e+00
-2.16596231e-01 8.98916841e-01 -3.84912074e-01 3.93709950e-02
-1.01810765e+00 -3.93800676e-01 5.89182496e-01 3.58199686e-01
8.26224625e-01 -1.09832561e+00 -2.74985075e-01 3.16657990e-01
-8.02084982e-01 -1.83864939e+00 -4.20594603e-01 2.15352461e-01
-5.42363167e-01 -1.06667161e+00 -1.25715807e-01 -6.80498719e-01
2.30428800e-01 3.85688134e-02 1.37857008e+00 -9.78138112e-03
3.11841190e-01 -9.85767394e-02 -7.20137239e-01 2.21184902e-02
-6.73308611e-01 3.70044947e-01 -4.56495017e-01 -3.71291161e-01
5.77160537e-01 -1.63275287e-01 -1.39461294e-01 2.79066116e-01
-2.79554099e-01 2.39150554e-01 1.48842916e-01 9.63425756e-01
-1.56014226e-02 -1.14099897e-01 5.88044167e-01 -1.55665505e+00
1.16089642e+00 -5.74612021e-01 4.98140939e-02 5.51061213e-01
-4.16378975e-01 3.32597852e-01 6.99246585e-01 -1.74246654e-01
-1.47632563e+00 -2.91513026e-01 -2.59504020e-01 4.52546656e-01
-2.83656180e-01 7.93729424e-01 -3.92343342e-01 5.25564253e-01
4.23907787e-01 -1.24725245e-01 -4.19919789e-01 -5.06753206e-01
6.04998589e-01 8.77514303e-01 4.76209372e-01 -6.30027235e-01
5.10969222e-01 -3.01933050e-01 -1.77364379e-01 -7.01753914e-01
-1.42173302e+00 -6.93221629e-01 -8.03212047e-01 8.29527900e-02
8.01089346e-01 -7.35215068e-01 -9.27873969e-01 1.38400331e-01
-1.60343659e+00 -5.56716084e-01 -2.78280824e-01 1.93914771e-01
-2.00480923e-01 2.93482035e-01 -1.11560798e+00 -8.60414267e-01
-3.79466057e-01 -6.15418315e-01 7.64457285e-01 4.75637466e-01
-8.61405134e-01 -1.11929345e+00 -8.23280364e-02 5.99130154e-01
-9.33066159e-02 -7.26448745e-02 9.17756915e-01 -1.38531017e+00
-6.66489452e-02 1.29354730e-01 -3.98118436e-01 8.10124055e-02
5.53667068e-01 -2.89646953e-01 -6.68281436e-01 2.66622871e-01
-1.27368107e-01 -6.06943429e-01 5.35932958e-01 -3.57329786e-01
6.54446900e-01 -7.67156959e-01 -3.78711402e-01 -3.08656879e-02
6.52020395e-01 5.45186102e-01 6.02129757e-01 2.19568133e-01
6.07804120e-01 9.95631933e-01 1.00000739e+00 3.96149457e-01
1.09504068e+00 9.09083009e-01 -2.74670452e-01 -1.17081173e-01
-6.99920729e-02 -2.94913530e-01 2.59848595e-01 9.16035831e-01
9.76191014e-02 -4.64793921e-01 -7.59256363e-01 2.26804361e-01
-1.98068511e+00 -8.27056110e-01 -3.25058728e-01 1.57781148e+00
1.67554164e+00 2.04201296e-01 1.96325570e-01 -1.64699674e-01
7.18485475e-01 2.66822904e-01 -2.33956367e-01 -6.95233107e-01
-1.15342498e-01 2.97783792e-01 1.27548799e-01 1.04889059e+00
-1.06966007e+00 1.57661760e+00 6.04108143e+00 6.86371744e-01
-3.37508112e-01 -1.09191619e-01 4.83399838e-01 6.37979925e-01
-1.17966838e-01 2.45528460e-01 -1.33261847e+00 -7.83937201e-02
1.03606498e+00 -4.05484438e-01 1.42927304e-01 4.09679681e-01
-2.67507620e-02 -3.05908561e-01 -1.14603412e+00 4.97314960e-01
-7.07728788e-02 -1.24522901e+00 1.25760183e-01 -3.89008313e-01
1.61072716e-01 -5.36979079e-01 -6.40859306e-01 5.64224124e-01
9.32507396e-01 -1.03097391e+00 -2.35593133e-02 2.69236028e-01
6.06481194e-01 -6.13709509e-01 1.04702735e+00 3.42551172e-01
-1.39260173e+00 1.85093999e-01 -6.19174726e-02 -1.72762766e-01
4.50423688e-01 1.78204969e-01 -1.63176501e+00 9.86688197e-01
2.40171254e-01 6.89704895e-01 -5.18934786e-01 2.94301867e-01
-9.18830752e-01 5.61779380e-01 -2.36293390e-01 -2.42845416e-01
1.11066148e-01 -1.16990149e-01 6.94351912e-01 1.61253166e+00
-5.51238835e-01 8.56709361e-01 2.47930005e-01 5.48191905e-01
-3.31653804e-01 9.32002589e-02 -3.07295322e-01 -2.90418006e-02
7.90267706e-01 1.50256491e+00 -4.91991013e-01 -5.05842805e-01
-3.99278134e-01 1.20725000e+00 7.28839517e-01 2.26672038e-01
-4.53633606e-01 -6.00465059e-01 4.25339133e-01 -3.92198354e-01
1.48389995e-01 -2.19863862e-01 -2.07446255e-02 -1.19902575e+00
-1.76420882e-01 -8.56821239e-01 7.84535706e-01 -4.48355973e-01
-1.38061857e+00 1.01697230e+00 2.19328448e-01 -6.86729610e-01
-5.73824108e-01 -3.98785740e-01 -5.47119260e-01 1.03636110e+00
-1.34484315e+00 -1.02301693e+00 9.53405797e-02 4.99553174e-01
9.41091776e-01 6.59696467e-04 1.17918515e+00 -1.65221006e-01
-7.57650495e-01 6.23704851e-01 -7.05059409e-01 7.36269712e-01
7.64595866e-01 -1.57306814e+00 8.08655083e-01 4.85298842e-01
1.50788859e-01 8.53682280e-01 6.95031583e-01 -9.09940243e-01
-7.11201787e-01 -8.85394812e-01 1.82039618e+00 -5.32644868e-01
8.52440000e-01 -5.16131639e-01 -1.15853167e+00 8.53392780e-01
6.16403043e-01 -4.82636184e-01 8.89451861e-01 6.28541827e-01
-4.61591095e-01 4.59754437e-01 -9.80623722e-01 7.77161062e-01
9.73603308e-01 -7.23384976e-01 -1.40896285e+00 3.00792754e-01
1.13516569e+00 -8.67183328e-01 -1.27748621e+00 3.60498399e-01
7.05434904e-02 -4.74434882e-01 8.17677498e-01 -9.86414313e-01
6.37046933e-01 6.75224587e-02 8.77467692e-02 -1.45122480e+00
-5.22472663e-04 -1.06267226e+00 -6.87663734e-01 1.69032466e+00
9.07888412e-01 -6.14140570e-01 2.71042585e-01 1.06713295e+00
-1.15460157e-01 -9.86348331e-01 -6.90343201e-01 -6.24733686e-01
-1.05804764e-01 1.09050848e-01 5.31248987e-01 1.00584269e+00
8.99348259e-01 1.56530380e+00 -2.01530427e-01 -2.01724041e-02
-6.41207024e-02 1.71554238e-01 6.72635436e-01 -1.12094212e+00
-3.41458827e-01 -1.59869447e-01 1.39267311e-01 -1.67916787e+00
6.55177951e-01 -7.26888657e-01 2.11737454e-01 -1.83003974e+00
1.70944393e-01 -4.76672173e-01 1.34216458e-01 5.99084556e-01
-8.42265964e-01 -3.90315235e-01 1.35269940e-01 6.12119026e-02
-9.31896567e-01 9.10259485e-01 1.25611722e+00 2.68678516e-02
-5.42620838e-01 1.69291168e-01 -7.14448810e-01 7.97919512e-01
7.08137870e-01 -3.06851238e-01 -3.36205512e-01 7.51409084e-02
-5.57985390e-03 5.96453607e-01 -2.04588547e-01 -1.71827033e-01
3.56964260e-01 -3.08810145e-01 -1.84659064e-01 -4.63314950e-01
4.74641293e-01 -3.57491076e-01 -4.74118620e-01 -4.18448411e-02
-8.91159952e-01 -6.57177269e-02 -8.03465322e-02 2.74944276e-01
-4.17292655e-01 -3.84157121e-01 4.05965060e-01 -8.35429728e-02
-5.64669847e-01 2.18119286e-02 -4.43462908e-01 5.40869772e-01
6.23534203e-01 2.32166335e-01 -4.83854115e-01 -5.97332239e-01
-7.53644347e-01 5.85459530e-01 -1.40805975e-01 5.28516889e-01
4.99245435e-01 -9.36069608e-01 -6.58210933e-01 -6.01474047e-01
1.49971858e-01 3.96762401e-01 -9.39488932e-02 3.79894614e-01
-2.20669843e-02 5.87621212e-01 3.08412343e-01 9.25465152e-02
-1.67961502e+00 1.17641822e-01 3.55304480e-01 -1.06790853e+00
-9.32750821e-01 1.14211988e+00 -5.33358045e-02 -6.95667386e-01
6.22008070e-02 -2.74102002e-01 -1.00418770e+00 1.38893902e-01
5.59898078e-01 2.58049723e-02 -8.24544355e-02 -7.61430860e-01
-3.73310357e-01 -1.83152527e-01 -5.25571048e-01 -2.90823638e-01
1.10433996e+00 -5.92017949e-01 -2.74727881e-01 7.30046093e-01
7.40211070e-01 1.96385249e-01 -9.58119392e-01 -7.90492058e-01
6.37835860e-01 -1.27244428e-01 -4.53136742e-01 -9.19094920e-01
-4.54654455e-01 3.12910050e-01 -5.67993283e-01 5.71229935e-01
6.73575759e-01 4.10172641e-01 1.16076982e+00 8.12028170e-01
2.12326050e-01 -9.92394805e-01 -4.35840636e-02 1.34833896e+00
1.08280790e+00 -1.18224251e+00 3.70300487e-02 -1.24479651e+00
-1.29371524e+00 1.09623051e+00 9.92817879e-01 2.54297078e-01
3.07155818e-01 3.87445003e-01 7.37684891e-02 -3.62932473e-01
-1.09016621e+00 -4.56222743e-01 2.90047139e-01 4.38899487e-01
8.90863717e-01 -1.42648399e-01 -5.72421670e-01 8.84695888e-01
-4.27467853e-01 -4.66879517e-01 5.63393652e-01 7.64265001e-01
-1.41110241e-01 -1.50176966e+00 2.59035140e-01 3.79575372e-01
-3.38683426e-01 -4.10530448e-01 -1.11314929e+00 5.81585288e-01
-4.30958241e-01 1.57131088e+00 -1.11122146e-01 -5.63938856e-01
5.96120775e-01 2.13186979e-01 3.85756493e-01 -1.20202196e+00
-1.21340251e+00 -3.06279033e-01 1.34119332e+00 -4.29398865e-01
-4.32393253e-01 -4.06629682e-01 -1.72435689e+00 -1.78185567e-01
-5.95303893e-01 7.07943976e-01 -1.45701438e-01 1.43296850e+00
1.99503466e-01 6.79764390e-01 8.04002047e-01 -1.43000349e-01
-3.99801791e-01 -1.25431848e+00 -3.53271574e-01 4.34531480e-01
8.61104503e-02 -5.82654357e-01 4.44457904e-02 -2.19947994e-01] | [12.43798542022705, 8.127409934997559] |
1b4581de-7c46-4777-ab7e-c0908750c49c | 190600734 | 1906.00734 | null | https://arxiv.org/abs/1906.00734v3 | https://arxiv.org/pdf/1906.00734v3.pdf | Separate In Latent Space: Unsupervised Single Image Layer Separation | Many real world vision tasks, such as reflection removal from a transparent surface and intrinsic image decomposition, can be modeled as single image layer separation. However, this problem is highly ill-posed, requiring accurately aligned and hard to collect triplet data to train the CNN models. To address this problem, this paper proposes an unsupervised method that requires no ground truth data triplet in training. At the core of the method are two assumptions about data distributions in the latent spaces of different layers, based on which a novel unsupervised layer separation pipeline can be derived. Then the method can be constructed based on the GANs framework with self-supervision and cycle consistency constraints, etc. Experimental results demonstrate its successfulness in outperforming existing unsupervised methods in both synthetic and real world tasks. The method also shows its ability to solve a more challenging multi-layer separation task. | ['Yunfei Liu', 'Feng Lu'] | 2019-06-03 | null | null | null | null | ['intrinsic-image-decomposition', 'reflection-removal'] | ['computer-vision', 'computer-vision'] | [ 5.18053472e-01 1.73764750e-01 -1.80701986e-02 -3.65772009e-01
-4.77896839e-01 -2.27034122e-01 4.71723288e-01 -8.15763116e-01
-1.18119538e-01 7.40546644e-01 -1.91351265e-01 8.07776526e-02
-4.15593497e-02 -4.54586178e-01 -7.19801068e-01 -1.19683278e+00
7.04136014e-01 2.21363366e-01 6.62730709e-02 2.42831483e-01
8.27888250e-02 1.07086137e-01 -1.46828604e+00 1.52871758e-01
1.04133523e+00 9.46829438e-01 4.34510529e-01 2.42902264e-01
-2.94408858e-01 7.43874013e-01 -2.99498260e-01 -2.47416526e-01
4.74548340e-01 -5.28535724e-01 -5.52228570e-01 6.44352674e-01
4.80027020e-01 -2.90390342e-01 8.55451599e-02 1.08202958e+00
2.08774209e-01 -2.13046715e-01 5.89974165e-01 -1.28167057e+00
-5.43871343e-01 2.74080306e-01 -8.58722746e-01 -4.52942610e-01
-2.70328015e-01 -6.39888495e-02 8.10329616e-01 -6.65480375e-01
1.66625202e-01 9.32232559e-01 5.35776258e-01 7.66266227e-01
-1.35447562e+00 -7.96307266e-01 2.97483146e-01 -2.13425547e-01
-1.00812185e+00 -6.07426107e-01 9.92334604e-01 -6.07101023e-01
5.59999049e-01 8.87435153e-02 6.00605011e-01 1.24756348e+00
-1.75641760e-01 7.48617589e-01 1.55873728e+00 -5.39476693e-01
8.34614038e-02 2.77422935e-01 8.74363929e-02 7.84010112e-01
5.51566780e-01 -1.50365859e-01 -7.16718972e-01 2.55243391e-01
9.89342332e-01 -5.11463210e-02 -2.75597662e-01 -3.65848035e-01
-1.03375459e+00 4.73651081e-01 3.59180212e-01 1.80295363e-01
-2.60385480e-02 -7.13667572e-02 -1.31649271e-01 1.48578912e-01
6.69431150e-01 1.44649327e-01 -3.64073038e-01 4.48551327e-01
-8.96668911e-01 -2.13257790e-01 6.36383057e-01 1.05654275e+00
1.09050035e+00 3.21475178e-01 1.16723955e-01 7.67164946e-01
6.78827167e-01 4.80942309e-01 5.27447283e-01 -9.00313497e-01
4.55846310e-01 7.70672321e-01 -9.97942314e-03 -6.60249770e-01
1.13416985e-02 -4.24148887e-01 -1.22479153e+00 4.49563652e-01
2.07991689e-01 -8.89877602e-02 -1.30687678e+00 1.71196342e+00
3.07069868e-01 4.83173341e-01 2.61687458e-01 7.10896969e-01
7.75665820e-01 5.90291321e-01 -5.07742167e-01 -4.87312555e-01
9.87297654e-01 -1.34118950e+00 -8.96216393e-01 -5.01618803e-01
1.98908895e-01 -6.72387302e-01 1.02790403e+00 5.62543452e-01
-9.55297768e-01 -6.44421220e-01 -1.23187828e+00 -1.72215298e-01
-7.14543182e-03 6.05044305e-01 8.84799957e-01 6.17307782e-01
-9.34034705e-01 2.18264490e-01 -1.06229234e+00 -1.55882999e-01
4.39491957e-01 3.85637492e-01 -3.92125458e-01 -2.31911372e-02
-5.30342340e-01 4.46514875e-01 4.63275522e-01 6.72636032e-01
-9.11391914e-01 -4.25492495e-01 -8.78790140e-01 -3.28974932e-01
3.31970662e-01 -9.61288095e-01 9.64469612e-01 -1.42203569e+00
-1.99124384e+00 9.23237920e-01 -3.39166075e-01 -2.39098728e-01
3.98773998e-01 -4.46531653e-01 -1.00197963e-01 2.65116468e-02
5.19780703e-02 2.91183025e-01 1.30059087e+00 -1.62229025e+00
-3.37273180e-01 -3.13549846e-01 -1.52501240e-01 2.56245583e-01
-4.53589171e-01 -1.03318319e-01 -5.74645698e-01 -5.15220344e-01
4.63148594e-01 -9.33700085e-01 -1.68823078e-01 -1.01681657e-01
-5.88318229e-01 6.58338293e-02 7.29393721e-01 -5.72377443e-01
6.13221645e-01 -2.18335438e+00 2.68049389e-01 -2.26203501e-02
4.33473110e-01 1.75914168e-01 -8.01608860e-02 1.52236164e-01
-6.67714924e-02 -3.71057987e-02 -5.19035757e-01 -8.35865736e-01
-1.06057316e-01 4.00022119e-01 -3.36829811e-01 4.63123113e-01
2.49242797e-01 7.31031001e-01 -5.20657420e-01 -4.18957144e-01
7.05078989e-02 3.65382910e-01 -2.91773647e-01 6.03294134e-01
-2.27283701e-01 9.50512230e-01 -3.69069785e-01 3.48799765e-01
8.91427279e-01 -4.42982048e-01 1.93605870e-01 -3.99712056e-01
-1.31980494e-01 2.64416546e-01 -1.19884086e+00 1.87289846e+00
-3.08352232e-01 6.63859010e-01 2.50542223e-01 -1.39214134e+00
9.57330406e-01 3.04205626e-01 4.01911259e-01 -5.52191317e-01
7.97508061e-02 2.16942579e-01 -6.54665753e-02 -6.51533425e-01
-4.68591638e-02 -2.74766922e-01 3.60763699e-01 3.77065390e-01
1.21179931e-01 -1.03908338e-01 4.42984924e-02 -5.78673109e-02
5.26263475e-01 4.65887308e-01 -2.19512597e-01 -3.40543002e-01
4.77006316e-01 -3.33837032e-01 9.82713699e-01 4.38196391e-01
3.97344112e-01 9.29622769e-01 1.83965668e-01 -3.13755810e-01
-8.58699799e-01 -9.81702805e-01 8.82177949e-02 2.12117210e-01
2.83100367e-01 -2.76102096e-01 -9.50998247e-01 -6.10320747e-01
-5.48824131e-01 1.76082566e-01 -5.25966406e-01 -6.00786395e-02
-4.20815289e-01 -9.66575205e-01 1.74214274e-01 3.69543552e-01
1.04862511e+00 -8.19421828e-01 -3.51845592e-01 -1.85532436e-01
-2.38336027e-01 -1.60455179e+00 -2.64246054e-02 8.30227435e-02
-9.32369769e-01 -1.04360461e+00 -5.61417162e-01 -1.03079331e+00
1.16599631e+00 6.02909684e-01 9.13231492e-01 6.52354062e-02
-1.67939261e-01 3.28444749e-01 -1.44979060e-01 -5.39339125e-01
-4.48139198e-03 -4.23419140e-02 1.10021263e-01 7.06380785e-01
-1.88739481e-03 -7.16974735e-01 -4.87968445e-01 2.88576216e-01
-9.88763928e-01 6.44967198e-01 7.02321410e-01 8.40455174e-01
6.38724029e-01 2.92615443e-01 2.58639365e-01 -9.49845552e-01
2.34098628e-01 -2.97351498e-02 -7.55708933e-01 4.54307318e-01
-6.53775215e-01 1.70090914e-01 4.54622447e-01 -3.72565359e-01
-1.57378495e+00 1.02605604e-01 1.90099046e-01 -4.59733725e-01
-1.96716264e-01 3.70660126e-01 -6.82248831e-01 -8.83743986e-02
1.75883383e-01 4.15708095e-01 1.87130764e-01 -6.36604846e-01
1.41325027e-01 4.64802682e-01 4.07879859e-01 -5.77432275e-01
1.21565294e+00 8.05460751e-01 -1.09534390e-01 -9.43440020e-01
-1.18654001e+00 -3.40713024e-01 -7.22613156e-01 -9.78762358e-02
1.11694336e+00 -1.16476536e+00 -3.19139391e-01 1.01979709e+00
-1.09191048e+00 -5.89550316e-01 -6.07999526e-02 5.72710454e-01
-4.86948133e-01 4.83420193e-01 -3.95869046e-01 -7.89357007e-01
-1.99606016e-01 -1.07838643e+00 1.04700184e+00 4.63641554e-01
5.41002691e-01 -1.08613825e+00 1.22430317e-01 6.78535402e-01
1.30241096e-01 2.44211406e-01 5.88733017e-01 5.69880605e-02
-9.62672412e-01 1.61282957e-01 -2.41352648e-01 9.68650043e-01
5.06437421e-01 2.23371103e-01 -1.27672708e+00 -4.53651488e-01
7.66295731e-01 -4.03262436e-01 1.11934209e+00 2.09374830e-01
1.17138946e+00 -3.31718057e-01 -2.42789954e-01 1.09460807e+00
1.53545380e+00 -1.21845245e-01 9.19171572e-01 1.17538251e-01
1.15953970e+00 7.89214194e-01 3.47560644e-01 -2.05267295e-02
3.39759767e-01 4.87961709e-01 5.27226090e-01 -4.77320820e-01
-2.59600252e-01 -8.39586183e-02 5.90962589e-01 1.35296762e+00
-2.28153110e-01 -1.35460392e-01 -6.60090089e-01 3.02085429e-01
-2.01247478e+00 -6.87511623e-01 -2.80133814e-01 2.13370609e+00
7.61529028e-01 -1.48454076e-02 -3.29370022e-01 7.38866329e-02
4.35546786e-01 3.83004285e-02 -5.29328763e-01 8.59027132e-02
-2.52823770e-01 3.38154472e-02 3.24751943e-01 4.03521121e-01
-9.60610628e-01 8.26397955e-01 5.90679979e+00 4.93230969e-01
-1.28251779e+00 -5.56967035e-03 6.33909047e-01 2.24236116e-01
-3.73394281e-01 2.44074896e-01 -8.87587965e-01 4.82345045e-01
4.01428849e-01 5.05897224e-01 4.01971132e-01 2.55888343e-01
-7.16511235e-02 -1.59885541e-01 -1.08611917e+00 1.30305207e+00
2.74005324e-01 -1.01444733e+00 3.57427970e-02 1.90540925e-01
1.09403265e+00 2.14089155e-02 1.29108891e-01 -1.86357915e-01
3.11772376e-01 -9.23539996e-01 4.64561224e-01 7.27624953e-01
6.63408637e-01 -2.77894139e-01 3.70486140e-01 6.34374261e-01
-9.04877543e-01 1.75528646e-01 -3.96086216e-01 -3.43547389e-02
-1.45881131e-01 8.42826366e-01 -4.74801868e-01 9.31220233e-01
7.91522682e-01 1.09092844e+00 -4.12295997e-01 8.13919425e-01
-6.26413524e-01 7.15872943e-01 -3.10735792e-01 4.67094094e-01
-5.36516681e-02 -7.59892762e-01 2.59421915e-01 7.41209507e-01
3.27612311e-01 -1.69902280e-01 3.59121151e-02 1.08134425e+00
1.29600540e-01 -3.27083200e-01 -6.51228607e-01 8.45450088e-02
3.04766838e-02 1.52321756e+00 -6.95770919e-01 -8.04875866e-02
-5.95305443e-01 1.13166606e+00 3.68388176e-01 8.09486985e-01
-8.57309580e-01 6.07704595e-02 3.87375414e-01 -9.10348445e-02
2.43834615e-01 -5.12140751e-01 -3.72847289e-01 -1.73849106e+00
5.18594444e-01 -9.27236855e-01 -1.34816226e-02 -8.56523693e-01
-1.39993119e+00 4.55920517e-01 -1.02601722e-01 -1.39009786e+00
7.63889402e-02 -9.55839813e-01 -8.34796250e-01 8.35687935e-01
-2.00187802e+00 -1.59990013e+00 -6.62883997e-01 7.89596379e-01
3.28256696e-01 -2.91517258e-01 6.19182527e-01 1.62468269e-01
-1.06482220e+00 3.59958291e-01 1.78409919e-01 1.43666968e-01
6.25965476e-01 -1.22128952e+00 4.52099964e-02 1.24213815e+00
3.25972825e-01 5.33376992e-01 2.92190969e-01 -4.66907203e-01
-1.47045779e+00 -1.08290374e+00 3.82255793e-01 -2.10684597e-01
5.42879879e-01 -7.22074687e-01 -8.20405483e-01 8.07674408e-01
5.11730790e-01 -4.89504263e-02 8.05616021e-01 -8.28926861e-02
-4.41516936e-01 -4.62329119e-01 -7.37858772e-01 3.96530509e-01
9.81607854e-01 -4.98250067e-01 -2.79537588e-01 4.83452260e-01
6.42961502e-01 -2.72628278e-01 -6.26106024e-01 5.66453815e-01
3.11481059e-01 -1.06498659e+00 8.16300273e-01 -3.28378677e-01
5.55816710e-01 -4.65254962e-01 3.94912884e-02 -1.00293112e+00
-1.32436082e-01 -7.38029897e-01 -5.53789698e-02 1.62171459e+00
3.74594182e-01 -8.73566151e-01 7.37120032e-01 5.50948977e-01
-1.14066176e-01 -3.48768353e-01 -5.27184665e-01 -7.72255063e-01
-1.83521673e-01 -1.11442462e-01 3.22235912e-01 9.72664654e-01
-6.58892453e-01 6.35586917e-01 -5.73667705e-01 4.93079424e-01
1.12702811e+00 4.63509828e-01 1.05118525e+00 -1.38463759e+00
-5.18888414e-01 -3.11964810e-01 -5.26885018e-02 -1.11326659e+00
2.72947192e-01 -6.95133924e-01 3.18850815e-01 -1.71876955e+00
2.69995987e-01 -4.40053672e-01 -2.00116560e-01 4.87118065e-01
-1.09706379e-01 3.22486967e-01 -8.24635103e-02 5.49022794e-01
-4.26733971e-01 7.95385242e-01 1.31461382e+00 -2.48900741e-01
-2.97359265e-02 3.84560227e-02 -6.52662516e-01 8.90250683e-01
1.02587569e+00 -3.98024887e-01 -7.84744918e-01 -9.68322575e-01
3.94156367e-01 -3.79226416e-01 4.19409573e-01 -1.05482841e+00
2.02227101e-01 -2.22912416e-01 1.51284501e-01 -3.32180679e-01
2.34250486e-01 -9.87506211e-01 2.13765293e-01 1.28813282e-01
5.30571565e-02 -4.20569479e-01 6.80814683e-02 5.32076001e-01
-3.90977293e-01 -2.31558263e-01 7.24076390e-01 -1.58195674e-01
-3.75102699e-01 3.48073423e-01 7.33782649e-02 -9.15626138e-02
8.73900652e-01 -4.40843314e-01 -1.49118230e-01 8.81038085e-02
-5.12599170e-01 1.22523598e-01 5.43637812e-01 2.38462389e-01
6.28550410e-01 -1.35581994e+00 -6.06560767e-01 6.33749068e-01
-6.35734200e-03 7.92942762e-01 1.60049081e-01 1.01206183e+00
-3.36353540e-01 6.20068312e-02 -3.29608291e-01 -8.91830921e-01
-1.13240588e+00 3.28557760e-01 3.90764952e-01 -2.34702617e-01
-5.89053452e-01 8.10682416e-01 7.87932873e-01 -3.62719029e-01
1.98022306e-01 -3.15706104e-01 -1.14770271e-01 -1.71508566e-01
4.14260149e-01 -4.72664870e-02 3.43627371e-02 -3.74024421e-01
4.04840335e-02 9.29018080e-01 -9.76229236e-02 -2.64903530e-02
1.47544527e+00 -2.59463817e-01 -6.31180108e-01 6.83237135e-01
1.07000315e+00 -3.26946042e-02 -1.65226102e+00 -3.81967455e-01
-2.65153736e-01 -2.19137028e-01 -6.47503242e-04 -3.44610393e-01
-1.46720266e+00 1.11005974e+00 3.68727088e-01 2.36247405e-01
1.32316720e+00 -1.85619563e-01 5.53001106e-01 3.39790851e-01
2.06287414e-01 -1.02266538e+00 3.21287930e-01 2.38556579e-01
8.35747361e-01 -1.42338717e+00 -5.51001802e-02 -6.81730151e-01
-3.26173484e-01 1.06406891e+00 9.97422099e-01 -2.28027031e-01
6.12370312e-01 3.36526245e-01 2.50075430e-01 -1.88606292e-01
-5.88541627e-01 -1.62678301e-01 4.27893102e-01 5.47199845e-01
3.34202290e-01 -2.86320537e-01 5.61273098e-02 4.48054075e-01
7.11851791e-02 -2.12019876e-01 3.85574341e-01 7.73499966e-01
-1.33908167e-01 -1.22292995e+00 -1.22894272e-01 2.18318328e-01
-2.38942206e-01 1.72256154e-03 -3.87942195e-01 4.47103143e-01
1.69203386e-01 7.63585210e-01 -6.48039207e-02 -2.38366395e-01
-2.10151181e-01 -2.09085628e-01 6.93835974e-01 -7.12940693e-01
-8.36982727e-02 4.00613606e-01 -2.38791332e-01 -4.32947725e-01
-1.20569909e+00 -3.90439719e-01 -9.87038136e-01 3.94016594e-01
-5.51872849e-01 5.18167354e-02 6.23679876e-01 1.07600617e+00
2.62611210e-01 5.54480493e-01 7.34804451e-01 -8.19334030e-01
-2.96832740e-01 -9.62810695e-01 -5.03126323e-01 4.20028508e-01
4.36624736e-01 -5.74443221e-01 -4.72312510e-01 4.99774724e-01] | [10.030004501342773, -2.6978225708007812] |
1fea16f7-4f88-473e-b8af-6ac42f194540 | zoom-in-net-deep-mining-lesions-for-diabetic | 1706.04372 | null | http://arxiv.org/abs/1706.04372v1 | http://arxiv.org/pdf/1706.04372v1.pdf | Zoom-in-Net: Deep Mining Lesions for Diabetic Retinopathy Detection | We propose a convolution neural network based algorithm for simultaneously
diagnosing diabetic retinopathy and highlighting suspicious regions. Our
contributions are two folds: 1) a network termed Zoom-in-Net which mimics the
zoom-in process of a clinician to examine the retinal images. Trained with only
image-level supervisions, Zoomin-Net can generate attention maps which
highlight suspicious regions, and predicts the disease level accurately based
on both the whole image and its high resolution suspicious patches. 2) Only
four bounding boxes generated from the automatically learned attention maps are
enough to cover 80% of the lesions labeled by an experienced ophthalmologist,
which shows good localization ability of the attention maps. By clustering
features at high response locations on the attention maps, we discover
meaningful clusters which contain potential lesions in diabetic retinopathy.
Experiments show that our algorithm outperform the state-of-the-art methods on
two datasets, EyePACS and Messidor. | ['Hongsheng Li', 'Jianping Shi', 'Yanxin Yin', 'Wei Fang', 'Zhe Wang', 'Xiaogang Wang'] | 2017-06-14 | null | null | null | null | ['diabetic-retinopathy-detection'] | ['medical'] | [ 2.66501844e-01 3.66453201e-01 -6.38528243e-02 -4.89540607e-01
-5.01550078e-01 -1.68580078e-02 7.25935213e-03 9.14488435e-02
-1.29393294e-01 3.93099546e-01 2.49428824e-01 -3.07293952e-01
-1.89319357e-01 -6.73333466e-01 -6.06899440e-01 -6.24928534e-01
-1.32556662e-01 2.81112254e-01 5.59752703e-01 1.48750812e-01
3.97332013e-01 5.00385880e-01 -1.70183837e+00 7.56793857e-01
1.37499523e+00 9.37206149e-01 3.28175813e-01 8.48126292e-01
3.62009145e-02 1.33353937e+00 -4.65797246e-01 -5.68081401e-02
3.76917511e-01 -4.64110136e-01 -6.92603171e-01 4.08318549e-01
8.94000053e-01 -4.74843174e-01 -4.13929522e-01 1.36759281e+00
5.69846869e-01 -2.26674363e-01 5.09959161e-01 -6.74242914e-01
-1.07039404e+00 1.35975167e-01 -8.51662755e-01 8.43064666e-01
-2.59471714e-01 4.53437448e-01 5.93691826e-01 -8.25345516e-01
6.59041345e-01 1.11635494e+00 3.50286901e-01 4.12778676e-01
-9.48612690e-01 -3.45969319e-01 -6.51786774e-02 3.40485960e-01
-1.17486370e+00 -2.37688452e-01 1.87743962e-01 -6.33441627e-01
6.62024558e-01 3.42024475e-01 8.10583115e-01 3.41154933e-01
3.38514149e-01 6.58812463e-01 1.22588384e+00 -4.73563164e-01
8.60341564e-02 1.55870661e-01 1.94382742e-01 8.83920789e-01
2.35766664e-01 1.52674377e-01 -6.04799651e-02 5.71263433e-02
1.25847960e+00 3.92392308e-01 -6.04437768e-01 -1.91077590e-01
-1.17412174e+00 8.40620041e-01 1.05128348e+00 -1.60897702e-01
-5.46649396e-01 -1.60114169e-02 1.07760750e-01 -1.32523298e-01
4.47491318e-01 4.44725275e-01 -7.97067583e-02 3.87219667e-01
-6.76838636e-01 -1.71223640e-01 1.98279563e-02 7.79953182e-01
4.68020529e-01 -3.31280380e-01 -6.07741058e-01 7.31081486e-01
5.19641154e-02 6.88297451e-02 2.51825392e-01 -9.74807262e-01
1.91685557e-02 1.29526949e+00 1.63387775e-01 -5.28451622e-01
-5.26195228e-01 -5.47205806e-01 -9.11968768e-01 9.72492516e-01
4.63441491e-01 -4.09398675e-01 -1.54354560e+00 6.66292548e-01
1.99508503e-01 2.69297272e-01 -1.93517461e-01 1.30437410e+00
1.07109034e+00 2.11696938e-01 -1.52273662e-02 9.90914777e-02
1.62773097e+00 -1.19042408e+00 -5.66081047e-01 -2.36158326e-01
4.62322354e-01 -4.93306875e-01 1.19049895e+00 2.53777564e-01
-1.23198736e+00 -4.46001858e-01 -7.68036723e-01 -3.23178858e-01
-2.89276540e-01 6.41216576e-01 6.75162494e-01 1.28904194e-01
-1.38669574e+00 2.08496302e-01 -4.94799554e-01 -4.74303216e-01
1.15893781e+00 2.93407440e-01 -1.88431174e-01 -2.70118386e-01
-5.36798239e-01 9.51568961e-01 1.28364056e-01 1.37392297e-01
-7.01225996e-01 -8.50627422e-01 -6.48933291e-01 2.88011312e-01
-1.36804823e-02 -7.89057910e-01 1.02876019e+00 -1.08148932e+00
-1.05926204e+00 1.31517673e+00 -3.74900639e-01 -7.41189063e-01
5.71582496e-01 -1.43020466e-01 -4.11715597e-01 6.37917519e-01
1.45275846e-01 1.05390275e+00 5.88843405e-01 -1.02412856e+00
-1.21970820e+00 -2.86564678e-01 7.93183446e-02 1.18476994e-01
8.49887580e-02 5.72507322e-01 -6.34039283e-01 -3.64829689e-01
-1.11231178e-01 -5.19457102e-01 -6.78061724e-01 4.81812268e-01
-7.41274655e-01 -2.35997930e-01 6.23876572e-01 -4.38204706e-01
1.14156950e+00 -2.09261060e+00 -4.52294141e-01 1.42175183e-01
9.54566061e-01 5.28046131e-01 3.17256935e-02 -2.87715882e-01
-3.15633684e-01 3.06278635e-02 2.42460147e-01 1.64068401e-01
-6.12301648e-01 -9.58779976e-02 -2.06753314e-01 6.13349020e-01
5.57781160e-01 1.01497996e+00 -8.95217240e-01 -7.00118303e-01
4.80690002e-01 5.41689157e-01 -3.92656058e-01 2.29622442e-02
-6.74960613e-02 4.67094511e-01 -3.10728580e-01 8.92417192e-01
5.23319423e-01 -1.03103960e+00 -1.90053016e-01 -1.48516566e-01
-1.94791317e-01 4.29463666e-03 -5.86655498e-01 9.60515976e-01
1.46452785e-01 1.09203053e+00 -3.45339864e-01 -5.22226989e-01
7.78477192e-01 8.90684947e-02 2.56483436e-01 -5.64243436e-01
2.81057149e-01 -1.01028860e-01 3.15984577e-01 -1.02724671e+00
-1.46001264e-01 2.20857114e-01 6.56429291e-01 1.02139585e-01
-1.70476064e-01 6.30183399e-01 2.23677650e-01 1.14555195e-01
1.08859301e+00 -4.56879288e-01 3.35622102e-01 -1.41407356e-01
1.74335763e-01 7.93986842e-02 5.51442146e-01 8.56973648e-01
-5.49857974e-01 8.83158803e-01 9.12893176e-01 -9.61057186e-01
-9.67050374e-01 -7.69266903e-01 -5.27917802e-01 7.48055875e-01
1.94157749e-01 -2.89709996e-02 -7.93653369e-01 -7.94296920e-01
2.91105313e-03 2.48762146e-01 -1.28873301e+00 1.87128499e-01
-2.70055383e-01 -8.16364110e-01 -6.02055639e-02 7.51640141e-01
3.32182407e-01 -1.25172091e+00 -9.39059019e-01 -2.48201385e-01
2.19862372e-01 -6.81580067e-01 -4.75582808e-01 -2.89297048e-02
-8.20181489e-01 -1.55778027e+00 -9.14399743e-01 -1.24710667e+00
1.31284344e+00 4.32467878e-01 1.19621360e+00 1.80156320e-01
-9.77177262e-01 -2.99683154e-01 -1.03027135e-01 -7.95482814e-01
-1.27185866e-01 -5.80900729e-01 -5.12293279e-01 3.31156611e-01
9.24420595e-01 -3.38826776e-01 -1.21885061e+00 2.24973619e-01
-3.43480617e-01 1.70467287e-01 1.15909612e+00 6.82070494e-01
8.21205735e-01 -2.06609908e-02 2.85491914e-01 -1.09262729e+00
3.18345904e-01 -1.98513567e-01 -7.62086451e-01 2.92831749e-01
-3.49830061e-01 -3.38326335e-01 4.69365083e-02 -3.86021197e-01
-6.20520949e-01 2.94556320e-01 6.37052476e-01 -7.74959385e-01
-5.15213490e-01 -7.24217966e-02 4.50040996e-01 -3.23687285e-01
1.05077851e+00 3.33605260e-02 2.58705646e-01 -4.23704326e-01
3.20639074e-01 8.50678146e-01 7.84457147e-01 4.03402030e-01
2.24957451e-01 7.25336790e-01 -2.26916894e-01 -4.88791257e-01
-1.26017094e+00 -6.81352735e-01 -5.66568851e-01 -2.59364575e-01
1.16052473e+00 -7.71044672e-01 -9.72500682e-01 1.59642875e-01
-9.94193316e-01 -2.12621197e-01 -4.78315860e-01 4.64473933e-01
-3.26839864e-01 -1.03327744e-02 -6.10118747e-01 -5.40667057e-01
-4.32787836e-01 -1.16269994e+00 1.07144475e+00 8.91803205e-01
9.87337902e-02 -6.99900985e-01 -2.65938491e-01 2.85938233e-01
3.05790782e-01 4.44120437e-01 1.14991546e+00 -2.86099643e-01
-1.05782676e+00 -2.05507562e-01 -1.01000702e+00 2.16403857e-01
6.15887642e-02 3.38631988e-01 -9.89372194e-01 -3.17891091e-02
-5.45112431e-01 -1.08784195e-02 1.13363874e+00 1.33734953e+00
1.59005439e+00 -3.23916972e-01 -6.73470020e-01 8.91766548e-01
1.35567760e+00 2.26357996e-01 1.04574621e+00 2.84763426e-01
4.10834849e-01 5.52477181e-01 3.77345920e-01 1.26790971e-01
1.43370807e-01 2.37885177e-01 7.97240853e-01 -1.13946831e+00
-5.45691967e-01 1.57152817e-01 -2.58751154e-01 -1.51483208e-01
-3.55056018e-01 1.15455382e-01 -9.00938451e-01 1.02542627e+00
-1.81539381e+00 -7.43307769e-01 -3.74312043e-01 2.16729188e+00
8.41006041e-01 1.92710832e-02 2.17442587e-01 -4.22734290e-01
1.05551720e+00 -3.12938362e-01 -8.11399043e-01 -1.28461748e-01
-6.21023998e-02 1.68999344e-01 6.00821078e-01 2.80939817e-01
-1.32704949e+00 6.79613233e-01 7.03801298e+00 2.18778178e-01
-1.05653071e+00 -1.55754864e-01 1.04819572e+00 -5.62075198e-01
3.32956016e-01 -2.38425016e-01 -7.25832403e-01 6.08135700e-01
4.32766259e-01 7.61708990e-02 -9.10646166e-04 8.32208216e-01
5.33990443e-01 -9.14327949e-02 -1.02460122e+00 8.75230551e-01
1.23926103e-01 -1.75160837e+00 2.17394382e-01 1.92821920e-01
9.82101083e-01 3.29313457e-01 3.44671100e-01 -2.77668625e-01
5.59734285e-01 -1.53739178e+00 -1.34209946e-01 9.70866561e-01
1.21462488e+00 -5.87505937e-01 9.79680538e-01 -2.54683107e-01
-4.88482416e-01 -4.02075648e-01 -6.78066790e-01 2.27904961e-01
-1.81944996e-01 6.13966107e-01 -1.26917338e+00 -2.93744773e-01
1.05134833e+00 9.46104467e-01 -9.90514994e-01 1.92102933e+00
-2.32034624e-01 4.99720186e-01 2.01505929e-01 1.67449847e-01
3.47439051e-01 -1.36824355e-01 5.17508209e-01 1.15515125e+00
8.34081024e-02 1.93803519e-01 -1.25890896e-01 1.22793245e+00
1.31081283e-01 8.37692022e-02 -4.20806140e-01 3.35269511e-01
2.23665118e-01 1.31509066e+00 -5.41795075e-01 -3.66301328e-01
-4.82150137e-01 6.83685780e-01 1.67320222e-01 7.03265607e-01
-5.32650530e-01 -8.11599076e-01 7.50527442e-01 5.71527481e-01
4.90714461e-01 6.72914147e-01 -5.46535850e-01 -7.10898459e-01
-7.62724131e-02 -6.38102651e-01 4.78786498e-01 -1.33193409e+00
-1.29045844e+00 6.88104510e-01 -9.23642993e-01 -1.62262392e+00
3.15439045e-01 -7.45151341e-01 -1.08828259e+00 1.07249069e+00
-1.80080557e+00 -1.06060743e+00 -8.60604703e-01 6.55376375e-01
3.51847410e-01 -3.70026916e-01 6.36102796e-01 1.64957747e-01
-7.21416652e-01 3.19449902e-01 -1.00792758e-02 4.69918817e-01
8.29403460e-01 -1.63332951e+00 2.37840056e-01 8.31478953e-01
-2.02027634e-01 5.65260172e-01 2.99986780e-01 -4.30378318e-01
-3.82827133e-01 -1.58415473e+00 8.92220378e-01 -5.79765320e-01
4.46953118e-01 2.29099169e-01 -8.73110354e-01 8.12609375e-01
2.63646156e-01 6.80598080e-01 5.46709061e-01 1.63235813e-02
-1.80187840e-02 4.48832847e-03 -9.45265412e-01 6.12735569e-01
6.92052841e-01 -2.00812966e-01 -4.83983099e-01 9.53519166e-01
5.82629740e-01 -7.39781201e-01 -7.02537894e-01 2.37786442e-01
1.80322692e-01 -1.15135264e+00 8.93874764e-01 -1.13436484e+00
8.59476089e-01 -3.97751182e-01 4.59934860e-01 -1.01423323e+00
-5.61009586e-01 -6.81126773e-01 -5.91657199e-02 5.29255509e-01
3.80467743e-01 -6.41785979e-01 7.47719169e-01 3.21577966e-01
-2.19837025e-01 -1.15537047e+00 -5.31875134e-01 7.93763921e-02
-2.63478994e-01 4.21133667e-01 3.45288336e-01 7.61109233e-01
-1.13580130e-01 7.89500400e-02 1.50474217e-02 6.19939983e-01
5.92003465e-01 2.69617498e-01 4.33749080e-01 -1.36812079e+00
4.30556871e-02 -5.17175078e-01 -9.07129943e-01 -7.72369742e-01
-6.81248307e-01 -6.77014530e-01 -2.30461225e-01 -1.83864999e+00
5.26518404e-01 -2.68804908e-01 -6.32952929e-01 7.75590122e-01
-4.54505503e-01 4.65744257e-01 -1.69194281e-01 2.20425129e-01
-8.09885025e-01 -1.00315548e-01 1.65082121e+00 -1.14649087e-01
-4.11519855e-01 1.91891864e-01 -1.14092755e+00 9.29962516e-01
5.74204445e-01 -1.29200146e-01 -1.83937654e-01 -3.06597471e-01
-1.74914405e-01 -1.92281231e-01 7.79317558e-01 -9.74865913e-01
4.51204538e-01 -5.34720533e-02 7.89218307e-01 -7.28168309e-01
-1.40768498e-01 -3.52945507e-01 -6.09655917e-01 2.08453298e-01
-4.43542600e-01 -4.20506150e-01 9.21261217e-03 5.33351183e-01
-2.11271062e-01 3.03758889e-01 1.35464120e+00 -2.51049310e-01
-6.80426478e-01 3.23017329e-01 -1.81236148e-01 9.25130919e-02
1.15362120e+00 -4.08048004e-01 -1.08016562e+00 -2.23211393e-01
-9.50522125e-01 3.94758791e-01 4.29479629e-01 2.34938204e-01
6.30895555e-01 -9.40178931e-01 -9.76634264e-01 4.44768429e-01
3.53008449e-01 4.05281723e-01 5.27775705e-01 1.24119341e+00
-8.22135329e-01 5.77582121e-01 -3.43215615e-01 -9.22759533e-01
-1.47351170e+00 6.34639144e-01 9.22179580e-01 1.97280511e-01
-1.32654476e+00 1.02182591e+00 7.89422870e-01 2.26736382e-01
5.11814892e-01 -8.02684724e-01 -5.27040660e-01 -3.24817777e-01
1.20462871e+00 2.19245002e-01 -8.19557086e-02 -7.37057626e-02
-1.35386288e-01 6.62404597e-01 -4.55654472e-01 7.47360587e-01
1.28894949e+00 4.47655506e-02 -2.53598154e-01 -1.43055066e-01
6.35126829e-01 -2.00199619e-01 -1.56746316e+00 -3.72559756e-01
-4.37927604e-01 -8.39238703e-01 4.75166500e-01 -1.35191619e+00
-1.31151092e+00 9.98347282e-01 1.14381003e+00 1.68247342e-01
1.34215927e+00 1.27556294e-01 5.22777498e-01 -5.70747517e-02
-2.69322634e-01 -7.01780975e-01 -8.53530616e-02 -1.14096232e-01
8.48679900e-01 -1.42643428e+00 -1.14254124e-01 -5.88701367e-01
-9.85043406e-01 1.00803864e+00 8.89518678e-01 -3.05929840e-01
4.34955895e-01 1.19563870e-01 6.16836786e-01 -6.24467611e-01
-7.32943773e-01 -5.50202549e-01 6.78778529e-01 9.66355264e-01
1.58544421e-01 1.44398473e-02 2.76049972e-01 3.40563834e-01
2.02352956e-01 2.52219826e-01 8.20248663e-01 3.96109879e-01
-9.78633046e-01 -3.60462815e-01 -1.76814005e-01 1.07460272e+00
-4.52114463e-01 -3.57674271e-01 -4.87422913e-01 6.86635733e-01
3.08165789e-01 6.28692448e-01 5.14880538e-01 -2.40551177e-02
2.13061213e-01 -2.68483073e-01 1.11898310e-01 -8.12092304e-01
-4.41909611e-01 1.96790352e-01 -8.51344168e-02 -8.05528402e-01
1.22595830e-02 -2.42629692e-01 -1.11157215e+00 1.16901025e-01
-1.67267188e-01 -3.47182155e-01 4.54465188e-02 4.65409547e-01
8.45046282e-01 8.87791812e-01 4.09598887e-01 -5.87435067e-01
-1.65871158e-01 -1.06682467e+00 -9.08139944e-01 1.75570682e-01
8.14755976e-01 -5.35733223e-01 -1.35370106e-01 4.22349244e-01] | [15.789183616638184, -3.955566883087158] |
e6522313-8057-4e02-9663-c622a356d9f1 | traffic-surveillance-camera-calibration-by-3d | 1702.06451 | null | http://arxiv.org/abs/1702.06451v2 | http://arxiv.org/pdf/1702.06451v2.pdf | Traffic Surveillance Camera Calibration by 3D Model Bounding Box Alignment for Accurate Vehicle Speed Measurement | In this paper, we focus on fully automatic traffic surveillance camera
calibration, which we use for speed measurement of passing vehicles. We improve
over a recent state-of-the-art camera calibration method for traffic
surveillance based on two detected vanishing points. More importantly, we
propose a novel automatic scene scale inference method. The method is based on
matching bounding boxes of rendered 3D models of vehicles with detected
bounding boxes in the image. The proposed method can be used from arbitrary
viewpoints, since it has no constraints on camera placement. We evaluate our
method on the recent comprehensive dataset for speed measurement BrnoCompSpeed.
Experiments show that our automatic camera calibration method by detection of
two vanishing points reduces error by 50% (mean distance ratio error reduced
from 0.18 to 0.09) compared to the previous state-of-the-art method. We also
show that our scene scale inference method is more precise, outperforming both
state-of-the-art automatic calibration method for speed measurement (error
reduction by 86% -- 7.98km/h to 1.10km/h) and manual calibration (error
reduction by 19% -- 1.35km/h to 1.10km/h). We also present qualitative results
of the proposed automatic camera calibration method on video sequences obtained
from real surveillance cameras in various places, and under different lighting
conditions (night, dawn, day). | ['Roman Juránek', 'Adam Herout', 'Jakub Sochor'] | 2017-02-21 | null | null | null | null | ['vehicle-speed-estimation'] | ['computer-vision'] | [ 1.82927903e-02 -2.66038626e-01 4.17995788e-02 -2.59778380e-01
-5.75679302e-01 -6.74129844e-01 6.16472363e-01 -5.52352779e-02
-6.99361801e-01 3.64709377e-01 -3.25954944e-01 -5.87310910e-01
2.45679468e-01 -7.29102075e-01 -8.25190604e-01 -6.00479603e-01
3.16905409e-01 4.66547966e-01 1.02075815e+00 -2.32144892e-01
1.74187243e-01 7.60586321e-01 -1.29239976e+00 -3.73350769e-01
4.79846776e-01 7.69513011e-01 -1.37605384e-01 1.11057293e+00
3.62090588e-01 4.26937997e-01 -6.38167799e-01 -6.97638690e-01
4.46018696e-01 -8.56581703e-02 -2.02902734e-01 3.48998696e-01
9.12471294e-01 -5.94784617e-01 -3.11783373e-01 1.05772746e+00
-1.16459411e-02 -3.20355929e-02 4.95501816e-01 -1.33846509e+00
1.21344596e-01 -4.68823612e-01 -7.81483948e-01 4.40499336e-01
3.15222770e-01 -7.42897717e-03 2.26228863e-01 -2.63444006e-01
5.29215753e-01 9.50367093e-01 8.63528907e-01 5.41260004e-01
-1.05250025e+00 -8.09255838e-01 -1.55934870e-01 2.10114732e-01
-1.66604042e+00 -5.14374077e-01 8.05513918e-01 -6.70543790e-01
7.02102184e-01 1.50214925e-01 6.70480430e-01 7.58014977e-01
2.08034903e-01 2.26058830e-02 8.91331911e-01 -3.83614242e-01
1.83819607e-01 4.71980691e-01 2.45215535e-01 9.48187590e-01
7.98128843e-01 2.15392321e-01 2.34619886e-01 -6.99849054e-03
9.29030240e-01 2.31358670e-02 -4.08112369e-02 -4.62014824e-01
-1.06884181e+00 8.40154469e-01 3.02378573e-02 3.04987524e-02
2.21617222e-02 1.99221417e-01 1.74540609e-01 -3.42425555e-02
4.49858129e-01 -5.52559234e-02 -2.50175506e-01 -4.23469931e-01
-9.77667153e-01 3.11840981e-01 5.64962983e-01 1.32351172e+00
6.28349006e-01 -9.49300304e-02 1.81685194e-01 3.90262604e-01
1.56579986e-01 1.30337310e+00 -2.08166197e-01 -1.08236182e+00
6.61082625e-01 3.98919612e-01 4.19456780e-01 -1.19865787e+00
-5.92715263e-01 4.04986320e-03 -5.79930007e-01 1.96461678e-01
5.93674183e-01 -3.25397372e-01 -4.32119519e-01 1.44247699e+00
6.05812728e-01 3.88408840e-01 2.62335669e-02 9.42257345e-01
5.51335573e-01 5.83237469e-01 -1.82551622e-01 -4.52668190e-01
1.34275711e+00 -7.61201262e-01 -5.28393149e-01 7.39645436e-02
7.73637116e-01 -8.10440123e-01 4.90539134e-01 4.27143782e-01
-7.46343136e-01 -7.00104833e-01 -8.21088195e-01 4.59793478e-01
-2.25452304e-01 3.35562617e-01 1.50359645e-01 1.22347498e+00
-9.46313679e-01 4.19770032e-02 -7.27874279e-01 -7.27046311e-01
3.78802195e-02 2.19255388e-01 -3.73326361e-01 1.19501069e-01
-7.59842098e-01 1.00909996e+00 4.74625034e-03 -2.13011354e-01
-8.30803990e-01 -5.42452335e-01 -8.20028782e-01 -4.25147057e-01
4.89862174e-01 -5.77295363e-01 1.36388934e+00 -6.95807517e-01
-1.67456079e+00 8.63652706e-01 -3.89057070e-01 -6.04185760e-01
8.63675535e-01 -3.55498880e-01 -8.48187804e-01 5.82608402e-01
1.21118084e-01 5.10688245e-01 5.26150465e-01 -1.18721175e+00
-8.71629894e-01 -2.33559087e-01 1.77306548e-01 -1.32133946e-01
-4.95163426e-02 3.45396757e-01 -7.04912424e-01 -2.71399349e-01
-2.15568617e-01 -1.24526072e+00 -1.30750105e-01 1.57628670e-01
-1.27110234e-03 1.57481015e-01 1.16258502e+00 -6.17755115e-01
1.05480027e+00 -1.84149015e+00 -3.52268875e-01 1.16176710e-01
1.15730606e-01 7.28286505e-01 1.45777017e-01 8.05794150e-02
2.95270890e-01 -1.68281883e-01 -1.37271583e-01 -4.00073498e-01
-4.06080574e-01 1.19105607e-01 4.03826684e-02 1.03473711e+00
-6.38085082e-02 3.39211285e-01 -8.90324950e-01 -5.93480349e-01
8.36342692e-01 6.91970050e-01 -3.59769106e-01 2.34097689e-01
5.32396257e-01 3.55373085e-01 -1.20873809e-01 5.18534303e-01
1.18325233e+00 4.13789958e-01 -1.61972180e-01 -2.03910053e-01
-5.58445752e-01 -3.66971463e-01 -1.12348258e+00 9.90449607e-01
-3.51778686e-01 1.05632818e+00 -9.24932137e-02 -7.40779281e-01
9.97812390e-01 3.68508101e-01 3.99925470e-01 -4.79904711e-01
3.80521595e-01 -1.20542385e-01 -3.19464833e-01 -6.34908259e-01
4.49330270e-01 -5.02682067e-02 3.38035896e-02 -1.77227780e-01
-1.70369178e-01 -3.27750683e-01 5.40711105e-01 1.66297674e-01
7.51978338e-01 -2.68554270e-01 2.73541540e-01 -4.11182284e-01
8.85788858e-01 5.52416071e-02 4.85440373e-01 5.19030213e-01
-6.52843893e-01 4.77762312e-01 4.84079778e-01 -6.52725816e-01
-1.39255011e+00 -9.76062000e-01 -2.89461017e-01 5.08526742e-01
4.26035434e-01 -3.41328710e-01 -1.28987265e+00 -6.41331136e-01
-2.16754615e-01 5.61920822e-01 -6.49356604e-01 1.97697431e-01
-7.72623301e-01 -4.39849585e-01 5.49148619e-01 7.07150578e-01
7.54051149e-01 -1.26709044e-01 -9.52384412e-01 -1.11561365e-01
3.79026891e-03 -1.97709870e+00 -4.50968385e-01 -8.47992063e-01
-7.93998539e-01 -1.46867061e+00 -6.34878516e-01 -3.44563067e-01
9.76232290e-01 9.90723193e-01 8.97046983e-01 3.27450149e-02
-2.47652289e-02 8.49542022e-01 -3.92848551e-01 -5.25064647e-01
-5.49331844e-01 -2.85484821e-01 2.85667211e-01 2.25105986e-01
4.79738146e-01 1.69954330e-01 -5.30636430e-01 9.59639490e-01
-6.46904826e-01 2.17164591e-01 2.39778563e-01 1.59496263e-01
3.03398103e-01 1.38500743e-02 -1.59385502e-01 -7.03319967e-01
4.64508077e-03 -1.93166301e-01 -1.59299982e+00 3.16696577e-02
-2.97532678e-01 -3.43322545e-01 6.27980649e-01 -4.10772532e-01
-8.95707548e-01 1.96513861e-01 8.97244439e-02 -6.28566504e-01
-5.00293016e-01 -4.15419787e-01 1.85039863e-02 -3.94869417e-01
5.24483562e-01 -7.29083642e-02 -1.08059816e-01 -7.43980482e-02
1.65454492e-01 5.81860542e-01 7.25369513e-01 -1.09988181e-02
1.13003707e+00 9.25519109e-01 4.61443007e-01 -1.34117961e+00
-3.64892900e-01 -8.32604766e-01 -9.80611801e-01 -6.99383020e-01
1.21285760e+00 -8.41696560e-01 -7.62150049e-01 4.43728268e-01
-1.27054214e+00 -9.10325125e-02 3.82873029e-01 7.80942023e-01
-3.12229425e-01 6.88932180e-01 -3.09841812e-01 -9.62531388e-01
-9.36384872e-02 -1.21756577e+00 1.28668427e+00 2.47474968e-01
1.68522120e-01 -1.22585535e+00 2.01147243e-01 5.83789885e-01
4.61922526e-01 4.17910784e-01 -8.73658285e-02 -7.20597431e-02
-4.17497814e-01 -6.74080968e-01 -2.69714713e-01 3.13378721e-01
-1.34918004e-01 4.36692566e-01 -7.64387190e-01 -3.05299938e-01
-1.26044199e-01 6.24692917e-01 6.14864111e-01 5.60096264e-01
6.09429598e-01 -7.53079876e-02 -5.25349557e-01 5.68334997e-01
1.44755340e+00 3.60238582e-01 5.85707426e-01 3.39127064e-01
7.04524100e-01 3.84374708e-01 7.02658832e-01 3.19625795e-01
4.99725133e-01 1.11505735e+00 4.81106073e-01 -1.94052652e-01
-5.35560735e-02 1.01689612e-02 4.28711325e-01 4.64187622e-01
-3.89554530e-01 -7.44751394e-02 -9.60396230e-01 3.82457733e-01
-1.68878186e+00 -1.01882958e+00 -7.44753063e-01 2.58544803e+00
-1.14542618e-01 3.71079415e-01 5.93703985e-01 1.98166698e-01
1.05019534e+00 -1.63700730e-01 2.08131522e-01 -4.38564837e-01
2.84507334e-01 -4.18097258e-01 1.11557150e+00 8.50542545e-01
-1.44507587e+00 8.02664876e-01 5.89167690e+00 1.98737636e-01
-1.02341413e+00 1.82572350e-01 3.89956385e-01 9.40919146e-02
4.07832503e-01 1.70257706e-02 -1.18905759e+00 5.57855725e-01
1.13047993e+00 1.74512595e-01 1.19564876e-01 8.69488418e-01
4.56128478e-01 -5.09804606e-01 -8.18515718e-01 1.32295966e+00
4.55801278e-01 -1.01604903e+00 -1.45051837e-01 4.67683226e-02
6.98425829e-01 -1.23085625e-01 -3.45588654e-01 -3.49916965e-02
-1.41945243e-01 -2.47466058e-01 8.26192379e-01 3.91771764e-01
7.77531803e-01 -8.38583291e-01 1.07994533e+00 3.16841692e-01
-1.56393492e+00 1.54825807e-01 -5.60541451e-01 -6.77891672e-02
4.65167463e-01 5.13005435e-01 -8.48577261e-01 3.40228468e-01
5.73619783e-01 4.85763073e-01 -8.13373685e-01 1.15509546e+00
-1.72369376e-01 6.58078372e-01 -4.05674130e-01 2.37212269e-04
2.51297265e-01 -3.99565220e-01 6.77700877e-01 1.71220243e+00
3.57753068e-01 2.85623699e-01 -2.58802939e-02 3.21121097e-01
3.63527000e-01 -1.51243344e-01 -8.64302397e-01 8.38052392e-01
3.06594461e-01 1.25760341e+00 -7.80419409e-01 -7.50489771e-01
-6.44243479e-01 6.60145402e-01 -2.07212299e-01 2.08268985e-01
-1.58676672e+00 -4.37247574e-01 5.91818273e-01 5.05708277e-01
3.58361214e-01 -4.70504820e-01 1.98080063e-01 -1.14546943e+00
2.84159705e-02 -9.85299945e-02 1.07067622e-01 -6.02328241e-01
-5.42886913e-01 5.61524272e-01 6.65201485e-01 -1.53273404e+00
-7.03947246e-02 -8.09425116e-01 -7.06322730e-01 1.87661171e-01
-1.29460955e+00 -1.02294528e+00 -7.40009606e-01 5.56517720e-01
5.72449028e-01 -1.61446840e-01 3.38816583e-01 5.05586445e-01
-6.78868175e-01 5.56036353e-01 -1.30095892e-02 3.43494415e-01
4.41173553e-01 -7.84269810e-01 5.36131561e-01 1.27216625e+00
-2.38941133e-01 1.24495372e-01 7.46192932e-01 -3.78525108e-01
-1.32531261e+00 -1.22259009e+00 9.29224253e-01 -8.66300642e-01
6.25230432e-01 -5.14517486e-01 -4.10133123e-01 4.91519481e-01
-2.84547894e-03 2.46494979e-01 3.25429440e-01 -4.63168383e-01
-3.63604933e-01 -5.76448679e-01 -1.33274186e+00 4.14242834e-01
7.61760235e-01 -2.22246245e-01 -4.17473137e-01 2.17942148e-01
5.45809627e-01 -3.11225355e-01 -4.61604506e-01 3.21876556e-01
6.56154394e-01 -1.08607566e+00 8.52072358e-01 -1.46271214e-01
-3.75242680e-01 -6.48087025e-01 -1.86851129e-01 -8.36796224e-01
-2.64622957e-01 -6.28645599e-01 1.62350670e-01 1.01558244e+00
1.17033571e-01 -7.18091667e-01 5.83135426e-01 5.73480427e-01
-1.40603065e-01 -1.48050323e-01 -1.20982969e+00 -9.38346803e-01
-1.90789476e-01 -7.67178833e-01 1.29331112e-01 5.69501042e-01
-2.97413170e-01 2.46334970e-01 -4.42038774e-01 5.79103947e-01
9.28831816e-01 -4.74844724e-01 1.35902083e+00 -1.20834148e+00
5.78842908e-02 -3.89643610e-01 -9.64803576e-01 -9.79012907e-01
-1.01310693e-01 6.15063636e-03 -3.41200262e-01 -1.24013841e+00
3.38788956e-01 9.28414762e-02 2.52736211e-01 -1.15394324e-01
1.12233207e-01 3.41696888e-01 3.47986102e-01 -1.44444421e-01
-8.12682271e-01 -7.59718269e-02 7.58028150e-01 1.13068841e-01
6.67500570e-02 1.15972847e-01 -9.69399810e-02 9.10387516e-01
9.51980948e-01 -4.51887846e-01 -2.68805802e-01 -4.19272542e-01
-1.87843919e-01 1.92564726e-01 7.01761961e-01 -1.58285058e+00
2.09355533e-01 -1.44111812e-01 1.58147708e-01 -5.36516905e-01
3.49536806e-01 -1.17208183e+00 2.01579064e-01 7.02027082e-01
2.20996976e-01 4.42093641e-01 4.16392684e-01 4.24878001e-01
1.75648883e-01 -9.23706889e-02 1.07488322e+00 1.12786412e-01
-5.44765651e-01 7.28188083e-02 -6.72617257e-01 -8.73052105e-02
1.45245516e+00 -4.97465640e-01 -4.35204327e-01 -4.12803978e-01
-1.74107581e-01 -2.00963065e-01 6.42468929e-01 3.85186285e-01
4.11081880e-01 -1.18981147e+00 -6.62078083e-01 3.38457882e-01
9.37877744e-02 -3.81311625e-01 4.09140252e-02 1.04119122e+00
-1.09197867e+00 8.55896950e-01 -2.84574237e-02 -9.58503366e-01
-1.70791709e+00 8.73504519e-01 3.08983564e-01 1.41136199e-01
-3.23405236e-01 1.82249561e-01 1.38004050e-01 -5.17874360e-02
8.34892911e-04 -6.66832030e-01 -1.00725368e-01 -4.10754710e-01
6.50142312e-01 9.70001757e-01 8.21620896e-02 -1.25192416e+00
-6.11952305e-01 1.46157825e+00 2.10481659e-01 -5.65611199e-02
7.80030131e-01 -1.75141156e-01 3.40264618e-01 7.27622956e-02
1.33334100e+00 1.38563693e-01 -1.43952310e+00 2.95491904e-01
-5.05472898e-01 -7.02664435e-01 -2.42329706e-02 -1.92723036e-01
-1.07031572e+00 7.47852206e-01 9.44475591e-01 2.38785058e-01
1.05850923e+00 -7.81478882e-02 5.72048545e-01 4.11360741e-01
5.21929204e-01 -9.90249813e-01 -3.04368079e-01 2.34323233e-01
4.53231961e-01 -1.48737001e+00 7.50027448e-02 -6.66117430e-01
-5.15504539e-01 1.14164710e+00 6.84549093e-01 -1.45015582e-01
5.88749528e-01 1.25006109e-01 1.51240721e-01 6.57514036e-02
-4.15868878e-01 -1.14801310e-01 2.35529691e-01 7.26075172e-01
1.74873531e-01 2.54356086e-01 -2.29804695e-01 -2.05742776e-01
3.73205021e-02 -2.07924455e-01 7.51210451e-01 7.63408124e-01
-5.61493337e-01 -7.24695146e-01 -8.43104959e-01 -1.20798334e-01
-1.58393115e-01 3.69033217e-01 -1.86608195e-01 1.24616551e+00
1.52478889e-01 1.39365184e+00 3.74060392e-01 -4.33222950e-01
6.84826255e-01 -5.07859290e-01 3.99519652e-01 1.14646390e-01
-1.03945099e-01 6.60666600e-02 1.84177980e-01 -5.80568671e-01
-8.24980855e-01 -7.09487140e-01 -8.90952051e-01 -6.64601862e-01
-4.72332656e-01 1.82835817e-01 7.80950487e-01 5.61555624e-01
-2.71374881e-02 6.79879263e-03 7.84676075e-01 -9.19494510e-01
3.67629938e-02 -6.44165993e-01 -2.15628132e-01 4.78206277e-01
4.89564180e-01 -9.07158911e-01 -4.37776119e-01 3.53021413e-01] | [8.025541305541992, -1.2039401531219482] |
eafc5604-894d-4e5e-b30f-4790db3fabce | injecting-domain-knowledge-in-language-models | 2212.08120 | null | https://arxiv.org/abs/2212.08120v1 | https://arxiv.org/pdf/2212.08120v1.pdf | Injecting Domain Knowledge in Language Models for Task-Oriented Dialogue Systems | Pre-trained language models (PLM) have advanced the state-of-the-art across NLP applications, but lack domain-specific knowledge that does not naturally occur in pre-training data. Previous studies augmented PLMs with symbolic knowledge for different downstream NLP tasks. However, knowledge bases (KBs) utilized in these studies are usually large-scale and static, in contrast to small, domain-specific, and modifiable knowledge bases that are prominent in real-world task-oriented dialogue (TOD) systems. In this paper, we showcase the advantages of injecting domain-specific knowledge prior to fine-tuning on TOD tasks. To this end, we utilize light-weight adapters that can be easily integrated with PLMs and serve as a repository for facts learned from different KBs. To measure the efficacy of proposed knowledge injection methods, we introduce Knowledge Probing using Response Selection (KPRS) -- a probe designed specifically for TOD models. Experiments on KPRS and the response generation task show improvements of knowledge injection with adapters over strong baselines. | ['Saab Mansour', 'Yi Zhang', 'Sawsan Alqahtani', 'Daniele Bonadiman', 'Denis Emelin'] | 2022-12-15 | null | null | null | null | ['response-generation', 'task-oriented-dialogue-systems'] | ['natural-language-processing', 'natural-language-processing'] | [-1.37851700e-01 5.94401658e-01 -4.41783696e-01 -3.55116129e-01
-1.00644779e+00 -9.66796160e-01 7.41769075e-01 1.28649563e-01
-5.53813040e-01 1.20282376e+00 5.63838363e-01 -3.22593838e-01
5.09439744e-02 -7.57817984e-01 -7.88831651e-01 -1.19879082e-01
2.32905328e-01 9.85692620e-01 6.53842568e-01 -6.29904926e-01
2.26934105e-02 2.01578259e-01 -7.64887094e-01 8.25979531e-01
9.52060223e-01 5.83235681e-01 3.33445430e-01 4.16886628e-01
-4.72726792e-01 1.06384063e+00 -8.35617304e-01 -6.16202474e-01
-7.96893165e-02 -2.21759453e-01 -1.51205194e+00 -5.19988656e-01
2.26155102e-01 -4.22473967e-01 -2.21312955e-01 6.21201634e-01
5.68022430e-01 2.69091398e-01 6.58333719e-01 -9.10475850e-01
-9.97739971e-01 1.14997363e+00 3.53722386e-02 1.89960822e-01
4.05215412e-01 2.75901794e-01 1.02067673e+00 -8.69315147e-01
9.00938332e-01 1.43937683e+00 5.86387634e-01 9.08868253e-01
-1.23956168e+00 -3.38335335e-01 1.35342419e-01 3.37613523e-01
-9.64431524e-01 -5.98844469e-01 6.18420541e-01 -2.18974322e-01
1.59893167e+00 3.66155873e-03 2.12616742e-01 1.50161397e+00
-3.52905959e-01 1.11416733e+00 9.99011278e-01 -6.84258342e-01
-4.82080579e-02 8.46002281e-01 4.23196256e-01 5.03712058e-01
5.22891479e-03 -3.59806806e-01 -9.58842993e-01 -3.05040985e-01
6.86185241e-01 -6.76776707e-01 -5.08143008e-01 -2.41032809e-01
-1.18037319e+00 7.66913176e-01 6.60104826e-02 3.05580944e-01
-4.96090263e-01 -2.72682250e-01 5.35564601e-01 5.23251414e-01
5.07994950e-01 9.69325781e-01 -1.15322030e+00 -4.57445711e-01
-5.09011865e-01 2.38448352e-01 1.35458291e+00 1.13034189e+00
6.89429700e-01 -1.18345194e-01 -7.56975114e-01 1.23379087e+00
1.22014089e-02 4.57046658e-01 7.55271852e-01 -8.60178471e-01
8.66521180e-01 8.65325093e-01 3.50117147e-01 -3.70650411e-01
-3.11876833e-01 -1.80746004e-01 -3.81040841e-01 -5.79639554e-01
5.94856679e-01 -4.66006905e-01 -7.76134968e-01 1.88367283e+00
4.86981034e-01 -1.04818396e-01 7.74572432e-01 7.23410726e-01
1.06229818e+00 6.72442138e-01 2.00483486e-01 -1.30089521e-01
1.39009333e+00 -1.24466896e+00 -6.61795855e-01 -3.63536030e-01
1.03837991e+00 -4.67240661e-01 1.51932597e+00 9.70516875e-02
-1.09620249e+00 -3.09904218e-01 -5.53946972e-01 -2.76648313e-01
-6.11296058e-01 9.38999355e-02 5.52962780e-01 4.95405018e-01
-1.12692940e+00 2.16976702e-01 -4.06598717e-01 -5.59260488e-01
-1.44231822e-02 6.85312226e-02 -2.35570922e-01 -6.84941635e-02
-1.85624468e+00 1.41080427e+00 5.77576816e-01 -1.20443106e-01
-1.07664657e+00 -9.56507981e-01 -6.89971209e-01 2.50374638e-02
8.03525567e-01 -8.12100768e-01 1.94820440e+00 -7.21696973e-01
-2.06454563e+00 6.92326963e-01 -1.22806117e-01 -6.52905524e-01
3.53238553e-01 -5.12970030e-01 -2.40115210e-01 1.64556742e-01
2.16490752e-03 6.13980293e-01 6.17857754e-01 -1.05680144e+00
-4.73872960e-01 1.15430638e-01 7.32311547e-01 3.53044182e-01
-4.30433005e-01 1.11457311e-01 -3.37664068e-01 -3.50161701e-01
-5.15835106e-01 -7.65123785e-01 4.78066653e-02 -6.92632973e-01
-4.46399540e-01 -4.37054962e-01 6.40987694e-01 -7.70756602e-01
1.35132790e+00 -1.72940493e+00 1.68532640e-01 -1.40509591e-01
-2.71134764e-01 9.73001242e-01 -2.82710582e-01 8.51377368e-01
3.98962647e-01 1.30712271e-01 5.14433905e-02 -7.37494230e-02
1.70393869e-01 5.09025037e-01 -5.94143391e-01 -4.95344907e-01
4.67133820e-01 1.12507331e+00 -1.10647309e+00 -4.53298330e-01
-9.56424847e-02 8.36431012e-02 -3.82162213e-01 3.03185999e-01
-9.98360038e-01 4.73309159e-01 -6.93359494e-01 4.27172244e-01
1.47045717e-01 -3.23084235e-01 2.41524741e-01 -1.92890763e-01
2.09391877e-01 1.04532766e+00 -7.04822481e-01 1.74008965e+00
-7.73093283e-01 2.44435877e-01 -1.38998255e-01 -7.82795191e-01
7.61268198e-01 6.80199742e-01 -2.59806752e-01 -5.70872009e-01
-2.47050226e-01 1.47531047e-01 2.75683179e-02 -5.16844332e-01
7.48772323e-01 -6.40850887e-02 -2.84692124e-02 5.70759118e-01
5.67250967e-01 -1.23566978e-01 2.44124502e-01 4.66719389e-01
1.21382856e+00 7.44462833e-02 4.22413528e-01 4.28414196e-02
7.49149263e-01 3.97890031e-01 3.82261217e-01 1.01361060e+00
8.75307433e-03 5.33761308e-02 4.77961153e-01 1.69139449e-02
-6.79453313e-01 -7.47335732e-01 9.75345299e-02 1.42810524e+00
-2.83181787e-01 -4.14020509e-01 -7.65008390e-01 -1.10256612e+00
7.12822098e-03 1.17173493e+00 -3.04895908e-01 -2.69415349e-01
-6.43922865e-01 -5.03208399e-01 9.60268497e-01 5.84738314e-01
5.86281478e-01 -1.38665307e+00 -1.56195015e-01 4.96696919e-01
-4.28259611e-01 -1.26618612e+00 -2.07731932e-01 1.55667186e-01
-7.37716019e-01 -8.78521919e-01 -5.53850234e-01 -5.71158707e-01
3.08960110e-01 7.21712559e-02 1.48686433e+00 -3.95121455e-01
2.85724908e-01 7.84370840e-01 -6.80300653e-01 -4.69300002e-01
-8.22601318e-01 5.60906529e-01 -4.32342440e-02 -2.43946314e-01
6.18181765e-01 -2.98478723e-01 -1.93126842e-01 2.93950975e-01
-6.18607461e-01 7.44683817e-02 7.54258275e-01 9.98138309e-01
3.06476623e-01 -4.12089765e-01 1.06263828e+00 -1.19124079e+00
1.07503974e+00 -4.56996799e-01 -3.24416190e-01 7.04018950e-01
-3.93702000e-01 3.31478357e-01 6.79286420e-01 -6.21523380e-01
-1.74969554e+00 -4.46465820e-01 -6.85345288e-03 -5.04943170e-02
-2.35645369e-01 9.79364216e-01 -1.07946649e-01 9.04423445e-02
9.10801888e-01 2.47806415e-01 -3.84010673e-01 -7.96958685e-01
9.01633561e-01 8.19723725e-01 2.53009826e-01 -1.14231503e+00
4.18873370e-01 -9.95022506e-02 -6.23815715e-01 -7.20748127e-01
-1.29777431e+00 -5.77140391e-01 -5.18255413e-01 2.25565493e-01
4.28681016e-01 -9.48698997e-01 -3.66562724e-01 3.93638968e-01
-1.44776177e+00 -8.72732580e-01 -5.13828397e-01 3.59998077e-01
-4.17978376e-01 4.75253463e-02 -9.34483230e-01 -7.38683462e-01
-5.54661870e-01 -6.50219262e-01 8.11747372e-01 1.83138788e-01
-4.32979167e-01 -1.27001500e+00 2.83176661e-01 5.68320751e-01
5.72128296e-01 -6.05032384e-01 1.25087464e+00 -1.30372787e+00
-4.51047927e-01 1.24013320e-01 -1.69382364e-01 6.19489074e-01
8.25290829e-02 -4.60468143e-01 -1.06819010e+00 5.06435707e-02
-1.56424761e-01 -1.06984878e+00 6.92280352e-01 -1.23276472e-01
5.44315875e-01 -6.75977111e-01 -4.39230442e-01 -1.05784982e-01
8.50197196e-01 -1.04929790e-01 2.39714161e-01 4.52444643e-01
4.05417055e-01 8.50652456e-01 7.40158737e-01 1.73499331e-01
7.52883792e-01 8.17400515e-01 -2.83918470e-01 1.91479862e-01
-4.11864966e-01 -3.93146068e-01 5.56684494e-01 1.02461112e+00
1.50854543e-01 -2.85413206e-01 -1.09575713e+00 8.69794428e-01
-2.02333260e+00 -6.32992268e-01 1.42210275e-01 1.87727988e+00
1.95537055e+00 5.26999421e-02 -8.63532126e-02 -5.77549279e-01
4.67323869e-01 -8.53642821e-02 -4.94351178e-01 -5.17721653e-01
-1.18425772e-01 2.42067650e-01 2.20184520e-01 5.93193829e-01
-7.48046875e-01 1.43045878e+00 5.66126585e+00 9.26126301e-01
-9.71718550e-01 3.84411991e-01 5.95938601e-03 6.31221905e-02
-2.21351802e-01 1.08072266e-01 -1.40341818e+00 2.01245427e-01
1.21350348e+00 -3.59256715e-01 1.84202671e-01 1.00231695e+00
-3.04520130e-04 -5.42592779e-02 -1.29616833e+00 3.27385992e-01
-2.30776876e-01 -1.26810062e+00 3.48420233e-01 -4.76891607e-01
6.20447099e-01 2.05376133e-01 -1.55941531e-01 9.65373039e-01
8.89476180e-01 -6.02705359e-01 3.23004007e-01 4.08745646e-01
4.91229415e-01 -2.40435779e-01 9.07271206e-01 7.09715843e-01
-5.60583711e-01 1.03410333e-01 -3.79971832e-01 8.51499662e-02
3.36572498e-01 3.88863236e-01 -1.72546542e+00 7.29357660e-01
3.89184505e-01 2.71348596e-01 -3.81528944e-01 5.77074945e-01
-7.81784475e-01 9.94969845e-01 -4.22486573e-01 -2.19330266e-01
3.92038465e-01 3.18472236e-01 3.94517869e-01 1.66282868e+00
-1.57710612e-01 1.18489586e-01 1.25772476e-01 8.05017889e-01
-3.89085412e-01 2.33151317e-01 -4.63689327e-01 -2.70877033e-01
7.67648935e-01 1.07378137e+00 3.75675336e-02 -6.52170718e-01
-6.27089262e-01 9.20768797e-01 7.85868585e-01 6.14929199e-01
-4.36490387e-01 -2.03515038e-01 5.28418601e-01 -1.33153483e-01
1.81445032e-01 -8.53267163e-02 2.20984221e-01 -1.25397122e+00
7.38167465e-02 -9.87312794e-01 4.57740635e-01 -6.26463234e-01
-1.66921771e+00 4.21802282e-01 3.08448434e-01 -6.35831714e-01
-6.94171846e-01 -5.06140947e-01 -1.93720266e-01 1.12369847e+00
-1.83428979e+00 -1.39952719e+00 1.25381231e-01 7.01509118e-01
7.50727475e-01 -1.93297774e-01 1.19200838e+00 3.90364751e-02
-4.40954059e-01 7.09420621e-01 5.68634607e-02 4.09135520e-01
1.17363274e+00 -1.26320016e+00 3.88612568e-01 5.71499467e-01
9.17180702e-02 9.93990064e-01 4.55835819e-01 -8.21805298e-01
-1.31404889e+00 -1.11331964e+00 1.09828579e+00 -7.91785419e-01
9.57791805e-01 -2.65578985e-01 -1.41892827e+00 9.68745232e-01
3.07358623e-01 -2.19867110e-01 6.85687184e-01 5.39021671e-01
-5.45306027e-01 3.64386253e-02 -1.03246474e+00 6.02280974e-01
9.11290705e-01 -7.42278814e-01 -1.18992722e+00 4.46996778e-01
1.04881191e+00 -5.41148067e-01 -8.49906683e-01 3.34128946e-01
2.48462752e-01 -3.96709561e-01 9.90424633e-01 -1.08370292e+00
1.61555126e-01 -6.57746866e-02 4.32927907e-02 -1.64202297e+00
-6.61992431e-02 -6.00399077e-01 -5.13867080e-01 1.67752123e+00
8.51882458e-01 -6.66012406e-01 5.21350741e-01 9.20303762e-01
-3.08375627e-01 -4.05560702e-01 -8.00168216e-01 -9.74737942e-01
2.43414924e-01 -2.26486504e-01 3.52119088e-01 1.19957697e+00
5.69883764e-01 9.67320323e-01 -2.32045159e-01 9.19219702e-02
-2.82876194e-02 -1.13139816e-01 9.48904216e-01 -8.20686996e-01
-5.94956100e-01 -1.83637917e-01 2.58153081e-01 -1.26231337e+00
3.77613664e-01 -9.19686854e-01 7.17436969e-02 -1.60027683e+00
-8.02644435e-03 -6.58481538e-01 -1.60926953e-01 1.04326510e+00
-3.29580456e-01 -3.92947733e-01 2.96963807e-02 1.74762651e-01
-6.90461457e-01 6.61866724e-01 1.14712059e+00 -8.37265179e-02
-6.09725714e-01 -2.70818889e-01 -5.14859796e-01 6.34280205e-01
7.17113435e-01 -4.97352391e-01 -7.86234140e-01 -6.01858795e-01
1.76720440e-01 1.37128294e-01 1.17465600e-01 -5.12915611e-01
4.95979428e-01 -1.88556597e-01 -3.40277284e-01 -1.83978155e-01
5.04511297e-01 -4.39982712e-01 -2.72505522e-01 -7.13062361e-02
-5.33594191e-01 -5.08724570e-01 4.86343056e-01 3.69844377e-01
-4.16852534e-01 -5.09419680e-01 3.64968836e-01 -4.81330603e-01
-9.87337470e-01 -2.23691866e-01 -3.56892705e-01 6.28819346e-01
5.79084992e-01 2.16278911e-01 -9.19237792e-01 -3.33527744e-01
-4.92922336e-01 3.34773451e-01 1.06688730e-01 4.64373529e-01
3.55525047e-01 -8.43230903e-01 -6.81997061e-01 -1.99862465e-01
4.21584040e-01 1.69150203e-01 9.25114080e-02 8.01501393e-01
-4.84025851e-02 8.39014232e-01 -8.12919531e-03 -1.52671978e-01
-9.06124830e-01 4.50871438e-01 1.79581657e-01 -7.53786147e-01
-4.43344742e-01 1.05583572e+00 8.99594128e-02 -9.48987603e-01
3.23622882e-01 -3.12590688e-01 -2.50599444e-01 1.05206192e-01
5.01536846e-01 8.57898965e-02 2.17469171e-01 -2.56021731e-02
-2.78479785e-01 -5.96621707e-02 -5.35302579e-01 -2.71731347e-01
1.03891397e+00 -1.38039142e-01 -5.92992604e-02 6.04863465e-01
6.16045892e-01 -1.73511244e-02 -9.39805269e-01 -9.61010158e-01
3.22471440e-01 -5.50304130e-02 -2.55941570e-01 -1.59189951e+00
-3.90917689e-01 7.77609587e-01 -1.93652645e-01 4.14571241e-02
6.31072044e-01 1.95103258e-01 8.28201711e-01 1.13210905e+00
6.48764789e-01 -1.19357395e+00 1.60206959e-01 9.89807606e-01
1.02405739e+00 -1.13806164e+00 -4.76644009e-01 -4.00117129e-01
-1.04488468e+00 7.84497857e-01 9.54805017e-01 5.56143582e-01
2.54407346e-01 4.93774861e-02 3.47272128e-01 -2.83356421e-02
-1.23208034e+00 -3.20817679e-01 1.56867608e-01 7.92071760e-01
5.33755660e-01 -1.24936961e-01 -2.26745144e-01 9.41622794e-01
-3.52419764e-02 1.93977281e-01 4.44082081e-01 1.10311055e+00
-4.20266181e-01 -1.37859607e+00 -5.09513803e-02 2.84170985e-01
-3.15052450e-01 -5.80395877e-01 -7.58951426e-01 6.94104254e-01
-4.30493414e-01 1.04598534e+00 -5.20635247e-01 -6.44760430e-02
5.91176569e-01 7.20400929e-01 3.55937392e-01 -1.09415770e+00
-9.38858509e-01 -5.23916900e-01 9.43255067e-01 -3.98407608e-01
-4.16924506e-01 -4.23615336e-01 -1.06222558e+00 -6.23623431e-02
-4.61313993e-01 4.40637231e-01 3.52587223e-01 9.90519404e-01
6.05348706e-01 2.33538881e-01 -1.03539653e-01 -2.85402328e-01
-8.56832922e-01 -1.49723828e+00 -2.24234417e-01 2.95817316e-01
7.47134089e-02 -6.38901591e-01 2.26255711e-02 2.35797256e-01] | [11.823530197143555, 8.167834281921387] |
09c3657d-1f6f-4492-a58f-542612156710 | multi-modal-multi-class-parkinson-disease | 2307.02978 | null | https://arxiv.org/abs/2307.02978v1 | https://arxiv.org/pdf/2307.02978v1.pdf | Multi-modal multi-class Parkinson disease classification using CNN and decision level fusion | Parkinson disease is the second most common neurodegenerative disorder, as reported by the World Health Organization. In this paper, we propose a direct three-Class PD classification using two different modalities, namely, MRI and DTI. The three classes used for classification are PD, Scans Without Evidence of Dopamine Deficit and Healthy Control. We use white matter and gray matter from the MRI and fractional anisotropy and mean diffusivity from the DTI to achieve our goal. We train four separate CNNs on the above four types of data. At the decision level, the outputs of the four CNN models are fused with an optimal weighted average fusion technique. We achieve an accuracy of 95.53 percentage for the direct three class classification of PD, HC and SWEDD on the publicly available PPMI database. Extensive comparisons including a series of ablation studies clearly demonstrate the effectiveness of our proposed solution. | ['Ananda S. Chowdhury', 'Sushanta Kumar Sahu'] | 2023-07-06 | null | null | null | null | ['classification-1'] | ['methodology'] | [-2.09751842e-03 4.55404036e-02 -2.93166012e-01 -3.89826775e-01
-6.26474857e-01 -1.91375360e-01 5.54374337e-01 -2.41945893e-01
-6.53936327e-01 1.14018047e+00 2.91036397e-01 -7.27005079e-02
4.89373840e-02 -6.21058583e-01 -1.44297495e-01 -9.39105332e-01
-3.61908644e-01 6.58767700e-01 3.67758870e-01 2.16164693e-01
5.01374379e-02 5.60314298e-01 -1.06513822e+00 3.19885284e-01
1.28063166e+00 1.28412259e+00 4.42325920e-01 3.49084526e-01
1.18630841e-01 5.00529408e-01 -5.50813437e-01 -6.92860857e-02
3.01020265e-01 -1.33638814e-01 -6.21469140e-01 -2.35453527e-02
3.51520509e-01 -9.12985921e-01 -3.84254843e-01 1.14007854e+00
8.20137322e-01 -4.71989632e-01 7.08696246e-01 -1.02278435e+00
-7.34193861e-01 4.27985579e-01 -6.07188344e-01 7.97835231e-01
-2.41039291e-01 6.71389878e-01 6.78410649e-01 -3.86671692e-01
7.02588737e-01 1.30009735e+00 6.48437142e-01 6.01358235e-01
-1.12798882e+00 -6.36104763e-01 -5.09598069e-02 6.47438705e-01
-7.36444831e-01 -1.34624377e-01 4.67176348e-01 -7.32552767e-01
1.13299537e+00 -2.83448726e-01 9.87454474e-01 1.16095603e+00
5.42336583e-01 7.55513966e-01 1.50812256e+00 8.76627192e-02
3.58529598e-01 -5.75250626e-01 5.16907632e-01 5.74848413e-01
4.58161980e-01 2.12620109e-01 2.99634170e-02 -4.38668311e-01
7.23859727e-01 8.11036974e-02 -5.03143907e-01 -1.17804399e-02
-1.58064580e+00 8.68455648e-01 5.23106158e-01 4.95793700e-01
-7.31053352e-01 4.55358066e-02 4.94945288e-01 1.24431364e-01
2.75708020e-01 -9.80179161e-02 -3.09712768e-01 2.20502004e-01
-1.00412786e+00 4.49416935e-01 2.16017619e-01 4.60830510e-01
-5.34445271e-02 1.40487719e-02 -3.92393887e-01 1.03889263e+00
6.95161343e-01 9.09285724e-01 1.04943252e+00 -1.22584343e+00
6.28023386e-01 5.93894541e-01 3.45216468e-02 -4.87499535e-01
-8.74684691e-01 -3.42586458e-01 -9.74963605e-01 7.16487885e-01
5.86852670e-01 -1.90842584e-01 -1.31433618e+00 1.60862243e+00
-8.49224254e-02 -2.99045593e-01 -4.13592570e-02 1.03141987e+00
6.00756764e-01 -8.93700570e-02 2.76369005e-01 -1.13821827e-01
1.41927814e+00 -8.48283172e-01 -8.53501499e-01 -2.58762568e-01
5.66153824e-01 6.44346774e-02 5.66662550e-01 4.29400861e-01
-1.02124166e+00 -6.11040369e-02 -1.32659888e+00 -8.72698873e-02
-3.10834706e-01 4.73802865e-01 3.07297021e-01 3.20324123e-01
-1.13483000e+00 7.28530109e-01 -1.20464683e+00 -4.07640219e-01
1.03694093e+00 4.65648592e-01 -4.84126627e-01 -1.58697486e-01
-1.17544127e+00 1.46112645e+00 3.00233513e-01 1.20903794e-02
-1.20586669e+00 -6.56799495e-01 -2.34799027e-01 -3.84485573e-01
-5.70424259e-01 -8.25585961e-01 9.47193384e-01 -3.78693014e-01
-9.93947744e-01 1.07121527e+00 2.60354183e-03 -8.62981975e-01
7.64416456e-01 8.84696096e-02 -4.19618160e-01 3.27928275e-01
3.97324741e-01 9.37968552e-01 6.77200079e-01 -8.60808969e-01
-8.73921812e-01 -1.15684569e+00 -1.58168942e-01 6.29666299e-02
1.00181580e-01 -1.99559350e-02 2.27224886e-01 -5.10881007e-01
3.13718796e-01 -1.06116033e+00 -1.48407757e-01 4.95893270e-01
-6.53668106e-01 -2.13047981e-01 1.04598475e+00 -1.23442721e+00
6.65081680e-01 -1.89082062e+00 2.08871946e-01 -6.12331331e-02
1.02843952e+00 2.72948980e-01 1.28530487e-01 -5.65112233e-01
-2.93966442e-01 4.73958775e-02 -6.17483437e-01 -6.00634776e-02
4.44002748e-02 2.77205855e-01 2.42788881e-01 8.33993554e-01
2.29060844e-01 9.29255664e-01 -6.81453109e-01 -2.24713802e-01
7.40957335e-02 5.30774117e-01 -1.83576226e-01 -7.60098025e-02
1.05772667e-01 8.02864194e-01 -4.18770403e-01 8.25294614e-01
6.30507112e-01 -2.30644539e-01 2.67568558e-01 -5.37262142e-01
-1.71766449e-02 1.72924727e-01 -4.16488379e-01 1.48281896e+00
2.80664772e-01 3.67309183e-01 1.28542140e-01 -1.10005951e+00
5.62700987e-01 3.26824307e-01 7.16806054e-01 -9.20417547e-01
3.26497436e-01 6.66863084e-01 7.71355271e-01 -4.66298491e-01
-4.24679428e-01 -9.74743217e-02 3.25709909e-01 4.43752915e-01
2.26460278e-01 4.81688619e-01 4.61843729e-01 -3.28562647e-01
1.41616476e+00 -2.16525927e-01 1.80656478e-01 -4.06250536e-01
3.88955534e-01 -7.29323481e-04 8.29067409e-01 4.23086405e-01
-1.09153807e+00 6.27426565e-01 6.46393001e-01 -4.48201954e-01
-1.05945766e+00 -1.30245578e+00 -2.43218124e-01 2.00366572e-01
-3.46840888e-01 3.45990807e-01 -7.43613005e-01 -8.67338598e-01
2.09518015e-01 2.89889902e-01 -7.58123755e-01 -1.84871912e-01
-7.89692461e-01 -1.39658248e+00 7.10682690e-01 7.02559888e-01
1.00474870e+00 -9.75908041e-01 -7.74878085e-01 2.57044017e-01
-1.58619642e-01 -9.39532042e-01 -1.14521645e-01 3.44876975e-01
-1.40733361e+00 -1.46126282e+00 -1.54542744e+00 -6.86473489e-01
5.49944818e-01 -9.66888815e-02 7.59312093e-01 -3.07609200e-01
-2.28117228e-01 5.92238829e-03 1.71888601e-02 -8.58811513e-02
-2.98445582e-01 -8.84915590e-02 2.61178166e-01 -3.37118328e-01
3.94779891e-01 -9.83689129e-01 -9.60929394e-01 -4.09681350e-03
-5.67456126e-01 -1.64078504e-01 8.51125896e-01 5.26097715e-01
8.00867498e-01 -1.64775893e-01 6.98394656e-01 -1.69438303e-01
9.60696876e-01 -5.30029833e-01 -1.17252193e-01 1.58721432e-01
-6.12263560e-01 -1.46410331e-01 1.76962614e-01 -4.33572799e-01
-6.13014340e-01 1.04300104e-01 -1.13598444e-01 -5.26289046e-01
-3.68715942e-01 1.66471720e-01 -3.23890746e-01 -2.87619084e-02
4.91716057e-01 1.86265558e-01 3.31687212e-01 -4.71704394e-01
2.65742123e-01 6.25569761e-01 5.57637453e-01 -1.10059261e-01
1.24528281e-01 7.96364486e-01 -2.64431119e-01 -4.17861849e-01
-3.98771197e-01 7.12073455e-03 -8.83862257e-01 -1.29066601e-01
1.23085642e+00 -7.81723797e-01 -3.94407272e-01 1.06318510e+00
-1.23530626e+00 -3.60722631e-01 -1.07245229e-01 7.19297767e-01
-6.95052207e-01 5.12458622e-01 -5.34616709e-01 -2.83577710e-01
-9.69051719e-01 -1.61404026e+00 7.78257668e-01 -4.81786132e-02
-2.34280407e-01 -9.04533803e-01 1.70016915e-01 2.34535709e-01
4.73549604e-01 3.81507397e-01 1.39592111e+00 -7.44619906e-01
3.36744599e-02 1.18315853e-01 -5.68449080e-01 4.14037973e-01
3.72277617e-01 -4.67416793e-01 -7.50085294e-01 -2.05331042e-01
2.68970102e-01 -1.91621616e-01 1.07282495e+00 7.84042358e-01
6.92951024e-01 2.76573151e-02 -4.57019776e-01 2.34518632e-01
1.24786782e+00 8.32411408e-01 8.20757031e-01 5.46890557e-01
6.72571719e-01 2.31732160e-01 3.17061990e-02 3.02238941e-01
4.25873905e-01 5.01158893e-01 5.55760860e-01 1.65690482e-01
-5.35469353e-01 6.37836874e-01 2.88817883e-01 6.66406929e-01
-3.16088557e-01 1.95919871e-01 -1.27767754e+00 7.36951351e-01
-1.76398027e+00 -9.71708417e-01 -2.73868084e-01 1.93088436e+00
7.11906850e-01 6.45304471e-02 2.70260066e-01 6.39970079e-02
9.29606795e-01 -5.90303726e-02 -1.19152069e+00 2.55041361e-01
-3.34446818e-01 1.62206590e-02 6.88124359e-01 7.35553429e-02
-1.16162980e+00 2.74443299e-01 7.32537651e+00 2.90794134e-01
-1.23578954e+00 5.72339833e-01 5.38052261e-01 -4.16516304e-01
3.24429780e-01 -6.64663374e-01 -6.06909215e-01 7.72584081e-01
9.16301847e-01 1.70190800e-02 5.55769980e-01 3.93012524e-01
5.50170004e-01 -1.91030204e-01 -9.19429004e-01 9.55539703e-01
-2.08843037e-01 -9.91496682e-01 -1.25084966e-01 3.47728848e-01
4.14804459e-01 9.37166452e-01 5.46030588e-02 9.35723186e-02
3.88072550e-01 -9.75923777e-01 5.43030500e-01 8.60611737e-01
6.28890276e-01 -2.84401953e-01 6.73919320e-01 -1.06531475e-02
-6.59937859e-01 -2.72611141e-01 -3.80958291e-03 5.10174751e-01
2.22992912e-01 9.14357007e-01 -2.79286563e-01 5.58620505e-02
9.19660509e-01 9.68878508e-01 -4.58706021e-01 1.17625403e+00
-1.48560762e-01 4.40349340e-01 -3.76154453e-01 5.86239696e-01
1.81293055e-01 -1.27796784e-01 5.76155365e-01 7.56159961e-01
5.36654711e-01 -1.62785396e-01 -9.63211730e-02 1.33054507e+00
4.22312506e-03 -4.83776033e-01 -4.04129654e-01 -1.32309690e-01
2.43952703e-02 1.06529284e+00 -5.11387706e-01 -4.26767528e-01
-3.41090173e-01 7.76550293e-01 2.87628651e-01 7.11283162e-02
-8.25490713e-01 -1.26846418e-01 6.89661324e-01 -2.58552343e-01
2.45589346e-01 -1.77815184e-01 -4.30044711e-01 -1.13033104e+00
9.24346149e-02 -8.49323869e-01 1.82245880e-01 -7.16428876e-01
-1.60062647e+00 5.54732680e-01 -1.71840906e-01 -1.04492342e+00
2.57028602e-02 -9.10066664e-01 -7.76129425e-01 1.08743596e+00
-1.70928943e+00 -7.79673934e-01 -1.30597919e-01 7.36714065e-01
1.69328272e-01 -3.58897418e-01 6.70190513e-01 7.00560272e-01
-6.56369150e-01 -1.43195212e-01 3.08524668e-01 2.29193762e-01
5.36875486e-01 -1.32055914e+00 2.19813377e-01 3.00690055e-01
-9.33331192e-01 4.59739476e-01 2.93729812e-01 -1.12094629e+00
-7.98612237e-01 -1.29362679e+00 6.28537595e-01 1.81456134e-02
8.31569791e-01 2.64177352e-01 -7.79792190e-01 5.58578432e-01
2.05286324e-01 5.64033762e-02 4.16579127e-01 -5.08094668e-01
-1.41575754e-01 8.44745561e-02 -1.78237653e+00 2.38314465e-01
8.83582115e-01 -2.59286255e-01 -9.73739803e-01 6.25427186e-01
2.66917199e-01 -9.66716930e-02 -1.48194027e+00 4.83147293e-01
6.40179038e-01 -8.35806787e-01 9.25807536e-01 -5.00718117e-01
4.24391806e-01 -1.33340642e-01 -2.61153877e-01 -1.57387745e+00
-4.21631217e-01 4.07938421e-01 -4.55570459e-01 7.24062324e-01
2.69605041e-01 -8.86756480e-01 5.63365042e-01 5.15535057e-01
-3.54455203e-01 -7.94801593e-01 -1.05274439e+00 -9.05816138e-01
6.57319248e-01 -1.12026393e-01 3.65458757e-01 8.57077539e-01
-1.79649785e-01 -3.50075103e-02 5.82699431e-03 2.52755675e-02
9.20480013e-01 -2.10131675e-01 -3.54134381e-01 -1.60990226e+00
1.50853798e-01 -6.48314178e-01 -5.12307584e-01 -2.29185134e-01
1.04211971e-01 -1.21704650e+00 -2.56202519e-01 -2.04557395e+00
3.61872464e-01 -2.84279972e-01 -5.37707806e-01 8.66606414e-01
3.74286354e-01 3.15984070e-01 5.69649078e-02 6.85531378e-01
-1.58471316e-01 5.35917521e-01 1.54068220e+00 -7.81837404e-01
5.30036651e-02 -4.24956203e-01 -5.29504597e-01 8.96721363e-01
1.41609585e+00 -4.10470426e-01 -3.30396920e-01 -5.50202668e-01
-5.39003789e-01 -1.93080127e-01 6.40190244e-01 -1.44145989e+00
-1.93444729e-01 1.43291384e-01 7.35586166e-01 -6.70035660e-01
4.71732080e-01 -6.68944836e-01 -1.01930521e-01 8.81617188e-01
-1.15818053e-01 -1.03587322e-01 -4.99597564e-02 3.71145606e-01
6.91969246e-02 3.30014825e-02 1.27163327e+00 -4.72272575e-01
-3.73133481e-01 6.37219012e-01 -5.89812577e-01 -5.63318096e-02
8.65528166e-01 -5.08266613e-02 -9.32941318e-01 1.79218307e-01
-1.47275651e+00 1.65651724e-01 1.01672020e-03 2.38665149e-01
5.19644558e-01 -1.65092707e+00 -6.79286003e-01 -2.08352327e-01
-1.73263103e-01 -5.57620525e-01 1.73653483e-01 1.65255451e+00
-2.95186728e-01 7.66085327e-01 -1.11159289e+00 -6.38940811e-01
-8.30038965e-01 1.33102432e-01 9.05423760e-01 -5.31098127e-01
-9.72951889e-01 8.53250548e-02 -4.24279384e-02 -4.20416951e-01
1.39835492e-01 -6.50166750e-01 -4.86155301e-01 2.14504555e-01
7.55637467e-01 5.56910932e-01 3.46796393e-01 -7.81464875e-01
-4.11556572e-01 1.77192330e-01 -2.57371217e-01 -1.03570543e-01
1.54250002e+00 -9.31142271e-02 -3.44296247e-01 2.26903901e-01
1.15391457e+00 -8.06145489e-01 -1.18043339e+00 1.62178099e-01
3.38845402e-02 9.11099091e-02 3.98738235e-01 -1.30619442e+00
-1.67620814e+00 1.02872288e+00 1.67450678e+00 1.17325559e-01
7.72060215e-01 -9.11590680e-02 1.15575159e+00 2.48521924e-01
4.25605804e-01 -9.80996251e-01 -6.37621105e-01 4.54031616e-01
1.09322417e+00 -1.03295517e+00 -1.88767761e-01 1.55621022e-01
-4.53855395e-01 9.69829381e-01 5.45889795e-01 -4.86308545e-01
5.89816511e-01 1.36550754e-01 1.20041333e-01 -3.55676562e-01
-4.28742528e-01 -2.12090701e-01 -1.59550160e-02 1.16059244e+00
2.65189022e-01 2.61949096e-02 -8.30892861e-01 8.63674104e-01
1.61097556e-01 3.01329195e-01 2.55501628e-01 9.53976035e-01
-6.38930857e-01 -8.93129587e-01 -3.35158855e-01 1.20591497e+00
-5.20132780e-01 1.24686666e-01 -4.90774482e-01 6.34354889e-01
4.03793424e-01 5.43089330e-01 -2.34223399e-02 -1.94902182e-01
2.40724400e-01 2.25388318e-01 5.58115602e-01 -1.38754889e-01
-3.98645669e-01 -2.84937054e-01 1.49841815e-01 -6.36526942e-01
-6.48410082e-01 -8.64718735e-01 -1.62526667e+00 -3.14748026e-02
2.46772483e-01 -5.19943833e-01 5.21761835e-01 1.16813028e+00
4.39568490e-01 6.22723877e-01 7.91593045e-02 -1.09318697e+00
-5.12950063e-01 -1.24841046e+00 -8.44786704e-01 1.65984958e-01
3.61398727e-01 -1.25026846e+00 -2.31887266e-01 -2.88727488e-02] | [14.140833854675293, -1.7807226181030273] |
9c1ef9f3-75c9-4ed2-ae47-d1132dc48baf | self-supervised-learning-of-a-facial | 1808.06882 | null | http://arxiv.org/abs/1808.06882v1 | http://arxiv.org/pdf/1808.06882v1.pdf | Self-supervised learning of a facial attribute embedding from video | We propose a self-supervised framework for learning facial attributes by
simply watching videos of a human face speaking, laughing, and moving over
time. To perform this task, we introduce a network, Facial Attributes-Net
(FAb-Net), that is trained to embed multiple frames from the same video
face-track into a common low-dimensional space. With this approach, we make
three contributions: first, we show that the network can leverage information
from multiple source frames by predicting confidence/attention masks for each
frame; second, we demonstrate that using a curriculum learning regime improves
the learned embedding; finally, we demonstrate that the network learns a
meaningful face embedding that encodes information about head pose, facial
landmarks and facial expression, i.e. facial attributes, without having been
supervised with any labelled data. We are comparable or superior to
state-of-the-art self-supervised methods on these tasks and approach the
performance of supervised methods. | ['Olivia Wiles', 'A. Sophia Koepke', 'Andrew Zisserman'] | 2018-08-21 | null | null | null | null | ['unsupervised-facial-landmark-detection'] | ['computer-vision'] | [ 2.52894044e-01 6.25609398e-01 -2.78016239e-01 -1.03472769e+00
-7.63981700e-01 -2.85694569e-01 6.97253644e-01 -3.37654680e-01
-2.49629602e-01 4.74032342e-01 5.34886777e-01 5.51069319e-01
1.74861729e-01 -2.99662411e-01 -9.39431071e-01 -6.45933807e-01
-4.06445742e-01 3.87686431e-01 -3.79355550e-01 3.10751855e-01
-1.08111113e-01 4.50523347e-01 -1.91186225e+00 1.60848320e-01
1.80833973e-02 1.16702449e+00 -3.95541936e-01 6.58783317e-01
2.91308165e-01 1.15098834e+00 -1.89629197e-01 -5.84194243e-01
9.69304964e-02 -4.62316424e-01 -7.86885500e-01 4.20084924e-01
1.08793771e+00 -8.19798112e-01 -4.20324445e-01 8.88926268e-01
2.41778791e-01 1.85369208e-01 8.64888906e-01 -1.66702747e+00
-6.35979772e-01 1.44572064e-01 -4.55278456e-01 -8.06745142e-02
5.05192518e-01 2.54659750e-03 1.10600626e+00 -1.09861350e+00
8.46224308e-01 1.55576825e+00 6.92540824e-01 1.16787338e+00
-1.40950179e+00 -8.02418172e-01 2.39129528e-01 1.44996375e-01
-1.24511194e+00 -1.17393506e+00 9.10775721e-01 -4.14821357e-01
5.64711988e-01 -9.22000483e-02 6.18741572e-01 1.46317220e+00
-3.05710405e-01 9.84206915e-01 8.58374655e-01 -3.08598548e-01
5.14563248e-02 -1.26349637e-02 -2.22659290e-01 1.01844823e+00
-5.11319041e-01 2.91125119e-01 -7.75688946e-01 -1.19120017e-01
5.34089565e-01 -6.56637102e-02 -1.05215929e-01 -6.20124400e-01
-9.26175117e-01 9.38838065e-01 2.44218215e-01 -3.84873226e-02
-2.49883503e-01 5.10724962e-01 3.50866258e-01 2.57751256e-01
8.49187076e-01 -5.31594194e-02 -4.23699021e-01 -1.21253595e-01
-1.04117799e+00 1.23223521e-01 8.27254474e-01 1.07183480e+00
1.07348204e+00 1.64328024e-01 2.74424125e-02 5.70956826e-01
3.53762895e-01 5.19430280e-01 2.16253594e-01 -1.53455067e+00
-1.94170609e-01 1.01109788e-01 -2.60995567e-01 -8.27898264e-01
-3.99347186e-01 2.47061521e-01 -5.08401096e-01 4.64855403e-01
2.63362139e-01 -2.95724273e-01 -7.97008991e-01 2.30422235e+00
4.45858866e-01 7.18255460e-01 6.33951128e-02 6.50643826e-01
7.69501626e-01 4.65417594e-01 2.90688187e-01 -3.49596739e-01
1.17055845e+00 -1.08251476e+00 -8.82908821e-01 -1.11763179e-01
3.19688380e-01 -5.17229497e-01 7.54265904e-01 1.19179115e-01
-1.22923267e+00 -7.16149390e-01 -7.87435889e-01 -6.16219118e-02
-1.69956073e-01 2.35833049e-01 6.47373796e-01 5.17531157e-01
-1.36231303e+00 8.10365021e-01 -9.92922008e-01 -5.67070067e-01
8.04084182e-01 4.42903459e-01 -8.67164791e-01 -7.65989593e-04
-6.63619816e-01 6.95788324e-01 -3.97861972e-02 -1.90542087e-01
-1.36206913e+00 -7.92212844e-01 -1.27374744e+00 2.44415570e-02
2.59323061e-01 -5.76239705e-01 1.40662086e+00 -1.68621409e+00
-1.75860786e+00 1.20273840e+00 -4.67928410e-01 -2.80458748e-01
2.07085475e-01 -4.30042833e-01 -3.32772911e-01 5.89791417e-01
-8.80636834e-03 1.26182044e+00 1.48038697e+00 -1.13236725e+00
-4.26180929e-01 -5.53047478e-01 1.40426271e-02 5.28711490e-02
-5.56953788e-01 2.76158810e-01 -3.63825887e-01 -6.31100476e-01
-4.53455865e-01 -1.05709362e+00 2.25296184e-01 6.09764159e-01
-1.69198081e-01 -3.12992871e-01 1.13677847e+00 -4.59268302e-01
6.72150910e-01 -2.57798481e+00 3.10098261e-01 -5.01925545e-03
3.31110984e-01 1.05618395e-01 -3.64930004e-01 -2.06332300e-02
-4.04589921e-01 -4.11090627e-02 -5.86080067e-02 -1.04868150e+00
-1.33187743e-02 4.89133179e-01 -1.19773403e-01 7.53260434e-01
7.16608167e-01 8.30293655e-01 -9.31469321e-01 -6.47175550e-01
6.29674792e-02 8.70548844e-01 -7.97248662e-01 5.43782473e-01
-3.08429688e-01 7.07982540e-01 -2.11098000e-01 6.55151784e-01
4.85495657e-01 -1.75551951e-01 6.44059628e-02 -3.58663350e-01
2.67122269e-01 -2.76326407e-02 -8.86484742e-01 1.83592534e+00
-4.01908368e-01 8.09661686e-01 2.60870606e-01 -9.62756336e-01
5.92188835e-01 7.27639258e-01 5.85318863e-01 -2.41314828e-01
2.21879736e-01 -1.19654268e-01 -5.96494675e-01 -6.50579095e-01
-1.39944911e-01 -1.43548712e-01 3.31428677e-01 6.73479676e-01
9.58073497e-01 1.29600480e-01 -6.88958243e-02 1.67455330e-01
9.71337736e-01 5.10016501e-01 9.32475552e-02 -1.31902829e-01
5.10880053e-01 -6.60255849e-01 4.93007928e-01 1.14647612e-01
-4.05150890e-01 6.07369542e-01 6.09026551e-01 -5.71327031e-01
-1.03884757e+00 -9.08733010e-01 -3.35811786e-02 1.63118279e+00
-4.96700495e-01 -4.99585092e-01 -1.11779165e+00 -7.76829720e-01
1.38448879e-01 2.32025206e-01 -1.10567236e+00 -1.30475685e-01
-4.58385795e-01 -8.93623978e-02 4.05888170e-01 8.42084527e-01
1.45912588e-01 -1.15943456e+00 -3.54960769e-01 -9.27108228e-02
-1.17979078e-02 -1.40693367e+00 -6.35234654e-01 1.73852324e-01
-6.26842320e-01 -1.02939415e+00 -5.98297656e-01 -7.52655029e-01
9.29191530e-01 -9.31994021e-02 1.28186619e+00 2.84937918e-01
-2.44779617e-01 1.07847381e+00 -3.12372684e-01 -3.96283925e-01
-2.31008977e-01 -7.69441351e-02 3.14503729e-01 5.52129030e-01
5.57575345e-01 -8.23615849e-01 -4.44868207e-01 8.99056271e-02
-6.44004166e-01 -1.10605851e-01 3.90912294e-01 7.79878318e-01
3.87624532e-01 -5.85836709e-01 4.06290650e-01 -9.47785676e-01
4.89406548e-02 -4.21618730e-01 -3.79106045e-01 -5.26050180e-02
-2.75442541e-01 1.46288544e-01 2.81116933e-01 -4.42540646e-01
-9.12913859e-01 6.22989655e-01 -7.98942223e-02 -8.20217609e-01
-5.50128222e-01 -5.53692318e-02 -1.71302110e-01 -2.94739217e-01
3.60059261e-01 -1.08502030e-01 6.35853291e-01 -2.15710104e-01
6.84486210e-01 3.82827759e-01 8.78062844e-01 -6.14778101e-01
9.30591464e-01 8.65314424e-01 4.92280386e-02 -7.90687919e-01
-1.21028924e+00 -4.66866672e-01 -1.05399883e+00 -3.05397719e-01
1.14154792e+00 -1.13246346e+00 -1.06211925e+00 3.39975446e-01
-1.12935495e+00 -4.79753733e-01 -3.08411688e-01 3.31045270e-01
-1.13276982e+00 6.14467338e-02 -6.86205626e-01 -8.54631662e-01
-2.59067807e-02 -7.95967281e-01 1.40571630e+00 2.50618130e-01
-3.90935779e-01 -1.08982837e+00 -1.76909883e-02 1.12310596e-01
2.80607730e-01 4.77973193e-01 5.00047028e-01 -6.60904706e-01
-3.79115522e-01 5.78346960e-02 -1.14116140e-01 5.13510108e-01
2.27375433e-01 1.38133287e-01 -1.55036008e+00 -4.32970524e-01
-1.03806444e-01 -1.02075779e+00 6.61323249e-01 2.94285268e-01
1.32372546e+00 -6.05525851e-01 -2.09153235e-01 1.04172492e+00
9.20217574e-01 -2.71804661e-01 4.60286438e-01 1.65466759e-02
6.18888497e-01 9.69164610e-01 3.21941197e-01 4.14783984e-01
4.59888637e-01 6.70304537e-01 4.72143710e-01 -3.00771505e-01
-2.84975648e-01 -3.08910459e-01 5.63279808e-01 3.59939605e-01
-1.99200273e-01 3.06469232e-01 -3.23002934e-01 5.20424545e-01
-1.78471208e+00 -1.08339643e+00 6.26960814e-01 1.80479562e+00
8.78097117e-01 -3.76987159e-01 3.71627092e-01 -9.70035493e-02
4.66233045e-01 2.97778100e-01 -6.89618468e-01 -1.60490945e-01
-8.70370399e-03 4.40877467e-01 7.87517801e-02 5.73587000e-01
-1.42604494e+00 1.15843725e+00 7.17908049e+00 1.96974427e-01
-1.04081357e+00 1.74935207e-01 7.43701816e-01 -3.09760362e-01
2.01292317e-02 -2.11307302e-01 -7.59034395e-01 1.41201794e-01
1.16553402e+00 4.91308533e-02 5.74918151e-01 1.10556734e+00
-1.49752364e-01 3.16049844e-01 -1.76060963e+00 9.99391377e-01
7.04335332e-01 -1.09640622e+00 3.24589922e-03 2.03232672e-02
5.94932497e-01 -2.31043115e-01 2.80951202e-01 1.77408010e-01
4.52783048e-01 -1.25925481e+00 6.58327758e-01 4.76056725e-01
1.05618596e+00 -7.12517083e-01 3.77597362e-01 -1.71842039e-01
-1.09672296e+00 -3.03672515e-02 -1.90919682e-01 1.05787151e-01
9.44669247e-02 1.29287034e-01 -5.38074255e-01 7.25669265e-02
7.51071632e-01 1.07467782e+00 -4.45713222e-01 4.74251062e-01
-2.95070559e-01 5.73773921e-01 -1.35224476e-01 5.63526571e-01
1.65064782e-01 4.77734692e-02 2.71944106e-01 1.11297083e+00
1.58511862e-01 2.30173841e-01 1.70072094e-01 6.56745136e-01
-3.12460363e-01 -6.76522106e-02 -7.31081724e-01 -1.48269115e-02
1.59600869e-01 1.46395218e+00 -1.30452782e-01 -2.58568287e-01
-7.49516845e-01 1.18872106e+00 4.46067065e-01 4.41686332e-01
-6.05440378e-01 5.82235046e-02 1.10939765e+00 7.50230700e-02
3.95656735e-01 3.68709676e-02 4.43019420e-01 -1.01370633e+00
-1.62140906e-01 -8.83617878e-01 3.94284159e-01 -9.40836251e-01
-1.38370514e+00 7.22935796e-01 -8.31368268e-02 -9.96784449e-01
-6.46118701e-01 -7.73713648e-01 -4.98940021e-01 4.97647673e-01
-1.79280674e+00 -1.46402884e+00 -4.53409493e-01 1.03993213e+00
3.98051202e-01 -4.75885510e-01 1.16363871e+00 1.75660878e-01
-3.64605159e-01 8.57977271e-01 -2.51586050e-01 4.90198582e-01
1.08794212e+00 -1.20773816e+00 2.98642665e-01 3.46046060e-01
4.45009559e-01 4.83278275e-01 4.84063745e-01 -1.54861107e-01
-1.38276041e+00 -1.13223422e+00 7.12320030e-01 -6.12579644e-01
6.21547401e-01 -6.11167669e-01 -8.41905951e-01 1.20039916e+00
2.44458839e-01 8.63619030e-01 1.00274265e+00 2.15152696e-01
-6.77452385e-01 -1.88915491e-01 -1.19497550e+00 3.50444168e-01
1.19207525e+00 -9.09195840e-01 -3.45005423e-01 3.56548935e-01
6.57362163e-01 -3.51446271e-01 -7.45712519e-01 1.38313591e-01
7.05735743e-01 -9.01858330e-01 9.55869973e-01 -9.99142945e-01
4.60269719e-01 3.48309055e-02 -2.31113613e-01 -1.27916288e+00
-2.56159723e-01 -9.54965174e-01 -3.94108027e-01 1.36570990e+00
1.59224346e-01 -2.79864937e-01 9.33429360e-01 6.09789133e-01
1.39222100e-01 -5.84429026e-01 -9.64334965e-01 -5.05274653e-01
-2.61970647e-02 -1.34682566e-01 6.11574352e-01 9.18815076e-01
-1.25176564e-01 3.46190780e-01 -6.53415680e-01 1.34678915e-01
8.89177561e-01 -3.00628930e-01 9.40501451e-01 -1.43177342e+00
-3.05263512e-02 -1.71097755e-01 -7.57984877e-01 -8.07841063e-01
1.10440350e+00 -6.72017217e-01 -6.73006848e-02 -6.41892135e-01
3.80892783e-01 -3.74082066e-02 -1.09439731e-01 7.78359950e-01
1.69048812e-02 5.97697377e-01 1.95412636e-01 1.67338282e-01
-7.31906891e-01 7.24419892e-01 9.08174634e-01 -1.09036705e-02
3.77147734e-01 -2.18131542e-01 -6.10773563e-01 9.63282228e-01
3.51928234e-01 -4.79668587e-01 -3.72390777e-01 -4.30579334e-01
-2.12998241e-01 5.04774041e-02 5.14391899e-01 -9.39713538e-01
2.12945268e-01 8.93929377e-02 7.01611578e-01 -4.33058329e-02
5.30230403e-01 -9.61474836e-01 -2.35032991e-01 -5.85369654e-02
-5.52087128e-01 -7.54780322e-02 1.74370378e-01 6.62477970e-01
-2.43122861e-01 -4.80670445e-02 9.73479450e-01 1.10543862e-01
-7.25887775e-01 6.89750493e-01 -1.38150185e-01 3.28068882e-02
1.09751177e+00 1.40646740e-03 -5.74694993e-03 -9.07404900e-01
-1.12085569e+00 1.00071236e-01 5.52471161e-01 4.63836402e-01
5.10376871e-01 -1.72468972e+00 -6.74811780e-01 5.07674754e-01
2.64811248e-01 -3.22249770e-01 5.15803602e-03 6.00936949e-01
-5.23387641e-02 1.10550649e-01 -4.47966516e-01 -7.91657448e-01
-1.47309852e+00 5.92629373e-01 1.53594941e-01 3.70558411e-01
-5.84708691e-01 8.47722411e-01 2.86330938e-01 -4.14595187e-01
5.45167446e-01 2.42055997e-01 -2.66648352e-01 1.90780818e-01
8.69087577e-01 -1.16813727e-01 -1.72139034e-01 -1.14396691e+00
-2.94156492e-01 8.60892594e-01 -1.11387566e-01 -2.24945068e-01
1.53354132e+00 -6.08202852e-02 -2.32130736e-02 3.19018930e-01
1.74932063e+00 -1.41961396e-01 -1.95341480e+00 -4.03194577e-01
-1.86861962e-01 -4.19959277e-01 8.96494836e-02 -3.21922183e-01
-1.42932606e+00 8.35858881e-01 6.65455759e-01 -3.51330131e-01
1.14264536e+00 2.69373655e-01 5.02496779e-01 3.42188686e-01
5.79034649e-02 -1.00581813e+00 6.31325781e-01 3.91458720e-01
7.19494820e-01 -1.49571061e+00 -1.33582056e-01 -3.00581753e-01
-7.05272973e-01 1.28371060e+00 4.09875005e-01 -1.37615725e-01
9.40946102e-01 3.80687147e-01 2.83425301e-01 -3.26727509e-01
-8.09321523e-01 -2.94296890e-01 2.98757792e-01 9.66112792e-01
5.50018251e-01 -2.61300325e-01 5.65516651e-01 3.00503433e-01
-1.90225691e-01 1.28959060e-01 2.11151287e-01 8.55108917e-01
-1.13035955e-01 -9.71411407e-01 -9.69621167e-02 1.58486009e-01
-5.82974911e-01 4.20629308e-02 -3.36582363e-01 6.64723277e-01
-1.44726448e-02 6.49687231e-01 5.19076109e-01 -1.63926795e-01
1.65144131e-02 4.36471999e-01 6.89197004e-01 -7.04998553e-01
-8.61210972e-02 5.85299246e-02 -6.97861612e-02 -9.72029150e-01
-9.17280734e-01 -8.77140880e-01 -9.46149230e-01 -2.67959803e-01
6.39583319e-02 -4.28066552e-02 4.55663472e-01 8.99951220e-01
3.34102601e-01 1.51635468e-01 8.65553200e-01 -1.39002585e+00
-4.06421602e-01 -8.70361269e-01 -5.40742934e-01 7.44026065e-01
8.14401865e-01 -9.49587464e-01 -6.43955052e-01 7.36400247e-01] | [13.462626457214355, 1.2177560329437256] |
4d684b6f-4406-4574-b29f-e1a8702bd88a | assisted-sound-sample-generation-with-musical | 1904.06215 | null | https://arxiv.org/abs/1904.06215v2 | https://arxiv.org/pdf/1904.06215v2.pdf | Assisted Sound Sample Generation with Musical Conditioning in Adversarial Auto-Encoders | Generative models have thrived in computer vision, enabling unprecedented image processes. Yet the results in audio remain less advanced. Our project targets real-time sound synthesis from a reduced set of high-level parameters, including semantic controls that can be adapted to different sound libraries and specific tags. These generative variables should allow expressive modulations of target musical qualities and continuously mix into new styles. To this extent we train AEs on an orchestral database of individual note samples, along with their intrinsic attributes: note class, timbre domain and extended playing techniques. We condition the decoder for control over the rendered note attributes and use latent adversarial training for learning expressive style parameters that can ultimately be mixed. We evaluate both generative performances and latent representation. Our ablation study demonstrates the effectiveness of the musical conditioning mechanisms. The proposed model generates notes as magnitude spectrograms from any probabilistic latent code samples, with expressive control of orchestral timbres and playing styles. Its training data subsets can directly be visualized in the 3D latent representation. Waveform rendering can be done offline with GLA. In order to allow real-time interactions, we fine-tune the decoder with a pretrained MCNN and embed the full waveform generation pipeline in a plugin. Moreover the encoder could be used to process new input samples, after manipulating their latent attribute representation, the decoder can generate sample variations as an audio effect would. Our solution remains rather fast to train, it can directly be applied to other sound domains, including an user's libraries with custom sound tags that could be mapped to specific generative controls. As a result, it fosters creativity and intuitive audio style experimentations. | ['Adrien Bitton', 'Antoine Caillon', 'Martin Fouilleul', 'Philippe Esling'] | 2019-04-12 | null | null | null | null | ['audio-generation'] | ['audio'] | [ 4.28700089e-01 1.12845793e-01 3.37492734e-01 -1.48368359e-01
-8.84719431e-01 -1.22489929e+00 6.52479351e-01 -4.76643980e-01
-6.88332692e-03 3.96397829e-01 4.05860730e-02 5.00780419e-02
-1.17985778e-01 -9.25483704e-01 -6.64218307e-01 -8.02664638e-01
-1.87958777e-01 6.62808657e-01 2.03125700e-01 -2.88652003e-01
1.89206805e-02 7.02530444e-01 -2.05046368e+00 6.05669558e-01
5.06138325e-01 8.18286836e-01 2.15703860e-01 1.31454945e+00
-1.92622334e-01 3.49529892e-01 -1.11224091e+00 -1.86045110e-01
3.91206503e-01 -4.40274686e-01 -1.67324498e-01 -3.34283233e-01
3.74363571e-01 -1.09089695e-01 1.81077391e-01 9.39495981e-01
8.29841197e-01 -1.29299596e-01 6.52723253e-01 -1.30011308e+00
-4.34339076e-01 9.14122045e-01 2.07196087e-01 -5.19003570e-01
4.48304981e-01 6.80538476e-01 7.66397476e-01 -3.76848549e-01
5.12668788e-01 1.25272465e+00 6.19008720e-01 7.48336554e-01
-1.71410692e+00 -1.03458226e+00 -3.79049569e-01 -9.78375450e-02
-1.08526015e+00 -3.52220625e-01 1.00790524e+00 -5.30766249e-01
5.00473917e-01 7.44504988e-01 9.85896707e-01 1.53104889e+00
-1.74364373e-01 4.89208639e-01 1.25398207e+00 -6.99174523e-01
4.31535423e-01 3.22688311e-01 -4.89481360e-01 3.92276645e-01
-5.55263817e-01 2.55360067e-01 -6.43050730e-01 -1.36924401e-01
1.04084551e+00 -6.13396406e-01 -3.53949457e-01 -2.48585805e-01
-1.10088670e+00 7.47207403e-01 7.34388009e-02 1.87653974e-01
-1.05378255e-02 5.69223940e-01 3.17838579e-01 4.24917012e-01
-6.22947328e-02 9.54097509e-01 -4.72988009e-01 -4.03088599e-01
-1.11167455e+00 5.84452271e-01 9.74145412e-01 9.64271963e-01
6.70225263e-01 4.57624674e-01 -5.36869764e-01 8.57487082e-01
2.48449013e-01 7.35982597e-01 4.63378072e-01 -1.20987189e+00
-7.25210831e-02 2.23808363e-01 -1.28395215e-01 -6.20831311e-01
-1.29686132e-01 -4.19566154e-01 -4.49604154e-01 9.45598781e-01
3.63262147e-01 -9.88252982e-02 -1.06034100e+00 1.81922090e+00
1.28532767e-01 1.33839250e-01 -1.88534230e-01 7.96663702e-01
5.47904134e-01 7.22532153e-01 -1.60929710e-02 1.43182054e-01
1.43084788e+00 -4.33352470e-01 -6.76775217e-01 1.29812762e-01
5.03758602e-02 -1.17728519e+00 1.73224139e+00 7.82752872e-01
-1.24934065e+00 -9.74731445e-01 -1.13887405e+00 1.32706128e-02
-4.07109231e-01 3.07454020e-01 6.40345097e-01 1.11251020e+00
-1.03511798e+00 8.13956857e-01 -7.29065001e-01 1.79694846e-01
1.03969850e-01 6.75939798e-01 1.47867724e-01 6.16097033e-01
-1.14720881e+00 4.92016822e-01 3.32517177e-01 -1.92941397e-01
-1.19528556e+00 -1.09949100e+00 -6.30828917e-01 1.20396078e-01
-1.39997855e-01 -1.01549423e+00 1.17521107e+00 -1.19437933e+00
-2.31636238e+00 6.40102863e-01 5.27547061e-01 -2.85318255e-01
6.14006698e-01 -1.44139409e-01 -4.27986175e-01 -7.12200403e-02
-1.35272071e-01 8.82079363e-01 1.29878700e+00 -1.39722753e+00
-3.21711659e-01 1.23288259e-01 9.47289392e-02 -8.09608027e-02
-3.79000664e-01 -6.85326457e-02 -4.99073386e-01 -9.87359881e-01
-3.88868302e-01 -1.16296411e+00 8.71555656e-02 -8.07771534e-02
-4.57738072e-01 4.63182300e-01 7.71767676e-01 -3.56857240e-01
1.00653183e+00 -2.26601672e+00 3.06347460e-01 3.19750100e-01
-3.23491335e-01 5.05112298e-02 -1.48559183e-01 2.22625822e-01
-2.69685358e-01 1.53860718e-01 -2.55594462e-01 -4.56945211e-01
4.75613296e-01 2.77310759e-02 -5.90251505e-01 -1.04923263e-01
1.76477030e-01 5.51480174e-01 -6.22346520e-01 -2.06080794e-01
3.62062156e-01 7.33610272e-01 -1.04605329e+00 4.42905754e-01
-6.63434446e-01 6.01635396e-01 -2.96130955e-01 2.88478702e-01
4.61432278e-01 4.94893372e-01 -7.64780417e-02 -4.36032295e-01
-1.63283795e-01 2.31233701e-01 -1.40704536e+00 2.10824156e+00
-9.28074539e-01 6.19710505e-01 1.52670816e-01 -2.69821376e-01
1.31062329e+00 2.89898664e-01 1.41064182e-01 -3.84123102e-02
8.51406381e-02 1.02037422e-01 -5.50837889e-02 -3.75813752e-01
4.36781913e-01 -4.05872375e-01 -2.99186766e-01 2.84811705e-01
2.27166504e-01 -9.38905120e-01 -1.26600802e-01 -1.59386799e-01
8.58884931e-01 7.89837062e-01 -3.12022150e-01 -5.84000126e-02
2.76479483e-01 -3.11395377e-01 5.22552989e-02 6.62861824e-01
6.11668706e-01 9.88885880e-01 5.95895946e-01 -6.73318207e-02
-1.27452981e+00 -1.54148495e+00 -9.36800241e-02 1.34310603e+00
-6.24747992e-01 -5.09169281e-01 -8.89576852e-01 -1.90342709e-01
-3.75963449e-01 1.03764200e+00 -5.90838432e-01 -3.22842956e-01
-6.50877953e-01 -4.44529802e-01 9.61540997e-01 2.21845120e-01
2.24005058e-01 -1.38840473e+00 -6.35395229e-01 1.29877701e-01
3.17475528e-01 -6.37766182e-01 -3.99281591e-01 4.75406915e-01
-6.14859045e-01 -4.42373991e-01 -4.76144254e-01 -5.09805322e-01
1.40469462e-01 -8.96992803e-01 1.13401198e+00 -4.44541723e-01
-6.00653648e-01 4.99708235e-01 -3.16948980e-01 -6.83699131e-01
-1.00822544e+00 1.74246281e-01 -4.27885056e-02 8.52779299e-02
-4.11994159e-01 -1.28004241e+00 -5.92620552e-01 -9.47856635e-04
-1.13070965e+00 2.10232437e-01 1.63786963e-01 5.25188625e-01
4.54138100e-01 -1.04795173e-01 4.18265760e-01 -8.88654649e-01
8.82345200e-01 4.89871204e-02 -5.83116293e-01 7.98736140e-02
-1.94522038e-01 2.73021877e-01 8.01889896e-01 -8.44600797e-01
-9.84518766e-01 2.75679946e-01 -3.76988709e-01 -6.54008687e-01
-4.01289999e-01 4.91677225e-03 -6.48177147e-01 3.63877624e-01
8.90751362e-01 -7.95450434e-03 -1.30632207e-01 -3.86305571e-01
8.07938337e-01 5.78143060e-01 7.20147133e-01 -1.04527617e+00
9.75495100e-01 2.32913896e-01 -2.17927601e-02 -5.56523740e-01
-2.12693095e-01 5.04001021e-01 -3.96762371e-01 -2.02980906e-01
8.84250164e-01 -5.46026826e-01 -7.81467438e-01 2.71601111e-01
-1.04563069e+00 -6.63137257e-01 -8.21208537e-01 4.25842106e-01
-1.05774426e+00 -9.72003490e-02 -5.00012696e-01 -6.99055493e-01
-2.31482282e-01 -1.41422534e+00 1.19091821e+00 1.82651699e-01
-6.22776926e-01 -7.57913172e-01 2.09396124e-01 -1.79644063e-01
4.45652783e-01 4.45117325e-01 1.02081382e+00 -2.54226327e-01
-7.06751466e-01 -1.26042932e-01 5.39989889e-01 3.78623754e-01
4.48683947e-02 5.80101132e-01 -1.42839837e+00 -8.90884828e-03
-1.43661320e-01 -1.70979723e-01 5.62775612e-01 2.21736357e-01
1.34286249e+00 -3.01761210e-01 1.26675218e-01 1.06087780e+00
1.15635002e+00 2.52974123e-01 7.88972914e-01 3.35253716e-01
6.08928740e-01 4.61500555e-01 8.64110813e-02 4.78833497e-01
-3.69810015e-01 9.86266136e-01 3.68299216e-01 8.05276539e-03
-4.33871329e-01 -4.42221403e-01 6.26637638e-01 5.33208191e-01
-1.96079105e-01 -1.36794597e-01 -5.35349905e-01 -6.29767999e-02
-1.28645110e+00 -1.10658002e+00 1.70967039e-02 2.15930390e+00
1.19025922e+00 1.90119892e-01 2.15536639e-01 3.71426880e-01
4.89184439e-01 -5.35213389e-02 -3.37118655e-01 -7.89091051e-01
-2.92842630e-02 1.01528549e+00 2.02724010e-01 5.08137524e-01
-7.88647890e-01 9.23195302e-01 6.15406084e+00 1.11927700e+00
-1.43479180e+00 -6.41919523e-02 3.51372838e-01 -5.67105234e-01
-9.13620710e-01 4.38911803e-02 -6.99220896e-01 3.97586435e-01
9.54036057e-01 1.83470130e-01 8.27206314e-01 7.64154255e-01
2.07559049e-01 4.80176270e-01 -1.23981905e+00 9.13065434e-01
-1.49636567e-01 -1.39207816e+00 3.61393481e-01 -5.77011704e-02
3.54855001e-01 -5.46654522e-01 5.37880361e-01 4.82253850e-01
3.87225449e-01 -1.16883421e+00 1.16604829e+00 8.11739564e-01
1.41206670e+00 -9.26434577e-01 -8.74409899e-02 1.55290300e-02
-8.47423434e-01 -3.20199244e-02 -6.53118640e-02 1.51256159e-01
1.65559083e-01 2.41193920e-01 -9.29322481e-01 1.60058245e-01
6.21438920e-01 1.54448859e-02 -4.62996006e-01 6.92693293e-01
-2.92424709e-01 8.91892016e-01 -5.10139763e-01 6.71422407e-02
-2.78724432e-01 -1.81814134e-01 5.88576555e-01 1.42950583e+00
6.64609790e-01 -2.87663877e-01 -1.05325527e-01 1.23799133e+00
2.53986269e-01 6.27346933e-02 -4.44047511e-01 -2.84454431e-02
4.22835886e-01 1.38919795e+00 -6.24958754e-01 -7.21265897e-02
3.91946346e-01 1.03101218e+00 -1.57911688e-01 3.28032047e-01
-1.18870294e+00 -4.28758472e-01 7.29289055e-01 2.94390798e-01
2.05526426e-01 -6.11808933e-02 -3.69076937e-01 -8.34041357e-01
-3.93470556e-01 -1.18773019e+00 -3.69845182e-02 -1.23753989e+00
-1.00552285e+00 8.22892785e-01 5.60700446e-02 -1.23858142e+00
-5.59031248e-01 -5.73460877e-01 -6.88070297e-01 1.07142758e+00
-7.05096185e-01 -1.37945902e+00 -1.76925614e-01 6.43705904e-01
3.11828256e-01 -4.86931145e-01 1.35959089e+00 2.89997935e-01
2.65780501e-02 5.93224406e-01 -3.34330529e-01 1.97288580e-02
8.24140847e-01 -1.49174368e+00 2.12443203e-01 4.00640607e-01
5.14983714e-01 5.49246192e-01 1.16662526e+00 -2.63527960e-01
-1.05908144e+00 -7.06358850e-01 5.41593097e-02 -5.02953053e-01
6.50727868e-01 -8.00324857e-01 -6.27980947e-01 3.45481008e-01
2.92270839e-01 -4.01492745e-01 8.94701183e-01 4.11972404e-02
-3.54575843e-01 -3.01385522e-01 -8.22135210e-01 8.19426298e-01
8.44455957e-01 -6.58351541e-01 -2.64948100e-01 8.11827555e-02
7.41953075e-01 -4.54108298e-01 -8.81294787e-01 3.23329479e-01
7.22594678e-01 -1.20177352e+00 1.13762367e+00 -3.46517026e-01
4.43119824e-01 -6.18560076e-01 -1.29950568e-01 -1.45574498e+00
-1.94842860e-01 -9.77610290e-01 2.62666047e-01 1.75041819e+00
4.04025912e-01 -6.83175027e-02 5.54144382e-01 3.38806182e-01
-3.19408536e-01 -3.58527273e-01 -5.81067085e-01 -4.64172453e-01
-1.92023888e-02 -1.00367904e+00 1.03464699e+00 6.33725047e-01
-3.36988479e-01 2.36406550e-01 -3.00594687e-01 2.90056020e-01
3.11435372e-01 2.94873953e-01 1.11740005e+00 -9.68205690e-01
-1.18686521e+00 -7.27748692e-01 -2.00890929e-01 -5.65788686e-01
5.45513146e-02 -9.69116390e-01 3.49266542e-04 -9.84133482e-01
-4.61173296e-01 -8.53919744e-01 -2.15909048e-03 4.13287818e-01
1.01835765e-01 4.31795448e-01 6.44691527e-01 -8.62955488e-03
2.32796550e-01 4.98515218e-01 1.23994184e+00 -1.41580865e-01
-4.72707212e-01 2.19555333e-01 -1.87805429e-01 7.65520453e-01
8.77839148e-01 -5.53731203e-01 -6.44680083e-01 -9.85696986e-02
4.36423540e-01 2.01701596e-01 5.33199430e-01 -1.46453798e+00
-1.78028747e-01 4.88996543e-02 5.69158137e-01 -1.04018621e-01
6.48520589e-01 -6.35212302e-01 7.95853257e-01 1.24793634e-01
-6.12036824e-01 -2.70549685e-01 3.51685584e-01 6.07051924e-02
5.67572974e-02 -3.63028020e-01 7.17988193e-01 -1.29434168e-01
-4.18801755e-02 1.12304818e-02 -3.33258003e-01 -1.35611340e-01
6.45171762e-01 -4.52874631e-01 1.84607878e-01 -4.41199213e-01
-1.33572471e+00 -5.85227966e-01 6.05803013e-01 3.92693251e-01
2.61285335e-01 -1.29498684e+00 -6.32768273e-01 5.00813067e-01
-4.58723828e-02 -8.82812515e-02 2.66781658e-01 4.43104506e-02
-6.45234346e-01 -2.98601329e-01 -4.93556112e-01 -6.42398715e-01
-1.08048165e+00 7.53567100e-01 4.63340133e-01 3.66590545e-02
-4.43811566e-01 7.61247993e-01 3.61998975e-01 -4.06726360e-01
1.35740519e-01 -3.77753198e-01 -1.32747933e-01 2.25179195e-01
3.25239509e-01 -2.90041957e-02 -1.17707178e-01 1.78341735e-02
4.01749909e-02 6.55182421e-01 7.46764600e-01 -7.89518297e-01
1.43764699e+00 3.56065780e-01 1.24711797e-01 6.82275414e-01
1.10659266e+00 7.56848395e-01 -1.48066723e+00 7.03787506e-01
-5.05075097e-01 -2.60702014e-01 -2.29133576e-01 -1.10870790e+00
-9.71208870e-01 9.09850538e-01 8.11346650e-01 2.08060309e-01
1.28339052e+00 -1.27700254e-01 2.67571926e-01 -1.78015932e-01
3.79188478e-01 -7.97509491e-01 2.13008016e-01 1.24740750e-01
1.08787954e+00 -2.51348108e-01 -4.12837803e-01 -2.47273564e-01
-5.59143245e-01 1.34615171e+00 1.65117636e-01 -3.49362314e-01
5.03728330e-01 1.03707445e+00 4.29484993e-01 3.99893075e-02
-5.81365228e-01 3.62391621e-02 4.70789492e-01 8.35808873e-01
6.26829565e-01 1.67870015e-01 2.03672573e-01 6.73527002e-01
-1.16264331e+00 -6.85322061e-02 4.07823920e-01 1.44681171e-01
-1.29640490e-01 -1.63264406e+00 -6.80730879e-01 3.37132327e-02
-3.91453624e-01 -2.33369038e-01 -1.37869090e-01 5.27652085e-01
7.39165008e-01 4.82886881e-01 1.28647238e-01 -3.73024762e-01
2.89601713e-01 3.68166894e-01 8.12651753e-01 -7.11178422e-01
-9.96305525e-01 4.23248559e-01 1.01577044e-01 -2.72834867e-01
-3.45195383e-02 -6.22216105e-01 -1.32022250e+00 5.32492958e-02
2.52313521e-02 9.94572714e-02 8.20407212e-01 2.75852531e-01
2.13309780e-01 1.01056778e+00 3.32311392e-01 -1.11663592e+00
-3.61059368e-01 -8.83919120e-01 -5.25965631e-01 3.52464914e-01
5.18031884e-03 -3.38524222e-01 -3.66974652e-01 5.38791716e-01] | [15.759958267211914, 5.804262638092041] |
2c4c370c-b0bf-4d36-bf63-00b5effee826 | harnessing-multilinguality-in-unsupervised | 2009.11201 | null | https://arxiv.org/abs/2009.11201v2 | https://arxiv.org/pdf/2009.11201v2.pdf | Harnessing Multilinguality in Unsupervised Machine Translation for Rare Languages | Unsupervised translation has reached impressive performance on resource-rich language pairs such as English-French and English-German. However, early studies have shown that in more realistic settings involving low-resource, rare languages, unsupervised translation performs poorly, achieving less than 3.0 BLEU. In this work, we show that multilinguality is critical to making unsupervised systems practical for low-resource settings. In particular, we present a single model for 5 low-resource languages (Gujarati, Kazakh, Nepali, Sinhala, and Turkish) to and from English directions, which leverages monolingual and auxiliary parallel data from other high-resource language pairs via a three-stage training scheme. We outperform all current state-of-the-art unsupervised baselines for these languages, achieving gains of up to 14.4 BLEU. Additionally, we outperform a large collection of supervised WMT submissions for various language pairs as well as match the performance of the current state-of-the-art supervised model for Nepali-English. We conduct a series of ablation studies to establish the robustness of our model under different degrees of data quality, as well as to analyze the factors which led to the superior performance of the proposed approach over traditional unsupervised models. | ['Aditya Siddhant', 'Xavier Garcia', 'Orhan Firat', 'Ankur P. Parikh'] | 2020-09-23 | null | https://aclanthology.org/2021.naacl-main.89 | https://aclanthology.org/2021.naacl-main.89.pdf | naacl-2021-4 | ['unsupervised-machine-translation'] | ['natural-language-processing'] | [-1.83665708e-01 -2.76025832e-01 -4.23719496e-01 -3.50534558e-01
-1.60872841e+00 -8.95543873e-01 8.62968326e-01 -9.53826308e-02
-7.60691941e-01 1.08622885e+00 5.11715412e-01 -8.09554875e-01
3.13049883e-01 -2.68809319e-01 -6.83179438e-01 -2.91686237e-01
1.20813221e-01 7.07123876e-01 -1.64724976e-01 -4.78115588e-01
-8.00962895e-02 8.38253871e-02 -7.18061030e-01 1.44879758e-01
1.30835843e+00 -1.32340845e-02 6.04562908e-02 4.27427530e-01
7.82752931e-02 3.28345299e-01 -4.27600920e-01 -7.91694403e-01
4.05689389e-01 -4.78377581e-01 -7.62807786e-01 -3.20114315e-01
4.31338161e-01 -3.38191032e-01 -2.12222636e-01 7.69673824e-01
7.53665328e-01 -2.55948693e-01 6.30999327e-01 -7.60891557e-01
-9.40430939e-01 1.01205230e+00 -5.59753656e-01 1.84925109e-01
3.89745116e-01 3.56469825e-02 1.08042264e+00 -1.42433155e+00
8.48496675e-01 1.02253366e+00 7.67537713e-01 3.40475738e-01
-1.17147779e+00 -7.84093857e-01 -5.16874436e-03 -2.80327529e-01
-1.56141901e+00 -9.13109541e-01 2.27740020e-01 -1.70095116e-01
1.34573424e+00 -2.90922988e-02 2.08039537e-01 1.24914670e+00
2.93436915e-01 8.36272597e-01 1.52995503e+00 -8.44497263e-01
-3.39849055e-01 2.41416425e-01 -1.82975203e-01 4.52765226e-01
2.93783218e-01 1.59633383e-02 -7.05203712e-01 -5.41992560e-02
3.26166391e-01 -3.81638497e-01 -1.25637114e-01 1.44614637e-01
-1.72900605e+00 6.51842475e-01 -3.36874165e-02 5.16277313e-01
-1.61468476e-01 -2.48913109e-01 4.35768813e-01 7.48324573e-01
7.87924409e-01 5.02466321e-01 -7.67754912e-01 -4.30860102e-01
-1.12941420e+00 -1.11184955e-01 9.04303908e-01 1.41325915e+00
6.51668549e-01 6.41963035e-02 -8.33951533e-02 1.10133803e+00
2.15177417e-01 8.74194145e-01 5.09947717e-01 -3.56809139e-01
1.18732512e+00 6.14073351e-02 1.42433867e-01 -3.19825947e-01
-1.46471083e-01 -5.23142457e-01 -6.99770808e-01 -5.44576168e-01
5.44872105e-01 -3.99193466e-01 -8.35868359e-01 1.83703637e+00
-6.66619539e-02 -4.41864818e-01 4.60096568e-01 7.55902171e-01
5.72608113e-01 6.42792463e-01 -9.31826606e-02 -3.34878623e-01
1.02334809e+00 -1.33528590e+00 -6.57653987e-01 -4.34994519e-01
9.61153150e-01 -1.48903036e+00 1.47334075e+00 5.69269806e-02
-1.12743211e+00 -4.40512657e-01 -9.45582867e-01 -9.16750059e-02
-2.64122486e-01 3.95401537e-01 5.76850772e-01 7.08598912e-01
-1.10529244e+00 3.39511186e-01 -8.80418301e-01 -9.97269213e-01
2.80992966e-03 3.08980286e-01 -5.34829497e-01 -3.15805256e-01
-1.25675726e+00 1.10922027e+00 3.37645918e-01 7.98478499e-02
-7.39616573e-01 -5.45178056e-01 -6.41569316e-01 -4.11614150e-01
-3.64006646e-02 -3.37254375e-01 1.18588722e+00 -9.30449486e-01
-1.52050161e+00 9.73211765e-01 -2.38542035e-01 -2.35113844e-01
5.98347068e-01 -6.20844662e-01 -6.70786679e-01 -1.56402573e-01
4.10327852e-01 4.55278903e-01 2.83702999e-01 -9.22239840e-01
-4.99652565e-01 -9.53020304e-02 -1.37572274e-01 5.81732631e-01
-5.65983772e-01 6.71379387e-01 -8.62132370e-01 -8.46202374e-01
5.32162786e-02 -1.13956738e+00 -9.53603387e-02 -8.40941906e-01
-3.55050147e-01 1.21478125e-01 3.27419132e-01 -1.05145037e+00
1.23527586e+00 -2.10476089e+00 -2.14675572e-02 1.04934536e-02
-2.52040803e-01 4.58833948e-02 -4.29006398e-01 8.98800492e-01
2.76677310e-01 3.04971814e-01 -1.52605712e-01 -4.40797359e-01
-5.06476238e-02 1.91727638e-01 -1.19601727e-01 4.69567835e-01
3.05959642e-01 1.02263725e+00 -1.05979848e+00 -4.36722934e-01
-2.55642653e-01 1.80495828e-01 -3.32511127e-01 8.56692195e-02
1.52046755e-01 5.62379777e-01 -5.68335019e-02 9.09355640e-01
4.25556421e-01 1.23977244e-01 5.39117932e-01 1.45463422e-01
-2.85266519e-01 9.19009209e-01 -4.93006647e-01 1.90814364e+00
-5.76514006e-01 6.41021848e-01 -1.48043841e-01 -3.28557879e-01
8.03269982e-01 4.33528900e-01 1.57858282e-01 -8.56166840e-01
-1.36374697e-01 9.21020746e-01 3.55243713e-01 -6.99423254e-02
8.47302079e-01 -5.68415672e-02 -3.20665210e-01 7.45895028e-01
1.85306028e-01 -1.55396819e-01 4.34722543e-01 3.33596408e-01
8.60893846e-01 3.31501186e-01 1.11425310e-01 -6.60324216e-01
1.69651881e-01 1.25449896e-01 5.76189160e-01 7.83314824e-01
-1.60703421e-01 7.06769764e-01 -4.12390456e-02 9.10545420e-03
-1.36229837e+00 -1.21459484e+00 -2.16657892e-02 1.19630766e+00
-2.50543565e-01 -5.77329516e-01 -6.35573864e-01 -7.85051823e-01
-3.65277827e-01 6.71637774e-01 -3.39097679e-02 1.10498063e-01
-8.74589503e-01 -1.09522033e+00 1.03916621e+00 2.86431462e-01
5.79391539e-01 -7.07980514e-01 2.53738225e-01 2.89799452e-01
-6.45690978e-01 -1.41342092e+00 -7.75364041e-01 1.98776990e-01
-9.26936805e-01 -5.21613896e-01 -8.62021089e-01 -1.03700089e+00
5.46789050e-01 2.72298336e-01 1.52133155e+00 -3.22168887e-01
4.47627485e-01 1.38644412e-01 -5.59460580e-01 -2.79527247e-01
-6.47629797e-01 7.96442509e-01 5.66134691e-01 -4.17014658e-01
6.84328437e-01 -3.12009037e-01 -1.84852868e-01 3.43612969e-01
-6.10038161e-01 1.17350779e-02 9.24277544e-01 8.76944005e-01
4.85575855e-01 -3.72645885e-01 6.82229519e-01 -1.08602977e+00
7.70519137e-01 -5.16767025e-01 -1.94428459e-01 4.69046861e-01
-7.18281627e-01 -7.10143745e-02 8.60295892e-01 -4.29223955e-01
-1.06787193e+00 -2.45860010e-01 -3.21729891e-02 1.81082547e-01
2.13065058e-01 8.35698247e-01 -1.68122262e-01 2.42591798e-01
7.11346149e-01 3.23026568e-01 -3.04084539e-01 -5.42164326e-01
4.45081085e-01 1.15281057e+00 4.23515290e-01 -8.50954950e-01
1.07116973e+00 5.62327914e-05 -6.33209348e-01 -6.14195764e-01
-5.33199072e-01 -4.78110433e-01 -9.76768136e-01 1.43657565e-01
3.11279267e-01 -1.49870193e+00 3.68792206e-01 6.01688743e-01
-8.78885984e-01 -5.55583537e-01 1.35888025e-01 9.63762164e-01
-3.71926039e-01 2.40167767e-01 -1.07285237e+00 -3.49178672e-01
-5.73124290e-01 -1.24134862e+00 9.02308464e-01 -1.79351509e-01
-3.63144219e-01 -1.17705953e+00 3.03215742e-01 3.39055449e-01
4.85785544e-01 -2.53492892e-01 8.74116838e-01 -8.12498927e-01
-2.64096200e-01 4.14850637e-02 -2.13484287e-01 3.17326635e-01
3.38687420e-01 -4.20077741e-02 -5.31472206e-01 -8.32267761e-01
-3.41680676e-01 -6.16458833e-01 4.70345080e-01 -1.59783185e-01
4.74812165e-02 -1.53125837e-01 -1.11696124e-01 6.07803106e-01
1.30908382e+00 -2.15597808e-01 4.57872361e-01 4.48210180e-01
6.38001561e-01 2.84125358e-01 8.00061524e-01 -1.37354091e-01
6.40630603e-01 6.20884180e-01 -4.53727990e-01 -3.84037495e-01
-1.72700539e-01 -4.49643821e-01 9.35125113e-01 2.04127359e+00
-8.57546031e-02 -1.15691565e-01 -1.24300492e+00 8.74323964e-01
-1.61194253e+00 -4.94682968e-01 1.61518380e-02 2.45133734e+00
1.33085036e+00 1.89123943e-01 1.26207083e-01 -5.54514647e-01
4.59012002e-01 3.52306776e-02 -1.32273465e-01 -5.96825957e-01
-5.13060689e-01 2.52761573e-01 7.67328799e-01 5.13469279e-01
-9.12119091e-01 1.63161850e+00 7.05032110e+00 7.47316003e-01
-1.10612452e+00 4.02743101e-01 4.44498688e-01 -1.35520577e-01
-3.95876437e-01 2.47281790e-01 -8.43484640e-01 2.88806498e-01
1.38700461e+00 -2.17565686e-01 6.41797900e-01 3.52954686e-01
3.30083489e-01 2.19520360e-01 -1.12536955e+00 6.64102793e-01
1.51983202e-01 -8.16465318e-01 2.29668915e-02 2.54279934e-02
1.23880827e+00 9.28981900e-01 7.75191374e-03 5.43422818e-01
6.74376845e-01 -9.18012917e-01 6.43674195e-01 -6.34726323e-03
1.28474844e+00 -6.72675490e-01 7.29792297e-01 3.69874120e-01
-9.47396994e-01 4.23262239e-01 -4.03500795e-01 4.89037447e-02
9.56902653e-02 3.98988098e-01 -8.44817519e-01 8.04903269e-01
5.60191393e-01 7.13102102e-01 -6.50887966e-01 6.08809471e-01
-3.86994243e-01 1.06752753e+00 -5.04644394e-01 1.18571647e-01
5.53041935e-01 -2.52214670e-01 3.72675240e-01 1.68234801e+00
4.70688522e-01 -2.68581092e-01 2.01044992e-01 2.38961801e-01
-4.50775146e-01 7.24943876e-01 -7.76066303e-01 -3.99045706e-01
5.69056690e-01 9.61233556e-01 -4.65539306e-01 -2.39186794e-01
-8.74979556e-01 1.20866609e+00 7.26333141e-01 6.63313985e-01
-6.03985310e-01 -3.84795725e-01 4.83996153e-01 -1.08711094e-01
3.04233395e-02 -6.77209496e-01 -9.79141742e-02 -1.59043956e+00
2.26459071e-01 -1.51890755e+00 2.11161822e-01 -3.97149086e-01
-1.29774261e+00 7.68857360e-01 -1.85100645e-01 -1.42994142e+00
-4.93793964e-01 -3.96493316e-01 -3.57528895e-01 1.25899661e+00
-1.53688252e+00 -1.47452080e+00 3.36083412e-01 5.29792309e-01
7.04050183e-01 -6.16185904e-01 9.94917512e-01 5.76170802e-01
-5.18039465e-01 1.20549417e+00 8.03440928e-01 3.81233335e-01
1.49362421e+00 -1.04401278e+00 9.65386689e-01 1.33786094e+00
5.33801973e-01 1.07361174e+00 4.05311257e-01 -7.68377900e-01
-1.59621954e+00 -1.06378734e+00 1.55380821e+00 -6.88506842e-01
8.81476879e-01 -5.75644910e-01 -4.94512826e-01 9.28656220e-01
5.91894746e-01 -3.49015206e-01 7.82184243e-01 5.34610748e-01
-5.11216521e-01 5.35869226e-02 -7.15044379e-01 9.41795588e-01
1.08573449e+00 -9.41875935e-01 -5.04995406e-01 3.78230631e-01
5.71209252e-01 -3.53800923e-01 -1.06767166e+00 3.22670817e-01
6.68586493e-01 -5.21700025e-01 5.83739161e-01 -7.54893184e-01
5.58867455e-01 -1.33117259e-01 -3.85102540e-01 -1.64784229e+00
-3.00470769e-01 -1.00142360e+00 2.36157835e-01 1.41705179e+00
1.01859152e+00 -6.26205802e-01 3.80072534e-01 1.38267010e-01
-1.34554371e-01 -4.91369218e-01 -7.28219748e-01 -1.06917655e+00
6.62017226e-01 -3.33895028e-01 3.96900952e-01 1.13267446e+00
7.74162263e-02 7.56877363e-01 -5.61332405e-01 -9.37531330e-03
3.14553797e-01 -8.49432647e-02 9.45262849e-01 -5.25669515e-01
-3.06688935e-01 -2.46864513e-01 1.12543777e-01 -1.13345695e+00
2.43939757e-01 -1.20259786e+00 9.13876519e-02 -1.47695088e+00
4.37255472e-01 -6.93013847e-01 -3.64702940e-01 4.09252197e-01
-4.47844118e-01 6.94773972e-01 1.95016339e-01 7.37103403e-01
-4.87562776e-01 3.51636589e-01 9.97478843e-01 -1.35594070e-01
-3.43357682e-01 -2.97273636e-01 -8.32879424e-01 3.85220528e-01
8.75564456e-01 -5.36951005e-01 -2.44387671e-01 -1.21372211e+00
2.23442152e-01 -4.06240970e-01 -5.56801081e-01 -5.70859432e-01
1.22128926e-01 -4.54665944e-02 2.18942434e-01 -3.93066078e-01
-1.12122402e-01 -5.44169664e-01 -8.60977322e-02 1.28683507e-01
-1.97606578e-01 6.29745722e-01 2.63409495e-01 1.53404465e-02
-3.66918027e-01 2.54080351e-03 4.19772893e-01 3.47341225e-03
-4.21024740e-01 1.70557350e-01 -4.30897176e-01 4.68811691e-01
5.05042911e-01 8.54882151e-02 -3.16564590e-01 -3.38370949e-01
-1.43607423e-01 1.67032301e-01 7.97618389e-01 6.55038953e-01
1.56615525e-01 -1.40497088e+00 -1.39391184e+00 2.55597860e-01
4.95710135e-01 -4.63837951e-01 -4.50196087e-01 1.06501806e+00
-6.79810524e-01 4.92759347e-01 -1.02146685e-01 -4.59568024e-01
-8.54169309e-01 8.04569051e-02 -2.97027826e-03 -5.99250913e-01
-3.11511874e-01 4.71180975e-01 -2.30711848e-01 -1.13661098e+00
-9.08911303e-02 8.48511159e-02 3.33955228e-01 -2.01211423e-01
1.77862734e-01 9.54834446e-02 2.91415215e-01 -9.54724371e-01
-4.89846289e-01 3.82554740e-01 -5.34767747e-01 -6.02301002e-01
9.42947447e-01 -3.50093573e-01 -2.33530268e-01 5.54989457e-01
1.14759588e+00 7.09602952e-01 -6.56417310e-01 -4.48125899e-01
1.05255216e-01 -2.97223926e-01 -2.15924948e-01 -1.06186175e+00
-5.34412801e-01 6.23672962e-01 1.87148750e-01 -5.30507565e-01
1.04796624e+00 -1.91704631e-01 8.67144108e-01 5.36170542e-01
6.70813799e-01 -1.34416032e+00 -4.52584147e-01 1.07163978e+00
4.61066246e-01 -1.46703506e+00 -2.21917573e-02 -1.58875316e-01
-6.07586384e-01 8.42607141e-01 3.02408040e-01 2.78631270e-01
2.89414108e-01 3.43994290e-01 6.32965565e-01 3.47877383e-01
-6.31116629e-01 -1.30428240e-01 3.16164404e-01 3.89910221e-01
1.07304835e+00 2.43698910e-01 -6.71920419e-01 4.04361427e-01
-5.01039743e-01 -2.06272691e-01 3.66223961e-01 9.98754978e-01
1.33272573e-01 -1.52543771e+00 -2.34588116e-01 3.55031997e-01
-9.30663288e-01 -8.09768200e-01 -4.94352192e-01 1.12818110e+00
-3.64597261e-01 1.07190990e+00 -1.06238000e-01 -3.74306381e-01
1.30827978e-01 1.72033057e-01 4.94078517e-01 -7.69603133e-01
-9.12682533e-01 4.22132969e-01 4.94283259e-01 -1.95666581e-01
-4.07390267e-01 -7.22776294e-01 -8.64242911e-01 -4.95522141e-01
-3.37671787e-01 2.99774796e-01 5.11482060e-01 9.91373003e-01
2.18590245e-01 -1.12434112e-01 7.03947723e-01 -4.57143903e-01
-5.59191048e-01 -1.40356457e+00 -1.97482750e-01 3.83492768e-01
-1.07898951e-01 -1.71072818e-02 -2.15840623e-01 -2.76673250e-02] | [11.559585571289062, 10.341249465942383] |
6f778f7b-1a1a-4952-8abf-5cbe325fbd7c | analysis-of-adversarial-attacks-against-cnn | 1808.08426 | null | http://arxiv.org/abs/1808.08426v1 | http://arxiv.org/pdf/1808.08426v1.pdf | Analysis of adversarial attacks against CNN-based image forgery detectors | With the ubiquitous diffusion of social networks, images are becoming a
dominant and powerful communication channel. Not surprisingly, they are also
increasingly subject to manipulations aimed at distorting information and
spreading fake news. In recent years, the scientific community has devoted
major efforts to contrast this menace, and many image forgery detectors have
been proposed. Currently, due to the success of deep learning in many
multimedia processing tasks, there is high interest towards CNN-based
detectors, and early results are already very promising. Recent studies in
computer vision, however, have shown CNNs to be highly vulnerable to
adversarial attacks, small perturbations of the input data which drive the
network towards erroneous classification. In this paper we analyze the
vulnerability of CNN-based image forensics methods to adversarial attacks,
considering several detectors and several types of attack, and testing
performance on a wide range of common manipulations, both easily and hardly
detectable. | ['Luisa Verdoliva', 'Diego Gragnaniello', 'Giovanni Poggi', 'Francesco Marra'] | 2018-08-25 | null | null | null | null | ['image-forensics'] | ['computer-vision'] | [ 2.06660405e-01 -6.97788596e-02 6.08245470e-02 1.50710434e-01
-2.56404966e-01 -8.03195238e-01 9.04576242e-01 1.76454440e-01
-5.16638875e-01 5.63148558e-01 -1.22235872e-01 -3.48478705e-01
3.56750399e-01 -9.58123744e-01 -7.54012883e-01 -6.07631743e-01
-2.57574469e-01 -9.02093500e-02 5.48475146e-01 -4.56944436e-01
4.12814558e-01 7.00582683e-01 -1.25524604e+00 2.56008685e-01
3.02025139e-01 8.61825705e-01 -4.50790316e-01 8.57368529e-01
2.01754764e-01 9.56515312e-01 -1.05577266e+00 -1.14138103e+00
3.18096966e-01 -4.24677968e-01 -5.66136599e-01 -1.86385978e-02
4.01180297e-01 -4.96649563e-01 -1.00977051e+00 1.60656703e+00
5.44382691e-01 -4.28188235e-01 6.83903247e-02 -1.37612712e+00
-7.70976126e-01 4.27987903e-01 -6.06647193e-01 7.08685398e-01
3.17031056e-01 6.21472061e-01 4.21676338e-01 -5.16559362e-01
8.10939074e-01 1.41542625e+00 7.44575918e-01 5.52383244e-01
-8.68316591e-01 -9.30382252e-01 -4.36364651e-01 4.08159733e-01
-1.03797424e+00 -4.86393869e-01 8.44553292e-01 -1.50426403e-01
6.02254987e-01 2.81787872e-01 6.03472948e-01 1.65059185e+00
5.60987294e-01 7.55908608e-01 1.23088622e+00 -3.86400074e-01
7.77291134e-02 2.77126074e-01 -4.25938457e-01 6.67058647e-01
4.76102144e-01 4.13643390e-01 -4.33388859e-01 -3.73087615e-01
5.91828644e-01 -9.89296362e-02 -3.84813756e-01 3.40145491e-02
-9.89784002e-01 9.57038641e-01 5.40383399e-01 4.54718649e-01
-1.59220964e-01 3.93986672e-01 6.72773659e-01 6.73145235e-01
5.11252582e-01 5.69691002e-01 5.15146889e-02 -5.61410040e-02
-8.49853873e-01 3.74370784e-01 7.74317801e-01 4.06059980e-01
2.28037700e-01 1.73365965e-01 2.16422319e-01 4.50935870e-01
4.78370748e-02 4.48891938e-01 4.76855934e-01 -6.74718320e-01
3.26779038e-01 3.26436996e-01 -2.05760956e-01 -2.00020766e+00
-1.28423676e-01 -4.29838508e-01 -1.11312807e+00 4.02273566e-01
7.44650185e-01 -9.53573883e-02 -7.34337449e-01 1.37265337e+00
1.83155686e-01 1.36495575e-01 -9.02967155e-02 7.99310565e-01
3.70001316e-01 4.09685820e-01 -3.70264761e-02 4.93631996e-02
1.04347730e+00 -4.24880147e-01 -6.66008770e-01 -2.14386508e-01
3.40702951e-01 -8.46307874e-01 5.12506664e-01 6.95407391e-01
-1.11466229e+00 -2.65321642e-01 -1.21703124e+00 1.88697591e-01
-6.33036613e-01 -7.38068461e-01 5.52943707e-01 1.33134747e+00
-9.66979802e-01 9.13717508e-01 -5.27705252e-01 -2.31701985e-01
1.13489497e+00 2.06935927e-01 -4.74797428e-01 -3.85263115e-01
-1.40234387e+00 8.54681015e-01 2.91847885e-01 1.62358880e-01
-1.02821171e+00 -2.89124161e-01 -3.11536223e-01 -1.57025218e-01
2.74211198e-01 -2.29776829e-01 7.13143170e-01 -1.28252268e+00
-1.11722445e+00 1.11468172e+00 5.76481521e-01 -1.01942909e+00
1.15321088e+00 1.38337016e-01 -9.64873075e-01 4.62006032e-01
-2.31119275e-01 5.57611227e-01 1.46486390e+00 -1.08950138e+00
-3.13116431e-01 -3.99872541e-01 2.02155516e-01 -4.36949998e-01
-7.57709563e-01 4.51515406e-01 5.19376388e-03 -8.86031270e-01
-2.45008364e-01 -8.33449900e-01 -2.04866603e-01 2.74662137e-01
-6.95298851e-01 3.04433137e-01 1.19271564e+00 -6.38599634e-01
9.61971521e-01 -2.31394839e+00 -2.08661303e-01 1.78939819e-01
5.68774104e-01 8.05139422e-01 5.27754985e-03 3.53297979e-01
-7.80992955e-02 6.03045583e-01 -4.98681694e-01 -9.44317654e-02
-3.68041247e-01 1.18763745e-01 -5.70553899e-01 1.07553172e+00
3.29731911e-01 1.16286027e+00 -1.06591582e+00 -2.49608308e-01
1.11214116e-01 4.75251406e-01 -3.51088643e-01 -1.75905392e-01
-1.51009679e-01 4.26328897e-01 -2.23337233e-01 7.28247821e-01
7.56233275e-01 -2.63162963e-02 -1.26316339e-01 1.15484834e-01
2.13296324e-01 -2.43777037e-01 -8.02283645e-01 9.85277534e-01
1.35104522e-01 1.38409138e+00 1.24921650e-01 -1.16450846e+00
5.43458164e-01 2.92243600e-01 1.85165048e-01 -7.50429273e-01
6.17899477e-01 2.55745620e-01 2.95668125e-01 -6.21577322e-01
5.45350194e-01 -7.77402297e-02 8.78920481e-02 4.42375213e-01
4.45861556e-02 -1.30356336e-02 -1.21583864e-01 4.00731027e-01
1.58617949e+00 -4.94086981e-01 8.16631913e-02 1.40358597e-01
3.62993002e-01 -5.34247188e-03 8.64614621e-02 8.69233847e-01
-5.11706471e-01 6.71537399e-01 6.61592782e-01 -7.24204957e-01
-1.31915188e+00 -6.88095987e-01 2.43583880e-02 3.03673685e-01
1.95646659e-01 -2.13367306e-02 -9.85017359e-01 -7.26899564e-01
9.01263282e-02 2.56564468e-01 -5.43639600e-01 -5.98009825e-01
-5.41795731e-01 -8.04341733e-01 1.42827535e+00 3.51230949e-02
1.01943254e+00 -1.30228710e+00 -5.40392220e-01 2.57677585e-01
1.52020067e-01 -1.45955479e+00 -6.79305121e-02 -3.28251839e-01
-7.55590856e-01 -1.30971730e+00 -7.71013796e-01 -4.71839577e-01
4.55652773e-01 5.14797270e-01 8.83997977e-01 7.57118821e-01
-6.46226168e-01 3.38896394e-01 -3.47054422e-01 -5.96846759e-01
-1.07101846e+00 -2.11248353e-01 -1.51642710e-01 2.57564098e-01
1.57386094e-01 -5.65622866e-01 -4.95720446e-01 1.93652555e-01
-1.50889742e+00 -5.11464715e-01 5.16121268e-01 5.26018441e-01
-2.43813083e-01 3.83939892e-01 4.29591894e-01 -9.35879052e-01
8.39011490e-01 -7.14531124e-01 -4.54157978e-01 -1.47958085e-01
-4.26238388e-01 -4.31630433e-01 7.16655672e-01 -6.36040330e-01
-6.60223901e-01 -4.17320758e-01 -1.28966942e-01 -5.52497089e-01
-2.88207889e-01 4.14167881e-01 6.95024431e-02 -7.09470212e-01
1.07428873e+00 1.93812504e-01 1.46179929e-01 2.60313060e-02
2.46462449e-01 4.95436251e-01 5.46135247e-01 6.90607820e-03
1.26613748e+00 9.54372346e-01 1.12960063e-01 -1.36424136e+00
-5.36650419e-01 8.77742656e-03 -2.27903008e-01 -5.98743677e-01
6.31435335e-01 -5.23218691e-01 -5.44037342e-01 1.28675568e+00
-1.39154482e+00 1.43731236e-01 1.25615105e-01 -2.48898123e-03
-1.45257218e-02 8.95693362e-01 -7.95154274e-01 -6.45194113e-01
4.86251973e-02 -1.07068050e+00 4.12426591e-01 1.43075855e-02
4.99346219e-02 -1.03930616e+00 -2.65259802e-01 3.23114038e-01
6.73547566e-01 7.20387161e-01 6.09854281e-01 -6.17293239e-01
-7.60966778e-01 -7.55517066e-01 -3.81274641e-01 7.77144909e-01
-2.02649802e-01 1.09492399e-01 -1.11217308e+00 -4.33390021e-01
2.18766302e-01 -3.07771564e-01 9.24508750e-01 5.07509895e-02
1.12878549e+00 -4.57799882e-01 -2.77713150e-01 5.13729334e-01
1.38639247e+00 4.53134887e-02 1.25565743e+00 4.85868424e-01
5.84763467e-01 5.13369381e-01 -2.01329566e-03 1.97641850e-01
-2.45425209e-01 3.22507411e-01 1.08451641e+00 -1.16609953e-01
-6.53695464e-02 -5.24892919e-02 3.90853047e-01 3.49382371e-01
-3.87731008e-02 -7.31435895e-01 -8.07587683e-01 4.14283156e-01
-1.27780616e+00 -1.35273564e+00 -3.46373260e-01 2.04768682e+00
5.63588798e-01 6.46662474e-01 8.87453705e-02 4.22111958e-01
9.61969733e-01 4.79149252e-01 -4.56508934e-01 -2.68816680e-01
-5.21299660e-01 2.91719049e-01 8.83185923e-01 8.69224686e-03
-1.24356520e+00 1.01903534e+00 6.61124945e+00 7.50273168e-01
-1.25891137e+00 8.71223584e-02 9.53968108e-01 1.79541156e-01
1.05672225e-01 -2.94294924e-01 -1.97228357e-01 6.90314174e-01
8.75755727e-01 -5.60600031e-03 3.75964791e-01 6.66210473e-01
-1.43773835e-02 -1.72238961e-01 -6.34743750e-01 1.01570272e+00
2.51564324e-01 -1.66164100e+00 3.58599313e-02 3.93472522e-01
6.92878425e-01 -6.56948984e-03 4.27111745e-01 -1.69954136e-01
2.63829440e-01 -1.11892092e+00 7.73401201e-01 5.64351231e-02
5.23398459e-01 -7.88135231e-01 8.70472252e-01 3.42448294e-01
-3.74094486e-01 -1.69764712e-01 -5.34047127e-01 -1.41782925e-01
7.49724954e-02 6.88759267e-01 -4.83764529e-01 1.59149766e-01
8.10273707e-01 5.94444752e-01 -6.77353144e-01 1.04869258e+00
-1.55625939e-01 7.55323231e-01 -1.11032695e-01 -2.67983656e-02
4.24781293e-01 1.46192327e-01 7.93081582e-01 1.09466136e+00
1.03978164e-01 -1.19075604e-01 -3.81322682e-01 7.76622236e-01
-4.15969938e-01 -2.97943741e-01 -1.02314353e+00 -4.24899846e-01
1.25211656e-01 1.05273080e+00 -1.18857563e+00 -1.50561735e-01
-2.51987100e-01 1.31046915e+00 -2.79334350e-03 1.21550687e-01
-9.73469973e-01 -9.35710222e-02 7.37031221e-01 1.63621917e-01
1.58593342e-01 -3.80754590e-01 2.07530856e-01 -1.02150023e+00
-2.30137981e-03 -1.24589014e+00 1.12275267e-02 -5.32528520e-01
-1.30035400e+00 5.13460279e-01 -4.27927196e-01 -1.08185291e+00
2.35623866e-01 -5.88093042e-01 -5.46169996e-01 4.22573775e-01
-1.45382082e+00 -7.58398771e-01 -3.23574394e-02 6.27773583e-01
3.64102453e-01 -3.92394096e-01 4.52125371e-01 5.01452208e-01
-3.44656885e-01 5.54129303e-01 -8.91477708e-03 5.84577322e-01
5.49567759e-01 -5.83446860e-01 8.00709665e-01 9.80592906e-01
2.79024005e-01 4.57858860e-01 9.38276112e-01 -6.67608976e-01
-1.35749078e+00 -8.49856436e-01 4.87453014e-01 -3.64991069e-01
9.22520399e-01 -3.42660874e-01 -1.06663752e+00 3.63490224e-01
3.74113739e-01 1.73442826e-01 2.12636486e-01 -7.39502549e-01
-6.45241678e-01 2.82393605e-01 -1.45530093e+00 6.58818722e-01
8.65115345e-01 -7.54733086e-01 -2.04132095e-01 4.15129095e-01
4.70237136e-01 -2.89817274e-01 -2.79876828e-01 -5.58258370e-02
5.30866385e-01 -1.52027309e+00 1.17736435e+00 -6.41364694e-01
7.75839329e-01 6.15440719e-02 2.86103755e-01 -1.21336818e+00
-8.28455482e-03 -8.81030798e-01 -8.19771066e-02 8.70107532e-01
9.35320407e-02 -7.00771272e-01 9.69261408e-01 1.12615064e-01
2.27214262e-01 -2.14553669e-01 -1.02490139e+00 -8.80222559e-01
3.97924101e-03 -4.93253261e-01 2.11243570e-01 1.32231343e+00
-3.78259778e-01 -1.90207958e-01 -7.25461841e-01 2.45094255e-01
6.98910296e-01 -7.32366979e-01 7.29219556e-01 -1.05060804e+00
-1.57190487e-01 -6.70194745e-01 -1.04545462e+00 -6.28390074e-01
-2.12374777e-01 -6.49303913e-01 -3.45207393e-01 -6.66170835e-01
-1.62985101e-01 -2.99199596e-02 -9.25206169e-02 2.14169845e-02
2.73562521e-02 8.72636974e-01 2.58166045e-01 3.58109266e-01
-3.42007577e-01 8.73588994e-02 1.26665878e+00 -4.76267546e-01
4.46057260e-01 1.46957217e-02 -5.48492789e-01 8.17022443e-01
8.42511833e-01 -8.37758482e-01 6.30615801e-02 -3.51346165e-01
7.75775611e-01 -2.92856038e-01 7.85573781e-01 -1.30300903e+00
1.38785347e-01 3.10352296e-01 3.20910364e-01 -1.39193997e-01
2.25602657e-01 -8.50922227e-01 -2.47246753e-02 8.40899765e-01
-2.37865806e-01 1.11849554e-01 1.77025884e-01 8.48027647e-01
-3.58108282e-01 -1.67099550e-01 1.03802133e+00 -4.23707575e-01
-5.84868610e-01 3.06232840e-01 -5.25864661e-01 5.45546673e-02
1.07537806e+00 -4.07448143e-01 -4.43232864e-01 -7.71816492e-01
-5.34851968e-01 -3.67068529e-01 5.16898096e-01 5.27339876e-01
7.79595017e-01 -9.34348226e-01 -6.10756397e-01 5.88544905e-02
-2.23347276e-01 -4.54288781e-01 8.41600895e-02 5.10987043e-01
-9.79617715e-01 -5.24512343e-02 -4.68758792e-01 -3.77020359e-01
-1.30452514e+00 9.71020877e-01 2.72174716e-01 1.15097575e-02
-5.51088572e-01 9.62187529e-01 -3.34225029e-01 1.64900973e-01
1.28481954e-01 8.23413953e-02 4.84366417e-02 2.61348635e-02
7.20577717e-01 6.18008733e-01 7.71704316e-02 -6.92125559e-01
-1.21833794e-01 5.53988032e-02 -8.80161151e-02 1.84915274e-01
1.25835931e+00 1.01049736e-01 -2.11603269e-01 -2.59266794e-02
1.25980449e+00 -5.30931801e-02 -1.01662648e+00 -1.46698982e-01
1.84260234e-01 -6.96801007e-01 1.02749884e-01 -4.80126649e-01
-1.47189248e+00 9.91770208e-01 5.18045962e-01 1.01123834e+00
8.77855599e-01 -2.09324807e-01 1.05456626e+00 1.82266697e-01
3.77944857e-01 -8.00410509e-01 5.34137964e-01 3.27380061e-01
7.03814268e-01 -1.22430086e+00 1.31937072e-01 -3.82618576e-01
-2.11637095e-01 1.16670525e+00 7.77257886e-03 -5.47839820e-01
6.54962003e-01 3.17899197e-01 8.26605260e-02 -3.41486752e-01
-1.60852656e-01 3.35346133e-01 -3.04332972e-01 7.08529949e-01
-4.83332723e-02 -2.91813791e-01 -4.81970534e-02 -3.48636836e-01
-1.12569422e-01 -1.09569699e-01 9.51504886e-01 8.19667935e-01
-4.06972557e-01 -1.05899227e+00 -6.34341836e-01 2.59671152e-01
-9.91764605e-01 -6.87342957e-02 -7.09706724e-01 9.56239700e-01
9.06564370e-02 9.14949536e-01 -1.46132037e-01 -4.12963837e-01
1.20428815e-01 -2.14968204e-01 4.63028908e-01 -5.04508540e-02
-9.19166327e-01 -5.06592870e-01 -1.77272230e-01 -7.40156889e-01
-4.00615513e-01 -6.53596699e-01 -6.15189791e-01 -8.87840033e-01
-4.88870353e-01 -3.70130509e-01 7.96271205e-01 9.66755986e-01
1.80576444e-01 2.84697115e-01 7.25374937e-01 -8.76208603e-01
-5.90760827e-01 -6.73528492e-01 -5.12715518e-01 6.78707063e-01
4.56928223e-01 -4.82339323e-01 -7.64720619e-01 -3.82690802e-02] | [12.500328063964844, 1.1254485845565796] |
25b3d753-2cb9-4181-82b0-68d2e0910bf9 | thor-net-end-to-end-graformer-based-realistic | 2210.13853 | null | https://arxiv.org/abs/2210.13853v1 | https://arxiv.org/pdf/2210.13853v1.pdf | THOR-Net: End-to-end Graformer-based Realistic Two Hands and Object Reconstruction with Self-supervision | Realistic reconstruction of two hands interacting with objects is a new and challenging problem that is essential for building personalized Virtual and Augmented Reality environments. Graph Convolutional networks (GCNs) allow for the preservation of the topologies of hands poses and shapes by modeling them as a graph. In this work, we propose the THOR-Net which combines the power of GCNs, Transformer, and self-supervision to realistically reconstruct two hands and an object from a single RGB image. Our network comprises two stages; namely the features extraction stage and the reconstruction stage. In the features extraction stage, a Keypoint RCNN is used to extract 2D poses, features maps, heatmaps, and bounding boxes from a monocular RGB image. Thereafter, this 2D information is modeled as two graphs and passed to the two branches of the reconstruction stage. The shape reconstruction branch estimates meshes of two hands and an object using our novel coarse-to-fine GraFormer shape network. The 3D poses of the hands and objects are reconstructed by the other branch using a GraFormer network. Finally, a self-supervised photometric loss is used to directly regress the realistic textured of each vertex in the hands' meshes. Our approach achieves State-of-the-art results in Hand shape estimation on the HO-3D dataset (10.0mm) exceeding ArtiBoost (10.8mm). It also surpasses other methods in hand pose estimation on the challenging two hands and object (H2O) dataset by 5mm on the left-hand pose and 1 mm on the right-hand pose. | ['Didier Stricker', 'Nadia Robertini', 'Ahmed Elhayek', 'Jameel Malik', 'Ahmed Tawfik Aboukhadra'] | 2022-10-25 | null | null | null | null | ['object-reconstruction'] | ['computer-vision'] | [-1.30167171e-01 2.48086601e-01 3.63851815e-01 -2.06101477e-01
-4.67667818e-01 -3.75049978e-01 4.25758868e-01 -2.21520007e-01
-2.60864586e-01 5.25040030e-01 -1.40821800e-01 2.12527573e-01
-1.28489425e-02 -8.23743701e-01 -1.04714084e+00 -5.14981568e-01
2.62139648e-01 1.26430547e+00 4.22457188e-01 -3.07756096e-01
-2.70954162e-01 1.07979703e+00 -1.58299363e+00 1.14199094e-01
2.69061238e-01 1.44894981e+00 5.80142289e-02 6.85249984e-01
1.13264576e-01 4.59855258e-01 -1.42133594e-01 -6.02596343e-01
3.80570531e-01 9.89134312e-02 -5.61945319e-01 8.28344598e-02
7.40031421e-01 -6.13346398e-01 -5.45731306e-01 5.56782484e-01
7.91352689e-01 3.94951552e-02 5.00090718e-01 -1.04136419e+00
-1.68990329e-01 2.11401567e-01 -6.54421687e-01 -5.74139237e-01
5.54447234e-01 5.99853583e-02 7.18916833e-01 -1.05148113e+00
1.02742469e+00 1.44501090e+00 9.93422091e-01 4.75477129e-01
-1.23667312e+00 -5.88764131e-01 9.05390903e-02 -1.70074999e-01
-1.61100543e+00 -1.44872174e-01 8.62271428e-01 -5.56394279e-01
7.70196795e-01 4.01485860e-01 1.20495296e+00 1.24417531e+00
1.69916376e-01 4.71335679e-01 1.00896668e+00 -2.62682557e-01
3.94835621e-02 4.85177860e-02 -3.20050448e-01 9.81924832e-01
-1.47268072e-01 1.53342158e-01 -4.71283585e-01 -7.37367868e-02
1.46217167e+00 1.33677378e-01 -1.39779970e-01 -6.10398829e-01
-9.95161235e-01 3.61945152e-01 8.07424545e-01 -5.22101037e-02
-4.97521371e-01 4.60440248e-01 -8.65987316e-02 -1.44689009e-01
5.58790922e-01 -8.17746073e-02 -5.73360980e-01 6.22598752e-02
-6.95589900e-01 5.31443477e-01 8.46397579e-01 1.05405414e+00
5.91658294e-01 -2.53220975e-01 -6.00816347e-02 4.98696804e-01
5.58286011e-01 7.62309372e-01 -2.70904630e-01 -6.74273133e-01
3.51381660e-01 7.15680480e-01 8.88917372e-02 -1.11877120e+00
-6.73002005e-01 -2.99421608e-01 -7.60349035e-01 3.56505543e-01
5.58148265e-01 6.10608272e-02 -1.23332322e+00 1.32293630e+00
7.15486884e-01 -6.50287345e-02 -7.92423666e-01 1.24641323e+00
1.04771364e+00 8.59461874e-02 -3.18457663e-01 3.24501544e-01
1.38402295e+00 -7.03975022e-01 -3.36847901e-01 -7.23941904e-03
-1.61066175e-01 -7.43085980e-01 9.95194077e-01 3.94863456e-01
-1.40581977e+00 -5.15743911e-01 -7.36947238e-01 -4.13288236e-01
-1.78428009e-01 1.48878232e-01 4.73908603e-01 3.39725375e-01
-8.56083095e-01 8.52381825e-01 -9.39629436e-01 -1.14601865e-01
6.02936149e-01 5.68239570e-01 -4.83315200e-01 5.11125028e-02
-6.61139727e-01 8.16676438e-01 -4.19335254e-03 4.73443270e-01
-9.17460561e-01 -8.47578228e-01 -7.83901632e-01 -1.68925732e-01
3.91684473e-01 -8.98248553e-01 8.12466264e-01 -6.83906317e-01
-1.80857563e+00 1.05385518e+00 2.91299582e-01 5.48005961e-02
1.21060455e+00 -2.79818892e-01 1.37834981e-01 -6.34319112e-02
-2.32610032e-01 5.26037991e-01 1.11501765e+00 -1.50872934e+00
-2.16969013e-01 -1.00952220e+00 7.49325678e-02 2.28944972e-01
3.56889248e-01 -3.82218599e-01 -8.51149619e-01 -5.90562403e-01
5.14305532e-01 -1.12000132e+00 1.20677026e-02 5.55140436e-01
-6.17005348e-01 -4.90608625e-02 7.19171941e-01 -1.27443933e+00
5.09361565e-01 -1.95964098e+00 6.58500731e-01 6.55903518e-01
4.62357640e-01 -1.43195495e-01 2.59309784e-02 3.37920412e-02
1.25685707e-01 -3.49716753e-01 1.58412531e-02 -8.69052649e-01
6.96766824e-02 8.65815058e-02 1.70612484e-01 7.66548157e-01
-2.06652433e-01 1.00669765e+00 -6.71749353e-01 -3.97663116e-01
4.10423964e-01 1.15633357e+00 -4.15144801e-01 3.63390982e-01
-4.10104394e-01 6.96911573e-01 -2.35152513e-01 7.38374293e-01
8.72910559e-01 -1.38707802e-01 3.68503243e-01 -5.36223590e-01
5.70369922e-02 1.10151567e-01 -1.53915000e+00 1.97528589e+00
-6.70535862e-01 2.10645899e-01 3.38563353e-01 -2.40538061e-01
7.69157231e-01 3.05149168e-01 4.17621821e-01 -4.79817450e-01
5.09763658e-01 2.03140020e-01 -4.72301275e-01 -1.15719870e-01
1.52558777e-02 -2.17109352e-01 2.10681930e-01 3.00063014e-01
1.04441062e-01 -5.60137510e-01 -4.82943237e-01 -1.42853990e-01
6.51042819e-01 7.04381406e-01 -2.71168947e-01 8.31840187e-02
2.88380682e-01 -4.40634131e-01 1.11211158e-01 2.22995654e-01
4.27320600e-01 8.60494494e-01 2.93696553e-01 -7.32529938e-01
-1.13031101e+00 -1.29434180e+00 -8.64840150e-02 8.93197358e-01
1.17094763e-01 -1.93010837e-01 -7.67248929e-01 -6.35945439e-01
5.90788722e-01 1.98709041e-01 -8.90995681e-01 1.92853510e-01
-6.44953847e-01 -3.94571796e-02 8.94030631e-02 8.30335140e-01
3.44391584e-01 -9.64825869e-01 -5.48882961e-01 -1.27077103e-01
-3.51950377e-02 -1.09330368e+00 -4.10761356e-01 1.16799744e-02
-6.43034577e-01 -1.04760301e+00 -6.95056379e-01 -6.84232056e-01
7.99149930e-01 -3.12239796e-01 1.02420831e+00 2.19938345e-03
-4.75370288e-01 3.55008990e-01 6.53041899e-02 -2.90148765e-01
-8.13142955e-03 1.08122222e-01 5.73648736e-02 1.49522960e-01
-4.40992892e-01 -9.13596690e-01 -7.63848960e-01 4.62419629e-01
-3.61242235e-01 1.81528941e-01 2.65128374e-01 4.70863134e-01
1.08962917e+00 -3.70504290e-01 -2.02194259e-01 -5.72950900e-01
5.35520315e-02 -2.95338091e-02 -8.53447497e-01 1.91344246e-01
-1.02430739e-01 -1.22797251e-01 4.49025571e-01 -5.24791539e-01
-9.81033862e-01 6.58778429e-01 -1.75934523e-01 -8.22420955e-01
2.57590618e-02 -3.75184081e-02 -3.96091580e-01 -4.17139381e-01
6.03034735e-01 -1.99989557e-01 5.68911135e-02 -8.24146450e-01
4.10183966e-01 5.22885263e-01 4.37813580e-01 -5.18911719e-01
8.88603091e-01 7.85140455e-01 4.00097787e-01 -8.02543342e-01
-6.36202693e-01 -1.62218824e-01 -1.19567513e+00 -5.06029129e-01
8.87281477e-01 -6.68903053e-01 -1.22812581e+00 5.79350173e-01
-1.29855597e+00 -4.77162540e-01 -4.28629011e-01 2.20632315e-01
-6.92227304e-01 6.62750229e-02 -5.96402109e-01 -8.37011397e-01
-4.71050352e-01 -1.13824844e+00 1.63685727e+00 -1.35863781e-01
-3.11672315e-02 -7.23981082e-01 -4.47435752e-02 5.72521508e-01
4.24645329e-03 6.36894882e-01 4.48812187e-01 8.91135484e-02
-7.18755186e-01 -4.07403231e-01 -2.14613214e-01 1.91324219e-01
-8.39966312e-02 -1.01141810e-01 -1.21503627e+00 -3.70923877e-01
-2.99536943e-01 -3.36468309e-01 3.64414424e-01 4.55474555e-01
1.25190985e+00 -9.93101895e-02 -2.27036402e-01 9.27802205e-01
1.31840050e+00 -3.26814383e-01 6.37826085e-01 -6.09628372e-02
1.30658364e+00 6.51896656e-01 2.64158636e-01 5.53453326e-01
4.98216480e-01 9.62040842e-01 8.73293638e-01 -1.89506322e-01
-4.85748112e-01 -3.60949904e-01 -1.25726089e-01 6.27009034e-01
-8.32707167e-01 2.48518482e-01 -9.09220517e-01 1.08956881e-01
-1.69912601e+00 -3.89753789e-01 -2.72111177e-01 2.40492272e+00
6.40845835e-01 8.74129757e-02 3.14912170e-01 2.08227783e-02
4.60721433e-01 -1.65739954e-01 -6.62804484e-01 1.57478407e-01
1.09833024e-01 5.85252523e-01 2.86390603e-01 5.20151913e-01
-8.28048110e-01 9.75407064e-01 4.89922047e+00 4.28250998e-01
-1.08423567e+00 1.41050667e-01 2.05050677e-01 -3.50818306e-01
-1.18082173e-01 -2.12485865e-01 -5.46031415e-01 1.54292107e-01
4.26644534e-01 6.25791669e-01 8.35931659e-01 8.12435210e-01
-2.13499188e-01 5.39603345e-02 -1.16292214e+00 1.20701277e+00
1.03085265e-01 -1.13420284e+00 -4.13117483e-02 1.82401374e-01
5.32905340e-01 -7.75509849e-02 -2.26782262e-02 -1.91519275e-01
1.13576576e-01 -1.16368306e+00 1.23628402e+00 1.03994679e+00
1.28499126e+00 -8.51403952e-01 6.24837160e-01 3.14526439e-01
-1.40923250e+00 2.95760661e-01 -1.97692901e-01 9.39492956e-02
1.35991588e-01 4.38225538e-01 -7.80981302e-01 6.21896088e-01
9.34920251e-01 3.72877568e-01 -3.31801176e-01 6.44017279e-01
-2.72256702e-01 -1.27269909e-01 -5.69175065e-01 2.48192325e-01
-3.16776365e-01 -1.23265438e-01 4.48303401e-01 7.60292292e-01
-1.18686520e-01 -1.48727531e-02 1.78910106e-01 9.78650749e-01
-2.54740626e-01 -8.97223055e-02 -2.07374319e-01 3.27953190e-01
1.94212049e-01 1.30397928e+00 -8.34508836e-01 -7.83758610e-02
1.20463498e-01 1.34298825e+00 5.31220973e-01 1.09745227e-01
-8.59450936e-01 -2.23933503e-01 3.69919956e-01 7.18678176e-01
3.85978878e-01 -2.94258207e-01 -4.00638163e-01 -9.54080224e-01
4.61705059e-01 -5.24298251e-01 -3.65037508e-02 -9.33173239e-01
-1.16917491e+00 6.24648273e-01 -2.02608570e-01 -6.73633993e-01
1.15994625e-01 -7.45150626e-01 2.87557729e-02 1.04602504e+00
-1.01484787e+00 -1.81962168e+00 -9.45447922e-01 7.14243412e-01
2.05040619e-01 2.53867358e-01 8.54958534e-01 1.22046478e-01
-3.19270670e-01 5.08846045e-01 -2.26517931e-01 1.41437709e-01
3.57742548e-01 -1.24773753e+00 5.90403080e-01 6.16499856e-02
9.07303542e-02 2.95281291e-01 2.97854751e-01 -9.22652304e-01
-1.75409460e+00 -8.60550165e-01 5.19660950e-01 -8.68161321e-01
2.75856435e-01 -7.91188419e-01 -6.08121216e-01 8.60052645e-01
-6.21317387e-01 5.31381071e-01 2.63054501e-02 3.48939700e-03
-4.44261491e-01 -8.11102018e-02 -1.38063145e+00 4.44703281e-01
1.43627179e+00 -6.92185462e-01 -2.44674608e-01 5.29982984e-01
3.41898739e-01 -1.16649604e+00 -1.13618135e+00 2.18032762e-01
1.29686821e+00 -8.95590484e-01 1.24750996e+00 -4.96677935e-01
2.75721401e-01 -9.95023027e-02 -1.93044141e-01 -9.52403367e-01
-1.55442163e-01 -5.88641524e-01 -3.42896521e-01 7.24199116e-01
-1.03803806e-01 -2.20621213e-01 1.05415225e+00 5.45256972e-01
1.46056160e-01 -9.88447607e-01 -1.04561901e+00 -5.36798358e-01
-1.52546585e-01 -4.86110359e-01 7.00244248e-01 6.73803926e-01
-5.17035663e-01 2.00818658e-01 -1.57707721e-01 2.32428446e-01
9.21554148e-01 6.90848604e-02 8.90094936e-01 -1.67238176e+00
-2.82463133e-01 -1.21823408e-01 -5.11656284e-01 -9.59948838e-01
1.92166135e-01 -9.35494840e-01 -1.35696679e-02 -1.53597271e+00
2.90277116e-02 -5.16900301e-01 2.90522844e-01 4.09436345e-01
3.64195466e-01 4.57933635e-01 1.45688921e-01 2.75640246e-02
-1.74428314e-01 4.32847857e-01 1.53076458e+00 9.45990309e-02
-3.09051245e-01 1.14325322e-01 8.02728236e-02 9.90296423e-01
3.61643463e-01 -2.37616807e-01 1.04583018e-02 -4.36326206e-01
3.03536534e-01 2.20289692e-01 9.03537989e-01 -8.81340086e-01
2.66759712e-02 7.10321143e-02 8.42393637e-01 -8.17897141e-01
8.19762588e-01 -1.26562369e+00 6.73953950e-01 4.20345157e-01
9.75956544e-02 -6.02514185e-02 9.84322876e-02 4.17434931e-01
7.19838798e-01 3.82219821e-01 8.73951256e-01 -3.88085335e-01
5.60961924e-02 6.95196033e-01 4.65202689e-01 -1.62184343e-01
9.27675962e-01 -3.13426942e-01 2.44595975e-01 -2.43043557e-01
-1.25332844e+00 -1.10040195e-01 7.00030208e-01 3.11749458e-01
9.45100367e-01 -1.32019031e+00 -5.62358022e-01 4.85434890e-01
-3.10633242e-01 7.44717181e-01 3.11805129e-01 8.86413395e-01
-8.37785363e-01 1.82683114e-02 -2.93749273e-01 -8.18066299e-01
-1.42455578e+00 1.91138968e-01 7.26062596e-01 -5.19578308e-02
-8.78868639e-01 1.07603800e+00 -5.93087524e-02 -7.53983498e-01
6.10123873e-01 -4.17714089e-01 2.90301830e-01 2.47742981e-02
2.15117931e-01 7.59043396e-01 3.64383370e-01 -9.27385151e-01
-4.58851814e-01 8.40961635e-01 1.46418974e-01 -8.98760408e-02
1.61291528e+00 2.51733601e-01 -3.39618236e-01 4.13876086e-01
1.13807130e+00 4.74600457e-02 -1.57364857e+00 -7.02439621e-02
-5.30977488e-01 -5.56070149e-01 8.02115127e-02 -9.60649490e-01
-1.33332479e+00 9.01938319e-01 7.05386519e-01 -2.80666918e-01
6.54550254e-01 3.50835800e-01 8.04404616e-01 3.90687063e-02
6.74567699e-01 -9.58920121e-01 1.94970772e-01 3.78394544e-01
1.52931154e+00 -7.48677850e-01 2.51923323e-01 -7.39242196e-01
-2.35018745e-01 1.10754871e+00 5.96832395e-01 -2.05142200e-01
7.86016464e-01 2.41949663e-01 -1.41776323e-01 -5.57919085e-01
3.46400402e-02 9.47342545e-04 8.46555710e-01 6.66353226e-01
1.41992852e-01 3.30952227e-01 4.09106135e-01 5.07187963e-01
-5.38736701e-01 -8.83255452e-02 -2.47236758e-01 9.00642693e-01
1.57724291e-01 -7.60308981e-01 -5.62826931e-01 4.12537575e-01
-4.52901721e-02 2.22751886e-01 -7.25813091e-01 8.14839721e-01
3.45317572e-01 2.38228336e-01 2.29841083e-01 -6.44015789e-01
9.13859904e-01 -1.83468774e-01 1.19527662e+00 -5.98879814e-01
-8.17636967e-01 1.43943682e-01 -3.67610380e-02 -9.27957714e-01
-1.33267134e-01 -4.44113910e-01 -1.21659899e+00 -5.83673596e-01
-3.61715227e-01 -5.63305080e-01 1.01620948e+00 8.06717992e-01
2.84517884e-01 5.08681834e-01 2.84521997e-01 -1.71119595e+00
-4.68063027e-01 -7.37727702e-01 -8.66047740e-01 3.67509395e-01
2.00972855e-01 -9.80807543e-01 -1.01936787e-01 -2.78635383e-01] | [6.903442859649658, -1.212636947631836] |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.